Developing Hydrogel Systems for Biofabrication“3D Printing in Medicine Summer Course” University...
Transcript of Developing Hydrogel Systems for Biofabrication“3D Printing in Medicine Summer Course” University...
Developing Hydrogel Systems for Biofabrication
Project 1: Development of a 3D HTS Platform to Extend the Cell
Adhesion Peptide Library in Biofabrication
Supervisor: A/Prof. Tim R. Dargaville
School of Chemistry, Physics and Mechanical Engineering
Faculty of Science and Engineering
Queensland University of Technology
Project 2: Block Copolymers of 2-Oxazolines and 2-Oxazines: The
Influence of Polymer Architecture on the Rheological Properties of
a Potential Thermogelling Bioink
Supervisor: Prof. Robert Luxenhofer
Functional Polymer Materials, Chair for Advanced Materials Synthesis
Department of Chemistry and Pharmacy and Bavarian Polymer Institute
Julius-Maximilians-University Würzburg
Nick Huettner
BSc (Chemistry)
Submitted in fulfilment of the requirements for the degree
Masters of Applied Science (Research), Biofabrication
2019
i
Statement of Original Authorship The work contained in this thesis undertaken between Queensland University of
Technology (QUT) and Julius-Maximilians-University Würzburg (JMU) has not been
previously submitted to meet requirements for an award at these or any other higher
education institution. To the best of my knowledge and belief, the thesis contains no
material previously published or written by another person except where due reference
is made.
Signature: QUT Verified Signature
Date: _______________________________
N.
how
05/02/2019
ii
Keywords Biofabrication, high-throughput screening, cell-adhesion peptide, ECM, tissue
engineering, 3D cell culture, bioink, poly(2-oxazoline), poly(2-oxazine), hydrogel
iii
Preface
This thesis is submitted as part of the Masters of Applied Science (Research) degree
in Biofabrication, which is a collaboration between the Queensland University of
Technology (QUT) in Brisbane, Australia and the Julius-Maximilians-University (JMU)
in Würzburg, Germany. The thesis is submitted to both universities to fulfil the
respective degree requirements. For a better understanding of the structure of the
degree a timeline is included below.
University Research Stage Time Period
JMU Full-time coursework October 2016 –
February 2017
QUT Research and writing
Report 1 (R1)
May 2017 –
September 2017
QUT Research and writing
Report 2 (R2), QUT Thesis
October 2017 –
February 2018
JMU Research and writing
Würzburg Thesis (T1)
April 2018 –
September 2018
This thesis is a combined manuscript of R1, R2 (first research project) and T1 (second
research project). The two projects, conducted at QUT and JMU, follow the overall
topic of biofabrication. Despite that, they must not be seen as a continuous project, but
rather two separated ones, contributing to the same field of research.
iv
Publications, Conferences, Symposiums & Summer Schools
The following manuscript was published within the time frame of this degree
Opinion Article
N. Huettner, T.R. Dargaville, A. Forget, Discovering Cell-Adhesion Peptides in Tissue
Engineering: Beyond RGD, Trends Biotechnol. 36 (2018) 372–383.
doi:10.1016/j.tibtech.2018.01.008.
IF: 13.578 (2017)
The following conferences, symposiums and summer schools were attended in the
frame of this degree
Conferences
“International Conference on Biofabrication 2018” Würzburg, Germany (28 - 31/10/2018)
Symposiums
“Nanotechnology and Molecular Science Symposium” Queensland University of Technology, Brisbane, Australia (14/07/2017)
Summer Schools
“3D Printing in Medicine Summer Course” University of Otago, Christchurch, New Zealand (20 - 22/11/2017)
“3D Printing Technologies - Photochemical and Electrohydrodynamic Techniques” Julius-Maximilians-University Würzburg, Würzburg, Germany (16 - 20/07/2018)
v
Acknowledgements Queensland University of Technology
Firstly, regarding my work in Australia, I want to thank my principal supervisor
A/Professor Tim R. Dargaville for giving me the possibility to conduct my project in his
research group. His supervision over the whole time was appreciated and helped me
learning new and useful techniques for a more efficient and sustainable work ethic.
His advice in several questions regarding the project was very much appreciated.
I also want to thank my secondary supervisor Dr Aurelien Forget for giving me advice
and ideas for all kinds of experiment related to my topic. His expertise in tissue
engineering was from great help in all my cell related experiments, as well as his
expertise in hydrogels and high-throughput screening.
I would also like to express my gratitude to the whole Dargaville research group,
especially Eleonore Bolle for supplying me with great advice on cell culture, cell
staining and cell culture related consumables. On the other hand, I would like to thank
the members of the Hutmacher research group at the Institute of Health and
Biomedical Innovation (IHBI) of QUT and David van der Heide and Jacob Tickner from
the Queensland Institute of Medical Research (QIMR) for sharing their knowledge with
me.
I also appreciate the support of the people working in the Central Analytical Research
Facility (CARF) of QUT. Especially, David Marshall for helping with the mass
spectrometry experiments, Sanjleena Singh for her help at the confocal microscope
and Elizabeth Graham for assisting me with rheology measurements.
Julius-Maximilians-University Würzburg
In the last 6 months of my degree, I had the pleasure to conduct research for my thesis
in the Luxenhofer group at JMU. Therefore, I would firstly like to thank my supervisor
Prof. Robert Luxenhofer for the opportunity to do so. His guidance and advice
throughout the last six months of my degree have been helpful and helped making this
thesis what it is.
vi
I would also like to thank Thomas Lorson for his constant input, advice in experimental
questions and the many conversations in between experiments. His support in the
whole time of the thesis was very appreciated and guided me when planning
experiments.
I would also like to express my gratitude to all other members of the Luxenhofer
research group. I appreciated the nice atmosphere in the lab very much. Thank you
as well for measuring countless NMR samples and the feedback in group meeting
presentations. Apart from that, I would like to thank Matthias Beudert for his assistance
with my cell work experiments and Naomi Paxton for providing her knowledge and
support for the determination of the “bioprinting window” for my samples.
I also appreciated the many conversations with the other Bachelor/Master students in
the office, which made time at work even more interesting.
Since this is the end of two interesting and challenging years, but also full of growth
and fun, I also express my gratitude to all the other Biofabrication students from
Würzburg, Utrecht, Wollongong and Brisbane. It was a pleasure to meet all of them,
to work with them and to go through the same challenges with them.
I want to thank especially Prof. Jürgen Groll and Prof. Paul Dalton from JMU for the
opportunity of being part of this degree and for the opportunity to go to Brisbane to
conduct research, which made this time probably the best time in my life.
Finally, I would like to thank David Pershouse and Deanna Nicdao for all the good
conversations and gatherings throughout this degree. Also, again, Deanna, thank you
for the long hours of proof reading, as well as for illustrating some of the beautiful
schematic figures in this thesis. I thank my parents Ines and Joerg Huettner for their
constant support, not only financially but also emotionally in all times of this degree.
I
Project 1 Development of a 3D HTS Platform to Extend
the Cell Adhesion Peptide Library in Biofabrication
II
Abstract As an alternative to natural extracellular matrix macromolecules, cell adhesion
peptides have had a tremendous impact on the design of cell culture platforms,
implants and wound dressings. However, only a handful of cell adhesion peptides
have been utilized. The discrepancy in extracellular matrix composition strongly affects
cell behaviour, so it is paramount to reproduce such differences in synthetic systems.
This can be done by controlling matrix properties like stiffness and composition in vitro.
With this thesis cell adhesion of a murine 3T3 cell line within a hydrogel matrix of
different stiffness and different embedded cell adhesion peptides (CAPs) was
assessed. The CAPs used were derived from different ECM macromolecules to
consider the complex environment of the natural cell environment. It could be found
that CAPs, other than the in literature commonly used RGD, IKVAV, YIGSR promote
cell adhesion to the hydrogel. In another experiment it could be demonstrated that
hydrogel solid contents of over 25% still promotes cell adhesion, when an RGD
sequence was introduced to the system. However, this condition forces the cells into
larger cell constructs, rather than building single cell-cell contacts and the influence of
RGD got lower with higher matrix stiffness. Attempts to quantify the actual CAP
content in the hydrogels were conducted using the bicinchoninic acid (BCA) and
ninhydrin assay. Both assays were not viable for the use in the hydrogel system for
side reactions occurred, which resulted either in a false positive colouring of the
sample or dissolving of the sample. In a next step using a robotic liquid dispensing
robot, it could be shown that cell spreading is not affected by handling of the cells with
the robot. Furthermore, it was possible to deposit a larger number of hydrogels
automatically using the system. Therefore, this thesis provides the development of a
high-throughput screening approach for the engineering of synthetic ECM with tailored
physical and biochemical properties.
III
Table of Contents Abstract ...................................................................................................................... II
List of Abbreviations & Glossary ................................................................................ V
1. Introduction & Literature Review .......................................................................... 1
1.1 Cell Adhesion Peptides in Tissue Engineering .................................................. 1
1.2 Untapped Knowledge ........................................................................................ 3
1.3 Screening of Cell Adhesion Peptides ................................................................ 7
1.4 Thesis Outline ................................................................................................. 14
2. Materials & Methods .......................................................................................... 16
2.1 Materials ..................................................................................................... 16
2.2 Methods ...................................................................................................... 16
2.2.1 Chemical Crosslinking of 4-Arm PEG Macromonomers ....................... 16
2.2.2 Manufacturing of PEG-Based Hydrogel Macrowells ............................. 17
2.2.3 Cell Culture and 3D Cell Seeding Experiments .................................... 17
2.2.4 Cell Staining and Confocal Imaging ...................................................... 18
2.2.5 Rheology Measurements ...................................................................... 19
2.2.6 Swelling Study of PEG-Based Hydrogels ............................................. 19
2.2.7 Mass Spectrometry ............................................................................... 20
2.2.8 Bicinchoninic Acid (BCA) Peptide/Protein Assay .................................. 20
2.2.9 Ninhydrin Peptide/Protein Assay .......................................................... 20
3. Results & Discussion ......................................................................................... 22
3.1 Screening on Macrowells ............................................................................ 22
3.2 Screening in Hydrogel Droplets ................................................................... 27
3.2.1 Rheology of the PEG-4VS / PEG-4SH-System .................................... 29
3.2.2 Hydrogel Swelling and Polymer Release .............................................. 30
3.2.3 Peptide Quantification Assays .............................................................. 33
3.2.4 2D versus 3D Cell Culture .................................................................... 35
IV
3.2.5 Screening of CAP Concentrations in Stiff Hydrogel Matrices ............... 37
3.2.6 CAP Screening ..................................................................................... 39
3.2.7 Cell Deposition with the Robot .............................................................. 44
3.2.8 Blank Gel Deposition with the Robot .................................................... 45
4. Conclusion & Outlook ........................................................................................... 46
Bibliography ............................................................................................................. 48
V
List of Abbreviations & Glossary 2D cell culture (2-dimensional cell culture): cell culture on a flat x,y-directed surface,
such as a tissue culture plate.
2.5D cell culture (2.5-dimensional cell culture): cell culture that has a 3-dimensional
shape, e.g. wells on a hydrogel surface.
3D cell culture (3-dimensional cell culture): cell culture in x,y- and z- direction, e.g.
within a hydrogel matrix.
3T3 fibroblasts: fibroblast cell line acquired from mouse embryos.
4D cell culture (4-dimensional cell culture): modification of the 3D cell culture, in which
the cell behaviour is observed over time.
AA (amino acid): The smallest molecular building blocks of macromolecules in the
body like proteins or enzymes.
BCA (bicinchoninic acid): Reagent that chelates Cu(I) ions to form a purple coloured
complex. The coloured complex is used for a peptide/protein quantification assay.
BSA (bovine serum albumin): Globular protein derived from cow blood plasma. It is
used as a protein standard in protein assays and used for blocking of unspecific
binding sites in the cell staining process.
CAP (cell adhesion peptide): peptide sequences derived from ECM macromolecular proteins involved in the cell adhesion process through cell receptor binding.
COL I (collagen I): ECM protein that is abundant in the late tissue healing process and
in scars.
COL IV (collagen IV): ECM protein that builds a layer in the basal laminar.
DAPI (4',6-diamidino-2-phenylindole): blue fluorescent marker molecule that is used
to stain the nuclei of cells.
DMEM (Dulbeccos modified Eagle medium): standardized medium for cell culture,
containing different AAs, inorganic salts and vitamins.
dPBS (Dulbecco phosphate buffered saline): sterilized phosphate buffered saline
solution.
VI
ECM (extracellular matrix): the molecules excreted by cells forming the cell environment and providing support for cells to build tissues and organs.
ELN (elastin): ECM protein, which is responsible for the elasticity of tissue.
ESI (electron-spray ionisation): Sample injections probe for mass spectrometry, based
on high voltage ionization of an aerosol.
FC (formal charge): formal charge of peptides at pH 7.
FITC (fluorescein isothiocyanate): green fluorophore that can be attached to other
molecules, like CAPs to make them visible under the fluorescence microscope.
FN (fibronectin): ECM protein that is abundant in several native tissues and plays a
role in biological processes like tissue repair, embryogenesis, homeostasis or the cell
adhesion process.
G’ (storage modulus): measure of the stored deformation energy in the sample during
rheological measurements
G’’ (loss modulus): measure of the dissipated energy of the sample, which is lost in
the form of heat in rheological measurements.
HTS (high-throughput screening): a robot-assisted method allowing the testing of
100,000s of molecules on tissue models to identify future drugs.
IEP (isoelectric point): pH value, at which the average of positive and negative charges
in a peptide/protein are even.
In vitro model: A tissue replicate that was made in the laboratory under sterile
conditions.
LAM (laminin): ECM protein present in several tissues, especially in basal lamina
together with nidogen-1 and collagen IV.
LAMJ1 (laminin J1): J-chain of the laminin macromolecule.
mRNA (messenger ribonucleic acid): molecule that transfers genetic information from
the DNA to the ribosome, where the AA sequence for protein building is specified.
NID1 (nidogen-1): ECM macromolecule that serves as a connective molecule in the
basal laminar for collagen IV and laminin.
VII
OPN (osteopontin): ECM macromolecule that is mainly abundant in bone tissue.
PBS (phosphate buffered saline): Saline solution, buffered by phosphate with a pH
value of constant 7.3. The osmotic pressure is like the one in the human body.
PEG (poly(ethylene glycol)): non-cytotoxic polymer synthesized from ethylene oxide
PEG-4MAL (4-arm poly(ethylene glycol) maleimide): star-shaped PEG with four arms,
modified with a maleimide group as a terminal group on each arm.
PEG-4SH (4-arm poly(ethylene glycol) thiol): star-shaped PEG with four arms,
modified with a thiol group as a terminal group on each arm.
PEG-4VS (4-arm poly(ethylene glycol) vinyl sulfone): star-shaped PEG with four arms,
modified with a vinyl sulfone group as a terminal group on each arm.
PFA (paraformaldehyde): a polymer that is used to fix cells and their morphology for
the cell staining process.
PTFE (polytetrafluoroethylene): inert polymer that can be used as a hydrophobic
surface.
RT-PCR (real-time polymerase chain reaction): method to quantify gene expression
in real time by monitoring the amplification of a target DNA.
THBS1 (thrombospondin-1): ECM macromolecule that binds to other ECM proteins to
change their biological properties.
TRITC (tetramethylrhodamine): red fluorophore that can be attached to other
molecules, like CAPs to make them visible under the fluorescence microscope.
VTN (vitronectin): ECM macromolecule abundantly found in bone tissue and serum.
1. Introduction & Literature Review
1
1. Introduction & Literature Review
1.1 Cell Adhesion Peptides in Tissue Engineering
Cell-adhesion peptides (CAPs) are short amino acid (AA) sequences that are the
minimal motif required to specifically bind to a cell receptor responsible for the cell
adhesion. In the past decades, several CAPs have been utilized in synthetic cell
culture substrates to recapitulate the cell-binding properties of expensive animal-
based macromolecules of the extracellular matrix (ECM) (Figure 1).
Figure 1: Comparison between natural tissue and synthetic cell microenvironment. (A) In the native
extracellular matrix, cells attach through cell receptors to polysaccharides and proteins. (B) 3D synthetic
cell microenvironment made of a substrate with defined mechanical properties functionalized with
peptide sequences able to bind to cell receptors. Addition of soluble biological signals, like growth
factors, embedded in the artificial matrix can further assist reproducing the original tissue.
The natural cell microenvironment is composed of a variety of proteins and
polysaccharide macromolecules. In addition to providing a mechanical support, these
macromolecules link to the cells through receptors located on the cell membrane
(Figure 1 A). Receptors, such as integrin, participate in connecting the cell's
cytoskeleton to the ECM macromolecules. These connections allow the cell to migrate,
differentiate and organize. To recreate the natural cell environment in vitro, synthetic
substrates are used to mimic the mechanical properties of the natural ECM
(Figure 1 B). Biocompatible substrates utilized for cell cultures such as alginate,
1. Introduction & Literature Review
2
agarose or PEG lack the specific adhesive motifs allowing for the precise control of
cell fate and function. As an alternative to natural proteins, short peptides that
specifically target cell receptors can be used to create a biological link between the
synthetic substrate and the cells. These short peptides, so called cell adhesion
peptides (CAPs) can indicate the composition of the environment to the cell. The CAPs
are immobilized on the hydrogel-forming polymer by using optimized coupling
chemical reactions. In parallel, careful selection of the polymers and their processing
into a cell culture substrate allow to match the cell microenvironment mechanical
properties to the properties of natural tissues [1].
Several types of artificial substrates, such as implants[2], scaffolds[3], fibers[4] and
hydrogels[1] have been conjugated with CAPs. These functionalized synthetic
systems presenting CAPs have been shown to improve tissue integration of titanium
implants[5], induce cell spreading in 3D cell culture[6], and reduce scar formation[7].
While CAPs have allowed tremendous advances in biomedical materials, the average
number of publications since 1970 returned from a search on three databases
(Pubmed, Scopus, and Web of Science) for CAPs used for the functionalization of
scaffolds, hydrogels, implants, and fibers reveal paucity in the number of investigated
peptides (Figure 2).
Figure 2: Usage of cell-adhesion peptides (CAPs) reported in biomedical articles for the
functionalization of scaffolds, hydrogels, implants, fibers, or hydrogels. Data were obtained through a
Boolean search in Scopus, PubMed, and Web of Science for articles published between 1970 and
1. Introduction & Literature Review
3
2018, returning a total of 6919 entries combined on the three databases. These results were classified
by peptide sequence and averaged across the three databases.
From this search, we calculated that the majority (89%) of the literature reports use of
the RGD AAs sequence that targets the integrin receptor. The second and third most
reported, IKVAV and YIGSR, both isolated from the laminin sequence, represented
only 6% and 4% of the publications, respectively. Other marginal CAPs such as
DGEA, PHRSN, and PRARI represented less than 1% of the literature. The over-
representation of RGD AAs sequences in biomaterial composition is staggering.
Several factors can explain the lack of variety in the CAPs used in tissue engineering:
these CAPs work (they induce cell spreading and adhesion), they are soluble in
aqueous media, they are short and easy to manufacture at a high purity in automated
peptide synthesizers, and they are well characterized. In contrast, implementing new
CAPs can be a challenging task where many parameters need to be optimized
including peptide solubility, concentration, stability and chemical binding to a medical
device or a cell culture substrate.
This opinion paper presents several CAPs yet to be implemented on cell culture
substrates and proposes different strategies to efficiently trial these CAPs for their use
in medical devices, implants or 3D cell culture applications.
1.2 Untapped Knowledge Engineering tissue models or medical devices requires a comprehensive
understanding of the interactions between cells and their environment. The discovery
of fibronectin and its role in cell adhesion has paved the way for a greater appreciation
of the role played by the ECM macromolecules in cell adhesion[8]. Subsequently, the
minimal binding domain of fibronectin to integrin receptor, RGD, was isolated. Ever
since, this short peptide has been abundantly utilized to induce cell adhesion in
synthetic systems. However, the ECM includes other proteins and polysaccharides in
different ratios that are dependent on the type of tissue. Therefore, to accurately
replicate the complexity of the ECM, more than one CAP with non-specific integrin
binding could be used. Beyond the well characterized and utilized RGD, IKVAV, and
YIGSR peptides, many CAPs derived from ECM proteins have been identified, and
their receptor binding specificity described (Table 1).
1. Introduction & Literature Review
4
Table 1: Physicochemical properties of cell-adhesion peptide sequences isolated from ECM
macromolecules proteins binding to specific cell receptors.
Protein
Sequence Receptor FC IEP Ref.
COL I GFOGER α1β1, α2β1 0 6.7 [9]
DGEA α2β1 -2 3.6 [10]
LAM
YIGSR α4β1, 67 kDa, 38 kDa, 36 kDa 1 9.3 [11,1
2]
YIGSR (cyclic) 67 kDa 1 9.3 [13]
SIKVAV α3β1, α6β1 1 9.7 [14]
IKVAV α3β1, 110kDa , 67 kDa, 45
kDa, 32 kDa 1 9.7
[15–
17]
IKLLI α3β1 1 9.7 [18]
LRGDN αvβ1 0 6.2 [19]
SINNNR α6β1 1 10.6 [20]
LAM
J1
LRE - 0 6.3 [21]
PDGSR - 0 6.2 [13]
GTFALRGDNGQ VLA-6
0 6.1 [18]
CFALRGDNP 0 6.2 [18]
NPWHSIYITRFG α6β1
1.1 9.3 [22]
TWYKIAFQRNRK 4 11.6 [22]
KAFDITYVRLKF α5β1, αvβ3 2 10.2 [23]
LGTIPG 67 kDa 0 6.0 [24]
FN
GRGDS αvβ3, αvβ5 0 6.2 [25]
PKRGDL αvβ5, αvβ1 1 9.7 [25]
NGRAHA
α5β1, αvβ3, αvβ5, αvβ1
1 10.5 [25]
GACRGDCLGA
(cyclic) 0
6.0 [25]
IDAPS α4β1 -1 3.7 [26]
REDV α4β1 -1 4.2 [27]
PHSRN α5β1, αIIbβ3 1 10.5 [28]
KQAGDV αIIbβ3, α5β1 0 6.2 [29]
1. Introduction & Literature Review
5
LDV α4β1, α4β7, α4βP -1 3.7 [30,3
1]
WQPPRARI α4β1 2 12.5 [32]
SPPRRARV Heparin 3 12.8 [33]
LIGRKK Heparin 3 11.8 [34]
IWKHKGRDVILKKDVRFYC
CD157 4 10.4
[35]
KLDAPT (FN5) α4β7, α4β1 0 6.2 [36]
PRARI (FN12-14) α4β1 2 12.5 [37]
VTN CKKQRFRHRNRKG CD157, Heparin, αvβ5 7 12.5 [38]
OPN
KRSR Heparin 3 12.5 [39]
FHRRIKA Heparin 3 12.5 [40]
CGGNGEPRGDTYRAY
α5β3, α2β1 0 6.2
[41]
SVVYGLR α4β1 1 9.3 [42]
ELVTDFPTDLPAT α4β1 -3 3.4 [42]
ELN VPGIG -- 0 6.0 [43]
VGVAPG 67 kDa 0 6.0 [44]
COL IV
MNYYSNS αvβ3
0 6.0 [45]
CNYYSNS 0 6.0 [45]
THB
S1
CSVTCG Heparin 0 6.0 [46]
GRGDAC αvβ3, αIIbβ3 0 6.2 [47]
FQGVLQNVRFVF α3β1 1 10.6 [48]
AELDVP α4β1
-2 3.6 [49]
VALDEP -2 3.6 [49]
NID 1
GFRGDGQ -- 0 6.2 [50]
SIGFRGDGQTC Leukocyte response integrin
(LRI) 0
6.2 [51]
Hydrophobic, polar uncharged, polar charged and cysteine / FC: formal charge at pH
7; IEP: isoelectric point; Protein: COL I: collagen I; LAM: laminin; LAMJ1 : laminin J1;
FN: fibronectin; VTN: vitronectin; OPN: osteopontin; ELN: elastin; COLIV: collagen IV;
THBS1: thrombospondin-1; NID1: nidogen-1, closed at the * interface.
1. Introduction & Literature Review
6
These CAPs can mimic the cell adhesion of different ECM macromolecules by
explicitly targeting a cell receptor. Applying these CAPs alone or as a combination
would provide new avenues for the design of complex ECM-replicates able to target
diverse cell receptors. Nonetheless, CAP sequences have to be chosen carefully as
their physicochemical properties might be challenging for their implementation in
tissue engineering applications.
One considerable concern for the functionalization of synthetic systems, and
particularly hydrogels, is the solubility of the peptide sequence. For instance, aqueous-
insoluble peptides might be difficult to conjugate to hydrogels. Therefore, estimation
of the peptides’ properties could help in the selection of suitable CAPs for any
particular application. One way to predict the aqueous solubility of a peptide is by
assessing the chemical properties of the AAs in the peptide sequence. AAs with
aliphatic and aromatic hydrophobic side chains will lower the water solubility, while
acidic and basic groups like histidine and glutamic acid will have the opposite
effect[52]. Likewise, charges on the peptide will have an impact on the solubility and
might require utilizing acidic or basic aqueous media for the peptide solubilization[53].
Furthermore, cysteine residues in the sequence can enhance its stability against
proteases due to spontaneous formation of disulfide-bonds between cysteine
residues[54]. To assist in the selection and utilization of these CAPs, the
physicochemical properties of the AAs were classified and coded for their
hydrophobicity, charges, and polarity (Table 1). This classification helps to rapidly
identify CAPs that might be difficult to solubilize, such as elastin CAPs mainly
composed of hydrophobic AAs[55,56]. Therefore, taking the peptide’s
physicochemical properties into consideration could help to generate efficient CAPs
libraries and identify compatible CAP combinations.
For biomaterial design, CAPs can be used alone or as a combination of several CAPs,
each uniquely binding to one cell receptor[57]. Such combinations can open new
avenues for the precise control of cell function and the identification of synergistic
effects across CAPs as demonstrated with RGD and YIGSR for the regeneration of
the sciatic nerve[58]. In another example, seven CAPs immobilized on a hydrogel were
tested, both alone and as a combination, for the encapsulation of MIN6 mouse insulin
producing cells and revealed that specific CAP or combinations of CAPs could
increase insulin production[59]. Going beyond integrin and laminin receptors targeting
1. Introduction & Literature Review
7
could lead to new ways of controlling cell fate and function. CAPs explicitly targeting
one integrin receptor, heparin receptor or CD157 - a leukocyte surface receptor[60]
could lead to new in vitro tissue models.
Although the incorporation of new CAPs in tissue engineering or medical device
design is an uncertain and challenging task, low-risk strategies to efficiently identify
the appropriate CAP or combinations of CAPs for a specific application can be
implemented through screening methodologies.
1.3 Screening of Cell Adhesion Peptides The manufacturing of tissue-replicates consists of organizing cells of a specific tissue
in a synthetic cell microenvironment that reproduces the main characteristics of ECM,
namely, mechanical support, soluble growth factors, and cell–ECM interactions. The
optimal combination of these essential ECM features allows control over fate,
organization, and function of cells to build up functional tissues. For example, the
mechanical properties of the cell culture substrate in combination with variable CAP
concentrations can be used to regulate the fate of mesenchymal stem cells[61]. To
study cell–CAP interactions, CAPs are covalently bound to the cell culture substrate.
The functionalization of 2D or 3D synthetic substrates with CAPs requires the
utilization of coupling chemistries such as carboxylic acid activation for its reaction with
the N-terminal of the peptides. As an example, the fibronectin RGD peptide motif was
immobilized on agarose polysaccharides chemically modified to bear carboxylic acid
functional groups available for peptide coupling. This system was used to induce the
3D luminal organization of endothelial cells[1]. Similarly, alginate, which has native
carboxylic acid groups can be directly functionalized with different CAPs[62].
Alternatively, functional groups can be directly introduced into synthetic polymers such
as poly(ethylene glycol) (PEG), allowing for the direct conjugation of CAPs to engineer
cell microenvironments that direct epithelial tubulogenesis[63]. Because of the broad
parameters to be tested, the identification of the optimal combination of mechanical
properties, growth factors, CAPs and cells is a complex task. Therefore, methods that
allow for the rapid manufacturing of a large number of various cell microenvironments
could help to fasten the identification of CAPs, or combinations of CAPs, relevant for
a specific cell type.
1. Introduction & Literature Review
8
Inspired by drug discovery methods, a high-throughput screening (HTS) approach can
be applied to the fabrication and identification of intricate cell microenvironments. At
the beginning of the drug development process, a library of molecules is screened for
their activity on models of healthy or diseased tissue. This optimized screening
process can be scaled up to test hundreds of thousands of molecules per day in
automated systems. The uptake of this research paradigm has had a tremendous
impact on drug development with between 20-30% of drugs in clinical development
being identified through HTS[64]. Recent developments in chemistry and materials
science have afforded materials that can be processed and functionalized in
automated systems. As such, liquid handling robots that can fabricate a variety of cell
culture substrates are particularly well suited for the HTS of the cell environment
(Figure 3).
Figure 3: Proposed platforms for the screening of cell-adhesion peptides allowing for the assembly of
cells, substrate, peptides and growth factors by automated liquid handling robots.
One of the best examples of the application of these systems is the utilization of
surface-activated glass slides and tissue culture plastic that can be covalently linked
with ECM macromolecules to create microarrays of the functionalized surfaces[65].
Such an approach, translated to peptides could help to identify CAPs for the
functionalization of orthopedic implants made of polymers or metal alloys.
1. Introduction & Literature Review
9
Nevertheless, screening platforms are not only limited to 2D geometries. Microwell
platforms (‘2.5D’ geometry) can also be manufactured and functionalized in a single
automated step. As an example, poly(dimethylsiloxane) microwells with localized
reactive anchors[66] and growth factor functionalized collagen microwells were both
made with a liquid handling robot[67]. Further iterations of these systems can lead to
multi-CAP functionalized microwell platforms for the screening of cell-CAP
interactions.
Extending on the microwell concept is 3D cell culture. The use of rapid aqueous-
compatible addition reactions based on thiol-ene[68], Diels-Alder[69], or Huisgen
cycloaddition[70] opens the possibility to automatize the functionalization of polymer
with CAPs under physiological conditions[71]. These chemical reactions can also be
used to crosslink natural[72], and synthetic polymers in situ[73] to afford hydrogels of
different mechanical properties. CAPs terminated with orthogonal chemical moieties,
such as amines and methyl sulfone can be immobilized onto polymers bearing thiol
and carboxylic acid groups[74]. Alternatively, thiol-terminated CAPs can be
immobilized onto maleimide functionalized hydrogel-forming polymers such as
PEG[75]. Since these reactions occur under physiological conditions and do not
produce any byproducts, no purification steps are required, and liquid handling robots
can mix different CAPs, crosslinkers and hydrogel precursors to form CAP-
functionalized hydrogels. The automation of these fabrication steps allows for a rapid
and precise investigation of the concentration response of specific cell type to a given
library of CAPs. Furthermore, an automation platform can rapidly investigate the
synergistic effects of CAPs for a given cell type by mixing CAPs in different
concentrations and combinations. From these cell culture platforms, more complex
systems can be developed by integrating growth factors or substrates of various
mechanical properties[76]. As such, materials that can be blended to afford substrates
with different mechanical properties can be combined with coupling chemistries to
automatize the screening of both mechanical environments and CAPs. As an example,
molecular alloying of carboxylated agarose with native agarose in different ratios
afforded hydrogels with different mechanical properties and could be implemented in
an automated platform where a liquid handling robot blends the different hydrogel
precursors[1,77]. Systems with tuneable mechanical properties would allow the rapid
manufacturing of a variety of cell microenvironments and identify the role played by
1. Introduction & Literature Review
10
growth factors as a function of CAPs and mechanical properties. In turn, the HTS of
the main features of the ECM is expected to lead to accurate models of the natural
ECM.
Implementing an automated platform to fabricate the cellular microenvironments and
screening of the CAP impact is only one part of the challenge. Once manufactured,
methods to determine the role of the CAPs need to be applied. While the cell viability
on 2D substrate can provide a simple readout[78], more advanced characterization
techniques are often required and could be translated from drug discovery platforms
to the characterization of CAP-functionalized substrates. CAPs are involved in cell
migration, spreading and differentiation, thus readout techniques for these functions
need to be applied for the assessment of the CAPs. Cell spreading requires
microscopic techniques able of rapidly and automatically measuring the cell elongation
upon binding to CAPs. On 2D substrates, cell spreading can be determined by
conventional microscopic techniques. However, on opaque or 3D substrates different
instruments are required. Because opaque samples do not allow the use of
conventional microscopy, upright fluorescent microscopes or scanning electron
microscopes are required. In contrast, on 3D systems, the imaging of hydrogels can
be obtained using confocal laser microscopy able to acquire successive focal layers
through the samples. Comparatively, to acquire dynamic characteristics, such as cell
migration, requires advanced living cell culture imaging capabilities. Such cell
migration tracking techniques are now available on laboratory microscopes and even
smartphone-based systems. As an example, the movement of living spermatozoa on
2D substrate can be monitored and analyzed on a smartphone[79]. However, fully
characterizing cell movement in 3D environments requires imaging systems able to
track cell motility in space and over time, so-called 4D[80]. These cutting-edge
experiments result in a considerable amount of data that requires automatized
analysis of cell movements when comparing the behavior of a particular type of cells
in different cell microenvironments. Yet, alternatives to the microscopic imaging can
be implemented for the characterization of cells in their 3D cell microenvironment.
Because the binding of CAPs to a cell receptor induces downstream signaling pathway
activation, mRNA and protein analysis can be utilized to characterize the CAPs. One
convenient way to rapidly monitor protein synthesis is to utilize transfected cells with
a reporter gene that adds a fluorescent marker on the protein of interest[81]. While this
1. Introduction & Literature Review
11
technique is suitable when working with one specific cell type, it can be challenging to
expand to the screening of large cell type libraries in different cell microenvironments.
Alternatively, the change in gene expression can be gathered by real-time polymerase
chain reaction (RT-PCR). This technique requires isolation and purification of mRNA
of the cells and it is widely used and established across laboratories. RT-PCR provides
information on the gene profile and thus the cell fate in a process that can be fully
automatized. Primarily developed for 2D cell cultures, this technique has been
successfully translated into 3D cell culture platforms such as cell embedded
hydrogels[82]. Like the imaging of fixed samples, RT-PCR can only be applied for one
time point per sample and thereby considerably enlarge the number of samples
needed when conducting experiments over several time points.
Despite the challenges to implement a completely automated platform for cell culture
and their subsequent analysis, the HTS of cell microenvironment has been reported
in several examples. For instance, PEG hydrogels were used to identify 3D cell
microenvironments to reproduce the stem cell niche as a function of material
mechanical properties, cell concentration, soluble molecules and macromolecules of
the ECM[83]. Recently, photo-cured hydrogels were used to screen different CAPs as
a function of cell density for ten cell types on a 3D microarray platform, demonstrating
the feasibility of the HTS cell microenvironment approach[84].
The techniques applied in drug discovery HTS for experimental setup and
characterization can be translated into the screening of CAPs and the cell
microenvironment. Inspired by HTS platforms, the findings resulting from the
screening of the cell microenvironment could feed back into the drug discovery
process. This could offer tissue models for drug screening that reproduce healthy and
diseased ECM environments of the same tissue by changing the composition of the
synthetic ECM (Table 2).
1. Introduction & Literature Review
12
Table 2: Example of composition changes in the ECM during injury or diseases.
Disease / Injury Tissue Trend Microenvironment Ref
Cancer Breast ↑ Collagen I,II,III, IV, IX [85]
Glial scar Brain ↑ α1-Laminin-1
Collagen IV [86]
Hepatic fibrosis Liver ↑
Collagen I, III, and IV
Laminin [87]
↓ Elastin
Ehler-Danlos syndrome
Skin ↑ Collagen III [88]
Chronic obstructive pulmonary disease
Lung ↑ Collagen [89,
90] ↓ Elastin
↑Upregulation / ↓ Downregulation
From the smoothness of the brain tissues to the toughness of bones, the mechanical
properties of organs are considerably different. Such diverse environments are
fabricated by cells by secreting the various ECM macromolecules in different
proportions. These environments are tightly regulated between macromolecule
deposition and degradation in a process called homeostasis. However, certain
diseases or injuries can dysregulate this balance and macromolecules in the diseased
ECM can be over-represented (upregulation) or drastically lacking (downregulation).
Therefore, engineering of tissue models that recapitulate the unbalanced ECM
macromolecule composition would enable further investigation of the underlying
mechanism of disease development such as cancer metastasis. Reproduction of the
unbalanced cell-adhesion signals could allow for diverse diseases model to be
developed. So, to go beyond the integrin binding RGD amino acid sequence,
identification of cell-adhesion peptides (CAPs) that reproduce disease-specific ECM
is required to develop such synthetic models, Table 2. These models will have the
potential to further improve the drug discovery process by allowing a more precise
testing of drug candidate libraries.
However, despite the many advances in material processing, coupling chemistries and
CAP discoveries, the screening of cell microenvironments remains on the periphery
when engineering artificial tissue or medical devices. Additionally, development in
1. Introduction & Literature Review
13
automated systems for liquid handling, microscopic imaging and sample processing
for mRNA extraction now provide many solutions to aid in the implementation of CAP
screening. While automated systems can drastically reduce the time and manpower
required for such experiments, the high costs of such systems might hinder
widespread adoption of this approach. Nevertheless, the screening can be done
manually on a smaller scale with promising results as demonstrated for MIN6 cells[59].
The prospect of developing a platform for the identification of CAPs beyond RGD, has
the potential to impact several fields in biomedical research (Figure 4).
Figure 4: Potential applications of screening cell-adhesion peptides in medical devices design (short-
term), drug discovery (mid-term) and bioprinting of tissues (long-term) by allowing the biofabrication of
more complex and accurate cell microenvironments.
In the short term, results obtained from screening platforms can help to develop
medical devices such as surfaces or scaffolds used in wound dressings or to develop
implants with improved tissue integration[5]. Combining the results of CAP screening
with recent advances in hydrogel design would have the potential to help answering
fundamental questions. As an example, dynamic activation of peptides in a hydrogel
matrix could further the understanding of these new CAPs[91], design of hydrogel
matrices with specific CAPs could help to recapitulate different stages of organ
formation[63], and specifically targeting a family of integrin receptors could allow for
the precise direction of organ formation such as blood vessels[57]. Because 3D
experiments are complex to design and characterize, the screening of 3D cell
environments that could afford disease-specific models would require more time to
1. Introduction & Literature Review
14
feedback into the drug discovery process. Once established, these HTS platforms
would have the potential to allow screening of drug candidate libraries on substrates
mimicking different diseases or the comparison of diseases and healthy tissue on the
same platform. Underlying these developments is the aim of reducing high costs and
long development times associated with the drug discovery process and eventually
reduce the use of animal models for pre-clinical testing. On a longer perspective, the
field of 3D bioprinting aiming at manufacturing complex functional tissue as an
alternative to an allograft organ transplant could benefit from the proposed cell
microenvironment screening for the development of bioinks[92].
1.4 Thesis Outline As pointed out in the previous literature review and introduction the natural tissue
consists of a variety of different ECM components and, therefore, many binding sites
for the cells within a tissue. Replicating the specific properties of different tissues is a
main interest in tissue engineering to get a clearer view on native processes like
disease development or cancer metastasis in the body. Rebuilding native tissue
involves the use of all the different ECM components of the native tissue. However,
isolating whole proteins and processing them to tissue models can be time intensive
and expensive. As an alternative, parts of these ECM proteins, that have been proven
to be active in the cell adhesion process to the ECM, can be used. Unfortunately, in
current models there is only a very limited number of cell adhesion peptides used[93].
The most common one is the RGD sequence of the fibronectin protein, which can
support the cell adhesion and spreading of a broad spectrum of cell types[94–96].
Given its application range, tissue models built up with only this sequence can be
considered anything but tissue specific. There have been approaches in literature,
using a limited number of other CAPs and combinations of those to build up more
tailored tissue models[59]. And yet, to better imitate the native environment of cells in
the body it is important to expand the library of CAPs and find a way to effectively
screen combinations and concentration gradients of the tissue specific CAPs with
respect to the behaviour of the cells in the tissue model.
To address this problem, this thesis provides an approach for the effective HTS of
several uncommon CAP sequences found in literature and presented in Table 1. Four
1. Introduction & Literature Review
15
of those CAPs could be screened on their ability to support cell adhesion of a 3T3
fibroblast cell line. Furthermore, the influence of matrix stiffness on cell adhesion with
and without CAP in the system was assessed. First experiments for up-scaling the
proposed method was done by using a QIAgility liquid dispensing robot and a
poly(ethylene glycol) hydrogel system.
2. Materials & Methods
16
2. Materials & Methods
2.1 Materials All chemicals, unless noted otherwise, were used as supplied by the manufacturer.
4-arm poly(ethylene glycol) thiol (PEG-4SH) and 4-arm poly(ethylene glycol)
vinylsulfone (PEG-4VS) were purchased by Jenkem, USA. Phosphate buffered saline
pellets were purchased from Thermo Fisher Scientific, Australia. All cell adhesion
peptide sequences were purchased from GenScript Biotech, USA. DAPI and FITC-
and TRITC-tagged Phalloidin were purchased from Merck, Australia. Dulbecco’s
modified Eagle medium, dPBS, Bovine Serum Albumin (BSA) and 0.05 % Trypsin
were purchased from Gibco, Life Technologies in Australia, USA and New Zealand.
BCA and Ninhydrin assay kits were purchased by Thermo Fisher Scientific, Australia.
Rheology measurements were conducted on an Anton Paar M302 rheometer (Anton
Paar, Austria). Mass spectroscopy experiments were conducted on a LTQ Orbitrap
Elite mass spectrometer (Thermo Fisher Scientific, USA). Confocal images were taken
on a Nikin A1R confocal microscope (Nikon, Japan). For stereomicroscopy imaging a
Nikon SMZ25 stereomicroscope (Nikon, Japan) was used. Furthermore, for acquiring
brightfield images of cells a Zeiss Axio Vert.A1 (Zeiss, Germany) microscope was
used. Automated hydrogel pipetting experiments were conducted using a QIAgility
liquid handling robot (Qiagen, Germany).
2.2 Methods
2.2.1 Chemical Crosslinking of 4-Arm PEG Macromonomers Hydrogels of poly(ethylene glycol) thiol (PEG-4SH) and poly(ethylene glycol) vinyl
sulfone (PEG-4VS) were synthesized by chemical crosslinking in phosphate buffered
saline (PBS, pH = 7.3) at room temperature. For example, for making a 30 µL hydrogel
with a solid content of 10 % (w/v), 11 % (w/v) stock solutions of PEG-4VS (i.e. 11 mg
in 91.34 µL) and PEG-4SH (i.e. 2 mg in 16.54 µL) were prepared. This amount
considers the amount of volume the polymers occupy in solution. From these stock
solutions 23.38 µL (i.e. 2.57 mg PEG-4VS) and 3.90 µL (i.e. 0.43 mg PEG-4SH) were
2. Materials & Methods
17
aliquoted in 0.2 mL low-binding Eppendorf tubes, respectively. The spare solvent
volume of 2.73 µL, which could be used as a volume for cell and peptide addition in
cell culture experiments, was added to the aliquoted PEG-4VS stock. Both solutions,
PEG-4VS (+spare volume) and PEG-4SH, were mixed and vortexed at maximum
speed for 10 seconds. The mixture was centrifuged for another 5 seconds and
deposited on a 12-well plate.
2.2.2 Manufacturing of PEG-Based Hydrogel Macrowells 40 % (w/v) Pluronic® F-127/PBS solution droplets were deposited onto a PTFE
surface. The droplets were covered by a hydrogel precursor mixture of poly(ethylene
glycol) maleimide (PEG-4MAL) and poly(ethylene glycol) thiol (PEG-4SH). After a
gelation time of 5 minutes the hydrogel was inverted and the Pluronic® was rinsed out
of the well with cold MilliQ water. For the introduction of peptides onto the well surface
the peptides were mixed into the Pluronic® solution. In this case the gelation time
given for the hydrogel system was 15 minutes, to ensure diffusion of the peptides into
the hydrogel and, therefore, enabling covalent binding of the peptides within the gel.
2.2.3 Cell Culture and 3D Cell Seeding Experiments 3T3 cells were maintained by cell splitting every 2-3 days. For cell splitting the medium
of the T75 tissue culture flask was removed using a vacuum pump. The cells were
washed with dPBS (5 mL) and afterwards covered with 0.05 % Trypsin solution (1 mL).
After incubation at 37 °C and 5% CO2 for 2 minutes, the flask was topped up with
DMEM medium (9 mL) and centrifuged at 500 g for 5 minutes. The medium was
discarded, and the cell pellet resuspended in DMEM medium (10 mL), before
transferring the cells into a T75 tissue culture flask.
For cell seeding experiments onto glass coverslips, cells were detached with 0.05 %
Trypsin solution from the tissue culture flask and counted using a hemocytometer and
trypan blue. 10,000 cells each were seeded on sterile glass coverslips in a 6-well plate.
To allow adhesion of the cells, medium (3 mL) was added 20 minutes after seeding.
The cells were cultured for 2 days, fixed using 4 % paraformaldehyde (PFA) and
2. Materials & Methods
18
stained using FITC-tagged Phalloidin (1:50 in 1 % BSA in PBS) and DAPI (1:1000 in
PBS) for 45 minutes.
For cell embedding and CAP modification of PEG hydrogels, the polymers were
sterilized by freeze-drying. Cells were detached from the tissue culture flask and
counted. 10,000 cells were used per hydrogel. The cell suspension was added to the
PEG-4SH precursor solution. For the introduction of CAPs into the hydrogel, CAP was
added to the PEG-4VS precursor. Both polymer precursor solutions were mixed,
vortexed for 2 seconds and centrifuged for another 5 seconds before depositing 3 µL
droplets onto a 6-well plate.
2.2.4 Cell Staining and Confocal Imaging For actin and nuclei staining of the 2D cell cultures, the cells were washed three times
with PBS (+ additives Ca2+ and Mg2+) at room temperature under sterile conditions.
The cells were then fixed by covering them with 4 % PFA solution for 30 minutes at
room temperature. After washing the samples one time with PBS, the cells were
permeabilised for 5 minutes by incubating with 0.2 % Triton X-100/PBS solution at
room temperature on the shaker. The samples were washed three times with PBS and
incubated in a solution of DAPI (nuclei, ratio 1:1000) and FITC-tagged phalloidin (actin
filament, ratio 1:50) in 1 % BSA/PBS for 45 minutes at room temperature on the
shaker. Samples were washed two times with PBS and then stored at 4 °C until
confocal imaging.
For staining of the 3D cell cultures, the cells were washed three times with PBS
(+ additives Ca2+ and Mg2+) at room temperature under sterile conditions. The cells
were then fixed by covering them with 4 % PFA solution for 1 hour at room
temperature. After washing the samples one time with PBS, the cells were
permeabilised for 45 minutes by incubating with 0.2 % Triton X-100/PBS solution at
room temperature on the shaker. The samples were washed three times with PBS and
incubated in a solution of FITC- or TRITC-tagged phalloidin (actin filament, ratio 1:50)
in 1 % BSA/PBS overnight at room temperature on the shaker. After washing the
samples one time they were incubated in a solution of DAPI (nuclei, ratio 1:1000) in
PBS at room temperature for 30 minutes on the shaker. Samples were washed two
times with PBS and then stored at 4 °C until confocal imaging.
2. Materials & Methods
19
Confocal imaging of the 2D cell cultures was conducted by lifting the glass coverslips
on a microscopy slide and imaging it at 10x magnification using a 405 nm (DAPI) and
488 nm (FITC) laser. 3D cultures were imaged by lifting the hydrogels from the well-
plate onto a microscopy slide and imaging it at 10x magnification. For excitation of the
stain molecules a 405 nm (DAPI) and 488 nm (FITC)/ 561 nm (TRITC) laser was used.
2.2.5 Rheology Measurements For rheology measurements an oscillatory shear time test was conducted. A PP10
plate with a diameter of 10 mm was used. For the experiment γ was set to 0.05, the
frequency was 0.5 Hz. Both values were constant throughout the experiment. The gap
between plate and Sigmacote® coated glass platform was set to 0.2 mm with the
moving profile “5 Gel Set Gap”, which was pre-set by the manufacturer. The
temperature of the platform was maintained at 25°C by using a Peltier element. For
each replicate a total volume of 30 µL of polymer precursor solution was used. The
precursor solutions were mixed by using an Eppendorf pipette, vortexed for 5 seconds
and centrifuged for another 5 seconds. The sample was applied in the middle of the
glass platform and the storage and loss modulus were measured for a total time of
23 minutes.
2.2.6 Swelling Study of PEG-Based Hydrogels Three PEG-4VS/PEG-4SH hydrogels with a solid content of 10 % (w/v) were
crosslinked. Directly after crosslinking they were weighed on a balance and soaked in
a 2 mL Eppendorf tube with 1 mL of MilliQ water. The water was changed after 1, 3, 7
and 14 days of swelling. Therefore, the hydrogels were removed from the MilliQ water
using a spatula and placed onto a plastic tray. The water of each swelling step was
collected for mass spectrometry analysis. Hydrogels were weighed using the same
balance. After placing the hydrogels into new 2 mL Eppendorf tubes, they were filled
up with fresh 1 mL of MilliQ water. Between water changes the Eppendorf tubes
containing the hydrogels and the ones containing the residue swelling water were
stored at 4 °C.
2. Materials & Methods
20
2.2.7 Mass Spectrometry Mass spectrometry experiments were conducted by mixing the water samples
collected after each soaking step of the hydrogel with methanol to a total concentration
of 50 % (v/v). A reserpine standard was added to each sample, so that every sample
would contain a total concentration of 80 nM. 200-400 µL of the sample was injected
into the mass spectrometer using an electron-spray ionisation (ESI) probe (Thermo
Scientific, USA). Each spectrum was run for 60 seconds. After each run the capillary
was washed by flushing it with two times 1 mL of methanol/water (50:50 v/v) solution,
until the total signal intensity was lower than 100,000. Acquired Spectra were analysed
using the software Sublime Text 3 and Origin Pro 2018.
2.2.8 Bicinchoninic Acid (BCA) Peptide/Protein Assay For the BCA assay a kit by Thermo Fisher Scientific was used. It contained two
reagents: Reagent A contained sodium carbonate, sodium bicarbonate, bicinchoninic
acid and sodium tartrate in 0.1 M sodium hydroxide solution. Reagent B contains 4 %
cupric sulphate. For the experiment hydrogel samples with covalently bound
CRGDSGK sequence and hydrogels without any peptide, as well as a positive control
hydrogel, on which BSA (1 µL) was applied, were used. All hydrogel samples were
washed in 1 mL MilliQ water for at least 3 days before conducting the experiment, to
get rid of unbound peptides in the hydrogels. The water was changed at least 4x in
this time. For the assay, hydrogel samples were placed into one well of a 96-well plate
each. Reagent A and B were mixed in a ratio of 50:1 and 100 µL of the mixture was
applied to each well. The samples were incubated for 20 minutes at 50 °C. After
incubation the samples in the well plate could be analysed using a UV/Vis-plate
reader.
2.2.9 Ninhydrin Peptide/Protein Assay The ninhydrin reagent was purchased by Thermo Fisher Scientific. For the assay it
was diluted to a 1 % solution using ethanol. For the experiment hydrogel samples with
covalently bound CRGDSGK sequence and hydrogels without any peptide, as well as
a positive control hydrogel, on which BSA (1 µL) was applied, were used. All hydrogel
samples were washed in 1 mL MilliQ water for at least 3 days before conducting the
2. Materials & Methods
21
experiment, to get rid of unbound peptides in the hydrogels. The water was changed
at least 4x in this time. For the assay, hydrogel samples were placed into one well of
a 96-well plate each. The ninhydrin solution was applied to each well. The samples
were incubated for 45 minutes at 50 °C.
3. Results & Discussion
22
3. Results & Discussion In this thesis a 2.5D and a 3D approach for screening the CAPs in HTS platform was
used. For the 2.5D environment of the cells a hydrogel platform with CAP coated
macrowells on its surface was manufactured. Due to manufacturing problems of the
wells, when upscaling to robotic system, a second approach had to be developed. This
was to embed cells and CAPs into a droplet of PEG hydrogel. With this system four
different CAPs could be screened on their ability to support cell adhesion of a 3T3
fibroblast cell line.
3.1 Screening on Macrowells The first approach to address the HTS of CAPs was using macrowells on a hydrogel
surface. It was thought to use thiol-ene click-chemistry to attach peptide sequences
while manufacturing the macrowells. To manufacture these wells, droplets of
Pluronic® F-127 containing CAPs were deposited onto a surface and covered with a
hydrogel precursor solution. After inverting the sample and rinsing out the Pluronic®,
a microwell in the hydrogel in the negative shape of the Pluronic® droplet could be
obtained (Figure 5).
Figure 5: Schematic of the manufacturing of CAP coated macrowells on a PEG-based hydrogel. CAPs
were dissolved in Pluronic® F-127 and deposited as droplets on a PTFE surface. A hydrogel precursor
solution containing poly(ethlylene glycol) maleimide (PEG-4MAL) and PEG-4SH was applied. After
curing, the CAPs attached to the interface of Pluronic® F-127 and hydrogel via a Michael Addition
mechanism.
3. Results & Discussion
23
Pluronic® F-127 is a tri-block copolymer of a poly(propylene glycol) block flanked by
two poly(ethylene glycol) blocks (Figure 6) and is commercially distributed by the
company BASF, Germany.
In aqueous solutions between 20 and 40 % (w/v) it forms a physically cross-linked
hydrogel at 10 °C or higher, while being a liquid at lower temperatures. This behaviour
is caused by the physicochemical property called lower-critical-solution-temperature
(LCST), which often occurs in block-copolymers built of a hydrophobic and hydrophilic
polymer block. These kinds of polymers build micelles in aqueous solutions, with the
hydrophilic block presented to the solvent. This block interacts with the solvent, until
above a given temperature, LCST, the thermal energy has overcome the binding
energy between solvent and polymer. This promotes the interaction between
polymers, forcing the polymer into a random coil structure and to precipitate out of the
solution. For the microwell approach the CAPs were dissolved in 40 % (w/v) Pluronic®
F-127 solution. This concentration was used since the resulting gel was strong enough
to not deform when the PEG-4VS/PEG-4SH mixture was applied. The Pluronic®
solutions were to be deposited by a QIAgility liquid handling robot. Therefore, the
Pluronic® solution, as well as the environment of the robot and pipette tips had to be
cooled down while the robot was depositing droplets. This was important to maintain
a low viscosity of the polymer solution. If the temperature was too high, Pluronic®
would partially gel and stick to the pipette tip of the robot, causing issues when
depositing the droplets. The time available to the robot to manipulate the Pluronic®
before gelation and the temperature for precooling the robot equipment was
determined by measuring the temperature of the Pluronic® solution in the robot
equipment after cooling the equipment at -80 °C or -20 °C overnight. Figure 7 shows
the obtained graph.
Figure 6: Chemical structure of Pluronic® F-127. The polymer is built of a poly(propylene glycol) block flanked by two poly(ethylene glycol) blocks. x+z is approximately 70% of the whole molecular weight. The molecular weight of the poly(propylene glycol) unit is 4000 g/mol. Therefore, the average molecular mass of the polymer is approximately 13,000 g/mol.
3. Results & Discussion
24
Figure 7: Temperature curves of a 40% (w/v) Pluronic® F-127 solution in a stock solution block of the QIAgility robot. The block was precooled at -80 °C and -20°C overnight before the measurements. Higher and lower reference constants were set at 5 and -5 °C, since in this temperature region neither is the Pluronic® frozen, nor is it warm enough to show partially gelling. All measurements were done in three times replicates.
The suitable temperature range for processing was between 5 and -5 °C, since in this
range the Pluronic® was neither frozen, nor warm enough to cause partial gelling. It
could be seen that both methods keep the Pluronic® solution in the desired
temperature range for approximately 13 minutes. However, when precooling the
equipment at -20 °C, the Pluronic® did not freeze like it was the case at -80 °C and
could, therefore, be processed earlier. Thus, this temperature was used to precool the
robot equipment.
Next, the influence of the surface substrate on the droplet shape was investigated.
Droplets were deposited onto a PTFE foil, to achieve a high contact angle for the
droplet and, therefore, a hemispherical droplet shape. With this all macrowells made
with these droplets were symmetrical. This was important to ensure that every cell
would have the same environmental symmetry in the well. As a next step, the robot
should further apply a hydrogel precursor solution, which then gels at room
temperature on top of the gelled Pluronic® droplets. The PEG hydrogel system used
is shown in Figure 8.
3. Results & Discussion
25
Figure 8: Polymers and CAP example used in the macrowell approach. The crosslinking compounds
were PEG-4MAL and PEG-4SH. CAPs were attached covalently to the polymer network by Michael
Addition to free PEG-4MAL binding sites. These were introduced into the system by using a molar
excess of PEG-4MAL.
4-arm PEGs were chosen for this approach since they are well known in literature to
build reliable hydrogels[97]. Due to their binding sites at the end of each arm, they can
potentially link to four other polymers, therefore, ensuring a highly crosslinked polymer
network. Furthermore, the Pluronic®, as well as the PEG polymers were dissolved in
PBS to build a cell compatible environment. The gelation time of this system was
around 2 minutes, which was fast enough to keep the CAPs in the Pluronic® from
diffusing into the polymer solution. The CAPs present in the Pluronic® droplet
contained a cysteine AA unit and, therefore, a thiol (-SH) binding site. This reaction
site can bind to groups like maleimide or other carbon-carbon double bonds. Thus, the
-SH group of the CAP in a Pluronic® droplet can react with the maleimide component
of the hydrogel at the interface of PEG hydrogel and Pluronic®, due to an excess of
maleimide in the hydrogel. The viability of this approach could be shown in a smaller
scale, using one single hydrogel and binding the FITC-tagged sequence CRGDSGK-
FITC onto it (Figure 9).
3. Results & Discussion
26
Figure 9: Stereomicroscopy images of hydrogel macrowells. (A,B) Brightfield images of the macrowell
from the side and top. (B, C) Fluorescence images taken of a macrowell that was made by using a 3 µL
Pluronic® droplet without CAP. The CAP was introduced after manufacturing the well by applying 3 µL
Pluronic® solution with a FITC-tagged CAP into the well and letting it soak in for 15 minutes. The well
was rinsed extensively with PBS. (E, F) Fluorescence images taken of a macrowell that was made by
using a 3 µL Pluronic® droplet with FITC-tagged CAP inside. The scale bar in all images equals a length
of 2 mm.
The proof of concept study was done by coating the macrowell with CAP after
manufacturing it with a blank Pluronic® droplet and then fill the well with another
droplet with the CAP in it. The reaction with 2 mM Pluronic®-CAP-solution after
15 minutes showed a strong fluorescence in the well with few diffusion through the
gel. However, for the up-scaling of the method it is useful to get rid of the extra step of
manufacturing the droplet first, before introducing the CAP-Pluronic® solution. It would
be faster to link the CAP to the hydrogel, while the gel is crosslinking, also in respect
to the limited time the Pluronic® solution stays in its desired temperature range. So,
when doing this, it could be seen that with this method there was still CAP attached to
the hydrogel, but far less, then when the CAP was introduced after crosslinking. This
could be due to the diffusion of the polymer precursor solution. When a CAP molecule
binds to a precursor PEG-4VS, while the solution is still liquid it can occur what this
PEG-4VS-CAP molecule diffuses into the gel and does not stay at the interface of gel
and Pluronic®. The fluorescence of these molecules within the gel could be, due to
3. Results & Discussion
27
the distribution throughout the gel, not strong enough to be imaged with the here used
exposure of 100 ms. Also because of the enhanced number of degrees of freedom of
the polymers before crosslinking it is likely that several CAPs adhere to one PEG-4VS
molecule, which then is free to move in the polymer solution, giving a spatial uneven
distribution of CAP on the macrowell, which can be seen in Figure 9F. However, the
concept of coating macrowells with CAPs could be confirmed and brought to
upscaling.
When depositing droplets, every droplet would be slightly sucked up on the side of the
tip instead of being deposited accurately. Therefore, the main issue with this approach
was that the Pluronic® could not reproducibly be deposited with the robot. The pipette
tips would get too warm while handling the liquid, causing a partial gelation on the
pipette tip. The tip was not able to hold the temperature for a longer period.
After extensively testing this method, it was decided that this approach would take too
long to optimize within the project time frame.
3.2 Screening in Hydrogel Droplets Thus, a new approach was chosen, which involved seeding cells into hydrogels, while
the hydrogel is crosslinking (Figure 10). A hydrogel system based on thiol-ene
chemistry was chosen. Precisely, a 4-arm PEG with a vinyl sulfone group at the end
of each arm replaced the maleimide (Figure 11).
3. Results & Discussion
28
Figure 10: Schematic manufacturing process of cells embedded in CAP modified hydrogels.
Figure 11: Polymers and CAP example used in the cell droplet approach. The crosslinking compounds
were PEG-4VS and PEG-4SH. CAPs were attached covalently to the polymer network by Michael
Addition to free PEG-4MAL binding sites. These were introduced into the system by using a molar
excess of PEG-4VS.
The crosslinking kinetic of this hydrogel system was much slower, approximately 16
minutes, because of the lower electron affinity of the vinyl sulfone group compared to
the maleimide. This is due to the two adjacent carbonyl groups of the double bond of
the maleimide, which are more electron deficient than the single sulphur of the vinyl
sulfone. Given this longer gelation time, the processing time with the robot could be
higher and, for there is no Pluronic® involved, all the used reagents were low-viscosity
fluids.
The following section will show the characterization methods done with this
experiment, as well as results for the gels with embedded cells that could be obtained.
3. Results & Discussion
29
3.2.1 Rheology of the PEG-4VS / PEG-4SH-System Stock solutions of both polymers were prepared and mixed shortly before the
measurement. Under the background of HTS it is important that the hydrogel system
has a gelation time that makes it reasonable to process, but not too long that the cells
set down to form a bottom layer within the gel. Therefore, the storage and loss modulus
(G’ and G’’) of the PEG-4VS/PEG-4SH hydrogel system were assessed. Figure 12
shows a diagram of both moduli over a total time of 23 minutes.
Figure 12: Storage (red) and loss (blue) modulus of the PEG-4VS/PEG-4SH hydrogel system. The
gelation point could be identified at 16.13 ± 1.65 minutes as an average over three measurements.
The storage and loss modulus are a stable constant curve until the gelation of the
system. The gelation point was defined as the point in the diagram, where the both
curves of storage and loss modulus meet. Averaged over three measurements the
gelation time of the system could be determined as 16.13 ± 1.65 minutes.
The processing of the polymer solutions by the QIAgility robot took between 4 and 11
minutes, depending on the experimental setup. Therefore, the found gelation time
shows that the gel system is viable for this approach. Since both curves stay constant
over this period of processing, it can be assumed that the mixture of both polymers
will not change its viscosity until the gelation occurs. This is an important criterion for
the automatic processing of the gels, since the liquids might, if they are too viscous,
stick to the pipette tip of the robot and therefore influence the precise deposition of
volumes.
3. Results & Discussion
30
3.2.2 Hydrogel Swelling and Polymer Release When cells are seeded into the hydrogels, it is important to have a platform that can
analyse the proteins, markers and growth factors the cell excretes to assess the
viability of the hydrogel as a replacement for the natural tissue. If the cells do not
produce the same proteins or amounts of proteins in the synthetic system and the
natural system, then the cell microenvironment is not yet optimised. Mass
spectrometry has been proven to be a viable approach to screen the gels for those
excreted molecules.[98,99] However, non-crosslinked polymers that are diffusing out
of the hydrogel in the swelling process can superimpose the signals of the proteins.
Therefore, the water of each soaking step was stored and analysed using mass
spectrometry. Figure 13 shows the graph of the soaking steps.
Figure 13: Spectra of the soaking solutions of the hydrogels obtained by mass spectrometry analysis.
(A) Spectrum of the soaking solution, in which the hydrogels were soaked between synthesis and day 1.
(B) Spectrum of the soaking solution, in which the hydrogels were soaked between day 1 and day 3.
(C) Spectrum of the soaking solution, in which the hydrogels were soaked between day 3 and day 7.
(D) Spectrum of the soaking solution, in which the hydrogels were soaked between day 7 and day 14.
3. Results & Discussion
31
The red arrow in each image indicates the main signal of the reserpine standard (80 nM) that was added
to the sample. The chemical structure of reserpine is shown in (B).
The spectra show the excreted polymer as an accumulation of signals, given by the
high molecular weight of the polymers used (5 kDa for PEG-4SH and 20 kDa for
PEG-4VS) and the superposition of their signals. Reserpine, a common standard used
in mass spectrometry[100,101], derived from the Rauvolfia serpentine plant was
added in a concentration of 80 nM to each of the measured samples. It showed a
consistent signal at m/z = 609.277 and could thus be used as a reference to qualify
the amount of polymer signal in the sample. A shows the spectrum of the MilliQ water,
in which the hydrogel was soaked from day 0 to day 1. It is the only sample, in which
the polymer signal has a higher intensity than the standard, since too much
non-crosslinked polymer diffused out of the gel. Knowing that the reserpine standard
is present in a concentration of 80 nM, the concentration of diffused polymer can be
estimated by comparing the highest signal peak of the polymers to the standard. In A
the reserpine standard has an intensity of 12,000, while the highest peak of the
polymer signals is approximately 16,000. Therefore, the polymer concentration of
PEG-4SH and PEG-4VS can be estimated as 107 nM, assuming a similar ionization
of the different species. This result suggests, that after one day of soaking the gel no
proteomic analysis with mass spectrometry could have been undertaken, since the
signals of the non-crosslinked polymers would make a differentiated characterisation
of low excreted molecules difficult. Comparing B to A, the reserpine signal has a higher
intensity because the amount of polymer in the sample has decreased. Applying the
same calculation as for A the amount of polymer in the sample can be estimated as
40 nM, which is only 37 % of the amount found A. B was the water used for soaking
from day 1 to day 3. That indicates that, depending on the abundancy of the proteins
expressed by the cells, mass spectrometry analysis could be used to analyse the cell
microenvironment after day 3. C and D show the spectra of the soaking from day 3 to
day 7 and day 7 to day 14, respectively. In both cases the reserpine signal at a
concentration of 80 nM is much higher than the polymer signal, giving a concentration
of approximately 3 nM, which is only 3 % of the polymer amount in A.
Considering all these results, it can be assumed that after soaking the hydrogels for
7 days the sample would be suitable for proteomic analysis by mass spectrometry,
since the amount of free residue polymer is so low, that molecules with a concentration
3. Results & Discussion
32
of 80 nM in the gel can be observed without interference with the polymer signals.
Macromolecules with expected high concentrations in the hydrogel, i.e. strongly
expressed proteins by the cell, could be even analysed after only three days of
soaking. The soaking in this experiment was conducted in MilliQ water, which would
not be viable for gels with cells in them, for the cells would die or at least change their
expressed microenvironment. For this kind of experiment the cells in the hydrogels
can just be cultured for seven days, including regular media changes. Through that,
the cells would not lose their nutrients given by the medium, but the medium change
would wash out the non-crosslinked polymers in the gel.
Furthermore, a swelling study of the hydrogels was conducted to determine the
change in the hydrogel size, and, therefore, the space cells will gain, when seeded
into the hydrogel. The hydrogels were swollen in MilliQ water for two weeks. In this
time the water was exchanged frequently. The weight of the hydrogels was taken after
each of these water changes. Figure 14 shows the mass of the hydrogel as a function
of swelling time.
Figure 14: Mass of 30 µL 10 % solid content hydrogels directly after synthesis and 1, 3, 7 and 14 days
of swelling in 1 mL MilliQ water. The water was changed after each weighing step and kept for mass
spectrometry analysis. The obtained graph shows that the hydrogel reaches an equilibrium mass after
3 days of swelling.
In total, three replicates were manufactured. In the first 24 h the hydrogels gain almost
250 % of their original weight (25 mg) by taking up water. After this period the weight
of the hydrogel stabilizes to give a total equilibrium weight of 115 mg (365 % of the
initial weight.). Translating this swelling behaviour into cell culture shows that cells will
3. Results & Discussion
33
gain space to proliferate and spread, when cultured in this hydrogel system. This gain
of mass, and, therefore, volume, can then be used by the cell to fill it up with expressed
ECM and marker molecules.
3.2.3 Peptide Quantification Assays To imitate natural tissue using CAPs it is important to have control over the content of
different CAPs in the tissue model. Calculated concentrations of a CAP in a hydrogel
system are a theoretical value that most of the time does not take processes like e.g.
the self-deactivation of the polymer precursors in solution and due to storage into
account, which then affects the ability of the polymer to bind the peptide. Due to that
the actual concentration of CAP in the hydrogel will differ from the calculated one.
Therefore, determining the difference between calculated and actual CAP
concentration is important. To achieve this, the peptide content within the hydrogel
system of this thesis must be determined and compared to the calculated amount of
CAP. In literature, there are methods, involving e.g. measuring the fluorescence
emitted by fluorophore tagged peptides[102] or tagging the peptides with radioactive
isotopes[103]. However, these methods are expensive, and not suitable, when a high
number of CAPs should be screened in a short amount of time.
Alternatively, colorimetric assays can be used. Those assays have a high sensitivity
for lower concentrations. Thus, two different assays were tested: the BCA assay and
the ninhydrin assay.
The BCA assay is based on two main compounds: bicinchoninic acid (BCA) and
copper(II)-sulphate in a highly alkaline solution. A stock solution together with sodium
carbonate, sodium bicarbonate and sodium tartrate is prepared and applied on the
sample. The sample then is incubated for 20 minutes at 50 °C. The occurring reaction
consists of two sub-reactions. First the peptide bonds in the peptide backbone reduce
Cu(II) to Cu(I). This ion can then be chelated by the BCA, giving a purple coloured
complex. By spectrophotometric analysis the amount of peptide in the sample can
then be quantified by using a calibration curve, acquired by a concentration series of
the to be examined CAP.
This assay was used on the PEG-4VS/PEG-4SH hydrogel system, loaded with 2 mM
of CRGDSGK CAP. Figure 15 shows the result of this assay.
3. Results & Discussion
34
Figure 15: Results of the colorimetric BCA assay. The reagent stock was applied on two hydrogels that were loaded with 2 mM CAP (CRGDSGK), two hydrogels without CAP, one positive control containing BSA and one negative control with a blank hydrogel. All samples were incubated for 20 minutes at 50 °C. A purple colour indicates a positive result of the test.
The assay was used on two hydrogel samples with CAP, two blank samples without
CAP and a positive control with BSA, which was applied onto the hydrogel just before
the assay was conducted. All the hydrogels show a positive test result, even when no
CAP was involved in the synthesis. After incubation the BCA reagent in the wells with
samples containing CAP, as well as the positive control with BSA had a purple colour,
while all the other BCA stock solutions kept their initial green colour. However, the
hydrogels themselves were stained in all cases. It can be deducted that there must be
an active reducing site for Cu(II), like e.g. the vinyl sulfonyl groups that are covalently
bound to a thiol within the hydrogel or even free vinyl sulfone groups, that are normally
used for the CAP adhesion, but still available after CAP adhesion. The fact that the
hydrogel itself builds the purple complex, even without a peptide, makes it difficult to
quantitatively analyse this assay, since there are fragment BCA complexes from the
hydrogel itself. As the double bond of the PEG-4VS, which is available in excess after
the hydrogel synthesis, is a potential reduction site for the copper(II) another
experiment was conducted (data not shown), in which these binding sites were
quenched by soaking the hydrogel in 2-mercaptoethanol for 20 minutes. The result of
this experiment did not differ from the previous result.
As an alternative to this method, a ninhydrin assay was conducted. The ninhydrin
reacts with the free terminal amines of the peptide, building a blue coloured dimer with
another ninhydrin molecule. The colour of this complex then could be analysed by
spectrophotometric methods. However, when the ninhydrin reagent was applied on
the hydrogel, the hydrogel started dissolving and no change in colour could be
3. Results & Discussion
35
observed. This could be due to the ninhydrin molecule nucleophilic attacking the
sulfonyl group of the hydrogel, when applied.
Given these results, both assays are not suitable to quantify the amount of peptide in
the hydrogel. Chemical reactions between the assay reagents and the hydrogel are
interfering with the result of the assay. To quantify the peptide, it is important to either
find reagents for colorimetric assays, which do not react with the hydrogel system or
to find another, non-colorimetric approach. A platform for this could be given by
elemental analysis, with which the peptide amount could be quantified by the
comparison of the elemental ratios of a blank gel and one with CAP. Peptide specific
elements like nitrogen can be used to determine a percentage of the CAP in the
hydrogel from which then, with a known total hydrogel mass, the peptide mass can be
determined. This, however, to my extent of knowledge has not been tried in literature.
On the other hand, this approach would also not be suitable for scale up, since the
samples would always need to be transferred to the instrument and prepared
previously to that by lyophilizing. Finding another colorimetric assay would be more
suitable. There are several different typed of assays based on different chemical
reagents available in literature[104–106], which can be assessed on their ability to
work with this hydrogel system. One example is the Lowry assay, based on the
reaction of phosphotungstic and phosphomolybdic acid with Cu(I). This could be part
of the future work for this project.
3.2.4 2D versus 3D Cell Culture To prove the suitability of the chosen hydrogel system for the encapsulation of 3T3
cells, an experiment embedding 10,000 cells in a hydrogel of the volume of 3 µL was
conducted. The hydrogel was loaded with 2 mM of the peptide sequence CRGDSGK
to support the cell adhesion to the gel and to also promote spreading of such. The
hydrogel had a solid content of 10 % (w/v) and the binding sites of the PEG-4VS was
not saturated with the used amount of peptide. As a control, the same number of cells
was also seeded on glass coverslips. After incubating the cells for 24 h the cells were
fixed and stained, using DAPI for the cell nuclei and FITC-tagged Phalloidin for the
actin filament of the cell. Figure 16 shows the result of this staining.
3. Results & Discussion
36
Figure 16: Comparison of cell spreading of 3T3 cells after 24 h of 2D and 3D culture in DMEM medium.
All samples were fixed with 4 % PFA and stained using FITC-tagged Phalloidin (Actin filament of the
cell) and DAPI (cell nuclei). (A-C) Confocal images of 3T3 cells seeded on glass coverslips and stained
with Phalloidin and DAPI. (D-F) Confocal images of 3T3 cells in PEG-4VS/PEG-4SH hydrogels modified
with 2 mM of the CAP CRGDSGK and stained by Phalloidin and DAPI. The images were acquired by
running Z-stacks of 120-160 µm through the hydrogel and superimposing them with a maximum
intensity stack. The scale bar on each image equals 100 µm.
Besides the staining of the cell nuclei in B and C, the staining of the actin filament in
A and B gives the most information about the cell spreading and adhesion in the
hydrogel and the coverslips. The fact that 2D and 3D cell culture lead to different cell
morphologies and behaviour are long known[107]. The same difference in both
methods can be seen in this experiment. The 2D cell culture gives a highly connected
network of cells, which covers almost the whole surface of the glass slide. The 3D cell
spreading is much less dense, shows, however, a connection between the different
cells in the 3D environment. Cells cultured in a hydrogel must migrate through the gel,
slowing their proliferation rate down and giving them more natural ways to spread
compared to the 2D culture. Cells in these gels use the CRGDSGK sequence to
adhere once they get embedded into the gel.
3. Results & Discussion
37
Figure 17: Volume view of the hydrogel modified with CRGDSGK from a diagonal view (A) and a side
view (B). To acquire this, a Z-stack through the whole hydrogel was run. A dome structure with cells
throughout the whole hydrogel could be observed.
Looking at the volume view (Figure 17) of the CRGDSGK modified hydrogels obtained
by confocal microscopy imaging shows that the cells are distributed throughout the
whole gel (A and B). The gelation time of the hydrogel system is fast enough to prevent
the cells from migrating to the bottom of the hydrogel and, therefore, building a
monolayer. This can be seen in B, since it shows that the bottom part of the gel has
less cells than on the upper part.
This demonstrates that this approach is viable for cell culture and it can be used for
the screening of CAPs rarely or not at all used in tissue engineering.
3.2.5 Screening of CAP Concentrations in Stiff Hydrogel Matrices The stiffness of natural tissue throughout the body is highly different and dependent
on the type of tissue available. Processes, like the development of diseases, e.g.
breast cancer, can change the ECM stiffness.[108] In the case of breast cancer, this
can cause cells like fibroblasts, which are common in the tissue[109], to react
differently to this change[110]. The change of ECM stiffness can be modelled in vitro
by introducing different solid contents of the hydrogels. Therefore, an experiment was
conducted, which involved high hydrogel stiffnesses. The disease driven change in
molecular composition of the ECM[111], was considered by introducing different
concentrations of the CRGDSGK CAP into the hydrogel system. Figure 18 shows the
confocal images of the DAPI and TRITC-tagged phalloidin stained 3T3 fibroblast cells
after 48 h of culture in DMEM medium.
3. Results & Discussion
38
Figure 18: Comparison of cell spreading of 3T3 cells after 48 h 3D culture in DMEM medium for different CRGDSGK concentrations (2, 0.5 and 0.05 mM) and hydrogel solid contents (25, 39 and 49 % (w/v)). All samples were fixed with 4 % PFA and stained using TRITC-tagged Phalloidin (Actin filament of the cell) and DAPI (cell nuclei). The images were acquired by running Z-stacks of 120-160 µm through the hydrogel and superimposing them with a maximum intensity stack. The scale bar on each image equals 100 µm
All the samples showed that even with high hydrogel percentages until 49 % (w/v) cells
within the hydrogel system still show spreading. In some of the samples DAPI stained
fragments could be observed that were present in the whole hydrogel. Those were
cracks, caused by the high stiffness of the gels. While the 25 % hydrogel shows cell-
cell interactions across the hydrogel, the higher percentages 39 and 49 % show more
accumulation of cells into a bigger cell construct. The 25 % gel showed a similar cell
spreading regardless of the amount of CAP in the gel. The 39 % hydrogel, however,
seems only to promote cell spreading at 2 mM CAP concentration, but not below. If
less than this concentration was used, the cells in the hydrogel did not even
accumulate anymore and only spread their actin filament very little, which could be
predicted, given the high hydrogel stiffness. Without a certain concentration of CAP in
the hydrogel, to which the cell can adhere, the cell migration through the gel is not
promoted. The 49 % hydrogel showed at all CAP concentrations accumulation of cells
into a bigger construct, however, less single cell-cell interactions than the previous two
3. Results & Discussion
39
stiffnesses can be observed and the cell shape is more spherical, indicating that the
cells cannot spread through the matrix anymore. The accumulation of cells is not as
big as in 39 %, which can be explained with the higher stiffness of this hydrogel,
preventing the spreading of the actin filament. The samples with CAP seem to promote
the cell spreading the same way as the control sample without CAP, indicating that at
this high stiffness the CAP influence on the cell adhesion process is limited. Negative
control hydrogels, without CAP attached to the PEG-4VS showed consistently less
cell spreading than their counterparts with CAPs, except at 49 %.
With this experiment the influence of the matrix stiffness on cells in environments with
different CAP concentrations could be assessed. It could be shown that cells spread
even at a hydrogel solid content of 49 %, where the concentration of CAP does not
seem to have influence on the cell spreading anymore. It could be shown that the
matrix stiffness alters the cell adhesion behaviour of the fibroblasts, also in
dependence of the CAP concentration. The stiffer the matrix, the less dependent the
cells are on the CAP concentration in their spreading behaviour. In future experiments,
proteomics of these fibroblast cells could be conducted, and the results compared to
cells in a healthy breast tissue and a cancer tissue, to see whether a stiffer matrix
causes the cell to act abnormal and, therefore, support processes like tumour growth.
3.2.6 CAP Screening The natural ECM is build up by cells and highly different in composition based on the
tissue of the body. Replicating the ECM of a specific tissue in tissue engineered
models requires the combination of many different ECM proteins. The use and
combination of these into one tissue model can be expensive, hard to process without
destroying the protein structure and time intensive. Therefore, CAPs of those proteins
can be used, instead of the whole protein structures. The RGD sequence is the most
abundantly used CAP for tissue engineered models, even though it is not a CAP that
is specific for one ECM protein. Therefore, the use of other and more specific CAPs is
necessary to build up an environment for cells, which imitates the natural tissue
composition. Therefore, different CAPs must be combined in different concentrations
and screened for the cell behaviour in comparison to the natural tissue.
3. Results & Discussion
40
With this study it is aimed to address the problem of having a limited CAP library
available by assessing cell adhesion of CAPs of different ECM proteins and comparing
the results to the CRGDSGK sequence from section 3.2.4. A screen of the four peptide
sequences CDGEA, available in collagen I, CKRSR, which can be found in
osteopontin, CLDV, from the fibronectin protein and CAELDVP from the
thrombospondin-1 protein was conducted. Six hydrogels with a volume of 3 µL each
were modified with the CAPs giving a total concentration of 2 mM per gel. Since it was
not known how well the peptides support the cell adhesion process, the 3T3 cells
embedded in the hydrogels were cultured for 48 h, before they were fixed and stained
with FITC-tagged Phalloidin and DAPI. Figure 19 shows the result of the screening.
3. Results & Discussion
41
Figure 19: Confocal images of 3T3 cells stained with Phalloidin and DAPI in 10 % (w/v) hydrogels
modified with different CAPs. The images were acquired by running Z-stacks of 120-160 µm through
the hydrogel and superimposing them with a maximum intensity stack. (A-C) Positive control for cell
spreading, using the CRGDSGK sequence. Compared to this positive control were the CAPs CDGEA
(D-F), CKRSR (G-J), CLDV (K-M) and CAELDVP (N-P). The scale bar on each image equals 100 µm.
3. Results & Discussion
42
None of the peptides supports the cell spreading more, or even in a similar way than
CRGDSGK, given the lower number of cell-cell connections and actin spreading.
However, in all samples a non-spherical cell shape can be observed, indicating that
cell adhesion was supported by all the CAPs. The CDGEA CAP is a modification of
the Collagen I adhesion peptide sequence DGEA. The terminal cysteine unit was
introduced to covalently bind the peptide to the hydrogel. Cells show low spreading for
this CAP. In a recent study by Bi et al. they could show that DGEA together with a
collagen coating on scaffolds could improve skull defect healing.[112] Therefore, a
reason for the lack in spreading could be found in the type of cells used. However, the
use of 3T3 fibroblasts was interesting in that sense that these cells express collagen I
as part of their ECM[113]. Therefore, it could be seen whether the CAP, which the
fibroblasts normally express feed back into the cell adhesion of fibroblasts to their own
ECM. In another study by Weber et al., in which the cell viability and insulin expression
of murine pancreatic MIN6 β-cells were screened as a function of different CAP
combinations and concentrations, they could show that DGEA was also not able to
support cell viability and insulin production in contrast to peptides like IKVAV and
IKLLI.[59] They also could show that combinations of CAPs like PDSGR and YIGSR
were more viable to enhance cell-matrix interactions and insulin secretion than the
single CAPs. DGEA alone was not able to enhance cell viability. Insulin production of
the cells was neither supported by the used RGD sequence, nor by DGEA. All in all,
this shows limitations of these CAPs in pancreatic tissue models. On the other hand,
Wu et al. have already shown that DGEA can support adhesion and ECM deposition
of valve interstitial cells, when combined with the CAP RGDS.[114] Besides, they also
found out that the choice of CAP combination is affecting interaction of cells with the
environment. So, when they combined DGEA with YIGSR, it did promote ECM
expression of the cell, but inhibited cell differentiation.
The same can be applied for the result given by the CKRSR peptide. KRSR, which is
available in the ECM macromolecule osteopontin is known to promote adhesion of
osteoblasts through proteoglycans like heparin sulphate.[115] Despite that, a study by
Sawyer et al. showed that KRSR does not show enhanced mesenchymal stem cell or
osteoblast attachment on hydroxyapatite, relative to GRGDSPCA.[116] Therefore, this
CAP alone is not as good to promote the cell adhesion process as the RGD sequence
3. Results & Discussion
43
is. This behaviour was found here as well with fibroblast cells. Therefore, this CAP
alone might not be suitable to promote the cell adhesion process.
The last two CAPs that were screened in this experiment, CLDV and CAELDVP,
demonstrate that there might be also a correlation between CAP lengths. Both
peptides show more cell spreading than CKRSR and CDGEA, but still less than the
hydrogel with CRGDSGK. Further, CLDV, which is present in fibronectin, shows a
lower spreading of the cells than the longer thrombospondin-1 sequence CAELDVP.
Thrombospondin-1 is a ECM protein that is involved in a variety of cellular and
molecular mechanisms like binding to cell-surface receptors like integrins and
proteoglycans or participating in embryonic development or tumour metastasis.[117]
Therefore, it can be assumed that the difference in spreading compared to CKRSR
and CDGEA is due to their less specific protein origin. CLDV is abundant in fibronectin,
CAELDVP in thrombospondin-1, which binds to a variety of ECM macromolecules, i.e.
laminin, collagens, fibronectin or even some proteases.[118] This binding ability to
several proteins controls indirectly many cell-matrix interactions, since it changes the
biological properties of the EMC protein it is attached to. This binding to other ECM
macromolecules could indicate a support role of CAELDVP in terms of cell adhesion
promotion, which offers the opportunity of combinations with already proven CAPs like
IKVAV or YIGSR.
These results are a first step to expand the library of CAPs that can be used in tissue
engineering. In general, it should be considered that in this experiment only one
concentration of each CAP was assessed. The composition of the natural ECM is
highly complex and changes from tissue to tissue and, therefore, for every cell type as
well. Thus, in future experiments multiple concentration should be screened with the
aim to imitate a specific natural tissue.
Nevertheless, the conducted CAP screening experiment could show that it is useful to
think beyond RGD and establish new CAPs that lead to more tailorable in vitro tissue
models that mimic the natural ECM of the corresponding cell type accurately.
3. Results & Discussion
44
3.2.7 Cell Deposition with the Robot After demonstrating the screening of a variety of CAPs manually, the next step to
screen their influence on cell adhesion in an efficient manner is the automating by
using a robotic system. It is possible to screen different CAPs for cell adhesion by
pipetting the hydrogel system by hand. However, when using combinations of different
CAPs, cell types and a variety of concentrations, the amount of time quickly exceeds
a reasonable timeframe. For engineering tissue models, it is important that
concentration gradients, different CAP combinations and cell types can be screened
effectively in a matter of hours instead of days or weeks. To scale up the method
proposed in this thesis, a QIAgility robot by QIAGEN, initially used for RT-PCR
experiments, was adjusted to be able to deposit hydrogels, which are reproducible in
size and composition.
As a first test to assess the cell viability and recovery after experiencing mechanical
stress, when they are sheared through the head of the robot pipette tips, 3T3 cells
were deposited by the robot and the spreading and morphology of the cells compared
to manually deposited cells. Both cell samples were cultured for 24 h before imaging.
Figure 20 shows the result of this test.
Figure 20: Comparison of cell spreading between 3T3 cells deposited on a 96-well plate by hand (A)
and the QIAgility robot (B). The scale bar on each image equals 100 µm.
As the image shows, the cell morphology, as well as the spreading is in both cases
similar. It can be assumed that the cells recover after 24 h from their stress and start
adhering to the tissue culture plate they were seeded in. It was not possible to conduct
quantitative measurements, since the robot pipette head, due to the age of the robot,
was not calibrated, which made it impossible, especially when using small volumes,
to accurately measure volumes.
3. Results & Discussion
45
However, it can be predicted that the handling of cells by the robot does not affect their
ability to adhere and spread.
3.2.8 Blank Gel Deposition with the Robot It was now important to see whether the hydrogel system was compatible with the
processing by the robot. The criterion were the rheological properties of the hydrogel
system. As previously determined, the gelation time of the system is approximately
16 minutes, which is in average 6 minutes more than the robot would need to run a
standard program, which involved the automated deposition of hydrogels. On the other
hand, the rheology data (Figure 12) showed a constant viscosity of the precursor
solutions, which was comparable to water. Thus, the robot should be able to deposit
consistent hydrogels, without experiencing problems like partial gelation at the pipette
tip or in the stock solution. To verify this prediction, a large stock containing material
for 20 hydrogels (3 µL) was prepared by the robot using stock solutions of both
polymers, PEG-4VS and PEG-4SH. Figure 21 shows the row of hydrogels deposited
and coloured in blue food colorant.
Figure 21: Image of a row of PEG-4VS/PEG-4SH hydrogels deposited by the QIAgility robot into a 96-
well plate. All droplets had a theoretical volume of 3 µL. The droplets were coloured using a blue food
colorant to enhance the visibility on the white background.
All the hydrogel droplets were similar in size, measured in ImageJ. Therefore, the robot
is also able to deposit an array of reproducible hydrogels, which makes it to a viable
candidate for the automation of the CAP screening.
4. Conclusion & Outlook
46
4. Conclusion & Outlook With the work conducted in this project it was a first step in thinking beyond the
well-established RGD sequence. It could be shown that the approach of using
hydrogel droplets with covalently attached CAPs is a viable approach to screen for cell
adhesion and spreading in a 3D environment. Furthermore, CAPs were assessed
regarding their ability to promote the cell adhesion of a 3T3 fibroblast cell line by
comparison to an RGD positive control. All these four CAPs were able to promote cell
spreading to a certain extend. The thrombospondin-1 sequence CAELDVP
demonstrated the best cell spreading amongst all CAPs. By using different cell types
for the screening, depending on the mimicked tissue, specific CAPs, like KRSR for
bone tissue engineering, could be applied.
It could be also shown that the stiffness of the matrix has an influence on the cell
spreading and that the influence of CAPs on this process gets less dominant the higher
the hydrogel stiffness is. This could be a base for future experiments on diseased
tissue engineering, featuring stiffer hydrogel matrices that would normally not be used
for cell culture or tissue engineering because of their high stiffness. Higher stiffness
also resulted in forcing the cells into larger cell constructs, rather than supporting
single cell-cell interactions.
For future experiments, CAPs should be chosen carefully, corresponding to the ECM
environment that is to be mimicked. Therefore, it is also necessary to use
combinations and different concentrations of these CAPs, to take the complexity of the
natural tissue into account. On the other hand, it will be important to find a viable way
to quantify the amount of CAP, which is covalently bound to the hydrogel. This is
important to make sure that the theoretical amounts of peptide are the same as the
actual ones. One way to do this would be to find a more reliable colorimetric assay
than the BCA or the Ninhydrin assay, which were assessed in this thesis. Both assays
did not work properly because of side reactions of the reagents with the hydrogel.
Another possibility is to use elemental analysis of a blank hydrogel without CAP and
then one with bound CAP, to then calculate the actual amount from the ratio of the
elements. This approach, however, would be not suitable for HTS and could only
analyse one sample per CAP.
4. Conclusion & Outlook
47
In summary, this project showed the viability of novel, mainly unused CAPs derived
from ECM proteins in controlling the cell adhesion process on a PEG-based hydrogel
system. In future, by screening a library of CAPs[93], this could pave a new path to
tailored and cell specific microenvironments in in vitro models.
Bibliography
48
Bibliography [1] A. Forget, J. Christensen, S. Ludeke, E. Kohler, S. Tobias, M. Matloubi, R.
Thomann, V.P. Shastri, Polysaccharide hydrogels with tunable stiffness and
provasculogenic properties via α-helix to β-sheet switch in secondary structure,
Proc. Natl. Acad. Sci. 110 (2013) 12887–12892. doi:10.1073/pnas.1222880110.
[2] K. Wang, X. Zhang, L. Zhang, L. Qian, C. Liu, J. Zheng, Y. Jiang, Development
of biodegradable polymeric implants of RGD-modified PEG-PAMAM-DOX
conjugates for long-term intratumoral release, Drug Deliv. 22 (2015) 389–399.
doi:10.3109/10717544.2014.895457.
[3] J. Li, L. Zheng, L. Zeng, Y. Zhang, L. Jiang, J. Song, RGD Peptide-Grafted
Graphene Oxide as a New Biomimetic Nanointerface for Impedance-Monitoring
Cell Behaviors, J. Nanomater. 2016 (2016). doi:10.1155/2016/2828512.
[4] J. Yu, A.-R. Lee, W.-H. Lin, C.-W. Lin, Y.-K. Wu, W.-B. Tsai, Electrospun PLGA
Fibers Incorporated with Functionalized Biomolecules for Cardiac Tissue
Engineering, Tissue Eng. Part A. 20 (2014) 1896–1907.
doi:10.1089/ten.tea.2013.0008.
[5] R. Fraioli, K. Dashnyam, J.H. Kim, R.A. Perez, H.W. Kim, J. Gil, M.P. Ginebra,
J.M. Manero, C. Mas-Moruno, Surface guidance of stem cell behavior:
Chemically tailored co-presentation of integrin-binding peptides stimulates
osteogenic differentiation in vitro and bone formation in vivo, Acta Biomater. 43
(2016) 269–281. doi:10.1016/j.actbio.2016.07.049.
[6] S.J. Attwood, E. Cortes, A.W.M. Haining, B. Robinson, D. Li, J. Gautrot, A. del
Río Hernández, Adhesive ligand tether length affects the size and length of focal
adhesions and influences cell spreading and attachment, Sci. Rep. 6 (2016)
34334–34345. doi:10.1038/srep34334.
[7] S. Sridar, M.A. Churchward, V.K. Mushahwar, K.G. Todd, A.L. Elias, Peptide
modification of polyimide-insulated microwires: Towards improved
biocompatibility through reduced glial scarring, Acta Biomater. 60 (2017) 154–
166. doi:10.1016/j.actbio.2017.07.026.
[8] R. Mobasseri, L. Tian, M. Soleimani, S. Ramakrishna, H. Naderi-Manesh, Bio-
Bibliography
49
active molecules modified surfaces enhanced mesenchymal stem cell adhesion
and proliferation, Biochem. Biophys. Res. Commun. 483 (2017) 312–317.
doi:10.1016/j.bbrc.2016.12.146.
[9] C.G. Knight, L.F. Morton, a R. Peachey, D.S. Tuckwell, R.W. Farndale, M.J.
Barnes, The collagen-binding A-domains of integrins alpha(1)beta(1) and
alpha(2)beta(1) recognize the same specific amino acid sequence, GFOGER,
in native (triple-helical) collagens., J. Biol. Chem. 275 (2000) 35–40.
doi:10.1074/jbc.275.1.35.
[10] W.D. Staatz, K.F. Fok, M.M. Zutter, S.P. Adams, B.A. Rodriguez, S.A. Santoro,
Identification of a tetrapeptide recognition sequence for the α2β1 integrin in
collagen, J. Biol. Chem. 266 (1991) 7363–7367.
[11] T. Maeda, K. Titani, K. Sekiguchi, Cell-adhesive activity and receptor-binding
specificity of the laminin-derived YIGSR sequence grafted onto Staphylococcal
protein A., J. Biochem. 115 (1994) 182–189.
[12] B. Clement, B. Segui-Real, P. Savagner, H.K. Kleinman, Y. Yamada,
Hepatocyte attachment to laminin is mediated through multiple receptors, J. Cell
Biol. 110 (1990) 185–192. doi:10.1083/jcb.110.1.185.
[13] H.K. Kleinman, J. Graf, Y. Iwamoto, M. Sasaki, C.S. Schasteen, Y. Yamada,
G.R. Martin, F.A. Robey, Identification of a second active site in laminin for
promotion of cell adhesion and migration and inhibition of in vivo melanoma lung
colonization, Arch. Biochem. Biophys. 272 (1989) 39–45. doi:10.1016/0003-
9861(89)90192-6.
[14] V.M. Freitas, V.F. Vilas-Boas, D.C. Pimenta, V. Loureiro, M.A. Juliano, M.R.
Carvalho, J.J. V Pinheiro, A.C.M. Camargo, A.S. Moriscot, M.P. Hoffman, R.G.
Jaeger, SIKVAV, a Laminin α1-Derived Peptide, Interacts with Integrins and
Increases Protease Activity of a Human Salivary Gland Adenoid Cystic
Carcinoma Cell Line through the ERK 1/2 Signaling Pathway, Am. J. Pathol. 171
(2007) 124–138. doi:http://dx.doi.org/10.2353/ajpath.2007.051264.
[15] K.J. Tomaselli, D.E. Hall, L.A. Flier, K.R. Gehlsen, D.C. Turner, S. Carbonetto,
L.F. Reichardt, A neuronal cell line (PC12) expresses two β1-class integrins—
α1β1, and α3β1—that recognize different neurite outgrowth-promoting domains
Bibliography
50
in laminin, Neuron. 5 (1990) 651–662. doi:http://dx.doi.org/10.1016/0896-
6273(90)90219-6.
[16] H.K. Kleinman, B.S. Weeks, F.B. Cannon, T.M. Sweeney, G.C. Sephel, B.
Clement, M. Zain, M.O.J. Olson, M. Jucker, B.A. Burrous, Identification of a 110-
kDa nonintegrin cell surface laminin-binding protein which recognizes an a chain
neurite-promoting peptide, Arch. Biochem. Biophys. 290 (1991) 320–325.
doi:10.1016/0003-9861(91)90547-V.
[17] L. Romagnano, B. Babiarz, Mechanisms of murine trophoblast interaction with
laminin., Biol. Reprod.49 (1993) 374-80. http://www.ncbi.nlm.nih.gov/pubmed/8
373963.
[18] K. ‐I Tashiro, G.C. Sephel, D. Greatorex, M. Sasaki, N. Shirashi, G.R. Martin,
H.K. Kleinman, Y. Yamada, The RGD containing site of the mouse laminin A
chain is active for cell attachment, spreading, migration and neurite outgrowth,
J. Cell. Physiol. 146 (1991) 451–459. doi:10.1002/jcp.1041460316.
[19] Y. Ohga, F. Katagiri, K. Takeyama, K. Hozumi, Y. Kikkawa, N. Nishi, M. Nomizu,
Design and activity of multifunctional fibrils using receptor-specific small
peptides, Biomaterials. 30 (2009) 6731-6738. doi:http://dx.doi.org/10.1016/j.bio
materials.2009.08.044.
[20] M.L. Matter, G.W. Laurie, A novel laminin E8 cell adhesion site required for lung
alveolar formation in vitro, J. Cell Biol. 124 (1994) 1083–1090.
[21] D.D. Hunter, N. Cashman, J.F. Sanes, R. Morris-valero, P. Adams, W. Bulock,
An LRE ( Leucine-Arginine-Glutamate ) -dependent Adhesion of Neurons to S-
laminin Mechanism, J. Neurosci. 11 (1991) 3960–3971.
[22] H. Nakahara, M. Nomizu, S.K. Akiyama, Y. Yamada, Y. Yeh, W.-T. Chen, A
Mechanism for Regulation of Melanoma Invasion LIGATION OF α6β1
INTEGRIN BY LAMININ G PEPTIDES, J. Biol. Chem. 271 (1996) 27221–27224.
[23] M.L. Ponce, M. Nomizu, H.K. Kleinman, An angiogenic laminin site and its
antagonist bind through the αvβ3 and α5β1 integrins, FASEB J. 15 (2001) 1389–
1397. doi:10.1096/fj.00-0736com.
[24] R.P. Mecham, A. Hinek, G.L. Griffin, R.M. Senior, L.A. Liotta, The elastin
Bibliography
51
receptor shows structural and functional similarities to the 67-kDa tumor cell
laminin receptor, J. Biol. Chem. 264 (1989) 16652–16657.
[25] E. Koivunen, D. a Gay, E. Ruoslahti, Selection of peptides binding to the alpha
5 beta 1 integrin from phage display library., J. Biol. Chem. 268 (1993) 20205–
20210. doi:#.
[26] a P. Mould, M.J. Humphries, Identification of a novel recognition sequence for
the integrin alpha 4 beta 1 in the COOH-terminal heparin-binding domain of
fibronectin., EMBO J. 10 (1991) 4089–4095.
[27] M.J. Humphries, A. Komoriya, S.K. Akiyama, K. Olden, K.M. Yamada,
Identification of two distinct regions of the type III connecting segment of human
plasma fibronectin that promote cell type-specific adhesion., J. Biol. Chem. 262
(1987) 6886–6892.
[28] S. Aota, M. Nomizu, K.M. Yamada, The short amino acid sequence Pro-His-Ser-
Arg-Asn in human fibronectin enhances cell-adhesive function., J. Biol. Chem.
269 (1994) 24756–24761.
[29] M. Kloczewiak, S. Timmons, T.J. Lukas, J. Hawiger, Platelet receptor
recognition site on human fibrinogen. Synthesis and structure-function
relationship of peptides corresponding to the carboxy-terminal segment of the.
gamma. chain, Biochemistry. 23 (1984) 1767–1774.
[30] A. Komoriya, L.J. Green, M. Mervic, S.S. Yamada, K.M. Yamada, M.J.
Humphries, The minimal essential sequence for a major cell type-specific
adhesion site (CS1) within the alternatively spliced type III connecting segment
domain of fibronectin is leucine-aspartic acid-valine, J. Biol. Chem. 266 (1991)
15075–15079.
[31] C. Rüegg, A.A. Postigo, E.E. Sikorski, E.C. Butcher, R. Pytela, D.J. Erle, Role
of integrin α4β7/α4βP in lymphocyte adherence to fibronectin and VCAM-1 and
in homotypic cell clustering, J. Cell Biol. 117 (1992) 179–189.
[32] F. Ouchani, J. Devy, A. Rusciani, J.J. Helesbeux, S. Salesse, I. Letinois, D.
Gras-Billart, L. Duca, O. Duval, L. Martiny, E. Charpentier, Targeting focal
adhesion assembly by ethoxyfagaronine prevents lymphoblastic cell adhesion
Bibliography
52
to fibronectin, Anal. Cell. Pathol. 35 (2012) 267–284. doi:10.3233/ACP-2012-
0055.
[33] A. Woods, J.B. McCarthy, L.T. Furcht, J.R. Couchman, A synthetic peptide from
the COOH-terminal heparin-binding domain of fibronectin promotes focal
adhesion formation., Mol. Biol. Cell. 4 (1993) 605–613.
[34] S.L. Drake, J. Varnum, K.H. Mayo, P.C. Letourneau, L.T. Furcht, J.B. McCarthy,
Structural features of fibronectin synthetic peptide FN-C/H II, responsible for cell
adhesion, neurite extension, and heparan sulfate binding., J. Biol. Chem. 268
(1993) 15859–15867.
[35] S. Morone, S. Augeri, M. Cuccioloni, M. Mozzicafreddo, M. Angeletti, N. Lo
Buono, A. Giacomino, E. Ortolan, A. Funaro, Binding of CD157 protein to
fibronectin regulates cell adhesion and spreading, J. Biol. Chem. 289 (2014)
15588–15601.
[36] J. V Moyano, B. Carnemolla, C. Domınguez-Jiménez, M. Garcıa-Gila, J.P.
Albar, P. Sánchez-Aparicio, A. Leprini, G. Querzé, L. Zardi, A. Garcia-Pardo,
Fibronectin type III5 repeat contains a novel cell adhesion sequence, KLDAPT,
which binds activated α4β1 and α4β7 integrins, J. Biol. Chem. 272 (1997)
24832–24836.
[37] A. Sharma, J.A. Askari, M.J. Humphries, E.Y. Jones, D.I. Stuart, Crystal
structure of a heparin‐and integrin‐binding segment of human fibronectin, EMBO
J. 18 (1999) 1468–1479.
[38] B.E. Vogel, S.J. Lee, A. Hildebrand, W. Craig, M.D. Pierschbacher, F. Wong-
Staal, E. Ruoslahti, A novel integrin specificity exemplified by binding of the αvβ5
integrin to the basic domain of the HIV Tat protein and vitronectin, J. Cell Biol.
121 (1993) 461–468. doi:10.1083/jcb.121.2.461.
[39] K.C. Dee, T.T. Andersen, R. Bizios, Design and function of novel osteoblast-
adhesive peptides for chemical modification of biomaterials, J. Biomed. Mater.
Res. 40 (1998) 371-377. doi:10.1002/(SICI)1097-4636(19980605)40:3<371::AI
D-JBM5>3.0.CO;2-C.
[40] A. Rezania, K.E. Healy, Biomimetic peptide surfaces that regulate adhesion,
Bibliography
53
spreading, cytoskeletal organization, and mineralization of the matrix deposited
by osteoblast‐like cells, Biotechnol. Prog. 15 (1999) 19–32.
[41] G.M. Harbers, K.E. Healy, The effect of ligand type and density on osteoblast
adhesion, proliferation, and matrix mineralization, J. Biomed. Mater. Res. - Part
A. 75 (2005) 855–869. doi:10.1002/jbm.a.30482.
[42] K.J. Bayless, G.E. Davis, Identification of dual α4β1 integrin binding sites within
a 38 amino acid domain in the N-terminal thrombin fragment of human
osteopontin, J. Biol. Chem. 276 (2001) 13483–13489.
[43] C. García-Arévalo, M. Pierna, A. Girotti, F.J. Arias, J.C. Rodríguez-Cabello, A
comparative study of cell behavior on different energetic and bioactive polymeric
surfaces made from elastin-like recombinamers, Soft Matter. 8 (2012) 3239.
doi:10.1039/c2sm06893a.
[44] R.P. Mecham, A. Hinek, R. Entwistle, D.S. Wrenn, G.L. Griffin, R.M. Senior,
Elastin binds to a multifunctional 67-kilodalton peripheral membrane protein,
Biochemistry. 28 (1989) 3716–3722.
[45] N. Floquet, S. Pasco, L. Ramont, P. Derreumaux, J.Y. Laronze, J.M. Nuzillard,
F.X. Maquart, A.J.P. Alix, J.C. Monboisse, The antitumor properties of the α3
(IV)-(185-203) peptide from the NC1 domain of type IV collagen (tumstatin) are
conformation-dependent, J. Biol. Chem. 279 (2004) 2091–2100.
[46] N.-H. Guo, H.C. Krutzsch, E. Negre, T. Vogel, D.A. Blake, D.D. Roberts,
Heparin- and Sulfatide-Binding Peptides from the Type I Repeats of Human
Thrombospondin Promote Melanoma Cell Adhesion, Proc. Natl. Acad. Sci. U.
S. A. 89 (1992) 3040–3044. http://www.jstor.org/stable/2359029.
[47] J. Lawler, R. Weinstein, R.O. Hynes, Cell attachment to thrombospondin: the
role of ARG-GLY-ASP, calcium, and integrin receptors., J. Cell Biol. 107 (1988)
2351–2361.
[48] H.C. Krutzsch, B.J. Choe, J.M. Sipes, N. h Guo, D.D. Roberts, Identification of
an α3β1 integrin recognition sequence in thrombospondin-1, J. Biol. Chem. 274
(1999) 24080–24086. doi:10.1074/JBC.274.34.24080.
[49] Z. Li, M.J. Calzada, J.M. Sipes, J.A. Cashel, H.C. Krutzsch, D.S. Annis, D.F.
Bibliography
54
Mosher, D.D. Roberts, Interactions of thrombospondins with α4β1 integrin and
CD47 differentially modulate T cell behavior, J. Cell Biol. 157 (2002) 509–519.
doi:10.1083/jcb.200109098.
[50] F.D. Yelian, N.A. Edgeworth, L.J. Dong, A.E. Chung, D.R. Armant, Recombinant
entactin promotes mouse primary trophoblast cell adhesion and migration
through the Arg-Gly-Asp (RGD) recognition sequence [published erratum
appears in J Cell Biol 1993 Jul;122(1):279], J Cell Biol. 121 (1993) 923–929.
[51] R.M. Senior, H.D. Gresham, G.L. Griffin, E.J. Brown, A.E. Chung, Entactin
stimulates neutrophil adhesion and chemotaxis through interactions between its
Arg-Gly-Asp (RGD) domain and the leukocyte response integrin., J. Clin. Invest.
90 (1992) 2251–2257.
[52] S. Pacheco, T. Kanou, S.-Y. Fung, K. Chen, D. Lee, X. Bai, S. Keshavjee, M.
Liu, Formulation of hydrophobic therapeutics with self-assembling peptide and
amino acid: A new platform for intravenous drug delivery, J. Control. Release.
239 (2016) 211–222. doi:10.1016/j.jconrel.2016.08.038.
[53] A.N. Hoofnagle, J.R. Whiteaker, S.A. Carr, E. Kuhn, T. Liu, S.A. Massoni, S.N.
Thomas, R. Reid Townsend, L.J. Zimmerman, E. Boja, J. Chen, D.L. Crimmins,
S.R. Davies, Y. Gao, T.R. Hiltke, K.A. Ketchum, C.R. Kinsinger, M. Mesri, M.R.
Meyer, W.J. Qian, R.M. Schoenherr, M.G. Scott, T. Shi, G.R. Whiteley, J.A.
Wrobel, C. Wu, B.L. Ackermann, R. Aebersold, D.R. Barnidge, D.M. Bunk, N.
Clarke, J.B. Fishman, R.P. Grant, U. Kusebauch, M.M. Kushnir, M.S. Lowenthal,
R.L. Moritz, H. Neubert, S.D. Patterson, A.L. Rockwood, J. Rogers, R.J. Singh,
J.E. Van Eyk, S.H. Wong, S. Zhang, D.W. Chan, X. Chen, M.J. Ellis, D.C.
Liebler, K.D. Rodland, H. Rodriguez, R.D. Smith, Z. Zhang, H. Zhang, A.G.
Paulovich, Recommendations for the generation, quantification, storage, and
handling of peptides used for mass spectrometry-based assays, Clin. Chem. 62
(2016) 48–69. doi:10.1373/clinchem.2015.250563.
[54] M. Gallego, L. Mora, M. Reig, F. Toldrá, Stability of the potent antioxidant
peptide SNAAC identified from Spanish dry-cured ham, Food Res. Int. 105
(2018) 873–879. doi:10.1016/j.foodres.2017.12.006.
[55] N. Pawar, P. Kaushik, H.B. Bohidar, Hydrophobic hydration and anomalous
Bibliography
55
diffusion of elastin in an ethanolic solution, Phys. Chem. Chem. Phys. 19 (2017)
13994–14000. doi:10.1039/C7CP01384A.
[56] A.C. Mora, C.E.H. Schmelzer, C. Luise, W. Sippl, M. Pietzsch, W.
Hoehenwarter, A. Heinz, Degradation of tropoelastin and skin elastin by
neprilysin, Biochimie. 146 (2018) 73–78. doi:10.1016/j.biochi.2017.11.018.
[57] S. Li, L.R. Nih, H. Bachman, P. Fei, Y. Li, E. Nam, R. Dimatteo, S.T. Carmichael,
T.H. Barker, T. Segura, Hydrogels with precisely controlled integrin activation
dictate vascular patterning and permeability, Nat. Mater. 16 (2017) 953–961.
doi:10.1038/nmat4954.
[58] L. Zhu, K. Wang, T. Ma, L. Huang, B. Xia, S. Zhu, Y. Yang, Z. Liu, X. Quan, K.
Luo, D. Kong, J. Huang, Z. Luo, Noncovalent Bonding of RGD and YIGSR to an
Electrospun Poly(ε-Caprolactone) Conduit through Peptide Self-Assembly to
Synergistically Promote Sciatic Nerve Regeneration in Rats, Adv. Healthc.
Mater. 6 (2017) 1600860–1600874. doi:10.1002/adhm.201600860.
[59] L.M. Weber, K.N. Hayda, K. Haskins, K.S. Anseth, The effects of cell-matrix
interactions on encapsulated β-cell function within hydrogels functionalized with
matrix-derived adhesive peptides, Biomaterials. 28 (2007) 3004–3011.
doi:10.1016/j.biomaterials.2007.03.005.
[60] H. Higashida, M. Liang, T. Yoshihara, S. Akther, A. Fakhrul, C. Stanislav, T.-S.
Nam, U.-H. Kim, S. Kasai, T. Nishimura, N. Al Mahmuda, S. Yokoyama, K.
Ishihara, M. Gerasimenko, A. Salmina, J. Zhong, T. Tsuji, C. Tsuji, O. Lopatina,
An immunohistochemical, enzymatic, and behavioral study of CD157/BST-1 as
a neuroregulator, BMC Neurosci. 18 (2017) 35–47. doi:10.1186/s12868-017-
0350-7.
[61] J.H. Wen, L.G. Vincent, A. Fuhrmann, Y.S. Choi, K.C. Hribar, H. Taylor-Weiner,
S. Chen, A.J. Engler, Interplay of matrix stiffness and protein tethering in stem
cell differentiation, Nat. Mater. 13 (2014) 979–987. doi:10.1038/nmat4051.
[62] A. Llacua, B.J. De Haan, S.A. Smink, P. De Vos, Extracellular matrix
components supporting human islet function in alginate-based
immunoprotective microcapsules for treatment of diabetes, J. Biomed. Mater.
Res. - Part A. 104 (2016) 1788–1796. doi:10.1002/jbm.a.35706.
Bibliography
56
[63] N. Gjorevski, N. Sachs, A. Manfrin, S. Giger, M.E. Bragina, P. Ordóñez-Morán,
H. Clevers, M.P. Lutolf, Designer matrices for intestinal stem cell and organoid
culture, Nature. 539 (2016) 560–564. doi:10.1038/nature20168.
[64] R. Macarron, M.N. Banks, D. Bojanic, D.J. Burns, D.A. Cirovic, T. Garyantes,
D.V.S. Green, R.P. Hertzberg, W.P. Janzen, J.W. Paslay, U. Schopfer, G.S.
Sittampalam, Impact of high-throughput screening in biomedical research, Nat.
Rev. Drug Discov. 10 (2011) 188–195. doi:10.1038/nrd3368.
[65] S. Rasi Ghaemi, F. Harding, B. Delalat, R. Vasani, N.H. Voelcker, Surface
engineering for long-term culturing of mesenchymal stem cell microarrays,
Biomacromolecules. 14 (2013) 2675–2683. doi:10.1021/bm400531n.
[66] A. Forget, A.L.S. Burzava, B. Delalat, K. Vasilev, F.J. Harding, A. Blencowe,
N.H. Voelcker, Rapid fabrication of functionalised poly (dimethylsiloxane)
microwells for cell aggregate formation, Biomater. Sci. 5 (2017) 828–836.
[67] A. Forget, M. Waibel, D.M. Rojas-Canales, S. Chen, N. Kawazoe, F.J. Harding,
T. Loudovaris, P.T.H. Coates, A. Blencowe, G. Chen, IGF-2 coated porous
collagen microwells for the culture of pancreatic islets, J. Mater. Chem. B. 5
(2017) 220–225.
[68] J.C. Grim, I.A. Marozas, K.S. Anseth, Thiol-ene and photo-cleavage chemistry
for controlled presentation of biomolecules in hydrogels, J. Control. Release.
219 (2015) 95–106.
[69] M. Pagel, R. Meier, K. Braun, M. Wiessler, A.G. Beck-Sickinger, On-resin Diels–
Alder reaction with inverse electron demand: an efficient ligation method for
complex peptides with a varying spacer to optimize cell adhesion, Org. Biomol.
Chem. 14 (2016) 4809–4816. doi:10.1039/C6OB00314A.
[70] F. Lin, J. Yu, W. Tang, J. Zheng, A. Defante, K. Guo, C. Wesdemiotis, M.L.
Becker, Peptide-functionalized oxime hydrogels with tunable mechanical
properties and gelation behavior, Biomacromolecules. 14 (2013) 3749–3758.
doi:10.1021/bm401133r.
[71] W. Tang, M.L. Becker, “Click” reactions: a versatile toolbox for the synthesis of
peptide conjugates, Chem. Soc. Rev. 43 (2014) 7013-7039. doi:10.1039/C4CS
Bibliography
57
00139G.
[72] F. Yu, X. Cao, Y. Li, L. Zeng, B. Yuan, X. Chen, An injectable hyaluronic
acid/PEG hydrogel for cartilage tissue engineering formed by integrating
enzymatic crosslinking and Diels–Alder “click chemistry,” Polym. Chem. 5
(2014) 1082–1090. doi:10.1039/C3PY00869J.
[73] M.W. Toepke, N.A. Impellitteri, J.M. Theisen, W.L. Murphy, Characterization of
thiol-ene crosslinked PEG hydrogels, Macromol. Mater. Eng. 298 (2013) 699–
703. doi:10.1002/mame.201200119.
[74] A. Farrukh, J.I. Paez, M. Salierno, W. Fan, B. Berninger, A. Del Campo,
Bifunctional Poly(acrylamide) Hydrogels through Orthogonal Coupling
Chemistries, Biomacromolecules. 18 (2017) 906-913. doi:10.1021/acs.biomac.
6b01784.
[75] E.A. Phelps, K.L. Templeman, P.M. Thulé, A.J. García, Engineered VEGF-
releasing PEG-MAL hydrogel for pancreatic islet vascularization, Drug Deliv.
Transl. Res. 5 (2015) 125–136. doi:10.1007/s13346-013-0142-2.
[76] A. Forget, A. Blaeser, F. Miessmer, M. Köpf, D.F.D. Campos, N.H. Voelcker, A.
Blencowe, H. Fischer, V.P. Shastri, Mechanically Tunable Bioink for 3D
Bioprinting of Human Cells, Adv. Healthc. Mater. 6 (2017) 1700255.
doi:10.1002/adhm.201700255.
[77] A. Forget, J. Christensen, S. Lüdeke, E. Kohler, S. Tobias, M. Matloubi, R.
Thomann, V.P. Shastri, Polysaccharide hydrogels with tunable stiffness and
provasculogenic properties via α-helix to β-sheet switch in secondary structure,
Proc. Natl. Acad. Sci. 110 (2013) 12887–12892.
[78] O. Sirenko, J. Hesley, I. Rusyn, E.F. Cromwell, High-Content High-Throughput
Assays for Characterizing the Viability and Morphology of Human iPSC-Derived
Neuronal Cultures, Assay Drug Dev. Technol. 12 (2014) 536–547.
doi:10.1089/adt.2014.592.
[79] M.K. Kanakasabapathy, M. Sadasivam, A. Singh, C. Preston, P. Thirumalaraju,
M. Venkataraman, C.L. Bormann, M.S. Draz, J.C. Petrozza, H. Shafiee, An
automated smartphone-based diagnostic assay for point-of-care semen
Bibliography
58
analysis, Sci. Transl. Med. 9 (2017) eaai7863.
doi:10.1126/scitranslmed.aai7863.
[80] C.C.F. Homem, I. Reichardt, C. Berger, T. Lendl, J.A. Knoblich, Long-term live
cell imaging and automated 4D analysis of Drosophila neuroblast lineages,
PLoS One. 8 (2013) 1–10. doi:10.1371/journal.pone.0079588.
[81] M.E. Tanenbaum, L.A. Gilbert, L.S. Qi, J.S. Weissman, R.D. Vale, A protein-
tagging system for signal amplification in gene expression and fluorescence
imaging, Cell. 159 (2014) 635–646. doi:10.1016/j.cell.2014.09.039.
[82] T. Ogura, A. Tsuchiya, T. Minas, S. Mizuno, Methods of high integrity RNA
extraction from cell/agarose construct, BMC Res. Notes. 8 (2015) 644.
doi:10.1186/s13104-015-1627-5.
[83] A. Ranga, S. Gobaa, Y. Okawa, K. Mosiewicz, A. Negro, M.P. Lutolf, 3D niche
microarrays for systems-level analyses of cell fate, Nat. Commun. 5 (2014) 1–
10. doi:10.1038/ncomms5324.
[84] S. Sharma, M. Floren, Y. Ding, K.R. Stenmark, W. Tan, S.J. Bryant, A
photoclickable peptide microarray platform for facile and rapid screening of 3-D
tissue-microenvironments, Biomaterials. 143 (2017) 17-28. doi:10.1016/j.biom
aterials.2017.07.025.
[85] P. Lu, V.M. Weaver, Z. Werb, The extracellular matrix: A dynamic niche in
cancer progression, J. Cell Biol. 196 (2012) 395-406. doi:10.1083/jcb.2011021
47.
[86] P. Liesi, T. Kauppila, Induction of Type IV Collagen and Other Basement-
Membrane-Associated Proteins after Spinal Cord Injury of the Adult Rat May
Participate in Formation of the Glial Scar, Exp. Neurol. 173 (2002) 31–45.
doi:10.1006/exnr.2001.7800.
[87] R. Bataller, D. Brenner, Liver fibrosis, J. Clin. Invest. 115 (2005) 209–218.
doi:10.1172/JCI200524282.The.
[88] P.H. Byers, M.L. Murray, Ehlers-Danlos syndrome: A showcase of conditions
that lead to understanding matrix biology, Matrix Biol. 33 (2014) 10–15.
doi:10.1016/j.matbio.2013.07.005.
Bibliography
59
[89] M.J. Merrilees, P.S. Ching, B. Beaumont, A. Hinek, T.N. Wight, P.N. Black,
Changes in elastin, elastin binding protein and versican in alveoli in chronic
obstructive pulmonary disease, Respir. Res. 9 (2008) 41–50. doi:10.1186/1465-
9921-9-41.
[90] R. Annoni, T. Lanças, R. Yukimatsu Tanigawa, M. de Medeiros Matsushita, S.
de Morais Fernezlian, A. Bruno, L. Fernando Ferraz da Silva, P.J. Roughley, S.
Battaglia, M. Dolhnikoff, P.S. Hiemstra, P.J. Sterk, K.F. Rabe, T. Mauad,
Extracellular matrix composition in COPD, Eur. Respir. J. 40 (2012) 1362–1373.
doi:10.1183/09031936.00192611.
[91] T.T. Lee, J.R. García, J.I. Paez, A. Singh, E.A. Phelps, S. Weis, Z. Shafiq, A.
Shekaran, A. del Campo, A.J. García, Light-triggered in vivo activation of
adhesive peptides regulates cell adhesion, inflammation and vascularization of
biomaterials, Nat. Mater. 14 (2015) 352–360. doi:10.1038/nmat4157.
[92] M.M. Laronda, A.L. Rutz, S. Xiao, K.A. Whelan, F.E. Duncan, E.W. Roth, T.K.
Woodruff, R.N. Shah, A bioprosthetic ovary created using 3D printed
microporous scaffolds restores ovarian function in sterilized mice, Nat.
Commun. 8 (2017) 15261. doi:10.1038/ncomms15261.
[93] N. Huettner, T.R. Dargaville, A. Forget, Discovering Cell-Adhesion Peptides in
Tissue Engineering: Beyond RGD, Trends Biotechnol. 36 (2018) 372–383.
doi:10.1016/j.tibtech.2018.01.008.
[94] J.A. Burdick, A. Khademhosseini, R. Langer, Fabrication of gradient hydrogels
using a microfluidics/photopolymerization process, Langmuir. 20 (2004) 5153–
5156. doi:10.1021/la049298n.
[95] J. Chen, G.H. Altman, V. Karageorgiou, R. Horan, A. Collette, V. Volloch, T.
Colabro, D.L. Kaplan, Human bone marrow stromal cell and ligament fibroblast
responses on RGD-modified silk fibers, J. Biomed. Mater. Res. 67A (2003) 559–
570. doi:10.1002/jbm.a.10120.
[96] J.A. Burdick, K.S. Anseth, Photoencapsulation of osteoblasts in injectable RGD-
modified PEG hydrogels for bone tissue engineering, Biomaterials. 23 (2002)
4315–4323. doi:10.1016/S0142-9612(02)00176-X.
Bibliography
60
[97] S. Wang, P.D. Dalton, T.R. Dargaville, Spatial Patterning of Hydrogels via 3D
Covalent Transfer Stamping from a Fugitive Ink, Macromol. Rapid Commun.
1700564 (2017) 3–7. doi:10.1002/marc.201700564.
[98] M.D. Swartzlander, C.A. Barnes, A.K. Blakney, J.L. Kaar, T.R. Kyriakides, S.J.
Bryant, Linking the foreign body response and protein adsorption to PEG-based
hydrogels using proteomics, Biomaterials. 41 (2015) 26–36.
doi:10.1016/j.biomaterials.2014.11.026.
[99] X. Wang, D.R. Schmidt, E.J. Joyce, W.J. Kao, Application of MS-Based
Proteomics to Study Serum Protein Adsorption/Absorption and Complement C3
Activation on Poly(ethylene glycol) Hydrogels, J. Biomater. Sci. Polym. Ed. 22
(2011) 1343–1362. doi:10.1163/092050610X508400.
[100] B. Hu, Z.-P. Yao, Detection of native proteins using solid-substrate electrospray
ionization mass spectrometry with nonpolar solvents, Anal. Chim. Acta. 1004
(2018) 51–57. doi:https://doi.org/10.1016/j.aca.2017.11.079.
[101] T. Ito, K. Okazaki, D. Nakajima, D. Shibata, S. Murakami, S. Schaffer, Mass
spectrometry-based metabolomics to identify taurine-modified metabolites in
heart, Amino Acids. 50 (2018) 117–124. doi:10.1007/s00726-017-2498-y.
[102] J.E. Leslie-Barbick, C. Shen, C. Chen, J.L. West, Micron-Scale Spatially
Patterned, Covalently Immobilized Vascular Endothelial Growth Factor on
Hydrogels Accelerates Endothelial Tubulogenesis and Increases Cellular
Angiogenic Responses, Tissue Eng. Part A. 17 (2011) 221–229.
doi:10.1089/ten.tea.2010.0202.
[103] J.A. Rowley, G. Madlambayan, D.J. Mooney, Alginate hydrogels as synthetic
extracellular matrix materials, Biomaterials. 20 (1999) 45–53.
doi:10.1016/S0142-9612(98)00107-0.
[104] B. Duan, Z. Yin, L. Hockaday Kang, R.L. Magin, J.T. Butcher, Active tissue
stiffness modulation controls valve interstitial cell phenotype and osteogenic
potential in 3D culture, Acta Biomater. 36 (2016) 42–54.
doi:10.1016/j.actbio.2016.03.007.
[105] J.S. Miller, C.J. Shen, W.R. Legant, J.D. Baranski, B.L. Blakely, C.S. Chen,
Bibliography
61
Bioactive hydrogels made from step-growth derived PEG-peptide macromers,
Biomaterials. 31 (2010) 3736–3743. doi:10.1016/j.biomaterials.2010.01.058.
[106] E. V. Alakpa, V. Jayawarna, K.E.V. Burgess, C.C. West, B. Péault, R. V. Ulijn,
M.J. Dalby, Improving cartilage phenotype from differentiated pericytes in
tunable peptide hydrogels, Sci. Rep. 7 (2017) 1–11. doi:10.1038/s41598-017-
07255-z.
[107] A.I. Khoruzhenko, 2D- and 3D-cell culture, Biopolym. Cell. 27 (2011) 17–24.
doi:10.7124/bc.00007D.
[108] M.F. Pang, M.J. Siedlik, S. Han, M. Stallings-Mann, D.C. Radisky, C.M. Nelson,
Tissue stiffness and hypoxia modulate the integrin-linked kinase ilk to control
breast cancer stem-like cells, Cancer Res. 76 (2016) 5277–5287.
doi:10.1158/0008-5472.CAN-16-0579.
[109] K.E. Shafer-Peltier, A.S. Haka, M. Fitzmaurice, J. Crowe, J. Myles, R.R. Dasari,
M.S. Feld, Raman microspectroscopic model of human breast tissue:
Implications for breast cancer diagnosis in vivo, J. Raman Spectrosc. 33 (2002)
552–563. doi:10.1002/jrs.877.
[110] F. Grinnell, C. Ho, G. Skuta, F. Grinnell, C. Ho, Y. Lin, G. Skuta, CELL BIOLOGY
AND METABOLISM : Differences in the Regulation of Fibroblast Contraction of
Floating Versus Stressed Collagen Matrices Differences in the Regulation of
Fibroblast Contraction of Floating Versus Stressed Collagen Matrices *, 274
(1999) 918–923.
[111] E. Ioachim, A. Charchanti, E. Briasoulis, V. Karavasilis, H. Tsanou, D.L.
Arvanitis, N.J. Agnantis, N. Pavlidis, Immunohistochemical expression of
extracellular matrix components tenascin, fibronectin, collagen type IV and
laminin in breast cancer: Their prognostic value and role in tumour invasion and
progression, Eur. J. Cancer. 38 (2002) 2362–2370. doi:10.1016/S0959-
8049(02)00210-1.
[112] M. Bi, H. Han, S. Dong, Y. Zhang, W. Xu, B. Zhu, J. Wang, Y. Zhou, J. Ding,
Collagen-Coated Poly(lactide-co-glycolide)/Hydroxyapatite Scaffold
Incorporated with DGEA Peptide for Synergistic Repair of Skull Defect,
Polymers (Basel). 10 (2018) 109. doi:10.3390/polym10020109.
Bibliography
62
[113] M. Ivarsson, a McWhirter, T.K. Borg, K. Rubin, Type I collagen synthesis in
cultured human fibroblasts: regulation by cell spreading, platelet-derived growth
factor and interactions with collagen fibers., Matrix Biol. 16 (1998) 409–425.
[114] Y. Wu, K. Jane Grande-Allen, J.L. West, Adhesive Peptide Sequences Regulate
Valve Interstitial Cell Adhesion, Phenotype and Extracellular Matrix Deposition,
Cell. Mol. Bioeng. 9 (2016) 479–495. doi:10.1007/s12195-016-0451-x.
[115] G. Balasundaram, T.J. Webster, Increased osteoblast adhesion on nanograined
Ti modified with KRSR, J. Biomed. Mater. Res. Part A. 80A (2007) 602–611.
doi:10.1002/jbm.a.30954.
[116] A.A. Sawyer, K.M. Hennessy, S.L. Bellis, The effect of adsorbed serum proteins,
RGD and proteoglycan-binding peptides on the adhesion of mesenchymal stem
cells to hydroxyapatite, Biomaterials. 28 (2007) 383–392.
doi:10.1016/j.biomaterials.2006.08.031.
[117] P. Bornstein, Diversity of function is inherent in matricellular proteins: An
appraisal of thrombospondin 1, J. Cell Biol. 130 (1995) 503–506.
doi:10.1083/jcb.130.3.503.
[118] J.M. Sipes, N.H. Guo, E. Negre, T. Vogel, H.C. Krutzsch, D.D. Roberts,
Inhibition of fibronectin binding and fibronectin-mediated cell adhesion to
collagen by a peptide from the second type I repeat of thrombospondin, J. Cell
Biol. 121 (1993) 469–477. doi:10.1083/jcb.121.2.469.
I
Project 2
Block Copolymers of 2-Oxazolines and 2-Oxazines: The Influence of Polymer
Architecture on the Rheological Properties of a Potential Thermogelling Bioink
II
Abstract Recently, there has been found a diblock copolymer of thermoresponsive poly(2-n-
propyl-2-oxazine) (PnPrOzi) and hydrophilic poly(2-methyl-2-oxazoline) (PMeOx),
which forms a physically crosslinked hydrogel above 20 wt% at temperatures above
the lower-critical-solution-temperature (LCST) of the system. It could be shown that
the properties of this hydrogel system are ideal for extrusion based bioprinting, given
through high shear thinning and recovery properties, as well as cytocompatibility.
The current thesis aimed to adjust and enhance the mechanical and printing properties
of the hydrogel system by blending different architectures of the named di block copol-
ymer. This was done by synthesizing and characterizing three different 4-arm star-
shaped block copolymers of PMeOx and PnPrOzi. Repeating units were 40/49
(PMeOx/PnPrOzi) for P3, 23/23 for P4 and 13/13 for P5, respectively.
It could be shown that blending the star-shaped copolymers P4 and P5 with 5 out of
total 20 wt% polymer content into the linear system resulted in loss of the shear thin-
ning and recovery properties. In contrast, P3 showed no influence on the shear thin-
ning and recovery properties of the material in all blends. At 2.5 wt% (total polymer
content 20 wt%) blending no significant influence of P3, P4 and P5 on the hydrogel
properties could be observed.
Applying a mathematical model, theoretical parameter combinations, in which extru-
sion would be successful, could be evaluated for the blends. It could be shown that
blending different architectures increases the amount of parameter combinations at
which the printing is successful, the so-called bioprinting window.
Relative cell viability was obtained by WST-1 assay, which indicated higher cell toxicity
with shorter arm length of the star-shaped copolymers.
With this thesis, a method was found to potentially increase the printability of a new
potential thermogelling bioink in biofabrication. Further blending experiments, with an
extended concentration range and other architectures, are needed to confirm the ad-
justability of the system.
III
Table of Contents
Abstract ...................................................................................................................... II
List of Abbreviations & Glossary ................................................................................. V
1. Introduction .......................................................................................................... 1
2. State of Knowledge ............................................................................................. 3
2.1 Tissue Engineering & Biofabrication .............................................................. 3
2.1.1 From Tissue Engineering to Biofabrication - A Summary ....................... 3
2.1.2 Bioinks in Biofabrication - An Overview .................................................. 5
2.2 Chemical Background of Poly(2-oxazoline)s and Poly(2-oxazine)s............... 7
2.2.1 Monomer Synthesis ................................................................................ 7
2.2.2 Living Cationic Ring Opening Polymerisation (LCROP) .......................... 9
2.2.3 Physicochemical Properties of Poly(2-oxazoline)s and ............................ Poly(2-oxazine)s in Aqueous Solution .................................................. 14
2.3 Poly(2-Oxazolines) and Poly(2-Oxazines): Applications .............................. 15
3. Materials & Methods .......................................................................................... 18
3.1 Materials & Instruments ............................................................................... 18
3.2 Methods ....................................................................................................... 19
3.2.1 Synthesis Protocols .............................................................................. 19
3.2.2 Rheological & Viscosity Measurements ................................................ 26
3.2.3 Differential Scanning Calorimetry Measurements ................................. 26
3.2.4 Nuclear Magnetic Resonance Spectroscopy Measurements ................ 26
3.2.5 Gel-Permeation Chromatography Measurements ................................. 27
3.2.6 WST-1 Assay ........................................................................................ 27
4. Results & Discussion ......................................................................................... 28
4.1 Motivation .................................................................................................... 28
4.2 Synthesis of a Linear 2-n-Propyl-2-Oxazine and 2-Methyl-2-Oxazoline ......... Block Copolymer ......................................................................................... 29
4.3 Synthesis of Star-Shaped Block Copolymers of 2-n-Propyl-2-Oxazine .......... and 2-Methyl-2-Oxazoline ........................................................................... 37
4.3.1 Polymer Design ..................................................................................... 38
4.3.2 Initiator Synthesis .................................................................................. 38
4.3.3 Polymerization of Star-Shaped Copolymers of 2-methyl-2-oxazoline ...... and 2-n-propyl-2-oxazine ...................................................................... 42
4.3.4 Physicochemical Properties .................................................................. 52
IV
4.3.5 Conclusion ............................................................................................ 63
4.4 Architectural Influence on Rheological Properties ....................................... 64
4.4.1 Rheological Assessment of Polymer Blends with Different ...................... Architectures ......................................................................................... 64
4.4.2 Conclusion ............................................................................................ 78
4.5 Assessment of Cytotoxicity of Star-Shaped Copolymers ............................. 79
4.5.1 WST-1 Assay ........................................................................................ 79
4.5.2 Conclusion ............................................................................................ 81
4.6 Process Upscaling .......................................................................................... 81
4.6.1 Up-scaled Synthesis of nPrOzi ............................................................. 81
5. Summary & Outlook .......................................................................................... 82
Bibliography .............................................................................................................. 85
Appendices ............................................................................................................. 104
Appendix I: Supplementary Data ........................................................................ 104 1H NMR spectra of P4 and P5 ......................................................................... 104 1H NMR spectra of I2 and I3 ............................................................................ 105
Amplitude sweep of 20 wt. % P1 compared to 20 wt. % P2 ............................ 106
Frequency sweep of 20 wt. % P1 compared to 20 wt. % P2 ........................... 106
Amplitude sweep of 30 & 40 wt. % of P3-P5 ................................................... 107
Amplitude sweep of polymer blends P1/P3, P1/P4, P1/P5 and references ..... 108
V
List of Abbreviations & Glossary 2D cell culture (2-dimensional cell culture): cell culture on a flat x,y-directed surface,
such as a tissue culture plate.
3D cell culture (3-dimensional cell culture): cell culture in x,y- and z- direction, e.g.
within a hydrogel matrix.
ACN (acetonitrile): Solvent in organic chemistry.
AM (additive manufacturing): Production methods for objects following a bottom-up
approach, e.g. 3D printing.
Amphiphilic: Substrate that has partial hydrophilic and partial hydrophobic character.
Biofabrication: Field of research that aims for the fabrication of functional tissues by
using AM methods.
Bioink: A hydrogel compound that can be used for the encapsulation of cells in the
bioprinting process.
Bioprinting: Utilization of inkjet printing, LIFT and extrusion-based printing for the for-
mation of functional tissue models with cell-laden matrices.
Boc-Piperazin (tert-butyl piperazin-1-carboxlate): Organic compound used as termina-
tion reagent in CROP.
nBuOx (2-n-butyl-2-oxazoline): 2-oxazoline with n-butyl-functionalized 2-position.
DLS (Dynamic light scattering): Analysis method of particles and aggregates in solution
based on light scattering of the sample.
DMEM (Dulbecco’s modified Eagle medium): standardized medium for cell culture,
containing different amino acids, inorganic salts and vitamins.
DMF (dimethylformamide): Eluent in GPC measurements and organic solvent.
DTT (dithiothreitol): Organic compound with alcohol and thiol units, used for the cross-
linking of polymers.
ECM (extracellular matrix): the molecules excreted by cells forming the cell environ-
ment and providing support for cells to form tissues and organs.
VI
EtOx (2-ethyl-2-oxazoline): 2-oxazoline with ethyl-functionalized 2-position.
FCS (foetal calf serum): Serum supplement for cell culture derived from the calf foetus.
GelMA (gelatine methacryloyl): Methacrylated gelatine that can be photochemically
crosslinked.
HFIP (1,1,1,3,3,3-hexafluoro-2-propanol): Eluent in the GPC instrument.
HUVECs (human umbilical vein endothelial cells): Endothelial cells that are obtained
from the umbilical vein of humans.
Hydrogel: A 3D covalently or physically crosslinked polymer network with high water
content.
IKVAV (isoleucine-lysine-valine-alanine-valine): Peptide sequence derived from the α-
1 chain of laminin.
In vitro model: A tissue replicate that was made in the laboratory under sterile condi-
tions.
iPrOx (2-isopropyl-2-oxazoline): 2-oxazoline with isopropyl-functionalized 2-position.
LCROP (cationic ring opening polymerization): Living polymerisation based on a cati-
onic propagation species.
LCST (lower-critical-solution-temperature): Temperature, above which an amphiphilic
polymer is not soluble in all concentrations in a given solvent anymore.
LIFT (laser-induced forward transfer): Bioprinting method, in which a laser heats up a
bioink reservoir to eject one droplet at a time to build up tissue constructs.
Loss modulus G’: Modulus obtained through oscillatory measurements in rheology. It
represents the viscous properties of a viscoelastic material.
LVE (linear viscoelastic) range: Rheological region, in which G’ and G’’ are constant
independently of the applied strain.
MeOx (2-methyl-2-oxazoline): 2-oxazoline with methyl-functionalized 2-position.
MMP (matrix metalloproteinase): Proteases that are responsible for ECM degradation.
VII
MSC (multipotent stromal cells): stromal cells that can differentiate into different cell
types.
MWCO (molecular weight cut-off): Size limitation for molecules in dialysis tubes. Above
the limitation the molecule cannot diffuse through the tube pores anymore.
NMR (nuclear magnetic resonance) spectroscopy: Spectroscopy method that exploits
the nuclear spin relaxation after stimulation.
NonOx (2-nonyl-2-oxazoline): 2-oxazoline with nonyl-functionalized 2-position.
PhOx (2-phenyl-2-oxazoline): 2-oxazoline with phenyl-functionalized 2-position.
nPrOzi (2-n-propyl-2-oxazine): 2-oxazine with n-propyl-functionalized 2-position.
OTf (triflate): Good leaving group in organic chemistry.
PBS (phosphate buffered saline): Saline solution, buffered by phosphate with a pH
value of constant 7.3. The osmotic pressure is like the one in the human body.
PCL (poly(ε-caprolacton)): thermoplastic polymer used in biomedical applications.
PEG (poly(ethylene glycol): non-cytotoxic polymer synthesized from ethylene oxide.
PhCN (benzonitrile): Organic solvent used in CROP.
Polymersome: Hollow spheres enclosing a solution, built by amphiphilic block copoly-
mers.
POx (poly(2-oxazoline)): Polymers from 2-oxazoline compounds.
POzi (poly(2-oxazine)): Polymers from 2-oxazine compounds.
RGD (arginine-glycine-aspartic acid): Peptide sequence, well known for its properties
in directing cell adhesion to a material.
SANS (small angle neutron scattering): Analysis method for supramolecular structures
based on neutron scattering.
SLS (selective laser sintering): AM process, in which metal powder is melted into an
object by a layer-by-layer approach using a laser.
VIII
Storage modulus G’: Modulus obtained through oscillatory measurements in rheology.
It represents the elastic properties of a viscoelastic material.
UCST (upper-critical-solution-temperature): Temperature above which a polymer is
soluble in a given solvent at any given concentration.
UV/Vis: Ultraviolet and visible part of the light spectrum.
TE (Tissue Engineering): Research field in biomedical sciences that focuses on the
engineering of functional tissue.
WST-1 (Sodium 4-[2-(4-iodophenyl)-3-(4-nitrophenyl)tetrazol-2-ium-5-yl]benzene-1,3-
disulfonate): Reagent for a cell proliferation assay.
1. Introduction
1
1. Introduction Tissue engineering has become a well-established life science through the aims of
enhancing rapidity and personalization of medical treatments. These aims have been
pursued by seeking knowledge about cell behaviour in their native environment and
the dependency on different chemical, mechanical and biological factors. Scientists
have come to realize that there is a major difference in cell behaviour dependent on
whether the cells are cultured in a matrix (3-dimensional, 3D) or on a substrate (2-
dimensional, 2D), e.g. a petri dish.[1,2] With this basic, yet revolutionizing, understand-
ing, researchers were able to produce 3D tissue models of skin and ultimately build up
an artificial skin model.[3] Today, the cosmetic industry in Europe uses only tissue
models to test their products.[4,5]
Even though it is possible to manufacture 3D cultures of cells that mimic the native
tissue, the architectures are limited to the shape of the mould the cells are cultured in.
Architecture, however, matters when it comes to personalizing medical treatment. Tis-
sue defects will differ from patient to patient. In recent years, to overcome this issue, a
research field called biofabrication emerged from TE, aiming to combine the
knowledge in TE with additive manufacturing (AM). Using the advantages of AM, hier-
archical structures with adjustable architectures can be printed by utilizing so called
bioinks, providing an artificial 3D environment for cells after printing. AM has already
shown its applicability in manufacturing several kinds of clinically relevant titanium hip
implants[6], maxillofacial reconstruction implants[7,8], or total joint replacements in the
veterinary medicine field[9].
However, until now these applications are mainly based on the simple manufacturing
of metal or polymer parts based on computer tomography images and without cells.
On the one hand, this already had an impact on personalized treatment of patients and
should be further pursued; on the other hand, it is still a big challenge to implement
cells and growth factors into these constructs to treat larger or more complex defects.
Especially the treatment of soft tissue, e.g. for heart or even cartilage reconstruction,
is far from being fully established at this time - let alone the possibility of printing a fully
functional organ. Native tissue is a complex and hierarchical construct with different
cell and extracellular matrix (ECM) protein gradients, as well as spatially changing me-
chanical properties, which need to be reproduced when a defect is approached.
1. Introduction
2
To overcome this, a process called bioprinting, in which cells are encapsulated in hy-
drogel matrices, termed as bioinks, has been proposed. The bioink is extruded, giving
the possibility to introduce gradients with different chemical and biological motifs, cells
or mechanical properties into one printed construct. This construct is then cultured in
a bioreactor to give the cells time to express their own native ECM while simultaneously
digesting the artificial matrix.
Many start-up companies have emerged, commercializing and distributing this pro-
cess, to increase the accessibility of this technology. Yet, the library of commercially
available bioinks is still limited and needs to be extended, to accelerate research in the
field of biofabrication. Therefore, there is a strong need to develop a bioink, which can
be modified in a fast and efficient way to meet mechanical and biochemical require-
ments tailored to the target tissue.
Recently, a di-block copolymer of hydrophilic poly(2-methyl-2-oxazoline) (PMeOx) and
thermoresponsive poly(2-n-propyl-2-oxazine) (PnPrOzi), developed in the group of
Prof. Robert Luxenhofer, has shown to form a physically crosslinked hydrogel at room
temperature with excellent shear thinning properties and cytocompatibility.[10] These
properties make it a promising candidate for the use in bioprinting. Nonetheless, shape
fidelity of the printed constructs is weak, with fusion being observed between layers.
To overcome this problem, this thesis provides an approach to modify the mechanical
properties of the hydrogel by introducing a different polymer architecture of the same
polymer into the system.
Showing the change in properties by blending two polymers of the same type with
different architectures, without additional synthesis steps, is an alternative approach to
addressing mechanical properties of current viable candidates.
This allows accessibility for researchers from interdisciplinary fields to tailor the me-chanics of their own bioink through mixing.
2. State of Knowledge
3
2. State of Knowledge
2.1 Tissue Engineering & Biofabrication
2.1.1 From Tissue Engineering to Biofabrication - A Summary
The term tissue engineering was coined in the late 1980s and gave rise to a field of
research, which almost 30 years later has become one of the biggest research fields
in regenerative medicine.[11] The focus of this highly relevant area of research is the
construction of functional tissues, with the long-term goal of replacing faulty native tis-
sue parts through the harvest, expansion and maturation of the patient’s own cells into
functional tissue replacements.[12] While this aim of personalized patient treatment is
still a long-term goal, tissue engineering has made tremendous progress in the field of
drug screening and the replacement of animal testing. Today, animal testing in the
European cosmetic industry is widely forbidden, which forced the involved companies
to invest in tissue models, increasing and advancing the availability of current models,
mainly skin models[4,5], replacing the use of animal models.
The need of reliable tissue models in the drug screening process is of urgency, with
95% of all anti-cancer drug candidates that manage to get into clinical trials fail in phase
1 - 3.[13] This tremendous number is likely due to pre-clinical models that are unable
to represent the highly complex tumour environment. Traditional cancer models are
often Xenografts with murine origin or 2D cultures, which are inaccurate representa-
tions of human physiology, when it comes, for instance, to drug resistance.[14] There-
fore, TE tackled this problem by the construction of spheroid based[15] and scaffold
based[16] approaches to develop a 3D environment that could be used for drug
screening. Subsequently, significant differences in tissue response to different drug
candidates were found.[15] However, despite the fabrication of more representative
models, issues concerning media perfusion and reproducibility emerged.[17]
These concerns eventually led to the establishment of the field of biofabrication, which
focuses on the construction of hierarchical tissue models by the convergence of AM
and TE.[18] To elaborate, cells are printed directly via 3D printing technologies through
either the process of bioprinting, where the cells are implemented into a bioink[19,20],
or bioassembly, where cells are encapsulated in spheroids to form larger con-
structs[21,22], followed by a maturation process into a functional lab-grown tissue.[23]
2. State of Knowledge
4
Given biofabrication is relatively new, and definitions are under constant revision in this
growing field of science[23,24], AM itself has recently become part of accepted medical
procedures. Implants with inbuilt porosities manufactured via 3D metal printing tech-
niques like selective laser sintering (SLS) are applied in surgeries without the use of
binders or solvents.[6,25–27] In addition, constructs manufactured by AM also assist
surgeons in pre-surgical procedure planning[28] or provide educational aids for medi-
cal staff or patients[29]. The establishment of AM is not only driven by enhanced fea-
tures and ease of fabrication but also by rapid translation with the rising number of
organ transplantations needed worldwide. For example, approximately 850,000 new
patients are diagnosed with liver cancer annually, of which only 30 - 40 % can receive
treatment by liver transplant or resection, with survival rates between 50 and 75 % after
5 years post-treatment.[30] Additionally, organ rejection must be taken into account,
which requires a daily routine for immunosuppressive drugs.[31] The ability to engineer
tissues provides promise for the regeneration, rather than replacement, of organs.[32–
34] It could be shown that different cell types, like stem cells, NIH 3T3 fibroblasts, hu-
man umbilical vein endothelial cells (HUVECs) or chondrocytes, in different matrices
were bioprinted and able to form structures resembling tissues, such as blood ves-
sels.[35]
Even though there has been made major progress in the field of bioprinting, there are
still major challenges to address before the rapid manufacturing of functional tissues
can be successfully translated into medical procedures. These include but are not lim-
ited to (1) the low number of commercial bioinks that is available on the market. (2)
Assessment of cytotoxicity and fate of used materials in the human body, which is still
not fully understood yet.[36] (3) The reproduction of mechanical properties of the native
cell environment through the bioink in vitro, which also includes to be able to introduce
gradients of mechanical properties and biological moieties into a system.[37] And (4)
bioreactors have to be designed that can be commercialized and made accessible for
on-site maturation of the functional tissue after the printing process.[38]
With all this in mind, bioprinting has demonstrated its value in research. However, to
advance this field from benchtop to bedside, suitable materials and protocols for print-
ing must first be established to deem this process clinically viable.
2. State of Knowledge
5
2.1.2 Bioinks in Biofabrication - An Overview
Bioinks are cell-laden polymer networks with high water content, namely hydrogels[39],
that are used in the extrusion-based bioprinting process in biofabrication.[40] This def-
inition by Malda et al. conveys the key role of bioinks as a matrix for cells during printing
and for the maturation process. Therefore, bioinks must be versatile in their properties,
in order to adapt to certain cell types, printing and maturation conditions.[41] It is im-
portant to note that in literature the term bioprinting is categorised into extrusion-based
bioprinting, inkjet printing[42,43] and laser-induced forward transfer (LIFT)[44,45].
Each process requires different bioink requirements, depending on the printing tech-
nique[46]. Based on the applicability of extrusion-based systems to this work, only bi-
oinks compatible with these systems will be further discussed.
Extrusion-based bioprinting allows for the production of cell structures with a high cell
density at desired positions within a bioprinted construct.[46] Bioink candidates for this
system require excellent shear thinning properties during the extrusion process and
rapid recovery directly after leaving the nozzle of the printer head. This ensures shape
fidelity of the printed construct, enabling layer-stacking into larger constructs.[41] Fur-
thermore, they must be non-cytotoxic and protect the cells in the printing process. Cell
survival is directly related to the shear stress the cells experience in the printer needle,
and therefore also to the flow profile of the bioink through the needle.[47] The bigger
the velocity gradient across the needle diameter, the more shear stress cells will ex-
perience in the printing process thereby increasing the risk for cellular damage.[48]
Depending on the application, the ink must be crosslinked via binding moieties, like
matrix metalloproteinase (MMP) cleavable peptides[49], that the cells can cleave in the
maturation process, while expressing their own native ECM. To allow for the expres-
sion of ECM, degradation time of the bioink[50] and introduction of biochemical moie-
ties should be considered. Latter could be achieved by implementing peptide se-
quences or motifs into the system by e.g. established click-chemistry.[49,51,52] As the
biofabrication and bioprinting fields evolved, many bioink candidates covering these
properties have been found and are now essential part of research. They can be cat-
egorised into two groups: natural and synthetic polymer systems.
Most naturally derived polymer systems contain biochemical motifs which provide cues
for cellular differentiation within the matrix. Natural systems are currently the most
abundantly used, including gelatine methacryloyl (GelMA)[53,54], agarose[55],
2. State of Knowledge
6
alginate[56,57], collagen[58], hyaluronic acid[59], chitosan[60] or silk[61]. However,
huge batch-to-batch variability occurs in these systems due to purification difficul-
ties.[62] The variability covers molecular mass and chemical motifs and, therefore,
changes the environment that is presented to the cells with each batch. In addition, the
printability of these inks becomes more variable, as shear thinning properties and re-
covery behaviour of the materials change. This variability limits their potential for up-
scaling and rapid manufacturing to cellular constructs, given printing parameters must
be redefined with each new batch.
In contrast, synthetic approaches provide the opportunity to manufacture more repro-
ducible systems.[63] Despite lacking the cytocompatibility of natural polymers, click-
chemistry has been proven to be a valid way of modifying the “naked” synthetic poly-
mers with biochemical moieties on demand, allowing a specific environment for cells
after printing.[64] However, until recently, only a handful of synthetic systems have
been developed to be utilised as a bioink. These include e.g. a member of the Plu-
ronic® family, Pluronic® F127[65], crosslinked peptide sequences[66] or a recently
found diblock copolymer of PnPrOzi and PMeOx[10]. However, an accommodation of
advantages and disadvantages arises with each of aforementioned examples. F127
for instance demonstrates excellent shear thinning and recovery properties at concen-
trations over 20 wt.%[65] however, forms hydrogels at 25 % (w/v), and expressed tox-
icity in multipotent stromal cells (MSCs) after three days in culture[67]. The diblock
copolymer of PnPrOzi and PMeOx on the other hand exhibits superb recovery, shear
thinning and cytocompatibility towards NIH 3T3 fibroblasts[10], while the shape fidelity
of the printed construct is yet to be optimised. However, of course cytotoxicity in a 3T3
cell line and in MSCs cannot be compared in this case, given their origin.
Recently, there has been a push in the development of natural/natural and syn-
thetic/natural copolymers, to combine properties of different materials. Two early ex-
amples, not yet related to bioprinting, were the following two copolymers based on
F127. Through the covalent coupling of the synthetic polymer to chitosan, the mechan-
ical properties and the formerly poor cytocompatibility, were enhanced.[68] Even better
control over the mechanical properties of the hydrogel could be obtained when F127
was covalently attached to fibrinogen.[69] The combination of physical, temperature-
dependent crosslinking of F127 and covalent photo crosslinking of F127 and fibrinogen
made the adjustment of the hydrogel strength a possibility, depending on the
2. State of Knowledge
7
crosslinking temperature. A more recent example revealed that the covalent crosslink-
ing of poly(glycidol)s with small amounts of hyaluronic acid had a major effect on chon-
drogenic differentiation.[70] After some approaches in combining natural and synthetic
polymers, recent studies mainly observe composite materials involving PEG[71–73],
whilst a lack in literature exists for other synthetic systems. One reason for this might
be found in the long approval periods for new materials, but also in the lack of synthetic
materials available. Rather, a trend in literature is found for natural polymer combina-
tions to modify bioink properties. Alginate hydrogels, for example, were blended with
either agarose or collagen and printed using extrusion-based bioprinting.[74] The re-
sulting scaffolds showed enhanced tensile strength and compression modulus, com-
bined with a decrease in swelling ratio. Furthermore, cell viability of articular chondro-
cytes in the alginate/collagen hydrogels was increased compared to normal alginate
culture. Another example is a blend of gelatine and oxidized dextran, which allowed
pH dependent physical crosslinking of the system via pH change in the range between
6.0 and 8.0.[75] High cell viability of above 95% was observed after first printing trials
with human dermal fibroblasts.
Even though the blending of polymers might be a solution to develop a library of ad-
justable bioinks, the use of only natural polymers might cause problems for up scaling
given the batch-to-batch variability. Therefore, as argued by Costa and colleagues[76]
on the field of bioinks, it is as important to find synthetic systems that can be either
combined with natural polymers or provide a competitive alternative. As mentioned
before, synthetic hydrogels like F127 can be utilized to form constructs with high shape
fidelity. If such properties can be combined with the high biochemical compatibility of
the natural polymers, then the development of bioinks with a broad applicability in bio-
fabrication can be accelerated.
2.2 Chemical Background of Poly(2-oxazoline)s and Poly(2-oxazine)s
2.2.1 Monomer Synthesis
The synthesis of Ox and Ozi in general is a similar procedure, since both compounds
only differ in one ring-membered CH2 group (Figure 1)
2. State of Knowledge
8
Figure 1: General chemical structures of 2-oxazolines (Ox) and 2-oxazines (Ozi).
Most of the procedures found in literature are described for different Ox systems. Given
the chemical similarity of both compounds, synthesis routes can be translated to Ozi,
as well. Therefore, for a better overview, only the synthesis for Ox compounds will be
shown. One common synthesis route addressing the synthesis of cyclic imino ethers
with a variety of R groups and through simple techniques, was established in the 1970s
by Witte and Seeliger.[77] The synthesis procedure involved the reaction of a nitrile
compound with an amino alcohol (Figure 2).
Figure 2: General synthesis protocol for cyclic imino ethers, as described by Witte and Seeliger (1972).
Both compounds were stirred at increased temperatures in the presence of a cadmium
or zinc catalyst. With this protocol a library of Ox and Ozi could be synthesized in only
25 h and with yields >62%.
Based on this approach, many different synthesis routes have been developed. Vor-
brüggen and Krolikiewicz[78] could show that the nitrile could be replaced by a carbox-
ylic acid. Without a loss in yield through this kind of reaction, highly functionalized Ox
could be synthesized. The reaction temperature could be dropped to room temperature
or even 2 °C, depending on the Ox compound. While the study utilized a phosphor-
based catalyst, another study from Hansel et al.[79] showed that palladium-based cat-
alysts can be used for a side-reaction-free synthesis, producing the same yields.
As it can be seen in these examples, the catalyst is essential for the reaction of an
amino alcohol with several other reagents to form the final Ox compound. Even rare
earth metals like lanthanum were utilized together with carboxylic esters.[80] These
compounds are well-established in literature. Ox were synthesized previously from car-
boxylic esters via a four-step synthesis.[81] Therefore, by choosing the right catalyst,
2. State of Knowledge
9
the synthesis steps could be reduced, while yields of up to 82 % could be achieved.
The process was later improved by Ohshima et al.[82] by using a cluster complex of
zinc as a catalyst. Interestingly, the conversion of different lactones to Ox was achieved
simultaneously. Depending on the lactone, the chain length of the side chain differed.
On contrary, there are synthesis procedures which deem the catalyst redundant.
Pirrung and Tumey[83] showed the intramolecular cyclisation of hydroxyamides in a
two-step synthesis. In the first step the hydroxy group of the reagent is converted into
a good leaving group. In a second step through simultaneous cyclization of the reagent
and cleavage of the leaving group, an Ox compound is built. The reaction uses the
high stability of the five-membered ring that is built in the reaction, as well as the gain
in entropy, when the leaving group is cleaved. Additionally, only the intended com-
pound will undergo cyclization, increasing the overall reaction efficiency as remaining
compounds are trapped on the resin. Through this method, it was also possible to
synthesize the six-membered ring of oxazines. Depending on the compound, yields
between 32 and 75 % with very high purities could be achieved. Furthermore, reaction
times could be set between 1 and 24 h for Ox and Ozi compounds, respectively.
All in all, it could be shown that Ox compounds have versatile synthesis routes with
and without initiator systems. The named procedures were either already or can be
translated into Ozi systems due to the high chemical similarities of the compounds and
the reagents. Given the well-established synthesis protocols, there is a large available
library of Ox and Ozi compounds, which opens the gateway to different polymeriza-
tions, resulting in versatile polymers.
2.2.2 Living Cationic Ring Opening Polymerisation (LCROP)
As is was pointed out before, the advantage of synthetic polymers in comparison to
natural systems is a higher reproducibility. However, to achieve this it is important to
have controlled polymerisation mechanisms that allow a precise tailoring of the poly-
mer itself. Living polymerisations are perfectly suited to accomplish that.
Living polymerisations are chain growth polymerisations, that do not show termination
reactions, even after all monomer is used up. The chain ends stay active until more
monomer is added or the reaction is terminated via another reagent. Hence, this type
of polymerisation allows access to well-defined block copolymers. Furthermore,
2. State of Knowledge
10
polymers, which are synthesized by a living polymerization, exhibit narrow molecular
mass distributions, given through a fast initiation of the reaction and relative to it slow
propagation.
There are two categories of living polymerizations: anionic and cationic polymerisation.
Besides these two, other mechanisms[84], based on controlled radical polymerisation
have been established. Those, however, per se cannot be called “living”, since the
active chain ends are in an equilibrium with a dormant species.
For the synthesis of poly(2-oxazoline)s (POx) and poly(2-oxazine)s (POzi), cationic
living polymerisation in the form of living cationic ring opening polymerization (LCROP,
Figure 3) is used.[85–87]
Figure 3: LCROP polymerisation on the example of 2-oxazoline and 2-oxazine monomers. The tech-nique can be used for synthesis of homo- and block copolymers.
In relevance for this thesis, the mechanism of the reaction is illustrated for CROP of
Ox and Ozi. To initiate the reaction, a good electrophile is needed, which can attach to
the electron-rich, ring-bound nitrogen atom in the first step and create electron defi-
ciency that can be delocalised over the neighbouring double bond between oxygen
and nitrogen. This delocalisation of the positive charge, as well as the counter ion of
the initiator, stabilizes the system and slows down the propagation. Most commonly
used initiator systems for Ox and Ozi are triflate[88,89] or tosylate[90] derivates. In the
propagation step, the nitrogen of the next monomer unit binds to the active cationic
chain end, inducing a positive charge into the new monomer system. After complete
monomer consumption, the active chain end in a living polymerization stays active and
can now be either terminated using a nucleophilic reagent, or a second monomer can
2. State of Knowledge
11
be added. This makes LCROP an ideal method for the synthesis of defined block co-
polymers.[91]
Despite it being established for many years with a well understood mechanism, publi-
cations on POx or POz regularly reveal issues of side reactions when polymerizing Ox
via LCROP. It is assumed that chain transfer and coupling reactions are the main rea-
sons for bimodalities and low-molecular tailing, observed in GPC elugrams.[92]
Figure 4: Reaction scheme of chain transfer and chain coupling reactions in LCROP, according to Litt
et al.[92]
The common theory (Figure 4) is that an α-proton of an active chain end is abstracted,
causing the initiation of another chain, with consecutively lower molecular mass. The
remaining olefinic chain can then potentially participate in chain coupling reactions.
Also, those side reactions are often accompanied by a colour change of the reaction
mixture.[93] The origin of this yellow colour is not clear, but seems to play a role in the
side reactions that occur.
To approach this reoccurring issue, research groups started evaluating the different
parameters involved in LCROP. Park et al.[94] tried to solve the problem by reducing
the polymerization temperature of 2-isopropyl-2-oxazoline (iPrOx). The polymer, when
synthesized at temperature of 85 °C, showed a high molecular signal, as well as a low
molecular tailing in the GPC elugrams. They deducted that the strong high molecular
signal must be due to coupling mechanisms and the tailing due to chain transfer reac-
tions, resulting in polymer chains of low molecular mass. After reducing the tempera-
ture to 42 °C they observed the absence of both signals caused by the side reactions.
2. State of Knowledge
12
However, because of the drastic change in temperature the reaction time for the
polymerization of 88 monomer units took 476.5 h (~20 d), which is not suitable for
industrial and academic time frames. Interestingly, Wiesbrock et al.[95] could find the
perfect temperature for a set of POx polymerisations to be 140 °C in acetonitrile, when
using a microwave for the polymerisation. This value, however, was chosen based on
the compliance of theoretical and measured kinetics of the polymerizations. In another
study of Hoogenboom et al.[96] an optimal temperature for CROP of 2-phenyl-2-oxa-
zoline (PhOx) of 130 °C could be observed. At this temperature the GPC elugram
shows the lowest dispersity (Đ = 1.31) however a significant low molecular tailing was
still observed. These findings stand in contrast to the temperatures found for optimal
polymerizations by Park et al. It is also noteworthy, that there has been a study[97] on
pressurized LCROPs of 2-ethyl-2-oxazoline (EtOx). One would expect higher side re-
action numbers when molecules are in closer proximity to each other under pressure
at temperatures of 120 or 140 °C. In contrast, Đ of 1.20 could be achieved for the
polymers and the GPC elugrams showed no significant side reaction-based signals.
Wiesbrock et al.[95] also approached the side reaction issues by adjusting the mono-
mer concentration in the reaction mixture. They observed the microwave-induced
LCROP mechanism of a set of 2-oxazolines at different concentrations and solvents
affected polydispersity. In all cases, they could see the development of high molecular
shoulders or signals and low molecular mass tailings in their GPC elugrams. However,
in all POx polymers with aliphatic side chains, both phenomena could be reduced,
when reducing the concentration from bulk conditions to around 60 % of the bulk con-
centration. In poly(2-phenyl-2-oxazoline) (PPhOx), a high molecular signal/shoulder in
the GPC elugram was not visible anymore at half of the bulk concentration. This shows
that the concentration of monomers can also help to control the discussed side reac-
tions.
Another recent approach has been presented by Hoogenboom et al. in the form of a
patent.[98] They propose the synthesis of high molecular mass POx compounds, with
a narrow molecular mass distribution by using in situ distillation. A small-scale polymer-
ization initiates in a separate flask with a low amount of monomer units. After the mon-
omer was consumed, the initiated oligomer is distilled from this flask into another flask
with the remaining, unreacted monomer. Therefore, impurities and side-products are
supposed to be left behind in the distillation swamp. However, there might be technical
2. State of Knowledge
13
and financial problems with up-scaling and transferring this system to other laborato-
ries, since living polymerisations are susceptible to air and water. Thus a fume hood
would be required to have a polymerization flask equipped and attached to distillation
equipment, which could be difficult for large batches of polymer.
In the end, even though these evaluations were conducted for POx systems, the oc-
curring side reactions and the reaction mechanism are the same for POzi systems.
Even though the synthesis of homopolymers can be controlled quite well, block copol-
ymers still show, difficulties to control side reactions without the use of expensive
equipment or a generous amount of time remain. Therefore, prospects require more
experimental investigations to gain a better insight on the mechanisms involved and
how to control them.
Besides the commonly synthesized linear system, other architectures have also been
established. Hoogenboom et al.[90] observed that the initiator choice is important when
synthesizing star-shaped systems. It was shown that using an initiator based on
(di)pentaerythritol for LCROP, only slow initiation rates were obtained, resulting in ill-
defined homopolymers of EtOx. They addressed the problem by replacing the initiator
by a porphyrin-based system. Other initiators used in literature include t-butyl-
calix[4]arene[99] or multi-arm triflates[100]. Together with other architectures, like
brushes[101] or bow-ties[102], POx provide a toolbox for versatile polymer architec-
tures. This makes POx systems more adjustable, given the possibilities of different
architectures.[103]
2. State of Knowledge
14
2.2.3 Physicochemical Properties of Poly(2-oxazoline)s and Poly(2-oxazine)s in Aqueous Solution
It is long known that polymers in solution undergo different behaviours compared to
single molecules. Polymer solutions can exhibit lower-critical-solution-temperature
(LCST) - a temperature above which the polymer is no longer soluble in a given sol-
vent.[104] Polymers start precipitating from solution. POx and POzi polymer systems,
polymerized via LCROP, have the advantage of tailoring this temperature, depending
on the hydrophobicity of the side-chain, polymer concentration, architecture, end
groups and molecular mass of the compounds.[105–108] Therefore, a plethora of pa-
rameters is available to customize a material. Polymers with thermoresponsive behav-
iour in solution are interesting for many fields of applications, like bioengineering[109]
or the development of drug delivery systems[110] and chromatographic separation
processes[111].
Besides this, POx/POzi in solution can form a variety of different structures, which can
be exploited for different applications. A recent example is the thermogelling hydrogel
from a block copolymer of PnPrOzi and PMeOx. The hydrophobic interactions of the
PnPrOzi block form a sponge like structure.[10] In another study of Hoogenboom et
al.[112], they were able to control the micelle formation, as well as the micelle structure
by modification of the solvent that was used. This example shows that the formation
process of a micelle can be controlled. Another study also exploits this behaviour by
forcing the micelle into a rigid-rod structure[113] and even polymersome architectures
could be established using tri-block copolymers of MeOx and PhOx.[114] Luxenhofer
et al.[115] could show that doubly amphiphilic di- and triblock POx copolymers exhibit
higher polarity than water, based on evaluation by pyrene assay, as well a drug loading
capacity of paclitaxel of 45 wt%. Schulz et al.[116] further investigates triblock copoly-
mers with a hydrophobic centre block of 2-nonyl-2-oxazoline (NonOx) or 2-n-butyl-2-
oxazoline (nBuOx) and could show that micelles loaded with paclitaxel changed mor-
phology from filomicelles to spherical ones.
In the end, this is just a first glance for the solution properties of POx derivates. Solution
behaviour can be adjusted using different compounds and environmental influences,
making these vehicles adaptable to different problems in applications, e.g. drug deliv-
ery.[117]
2. State of Knowledge
15
2.3 Poly(2-Oxazolines) and Poly(2-Oxazines): Applications
Due to the high adaptability of POx and POzi compounds, they can be used in versatile
ways, as described in recently published review articles by Lorson et al. and Dargaville
et al.[118,119].
The most relevant field for the application of POx and POzi polymers is in biomedical
sciences as drug delivery systems. Drugs like e.g. paclitaxel show in general low water
solubility, and therefore low bioavailability.[120] To address this, the drug can be
loaded into micelles, acting as a hydrophilic shell for the drug. Especially drug loading
capacities of micelles, formed by amphiphilic block copolymers of POx, are interesting.
Several anti-cancer drugs could be already solubilized in block copolymers of either
POx/POx or POx/POzi. As mentioned before, Luxenhofer et al.[115] could show load-
ing of micelles of a triblock copolymer (PMeOx-poly(nBuOx)-PMeOx), with up to
45 wt.% with paclitaxel. Another example is the study by Lübtow et al.[117], using mi-
celles of a similar triblock copolymer with PnPrOzi as core block and again PMeOx as
outer blocks. Curcumin could be successfully loaded with over 50 wt.% into the mi-
celles.
Another area within the biomedical field is the application of surface coating agents. It
was shown that PMeOx can be used to coat poly(L-lysine).[121] Compared to PEG-
modified surfaces, the stability of the coatings was increased, while the antibacterial
properties of the PMeOx coating could be proven to be similar to the ones of PEG. But
coatings of POx were not only used for the introduction of anti-fouling behaviour of a
surface. It was shown that a glass or silicon surface could be covalently modified with
poly(2-ethyl-2-oxazoline) (PEtOx) to induce controlled cell migration of HUVECs onto
the surface.[122] It was proposed that these findings, could lead to potential applica-
tions in the design of artificial heart valves.
Organic chemistry has found that the synthesis of complex organic molecules can be
achieved by enzymatic-directed synthesis methods.[123] However, a major drawback
in these syntheses is the solubility of enzymes and lysozymes in organic solvents. POx
found application in this field, because it was found that the solubility, enzymes and
lysozymes, covalently attached to PMeOx or PEtOx, resulted in better solubilities of
these compounds in organic solvents, on the cost of partial loss of enzyme activ-
ity.[124]
2. State of Knowledge
16
The applicability of POx systems was also shown by the development of several hy-
drogels from POx compounds, which could have positive impact on 3D cell culture
systems. The first known hydrogels of POx were pioneered by Chujo et al., who used
techniques like photo crosslinking[125], disulphide crosslinking[126] or metal mediated
crosslinking with bipyridyl compounds[127,128] to form POx-based hydrogels.
While the application range of these polymers is not extensively covered[119], there
have been approaches to form hydrogels that can be used in tissue engineering and
cell culture. In that regard, van der Heide et al.[129] were able to manufacture hydro-
gels by covalent crosslinking of block copolymers of PMeOx and poly(2-decyl-2-oxa-
zoline) (PDecOx) via dithiothreitol (DTT). The hydrogels were soaked in water, freeze
dried, and rehydrated, resulting in porous hydrogels. These samples could then be
used for the 3D cell culture of primary human stroma cells, when modified with an
arginine-glycine-aspartic acid (RGD) sequence. These hydrogels were supposed to be
utilized as biological inert scaffolds for the building of in vitro human epithelial barrier
structures.
That POx hydrogels can be used in combinations with cells could also be shown by
Farrugia et al.[130]. It was observed that “naked” POx hydrogels of block copolymers
of PMeOx and poly(2-(dec-9-enyl)-2-oxazoline) result in cell repulsion. As soon as an
RGD sequence was covalently attached to the system, it could be seen that fibroblasts
started spreading inside the hydrogel. Furthermore, using the same polymer blocks,
Dargaville et al.[131] did show that the hydrogels could be photocrosslinked with an
ester-group containing crosslinker, enabling the degradation of the hydrogel.
Using POx hydrogels for the 3D culture of cells has been further pursued in a recent
publication by Lorson et al.[10] Instead of chemical crosslinking POx, like it was the
case in the named previous publication, a physically, thermogelling hydrogel of PMeOx
and PnPrOzi could be found. It showed excellent shear thinning and recovery proper-
ties, high mechanical strength (~3 kPa), as well as superb cytocompatibility on NIH
3T3 fibroblasts. This first case physically hydrogel based on POx/POzi, was therefore
proposed as an ideal candidate for the extrusion-based bioprinting process in biofab-
rication.
POx hydrogels were also used in combination with fibres manufactured by melt elec-
trowritten poly(ε-caprolacton) (PCL)[132]. A block copolymer of PEtOx and poly(2-(3-
2. State of Knowledge
17
butenyl)-2-oxazoline) was chemically crosslinked via DTT to form a hydrogel, in which
a PCL scaffold was embedded. After dissolving the scaffold, hierarchically patterned
pores were left. It was proposed that this will open new ways for cell scaffolds in tissue
engineering.
In conclusion, the here named applications of POx and POzi compounds are mainly
based in the field of biomedical science. It could be shown that they are well suited
drug-delivery systems, as well as usable in 3D cell culture and biofabrication. However,
even though these areas are the major focus of these compounds, it is noteworthy that
they have been also used in other fields, like in production of microparticles.[133]
With all these application fields in mind, as well as, with respect to the growing interest
in these compounds[118], it can be assumed that the range of applications will further
increase in the next years.
3. Materials & Methods
18
3. Materials & Methods
3.1 Materials & Instruments
All chemicals, unless noted otherwise, were used as supplied by the manufacturer.
Butyronitrile, 3-amino-1-propanol, methyl triflate, pentaerythritol, chloroform, benzo-
nitrile, zinc acetate dihydrate, trifluoromethansulfonic acid anhydride, ethyl isonipeco-
tate, chloroform and Pluronic® F-127 were purchased from Sigma-Aldrich (Steinheim,
Germany). Tert-butyl piperazin-1-carboxylate (Boc-piperazine) was purchased from
TCI Chemicals (Zwijndrecht, Belgium). Potassium carbonate was purchased by Fisher
Scientific (Loughborough, United Kingdom). Pyridin and 2-methyl-2-oxazoline were
purchased from Acros Organics (Geel, Belgium). Acetonitrile was purchased from Carl
Roth GmbH & Co. KG (Karlsruhe, Germany). 1,1,1,3,3,3-Hexafluoro-2-propanol
(HFIP) was purchased by chemPUR Feinchemikalien und Forschungsbedarf GmbH
(Karlsruhe, Germany).
Deuterated methanol and dimethyl sulfoxide were purchased from Deutero GmbH
(Kastellaun, Germany). Deuterated Chloroform was purchased by Eurisotop
(Saint-Aubin, France) and deuterated acetonitrile was purchased from Sigma-Aldrich
(Steinheim, Germany).
Dublecco’s modified eagle medium (DMEM, high glucose) was purchased from Sigma-
Aldrich (Steinheim, Germany). Cell proliferation reagent WST-1 was purchased by
Roche Diagnostics GmbH (Mannheim, Germany).
Nuclear magnetic resonance measurements were conducted on a Bruker Fourier 300
(Bruker, USA). Gel permeation chromatography was conducted on 1260 Infinity Sys-
tem with HFIP as eluent (Agilent Technologies, USA). Synthesized polymers were di-
alyzed in Spectra/Por® 7 dialysis membranes (Spectrum Laboratories Inc.) and freeze-
dried using an ALPHA 1-2 LD plus system (Martin Christ Gefriertrocknungsanlagen
GmbH, Germany). Ultrapure water (conductibility: 56 mS/cm) was obtained from a
Barnstead™ MicroPure™ water purification system (Thermo Scientific, USA). Differ-
ential Scanning Calorimetry measurements were conducted on a DSC 204 F1 Phoenix
instrument (Netzsch GmbH & Co. KG, Germany). For rheology experiments an Anton
Paar Physica MCR 301 (Anton Paar, Austria) with conical plate-plate setup was used.
Viscosity measurements were done on a Lovis 2000M falling sphere viscosimeter
3. Materials & Methods
19
(Anton Paar, Austria). Corresponding densities of polymer solutions were determined
on a DMA 4100M system (Anton Paar, Austria). UV/Vis spectroscopy was conducted
on a Spectramax 250 microplate reader (Sunnyvale, USA).
For NMR spectra analysis MestreNova (Mestrelab, Spain) and for graph plotting Origin
2016 (OriginLab, USA) was used. Chemical structures were illustrated, and molecular
mass calculations of polymers conducted with ChemDraw (PerkinElmer, USA). Fig-
ures were assembled with Adobe Photoshop and Adobe Illustrator (Adobe, USA).
3.2 Methods
3.2.1 Synthesis Protocols
All molecules or polymers were synthesized under inert argon atmosphere. Reagents
were stored under argon atmosphere, polymers under non-inert atmosphere.
Synthesis of 2-n-propyl-2-oxazine (M1)
The synthesis of 2-n-propyl-2-oxazine was adapted from the procedure of Sinnwell and
Ritter[134]. In a 5 L reactor under non-inert conditions butyronitrile (1.92 kg, 2.41 L,
27.73 mol, 1.2 eq.) was heated to 50 °C. After flushing the reactor with argon atmos-
phere, zinc acetate dihydrate (126.86 g, 0.58 mol, 0.025 eq.) and 3-amino-1-propanol
(1.74 kg, 1.75 L, 23.11 mol, 1.0 eq.) were added and the reaction mixture was stirred
at 130 °C for 67.5 h. To use up residue alcohol, observed with 1H NMR spectroscopy,
more butyronitrile (200 mL, 2.30 mol, 0.1 eq.) was added to the solution. After stirring
for another 28.5 h the temperature was cooled down to 25 °C and stirring was contin-
ued for 72 h. The crude product was isolated by direct distillation from the reactor. Final
purification of the compound was conducted through fine distillation over CaH2. 2-n-
propyl-2-oxazine (1.33 kg, 1.40 L, 10.46 mol, 45.2 %) could be isolated as a colourless
liquid.
3. Materials & Methods
20
1H NMR (300 MHz, Chloroform-d) δ 4.19 – 4.05 (m, 2H), 3.34 (t, J = 5.9 Hz, 2H), 2.08
(t, J = 7.6 Hz, 2H), 1.83 (p, J = 5.7 Hz, 2H), 1.57 (dq, J = 14.7, 7.4 Hz, 2H), 0.92 (t, J =
7.4 Hz, 3H).
Synthesis of poly(2-n-propyl-2-oxazine)-b-poly(2-methyl-2-oxazoline)-ethyl isonipeco-
tate [P[nPrOzi]83-b-P[MeOx]68] (P1)
The synthesis was adapted from a recently described procedure.[10] 2-n-propyl-oxa-
zine (7.44 g, 7.83 mL, 58.50 mmol, 80.1 eq.), benzonitrile (16.62 mL) and methyl tri-
flate (0.12 g, 0.08 mL, 0.73 mmol, 1.0 eq.) were added to a flame-dried flask and stirred
for 8 h at 120 °C. After the 1H NMR spectra showed complete monomer consumption,
the solution was cooled down to room temperature and 2-methyl-2-oxazoline (4.96 g,
4.94 mL, 58.33 mmol, 79.9 eq.) was added. The reaction was continued at 100 °C for
15 h. After no monomer signal was observed in the 1H NMR spectra, the reaction was
cooled down to 0 °C and ethyl isonipecotate (0.37 g, 0.36 mL, 2.35 mmol, 3.2 eq.) was
added. Following stirring for 21 h at 40 °C the reaction mixture was cooled to room
temperature. Potassium carbonate (0.37 g, 2.68 mmol, 3.7 eq.) was added, and the
solution was stirred for another 5 h. The solvent was removed under reduced pressure
and the pre-dried polymer was left in a vacuum oven (0 mbar, 50 °C) for 15 h. Finally,
the crude product was dialyzed against ultrapure water for 74 h (0-18 h: MWCO: 1
kDa; 18-74 h: MWCO: 10 kDa) and lyophilized. The final product (8.28 g, 0.48 mmol,
66.1 %) was isolated as a white solid.
3. Materials & Methods
21
1H NMR (300 MHz, Methanol-d4) δ 4.21 – 4.13 (m, 1H), 3.74 – 3.50 (m, 271H), 3.50 –
3.36 (m, 312H), 3.10 (d, J = 8.0 Hz, 2H), 2.96 (d, J = 7.3 Hz, 1H), 2.39 (h, J = 6.1, 5.3
Hz, 172H), 2.24 – 2.07 (m, 207H), 2.01 – 1.77 (m, 166H), 1.68 (ddp, J = 11.1, 7.3, 3.8
Hz, 174H), 1.01 (tt, J = 7.5, 2.6 Hz, 251H).
Mw: 5.3 kg/mol
Mn: 4.2 kg/mol
Đ: 1.29
Synthesis of Pentaerythritoltetrakistriflate (I1)
The reaction and purification were adapted from a previously described proce-
dure.[135] In a flame-dried flask pentaerythritol (0.42 g, 3.01 mmol, 1.0 eq.) was sus-
pended in a mixture of pyridin (2.46 g, 2.50 mL, 31.04 mmol, 10.3 eq.) and acetonitrile
(ACN, 3.15 g, 4.00 mL, 71.61 mmol, 25.5 eq.) and cooled down to 0 °C. Subsequently,
tetrafluoromethansulfonic acid anhydride (4.08 g, 2.43 mL, 14.45 mmol, 4.8 eq.) was
added over the period of 40 min, resulting in a dark-orange colour of the suspension.
After stirring for 3 h at room temperature ice-cold 1 M hydrochloric acid (20 mL) was
added and the precipitated product washed with ice-cold water (3x 1.5 mL). The crude
product was purified by recrystallization in ACN/ water. Finally, the product could be
isolated as a white-crystalline solid (1.48 g, 2.23 mmol, 74.0 %).
Melting point: 171.8 °C
3. Materials & Methods
22
Synthesis of 4-arm poly(2-methyl-2-oxazoline)-b-poly(2-n-propyl-2-oxazine)-tert-butyl
piperazine-1-carboxylate [(P[MeOx]n-b-P[nPrOzi]n)4] (P3, P4, P5)
General synthesis protocol
The synthesis and purification of the polymers was adapted from recently described
methods.[10] 2-methyl-2-oxazoline, benzonitrile (13.00 mL) and I1 were added to a
flame-dried flask and stirred for 4.5-6 h at 105 °C. After the 1H NMR showed full mon-
omer consumption, the solution was cooled down to room temperature and 2-n-pro-
pyl-oxazine was added. The reaction was continued at 120 °C for at least 16 h. After
no monomer signal was observed in the 1H NMR spectra, the reaction was cooled
down to 0 °C and tert-butyl piperazine-1-carboxylate was added. Following stirring for
18 h at 40 °C the reaction mixture was cooled to room temperature. Potassium car-
bonate was added, and the solution stirred for another 5 h. The solvent was removed
under reduced pressure and the pre-dried polymer was left in a vacuum oven (0 mbar,
50 °C) for 2 d. Finally, the crude product was dialyzed against ultrapure water for 24 h
(MWCO: 4 kDa) and lyophilized. The final product was isolated as white foam.
3. Materials & Methods
23
4-arm [P[MeOx]40-b-P[nPrOzi]49] (P3)
Reagents
Reagent Mass [g] Volume [mL] Moles [mmol] Equivalents
2-n-propyl-2-oxazine 3.59 3.78 28.22 188.1
2-methyl-2-oxazoline 2.63 2.62 30.90 206.0
I1 0.10 - 0.15 1.0
Boc-piperazine 0.10 - 0.56 3.7
K2CO3 0.32 - 2.32 16.0
Yield: 5.40 g (83.2 %)
1H NMR (300 MHz, Chloroform-d) δ 3.76 – 3.38 (m, 772H), 3.37 – 3.15 (m, 766H),
2.25 (q, J = 6.5 Hz, 381H), 2.18 – 2.02 (m, 588H), 1.73 (s, 603H), 1.70 – 1.58 (m,
389H), 1.46 (s, 36H), 0.95 (t, J = 5.9 Hz, 573H).
Mw: 7.9 kg/mol
Mn: 5.3 kg/mol
Đ: 1.50
3. Materials & Methods
24
4-arm P[MeOx]23-b-P[nPrOzi]23 (P4)
Reagents
Reagent Mass [g] Volume [mL] Moles [mmol] Equivalents
2-n-propyl-2-oxazine 4.30 4.52 33.81 96.6
2-methyl-2-oxazoline 2.87 2.86 33.72 96.4
I1 0.23 - 0.35 1.0
Boc-piperazine 1.47 - 7.89 22.6
K2CO3 1.12 - 8.10 23.2
Yield: 5.80 g (87.5 %)
1H NMR (300 MHz, Chloroform-d) δ 3.46 (s, 419H), 3.36 – 3.19 (m, 428H), 2.25 (q,
J = 6.6 Hz, 215H), 2.17 – 2.05 (m, 295H), 1.76 (s, 318H), 1.71 – 1.60 (m, 207H), 1.45
(s, 36H), 0.95 (t, J = 5.9 Hz, 315H).
Mw: 4.5 kg/mol
Mn: 2.9 kg/mol
Đ: 1.53
3. Materials & Methods
25
4-arm P[MeOx]14-b-P[nPrOzi]14 (P5)
Reagents
Reagent Mass [g] Volume [mL] Moles [mmol] Equivalents
2-n-propyl-2-oxazine 3.89 4.09 30.58 51.0
2-methyl-2-oxazoline 2.66 2.65 31.25 52.1
I1 0.40 - 0.60 1.0
Boc-piperazine 2.67 - 14.34 23.9
K2CO3 1.95 - 14.10 23.5
Yield: 4.62 g (73.9 %)
1H NMR (300 MHz, Chloroform-d) δ 3.45 (s, 245H), 3.29 (s, 233H), 2.30 – 2.18 (m,
123H), 2.15 – 2.01 (m, 181H), 1.78 (s, 111H), 1.70 – 1.59 (m, 120H), 1.44 (s, 36H),
0.94 (t, J = 5.7 Hz, 183H).
Mw: 3.3 kg/mol
Mn: 2.1 kg/mol
Đ: 1.55
3. Materials & Methods
26
3.2.2 Rheological & Viscosity Measurements
For rheological measurements a conical plate-plate setup of the above described rhe-
ometer was used. Approximately 500 µL of sample were applied onto the base plate
of the device. For preparation of the measurement a CP60-0.5 conical plate was low-
ered onto the sample plate to obtain a final gap of 0.058 mm. The sample chamber
was temperature sealed by using a water barrier.
For temperature dependent viscosity measurements densities of each sample were
determined at the start and end temperature (10 and 50 °C) of the viscosity measure-
ments. The samples then were filled into capillaries with 1.8 mm or 2.1 mm diameter,
depending on viscosity and polymer concentration of the samples. Temperature de-
pendent measurements were conducted between 10 and 50 °C, at angles between 20
and 70 °, depending on the sample viscosity. Measurement increments were chosen
as 0.5 °C. Each value is the mean average of four measurements.
3.2.3 Differential Scanning Calorimetry Measurements
For differential scanning calorimetry (DSC) measurements 10-20 mg of sample were
given into an aluminium pot with pierced lid. All samples were measured against an
empty pot as a reference. Per sample three temperature cycles were measured. The
first one was disregarded. Under nitrogen atmosphere, samples were cooled down to
-50 °C and heated up to 210 °C. The heating rate was chosen to be 5 K/min.
3.2.4 Nuclear Magnetic Resonance Spectroscopy Measurements
For NMR spectroscopy samples were prepared by dissolving 5-7 mg of sample in deu-
terated solvent. Magnetic field of the spectrometer was given as 300 MHz. 1H NMR
spectra of single molecules were acquired with 32 scans; 128 scans were used for
polymers.
3. Materials & Methods
27
3.2.5 Gel-Permeation Chromatography Measurements
Gel-permeation chromatography (GPC) measurements were conducted on the previ-
ously mentioned instrument. The column ran with 1,1,1,3,3,3-hexafluoro-2-propanol
(HFIP, 3 g/L potassium trifluoroacetic acid (KTFA)) as eluent. The applied pressure
was between 32-34 bar. Temperature for the measurements was 40 °C.
For sample preparation, 2 mg of sample was dissolved in 300 µL of HFIP and filtered
(200 µm) prior to transferring the sample solutions in GPC vials with corresponding
inlets.
3.2.6 WST-1 Assay
2000 NIH 3T3 fibroblasts were seeded into a 96-well plate and cultured in DMEM me-
dium (1 % Penicillin/Streptomycin, 10 % foetal calf serum (FCS)) for 24 h at 37 °C and
5% CO2. Polymer solutions were added to final concentrations of 10, 5, 1, 0.1 and 0.01
wt.% and a total volume of 250 µL per well. After another 24 h in culture, medium was
removed, the cells washed with PBS (1x 200 µL) and filled up with fresh medium (250
µL). WST-1 reagent (10 µL) was added to the wells and the cells were incubated for
4 h. The absorbance the samples was determined by UV/Vis spectroscopy at a wave-
length of 450 and 650 nm. Samples were briefly shaken before the measurement. As
a negative control, wells were filled up only with cell culture medium. A positive control
was prepared by adding methanol to cells. Each sample was measured in a triplicate.
4. Results & Discussion
28
4. Results & Discussion
4.1 Motivation
To achieve the aim of personalized medicine and ultimately the printing of functional
cell constructs, bioprinting has proven to be a promising way to print highly hierarchical
structures with different cells and material properties.[136–138] Given the novelty of
the field, recent technological advances of printers led to rapid progression relative to
the material development, thus the demand for a library of different bioinks that can be
adapted rapidly into available printing devices has risen. Following the call for new
materials in bioprinting, many bioink candidates have been developed recently.[139–
142] A promising example of such a candidate is a diblock copolymer of PMeOx and
PnPrOzi.[10] It forms a physically crosslinked hydrogel with high mechanical strength
(~3 kPa) at 20 wt.%, which possesses superb shear thinning and recovery properties,
as well as high cytocompatibility up to high polymer concentrations of 25 wt.%. NIH
3T3 cells that were encapsulated into the hydrogel and extruded using a bioprinter
showed excellent cell survival rates in the printing process.
Based on this work, this thesis aims to make the named system more adjustable for
the encapsulation of different cell types and printing devices. It is assumed that hydro-
gel properties, like hydrogel strength, shear thinning and recovery properties could be
adjusted by introducing the same kind of polymer in different architectures into the
system. With this it would be possible to print a hierarchical object with e.g. different
mechanical properties using a single bioink and without introducing alternative mole-
cules that could harm the hydrogel structure, which is formed up by hydrophobic inter-
actions between the PnPrOzi blocks of the polymer.
Therefore, linear and star-shaped block copolymers of the named monomers were
synthesized and blended together, to evaluate the effect of blending the two architec-
tures on the hydrogel properties.
By developing such a fully adjustable bioink, it is planned to contribute to a bigger
hydrogel library for bioprinting, that can be used in the future biofabrication of functional
tissues, and, therefore, contribute to the aim of personalized medicine.
4. Results & Discussion
29
4.2 Synthesis of a Linear 2-n-Propyl-2-Oxazine and 2-Methyl-2-Oxazoline Block Copolymer
To create a foundation for the synthesis of different polymer architectures of nPrOzi
and MeOx, a linear copolymer of the aforementioned monomers was synthesized us-
ing CROP (Figure 5).
Figure 5: Chemical structures of 2-methyl-2-oxazoline (MeOx), 2-n-propyl-2-oxazine (nPrOzi) and the
copolymer P1.
P1 was characterized using nuclear magnetic resonance (NMR) spectroscopy (Fig-ure 6), rheology (Figure 7) and gel permeation chromatography (GPC) (Figure 8).
Lorson et al.[10] could show that the hydrogel properties are not dependent on the
chain length of the copolymer. Therefore, the results of all measurements were com-
pared to an earlier batch of the polymer with a chain length of 100 monomer units per
polymer block (P2). P2 was provided by Thomas Lorson.
4. Results & Discussion
30
Figure 6: 1H NMR spectrum of P1 measured in deuterated methanol at 300 MHz.
In the 1H NMR spectrum of P1 (Figure 6) there is no characteristic monomer signal of
MeOx at 3.75 and 4.17 ppm or nPrOzi at 3.29 and 4.07 ppm visible, indicating complete
monomer conversion. The experimental chain length was determined by end group
analysis, referring to the methyl group fragment of the initiator (signal 1). The chain
length will then be compared to the theoretically calculated one (Table 1).
4. Results & Discussion
31
Table 1: Chemical structure of P1, theoretical and experimental block length of PMeOx and PnPrOzi
in P1 derived by theoretical and experimentally found values for 1H NMR spectrum integrals.
Signal No.
Theoretical Integral
NMR Integral
Theoretical block length
(PnPrOzi/PMeOx)
NMR spectra block length
(PnPrOzi/PMeOx)
2 320 312
80/80 83/68
3 160 166
4 320 271
5 160 172
6 160 174
7 240 250
8 240 207
The theoretical block length of PnPrOzi corresponds well with the experimentally found
one. However, the PMeOx block appeared to be 12 monomer units (15%) shorter than
theoretically calculated. This difference of chain lengths could be explained by mono-
mer conversion in the second step of the polymerization. If in the synthesis of the first
block (PnPrOzi) chain transfer reactions occur when there is not much monomer left
in solution, then these chains of low molecular mass will be much smaller than the
regular PnPrOzi chains. Therefore, if in the second step MeOx monomer gets con-
sumed by these chains, it could result in short chains, which then get washed out in
the dialysis process. Furthermore, MeOx chains could have been terminated too early
by impurities in the reaction system. These low molecular products were then sepa-
rated in dialysis, resulting in lower length of the PMeOx block in the remaining polymer.
To assess the rheological behaviour of the synthesized copolymer, a temperature
sweep in the linear viscoelastic range of the aqueous polymer solution (for amplitude
and frequency sweep see Appendix I: Figure 5 and 6) was deducted at 20 wt.%. At
this concentration the polymer solution should exhibit a sol-gel transition and conse-
quently form a hydrogel at room temperature.
4. Results & Discussion
32
Figure 7: Gelation temperatures and storage and loss modulus (G' and G'') of a 20 wt. % hydrogel of
P1 (dialysed in tube with MWCO of 1 kDa) (grey) and a reference linear copolymer of MeOx and
nPrOzi P2 (green) at temperatures between 1 and 50 °C.
With the storage modulus (G’) being higher than the loss modulus (G’’) in the region
above 21 °C, P1 does form a hydrogel at this polymer concentration. However, it can
be seen (Figure 7) that even after 18 h of dialysis against Millipore water (MWCO:
1 kDa) the hydrogel strength of the synthesized P1 (G’ = 490 Pa, tan δ = 0.14) com-
pared to P2 (G’ = 2810 Pa, tan δ = 0.03) is weak. Also, the gelation temperature, de-
fined as the intersection of G’ with G’’, is approximately 9 °C higher than for P2. This
different behaviour could originate from the difference in block lengths of a polymer
chain, since the ratio of PMeOx and PnPrOzi is not the ideal 1:1 anymore. It could also
be due to impurities that originate from the polymerization, i.e. low molecular side prod-
ucts in the reaction. To further analyse the polymer a GPC sample of P1 was measured
(Figure 8). Weight and number average molecular mass (Mw and Mn) as well as the
dispersity Đ of the sample were determined (Table 2).
4. Results & Discussion
33
Figure 8: GPC elugram of P1 after 18 hours of dialysis against Millipore water in a dialysis tube with
MWCO 1 kDa. The elugram shows low molecular tailing of the monodisperse polymer. GPC samples
were run with 1,1,1,3,3,3-hexafluoro-2-propanol (HFIP) as solvent.
Table 2: Dispersity and weight and number average molecular mass of P1 after 18h of dialysis in a
tube with MWCO 1 kDa.
Mw [kg/mol] Mn [kg/mol] Đ
5.3 4.2 1.29
The elugram shows a monomodal molecular mass distribution. The molecular mass
acquired by GPC measurements can only be a relative reference between measure-
ments of similar polymers and not as the absolute molecular mass of the polymer. The
GPC instrument that was used for measurements was calibrated with linear poly(eth-
ylene glycol), which has different retention times on the GPC column.
It is known from literature that a low amount of low molecular tailing can affect the
gelation of the polymer, which is based on supramolecular assembly of the polymer
chains into a sponge structure.[10] Therefore, it was hypothesized that the same could
affect P1, even though no pronounced low molecular tailing is visible in the GPC
elugram. To assess this possibility, the polymer was again dialysed against Millipore
water, this time with a dialysis tube with a MWCO of 10 kDa instead of 1 kDa and for
a total of 56 h. A sample of the polymer was collected after 28 and 56 h and analysed
via GPC and rheology.
4. Results & Discussion
34
Figure 9: GPC elugram of P1 after 18 hours of dialysis against Millipore water in a dialysis tube with
MWCO 1 kDa (grey), followed by 28 (red) and a final of 56 h (blue) in a dialysis tube with MWCO 10
kDa. The low molecular tailing observed after the first dialysis step decreased with higher MWCO and
longer dialysis time. GPC samples were run with 1,1,1,3,3,3-hexafluoro-2-propanol (HFIP) as solvent.
It seems like there was indeed a non-prominent low molecular tailing, which was re-
duced after 28 h of dialysis and even more after 56 h (Figure 9). This is also shown in
the trend of dispersity values of the polymer (Table 3).
Table 3: Dispersity and weight and number average molecular mass of P1 after 18 hours of dialysis
against Millipore water in a dialysis tube with MWCO 1 kDa., followed by 28 and a final of 56 h in a di-
alysis tube with MWCO 10 kDa.
Mw [kg/mol] Mn [kg/mol] Đ
18 h 1 kDa 5.3 4.2 1.29
28 h 10 kDa 5.8 4.7 1.23
56 h 10 kDa 5.9 4.9 1.21
Both, the relative increase of the molecular mass and the lower dispersity indicate the
absence of low molecular mass impurities in the system. To assess whether the hy-
drogel properties changed, temperature sweeps in the linear viscoelastic range were
measured to compare it to the initial P1 sample (Figure 10).
4. Results & Discussion
35
Figure 10: Gelation temperature, G' and G'' of a 20 wt.% hydrogel of P1 after 18 hours of dialysis
against Millipore water in a dialysis tube with MWCO 1 kDa (grey)., followed by 28 (red) and a final of
56 h (blue) in a dialysis tube with MWCO 10 kDa and a reference linear copolymer of PMeOx and
PnPrOzi P2 (green) at temperatures between 0 and 50 °C.
After 28 h of dialysis in a dialysis tube with MWCO 10 kDa G’ of P1 increased from 490
Pa to 1600 Pa (tan δ = 0.13). Also, the temperature of gelation decreased from 21.1 to
20.2 °C. After 56 h G’ is even slightly above the one of reference P2 (2950 Pa,
tan δ = 0.08). Nonetheless, even though the same hydrogel strength could be acquired
after dialysis it still shows a different gelation behaviour compared to P2. Both polymers
have a two-step gelation, indicating two different phase transitions above the gelation
temperature. The synthesized P1 remains, after reaching the first plateau of gelation,
between 20.5 to 27.5 °C in this state, while P2 shifts at 13.3 °C almost immediately
into the second plateau. This behaviour was also observed in P1, before dialysing it in
a MWCO 10 kDa dialysis tube. However, to assume the origins of this just from the
temperature sweep measurement is unjustified and not relevant for the focus of this
thesis.
In addition to the temperature sweeps performed on the rheometer, an oscillation re-
covery experiment with a 20 wt.% solution of P1 was conducted and the results were
compared to a solution of P2 of the same concentration (Figure 11)
4. Results & Discussion
36
Figure 11: Recovery properties of a 20 wt. % hydrogel of P1 (grey) compared to a hydrogel of the ref-
erence polymer P2 with the same solid content (red). G' and G'' in this case are dependent on the
shear stress that was applied. Applied shear stress for the relaxed state was chosen to be 0.5 % and
in the tension state 150 %. Samples were measured at 37 °C.
The red curve represents the reference polymer P2 and the black one the synthesized
P1. P2 shows excellent shear thinning properties at high strains of 150%, which can
be seen at the direct drop of G’, as soon as high strain is applied. As soon as the strain
was reduced to 0.5%, the polymer recovers directly back to its original G’ value. This
result was verified over three consecutive cycles of high and low strain. Compared to
that, P1 shows similar behaviour. Once high strain is applied, it loses its gel character
and becomes a sol (G’’ > G’). The same can be seen in terms of recovery. P1 instantly
recovers after strain is released. Nevertheless, in contrast to P2, P1 recovers over two
states. This, in combination with the large temperature region, in which P1 remains
within the first plateau of gelation might be a reason for the two-step recovery. If the
second plateau is reached at 37 °C, where the measurement takes place, it is possible
that the system recovers again over these two structural states, after high strains. P2,
on the other hand, showed almost instant transition into the second plateau state in
the temperature sweep (Figure 10). Apart from that, P1 also shows recovery back to
its initial hydrogel strength over the period of three consecutive measuring cycles. The
direct decrease of G’ after application of a high shear stress further shows the good
shear thinning properties of P1, that resemble those of P2.
4. Results & Discussion
37
In the end, a linear diblock copolymer P1 of PMeOx and PnPrOzi could be synthesized
and characterized using GPC, NMR spectroscopy and rheology. Compared to P2, an-
other copolymer of PMeOx and PnPrOzi with a longer chain length, it showed similar
hydrogel strength and recovery properties. The gelation temperature and the recovery
in two steps differed from the reference. Given the similar G’ and tan δ and good re-
covery over three cycles, it was aimed to use the polymer for blending purpose with
different polymer architectures of the same monomers.
4.3 Synthesis of Star-Shaped Block Copolymers of 2-n-Pro-pyl-2-Oxazine and 2-Methyl-2-Oxazoline
In the second part of this thesis the synthesis of the linear polymer should be trans-
ferred to more complex architectures.
Figure 12: Schematic figure of different polymer architectures of PMeOx (blue) and PnPrOzi (red). In-
stead of linear copolymers of the two monomers, miktoarm polymers, polymer brushes and di-block
copolymer star-shaped polymers are possible.
The assumption behind synthesizing different architectures, like miktoarm-polymers,
polymer brushes or star-shapes di-block copolymers with same chemical groups orig-
inates from the desire to modify material properties without adding other polymers,
nanoparticles, etc. Multiarm star-shaped copolymers could e.g. be able to act as cross-
linking points in the sponge-like network of the linear polymer and, therefore, enhance
the mechanical properties of the hydrogel. They could also be used as covalent cross-
linkers of the network, when modified with e.g. click-chemistry suitable moieties. All in
4. Results & Discussion
38
all, adding different architectures to the established linear system might offer the op-
portunity to tailor the bioink properties to cell type and printer, and make it a more
versatile solution for bioprinting problems.
For that purpose, star-shaped diblock copolymers of PMeOx and PnPrOzi with differ-
ent arm lengths were synthesized and later blended with the linear polymer P1.
4.3.1 Polymer Design
For the synthesis of star-shaped di-block copolymers two issues had to be addressed.
(1) In what order should the blocks be attached to the initiator? (2) What are suitable
arm lengths for star-shaped polymers, with respect to the monomer units of the estab-
lished linear system? For the order of monomer block it was decided to have the more
hydrophilic PMeOx as a first block close to the initiator. The initiator is a tetra function-
alized triflate of pentaerythritol, since it has been already used in literature[100,135].
The more hydrophobic PnPrOzi block forms the outer shell of the 4-arm star. Even
though the polymers will be dissolved in water, the PnPrOzi as a more hydrophobic
outer shell was chosen because of better interaction with the circumferential linear pol-
ymer chains. The PnPrOzi block is responsible for the hydrogel forming properties of
the linear polymer system at temperatures above 19 °C (P1). Therefore, having it as
the inner block, the PMeOx blocks could have inhibited the physicochemical interaction
between the hydrophobic polymer blocks.
4.3.2 Initiator Synthesis
The first step for synthesizing the star-shaped 4-arm copolymers was initiator I1, fol-
lowing a procedure previously described by Menger et al.[135] (Figure 13).
Figure 13: Synthesis route of initiator I1. Pentaerythritol reacts in a suspension of pyridine and ace-
tonitrile (ACN) with trifluoro methane sulfonic acid anhydride to tetrakistriflatepentaerythritol.
4. Results & Discussion
39
The sulfonic acid anhydride is added in excess, to ensure functionalization of each arm
of pentaerythritol. In literature, the product was characterized by 1H NMR spectroscopy
and melting point analysis. It shows good solubility in acetonitrile while being insoluble
in water. However, due to the spectrum only showing one signal of the CH2-groups at
the molecule, NMR spectroscopy seems to be not the analysis method of choice. This
can also be seen in the 1H NMR spectrum of I1.
Figure 14: 1H NMR spectrum of I1 measured in deuterated acetonitrile at 300 MHz.
The spectrum (Figure 14) shows one peak at 4.75 ppm. However, this peak cannot
be references against any other peak in the system. Pentaerythritol cannot be added
to the same NMR solvent, since it shows opposing solubility to the functionalized prod-
uct. The NMR signal is slightly high-field shifted, compared to literature
(5.15 ppm)[135] in the same solvent. This sole result in terms of characterization, how-
ever, leaves the degree of functionalization open. As the acetonitrile solvent used for
NMR spectroscopy measurement contains huge amount of water, this could potentially
lead to cleaving off triflate groups. On the other hand, this could then be observed as
different signals in the spectrum. To confirm the tetra functionalization of I1 the litera-
ture melting point difference of pentaerythritol and tetrakistriflatepentaerythritol was ex-
ploited. While pentaerythritol melts at 263 °C[143], tetrakistriflatepentaerythritol has a
melting point of 165-166 °C in literature[135]. This difference of melting points could
indicate that there is a melting point range from 263-165 °C depending on the degree
of functionalization of the alcohol units. Therefore, the same reaction was conducted
4. Results & Discussion
40
again, but instead of an excess of the anhydride, ½ and ¾ of the hydroxide equivalents
were added to the reaction to ensure partial functionalization of pentaerythritol.
Figure 15: Proposed chemical structure of initiators I2 and I3 from pentaerythritol with partially func-
tionalized hydroxide groups.
It is hypothesised that two species (Figure 15) will be synthesized with different melting
points in the range between 165 and 263 °C. I1, I2 and I3 got characterized using
differential scanning calorimetry (DSC) (Figure 16).
Figure 16: DSC elugram of I1 (grey), I2 (red) and I3 (blue). Samples were heated and cooled with a
rate of 5 K/min to a maximum of 210 °C and a minimum of -50 °C. Y-axis shows exothermal
processes as a negative signal and endothermal processes as a positive signal. A total of 3 cycles
was measured per sample, the first cycle was discarded. The second cycle is shown.
The melting points are all approximately at the same temperature (Figure 16). Rather
than indicating the presence of different functionalized species, this difference can also
occur due to differences in the cooling and heating cycle of the DSC instrument. The
recrystallization peaks in the cooling phase are more shifted. I2 and I3 have a
4. Results & Discussion
41
difference to I1 of maximal ± 1.6 °C. This could be interpreted as different substances
in the DSC. However, 1H NMR spectroscopy indicates that all samples are the same
material, given a same shift of signals in the spectrum for each sample I1, I2 and I3 (Appendix I: Figure 3 and 4).
To conclude, the conducted experiment suggests that only the tetra functionalized ini-
tiator was synthesized, since, independently of the trifluoromethyl sulfonic acid anhy-
dride equivalents, the melting point of the final product does not change significantly.
This assumes a melting point gradient between fully functionalized and non-function-
alized reagent. However, the less equivalents of the anhydride were added, the lower
was the yield of each reaction. Therefore, it could be hypothesized that I1 is the ener-
getically favoured form of functionalized triflates on a pentaerythritol core. On the other
hand, it is also possible that all functionalization degrees are synthesized in the reac-
tion. In this case, the partially functionalized compounds could still be soluble in water.
As the crude product of the reaction was washed with cold water, for purification, these
partially functionalized side products could be potentially washed out. Thus, only the
tetra functionalized product could be obtained. To confirm this theory, the washing so-
lution should be analysed via 1H NMR spectroscopy in future experiments.
Additionally, 1H NMR spectroscopy of I2 and I3 (Appendix I: Figure 3 and 4) was
conducted to compare the signal shift of the CH2 groups with the ones from I1.
Both spectra (I2, I3) showed the same shift as observed for I1. This, together with the
decrease in yield of the final product with lower anhydride equivalents, is another indi-
cation that only one species of initiator is synthesized, independently of the amount of
trifluoro sulfonic acid anhydride in the reaction.
Furthermore, Luxenhofer et al.[144] could previously show that a tetra functionalized
initiator, that was synthesized following the same procedure, showed a reaction kinetic
that was four times faster than a polymerization of 2-oxazolines initiated with methyl
triflate.
To sum it up, it can be assumed that initiator I1 was successfully synthesized. This
could be verified by DSC, 1H NMR NMR spectroscopy and previous conducted exper-
iments in literature.
4. Results & Discussion
42
4.3.3 Polymerization of Star-Shaped Copolymers of 2-methyl-2-ox-azoline and 2-n-propyl-2-oxazine
As a more complex structure of the di-block copolymers of PMeOx and PnPrOzi, 4-
arm star-shaped architectures (P1-P3) of different block lengths were synthesized us-
ing the protocol of the linear polymer synthesis (Figure 17).
Figure 17: Theoretical chemical structures of star-shaped copolymers of PMeOx and PnPrOzi. Block
length was halved from P3 to P4 and from P4 to P5.
P3 is the polymer with highest number of repeating units. With a theoretical length of
approximately 50 repeating units per block the direct comparison to one of the linear
polymers of the same block length, described by Lorson et al.[10], was aimed for. How-
ever, the molecular mass is four times higher than in the linear system. P5, on the other
hand, was supposed to mimic the linear polymer by having the 52 repeating units of
each block on the whole star, and therefore, a comparable molecular mass to the linear
system. P4 is an intermediate sized star-shaped polymer to fill in the big size gap be-
tween the aforementioned polymers. Similar as described in Section 4.2 the experi-
mentally found block length was supposed to be determined from the 1H NMR spectra
by end group analysis using the signal of the Boc-group. Exemplary, the 1H NMR spec-
trum of P3 is shown (Figure 18). (For P4 and P5 see Appendix I: Figure 1 and 2).
4. Results & Discussion
43
Figure 18: 1H NMR spectrum of P3 measured in deuterated chloroform at 300 MHz.
The experimental amounts of repeating units were determined by integration of the 1H
NMR signals, referenced to the Boc-signal at 1.44-1.46 ppm (36 H = 4x 9H), depending
on the polymer. (Table 4-6).
Table 4: Chemical structure, theoretical and experimental block length of PMeOx and PnPrOzi in P3.
Signal No. P3
Theoretical Integral
NMR Integral
Theoretical block length
(PMeOx/PnPrOzi)
NMR spectra block length
(PMeOx/PnPrOzi)
1 832 772
52/47 38/51
2 752 766
3 376 603
4 624 588
5 376 381
6 376 389
7 564 573
4. Results & Discussion
44
Table 5: Chemical structure, theoretical and experimental block length of PMeOx and PnPrOzi in P4.
Signal No. P4
Theoretical Integral
NMR Integral
Theoretical block length
(PMeOx/PnPrOzi)
NMR spectra block length
(PMeOx/PnPrOzi)
1 384 419
24/24 34/34
2 384 428
3 192 318
4 288 295
5 192 215
6 192 207
7 288 315
Table 6: Chemical structure, theoretical and experimental block length of PMeOx and PnPrOzi in P5.
Signal No. P5
Theoretical Integral
NMR Integral
Theoretical block length
(PMeOx/PnPrOzi)
NMR spectra block length
(PMeOx/PnPrOzi)
1 208 245
13/13 22/20
2 208 233
3 104 111
4 156 181
5 104 123
6 104 120
7 156 183
4. Results & Discussion
45
The experimentally derived chain lengths of P3, P4 and P5 for each block show a
deviation to the calculated values. Furthermore, signal 3 was for the calculation of the
chain length of P3 and P4 not assessed because of an underlying solvent signal, most
likely caused by water from the deuterated solvent or the polymer itself. The deviation
from the theoretical chain length is high in all copolymers. However, the trend of block
chain length can still be seen. It is noteworthy that the integration of the signals was
highly dependent on the integral chosen for the Boc-signal. Therefore, the high integral
values might be caused by that. It could also be argued that the protection group
cleaves off with time, when the polymer is in solution, and with this, the integral of the
Boc-group would have to be chosen as less than 36.
To get a better insight into the chain length, GPC measurements of the three polymers
P3-P5 were conducted (Figure 19).
Figure 19: GPC elugram of P3 (grey), P4 (red) and P5 (blue). All signals show a bimodal mass distri-
bution. GPC samples were run with 1,1,1,3,3,3-hexafluoro-2-propanol (HFIP) as solvent.
4. Results & Discussion
46
Table 7: Dispersity and weight and number average molecular mass of P3, P4 and P5.
Mw [kg/mol] Mn [kg/mol] Đ
P3 7.9 5.3 1.50
P4 4.5 2.9 1.53
P5 3.3 2.1 1.55
All elugrams show a significant bimodality that occurs in all three star-shaped poly-
mers. That caused high Đ values of up to 1.55 (Table 7). Measured molecular masses
are not representative for the actual molecular mass of each compound, since the GPC
instrument was calibrated with a PEG standard. However, it can be used to compare
coil sizes of the polymers with each other. As expected, P3 has its signal at lower
elution volumes, while P5 shows, due to the lower number of repeating units, longer
retention times on the column. Also, the dispersity of all polymers is comparable, indi-
cating that it could be the same phenomenon occurring in all three synthesis proce-
dures, that causes the bimodality of the copolymers. Furthermore, to ensure that the
bimodality is not only an artefact from the GPC analysis, a sample of each polymer
was additionally measured on another GPC system using dimethylformamide (DMF)
as eluent. Again, a bimodality could be observed for all copolymers (P3-P5).
Side reactions like chain transfer or coupling reactions for POx homopolymers have
been reported repetitively in literature.[92,94,145–147] Therefore, there is a high
chance that the bimodality is also caused by this kind of reactions, since the protocol
for linear block copolymers of PMeOx and PnPrOzi was used without change, even
though four times the monomer amount is present, when synthesizing a 4-arm star-
shaped copolymer of these monomers. As described in section 2.2.2, these bimodal-
ities could be caused by too high temperatures of the reaction or a too high concentra-
tion of the monomers. Based on that, three experiments, aiming to synthesize a uni-
modal star-shaped copolymer P4, were conducted. In the first two experiments the
monomer concentration in the reaction mixture was reduced by half and quarter of the
initial concentration, which was described by Lorson et al.[10]. The third experiment
was using reduced polymerization temperatures of 80 °C for the PMeOx block and
100 °C for the PnPrOzi block, instead of the previously used standard conditions of
100 °C for PMeOx and 120 °C for PnPrOzi.
4. Results & Discussion
47
1H NMR spectrum and GPC samples of the first experiment with half of the initial mon-
omer concentration were measured (Figure 20, Table 8).
Figure 20: 1H NMR spectrum of P4, polymerized with half of the initial monomer concentration meas-
ured in deuterated chloroform at 300 MHz. GPC elugram of the polymerized product, with a high mo-
lecular shoulder. GPC samples were run with 1,1,1,3,3,3-hexafluoro-2-propanol (HFIP) as solvent.
Table 8: Dispersity and weight and number average molecular mass of the modified P4 polymerization
with half the initial monomer concentration.
Mw [kg/mol] Mn [kg/mol] Đ
½ Concentration 3.1 2.1 1.45
The NMR spectrum shows significantly weaker signals of the PnPrOzi block of the
copolymer, indicating insufficient reaction of the monomer at this concentration. On the
other hand, also no Boc-signal in the spectrum can be found, indicating a termination
of the reaction before the termination step. Early termination of chains might explain
the absence of PnPrOzi signal in the spectrum, since it could not polymerize. That also
corresponds to the long reaction times that were observed for the PnPrOzi block. Fur-
thermore, a new peak at 1.94 ppm arises (red circle), which implies the presence of
side reactions. Resulting from this, the molecular mass of the polymer acquired from
the GPC measurement is 1.4 kg/mol lower than P4. Nevertheless, the elugram shows
a large reduction of the lower molecular peak in the GPC. This might indicate that the
bimodality is mainly cause by the PnPrOzi block of the polymer. In contrast, it could be
proven through reaction control after the polymerization of each block that the bimo-
dality at standard reaction parameters already occurs after polymerizing the PMeOx
4. Results & Discussion
48
block. Since it seems like early termination of the reaction occurred, a repetition of the
experiment was conducted.
Figure 21: 1H NMR spectrum of P4, polymerized with half of the initial monomer concentration meas-
ured in deuterated chloroform at 300 MHz. GPC elugram of the polymerized product, with a high mo-
lecular shoulder. GPC samples were run with 1,1,1,3,3,3-hexafluoro-2-propanol (HFIP) as solvent.
Table 9: Dispersity and weight and number average molar mass of the repeated modified P4 polymer-
ization with half the initial monomer concentration.
Mw [kg/mol] Mn [kg/mol] Đ
½ Concentration #2 4.3 2.9 1.45
In the repetition experiment (Figure 21, Table 9), the copolymer shows a similar proton
NMR spectrum to P4. nPrOzi seems to be fully consumed and polymers show termi-
nation through Boc-piperazine, indicated by the corresponding signal in the spectrum.
The peak that indicated side reactions in the system is not dominant anymore. How-
ever, the GPC elugram shows another bimodal distribution. Mw suggested by the GPC
measurement is higher compared to P4. Most importantly, the result of Figure 20 could
not be reproduced. It can be concluded that halving the concentration of the monomer
in the reaction mixture does not affect the bimodality caused in the process of polymer
synthesis.
To further assess the hypothesis that lower monomer concentrations prevent side re-
actions in the system, another P4-type polymer was aimed to be synthesized, with only
¼ of the initial concentration of monomer for the polymerization.
4. Results & Discussion
49
Figure 22: 1H NMR spectrum of P4, polymerized with quarter of the initial monomer concentration
measured in deuterated chloroform at 300 MHz. GPC elugram of the polymerized product, with a high
molecular shoulder. GPC samples were run with 1,1,1,3,3,3-hexafluoro-2-propanol (HFIP) as solvent.
Table 10: Dispersity and weight and number average molecular mass of the modified P4 polymeriza-
tion with quarter of the initial monomer concentration.
Mw [kg/mol] Mn [kg/mol] Đ
¼ Concentration 1.8 1.4 1.28
The data (Figure 22, Table 10) shows a similar result to the experiments, where mon-
omer concentration was halved. The NMR spectrum indicates a full conversion of the
nPrOzi monomer. However, the it can be seen in the elugram (Figure 22b) that there
is a high molecular tailing. Mw is 2.7 kg/mol lower than P4, indicating that not all mon-
omers reacted. To verify results, the experiment was conducted a second time under
the same conditions.
4. Results & Discussion
50
Figure 23: 1H NMR spectrum of P4, polymerized with quarter of the initial monomer concentration
measured in deuterated chloroform at 300 MHz. GPC elugram of the polymerized product, with a high
molecular shoulder. GPC samples were run with 1,1,1,3,3,3-hexafluoro-2-propanol (HFIP) as solvent.
Table 11: Dispersity and weight and number average molecular mass of the repeated modified P4
polymerization with quarter the initial monomer concentration.
Mw [kg/mol] Mn [kg/mol] Đ
¼ Concentration #2 3.9 2.8 1.38
The acquired data (Figure 23, Table 11) shows again a bimodality in the GPC elugram
similar to P4 or the second repetition of the halved concentration. The molecular mass,
however resembles the ones found in the GPC measurements of P4, which is indicat-
ing a full conversion of both monomers, MeOx and nPrOzi. Furthermore, the Boc-pi-
perazine signal can be observed. In the end, it can be said that due to reproducibility
issues, changing the monomer concentration does not have any influence on the bi-
modality found in the GPC elugram of P4.
As mentioned previously in this chapter, a second possibility to avoid a bimodal molar
mass distribution is the reduction of reaction temperature.[95] POzi are known for
needing higher reaction temperatures compared to POx.[117] The polymerization tem-
perature used in this thesis was 120 °C for the PnPrOzi block. Lowering the tempera-
ture could bring significantly higher reaction times, as previously shown for the synthe-
sis of POx[148]. A lower polymerization temperature could also give way for more side
reactions in the polymerization itself, because the reaction is much slower and active
cationic chain ends can experience chain transfer of termination reactions through
4. Results & Discussion
51
small amounts of impurities in the reaction flask. On the other hand, it could also sup-
press side reactions with these impurities. However, to evaluate the effect of tempera-
ture on the polymerization process, an experiment was conducted where the polymer-
ization temperature for the PMeOx block was 80 °C and 100 °C for PnPrOzi block
(Figure 24, Table 12).
Figure 24: 1H NMR spectrum of P4, polymerized with reduced reaction temperatures of 80 °C for the
PMeOx and 100 °C for the PnPrOzi block measured in deuterated chloroform at 300 MHz. GPC
elugram of the polymerized product, with a high molecular shoulder. GPC samples were run with
1,1,1,3,3,3-hexafluoro-2-propanol (HFIP) as solvent.
Table 12: Dispersity and weight and number average molecular mass of the modified P4 polymeriza-
tion with reduced reaction temperatures of 80 °C for the PMeOx and 100 °C for the PnPrOzi block.
Mw [kg/mol] Mn [kg/mol] Đ
80 °C PMeOx / 100 °C PnPrOzi
3.8 2.5 1.52
The 1H NMR spectrum shows a significant difference in signal intensity of the PnPrOzi
block to the PMeOx block, which was expected given the lower reaction rates at de-
creased temperatures. Due to the low temperatures, the reaction times for PMeOx had
to be prolonged to at least 12 h (overnight), compared to four or five hour, after which
full monomer conversion is usually observable, when polymerized at 100 °C. After 5 h,
there was still a strong signal of free monomer in the reaction mixture. nPrOzi had to
react even longer, given that it naturally is slower in its reaction rate. However, after
48 h the second block was terminated for time frame reasons, with almost no conver-
sion of nPrOzi in the reaction mixture. Due to this the resulting polymer is deficient in
4. Results & Discussion
52
its PnPrOzi block. Furthermore, the dispersity of the GPC elugram shows that, poten-
tially due to the long reaction times, a broad polymer distribution is seen in the final
product of the reaction. Interestingly the acquired values for the molecular mass in this
experiment resemble very well the values for P4, even though there was unpolymer-
ized nPrOzi in solution when the polymerization was terminated. This could be ex-
plained by similar coiling behaviour on the column. This shows the inaccuracy of the
results acquired from the measured GPC samples. For future experiments is sug-
gested to use a standard, which resembles the architecture of the polymers more, i.e.
a standard calibration curve of the GPC with star-shaped PEG molecules. Also, as
seen in the previous experiments regarding the concentration, no Boc-piperazine sig-
nal can be observed.
Alto conclude, it could be shown that the change of monomer concentration does not
reproducibly affect the bimodality of the resulting polymers. The first experiments of
each parameter change were conducted simultaneously. Due to the absence of Boc-
signal in the NMR spectrum, as well as incomplete monomer consumption, it can be
assumed that there had been a systematic mistake, when conducting the experiments.
Follow up experiments could in case of the concentration change show the presence
of termination reagent on the polymer, as well as full monomer conversion. However,
elugrams of these experiments showed a bimodal mass distribution. It is also worth
noticing that the yield of all the polymerization varieties was between 6 and 30 % com-
pared to yields of 74 to 87 % that are obtained for P3-P5 under normal conditions. This
might be due to early chain termination, resulting in separation of these molecules in
the purification process. For further investigation, viscosity detection in the GPC meas-
urements could be used, since the polymer architecture should affect the sample vis-
cosity. With a focus on the thesis time-frame and application the bimodality was ac-
cepted, since the same bimodality could be shown for the mass distribution of all three
different polymers. Therefore, P3-P5 were used for further characterization studies of
star-shaped block copolymers of POx and POzi.
4.3.4 Physicochemical Properties
After the synthesis of three star-shaped polymers with different arm lengths, charac-
terization through more than just 1H NMR spectroscopy and GPC analysis is an im-
portant factor in understanding this new polymer system. The corresponding linear
4. Results & Discussion
53
polymer has already undergone extensive studies, e.g. viscosity measurements, rhe-
ology or small angle neutron scattering (SANS) analysis, by the Luxenhofer research
group.[10] With the synthesis of the star-shaped analogous polymer systems it is
hoped to achieve similar properties to the linear system.
First P3-P5 were assessed on their gelation properties. Therefore, a concentration se-
ries in the range of 1-40 wt.% was generated for each of the polymers and its hydrogel
forming properties were assessed by incubating the solutions at 4 °C, room tempera-
ture (20 °C) and physiological temperature (37 °C) (Figure 25-27).
Figure 25: Photographs of a concentration series of the P3 polymer (1,5, 10, 20, 30 and 40 wt.%) at
different temperatures (4, 20 and 37 °C).
Figure 26: Photographs of a concentration series of the P4 polymer (1,5, 10, 20, 30 and 40 wt.%) at
different temperatures (4, 20 and 37 °C).
4. Results & Discussion
54
Figure 27: Photographs of a concentration series of the P5 polymer (1,5, 10, 20, 30 and 40 wt.%) at
different temperatures (4, 20 and 37 °C).
The observed behaviour was similar to the already described one for the linear system.
The lower concentrations tend to form transparent solutions at lower temperatures,
which turn turbid once the temperature reaches 20 °C. This is due to the lower critical
solution temperature (LCST) of the PnPrOzi block (11 °C[149]) in the polymer, which
affects the LCST of the whole polymer. It is notable that P3 forms turbid solutions at all
concentrations, except at 30 and 40 wt.%, where the formation of a turbid hydrogel
could be observed. The turbidity could be explained due to aggregation of polymers,
which then start scattering light. The aggregation below the LCST of PnPrOzi could be
explained by strong hydrophobic interactions of the four arms, resulting in a dense
hydrophobic shell and, therefore, lowering the solubility of the polymer in solution. The
hydrogels formed by P3 differ from the linear system, since a hydrogel is only formed
at higher polymer content and the hydrogels are turbid, which could be problematic for
cell culture and imaging applications.
Surprisingly, P4 shows a variety of transparent solutions in the given concentration
range. Due to the PnPrOzi block and its LCST it is expected that at some concentra-
tions and higher temperatures the polymer is not soluble anymore and precipitates.
This, however, only occurs in the 1 wt.% aqueous solution at 20 °C. The fact that this
solution gets transparent at higher temperatures again speaks for the presence of an
upper critical solution temperature (UCST) in this system or the formation of another
phase. P4 only forms a hydrogel at 40 wt.% and 37 °C. This is twice the mass of poly-
mer that is needed for a hydrogel than in the linear model. Since the polymers are
supposed to be used as bioinks in biofabrication, and high polymer contents can affect
4. Results & Discussion
55
cell behaviour, resulting in lower cell viability[150]. The hydrogels formed by P3 and
P4 alone are not suitable as bioinks, even though the polymers form hydrogels. Fur-
thermore, P3 has the major drawback of forming turbid hydrogel, which is undesirable
when it comes to imaging of cells in a hydrogel network or the introduction of photo-
chemical click-chemistry for the introduction of biological motifs like peptide sequences
into the system.
The solutions of P5 showed precipitation of polymer depending on the temperature.
With increasing temperature, the concentration, at which the polymer precipitates, de-
creases, so that at 37 °C all solutions appear turbid. P5 does not form hydrogels, pos-
sibly due to the short arm lengths of the polymer. Thermogelation properties of the
PMeOx-PnPrOzi block copolymer system are based on hydrophobic interactions be-
tween the PnPrOzi blocks. It is possible that the short PnPrOzi blocks of P5 coil to-
gether rather than interacting with each other, given the proximity the polymers need
to interact. To further investigate the nature of the network, structural analysis by e.g.
SANS is needed.
The hydrogels formed by polymers P3 and P4 were assessed using rheology and the
results compared to the previously synthesized linear polymer P1. To determine the
LVE region of the hydrogel an amplitude sweep was conducted (Appendix I: Fig-ure 7). To evaluate the hydrogels further, frequency sweeps of the thermogelling sam-
ples were conducted (Figure 28).
4. Results & Discussion
56
Figure 28: Frequency sweep from 0.05 1/s to 100 1/s for the polymer solutions that showed thermo-
gelation. This was the case for a 30 wt.% solution of P3 (A) and a 40 wt.% solution of P3 (B) and P4
(C). The graphs were compared to a 20 wt.% hydrogel of P1 (D) Graphs show G’ and G’’ of the sys-
tem plotted against the applied angular frequency. Samples were measured at 37 °C.
In the case of P3 (Figure 28A and B) all hydrogels are independently in their strength
of the angular frequency applied. The hydrogel of P4 (Figure 28C) shows a viscous
character at low frequencies, since G’’ is higher than G’. G’ increases from 77 Pa at
low to 694 Pa at high frequencies. It is also noteworthy that tan δ (G’’/G’) of the P4
hydrogel is close to 1, which is not the case for P3 hydrogels assessed
[tan δ (30 wt.%) = 0.47; tan δ (40 wt.%) = 0.14], as well as for P1 at 20 wt.%
[tan δ (20 wt.%) = 0.21]. Therefore, the P4 hydrogel shows viscous character very high
and very low frequencies. P3 shows G’ values up to 10 kPa, which is stronger than
many hydrogels that are currently available in literature[151,152], including the synthe-
sized P1, which is at approximately 1 kPa. However, as mentioned before the hydro-
gels are all turbid, which is a major drawback for biological applications.
Temperature sweeps from 1 to 50 °C were conducted to assess the gelation process
and gelation temperatures of the different hydrogels (Figure 29).
4. Results & Discussion
57
Figure 29: Temperature sweep from 1 to 50 °C for the polymer solutions that showed thermogelation.
This was the case for a 30 wt.% solution of P3 (grey) and a 40 wt.% solution of P3 (red) and P4 (blue).
The graphs were compared to a 20 wt.% hydrogel of P1 (green). G’ and G’’ of the system was plotted
against the temperature.
The gelation temperatures were given by the intersection of G’ and G’’ of each sample.
Regardless of the polymer concentration, P3 shows similar gelation temperatures as
P1 hydrogels. P4 exhibits a gelation temperature (34.4 °C) which exceeds the ones of
P3 and P1. In terms of the gelation temperatures, however, it can be summarized that
all investigated hydrogels could still be used in cell culture and biological applications,
since they all form a gel below physiological temperature of 37 °C.
G’ and G’’ of the different hydrogels, show significant differences in the gelation pro-
cess. As mentioned previously the gelation of the linear polymer can be separated into
two stages. One plateau is observable directly after G’ intersects with G’’, the other
one depending on the system at higher temperatures. P3 and P1 reach the second
plateau of gelation approximately 8 °C after the first plateau is reached. In both cases
G’ and G’’ rise simultaneously prior to reaching the gelation temperature. After reach-
ing the first plateau G’ stabilizes, while G’’ is rising with a lower slope than before. G’’
is increasing rapidly just prior to the second plateau, before it than stabilized together
with G’ after reaching the second plateau. This shows that the viscous properties of
the systems increase strongly when undergoing the second phase shift. This could be
4. Results & Discussion
58
due to breaking up of the structure that was formed after the first plateau, before form-
ing the structure after the second plateau. P4, on the other hand, shows a different
behaviour in gelation. Until 15.4 °C both, G’ and G’’, show a low decrease, which is
normally observed before a gelation occurs. Between 15.4 °C and 21.5 °C a steep
increase of G’ is observed, while G’’ remains constant. The elastic properties of the
system increase steeply, while the system has still overall viscous character. This could
indicate assembly of the star-shaped polymers in the solutions into bigger, more rigid
structures. The increase of G’ slows down, but is still present, until 32.7 °C, where now
G’ and G’’ are increasing rapidly until G’ overcomes G’’ and a plateau at 34.8 °C is
reached. This second, rapid increase of both moduli, that eventually leads to the gela-
tion of the system at 34.4 °C, could indicate an assembly of the constructs built after
the first increase of G’, when the system was still overall viscous.
To sum up, P1, P3 and P4 all reach their maximum G’ via an intermediate plateau.
However, the hydrogel formation process of P4 seems to be more complex than the
other polymers, given the different gelation temperature, G’ and curve trends.
For the use in bioprinting, either as hydrogels or blended with other compounds, the
star-shaped polymers were also assessed to compare their properties regarding re-
covery after high shear stress. Therefore, the hydrogel forming samples were exposed
to a total of three cycles of alternating high and low shear stress and compared to a
20 wt.% solution of P1 (Figure 30).
4. Results & Discussion
59
Figure 30: Recovery experiments shear stresses of 0.5% in the relaxed state and 150% in the
stressed state for the polymer solutions that showed thermogelation. This was the case for a 30 wt.%
solution of P3 (A) and a 40 wt.% solution of P3 (B) and P4 (C). The graphs were compared to a
20 wt.% hydrogel of P1 (D) Graphs show G’ and G’’ of the system plotted against time. Samples were
measured at 37 °C.
In general, recovery of P3 is rapid after shear stress (150%) is released from the sys-
tem. This is the same for the linear system. A hydrogel of 30 wt.% of P3 also shows
two step recovery behaviour as described for P1, however, more pronounced. For P1
it was hypothesized that this could be due to a ratio of the polymer blocks that is not
the ideal 1:1 ratio. This unsymmetrical ratio could trigger a delayed network forming
after high shear stress. Since there is also such a ratio present in the theoretical com-
position of P3 the two-step recovery could be also due to this. It is notable that the
recovery profile changes with higher polymer content. At 40 wt.% the recovery hap-
pens in one step and rapidly after the shear stress was released, which is comparable
to the recovery behaviour described by Lorson et al. All P3 hydrogels recover after
undergoing all cycles back to the initial G’ value [5120 Pa (30 wt.%), 9440 Pa
(40 wt.%)], indicating a full recovery of the hydrogel properties, like it was observed in
4. Results & Discussion
60
the linear system. P4, on the other hand, shows a different recovery behaviour. G’ does
not recover to the initial value (816 Pa) anymore. Like seen in the frequency sweep of
P4, G’ and G’’ are closer together than the two moduli in P3 or P1 hydrogels. This high
tan δ value could also be responsible for the recovery profile of the P4 hydrogel. The
trend of G’ and G’’ in the hydrogel of 30 wt.% P3 and 20 wt.% P1 is similar, probably
due to already discussed reasons. On the other hand, G’’ does not change at all in the
40 wt.% hydrogel of P3. That means, that the viscous character of the hydrogel stays
the same, regardless of the shear stress that is applied. Still, a gel-sol transition exists
at high shear rates. Therefore, G’ in this system seems more dependent on shear
stress than G’’.
Finally, it should be mentioned that all samples show good shear thinning properties.
G’ in all systems, as soon as high shear stress is applied, dropped instantly under the
value of G’’. This fast transition from elastic to viscous character of the material is an
important property for hydrogels that are used as bioinks. The gold standard in this
case is Pluronic® F127, as its aqueous solution shows excellent shear thinning and
recovery properties.[153]
Figure 31: Dynamic viscosity of P3, P4 and P5 between 10 and 50 °C at varying concentrations (1, 5,
10, 20, 30 and 40 wt.%) and Pluronic® F127 at 10 wt.%.
4. Results & Discussion
61
In addition to the rheology measurements, viscosity measurements of a concentration
series in the range between 1 wt.% and 40 wt.% of the polymers P3-P5 were con-
ducted (Figure 31). The results were compared to a 10 wt. % solution of the previously
mentioned Pluronic® F127. A concentration, where it does not show thermogelation
properties yet.
P3 and P4 show decreasing dynamic viscosity with increasing temperatures at con-
centration below 10 wt.%. At 10 wt.% and above, dynamic viscosities of P3 start to
increase above 16.5 °C. Lorson et al.[10] suggested that the increase in viscosity oc-
curs around the cloud point of the polymer solution. Therefore, this could be used to
obtain information about the cloud point of the synthesized polymers. Above 20 wt.%,
no data of P3 and P4 could be obtained, since the viscosity of the material was too
high for a falling sphere viscometer. Even though all solutions of P4, except for the
1 wt.% solution at 20 °C, are clear solutions, show however an increase in viscosity at
approximately 19 °C. This could indicate the formation of a secondary structure.
P3 shows an increase in viscosity at its gelation temperature, found in the rheology
experiments. However, the gelation temperature for P4 is much higher than this in-
crease in viscosity. This could indicate a preassembly of polymer in the solution prior
to the actual gelation. This would follow the found rapid rise of G’ in the temperature
sweep between 15.4 and 21.5 °C at constant G’’. P5 also shows a weak rise in viscosity
at around 22 °C for concentrations of 20 and 30 wt.%. When measuring a 5 wt.% so-
lution of P5 a phenomenon could be observed, where the viscosity experiences a rapid
rise between 15 and 20 °C, followed by a slow decrease of the viscosity with increasing
temperatures. This behaviour could be observed over several measurements and by
using different capillary diameters. When taking a closer look at the polymer in the
measurement, it is revealed that the polymer starts precipitating in the temperature
range where the rapid rise in viscosity was observed. After that, the polymer starts to
dissolve again, until full dissolution is achieved at 50 °C. It is hypothesized that the
polymer reaches its LCST between 15 and 20 °C and forms aggregates that precipi-
tate. By increasing the temperature further, the UCST is reached, resulting in dissolu-
tion of the polymer. It is notable that this behaviour only occurred with the polymer that
does not form a hydrogel and more importantly, only at a concentration of 5 wt.%. It
can be suggested that due to the low number of repeating units on each arm, larger
aggregates can be formed at this concentration by interaction with other P5 polymers.
4. Results & Discussion
62
Future experiments can analyse the aggregates further by SANS or dynamic light scat-
tering (DLS) experiments.
Compared to a 10 wt.% solution of Pluronic® F127 all measured polymers show a
similar viscosity at the same concentration. The solution of F127 also shows an in-
crease in viscosity that was observed in the star-shaped polymers P3-P5. F127, there-
fore, shows similarities to P3, which is the only measured compound that shows a
temperature dependent increase of viscosity at 10 wt.%. However, P3 shows the in-
crease 7.5 °C earlier than the F127 solution (22.5 °C). This could indicate a difference
in gelation process, since this temperature difference was already seen for the linear
diblock copolymer system of PMeOx and PnPrOzi.[10]
4. Results & Discussion
63
4.3.5 Conclusion
The synthesis of a linear diblock copolymer of PMeOx and PnPrOzi could be repro-
duced. Hydrogel properties of a former batch of the same polymer type (P2) could be
achieved after eliminating impurities, which originated from the polymerization process.
For purification the polymer was dialysed against Millipore water using a dialysis tube
with a MWCO of half the theoretical polymer weight (10 kDa). Through rheology meas-
urements it could be verified that the linear copolymer P1 shows hydrogel properties
at 20 wt.%, which correspond to P2. Recovery of G’ of P1, after high mechanical stress
was applied, happened over two steps, in contrast to P2, which recovers directly to its
initial G’ values. It was explained that this is due to an imperfect monomer ratio, which
delays the network rebuilding of the polymer system.
To modify hydrogel properties of P1, it is planned to blend other architectures of the
same copolymer into a solution of P1. Therefore, synthesis of star-shaped diblock co-
polymers of PMeOx and PnPrOzi was planned. The initiator synthesis revealed prob-
lems with identifying the degree of functionalization. A difference in melting points of
pentaerythritol and the functionalized version was used in addition to previous kinetic
studies to ensure full initiator functionalization.
Three star-shaped copolymers of PMeOx and PnPrOzi P3-P5 could be synthesized
and characterized using viscometry and rheology. Approaches to control the polymer-
ization by decreasing the temperature during the polymerization and reducing the mon-
omer concentration in the reaction mixture, after all polymers showed the same bimo-
dality, failed to reduce the bimodality due to irreproducibility or affected the monomer
conversion. Thermogelation at 30 wt.% (P3) and 40 wt.% (P3, P4) could be observed
at 37 °C. Turbid (P3) and transparent hydrogels (P4) could be observed. Temperature
sweeps of the hydrogel forming samples indicated a two-step gelation of all samples.
Gelation temperatures of P3 (18.3 °C) are similar to P1 (19.4 °C), while P4 forms a
hydrogel at 34.4 °C. P3 hydrogels showed good shear thinning and recovery proper-
ties, while the P4 hydrogel has weak recovery properties.
Viscosity analysis of a concentration range between 1 and 40 wt.% of P3 and P4
showed similar behaviour for low concentrations until 5 wt.%, while P5 showed two
phase changes in the temperature range of a measurement at this concentration.
Higher concentrations of polymer showed an increase of viscosity, which should be
around the cloud point of the system. In comparison to a 10 wt.% solution of an
4. Results & Discussion
64
established standard in biofabrication, Pluronic® F127, solutions of P3-P5 have similar
viscosities.
To sum up, linear and star-shaped copolymers consisting of PMeOx and PnPrOzi
could be synthesized. Through hydrogel formation, rapid recovery and shear thinning
properties a base for further experiments could be built.
4.4 Architectural Influence on Rheological Properties
4.4.1 Rheological Assessment of Polymer Blends with Different Architectures
The hydrogel properties of P1 make is a suitable candidate for the process of bioprint-
ing in biofabrication. However, with different cell types and printers that are used, a
bioink must cover a range of properties, that can be adapted swiftly and without much
synthetical issue. To realize this for the bioink candidate P1 it was hypothesized that
by blending another polymer architecture into the linear system it would be possible to
modify hydrogel properties like G’ or shear thinning behaviour. This approach would
avoid the use of other, unrelated, molecules in the system, which would potentially
counteract the gelation of the system, that is built up by the hydrophobic interactions
the PnPrOzi block. Therefore, the synthesized star-shaped copolymers of PMeOx and
PnPrOzi were blended with solutions of the linear equivalent polymer. To have a better
comparison to the hydrogel of P1 at 20 wt.%, all mixtures (Table 13) were calculated
to have a total polymer concentration of 20 wt.%.
Table 13: Overview over all polymer blends made by blending P1 with P3, P4 or P5 to a total solid
content of 20 wt. %.
17.5 wt.% P5 + 2.5 wt.% P3 15 wt.% P5 + 5 wt.% P3
17.5 wt.% P5 + 2.5 wt.% P4 15 wt.% P5 + 5 wt.% P4
17.5 wt.% P5 + 2.5 wt.% P5 15 wt.% P5 + 5 wt.% P5
20 wt.% P1
Now a plan for rheological measurement had be developed. Therefore, the extrusion
process of a bioink in the bioprinting process is observed in more detail (Figure 32).
4. Results & Discussion
65
Figure 32: Schematic illustration of rheological measurements that can be conducted to simulate the
extrusion process in bioprinting. Essentially it consists of three processes: flow initiation, shear thin-
ning and post-printing recovery.
The extrusion can be categorised into three processes. First there is the flow initiation,
which is characterized by the flow point a material has. The second process is the
shear thinning of the bioink, while it is extruded and last the printed ink must recover
as soon as possible after the extrusion, to ensure a proper fibre building on the collector
plate. All these different processes can be simulated by rheology measurements (Fig-ure 32). This gives the opportunity of acquiring qualitative and quantitative data of the
bioink, before buying an expensive printing system. For assessing the previously
named blends of linear and shar-shaped polymers, the focus of the measurements
conducted was set on the shear thinning properties of the material and the post-print
recovery, since the printed result is highly dependent on these parameters. Addition-
ally, an amplitude sweep, to determine the LVE range of the materials, was conducted
(Appendix I: Figure 8). Samples were assessed using a frequency sweep in the LVE
region (Figure 33).
4. Results & Discussion
66
Figure 33: Frequency sweeps with angular frequencies from 0.01 to 100 rad/s. Samples were the pol-
ymer blends of P1/P3 (A, B), P1/P4 (C, D) and P1/P5 (E, F) in comparison to a 20 wt.% hydrogel of
P1 (G) and a 20 wt.% solution of Pluronic ® F127 (H). The graphs show G’ and G’’ plotted against the
angular frequency. Samples were measured at 35 °C.
The effect on the blending of two different architectures of the same polymers are quite
divers. Almost all samples formed a hydrogel, which indicates a good implementation
of the star-shaped polymers into the linear system. Polymer blends, including P3 (A,
B), showed a constant trend of G’ and G’’ and, therefore, a hydrogel strength that is
4. Results & Discussion
67
independent of the applied frequency. G’ did not increase compared to the pure linear
system. Small variations that can be seen in G’’ at 2.5 wt.% P3 (A) disappear at 5 wt.%
(B). This indicated a stabilization of the network built up by P1, since the reference
sample of P1 (G) also shows variations of G’’, similar to A. Replacement of 2.5 wt.%
of P1 by P4 (C) shows no change in hydrogel strength or trend of G’ and G’’ compared
to reference G. However, when replacing 5 wt.% of P1 by P4 (D) the hydrogel shows
a weaker hydrogel strength of almost 100 Pa (tan δ = 1.05) at 0.01 rad/s, instead of
the normally achieved values around 2-3 kPa. G’ increases with higher frequencies.
This behaviour is similar to the one that is observed, when measuring a 20 wt.% sam-
ple of Pluronic® F127 (H). At a frequency of 0.01 rad/s, the material shows a G’ of
5510 Pa (tan δ = 0.87), while at a higher frequency of 100 rad/s a G’ of 21500 Pa
(tan δ = 0.02) can be achieved. This indicates that the structure stability of the polymer
system depends on the mechanical stress. Since the replacement of 5 wt.% by P4 is
the only sample that shows this behaviour it can be hypothesized that the system that
is built up by P4 and P1 molecules is different from a pure P1 sample.
A sample of 2.5 wt.% P5 (E) showed no significant difference in G’ or G’’ to the refer-
ence sample (G). As soon as this polymer amount is doubled (F) the sample has a
more dominant viscous character at very high and very low frequencies, since G’’ has
a higher value then G’. In the range of frequencies between 0.03 and 6.8 rad/s the
sample shows a more elastic character, which indicated existence of a hydrogel. How-
ever, the strength is in that region not above 50 Pa, which cannot be considered a
usable sample for bioprinting anymore. High concentrations of P5 seem to destroy the
network formed by the P1 chains.
It seems like replacements of 2.5 wt.% of the linear P1 by star-shaped polymers is not
affecting the hydrogel strength of the system. However, for the replacement of 5 wt.%
by P4 and P5, samples show a significant change in the G’, also introducing a fre-
quency dependency into the system. Bioinks should, to be printed accurately and re-
producibly, have a constant G’ value in a range of frequencies. P3 is the only star-
shaped polymer that even stabilizes G’’ of the hydrogel system more than in the refer-
ence P1. Even though hydrogel strength itself could not be increased, this shows the
potential application of P3 as an additive in P1 systems. The only drawback of this
polymer is that samples get slightly turbid after the addition of P3. This correlates with
the behaviour of pure P3 solutions. There might be an issue in regard of imaging of
4. Results & Discussion
68
e.g. cells inside the hydrogels, however, as a sacrificial bioink, which is to be washed
out over time or degrades, this system could still be used if the other measurements
show similar good properties of the P1/P3 system.
To now evaluate how the blends behave in the extrusion process, the shear thinning
properties were assessed. Therefore, the viscosity of the materials was plotted against
the applied shear rate (Figure 34).
Figure 34: Graph showing the shear rate dependency of the viscosity. Measured samples were the
polymer blends of P1/P3, P1/P4 and P1/P5 in comparison to a 20 wt.% hydrogel of P1 and a 20 wt.%
solution of Pluronic ® F127. Samples were measured at 35 °C.
The trend for the viscosity of each blend was compared to a 20 wt.% solution of P1
and a 20 wt.% solution of Pluronic® F127, which both show excellent shear thinning
properties. Overall it can be seen that all samples, except for 5 wt.% P4 and P5 show
similar shear thinning properties like the reference. This result corresponds to the in-
formation that could be obtained from the frequency sweeps, where only the 5 wt.%
samples of P4 and P5 affected the hydrogel strength significantly. To get a better in-
sight in the shear thinning properties of each blend, the linear regions of the viscosity
curves were fitted by the following power law in Equation (1).[47]
4. Results & Discussion
69
𝜂 = 𝐾 ∙ ��𝑛−1 (1)
(with the viscosity η, the shear rate ��, the flow behaviour index n and the flow con-
sistency index K)
From this equation values for K and n can be determined (Table 14). In terms of shear
thinning properties of a material, especially n is an important value. Is 0<n<1 then the
assessed material has shear thinning properties, while n=1 indicates a Newtonian
fluid.[47] Therefore, the lower n, the better the shear thinning properties of the material.
Table 14: K and n values derived from the power law fit of the viscosity-shear rate plot. Values were
calculated for polymer blends of P1/P3, P1/P4 and P1/P5 in comparison to a 20 wt.% hydrogel of P1
and a 20 wt.% solution of Pluronic ® F127.
Sample n [-] K [𝐏𝐚 ∙ 𝐬𝐧]
17.5 wt.% P1 + 2.5 wt.% P3 0.26 119.82
17.5 wt.% P1 + 2.5 wt.% P4 0.31 94.26
17.5 wt.% P1 + 2.5 wt.% P5 0.35 102.51
15 wt.% P1 + 5 wt.% P3 0.27 117.95
15 wt.% P1 + 5 wt.% P4 0.45 44.80
15 wt.% P1 + 5 wt.% P5 0.76 8.63
20 wt.% P1 0.27 87.10
20 wt.% F127 0.22 103.89
K and n values for the reference F127 could be compared to the ones found by Paxton
et al. It could be shown that their n value (0.12) is approximately half of the determined
value in this thesis and K (222 Pa ∙ sn) is approximately twice as high. However, K and
n values are still in the same range, indicating superb shear thinning of the hydrogel.
The values for n underline what was found before for the different blends. The hydrogel
formed by only P1 polymers has shear thinning properties, which are similar to Plu-
ronic® F127 at the same concentration. This shear thinning is almost not influenced
by the addition of 2.5 wt.% P3, P4 or P5, which can be seen in the similarity of the n
values. However, there is a small trend of the values observable that showed that the
shear thinning properties at this percentage already slightly decrease, the smaller the
arm length of the star. This trend gets more significant, once 5 wt.% of P1 is replaced.
While P3 does not alter the n value, and, therefore, shear thinning properties of the
4. Results & Discussion
70
system, P4 and P5 seem to destroy the supramolecular structure of the polymer and
with it the shear thinning properties.
The K values are not as easily accessible for the understanding of the material prop-
erties. Nevertheless, the K values can be used in combination with the acquired n val-
ues to model a so called bioprinting window, using a mathematical model recently es-
tablished by Paxton et al.[47]. The bioprinting window is a combination of different pa-
rameters that are important for the extrusion, namely needle diameter, needle length
and pressure, which enable a fibre formation on the collector plate. To achieve this,
the following Equation (2) was deducted by Paxton et al.[47]:
�� = (−∆𝑝2𝐿𝐾
) ( 𝑛3𝑛+1
)𝑅𝑛+1𝑛 (2)
(With the average extrusion velocity ��, the applied printing pressure ∆𝑝, the needle
length L, the flow behaviour index n and the flow consistency index K and the needle
radius R)
With this equation it is possible to determine an average extrusion velocity of the hy-
drogel in the syringe for different needle diameters, needle lengths and printing pres-
sures by inserting n and K values. This average velocity is then compared to experi-
mentally found data of collector velocities. If the average extrusion velocity is lower
than the velocity range of the collector speed, not enough material gets extruded onto
the collector and fibres get stretched. In case of a higher velocity than the collector
velocity, more material gets extruded and the fibres swell on the collector. With this
information a mapping of ideal extrusion parameters can be calculated, which can
serve as a guideline for experiments.
However, it is important to say that this model is purely based on the K and n values
of a material. It cannot give advice on how the fibres recover, and therefore about the
shape fidelity of printed constructs, or whether the material even forms a stable hydro-
gel. Nevertheless, to get preliminary data on extrusion of the polymer blends this model
was used, and mapping of all samples, including references was conducted for printing
pressures from 0-5 bar and needle diameters of 200, 250 and 330 µm (Figure 35). A
standard needle length was chosen as 12.22 mm (1/2 inch). These parameters are in
agreeance with the screening done by Paxton et al.[47].
4. Results & Discussion
71
Figure 35: Theoretically calculated bioprinting windows of the polymer blends of P1/P3, P1/P4 and
P1/P5 in comparison to a 20 wt.% hydrogel of P1 and a 20 wt.% solution of Pluronic ® F127. Blue
indicates that the extrusion velocity at the given parameters is too slow compared to the collector
speed, resulting in fibres getting stretched on the collector plate. Red indicates that the extrusion
velocity is too high at the given parameters compared to the collector speed, resulting in a swelling of
the fibres on the collector plate. White shows optimal printing parameters, since the extrusion velocity
is in the range of experimentally achievable collector speeds. For the calculations a needle length of
12.22 mm (1/2 inch) was assumed.
The both reference samples (G, H) seem to have narrow bioprinting windows. With the
addition of star-shaped polymers, this window broadens in all cases A-F. Interestingly
the bioprinting window seems to be the largest, which means that the ink is extrudable
at most parameters. This is the case, when P4 and P5 were included into the polymer
system at 5 wt.% (D and F). The previous experiments showed, that neither those
samples form a hydrogel with low tan δ, but also the systems lose shear thinning prop-
erties. This shows perfectly the limitations of the model, since D and F are not printable,
given their high tan δ values.
In combination with the previous experiments it can be said that P3 is the most prom-
ising candidate for a blend with P1, since it does not only not affect shear thinning or
G’, but also seems to make the hydrogel extrudable at a broader range of parameter
4. Results & Discussion
72
combinations. P4 and P5 on the other hand, open the bioprinting window significantly,
but show a decrease in hydrogel quality when mixed with P1.
The same parameters also allow to calculate a flow profile of the hydrogel in the nee-
dle, hence, giving information about cell viability in the printing process. Flow profiles
were calculated with Equation (3).
𝑣 = 𝑛𝑛+1
( ∆𝑝2𝐿𝐾
)1𝑛 (𝑅
𝑛+1𝑛 − 𝑟
𝑛+1𝑛 ) (3)
(With the extrusion velocity v, the applied printing pressure ∆𝑝, the needle length L, the
flow behaviour index n and the flow consistency index K, the incremental needle radius
r and the total needle radius R)
By plotting the extrusion velocity against the increment of the needle radius, flow pro-
files could be plotted (Figure 36).
4. Results & Discussion
73
Figure 36: Calculated theoretical extrusion velocities and shear rates as a function of the distance
from the needle centre. Samples were the polymer blends of P1/P3 (A, B), P1/P4 (C, D) and P1/P5
(E, F) in comparison to a 20 wt.% hydrogel of P1 (G) and a 20 wt.% solution of Pluronic ® F127 (H).
For the calculations parameters that were within the common theoretical bioprinting window of all
samples were chosen: 1 bar extrusion pressure, 12.22 mm (1/2 inch) needle length, 250 µm needle
diameter.
4. Results & Discussion
74
All modelling of the velocity profile was conducted with parameters that were within the
theoretical bioprinting window of every sample. The chosen parameters were a pres-
sure of 1 bar, a needle length of 12.22 mm (1/2 inch) and a needle diameter of 250
µm. The radius r from 0 to 125 µm corresponds to the way from the needle centre to
the inner wall of the needle. A hydrogel of P1 (G) shows very even velocity profile from
the centre of the needle approximately 45 µm from the needle wall. That corresponds
well to the flow profile of the reference Pluronic® F127 (H). An even velocity profile
means that cell in the hydrogel would experience no velocity gradient, which would
induce mechanical stress on the cells that could potentially lead to damage of the
cells.[154] Cells in the P1 hydrogel experience a much lower velocity of only
1200 mm/min compared to 2000 mm/min in F127. Less velocity also means less ve-
locity gradients and therefore, less damage for the cells in the printing process. The
velocity decreases further with a factor of three as soon as P3 or P4 are mixed into the
system. When 2.5 wt.% of P1 get replaced by P3 or P4 (A, C) the velocity of extrusion
in the needle centre is only around 400 mm/min. Without changing the velocity profile,
compared to F127 (H) or P1 (G), this shows that cells would be extruded slower
through the needle at the same velocity profile. An implementation of 5 wt.% of P3 (B)
does neither change the velocity profile towards the needle wall, not the extrusion ve-
locity, which follows the results that were already found for P1/P3 blends. The same
can be observed for the blend with 5 wt.% of P4 (D). Here the velocity in the needle
diameter starts to decrease earlier than in the references, which causes a velocity gra-
dient that starts directly at the needle centre. Therefore, cells in the ink would experi-
ence stress in any point in the needle, which could cause cell damage and affect the
printing result. Blends with P5 seems to fit into this scheme, since the blend with
2.5 wt. % P5 (E) does not seem to be affected in its velocity profile, while a 5 wt.%
solution (F) shows a strong velocity gradient at each point of the needle. Furthermore,
it seems like a higher concentration of P5 is causing an increase in velocity at the
needle centre from only 200 mm/min to 700 mm/min, the highest velocity within the
blends. This could be due to the mainly Newtonian character of the solution with high
P5 content.
It is also noteworthy that the shear rate that cells experience within the extrusion pro-
cess could be decreased significantly by blending in the star-shaped polymers. All
blends show a shear rate of around 200 1/s, compared to the P1 system with 700 1/s
and the gold standard in bioprinting Pluronic® F127 with 1300 1/s. These are the shear
4. Results & Discussion
75
rates that cells experience at the needle wall, where the velocity gradient is the highest.
Therefore, if there is a lower shear rate cells are more likely to survive the extrusion
process.
Another diagram, described by Paxton et al., could be obtained. By using the previ-
ously calculated data, it was also possible to calculate residence time of the cells in
the needle. It is desirable that the cells do not spend longer times under shear rates in
the needle, since the longer the cells are exposed to it, the higher the risk of cell dam-
age. Therefore, residence time profiles were calculated (Figure 37).
4. Results & Discussion
76
Figure 37: Calculated theoretical residence times of materials/cells in the needle as a function of the
distance from the needle centre. Samples were the polymer blends of P1/P3 (A, B), P1/P4 (C, D) and
P1/P5 (E, F) in comparison to a 20 wt.% hydrogel of P1 (G) and a 20 wt.% solution of Pluronic ® F127
(H). For the calculations parameters that were within the common theoretical bioprinting window of all
samples were chosen: 1 bar extrusion pressure, 12.22 mm (1/2 inch) needle length, 250 µm needle
diameter.
Cells that are implemented into a system of P1 (G) or F127 (H), independently of their
position in the needle, are exiting the needle tip almost immediately. This also applies
to cells, encapsulated in the blends, if the cells are located around 20 µm away from
4. Results & Discussion
77
the needle wall. In this region the residence time is almost constant and not higher
than four seconds at a time. However, compared to the references G and H, if cells
are close to the needle wall, this residence time increases. While P1/P3 and P1/P4
blends (A, B and C, D) show a residence time of cells at the needle wall of 20s, this
value increased to 40 s, in a P1/P5 blend where 2.5 wt.% of P5 (E) are present. This
time decreases again with a more Newtonian character of the fluid at 5 wt.% (F).
Concluding, these simulations could show that blending the P1 with P3-P5 could in-
crease the window of parameters, in which the material can be extruded successfully
onto a collector plate. Furthermore, extrusion velocities could be reduces compared to
P1, with only minor influence on the residence time of cells in the needle. P1/P3 blends
could prove a stable flow profile, also at higher concentrations of P3, which was not
possible to obtain by blending P4 or P5 into P1 systems. Furthermore, the blending
with P3 does not seem to affect the shear thinning properties of the material, even in
concentrations of 5 wt.%, while P4 and P5 blends with P1 did lower the shear thinning
properties significantly at higher concentrations in the blends.
Finally, the recovery of the blends was assessed. Recovery of a bioink after extrusion
is an essential factor for the formation of hierarchical structures. If the recovery is not
instantly after the extrusion of the ink, then fibres and fibre layers will fuse together, not
allowing the building of a hierarchical construct. To assess the recovery in the different
blends, viscosity was measured over a total of five cycles of high (100 1/s) and low
(0.1 1/s) shear rates (Figure 38).
4. Results & Discussion
78
Figure 38: Recovery tests of the polymer blends P1/P3 (A), P1/P4 (B) and P1/P5 (C). As shear rate
for the relaxed state 0.1 1/s, for the tension state 100 1/s were chosen. A total of five recovery cycles
was used to ensure reproducibility of the results. Viscosity is plotted against the experimental time.
Samples were measured at 35 °C.
As much as P3 did not influence the shear thinning properties of P1 hydrogels, it also
did not influence recovery of the system. The recovery is instantly after the shear rate
is decreased and directly back to the initial viscosity of the relaxed state
(2.5 wt.%: 774 Pa ∙ s; 5 wt.%: 501 Pa ∙ s). In the previous section, blends of P4 and
P1, showed a loss in shear thinning properties at high concentrations. This can be also
observed here, since the viscosity at high shear rate decreases to approximately
4 Pa ∙ s, independent of the viscosity of the system in relaxed state (2.5 wt.% P4:
456 Pa ∙ s; 5 wt.% P4: 162 Pa ∙ s). This effect increases in P5 blends, where viscosity
does only decrease minorly at high polymer contents of P5, in addition to a non-com-
plete recovery to the initial state of the material.
4.4.2 Conclusion
After assessment of the polymer blends of P1 with P3, P4 and P5 it can be concluded
that the blends with P3 are the most suitable ones, in terms of properties, that are
important in the printing process. Even at high concentrations of 5 wt.%, which would
4. Results & Discussion
79
normally be considered as impurities in the system, the shear thinning, the extrusion
velocity profile, the residence time of cell in the needle and the recovery properties of
the material stay constant. In addition to that the window of parameters, with which the
material can be printed is extended, given the possibility to print the hydrogel system
with different needles, pressures and even printing systems. Furthermore, the theoret-
ical velocity of cells in the needle is reduced, which could lead to higher cell survival in
the printing process, even with very vulnerable cell types. The drawback of having a
turbid hydrogel may be bad for direct cell imaging in the hydrogel, however, it does not
affect its usage as a degradable matrix, when modified with cleavable peptide se-
quences. P4 and P5 at high concentrations in the P1 system, resulted in a major loss
in shear thinning and G’, a constant theoretical velocity gradient in the needle, inducing
shear stress to printed cells, as well as reduced recovery properties of the hydrogel
system.
4.5 Assessment of Cytotoxicity of Star-Shaped Copolymers
4.5.1 WST-1 Assay
As recently shown[10], the linear system of P1 does not show cytotoxicity on NIH 3T3
fibroblasts, when incubated in concentrations up to 25 wt.% for 24 h. As the architec-
ture of the polymers P3, P4 and P5 was changed from the initial linear shape to a star-
like shape, it is possible that the cytotoxicity of these polymers changed compared to
P1. To evaluate cytotoxicity of the polymers, a WST-1 assay, as described by Lorson
et al.[10] was conducted, using the same cell type and incubation times to get compa-
rable results. Fibroblasts were incubated in a concentration series of the star-shaped
polymers (0.01, 0.1, 1, 5 and 10 wt.%) for 24 h before the WST-1 reagent was added
to the solution. By a colorimetric readout, after 4 h of incubation, using an UV/Vis spec-
trometer, relative cell viability as a function of the polymer concentrations was as-
sessed (Figure 39).
4. Results & Discussion
80
Figure 39: Relative cell viabilities of NIH 3T3 fibroblasts incubated with P3-P5 for 24 h at concentra-
tions of 0.01, 0.1, 1, 5 and 10 wt.%. Data was acquired through WST-1 assay after 4 h of incubation.
Relative cell viabilities of over 100 % for most of the concentrations was observed.
These high values either indicate the case that the cells proliferated better than control
groups after the polymers were applied. Another explanation would be a high absorb-
ance of the polymers at the given wavelengths that were used to measure the colori-
metric assay. Samples were freed from the polymer solutions after 24 h incubation and
washed with PBS. However, it could be that there was residue polymer in the solution,
affecting the readout of the assay. On the other hand, no colour difference between
the different samples could be observed prior to incubating the samples with WST-1
reagent. The assay could have also been affected by contaminations in the cell culture.
Decreased viability of the control samples, which were not treated with polymer, would
result in higher relative viabilities of the polymer-treated samples.
Nevertheless, trends can be observed in the diagram. With less monomer units at each
arm of the polymers, the relative cell viability decreases. Also P5 is the only polymer
that decreases the relative cell viability in concentrations over 1 wt.% under 100%,
indicating cytotoxicity of the polymer at higher concentrations. P3 and P4 show a de-
crease in relative cell viability with higher concentration, even though these values do
not drop below 100%, even at high polymer concentrations of 10 wt.%.
4. Results & Discussion
81
4.5.2 Conclusion
In summary, it could be shown that the relative cell viability of cells that were incubated
with P5 polymer solution, show more cytotoxic behaviour, compared to cells incubated
with P3 and P4. Relative cell viabilities above 100 % were observed. Therefore, the
results can only be seen relative between the polymers. In the whole range of polymer
concentrations (0.01 to 10 wt.%), P3 with the most monomer units per arm shows
higher relative cell viability compared to P4 and P5. Therefore, it could be assumed
that less monomer units, result in higher cell toxicity of the polymers, relative to each
other.
Due to these complications with the protocol of the assay the data cannot be discussed
in comparison with results for the linear polymer.
4.6 Process Upscaling
4.6.1 Up-scaled Synthesis of nPrOzi
With the focus on commercializing polymer and monomer, the area of upscaling is an
important factor in overcoming the obstacle from flask size to industry size synthesis.
Synthesis of nPrOzi was conducted in a 5 L reactor with double coated oil bath around
the reactor for heating purpose. The synthesis was adapted by the procedure of Sinn-
well and Ritter[134]. 3-amino-1-propanol and butyronitrile were added into the reactor
and stirred for a total of 67.5 h at a constant temperature of 130 °C. The reactor was
equipped with a propeller mixer, ensuring a constant mixture of the solution. Further-
more, the temperature could be hold constantly at 130 °C for the time of the synthesis.
When distilling the crude product, a reaction swamp of approximately 1 L had to be left
in the reactor, given that the stirrer could not stir the mixture anymore. Despite that a
total yield of 45.24% could be achieved, which is only approximately 4% under the
yield described in literature. Therefore, it can be concluded that the upscaling of the
nPrOzi synthesis could be achieved without significant losses in yield.
5. Summary & Outlook
82
5. Summary & Outlook
In summary it could be shown that it is possible to synthesize a linear copolymer of
PMeOx and PnPrOzi P1. The hydrogel formed by P1 in aqueous solution above
20 wt.%, showed significantly weaker G’ compared to a former batch P2 of the poly-
mer. It was hypothesized that low molecular impurities could be responsible, which
was already known from previous synthesis of this polymer. After dialysing, mechanical
properties comparable to P2 could be achieved.
Furthermore, in this thesis three star-shaped copolymers of PMeOx and PnPrOzi P3-
P5 could be synthesized. All mass distributions of the polymers showed a bimodality
in GPC elugrams. It was assumed that this would be an issue caused by either the
degree of functionalization of the initiator I1 or the polymerization conditions. Through
a melting point analysis and comparison with previous studies on the same initiator
system, the initiator could be excluded as a potential reason for the bimodality. After
the polymerization was conducted at reduced monomer concentration and reduced
reaction temperature, without reproducible results, the standard polymerization
method was accepted. After characterization it could be seen that the two star-shaped
copolymers P3 and P4 form hydrogels at polymer contents of 30 or 40 wt. %. Hydro-
gels from P3 show similar shear thinning and recovery properties compared to the P1
system. P4 is lacking recovery properties and P5 is not forming hydrogels in the range
of 1 to 40 wt.%.
With the aim to modify the hydrogel properties of the linear P1 system, blends of a total
polymer content of 20 wt. % and 2.5 or 5 wt. % of star-shaped polymer P3-P5 were
analysed regarding their properties in the extrusion-based bioprinting process. It could
be shown that blends with P3 are the only ones enhancing the extrusion properties of
the hydrogel system, while blends with P4 and P5 decreased shear thinning and re-
covery properties of the hydrogel system.
WST-1 assay was used to evaluate the relative cell viability of P3-P5. It could be seen
that P5 showed relative to the other polymers a lower relative cell viability at concen-
trations over 1 wt.%. Data of this assay could also be used as a relative comparison
between the polymers, since relative cell viabilities at most polymer concentrations
exceeded 100%.
5. Summary & Outlook
83
All in all, this thesis could build the foundation for the adjustment of the P1 system by
blending with other architectures.
In future work the blended architectures can be extended toward structures like poly-
mer brushes and miktoarm polymers. Furthermore, the range of blending concentra-
tions can be varied to lower the amount of polymer that is necessary to form a hydrogel
with the same properties as P1. It is promising that a system of pure P3 forms hydro-
gels with a G’ of 10 kPa, which could indicate enhancement of the mechanical proper-
ties, of the linear system at high blending ratios. Before that, it is important to optimize
the polymerization of the multiarm polymers, to obtain monomodal copolymers, which
can be reproducibly synthesized.
In terms of the star-shaped polymers it would be also interesting to have a closer look
at the hydrogel structure that is built in these systems, since the process of gelation in
these systems seems to differ from the linear system, which forms a macromolecular
sponge-like system. SANS would be a possible way to do so. Furthermore, since the
star-shaped polymers should have different viscosities depending on their structure,
the bimodality could be further evaluated by using a viscosity detector at the GPC in-
strument.
The star-shaped diblock copolymers could also be used as covalent crosslinkers in the
sponge-like lattice of the linear system, when both sides are modified with suitable
binding sites, like thiol-ene or Diels-Alder binding sites. Since the star-shaped diblock
copolymers have four arms that can bind to the linear system a high degree of cross-
linking at low polymer concentrations could be achieved. Instead of using the star-
shaped polymers, also peptide sequences cleavable by the cells, like matrix metallo-
protease (MMP) peptides, could be used as crosslinking components, to introduce a
degradability into the system.
To further progress in the biological compatibility of the system cell adhesion peptide
sequences like RGD or IKVAV could be implemented into the system, also via the
same binding sites named for the crosslinking. To approach the printing of the materi-
als it is also necessary to find optimal printing parameters, which can then be compared
to the found theoretical bioprinting window of the blends. A final matter to investigate
would be the special viscosity behaviour of a 5 wt.% P5 solution by e.g. DLS.
5. Summary & Outlook
84
With this broad outlook, the polymer system that was assessed in this thesis has po-
tential to further improve and broaden the applicability of P1. This thesis could suc-
cessfully contribute, by showing that polymer blending can enhance the properties of
the hydrogel that are important in the bioprinting process.
Bibliography
85
Bibliography
[1] M.W. Tibbitt, K.S. Anseth, Hydrogels as extracellular matrix mimics for 3D cell
culture, Biotechnol. Bioeng. 103 (2009) 655–663. doi:10.1002/bit.22361.
[2] Y. Imamura, T. Mukohara, Y. Shimono, Y. Funakoshi, N. Chayahara, M.
Toyoda, N. Kiyota, S. Takao, S. Kono, T. Nakatsura, H. Minami, Comparison of
2D- and 3D-culture models as drug-testing platforms in breast cancer, Oncol.
Rep. 33 (2015) 1837–1843. doi:10.3892/or.2015.3767.
[3] J. Lange, F. Weil, C. Riegler, F. Groeber, S. Rebhan, S. Kurdyn, M. Alb, H.
Kneitz, G. Gelbrich, H. Walles, S. Mielke, Interactions of donor sources and
media influence the histo-morphological quality of full-thickness skin models,
Biotechnol. J. 11 (2016) 1352–1361. doi:10.1002/biot.201600360.
[4] H. Pageon, H. Zucchi, Z. Dai, D.R. Sell, C.M. Strauch, V.M. Monnier, D.
Asselineau, Biological Effects Induced by Specific Advanced Glycation End
Products in the Reconstructed Skin Model of Aging, Biores. Open Access. 4
(2015) 54–64. doi:10.1089/biores.2014.0053.
[5] S. Girardeau-Hubert, S. Teluob, H. Pageon, D. Asselineau, The reconstructed
skin model as a new tool for investigating in vitro dermal fillers: Increased
fibroblast activity by hyaluronic acid, Eur. J. Dermatology. 25 (2015) 312–322.
doi:10.1684/ejd.2015.2563.
[6] O.L.A. Harrysson, O. Cansizoglu, D.J. Marcellin-Little, D.R. Cormier, H.A.
West, Direct metal fabrication of titanium implants with tailored materials and
mechanical properties using electron beam melting technology, Mater. Sci.
Eng. C. 28 (2008) 366–373. doi:10.1016/j.msec.2007.04.022.
[7] A.A. Mueller, P. Paysan, R. Schumacher, H.F. Zeilhofer, B.I. Berg-Boerner, J.
Maurer, T. Vetter, E. Schkommodau, P. Juergens, K. Schwenzer-Zimmerer,
Missing facial parts computed by a morphable model and transferred directly to
a polyamide laser-sintered prosthesis: An innovation study, Br. J. Oral
Maxillofac. Surg. 49 (2011) e67–e71. doi:10.1016/j.bjoms.2011.02.007.
[8] A.L. Jardini, M.A. Larosa, R.M. Filho, C.A.D.C. Zavaglia, L.F. Bernardes, C.S.
Lambert, D.R. Calderoni, P. Kharmandayan, Cranial reconstruction: 3D
Bibliography
86
biomodel and custom-built implant created using additive manufacturing, J.
Cranio-Maxillofacial Surg. 42 (2014) 1877–1884.
doi:10.1016/j.jcms.2014.07.006.
[9] O.L.A. Harrysson, D.J. Marcellin-Little, T.J. Horn, Applications of Metal Additive
Manufacturing in Veterinary Orthopedic Surgery, Jom. 67 (2015) 647–654.
doi:10.1007/s11837-015-1295-x.
[10] T. Lorson, S. Jaksch, M.M. Lübtow, T. Jüngst, J. Groll, T. Lühmann, R.
Luxenhofer, A Thermogelling Supramolecular Hydrogel with Sponge-Like
Morphology as a Cytocompatible Bioink, Biomacromolecules. 18 (2017) 2161–
2171. doi:10.1021/acs.biomac.7b00481.
[11] F. Berthiaume, T.J. Maguire, M.L. Yarmush, Tissue Engineering and
Regenerative Medicine: History, Progress, and Challenges, Annu. Rev. Chem.
Biomol. Eng. 2 (2011) 403–430. doi:10.1146/annurev-chembioeng-061010-
114257.
[12] C.A. Vacanti, The history of tissue engineering, J. Cell. Mol. Med. 10 (2006)
569–576. doi:10.1111/j.1582-4934.2006.tb00421.x.
[13] C. Unger, N. Kramer, A. Walzl, M. Scherzer, M. Hengstschläger, H. Dolznig,
Modeling human carcinomas: Physiologically relevant 3D models to improve
anti-cancer drug development, Adv. Drug Deliv. Rev. 79 (2014) 50–67.
doi:10.1016/j.addr.2014.10.015.
[14] A. Kamb, What’s wrong with our cancer models?, Nat. Rev. Drug Discov. 4
(2005) 161–5. doi:10.1038/nrd1635.
[15] Y.C. Tung, A.Y. Hsiao, S.G. Allen, Y.S. Torisawa, M. Ho, S. Takayama, High-
throughput 3D spheroid culture and drug testing using a 384 hanging drop
array, Analyst. 136 (2011) 473–478. doi:10.1039/c0an00609b.
[16] C. Fischbach, R. Chen, T. Matsumoto, T. Schmelzle, J.S. Brugge, P.J.
Polverini, D.J. Mooney, Engineering tumors with 3D scaffolds, Nat. Methods. 4
(2007) 855–860. doi:10.1038/nmeth1085.
[17] R.N. Hardwick, C. Viergever, A.E. Chen, D.G. Nguyen, 3D bioengineered
tissues: From advancements in in vitro safety to new horizons in disease
Bibliography
87
modeling, Clin. Pharmacol. Ther. 101 (2017) 453–457. doi:10.1002/cpt.569.
[18] J. Groll et al., Biofabrication : reappraising the definition of an evolving field,
Biofabrication. 8 (2016). doi:10.1088/1758-5090/8/1/013001.
[19] H. Martínez Ávila, S. Schwarz, N. Rotter, P. Gatenholm, 3D bioprinting of
human chondrocyte-laden nanocellulose hydrogels for patient-specific auricular
cartilage regeneration, Bioprinting. 1–2 (2016) 22–35.
doi:10.1016/j.bprint.2016.08.003.
[20] D.B. Kolesky, K.A. Homan, M.A. Skylar-Scott, J.A. Lewis, Three-dimensional
bioprinting of thick vascularized tissues, Proc. Natl. Acad. Sci. 113 (2016)
3179–3184. doi:10.1073/pnas.1521342113.
[21] V. Beachley, V. Kasyanov, A. Nagy-Mehesz, R. Norris, I. Ozolanta, M. Kalejs,
P. Stradins, L. Baptista, K. da Silva, J. Grainjero, X. Wen, V. Mironov, The
fusion of tissue spheroids attached to pre-stretched electrospun polyurethane
scaffolds, J. Tissue Eng. 5 (2014) 204173141455656.
doi:10.1177/2041731414556561.
[22] N. V. Mekhileri, K.S. Lim, G.C.J. Brown, I. Mutreja, B.S. Schon, G.J. Hooper,
T.B.F. Woodfield, Automated 3D bioassembly of micro-tissues for
biofabrication of hybrid tissue engineered constructs, Biofabrication. 10 (2018).
doi:10.1088/1758-5090/aa9ef1.
[23] L. Moroni, T. Boland, J.A. Burdick, C. De Maria, B. Derby, G. Forgacs, J. Groll,
Q. Li, J. Malda, V.A. Mironov, C. Mota, M. Nakamura, W. Shu, S. Takeuchi,
T.B.F. Woodfield, T. Xu, J.J. Yoo, G. Vozzi, Biofabrication: A Guide to
Technology and Terminology, Trends Biotechnol. 36 (2018) 384–402.
doi:10.1016/j.tibtech.2017.10.015.
[24] J.G. and T.B. and T.B. and J.A.B. and D.-W.C. and P.D.D. and B.D. and G.F.
and Q.L. and V.A.M. and L.M. and M.N. and W.S. and S.T. and G. Malda,
Biofabrication: reappraising the definition of an evolving field, Biofabrication. 8
(2016) 13001. http://stacks.iop.org/1758-5090/8/i=1/a=013001.
[25] E. Berry, J.M. Brown, M. Connell, C.M. Craven, N.D. Efford, a Radjenovic, M.
a Smith, Preliminary experience with medical applications of rapid prototyping
Bibliography
88
by selective laser sintering., Med. Eng. Phys. 19 (1997) 90–96.
doi:10.1016/S1350-4533(96)00039-2.
[26] J.M. Williams, A. Adewunmi, R.M. Schek, C.L. Flanagan, P.H. Krebsbach, S.E.
Feinberg, S.J. Hollister, S. Das, Bone tissue engineering using
polycaprolactone scaffolds fabricated via selective laser sintering, Biomaterials.
26 (2005) 4817–4827. doi:10.1016/j.biomaterials.2004.11.057.
[27] P. Bertrand, F. Bayle, C. Combe, P. Goeuriot, I. Smurov, Ceramic components
manufacturing by selective laser sintering, Appl. Surf. Sci. 254 (2007) 989–992.
doi:10.1016/j.apsusc.2007.08.085.
[28] J. Giannatsis, V. Dedoussis, Additive fabrication technologies applied to
medicine and health care: A review, Int. J. Adv. Manuf. Technol. 40 (2009)
116–127. doi:10.1007/s00170-007-1308-1.
[29] J.C. Bernhard, S. Isotani, T. Matsugasumi, V. Duddalwar, A.J. Hung, E. Suer,
E. Baco, R. Satkunasivam, H. Djaladat, C. Metcalfe, B. Hu, K. Wong, D. Park,
M. Nguyen, D. Hwang, S.T. Bazargani, A.L. de Castro Abreu, M. Aron, O.
Ukimura, I.S. Gill, Personalized 3D printed model of kidney and tumor
anatomy: a useful tool for patient education, World J. Urol. 34 (2016) 337–345.
doi:10.1007/s00345-015-1632-2.
[30] J.M. Llovet, M. Pavel, J. Rimola, M.A. Diaz, J. Colmenero, D. Saavedra-Perez,
C. Fondevila, C. Ayuso, J. Fuster, P. Ginès, J. Bruix, J.C. Garcia-Valdecasas,
Pilot study of living donor liver transplantation for patients with hepatocellular
carcinoma exceeding Milan Criteria (Barcelona Clinic Liver Cancer extended
criteria), Liver Transplant. 24 (2018) 369–379. doi:10.1002/lt.24977.
[31] L.K. Henderson, B.J. Nankivell, J.R. Chapman, Surveillance protocol kidney
transplant biopsies: Their evolving role in clinical practice, Am. J. Transplant.
11 (2011) 1570–1575. doi:10.1111/j.1600-6143.2011.03677.x.
[32] W.J. Kim, M. Kim, G.H. Kim, 3D-Printed Biomimetic Scaffold Simulating
Microfibril Muscle Structure, Adv. Funct. Mater. 28 (2018) 1–12.
doi:10.1002/adfm.201800405.
[33] A. Athirasala, A. Tahayeri, G. Thrivikraman, C.M. Franca, N. Monteiro, V. Tran,
Bibliography
89
J. Ferracane, L.E. Bertassoni, A dentin-derived hydrogel bioink for 3D
bioprinting of cell laden scaffolds for regenerative dentistry, Biofabrication. 10
(2018). doi:10.1088/1758-5090/aa9b4e.
[34] Z. Wang, S.J. Lee, H.J. Cheng, J.J. Yoo, A. Atala, 3D bioprinted functional and
contractile cardiac tissue constructs, Acta Biomater. 70 (2018) 48–56.
doi:10.1016/j.actbio.2018.02.007.
[35] H. Jian, M. Wang, S. Wang, A. Wang, S. Bai, 3D bioprinting for cell culture and
tissue fabrication, Bio-Design Manuf. 1 (2018) 45–61. doi:10.1007/s42242-018-
0006-1.
[36] S. Ahadian, A. Khademhosseini, A perspective on 3D bioprinting in tissue
regeneration, Bio-Design Manuf. 1 (2018) 157–160. doi:10.1007/s42242-018-
0020-3.
[37] D. Williams, P. Thayer, H. Martinez, E. Gatenholm, A. Khademhosseini, A
perspective on the physical, mechanical and biological specifications of bioinks
and the development of functional tissues in 3D bioprinting, Bioprinting. 9
(2018) 19–36. doi:10.1016/j.bprint.2018.02.003.
[38] I. Martin, D. Wendt, M. Heberer, The role of bioreactors in tissue engineering,
Trends Biotechnol. 22 (2004) 80–86. doi:10.1016/j.tibtech.2003.12.001.
[39] D. Buenger, F. Topuz, J. Groll, Hydrogels in sensing applications, Prog. Polym.
Sci. 37 (2012) 1678–1719. doi:10.1016/j.progpolymsci.2012.09.001.
[40] J. Malda, J. Visser, F.P. Melchels, T. Jüngst, W.E. Hennink, W.J.A. Dhert, J.
Groll, D.W. Hutmacher, 25th anniversary article: Engineering hydrogels for
biofabrication, Adv. Mater. 25 (2013) 5011–5028.
doi:10.1002/adma.201302042.
[41] J.H.Y. Chung, S. Naficy, Z. Yue, R. Kapsa, A. Quigley, S.E. Moulton, G.G.
Wallace, Bio-ink properties and printability for extrusion printing living cells,
Biomater. Sci. 1 (2013) 763–773. doi:10.1039/c3bm00012e.
[42] C. Xu, W. Chai, Y. Huang, R.R. Markwald, Scaffold-free inkjet printing of three-
dimensional zigzag cellular tubes, Biotechnol. Bioeng. 109 (2012) 3152–3160.
doi:10.1002/bit.24591.
Bibliography
90
[43] G. Gao, A.F. Schilling, K. Hubbell, T. Yonezawa, D. Truong, Y. Hong, G. Dai,
X. Cui, Improved properties of bone and cartilage tissue from 3D inkjet-
bioprinted human mesenchymal stem cells by simultaneous deposition and
photocrosslinking in PEG-GelMA, Biotechnol. Lett. 37 (2015) 2349–2355.
doi:10.1007/s10529-015-1921-2.
[44] F. Guillemot, A. Souquet, S. Catros, B. Guillotin, J. Lopez, M. Faucon, B.
Pippenger, R. Bareille, M. Rémy, S. Bellance, P. Chabassier, J.C. Fricain, J.
Amédée, High-throughput laser printing of cells and biomaterials for tissue
engineering, Acta Biomater. 6 (2010) 2494–2500.
doi:10.1016/j.actbio.2009.09.029.
[45] J.A. Barron, P. Wu, H.D. Ladouceur, B.R. Ringeisen, Biological laser printing: A
novel technique for creating heterogeneous 3-dimensional cell patterns,
Biomed. Microdevices. 6 (2004) 139–147.
doi:10.1023/B:BMMD.0000031751.67267.9f.
[46] K. Holzl, S. Lin, L. Tytgat, S. Van Vlierberghe, L. Gu, A. Ovsianikov, Bioink
properties before, during and after 3D bioprinting, Biofabrication. 8 (2016)
32002. doi:10.1088/1758-5090/8/3/032002.
[47] N. Paxton, W. Smolan, T. Böck, F. Melchels, J. Groll, T. Jungst, Proposal to
assess printability of bioinks for extrusion-based bioprinting and evaluation of
rheological properties governing bioprintability, Biofabrication. 9 (2017).
doi:10.1088/1758-5090/aa8dd8.
[48] A. Blaeser, D.F. Duarte Campos, U. Puster, W. Richtering, M.M. Stevens, H.
Fischer, Controlling Shear Stress in 3D Bioprinting is a Key Factor to Balance
Printing Resolution and Stem Cell Integrity, Adv. Healthc. Mater. 5 (2016) 326–
333. doi:10.1002/adhm.201500677.
[49] L. Ouyang, C.B. Highley, W. Sun, J.A. Burdick, A Generalizable Strategy for
the 3D Bioprinting of Hydrogels from Nonviscous Photo-crosslinkable Inks,
Adv. Mater. 29 (2017). doi:10.1002/adma.201604983.
[50] Z. Wu, X. Su, Y. Xu, B. Kong, W. Sun, S. Mi, Bioprinting three-dimensional cell-
laden tissue constructs with controllable degradation, Sci. Rep. 6 (2016) 1–10.
doi:10.1038/srep24474.
Bibliography
91
[51] C. Echalier, R. Levato, M.A. Mateos-Timoneda, O. Castaño, S. Déjean, X.
Garric, C. Pinese, D. Noël, E. Engel, J. Martinez, A. Mehdi, G. Subra, Modular
bioink for 3D printing of biocompatible hydrogels: sol-gel polymerization of
hybrid peptides and polymers, RSC Adv. 7 (2017) 12231–12235.
doi:10.1039/c6ra28540f.
[52] N. Huettner, T.R. Dargaville, A. Forget, Discovering Cell-Adhesion Peptides in
Tissue Engineering: Beyond RGD, Trends Biotechnol. 36 (2018).
doi:10.1016/j.tibtech.2018.01.008.
[53] J. Lee, C.H. Park, C.S. Kim, Microcylinder-laden gelatin-based bioink
engineered for 3D bioprinting, Mater. Lett. 233 (2018) 24–27.
doi:10.1016/j.matlet.2018.08.138.
[54] M. de Ruijter, A. Ribeiro, I. Dokter, M. Castilho, J. Malda, Simultaneous
Micropatterning of Fibrous Meshes and Bioinks for the Fabrication of Living
Tissue Constructs, Adv. Healthc. Mater. 1800418 (2018) 1800418.
doi:10.1002/adhm.201800418.
[55] C.L. Hedegaard, E.C. Collin, C. Redondo-Gómez, L.T.H. Nguyen, K.W. Ng,
A.A. Castrejón-Pita, J.R. Castrejón-Pita, A. Mata, Hydrodynamically Guided
Hierarchical Self-Assembly of Peptide–Protein Bioinks, Adv. Funct. Mater. 28
(2018) 1–13. doi:10.1002/adfm.201703716.
[56] T.G. and G.J.G. and J.S.C. and A.K.P.R. and Y.-J.S. and A.A. and J.J.Y. and
S.J. Lee, L. Ouyang, R. Yao, Y. Zhao, Optimization of gelatin–alginate
composite bioink printability using rheological parameters: a systematic
approach, Biofabrication. 10 (2018) 34106. doi:10.1088/1758-5090/aacdc7.
[57] H.W. Ooi, C. Mota, A. Tessa Ten Cate, A. Calore, L. Moroni, M.B. Baker, Thiol-
Ene Alginate Hydrogels as Versatile Bioinks for Bioprinting,
Biomacromolecules. 19 (2018) 3390–3400. doi:10.1021/acs.biomac.8b00696.
[58] Y.W. Koo, E.J. Choi, J.Y. Lee, H.J. Kim, G.H. Kim, S.H. Do, 3D printed cell-
laden collagen and hybrid scaffolds for in vivo articular cartilage tissue
regeneration, J. Ind. Eng. Chem. 66 (2018) 343–355.
doi:10.1016/j.jiec.2018.05.049.
Bibliography
92
[59] N. Law, B. Doney, H. Glover, Y. Qin, Z.M. Aman, T.B. Sercombe, L.J. Liew,
R.J. Dilley, B.J. Doyle, Characterisation of hyaluronic acid methylcellulose
hydrogels for 3D bioprinting, J. Mech. Behav. Biomed. Mater. 77 (2018) 389–
399. doi:10.1016/j.jmbbm.2017.09.031.
[60] M. Dash, F. Chiellini, R.M. Ottenbrite, E. Chiellini, Chitosan - A versatile semi-
synthetic polymer in biomedical applications, Prog. Polym. Sci. 36 (2011) 981–
1014. doi:10.1016/j.progpolymsci.2011.02.001.
[61] S. Das, F. Pati, Y.J. Choi, G. Rijal, J.H. Shim, S.W. Kim, A.R. Ray, D.W. Cho,
S. Ghosh, Bioprintable, cell-laden silk fibroin-gelatin hydrogel supporting
multilineage differentiation of stem cells for fabrication of three-dimensional
tissue constructs, Acta Biomater. 11 (2015) 233–246.
doi:10.1016/j.actbio.2014.09.023.
[62] O. Pillai, R. Panchagnula, Polymers in Drug Delivery Omathanu Pillai and
Ramesh Panchagnula, Curr. Opin. Chem. Biol. 5 (2001) 447–451.
doi:https://doi.org/10.1016/S1367-5931(00)00227-1.
[63] R. Langer, Tissue engineering: A new field and its challenges, Pharm. Res. 14
(1997) 840–841. doi:10.1023/A:1012131329148.
[64] R. Kurzrock, D.J. Stewart, Click chemistry, 3D-printing, and omics: the future of
drug development, Oncotarget. 7 (2016) 2155–2158.
doi:10.18632/oncotarget.6787.
[65] E. Gioffredi, M. Boffito, S. Calzone, S.M. Giannitelli, A. Rainer, M. Trombetta,
P. Mozetic, V. Chiono, Pluronic F127 Hydrogel Characterization and
Biofabrication in Cellularized Constructs for Tissue Engineering Applications,
Procedia CIRP. 49 (2016) 125–132. doi:10.1016/j.procir.2015.11.001.
[66] R. Murphy, D.P. Walsh, C.A. Hamilton, S.A. Cryan, M. In Het Panhuis, A.
Heise, Degradable 3D-Printed Hydrogels Based on Star-Shaped
Copolypeptides, Biomacromolecules. 19 (2018) 2691–2699.
doi:10.1021/acs.biomac.8b00299.
[67] N.E. Fedorovich, I. Swennen, J. Girones, L. Moroni, C.A. Van Blitterswijk, E.
Schacht, J. Alblas, W.J.A. Dhert, Evaluation of photocrosslinked lutrol hydrogel
Bibliography
93
for tissue printing applications, Biomacromolecules. 10 (2009) 1689–1696.
doi:10.1021/bm801463q.
[68] K.M. Park, S.Y. Lee, Y.K. Joung, J.S. Na, M.C. Lee, K.D. Park,
Thermosensitive chitosan-Pluronic hydrogel as an injectable cell delivery
carrier for cartilage regeneration, Acta Biomater. 5 (2009) 1956–1965.
doi:10.1016/j.actbio.2009.01.040.
[69] Y. Shachaf, M. Gonen-Wadmany, D. Seliktar, The biocompatibility of
Pluronic®F127 fibrinogen-based hydrogels, Biomaterials. 31 (2010) 2836–
2847. doi:10.1016/j.biomaterials.2009.12.050.
[70] S. Stichler, T. Bock, N.C. Paxton, S. Bertlein, R. Levato, V. Schill, W. Smolan,
J. Malda, J. Tessmar, T. Blunk, J. Groll, Double printing of hyaluronic acid /
poly(glycidol) hybrid hydrogels with poly(epsilon-caprolactone) for MSC
chondrogenesis., Biofabrication. (2017). doi:10.1088/1758-5090/aa8cb7.
[71] S. Ji, M. Guvendiren, Recent Advances in Bioink Design for 3D Bioprinting of
Tissues and Organs, Front. Bioeng. Biotechnol. 5 (2017) 1–8.
doi:10.3389/fbioe.2017.00023.
[72] A.C. Daly, S.E. Critchley, E.M. Rencsok, I.A.D. Mancini, V.H.M. Mouser, A.
Abbadessa, R. Levato, W.E. Hennink, T. Vermonden, D. Gawlitta, J. Malda,
Development of a thermosensitive HAMA- containing bio-ink for the fabrication
of composite cartilage repair constructs Fixation of Hydrogel Constructs for
Cartilage Repair in the Equine Model: A Challenging Issue Fixation of hydrogel
constructs for cartil, Biofabrication. 9 (2017). https://doi.org/10.1088/1758-
5090/aa6265.
[73] A. Skardal, J. Zhang, G.D. Prestwich, Bioprinting vessel-like constructs using
hyaluronan hydrogels crosslinked with tetrahedral polyethylene glycol
tetracrylates, Biomaterials. 31 (2010) 6173–6181.
doi:10.1016/j.biomaterials.2010.04.045.
[74] X. Yang, Z. Lu, H. Wu, W. Li, L. Zheng, J. Zhao, Collagen-alginate as bioink for
three-dimensional (3D) cell printing based cartilage tissue engineering, Mater.
Sci. Eng. C. 83 (2018) 195–201. doi:10.1016/j.msec.2017.09.002.
Bibliography
94
[75] Z. Du, N. Li, Y. Hua, Y. Shi, C. Bao, H. Zhang, Y. Yang, Q. Lin, L. Zhu,
Physiological pH-dependent gelation for 3D printing based on the phase
separation of gelatin and oxidized dextran, Chem. Commun. 53 (2017) 13023–
13026. doi:10.1039/c7cc08225h.
[76] R.M. Costa, S. Rauf, C.A.E. Hauser, Towards biologically relevant synthetic
designer matrices in 3D bioprinting for tissue engineering and regenerative
medicine, Curr. Opin. Biomed. Eng. 2 (2017) 90–98.
doi:10.1016/j.cobme.2017.05.001.
[77] H. Witte, W. Seeliger, Simple Synthesis of 2‐Substituted 2‐Oxazolines and 5,6‐
Dihydro‐4H‐1,3‐oxazines, Angew. Chemie Int. Ed. English. 11 (1972) 287–288.
doi:10.1002/anie.197202871.
[78] H. Vorbrüggen, K. Krolikiewicz, a Simple Synthesis of Oxazolines, Oxazines,
Thiazolines and 2-Substituted Benzoxazoles, Tetrahedron. 49 (1993) 9353–
9372.
[79] J. Hansel, S.O. Hogan, S. Lensky, A.R. Ritter, M.J. Miller, Oxazoline Formation
via a Palladium-catalyzed Cyclization: A Direct, Stereoselective Approach to
cis-5-Amino-2-cyclopenten-1-ol Derivatives, Tetrahedron Lett. 36 (1995) 2913–
2916.
[80] P. Zhou, J.E. Blubaum, C.T. Burns, N.R. Natale, The direct synthesis of 2-
Oxazolines from carboxylic esters using lanthanide chloride as catalyst,
Tetrahedron Lett. 38 (1997) 7019–7020. doi:10.1016/S0040-4039(97)01641-9.
[81] N.R. Natale, J.I. Mckenna, C. Niou, M. Borth, H. Hope, Metalation of
Isoxazolyloxazolines, a Facile Route to Functionally Complex Isoxazoles:
Utility, Scope, and Comparison to Dianion Methodology, J. Org. Chem. 50
(1985) 5660–5666. doi:10.1021/jo00350a047.
[82] T. Ohshima, T. Iwasaki, K. Mashima, Direct conversion of esters, lactones, and
carboxylic acids to oxazolines catalyzed by a tetranuclear zinc cluster, Chem.
Commun. (2006) 2711–2713. doi:10.1039/b605066b.
[83] M.C. Pirrung, L.N. Tumey, Oxazoline synthesis from hydroxyamides by resin
capture and ring-forming release, J. Comb. Chem. 2 (2000) 675–680.
Bibliography
95
doi:10.1021/cc000047g.
[84] C. Boyer, N.A. Corrigan, K. Jung, D. Nguyen, T.K. Nguyen, N.N.M. Adnan, S.
Oliver, S. Shanmugam, J. Yeow, Copper-mediated living radical polymerization
(atom transfer radical polymerization and copper(0) mediated polymerization):
From fundamentals to bioapplications, Chem. Rev. 116 (2016) 1803–1949.
doi:10.1021/acs.chemrev.5b00396.
[85] M.E. Kourti, E. Fega, M. Pitsikalis, Block copolymers based on 2-methyl- and
2-phenyl-oxazoline by metallocene-mediated cationic ring-opening
polymerization: Synthesis and characterization, Polym. Chem. 7 (2016) 2821–
2835. doi:10.1039/c6py00405a.
[86] A. Krieg, C. Weber, R. Hoogenboom, C.R. Becer, U.S. Schubert, Block
copolymers of poly(2-oxazoline)s and poly(meth)acrylates: A crossover
between cationic ring-opening polymerization (CROP) and reversible addition-
fragmentation chain transfer (RAFT), ACS Macro Lett. 1 (2012) 776–779.
doi:10.1021/mz300128p.
[87] M. Hartlieb, D. Pretzel, M. Wagner, S. Hoeppener, P. Bellstedt, M. Görlach, C.
Englert, K. Kempe, U.S. Schubert, Core cross-linked nanogels based on the
self-assembly of double hydrophilic poly(2-oxazoline) block copolymers, J.
Mater. Chem. B. 3 (2015) 1748–1759. doi:10.1039/c4tb02069c.
[88] J.F. Nawroth, J.R. McDaniel, A. Chilkoti, R. Jordan, R. Luxenhofer, Maleimide-
Functionalized Poly(2-Oxazoline)s and Their Conjugation to Elastin-Like
Polypeptides, Macromol. Biosci. 16 (2016) 322–333.
doi:10.1002/mabi.201500376.
[89] M.M. Lübtow, L. Keßler, A. Appelt-Menzel, T. Lorson, N. Gangloff, M. Kirsch, S.
Dahms, R. Luxenhofer, More Is Sometimes Less: Curcumin and Paclitaxel
Formulations Using Poly(2-oxazoline) and Poly(2-oxazine)-Based Amphiphiles
Bearing Linear and Branched C9 Side Chains, Macromol. Biosci. 1800155
(2018) 1800155. doi:10.1002/mabi.201800155.
[90] R. Hoogenboom, M.W.M. Fijten, G. Kickelbick, U.S. Schubert, Synthesis and
crystal structures of multifunctional tosylates as basis for star-shaped poly(2-
ethyl-2-oxazoline)s, Beilstein J. Org. Chem. 6 (2010) 773–783.
Bibliography
96
doi:10.3762/bjoc.6.96.
[91] V.R. De La Rosa, S. Tempelaar, P. Dubois, R. Hoogenboom, L. Mespouille,
Poly(2-ethyl-2-oxazoline)-block-polycarbonate block copolymers: From
improved end-group control in poly(2-oxazoline)s to chain extension with
aliphatic polycarbonate through a fully metal-free ring-opening polymerisation
process, Polym. Chem. 7 (2016) 1559–1568. doi:10.1039/c5py01913c.
[92] M. Litt, A. Levy, J. Herz, Polymerization of Cyclic Imino Ethers. X. Kinetics,
Chain Transfer, and Repolymerization, J. Macromol. Sci. Part A - Chem. 9
(1975) 703–727. doi:10.1080/00222337508065890.
[93] R. Hoogenboom, Poly(2-oxazoline)s: Alive and kicking, Macromol. Chem.
Phys. 208 (2007) 18–25. doi:10.1002/macp.200600558.
[94] J.S. Park, K. Kataoka, Precise control of lower critical solution temperature of
thermosensitive poly(2-isopropyl-2-oxazoline) via gradient copolymerization
with 2-ethyl-2-oxazoline as a hydrophilic comonomer, Macromolecules. 39
(2006) 6622–6630. doi:10.1021/ma0605548.
[95] F. Wiesbrock, R. Hoogenboom, M.A.M. Leenen, M.A.R. Meier, U.S. Schubert,
Investigation of the living cationic ring-opening polymerization of 2-methyl-, 2-
ethyl-, 2-nonyl-, and 2-phenyl-2-oxazoline in a single-mode microwave reactor,
Macromolecules. 38 (2005) 5025–5034. doi:10.1021/ma0474170.
[96] R. Hoogenboom, M.W.M. Fijten, U.S. Schubert, The Effect of Temperature on
the Living Cationic Polymerization of 2-Phenyl-2-oxazoline Explored Utilizing
an Automated Synthesizer, Macromol. Rapid Commun. 25 (2004) 339–343.
doi:10.1002/marc.200300233.
[97] R. Hoogenboom, M.W.M. Fijten, R.M. Paulus, H.M.L. Thijs, S. Hoeppener, G.
Kickelbick, U.S. Schubert, Accelerated pressure synthesis and characterization
of 2-oxazoline block copolymers, Polymer (Guildf). 47 (2006) 75–84.
doi:10.1016/j.polymer.2005.11.025.
[98] R. Hoogenboom, B.D. Monnery, Method for the preparation of uniform, High
molar mass cyclic imino ether polymers, WO2016008817 A1, 2017.
[99] A.I. Amirova, O. V. Golub, T.U. Kirila, A.B. Razina, A. V. Tenkovtsev, A.P.
Bibliography
97
Filippov, Influence of arm length and number on star-shaped poly(2-isopropyl-
2-oxazoline) aggregation in aqueous solutions near cloud point, Soft Mater. 14
(2016) 15–26. doi:10.1080/1539445X.2015.1098703.
[100] R. Luxenhofer, M. Bezen, R. Jordan, Kinetic Investigations on the
Polymerization of 2-Oxazolines Using Pluritriflate Initators, Macromol. Rapid
Commun. 29 (2008) 1509–1513. doi:10.1002/marc.200800293.
[101] N. Zhang, R. Luxenhofer, R. Jordan, Thermoresponsive poly(2-oxazoline)
molecular brushes by living ionic polymerization: Kinetic investigations of
pendant chain grafting and cloud point modulation by backbone and side chain
length variation, Macromol. Chem. Phys. 213 (2012) 973–981.
doi:10.1002/macp.201200015.
[102] A. Kowalczuk, J. Kronek, K. Bosowska, B. Trzebicka, A. Dworak, Star poly(2-
ethyl-2-oxazoline)s-synthesis and thermosensitivity, Polym. Int. 60 (2011)
1001–1009. doi:10.1002/pi.3103.
[103] A. V Tenkovtsev, A.I. Amirova, A.P. Filippov, Star‐shaped Poly (2‐alkyl‐2‐
oxazolines): Synthesis and Properties, Temp. Polym. Chem. Prop. Appl. (2018)
67–92.
[104] P.I. Freeman, J.S. Rowlinson, Lower critical points in polymer solutions,
Polymer (Guildf). 1 (1960) 20–26. doi:10.1016/0032-3861(60)90004-5.
[105] Q. Zhang, C. Weber, U.S. Schubert, R. Hoogenboom, Thermoresponsive
polymers with lower critical solution temperature: From fundamental aspects
and measuring techniques to recommended turbidimetry conditions, Mater.
Horizons. 4 (2017) 109–116. doi:10.1039/c7mh00016b.
[106] M. Glassner, M. Vergaelen, R. Hoogenboom, Poly(2-oxazoline)s: A
comprehensive overview of polymer structures and their physical properties,
Polym. Int. 67 (2018) 32–45. doi:10.1002/pi.5457.
[107] R. Hoogenboom, H.M.L. Thijs, M.J.H.C. Jochems, B.M. Van Lankvelt, M.W.M.
Fijten, U.S. Schubert, Tuning the LCST of poly(2-oxazoline)s by varying
composition and molecular weight: Alternatives to poly(N-
isopropylacrylamide)?, Chem. Commun. (2008) 5758–5760.
Bibliography
98
doi:10.1039/b813140f.
[108] A. Dworak, B. Trzebicka, A. Kowalczuk, C. Tsvetanov, S. Rangelov,
Polyoxazolines — mechanism of synthesis and solution properties, Polimery.
59 (2014) 88–94. doi:10.14314/polimery.2014.088.
[109] Z.M.O. Rzaev, S. Dinçer, E. Pişkin, Functional copolymers of N-
isopropylacrylamide for bioengineering applications, Prog. Polym. Sci. 32
(2007) 534–595. doi:10.1016/j.progpolymsci.2007.01.006.
[110] D. Schmaljohann, Thermo- and pH-responsive polymers in drug delivery, Adv.
Drug Deliv. Rev. 58 (2006) 1655–1670. doi:10.1016/j.addr.2006.09.020.
[111] A. Kikuchi, T. Okano, Intelligent thermoresponsive polymeric stationary phases
for aqueous chromatography of biological compounds, Prog. Polym. Sci. 27
(2002) 1165–1193. doi:10.1016/S0079-6700(02)00013-8.
[112] R. Hoogenboom, H.M.L. Lambermont-Thijs, M.J.H.C. Jochems, S. Hoeppener,
C. Guerlain, C.A. Fustin, J.F. Gohy, U.S. Schubert, A schizophrenic gradient
copolymer: Switching and reversing poly(2-oxazoline) micelles based on UCST
and subtle solvent changes, Soft Matter. 5 (2009) 3590–3592.
doi:10.1039/b912491h.
[113] P. Wilson, P.C. Ke, T.P. Davis, K. Kempe, Poly(2-oxazoline)-based micro- and
nanoparticles: A review, Eur. Polym. J. 88 (2017) 486–515.
doi:10.1016/j.eurpolymj.2016.09.011.
[114] C. Krumm, C.P. Fik, M. Meuris, G.J. Dropalla, H. Geltenpoth, A. Sickmann,
J.C. Tiller, Well-Defi ned Amphiphilic Poly ( 2-oxazoline ) ABA-Triblock
Copolymers and Their Aggregation Behavior in Aqueous Solution, (n.d.) 1677–
1682. doi:10.1002/marc.201200192.
[115] R. Luxenhofer, A. Schulz, C. Roques, S. Li, T.K. Bronich, E. V. Batrakova, R.
Jordan, A. V. Kabanov, Doubly amphiphilic poly(2-oxazoline)s as high-capacity
delivery systems for hydrophobic drugs, Biomaterials. 31 (2010) 4972–4979.
doi:10.1016/j.biomaterials.2010.02.057.
[116] A. Schulz, S. Jaksch, R. Schubel, E. Wegener, Z. Di, Y. Han, A. Meister, J.
Kressler, A. V. Kabanov, R. Luxenhofer, C.M. Papadakis, R. Jordan, Drug-
Bibliography
99
induced morphology switch in drug delivery systems based on poly(2-
oxazoline)s, ACS Nano. 8 (2014) 2686–2696. doi:10.1021/nn406388t.
[117] M.M. Lübtow, L. Hahn, M.S. Haider, R. Luxenhofer, Drug Specificity, Synergy
and Antagonism in Ultrahigh Capacity Poly(2-oxazoline)/Poly(2-oxazine) based
Formulations, J. Am. Chem. Soc. 139 (2017) 10980–10983.
doi:10.1021/jacs.7b05376.
[118] T. Lorson, M.M. Lübtow, E. Wegener, M.S. Haider, S. Borova, D. Nahm, R.
Jordan, M. Sokolski-Papkov, A. V. Kabanov, R. Luxenhofer, Poly(2-oxazoline)s
based biomaterials: A comprehensive and critical update, Biomaterials. 178
(2018) 204–280. doi:10.1016/j.biomaterials.2018.05.022.
[119] T.R. Dargaville, J.R. Park, R. Hoogenboom, Poly(2-oxazoline) Hydrogels:
State-of-the-Art and Emerging Applications, Macromol. Biosci. 18 (2018) 1–15.
doi:10.1002/mabi.201800070.
[120] R. Luxenhofer, Y. Han, A. Schulz, J. Tong, Z. He, A. V. Kabanov, R. Jordan,
Poly(2-oxazoline)s as polymer therapeutics, Macromol. Rapid Commun. 33
(2012) 1613–1631. doi:10.1002/marc.201200354.
[121] R. Konradi, C. Acikgoz, M. Textor, Polyoxazolines for nonfouling surface
coatings - A direct comparison to the gold standard PEG, Macromol. Rapid
Commun. 33 (2012) 1663–1676. doi:10.1002/marc.201200422.
[122] B.J. Chang, O. Prucker, E. Groh, A. Wallrath, M. Dahm, J. Rühe, Surface-
attached polymer monolayers for the control of endothelial cell adhesion,
Colloids Surfaces A Physicochem. Eng. Asp. 198–200 (2002) 519–526.
doi:10.1016/S0927-7757(01)00952-9.
[123] M. Schober, K. Faber, Inverting hydrolases and their use in enantioconvergent
biotransformations, Trends Biotechnol. 31 (2013) 468–478.
doi:10.1016/j.tibtech.2013.05.005.
[124] S. Konieczny, C.P. Fik, N.J.H. Averesch, J.C. Tiller, Organosoluble enzyme
conjugates with poly(2-oxazoline)s via pyromellitic acid dianhydride, J.
Biotechnol. 159 (2012) 195–203. doi:10.1016/j.jbiotec.2012.01.016.
[125] Y. Chujo, K. Sada, R. Nomura, A. Naka, T. Saegusa, Photogelation and Redox
Bibliography
100
Properties of Anthracene-Disulfide-Modified Polyoxazolines, Macromolecules.
26 (1993) 5611–5614. doi:10.1021/ma00073a012.
[126] Y. Chujo, K. Sada, A. Naka, R. Nomura, T. Saegusa, Synthesis and Redox
Gelation of Disulfide-Modified Polyoxazoline, Macromolecules. 26 (1993) 883–
887. doi:10.1021/ma00057a001.
[127] Y. Chujo, K. Sada, T. Saegusa, Cobalt(III) Bipyridyl-Branched Polyoxazoline
Complex as a Thermally and Redox Reversible Hydrogel, Macromolecules. 26
(1993) 6320–6323. doi:10.1021/ma00076a002.
[128] Y. Chujo, K. Sada, T. Saegusa, Synthesis of Bipyridyl-Branched Polyoxazoline
and Its Gelation by Means of Metal Coordination, Polym. J. 25 (1993) 599–608.
doi:10.1295/polymj.25.599.
[129] D.J. van der Heide, B. Verbraeken, R. Hoogenboom, T.R. Dargaville, D.K.
Hickey, Porous poly (2-oxazoline) scaffolds for developing 3D primary human
tissue culture, Biomater Tissue Technol. 1 (2017) 1–5.
doi:10.15761/BTT.1000104.
[130] B.L. Farrugia, K. Kempe, U.S. Schubert, R. Hoogenboom, T.R. Dargaville,
Poly(2-oxazoline) hydrogels for controlled fibroblast attachment,
Biomacromolecules. 14 (2013) 2724–2732. doi:10.1021/bm400518h.
[131] T.R. Dargaville, R. Forster, B.L. Farrugia, K. Kempe, L. Voorhaar, U.S.
Schubert, R. Hoogenboom, Poly(2-oxazoline) hydrogel monoliths via thiol-ene
coupling, Macromol. Rapid Commun. 33 (2012) 1695–1700.
doi:10.1002/marc.201200249.
[132] J.N. Haigh, Y. Chuang, B. Farrugia, R. Hoogenboom, P.D. Dalton, T.R.
Dargaville, Hierarchically Structured Porous Poly ( 2-oxazoline ) Hydrogels,
Macromol. Rapid Commun. 37 (2016) 93–99. doi:10.1002/marc.201500495 93.
[133] S. Koybayashi, H. Uyama, I. Yamamoto, Y. Mastumoto, Preparation of
Monodispersed Poly(methyl methacrylate) Particle in the Size of Micron Range,
Polym. J. 22 (1990) 759–761. doi:10.1295/polymj.22.759.
[134] S. Sinnwell, H. Ritter, Microwave accelerated polymerization of 2-phenyl-5,6-
dihydro-4H-1,3- oxazine: Kinetics and influence of end-groups on glass
Bibliography
101
transition temperature, Macromol. Rapid Commun. 27 (2006) 1335–1340.
doi:10.1002/marc.200600231.
[135] F.M. Menger, V.A. Migulin, Synthesis and properties of multiarmed geminis, J.
Org. Chem. 64 (1999) 8916–8921. doi:10.1021/jo9912350.
[136] L. Ning, 3D bioprinting of scaffolds with living Schwann cells for potential nerve
tissue engineering applications, Biofabrication. In press (2018).
doi:10.1088/1758-5090/aacd30.
[137] W. Liu, Z. Zhong, N. Hu, Y. Zhou, L. Maggio, A.K. Miri, A. Fragasso, X. Jin, A.
Khademhosseini, Y.S. Zhang, Coaxial extrusion bioprinting of 3D microfibrous
constructs with cell-favorable gelatin methacryloyl microenvironments,
Biofabrication. 10 (2018). doi:10.1088/1758-5090/aa9d44.
[138] D.B. Kolesky, R.L. Truby, A.S. Gladman, T.A. Busbee, K.A. Homan, J.A. Lewis,
3D Bioprinting of Vascularized , Heterogeneous Cell-Laden Tissue Constructs,
(2014) 3124–3130. doi:10.1002/adma.201305506.
[139] Z. Zheng, J. Wu, M. Liu, H. Wang, C. Li, M.J. Rodriguez, G. Li, X. Wang, D.L.
Kaplan, 3D Bioprinting of Self-Standing Silk-Based Bioink, Adv. Healthc. Mater.
7 (2018) 1–12. doi:10.1002/adhm.201701026.
[140] N. Celikkin, J.S. Padial, M. Costantini, H. Hendrikse, R. Cohn, C.J. Wilson, A.E.
Rowan, W. Świȩszkowski, 3D printing of thermoresponsive polyisocyanide
(PIC) hydrogels as bioink and fugitive material for tissue engineering, Polymers
(Basel). 10 (2018). doi:10.3390/polym10050555.
[141] I. Pepelanova, K. Kruppa, T. Scheper, A. Lavrentieva, Gelatin-Methacryloyl
(GelMA) Hydrogels with Defined Degree of Functionalization as a Versatile
Toolkit for 3D Cell Culture and Extrusion Bioprinting, Bioengineering. 5 (2018)
55. doi:10.3390/bioengineering5030055.
[142] K. Kang, Y. Kim, S.B. Lee, J.S. Kim, S. Park, W. Kim, H.-M. Yang, S.-J. Kim, J.
Jeong, D. Choi, Three-dimensional bio-printing of hepatic structures with direct-
converted hepatocyte-like cells, Tissue Eng. Part A. 24 (2017)
ten.TEA.2017.0161. doi:10.1089/ten.TEA.2017.0161.
[143] S.F. Marrian, The Chemical Reactions of Pentaerythritol and its Derivatives,
Bibliography
102
Chem. Rev. 43 (1948) 149–202. doi:10.1021/cr60134a004.
[144] R. Luxenhofer, M. Bezen, R. Jordan, Kinetic investigations on the
polymerization of 2-oxazolines using pluritriflate initators, Macromol. Rapid
Commun. 29 (2008) 1509–1513. doi:10.1002/marc.200800293.
[145] F. Wiesbrock, R. Hoogenboom, M. Leenen, S.F.G.M. Van Nispen, M. Van Der
Loop, C.H. Abeln, A.M.J. Van Den Berg, U.S. Schubert, Microwave-assisted
synthesis of a 42-membered library of diblock copoly(2-oxazoline)s and chain-
extended homo poly(2-oxazoline)s and their thermal characterization,
Macromolecules. 38 (2005) 7957–7966. doi:10.1021/ma050437x.
[146] M.W.M. Fijten, C. Haensch, B.M.V. Lankvelt, R. Hoogenboom, U.S. Schubert,
Clickable poly(2-oxazoline)s as versatile building blocks, Macromol. Chem.
Phys. 209 (2008) 1887–1895. doi:10.1002/macp.200800226.
[147] T.X. Viegas, M.D. Bentley, J.M. Harris, Z. Fang, K. Yoon, B. Dizman, R.
Weimer, A. Mero, G. Pasut, F.M. Veronese, Polyoxazoline: Chemistry,
properties, and applications in drug delivery, Bioconjug. Chem. 22 (2011) 976–
986. doi:10.1021/bc200049d.
[148] M.M. Bloksma, U.S. Schubert, R. Hoogenboom, Poly(cyclic imino ether)s
beyond 2-substituted-2-oxazolines, Macromol. Rapid Commun. 32 (2011)
1419–1441. doi:10.1002/marc.201100138.
[149] M.M. Bloksma, R.M. Paulus, H.P.C. Van Kuringen, F. Van Der Woerdt, H.M.L.
Lambermont-Thijs, U.S. Schubert, R. Hoogenboom, Thermoresponsive poly(2-
oxazine)s, Macromol. Rapid Commun. 33 (2012) 92–96.
doi:10.1002/marc.201100587.
[150] Y. Zhao, Y. Li, S. Mao, Effect of bioink properties on printability and cell viability
for 3D bioplotting of embryonic stem cells The influence of printing parameters
on cell survival rate and printability in microextrusion-based 3D cell printing
technology, Biofabrication. 8 (2016). doi:10.1088/1758-5090/8/3/035020.
[151] C. Li, N.J. Buurma, I. Haq, C. Turner, S.P. Armes, V. Castelletto, I.W. Hamley,
A.L. Lewis, Synthesis and characterization of biocompatible, thermoresponsive
ABC and ABA triblock copolymer gelators, Langmuir. 21 (2005) 11026–11033.
Bibliography
103
doi:10.1021/la0515672.
[152] S. Xuan, C.U. Lee, C. Chen, A.B. Doyle, Y. Zhang, L. Guo, V.T. John, D.
Hayes, D. Zhang, Thermoreversible and Injectable ABC Polypeptoid
Hydrogels: Controlling the Hydrogel Properties through Molecular Design,
Chem. Mater. 28 (2016) 727–737. doi:10.1021/acs.chemmater.5b03528.
[153] M. Müller, J. Becher, M. Schnabelrauch, M. Zenobi-Wong, Nanostructured
Pluronic hydrogels as bioinks for 3D bioprinting, Biofabrication. 7 (2015).
doi:10.1088/1758-5090/7/3/035006.
[154] D.B. Kolesky, R.L. Truby, A.S. Gladman, T.A. Busbee, K.A. Homan, J.A. Lewis,
3D bioprinting of vascularized, heterogeneous cell-laden tissue constructs,
Adv. Mater. 26 (2014) 3124–3130. doi:10.1002/adma.201305506.
Appendices
104
Appendices
Appendix I: Supplementary Data
1H NMR spectra of P4 and P5
Appendix 1: 1H NMR spectrum of P4 measured in deuterated chloroform at 300 MHz.
Appendix 2: 1H NMR spectrum of P5 measured in deuterated chloroform at 300 MHz.
Appendices
105
1H NMR spectra of I2 and I3
Appendix 3: 1H NMR spectrum of I2 measured in deuterated acetonitrile at 300 MHz.
Appendix 4: 1H NMR spectrum of I3 measured in deuterated acetonitrile at 300 MHz.
Appendices
106
Amplitude sweep of 20 wt. % P1 compared to 20 wt. % P2
Appendix 5: Amplitude sweep with a strain from 0.01 to 100% of P1 samples. G' and G'' of a 20 wt%
hydrogel of P1 after 18 hours of dialysis against Millipore water in a dialysis tube with MWCO 1 kDa
(grey)., followed by 28 (red) and a final of 56 h (blue) in a dialysis tube with MWCO 10 kDa and a ref-
erence linear copolymer of PMeOx and PnPrOzi P2 (green) at a temperature of 37 °C.
Frequency sweep of 20 wt. % P1 compared to 20 wt. % P2
Appendix 6: Frequency sweep from 0.05 to 100 1/s. G' and G'' of a 20 wt.% hydrogel of P1 after 18
hours of dialysis against Millipore water in a dialysis tube with MWCO 1 kDa (grey)., followed by 28
Appendices
107
(red) and a final of 56 h (blue) in a dialysis tube with MWCO 10 kDa and a reference linear copolymer
of PMeOx and PnPrOzi P2 (green) at a temperature of 37 °C.
Amplitude sweep of 30 & 40 wt. % of P3-P5
Appendix 7: Amplitude sweep with a strain from 0.01 to 100%. Samples were the star-shaped copoly-
mers of PMeOx and PnPrOzi, namely P3 at 30 and 40 wt% (A, B), P4 at 30 and 40 wt% (C, D) and P5
at 30 and 40 wt% (E, F) in comparison to a 20 wt.% hydrogel of P1 (G) and a 20 wt.% solution of
Pluronic ® F127 (H). The graphs show the trend of G’ and G’’ plotted against the applied strain.
Samples were measured at 37 °C.
Appendices
108
Amplitude sweep of polymer blends P1/P3, P1/P4, P1/P5 and references
Appendix 8: Amplitude sweep with a strain from 0.01 to 500%. Samples were the polymer blends of
P1/P3 (A, B), P1/P4 (C, D) and P1/P5 (E, F) in comparison to a 20 wt.% hydrogel of P1 (G) and a
20 wt.% solution of Pluronic ® F127 (H). The graphs show the trend of G’ and G’’ plotted against the
angular frequency. Samples were measured at 35 °C.