Post on 28-Jun-2020
Enhanced Bioactivity and Sustained Release of NT-3 and Anti-NogoA from a
Polymeric Drug Delivery System for Treatment of Spinal Cord Injury
by
Jason Stanwick
A thesis submitted in conformity with the requirements
for the degree of Master of Applied Science
Department of Chemical Engineering and Applied Chemistry
University of Toronto
© Copyright by Jason Stanwick 2011
ii
Enhanced Bioactivity and Sustained Release of Anti-NogoA and NT-3 from a
Composite Polymeric Drug Delivery System for Treatment of Spinal Cord Injury
Jason Stanwick
M.A.Sc
Department of Chemical Engineering and Applied Chemistry
University of Toronto
2011
Abstract
Neurotrophin-3 (NT-3) and anti-NogoA have shown promise in regenerative strategies after
spinal cord injury; however, conventional methods for localized release to the injured spinal
cord are either prone to infection or not suitable for sustained release. To address these issues,
we have designed a composite drug delivery system that is comprised of poly(lactic-co-glycolic
acid) (PLGA) nanoparticles dispersed in an injectable hydrogel of hyaluronan and methyl
cellulose (HAMC). Achieving sustained and bioactive protein release from PLGA particles is a
known challenge; consequently, we studied the effects of processing parameters and excipient
selection on protein release, stability, and bioactivity. We found that embedding PLGA
nanoparticles in HAMC results in more linear drug release likely due to the formation of a
diffusion-limiting layer of methyl cellulose on the particle surface. Co-encapsulated MgCO3 was
able to significantly improve NT-3 bioactivity, while trehalose + hyaluronan was able to
improve anti-NogoA bioactivity and release.
iii
Acknowledgments
I am grateful for financial support from the Canadian Institutes of Health Research (MSS) and
for fellowship support from both the Ontario Graduate Scholarships in Science and Technology
(JS) and the Natural Sciences and Engineering Research Council of Canada (JS). I would like to
express my gratitude to Dr. Ying Fang Chen for assistance with the dorsal root ganglia bioassay,
Dr. Philip Y.K. Choi for a helpful discussion regarding the mathematical model, Dr. Douglas
Baumann for insightful comments, and Dr. Molly Shoichet for her support and guidance.
iv
Declaration of Co-Authorship
The original scientific content of the thesis is comprised of one article that is submitted to a
peer-reviewed internationally recognized journal and a second article that is in preparation for
the same. In both cases these contributions were primarily the work of Jason Stanwick. The
contributions of the co-authors are declared in the following section in conformity with the
requirements for the degree of Master‟s of Applied Science.
v
Table of Contents Abstract ............................................................................................................................................ii Acknowledgments ......................................................................................................................... iii Declaration of Co-Authorship ........................................................................................................ iv
Table of Contents ............................................................................................................................. v Abstracts of Articles Appearing in the Thesis ...............................................................................vii List of Tables .................................................................................................................................. ix List of Figures .................................................................................................................................. x List of Appendices ......................................................................................................................... xv
1 Introduction ................................................................................................................................. 1 1.1 Overview.............................................................................................................................. 1 1.2 Spinal Cord Injury ............................................................................................................... 1
1.3 Hypothesis and Objectives .................................................................................................. 2 1.4 Anatomy of the Spinal Cord ................................................................................................ 2 1.5 Tissue Response to Spinal Cord Injury ............................................................................... 3 1.6 Current Treatments and Clinical Trials for Spinal Cord Injury ........................................... 4
1.7 Neuroregenerative Molecules .............................................................................................. 5 1.8 Drug Delivery to the Injured Spinal Cord ........................................................................... 6
1.9 Proposed Drug Delivery System ......................................................................................... 6 1.10 Sources of Protein Instability............................................................................................... 7 1.11 Obstacles to Sustained Release ............................................................................................ 8
1.12 Modeling Protein Release .................................................................................................. 10 1.13 Summary ............................................................................................................................ 13
1.14 Scope of Thesis .................................................................................................................. 14 2 Enhanced Neurotrophin-3 Bioactivity and Release from a Nanoparticle-Loaded
Composite Hydrogel ................................................................................................................. 15 2.1 Introduction........................................................................................................................ 15 2.2 Materials and Methods ...................................................................................................... 17
2.2.1 Materials ................................................................................................................ 17 2.2.2 Nanoparticle Processing and Hydrogel Preparation .............................................. 18
2.2.3 Nanoparticle Characterization ............................................................................... 18 2.2.4 NT-3 Release ......................................................................................................... 19 2.2.5 Mathematical Model .............................................................................................. 19
2.2.6 PLGA Degradation ................................................................................................ 19 2.2.7 Detection of NT-3 .................................................................................................. 20 2.2.8 NT-3 Bioactivity by Dorsal Root Ganglia (DRG) Bioassay ................................. 21 2.2.9 Statistical Analysis................................................................................................. 22
2.3 Results ............................................................................................................................... 22 2.3.1 Effect of Embedding PLGA Nanoparticles in HAMC .......................................... 22 2.3.2 NT-3 Stability Improvement.................................................................................. 25 2.3.3 Effect of Processing Parameters on NT-3 Release Kinetics .................................. 30 2.3.4 In Vitro Bioactivity ................................................................................................ 31
2.4 Discussion .......................................................................................................................... 33 3 In Vitro Sustained Release of Bioactive Anti-NogoA, a Molecule in Clinical
Development for Treatment of Spinal Cord Injury................................................................... 37 3.1 Introduction........................................................................................................................ 37
vi
3.2 Materials and Methods ...................................................................................................... 38
3.2.1 Materials ................................................................................................................ 38 3.2.2 Nanoparticle Processing and Hydrogel Preparation .............................................. 39 3.2.3 Particle Characterization ........................................................................................ 39 3.2.4 Drug Release Studies ............................................................................................. 40
3.2.5 Mathematical model .............................................................................................. 40 3.2.6 Statistical Analysis................................................................................................. 41
3.3 Results ............................................................................................................................... 41 3.3.1 Anti-NogoA bioactivity was enhanced by trehalose and hyaluronan, but unaffected
by co-encapsulated bases relative to no co-encapsulants .................................................... 41
3.3.2 Anti-NogoA release kinetics were influenced by trehalose and hyaluronan together
and the presence of bases, but not by trehalose alone ......................................................... 44 3.4 Discussion .......................................................................................................................... 48
4 Discussion ................................................................................................................................. 53
4.1 Achieving Sustained and Bioactive NT-3 and anti-NogoA Release ................................. 53 4.2 Why do NT-3 and anti-NogoA behave differently? .......................................................... 57
5 Conclusions ............................................................................................................................... 59 6 Recommendations for Future Work ......................................................................................... 60
6.1 In Vitro Optimization ........................................................................................................ 60 6.2 In Vivo Efficacy ................................................................................................................ 61
7 References ................................................................................................................................. 62
8 Appendix A ............................................................................................................................... 69
vii
Abstracts of Articles Appearing in the Thesis
Enhanced Neurotrophin-3 Bioactivity and Release from a Nanoparticle-Loaded
Composite Hydrogel
Jason C. Stanwick, M. Douglas Baumann, and Molly S. Shoichet
Neurotrophin-3 (NT-3) has shown promise in regenerative strategies after spinal cord injury;
however, sustained local delivery is difficult to achieve by conventional methods. Controlled
release from poly(lactic-co-glycolic acid) (PLGA) nanoparticles has been studied for numerous
proteins, yet achieving sustained release of bioactive proteins remains a challenge. To address
these issues, we designed a composite drug delivery system comprised of NT-3 encapsulated in
PLGA nanoparticles dispersed in an injectable hydrogel of hyaluronan and methyl cellulose
(HAMC). A continuum model was used to fit the in vitro release kinetics of an NT-3 analog
from a nanoparticle formulation. Interestingly, the model suggested that the linear drug release
observed from composite HAMC was likely due to a diffusion-limiting layer of methyl cellulose
on the particle surface. We then studied the effects of processing parameters and excipient
selection on NT-3 release, stability, and bioactivity. Trehalose was shown to be the most
effective additive for stabilizing NT-3 during sonication and lyophilization and PLGA itself was
shown to stabilize NT-3 during these processes. Of four excipients tested, PEG 400 was the
most effective during nanoparticle fabrication, with 74% of NT-3 detected by ELISA.
Conversely, co-encapsulation of magnesium carbonate with NT-3 was most effective in
maintaining NT-3 bioactivity over 28 days according to a cell-based axonal outgrowth assay.
Together, the modeling and optimized processing parameters provide insight critical to testing
the formulation in vivo.
JCS conceived, designed, and executed the experiments and wrote the manuscript. MDB
conducted a drug release study of α-chymotrypsin (Data points in Figure 3) and edited the
manuscript. MSS conceived the project and edited the manuscript.
viii
In Vitro Sustained Release of Bioactive Anti-NogoA, a Molecule in Clinical Development
for Treatment of Spinal Cord Injury
Jason C. Stanwick, M. Douglas Baumann, and Molly S. Shoichet
NogoA is a promising target for enhancing neuroregeneration after spinal cord injury (SCI) and
was the subject of a recently completed phase I clinical trial. This trial was prompted by
multiple reports of functional recovery in rat and non-human primate models of SCI following
continuous intrathecal infusion of anti-NogoA antibodies for 2-4 weeks. These reports utilized
internal pump and intrathecal catheter systems which are not clinically approved for treatment
of SCI, and the trial sponsor therefore employed existing delivery technologies with known
limitations. We previously reported the development of a drug delivery system (DDS) designed
for local delivery to the spinal cord which combined the safety of bolus injection with the
continuous release profile of catheter based systems. The DDS is an injectable composite of
drug loaded poly(lactic-co-glycolic acid) nanoparticles dispersed in a hydrogel matrix. We
presently report the in vitro formulation of this DDS for release of the anti-NogoA mAb 11c7,
including the effect of select co-encapsulated excipients on the release of bioactive 11c7. Co-
encapsulation of MgCO3 or CaCO3 with 11c7 slowed the rate of anti-NogoA release but did not
influence anti-NogoA bioactivity. Co-encapsulation of trehalose significantly improved 11c7
bioactivity at early times, while co-encapsulating trehalose and hyaluronan improved bioactivity
up to 28 days and also resulted in 2-3 fold greater fractional release at 28 days relative to all
other formulations.
JCS conceived, designed, and executed the experiments and wrote the manuscript. MDB edited
the manuscript. MSS conceived the project and edited the manuscript.
ix
List of Tables
Table 1 - A summary of select current clinical studies for spinal cord injury............................... 5
Table 2 – A summary of several causes of protein destabilization during in vitro processing,
drug release, and storage.............................................................................................................. 16
Table 3 – A summary of the first-order bioactivity loss model parameters for the five
formulations described in Figure 10, Figure 11, and Figure 12. ................................................. 42
Table 4 – Mathematical model parameters and particle characterization for selected
formulations ................................................................................................................................. 46
Table 5 – A comparison of the proposed drug delivery system to in vivo NT-3 and anti-NogoA
drug release studies ...................................................................................................................... 54
x
List of Figures
Figure 1– Gross anatomy of the spinal cord. Neural tissue is surrounded by the pia mater,
arachnoid mater, dura, and vertebrae. Cerebrospinal fluid flows through the intrathecal space.
Image copyright 2005 by Michael Corrin. ...................................................................................... 3
Figure 2 – The proposed drug delivery system. The nanoparticle/hydrogel composite is injected
into the intrathecal space. Image copyright by Michael Corrin...................................................... 7
Figure 3- Embedding PLGA nanoparticles in HAMC reduces the burst release and supports
sustained delivery of α-chymotrypsin, an analog for NT-3. Release of α-chymotrypsin from ()
PLGA nanoparticles embedded in HAMC had a lower burst release and more sustained delivery
than from () PLGA nanoparticles alone (n=3, mean ± standard deviation). These two data sets
were previously published by Baumann et al. [3]. α-Chymotrypsin release from (Δ) HAMC
alone occurs over the span of hours. A continuum model based on Fickian diffusion was able to
predict release from ( ) PLGA nanoparticles in aCSF and from ( ) HAMC; however, a
similar model that incorporated diffusion through the HAMC gel was not able to predict release
from nanoparticle embedded in HAMC ( ). Only when model variables associated with the
particles themselves were augmented was an accurate fit obtained ( ). Release of α-
Chymotrypsin and its model fit were in close agreement with a similar formulation with
encapsulated () NT-3 in PLGA nanoparticles, embedded in HAMC (n=3, mean ± standard
deviation) ....................................................................................................................................... 24
Figure 4- Attenuation of the burst release from composite HAMC is not the result of altered
PLGA nanoparticle degradation. PLGA degradation was monitored over 30 d by organic GPC
for () PLGA nanoparticles in aCSF and () PLGA nanoparticles embedded in HAMC (n=3,
mean ± standard deviation). Both traces were similar to each other and to a first order
degradation model using kdeg = 0.086 days-1
( ). Mass loss for () PLGA nanoparticles in
aCSF and () PLGA nanoparticles embedded in HAMC were indistinguishable (n=3, mean ±
standard deviation)......................................................................................................................... 25
Figure 5 – The effects of three processing steps on the stability of NT-3 were investigated by
ELISA. (a) Encapsulation of NT-3 within PLGA nanoparticles stabilized the protein during the
double emulsion synthesis, retaining approximately 40% NT-3 detectability using the following
xi
co-encapsulants: trehalose + hyaluronan, MgCO3, or no additives. Co-encapsulated PEG 400
significantly improved NT-3 stability (p<0.001, n=3, mean ± standard deviation), resulting in
74% detection after processing. (b) After sonication 400 mM trehalose significantly improved
detectability from 25% to 39% (p<0.001, n=3, mean ± standard deviation). (c) The addition of
400 mM trehalose prior to lyophilization improved NT-3 detectability significantly compared to
all other additives (p<0.001, n=3, mean ± standard deviation). .................................................... 27
Figure 6– NT-3 was not detected by ELISA after exposure to low pH and steadily lost
detectability after incubation at 37⁰C. (a) NT-3 detectability by ELISA was fairly stable between
pH 7.4 and pH 3 when incubated for 24 h; however, below pH 3 NT-3 was not detected(n=3,
mean ± standard deviation). (b) NT-3 steadily lost ELISA detectability at approximately 2.5%
per day over the first 23 d of incubation at 37 ⁰C in aCSF (n=3, mean ± standard deviation). .... 28
Figure 7 – NT-3 stored at 4 ⁰C or -80 ⁰C remained stable, but not when stored at 4ºC with 1
wt% BSA for 7 d. When NT-3 was stored in aCSF for 7 d at -80 ⁰C, the NT-3 concentration
measured by ELISA was similar to the initial concentration. Similarly storage at 4 ⁰C only
resulted in a modest 19% loss in detection compared to the initial concentration (p<0.05, n=3,
mean ± standard deviation). However, when stored at these temperatures in the presence of 1
wt% BSA, more than half of the initial NT-3 detected was lost (p<0.001, n=3, mean ± standard
deviation). A fresh sample in 1 wt% BSA (Initial Concentration + BSA) did not exhibit this
same loss in detection, which indicates that this phenomenon is not simply due to the BSA
blocking the ELISA plate. ............................................................................................................. 29
Figure 8 – NT-3 in vitro release from PLGA nanoparticles was fine-tuned by incorporating
excipients and adjusting polymer properties. (a) NT-3 release from PLGA nanoparticles
embedded in () HAMC was not considerably changed by co-encapsulation with () trehalose
and hyaluronan. Co-encapsulation with () MgCO3 resulted in a reduced burst and reduced
cumulative release of NT-3. Co-encapsulation with (Δ) PEG 400 led to a 7 d release profile,
with only 1% released thereafter (n=3, mean ± standard deviation). (b) Release amounts of NT-3
per mg of PLGA nanoparticle for all four formulations, as measured by ELISA, shows the
largest release amount from PLGA alone and PLGA with PEG 400 (n=3, mean ± standard
deviation). This demonstrates that approximately 100 ng of NT-3 can be delivered over 7 d from
xii
the formulation with PEG 400, while 110 ng of NT-3 can be delivered over 28 d from the
formulation without additives. ....................................................................................................... 31
Figure 9 – Released NT-3 is bioactive in a rat dorsal root ganglia neurite outgrowth assay. (a)
NT-3 standards in 0.5 mL aCSF and 0.5 mL differentiation media. The increase in average
number of neurites/DRG with increased NT-3 suggests a correlation in amount of NT-3 present
and number of neurites. (b) The NT-3 released from PLGA nanoparticles was followed in terms
of the following co-encapsulants: ( ) no additives, ( ) trehalose and hyaluronan, () PEG
400, and () MgCO3. All samples up to 28 days stimulated neurite outgrowth from rat dorsal
root ganglia, with the exception of the PEG 400 batch at day 28. Batches with co-encapsulated
MgCO3 exhibited more robust neurite outgrowth with significant differences relative to all other
variables at 1d, 14 d, and 28 d (p<0.001, n=10). ........................................................................... 32
Figure 10– Co-encapsulated trehalose significantly improves the initial bioactivity of released
anti-Nogo-A. (a) The percentage of anti-NogoA that is bioactive during release is significantly
higher (p<0.05, n=3, mean ± standard deviation) at 1, 2, and 7 days when () trehalose is co-
encapsulated with anti-NogoA compared to a formulation with () no additives. At 14, 21, and
28 d, there was no measurable bioactivity for either formulation. A first-order bioactivity loss
model was used to simulate anti-NogoA bioactivity for ( ) co-encapsulated trehalose and (
) no additives. (b) The first 7 d of data were plotted on a semi-log plot to demonstrate that the
improvement to bioactivity is a result of increased initial bioactivity, rather than a change in the
rate of bioactivity loss. . ................................................................................................................ 42
Figure 11– Co-encapsulated hyaluronan with trehalose significant improves bioactivity of
released anti-NogoA at late time points. (a) Anti-NogoA bioactivity is similar over the first 7 d
comparing () co-encapsulated trehalose to () co-encapsulated trehalose and hyaluronan, but
the latter formulation has significantly higher (p<0.05, n=3, mean ± standard deviation)
bioactivity at 14, 21, and 28 d. First-order bioactivity loss models for ( ) co-encapsulated
trehalose and ( ) co-encapsulated hyaluronan were plotted. (b) The bioactivity data was
plotted on a semi-log plot to illustrate the improvement to bioactivity garnered by () co-
encapsulating trehalose and hyaluronan. The first-order model for anti-NogoA bioactivity from
a formulation with ( ) co-encapsulated trehalose was only taken out to 7 d because bioactivity
for this formulation was undetectable at 14 d and beyond.. .......................................................... 43
xiii
Figure 12 – Anti-NogoA bioactivity is similar at early time points with and without co-
encapsulated bases. (a) Trehalose + Hyaluronan nanoparticle formulations with () co-
encapsulated CaCO3 or () co-encapsulated MgCO3 demonstrated similar bioactivity up to 3 d
compared to a formulation with () co-encapsulated trehalose and hyaluronan (n=3, mean ±
standard deviation). There was no detectable bioactive anti-NogoA for the base-encapsulated
formulations at from 7 d onward. First-order bioactivity loss models were identical for ( ) co-
encapsulated CaCO3 and the no base formulation, which were also similar to ( ) co-
encapsulated MgCO3. (b) A semi-log plot of bioactivity data up to 28 d demonstrates that co-
encapsulated bases do no alter early bioactivity, but surprisingly do not improve anti-NogoA
bioactivity at later time points.. ..................................................................................................... 44
Figure 13 – Co-encapsulated trehalose does not influence anti-NogoA release kinetics, while
hyaluronan and trehalose enhance sustained anti-NogoA delivery. When () trehalose was co-
encapsulated in a formulation, a total anti-NogoA release profile was obtained similar to an ()
additive-free formulation. On the other hand, () co-encapsulated trehalose and hyaluronan
increased the burst amount and long-term release rate of anti-NogoA (n=3, mean ± standard
deviation). All traces in this figure are simulations developed using the model, parameters
available in Table 4. ....................................................................................................................... 46
Figure 14 – Co-encapsulated bases reduce the release rate of anti-NogoA. When () CaCO3 or
() MgCO3 were co-encapsulated with anti-NogoA, trehalose and hyaluronan, the total anti-
NogoA release profiles were dramatically reduced compared to () co-encapsulated trehalose
and hyaluronan (n=3, mean ± standard deviation). All traces in this figure are simulations
developed using the model, parameters available in Table 4. ....................................................... 48
Figure 15 - The effect of NT-3 and NogoA on neurite outgrowth viewed in a non-competitive
inhibition model. a) A diagram illustrating the interaction between the inhibitory protein NogoA
with the nogo receptor and NT-3 with the TrkC receptor on neuronal cells. The former inhibits
neurite outgrowth and the latter improves neurite outgrowth. b) The equilibrium equations
describing the receptor/ligand interactions. ................................................................................... 55
Figure 16 – Rate of neurite outgrowth as a function of NT-3 concentration for two values of
NogoA concentration, as simulated by non-competitive ligand-receptor kinetics. This model
xiv
suggests that delivery of anti-NogoA in combination with NT-3 would provide faster neurite
regeneration compared to simply increasing the dosage of NT-3. ................................................ 56
xv
List of Appendices
Appendix A – Summary of the governing equations for the drug release from particles ...69
1
1 Introduction
1.1 Overview
This introduction provides context for the papers presented in Chapter 2 and Chapter 3, which
deal with strategies for enhancing bioactivity and sustained release of neurotrophin-3 (NT-3)
and anti-NogoA from a composite polymer drug delivery system (DDS). This system is
designed to provide an improved treatment option for spinal cord injury. The complications
associated with treating this condition are outlined in Sections 1.2 – 1.7. In Section 1.9, the
proposed DDS is described. The obstacles related to implementing the proposed system are
outlined in Sections 1.10-1.11. In Chapter 2, the processing conditions that lead to NT-3
instability were isolated and treated, allowing for 28 day bioactive drug delivery in vitro. In
Chapter 3, the effects of a variety of excipients on anti-NogoA encapsulation, release, and
bioactivity were investigated, which resulted in 28 day bioactive anti-NogoA delivery in vitro.
1.2 Spinal Cord Injury
Spinal cord injury (SCI) affects 130,000 people each year worldwide [1], of which
approximately 11,000 are from North America [2]. Depending on the location of injury, SCI can
affect the function of different organs. Quadriplegics often report that they would most like to
recover their arm and hand function, while paraplegics rank recovery of sexual function as most
valuable, followed by recovery of bladder/bowel function and eliminating chronic pain, among
other disabilities [1]. The consequences of SCI are not limited to physical suffering. This injury
can also pose a significant economic burden, as the estimated lifetime costs by age of injury can
range from $400,000 to 2.7 million dollars, depending on the age of the individual and the
severity of injury [3]. SCI generally affects younger male populations, as 82% of patients are
male with an average age of 31, which is understandable considering that the predominant
causes of SCI are vehicular accidents, violence, falls, and sports related injuries [4]. Prof.
Shoichet‟s group is developing a treatment option for this devastating condition, which involves
pharmaceutical delivery via drug-loaded poly(lactic-co-glycolic acid) nanoparticles embedded
in a gel of hyaluronan and methyl cellulose (HAMC).
2
1.3 Hypothesis and Objectives
We hypothesize that bioactive neurotrophin-3 and anti-NogoA can be delivered over 28 days in
vitro from PLGA nanoparticles embedded in a HAMC hydrogel. To test this hypothesis, three
objectives were pursued:
1. Identify and minimize sources of NT-3 and anti-NogoA bioactivity loss over the life
cycle of the drug delivery system
2. Develop a mathematical model to guide the choice of processing parameters for
sustained release
3. Deliver NT-3 and anti-NogoA in vitro and quantify bioactive release
1.4 Anatomy of the Spinal Cord
The central nervous system is comprised of the brain and spinal cord. The spinal cord permits
motor and sensory communication between the brain and the peripheral nervous system. The
spinal cord consists of grey matter surrounded by white matter, which are at a cellular level
neuronal cell bodies and myelinated axons, respectively. A cross-sectional view of the spinal
cord is presented in Figure 1. Starting in the centre of the diagram, we see the butterfly-shaped
grey matter, which is surrounded by the white matter. This section of the cord is covered by a
layer known as the pia mater, a highly vascularized, protective matrix. The dura and arachnoid
mater are similarly protective meningeal membranes. In between the arachnoid mater and pia
mater is the intrathecal space, which carries cerebrospinal fluid (CSF), an acellular buffer
solution that protects the spinal cord from sudden motion, while providing a pathway for
metabolic waste removal and delivery of nutrition to the spinal cord. The final layer surrounding
the spinal cord is the epidural space, which consists of fatty tissue and sits beside the vertebra.
One important property of the spinal cord is the blood-spinal cord barrier (BSCB), which
consists of endothelial cells and astrocytes and lines blood vessels within the cord. This barrier
restricts diffusion of molecules and ions into the cord, which is one of the primary obstacles that
prevent systemic delivery of therapeutics to the injured spinal cord.
3
Figure 1– Gross anatomy of the spinal cord. Neural tissue is surrounded by the pia mater,
arachnoid mater, dura, and vertebrae. Cerebrospinal fluid flows through the intrathecal space.
Image copyright 2005 by Michael Corrin.
1.5 Tissue Response to Spinal Cord Injury
SCI results from either compression or laceration of the spinal cord. In both cases, this damage
results in neuronal cell death and severing of axons. This primary insult is following by a
secondary cascade that exacerbates the original injury. This cascade begins with an
inflammatory response, which is characterized by ischemia and hypoxia in the grey matter [5],
which leads to apoptosis in local neurons and oligodendrocytes for weeks after the initial injury.
A cavity forms, which is later isolated by reactive astocytes in what is known as the glial scar
[6].
4
1.6 Current Treatments and Clinical Trials for Spinal Cord Injury
There is no treatment for spinal cord injury that has shown significant functional recovery.
Delivery of methylprednisolone sodium succinate (MPSS) offers modest functional benefit, but
its use is contentious because of the wide range of serious side effects [7]. Surgical
decompression of the spinal cord and physical rehabilitation immediately after injury has
become widespread in North America [8], but offers only modest recovery. In light of these
ineffective treatment options, numerous clinical trials for promising therapeutics and cell
delivery strategies are underway.
Anti-NogoA, an anti-inhibitory molecule, entered a clinical trial in 2006. This therapeutic was
delivered by an external pump, until complications, likely associated with infection, led to
delivery by repeated bolus injections. A Phase II clinical trial for anti-NogoA is planned. A
summary of some clinical trials are listed in Table 1. Cethrin, a Rho agonist, completed Phase
IIa clinical trials, but was terminated as a treatment option in 2010. This molecule was delivered
by extra-dural administration with a fibrin matrix used to supply Cethrin to the dural mater. Cell
delivery strategies are also being investigated clinically. The Procord trial for autologous
macrophages did not progress past Phase II trials [9], while a pilot study for bone marrow
stromal cell injection has entered Phase II clinical trials [10]. An emergent trend in SCI
treatment is the use of drug delivery systems that reduce systemic side effects and avoid the
blood-spinal cord barrier.
5
Table 1 - A summary of select current clinical studies for spinal cord injury
Approach Description Current
Stage of
Clinical
Trial
Delivery of
Therapeutics
Anti-Nogo-A A monoclonal antibody that overcomes neurite outgrowth
inhibition caused by Nogo-A. Delivered by osmotic pump or
repeated bolus delivery by lumbar puncture.
Phase I
Cethrin® A Rho antagonist that promotes neuroprotection and
neuroregeneration. Administered to the dura of the injured spinal
cord with a fibrin sealant.
Phase IIb
-
suspended
Minocycline A broad spectrum antibiotic that has demonstrated neuroprotection
in animal models. Injected intravenously.
Phase I/II
Fampridine® A voltage-dependent potassium channel blocker that improves
sensory and motor function. Delivered intravenously or orally.
Phase III
Cell-based
Therapies
Procord® Autologous macrophages that elevate the release of protective
molecules. Delivered by injection onto the spinal cord.
Phase II -
suspended
Bone
marrow
stromal cells
Contains pluripotent cells that release growth factors to improve
regeneration. Delivered by transplantation onto the spinal cord.
Phase II
1.7 Neuroregenerative Molecules
Pharmaceuticals for SCI can be classified as either neuroprotective or neuroregenerative,
namely, they either protect neurons from apoptosis or promote the regrowth or repair of nervous
tissues or cells. Neurotrophin-3 (NT-3) and anti-NogoA are two promising neuroregenerative
molecules and are the focus of this thesis. NT-3 is responsible for the maintenance,
proliferation, and differentiation of tyrosine kinase C-positive (TrkC) neurons [11]. It is a
growth factor that belongs to the neurotrophin family, which also includes nerve growth factor,
brain-derived neurotrophic factor, neurotrophin-4, and neurotrophin-6. NT-3 is known to bind to
two receptors with nanomolar affinity: TrkC and p75NTR
[12]. NT-3 has been delivered both in
vitro and in vivo using various strategies, including fibrin scaffolds [13], lipid microtubes
embedded in agarose gels formed in situ [14], transfected olfactory ensheathing cells [15],
poly(N-isopropylacrylamide)-co-poly(ethylene glycol) gels [16], and poly(lactic-co-glycolic
acid) (PLGA) microspheres embedded in a poly(ethylene glycol) gel [17]. NT-3 has been shown
6
to be particularly effective in combination with brain-derived neurotrophic factor [18], cyclic
adenosine monophosphate [19], and chondroitinase ABC [14]. Sustained release of NT-3 for 14
days [20] and up to one month [21] has been shown to promote axonal regeneration and
functional recovery.
Anti-NogoA is an antagonist of NogoA, which is a myelin inhibitor known to cause growth
cone collapse and reduce neurite outgrowth [22]. Anti-NogoA has been shown to improve
functional recovery in rat models when delivered by an intrathecal catheter over 14 d [23] or 28
d [24] and is currently being studied clinically [7]. To avoid the blood-spinal cord barrier in
clinical trials, osmotic minipumps have been used; however, complications likely associated
with the infection-prone external minipumps [25], led to the use of repeated bolus injections.
1.8 Drug Delivery to the Injured Spinal Cord
The BSCB inhibits the transfer of therapeutics from the blood stream to the CNS, eliminating
systemic drug delivery as a desirable treatment option. Local drug release strategies to the spinal
cord involve either epidural delivery or intrathecal delivery and are carried out by either bolus
injection or by continuous infusion from a minipump-catheter system. Intrathecal delivery
avoids the need for molecular diffusion through the dura mater into the spinal cord. Bolus
injections allow for local delivery to the spinal cord, but do not offer the sustained delivery
required by some therapies [21]. On the other hand, external minipump-catheter systems offer
sustained and local delivery, but are prone to infection over long time periods [25].
1.9 Proposed Drug Delivery System
Considering 14 to 28 day delivery regimes required for NT-3 and anti-NogoA, an ideal drug
delivery system would combine the local delivery afforded by bolus injections with the
sustained delivery associated with minipump-catheter systems. To this end, Prof. Shoichet‟s
group has developed a drug delivery system that consists of drug-loaded poly(lactic-co-glycolic
acid) (PLGA) nanoparticles embedded in a hydrogel of hyaluronan and methyl cellulose
(HAMC). The nanoparticles are formed by double-emulsion synthesis and slow the rate of drug
release, while the hydrogel localizes the particles at the site of injury in the intrathecal space
(Figure 2). This approach is minimally invasive, biocompatible over 28 d [26], and has been
shown to release dbcAMP, EGF, α-chymotrypsin, and IgG over 28 d in vitro [27] and fibroblast
7
growth factor 2 over 24 hours in vivo [28]. Achieving sustained and bioactive protein release
from a polymeric drug delivery system is a known challenge. In this thesis, strategies for
achieving sustained and bioactive release of NT-3 and anti-NogoA from this system are
explored.
Figure 2 – The proposed drug delivery system. The nanoparticle/hydrogel composite is
injected into the intrathecal space. Image copyright by Michael Corrin.
1.10 Sources of Protein Instability
Proteins in general are susceptible to structural damage from a variety of environmental sources,
especially during formulation, release, and storage. For instance, during particle synthesis a
variety of proteins have been shown to become damaged during sonication [29], lyophilization
[30], freeze/thaw cycles [31], exposure to organic solvents [32] and at low pH [33]. Sonication-
induced damaged is caused by the local thermal and shear stresses caused by the instrument.
Protein instability due to lyophilization is caused by a variety of stresses, which include low
temperature stress; freezing stresses, such as, the formation of dendritic ice crystals, increased
8
ionic strength, altered pH, and phase separation; and drying stresses, namely, the removal of
the protein hydration shell [30]. Exposure to organic solvents causes protein unfolding by
presenting their hydrophobic area to the organic phase of the water/oil emulsion [34].
Strategies for mitigating damage to proteins in polymeric particles involve the adjustment of
polymer properties, manipulation of proteins, or the use of additives. Polymer hydrophobicity
can influence protein stability and release kinetics, as it has been shown that a blend of
poly(lactic acid) (PLA) and poly(ethylene glycol) (PEG) can improve the stability of
encapsulated bovine serum albumin (BSA) by reducing the rate of acidic oligomer formation
(because PLA degrades slower than PLGA), while increasing water uptake because of the
hydrophilic PEG component, which resulted in the more rapid removal of acidic degradation
products [35]. Polymer concentration can affect protein activity, as lysozyme activity was
increased from 59% to 83% by increasing PLGA concentration in the organic phase from 4.5%
to 37%. This improvement was attributed to the faster particle solidification, which results in
reduced exposure time between the protein and the water/oil interface [36]. In terms of protein
manipulation, covalent bonding of PEG to therapeutic proteins (pegylation) has shown the most
promise [34]. For example, methoxypoly(ethylene glycol)-conjugated lysozyme demonstrated
enhanced stability during exposure to organic solvents and homogenization [37], pegylated
interferon-α was protected from the organic solvent/water interface during PLGA particle
processing [38], and nerve growth factor (NGF) was stabilized during encapsulation and release
from PLGA microspheres through pegylation [39]. The use of excipients is the most widely
employed strategy for improving protein stability [40]. Lyoprotectants such as trehalose are
often used to minimize damage during lyophilization [30]; viscosity-controlling agents such as
hyaluronan have been used to minimize protein denaturation during emulsion processing [41];
and basic additives have been used to neutralize the low pH environment often found inside
PLGA particles [42].
1.11 Obstacles to Sustained Release
Attaining sustained and complete release of proteins from PLGA particles is a widely reported
challenge [43]. Since all reports of NT-3 and anti-NogoA delivery that demonstrated functional
recovery required sustained delivery either for 14 days or 28 days (see Section 1.7), PLGA
formulation design is particularly important. The PLGA nanoparticles described in this thesis
9
are formed using a double emulsion process [44]. This technique involves dissolving the
protein of interest in an inner aqueous phase, which is then added to a solution of PLGA in an
organic solvent. This suspension is sonicated to form a primary (w/o) emulsion. This is then
added to an outer aqueous phase containing surfactant and sonicated to produce the double
(w/o/w) emulsion. As the organic solvent partitions into the outer aqueous phase and evaporates,
the final solid PLGA nanoparticles are formed.
Protein release from PLGA particles occurs through diffusion through pores in the particle
matrix. These pores are formed either through dissolution of water-soluble molecules within the
particles or through the hydrolytic degradation of the polymer‟s ester bond linkage [43]. A
typical release profile from PLGA particles consists of an initial burst release for the first 1 to 3
days, followed by a slow release phase, which persists until a threshold of polymer degradation
has occurred, at which point the remaining encapsulated protein is released in a second burst.
The onset of this second burst generally occurs between 30 to 90 days. Interestingly, the
degradation of the PLGA is autocatalytic; as the polymer degrades, acidic degradation products
(oligomers of lactic and glycolic acid) are released, which increase the rate of acid-catalyzed
hydrolysis. Since water uptake is known to be faster than the rate of polymer degradation [45],
bulk erosion is the mechanism of particle deterioration. Clearly, any strategy used to adjust
release from these particles must account for both protein diffusion and PLGA degradation.
Protein delivery from PLGA nanoparticles is a complex process that is influenced by a wide
range of parameters, including PLGA molecular weight (MW), PLGA concentration, particle
size, and particle morphology [46]. A principle consideration when designing particle
formulations is the solidification time, namely, the time it takes for the organic solvent to be
extracted from the particle. As this time increases, there is more opportunity for entrapped drug
to diffuse out, reducing encapsulation efficiency. Further, increased solidification time permits
more water uptake, which increases particle porosity, increasing the subsequent rate of drug
release [47]. As PLGA MW increases, its solubility in many organic solvents decreases, which
promotes faster particle solidification. Similarly, as PLGA concentration in the organic phase
increases, the solubility limit is reached more rapidly as the organic solvent is extracted, which
promotes more rapid solidification [47]. Particle size can adjust release kinetics as a
consequence of the surface area to volume ratio of different sized particles. Moreover, particle
size can influence whether the particles are injectable or available to be sterile filtered. Size can
10
be controlled by adjusting the PLGA concentration [48, 49], PLGA MW [50], surfactant
concentration [51], water/oil ratio during synthesis [52], or the intensity or duration of
sonication [52]. The onset of the degradation stage of PLGA nanoparticles is generally
controlled by the selection of PLGA molecular weight [14] and drug loading [16].
The excipients used for improvement of protein stability can also unintentionally alter the
release profile of proteins from PLGA particles. Water-soluble additives can act as pore-forming
agents and result in faster drug release [53]. Excipients can also modify the solubility of PLGA
in the organic phase during processing or alter the osmolarity of the inner aqueous phase, both
of which can affect encapsulation efficiency and release kinetics as a result of modified water
uptake during nanoparticle solidification [47]. Clearly, it is non-trivial to improve the bioactivity
of encapsulated proteins without modifying release kinetics.
1.12 Modeling Protein Release
In order to understand and control the myriad of factors that affect release, some groups have
developed mathematical models of protein release from PLGA particles. This section will only
outline mechanistic mathematical models, which can be used to understand the mechanism of
release, while empirical models, which are purely descriptive, are omitted. Depending on the
composition, geometry, and preparation method of a drug delivery system, different transport
phenomena may be controlling release rate, including [45]:
water penetration into the system
drug dissolution
dissolution/degradation of the matrix former
precipitation and re-dissolution of degradation products
structural changes within the system occurring during drug
release, such as the creation/closure of water-filled pores
changes in the microenvironmental pH(e.g., creation of acidic microclimates in PLGA-
based delivery systems and subsequent autocatalysis of the polyester)
diffusion of drug and/or degradation products of the matrix material out of the device
with constant or time-dependent diffusion coefficients
osmotic effects
convection processes, and
11
adsorption/desorption phenomena
It is usually not practical to incorporate all of these phenomena into a mechanistic model, so
only the dominant (rate limiting) physical processes are considered. In the case of drug release
from polymeric particles, the physical processes incorporated into most published models are
protein diffusion and polymer degradation.
Charlier et al. [54] reported a model for drug release from bulk eroding PLGA films. In this
model, polymer degradation and drug diffusion are considered simultaneously using a pseudo-
steady state approach, similar to that used in the well-known square root of time Higuchi model
[55]. Also, it is assumed that the initial drug loading is much greater than the solubility of the
drug within the system. The diffusivity term is taken as:
where D0 is the initial diffusivity of the drug, k is the first-order degradation rate constant of
PLGA, and t is elapsed time. They then derived the following formula to describe the absolute
amount of drug released, Q:
where S is the surface area exposed to the release buffer, Co and Cs are the initial drug
concentration and the solubility of drug in the system, respectively. They were able to achieve
good agreement between theoretical and experimental release of mifepristone from PLGA-based
films.
Raman et al. [56] developed an alternate model for drug release from PLGA microparticles.
Similar to the Charlier model, this group considered both drug diffusion and polymer
degradation. The model was based on a Fickian diffusion model in spherical co-ordinates with a
time-dependent diffusivity term:
12
where c is the concentration of drug in the particles, t is time, r is the radial co-ordinate, and
D(Mw) is the molecular weight-dependent diffusivity term.
The D(Mw) term was experimentally determined for piroxicam-loaded PLGA microparticles
and the following boundary conditions were imposed upon the governing equation:
0
where R represents the radius of the particles and f(r) is the initial drug distribution within the
spheres. This model was solved numerically and accurately fitted to experimentally measured
release of piroxicam-loaded PLGA microspheres.
Faisant et al. [57] developed a similar model that used the same governing equation and
boundary conditions as the Raman model, but incorporated a fit parameter in the diffusivity
term:
where D0 is the initial diffusivity of 5-Flourouracil from their 125µm PLGA microparticles, k is
a fit parameter, kdegr is the first-order degradation rate constant of PLGA, and t is time. Their
model led to good agreement between experimental and simulated drug release.
An alternative to the Mw-dependent diffusivity parameters used in the models by Charlier,
Raman, and Faisant is considering polymer degradation based on Monte Carlo simulations.
Siepmann et al. [58] developed a model that used Fick‟s second law of diffusion combined with
Monte Carlo simulations to determine the diffusivity term, which allowed for the numerical
determination of release kinetics. The theory behind this approach is that PLGA particles are
known to degrade by bulk erosion, but due to the complexity of these systems, it is unclear at
what point a particular ester bond in a defined location will be cleaved. This approach
13
theoretically divides the particles into pixels, which are each given a randomly distributed
“lifetime”. When a pixel has passed its lifetime, a status function combines this information with
the status of other pixels to produce a diffusivity term.
1.13 Summary
Spinal cord injury is a devastating condition that affects 130,000 people each year worldwide.
Beyond the physical and emotional trauma, patients face a significant financial burden. The
injury itself is caused by either compression or transection of the spinal cord, which is followed
by a secondary cascade, which causes further cell death. Currently, no treatment for spinal cord
injury has shown substantial functional recovery in a clinical setting. A promising new approach
is the delivery of neuroregenerative medicine like neurotrophin-3 (NT-3) or anti-NogoA over
the span of 2 to 4 weeks. In particular, anti-NogoA has shown promise in Phase I clinical trials;
however, there is no technique that is capable of sustained protein release that is both safe and
minimally invasive. The minipump/catheter system used in the anti-NogoA trial had to be
discontinued, likely as a result from the system‟s known predisposition toward infection.
Instead, repeated bolus injections to the spinal cord were implemented, which do not offer
sustained delivery. Clearly, there is a need for technology which is capable of delivering these
promising neuroregenerative agents to the spinal cord over sustained periods in a safe and
minimally invasive manner.
Towards this goal, our lab has developed a drug delivery system that is comprised of drug-
loaded PLGA nanoparticles dispersed in a hydrogel of hyaluronan and methyl cellulose. The
nanoparticles offer sustained release of the entrapped drug, while the hydrogel localizes the
particles at the site of injury. When delivered by injection to the intrathecal space during
decompressive surgery, this system is minimally invasive. It has also been shown to be
biocompatible in rat models. Yet, achieving sustained and bioactive release of proteins from
PLGA particles is a known obstacle. The tertiary structure of proteins can be altered during
particle processing, drug release, or storage. Excipients are commonly used to improve protein
stability in similar systems, but these additives can alter release kinetics of the original
therapeutic.
14
The current work describes the development of PLGA nanoparticles embedded in a HAMC
hydrogel for sustained and bioactive in vitro release of NT-3 and anti-NogoA. The processing
conditions that affect NT-3 stability were isolated and treatments were identified. A
mathematical model was developed to quantify the effect of excipients on release kinetics. For
both NT-3 and anti-NogoA, bioactive delivery was achieved up to 28 days.
1.14 Scope of Thesis
This thesis describes the development of the PLGA nanoparticles embedded in HAMC to
achieve sustained and bioactive release of NT-3 and anti-NogoA over 28 days. These original
contributions are divided into two chapters:
Chapter 2 – NT-3 was encapsulated in PLGA nanoparticles by the double emulsion
method and its bioactivity studied after release from composite HAMC as a function of
processing parameters and choice of excipients. A continuum model was used to fit the
in vitro release kinetics of an NT-3 analog from a nanoparticle formulation. Sources of
NT-3 bioactivity loss were isolated and ELISA was used to quantify the effect on NT-3
of: sonication; lyophilization; incubation at low pH; and storage at -80, 4, or 37 °C. The
effects of the excipients trehalose, hyaluronan, PEG 400, and MgCO3 were also
evaluated.
Chapter 3 – In vitro anti-NogoA release studies were conducted from the proposed drug
delivery system and the efficacy of several formulations in improving anti-NogoA
bioactivity and sustained release in vitro were evaluated. Formulations were based on
combinations of co-encapsulated trehalose, hyaluronan, MgCO3, and CaCO3. A
continuum model was used to fit the in vitro release kinetics of an NT-3 analog from a
nanoparticle formulation to allow for quantitative comparisons of formulations.
15
2 Enhanced Neurotrophin-3 Bioactivity and Release from a
Nanoparticle-Loaded Composite Hydrogel
2.1 Introduction
Spinal cord injury is a devastating condition that affects more than 130,000 people each year
worldwide and often results in permanent functional and sensory deficits [1]. Pharmaceutical
therapy is promising because many targets for neuroprotection and neuroregeneration have been
identified; however, systemic administration is only possible for very few molecules because the
blood-spinal cord barrier (BSCB) limits diffusion into the spinal cord. Local delivery is
attractive because it bypasses the BSCB, but strategies used clinically are not ideal: external
minipumps are prone to infection [25] and bolus injections offer only transient delivery. A
localized drug delivery system comprised of drug-loaded PLGA nanoparticles dispersed within
a hydrogel of hyaluronan and methyl cellulose (HAMC) and injected into the intrathecal space
that surrounds the spinal cord has been reported [27]. The nanoparticles offer sustained release
while the HAMC gel localizes the nanoparticles at the site of injection. The strategy is designed
to combine the simplicity and safety of bolus injection with the sustained release offered by
pump and catheter systems. Composite HAMC (PLGA nanoparticles embedded in HAMC) is
biodegradable, injectable, and biocompatible in the intrathecal space over 28 days [26].
The neurotrophins are a family of regenerative proteins that modulate the survival, development,
and function of neurons in the central nervous system [59]. A foremost example is neurotrophin-
3 (NT-3), which is responsible for the maintenance, proliferation, and differentiation of tyrosine
kinase C-positive neurons [11]. NT-3 has been delivered both in vitro and in vivo using various
strategies, including: fibrin scaffolds [13], lipid microtubes embedded in agarose gels [14],
transfected olfactory ensheathing cells [15], poly(N-isopropylacrylamide)-co-poly(ethylene
glycol) (PNIPAAm-PEG) gels [16], and PLGA microspheres embedded in a PEG gel [17]. NT-
3 has been shown to be particularly effective in combination with brain-derived neurotrophic
factor (BDNF) [18], cyclic adenosine monophosphate (cAMP) [19], and chondroitinase ABC
[14]. Sustained release of NT-3 for 14 days [20] and up to one month [21] has been shown to
promote axonal regeneration and functional recovery. The critical challenge when formulating
16
NT-3 for sustained release from PLGA particles is to retain bioactivity throughout the
treatment term.
Proteins in general are susceptible to structural damage when exposed to harsh conditions,
including those experienced during encapsulation, release, and storage. For example, during
particle synthesis a variety of proteins have been shown to become damaged during sonication
[29], lyophilization [30], freeze/thaw cycles [31], and at low pH [33], as summarized in Table 2
(for an extensive review see [34]). Excipients are often added to drug delivery systems to
minimize the damage caused by these processes [30]. For example, lyoprotectants such as
trehalose are often used to minimize damage during lyophilization [30]; viscosity-controlling
agents such as hyaluronan have been used to minimize protein denaturation during emulsion
processing [41]; and basic additives such as magnesium carbonate have been used to neutralize
the low pH environment often found inside PLGA particles [42]. These excipients can also
unintentionally alter the release profile of proteins from PLGA particles formed by double
emulsion solvent evaporation. Water-soluble additives can act as pore-forming agents and result
in faster drug release [53]. Excipients can also modify the solubility of PLGA in the organic
phase during processing or alter the osmolarity of the inner aqueous phase, both of which can
affect encapsulation efficiency and release kinetics as a result of modified water uptake during
nanoparticle solidification [47]. Clearly, it is non-trivial to improve the bioactivity of
encapsulated proteins without influencing release kinetics.
Table 2 – A summary of several causes of protein destabilization during in vitro processing,
drug release, and storage
Processing Drug Release Storage
Potential Causes of
Protein Instability
Lyophilization
Sonication
Organic solvents
Aggregation
Adsorption
Denaturation
Degradation
Incubation
Freeze/Thaw cycles
17
Towards the development of a drug delivery system for spinal cord injury, we explore the
influence of processing parameters on NT-3 stability, release kinetics, and bioactivity in the
context of proposed PLGA nanoparticle/HAMC hydrogel composite drug delivery system. By
fitting experimental data points to a theoretical model of release, we provide insight into the
mechanism of NT-3 release from the composite HAMC. NT-3 detection by ELISA was used to
assess structural damage during processes associated with nanoparticle fabrication, in vitro
release, and storage. ELISA and a rat dorsal root ganglion bioassay were then used to assess
NT-3 release kinetics and bioactivity, respectively, from various PLGA nanoparticle
formulations.
2.2 Materials and Methods
2.2.1 Materials
Recombinant human NT-3 was purchased from R&D Systems (Minneapolis, USA). Trehalose,
MgCO3, lactose, glucose, glycerol, poly-D-lysine, sodium dodecyl sulfate (SDS), bovine serum
albumin (BSA), and α-chymotrypsin (type II from bovine pancreas), were purchased from
Sigma-Aldrich (Oakville, CA). Poly(DL-lactic-co-glycolic acid) 50:50 of inherent viscosity 0.67
dL/g (Mn = 30000, Mw = 45000) was purchased from Durect (Cupertino, USA). Poly(vinyl
alcohol), 6000 g/mol was purchased from Polysciences Inc. (Warrington, USA). Sodium
hyaluronate, 2600 kg/mol was purchased from Lifecore (Chaska, USA). Methyl cellulose, 300
kg/mol, was purchased from Shin-Etsu (Tokyo, Japan). Sodium hydroxide was purchased from
EMD Chemicals (Gibbstown, USA). Pluronic F-127 was purchased from BASF (Mississauga,
CA). Fetal bovine serum (FBS), B-27 serum-free supplement, penicillin-streptomycin, and
laminin were purchased from Invitrogen (Burlington, CA).
Artificial cerebrospinal fluid (aCSF) at a pH of 7.4 was prepared as described by Gupta et al.
[60]. HPLC grade dichloromethane (DCM), dimethyl sulfoxide (DMSO), tetrahydrofuran
(THF), and hydrochloric acid (HCl) were purchased from Caledon Labs (Georgetown, CA).
Dulbecco‟s phosphate buffered saline (pH 7.4, 9.55 g/L) was purchased from Wisent Inc. (St-
Bruno, CA). All buffers were prepared using water distilled and deionized using a Millipore
Milli-RO 10 Plus and Milli-Q UF Plus at 18 MΩ resistance (Millipore, Bedford, USA). Neural
basal media and glutamine 200 mM were purchased from Gibco (Burlington, CA).
18
2.2.2 Nanoparticle Processing and Hydrogel Preparation
NT-3 loaded nanoparticles were prepared using a water/oil/water (w/o/w) double emulsion
solvent evaporation technique, as described previously [27]. Briefly, an inner aqueous phase of
100 μL aCSF containing 5 μg NT-3 and 12 mg BSA was mixed with an organic phase of 0.9
mL DCM, 120 mg PLGA and 0.45 mg Pluronic F-127. This mixture was sonicated using a
Vibra-Cell (Sonics, Newtown, USA) on ice for 10 minutes at 26 W and 20 kHz to create the
primary emulsion, which was subsequently mixed with the outer aqueous phase of 5.5 mL of 25
mg/mL PVA. The secondary emulsion was formed through sonication on ice for an additional
10 min at 39 W and 20 kHz. This double emulsion was then added to 34.5 mL of 25 mg/mL
PVA and stirred gently for 20 h at room temperature. PLGA nanoparticles were then isolated
and washed 4 times by ultracentrifugation (Beckman, Mississauga, CA), lyophilized (Labconco,
Kansas City, USA), and stored at -20 ⁰C. Various excipients were also incorporated in modified
formulations: (a) 14 mg trehalose and 1.3 mg hyaluronan were dissolved in the inner aqueous
phase; (b) 5 μL of PEG 400 was added to the inner aqueous phase; (c) 12 mg of α-chymotrypsin
was used in place of NT-3 and BSA in the inner aqueous phase; (d) 4 mg MgCO3 was added to
the organic phase.
HAMC hydrogels were prepared through the physical blending of hyaluronan and methyl
cellulose in aCSF for a final composition of 1 wt% 2600 kg/mol hyaluronan and 3 wt% 300
kg/mol methyl cellulose. Methyl cellulose and hyaluronan were sequentially dispersed in aCSF
using a dual asymmetric centrifugal mixer (Flacktek Inc., Landrum, USA) and left to dissolve
overnight at 4 ⁰C.
2.2.3 Nanoparticle Characterization
Particle size was measured using dynamic light scattering (Zetasizer Nano ZS, Malvern
Instruments, Malvern, UK). Particle yield was defined as the total mass of particles produced
divided by mass of the initial mass of PLGA used, adjusted for protein content. Drug loading is
the mass fraction of NT-3 (or α-chymotrypsin) in the particles, while encapsulation efficiency is
the measured drug loading of the particles divided by the theoretical maximum drug loading. To
determine the total protein encapsulation efficiency, 1 mg of nanoparticles was dissolved in 5
mL DMSO and added to 5 mL of 0.05 M NaOH containing 0.05 wt% SDS and analyzed using
19
the total protein BCA assay according to the manufacturer‟s instructions (Thermo Scientific,
Nepean, CA). To determine NT-3 encapsulation efficiency, 1 mg of particles was dissolved in 1
mL DCM for 1 h. The protein was then extracted into a liquid phase of 10.5 mL reagent diluent
and analyzed using an NT-3 ELISA (R&D Systems, Minneapolis, USA) according to the
manufacturer‟s protocol.
2.2.4 NT-3 Release
Release profiles of α-chymotrypsin or NT-3 from each particle batch were obtained by
dispersing 10 mg of particles in 0.1 mL of HAMC in a 2 mL microcentrifuge tube (Axygen,
Union City, USA) using a dual asymmetric centrifugal mixer at 3300 rpm for 4 min to produce a
final composition of 8 wt% particles, 1 wt% hyaluronan, and 3 wt% methyl cellulose. The
composite was then warmed to 37 ⁰C and 0.9 mL pre-warmed aCSF was added to the sample
tubes. The supernatant was removed and replaced completely at the following time points: 3, 6
h, 1, 3, 7, 14, 21, and 28 d. The protein content of the supernatant was determined by ELISA
(NT-3) or BCA (α-chymotrypsin). After 28 d, the NT-3 remaining inside the particles was
quantified by dissolving the particles in 0.1 mL DCM and extracting the remaining protein into
1 mL reagent diluent for protein quantification by ELISA.
2.2.5 Mathematical Model
A mathematical model constructed in Matlab (MathWorks, Natick, USA) was used to
quantitatively describe the effect of various processing parameters on the release kinetics of α-
chymotrypsin, an analog for NT-3. Applying the models developed by Faisant et al. [57] and
Raman et al. [56], with minor modifications (Supplemental Information), release from
composite HAMC was simulated in two parts: release from PLGA particles was simulated using
a one-dimensional Fickian diffusion model in spherical coordinates and release from the HAMC
hydrogel was simulated using a one-dimensional Fickian diffusion model in Cartesian
coordinates.
2.2.6 PLGA Degradation
To determine whether dispersion in HAMC altered the rate of PLGA degradation relative to
dispersion in aCSF, 10 mg of PLGA particles (without encapsulated protein) were dispersed in
20
0.1 mL concentrated HAMC or aCSF in 2 mL microcentrifuge tubes. In both cases, 0.9 mL
aCSF was added to all tubes and incubated at 37 ⁰C on a rotary shaker at 2 Hz. Samples were
washed 6 times with ice-cold distilled water, isolated by ultracentrifugation, and lyophilized at
the following time points: 0, 2, 4, 8, 21, and 30 d. The molecular weight of the PLGA samples
was determined by gel permeation chromatography in THF relative to polystyrene standards on
a system comprised of two-column sets, GMHHR-M and GMHHR-H (Viscotek,
Worcestershire, UK), and a triple detector array (TDA302) at room temperature with an eluent
flow rate of 0.6 mL/min.
To determine fractional mass loss of PLGA, each sample was weighed before and after
incubation and processing. Mass loss values were corrected by the amount of mass loss at day 0
to account for losses resulting from the isolation process.
2.2.7 Detection of NT-3
The capture antibody used in the NT-3 ELISA kit binds an epitope of recombinant human NT-3
in its bioactive site. Concentrations measured by ELISA were therefore interpreted as a measure
of the tertiary structure of NT-3.
2.2.7.1 The Effect of Lyophilization, Sonication, and Nanoparticle
Fabrication on NT-3
To determine the effect of double emulsion processing on NT-3 detection, nanoparticle
formulations were dissolved and analyzed by NT-3 ELISA and BCA. The fraction of NT-3
which retained its native conformation was taken as the ratio of the encapsulation efficiency in
PLGA of NT-3 (ELISA) to that of BSA + NT-3 (BCA) under the assumption that total NT-3
and BSA encapsulation efficiencies were similar.
To study the effect of sonication on NT-3 detection, samples of 1 mL of 1 ng/mL NT-3 in aCSF
were sonicated for 5 min, 10 min, or 10 min with 400 mM trehalose at 26 W and 20 kHz to
simulate the conditions used to create the primary emulsion. The concentration of NT-3 was
then measured by ELISA.
21
To assess the effect of lyophilization on NT-3, 0.9 mL of 1 ng/mL NT-3 were combined with
different potential lyoprotectants at 400 mM in 0.5 mL aCSF in 2 mL microcentrifuge tubes,
lyophilized for 3 d, reconstituted in reagent diluent and assayed by ELISA. The following agents
were studied: trehalose, lactose, cyclodextrin, galactose, glucose, and glycerol.
2.2.7.2 Effect of pH on Bioactivity of NT-3
Samples of 1 ng/mL NT-3 were incubated for 24 h at 37 ºC in aCSF with a pH of: 7.4, 6.0, 4.0,
3.0, or 2.0. The concentration of NT-3 in the samples was then determined by ELISA and the
results normalized to the sample at pH 7.4.
2.2.7.3 Effect of Storage Conditions on Bioactivity of NT-3
The effect of different storage conditions on NT-3 bioactivity was examined by storing 1 mL of
667 pg/mL NT-3 in aCSF at 4 ⁰C or at -80 ⁰C, with and without the addition of 1 wt% BSA for
7 days. An incubation study at 37 ⁰C of 1 ng/mL NT-3 in aCSF was carried out for a total of 30
d with samples collected after: 1, 2, 4, 7, 14, 23, and 30 d.
2.2.8 NT-3 Bioactivity by Dorsal Root Ganglia (DRG) Bioassay
The bioactivity of NT-3 released over 28 d was determined using a DRG bioassay performed as
described by Hurtado et al. [61] and Blits et al. [62], with modifications described below. All
animal procedures were performed in accordance with the Guide to the Care and Use of
Experimental Animals (Canadian Council on Animal Care) and protocols were approved by the
Animal Care Committee of the Research Institute of the University Health Network. Rat embryo
DRG (E17 Female Sprague-Dawley Rats) were removed and pooled in differentiation media
comprised of neural basal media with 1 vol% fetal bovine serum, 2 vol% B-27 serum-free
supplement, 1 vol% penicillin-streptomycin, 1 vol% L-glutamine. The DRG were then placed
on 12 mm diameter glass cover slips coated with poly-D-lysine (50 μg/mL in sterile water) and
laminin (5 μg/mL in PBS) in a 24-well plate. All wells were treated with 0.5 mL of
differentiation media and 0.5 mL of the NT-3 release study supernatant, which was collected at
3, 6 h, 1, 3, 7, 14, 21, and 28 d. For the controls, 0.5 mL differential media and 0.5 mL of aCSF
with appropriate concentrations of NT-3 were added to the wells. The DRG were grown for 48 h
at 37 ⁰C and 5% CO2, and imaged using a CoolSnap HQ camera (Photometrics, Tucson, USA)
22
mounted on an Axiovert S100 microscope (Zeiss, Toronto, CA). Neurites greater than 50 μm
were counted for 10 DRG per group. Treatment groups were compared to NT-3 controls to
assess bioactivity as previously reported [61-63].
2.2.9 Statistical Analysis
All data are presented as mean ± standard deviation. For pair-wise comparison of these
averages, t-tests were carried out. For comparison of multiple groups, ANOVA comparisons
were conducted and when differences were found between groups, Bonferroni post-hoc analysis
was performed. Significance was assigned at p<0.05 unless otherwise specified.
2.3 Results
2.3.1 Effect of Embedding PLGA Nanoparticles in HAMC
The release of α-chymotrypsin, a model protein for NT-3, from the composite hydrogel was
compared to each of PLGA alone and HAMC alone (Figure 3). Release from HAMC alone was
fastest and near completion within 1 d, demonstrating a diffusion-controlled mechanism.
Release from PLGA nanoparticles showed the typical burst release within the first 2-3 d,
followed by a plateau. Unexpectedly, release from the composite hydrogel deviated significantly
from the controls, having a significantly reduced burst release followed by a linear release
profile, suggesting an interaction between PLGA and HAMC. We hypothesized that one of two
mechanisms was causing this behavior: (a) embedding the PLGA in HAMC reduced the
degradation rate of the particles, resulting in an altered release profile; or (b) the MC in HAMC
adsorbed to the surface of the PLGA particles, thereby resulting in reduced diffusion across the
PLGA-hydrogel boundary and an altered release profile.
To better understand how embedding the particles in HAMC reduced drug release, the release
data was simulated in Matlab. Release of α-chymotrypsin from a 3 mm slab of HAMC alone
was described with a one-dimensional Fickian diffusion model in Cartesian coordinates with a
good fit (R2
= 0.99) and molecular diffusivity of 8.6 x 10-7
cm2/s (Figure 3). Similarly, release
from PLGA nanoparticles in aCSF was fit using a modified one-dimensional Fickian diffusion
model in spherical coordinates (R2
= 0.96) by the following model parameters: a burst fraction
(Fburst) of 0.85, an initial diffusivity (Do) of 8.9 x 10-18
cm2/s, and a fit parameter (k) of 1. When
23
these two models were combined to simulate release from PLGA nanoparticles embedded in
HAMC, the resulting fit was poor (R2
= -1.05) and resembled a slightly delayed release from
PLGA nanoparticles. Only when the model parameters of the nanoparticles were adjusted to
Fburst= 0.3, Do=8.9 x 10-17
cm2/s, and k= 1.9 x 10
15 were the experimental data well described
(R2
= 0.99). The diffusivity values obtained using the model are in close agreement with
previously published drug diffusivity values between 8 x 10-18
cm2/s and 4 x 10
-17 cm
2/s from
PLGA nanoparticles [64]. The release kinetics of α-chymotrypsin and the fit produced by the
mathematical model were similar to release kinetics of NT-3 under identical formulation and
release conditions (Figure 3). The similarity in release profile for the model protein α-
chymotrypsin and NT-3 was expected because α-chymotrypsin and NT-3 have similar
molecular weights (25 kDa and 29 kDa) and isoelectric points (8.8 and 9.4).
To elucidate the impact of the HAMC on the degradation rate of PLGA, the changes in molar
mass (by GPC) and mass were followed over 30 d for PLGA in HAMC vs. PLGA in aCSF
buffer. As shown in Figure 4, the change in molar mass for PLGA was unaffected by the
presence of HAMC and both degradation profiles are well described by a first-order degradation
model using a degradation rate constant (kdeg) of 0.086 d-1
, in close agreement with published
values of PLGA degradation between 0.075 d-1
and 0.093 d-1
[65]. Similarly, mass loss of PLGA
particles in aCSF was indistinguishable from PLGA particles dispersed in HAMC over 30 d.
24
Figure 3- Embedding PLGA nanoparticles in HAMC reduces the burst release and
supports sustained delivery of α-chymotrypsin, an analog for NT-3. Release of α-
chymotrypsin from () PLGA nanoparticles embedded in HAMC had a lower burst release and
more sustained delivery than from () PLGA nanoparticles alone (n=3, mean ± standard
deviation). These two data sets were previously published by Baumann et al. [3]. α-
Chymotrypsin release from (Δ) HAMC alone occurs over the span of hours. A continuum model
based on Fickian diffusion was able to predict release from ( ) PLGA nanoparticles in aCSF
and from ( ) HAMC; however, a similar model that incorporated diffusion through the
HAMC gel was not able to predict release from nanoparticle embedded in HAMC ( ). Only
when model variables associated with the particles themselves were augmented was an accurate
fit obtained ( ). Release of α-Chymotrypsin and its model fit were in close agreement with
a similar formulation with encapsulated () NT-3 in PLGA nanoparticles, embedded in HAMC
(n=3, mean ± standard deviation)
.
25
Figure 4- Attenuation of the burst release from composite HAMC is not the result of
altered PLGA nanoparticle degradation. PLGA degradation was monitored over 30 d by
organic GPC for () PLGA nanoparticles in aCSF and () PLGA nanoparticles embedded in
HAMC (n=3, mean ± standard deviation). Both traces were similar to each other and to a first
order degradation model using kdeg = 0.086 days-1
( ). Mass loss for () PLGA nanoparticles
in aCSF and () PLGA nanoparticles embedded in HAMC were indistinguishable (n=3, mean ±
standard deviation).
2.3.2 NT-3 Stability Improvement
Beyond sustained release, NT-3 stability was investigated because damage to tertiary protein
structure is a common concern in polymeric sustained release devices. Excipients are often
added to PLGA particles to stabilize encapsulated proteins against the conditions encountered in
particle synthesis [34]. To maximize the fraction of NT-3 with native tertiary structure upon
release from PLGA nanoparticles, we conducted a life-cycle analysis and isolated the
fabrication steps expected to affect NT-3 structure.
2.3.2.1 Structural Damage during Nanoparticle Fabrication
We first examined the effect of excipients on NT-3 detection by ELISA after nanoparticle
fabrication. After double emulsion synthesis, 40% of the NT-3 encapsulated in PLGA
26
nanoparticles was detected (Figure 5a). Trehalose and hyaluronan were co-encapsulated
within PLGA nanoparticles because trehalose is a known lyoprotectant and hyaluronan is known
to improve encapsulated protein stability by increasing the viscosity of the inner aqueous phase
of the emulsion, reducing contact between the protein and the organic solvent during processing
[41]. Surprisingly, there was no improvement in NT-3 detection after co-encapsulation, likely
because PLGA masks any improvement by acting through the same stabilization mechanisms as
trehalose and hyaluronan. Magnesium carbonate was also screened as a co-encapsulant to
improve the long-term stability of NT-3, where the carbonate ion is known to buffer the acidic
degradation products of PLGA [66]. Magnesium carbonate did not affect initial NT-3 stability,
as expected, because the acidic microenvironment within PLGA takes time to develop. Neither
trehalose + hyaluronan nor magnesium carbonate impacted the amount of NT-3 detected;
however, addition of PEG 400 had a profound impact, with 74% of NT-3 detected, likely
because PEG 400 is a surfactant that minimizes contact between NT-3 in the inner aqueous
phase and the organic phase during the high energy sonication required for particle formation
[67].
The effects of sonication and lyophilization on NT-3, two key operations used to create
nanoparticles in the double emulsion procedure, were then investigated. NT-3 was particularly
sensitive to sonication, as only 32% and 25% of NT-3 was detected after sonicating for 5 and 10
min, respectively (Figure 5b). Sonication of an NT-3 solution containing 400 mM trehalose for
10 min yielded 39% detection (compared to 25% without trehalose), an effect which is likely
caused by the reduction in cavitation associated with the increased viscosity afforded by
trehalose [68]. NT-3 was also highly susceptible to lyophilization. In aCSF, only 3% of the
lyophilized NT-3 was measured by ELISA. Of the several lyoprotectants investigated (Figure
5c), only 400 mM trehalose was able to improve NT-3 detection from 3% to 22%, whereas all
other agents yielded detection below 10% (Figure 5c). The most likely mechanism for this
protective effect is that trehalose is able to satisfy the hydrogen bonding requirements of polar
groups on the exposed surface of NT-3 after the sublimation of water [69]
27
Figure 5 – The effects of three processing steps on the stability of NT-3 were investigated
by ELISA. (a) Encapsulation of NT-3 within PLGA nanoparticles stabilized the protein during
the double emulsion synthesis, retaining approximately 40% NT-3 detectability using the
following co-encapsulants: trehalose + hyaluronan, MgCO3, or no additives. Co-encapsulated
PEG 400 significantly improved NT-3 stability (p<0.001, n=3, mean ± standard deviation),
resulting in 74% detection after processing. (b) After sonication 400 mM trehalose significantly
improved detectability from 25% to 39% (p<0.001, n=3, mean ± standard deviation). (c) The
addition of 400 mM trehalose prior to lyophilization improved NT-3 detectability significantly
compared to all other additives (p<0.001, n=3, mean ± standard deviation).
a) b)
c)
a) b)
c)
28
2.3.2.2 Structural Damage during Drug Release
The effects of pH and long-term incubation on NT-3 were studied to determine whether the
tertiary structure was sensitive to environments encountered during in vitro release studies. NT-
3 dissolved in aCSF and incubated at 37 ⁰C was stable for 24 h at pH 7.4. Moreover, 80% of
NT-3 was detected between pH 3 and pH 6 for 24 h; however, at pH 2, effectively all NT-3 was
denatured, as only 1% of the initial concentration was detected by ELISA (Figure 6a). During
release studies, NT-3 diffused from composite HAMC into the supernatant where it remained at
pH 7.4 and 37 ⁰C for up to 7 d (the maximum interval between sampling). Under these
conditions, 80% of NT-3 was detected after 7 d. This value decreased to 35% after 23 d (Figure
6b).
Figure 6– NT-3 was not detected by ELISA after exposure to low pH and steadily lost
detectability after incubation at 37⁰C. (a) NT-3 detectability by ELISA was fairly stable
between pH 7.4 and pH 3 when incubated for 24 h; however, below pH 3 NT-3 was not
detected(n=3, mean ± standard deviation). (b) NT-3 steadily lost ELISA detectability at
approximately 2.5% per day over the first 23 d of incubation at 37 ⁰C in aCSF (n=3, mean ±
standard deviation).
2.3.2.3 Structural Damage during Storage
Samples of NT-3 containing 0.1 wt% BSA in solution were stored for 7 d at 4 ⁰C and at -80 ⁰C
with and without the addition of additional 1 wt% BSA. Storage at both temperatures largely
maintained NT-3 conformation over this time period, but the addition of 1 wt% BSA
significantly reduced the detection of the NT-3 (Figure 7). BSA was added to the NT-3 based
a) b)
29
on the assumption that BSA would improve the stability of NT-3 by preventing adsorption
onto the surface of the storage vessel; however, surprisingly, approximately half of the NT-3
originally detected was lost. This was true at both 4 and -80 °C. Importantly, we know that the
additional BSA did not block the ELISA capture antibody because when 1 wt% BSA was added
to a fresh sample of NT-3, all of the original NT-3 was detected. Given that globular proteins
are known to aggregate in the presence of salts [70], and that BSA is known to aggregate
through thiol-disulfide interchange reaction [71], it is possible that BSA and NT-3 aggregated
faster at a higher BSA concentration, thereby accounting for the reduced NT-3 detected.
Figure 7 – NT-3 stored at 4 ⁰C or -80 ⁰C remained stable, but not when stored at 4ºC with
1 wt% BSA for 7 d. When NT-3 was stored in aCSF for 7 d at -80 ⁰C, the NT-3 concentration
measured by ELISA was similar to the initial concentration. Similarly storage at 4 ⁰C only
resulted in a modest 19% loss in detection compared to the initial concentration (p<0.05, n=3,
mean ± standard deviation). However, when stored at these temperatures in the presence of 1
wt% BSA, more than half of the initial NT-3 detected was lost (p<0.001, n=3, mean ± standard
deviation). A fresh sample in 1 wt% BSA (Initial Concentration + BSA) did not exhibit this
same loss in detection, which indicates that this phenomenon is not simply due to the BSA
blocking the ELISA plate.
30
2.3.3 Effect of Processing Parameters on NT-3 Release Kinetics
Excipients can alter release kinetics by acting as pore-forming agents or by changing the particle
formation process [72]. Consequently, the effects of additives on NT-3 release kinetics from
PLGA nanoparticles embedded in HAMC were investigated. NT-3 was individually co-
encapsulated with one of: trehalose and hyaluronan, MgCO3, or PEG 400. In each synthesis,
particle yield was between 67% and 85% and nanoparticles were 200-300 nm in diameter. None
of the excipients tested were found to affect nanoparticle yield or diameter; however, the
encapsulation efficiency was affected. Total protein encapsulation was: 97% without an
excipient, 98% when co-encapsulated with PEG 400, 70% with MgCO3, and 30% with trehalose
and hyaluronan. These encapsulation efficiencies correspond to total protein loadings of: 8.8
wt%, 8.9 wt%, 6.4 wt%, and 2.7 wt%, respectively.
Notwithstanding significant differences in encapsulation efficiency, the release profile of NT-3
from PLGA nanospheres dispersed in HAMC when co-encapsulated with trehalose +
hyaluronan vs. excipient-free control nanoparticles was similar, with 63% ± 9% and 57% ± 10%
cumulative release after 28 d, respectively (Figure 8a). Interestingly, NT-3 co-encapsulated
with either MgCO3 or PEG 400 resulted in a lower cumulative release of only 28% ± 10% over
28 d, with very little additional NT-3 released from co-encapsulated PEG 400 after 7 d. The
total mass of NT-3 released was 113 ± 19 ng NT-3 / mg particle for the formulation with no
additives, 34 ± 5 ng NT-3 / mg particle for the trehalose + hyaluronan formulation, 105 ± 15 ng
NT-3 / mg particle for the PEG 400 formulation, and 36 ± 6 ng NT-3 / mg particle for the
MgCO3 batch (Figure 8b). These data suggest that the most NT-3 released over a 28 d period is
achieved from PLGA nanoparticles encapsulated with no additives.
31
Figure 8 – NT-3 in vitro release from PLGA nanoparticles was fine-tuned by incorporating
excipients and adjusting polymer properties. (a) NT-3 release from PLGA nanoparticles
embedded in () HAMC was not considerably changed by co-encapsulation with () trehalose
and hyaluronan. Co-encapsulation with () MgCO3 resulted in a reduced burst and reduced
cumulative release of NT-3. Co-encapsulation with (Δ) PEG 400 led to a 7 d release profile,
with only 1% released thereafter (n=3, mean ± standard deviation). (b) Release amounts of NT-3
per mg of PLGA nanoparticle for all four formulations, as measured by ELISA, shows the
largest release amount from PLGA alone and PLGA with PEG 400 (n=3, mean ± standard
deviation). This demonstrates that approximately 100 ng of NT-3 can be delivered over 7 d from
the formulation with PEG 400, while 110 ng of NT-3 can be delivered over 28 d from the
formulation without additives.
2.3.4 In Vitro Bioactivity
The bioactivity of NT-3 released from composite HAMC was measured by neurite outgrowth
from embryonic rat DRGs. These DRG demonstrated a dose response to NT-3 between 1-100
ng/mL where the number of neurites increased with increasing NT-3 concentration, which is
consistent with previous reports [73]. At days 1, 14, and 28, all nanoparticle formulations
promoted neurite outgrowth except for the formulation that co-encapsulated PEG 400 (Figure
9). In this case, no neurite outgrowth was observed at 28 d because there was no NT-3 released
between 14 and 28 d for this formulation, which is substantiated by the ELISA data.
a) b) a) b)
32
The control nanoparticle formulation and the batch that had trehalose and hyaluronan co-
encapsulated were statistically indistinguishable at all time points, which suggests that any
bioactivity improvement caused by these additives was marginal and undetectable by the DRG
bioassay. The batch containing PEG 400 also elicited similar DRG neurite outgrowth up to 14 d
compared to the aforementioned samples, but at 28 d the bioactivity of the NT-3 in the
supernatant was identical to the control. The nanoparticle batch containing co-encapsulated
MgCO3 exhibited significantly more robust neurite outgrowth than all other nanoparticle batches
at all time points, which demonstrates that even with the reduced amount of NT-3 released,
more of it is bioactive. This elucidates the importance of following bioactivity (and not just the
amount) of released proteins.
Figure 9 – Released NT-3 is bioactive in a rat dorsal root ganglia neurite outgrowth assay.
(a) NT-3 standards in 0.5 mL aCSF and 0.5 mL differentiation media. The increase in average
number of neurites/DRG with increased NT-3 suggests a correlation in amount of NT-3 present
and number of neurites. (b) The NT-3 released from PLGA nanoparticles was followed in terms
of the following co-encapsulants: ( ) no additives, ( ) trehalose and hyaluronan, () PEG
400, and () MgCO3. All samples up to 28 days stimulated neurite outgrowth from rat dorsal
root ganglia, with the exception of the PEG 400 batch at day 28. Batches with co-encapsulated
MgCO3 exhibited more robust neurite outgrowth with significant differences relative to all other
variables at 1d, 14 d, and 28 d (p<0.001, n=10).
a) b)
33
2.4 Discussion
NT-3 is a neuroregenerative protein that modulates the maintenance, proliferation, and
differentiation of neurons that express TrkC receptors [11] and has been investigated pre-
clinically as a treatment for spinal cord injury and stroke. Sustained release of NT-3 has resulted
in improved tissue and functional benefit relative to instantaneous release in the treatment of
spinal cord injury, where NT-3 was released for 14 d [20] and 28 d [21]. With reference to the
lack of FDA-approved methods for localized and sustained release for treatment of acute SCI
we sought to formulate our experimental drug delivery system [27] for release of NT-3.
A principal challenge of formulating proteins for sustained release is maintaining bioactivity
over the desired treatment term. There is substantial literature on sustained release from PLGA
particles [43], and while sustained NT-3 release from PLGA microparticles embedded in a PEG
gel has been published, bioactivity was not measured [17]. One property of PLGA particles that
is well described is the formation of an acidic microenvironment within the particle during
hydrolytic degradation [66]. We were concerned that embedding PLGA particles in HAMC may
increase the acidity of the PLGA environment by impeding the diffusive release of acidic
oligomers. To better understand the interaction between PLGA particles and HAMC, we fitted
experimental release data using a mathematical model and monitored the degradation kinetics
and mass loss of PLGA.
The release of dissolved α-chymotrypsin from HAMC alone was well described by Fickian
diffusion in Figure 3 and release from PLGA particles was likewise well described by a one-
dimensional Fickian diffusion model in spherical coordinates with a time–dependent diffusivity
term. However, the release of encapsulated α-chymotrypsin from composite HAMC could not
be described using a model incorporating the best fit parameters from the freely-suspended
particles and Fickian diffusion in HAMC, suggesting diffusion through the PLGA nanoparticles
and HAMC are not distinct processes. This model only accurately described the composite
HAMC release profile of α-chymotrypsin and NT-3 after the particle model parameters were
adjusted dramatically. The revised model parameters suggest that because such a large empirical
fit parameter is required to describe release from the system, there is another physical
mechanism controlling release besides the diffusion and particle degradation considered by the
model. Together with the observation in Figure 4 that the molecular weight and mass loss of
34
PLGA are not influenced by the presence of HAMC, and that the mathematical models
demonstrate that diffusion through HAMC is not the source of delayed release, we suggest that
an interaction between HAMC and PLGA nanoparticles is creating a diffusive barrier at the
particle-hydrogel interface. Since methyl cellulose gels through hydrophobic junctions [74] and
methyl cellulose and PLGA particles associate through hydrophobic interactions over the span
of hours [26], we propose that this association slows protein diffusion through a restrictive
membrane-like mechanism, thereby attenuating release [75, 76]. This interpretation suggests
that an enhanced sustained release profile is possible for a wide range of molecules from
composite HAMC relative to freely suspended PLGA particles because the mechanism is
independent of the properties of the drug. This view is supported by the report of attenuated
release of PLGA-encapsulated dbcAMP (Mw = 0.46 kDa), EGF (Mw = 6.2 kDa), and IgG (Mw
= 150 kDa) from composite HAMC [27]. This explanation is further strengthened by the
identical degradation rate of PLGA whether dispersed in HAMC or not. Importantly, the similar
degradation profile elucidated in Figure 4 for PLGA nanospheres in HAMC vs. in buffer,
suggests that the environment within PLGA particles in composite HAMC is similar to that of
PLGA in suspension. The improved sustained release of α-chymotrypsin from composite
HAMC relative to PLGA alone and the conclusion that the particle microenvironment was
similar in both cases led us to formulate NT-3 in composite HAMC. Since this system
demonstrated the capacity for sustained release, we next sought to assess the stability of NT-3,
which is a concern in PLGA particles [34].
We explored strategies to improve the stability of NT-3 during processes associated with
nanoparticle fabrication, drug release, and storage. The process of encapsulation in PLGA
negatively affects protein stability at multiple points in the synthesis to the extent that only 40%
of encapsulated NT-3 was detected in initial particle formulations (Figure 5a). Testing the
effects of sonication and lyophilization on NT-3 in solution failed to isolate the source of
bioactivity loss because these operations damaged NT-3 more than encapsulation itself (Figure
5b,c), indicating that PLGA is an important stabilizer for NT-3. Dissolved PLGA may preserve
protein activity by increasing solution viscosity during sonication [68] and providing a hydrogen
bonding partner for encapsulated proteins during lyophilization [69].
Considering the NT-3 release profiles (Figure 8) and the corresponding NT-3 bioactivity data
(Figure 9), it is clear that the nanoparticle formulation with no excipients demonstrated
35
sustained, bioactive release over 28 d, greater than that observed by co-encapsulation with
trehalose and hyaluronan. These data suggest that any pores formed by these co-encapsulants
are likely too small or too poorly interconnected to increase NT-3 release rate, as has been
observed in other systems [72]. ELISA and DRG results indicate that co-encapsulation of PEG
400 reduced the duration of bioactive NT-3 release from 28 d to 14 d, with the bulk of release
occurring in the first 7 d. This reduced release duration precluded PEG 400 from being co-
encapsulated within formulations optimized for more sustained delivery, such as the MgCO3
preparation, which demonstrated improved bioactivity up to 28 d compared to the control.
Interestingly, 28 d release profiles using PEG 400 as a co-encapsulant has only been previously
reported in significantly larger, 30 μm PLGA microspheres for another neurotrophin, NGF [67].
Release can be extended from these larger particles because their surface area to volume ratio is
150 times smaller than the 200 nm particles used in the current work. This smaller ratio slows
water uptake, drug diffusion, and matrix degradation.
Co-encapsulation with MgCO3 reduced the released fraction of NT-3 and significantly improved
its bioactivity over 28 d relative to all other groups. Magnesium carbonate crystals located near
the surface of the PLGA particles likely contribute to the increased burst release of NT-3, as this
salt can rapidly dissolve in solution, resulting in pore formation near the surface of the particles.
The microenvironment in PLGA particles has been reported to be as low as pH 1.5 to pH 3 due
to the acidic oligomers produced by the hydrolytic degradation of ester bonds in PLGA [77], an
environment in which NT-3 was shown to be particularly sensitive. The release profile and
bioactivity of NT-3 from PLGA nanoparticles co-encapsulated with MgCO3 suggest that
MgCO3 neutralizes acidic PLGA degradation products in the same way as has been reported for
microspheres [78]. The autocatalytic degradation mechanism of PLGA is slowed, thereby
reducing the rate of NT-3 release. Given the susceptibly of NT-3 to low pH, the MgCO3 likely
maintains a higher pH within the PLGA nanoparticles than would be expected in PLGA alone.
These results demonstrate that sustained release of bioactive NT-3 can be achieved from the
proposed composite HAMC drug delivery system. It was discovered that the methyl cellulose in
HAMC adheres to dispersed PLGA particles through hydrophobic interactions, a previously
unknown mechanism. This understanding could lead to the sustained delivery of a wide range of
hydrophilic molecules from this system without disturbing the degradation kinetics of the
PLGA. Co-encapsulation of PEG 400 significantly improved NT-3 detection immediately after
36
encapsulation, highlighting the value of this additive during the processing. Yet, PEG 400
decreased the release duration of NT-3 from 28 d to 7 d, which prevented its addition to other
long-term release formulations. The trehalose + hyaluronan formulation was also eliminated
from further evaluation because it was outperformed by the excipient-free formulation.
Specifically, the excipient-free formulation had a higher total release amount (Figure 8b) with
similar release kinetics (Figure 8a) and bioactivity (Figure 9). Co-encapsulated MgCO3
dramatically enhanced the bioactivity of NT-3 up to 28 d, which was attributed to the ability of
MgCO3 to neutralize the low pH inside PLGA particles, an environment in which NT-3 was
shown to be particularly vulnerable to structural damage. Since the MgCO3 formulation
demonstrated measurable release over 28 d and elicited the most neurite outgrowth from DRG,
it is the preferred preparation for 28 d delivery. Importantly, the proposed drug delivery system
offers a total deliverable NT-3 amount that is comparable to similar systems, including NT-3
delivery from a fibrin gel [79] and from microtubes embedded in an agarose gel [14]. We are
encouraged by the previously reported in vivo biocompatibility of this system [26] in addition to
the pharmacologically relevant dose, and the sustained and bioactive release of NT-3 in the
MgCO3 formulation. Future studies will assess the efficacy of this formulation in vivo.
37
3 In Vitro Sustained Release of Bioactive Anti-NogoA, a
Molecule in Clinical Development for Treatment of
Spinal Cord Injury
3.1 Introduction
Spinal cord injury affects 130,000 0people each year worldwide [1] and often results in
permanent sensory and motor deficiencies. One promising treatment option involves the
delivery of anti-NogoA, an antagonist of NogoA, which is a myelin inhibitor known to cause
growth cone collapse and reduce neurite outgrowth [22]. Anti-NogoA has been shown to
improve functional recovery in rat models when delivered by an intrathecal catheter over 14 d
[23] or 28 d [24] and is being studied clinically [7]. To avoid the blood-spinal cord barrier in
clinical trials, osmotic minipumps have been used; however, external minipumps are prone to
infection [25], which provides the motivation to develop technology that is capable of safely
delivering this promising molecule over sustained durations to patients.
We have developed a drug delivery system that consists of drug-loaded poly(lactic-co-glycolic
acid) (PLGA) nanoparticles embedded in a hydrogel of hyaluronan and methyl cellulose
(HAMC). The nanoparticles are formed by double-emulsion synthesis and slow the rate of drug
release, while the hydrogel localizes the particles at the site of injury in the intrathecal space.
This approach is minimally invasive, biocompatible over 28 d [26], and has been shown to
release dbcAMP, EGF, α-chymotrypsin, and IgG over 28 d in vitro [27] and fibroblast growth
factor 2 over 24 hours in vivo [28]. Achieving sustained delivery of anti-NogoA from PLGA
particles that remains bioactive over the release duration is a primary concern.
Proteins in general are susceptible to conformational damages caused by the deleterious
environments associated with processing or release. For example, Han et al. observed
degradation and aggregation of recombinant human serum albumin after lyophilization, a
problem that was only avoided with the addition of sugar excipients [80], which are known to
stabilize proteins by satisfying the hydrogen bonding requirements of the protein after water
sublimation [69]. Proteins encapsulated within PLGA particles are also known to become
damaged because of exposure to the water-organic solvent interface and the acidic environment
38
associated with PLGA particles [34]. The problem raised by the presence of the organic solvent
has been addressed with co-encapsulated hyaluronan, which has been suggested to improve
protein stability by creating a viscous microenvironment that slows the interaction between the
protein and the water/organic solvent interface [41]. Co-encapsulated bases have been shown to
neutralize the acidic environment inside PLGA particles, and have stabilized encapsulated
bovine serum albumin [78]. These additives, however, can affect the release kinetics of the
encapsulated proteins. Water soluble additives can act as porogens, which results in accelerated
drug release because of the formation of new pore networks upon particle wetting [72]. Co-
encapsulated bases slow acid-catalyzed PLGA degradation, which retards long-term protein
delivery [81]. With this in mind, it is important that strategies used to improve bioactivity also
be engineered to achieve suitable release kinetics.
In this chapter, we explore additive formulations for improving anti-NogoA bioactivity without
sacrificing sustained release kinetics. The effects of co-encapsulating several additives on anti-
NogoA bioactivity, encapsulation efficiency, and release kinetics were studied. Formulations
were based on combinations of the following excipients: trehalose, hyaluronan, MgCO3, and
CaCO3.
3.2 Materials and Methods
3.2.1 Materials
The anti-NogoA mAb 11c7 was generously donated by Novartis (Basel, CH). Trehalose,
MgCO3, sodium dodecyl sulfate, poly(DL-lactic-co-glycolic acid) 50:50 of inherent viscosity
0.15-0.25 dL/g, and IgG from human serum of reagent grade were purchased from Sigma-
Aldrich (Oakville, CA). Poly(vinyl alcohol), 6000 g/mol,) was purchased from Polysciences Inc.
(Warrington, USA). Sodium hyaluronate, 2600 kg/mol was purchased from Lifecore (Chaska,
USA). Methyl cellulose, 300 kg/mol, was purchased from Shin-Etsu (Tokyo, Japan). Sodium
hydroxide was purchased from EMD Chemicals (Gibbstown, USA). Pluronic F-127 was
purchased from BASF (Missisauga, CA).
Artificial cerebrospinal fluid (aCSF) at pH of 7.4 was prepared as described by Gupta et al. [60].
HPLC grade dichloromethane (DCM), dimethyl sulfoxide (DMSO), and hydrochloric acid
(HCl) were purchased from Caledon Labs (Georgetown, CA). Dulbecco‟s phosphate buffered
39
saline (pH 7.4, 9.55 g/L) was purchased from Wisent Inc. (St-Bruno, CA). All buffers were
prepared using water distilled and deionized using a Millipore Milli-RO 10 Plus and Milli-Q UF
Plus at 18 MΩ resistance (Millipore, Bedford, USA).
3.2.2 Nanoparticle Processing and Hydrogel Preparation
Nanoparticles loaded with anti-NogoA were produced using a water/oil/water (w/o/w) double
emulsion solvent evaporation technique, described elsewhere [27]. Briefly, an inner aqueous
phase of 178 μL aCSF containing 0.72 mg anti-NogoA and 1.36 mg IgG was mixed with an
organic phase of 1.6 mL DCM, 80 mg PLGA and 0.8 mg Pluronic F-127. This mixture was
sonicated using a Vibra-Cell (Sonics, Newtown, USA) on ice for 10 minutes at 26 watts and 20
kHz to create the primary emulsion, which was subsequently mixed with the outer aqueous
phase of 5.3 mL of 25 mg/mL PVA. The secondary emulsion was formed through sonication on
ice for an additional 10 minutes at 39 watts and 20 kHz. This double emulsion was then added
to 53 mL of 25 mg/mL PVA and stirred gently for 20 hours at room temperature. PLGA
nanoparticles were isolated, washed 4 times by ultracentrifugation (Beckman, Missisauga, CA),
lyophilized (Labconco, Kansas City, USA), and stored at -20 ⁰C. In syntheses with co-
encapsulated trehalose, 25 mg trehalose was added to the inner aqueous phase of the emulsion in
place of IgG. In the batches containing trehalose and hyaluronan, 25 mg trehalose and 1.8 mg
hyaluronan were added to the inner aqueous phase. In the formulations which included base, 35
mg CaCO3 or 6 mg MgCO3 were added to the organic phase in addition to 25 mg trehalose and
1.8 mg hyaluronan in the inner aqueous phase.
HAMC hydrogels were prepared through the physical blending of hyaluronan and methyl
cellulose in aCSF to achieve a final composition of 1 wt% 2600 kg/mol hyaluronan and 3 wt%
300 kg/mol methyl cellulose after addition of the PLGA nanoparticles. Methyl cellulose was
first dispersed in the aCSF using a dual asymmetric centrifugal mixer (Flacktek Inc., Landrum,
USA) and left to dissolve overnight at 4 ⁰C, followed by hyaluronan which was dissolved in the
same manner.
3.2.3 Particle Characterization
Particle size was measured using dynamic light scattering (Zetasizer Nano ZS, Malvern
Instruments, Malvern, UK). Particle yield was defined as the total mass of particles produced
40
divided by mass of the initial mass of PLGA used, adjusted for protein content. Drug loading is
the mass fraction of anti-NogoA in the particles and encapsulation efficiency is the measured
protein loading of the particles divided by the theoretical maximum drug loading. To determine
the total protein encapsulation efficiency, 1 mg nanoparticles were dissolved in 5 mL DMSO
and added to 5 mL of 0.05 M NaOH containing 0.05 wt% SDS and analyzed using the total
protein BCA assay according to the manufacturer‟s instructions (Thermo Scientific, Nepean,
CA). To determine anti-NogoA encapsulation efficiency, 1 mg of particles was dissolved in 1
mL DCM for 1 hour. The protein was then extracted into a liquid phase of 10.5 mL reagent
diluent and analyzed using an anti-NogoA ELISA.
3.2.4 Drug Release Studies
Release profiles of anti-NogoA from each formulation weere obtained by dispersing 10 mg of
particles in 0.1 mL of concentrated HAMC in a 2 mL microcentrifuge tube (Axygen, Union
City, USA) using a dual asymmetric centrifugal mixer at 3300 rpm for 4 minutes to produce a
final composition of 8 wt% particles, 1 wt% hyaluronan, and 3 wt% methyl cellulose. The
composite was then warmed to 37 ⁰C and 0.9 mL pre-warmed aCSF was added to the sample
tubes. The supernatant was removed and replaced completely at 3 and 6 hours, and 1, 3, 7, 14,
21, and 28 d. The protein content of the supernatant was determined by BCA assay the bioactive
11c7 concentration by ELISA. After the drug release studies were stopped, the unreleased drug
inside the particles was quantified by dissolving the remaining particles in HAMC in 0.1 mL
DCM and extracting the remaining protein into 1 mL reagent diluent for protein quantification
by BCA assay. Bioactive anti-NogoA was measured by a custom ELISA that uses a fragment of
NogoA containing the sequence against which 11c7 was raised as the capture antibody, ensuring
that only biologically active anti-NogoA is detected [82].
3.2.5 Mathematical model
A mathematical model constructed in Matlab (MathWorks, Natick, USA) was used to
quantitatively describe the effect of various processing parameters on the kinetics of anti-NogoA
release. Based on the models developed by Faisant et al. [57] and Raman et al. [56], with minor
modifications (Appendix A), release from composite HAMC was simulated in two parts:
release from PLGA particles was simulated using a one-dimensional Fickian diffusion model in
41
spherical coordinates and release from the HAMC hydrogel was simulated using a one-
dimensional Fickian diffusion model in Cartesian coordinates.
3.2.6 Statistical Analysis
All data are presented as mean ± standard deviation. To assess statistical differences between
these averages Student‟s t-tests were conducted and significance was assigned at p<0.05 unless
otherwise specified.
3.3 Results
3.3.1 Anti-NogoA bioactivity was enhanced by trehalose and
hyaluronan, but unaffected by co-encapsulated bases relative to no co-
encapsulants
Various additives were co-encapsulated within PLGA nanoparticles and the bioactive fraction of
the anti-NogoA released in vitro was calculated by comparing the amount measured by ELISA
vs. that measured by BCA. Co-encapsulated trehalose improved the bioactivity of released anti-
NogoA significantly at 1, 2, and 7 days compared to PLGA nanoparticles with no co-
encapsulants (Figure 10). When both traces were fitted using a first-order degradation model,
the initial bioactivity (F0) of the trehalose formulation was 100% compared to 55% for the
control formulation. The first-order degradation term was similar for both cases, 0.010 hours-1
and 0.012 hours-1
for the trehalose and control formulations respectively, as reported in Table 3.
42
Figure 10– Co-encapsulated trehalose significantly improves the initial bioactivity of
released anti-Nogo-A. (a) The percentage of anti-NogoA that is bioactive during release is
significantly higher (p<0.05, n=3, mean ± standard deviation) at 1, 2, and 7 days when ()
trehalose is co-encapsulated with anti-NogoA compared to a formulation with () no additives.
At 14, 21, and 28 d, there was no measurable bioactivity for either formulation. A first-order
bioactivity loss model was used to simulate anti-NogoA bioactivity for ( ) co-encapsulated
trehalose and ( ) no additives. (b) The first 7 d of data were plotted on a semi-log plot to
demonstrate that the improvement to bioactivity is a result of increased initial bioactivity, rather
than a change in the rate of bioactivity loss.
Table 3 – A summary of the first-order bioactivity loss model parameters for the five
formulations described in Figure 10, Figure 11, and Figure 12.
F = F0e-kt
Additive F0 k (hours-1
)
None 0.55 0.012
Trehalose (T) 1.00 0.010
T + Hyaluronan (H) 1.00 0.011
T + H + CaCO3 1.00 0.011
T + H + MgCO3 1.00 0.017
a) b)
43
When hyaluronan and trehalose were both co-encapsulated within PLGA nanoparticles, 11c7
bioactivity was maintained at early time points as in the trehalose only formulation, and
significantly improved at 14, 21, and 28 (Figure 11).
Figure 11– Co-encapsulated hyaluronan with trehalose significant improves bioactivity of
released anti-NogoA at late time points. (a) Anti-NogoA bioactivity is similar over the first 7
d comparing () co-encapsulated trehalose to () co-encapsulated trehalose and hyaluronan,
but the latter formulation has significantly higher (p<0.05, n=3, mean ± standard deviation)
bioactivity at 14, 21, and 28 d. First-order bioactivity loss models for ( ) co-encapsulated
trehalose and ( ) co-encapsulated hyaluronan were plotted. (b) The bioactivity data was
plotted on a semi-log plot to illustrate the improvement to bioactivity garnered by () co-
encapsulating trehalose and hyaluronan. The first-order model for anti-NogoA bioactivity from
a formulation with ( ) co-encapsulated trehalose was only taken out to 7 d because bioactivity
for this formulation was undetectable at 14 d and beyond.
Anti-NogoA bioactivity over the first 3 days of release was not influenced by the presence of the
co-encapsulated bases in addition to co-encapsulated trehalose + hyaluronan (Figure 12). At
day 7, however, no additional bioactive anti-NogoA could be detected by ELISA. The initial
bioactivity F0 was 100% for all three formulations. The first-order degradation term was similar
for all cases at 0.011 hours-1
, 0.017 hours-1
, and 0.011 hours-1
for the CaCO3, MgCO3, and no
base formulations, respectively, as outlined in Table 3.
a) b)
44
Figure 12 – Anti-NogoA bioactivity is similar at early time points with and without co-
encapsulated bases. (a) Trehalose + Hyaluronan nanoparticle formulations with () co-
encapsulated CaCO3 or () co-encapsulated MgCO3 demonstrated similar bioactivity up to 3 d
compared to a formulation with () co-encapsulated trehalose and hyaluronan (n=3, mean ±
standard deviation). There was no detectable bioactive anti-NogoA for the base-encapsulated
formulations at from 7 d onward. First-order bioactivity loss models were identical for ( ) co-
encapsulated CaCO3 and the no base formulation, which were also similar to ( ) co-
encapsulated MgCO3. (b) A semi-log plot of bioactivity data up to 28 d demonstrates that co-
encapsulated bases do no alter early bioactivity, but surprisingly do not improve anti-NogoA
bioactivity at later time points.
3.3.2 Anti-NogoA release kinetics were influenced by trehalose and
hyaluronan together and the presence of bases, but not by trehalose alone
Anti-NogoA release kinetics were followed by BCA assay to assess the impact of bioactivity-
preserving excipients. Co-encapsulated trehalose did not affect anti-NogoA release kinetics over
77 days compared to a control without any co-encapsulants (Figure 13). In both cases, a burst of
10% was observed over the first 3 days, followed by a slower release phase, which resulted in
21% release amount after 77 days. When hyaluronan and trehalose were co-encapsulated, anti-
NogoA was released at a faster rate over 54 days compared to formulations without hyaluronan.
a) b) a)
45
This release profile was characterized by a 22% burst release over the first 3 days, followed by
a slower release rate up to 54 days, at which point 66% of the initial anti-NogoA was released.
The formulation with no additives was fit adequately (R2 = 0.96) with the following model
parameters: initial diffusivity (D0) = 2 x 10-17
cm2/s, burst fraction (Fburst) = 17%, and
degradation fit constant (k) = 0.012. The trehalose formulation was also simulated (R2=0.98)
with the following fit parameters: D0 = 3 x 10-17
cm2/s, Fburst = 16%, k = 0.005. Co-
coencapsulated hyaluronan with trehalose was fitted (R2 = 0.99) with the following parameters:
D0 = 2 x 10-16
cm2/s, Fburst = 21%, k = 0.055. Encapsulation efficiencies were measured for these
three formulations: 97% for no additives, 89% for trehalose, 43% for trehalose and hyaluronan,
as summarized in Table 4.
46
Figure 13 – Co-encapsulated trehalose does not influence anti-NogoA release kinetics,
while hyaluronan and trehalose enhance sustained anti-NogoA delivery. When ()
trehalose was co-encapsulated in a formulation, a total anti-NogoA release profile was obtained
similar to an () additive-free formulation. On the other hand, () co-encapsulated trehalose
and hyaluronan increased the burst amount and long-term release rate of anti-NogoA (n=3,
mean ± standard deviation). All traces in this figure are simulations developed using the model,
parameters available in Table 4.
Table 4 – Mathematical model parameters and particle characterization for selected
formulations
Co-encapsulants Fburst
(%)
D0 (cm2/s) k R
2
No additives
17 2 x 10-17
0.012 0.96
Trehalose (T)
16 3 x 10-17
0.005 0.98
T + Hyaluronan (H)
21 2 x 10-16
0.055 0.99
T + H + CaCO3
19 2 x 10-17
0.005 0.95
T + H + MgCO3 11 3 x 10-17
0.035 0.99
47
MgCO3 and CaCO3 were co-encapsulated within PLGA nanoparticles with trehalose and
hyaluronan and in vitro release kinetics and bioactivity were measured. Reduced release kinetics
were observed in the samples with bases (Figure 14). A 7%, 15%, and 22% burst release over
the first 3 days was observed for the MgCO3, CaCO3, and base-free formulations, respectively.
Both co-encapsulated base formulations demonstrated reduced long-term delivery; compared to
the 66% release after 54 days in the base-free formulation, only 19% and 23% were released in
the same time frame for the MgCO3 and CaCO3 formulations, respectively. The formulation
without co-encapsulated bases was fitted (R2 = 0.99) with the following parameters: D0 = 2 x 10
-
16 cm
2/s, Fburst = 21%, k = 0.055. The MgCO3 formulation was fitted (R
2 = 0.99) with the
following parameters: D0 = 3 x 10-17
cm2/s, Fburst = 11%, k = 0.035. The CaCO3 formulation was
fitted (R2 = 0.95) with the following parameters: D0 = 2 x 10
-17 cm
2/s, Fburst = 19%, k = 0.005.
Encapsulation efficiencies were measured for each of the formulations: base-free was 43%,
MgCO3 was 80%, and CaCO3 was 90%, as summarized in Table 4.
48
Figure 14 – Co-encapsulated bases reduce the release rate of anti-NogoA. When ()
CaCO3 or () MgCO3 were co-encapsulated with anti-NogoA, trehalose and hyaluronan, the
total anti-NogoA release profiles were dramatically reduced compared to () co-encapsulated
trehalose and hyaluronan (n=3, mean ± standard deviation). All traces in this figure are
simulations developed using the model, parameters available in Table 4.
3.4 Discussion
Spinal cord injury is a devastating condition, which results in permanent sensory or motor
deficiencies as a result of neuronal tissue damage. Although the regenerative capacity of the
central nervous system when presented with a suitable environment has been known for almost
thirty years [83], there is no cure for this condition. One promising approach is the delivery of
anti-NogoA, which has been shown to improve functional recovery in rat models and is
currently being evaluated clinically. This molecule is known to act as an antagonist for NogoA,
a myelin-associate inhibitory protein, which has been shown to cause growth cone collapse and
inhibit neuronal outgrowth. All reports that exhibit functional recovery with anti-NogoA require
sustained release between 14 and 28 d and use a minipump-catheter system; however, these
systems are prone to infection over long time periods [25]. Consequently, we developed an
49
injectable composite drug delivery system that is capable of delivering regenerative proteins to
the intrathecal space, while being minimally invasive, biodegradable, and biocompatible. It was
previously unknown whether anti-NogoA could be released up to 14 or 28 d, and if its
bioactivity could be retained over these time frames. In this paper, we investigated the ability of
combinations of various excipients on anti-NogoA encapsulation, release, and bioactivity in the
context of our in vitro system.
To improve the bioactivity of encapsulated anti-NogoA, excipients were co-encapsulated within
PLGA nanoparticles and the bioactive fraction of the anti-NogoA released over 28 d in vitro
was measured by ELISA. When trehalose was used as an excipient, anti-NogoA bioactivity was
significantly improved for up to one week of release, without dramatically affecting the rate at
which bioactivity was lost. These results confirm the well-documented ability of trehalose to
protect proteins during lyophilization [30] by providing a hydrogen bonding partner for the
proteins as water is sublimed [69]. Further, the ability of co-encapsulated trehalose to maintain
all of the bioactivity of encapsulated anti-NogoA suggests that anti-NogoA is only sensitive to
lyophilization during processing, and is unaffected by sonication and exposure to organic
solvents, which is promising since these processes are generally known to cause damage to the
tertiary structure of proteins [34].
We then wanted to know if the co-encapsulated trehalose would alter the release kinetics of anti-
NogoA, which is a potentially undesirable consequence of encapsulating water soluble
additives. Interestingly, co-encapsulated trehalose did not change the release kinetics of anti-
NogoA, which is likely due to the small size of trehalose (Mw = 0.38 kDa) compared to anti-
NogoA (Mw = 150 kDa). The pores created by the dissolution of trehalose are likely too small or
poorly interconnected to create a pathway for anti-NogoA diffusion and release. In an effort to
achieve more sustained release, hyaluronan was co-encapsulated with trehalose and anti-NogoA
and a higher burst release and more sustained long-term delivery was observed. The larger
hyaluronan (Mw = 2600 kDa) likely creates a large pore network upon dissolution that allows
for faster and more complete diffusive anti-NogoA release. This view is further supported by the
model fit parameters, as discussed below. These results suggest that additives or complementary
drugs below a threshold size may be co-encapsulated with larger molecules without affecting
the release kinetics of the larger drugs. Also, co-encapsulated hyaluronan leads to a more
sustained and complete anti-NogoA release.
50
In light of the beneficial effect of hyaluronan on release kinetics, we investigated the impact of
this co-encapsulant on anti-NogoA bioactivity. When hyaluronan was co-encapsulated with
trehalose and anti-NogoA, There was no difference in bioactivity during early time points, but at
14, 21, and 28 d, there was a significant improvement in bioactivity. Lee et al. previously
suggested that the ability of co-encapsulated hyaluronan to improve protein stability was caused
by the formation of a viscous aqueous phase that inhibits the interaction between the protein and
the organic phase during double emulsion synthesis [41]. Our results suggest that this viscous
aqueous phase also protects the protein during drug release, as it may inhibit adsorption of
protein onto the surface of the polymer, a known problem that affects protein stability in PLGA
particles [84]. These results demonstrate that trehalose and hyaluronan can stabilize anti-
NogoA at early and late release times, respectively.
In an effort to further improve bioactivity at late time points, the effect of co-encapsulated
MgCO3 and CaCO3 on anti-NogoA release and bioactivity was assessed. Co-encapsulated bases
are known to neutralize the acidic environment inside PLGA particles [66] and this strategy has
been used to stabilize bovine serum albumin (BSA) [78], basic fibroblast growth factor [78],
bone morphogenetic protein-2, [78], and tissue plasminogen activator [85]. MgCO3 was chosen
based on results that demonstrate its ability to neutralize the acidic environment inside PLGA
particles [66] and stabilize tetanus toxoid in PLGA particles [86]. CaCO3 was chosen because it
is similar to MgCO3 in terms of size and basicity, yet dissimilar in terms of solubility (ksp CaCO3 =
3.36 x 10-9
, ksp MgCO3 = 6.82 x 10-6
). In contrast to work done by Zhu et al. [78], we observed
attenuated protein release kinetics with co-encapsulated bases. This discrepancy is likely a result
of the reported aggregation of their model protein (BSA). It is likely that the co-encapsulated
bases slowed the rate of acid-induced PLGA hydrolysis, which reduced the rate of pore network
formation, resulting in attenuated long-term protein release. Ara et al. previously reported this
reduction in PLGA degradation as a result of co-encapsulated basic additives [81]. At early time
points, there was no appreciable difference in anti-NogoA bioactivity, which was expected since
the acidic environment inside PLGA particles takes several days to form [66]. The lack of
observable benefit to the long-term bioactivity of anti-NogoA is likely due to the insignificant
total protein release amounts between 7 d and 28 d, rather than poor bioactivity.
The mathematical model applied in this study allows for the quantitative comparison of the
release profiles of the PLGA nanoparticles presented in this paper. As summarized in Table 4,
51
the model was able to fit all release profiles with R2 values ranging from 0.95 to 0.99 over 54 d
of release. The initial diffusivity (D0) terms ranged from 2 x 10-17
to 2 x 10
-16 cm
2/s, in close
agreement with published diffusivity values from 200 nm poly(lactic acid) particles [64, 87, 88],
which are a comparable size to our approximately 300 nm particles. The formulation containing
trehalose and hyaluronan had the highest Fburst, D0, and k values, which suggests that this
molecule is acting as a porogen, forming a large pore network after dissolution [72]. This
increased porosity then increases the amount of anti-NogoA that is available for burst release
and increases the rate at which it is released because of the additional pathways for escape.
Interestingly, the high k parameter suggests that pore network formation due to polymer
degradation is highest for this formulation, which is expected because increased porosity allows
for more water uptake, which increases the rate of PLGA hydrolysis. One drawback of this
formulation is the lower encapsulation efficiency of 43%, when compared to the other
formulations that encapsulated between 80 to 97%. It is possible that the anionic hyaluronan is
interfering with the ionic attraction between anti-NogoA and the free carboxylic end groups on
the PLGA [47], which results in lower encapsulation. This interference is prevented when the
acidic hyaluronan is neutralized by the co-encapsulated bases.
The rate of anti-NogoA release from our system compares favorably with similar preclinical
systems that resulted in functional recovery in vivo. The formulation containing trehalose and
hyaluronan is suitable for application to spinal cord injury treatment because it has sustained
bioactivity up to 28 d and desirable release kinetics over this duration. While the bioactivity of
released anti-NogoA has been significantly improved, the 0.1% bioactive fraction after 28 d of
release may not be sufficient for some applications. In those cases, the presented formulation
should be further modified with protein stability strategies that have been published previously
[34]. Composite HAMC with co-encapsulated trehalose and hyaluronan can deliver up to 60 μg
of anti-NogoA per rat over 28 d (or 0.09 μg /hr), assuming 10 μL composite volume, a particle
loading of 200 mg/mL, and an anti-NogoA loading of 3 wt%. Anti-NogoA has only been
delivered by minipump systems in vivo, so it is important to normalize with respect to volume
the amount delivered in those studies to allow for direct comparison with our system. The
intrathecal space in a rat is approximately 200 μL and approximately 30 mL in a macaque
monkey [89], while the composite delivery volume is 10 μL. Consequently, to achieve the same
anti-NogoA concentration at the site of injury using a minipump system requires 20 fold more
52
protein in a rat and 3000 fold more protein in a monkey compared to our localized system for
this theoretical comparison. Adjusted to the 10 μL intrathecal volume surrounding the injury,
Liebscher et al. delivered anti-NogoA at a rate of 0.75 μg/h over 14 d in a rat model and
observed enhanced regeneration of corticospinal tract axons and improved motor recovery [23].
Similarly adjusted, Wu et al. delivered anti-NogoA at a rate of 0.004 μg/h over 28 d in a rat
model and observed significant functional recovery [24]. Adjusted, Freund et al. delivered anti-
NogoA at 0.014 ug/h over 14 d in a primate model and observed improved recovery of manual
dexterity and sprouting of corticospinal axons [90]. We are encouraged by the
pharmacologically relevant dose, sustained release, bioactivity retention, and biocompatibility
[26] of this system, which encourages future studies to investigate its efficacy in vivo.
53
4 Discussion
4.1 Achieving Sustained and Bioactive NT-3 and anti-NogoA
Release
In this work, the 28 day delivery of bioactive NT-3 and anti-NogoA from the composite
hydrogel drug delivery system was demonstrated. It is important to know whether the amount of
deliverable drug from this system would elicit functional recovery. To do this, a comparison was
conducted of drug delivery rate and duration between the proposed DDS and similar systems
from the literature.
The NT-3 payload of this DDS compares favorably with alternative strategies when differences
in distribution volume are accounted for. NT-3 delivered via minipump is distributed throughout
the intrathecal space, approximately 200 μL in the rat; whereas 10 μL of composite HAMC is
delivered to the site of injury. To achieve equivalent NT-3 concentrations adjacent to the injury
therefore requires 20 fold more protein delivered by minipump than from a localized strategy.
Composite HAMC with no excipients can deliver up to 100 μg of NT-3 per rat over 28 days (or
0.16 μg/hr), assuming 10 μL composite volume and a particle loading of 200 mg/mL and an
NT-3 loading of 5 wt%. These conditions can be achieved by replacing the co-encapsulated
BSA used in the current work with NT-3. Adjusted to the 10 μL intrathecal volume surrounding
the injury, minipump based delivery of 0.025 μg/hr has been reported, which resulted in
functional recovery in a rat model [20]. Among localized strategies, a fibrin gel was used to
deliver 200 ng NT-3 over 9 days (0.93 ng/hr) [79] and microtubes embedded in an agarose gel
were used to delivered 100 ng NT-3 per rat over 14 days (0.3 ng/hr) [14].
As outlined in Section 3.4, composite HAMC with co-encapsulated trehalose and hyaluronan
can deliver up to 60 μg of anti-NogoA per rat over 28 d (or 0.09 μg /hr), assuming 10 μL
composite volume, a particle loading of 200 mg/mL, and an anti-NogoA loading of 3 wt%.
Adjusted to the 10 μL intrathecal volume surrounding the injury, Liebscher et al. delivered anti-
NogoA at a rate of 0.75 μg/h over 14 d [23], Wu et al. delivered anti-NogoA at a rate of 0.004
μg/h over 28 d [24], and Freund et al. delivered anti-NogoA at 0.014 ug/h over 14 d [90]. These
finding are summarized in Table 5.
54
Table 5 – A comparison of the proposed drug delivery system to in vivo NT-3 and anti-NogoA
drug release studies
Authors Rate of
Delivery
(ng/hr)
Duration of
Delivery
(days)
Method of
Delivery
Animal
Model
Result
NT-3
Stanwick et al. 160 28 Composite
hydrogel
N/A N/A
Coumans et al. 25* 14 Minipump Rat axon regeneration,
functional recovery
Taylor et al. 0.93 9 Fibrin gel Rat axon sprouting,
Lee et al. 0.3 14 Microtubes in
agarose
Rat axon sprouting,
functional recovery
Tuszynski et al. Unknown 28 Cellular Rat axon growth,
functional recovery
Anti-NogoA
Stanwick et al. 90 28 Composite
hydrogel
N/A N/A
Liebscher et al. 750* 14 Minipump Rat axon regeneration,
motor recovery
Wu et al. 4* 28 Minipump Rat functional recovery
Freund et al. 14* 14 Minipump Primate axon sprouting,
manual dexterity
*Adjusted to a 10 μL intrathecal volume surrounding the injury
We are encouraged by the capacity of composite HAMC to deliver a therapeutically relevant
payload of NT-3 and anti-NogoA, which will likely lead to functional recovery that is at least as
efficacious as seen by other groups. In fact, the simultaneous delivery of these agents is
expected to elicit more axonal regeneration compared to increasing the dosage of either
individual therapeutic agent. This expectation is based on a ligand-receptor analysis of the
system. First, it was assumed that NogoA and NT-3 act on neuronal receptors non-competitively
(Figure 15a), which is reasonable given that they act on two different receptors (Nogo receptor
and TrkC, respectively). In this model, neurite outgrowth is a function of the rate of formation
of the TrkC-NT-3 binding, while NogoA prevents outgrowth when bound (Figure 15b).
55
a)
b)
These equilibrium equations can be rearranged to solve for neurite outgrowth rate using the
following equation:
Where Vm is the maximum neurite growth rate, [I] is the concentration of NogoA, [E] is the
concentration of NT-3, and Ki and Km are equilibrium constants.
NogoA
NT-3 Neuron TrkC
Nogo receptor
-
+ Neurite outgrowth
Neuron + NT-3 Neuron-NT-3 Complex Neurite Outgrowth
+ +
NogoA NogoA
Neuron-NogoA + NT-3 Neuron-NogoA-NT-3
KI
KM
K2
KI
KM
Figure 15 - The effect of NT-3 and NogoA on neurite outgrowth viewed in a non-
competitive inhibition model. a) A diagram illustrating the interaction between the
inhibitory protein NogoA with the nogo receptor and NT-3 with the TrkC receptor on
neuronal cells. The former inhibits neurite outgrowth and the latter improves neurite
outgrowth. b) The equilibrium equations describing the receptor/ligand interactions.
56
Substituting arbitrary values of Vm = 1, Ki = 1, and Km = 1 into this equation reveals that the
rate of neurite outgrowth reaches a plateau with increasing NT-3 concentration, which can only
be further increased by reducing the concentration of NogoA (Figure 16), namely, by
administering anti-NogoA to the system. This analysis suggests that delivery of both NT-3 and
anti-NogoA could provide more robust neurite regeneration than simply increasing the dosage
of NT-3.
Figure 16 – Rate of neurite outgrowth as a function of NT-3 concentration for two values
of NogoA concentration, as simulated by non-competitive ligand-receptor kinetics. This
model suggests that delivery of anti-NogoA in combination with NT-3 would provide faster
neurite regeneration compared to simply increasing the dosage of NT-3.
0
0.2
0.4
0.6
0.8
1
1.2
0 5 10 15 20 25
Ra
te o
f N
eu
rite
Ou
tgro
wth
NT-3 Concentration
Nogo-A = 0
Nogo-A = 1
57
4.2 Why do NT-3 and anti-NogoA behave differently?
Comparing the results presented in Chapters 2 and 3, one interesting observation is that NT-3
and anti-NogoA behave very differently. Specifically, these drugs display disparate
encapsulation, release, and bioactivity when trehalose + hyaluronan are co-encapsulated within
the PLGA particles. If these molecules acted similarly, they could reasonably be combined into
one chapter dealing with issues relating to protein stability. Instead a two chapter treatment was
necessary.
In Chapter 2, it was shown that the co-encapsulation of trehalose and hyaluronan had mild
effects on encapsulated NT-3. It did not dramatically alter the release kinetics of NT-3, when
compared to an excipient-free control (Figure 8a). This additive had no effect on NT-3
detection by ELISA after nanoparticle processing nor did it affect NT-3 bioactivity by DRG
bioassay up to 28 d. This co-encapsulant did, however, reduce the total protein encapsulation of
NT-3 and BSA from 97% to 30%.
In Chapter 3, on the other hand, it was shown that the co-encapsulation of trehalose +
hyaluronan had a drastic effect on anti-NogoA. This co-encapsulant significantly changed the
release profile of anti-NogoA, allowing for more sustained release (Figure 13). Further, this
additive improved bioactivity of anti-NogoA up to 28 d compared to an excipient-free control,
but reduced the total protein encapsulation of anti-NogoA and IgG from 97% to 43%.
So why does the co-encapsulation of trehalose + hyaluronan increase the release kinetics of the
larger anti-NogoA (Mw = 150 kDa), but not affect the release of NT-3 (Mw = 25 kDa). This
discrepancy is likely caused by the difference in PLGA MW used to produce these two
nanoparticle formulations. The NT-3 formulation used 0.67 dl/g (or approximately 30 kDa)
PLGA, while the anti-NogoA formulation used 0.2 dl/g (approximately 5 kDa) PLGA. The
choice to use PLGA of two different molecular weights was based on optimized release profiles
using model drugs (IgG and α–chymotrypsin). The longer PLGA chains in the NT-3
formulation are less soluble in the organic solvent used during synthesis, which results in faster
nanoparticle solidification upon solvent extraction. This faster solidification is associated with
smaller average pore size. This reduced pore size likely interferes with the ability of trehalose
58
and hyaluronan to increase pore interconnectivity. Consequently, these co-encapsulants have
no measurable effect on NT-3 release, while they are able to increase anti-NogoA release.
Another interesting phenomenon was that trehalose + hyaluronan was able to stabilize anti-
NogoA up to 28 d, but did not have any impact on NT-3 detectability or bioactivity. The ability
of this additive to stabilize anti-NogoA was anticipated because trehalose is a known
lyoprotectant [34] and hyaluronan has been shown to stabilize proteins from the
organic/aqueous interface during processing by increasing the viscosity in the inner aqueous
phase [41]. Consequently, the inability of this co-encapsulant to improve NT-3 stability was
initially surprising. This incongruity is explained again by the different molecular weights of the
PLGA used in the NT-3 formulations compared to the anti-NogoA formulation. As discussed in
Chapter 2, the effect of co-encapsulated trehalose + hyaluronan on NT-3 stability was likely
masked by the protective effect of PLGA, which can stabilize the protein through protein-
polymer hydrogen bond formation during lyophilization (similar to the mechanism of trehalose)
and through increasing the viscosity of the emulsion (similar to the mechanism of hyaluronan).
The protective effect of PLGA did not mask the co-encapsulant-mediated improvement to
bioactivity in the case of anti-NogoA because a lower PLGA MW and concentration was used
(50 mg/mL, 0.2 dl/g vs. 132 mg/mL, 0.67 dl/g). The lower PLGA concentration reduced the
amount of hydrogen bonding partners available to the protein during lyophilization, while the
lower molecular weight did not increase the viscosity of the emulsion substantially during
processing. These phenomena reduce the efficacy of PLGA as a stabilizer, which allows for the
observation of the effect of trehalose and hyaluronan in the anti-NogoA formulations.
In both NT-3 and anti-NogoA formulations, co-encapsulated trehalose + hyaluronan reduced
encapsulation efficiency dramatically. One possible explanation is that the increased viscosity of
the particles in emulsion slowed down particle solidification, allowing time for additional
entrapped drug to escape.
59
5 Conclusions
The primary contributions of this thesis have been:
1. A likely mechanism by which HAMC is able to linearize the release of drugs from
PLGA particles has been proposed:
Methylcellulose binds to the surface of the particles over the first few hours
through hydrophobic interactions, which creates a barrier to diffusion
This barrier reduces burst release, but does not influence particle degradation or
mass loss
2. Sustained and bioactive release of NT-3 has been achieved
Delivery over 28 d is achievable with no additives
Co-encapsulated PEG 400 improves stability during processing
Co-encapsulated MgCO3 substantially improves bioactivity over 28 d
3. Sustained and bioactive release of anti-NogoA has been achieved
Co-encapsulated trehalose improves anti-NogoA stability during processing
Further co-encapsulation of hyaluronan increases the release rate substantially
and increases the bioactivity of anti-NogoA up to 28 d
60
6 Recommendations for Future Work
In this thesis, 28 d bioactive delivery of NT-3 and anti-NogoA was achieved from PLGA
nanoparticle embedded in HAMC. Future work motivated by this thesis is classified as either
further in vitro optimization or evaluating in vivo efficacy.
6.1 In Vitro Optimization
Final in vitro optimization is recommended to prepare this drug delivery system for in vivo
studies. The recommended experiments were not conducted in conjunction with the in vitro
studies described in this thesis because they are beyond the scope of this Master‟s project.
The PLGA nanoparticle formulations recommended in Chapters 2 and 3 should be reduced in
size to allow for sterile filtration. PLGA and/or encapsulated proteins are known to degrade
during alternative sterilization techniques, including steam sterilization [91] and gamma
irradiation [92]. Consequently, filtration is the preferred method of sterilization for PLGA
particles; however, when these nanoparticles were sterile filtered, 90-95% of the particles were
lost. This loss is attributed to the large diameter and size distribution of the particles (200-300
nm) relative to the pore size of the filter (220 nm). It has been shown that particles closer to 150
nm in diameter have been successfully sterile filtered with 90% retention of particle mass [93].
This reduction in particle size can be accomplished by increasing the surfactant concentration in
the outer aqueous phase, increasing sonication duration, or increasing sonication intensity. We
recommend that these parameters be screened to reduce the average particle diameter to below
150 nm to allow for sterile filtration.
It is also recommended that the release kinetics of the NT-3 + MgCO3 formulation be further
improved. While this preparation yielded the best bioactivity as measured by DRG neurite
outgrowth, the release was not as complete as the control formulation. There are a number of
published strategies for achieving more complete release from PLGA particles, including:
reducing PLGA molecular weight, reducing PLGA concentration, or incorporating pore-forming
excipients. These factors should be screened in 28 d release studies to determine if more
complete 28 d release is achievable without sacrificing encapsulation efficiency or bioactivity.
61
Conversely, anti-NogoA exhibited desirable release kinetics over 28 d, but less than 1% of the
released drug was bioactive after 14 d. While the work presented in Chapter 3 represents a
significant improvement to anti-NogoA bioactivity, further work to increase activity is
recommended. To accomplish this task, we recommend a thorough investigation of the life
cycle of the system to determine the cause of anti-NogoA damage, similar to how NT-3 stability
was treated in Chapter 2. Once the primary sources of damage during processing, drug release,
and storage have been identified, appropriate published strategies [34] should be screened to
evaluate their efficacy in the stabilization of anti-NogoA.
6.2 In Vivo Efficacy
Once the nanoparticles have been reduced in size for facile sterilization and the release kinetics
and bioactivity have been further fine-tuned, the DDS will be prepared for in vivo evaluation.
Since this system has already demonstrated biocompatibility in the intrathecal space over 28 d in
a rat model [26], the next step should be evaluating the ability of this DDS to regenerate
damaged neurites and enhance functional recovery. We recommend measuring neuronal
sprouting, glial scar formation, and locomotor function in vivo. A compression injury model is
recommended because we have previously used this model for this system [26]. Treatment
groups should include: (a) Composite HAMC; (b) NT-3 in composite HAMC; (c) anti-NogoA
in composite HAMC; (d) NT-3 and anti-NogoA in composite HAMC. Neuronal sprouting could
be followed by identifying the lesion borders with a GFAP stain and then using the neural fiber
stain, Tuj1, to quantify the „% area covered by neural fibers‟. These measurements could be
divided into sections caudal, in between, and rostral to the injury site. To assess glial scar
formation, GFAP staining should be performed to quantify astrocyte density. To measure the
impact of the treatment groups on functional recovery, BBB open field motor testing should be
conducted weekly on the animals. These experiments will test the hypotheses that the proposed
drug delivery system will enhance neuronal sprouting, reduce the glial scar, and improve
functional recovery. If successful, this proposed study could provide the motivation to
incorporate a more complex „cocktail‟ of therapeutics into the DDS, such as cAMP,
chondroitinase ABC, and BDNF.
62
7 References
[1] S. Thuret, Moon, L.D.F., and Gage, F.H., Therapeutic interventions after spinal cord injury,
Nature, 7 (2006) 628-643.
[2] J. Wosnick, Baumann, M.D., Shoichet, M.S., Tissue therapy in the central nervous system,
Elsevier, New York, 2007.
[3] The University of Alabama National Spinal Cord Injury Statistical Centre, (2002).
[4] Rick Hansen Institute SCI Facts, (2011).
[5] A. Amar, Levy, M.L., Pathogenesis and pharmacological strategies for mitigating secondary
damage in acute spinal cord injury, Neurosurgery, 44 (1999) 1027-1039.
[6] O.N. Hausmann, Post-traumatic inflammation following spinal cord injury, Spinal Cord, 41
(2003) 369-378.
[7] G.W.J. Hawryluk, Rowland, J., Kwon, B.K., Fehlings, M.G., Protection and repair of the
injured spinal cord: a review of completed, ongoing, and planned clinical trials for acute spinal
cord injury, Neurosurg Focus, 25 (2008) E14.
[8] H. Barbeau, Rossignol, S., Enhancement of locomotor recovery following spinal cord injury,
Current Opinion in Neurology, 7 (1994) 517-524.
[9] D.C. Baptiste, Fehlings, M.G., Pharmacological approaches to repair the injured spinal cord,
J. Neurotrauma, 23 (2006) 318-334.
[10] R. Pal, Venkataramana, N.K., Bansal, A., et al., utologous bone marrow transplantation in
patients with subacute and chronic spinal cord injury, Cell Transplantation, 15 (2006) 675-687.
[11] R.C. Robinson, Radziejewski, C., Spragon, G., Greenwald, J., Kostura, M.R., Burtnick,
L.D., Stuart, D.I., Choe, S., and Jones, E.Y., The structures of the neurotrophin 4 homodimer
and the brain-derived neurotrophic factor/neurotrophin 4 heterodimer reveal a common Trk-
binding site, Protein Sci., 8 (1999) 2589-2597.
[12] M.J. Butte, et al., Crystal structure of Neurotrophin-3 homodimer shows distinct regions are
used to bind its receptors, Biochemistry, 37 (1998) 16846-16852.
[13] S.J. Taylor, and Sakiyama-Elbert, S.E., Effect of controlled delivery of neurotrophin-3
from fibrin on spinal cord injury in a long term model, J. Controlled Release, 116 (2006) 204-
210.
[14] H. Lee, McKeon, R.J., and Bellamkonda, R.V., Sustained delivery of thermostabilized
chABC enhances axonal sprouting and functional recovery after spinal cord injury, PNAS, 107
(2010) 3340-3345.
63
[15] M. Yu-Hai, Zhang, Y., Cao, L., Su, J.-C., Wang, Z.-W., Xu, A.-B., Zhang, S.-C., Effect of
neurotrophin-3 genetically modified olfactory ensheathing cells transplantation on spinal cord
injury, Cell Transplantation, 19 (2010) 167-177.
[16] N. Comolli, Neuhuber, B., Fischer, I., Lowman, A., In vitro analysis of PNIPAAm-PEG, a
novel, injectable scaffold for spinal cord repair, Acta Biomaterialia, 5 (2009) 1046-1055.
[17] J.A. Burdick, Ward, M., Liang, E., Young, M.J., Langer, R., Stimulation of neurite
outgrowth by neurotrophins delivered from degradable hydrogels, Biomaterials, 27 (2006) 452-
459.
[18] J.H. Brock, Rosenzweig, E.S., Moseanko, R., Havton, L.A., Edgerton, V.R., and
Tuszynski, M.H., Local and remote growth factor effects after primate spinal cord injury, J.
Neurosci., 30 (2010) 9728-9737.
[19] P. Lu, Yang, H., Jones, L.L., Filbin, M.T., Tuszynski, M.H., Combinatorial therapy with
neurotrophins and cAMP promotes axonal regeneration beyong sites of spinal cord injury, J.
Neurosci., 24 (2004) 6402-6409.
[20] J.V. Coumans, Lin, T.T.S., Dai, H.N., MacArthur, L., McAtee, M., Nash, C., and Bregman,
B.S., Axonal regeneration and functional recovery after complete spinal cord transection in rats
by delayed treatment with transplants and neurotrophins, J. Neurosci., 21 (2001) 9334-9344.
[21] M.H. Tuszynski, Cellular delivery of neurotrophin-3 promotes corticospinal axonal growth
and partial functional recovery after spinal cord injury, J. Neurosci., 17 (1997) 5560-5572.
[22] D.H.S. Lee, Strittmatter, S.M., Sah, D.W.Y., Targeting the nogo receptor to treat central
nervous system injuries, Nature Reviews Drug Discovery, 2 (2003) 872-878.
[23] T. Liebscher, Schnell, L., Schnell, D., Scholl, J., Schneider, R., Gullo, M., Fouad, K., Mir,
A., Rausch, M., Kindler, D., Hamers, F.P.T., Schwab, M.E., Nogo-A antibody improves
regeneration and locomotion of spinal cord-injured rats, Ann. Neurol., 58 (2005) 706-719.
[24] J. Wu, Yang, H., Qiu, Z., Zhang, Q., Ding, T., Geng, D., Effect of combined treatment with
methylprednisolone and nogo-A monoclonal antibody after rat spinal cord injury, J. Int. Med.
Res., 38 (2010) 570-582.
[25] D. Aprili, Bandschapp, O., Rochlitz, C., Urwyler, A., Ruppen, W., Serious complications
associated with external intrathecal catheters used in cancer pain patients: A systematic review
and meta-analysis, Anesthesiology, 111 (2009) 1346-1355.
[26] M.D. Baumann, Kang, C.E., Tator, C.H., and Shoichet, M.S., Intrathecal delivery of a
polymeric nanocomposite hydrogel after spinal cord injury, Biomaterials, 31 (2010) 7631-7639.
[27] M.D. Baumann, Kang, C.E., Stanwick, J.C., Wang, Y., Kim, H., Lapitsky, Y., Shoichet,
M.S., An injectable drug delivery platform for sustained combination therapy, J. Controlled
Release, 138 (2009) 205-213.
64
[28] C.E. Kang, Clarkson, R., Tator, C.H., Yeung, I.W.T., Shoichet, M.S., Spinal cord blood
flow and blood vessel permeability measured by dynamic computed tomography imaging in rats
after localized delivery of fibroblast growth factor, J. Neurotrauma, 27 (2010) 2041-2053.
[29] T. Peternel, and Komel, R., Isolation of biologically active nanomaterial (inclusion bodies)
from bacterial cells, Microbial Cell Factories, 9 (2010).
[30] W. Wang, Lyophilization and development of solid protein pharmaceuticals, Int. J. Pharm.,
203 (2000) 1-60.
[31] A. Hawe, Kasper, J.C., Friess, W., and Jiskoot, W., Structural properties of monoclonal
antibody aggregates induced by freeze-thawing and thermal stress, Eur. J. Pharm. Sci., 38
(2009) 79-87.
[32] R.S. Raghuvanshi, et al., Stabilization of dichloromethane-induced protein denaturation
during microencapsulation, Pharm. Dev. Technol., 3 (1998) 269-276.
[33] M.A. Ibrahim, Ismail, A., Fetouh, M.I., Gopferich, A., Stability of insulin during the
erosion of poly(lactic acid) and poly(lactic-co-glycoilic acid) microspheres., J. Controlled
Release, 106 (2005) 241-252.
[34] U. Bilati, Allemann, E., Doelker, E., Strategic approaches for overcoming peptide and
protein instability within biodegradable nano- and microparticles, Eur. J. Pharm. Sci., 59 (2005)
375-388.
[35] W. Jiang, Schwendeman, S.P., Stabilization and controlled release of bovine serum
albumin encapsulated in poly(D,L-lactide) and poly(ethylene glycol) microsphere blends,
Pharm. Res., 18 (2001) 878-885.
[36] R. Ghaderi, Carlfors, J., Biological activity of lysozyme after entrapment in poly(D,L-
lactide-co-glycolide)-microspheres, Pharm. Res., 14 (1997) 1556-1562.
[37] M. Diwan, Park, T.G., Pegylation enhances protein stability during encapsulation in PLGA
microspheres, J. Controlled Release, 73 (2001) 233-244.
[38] M. Diwan, Park, T.G., Stabilization of recombinant interferon-a by pegylation for
encapsulation in PLGA microspheres, Int. J. Pharm., 252 (2003) 111-122.
[39] X.M. Lam, et al., Encapsulation and stabilization of nerve growth factor into poly(lactic-co-
glycolic) acid microspheres, J. Pharm. Sci., 90 (2001) 13-24.
[40] D. Wong, Parasrampuria, J., Pharmaceutical excipients for the stabilization of proteins,
Pharm. Technol., 21 (1997) 34-50.
[41] E.S. Lee, Kwon, M.J., Lee, H., Kim, J.J., Stabilization of protein encapsulated in
poly(lactide-co-glycolide) microspheres by novel viscous S/W/O/W method, Int. J. Pharm., 331
(2007) 27-37.
[42] M.L. Houchin, and Topp, E.M., Chemical degradation of peptides and proteins in PLGA: a
review of reactions and mechanisms, J. Pharm. Sci., 97 (2008) 2395--2404.
65
[43] A. Giteau, Venier-Julienne, M.C., Aubert-Pouessel, A., Benoit, J.P., How to achieve
sustained and complete protein release from PLGA-based microparticles, Int. J. Pharm., 350
(2008) 14-26.
[44] S. Cohen, Yoshioka, T., Lucarelli, M., Hwang, L.H., Langer, R., Controlled delivery
systems for proteins based on poly(lactic/glycolic acid) microspheres, Pharmaceutical Research,
8 (1991) 713-720.
[45] J. Siepmann, Siepmann, F., Florence, A.T., Local controlled drug delivery to the brain:
Mathematical modeling of the underlying mass transport mechanisms, Int. J. Pharm., 314 (2006)
101-119.
[46] S. Freiberg, Zhu, X., Polymer microspheres for controlled drug release, Int. J. Pharm., 282
(2004) 1-18.
[47] Y. Yeo, and Park, K., Control of encapsulation efficiency and initial burst in polymeric
microparticle systems, Arch. Pharmacal Res., 27 (2004) 1-12.
[48] A. Budhian, Siegel, S.J., Winey, K.I., Haloperidol-loaded PLGA nanoparticles: Systematic
study of particle size and drug content, Int. J. Pharm., 336 (2007) 367-375.
[49] A. Lemprecht, et al., Influences of process parameters on nanoparticle preparation
performed by double emulsion pressure homogenization technique, Int. J. Pharm., 196 (2000)
177-182.
[50] C. Astete, Kumar, C., Sabliov, C., Size control of poly(D,L-lactide-co-glycolide) and
poly(D,L-lactide-co-glycolide)-magnetite nanoparticles synthesized by emulsion evaporation
technique, Colloids and Surfaces A: Physicochem. Eng. Aspects, 299 (2007) 209-216.
[51] A. Lamprecht, et al., Influences of process parameters on nanoparticle preparation
performed by double emulsion pressure homogenization technique, Int. J. Pharm., 196 (2000)
177-182.
[52] Y.y. Hsu, et al., Comparison of process parameters for microencapsulation of plasmid DNA
in poly(D,L-lactic-co-glycolic) acid microspheres, Journal of Drug Targeting, 7 (1999) 313-323.
[53] X. Wen, and Tresco, P.A., Fabrication and characterization of permeable degradable
poly(DL-lactide-co-glycolide) (PLGA) hollow fiber phase inversion membranes for use as nerve
tract guidance channels, Biomaterials, 27 (2006) 3800-3809.
[54] A. Charlier, et al., Release of mifepristone from biodegradable matrices: experimental and
theoretical evaluations, Int. J. Pharm., 200 (2000) 115-120.
[55] T. Higuchi, Rate of release of medicaments from ointment bases containing drugs in
suspension, Journal of phamaceutical sciences, 50 (1961) 874-875.
[56] C. Raman, Berkland, C., Kim, K., Pack, D.W., Modeling small-molecule release from PLG
microspheres: effects of polymer degradation and nonuniform drug distribution, J. Controlled
Release, 103 (2005) 149-158.
66
[57] N. Faisant, Siepmann, J., Benoit, J.P., PLGA-based microparticles: elucidation of
mechanisms and a new, simple mathematical model quantifying drug release, Eur. J. Pharm.
Sci., 15 (2002) 355-366.
[58] J. Siepmann, Faisant, N., Akiki, J., Richard, J., Benoit, J.P., Effect of the size of
biodegradable microparticles on drug release: experiment and theory, J. Controlled Release, 96
(2004) 123-134.
[59] L.F. Reichardt, Neurotrophin-regulated signalling pathways, Philosophical Transactions of
the Royal Society, 361 (2006) 1545-1564.
[60] D. Gupta, Tator, C.H., Shoichet, M.S., Fast-gelling injectable blend of hyaluronan and
methylcellulose for intrathecal, localized delivery to the injured spinal cord, Biomaterials, 27
(2006) 2370-2379.
[61] A. Hurtado, Moon, L.D.F., Maquet, V., Blits, B., Jerome, R., Oudega, M., Poly(D,L-lactic
acid) macroporous guidance scaffolds seeded with Schwann cells genetically modified to
secrete a bi-functional neurotrophin implanted in the completely transected adult rat thoracic
spinal cord, Biomaterials, 27 (2006) 430-442.
[62] B. Blits, Dijkhuizen, P.A., Boer, G.J., and Verhaagen, J., Intercostal nerve implants
transduced with an adenoviral vector encoding neurotrophin-3 promote regrowth of injured rat
corticospinal tract fibers and improve hindlimb function, Experimental Neurology, 164 (2000)
25-37.
[63] C.P. Riley, Cope, T.C., and Buck, C.R., CNS neurotrophins are biologically active and
expressed by multiple cell types, J. Mol. Histol., 35 (2004) 771-783.
[64] A. Budhian, Siegel, S.J., Winey, K.I., Controlling the in vitro release profiles for a system
of haloperidol-loaded PLGA nanoparticles, Int. J. Pharm., 346 (2008) 151-159.
[65] L.K. Chiu, Chiu, W.J., Cheng, Y.-L., Effects of polymer degradation on drug release - a
mechanistic study of morphology and transport properties in 50:50 poly(dl-lactide-co-
glycolide), Int. J. Pharm., 126 (1995) 169-178.
[66] L. Li, Schwendeman, S.P., Mapping neutral microclimate pH in PLGA microspheres, J.
Controlled Release, 101 (2005) 163-173.
[67] J.-M. Pean, Boury, F., Venier-Julienne, M.-C., Menei, P., Proust, J.-E., Benoit, J.-P., Why
does PEG 400 co-encapsulation improve NGF stability and release from PLGA biodegradable
microspheres? Pharm. Res., 16 (1999) 1294-1299.
[68] L.B. Feril Jr., and Kondo, T., Major factors involved in the inhibition of ultrasound-induced
free radical production and cell killing by pre-sonication incubation or by high cell density,
Ultrason. Sonochem., 12 (2005) 353-357.
[69] J.F. Carpenter, Crowe, J.H., Arakawa, T., Comparison of solute-induced protein
stabilization in aqueous solution and in frozen and dried state, J. Dairy Sci., 73 (1990) 3627-
3636.
67
[70] K. Ako, Nicolai, T., and Durand, D., Salt-induced gelation of globular protein aggregates:
structure and kinetics, Biomacromolecules, 11 (2010) 864-871.
[71] S.T. Kelly, Zydney, A.L., Effects of intermolecular thiol-disulfide interchange reactions on
BSA fouling during microfiltration, Biotechnol. Bioeng., 44 (1994) 972-982.
[72] R. Dorati, Colonna, C., Genta, I., Modena, T., Conti, B., Effect of porogen on the physico-
chemical properties and degradation performance of PLGA scaffolds, Polym. Degrad. Stabil.,
95 (2010) 694-701.
[73] S.J. Taylor, McDonald, J.W., and Sakiyama-Elbert, S.E., Controlled release of
neurotrophin-3 from fibrin gels for spinal cord injury, J. Controlled Release, 98 (2004) 281-294.
[74] N. Schupper, Rabin, Y., Rosenbluh, M., Multiple stages in the aging of a physical polymer
gel, Macromolecules, 41 (2008) 3983-3994.
[75] P. Arora, and Tandon, P.N., Mathematical models for drug delivery to chronic patients,
Applied Mathematical Modelling, 33 (2009) 692-705.
[76] B.-H. Chen, and Lee, D.J., Slow release of drug through deformed coating film: Effects of
morphology and drug diffusivity in the coating film, J. Pharm. Sci., 90 (2001) 1478-1496.
[77] K. Fu, Pack, D.W., Klibanov, A.M., Langer, R., Visual evidence of acidic environment
within degrading poly(lactic-co-glycolic acid) (PLGA) microspheres, Pharm. Res., 17 (2000)
100-107.
[78] G. Zhu, Mallery, S.R., and Schwendeman, S.P., Stabilization of proteins encapsulated in
injectable poly(lactide-co-glycolide), Nat. Biotechnol., 18 (2000) 52-57.
[79] S.J. Taylor, Rosenzweig, E.S., McDonald III, J.W., Sakiyama-Elbert, S.E., Delivery of
neurotrophin-3 from fibrin enhances neuronal fiber sprouting after spinal cord injury, J.
Controlled Release, 113 (2006) 226-235.
[80] Y. Han, Jin, B.-Y., Lee, S.-B., Sohn, Y., Joung, J.-W., Lee, J.-H., Effects of sugar additives
on protein stability of recombinant human serum albumin during lyophilization and storage,
Arch. Pharmacal Res., 30 (2007) 1124-1131.
[81] M. Ara, Watanabe, M., Imai, Y., Effect of blending calcium compounds on hydrolytic
degradation of poly(DL-lactic acid-co-glycolic acid), Biomaterials, 23 (2002) 2479-2483.
[82] M.D. Baumann, Austin, J.W., Fehlings, M.G., Shoichet, M.S., A quantitative ELISA for
bioactive anti-Nogo-A, a promising regenerative molecule for spinal cord injury repair,
Methods, 44 (2009) 104-108.
[83] S. David, Aguayo, A.J., Axonal elongation into peripheral nervous system "bridges" after
central nervous system injury in adult rats, Science, 214 (1981) 931-933.
[84] A. Giteau, Venier-Julienne, M.C., Aubert-Pouessel, A., Benoit, J.P., How to achieve
sustained and complete protein release from PLGA-based microparticles? Int. J. Pharm., 350
(2008) 14-26.
68
[85] J. Kang, Schwendeman, S.P., Comparison of the effects of Mg(OH)2 and sucrose on the
stability of bovine serum albumin encapsulated in injectable poly(D,L-lactide-co-glycolide)
implants, biomaterials, 23 (2002) 239-245.
[86] W. Jiang, Schwendeman, S.P., Stabilization of tetanus toxoid encapsulated in PLGA
microspheres, Mol. Pharm., 5 (2008) 808-817.
[87] M. Chorny, Fishbein, I., Danenberg, H.D., Golomb, G., Study of drug release mechanism
from tyrophostin AG-1295-loaded nanospheres by in situ and external sink methods, J.
Controlled Release, 83 (2002) 401-414.
[88] C. Rouzes, Leonard, M., Durand, A., Dellacherie, E., Influence of polymeric surfactants on
the properties of drug-loaded PLA nanospheres, Colloid Surf. B: Biointerfaces, 32 (2003) 125-
135.
[89] P. Matin, Goodwin, D.A., Cerebrospinal fluid scanning with 111In, Journal of Nuclear
Medicine, 12 (1971) 668-672.
[90] P. Freund, Schmidlin, E., Wannier, T., Bloch, J., Mir, A., Schwab, M.E., Rouiller, E.M.,
Nogo-A-specific antibody treatment enhances sprouting and functional recovery after cervical
lesion in adult primates, Nat. Med., 12 (2006) 790-792.
[91] A. Yaman, Alternative methods of terminal sterilization for biologically active
macromolecules, Current Opinion in Drug Discovery and Development, 4 (2001) 760-763.
[92] N. Faisant, Siepmann, J., Richard, J., Benoit, J.P., Mathematical modeling of drug release
from bioerodible microparticles: Effect of gamma-irradiation, European Journal of
Pharmaceutics and Biopharmaceutics, 56 (2003) 271-279.
[93] J. Wendorf, et al., A practical approach to the use of nanoparticles for vaccine delivery, J.
Pharm. Sci., 95 (2006) 2738-2750.
69
8 Appendix A
The governing equations and boundary conditions for release from PLGA nanoparticles are
summarized below:
Governing equation:
Boundary Conditions:
0
,
where c is concentration, r is radial position inside the nanoparticles, D(Mw) is
diffusivity, f(r) is the initial NT-3 distribution inside the nanoparticles, which was taken
as uniform.
The governing equations and boundary conditions for release from PLGA nanoparticles
embedded in HAMC are summarized below:
Governing equation:
Boundary Conditions:
c(t)
70
,
where x is position inside the HAMC hydrogel, f(x) is the initial NT-3 distribution
within the HAMC hydrogel, which was taken as 0, and c(t) is the predicted NT-3
concentration as a result of NT-3 released from nanoparticles alone.
In both cases, the diffusivity term is treated similarly,
When fraction released < fraction available for burst release (Fburst),
When fraction released > fraction available for burst release (Fburst),
,
Where Do is the initial diffusivity of NT-3 through the nanoparticles, k is a fit parameter
representing the degree to which degradation influences diffusivity, and kdeg is the first-
order degradation rate constant of PLGA.
R2 values were calculated based on the following formulae:
,
where yi is the experimental value, is the mean of all experimental values, and fi is the
value produced by the model.