A novel porous collagen scaffold around an implantable biosensor for improving biocompatibility. II. Long-term in vitro / in vivo sensitivity characteristics of sensors with NDGA- or GA-crosslinked collagen scaffolds

We have developed a new 3D porous and biostable collagen scaffold for implantable glucose sensors. The scaffolds were fabricated around the sensors and crosslinked using nordihydroguaiaretic acid (NDGA) or glutaraldehyde (GA) to enhance physical and
of 11
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  A novel porous collagen scaffold around an implantablebiosensor for improving biocompatibility. I.  In vitro  /  in vivo  stability of the scaffold and  in vitro  sensitivityof the glucose sensor with scaffold Young Min Ju, 1 Bazhang Yu, 2 Thomas J. Koob, 2 Yvonne Moussy, 3 Francis Moussy 1,2 1 Biomedical Engineering Program, University of South Florida, 4202 E. Fowler Avenue,ENB 118, Tampa, Florida 33620-5350 2 Department of Chemical Engineering, University of South Florida, 4202 E. Fowler Avenue,ENB 118, Tampa, Florida 33620-5350 3 Department of Mechanical Engineering, University of South Florida, 4202 E. Fowler Avenue,ENB 118, Tampa, Florida 33620-5350 Received 5 April 2007; revised 2 August 2007; accepted 20 August 2007Published online 17 December 2007 in Wiley InterScience ( DOI: 10.1002/jbm.a.31756 Abstract:  A new 3D porous and biostable collagen scaffoldhas been developed to improve the biocompatibility of implantable glucose sensors by minimizing tissue reactionswhile stimulating angiogenesis around the sensors. The novelcollagen scaffold was crosslinked using nordihydroguaiareticacid (NDGA) to enhance biostability. NDGA-treated collagenscaffolds were stable without physical deformation in thesubcutaneous tissue of rats for 4 weeks. In contrast, glutaral-dehyde (GA)-treated collagen scaffolds were extremely dam-aged following implantation. Both types of scaffolds (NDGA-and GA-crosslinked) were stable  in vitro  in the presence of collagenase with 70% retention of original weight after4 weeks of incubation. The response current (i.e. sensitivity)of sensors with porous scaffolds was not significantlychanged when compared with control sensors (no scaffold),while the response time ( T  95% ) was slightly delayed after aglucose concentration increase from 5 to 15 m  M . Above thisrange, the sensors coated with scaffolds had only a slightlylower sensitivity than the control sensors. These results indi-cate that we have developed a stable NDGA-crosslinked col-lagen scaffold for biosensors, and that the scaffold does notimpair the function of our sensor. We plan to use this scaf-fold to enhance the function and lifetime of the implantable biosensors by providing a controlled local environmentaround the sensors with the help of various drugs andgrowth factors (dexamethasone, VEGF, PDGF).   2007 WileyPeriodicals, Inc. J Biomed Mater Res 87A: 136–146, 2008 Key words:  implantable glucose sensor; collagen; porousscaffold; NDGA crosslinking INTRODUCTION To maintain near normal blood glucose levels (70–120 mg/dL), diabetic patients widely use over-the-counter glucose meters, which require finger prick-ing to obtain blood samples several times a day. Thepain, 1 inconvenience, and discomfort of self-monitor-ing of blood glucose are frequently obstacles to effec-tive patient compliance and optimal management of diabetes. During the past 20 years many kinds of continuous glucose monitoring systems have beenstudied including sensors implanted in the subcuta-neous tissue, 2–8 sensors implanted in the vascular bed, 9,10 and determining glucose concentration in in-terstitial fluid sampled using a microdialysis de-vice. 11–13 Although several studies of implantableglucose sensors have been reported, none of these biosensors are capable of continuously monitoringglucose levels during long-term implantation reli-ably. Progressive loss of sensor function occurs duein part to biofouling and to the consequences of aforeign body response such as inflammation, fibrosis,and loss of vasculature. 14–16 Many researchers have modified the surface of thesensors to reduce membrane biofouling  in vivo . In an Correspondence to:  F. Moussy, Brunel Institute for Bioen-gineering, Brunel University, Uxbridge, UB8 3PH, UK;e-mail: grant sponsor: National Institutes of Health(NIH/NIBIB), Bethesda, MD; contract grant number:R01EB001640  2007 Wiley Periodicals, Inc.  approach to reduce protein adsorption, Quinnet al. 17 used poly(ethylene glycol) (PEG) in a polyhy-droxyethylmethacrylate (PHEMA) matrix. Becausethe PEG chains tend to stand up perpendicular tothe membrane surface, they provide a water-richphase that resists binding of many protein mole-cules. Vadgama and coworkers 18,19 reduced proteinadsorption by using diamond-like carbon, so-called‘‘inert’’ materials. Shichiri et al. 20 incorporated an al-ginate/polylysine gel layer at the sensor. Shawet al. 21 reported improvement in biocompatibility of a biosensor coated with PHEMA/polyurethane (PU).Wilkins et al. 22 and Moussy et al. 2,23–25 introducedNafion TM (perfluorosulphonic acid) membrane, toreduce ‘‘biofouling’’ on the surface of the sensor andto reduce interference from urate and ascorbate.Armour et al. 9 coated their sensor tips with cross-linked albumin and Kerner et al. 26 developed cellu-lose-coated sensors to improve sensor blood compat-ibility. However, none of these approaches has beensuccessful for long term, stable glucose monitoring.Collagen and its derived matrices are used exten-sively as natural polymers in the biomedical fieldincluding tissue engineering because of its low anti-genicity, its biodegradability, and its good mechani-cal, haemostatic, and cell-binding properties. 27–31 Todevise strategies for using collagen in the develop-ment of advanced biomaterials for biomedical engi-neering, it is necessary to confer mechanical strengthand resistance to enzymatic (collagenase) degrada-tion resistance with chemical or physical crosslinkingstrategies. There are several strategies for crosslink-ing collagen-based biomaterials. Glutaraldehyde(GA) is the most widely used as a crosslinking agentfor collagen-based biomaterials. 27,32 However, GAand its reaction products are associated with cyto-toxicity  in vivo , because of the presence of crosslink-ing byproducts and the release of GA-linked colla-gen peptides during enzymatic degradation. 33,34 To avoid  in vivo  cytotoxicity and subsequent calci-fication of GA-crosslinked collagen, several alterna-tive compounds have been examined as potentialcollagen crosslinking agents, 35,36 such as polyepoxy,hexamethylene diisocyanate, 1-ethyl-3-(3-dimethyl-amino-propyl)carbodiimide, and ultraviolet or gamma-ray irradiation. Koob et al. 37–40 recently described aprocess for crosslinking of type I collagen fibers withnordihydroguaiaretic acid (NDGA), a plant com-pound with antioxidant properties. They showedthat NDGA significantly improved the mechanicalproperties of synthetic collagen fibers. In addition,they showed that NDGA-crosslinked collagenfibers neither elicit a foreign body response nor didthey stimulate an immune reaction during 6 weeks in vivo .The extent of crosslinking and choice of crosslink-ing agent may also affect the porosity and pore sizeof the scaffold and may greatly influence fibrouscapsule thickness, blood vessel density, and the loca-tion of vessels within the three-dimensional porousscaffold. 41 Large pore scaffolds (greater than 60  l mpore size) allow deep penetration of capillaries andsupporting extracellular matrix (ECM). Sharkawyet al. 42 recently showed that after 4 weeks of subcu-taneous implantation in rat, a well-organized colla-gen matrix typical of a foreign-body response encap-sulated nonporous implants, while the porous poly-vinyl alcohol (PVA) implants produced less fibrousand vascularized tissue capsules.The goal of this study was to develop a new po-rous collagen scaffold around implantable glucosesensors for improving their biocompatibility. We fab-ricated porous collagen scaffolds by using a freeze-drying method, followed by crosslinking usingNDGA or GA. We evaluated the resistance of NDGA- and GA-crosslinked collagen scaffolds todegradation using both  in vitro  and  in vivo  experi-ments. We also applied the scaffolds around a coil-type implantable glucose sensor and measured sen-sor function  in vitro . MATERIALS AND METHODSMaterials Type I collagen (purified from fetal bovine tendon) wasa generous gift from Shriners Hospital for Children(Tampa, FL). NDGA was purchased from Cayman Chemi-cal (Ann Arbor, MI). Glucose, bovine serum albumin(BSA), and 50% (w/w) GA were obtained from Fisher Sci-entific (Pittsburgh, PA). Glucose oxidase (GOD) (EC1.1.3.4., Type X-S,  Aspergillus niger , 157,500 U/g), epoxyadhesive (ATACS 5104), PU, tetrahydrofuran (THF), andcollagenase (EC, Type I, from  Clostridium histolyti-cum , 302 U/mg) were obtained from Sigma-Aldrich (St.Louis, MO). Sprague–Dawley out-bred rats (male, 375–399 g)were purchased from Harlan (Dublin, VA). Preparation and crosslinking of collagen scaffold The collagen scaffolds were prepared by a freeze-dryingmethod. Collagen was dissolved in 3% acetic acid to pre-pare a 1% (w/v) solution. The solution was applied to acylinder-shaped polypropylene mold ( F  10 mm, height8 mm) and then freeze-dried. A cylindrical 3D porous scaf-fold was obtained. The scaffolds were then crosslinkedwith NDGA or GA to minimize solubility and improve re-sistance to collagenase degradation.For NDGA crosslinking, dried collagen scaffolds were briefly soaked in absolute ethanol, followed by soaking in2  M  NaCl solution for 12 h at room temperature. Scaffoldswere resuspended in oxygen-sparged phosphate-bufferedsaline (PBS, 0.1  M  NaH 2 PO 4 , pH 9.0) for 30 min at roomtemperature. Scaffolds were then treated with 3 mg of POROUS COLLAGEN SCAFFOLD-COATED IMPLANTABLE BIOSENSOR 137  Journal of Biomedical Materials Research Part A  NDGA in 1 mL of PBS as follows: NDGA was dissolved in0.4 N   NaOH at a concentration of 30 mg/mL. One milliliterof the NDGA solution was added directly to PBS in whichthe scaffolds were suspended to a final concentration of 3 mg/mL. The scaffolds were agitated in the NDGA solutionfor 24 h at room temperature. The scaffolds were removed, brieflyrinsedwithwater,andfreeze-dried.For a comparative study of the effectiveness of theNDGA treatment, other scaffolds were treated with 0.5%GA for 2 h or 12 h in ethanol solution at room tempera-ture. To prevent the dissolution or loss of the matrix dur-ing the GA crosslinking process, we used 100% ethanolinstead of water. The crosslinked scaffolds were washedwith deionized water and freeze-dried again. The mor-phology of the scaffolds before/after crosslinking wasexamined using scanning electron microscopy (SEM) aftergold sputter coating of the samples in a metal evaporatoraccording to standard procedures.To evaluate the stability of the scaffold after crosslink-ing, the degree of crosslinking ( D c ) was estimated byweighing the dried samples before and after crosslinking. D c  was calculated using the following equation: D c  ð % Þ¼  sample mass after crosslinking =  sample mass before crosslinking   3 100The swelling property of the porous scaffolds wasexamined by measuring water absorption. The scaffoldswere weighed after thorough drying ( W  dry ) and immersedin purified water. After 24 h, the scaffolds were removedfrom the water and immediately weighed again ( W  wet ).Water absorption was calculated by using the followingequation:Water absorption  ð % Þ¼ ð W  wet  W  dry Þ = W  wet    3  100 In vitro  and  in vivo  evaluation ofthe collagen scaffolds To examine the biological stability of the crosslinkedscaffolds, we performed  in vitro  and  in vivo  biodegradationtests.  In vitro  biodegradation of NDGA- and GA-cross-linked scaffolds was tested using bacterial collagenase.Fabricated NDGA- and GA-crosslinked collagen scaffoldswere incubated in the collagenase solution (1 mg/mL inPBS at 37 8 C) for up to 4 weeks. Scaffolds were removedfrom the solution, rinsed with deionized water, and freeze-dried at given time intervals (weeks 1 to 4) during incuba-tion. The  in vitro  degradation was evaluated as the per-centage of weight difference of the dried scaffold beforeand after enzyme digestion.To determine the stability of the crosslinked scaffolds  invivo , we directly implanted NDGA- and GA-crosslinked col-lagen scaffolds in rats. The scaffolds were disinfected with70% ethanol solution for 2 h and implanted subcutaneouslyin the back of the rats. Scaffolds were explanted at 7, 14, 21,and 28 days after implantation. After explantation, thescaffolds were examined macroscopically. Preparation of porous collagen scaffoldsaround implantable glucose sensors We first fabricated coil - type glucose sensors loaded withcrosslinked enzyme (GOD) using a Platinum–Iridium (Pt/Ir) wire (Teflon coated,  F  0.125 mm, Pt:Ir  5  9:1, Medwire,Sigmund Cohn Corp.). Then, we applied bovine tendontype I collagen scaffolds around the sensors (Fig. 1).Briefly, to fabricate a glucose sensor, the Teflon coating of the top 10 mm of a Pt/Ir wire was removed and the wirewas wound up along a 30-gauge needle to form a coil-likecylinder. The cylinder unit had an outer diameter of 0.55mm and an inner diameter of 0.3 mm and a length of 1 mm. A cotton thread was inserted inside the coil chamberto retain the enzyme solution during enzyme coating of the electrodes. GOD was added and crosslinked to the sen-sors by dip coating in an aqueous solution containing 1%GOD, 4% BSA, and 0.6% (w/w) GA. The outer membraneof the sensor was coated with Epoxy-PU by dipping in Ep-oxy-PU solution [2.5% (w/v) in THF, Epoxy:PU  5  1:1].The sensor was dried at room temperature for at least24 h. The two ends of the sensing element were sealed byelectrically insulating sealant (Brush-On electrical tape,North American Oil Company). 43,44 To apply collagen scaffolds around the sensors, thesensors were dip-coated with 1% (w/v) collagen solutionand freeze-dried. The porous scaffolds around the glu-cose sensors were crosslinked with either NDGA or GAas previously described. Obtained sensors were storeddry at room temperature or in PBS at 4 8 C. The morphol-ogy of the sensors was observed using light microscopeand SEM.In addition, to evaluate sensitivity changes of the sen-sors with varying wall thickness of the scaffold around thesensors, we controlled the wall thickness of the scaffold bymultiple dipping/freezing cycles in collagen solution. Thescaffold with the sensor was then freeze-dried and cross-linked as previously described.Silver wires (Teflon coated,  F  0.125 mm, World Preci-sion Instruments) were used to fabricate the Ag/AgCl ref-erence electrodes. Silver wires were coiled and anodizedgalvanostatically at 1 mA overnight in stirred 0.1  M HCl. 43,44 Figure 1.  Schematic diagram of the scaffold-coated sens-ing element of the glucose electrode. 1 – Teflon-coveredPt-Ir wire; 2 – Ag/AgCl reference wire; 3 – collagen scaf-fold; 4 – electrically insulating sealant; 5 – Epoxy-PU outermembrane; 6 – enzyme layer; 7 – stripped and coiled Pt-Irwire; 8 – cotton fiber with GOD gel.138 JU ET AL.  Journal of Biomedical Materials Research Part A  In vitro  characterization of sensorscoated with scaffolds The glucose sensors were characterized in PBS (pH 7.4) at700 mV versus the incorporated Ag/AgCl reference electro-des. The working electrode (Pt/Ir wire) and Ag/AgCl refer-ence electrode of each sensor were connected to an Apollo4000 potentiostat (World Precision Instruments). The back-ground current was allowed to stabilize for 10 min, and thesensors were then exposed to a series of glucose solutions toexamine their sensitivities and linearities. The response sen-sitivity ( S ) was repeatedly assessed by (1) measuring theresponse current ( I  1 ) of a  C 1  glucose solution, (2) adding aconcentrated glucose solution into the measured solution toincrease the glucose concentration to  C 2 , and (3) measuringthe response current ( I  2 ) of the resulting solution. The sensi-tivity was expressed as the current increase caused by a1 m  M  glucose increase, that is,  S 5 ( I  2 2 I  1 )/( C 2 2 C 1 ). Figure 2.  Schematic mechanism for GA (A) and NDGA (B) crosslinking of the collagen scaffold.POROUS COLLAGEN SCAFFOLD-COATED IMPLANTABLE BIOSENSOR 139  Journal of Biomedical Materials Research Part A  RESULTS AND DISCUSSIONPreparation of porous crosslinkedcollagen scaffolds The chemistry of the NDGA crosslinking reactiondiffers from the reaction using the GA treatment(Fig. 2). GA is the most common crosslinking agentused for fixation of collagen scaffolds for tissue bio-engineering. Both aldehyde functional groups of theGA molecule react with amine groups between twoneighboring polypeptide chains, particularly lysineside chains. Unfortunately, GA crosslinking isencumbered with potential cytotoxicity problemscaused by the presence of unreacted residual groupsor the release of monomers and small polymers dur-ing enzymatic degradation. 33,34 NDGA treatment is an alternative crosslinkingagent, which possesses reactive catechols. Collagencrosslinking with NDGA mimics the quinine tanningmechanism in the skate egg capsule. Catechol–qui-none tanning systems are prevalent in a wide varietyof animals, which serve to strengthen vulnerableECMs (e.g. insect cuticle, mussel byssus threads). 37,45 NDGA, isolated from the creosote bush, is a low mo-lecular weight di-catechol containing two  ortho -cate-chols. The two catechols on NDGA undergo autooxi-dation at neutral or alkaline pH producing reactivequinones. Two quinones then couple via aryloxy freeradical formation and oxidative coupling, forming bisquinone crosslinks at each end. The NDGA con-tinues forming a large crosslinked bisquinone poly-mer network in which the collagen fibrils are embed-ded. The NDGA treatment does not form crosslinkswith amino acid side chains of collagen. 37,38,45 In this study, highly porous collagen scaffoldswere prepared by a freeze-drying method. We ascer-tained that the obtained scaffolds have an open celland interconnected pore structure based on SEM ob- Figure 3.  SEM morphology of the collagen scaffolds. Determination of the pore size of collagen scaffolds by SEM. (A) Nocrosslinking; (B) GA crosslinked; (C) NDGA crosslinked. Figure 4.  Bulk properties of GA- and NDGA-crosslinkedscaffolds. Results are shown as means 6  SD ( n  5 3).140 JU ET AL.  Journal of Biomedical Materials Research Part A
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks