lable at ScienceDirect

Biomaterials 30 (2009) 431–435

Contents lists avai

Biomaterials

journal homepage: www.elsevier .com/locate/biomateria ls

In situ polymerization of tropoelastin in the absence of chemical cross-linking

Suzanne M. Mithieux a, Yidong Tu a, Emine Korkmaz b, Filip Braet b, Anthony S. Weiss a,*

a School of Molecular and Microbial Biosciences, University of Sydney, Building G08, Sydney, NSW 2006, Australiab Australian Key Centre for Microscopy and Microanalysis, University of Sydney, NSW 2006, Australia

a r t i c l e i n f o

Article history:Received 10 September 2008Accepted 20 October 2008Available online 8 November 2008

Keywords:Biomimetic materialElastinElasticityExtracellular matrixHydrogelSelf-assembly

* Corresponding author. Tel.: þ61 2 9351 3434; faxE-mail address: a.weiss@usyd.edu.au (A.S. Weiss).

0142-9612/$ – see front matter � 2008 Elsevier Ltd.doi:10.1016/j.biomaterials.2008.10.018

a b s t r a c t

Tropoelastin, the polypeptide monomer precursor of elastin, is covalently cross-linked to give stableelastic structures. We show here that elastic biomaterials can be generated from tropoelastin in theabsence of the classically accepted cross-linking pathway. Under alkaline conditions tropoelastinproceeds through a sol–gel transition leading to the formation of an irreversible hydrogel. This does notoccur at neutral pH. The resulting biomaterial is stable, elastic and flexible. Scanning electron microscopyrevealed that the hydrogel forms through the coalescence of w1 mm quantized protein spheres. Thesespheres resemble the tropoelastin-rich globules that accumulate on cultured cell surfaces during elastinformation. In vitro cell culture studies demonstrate that the hydrogel can support human skin fibroblastproliferation. In vivo studies demonstrate that following injection, the tropoelastin solution undergoesrapid localized gelation to form a persistent mass. These subcutaneous rodent injection data establish thematerial’s potential as a novel cell-compatible elastic scaffold that can be formed in situ.

� 2008 Elsevier Ltd. All rights reserved.

1. Introduction

The protein polymer elastin contributes essential structural,mechanical and biological properties to the extracellular matrixand is therefore critical for the long-term function of a variety oftissues. Recognition of the value of these properties to tissue-engineered constructs has led to increasing interest in thedevelopment of elastin-based biomaterials [1]. Devitalized elastin-containing tissue, purified insoluble elastin and solubilised elastinpeptides from exogenous animal sources have formed the basis fora number of these constructs. However, many of these materialssuffer from heterogeneity and persistent biocompatibilityproblems. Recombinant technology has been used to produceelastin-like peptides, elastin-based polypeptides and tropoelastin,all of which contribute towards addressing these issues [2–5].

Macroassembled biomaterials that are made from connectivetissue proteins classically need to be cross-linked in order toprevent dissociation and impart stability. Soluble collagen andelastin-based molecules including tropoelastin and fragments arestabilised through the use of a variety of chemical cross-linkers thatinclude 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDC)and N-hydroxysuccinimide (NHS) [6], glutaraldehyde [7], bis(sulfo-succinimidyl)suberate [5], genipin and pyrroloquinoline quinone[8] and 1,6-diisocyanatohexane (HMDI) [9]. The use of chemicalcross-linkers in implantable materials impacts upon approval and

: þ61 2 9351 3467.

All rights reserved.

manufacturing processes, and can limit their application for humanuse.

Classically, elastin is formed in a multistage process thatincludes (1) the assembly of multiple soluble tropoelastinmolecules during coacervation, and (2) lysyl oxidase-mediatedcross-linking of the coacervated tropoelastin molecules. Thisprocess renders the resulting elastin insoluble and extremelydurable. Coacervation is recognised as a vital stage in elastogenesis,responsible for the alignment of tropoelastin molecules such thattheir lysine residues are situated in close proximity [10–12]. Thiscoacervate consists of amassed protein-rich spherical particles[13–15] and is fully reversible upon cooling. Dogma states that thecoacervate is rendered irreversible solely through the generation ofcovalent cross-links between the tropoelastin monomers.

Here we explore whether an elastic hydrogel can be made byassembling tropoelastin in the absence of a cross-linker. A single-component biomaterial would have potential as a tissue-bulkingagent that can be administered through a minimally invasiveprocedure. Elevated pH was assessed for its ability to effectirreversible macroassembly in order to generate a stablebiomaterial.

2. Materials and methods

2.1. Tropoelastin purification

Recombinant human tropoelastin isoform SHELD26A (Synthetic Human ELastinwithout domain 26A) corresponding to amino acid residues 27–724 of GenBankentry AAC98394 (gi 182020) was purified from bacteria on a multi-gram scale aspreviously described [16].

Fig. 1. (A) Biacore analysis of tropoelastin:tropoelastin interactions at increasingsolution pH. (B) Effect of pH on tropoelastin assembly.

Fig. 2. (A) CD spectra for tropoelastin solutions at pH 7 and 10.8 and 20 �C and 37 �C.(B) CDPro analysis of each solution’s predicted secondary structure content. (C)Difference spectra showing the classical curves expected for polyproline type II (PPII)structure.

S.M. Mithieux et al. / Biomaterials 30 (2009) 431–435432

2.2. Binding studies of tropoelastin under increasing pH conditions

Tropoelastin was immobilized onto the surface of a CM5 sensor chip at a level of6900 RU (Response unit) and Biacore 3000 studies were performed. Biacore runningbuffers (0.01 M HEPES, 0.2 M NaCl, 0.005% Tween 20) were prepared at 9 different pHlevels: 5, 5.5, 6, 6.5, 7, 7.5, 8, 8.5 and 9. Prior to each run the system was primed withthe appropriate running buffer. A 1 mM tropoelastin solution in the corresponding pHbuffer was then injected over the chip surface at a flow rate of 40 ml/min.Background non-specific binding was removed by subtracting values obtained fromblank chip surfaces that were run in parallel. Samples were injected for 2 min anddissociated for 3 min, followed by a 1-min injection of regeneration solution(1 M NaCl, 0.05% NaOH). All experiments were performed with 2–3 replicates at25 �C.

2.3. Circular dichroism spectroscopy

The effect of pH on the secondary structure of a 0.2 mg/ml tropoelastin solutionin water was monitored using 0.1 cm quartz cuvettes in a Jasco J-720spectropolarimeter. Twelve spectral accumulations were obtained for each sample.Spectra were measured from 180 to 260 nm. Temperature was controlled witha Neslab RTE-111 water bath. Secondary structure analysis of the CD spectra wasperformed with the CONTIN and CDSSTR programs through the CDPro softwarepackage using reference set SP43 [17].

2.4. Preparation of tropoelastin

For all further experiments tropoelastin was dissolved in PBS (10 mM Na phos-phate; 150 mM NaCl, pH 7.4). Solutions were filtered through a 0.22 mm MillexGPfilter and the pH was adjusted to pH 10 to 11 using sterile 0.1–1 M NaOH. Eachsolution was then diluted to the requisite final concentration.

2.5. Particle size analysis

Particle size analysis of a 2 mg/ml tropoelastin solution was performed on aMalvern Zetasizer. Following varying incubation times at 25 �C or 37 �C, thetropoelastin solution was allowed to equilibrate to the required temperature for5 min prior to measurement. Samples were measured at either 25 �C or 37 �C. Forphotography, the tropoelastin solution was observed at room temperature usinga Zeiss Axiovert 200 microscope and images were acquired using an HSm mono-chrome digital camera and 3i Stallion software (Zeiss).

Fig. 3. Particle size analysis of a 2 mg/ml tropoelastin solution at pH 10.4 over time.Inset: Stable spherical tropoelastin aggregates following a week-long incubation at37 �C.

Fig. 4. Stress–strain curves measured over 5 cycles. Inset: Tropoelastin-based hydrogelformed in a tubular mold then tied in a knot to demonstrate the flexibility of thematerial.

S.M. Mithieux et al. / Biomaterials 30 (2009) 431–435 433

2.6. Mechanical tests

Elastic sheets were formed by polymerizing 100 mg/ml tropoelastin in a glassmold for 16 h at 37 �C. Mechanical tests were performed on either dry or PBS-wetsheets. Stress–strain curves were generated with an EnduraTec model ELF3400biaxial materials’ testing machine (Bose Corporation, MN). Smoothed dynamicresponses were recorded over multiple cycles and used to calculate the Young’smodulus.

2.7. In vitro cell culture

Aliquots (0.5 ml) of filter-sterilized 100 mg/ml tropoelastin were polymerized inthe wells of 24-well tissue culture plates for 16 h at 37 �C. The polymer was equil-ibrated with cell culture medium (DMEM containing 10% fetal calf serum) prior to

Fig. 5. (A) Tropoelastin spherules coalesce to form a solid elastic mass. (B–D) Skin fibrobl0.1 mm).

the addition of 5�105 skin fibroblasts (GM02037, Coriell Research Institute, NJ) perwell. Media were replaced after 12 h and the cells were cultured for 24 h prior tofixing.

2.8. Scanning electron microscopy

Samples were fixed with 2% glutaraldehyde in 0.1 M Na-cacodylate buffer with0.1 M sucrose for 1 h at 37 �C. Samples underwent post-fixation with 1% osmium in0.1 M Na-cacodylate for 1 h and were then dehydrated in ethanol solutions of 70%,80%, 90% and 3 times 100%, 10 min each. For drying, the samples were thenimmersed for 3 min in 100% hexamethyldisilazane. The samples were thentransferred to a desiccator for 25 min to avoid water contamination. Finally thesamples were mounted on stubs, sputter coated with 10 nm gold and examined witha Philips SEM 505 at 30 kV [18].

2.9. In vivo animal tests

200 mg/ml tropoelastin was partially polymerized by incubation at 37 �C for 2 hand then kept at 4 �C prior to use. 0.1 ml was injected intradermally using a 25-gauge needle into female Sprague–Dawley rats. Normal rat physiology wasconfirmed over a period of 15 days. On day 15 the injection site and adjacent skinwere excised and fixed in 10% formalin for 48 h at 4 �C. Samples were then dehy-drated through a graded series of ethanol concentrations (70%, 80%, 95% �2, 100%�2) and transferred to xylene, then embedded in paraffin wax. Sections (5 mm) werecut on a rotary microtome and dried before being deparaffinized through xylene anda graded series of ethanols (100%�2, 95%�2, 70%) and transferred to water. Sectionswere stained with either hematoxylin & eosin or Milligan’s Trichrome. Immuno-histostaining used BA4 elastin-specific antibody.

3. Results and discussion

3.1. The effect of pH on tropoelastin association

We have previously shown that coacervation is facilitated byraising the pH of a tropoelastin solution from pH 5.4 to 8.4 [19].The effect of pH on tropoelastin association was further examinedusing surface plasmon resonance under conditions that areconducive to tropoelastin dimerization [20]. Tropoelastin was

ast attachment to the biomaterial (white scale bar in images A–C is 10 mm; in D it is

Fig. 6. Histological specimens showing injection site explants stained with (A)hematoxylin and eosin (40� magnification), (B) BA4 anti-elastin antibody (40�magnification)and (C) Milligan’s Trichrome (100� magnification); nuclei stain purpleand collagen stains blue/green. The elastic deposit is marked with an E.

S.M. Mithieux et al. / Biomaterials 30 (2009) 431–435434

immobilized on a Biacore CM5 sensor chip surface and challengedwith tropoelastin at increasing pH values. Tropoelastin:tropoelastin interactions increased as the pH was increased frompH 5 to 9 (Fig. 1A), consistent with promoted tropoelastin asso-ciation. We speculated that the trend would continue withincreasing pH but it would not be possible to investigate the effectof pH further with this methodology as buffers with pH >9 stripimmobilized protein off the chip surface [21]. To overcome thisconstraint, further investigations were conducted visually on10 mg/ml tropoelastin solutions at higher pH. These studiesrevealed that tropoelastin solutions were capable of generatinga solid, stable mass that was particularly evident at 37 �C at andabove pH 10. In stark contrast, the classical yellow oily coacervate[22] was seen below pH 9 (Fig. 1B).1

3.2. The secondary structure of tropoelastin at high pH

Circular dichroism was used to investigate the secondarystructure of tropoelastin at pH 7 and pH 10.8 at 20 �C and 37 �C(Fig. 2A). Secondary structure calculations revealed that thesecondary structure of tropoelastin was substantially conserved(Fig. 2B). However subtle differences were observed. An isodichroicpoint at 208 nm suggested a structural transition. Differencespectra at each of the temperatures (Fig. 2C) showed a consistentprofile that was characteristic of the classical curves expected forpolyproline type II (PPII) [23,24] with a substantial minimum atw195 nm and a slightly positive shoulder at 223 nm. These curvesreveal that the only change is a loss of some PPII during the tran-sition from pH 7 to pH 10.8.

3.3. High pH and low protein concentrations

Particle size analysis of 2 mg/ml tropoelastin at pH 10.4 wasperformed (Fig. 3). Initially at 25 �C, the average particle diameterin solution was w10 nm which corresponded to the tropoelastinmonomer [13]. Over 24 h at 37 �C, the assembling particlediameter grew to w1 mm and remained stable, even when re-equilibrated to 25 �C. Remarkably, when placed at 37 �C fora week, the particle size did not change (Fig. 3 inset). Thesetropoelastin spheres resemble the tropoelastin globules that areintermediates in elastin macroassembly [15]. This raises thepossibility that the spheres could be used as a building block forelastin synthesis.

3.4. Hydrogel formation

At concentrations of tropoelastin of at least10 mg/ml, a revers-ible sol–gel transition was seen where solutions were incubated forshort periods at pH 10. Stable solid hydrogels were formed from100 mg/ml solutions that were incubated for more than 5 h at37 �C. The resulting material was stable, elastic and flexible. Todemonstrate these properties, the wet biopolymer was formed ina tubular mold then tied in a knot. Mechanical tests were per-formed on wet material that was formed from a 100 mg/ml tro-poelastin solution at pH 10.8 for 16 h at 37 �C. The Young’s modulusaveraged 1.69� 0.04 MPa over 5 cycles (Fig. 4). This value compareswith the reported Young’s modulus for elastin of w1 MPa [25].

3.5. In vitro studies

Human skin fibroblasts colonized the hydrogel surface. SEMimages revealed an upper layer consisting of spheres that coalesced

1 For interpretation of colour in the sentence, the reader is referred to the webversion of the article.

to form a dense solid elastic lower layer (Fig. 5A). Individual cellscould immerse themselves within the layer of spheres (Fig. 5B). Theability of cells to attach and proliferate on the biomaterial wasdemonstrated by sheets of cells that were seen to populate thesurface (Fig. 5C and D).

3.6. In vivo animal tests

In vivo experiments demonstrated that subcutaneous injectionand in situ gelation were feasible. The resulting elastic depositpersisted for 2 weeks, elicited a mild foreign body response andpromoted collagen deposition and encapsulation of the material.

Hematoxylin and eosin staining revealed a substantial amor-phous material in the hypodermis (Fig. 6A) including the looseconnective tissue beneath the cutaneous muscle. Immunostainingwith the elastin-specific BA4 antibody demonstrated that this wasa large tropoelastin-based deposit (Fig. 6B). Curiously, it did notstain appropriately for elastin using a Verhoeff–Van Gieson stain.This reverse staining effect may be due to the high pH [26].Milligan’s Trichrome staining revealed the injected mass, sur-rounded by a fine fibrous capsule (Fig. 6C). This cellular and fibrillarresponse can confer stability and elicit a persistent augmentativeeffect from the biomaterial [27,28].

S.M. Mithieux et al. / Biomaterials 30 (2009) 431–435 435

4. Conclusion

Exposing tropoelastin to an alkaline environment initiated andpromoted irreversible self-assembly. This resulted in a sol–geltransition and the formation of a stable, elastic biomaterial withoutthe requirement for enzyme- or chemical-mediated cross-linking.Tropoelastin self-associated to give a slurry of stable proteinspheres that coalesced to generate solid, robust hydrogels.Mechanical, in vitro and in vivo studies reveal its potential as a cell-supportive material derived solely from recombinant humanprotein. This novel use of an alkaline treatment represents a newrange of tropoelastin-based formulations that can be either moldedor contemplated as injectable biomaterials.

Acknowledgement

A.S.W. acknowledges support from the Australian ResearchCouncil.

Appendix

Figures with essential colour discrimination. Fig. 6 of this articleis difficult to interpret in black and white. The full colour image canbe found in the on-line version, at doi:10.1016/j.biomaterials.2008.10.018.

References

[1] Daamen WF, Veerkamp JH, van Hest JC, van Kuppevelt TH. Elastin asa biomaterial for tissue engineering. Biomaterials 2007;28:4378–98.

[2] Chilkoti A, Christensen T, MacKay JA. Stimulus responsive elastin biopolymers:applications in medicine and biotechnology. Curr Opin Chem Biol2006;10:652–7.

[3] Lee J, Macosko CW, Urry DW. Elastomeric polypentapeptides cross-linked intomatrixes and fibers. Biomacromolecules 2001;2:170–9.

[4] Bellingham CM, Lillie MA, Gosline JM, Wright GM, Starcher BC, Bailey AJ, et al.Recombinant human elastin polypeptides self-assemble into biomaterialswith elastin-like properties. Biopolymers 2003;70:445–55.

[5] Mithieux SM, Rasko JE, Weiss AS. Synthetic elastin hydrogels derived frommassive elastic assemblies of self-organized human protein monomers.Biomaterials 2004;25:4921–7.

[6] Daamen WF, Nillesen ST, Wismans RG, Reinhardt DP, Hafmans T, Veerkamp JH,et al. A biomaterial composed of collagen and solubilized elastin enhancesangiogenesis and elastic fiber formation without calcification. Tissue Eng PartA 2008;14:349–60.

[7] Welsh ER, Tirrell DA. Engineering the extracellular matrix: a novel approach topolymeric biomaterials. I. Control of the physical properties of artificial proteinmatrices designed to support adhesion of vascular endothelial cells. Bio-macromolecules 2000;1:23–30.

[8] Vieth S, Bellingham CM, Keeley FW, Hodge SM, Rousseau D. Microstructuraland tensile properties of elastin-based polypeptides crosslinked with genipinand pyrroloquinoline quinone. Biopolymers 2007;85:199–206.

[9] Li M, Mondrinos MJ, Gandhi MR, Ko FK, Weiss AS, Lelkes PI. Electrospunprotein fibers as matrices for tissue engineering. Biomaterials 2005;26:5999–6008.

[10] O’Dell BL, Elsden DF, Thomas J, Partridge SM, Smith RH, Palmer R. Inhibition ofthe biosynthesis of the cross-links in elastin by a lathyrogen. Nature1966;209:401–2.

[11] Siegel RC, Pinnell SR, Martin GR. Cross-linking of collagen and elastin. Prop-erties of lysyl oxidase. Biochemistry 1970;9:4486–92.

[12] Rosenbloom J. Elastin: relation of protein and gene structure to disease. LabInvest 1984;51:605–23.

[13] Clarke AW, Arnspang EC, Mithieux SM, Korkmaz E, Braet F, Weiss AS. Tro-poelastin massively associates during coacervation to form quantized proteinspheres. Biochemistry 2006;45:9989–96.

[14] Osborne JL, Farmer R, Woodhouse KA. Self-assembled elastin-like polypeptideparticles. Acta Biomater 2008;4:49–57.

[15] Kozel BA, Rongish BJ, Czirok A, Zach J, Little CD, Davis EC, et al. Elastic fiberformation: a dynamic view of extracellular matrix assembly using timerreporters. J Cell Physiol 2006;207:87–96.

[16] Wu WJ, Vrhovski B, Weiss AS. Glycosaminoglycans mediate the coacervationof human tropoelastin through dominant charge interactions involving lysineside chains. J Biol Chem 1999;274:21719–24.

[17] Sreerama N, Woody RW. Estimation of protein secondary structure fromcircular dichroism spectra: comparison of CONTIN, SELCON, and CDSSTRmethods with an expanded reference set. Anal Biochem 2000;287:252–60.

[18] Braet F, De Zanger R, Wisse E. Drying cells for SEM, AFM and TEM by hex-amethyldisilazane: a study on hepatic endothelial cells. J Microsc 1997;186:84–7.

[19] Vrhovski B, Jensen S, Weiss AS. Coacervation characteristics of recombinanthuman tropoelastin. Eur J Biochem 1997;250:92–8.

[20] Tu Y, Weiss AS. Glycosaminoglycan-mediated coacervation of tropoelastinabolishes the critical concentration, accelerates coacervate formation, andfacilitates spherule fusion: implications for tropoelastin microassembly. Bio-macromolecules 2008.

[21] Choi JM, Hutson AM, Estes MK, Prasad BV. Atomic resolution structuralcharacterization of recognition of histo-blood group antigens by Norwalkvirus. Proc Natl Acad Sci U S A 2008;105:9175–80.

[22] Vrhovski B, Weiss AS. Biochemistry of tropoelastin. Eur J Biochem 1998;258:1–18.

[23] Ma K, Kan L, Wang K. Polyproline II helix is a key structural motif of the elasticPEVK segment of titin. Biochemistry 2001;40:3427–38.

[24] Bochicchio B, Pepe A, Flamia R, Lorusso M, Tamburro AM. Investigating theamyloidogenic nanostructured sequences of elastin: sequence encoded byexon 28 of human tropoelastin gene. Biomacromolecules 2007;8:3478–86.

[25] Fung YC. Biomechanics. Mechanical properties of living tissues. 2nd ed. NewYork: Springer-Verlag; 1993.

[26] Dickson CP, Robinson TF. Differentiating cardiac elastin, collagen andmicrofibrils with NaOH at the ultrastructural level. Histochemistry1988;89:105–7.

[27] Eppley BL, Summerlin DJ, Prevel CD, Sadove AM. Effects of a positively chargedbiomaterial for dermal and subcutaneous augmentation. Aesthetic Plast Surg1994;18:413–6.

[28] Lemperle G, Morhenn V, Charrier U. Human histology and persistence ofvarious injectable filler substances for soft tissue augmentation. AestheticPlast Surg 2003;27:354–66 [discussion 367].