Human Serum Albumin (HSA) Nanoparticles Reproducibility Of

Available online at www.sciencedirect.com

International Journal of Pharmaceutics 347 (2008) 109117

Pharmaceutical Nanotechno

Human serum albumin (HSA) nanopartf enser aV.

rsel, Jt amue-StJune007

Abstract

Nanoparticles prepared from human serum albumin (HSA) are versatile carrier systems for drug delivery and can be prepared by an establisheddesolvationIn the presendimers and hby size excluparticle prepAt pH 8.0 anparticles. Becomparablewere observ

For nanopbesides thewell as rHSAdegradation.of particle st 2007 Else

Keywords: N

1. Introdu

Nanopathe specificand Vauhtias syntheticle preparaalbumin (Htitude of stuSteinhause

CorresponE-mail ad

0378-5173/$doi:10.1016/jprocess. A reproducible process with a low batch-to-batch variability is required for transfer from the lab to an industrial production.t study the batch-to-batch variability of the starting material HSA on the preparation of nanoparticles was investigated. HSA can buildigher aggregates because of a free thiol group present in the molecule. Therefore, the quality of different HSA batches was analysedsion chromatography (SEC) and analytical ultracentrifugation (AUC). The amount of dimerised HSA detected by SEC did not affectaration. Higher aggregates of the protein detected in two batches by AUC disturbed nanoparticle formation at pH values below 8.0.d above monodisperse particles between 200 and 300 nm could be prepared with all batches, with higher pH values leading to smallersides human derived albumin a particle preparation was also feasible based on recombinant human serum albumin (rHSA). Underpreparation conditions monodisperse nanoparticles could be achieved and the same effects of protein aggregates on particle formationed.articulate drug delivery systems the enzymatic degradation is a crucial parameter for the release of an embedded drug. For this reason,

particle preparation process, particle degradation in the presence of different enzymes was studied. Under acidic conditions HSA asnanoparticles could be digested by pepsin and cathepsin B. At neutral pH trypsin, proteinase K, and protease were suitable for particle

It could be shown that the kinetics of particle degradation was dependent on the degree of particle stabilisation. Therefore, the degreeabilisation will influence drug release after cellular accumulation of HSA nanoparticles.vier B.V. All rights reserved.

anoparticles; Human serum albumin (HSA); Recombinant human serum albumin (rHSA); Particle size; Enzymatic degradation

ction

rticles have emerged as versatile carrier systems fordelivery of drugs to organs and tissues (Couvreur

er, 2006). Various macromolecular substances suchc and natural polymers can be used for nanoparti-tion (Kreuter, 2004). Among these, human serumSA) is a promising material and was used in a mul-dies for particle preparation (Michaelis et al., 2006;

r et al., 2006). Human serum albumin (molecular

ding author. Tel.: +49 69 798 29692; fax: +49 69 798 29694.dress: [email protected] (K. Langer).

weight of 65 kDa) belongs to a multigene family of proteins(He and Carter, 1992) and is the major soluble protein of the cir-culating system with a blood concentration of about 50 mg/ml.Human serum albumin consists of 585 amino acids containing 35cysteine residues which build 17 disulfide bridges. One free thiolgroup, namely Cys34, remains unbound. In circulating plasma30% of this free sulfhydryl Cys34 is oxidised by cysteine andglutathione (Carter and Ho, 1994). Oxidation can also occur bydimerisation during the isolation process. Some products con-tain up to 20% of dimerised albumin. This percentage increaseswith the age of the protein unless Cys34 has been blocked withcysteine or gluthathione. Since albumin dimers or higher aggre-gates may influence the preparation process of nanoparticles,the quality of the starting materials is of major importance.

see front matter 2007 Elsevier B.V. All rights reserved..ijpharm.2007.06.028preparation process and kinetics oK. Langer a,, M.G. Anhorn a, I. Steinhau

N. Schrickel a, S. Faust a,a Institut fur Pharmazeutische Technologie, Biozentrum Niederu

Max-von-Laue-Strae 9, D-60438 Frankfurb Institut fur Biophysik, Johann Wolfgang Goethe-Universitat, Max-von-La

Received 12 February 2007; received in revised form 18Available online 23 June 2logy

icles: Reproducibility ofzymatic degradation

, S. Dreis a, D. Celebi a,Vogel bohann Wolfgang Goethe-Universitat,Main, Germanyrae 1, D-60438 Frankfurt am Main, Germany2007; accepted 19 June 2007

110 K. Langer et al. / International Journal of Pharmaceutics 347 (2008) 109117

Besides oxidation leading to dimers, the human origin ofHSA is another drawback of this material. Consequently, thereis the potentis, CJD) anto these pro(rHSA). Sinhost organi2006). Actfor rHSA pBy this orgscale and wture. Neverhuman der

Despitenanoparticlby the FDAcytostatic din water, thpolyethylatThe drug inimprove drto the standet al., 2006

Althougof studies thfocus of mof particleassessmentthe presentwere prepaal., 2003).of albuminHSA werematerialsof the resuhand the biusing a varwere simulof HSA as

2. Materia

2.1. Reage

HSA (fr95% SDS8% solutioFour diffe035K7566,batches ofused for pof the nano(batch 010K067H1090)from Sigmachieved frother reage

many). All chemicals were of analytical grade and used asreceived.

ompatog

mo

s waPLCn anBioas m

as ca

rds.trati

jecteby Ulculatog

sam

ere

alytiere

atety exultraances c(f the

repa

A naious

ssolve sod 8.h a 0ermarmeof th

teg pud nan ofglut

nkint thaino g. Thsusp

resu

ntialpelleas petial risk of pathogen contamination (e.g. HIV, hepati-d the possibility of variability in quality. An answerblems could be seen in recombinant produced HSAce HSA is a non-glycosylated protein a wide range of

sms is appropriate for rHSA production (Kobayashi,ually, the most frequently used expression systemroduction is Pichia pastoris (Chuang et al., 2002).anism, it is possible to produce proteins in a largeith identical primary, secondary and tertiary struc-

theless, rHSA is still a very expensive alternative toived albumin.the problematic human origin, the first HSA-basede formulation, ABI 007 (Abraxane) was approvedin 2005. The 130 nm large nanoparticles contain the

rug paclitaxel. Due to the bad solubility of paclitaxele conventional drug preparation (Taxol) containsed castor oil (Cremophor EL) and ethanol as vehicles.corporation in nanoparticles follows a new concept toug solubility, with a variety of advantages conferredard paclitaxel therapy (Gradishar et al., 2005; Desai).h HSA nanoparticles were described in a multitudee variability of the preparation process was not in the

ost of these investigations. For a scaling-up processpreparation and for industrial particle preparation anof batch-to-batch variation is of major importance. Instudy nanoparticles based on human serum albuminred by a well-defined desolvation process (Langer etDifferent HSA batches containing different amountsdimers and higher aggregates as well as recombinantused for this process. The influence of the starting

quality on particle size, size distribution, and yieldlting nanoparticles was investigated. On the otherodegradability of the nanoparticles was assessed. Byiety of different enzymes, physiological conditionsated in order to evaluate the enzymatic degradationwell as rHSA nanoparticles.

ls and methods

nts and chemicals

action V, purity 9699%), recombinant HSA (rHSA;-PAGE, expressed in P. pastoris), and glutaraldehyde

n were obtained from Sigma (Steinheim, Germany).rent batches of human derived HSA (015K7535,

045K7535, 111K7614) as well as two differentrecombinant HSA (095K1258, 116K1451) were

article preparation. For the enzymatic degradationparticles proteinase K (batch 022K8620), protease7670), pepsin (batch 087H0163), pancreatin (batch

, and cathepsin B (batch 033K7685) were obtaineda (Steinheim, Germany). Trypsin (batch 9019) wasom BDH Laboratory supplies (Poole, England). Allnts were purchased from Merck (Darmstadt, Ger-

2.2. Cchrom(AUC)

Thebatcheon a Hcolum(TosohpH 7.0tem wstandaconcen

was initoredwas ca

chromThe

ples wthe anteins wphosphvelociXL-Aabsorbelled aused o

2.3. P

HSas prevwas dichlorid8.0, anthrougsel, Gtransfo4.0 mlat rooma tubinenableadditioof 8%crossliamoun

60 ammatrixof therature.

Thediffereof thestep wosition of HSA batches by size exclusionraphy (SEC) and analytical ultracentrifugation

lecular weight distribution of the four different HSAs analysed by size exclusion chromatography (SEC)system equipped with TSKgel G3000SWXL guard-

d a TSKgel G3000SWXL 7.8 mm 30 cm columnscience, Stuttgart, Germany) using phosphate bufferobile phase at a flow rate of 0.8 ml/min. The SEC sys-librated for molecular weight with globular protein

Aqueous solutions of the proteins were prepared at aon of 1 mg/ml and an aliquot (20.0l) of each sampled into the SEC system. The eluent fraction was mon-V detection (280 nm). The amount of dimeric HSA

ated relative to the total peak area in the respectiverams.e four different HSA batches and both rHSA sam-

investigated by sedimentation velocity analysis incal ultracentrifuge. Aqueous solutions of the pro-prepared at a concentration of 1 mg/ml in 20 mM

buffer containing 100 mM NaCl. The sedimentationperiments were performed using a Beckman Optimacentrifuge at rotor speeds of 35,000 rpm at 20 C. Theversus radius data (collected at 280 nm) were mod-

s) and c(M) distribution of non-interactive species,sedt program by Schuck et al. (2002).

ration of HSA nanoparticles

noparticles were prepared by a desolvation techniquely described (Langer et al., 2003). In principle, HSAed at a concentration of 100 mg/ml in 10 mM sodiumlution and the pH of the solution was titrated to 7.5,5, respectively. The resulting solutions were filtered.22m filtration unit (Schleicher und Schull, Das-ny). Aliquots (1.0 ml) of the HSA solutions wered into nanoparticles by the continuous addition ofe desolvating agent ethanol under stirring (550 rpm)

mperature. The ethanol addition was performed bymp (Ismatec IPN, Glattbrugg, Switzerland) which

noparticle preparation at a defined rate of ethanol1.0 ml/min. After the desolvation process 117.6l

araldehyde in water were added to induce particleg. This volume corresponds to 200% of the theoretict is necessary for the quantitative crosslinking of theroups present in the HSA molecules of the particle

e crosslinking process was performed under stirringension over a time period of 24 h at room tempe-

lting nanoparticles were purified by three cycles ofcentrifugation (16,100 g, 8 min) and redispersiont to the original volume in water. Each redispersionrformed in an ultrasonication bath (Elma Transsonic

K. Langer et al. / International Journal of Pharmaceutics 347 (2008) 109117 111

Digital T790/H) over 5 min. The resulting amount of HSAnanoparticles was determined gravimetrically.

For the epared onlyprepared asisation. Gluof a 8% gcrosslinkin

For eveformed in thgiven as m

2.4. Prepa

Recombas solutionride, sodiuN-acetyltryvation procthe rHSA ssettes (MWpurified wathe suppliewithout addcess a labscfreeze-drye

The resuticles bypreparationmatic degrarHSA batch

2.5. Deter

After pasity, and zspectroscopzetasizer 3The samplsured at a90.

As an asome of thanalytical2003). In pby addition20 mM sodsucrose, atand 0.7 at 4ultracentrifman Optimdescribed e(turbidity)elled as a don the resuused the lsRossmanith

Table 1Enzymes and conditions used for the degradation of HSA nanoparticles

se Kachium

fromeas, ty

tin froeas

in B

nzym

theenzy4, ancase

hlorthech woparg/mns w

andthealiq

es wparttedGerame ietriopa

ults

objet thealbuat th

noparticles.he first part of the study the main focus was on repro-lity of the particle formation under the aspect of differents of the starting material HSA. This part was based on ourwork (Langer et al., 2003) describing the optimisation ofparation process by the use of a pump-controlled systemdesolvation of HSA solutions. In the earlier study the pHf the HSA solution prior to the desolvation procedure was

to be the major parameter to control the resulting particlewas shown, that the preparation method applying a pump-lled system in combination with a defined pH adjustmentpresence of sodium chloride leads to well-defined meane sizes as well as to narrow particle size distributions. Itnzymatic degradation study nanoparticles were pre-from HSA batch 111K7614. Nanoparticles were

outlined above except for the extent of particle stabil-taraldehyde amounts of 23.5, 35.3, 47.0, and 58.8l

lutaraldehyde solution were employed resulting ing degrees of 40, 60, 80, and 100%, respectively.ry batch of HSA the particle preparation was per-ree independent samples. The analytical results were

ean value and standard deviation of these samples.

ration of rHSA nanoparticles

inant human serum albumin (rHSA) was obtainedcontaining the following excipients: sodium chlo-m phosphate buffer pH 7.4, sodium caprylate,ptophan. Since additives can interfere with the desol-ess of albumin these substances were removed fromolution by dialysis using Slide-A-Lyzer dialysis cas-CO 3500, Pierce, Rockford, USA). Dialysis againstter was performed according to the instruction of

r. After dialysis the rHSA solution was freeze-driedition of further excipients. For the freeze-drying pro-ale Lyovac GT2 (Leybold Heraeus, Hurth, Germany)r was used.lting lyophilised rHSA was used to prepare nanopar-the desolvation technique described above. Theprocess was performed at a pH of 8.5. For the enzy-dation study nanoparticles were prepared only from116K1451.

mination of particle size and size distribution

rticle purification average particle size, polydisper-etapotential were measured by photon correlationy (PCS) and microelectrophoresis using a Malvern

000HSA (Malvern Instruments Ltd., Malvern, UK).es were diluted 1:400 with purified water and mea-

temperature of 25 C and a scattering angle of

lternative method, size distribution was studied ine samples by sedimentation velocity analysis in theultracentrifuge (Vogel et al., 2002; Langer et al.,rinciple, the nanoparticle stock solution was brought,of appropriately concentrated solutions and water, toium phosphate (pH 7.0), 100 mM NaCl, 23.5% (w/v)a solute concentration giving turbidity between 0.620 nm in a cuvette with a 1 cm optical pathlength. Theugation experiments were performed using a Beck-a XL-A ultracentrifuge at rotor speeds of 3000 rpm asarlier (Vogel et al., 2002). The apparent absorbanceversus radius data (collected at 420 nm) were mod-istribution of non-diffusing spherical particles, basedlts described (Vogel et al., 2002). The calculations-g*(s)-variant of the sedt program by Schuck and

(2000).

Enzyme

ProteinaTritir

Proteasepancr

TrypsinPancrea

pancrPepsin

Catheps

2.6. E

Forferent5.4, 6.In thehydroc

Forcle batto nan1000pensio565 nm

Fortion anparticla nano

incuba5436,ious tiphotoming nan

3. Res

Thelook aserum

well asing na

In tducibibatcheearlierthe prefor thevalue ofoundsize. Itcontroin theparticlConcentration(g/ml)

Enzymatic activity(units/ml)

pH

fromalbum

2 0.060 7.5

bovinepe I

10 0.073 7.5

50 Unknown 7.5m porcine 50 1.25 7.5

5000 300 2.0

10 0.056 5.410 0.056 6.4

atic degradation of HSA nanoparticles

enzymatic degradation of the nanoparticles the dif-mes were used as listed in Table 1. The pH values ofd 7.5 were adjusted with phosphate buffer systems.of pepsin the pH value was adjusted with 0.01N

ic acid.calibration of the photometric assay each nanoparti-as diluted with the respective buffers of the enzymesticle concentrations of 0, 50, 100, 250, 500, andl, respectively. The turbidity of the nanoparticle sus-as determined photometrically at a wavelength ofwas used for the calculation of calibration curves.

determination of the kinetics of enzymatic degrada-uot of the nanoparticle suspension containing 2.0 mgere diluted with the respective enzyme solution toicle concentration of 1000g/ml. The mixture wasat 37 C under shaking (Eppendorf Thermomixertebau Eppendorf, Engelsdorf, Germany). After var-ntervals the turbidity of the samples was measuredcally at 565 nm and the concentration of the remain-rticles was calculated relative to the calibration curve.

and discussion

ctive of the present study was to take a more detailedreproducibility of the preparation process of humanmin (HSA) nanoparticles by protein desolvation ase kinetics of the enzymatic degradation of the result-


was a certain drawback of the earlier study that all of the exper-iments were performed with only one given batch of HSA and,therefore, tUnder theestablishmimportancematerials oto be faced

A furthefor humanmaterials.approved nrisk of patadditionallderived HSration was

In the sedegradationon in the pon some o

trypsin andof HSA nastudy a cleand enzymamounts ofto a reducelier study wquantitativquent quanthe enzymathe aspect othis reasonenzymes cowere focus

3.1. Purity

For thecommerciaused. Thesupplier wiis well knoits dimericis due to ature of the pmolecules (nanoparticltribution ofparticle prea double dhigh molecet al., 2005amount ofmay as weration procstudy we hthe differen

Table 2Amount of monomeric and dimeric protein in different HSA batches deter-mined by peak area in size exclusion chromatography (SEC) and qualitative by

al ultracentrifugation (AUC)tch Amount HSA by SEC (%) Higher molecular

aggregates by AUCMonomeric form Dimeric form

35 94.0 6.0 +66 93.2 6.8 +++35 94.9 5.1 +14 90.4 9.6

raphy (SEC) and by analytical ultracentrifugation (AUC).C method revealed that the different HSA batches con-

variable amounts of dimeric HSA (Table 2). By SEC ant between 5 and 10% dimeric HSA was detected for alls. The lowest content of 5.1% dimeric HSA was foundtch 045K7535, but AUC revealed a certain shift of theo lower molecular weights (Fig. 1). The highest amounttected within batch 111K7614. The results were in goodance with the data of AUC showing amounts between11% for the respective batches. Therefore, the chro-

raphic results were supported by the data of the AUCd (Fggre

ed wse ca. Ofor

ular wharacns o045Ked. Tmolorde

olecular weight distributions c(M) of HSA protein from different HSAThe protein was analysed at a concentration of 1 mg/ml in 20 mM

te buffer pH 7.0 containing 100 mM NaCl. Upper right: enlargementolecular weight distribution for detection of higher HSA aggregates.

s: () batch 035K7566; () batch 015K7535; () batch 111K7614; ()5K7535.he influence of different HSA batches was ignored.future aspect of industrial particle preparation theent of a reproducible preparation method is of major

and the influence of different batches of startingn the resulting physico-chemical characteristics has.r requirement for a pharmaceutical product intended

use is the application of safe and well defined startingAlthough human derived HSA is used for the FDAanoparticle preparation Abraxane, the remaining

hogen contamination is often discussed. Therefore,y in the first part of the study the exchange of humanA for recombinant HSA (rHSA) for particle prepa-evaluated.cond part of the study the kinetics of the enzymaticof HSA as well as rHSA nanoparticles was focused

resence of different enzymes. This part was basedf our preliminary results showing that the enzymes

proteinase K were well suited for the degradationnoparticles (Wartlick et al., 2004). In this earlier

ar dependency between glutaraldehyde crosslinkingatic particle degradation was observed with highercrosslinking during the preparation process leading

d ability for enzymatic particle degradation. This ear-as done in order to evaluate suitable conditions for

e particle degradation under the aspect of a subse-tification of matrix-bound drugs. In the present studytic degradation was evaluated in more detail underf intracellular degradation after particle uptake. Forbesides trypsin and proteinase K a range of furthervering intracellular enzymes of the cathepsin family

sed on.

of different HSA batches

preparation of HSA nanoparticles by desolvationlly available HSA isolated from human plasma waspurity of the starting material was declared by theth values between 96 and 99%. On the other hand itwn, that a certain amount of the protein is present inor higher aggregated form. The dimeric form of HSAfree sulfhydryl group (Cys34) in the primary struc-rotein which enables a covalent dimerisation of HSACarter and Ho, 1994). In our earlier work with gelatines we have found that a broad molecular weight dis-gelatin is responsible for problems with reproducibleparation. The problem was solved by establishing

esolvation method which enables the separation ofular gelatin fractions (Coester et al., 2000; Balthasar). Therefore, in the case of HSA nanoparticles thedimeric or higher molecular weight HSA impuritiesll have influence on the reproducibility of the prepa-ess. For this reason in the first part of the presentave analysed the molecular weight distributions oft commercial HSA batches by size exclusion chro-

analytic

HSA ba

015K75035K75045K75111K76

matogThe SEtainedamoun

batchefor bapeak twas deaccord8 andmatogmethoHSA anounc

for the360 kDpeaksmolecwas c

no sigbatchobservhigherbatch

Fig. 1. Mbatches.phosphaof the mSymbolbatch 04ig. 1). Moreover, the AUC method revealed highergates in the different batches. This was most pro-ithin batch 035K7566 showing two distinct peaksompounds at molecular weights of about 240 andn the other hand HSA batch 111K7614 only showedthe monomeric and dimeric HSA at the expected

eights of about 65 and 130 kDa. Although this batchterised by the highest amount of dimeric proteinf higher aggregates were observed. In the case of7535 a pronounced tailing of the dimer peak wasaking all these observations together the amount ofecular HSA aggregates decreased in the followingr: 035K7566 > 045K7535 = 015K7535 > 111K7614


Fig. 2. Molec() and afterat a concentr100 mM NaCfor detection

(Table 2).dimeric HSwas observ

Additiodrying ondeterminedwas providferent stabfreeze-dryinanoparticlples consistraces of habout 300 kafter dialysening of ththe hydratimolecules.lar weightrHSA, occminor tracever, rHSAweight dist

3.2. Repro

The diffnanoparticlpreparationtion was prof the soluThe additioa pump-coal., 2003).nanoparticlple were dpurity data

Influerepare

S.D.n: 10

zetndene nan).

previn pricon

atchighertheere

614theof

ter bndered c

VA,s 035ular weight distributions c(M) of recombinant HSA (rHSA) priorpurification and freeze-drying (). The protein was analysed

ation of 1 mg/ml in 20 mM phosphate buffer pH 7.0 containingl. Upper right: enlargement of the molecular weight distributionof higher HSA aggregates.

Therefore, no correlation between the amount ofA and amount of higher molecular HSA aggregatesed within the batches.nally, in this part of the study the influence of freeze-the molecular weight distribution of rHSA wasby analytical ultracentrifugation (AUC). As rHSA

ed in form of an aqueous solution containing dif-ilising agents, protein purification by dialysis andng of the purified protein was performed prior toe preparation. Before freeze-drying the rHSA sam-ted of a high amount of monomeric rHSA with onlyigher aggregates in the molecular weight range ofDa (Fig. 2). The situation was completely differentis and freeze-drying of rHSA. A pronounced broad-e monomer peak was observed, indicating change inon state of the protein or aggregation of the rHSAFurthermore, an additional small peak at a molecu-of about 190 kDa, representing the trimeric form of

Fig. 3.ticles p(mean centratio

ters theIndepetion th(n = 14

Assolutioeter toHSA bwith hUnderticles w111K7led toinsteadthe lateven u

increas(ANObatcheurred. In comparison to HSA of human origin, onlyes of these higher aggregates were detected. How-purification was crucial with regard to the molecularribution of the protein.

ducibility of HSA nanoparticle preparation

erent HSA batches were used for the preparation ofes by a well established desolvation method. Thewas done under defined conditions. The HSA solu-epared in the presence of 10 mM NaCl and the pHtions was adjusted to 7.5, 8.0, and 8.5, respectively.n of the desolvating agent ethanol was performed byntrolled system at a speed of 1.0 ml/min (Langer etAfter stabilisation and purification of the resultinges the particle size and polydispersity of each sam-etermined and the results were compared with theof the starting material. Besides these size parame-

of 278.1 On the othbatch 111Ka diameter

The samdispersity035K7566dispersity iwith a polyperse sizethe nanopabatch (Fig.the range bcorrelationmethod ofa parameteParticle yiein a respectinfluencednce of the HSA batch on particle diameter of HSA nanopar-d at different pH values in 10 mM sodium chloride solution; n = 3). Rate of ethanol addition: 1.0 ml/min; initial HSA con-0 mg/ml. Particle size measurement after particle purification.

apotential of the HSA nanoparticles was determined.t of the pH and HSA batch used for protein desolva-oparticles showed a zetapotential of 43.2 3.0 mV

ously described by our group the pH value of the HSAor to the desolvation procedure was the major param-trol the resulting particle size (Fig. 3). Within everyused the particle diameter was significantly reducedpH values of the HSA solution (ANOVA, p < 0.01).

chosen conditions and a pH value of 7.5 nanopar-only formed with the HSA batches 015K7535 andwhereas the other batches 035K7566 and 045K7535formation of aggregates in the micrometer scalenanoparticles. At higher pH values of 8.0 and 8.5atches led to the formation of nanoparticles butthese conditions their diameters were significantlyompared to the preparations of batch 111K6147

p < 0.01). For example at a pH value of 8.0 with theK7566 and 045K7535 nanoparticles with diameters

8.5 and 286.6 10.2 nm were achieved, respectively.er hand under the same preparation conditions HSA

7614 led to significantly smaller nanoparticles withof 198.1 2.3 nm.e is true for the batch dependency of the poly-

index (Fig. 4). At a pH value of 7.5 both batchesand 045K7535 led to aggregates with a high poly-ndex whereas with the other batches nanoparticlesdispersity index below 0.1, indicative for a monodis-distribution, were achieved. In contrast the yield ofrticles seems to be independent of the respective HSA5). Although slight variations of the particle yield inetween 18.1 and 23.1 mg/ml were observed, no clearto pH value or batch was observed. But due to the

analysis it has to be outlined that particle yield is notr as sensitive as particle diameter or polydispersity.ld is mainly controlled by the solubility of the proteinive solvent at a respective pH value and will hardly beby the formation of protein aggregates. Additionally,


Fig. 4. Influence of the HSA batch on polydispersity of HSA nanoparti-cles prepared at different pH values in 10 mM sodium chloride solution(mean S.D.; n = 3). Rate of ethanol addition: 1.0 ml/min; initial HSA con-centration: 100 mg/ml. Polydispersity measurement after particle purification.

due to the theory of PCS used for particle size analysis, alreadyminor traces of larger particles will lead to a deviation of thecorrelation function and, therefore, to a significant increase ofthe polydis

Based omolecularmation of ncomparedpolydispersobvious ththe resultinhighest amticle prepa045K7535and high poditions. Onhigher aggsize and po035K7566

Fig. 5. Influecles prepared(mean S.D.centration: 10

led to increased particle sizes at all pH values under evalua-tion. Furthermore at pH 7.5 both batches resulted in significantlyincreased polydispersity indices (ANOVA, p < 0.01). A certainexception can be observed for HSA batch 015K7535. Althoughthis batch is characterised by higher aggregates in a propor-tion comparable to batch 045K7535, nanoparticles instead ofmicroparticles were achieved at pH 7.5. Hence, other param-eters such as low molecular weight impurities (i.e. salts) ofHSA batches additionally may take influence on the formationof nanoparticles during the desolvation process. However, theseparticles prepared from batch 015K7535 were larger and showedan increased polydispersity compared to batch 111K6147.

Therefore, it has to be concluded that the purity of a HSAbatch is a fundamental precondition for the reproducible prepa-ration of His reasonabtein aggregpreparation

3.3. Prepa

sam

as albinand HSin q

al fodesclyticning

ular sationt pH

stabie parined

mma

eparwere

an thpersity index.n the idea that the amount of dimeric or higherweight HSA impurities takes influence on the for-anoparticles during the desolvation process we havethe amount of these impurities with the size andity data of the resulting nanoparticles. It becomes

at the amount of dimeric HSA has no influence ong nanoparticles. The HSA batch 111K7614 with theount of dimeric protein led to a reproducible par-ration even at a pH value of 7.5 whereas the batchwith the lowest amount of dimer led to aggregateslydispersity indices under the same preparation con-the other hand a correlation between the amount of

regates detected by AUC and the resulting particlelydispersity seems to exist. For example the batchesand 045K7535 with high amounts of these impurities

Thecles wrecom

deriveabilitymateri

Asby anabroademolecpreparticles aAftercles thdetermare su

cles prrHSAsize thnce of the HSA batch on particle yield of HSA nanoparti-at different pH values in 10 mM sodium chloride solution

; n = 3). Rate of ethanol addition: 1.0 ml/min; initial HSA con-0 mg/ml. Yield determination after particle purification.

human derparticle siztrast, the rHsize of 29probably b

Table 3Nanoparticlesresults of phybatches

Parameter

Particle diamePolydispersityZetapotentialParticle yieldSA nanoparticles by our desolvation procedure and itle to verify the HSA quality under the aspect of pro-ation and high molecular impurities prior to particle.

ration of rHSA nanoparticles

e desolvation process described for HSA nanoparti-so used for the preparation of nanoparticles based ont human serum albumin (rHSA). Since the human

A bears the risk of pathogen contamination and vari-uality, rHSA can be taken as an alternative startingr the preparation of drug carrier systems.ribed above the characterisation of the purified rHSAal ultracentrifugation (AUC) showed a significantof the monomer peak and some impurities in the

ize range of the trimeric protein (Fig. 2). For particlethe lyophilised product was transferred to nanopar-8.5 as previously outlined for HSA of human origin.

lisation and purification of the resulting nanoparti-ticle size, polydispersity, zetapotential, and yield was. The results of three independent particle samplesrised in Table 3. In comparison to HSA nanoparti-ed under the same conditions, the particles made ofsomewhat larger and showed a higher variability ine HSA particles. Nanoparticles prepared of the four

ived HSA batches of human origin were received ates between 179.9 7.2 and 248.2 14.5 nm. In con-

SA nanoparticles showed a significantly increased6.8 27.0 nm (ANOVA, p < 0.01). This effect cane attributed to aggregation during protein purification

prepared based on recombinant human serum albumin (rHSA):sico-chemical characterisation of three independent nanoparticle

Preparation Mean S.D.#1 #2 #3

ter (nm) 273.2 326.3 290.9 296.8 27.00.029 0.025 0.054 0.036 0.016

(mV) 52.8 47.7 47.6 49.4 3.0(mg/ml) 21.9 19.6 18.8 20.1 1.6


and freeze-drying detected by AUC. The parameters polydisper-sity and particle yield were comparable to HSA nanoparticles.The zetapoindicated s(Muller et aticles underinvestigatiorial againstorder to fin

3.4. Enzym

Besidessuitabilityof major imfavoured tothe particleprerequisituptake. Thics of the dnanoparticl

The conof trypsinfor a rapidbe usefulloading aftenzymaticthe degreeditions HSwere quantof nanopar9% of thea crosslink36.1% withof 60% weparable kindependencpresence oshown). Increatin wasonly 30.6 athe case ofWithin 24 h(100% glucontrast, inparticle deging degreetrypsin thistively highchosen pepof a matrixgastric enz1.5 and pHrequired.

The resumatic degrAs could b

Enzymatic degradation of HSA nanoparticles in the presence ofg/ml trypsin and (B) 5 mg/ml pepsin. Glutaraldehyde concentrations40 and 100% were used for particle crosslinking (mean S.D.; n = 3).

tion resulted in comparable degradation kinetics of theparticles (Fig. 7A and B). Therefore, the exchange ofgainst rHSA will be feasible for nanoparticle preparationt changing the degradation behaviour of the drug carrier.ides gastric and intestinal enzymes the intracellulare cathepsin B was focussed on within the present study.es of the cathepsin family are mainly localised in the

mes of cells. These organelles are part of the endo-ompartment and are involved in the final breakdown oflised material as well as the storage of indigestible mate-ukherjee et al., 1997). The pH of the lysosomes variesn 5.0 and 5.5. Nanoparticles are internalised into cellsh endocytosis and, therefore, eventually are delivered tomes. In order to enable an intracellular release of a drugped within the nanoparticle matrix, the particles have toraded by lysosomal enzymes such as cathepsin. For thegation of the biodegradability, cathepsin B was used atlues of 6.4 and 5.4. The higher pH value only led to aible degradation of HSA nanoparticles (Fig. 8A). Afterere was almost no digestion of the HSA nanoparticles,ndent of the glutaraldehyde amount used for particleation. In contrast, at pH 5.4 the degradation was morenced (Fig. 8B) because of the higher activity of cathep-tential of the rHSA nanoparticles of 49.4 3.0 mVufficient storage stability of the colloidal systemsl., 2000). In principle, preparation of rHSA nanopar-the chosen condition was feasible. Nevertheless, ourns showed that the exchange of one starting mate-a modified one requires systematic investigations ind optimal conditions for particle preparation.

atic degradation of HSA nanoparticles

the reproducible preparation of nanoparticles theof these systems for intracellular drug delivery isportance. As incorporative drug loading is mostlytake advantage of the high drug carrier capacity of

s, a controlled enzymatic degradation is a decisivee for intracellular drug release after cellular particleerefore, in the second part of the study the kinet-egradation of HSA nanoparticles as well as rHSAes in the presence of different enzymes was analysed.ditions for the enzymatic degradation in the presenceand pepsin were chosen in order to find conditionsdegradation of the HSA nanoparticles. This could

for a fast determination of the incorporative druger particle preparation. In the case of trypsin thedegradation of the particles was mainly influenced byof particle crosslinking (Fig. 6A). Under these con-A nanoparticles with a crosslinking degree of 40%itatively degraded within 90 min whereas in the caseticles crosslinked with 100% glutaraldehyde onlyparticles were degraded within 24 h. Particles withing degree of 80% showed a degraded fraction ofin 24 h and nanoparticles with a crosslinking degreere completely degraded within this time span. Com-etics of nanoparticle degradation with a significant

e on the crosslinking degree were observed in thef the enzymes proteinase K and protease (data notcontrast particle degradation in the presence of pan-not so pronounced. After an incubation period of 3 h

nd 11.2% of the HSA nanoparticles were degraded in40 and 100% crosslinked nanoparticles, respectively.

between 81.6% (40% glutaraldehyde) and 52.8%taraldehyde) of the nanoparticles were degraded. In

the presence of pepsin within 30 min a completeradation was observed independent of the crosslink-of the particle matrix (Fig. 6B). In comparison torapid degradation was mainly due to the compara-

enzymatic activity used for the assay. Therefore, thesin conditions are convenient for the determination-bound drug within a short time frame. But as theyme pepsin shows the highest activity between pH

3.0 drug stability under these acidic conditions is

lts of HSA nanoparticles were supported by the enzy-adation of recombinant HSA (rHSA) nanoparticles.e expected all of the enzymatic conditions under

Fig. 6.(A) 50between

evaluarHSAHSA awithousystem

BesenzymEnzymlysosocytic cinternarial (Mbetweethrouglysosoentrapbe deginvestipH vaneglig24 h thindepestabilispronou


Fig. 7. Enzymthe presencehyde concent(mean S.D.

sin at this pwith a crosincreased tglutaraldehcle degradaacidic pH cafter cellul

Takingto be conctease, and pand rHSAsis of drugnanoparticuthese enzycrosslinkinto longer timin the pressin B a quawas not pohighly stabenabled pafor cathepsmainly affeatic degradation of recombinant HSA (rHSA) nanoparticles inof (A) 50g/ml trypsin and (B) 5 mg/ml pepsin. Glutaralde-

rations between 40 and 100% were used for particle crosslinking; n = 3).

H value. After 180 min about 16.0% of the particlesslinking degree of 100% were degraded. This amounto 74.5% after 24 h. A difference between the distinctyde crosslinking degrees was not detectable. Parti-tion by enzymes of the cathepsin family at slightlyonfirms the biodegradability of HSA nanoparticles

ar uptake.the results of enzymatic degradation together, it hasluded that the enzymes trypsin, proteinase K, pro-epsin are well suited for a rapid degradation of HSAnanoparticles. This could be useful for the analy-compounds incorporated within the matrix of thelate drug delivery system. In the presence of all of

mes the degradation kinetics was dependent on theg degree of the particles with higher degrees leading

e spans of degradation. Under the chosen conditionsence of trypsin, proteinase K, protease, and cathep-ntitative degradation of 100% crosslinked particles

ssible over a time frame of 24 h. In contrast to theseilised particles a lower crosslinking degree of 40%rticle degradation within 24 h by all enzymes exceptin at pH 6.4. The choice of the respective enzyme iscted by the pH dependent stability of an incorporated

Fig. 8. Enzymcathepsin B abetween 40 an

drug. If thedegradationa suitable munder neutthe enzyme

4. Conclu

The prepared withbatch of theenced by thhigher amoof the nanoby the pHtion. HigheBesides HSalso be pre

Enzymapossible wdepends onrespectivewith pepsinK, cathepsatic degradation of HSA nanoparticles in the presence of 10g/mlt pH values of (A) 6.4 and (B) 5.4. Glutaraldehyde concentrationsd 100% were used for particle crosslinking (mean S.D.; n = 3).

drug is stable under acidic conditions, nanoparticlein the presence of pepsin and cathepsin at pH 5.4 isethod whereas if the respective drug is only stable

ral conditions trypsin, proteinase K or protease ares of choice for particle degradation.

sion

sent study shows that HSA nanoparticles can be pre-predictable and reproducible size dependent on thestarting material HSA used. The preparation is influ-e amount of high molecular HSA components, withunts leading to an increase of size and polydispersityparticles. These parameters can also be influenced

value of the HSA solution used for particle prepara-r pH values decrease particle size and polydispersity.A of human origin monodisperse nanoparticles can

pared under comparable conditions with rHSA.tic degradation of HSA and rHSA nanoparticles isith different enzymes. The kinetics of degradation

the crosslinking degree of the particles and theenzyme used. The fastest digestion can be achievedat pH 2. Under neutral conditions trypsin, proteinase

in B, and protease enable particle degradation over


24 h. Particle degradation in the presence of the intracellularenzyme cathepsin B confirms the biodegradability of HSA andrHSA nanoparticles as a prerequisite of drug release after cellularuptake.

Acknowledgements

This work was financially supported by the German Bun-desministerium fur Bildung und Forschung (BMBF) (Project13N8671). The authors acknowledge Tosoh Bioscience GmbHStuttgart, Germany, for the provided SEC column.

References

Balthasar, S., Michaelis, K., Dinauer, N., von Briesen, H., Kreuter, J., Langer, K.,2005. Preparation and characterisation of antibody modified nanoparticles asdrug carrier system for uptake in lymphocytes. Biomaterials 26, 27232732.

Carter, D.C., Ho, J.X., 1994. Structure of serum albumin. Adv. Protein Chem.45, 153203.

Chuang, V.T., Kragh-Hansen, U., Otagiri, M., 2002. Pharmaceutical strategiesutilizing recombinant human serum albumin. Pharm. Res. 19, 569577.

Coester, C.J., Langer, K., von Briesen, H., Kreuter, J., 2000. Gelatin nanopar-ticles by two step desolvationa new preparation method, surfacemodifications and cell uptake. J. Microencapsul. 17, 187193.

Couvreur, P., Vauhtier, C., 2006. Nanotechnology: intelligent design to treatcomplex disease. Pharm. Res. 23, 14171450.

Desai, N., Trieu, V., Yao, Z., Louie, L., Ci, S., Yang, A., Tao, C., De, T., Beals,B., Dykes, D., Noker, P., Yao, R., Labao, E., Hawkins, M., Soon-Shiong,P., 2006. Increased antitumor activity, intratumor paclitaxel concentrations,and endothelial cell transport of cremophor-free, albumin-bound paclitaxel,ABI-007, compared with cremophor-based paclitaxel. Clin. Cancer Res. 12,13171324.

Gradishar, WHawkins,

albumin-bound paclitaxel compared with polyethylated castor oil-basedpaclitaxel in women with breast cancer. J. Clin. Oncol. 23, 77947803.

He, X.M., Carter, D.C., 1992. Atomic structure and chemistry of human serumalbumin. Nature 358, 209215.

Kobayashi, K., 2006. Summary of recombinant human serum albumin develop-ment. Biologicals 34, 5559.

Kreuter, J., 2004. Nanoparticles as drug delivery systems. In: Nalwa, H.S. (Ed.),Encyclopedia of Nanoscience and Nanotechnology, vol. 7. American Scien-tific Publishers, Stevenson Ranch, USA, pp. 161180.

Langer, K., Balthasar, S., Vogel, V., Dinauer, N., von Briesen, H., Schubert, D.,2003. Optimization of the preparation process for human serum albumin(HSA) nanoparticles. Int. J. Pharm. 257, 169180.

Michaelis, K., Hoffmann, M.M., Dreis, S., Herbert, E., Alyautdin, R.N.,Michaelis, M., Kreuter, J., Langer, K., 2006. Covalent linkage of apolipopro-tein E to albumin nanoparticles strongly enhances drug transport into thebrain. J. Pharm. Exp. Ther. 317, 12461253.

Mukherjee, S., Richik, N.G., Maxfield, F.R., 1997. Endocytosis. Physiol. Rev.77, 759803.

Muller, R.H., Mader, K., Gohla, S., 2000. Solid lipid nanoparticles (SLN) forcontrolled drug deliverya review of the state of the art. Eur. J. Pharm.Biopharm. 50, 161177.

Schuck, P., Perugini, M.A., Gonzales, N.R., Howlett, G.J., Schubert, D.,2002. Size-distribution analysis of proteins by analytical ultracentrifugation:strategies and application to model systems. Biophys. J. 82, 10961111.

Schuck, P., Rossmanith, P., 2000. Determination of the sedimentation coefficientdistribution by least-squares boundary modeling. Biopolymers 54, 328341.

Steinhauser, I., Spankuch, B., Strebhardt, K., Langer, K., 2006. Trastuzumab-modified nanoparticles: optimisation of preparation and uptake in cancercells. Biomaterials 27, 49754983.

Vogel, V., Langer, K., Balthasar, S., Schuck, P., Machtle, W., Haase, W., vanden Broek, J.A., Tziatzios, C., Schubert, D., 2002. Characterization ofserum albumin nanoparticles by sedimentation velocity analysis and electronmicroscopy. Progr. Colloid Polym. Sci. 119, 3136.

Wartlick, H., Spankuch-Schmitt, B., Strebhardt, K., Kreuter, J., Langer, K., 2004.Tumour cell delivery of antisense oligonucleotides by human serum albumin

partic

.J., Tjulandin, S., Davidson, N., Shaw, H., Desai, N., Bhar, P.,M., OShaughnessy, J., 2005. Phase III trial of nanoparticle nano les. J. Controll. Rel. 96, 483495.

Human serum albumin (HSA) nanoparticles: Reproducibility of preparation process and kinetics of enzymatic degradationIntroductionMaterials and methodsReagents and chemicalsComposition of HSA batches by size exclusion chromatography (SEC) and analytical ultracentrifugation (AUC)Preparation of HSA nanoparticlesPreparation of rHSA nanoparticlesDetermination of particle size and size distributionEnzymatic degradation of HSA nanoparticles

Results and discussionPurity of different HSA batchesReproducibility of HSA nanoparticle preparationPreparation of rHSA nanoparticlesEnzymatic degradation of HSA nanoparticles

ConclusionAcknowledgementsReferences

Human Serum Albumin (HSA) Nanoparticles Reproducibility Of

Documents

Transcript of Human Serum Albumin (HSA) Nanoparticles Reproducibility Of