- Email: [email protected]
Controlled release of nerve growth factor enhances sciatic nerve regeneration
Available online at www.sciencedirect.com R
Experimental Neurology 184 (2003) 295–303
www.elsevier.com/locate/yexnr
Controlled release of nerve growth factor enhances sciatic nerve regeneration Annie C. Lee,a Vivian M. Yu,a James B. Lowe, III,a Michael J. Brenner,c Daniel A. Hunter,a Susan E. Mackinnon,a and Shelly E. Sakiyama-Elberta,b,* a
Division of Plastic and Reconstructive Surgery, Department of Surgery, Washington University School of Medicine, St. Louis, MO 63110, USA b Department of Biomedical Engineering, Washington University, St. Louis, MO 63130, USA c Department of Otolaryngology, Washington University School of Medicine, St. Louis, MO 63110, USA Received 16 January 2003; revised 21 April 2003; accepted 6 May 2003
Abstract Based on previous studies demonstrating the potential of growth factors to enhance peripheral nerve regeneration, we developed a novel growth factor delivery system to provide sustained delivery of nerve growth factor (NGF). This delivery system uses heparin to immobilize NGF and slow its diffusion from a fibrin matrix. This system has been previously shown to enhance neurite outgrowth in vitro, and in this study, we evaluated the ability of this delivery system to enhance nerve regeneration through conduits. We tested the effect of controlled NGF delivery on peripheral nerve regeneration in a 13-mm rat sciatic nerve defect. The heparin-containing delivery system was studied in combination with three doses of NGF (5, 20, or 50 ng/mL) and the results were compared with positive controls (isografts) and negative controls (fibrin alone, NGF alone, and empty conduits). Nerves were harvested at 6 weeks postoperatively for histomorphometric analysis. Axonal regeneration in the delivery system groups revealed a marked dose-dependent effect. The total number of nerve fibers at both the mid-conduit level and in the distal nerve showed no statistical difference for NGF doses at 20 and 50 ng/mL from the isograft (positive control). The results of this study demonstrate that the incorporation of a novel delivery system providing controlled release of growth factors enhances peripheral nerve regeneration and represents a significant contribution toward enhancing nerve regeneration across short nerve gaps. © 2003 Elsevier Science (USA). All rights reserved. Keywords: Peripheral nerve; Heparin; Fibrin; Nerve guide conduit; Tissue engineering; Drug delivery
Introduction Despite recent advances in the understanding of nerve regeneration and in surgical techniques, the complete functional recovery in a damaged nerve is rare (Seckel, 1990). It is estimated that more than 200,000 nerve repair procedures are performed annually in the United States alone (Archibald et al., 1991). The most common method of nerve repair is to directly suture the two severed nerve ends together; however, this is often not feasible if the gap
* Corresponding author. Department of Biomedical Engineering, Washington University, Campus Box 1097, One Brookings Drive, St. Louis, MO 63130-4899, USA. Fax: ⫹1-314-935-7448. E-mail address: [email protected] (S.E. SakiyamaElbert).
between the two ends is too large to allow a tension-free nerve repair. In these cases, a nerve autograft is indicated to improve functional recovery (Mackinnon, 1989). The autograft provides a scaffold for the regenerating nerve and guides it to its proper target. However, nerve grafting has its limitations, because of the associated donor site mobidity. Both of these procedures, direct repair and nerve autograft, are associated with less than optimum results (Lundborg, 1990). Less than 25% of patients who received direct repair or autograft repair of the median nerve at the wrist level regain full motor function and only 1–3% recovered normal sensation after 5 years (Beazley et al., 1984; Dellon and Mackinnon, 1988). One alternative method to autografting is to use a nerve guidance conduit (NGC) to connect the proximal and distal ends of the severed nerve and guide the regeneration of
0014-4886/03/$ – see front matter © 2003 Elsevier Science (USA). All rights reserved. doi:10.1016/S0014-4886(03)00258-9
296
A.C. Lee et al. / Experimental Neurology 184 (2003) 295–303
Mackinnon and Dellon, 1990). The use of biodegradable conduit is highly desirable, and our ultimate goal is to combine the “tissue-engineered” drug delivery matrices with biodegradable conduits. However, for this study we focused only on the drug delivery matrices to serve as conduit fillers. Neurotrophins have been shown to enhance peripheral nerve regeneration. NGF was reported to protect neurons from injury-induced death in lesioned sciatic nerves (Otto et al., 1987). Addition of NGF to saline-filled silicon tubes resulted in an increase in the number of myelinated axons in the regenerating nerve (Rich et al., 1989). NGF has also been shown to facilitate regeneration of hippocampal neurons across a peripheral nerve bridge (Varron and Conner, 1994). Neurotrophin 3 (NT-3) was found to increase the number of myelinated axons in a regenerating dorsal root through NGCs with NT-3 releasing rods in the wall (Borkenhagen, 1997).
Fig. 1. Dose-response relationship between matrix-bound -NGF concentration and number of nerve fibers regenerating across a 13-mm nerve gap bridged with a silicone conduit. The total number of nerve fibers were measured at midconduit (left) and at the distal end of the conduit (right) by quantitative histomorphometry (n ⫽ 6). Mean values and standard deviation are shown. *denotes P ⬍ 0.05 versus the positive isograft control. Fibrin ⫹ DS group contained peptide and heparin but no NGF. NGF (no DS) group contained free NGF in the fibrin matrix by no delivery system. NGF, nerve growth factor.
axons back to the appropriate target. These tubes allow the microenvironment of regeneration to be controlled by manipulating the contents of the NGC. Numerous studies have been performed to determine some of the fundamental mechanisms of regeneration by varying the conditions within such tubes. A comprehensive overview of the various materials used for NGC construction can be found in the review by Fields et al. (1989). Prior to 1989, silicone elastomer was the material of choice for NGC construction. More recently, biodegradable NGCs have been studied because an additional surgical procedure is not required to remove degradable NGCs. The use of poly(glycolic acid), poly(L-lactic acid), poly(lactic acid-co--caprolactone), poly(R-3-hydroxybutyric acid-co-(R)-3-hydroxyvaleric acid)-diol (PHB) conduits have been studied in short gaps in the rat sciatic nerve, and the degradable NGCs were found to promote comparable regeneration versus nerve isografts (Borkenhagen et al., 1998; den Dunnen et al., 1995; Evans et al., 1999;
Fig. 2. Dose-response relationship between matrix-bound -NGF concentration and percent nerve area in the tissue regenerating across a 13-mm nerve gap bridged with a silicone conduit. Percent nerve tissue was calculated as 100 times the neural area divided by the intrafascicular area at midconduit (left) and at the distal end of the conduit (right) by quantitative histomorphometry (n ⫽ 6). Mean values and standard deviation are shown. *denotes P ⬍ 0.05 versus the positive isograft control. Fibrin ⫹ DS group contained peptide and heparin but no NGF. NGF (no DS) group contained free NGF in the fibrin matrix by no delivery system. NGF, nerve growth factor.
A.C. Lee et al. / Experimental Neurology 184 (2003) 295–303
While the majority of growth factor delivery systems are based on diffusion of growth factors from degradable polymers, other researchers have also studied affinitybased delivery systems that immobilize and release growth factors based on noncovalent interactions. One example of an affinity-based delivery system is a heparinbased delivery system for heparin-binding growth factors, such as basic fibroblast growth factor (bFGF), developed by Edelman and coworkers (Edelman et al., 1991). This delivery system consists of heparinconjugated Sepharose beads that are encapsulated in alginate. Heparin-binding growth factors are immobilized within this delivery system based on electrostatic interactions between basic heparin-binding domains on the growth factors and sulfated groups on heparin. The growth factors are protected from degradation by heparin and are released slowly over time. This delivery system has been tested in a number of animal models and has recently completed phase I clinical trials successfully (Latham et al., 1999). Based on these studies, it appears that trophic factors from both the neurotrophin and fibroblast growth factor families have the potential to enhance peripheral nerve regeneration. However, in many of these studies, the authors point out difficulties in maintaining growth factor release over the long duration of nerve regeneration. One strategy for addressing this problem is to develop a delivery system in which neurotrophin release is mediated by cell-mediated processes, rather than passive release. Such a paradigm would mimic physiological conditions, in which growth factors often exist in the sequestered state and are released, by active, cell-mediated processes over the duration of wound healing. Toward this goal, we developed a fibrin-based delivery system that immobilizes growth factors based on moderate to high affinity interactions with heparinbinding sites located within the delivery system matrix. This drug delivery system consists of heparin-binding peptides covalently immobilized within a fibrin matrix, heparin bound to the immobilized peptides, and growth factor (Sakiyama-Elbert and Hubbell, 2000b). The noncovalent interactions serve to sequester growth factor by slowing its diffusion through the fibrin matrix and allow the rate of release to be increased when the matrix is degraded by infiltrating cells during nerve regeneration. We have previously demonstrated the ability of this fibrin-based delivery system to promote neurite outgrowth when delivering NGF, bFGF, NT-3, and brain-derived neurotrophin factor (BDNF) in vitro (Sakiyama-Elbert and Hubbell, 2000a,b). In this study we examined the effects of such a system on nerve regeneration across a 13-mm rat sciatic nerve defect that exceeds the critical defect size for spontaneous rat sciatic regeneration through silicone conduits.
297
Table 1 Experimental groupsa Group
Description
I II III IV V VI VII VIII
Empty conduit Conduit ⫹ fibrin Isograft Conduit ⫹ fibrin Conduit ⫹ fibrin Conduit ⫹ fibrin Conduit ⫹ fibrin Conduit ⫹ fibrin
a
⫹ ⫹ ⫹ ⫹ ⫹
DS DS DS DS GF
⫹ NGF—50 ng/mL ⫹ NGF—20 ng/mL ⫹ NGF—5 ng/mL (no GF) (no DS, NGF—50 ng/mL)
DS, delivery system; GF, growth factor.
Materials and methods Experimental design Adult male Wistar rats were randomized into nine experimental groups (n ⫽ 6). The sciatic nerve was transected in each rat by the excision of a 4 –5-mm segment just proximal to the trifurcation of the nerve. The experimental groups are summarized in Table 1. In Group I an empty silicone conduit was placed between the two ends of the transected nerve. In Group II, a conduit filled with unmodified fibrin was place between the two ends of the transected nerve. Group III animals were treated with a reversed nerve isograft from syngeneic donor animals. Animals were killed at 6 weeks postoperatively and their sciatic nerves were harvested for histomorphometric analysis. Preparation of fibrin matrices Fibrinogen solutions were prepared by dissolving human plasminogen-free fibrinogen (Sigma, St. Louis, MO) in deionized water at 8 mg/mL for 1 h and dialyzing versus 4 L of Tris-buffered saline (TBS) (33 mM Tris, 8 g/L NaCl, 0.2 g/L KCl, all from Sigma) at pH 7.4 overnight to exchange salts present in the protein solution. The resulting solution was sterilized by filtration through 5.0 m and 0.22 m syringe filters and the final fibrinogen concentration determined by measuring absorbance at 280 nm. Fibrinogen was kept in a separate Eppendorf tube until just prior to mixing to prevent premature polymerization. For the heparin-based delivery system, TBS, 50 mM CaCl2 (Sigma) in TBS, 20 U/mL thrombin (Sigma), 25 mg/mL ␣2PI1–7ATIII121–134 peptide (Sakiyama-Elbert and Hubbell, 2000b), 45 mg/mL heparin (Sigma), and 20 g/mL recombinant human NGF (Peprotech, Rocky Hill, NJ) were mixed to achieve final concentrations of 50, 20, and 5 ng/mL NGF. Silicone tubing (0.058 in. inside diameter ⫻ 0.009 in. wall thickness) was gas sterilized overnight, cut into 15-mm segments, and soaked in 70% ethyl alcohol. Prior to filling, the tubes were removed from the alcohol and rinsed with sterile saline solution. Fibrin and all other components were
298
A.C. Lee et al. / Experimental Neurology 184 (2003) 295–303
control, a 13-mm segment of sciatic nerve was harvested from a syngeneic donor animal and inserted into the recipient animal in reverse orientation. Muscle and skin were closed with 4-0 vicryl and 4-0 nylon sutures, respectively. Histomorphometric evaluation Sciatic nerve segments were harvested and fixed in an immersion of cold 3% glutaraldehyde solution in 0.1 M phosphate buffer (pH 7.2). The tissues were postfixed with 1% osmium tetraoxide, ethanol dehydrated, and embedded in Araldite 502. One-micrometer thin sections were made from the tissue with the conduits or isografts and distal host segments, and stained with 1% toluidine blue for examination under light microscopy (Fox et al., 1999; Lee et al., 2000). The slides were evaluated by an observer blinded to the experimental groups for overall nerve architecture, quantity of regenerated nerve fibers, degree of myelination, and Wallerian degeneration.
Fig. 3. Dose-response relationship between matrix-bound -NGF concentration and mean fiber width of nerve fibers regenerating across a 13-mm nerve gap bridged with a silicone conduit. Mean myelinated fiber width was measured for a minimum of 100 axons at midconduit (left) and at the distal end of the conduit (right) by quantitative histomorphometry (n ⫽ 6). Mean values and standard deviation are shown. *denotes P ⬍ 0.05 versus the positive isograft control. Fibrin ⫹ DS group contained peptide and heparin but no NGF. NGF (no DS) group contained free NGF in the fibrin matrix by no delivery system. NGF, nerve growth factor.
mixed, and the resulting solution was drawn up into the silicone tube using a pipette and set aside for 2–3 min to polymerize. Operative procedure Operative procedures were performed using aseptic technique and microsurgical dissection and repair with the aid of an operating microscope. Male Wistar rats were anesthetized with a subcutaneous injection of a mixture of ketamine (50 mg/kg) and xylazine (10 mg/kg), and acepromazine (1 mg/kg). The right sciatic nerve was exposed through a dorsolateral gluteal-muscle splitting incision and a 4 –5-mm segment excised just proximal to the trifurcation of the sciatic nerve; 15-mm silicone tubes with respective experimental contents were attached using 10-0 nylon interrupted microepineurial sutures to the proximal and distal stumps, resulting in a 13-mm gap. In animals receiving the isograft
Fig. 4. Dose-response relationship between matrix-bound -NGF concentration and mean fiber density of the nerve regenerating across a 13-mm nerve gap bridged with a silicone conduit. Fiber density (fiber number/ mm2) was measured at midconduit (left) and at the distal end of the conduit (right) by quantitative histomorphometry (n ⫽ 6). Mean values and standard deviation are shown. *denotes P ⬍ 0.05 versus the positive isograft control. Fibrin ⫹ DS group contained peptide and heparin but no NGF. NGF (no DS) group contained free NGF in the fibrin matrix by no delivery system. NGF, nerve growth factor.
A.C. Lee et al. / Experimental Neurology 184 (2003) 295–303
299
Fig. 5. Histological sections of regenerating nerves at the midline of the conduit (or graft). (A) 5 ng/mL NGF ⫹ delivery system (DS); (B) 20 ng/mL NGF ⫹ delivery system; (C) 50 ng/mL NGF ⫹ delivery system; and (D) isograft (positive) control. NGF, nerve growth factor.
Morphometric evaluation
Statistics
Proximal and distal cross sections from the host nerve, and section through the conduit or graft, were evaluated. At 1000⫻ magnification, six representative fields per nerve were evaluated with an automated digital image-analysis system linked to morphometry software (Fox et al., 1999; Lee et al., 2000; Midha et al., 1994, 1997). The system digitized the microscope image and displayed it on a video monitor with a calibration of 0.125 m/pixel. Analysis of the digitized information, based on gray and white scales, was performed by the computer software. Total fascicular area and total fiber number was measured. At least 500 myelinated fibers were measured to determine the myelin width, axon width, and fiber diameter. From these primary measurements the following morphometric indices were calculated: myelin area, nerve fiber area, nerve fiber density (fiber number/mm2), percent nerve tissue (100 ⫻ neural area/intrafascicular area), percent neural debris (100 ⫻ neural debris/intrafascicular area), and axon:myelin and axon: fiber area ratios. Morphometric indices from experimental neural grafts were compared to isograft controls.
These experiments were designed to demonstrate statistical significance based on expected differences and animal numbers. The computer program Statistica (version 5.5, Statsoft, Tulsa, OK) was used to perform all statistical analyses. Multiple comparisons (greater than two groups) were analyzed by an analysis of variance using Scheff´e’s post hoc test for multiple comparisons (all possible comparisons). Statistical significance is set at P ⬍ 0.05.
Results and discussion The ability of fibrin-based delivery systems to deliver NGF in a biologically active form over the course of nerve regeneration was tested. Silastic nerve guide conduits filled with the fibrin-based delivery system and one or three doses of NGF were implanted in a 13-mm rat sciatic nerve lesion and the progress of nerve regeneration was examined at 6 weeks post implantation using histomorphometric analysis. Sections of the sciatic nerve were evaluated for fiber den-
300
A.C. Lee et al. / Experimental Neurology 184 (2003) 295–303
A.C. Lee et al. / Experimental Neurology 184 (2003) 295–303
sity, percent nerve, total fiber number, and mean nerve fiber width at the midline (midpoint of the conduit) and 5 mm distal to the conduit (Figs. 1– 4, Table 2). Axonal regeneration in the NGF delivery system groups revealed a marked dose-dependent effect in all assessments. The total fiber numbers (Fig. 1) at both the midline and distal to the conduit in the two higher NGF dose (20 and 50 ng/mL) groups were not statistically different from the positive (isograft) control. Regeneration was observed in 4 of 6 animals for 20 and 50 ng/mL NGF doses. The isograft controls had an average of 9920 ⫾ 4177 fibers, while the 50 ng/mL NGF delivery system had an average of 6462 ⫾ 4177 fibers. The average number of fibers in normal rat sciatic nerve is approximately 7115 ⫾ 413 (Mackinnon et al., 1991), so both of these groups are relatively similar to the normal number of fibers. For the fibrin, fibrin ⫹ delivery system, and NGF (no delivery system) groups only 2 of 6 animals showed any regenerating nerve in the conduit, resulting in much lower fiber counts. (At the distal end of the nerve, regeneration had not progressed as rapidly as in the conduit, so the fiber counts were lower than at the conduit midline.) However, both the 20 and 50 ng/mL NGF delivery system groups were not statically different from the isograft control. This result suggests that the NGF delivery system can promote a similar level of nerve regeneration as the isograft control. The percentage of tissue area occupied by nerve provides a measure of the quality of the regenerating nerve. At the midline of the conduit (Fig. 2), only the empty conduit group had a significantly lower percentage of nerve than the isograft group, due to the large variability of the measurement and varying nerve cable thickness. However, at the distal end of the nerve differences between groups become more apparent. Only the 20 and 50 ng/mL NGF delivery system groups were not statistically different from the isograft control. The data suggest that regeneration had progressed more slowly in the other groups at the distal end of the nerve. The mean fiber width provides a measure of the maturity of the regenerating nerve fibers. At the midline of the conduit, the mean fiber width (Fig. 3) of all groups containing NGF (including the group lacking the delivery system) was not significantly different from the isograft control (2.93 m). For the NGF doses of 20 and 50 ng/mL groups (with the delivery system), the mean fiber width of nerve fibers distal to the conduit was not statistically different from the positive isograft control (2.82 m). For normal rat sciatic nerve, the mean myelinated fiber diameter is 6.5 m (Mackinnon et al., 1991). In nerves that are transected and immediately repaired (cut and repair) the average fiber diameters are 3.1 and 3.4 m at 1 and 3 months postopera-
301
tively, respectively (Mackinnon et al., 1991). These values are similar to values obtained at 6 weeks in our study at the conduit midline for nerve isograft and NGF/delivery system. Even at 24 months, the cut and repaired nerve still only has a mean fiber diameter of 4.7 m, which is statistically less than prior to injury (Mackinnon et al., 1991). This result suggests that NGF may play a role in the maturation of nerve fibers and that sustained release of NGF from the delivery system is important to promote maturation of fibers in the distal nerve. The nerve fiber density (fiber number/mm2) at the distal end of the conduit (Fig. 4) was not statistically different for the positive isograft control and NGF doses of 20 and 50 ng/mL (with the delivery system). For normal sciatic nerve, the fiber density is 11,882 fibers/mm2 (Mackinnon et al., 1991). Fiber densities of 25,382 and 28,934 fibers/mm2 are observed for cut and repair injuries at 1 and 3 months postoperatively, respectively (Mackinnon et al., 1991), which are similar to our NGF delivery system results at the midline of the conduit. This suggests that regeneration at the midline of the conduit is comparable to the cut and repair, but that regeneration in the distal segment of the conduit treatment groups lags that in the distal segment of the cut and repair treatment (due to the 13-mm length of the conduit gap). Images of the fibrin delivery system containing different NGF doses and of the isograft (positive) control are shown at the midline of the conduit in Fig. 5 and distal to the conduit in Fig. 6. Although there are a significant number of nerve fibers in the regenerating nerves, it is clear from the histological sections that the regenerating fibers have not fully matured and that further time will be required before functional recovery can occur. Regeneration has clearly progressed to a greater extent at the midline of the conduit than in the distal nerve. We selected the 6-week time point because, although axons have started to bridge the nerve gap, we can still see the effects of therapeutics on the time course of regeneration by looking at the maturity (diameter) of the nerve fibers. The fibrin matrix provides a biodegradable scaffold for fiber outgrowth in this nerve regeneration model, as well as the base for our drug delivery system. Based on our histology (Fig. 5), most if not all of the fibrin matrix has been resorbed by 6 weeks post implantation and has been replaced by the regenerating nerve cable. This is ideal because we do not want the fibrin matrix to inhibit the progress of regenerating fibers or to compress the nerve cable. The objective of this research was to develop and test a novel fibrin-based delivery system engineered to deliver biologically active growth factors in vivo. We have previously demonstrated the ability of such delivery systems to
Fig. 6. Histological sections of regenerating nerves at the distal end of the conduit (or graft). (A) 5 ng/mL NGF ⫹ delivery system (DS); (B) 20 ng/mL NGF ⫹ delivery system; (C) 50 ng/mL NGF ⫹ delivery system; (D) isograft (positive) control; (E) empty conduit; (F) fibrin alone; (G) Delivery system ⫹ fibrin (no NGF); and (H) NGF ⫹ fibrin (no delivery system). NGF, nerve growth factor.
302
A.C. Lee et al. / Experimental Neurology 184 (2003) 295–303
Table 2 Nerve histomorphometry dataa Group
5 ng/ml NGF 20 ng/ml NGF 50 ng/ml NGF Isograft Empty Fibrin Fibrin ⫹ DS NGF (no DS) a b
Midline of conduit
Distal to conduit
Number of nerve fibers
Percent nerve
Mean fiber diameter (m)
Density of nerve fibers
Number of nerve fibers
Percent nerve
Mean fiber diameter (m)
Density of nerve fibers
833 ⫾ 1518b 4527 ⫾ 4964 6462 ⫾ 4177 9920 ⫾ 4177 0 ⫾ 0b 1347 ⫾ 2343b 3179 ⫾ 4361b 4636 ⫾ 5654b
4.4 ⫾ 12.5 12.7 ⫾ 1.69 25.1 ⫾ 10 17.5 ⫾ 11.2 1.5 ⫾ 3.3b 7.1 ⫾ 13.2 14.64 ⫾ 20.1 14.45 ⫾ 16.7
2.8 ⫾ 0.24 3.01 ⫾ 0.17 2.50 ⫾ 1.23 2.93 ⫾ 0.29 0.64 ⫾ 1.42b 0.91 ⫾ 1.41b 1.23 ⫾ 1.69b 1.4 ⫾ 1.54
16276 ⫾ 13565 28678 ⫾ 6132 27237 ⫾ 8000 19122 ⫾ 9713 1404 ⫾ 3138b 9074 ⫾ 15821 14948 ⫾ 19000 17832 ⫾ 17000
0 ⫾ 0b 2370 ⫾ 3081 4094 ⫾ 24969 5315 ⫾ 3431 75 ⫾ 105b 103 ⫾ 253b 1303 ⫾ 2913b 1069 ⫾ 2349b
0 ⫾ 0b 2.8 ⫾ 3.9 5.1 ⫾ 8.1 9.3 ⫾ 9.9 0.03 ⫾ 0.07b 0.11 ⫾ 0.27b 2.0 ⫾ 4.5b 1.6 ⫾ 3.6b
0 ⫾ 0b 1.6 ⫾ 0.11 1.71 ⫾ 0.15 2.82 ⫾ 0.30 0.53 ⫾ 1.19b 0.45 ⫾ 1.10b 0.56 ⫾ 1.25b 0.81 ⫾ 1.26b
0 ⫾ 0b 7221 ⫾ 7361 10389 ⫾ 10669 10731 ⫾ 11112 80 ⫾ 110b 142 ⫾ 3478b 2453 ⫾ 5485 3055 ⫾ 7134
NGF, nerve growth factor; DS, delivery system. Denotes P ⬍ 0.05 versus the positive isograft control.
promote neurite extension in a dose-dependent manner in vitro. However, the biological environment surrounding nerve regeneration in vivo is much more complicated. Therefore, we sought to determine whether the materials we have previously tested in vitro could be successfully applied in vivo. The results of this study demonstrate that the incorporation of a novel delivery system providing controlled release of growth factors promotes peripheral nerve regeneration. Results in all morphometric measurements demonstrate that an optimal NGF dose, when combined with the delivery system, was not statistically different from the positive isograft control that provides the clinical “gold standard” for long gap peripheral nerve repair. Furthermore, the regeneration is comparable in terms of fiber diameter (maturity) and fiber density (at the conduit midline) to that observed in end-to end (short gap) nerve repairs. These results are rather remarkable considering that the conduit gap is 13 mm in length, significantly larger than the critical defect length (10 mm) for rat sciatic nerve and that the conduit lacks both the cellular components of nerve grafts and the basal lamina, both of which are hypothesized to facilitate regeneration in nerve grafts. Previous reports have demonstrated (Aebischer et al., 1989; Doolabh et al., 1996; Francel et al., 1997; Frostick et al., 1998; Maeda et al., 1993; Terenghi, 1999) bridging of short gaps (10 mm) using conduits (Jenq and Coggeshall, 1987; Lundborg et al., 1982; Seckel et al., 1984); but, empty conduits and fibrin alone have not proven sufficient to bridge gaps larger than 10 mm in a rat sciatic nerve. Clinically the most critical need is the treatment of nerve gaps greater than 10 mm to provide an alternative to nerve autografts. The materials investigated in this study may be useful in promoting peripheral nerve regeneration in vivo following transection due to trauma or surgery. This study also demonstrates the utility of a novel affinity-based delivery system for use in promoting tissue regeneration in vivo. This type of delivery system is extremely versatile and adaptable. It
can be modified to contain exogenous cell adhesion domains as well as drug delivery systems and has the potential for application in a variety of tissue-engineered constructs other than nerve, including bone, skin, vascular tissue, and cartilage.
References Aebischer, P., Salessiotis, A.N., Winn, S.R., 1989. Basic fibroblast growth factor released from synthetic guidance channels facilitates peripheral nerve regeneration across long nerve gaps. J. Neurosci. Res. 23, 232– 289. Archibald, S.J., Krarup, C., Shefner, J., Li, S.T., Madison, R.D., 1991. A collagen-based nerve guide conduit for peripheral nerve repair: an electrophysiological study of nerve regeneration in rodents and nonhuman primates. J. Comp. Neurol. 306, 685– 696. Beazley, W.C., Milek, M.A., Reiss, B.H., 1984. Results of nerve grafting in severe soft tissue injuries. Clin. Orthop. 208, 208 –212. Borkenhagen, M., 1997. Effect of NT-3 and BDNF released from nerve guidance channels on dorsal root regeneration, EPFL, Lausanne, Switzerland. Borkenhagen, M., Stoll, R.C., Neuenschwander, P., Suter, U.W., Aebischer, P., 1998. In vivo performance of a new biodegradable polyester urethane system used as a nerve guidance channel. Biomaterials 19, 2155–2165. Dellon, A.L., Mackinnon, S.E., 1988. An alternative to the classical nerve graft for the management of the short nerve gap. Plast. Reconstr. Surg. 82, 849 – 856. den Dunnen, W.F., van der Lei, B., Robinson, P.H., Holwerda, A., Pennings, A.J., Schakenraad, J.M., 1995. Biological performance of a degradable poly(lactic acid-epsilon-caprolactone) nerve guide: influence of tube dimensions. J. Biomed. Mater. Res. 29, 757–766. Doolabh, V.B., Hertl, M.C., Mackinnon, S.E., 1996. The role of conduits in nerve repair: a review. Rev. Neurosci. 7, 47– 84. Edelman, E., Mathiowitz, E., Langer, R., Klagsbrun, M., 1991. Controlled and modulated release of basic fibroblast growth factor. Biomaterials 12, 612– 626. Evans, G.R., Brandt, K., Widmer, M.S., Lu, L., Meszlenyi, R.K., Gupta, P.K., Mikos, A.G., Hodges, J., Williams, J., Gurlek, A., Nabawi, A., Lohman, R., Patrick Jr., C.W., 1999. In vivo evaluation of poly(L-lactic acid) porous conduits for peripheral nerve regeneration. Biomaterials 20, 1109 –1115.
A.C. Lee et al. / Experimental Neurology 184 (2003) 295–303 Fields, R.D., Le Beau, J.M., Longo, F.M., Ellisman, M.H., 1989. Nerve regeneration through artificial tubular implants. Prog. Neurobiol. 33, 87–134. Fox, D.J., Doolabh, V.B., Mackinnon, S.E., Genden, E.M., Hunter, D.A., 1999. Decreased cyclosporin A requirement with anti-ICAM-1 and anti LFA-1a in a peripheral nerve allotransplantation model. Restor. Neurol. Neurosci. 15, 319 –326. Francel, P.C., Francel, T.J., Mackinnon, S.E., Hertl, C., 1997. Enhancing nerve regeneration across a silicone tube conduit by using interposed short-segment nerve grafts. J. Neurosurg. 87, 887– 892. Frostick, S.P., Yin, Q., Kemp, G.J., 1998. Schwann cells, neurotrophic factors, and peripheral nerve regeneration. Microsurgery 18, 397– 405. Jenq, C.B., Coggeshall, R.E., 1987. Permeable tubes increase the length of the gap that regenerating axons can span. Brain Res. 408, 239 –242. Laham, R., Sellke, F., Edelman, E., Pearlman, J., Ware, J., Brown, D., Gold, J., Simons, M., 1999. Local perivascular delivery of basic fibroblast growth factor in patients undergoing coronary bypass surgery. Circulation 100, 1865–1871. Lee, M., Doolabh, V.B., Mackinnon, S.E., Jost, S., 2000. FK506 promotes functional recovery in crushed rat sciatic nerve. Muscle Nerve 23, 633– 640. Lundborg, G., 1990. Nerve regeneration problems in a clinical perspective. Restor. Neurol. Neurosci. 1, 297–302. Lundborg, G., Dahlin, L.B., Danielsen, N., Gelberman, R.H., Longo, F.M., Powell, H.C., Varon, S., 1982. Nerve regeneration in silicone chambers: influence of gap length and of distal stump components. Exp. Neurol. 76, 361–375. Mackinnon, S.E., 1989. Surgical management of the peripheral nerve gap. Clin. Plast. Surg. 16, 587– 603. Mackinnon, S.E., Dellon, A.L., 1990. Clinical nerve reconstruction with a bioabsorbable polyglycolic acid tube. Plast. Reconstr. Surg. 85, 419 – 424.
303
Mackinnon, S.E., Dellon, A.L., O’Brien, J.P., 1991. Changes in nerve fiber numbers distal to a nerve repair in the rat sciatic nerve model. Muscle Nerve 14, 1116 –1122. Maeda, T., Mackinnon, S.E., Best, T.J., Evans, P.J., Hunter, D.A., Midha, R.T., 1993. Regeneration across “stepping-stone” nerve grafts. Brain Res. 618, 196 –202. Midha, R., Mackinnon, S.E., Becker, L.E., 1994. The fate of Schwann cells in peripheral nerve allografts. J. Neuropathol. Exp. Neurol. 53, 316–322. Midha, R., Munro, C.A., Mackinnon, S.E., Ang, L.C., 1997. Motor and sensory specificity of host nerve axons influence nerve allograft rejection. J. Neuropathol. Exp. Neurol. 56, 421– 434. Otto, D., Unsicker, K., Grothe, C., 1987. Pharmacological effects of nerve growth factor and fibroblast growth factor applied to the transectioned sciatic nerve on neuron death in adult rat dorsal root ganglia. Neurosci. Lett. 83, 156 –160. Rich, R., Alexander, T., Pryor, J., Hollowell, J., 1989. Nerve growth factor enhances regeneration through silicone chambers. Exp. Neurol. 105, 162–170. Sakiyama-Elbert, S., Hubbell, J., 2000a. Controlled release of nerve growth factor from a heparin-containing fibrin-based cell ingrowth matrix. J. Controlled Rel. 69, 149 –158. Sakiyama-Elbert, S., Hubbell, J., 2000b. Development of fibrin derivatives for controlled release of heparin-binding growth factors. J. Controlled Rel. 65, 389 – 402. Seckel, B.R., 1990. Enhancement of peripheral nerve regeneration. Muscle Nerve 13, 785– 800. Seckel, B.R., Chiu, T.H., Nyilas, E., Sidman, R.L., 1984. Nerve regeneration through synthetic biodegradable nerve guides: regulation by the target organ. Plast. Reconstr. Surg. 74, 173–181. Terenghi, G., 1999. Peripheral nerve regeneration and neurotrophic factors. J. Anat. 194 (Pt 1), 1–14. Varon, S., Conner, J.M., 1994. Nerve growth factor in CNS repair. J. Neurotrauma 11, 473– 486.
Experimental Neurology 184 (2003) 295–303
www.elsevier.com/locate/yexnr
Controlled release of nerve growth factor enhances sciatic nerve regeneration Annie C. Lee,a Vivian M. Yu,a James B. Lowe, III,a Michael J. Brenner,c Daniel A. Hunter,a Susan E. Mackinnon,a and Shelly E. Sakiyama-Elberta,b,* a
Division of Plastic and Reconstructive Surgery, Department of Surgery, Washington University School of Medicine, St. Louis, MO 63110, USA b Department of Biomedical Engineering, Washington University, St. Louis, MO 63130, USA c Department of Otolaryngology, Washington University School of Medicine, St. Louis, MO 63110, USA Received 16 January 2003; revised 21 April 2003; accepted 6 May 2003
Abstract Based on previous studies demonstrating the potential of growth factors to enhance peripheral nerve regeneration, we developed a novel growth factor delivery system to provide sustained delivery of nerve growth factor (NGF). This delivery system uses heparin to immobilize NGF and slow its diffusion from a fibrin matrix. This system has been previously shown to enhance neurite outgrowth in vitro, and in this study, we evaluated the ability of this delivery system to enhance nerve regeneration through conduits. We tested the effect of controlled NGF delivery on peripheral nerve regeneration in a 13-mm rat sciatic nerve defect. The heparin-containing delivery system was studied in combination with three doses of NGF (5, 20, or 50 ng/mL) and the results were compared with positive controls (isografts) and negative controls (fibrin alone, NGF alone, and empty conduits). Nerves were harvested at 6 weeks postoperatively for histomorphometric analysis. Axonal regeneration in the delivery system groups revealed a marked dose-dependent effect. The total number of nerve fibers at both the mid-conduit level and in the distal nerve showed no statistical difference for NGF doses at 20 and 50 ng/mL from the isograft (positive control). The results of this study demonstrate that the incorporation of a novel delivery system providing controlled release of growth factors enhances peripheral nerve regeneration and represents a significant contribution toward enhancing nerve regeneration across short nerve gaps. © 2003 Elsevier Science (USA). All rights reserved. Keywords: Peripheral nerve; Heparin; Fibrin; Nerve guide conduit; Tissue engineering; Drug delivery
Introduction Despite recent advances in the understanding of nerve regeneration and in surgical techniques, the complete functional recovery in a damaged nerve is rare (Seckel, 1990). It is estimated that more than 200,000 nerve repair procedures are performed annually in the United States alone (Archibald et al., 1991). The most common method of nerve repair is to directly suture the two severed nerve ends together; however, this is often not feasible if the gap
* Corresponding author. Department of Biomedical Engineering, Washington University, Campus Box 1097, One Brookings Drive, St. Louis, MO 63130-4899, USA. Fax: ⫹1-314-935-7448. E-mail address: [email protected] (S.E. SakiyamaElbert).
between the two ends is too large to allow a tension-free nerve repair. In these cases, a nerve autograft is indicated to improve functional recovery (Mackinnon, 1989). The autograft provides a scaffold for the regenerating nerve and guides it to its proper target. However, nerve grafting has its limitations, because of the associated donor site mobidity. Both of these procedures, direct repair and nerve autograft, are associated with less than optimum results (Lundborg, 1990). Less than 25% of patients who received direct repair or autograft repair of the median nerve at the wrist level regain full motor function and only 1–3% recovered normal sensation after 5 years (Beazley et al., 1984; Dellon and Mackinnon, 1988). One alternative method to autografting is to use a nerve guidance conduit (NGC) to connect the proximal and distal ends of the severed nerve and guide the regeneration of
0014-4886/03/$ – see front matter © 2003 Elsevier Science (USA). All rights reserved. doi:10.1016/S0014-4886(03)00258-9
296
A.C. Lee et al. / Experimental Neurology 184 (2003) 295–303
Mackinnon and Dellon, 1990). The use of biodegradable conduit is highly desirable, and our ultimate goal is to combine the “tissue-engineered” drug delivery matrices with biodegradable conduits. However, for this study we focused only on the drug delivery matrices to serve as conduit fillers. Neurotrophins have been shown to enhance peripheral nerve regeneration. NGF was reported to protect neurons from injury-induced death in lesioned sciatic nerves (Otto et al., 1987). Addition of NGF to saline-filled silicon tubes resulted in an increase in the number of myelinated axons in the regenerating nerve (Rich et al., 1989). NGF has also been shown to facilitate regeneration of hippocampal neurons across a peripheral nerve bridge (Varron and Conner, 1994). Neurotrophin 3 (NT-3) was found to increase the number of myelinated axons in a regenerating dorsal root through NGCs with NT-3 releasing rods in the wall (Borkenhagen, 1997).
Fig. 1. Dose-response relationship between matrix-bound -NGF concentration and number of nerve fibers regenerating across a 13-mm nerve gap bridged with a silicone conduit. The total number of nerve fibers were measured at midconduit (left) and at the distal end of the conduit (right) by quantitative histomorphometry (n ⫽ 6). Mean values and standard deviation are shown. *denotes P ⬍ 0.05 versus the positive isograft control. Fibrin ⫹ DS group contained peptide and heparin but no NGF. NGF (no DS) group contained free NGF in the fibrin matrix by no delivery system. NGF, nerve growth factor.
axons back to the appropriate target. These tubes allow the microenvironment of regeneration to be controlled by manipulating the contents of the NGC. Numerous studies have been performed to determine some of the fundamental mechanisms of regeneration by varying the conditions within such tubes. A comprehensive overview of the various materials used for NGC construction can be found in the review by Fields et al. (1989). Prior to 1989, silicone elastomer was the material of choice for NGC construction. More recently, biodegradable NGCs have been studied because an additional surgical procedure is not required to remove degradable NGCs. The use of poly(glycolic acid), poly(L-lactic acid), poly(lactic acid-co--caprolactone), poly(R-3-hydroxybutyric acid-co-(R)-3-hydroxyvaleric acid)-diol (PHB) conduits have been studied in short gaps in the rat sciatic nerve, and the degradable NGCs were found to promote comparable regeneration versus nerve isografts (Borkenhagen et al., 1998; den Dunnen et al., 1995; Evans et al., 1999;
Fig. 2. Dose-response relationship between matrix-bound -NGF concentration and percent nerve area in the tissue regenerating across a 13-mm nerve gap bridged with a silicone conduit. Percent nerve tissue was calculated as 100 times the neural area divided by the intrafascicular area at midconduit (left) and at the distal end of the conduit (right) by quantitative histomorphometry (n ⫽ 6). Mean values and standard deviation are shown. *denotes P ⬍ 0.05 versus the positive isograft control. Fibrin ⫹ DS group contained peptide and heparin but no NGF. NGF (no DS) group contained free NGF in the fibrin matrix by no delivery system. NGF, nerve growth factor.
A.C. Lee et al. / Experimental Neurology 184 (2003) 295–303
While the majority of growth factor delivery systems are based on diffusion of growth factors from degradable polymers, other researchers have also studied affinitybased delivery systems that immobilize and release growth factors based on noncovalent interactions. One example of an affinity-based delivery system is a heparinbased delivery system for heparin-binding growth factors, such as basic fibroblast growth factor (bFGF), developed by Edelman and coworkers (Edelman et al., 1991). This delivery system consists of heparinconjugated Sepharose beads that are encapsulated in alginate. Heparin-binding growth factors are immobilized within this delivery system based on electrostatic interactions between basic heparin-binding domains on the growth factors and sulfated groups on heparin. The growth factors are protected from degradation by heparin and are released slowly over time. This delivery system has been tested in a number of animal models and has recently completed phase I clinical trials successfully (Latham et al., 1999). Based on these studies, it appears that trophic factors from both the neurotrophin and fibroblast growth factor families have the potential to enhance peripheral nerve regeneration. However, in many of these studies, the authors point out difficulties in maintaining growth factor release over the long duration of nerve regeneration. One strategy for addressing this problem is to develop a delivery system in which neurotrophin release is mediated by cell-mediated processes, rather than passive release. Such a paradigm would mimic physiological conditions, in which growth factors often exist in the sequestered state and are released, by active, cell-mediated processes over the duration of wound healing. Toward this goal, we developed a fibrin-based delivery system that immobilizes growth factors based on moderate to high affinity interactions with heparinbinding sites located within the delivery system matrix. This drug delivery system consists of heparin-binding peptides covalently immobilized within a fibrin matrix, heparin bound to the immobilized peptides, and growth factor (Sakiyama-Elbert and Hubbell, 2000b). The noncovalent interactions serve to sequester growth factor by slowing its diffusion through the fibrin matrix and allow the rate of release to be increased when the matrix is degraded by infiltrating cells during nerve regeneration. We have previously demonstrated the ability of this fibrin-based delivery system to promote neurite outgrowth when delivering NGF, bFGF, NT-3, and brain-derived neurotrophin factor (BDNF) in vitro (Sakiyama-Elbert and Hubbell, 2000a,b). In this study we examined the effects of such a system on nerve regeneration across a 13-mm rat sciatic nerve defect that exceeds the critical defect size for spontaneous rat sciatic regeneration through silicone conduits.
297
Table 1 Experimental groupsa Group
Description
I II III IV V VI VII VIII
Empty conduit Conduit ⫹ fibrin Isograft Conduit ⫹ fibrin Conduit ⫹ fibrin Conduit ⫹ fibrin Conduit ⫹ fibrin Conduit ⫹ fibrin
a
⫹ ⫹ ⫹ ⫹ ⫹
DS DS DS DS GF
⫹ NGF—50 ng/mL ⫹ NGF—20 ng/mL ⫹ NGF—5 ng/mL (no GF) (no DS, NGF—50 ng/mL)
DS, delivery system; GF, growth factor.
Materials and methods Experimental design Adult male Wistar rats were randomized into nine experimental groups (n ⫽ 6). The sciatic nerve was transected in each rat by the excision of a 4 –5-mm segment just proximal to the trifurcation of the nerve. The experimental groups are summarized in Table 1. In Group I an empty silicone conduit was placed between the two ends of the transected nerve. In Group II, a conduit filled with unmodified fibrin was place between the two ends of the transected nerve. Group III animals were treated with a reversed nerve isograft from syngeneic donor animals. Animals were killed at 6 weeks postoperatively and their sciatic nerves were harvested for histomorphometric analysis. Preparation of fibrin matrices Fibrinogen solutions were prepared by dissolving human plasminogen-free fibrinogen (Sigma, St. Louis, MO) in deionized water at 8 mg/mL for 1 h and dialyzing versus 4 L of Tris-buffered saline (TBS) (33 mM Tris, 8 g/L NaCl, 0.2 g/L KCl, all from Sigma) at pH 7.4 overnight to exchange salts present in the protein solution. The resulting solution was sterilized by filtration through 5.0 m and 0.22 m syringe filters and the final fibrinogen concentration determined by measuring absorbance at 280 nm. Fibrinogen was kept in a separate Eppendorf tube until just prior to mixing to prevent premature polymerization. For the heparin-based delivery system, TBS, 50 mM CaCl2 (Sigma) in TBS, 20 U/mL thrombin (Sigma), 25 mg/mL ␣2PI1–7ATIII121–134 peptide (Sakiyama-Elbert and Hubbell, 2000b), 45 mg/mL heparin (Sigma), and 20 g/mL recombinant human NGF (Peprotech, Rocky Hill, NJ) were mixed to achieve final concentrations of 50, 20, and 5 ng/mL NGF. Silicone tubing (0.058 in. inside diameter ⫻ 0.009 in. wall thickness) was gas sterilized overnight, cut into 15-mm segments, and soaked in 70% ethyl alcohol. Prior to filling, the tubes were removed from the alcohol and rinsed with sterile saline solution. Fibrin and all other components were
298
A.C. Lee et al. / Experimental Neurology 184 (2003) 295–303
control, a 13-mm segment of sciatic nerve was harvested from a syngeneic donor animal and inserted into the recipient animal in reverse orientation. Muscle and skin were closed with 4-0 vicryl and 4-0 nylon sutures, respectively. Histomorphometric evaluation Sciatic nerve segments were harvested and fixed in an immersion of cold 3% glutaraldehyde solution in 0.1 M phosphate buffer (pH 7.2). The tissues were postfixed with 1% osmium tetraoxide, ethanol dehydrated, and embedded in Araldite 502. One-micrometer thin sections were made from the tissue with the conduits or isografts and distal host segments, and stained with 1% toluidine blue for examination under light microscopy (Fox et al., 1999; Lee et al., 2000). The slides were evaluated by an observer blinded to the experimental groups for overall nerve architecture, quantity of regenerated nerve fibers, degree of myelination, and Wallerian degeneration.
Fig. 3. Dose-response relationship between matrix-bound -NGF concentration and mean fiber width of nerve fibers regenerating across a 13-mm nerve gap bridged with a silicone conduit. Mean myelinated fiber width was measured for a minimum of 100 axons at midconduit (left) and at the distal end of the conduit (right) by quantitative histomorphometry (n ⫽ 6). Mean values and standard deviation are shown. *denotes P ⬍ 0.05 versus the positive isograft control. Fibrin ⫹ DS group contained peptide and heparin but no NGF. NGF (no DS) group contained free NGF in the fibrin matrix by no delivery system. NGF, nerve growth factor.
mixed, and the resulting solution was drawn up into the silicone tube using a pipette and set aside for 2–3 min to polymerize. Operative procedure Operative procedures were performed using aseptic technique and microsurgical dissection and repair with the aid of an operating microscope. Male Wistar rats were anesthetized with a subcutaneous injection of a mixture of ketamine (50 mg/kg) and xylazine (10 mg/kg), and acepromazine (1 mg/kg). The right sciatic nerve was exposed through a dorsolateral gluteal-muscle splitting incision and a 4 –5-mm segment excised just proximal to the trifurcation of the sciatic nerve; 15-mm silicone tubes with respective experimental contents were attached using 10-0 nylon interrupted microepineurial sutures to the proximal and distal stumps, resulting in a 13-mm gap. In animals receiving the isograft
Fig. 4. Dose-response relationship between matrix-bound -NGF concentration and mean fiber density of the nerve regenerating across a 13-mm nerve gap bridged with a silicone conduit. Fiber density (fiber number/ mm2) was measured at midconduit (left) and at the distal end of the conduit (right) by quantitative histomorphometry (n ⫽ 6). Mean values and standard deviation are shown. *denotes P ⬍ 0.05 versus the positive isograft control. Fibrin ⫹ DS group contained peptide and heparin but no NGF. NGF (no DS) group contained free NGF in the fibrin matrix by no delivery system. NGF, nerve growth factor.
A.C. Lee et al. / Experimental Neurology 184 (2003) 295–303
299
Fig. 5. Histological sections of regenerating nerves at the midline of the conduit (or graft). (A) 5 ng/mL NGF ⫹ delivery system (DS); (B) 20 ng/mL NGF ⫹ delivery system; (C) 50 ng/mL NGF ⫹ delivery system; and (D) isograft (positive) control. NGF, nerve growth factor.
Morphometric evaluation
Statistics
Proximal and distal cross sections from the host nerve, and section through the conduit or graft, were evaluated. At 1000⫻ magnification, six representative fields per nerve were evaluated with an automated digital image-analysis system linked to morphometry software (Fox et al., 1999; Lee et al., 2000; Midha et al., 1994, 1997). The system digitized the microscope image and displayed it on a video monitor with a calibration of 0.125 m/pixel. Analysis of the digitized information, based on gray and white scales, was performed by the computer software. Total fascicular area and total fiber number was measured. At least 500 myelinated fibers were measured to determine the myelin width, axon width, and fiber diameter. From these primary measurements the following morphometric indices were calculated: myelin area, nerve fiber area, nerve fiber density (fiber number/mm2), percent nerve tissue (100 ⫻ neural area/intrafascicular area), percent neural debris (100 ⫻ neural debris/intrafascicular area), and axon:myelin and axon: fiber area ratios. Morphometric indices from experimental neural grafts were compared to isograft controls.
These experiments were designed to demonstrate statistical significance based on expected differences and animal numbers. The computer program Statistica (version 5.5, Statsoft, Tulsa, OK) was used to perform all statistical analyses. Multiple comparisons (greater than two groups) were analyzed by an analysis of variance using Scheff´e’s post hoc test for multiple comparisons (all possible comparisons). Statistical significance is set at P ⬍ 0.05.
Results and discussion The ability of fibrin-based delivery systems to deliver NGF in a biologically active form over the course of nerve regeneration was tested. Silastic nerve guide conduits filled with the fibrin-based delivery system and one or three doses of NGF were implanted in a 13-mm rat sciatic nerve lesion and the progress of nerve regeneration was examined at 6 weeks post implantation using histomorphometric analysis. Sections of the sciatic nerve were evaluated for fiber den-
300
A.C. Lee et al. / Experimental Neurology 184 (2003) 295–303
A.C. Lee et al. / Experimental Neurology 184 (2003) 295–303
sity, percent nerve, total fiber number, and mean nerve fiber width at the midline (midpoint of the conduit) and 5 mm distal to the conduit (Figs. 1– 4, Table 2). Axonal regeneration in the NGF delivery system groups revealed a marked dose-dependent effect in all assessments. The total fiber numbers (Fig. 1) at both the midline and distal to the conduit in the two higher NGF dose (20 and 50 ng/mL) groups were not statistically different from the positive (isograft) control. Regeneration was observed in 4 of 6 animals for 20 and 50 ng/mL NGF doses. The isograft controls had an average of 9920 ⫾ 4177 fibers, while the 50 ng/mL NGF delivery system had an average of 6462 ⫾ 4177 fibers. The average number of fibers in normal rat sciatic nerve is approximately 7115 ⫾ 413 (Mackinnon et al., 1991), so both of these groups are relatively similar to the normal number of fibers. For the fibrin, fibrin ⫹ delivery system, and NGF (no delivery system) groups only 2 of 6 animals showed any regenerating nerve in the conduit, resulting in much lower fiber counts. (At the distal end of the nerve, regeneration had not progressed as rapidly as in the conduit, so the fiber counts were lower than at the conduit midline.) However, both the 20 and 50 ng/mL NGF delivery system groups were not statically different from the isograft control. This result suggests that the NGF delivery system can promote a similar level of nerve regeneration as the isograft control. The percentage of tissue area occupied by nerve provides a measure of the quality of the regenerating nerve. At the midline of the conduit (Fig. 2), only the empty conduit group had a significantly lower percentage of nerve than the isograft group, due to the large variability of the measurement and varying nerve cable thickness. However, at the distal end of the nerve differences between groups become more apparent. Only the 20 and 50 ng/mL NGF delivery system groups were not statistically different from the isograft control. The data suggest that regeneration had progressed more slowly in the other groups at the distal end of the nerve. The mean fiber width provides a measure of the maturity of the regenerating nerve fibers. At the midline of the conduit, the mean fiber width (Fig. 3) of all groups containing NGF (including the group lacking the delivery system) was not significantly different from the isograft control (2.93 m). For the NGF doses of 20 and 50 ng/mL groups (with the delivery system), the mean fiber width of nerve fibers distal to the conduit was not statistically different from the positive isograft control (2.82 m). For normal rat sciatic nerve, the mean myelinated fiber diameter is 6.5 m (Mackinnon et al., 1991). In nerves that are transected and immediately repaired (cut and repair) the average fiber diameters are 3.1 and 3.4 m at 1 and 3 months postopera-
301
tively, respectively (Mackinnon et al., 1991). These values are similar to values obtained at 6 weeks in our study at the conduit midline for nerve isograft and NGF/delivery system. Even at 24 months, the cut and repaired nerve still only has a mean fiber diameter of 4.7 m, which is statistically less than prior to injury (Mackinnon et al., 1991). This result suggests that NGF may play a role in the maturation of nerve fibers and that sustained release of NGF from the delivery system is important to promote maturation of fibers in the distal nerve. The nerve fiber density (fiber number/mm2) at the distal end of the conduit (Fig. 4) was not statistically different for the positive isograft control and NGF doses of 20 and 50 ng/mL (with the delivery system). For normal sciatic nerve, the fiber density is 11,882 fibers/mm2 (Mackinnon et al., 1991). Fiber densities of 25,382 and 28,934 fibers/mm2 are observed for cut and repair injuries at 1 and 3 months postoperatively, respectively (Mackinnon et al., 1991), which are similar to our NGF delivery system results at the midline of the conduit. This suggests that regeneration at the midline of the conduit is comparable to the cut and repair, but that regeneration in the distal segment of the conduit treatment groups lags that in the distal segment of the cut and repair treatment (due to the 13-mm length of the conduit gap). Images of the fibrin delivery system containing different NGF doses and of the isograft (positive) control are shown at the midline of the conduit in Fig. 5 and distal to the conduit in Fig. 6. Although there are a significant number of nerve fibers in the regenerating nerves, it is clear from the histological sections that the regenerating fibers have not fully matured and that further time will be required before functional recovery can occur. Regeneration has clearly progressed to a greater extent at the midline of the conduit than in the distal nerve. We selected the 6-week time point because, although axons have started to bridge the nerve gap, we can still see the effects of therapeutics on the time course of regeneration by looking at the maturity (diameter) of the nerve fibers. The fibrin matrix provides a biodegradable scaffold for fiber outgrowth in this nerve regeneration model, as well as the base for our drug delivery system. Based on our histology (Fig. 5), most if not all of the fibrin matrix has been resorbed by 6 weeks post implantation and has been replaced by the regenerating nerve cable. This is ideal because we do not want the fibrin matrix to inhibit the progress of regenerating fibers or to compress the nerve cable. The objective of this research was to develop and test a novel fibrin-based delivery system engineered to deliver biologically active growth factors in vivo. We have previously demonstrated the ability of such delivery systems to
Fig. 6. Histological sections of regenerating nerves at the distal end of the conduit (or graft). (A) 5 ng/mL NGF ⫹ delivery system (DS); (B) 20 ng/mL NGF ⫹ delivery system; (C) 50 ng/mL NGF ⫹ delivery system; (D) isograft (positive) control; (E) empty conduit; (F) fibrin alone; (G) Delivery system ⫹ fibrin (no NGF); and (H) NGF ⫹ fibrin (no delivery system). NGF, nerve growth factor.
302
A.C. Lee et al. / Experimental Neurology 184 (2003) 295–303
Table 2 Nerve histomorphometry dataa Group
5 ng/ml NGF 20 ng/ml NGF 50 ng/ml NGF Isograft Empty Fibrin Fibrin ⫹ DS NGF (no DS) a b
Midline of conduit
Distal to conduit
Number of nerve fibers
Percent nerve
Mean fiber diameter (m)
Density of nerve fibers
Number of nerve fibers
Percent nerve
Mean fiber diameter (m)
Density of nerve fibers
833 ⫾ 1518b 4527 ⫾ 4964 6462 ⫾ 4177 9920 ⫾ 4177 0 ⫾ 0b 1347 ⫾ 2343b 3179 ⫾ 4361b 4636 ⫾ 5654b
4.4 ⫾ 12.5 12.7 ⫾ 1.69 25.1 ⫾ 10 17.5 ⫾ 11.2 1.5 ⫾ 3.3b 7.1 ⫾ 13.2 14.64 ⫾ 20.1 14.45 ⫾ 16.7
2.8 ⫾ 0.24 3.01 ⫾ 0.17 2.50 ⫾ 1.23 2.93 ⫾ 0.29 0.64 ⫾ 1.42b 0.91 ⫾ 1.41b 1.23 ⫾ 1.69b 1.4 ⫾ 1.54
16276 ⫾ 13565 28678 ⫾ 6132 27237 ⫾ 8000 19122 ⫾ 9713 1404 ⫾ 3138b 9074 ⫾ 15821 14948 ⫾ 19000 17832 ⫾ 17000
0 ⫾ 0b 2370 ⫾ 3081 4094 ⫾ 24969 5315 ⫾ 3431 75 ⫾ 105b 103 ⫾ 253b 1303 ⫾ 2913b 1069 ⫾ 2349b
0 ⫾ 0b 2.8 ⫾ 3.9 5.1 ⫾ 8.1 9.3 ⫾ 9.9 0.03 ⫾ 0.07b 0.11 ⫾ 0.27b 2.0 ⫾ 4.5b 1.6 ⫾ 3.6b
0 ⫾ 0b 1.6 ⫾ 0.11 1.71 ⫾ 0.15 2.82 ⫾ 0.30 0.53 ⫾ 1.19b 0.45 ⫾ 1.10b 0.56 ⫾ 1.25b 0.81 ⫾ 1.26b
0 ⫾ 0b 7221 ⫾ 7361 10389 ⫾ 10669 10731 ⫾ 11112 80 ⫾ 110b 142 ⫾ 3478b 2453 ⫾ 5485 3055 ⫾ 7134
NGF, nerve growth factor; DS, delivery system. Denotes P ⬍ 0.05 versus the positive isograft control.
promote neurite extension in a dose-dependent manner in vitro. However, the biological environment surrounding nerve regeneration in vivo is much more complicated. Therefore, we sought to determine whether the materials we have previously tested in vitro could be successfully applied in vivo. The results of this study demonstrate that the incorporation of a novel delivery system providing controlled release of growth factors promotes peripheral nerve regeneration. Results in all morphometric measurements demonstrate that an optimal NGF dose, when combined with the delivery system, was not statistically different from the positive isograft control that provides the clinical “gold standard” for long gap peripheral nerve repair. Furthermore, the regeneration is comparable in terms of fiber diameter (maturity) and fiber density (at the conduit midline) to that observed in end-to end (short gap) nerve repairs. These results are rather remarkable considering that the conduit gap is 13 mm in length, significantly larger than the critical defect length (10 mm) for rat sciatic nerve and that the conduit lacks both the cellular components of nerve grafts and the basal lamina, both of which are hypothesized to facilitate regeneration in nerve grafts. Previous reports have demonstrated (Aebischer et al., 1989; Doolabh et al., 1996; Francel et al., 1997; Frostick et al., 1998; Maeda et al., 1993; Terenghi, 1999) bridging of short gaps (10 mm) using conduits (Jenq and Coggeshall, 1987; Lundborg et al., 1982; Seckel et al., 1984); but, empty conduits and fibrin alone have not proven sufficient to bridge gaps larger than 10 mm in a rat sciatic nerve. Clinically the most critical need is the treatment of nerve gaps greater than 10 mm to provide an alternative to nerve autografts. The materials investigated in this study may be useful in promoting peripheral nerve regeneration in vivo following transection due to trauma or surgery. This study also demonstrates the utility of a novel affinity-based delivery system for use in promoting tissue regeneration in vivo. This type of delivery system is extremely versatile and adaptable. It
can be modified to contain exogenous cell adhesion domains as well as drug delivery systems and has the potential for application in a variety of tissue-engineered constructs other than nerve, including bone, skin, vascular tissue, and cartilage.
References Aebischer, P., Salessiotis, A.N., Winn, S.R., 1989. Basic fibroblast growth factor released from synthetic guidance channels facilitates peripheral nerve regeneration across long nerve gaps. J. Neurosci. Res. 23, 232– 289. Archibald, S.J., Krarup, C., Shefner, J., Li, S.T., Madison, R.D., 1991. A collagen-based nerve guide conduit for peripheral nerve repair: an electrophysiological study of nerve regeneration in rodents and nonhuman primates. J. Comp. Neurol. 306, 685– 696. Beazley, W.C., Milek, M.A., Reiss, B.H., 1984. Results of nerve grafting in severe soft tissue injuries. Clin. Orthop. 208, 208 –212. Borkenhagen, M., 1997. Effect of NT-3 and BDNF released from nerve guidance channels on dorsal root regeneration, EPFL, Lausanne, Switzerland. Borkenhagen, M., Stoll, R.C., Neuenschwander, P., Suter, U.W., Aebischer, P., 1998. In vivo performance of a new biodegradable polyester urethane system used as a nerve guidance channel. Biomaterials 19, 2155–2165. Dellon, A.L., Mackinnon, S.E., 1988. An alternative to the classical nerve graft for the management of the short nerve gap. Plast. Reconstr. Surg. 82, 849 – 856. den Dunnen, W.F., van der Lei, B., Robinson, P.H., Holwerda, A., Pennings, A.J., Schakenraad, J.M., 1995. Biological performance of a degradable poly(lactic acid-epsilon-caprolactone) nerve guide: influence of tube dimensions. J. Biomed. Mater. Res. 29, 757–766. Doolabh, V.B., Hertl, M.C., Mackinnon, S.E., 1996. The role of conduits in nerve repair: a review. Rev. Neurosci. 7, 47– 84. Edelman, E., Mathiowitz, E., Langer, R., Klagsbrun, M., 1991. Controlled and modulated release of basic fibroblast growth factor. Biomaterials 12, 612– 626. Evans, G.R., Brandt, K., Widmer, M.S., Lu, L., Meszlenyi, R.K., Gupta, P.K., Mikos, A.G., Hodges, J., Williams, J., Gurlek, A., Nabawi, A., Lohman, R., Patrick Jr., C.W., 1999. In vivo evaluation of poly(L-lactic acid) porous conduits for peripheral nerve regeneration. Biomaterials 20, 1109 –1115.
A.C. Lee et al. / Experimental Neurology 184 (2003) 295–303 Fields, R.D., Le Beau, J.M., Longo, F.M., Ellisman, M.H., 1989. Nerve regeneration through artificial tubular implants. Prog. Neurobiol. 33, 87–134. Fox, D.J., Doolabh, V.B., Mackinnon, S.E., Genden, E.M., Hunter, D.A., 1999. Decreased cyclosporin A requirement with anti-ICAM-1 and anti LFA-1a in a peripheral nerve allotransplantation model. Restor. Neurol. Neurosci. 15, 319 –326. Francel, P.C., Francel, T.J., Mackinnon, S.E., Hertl, C., 1997. Enhancing nerve regeneration across a silicone tube conduit by using interposed short-segment nerve grafts. J. Neurosurg. 87, 887– 892. Frostick, S.P., Yin, Q., Kemp, G.J., 1998. Schwann cells, neurotrophic factors, and peripheral nerve regeneration. Microsurgery 18, 397– 405. Jenq, C.B., Coggeshall, R.E., 1987. Permeable tubes increase the length of the gap that regenerating axons can span. Brain Res. 408, 239 –242. Laham, R., Sellke, F., Edelman, E., Pearlman, J., Ware, J., Brown, D., Gold, J., Simons, M., 1999. Local perivascular delivery of basic fibroblast growth factor in patients undergoing coronary bypass surgery. Circulation 100, 1865–1871. Lee, M., Doolabh, V.B., Mackinnon, S.E., Jost, S., 2000. FK506 promotes functional recovery in crushed rat sciatic nerve. Muscle Nerve 23, 633– 640. Lundborg, G., 1990. Nerve regeneration problems in a clinical perspective. Restor. Neurol. Neurosci. 1, 297–302. Lundborg, G., Dahlin, L.B., Danielsen, N., Gelberman, R.H., Longo, F.M., Powell, H.C., Varon, S., 1982. Nerve regeneration in silicone chambers: influence of gap length and of distal stump components. Exp. Neurol. 76, 361–375. Mackinnon, S.E., 1989. Surgical management of the peripheral nerve gap. Clin. Plast. Surg. 16, 587– 603. Mackinnon, S.E., Dellon, A.L., 1990. Clinical nerve reconstruction with a bioabsorbable polyglycolic acid tube. Plast. Reconstr. Surg. 85, 419 – 424.
303
Mackinnon, S.E., Dellon, A.L., O’Brien, J.P., 1991. Changes in nerve fiber numbers distal to a nerve repair in the rat sciatic nerve model. Muscle Nerve 14, 1116 –1122. Maeda, T., Mackinnon, S.E., Best, T.J., Evans, P.J., Hunter, D.A., Midha, R.T., 1993. Regeneration across “stepping-stone” nerve grafts. Brain Res. 618, 196 –202. Midha, R., Mackinnon, S.E., Becker, L.E., 1994. The fate of Schwann cells in peripheral nerve allografts. J. Neuropathol. Exp. Neurol. 53, 316–322. Midha, R., Munro, C.A., Mackinnon, S.E., Ang, L.C., 1997. Motor and sensory specificity of host nerve axons influence nerve allograft rejection. J. Neuropathol. Exp. Neurol. 56, 421– 434. Otto, D., Unsicker, K., Grothe, C., 1987. Pharmacological effects of nerve growth factor and fibroblast growth factor applied to the transectioned sciatic nerve on neuron death in adult rat dorsal root ganglia. Neurosci. Lett. 83, 156 –160. Rich, R., Alexander, T., Pryor, J., Hollowell, J., 1989. Nerve growth factor enhances regeneration through silicone chambers. Exp. Neurol. 105, 162–170. Sakiyama-Elbert, S., Hubbell, J., 2000a. Controlled release of nerve growth factor from a heparin-containing fibrin-based cell ingrowth matrix. J. Controlled Rel. 69, 149 –158. Sakiyama-Elbert, S., Hubbell, J., 2000b. Development of fibrin derivatives for controlled release of heparin-binding growth factors. J. Controlled Rel. 65, 389 – 402. Seckel, B.R., 1990. Enhancement of peripheral nerve regeneration. Muscle Nerve 13, 785– 800. Seckel, B.R., Chiu, T.H., Nyilas, E., Sidman, R.L., 1984. Nerve regeneration through synthetic biodegradable nerve guides: regulation by the target organ. Plast. Reconstr. Surg. 74, 173–181. Terenghi, G., 1999. Peripheral nerve regeneration and neurotrophic factors. J. Anat. 194 (Pt 1), 1–14. Varon, S., Conner, J.M., 1994. Nerve growth factor in CNS repair. J. Neurotrauma 11, 473– 486.