A neuroinductive biomaterial based on dopamine

Abstract

Chemical messengers such as neurotransmitters play an important role in cell communication, differentiation, and survival. We have designed and synthesized a bioactive biomaterial that derived its biological activity from dopamine. The resultant biodegradable polymer, PCD, has pendent groups bearing dopamine functionalities. Image analysis demonstrated that nerve growth factor-primed rat pheochromocytoma cells (PC12) and explanted rat dorsal root ganglions attached well and displayed substantial neurite outgrowth on the polymer surface. Furthermore, PCD promoted more vigorous neurite outgrowth in PC12 cells than tissue culture polystyrene, laminin, and poly(d-lysine). The histogram of neurite length of PC12 cells showed distinctive patterns on PCD that were absent on the controls. A subset of PC12 cells displayed high filopodium density on PCD. The addition of dopamine in culture medium had little effect on the differentiation of PC12 cells on tissue culture polystyrene. Tyrosine, the precursor of dopamine, did not exhibit this ability to impart specific bioactivity to an analogous polymer. Thus, the dopamine functional group is likely the origin of the inductive effect. PCD did not cause nerve degeneration or fibrous encapsulation when implanted immediately adjacent to the rat sciatic nerves. This work is a step toward creating a diverse family of bioactive materials using small chemical messengers as monomers.

Biomaterials are widely used in disease treatment and improving human well-being (1, 2). Recently, significant advances have been made to impart biological activity to biomaterials. Most of the existing bioactive materials are derived from extracellular matrix or are modified with extracellular matrix motifs (3–8). The most widely used extracellular matrix motifs include protein epitopes such as Arg-Gly-Asp (4, 5, 9), Tyr-Ile-Gly-Ser-Arg (10, 11), and Ile-Lys-Val-Ala-Val (6, 11) and glycosaminoglycans such as heparin (12, 13). In addition to cell–extracellular matrix interactions, cell differentiation and survival also depend on constant interactions with other cells through a plethora of messenger molecules (14–16). The biomaterial reported here is designed to use a chemical messenger to impart bioactivities to the resultant biodegradable polymers.

We chose to focus on using bioactive materials for functional restoration of damaged nerves because it is a major challenge in medicine (17). An important class of chemical messengers in the nervous system is the neurotransmitters (18, 19). They are essential to neuronal outgrowth during embryonic and neonatal development and after injury (20–23). The depletion of neurotransmitters during embryonic development results in developmental defects of the brain, suggesting that neurotransmitters play crucial roles as morphogens or neurotrophic factors (24). Dopamine in particular is vital in axon growth and synapse formation during the embryonic stage (25). We postulated that a biocompatible material containing dopamine functional groups would be neuroactive. Our specific design parameters address the following criteria: (i) degradability (we chose degradable material because prolonged presence of an implant may compress or hinder regenerating nerves); (ii) degradable functional groups (we chose ester for its versatile synthesis); (iii) dopamine functional groups to be polymerized (we chose the primary amine because the catechol is preserved among the catecholamine neurotransmitters whereas the amine group can be alkylated as in epinephrine); (iv) polymerization mechanism (we chose addition reaction between dopamine and an epoxide because it alkylates the amines under mild conditions).

Results

Polymer Synthesis and Degradation.

PCD was synthesized readily as a pale yellow powder that was soluble in N,N-dimethylforamide and low-molecular-weight alcohols and ketones, but not water (Fig. 1). NMR spectra revealed a change of chemical shift from ≈3.2 ppm in the diglycidyl ester to ≈4.0 ppm in PCD, consistent with the opening of the epoxide ring. The intense CO stretch at 1,730 cm⁻¹ in the FTIR spectra confirmed the formation of the ester bonds, whereas the bands at 866 and 814 cm⁻¹ revealed the presence of the catechol units in the polymer (Fig. 9, which is published as supporting information on the PNAS web site). The absence of absorption at 1,674, 1,662, and 1,564 cm⁻¹ indicated that oxidation of catechol to quinone was undetectable in PCD (26). The glass transition temperature was 37.2°C according to differential scanning calorimetry measurements. To verify whether the catecholamine functional groups were vital for bioactivity, we synthesized a control material (PCY) based on tyrosine, the precursor of dopamine. For PCY, the presence of phenol was confirmed by a medium-intensity absorption at 834 cm⁻¹ (Fig. 10, which is published as supporting information on the PNAS web site).

Fig. 1.

PCD was synthesized by using dopamine as a monomer. This synthesis strategy can be applied to a large number of diglycidyl esters and biomolecules containing primary amines.

Polymerization of dopamine converted its primary amine to a tertiary amine, which limited the formation of dopaminechrome, the oxidative intermediate to dopamine quinone (26). This increased the oxidation resistance of the catecholamine, thus minimizing the toxicity associated with dopamine quinone (27, 28). The ester bond in PCD rendered the polymer biodegradable, with a half-life in PBS solution (29, 30) of ≈50 days at 37°C (Fig. 2). The presence of low-molecular-weight (209–436) degradation products in the supernatant after 2 weeks was detected by MS. A major peak with an m/z ratio of 300 is consistent with the presence of N,N-bis(2,3-dihydroxypropyl)dopamine, the expected hydrolytic degradation product of PCD.

Fig. 2.

The degradation of PCD in vitro. Shown is degradation of the polymer expressed as a decrease of mass with time. Data are expressed as mean ± SD.

Effects of PCD on Cell Proliferation and Differentiation.

We evaluated the material's capability at promoting neurite outgrowth in vitro using explanted rat dorsal root ganglions (DRGs) and rat pheochromocytoma (PC12) cells that have been widely used in studying neuronal communication and interaction between biomaterials and neurons (31, 32). The PC12 cells were primed with nerve growth factor (NGF) for 24 h instead of 9 days (33) to better differentiate the materials' intrinsic ability to induce cell differentiation. The differentiated cells were cultured in 24-well plates coated homogeneously with PCD. Cell proliferation, differentiation, and morphology were compared with tissue culture polystyrene (TCPS), PCY, poly(d-lysine) (PDL), and laminin. The latter two are standard substrates for culturing many types of neurons. Culture medium containing dopamine at a concentration below its toxicity limit (34) was used to study the effect of dopamine on PC12 cells cultured on TCPS surfaces.

NGF-primed PC12 cells adhered on PCD and control surfaces within 4 h of seeding. The in vitro biocompatibility of PCD was evaluated by monitoring the number of adhered metabolically active cells using a 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay. PCD was noncytotoxic and appeared to be at least as biocompatible as standard neuron-culturing materials in vitro. Cell proliferation on PCD resembled that of PDL and was significantly better than that of TCPS with dopamine added to the culture medium (Fig. 3A). The introduction of free dopamine appeared to inhibit the proliferation of PC12 cells, which agrees with previous reports (34, 35). The experiments were terminated when cell aggregates appeared after 7 days, especially on TCPS and laminin surfaces.

Fig. 3.

NGF-primed PC12 cells grew longer neurites and exhibited higher differentiation rates on PCD versus PCY, PDL, laminin, TCPS, and TCPS with free dopamine (DA). (A) The number of metabolically active cells increased on all four materials as indicated by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assays (absorption measured at 570 nm). (B) Cells displayed the highest differentiation level on PCD, and it increased significantly with time. The control materials did not induce more differentiation as the cell culture time increased. (C) The median neurite length increased substantially on PCD over the culture period. In comparison, the change on the control surfaces was statistically insignificant over the entire culture period. ∗, P < 0.05 with respect to the changes of values on PCD surface with time. #, large standard deviation caused by the presence of a few cells with neurites up to 37 μm amid >89% of the cells with little neurites (Fig. 11). &, Data not representative of the whole cell population because of large numbers of cell aggregates on TCPS surfaces at day 7. The addition of dopamine appeared to further reduce the number of adherent cells.

PC12 cells with neurites longer than 20 μm are generally considered differentiated nerve cells (36). The differentiation level reached as high as 84 ± 12% on PCD at day 5 (Fig. 3B). In comparison, the highest level of differentiation was 30 ± 10%, 33 ± 13%, 16 ± 7.7%, 13 ± 3.7%, and 13 ± 3.8% on PCY, PDL, laminin, TCPS, and TCPS with 10 μM dopamine, respectively. PCD induced significantly more cell differentiation as culture time increased (ANOVA, P < 0.05). The differentiation level on TCPS surfaces also increased with time albeit at a much lower level. In contrast, the percentage of differentiated cells on PCY, PDL, and laminin did not change significantly throughout the entire culture period. The addition of free dopamine in culture medium had no significant effect on the differentiation of PC12 cells.

The differentiated cells on all materials exhibited neuronal morphology. However, the median neurite length on PCD was significantly higher than that on the controls. Moreover, the difference between PCD and the control increased with time (Fig. 3C). At day 7, the neurites were ≈150% longer than those on PCY, which yielded the longest neurites in the control group. The increase of median neurite length with time was significant on PCD, whereas the difference of neurite length on PCY, PDL, and laminin was statistically insignificant over the culture period. The cells on TCPS surfaces were poorly differentiated and displayed smaller median neurite length than that of the other four materials.

The histogram of the neurite length showed distinct patterns on PCD (Fig. 4A) that were absent in the controls (Fig. 4 B–D). The neurite length was divided into 10-μm increments up to 100 μm and larger increments as appropriate above 100 μm. The percentage of neurites with neurite length in a defined range was plotted (e.g., the data point at 50 μm represents the percentage of neurites 40–50 μm long). Distinct distribution patterns were visible on the PCD curves after 3 days of culture (Fig. 4A). In contrast, the neurites on the controls showed no clustering or grouping around any particular length; their distribution curves had no distinctive peaks. As culture time increased, the fraction of longer neurites increased considerably on PCD, shifting the curve further to the right. The percentage of neurites >20 μm long increased sharply at day 3 and reached as high as 77% on day 7. In comparison, most of the neurites on the controls were significantly shorter, <30% of which were longer than 20 μm throughout the culture period. This finding suggests that PC12 cells were more differentiated on PCD than on the control materials. The histogram of neurite length on TCPS resembles those of the other controls, and the addition of dopamine did not significantly alter the pattern (Figs. 11 and 12, which are published as supporting information on the PNAS web site).

Fig. 4.

The histogram of neurite length showed very different patterns on PCD (A) versus PCY (B), PDL (C), and laminin (D). The neurite length appeared to group into clusters on PCD, suggesting that the PC12 cell could be more differentiated. The weight of longer neurites increased sharply after day 3 on PCD. Cells in certain areas reached confluence on day 7, which likely contributed to the apparent decrease of longer neurites on PCD, PDL, and laminin surfaces.

Cell Morphology on PCD.

The number of PC12 cells exhibiting typical neuronal morphology was significantly higher on PCD (Fig. 5A) than on PCY, PDL, and laminin surfaces (Fig. 5 B–D). Compared with the above four surfaces, fewer cells displayed neuronal morphology on TCPS, and the addition of free dopamine had little effect (Figs. 13 and 14, which are published as supporting information on the PNAS web site). Neurites up to 180 μm long began to appear on PCD 3 days after seeding and grew to nearly 250 μm after 5 days of culture. In comparison, the longest neurites reached ≈150 μm on PCY, PDL, and laminin. The nerve growth cone is a highly motile structure at the tip of an extending axon that explores the extracellular environment and guides the extension of the axon. The primary morphological characteristic of a growth cone in vitro is a sheet-like lamellipodium with numerous fine processes called filopodia. The cells on PCD displayed structures resembling growth cones at the end of their neurites (Fig. 6A). Growth cone-like structures were also observed on PCY surfaces (Fig. 6B). A number of cells on PCD exhibited extensive filopodia (4.40 ± 1.66 filopodia per 10-μm neurite) along their neurites in a similar fashion to the spines on spiny neurons. In contrast, the filopodium density was lower for analogous neurites on PCY (1.38 ± 0.707 filopodia per 10-μm neurite). Low dopamine levels are known to cause a reduction of dendritic spine density in dopaminoceptive neurons (37). The higher filopodium density on PCD indicated that the catecholamine functional groups could induce unique responses in differentiated PC12 cells.

Fig. 5.

A significant number of PC12 cells differentiated and exhibited typical neuronal morphology on PCD. The cells are more differentiated and displayed longer neurites on PCD (A) than on PCY (B), PDL (C), and laminin (D). Data are from day 7. (Magnification: ×200. Scale bar: 100 μm.)

Fig. 6.

Scanning electron micrographs of PC12 cells grown on PCD (A) and PCY (B) surfaces with cell bodies partially visible on the top of the image. The cells exhibited growth cone-like structures at the tip of their neurites on both materials. However, the filopodia density along the neurites was higher on PCD surfaces, indicating potential inductive effects of the dopamine moieties. (Magnification: ×5,000. Scale bar: 5 μm.)

Biocompatibility with Nervous Tissue ex Vivo and in Vivo.

Preliminary study on rat DRG explants revealed that they adhered well and displayed significant neurite outgrowth on PCD (Fig. 7A). A large number of long, loosely bundled neurites sprouted out from the DRGs on both PCD and PDL surfaces (Fig. 7B). The neurites on PCD exhibited an “arborizing” mode of growth with extensive branching, whereas the neurites on PDL appeared more linear. An initial in vivo biocompatibility study indicated that PCD pellets did not cause nerve degeneration or fibrous encapsulation when implanted immediately adjacent to rat sciatic nerves. The brown remnants of the 20-mg implants were well circumscribed with a cavitated center containing inflammatory infiltrates including phagocytic cells up to 8 weeks after implantation (Fig. 8A). The cavitations of the implants were likely a result of both hydrolytic degradation and biodegradation caused by the inflammatory cells. Collagen deposition around the polymers was minimal (Fig. 8B), especially when compared with normal collagens surrounding the nerves (Fig. 8 C and D). The absence of fibrosis would yield a more permissive environment for regenerating axons, and multimonth degradation time of the material would be beneficial for nerve regeneration, which usually requires several months (38).

Fig. 7.

Extensive neurite outgrowth from rat DRG explants (postnatal day 3) cultured on PCD (A) and PDL (B) surfaces. Data are from day 9. (Magnification: ×200. Scale bar: 100 μm.)

Fig. 8.

Biocompatibility of PCD in vivo. (A) Hematoxylin and eosin staining showed brown residues of a degraded PCD implant (arrows) at 8 weeks after implantation. Cavities were formed inside the material with phagocytic inflammatory infiltrates. The implants caused little fibrosis, and no nerve degeneration was observed. (B) Mason's trichrome staining confirmed that collagen (blue fibers) deposition surrounding the polymer residues was minimal. (C and D) Hematoxylin and eosin staining (C) and Mason's trichrome staining (D) of the sham surgery site at 8 weeks after implantation. N, nerve. (Magnification: ×100. Scale bar: 200 μm.)

Discussion

Compared with standard neuron culture material, PDL and laminin, PCD can induce more PC12 cell differentiation, promote longer neurite outgrowth, and a present specific distribution pattern of neurite length. Primary rat DRGs extend long and arborized (branching repeatedly like a tree) neurites on PCD surfaces. The polymer is nontoxic in vitro and causes no nerve degeneration or fibrous encapsulation in vivo when implanted next to the sciatic nerve. The presence of catecholamine functional groups in PCD induces unique cellular responses, which are completely lost upon the substitution of one catechol hydroxyl group with a hydrogen atom. Therefore, the bioactivity of PCD can be attributed to the dopamine-like functional groups.

One question arising from this study is whether the cells can recognize the pendant functional groups in the polymer or whether the release of a dopamine-like degradation product is a prerequisite for the bioactivity. The ultimate degradation products are most likely N,N-bis(2,3-dihydroxypropyl)dopamine and 1,2-cyclohexane dicarboxylic acid because the weakest bond in the polymer is the ester bond. The former is a covalently modified dopamine, which might be recognized by dopamine receptors or transporters (39, 40). Structural analogs of dopamine can act as dopamine receptor agonists or antagonists and are known to interact with dopamine transporters. Treating PC12 cells with amphetamine, a structural analog of dopamine, is known to induce neurite outgrowth through interactions with the dopamine transporters (41, 42). A dopamine D₁ receptor antagonist, SCH-23390, has been shown to promote neurite outgrowth in PC12 cells (43). Cells displayed a distinct response to PCD in the initial hours after seeding when there was unlikely to be any appreciable amount of degradation. Thus cells probably recognize the catecholamine functional groups on the polymer through direct contact. This is consistent with reports of bioactivities derived from biomolecules tethered on polymers (4, 11, 44). Further, PCD most likely acted as a dopamine receptor antagonist because the addition of a nontoxic level of free dopamine had no positive effects on PC12 cells.

Biomaterials with catechol units are known to affect cell adhesion depending on the structure and composition of the polymer (45, 46). To the best of our knowledge, the interaction of catechol-containing polymers with neurons has not been reported. It is possible that catechol alone can induce the observed cellular response; however, dopamine functional groups are likely necessary because cell adhesion to the control materials was very efficient. Furthermore, acetylcholine-based polymers without any catechol functional groups induce similar responses in PC12 cells and rat DRGs (data not shown).

This study demonstrates that synthetic polymers with integrated dopamine functional groups can effectively induce neuronal differentiation. The modular design of the polymers is highly versatile. The diglycidyl ester can be synthesized from a large number of dicarboxylic acids allowing control of degradation, hydrophilicity, and structural rigidity of the polymer backbone. The concentration of the neurotransmitter functional groups in a polymer can be adjusted by copolymerization with an “inert” amine allowing further control of the structure and function of the resultant biomaterial. This simple design platform can lead to a large family of structurally and functionally diverse biomaterials useful in medicine, especially various areas of regenerative medicine.

Materials and Methods

Materials Synthesis and Characterization.

Diglycidyl 1,2-cyclohexanedicarboxylate was purchased from TCI (Portland, OR) and distilled under vacuum before use. All other chemicals were purchased from Alfa Aesar (Ward Hill, MA) and used without purification. Equimolar amounts of diglycidyl 1,2-cyclohexanedicarboxylate and dopamine were dissolved in N,N-dimethylformamide with 0.1% Mg(ClO₄)₂ under N₂ with constant stirring. The reaction mixture was heated at 90°C for 7 days, and the resultant viscous liquid was precipitated in diethyl ether. The precipitate was subjected to a quick wash with deionized water to remove the catalyst and byproducts. The purified PCD (71% yield) was lyophilized (Freezezone 4.5 freeze dryer; Labconco, Kansas City, MO) and stored under N₂ at −40°C. NMR spectra were recorded on a 400-MHz Mercury-400 NMR, and FTIR spectra were recorded on an IR-100 spectrometer from Thermo Nicolet (Waltham, MA). Differential scanning calorimetry was performed on a DSC Q100 from TA Instruments (New Castle, DE). PCD, ¹H NMR (methanol-d₄) δ 6.62–6.55 (br, 3H), 4.18–3.80 (br, 3H), 3.60–3.31 (br, 3H), 3.16–2.98 (br, 2H), 2.90–2.40 (br, 6H), 2.02–1.80 (br, 2H), 1.78–1.61 (br, 2H), 1.49–1.17 (br, 6H). FTIR: 1,727 cm⁻¹ (ester CO), 1,451 cm⁻¹ (CN), 1,289 cm⁻¹ (catechol CO), 814 and 866 cm⁻¹ (1,2,4-trisubstituted aromatic ring). The molecular weight of PCD was found to be 5,260 as determined by MALDI-MS on an ABI 4700 MALDI TOF Mass Spectrometer (Applied Biosystems, Foster City, CA). The molecular weight of the degradation products was measured by electrospray MS. When tyrosine ethyl ester was used in lieu of dopamine, PCY was obtained at 92% yield. PCY, ¹H NMR (methanol-d₄) δ 7.09–6.73 (br, 2H), 6.68–6.41 (br, 2H), 4.12–3.61 (br, 7H), 3.58–3.30 (br, 3H), 3.16–2.62 (br, 6H), 2.59–2.37 (br, 2H), 2.08–1.77 (br, 3H), 1.75–1.60 (br, 2H), 1.45–1.14 (br, 5H). FTIR: 1,732 cm⁻¹ (ester CO), 1,220 cm⁻¹ (phenol CO), 1,377 cm⁻¹ (phenol OH in-plane bending), 834 cm⁻¹ (para-substituted aromatic ring).

In Vitro Degradation and Coating of PCD.

The degradation study was conducted by using PCD pellets. PCD powders (10 mg) were compressed into 5-mm-diameter, 0.5-mm-thick pellets on a Carver press at 2,200 lb for 15 min under N₂ and submerged into 15 ml of sterile PBS solution in a centrifuge tube at 37°C. The disks were retrieved at a predetermined time point, dried in a vacuum oven at 50°C and 100 mTorr overnight, and weighed. To study cell interactions, TCPS surfaces were coated with a uniform layer of PCD ≈0.5 μm thick. Ethanol solution of PCD (0.7 mg/ml) was filtered through a 0.2-μm filter and added to a 24-well plate (100 μl per well). The solution was allowed to evaporate slowly in a N₂ atmosphere with the plate cover on to achieve a uniform coating without crystallization of the polymer. The resultant coating adhered firmly to the plate and was further dried under vacuum (100 mTorr) for 3 days and washed with PBS (6 × 1 ml per well) to remove any residual solvents. The plates were soaked in serum-free RPMI medium 1640 overnight. PCY-coated plates subjected to the same washing and soaking treatment and PDL- and laminin-coated plates (BD Biosciences, Bedford, MA) were used as controls.

Cytotoxicity and Cell Morphology.

All cell culture reagents were purchased from Mediatech unless noted otherwise. PC12 cells were cultured in 85% RPMI medium 1640, 10% heat-inactivated horse serum, and 5% FBS and maintained in a humid, 5% CO₂ incubator at 37°C. The cells were primed for 24 h in differentiation medium containing NGF (Promega, Madison, WI) (RPMI medium 1640, 1% heat-inactivated horse serum, 0.5% FBS, 50 ng/ml NGF, and 20 μg/ml gentamycin) before seeding at a density of 1 × 10⁴ per cm². The cells were maintained in an incubator with medium exchanged every 2 days. The medium for the dopamine control experiment contained 10 μM dopamine.

Cell morphology was monitored by using an inverted phase contrast microscope (TE-2000U microscope equipped with a 4 MP Diagnostics Spot Flex digital camera; Nikon, Melville, NY). Neurite length, defined as the distance from the tip of the neurite to the junction between the cell body and neurite base, was measured on days 1, 3, 5, and 7. In the case of branched neurites, the length of the longest branch was used. For each well, at least six images at ×200 were randomly acquired to ensure that >500 cells were captured. Differentiation level was quantified by the percentage of differentiated cells with neurites longer than 20 μm. The median neurite length was calculated by using Excel (Microsoft, Redmond, WA). At least 500 cells were measured in each group at any given time point. Image analysis was performed by using NIH ImageJ. The filopodia density was measured by counting the total number of filopodia divided by total length of neurites with filopodia along its shaft. Neurites with filopodium-like protrusions only at the very tip were excluded from the image analysis. The results are expressed as number of filopodia per 10 μm of neurite. The number of metabolically active cells was analyzed by using 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (Invitrogen, Carlsbad, CA) according to a modified protocol as previously described (47). The absorption of 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide was measured by using a microtiter-plate reader (TECAN SAFIRE, Durham, NC).

Isolation and Culture of Neonatal Rat DRG Explants.

DRGs from spinal levels L4–L6 of postnatal day-3 Sprague–Dawley rats were dissected and collected in Hanks' balanced salt solution. The ganglia were washed with Hanks' balanced salt solution twice and seeded in neurobasal medium supplemented with B-27 (2%; Invitrogen), 2 mM l-glutamine, and 50 ng/ml NGF in a precoated 24-well plate at one ganglion per well. Glial proliferation was inhibited with 5-fluorodeoxyuridine (7.5 μg/ml) and uridine (17.5 μg/ml) (MP Biomedicals, Solon, OH). The medium was first changed at day 4 and was changed every 3 days afterward. The culture was maintained in a humid 5% CO₂ incubator at 37°C.

In Vivo Biocompatibility.

PCD pellets (20 mg, 5-mm diameter, 1 mm thick) were sterilized by plasma for 40 min. The implants were soaked in sterile PBS containing 0.1 wt/vol% ascorbic acid on an orbital shaker for 2 h before implantation. Male Sprague–Dawley rats weighing 300–350 g were implanted with PCD under anesthesia maintained by halothane inhalation. Under sterile conditions, a gluteal muscle splitting technique was used to expose both sciatic nerves. Eight animals received two sterile implants with one implant placed directly underneath each sciatic nerve on the underlying muscle bed. Control animals (n = 4) underwent either sham surgery, with exposure of both sciatic nerves followed by closure (n = 3), or no surgery (n = 1). The muscle layer was closed with vicryl sutures, and the skin was closed with staples. Animals were cared for in compliance with protocols approved by the Committee on Animal Care of the Georgia Institute of Technology following National Institutes of Health guidelines for the care and use of laboratory animals (National Institutes of Health publication no. 85-23, rev. 1985).

Explants were harvested at 7, 14, 28, and 56 days after implantation. The surgical wounds were reopened; the gluteal musculature was removed en bloc with the sciatic nerve and remnants of the polymer and fixed in 10% formalin for 3 days before histological analysis. The control tissues were similarly harvested and fixed at the same time points. The tissues were dehydrated in a graded series of ethanol and xylenes and embedded in paraffin. Ten-micrometer-thick sections were cut along the longitudinal axis of each implant. Sections were stained by using a standard protocol for hematoxylin and eosin and Mason's trichrome stain. The histological samples were analyzed blindly and independently by a histopathologist (B.S.). Sections were analyzed for the degree of inflammation and fibrosis. The inflammatory response to each implant was assessed by rating the levels of lymphocytic and histiocytic infiltrate, and fibrosis was identified by collagen deposition.

Statistical Analysis.

At each time point for each test, three replicates were tested for the experimental and control samples. Multicomparison ANOVA was used to statistically compare experimental values at a given time point; P < 0.05 was considered statistically significant. The results are reported as mean values with standard deviations.

Abbreviations

PDL

poly(d-lysine)

TCPS

tissue culture polystyrene

NGF

nerve growth factor

DRG

dorsal root ganglion