Development and initial characterization of a chemically stabilized elastin-glycosaminoglycan-collagen composite shape-memory hydrogel for nucleus pulposus regeneration Jeremy Mercuri,1 Caroline Addington,1 Richard Pascal III,1 Sanjitpal Gill,1,2 Dan Simionescu1 1
Department of Bioengineering, Biocompatibility and Tissue Regeneration Laboratory, Clemson University, Rhodes Engineering Research Center, Clemson, South Carolina 29634 2 Department of Orthopaedic Surgery, The Village at Pelham, Spartanburg Regional Healthcare System, Greer, South Carolina 29650 Received 26 September 2013; revised 27 December 2013; accepted 29 January 2014 Published online 25 February 2014 in Wiley Online Library (wileyonlinelibrary.com). DOI: 10.1002/jbm.a.35104 Abstract: Nucleus pulposus (NP) is a resilient and hydrophilic tissue which plays a significant role in the biomechanical function of the intervertebral disc (IVD). Destruction of the NP extracellular matrix (ECM) is observed during the early stages of IVD degeneration. Herein, we describe the development and initial characterization of a novel biomaterial which attempts to recreate the resilient and hydrophilic nature of the NP via the construction of a chemically stabilized elastin-glycosaminoglycan-collagen (EGC) composite hydrogel. Results demonstrated that a resilient, hydrophilic hydrogel which displays a unique “shape-memory” sponge characteristic could be formed from a blend of soluble elastin aggregates, chondroitin-6-sulfate, hyaluronic acid and collagen following freeze-drying, stabilization with a carbodiimide and penta-galloyl glucose-based fixative, and subsequent partial degradation with glycosaminoglycan degrading
enzymes. The resultant material exhibited the ability to restore its original dimensions and water content following multi-cycle mechanical compression and illustrated resistance to accelerated enzymatic degradation. Preliminary in vitro studies utilizing human adipose derived stem cells (hADSCs) demonstrated that the material was cytocompatible and supported differentiation towards an NP cell-like phenotype. In vivo biocompatibility studies illustrated host cell infiltration and evidence of active remodeling following 4 weeks of implantation. Feasibility studies demonstrated that the EGC hydrogel could be delivered via minimally invaC 2014 Wiley Periodicals, Inc. J Biomed Mater Res sive methods. V Part A: 102A: 4380–4393, 2014.
Key Words: nucleus pulposus, tissue regeneration, biopolymer, hydrogel, stem cells
How to cite this article: Mercuri J, Addington C, Pascal III R, Gill S, Simionescu D. 2014. Development and initial characterization of a chemically stabilized elastin-glycosaminoglycan-collagen composite shape-memory hydrogel for nucleus pulposus regeneration. J Biomed Mater Res Part A 2014:102A:4380–4393.
INTRODUCTION
Low back pain originating from a degenerate intervertebral disc (IVD); termed “discogenic pain,” is a major source of disability and economic hardship in the U.S. and around the world. Freemont suggests that 40% of cases presenting with low back pain can be attributed in part to dysfunction of the IVD.1 Katz has determined that conservative estimates of direct costs associated with low back pain; including office visits, hospital admissions and associated procedures approach $20 billion per year in the U.S., and that total costs may approach $100 to $200 billion annually.2 Furthermore, nearly 5.7 million Americans are diagnosed with IVD disorders annually.3 Taken together; these staggering statistics provide motivation for research and development of therapeutic methodologies for combating IVD degeneration. The biological mechanisms underlying the degenerative process are multi-faceted and investigators continue their
efforts to understand the biochemical, mechanical, nutritional and genetic complexities thought to play a role in this pathological condition. It is clear however, that one of the resulting outcomes associated with the interplay of these factors is the development of an aberrant cell-mediated imbalance of extracellular matrix (ECM) production and enzymatic breakdown which results in IVD tissue degeneration.4 It has been noted that the first observable histologic signs of degeneration appear within the central hydrogel core of the IVD; a tissue structure known as the nucleus pulposus (NP).5,6 Composed predominantly of type II collagen, hyaluronic acid, aggrecan and chondrocyte-like NP cells, the NP plays a large role in maintaining IVD hydration and consequently its resiliency and biomechanics. Notably, the native NP ECM develops a swelling pressure due to its high fixed charge density and low permeability which together support compressive loading of the IVD.7 This phenomenon
Additional Supporting Information may be found in the online version of this article. Correspondence to: J. Mercuri; e-mail:
[email protected]
4380
C 2014 WILEY PERIODICALS, INC. V
ORIGINAL ARTICLE
also contributes to the observed diurnal fluctuations in water content - a hallmark of the NP; which gradually loses water and height throughout the day as the spine is loaded, but re-establishes these parameters when the load is reduced during nightly rest/unloading periods. Given the critical nature of healthy NP in maintaining IVD function and considering its apparent early demise, numerous interventional strategies targeting replacement and/or regeneration of the NP in an attempt to mitigate the downstream effects of IVD degeneration are under development. The development of biomaterials for use as NP replacements or scaffolds for NP tissue regeneration play a crucial role in therapeutic strategies addressing IVD degeneration. These biomaterials can be broadly classified based on their constituent components as being either derived from (1) synthetic polymers: that is, urethanes, acrylamides, acrylonitriles, methacrylates, and vinyl alcohols8–16 or (2) naturally occurring biopolymers: that is, collagens, glycosaminoglycans, elastin, alginate, chitosan, fibrin, calcium phosphate or decellularized tissue matrices.17–28 A major perceived advantage of utilizing biomaterials constructed of natural biopolymers over their synthetic counterparts includes their ability to be used in conjunction with autologous or allogeneic cell sources thus creating a “living” implant as opposed to an inanimate and inert synthetic material. Natural biopolymers have exhibited the ability to (1) support cell viability, (2) mimic native ECM biochemistry and mechanical properties, (3) allow for natural remodeling without the production of toxic byproducts, and (4) provide natural cues to cells which may stimulate a healing or regenerative response; including the potential to support autologous stem cell differentiation.19,21,23,26,29 Elastin is an ECM protein found within many tissues endowing resilience and permitting long-range deformability and passive recoil without energy input.30 Elastin molecules have been shown to be present in the human IVD, which are thought to play a crucial role in aiding the restoration of intervertebral disc matrix deformation.28,31 Accordingly, some have investigated the use of elastin-based materials for cartilage and IVD regeneration.32–34 Moss et al. have recently developed and evaluated a chemically modified hyaluronan-elastin-like peptide composite hydrogel for use as a scaffold for NP tissue engineering.35 Results indicate improved mechanical properties compared with gels composed of only hyaluronan and exhibited maintenance of human NP cell viability and phenotype over a three-week culture period.35 A recombinant protein copolymer of silk and elastin produced by genetically modified Escherichia coli bacteria, known as the NuCore Injectable Nucleus (Spine Wave), has completed early stage clinical trials and is undergoing assessment to characterize the material for use as a cell delivery vehicle.18,36 In consideration of the aforementioned points, we hypothesized that an elastin-based hydrogel chemically crosslinked with ECM molecules typically found in the human NP (i.e. chondroitin sulfate and hyaluronic acid) would result in a biomaterial which demonstrates the hydrophilicity and mechanical resiliency demonstrated by
JOURNAL OF BIOMEDICAL MATERIALS RESEARCH A | DEC 2014 VOL 102A, ISSUE 12
native NP. Attainment of such properties would result in advantages over other commonly used, non-resilient materials. The prescribed biomaterial should also exhibit the ability to (1) resist conditions conducive to accelerated enzymatic degradation (as can be observed within the degenerative IVD), (2) support human stem cell viability and differentiation towards an NP cell-like phenotype which will facilitate eventual healthy NP formation, (3) elicit an acceptable in vivo host response including evidence of biocompatibility and host remodeling, and (4) be able to be implanted in a minimally invasive fashion. MATERIALS AND METHODS
Elastin-glycosaminoglycan-collagen (EGC) hydrogel formation Hyaluronic acid (5 mg/mL) (Sigma Aldrich), chondroitin-6sulfate (27 mg/mL) (Sigma Aldrich), and soluble elastin (40 mg/mL) (Elastin Products Company) were combined sequentially in a neutralized solution of soluble type I collagen (3 mg/mL) (PureCol-Advanced BioMatrix) while mixed on ice. The resulting blend was aliquoted in 0.5 mL volumes into 2 mL microfuge tubes and subsequently gelled at 37 C for 2 h. The formed gels were frozen at 280 C and lyophilized. Dried samples were chemically stabilized with 1.5 mL of 60 mM 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide EDC/12 mM N-hydroxysuccinimide (EDC/NHS) for 24 h. Following thorough rinsing in water, hydrogels were submerged in 1.5 mL of 0.15% pentagalloyl glucose (PGG— Ajinmoto OmniChem) in PBS (pH 7.4) for 24 h for further chemical stabilization before thorough rinsing in water. Pilot studies indicated that chemical stabilization of the hydrogel improved its physical integrity and increased resistance to enzymatic degradation compared with hydrogels without crosslinking (data not shown). Enzymatic processing—Development of EGC hydrogel resiliency (“shape-memory”) In order to develop a hydrogel that exhibited resiliency, it was hypothesized that targeted partial enzymatic digestion of crosslinked ECM components was required. Pilot studies indicated that crosslinked hydrogels were brittle and nonresilient; however, partial degradation induced by non-specific bacterial enzymes resulted in a resilient material displaying sponge-like characteristics. Accordingly, pilot studies investigating the use of specific ECM degrading enzymes were performed. EGC hydrogels were incubated at 37 C for 19 h in either 1 or 5 U/mL of collagenase type I (Sigma) or elastase (Elastin Products Company) or in a mixture of glycosaminoglycan (GAG) degrading enzymes (5 U/mL Hyaluronidase (HAase:Sigma) 1 0.1 U/mL Chondroitinase ABC (CSase:Sigma) or 2.5 U/mL Hyaluronidase 1 0.05 U/mL Chondroitinase). Resultant hydrogels were subsequently rinsed in water and subjected to unconfined compression. Compression testing was performed at a rate of 1.5 mm/s to a strain of 75%. Resiliency was determined via real-time video recording of pre- and postcompression hydrogel heights using a metric ruler in addition to documenting the time required to obtain significant height recovery (defined as the re-establishment
4381
of greater than 90% of original sample height in 3 s). The enzyme solution which resulted in EGC hydrogel samples exhibiting greater than 90% height recovery in less than 3 s following compressive load removal was used for processing all subsequent hydrogels. The resulting resiliency phenomenon was defined as a “shape-memory” sponge characteristic due to the fact that the hydrogels could be compressed expelling their water content and subsequently allow for folding or rolling into a small dehydrated pellet. The material would remain in this formation until physiologic fluid was added to the pellet, at which time the EGC material would immediately absorb the fluid and re-establish its original geometry. To ensure controlled enzymatic degradation of the EGC hydrogels, enzymes were heat inactivated and hydrogels thoroughly rinsed in water. To confirm that the temperature required to heat inactivate the enzymes would not denature the hydrogel itself, differential scanning calorimetry (TA Instruments—Model 2920) was performed on the crosslinked EGC hydrogels to determine melt temperature. Briefly, crosslinked hydrogels (n 5 4) were heated from 225 C to 200 C at 10 C/min, heat flow versus temperature plots were recorded on Universal V3.9A TA Instrument analytical software and melt temperatures were determined at the apex of the endothermic peaks. Additionally, sample mass before and following testing was recorded to ensure no sample mass loss had occurred. To confirm heat inactivation of the enzyme solution, an enzyme kinetics study was performed. Briefly, a glycosaminoglycan (GAG) substrate solution (16 mg/mL chondroitin-6sulfate and 4 mg/mL hyaluronic acid) was made in PBE buffer (5 mM L-Cystein, 5 mM EDTA, 100 mM dibasic phosphate buffer, pH 7.5). A 100 mL aliquot of this solution was added to 1 mL aliquots of the GAG degrading enzyme mixture (5 U/mL Hyaluronidase and 0.1 U/mL Chondroitinase ABC). Active GAG degradation within this solution was determined by performing a dimethylmethylene blue (DMMB) assay on 50 mL aliquots (in triplicate) taken from the mixture of enzyme and substrate over progressive time intervals. Sample absorbances were plotted with respect to time to develop an enzyme digestion kinetic curve. GAG digestion was determined via a reduction in DMMB assay absorbance. We have determined in pilot studies (data not shown) that this multivalent cationic dye does not have the capability to bind degraded anionic glycosaminoglycan and therefore can be exploited to determine enzyme inactivation (as indicated by positive staining of intact, nondegraded glycosaminoglycan which results in a change in colorimetric absorbance as compared with degraded glycosaminoglycan samples). For comparison with the kinetic curve, 1 mL aliquots of the enzyme solution were heated to 55 C (the proposed inactivation temperature) for either 15, 30, or 45 min in a water bath before cooling and the addition of 100 mL aliquots of GAG substrate solution. The solutions were subsequently incubated at 37 C for 2 h in an oven to allow for digestion of the GAG substrate by any active enzymes. Again, DMMB analysis was performed in duplicate on 50 mL aliquots of these solutions for comparison of absorbance values to the kinetic curve and to nondigested GAG substrate control values.
4382
MERCURI ET AL.
Biomechanical characterization of EGC hydrogels Compressive mechanical properties were evaluated in unconfined conditions under displacement control with samples (n 5 6) submerged in 37 C PBS. The testing protocol included the application of a 0.05N preload immediately followed by pre-conditioning using 10 cycles of compression to 25% strain at a speed of 12.5 mm/min. The tenth loading cycle was used to calculate percent energy dissipation (hysteresis), determined as the percent difference between the area under the loading and unloading curves, and apparent linear region modulus, which was calculated from the linear region of the stress-strain loading curve between 0% and 25% compressive strain. Additionally, to illustrate their extreme resiliency, EGC samples underwent five additional hysteresis cycles to 50% strain, and percent energy dissipation and the modulus between 35% and 50% strain was calculated. Hydrogel wet weights were also measured before and immediately following testing to determine the change in water content to illustrate the hydrogels capacity to retain and re-absorb water. Next, hydrogels were unloaded and subjected to incremental stress relaxation testing to a final strain of 25% using 5% strain increments held for 20 min each. This test method was adapted from Cloyd et al.37 Equilibrium was reached when the observed change in stress was less than 0.001 kPa per second. Equilibrium modulus was determined from the slope of an equilibrium stress versus applied strain plot. Percent relaxation was determined by dividing the equilibrium stress by the peak stress and multiplying by 100. Additionally, semiquantitative testing of hydrogel resiliency following unconfined compression was performed using video capture and digital imaging software to track pre- and post-compression hydrogel height and time to recovery. EGC hydrogel resistance to enzymatic degradation EGC hydrogel component stabilization and overall resistance to enzymatic degradation was determined by evaluating the change in hydrogel dry mass following digestion in elastin, GAG and collagen degrading enzymes, respectively. Hydrogel samples (n 5 4 per study group per time point) were incubated in a solution of 1 U/mL collagenase or 1 U/mL elastase (both in 50 mM Tris, 10 mM CaCl3; pH 7.6), or a mixture of GAG degrading enzymes (0.1 U/mL chondroitinase ABC and 5.0 U/mL hyaluronidase in 100 mM ammonium acetate buffer; pH 7.4) at 37 C for 2 and 7 days. Additionally, noncrosslinked hydrogels digested in each respective enzyme solution and nondigested crosslinked hydrogels served as control groups for comparison. Percent mass loss was determined gravimetrically for each hydrogel by comparing pre- and postdigestion sample dry weights. Human adipose derived stem cell culture and differentiation on EGC hydrogels EGC hydrogels manufactured as described above, were sterilized via submersion in a solution of 0.1% peracetic acid in phosphate buffered saline (pH 7.4) for 4 h, before thorough rinsing in sterile water and incubation in a 50:50 mixture of fetal bovine serum and Dulbecco’s Modified Eagle’s Medium
ELASTIN-BASED BIOMATERIAL FOR COMBATING IVD DEGENERATION
ORIGINAL ARTICLE
(DMEM) containing 1% antibiotic/antimycotic. Human adipose derived stem cells (hADSCs—Passage 2–4: Invitrogen) were seeded onto the EGC hydrogels by exploiting the sponge-like characteristic of the material. Briefly, EGC hydrogels pre-incubated in media were partially dehydrated via manual compression with sterile forceps on a sterile field. Hydrogels were rehydrated via the drop-wise addition of 200 mL of differentiation media (DMEM with the addition of 1% fetal bovine serum, 1% antibiotic/antimycotic, 5 mg/ mL insulin, 5 mg/mL transferrin, 5 ng/mL sodium selenite, 10 ng/mL transforming growth factor beta, and 50 nM ascorbate-2-phosphate) containing approximately 800,000 hADSCs per sample. The cell suspension was absorbed into the material which concomitantly expanded and re-acquired its pre-dehydrated shape resulting in a targeted seeding density of 4 3 105 cells/cm3. Samples were cultured for a total of 14 days in 24-well plates containing differentiation media, which was changed every 2 to 3 days. For comparison, hADSCs cultured in monolayer in differentiation media were used as positive differentiation controls. Samples analyzed for gene expression (n 5 3 per timepoint) and biochemical evaluation (n 5 3 per time-point) were snap frozen in liquid nitrogen and stored at 280 C until time of analysis. Samples used for histological analysis (n 5 2 per study group) were stored in 10% neutral buffered formalin before standard histological processing. Samples for LIVE/DEAD fluorescent imaging (n 5 2 per study group) were thoroughly rinsed in sterile PBS before assay according to manufacturer’s instructions. Reverse transcription PCR analysis of hADSC seeded on EGC hydrogels Total RNA from hADSC seeded EGC hydrogels were isolated using Trizol reagent (Invitrogen) according to the manufacturer’s instructions. RNA integrity and quantification was assessed using an Agilent 2100 Bioanalyzer with RNA 6000 Nano microfluidics chips according to the manufacturer’s instructions. A total of