- Email: [email protected]
Peripheral nerve regeneration using a microporous polylactic acid asymmetric conduit in a rabbit long-gap sciatic nerve transection model
Biomaterials 32 (2011) 3764e3775
Contents lists available at ScienceDirect
Biomaterials journal homepage: www.elsevier.com/locate/biomaterials
Peripheral nerve regeneration using a microporous polylactic acid asymmetric conduit in a rabbit long-gap sciatic nerve transection model Shan-hui Hsu a, b, *, Shan-Ho Chan c, d, Chih-Ming Chiang d, Clayton Chi-Chang Chen d, Ching-Fen Jiang e a
Institute of Polymer Science and Engineering, National Taiwan University, Taipei, Taiwan Rehabilitation Engineering Research Center, National Taiwan University, Taipei, Taiwan c Department of Chemical Engineering, National Chung Hsing University, Taichung, Taiwan d Department of Radiology, Taichung Veterans General Hospital, Taichung, Taiwan e Department of Biomedical Engineering, I-Shou University, Kaohsiung, Taiwan b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 30 December 2010 Accepted 26 January 2011 Available online 10 March 2011
The performance of an asymmetric conduit made of microporous polylactic acid (PLA) in promoting the long-term peripheral nerve regeneration across a 20-mm-long sciatic nerve gap was evaluated by a rabbit sciatic nerve transection model. Magnetic resonance imaging (MRI) was employed to monitor the nerve regeneration process. The extents of nerve regeneration and conduit degradation were quantified by image analysis. Functional and histological analyses were followed to assess nerve reinnervation. MR images showed that the transected nerve was connected at about 4 months. The diameter of the regenerated nerve continued to increase while the conduit was gradually degraded. The conduit was completely degraded in 18 months. The degradation kinetics in vivo was estimated based on MR images. The functional recovery after 18 months was w82% based on electrophysiology. The extension range of the operated limb was slowly recuperated to w81% at 18 months. Histology showed that nerve bundles were self-assembled after 16e18 months, but the morphologies were still different from those of normal sciatic nerve. This was the first work on the long-term evaluation of peripheral nerve regeneration in a rabbit model, and the first to report the use of MRI to obtain the real-time images of regenerated nerve in a biomaterial conduit as well as to define the degradation rate of the conduit in vivo. The platform established in this study serves to evaluate the regeneration of larger-diameter (>3-mm) nerve across a long-gap bridged by a conduit. Ó 2011 Elsevier Ltd. All rights reserved.
Keywords: Nerve regeneration Asymmetric nerve conduit Magnetic resonance imaging (MRI) Functional recovery
1. Introduction Peripheral nerve injuries are frequently caused by trauma and may lead to a significant loss of sensory or motor functions. Wallerian degeneration occurs at the distal stump of the axonal injury site within 24 h of injury. The proximal end of the nerve fiber sends out sprouts toward the distal end [1e3]. If the gap defect is too wide, surgical treatments such as transplantation of an autologous nerve graft (“autograft”) can help to guide the direction of axon from the proximal to the distal ends [4,5]. Mackinnon and Dellon proposed the use of synthetic tube made of biodegradable polymers as the nerve guide (“nerve conduit”) [6]. The nerve conduit can provide an unperturbed environment for tissue repair. A fibrin
* Corresponding author. No. 1, Sec. 4 Roosevelt Road, Institute of Polymer Science and Engineering, National Taiwan University, Taipei 10617, Taiwan. Tel.: þ886 2 33665313; fax: þ886 4 22854734. E-mail address: [email protected] (S.-h. Hsu). 0142-9612/$ e see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.biomaterials.2011.01.065
bridge connecting the two stumps usually forms in one or two weeks, along which axons can migrate from one to the other end. The Schwann cells simultaneously participate in the remodeling of the endoneurial basal lamina [7]. Conduits made of different polymers have been developed to regenerate a nerve gap wider than 10 mm [8,9]. Especially, microporous polymer conduits with novel asymmetric membrane structure and permeability were found to enhance nerve regeneration significantly [10,11]. There are a few challenges encountered in studying the process of peripheral nerve reconstruction. The most popular animal model is the rat sciatic nerve transection, which provides a gap no more than 15 mm. Animal models for nerve regeneration across a large gap have been quite limited in literature [12]. In addition, few observations have been made on long-term recovery after an extensive period of time (e.g. over a year) [13e15]. Skepticism thus remains regarding if peripheral nerve regeneration observed in the rat sciatic nerve model is adaptable to that in human. Besides, the degradation rate is an important character of the nerve conduit and essential information for the product to be approved by the
S.-h. Hsu et al. / Biomaterials 32 (2011) 3764e3775
regulatory authorities, but little has been known with respect to the degradation rate of the conduit in vivo. Magnetic resonance imaging (MRI) is a method of providing detailed pictures of internal organs of human body without X-ray or other irradiation. In particular, MRI can visualize the brain, the spinal cord and nerves much more clearly than those by computed tomography [16,17]. MRI has been broadly applied in the diagnosis of neurological, musculoskeletal, cardiovascular and oncological diseases [18]. With the benefit from MRI, nerve regeneration in experimental animals can be monitored for a long period of time. We consider that MRI may be a potential tool for evaluation of the performance of tissue engineering scaffolds/conduits, as well as their degradation in vivo. However, to the best of our knowledge, there has been no report regarding the application of MRI in monitoring the peripheral nerve regeneration process in a biodegradable polymer conduit. A rabbit model of sciatic nerve transection was employed in this study. Regeneration of the larger-diameter (w3.4 mm) nerve across a 20 mm gap in an asymmetric poly(D,L-lactic acid) (PLA) conduit was evaluated. MRI was used to visualize the nerve regeneration during an 18-month period. The quantitative features of MRI could provide information regarding the dimension of the regenerated nerve and the degradation rate of the polymer conduit in vivo. Functional assessment and histological evaluation were used to verify the consistency of the results. 2. Materials and methods 2.1. Preparation of nerve conduits Microporous PLA conduits with novel asymmetric membrane structure and permeability were fabricated as previously described [10]. Briefly, 10% solution of PLA weight average molecular weight about 180 kDa (8300D, Cargill, USA) in 1,4dioxane was prepared and cast on a polydimethylsiloxane mold fit into glass dishes. The dishes were then placed in 40% alcohol that served as the nonsolvent. The substrates formed as a result of solvent-nonsolvent phase conversion. They were washed in water for 8 h and then dried in a 40 C oven for 24 h. The substrates were rolled to form nerve conduits and cut in 24 mm length. The average inner diameter of the conduits was 3.4 0.10 mm and the average wall thickness was 0.3 0.06 mm (Fig. 1A). The conduits were sterilized by 70% alcohol for 15 min and washed by phosphate buffered saline (PBS) for 5 min before surgery.
3765
2.2. Animal implantation 2.2.1. Animals and microsurgical procedures All animal procedures followed the ethical guidelines of the experimental animals and were approved by the Institutional Animal Care and Use Committee. A total number of twenty-six male adult New Zealand rabbits weighing 3000e3500 g were used. The rabbits were housed under temperature-controlled conditions at 22 3 C, with a cycle of 12-h light and darkness, and with free access to food and water. The animals were anesthetized and maintained with isoflurane (Halocarbon, USA) throughout the microsurgical procedures. The skin from clipped lateral thigh was scrubbed in a routine fashion with antiseptic solution. The incision extended from the greater trochanter to the mid-calf distally. A muscle splitting incision exposed the sciatic and posterior femur nerves. The sciatic nerve was originated from sacrum spine nerve creating an adequate distal segment. The conduit (w24 mm long) was placed into the nerve gap by 7-0 Nylon sutures. Nerve stumps at both ends were sutured into both ends of the conduit to a length of approximately 2 mm [19]. The nerve gap was thus 20 mm in length. Muscle and skin were closed with 4-0 Dexon sutures. The outcomes were evaluated by imaging techniques, functional assessment and histological analysis. Twenty-four animals were sacrificed before 18 months. Two spare animals were sacrificed after 2 years. 2.3. Imaging technique The MRI equipment used was a sonata 1.5 T system from Siemens (Germany). The image acquisition employed a 5-inch surface coil (small loop coil). The MRI protocol included the T1 weighted image (T1WI), T2 weighted image (T2WI), and T2 weighted short time of inversion recovery (STIR) [20]. T1WI for the anatomy of sciatic nerve (bright signal) was acquired inside the nerve conduit. The optimized parameters for T1WI in the current study were the repeat time (TR)/echo time (TE) 300/12 ms, field of view (FOV) 50e80 mm, and matrix pixel 256 256. T2WI was acquired to demonstrate the tissues (bright signals) near the region of the implanted conduit [21]. The optimized parameters for T2WI in the current study were TR/TE 3000e4000/100 ms, FOV 50e80 mm, and matrix pixel 256 256 [22]. T1WI and T2WI had different brightness and contrast. They were complementary to each other so better justification regarding soft tissues and non-soft tissues (e.g. nerve conduit) could be made. STIR did not provide images of better resolution in this case, and therefore was not presented in this study. MR images could be taken from the sagittal, axial, and coronal directions to show different views in vivo. The nerve and muscle signal ratios on each sequence were measured and processed with computers. The imaging parameters of MRI should be refined according to the different animals and experimental conditions. In addition to the qualitative feature of MRI mentioned previously, a computer program (MATLAB 7.6.0) was employed to convert the biomedical images into data regarding the volume of regenerated nerve and that of nerve conduits as described below. MRI images based on analog transmission and data traffic were converted into intelligent data. By the use of MATLAB in operation and multiplication of MR images, the DICOM file was changed into data. In accordance with the software function, the images of the interested regions were pricked. The whole construct was divided into a few regions where the volume of nerve conduits and that of regenerated nerve were calculated and then combined. The calculation of volume was based on T2WI. During the later period, the newly repaired regions of the regenerated nerve were expanded to occupy the areas of the degraded conduit in MR images. The pixels and matrix from interested regions were converted into computational data and functional values. Array 2D was used as the calculation method to compute the measurement of the area. Array 2D data were assembled to form 3D combinations. Finally, 3D images were assembled to form the relations between nerve volume (regenerated) and conduit volume (degraded). The volume of regenerated nerve obtained by this method was then normalized to the original volume of the initial sciatic nerve and represented by the percentage value. The volume of the remaining conduit was also normalized to the original volume of the initial nerve conduit and represented as percentage. 2.4. Functional assessment
Fig. 1. (A) The dimensions of the asymmetric PLA nerve conduits used in this study; and (B) the SEM images showing the microporosity in the cross-section of the asymmetric conduits. Scar bar ¼ 100 mm.
In the initial stage, rabbits limped with the operated legs. After five to eight days, the muscle of thigh and that of calf had obvious dystrophy and the animals acted on healthy extremity. No walking track analysis was available from literature for animals other than rats. For the reason, the differences in the muscle periphery of thigh and calf between the control (right) and the experimental (left) legs could be measured as a way to evaluate the functional recovery. The muscle periphery measurement was performed on rabbits before surgery and every two weeks after the implantation until no difference of muscle periphery between both legs was identify. Ten male rabbits were selected to measure the muscle periphery initially. Because some rabbits were sacrificed for the electrophysiological analysis and histology, after 16 months only six rabbits were used for data collection. Electrophysiological measurements were performed on the experimental animals before they were sacrificed for histological analysis. All animals were anesthetized. The amplifier used in the study had two channels and a ground line. The ground line was connected on the ear of the rabbit. Nerves were stimulated by a pair of needle electrodes. The needle electrode which was connected with a DC electrical
3766
S.-h. Hsu et al. / Biomaterials 32 (2011) 3764e3775
stimulator (PowerLab ML866, AD Instrument, Australia) was placed on the proximal end of the nerve. The action potential recorded from sciatic nerve in this study actually represented the aggregate electrical effect of individual action potentials generated by the electrical response of nerves to stimulation. The stimulation voltage
was generally from 1 to 10 mV and in 2 Hz pulse. The recording electrode had 10-s interval to avoid nerve fatigue and was avoided interference with radio wave. The pulses of muscle and nerve action potential were recorded by computer software (Scope for Windows, AD Instrument, Australia). Based on the nerve-to-nerve
Fig. 2. The exposed rabbit sciatic nerve (A) was transected and replaced by a nerve conduit (B). The abbreviation: P, the proximal end; D, the distal end. The anatomical position of the rabbit could be shown by the localizer image of MRI (C) to confirm the relative locations.Sciatic nerve was at 2-cm depth parallel to the side of the femur bone (D). After surgery, the nerve conduit was shown to be parallel to the side of the femur bone (E). Based on images of the early regeneration stage (<4 months, F, G and H), the regenerated nerve had either sprouted from the proximal end (F, 12 weeks) or sprouted from both ends (G, 10 weeks; H, 11 weeks; the sprouting rate from the proximal being faster). C, D, G, H: T1WI images; E, F: T2WI images.
S.-h. Hsu et al. / Biomaterials 32 (2011) 3764e3775 distances and time (from stimulation point to the maximum pulse amplitude), the nerve conduction velocity (NCV) for each group of conduits was evaluated. The peak amplitude was obtained from the muscle action potential (MAP). The dynamic motion behavior (jumping) of rabbits after nerve repair was continuously recorded as video and analyzed by a semi-automation program developed by us using MATLAB 7.4.0. In this analysis, the video with the mpg/mpeg format was converted into a sequence of static images to record the data (with the * frame rate of 40 frames per second). The vector a was defined by manually selecting two points horizontal to the ground in the first frame. If the video was taken steadily, * then the direction of vector a would be consistent in the rest of the frames. Another * vector b was defined as the central axis of the lower limb. The direction of * the vector b varied in accordance to the swing of the lower limb. The angle q * * between the two vectors a and b could be calculated according to equation
3767
** * * q ¼ cos1 ð a: bÞ=ðj aj:j bjÞ; 0+ q 180+ . A map indicating the variation of q versus time in terms of frame numbers could be used to evaluate the function of the lower limb. A complete swing cycle was given to show the changes of q corresponding to the movement of the lower limb. The extension range of the lower limb was denoted as Q and defined as the absolute value of the gap between the maximum and the minimum of the q during a swing cycle. Rabbit activities quantified by Q may be used to evaluate the degree of recovery of the limb. L Q values are supposed to indicate better and refined recovery. 2.5. Histological analysis After a period of time (4, 6, 8, 12, 16 and 18 months), four animals were euthanatized by CO2 overdose treatment for histological observation. The implanted
Fig. 3. The harvested conduits from sacrificed animals after 4 months (A), 12 months (C) and 18 months (E) and the corresponding sagittal MR images (B, D, F) before sacrifice are shown. The long black arrows in (B, D, F) indicated the regenerated nerve. The short black arrows in (B, D, F) indicated the conduit. The cross-section of the regenerated nerve became larger while the conduit was slowly degraded. The conduit in (F) was almost completely degraded. All MR images were T1WI images.
3768
S.-h. Hsu et al. / Biomaterials 32 (2011) 3764e3775
conduits were harvested and fixed in cold-buffered 4% glutaraldehyde solution immediately. After 48 h, the nerve conduits were cut open longitudinally. If the nerve gap was successfully connected, the regenerated nerve was in the form of a white string-like substance that connected the two ends of the initial gap. The specimens were washed in PBS several times and then transected into the proximal, medium and distal segments. The central portion of the medium segments (midconduit) was a completely regenerated region and was often used for comparison. All samples were postfixed in 1% osmium tetroxide (5 ml, Polysciences, USA), dehydrated in a graded series of ethanol solutions, and then embedded. The embedded samples were cut with 5 mm thickness, and stained with H&E, which did not stain the conduit (PLA). All nerve sections were observed under an optical microscope (Nikon Labophot, Japan), and photographs were taken using a digital camera (Nikon H666L, Japan).
2.6. Statistical analysis Twenty-six rabbits underwent the nerve conduit implantation.Five rabbits were evaluated by the imaging technique to trace the nerve regeneration every two weeks from 0 to 16 weeks, every four weeks after 16 weeks and every six to eight weeks after 64 weeks. Typical images were shown. The average diameters of the regenerated nerve based on the axial MR images at the midconduit as well as the computed volumes of the regenerated nerve and the degraded conduits were the mean values from five rabbits. Dynamic motion behavior was estimated at 4, 8, 12, 16 and 18 months (n ¼ 5 for each group) and the normal side was the control group. Electrophysiological measurement and histological analysis were analyzed at 8, 12 and 18 months (n ¼ 4). The results from multiple samples were expressed as mean standard deviation. Statistical differences were analyzed by one-way analysis
Fig. 4. The coronal views (A, C, E) and the axial views (B, D, F) of the rabbits receiving conduits for 4 months (A, B), 6 months (C, D) and 8 months (E, F). The long white arrows in images indicated the regenerated nerve and the short white arrows in images were nerve conduits (A, C, E). These images were clearer than those in Fig. 3 and could be used for quantification purposes. After 8 months, the regenerated nerve and the surrounding tissue became next to each other and less distinguishable in the image of axial view as a result of conduit degradation. All images were T1WI images.
S.-h. Hsu et al. / Biomaterials 32 (2011) 3764e3775 of variance (ANOVA). p < 0.05 was considered as statistically significant. After sacrificing (at 4, 6, 8, 12, 16 and 18 months, n ¼ 4 for each group), the nerve sections were retrieved for histological process. Typical histological data were shown.
3. Results The dimension of the PLA conduits used in the study is shown in Fig. 1A The cross-section of each conduit showed the typical asymmetric microporous structure, as illustrated in Fig. 1B During the surgery, the rabbit sciatic nerve (Fig. 2A) was replaced by the nerve conduit (Fig. 2B). The anatomical position shown in Fig. 2C was used to locate the sciatic nerve (2.0 cm below femur bone). Normal sciatic nerve in rabbits had a smooth linear structure with isointense MR images (Fig. 2D). Right after surgery, the nerve conduit was shown to be parallel to the side of the femur bone (Fig. 2E). The T2WI image (Fig. 2E) showed the black lumen of the conduit. Images taken at the early regeneration stage (<4 months) showed that in some rabbits the regenerated nerve sprouted from the proximal end (Fig. 2F, as commonly perceived. In other rabbits (two out of five, Fig. 2G,H)), the regenerated nerve sprouted from both ends. In the latter case, the rate of nerve sprouting from the proximal was much faster than that from the distal. All conduits maintained a stable supporting structure during the first four months. The newly regenerated nerve in the conduit was completely connected after about four months (Fig. 3, sagittal view at 4, 12, and 18 months). All subjects had successful connection. The gross appearance of the harvested nerve conduits showed that the regenerated nerve inside the nerve conduits gradually became thicker and angiogenesis was obvious as red color in the surrounding tissue (Fig. 3A,C,E). T1WI images obtained after 4 months showed the regenerated sciatic nerve (Fig. 3B,D,F). The bright signal of the regenerated area inside the tube was fluid component and those outside the tube were the neighboring blood vessels as indicated by the thin white arrows in Fig. 3D,F. During this period, degradation of the nerve conduits was visible from the MR images. The dark images around the conduits (indicated by the short black arrows in Fig. 3B,D,F) were the signals of nerve conduits.
3769
The regenerated nerve and repaired tissue replaced the spaces of the degraded conduits and had isointense images with soft tissue. The dark image lines of the nerve conduit in the early stage (as indicated by the short arrows in Fig. 3B) were blocked and replaced by the new regenerated nerve in the later stage. The size (diameter) of the regenerated nerve increased with time (from Fig. 3B to F). Images of the coronal and the axial views (4, 6, and 8 months) are shown in Fig. 4. During this period, the conduit and the regenerated nerve could be distinguished by the axial view of both T1WI images and T2WI images taken at the middle section of the conduit. The dark signals inside the ring area were the images of regenerated nerve, which became thicker gradually. The white signals of the ring were the fluid components inside the degraded conduits, which became narrower during this period. The diameter of the regenerated nerve at the midconduit quantified by the axial MR images was about 0.7 0.08 mm at 4 months (Fig. 4B), 1.8 0.13 mm at 6 months (Fig. 4D), and 2.8 0.17 mm at 8 months (Fig. 4F). The three-dimensional volume of the regenerated nerve and that of the nerve conduit was obtained from computation of the MR images. The tendency of changes for the volumes of the nerve conduit and the regenerated nerve along the implantation period is shown in Fig. 5. The changes of volumes for the nerve conduit and for the regenerated nerve were in the opposite trend. The regenerated nerve increased in diameter when the degradation of polymer continued in vivo. The two tendency curves crossed over each other at about 16e20 weeks, which was approximately when the nerve first connected. The nerve conduits were mostly (w95%) degraded after 52 weeks, and were almost completely (w99%) degraded after 18 months. Based on MR image analysis, the initial three-dimensional volume (V0) of the PLA nerve conduit in situ right after implantation was about 83.7 0.84 mm3. This value was in good agreement with that calculated based on the geometry and dimension shown in Fig.1A. The volume of the nerve conduit in vivo at each time point (V) and the original volume V0 were used to estimate the degradation kinetics by assuming that the volume was proportional to the amount of the reactant. The relationship of V0 vs. time appeared to be an exponential decay, as shown in Fig. 6A. The natural log of V0/V
Fig. 5. The volume change of the nerve conduit and that of the regenerated nerve during the period of 18 months. These data were obtained by MATLAB software based on MR images. The percent volume of nerve conduit was calculated based on the original volume of the nerve conduit. The percentage of nerve regeneration (including sprouted nerve before connection) was calculated based on the original volume of normal sciatic nerve. Nerve regeneration was accompanied by the conduit degradation and after 18 months the conduit was almost completely degraded.
3770
S.-h. Hsu et al. / Biomaterials 32 (2011) 3764e3775
Methods. The map of the variation of q versus time in terms of frame numbers as well as the maximum and minimum angles were indicated from the video image frames (Fig. 8B). The extension range of the lower limb (denoted as Q) was defined as the difference between the maximum and the minimum q during a swing cycle. Based on the map, Q was 29 2 at 4 months (Fig. 8C) and 68 3 for the control limb (Fig. 8D). The values of Q obtained at different periods after implantation are shown in Fig. 8E. The values of Q increased steadily from 4 to 18 months. This indicated that the motion function was slowly recovered from the nerve injury. The functional recovery was about 81% of that of the control after 18 months, based on the motion behavior analysis.
Fig. 6. (A) The exponential decay of the volume (V) of the nerve conduit. (B) The degradation kinetics of the conduit, showing the approximately linear relationship between natural log of V0/V and time.
vs. time was plotted in Fig. 6B, and was approximately linear. This suggested the pseudo first-order kinetics for the degradation of the PLA conduit. The slope of the linear curve reflected the rate constant, which was 0.0706 week1 or 0.0101 day1. The muscle peripheries of thigh and calf were measured and are shown in Fig. 7A. In the initial stage of injury, rabbits limped with the operated legs and had obvious muscular dystrophy. The decrease of muscle peripheries was an indirect proof of the early nerve injury in the rabbits. Both thigh muscle and calf muscle had smaller periphery that the original values before 36 weeks. The muscle peripheries were the lowest at 16e18 months and were slowly recovered to the original values. After 36 weeks, the muscular peripheries remained similar. The electrophysiological analysis was conducted at 8, 12, and 18 months before the animals were sacrificed. The nerve conduction velocity (NCV) was calculated based on NAP and the peak amplitude was obtained from MAP. The values of NCV at different time periods are shown in Fig. 7B. The NCV of the normal sciatic nerve was about 64.3 3 m/s, which was significantly higher (p < 0.05) than that of the regenerated nerve. The percent recovery of the rabbits based on NCV was about 62% at 8 months, 64% at 12 months, and 82% at 18 months. The peak amplitude for the normal side muscle was 0.0036 0.0001 mV, which was significantly higher (p < 0.05) than that of the experimental side (Fig. 7C). The recovery of the rabbits based on peak amplitude was about 58%, 69%, and 81% at 8 months,12 months, and 18 months respectively. Based on the electrophysiology, the percent functional recovery after 18 months had reached 80%. The jumping motion of the experimental rabbits was evaluated by a novel method involving the measurement of motion angle and extension range for the lower limb, as shown in Fig. 8. The motion * * angle q was defined from two vectors a and b obtained from the lower limb of each rabbit (Fig. 8A), as detailed in the Materials and
Fig. 7. (A) The changes in muscle periphery of lower limbs in experimental rabbits receiving conduits. At early implantation (before 36 weeks), the peripheries of the thigh and calf were obviously smaller (muscular dystrophy). After 36 weeks later, the peripheries of the thigh and calf were recovered to nearly normal so the measurement was done less frequently. (B, C) Electrophysiological data in the experimental rabbits receiving conduits recorded after 8, 12 and 18 months. The NCV values in (B) were obtained from the compound nerve action potential (NAP). The peak amplitude values in (C) were obtained from the compound muscle action potential (MAP).
S.-h. Hsu et al. / Biomaterials 32 (2011) 3764e3775
Histology of the nerve sections at the midconduit is shown in Fig. 9. All 24 subjects had successful connection (i.e. success rate ¼ 100%). Normal sciatic nerve of the rabbit (Fig. 9Aa) was compared with that of the regenerated nerve at 4 months (Fig. 9Ab), 6 months (Fig. 9Ac), 8 months (Fig. 9Ad), 12 months (Fig. 9Ae), 16 months (Fig. 9Af) and 18 months (Fig. 9Ag, h). The
3771
presence of myelins (M), blood vessels (V), regenerated nerves (N), connective tissues (CT), or fats (F) was indicated in the repaired tissues. At 4 months, the newly regenerated nerves showed two nerve bundles (Fig. 9Ab). At 6 months, the newly regenerated nerves had formed several nerve bundles (Fig. 9Ac). From 6 to 12 months, the bundles continued to expand their size (Fig. 9Acee).
.
.
Fig. 8. The dynamic behavior analysis for rabbits showing the extension range of the lower limbs (Q). The motion angle (q) was defined as the angle between the vectors a and b (panel A). During a swing (panel B; from (1), (2) to (3)), the motion angle was constantly recorded. The data curves of motion angle (q) after 4 months (C) and the control (normal side, D). The calculated data during 4e18 months were compared with that of the normal side (E).
3772
S.-h. Hsu et al. / Biomaterials 32 (2011) 3764e3775
Inside each bundle, the number of myelinated axons also increased. Before 12 months, the nerve bundles were in irregular shape (Fig. 9Abee). The structures of blood vessels, nerve fibers, myelins, and their distributions in newly regenerated nerves were different from those in the normal sciatic nerve. Blood vessel formation was found both inside the nerve bundles and outside the nerves near the connective tissues. After 16 months, the independent nerve bundles had self-organized into large nerve bundles (Fig. 9Af, h). At 18 months, the nerve bundles were merged or combined with each other. This self-assembly was accompanied by a decrease in the number of blood vessels. At this time, the morphology of the regenerated nerve tissue finally became more homogeneous (Fig. 9Ah), which was more like that of the normal sciatic nerve (Fig. 9Aa). The diameter of the regenerated nerve at 18 months was slightly larger than that of the normal nerve. Histology at larger magnification (400) is shown in Fig. 9B. Nerve sections at different periods were compared with that of the normal sciatic nerve (Fig. 9Ba). Thinner myelins were observed in 4e12 months (Fig. 9Bjem). The self-assembly of myelins into large bundles was clearly visualized after 16e18 months and the margins were indicated by the long black arrows in Fig. 9Bnep. 4. Discussion The rat model of sciatic nerve injury was the most common animal model for studying peripheral nerve regeneration. However, the longest gap that could be created in rats was 15 mm [9]. Rabbits are larger in size than rats and are more cost-efficient compared to larger animals such as dogs [12,23] and cats [24]. Sciatic nerve regeneration in rabbits across a 10 mm gap bridged by an autograft or an epineural flap has been examined for a period of 3 months [25,26]. So far there is no literature regarding large-gap sciatic nerve regeneration bridged by a conduit in rabbits. There was a study on bridging the gap of rabbit common peroneal nerve (thinner than sciatic nerve) with poly-3-hydroxybutyrate conduits, but nerve regeneration was only evaluated by histology after 4 months [27]. Our present study described nerve regeneration across a 20 mm sciatic nerve gap in rabbits during a period of 18 months, which could provide better translations to long-term clinical applications. The adequate nerve regeneration and functional recovery may be attributed to the asymmetric structure of our PLA conduits. The commercial Neurotube made of polyglycolic acid did not supply tubes longer than 20 mm with diameter similar to our conduits. Therefore, a comparison could not be made. However, we discovered that asymmetric PLA conduits with additional surface microgrooves (width/spacing/depth ¼ 20 mm/20 mm/3 mm) [19] could advance the nerve connection time from four months to three months in a parallel study, suggesting that the rabbit sciatic nerve model and MRI monitoring presented here may be a good platform for studying peripheral nerve regeneration and may help correlating the biomaterials design and device performance. MRI successfully monitored the regeneration progresses of transected peripheral nerve in the biomaterial conduit in this study. In a recent report, the acute crush injury (created by clamps) in rabbit sciatic nerve was diagnosed by MRI; but nerve repair was not followed [20]. Based on the MR images in the present study, we identified that some of the transected nerve sprouted from both the proximal and distal ends. The proximal stump of the transected nerve regenerated faster than the distal stump and the two sprouts finally connected not in the center but about 3/4 distance from the proximal end after about 4 months. This was completely surprising because it was different from the textbook concept that the transected nerve sprouts from the proximal end and that the distal end undergoes Wallerian degradation. It could be argued that the sprouted “nerve” from the distal end we observed may have been
the image of the organized fibrin bridge and not the nerve itself. Nevertheless, if it were the fibrin bridge, it should have linked both ends. Another argument could be that this phenomenon may only occur in acute nerve injury, which we have not had any clue yet. We suggest that this observation may deserve further confirmation and in-depth study for better understanding of the nerve regeneration mechanism inside a polymer conduit. Material degradation is essential information required by all tissue engineering medical products; however, few studies have fully characterized the biodegradation on site in vivo. It was indicated that the degradation rate of PLA differed according to the site of implantation [28]. Based on the present study, our nerve conduits were gradually degraded during the 18-month period (50% degraded within the first 20 weeks and 95% degraded within the first year). PLA was reported to degrade partially (56% or more based on molecular weight change) after 6 months in vivo (subcutaneous mice) [29]. The degradation rate of our PLA conduits in the first 6 months based on a different definition (volume reduction) roughly agreed with that reported. The proper degradation rate allowed the conduit to provide a stable structural support for the newly regenerated nerve during the first 4e6 months. This may account for the good success rate and functional outcome of our conduits after 18 months. Due to the heterogeneous nature of the PLA degradation, it is usually difficult to obtain the information regarding the degradation in vivo. Most studies measured the decrease in molecular weight [28], but it was less likely to obtain the degradation kinetics by measuring the average molecular weight of the remaining device. The advantage of MRI was to quantify the intensity of the conduit. The degradation kinetics estimated from the MR images indicated a pseudo first-order reaction. PLA was reported to undergo selfcatalysis in vitro [30]. The pseudo first-order kinetics of PLA was normally observed when the concentration of catalyst HCl was kept constant [30]. Therefore, the degradation information obtained from MR images suggested that in vivo the acid concentration near the conduit was relatively constant. PLA is not soluble in water unless dissociation of the acid end-group, which is expected to result in an acidic environment and a significant contribution to the hydrolysis. The degradation kinetics found in this study suggested a fast drainage of the acid generated near the conduit, which could be attributed to the special asymmetric design in the structure of the conduit. Results from the dynamic behavior analysis showed the rehabilitation of sciatic nerve in rabbits (81% after 18 months). Few papers have mentioned the analytical method for evaluation of the functional recovery in rabbits. The sciatic functional index (SFI) or dynamic cat walks (DCW) for rats are not applicable for rabbits because of the unique rabbit motivation (jumping, not walking) and size (much larger than rats) [31].The analysis of extension range is similar to SFI or DCW in a way that it analyzes the spontaneous coordinate activity. The percent functional recovery based on the extension range was very consistent with that based on the peak amplitude of compound muscle action potential obtained from electrophysiology. Therefore, the behavior analysis developed in this study was a novel and effective method to analyze the functional recovery during the period of nerve regeneration in the experimental animals. In sciatic nerve crushed rabbits, the myelin formation was immature and different from that of the normal sciatic nerve [20]. This was also observed during the early stage of nerve regeneration in our study. In the current model, nerves were connected at about 4 months, but the process of nerve bundle reorganization took much longer (more than a year). In the rat model of 10 mm sciatic nerve transection, these phenomena (including increase and decrease in the vessel number) were observed in 4e6 weeks [19]. The ligated sciatic nerves in rats recovered even faster and had the
S.-h. Hsu et al. / Biomaterials 32 (2011) 3764e3775
3773
Fig. 9. (A) Histology of the nerve sections at the midconduit after 4 months (b), 6 months (c), 8 months (d), 12 months (e), 16 months (f) and 18 months (g, h) of implantation, compared with that of the normal sciatic nerve (a). The abbreviation M: myelin, V: blood vessel, N: regenerated nerve, F: fat tissue and C: conduit. Scar bar ¼ 200 mm. Small nerve bundles inside the conduits were observed after 4e12 months (bee). Small bundles were self-assembled into large bundles after 16e18 months (with the margins indicated by the long black arrows in feh). (B) Images were displayed at larger magnification (400) for normal sciatic nerve (i), and for regenerated nerve after 4 months (j), 6 months (k), 8 months (l), 12 months (m), 16 months (n) and 18 months (o, p) of implantation.Scar bar ¼ 50 mm. Thinner myelins were observed within a year (jem). The self-assembly occurred after 16e18 months (nep). The margin of the self-assembled area was marked by the long black arrows (nep).
same histology as normal nerves after 12 weeks [32]. In our study, the evolutions of morphology for the regenerated nerves were consistent with the change in functional recovery demonstrated by the electrophysiology and behavior analysis. For minor injury in literature, the muscular function began to recover in some rabbits at six weeks following the crush of sciatic nerve [20]. Obviously it
took much more time for the large-gap sciatic nerve defected rabbits to recover from the injury, even in the presence of a protecting and guiding biomaterial conduit. On the other hand, since none of the rabbits died during the experimental period, and the average data collected from the two spare rabbits before their sacrifice at 24 months showed the promising NCV value at 61.5 m/s
3774
S.-h. Hsu et al. / Biomaterials 32 (2011) 3764e3775
Fig. 9. (continued).
(98% recovery) and extension range (Q) at 62.0 (91% recovery), we considered that the asymmetric PLA conduit may provide satisfactory long-term recovery for the injured peripheral nerve. 5. Conclusions The performance of an asymmetric PLA conduit could be evaluated by the rabbit sciatic nerve transection model during a much longer period compared with that observed in the rat model. The
PLA conduit was slowly degraded while the nerve regeneration proceeded. Some of the regenerated nerve sprouted from the proximal end, while others sprouted from both the proximal and distal stumps. In both cases, the nerve was connected after four months. After connection, the nerve continued to increase its diameter. The degradation of the conduit quantified by MR images showed pseudo first-order kinetics with a rate constant w0.01 day1. The functional recovery after 18 months was about 82% based on electrophysiology and was about 81% based on behavior
S.-h. Hsu et al. / Biomaterials 32 (2011) 3764e3775
analysis. Histology showed that the regenerated nerve inside the conduit was in the form of several small nerve bundles after six months and merged into a single large bundle after 18 months, which was close to the histology of normal sciatic nerve. By adjusting the parameters to display the nerve conduit and regenerated nerve, MRI could provide fine resolution to quantify the performance and the degradation of the nerve conduits in vivo. Acknowledgments The authors thank Dr. Su Jing Yang and the MRI center, Department of Radiology, Taichung Veterans General Hospital, for their group support. This work was supported by grants from the Taichung Veterans General. Hospital/National Chung Hsing University (grant number TCVGH-NCHU 977608) and the National Science Council (grant number NSC 99-2321-B-002-043). Appendix Figures with essential color discrimination. Certain figures in this article, particularly Figs. 2,3,8 and 9 are difficult to interpret in black and white. The full color images can be found in the online version, at doi:10.1016/j.biomaterials.2011.01.065. References [1] Faweett JW, Keynes RJ. Peripheral nerve regeneration. Annu Rev Neurosci 1990;13:43e60. [2] Vargas ME, Barres BA. Why is Wallerian degeneration in the CNS so slow? Annu Rev Neurosci 2007;30:153e79. [3] Coleman MP, Freeman MF. Wallerian degeneration, WldS, and Nmnat. Annu Rev Neurosci 2010;33:245e67. [4] Schmidt CE, Leach JB. Neural tissue engineering: strategies for repair and regeneration. Annu Rev Biomed Eng 2003;5:293e347. [5] Guertin AD, Zhang DP, Mak KS, Alberta JA, Kim HA. Microanatomy of axon/ glial signaling during Wallerian degeneration. J Neurosci 2005;25:3478e87. [6] Mackinnon SE, Dellon AL. A study of nerve regeneration across synthetic (Maxon) and biologic (collagen) nerve conduits for nerve gaps up to 5 cm in the primate. J Reconstr Microsurg 1990;6:117e21. [7] Cornbrooks CJ, Carey DJ, McDonald JA, Timpl R, Bunge RP. In vivo and in vitro observations on laminin production by Schwann cells. Proc Natl Acad Sci USA 1983;80:3850e4. [8] Chang JY, Lin JH, Yao CH, Chen JH, Lai TY, Chen YS. In vivo evaluation of a biodegradable EDC/NHS-cross-linked gelatin peripheral nerve guide conduit material. Macromol Biosci 2007;7:500e7. [9] Deumens R, Bozkurt A, Meek MF, Marcus MA, Joosten EA, Weis J, et al. Repairing injured peripheral nerves: bridging the gap. Prog Neurobiol 2010;92:245e76. [10] Chang CJ, Hsu SH. The effect of high outflow permeability in asymmetric poly (dl-lactic acid-co-glycolic acid) conduits for peripheral nerve regeneration. Biomaterials 2006;27:1035e42. [11] Oh SH, Kim JH, Song C, Jeon BH, Yoon JH, Seo TB, et al. Peripheral nerve regeneration within an asymmetrically porous PLGA/Pluronic F127 nerve guide conduit. Biomaterials 2008;29:1601e9.
3775
[12] Ichihara S, Inada Y, Nakada A, Endo K, Azuma T, Nakai R, et al. Development of new nerve guide tube for repair of long nerve defects. Tissue Eng Part C Methods 2009;15:387e402. [13] Meek MF, Den Dunnen WF, Schakenraad JM, Robinson PH. Long-term evaluation of functional nerve recovery after reconstruction with a thin-walled biodegradable poly (dl-lactide-epsilon-caprolactone) nerve guide, using walking track analysis and electrostimulation tests. Microsurgery 1999;19:247e53. [14] Ansselin AD, Fink T, Davey DF. Peripheral nerve regeneration through nerve guides seeded with adult Schwann cells. Neuropathol Appl Neurobiol 1997;23:387e98. [15] Kobayashi J, Mackinnon SE, Watanabe O, Ball DJ, Gu XM, Hunter DA, et al. The effect of duration of muscle denervation on functional recovery in the rat model. Muscle Nerve 1997;20:858e66. [16] Veraa S, Dijkman R, Meij BP, Voorhout G. Comparative imaging of spinal extradural lymphoma in a Bordeaux dog. Can Vet J 2010;51:519e21. [17] Bendszus M, Koltzenburg M, Wessig C, Solymosi L. Sequential MR imaging of denervated muscle: experimental study. AJNR Am J Neuroradiol 2002;23:1427e31. [18] Bendszus M, Wessig C, Reiners K, Bartsch AJ, Solymosi L, Koltzenberg M. MR imaging in the differential diagnosis of neurogenic foot drop. AJNR Am J Neuroradiol 2003;24:1273e4. [19] Hsu SH, Ni HC. Fabrication of the microgrooved/microporous polylactide substrates as peripheral nerve conduits and in vivo evaluation. Tissue Eng Part A 2009;15:1381e90. [20] Li X, Shen J, Chen J, Wang X, Liu Q, Liang B. Magnetic resonance imaging evaluation of acute crush injury of rabbit sciatic nerve: correlation with histology. Can Assoc Radiol J 2008;59:123e30. [21] Valentine HL, Does MD, Marshall V, Tonkin EG, Valentine WM. Multicomponent T2 analysis of dithiocarbamate-mediated peripheral nerve demyelination. Neurotoxicology 2007;28:645e54. [22] Stroman PW, Dorvil JC, Marois Y, Poddevin N, Guidoin R. In vivo time course studies of the tissue responses to resorbable polylactic acid implants by means of MRI. Magn Reson Med 1999;42:210e4. [23] Wang XD, Hu W, Cao Y, Yao J, Wu J, Gu XS. Dog sciatic nerve regeneration across a 30-mm defect bridged by a chitosan/PGA artificial nerve graft. Brain 2005;128:1897e910. [24] Suzuki Y, Tanihara M, Ohnishi K, Suzuki K, Endo K, Nishimura Y. Cat peripheral nerve regeneration across 50 mm gap repaired with a novel nerve guide composed of freeze-dried alginate gel. Neurosci Lett 1999;259:75e8. [25] Ignatiadis IA, Yiannakopoulos CK, Barbitsioti AD, Avram AM, Patralexis HG, Tsolakis CK, et al. Diverse types of epineural conduits for bridging short nerve defects. An experimental study in the rabbit. Microsurgery 2007;27:98e104. [26] Ignatiadis IA, Tsiampa VA, Yiannakopoulos CK, Xeinis SF, Papalois AE, Xenakis TH, et al. A new technique of autogenous conduits for bridging short nerve defects. An experimental study in the rabbit. Acta Neurochir Suppl 2007;100:73e6. [27] Young YC, Wiberg M, Terenghi G. Poly-3-hydroxybutyrate (PHB): a resorbable conduit for long-gap repair in peripheral nerves. Br J Plast Surg 2002;55:235e40. [28] Tschakaloff A, Losken HW, Oepen RV, Michaeli W, Moritz O, Mooney MP, et al. Degradation kinetics of biodegradable dl-polylactic acid biodegradable implants depending on the site of implantation. Int J Oral Maxillofac Surg 1994;23:443e5. [29] Gogolewski S, Jovanovic M, Perren SM, Dillon JG, Hughes MK. Tissue response and in vivo degradation of selected polyhydroxyacids: polylactides (PLA), poly (3-hydroxybutyrate) (PHB), and poly(3-hydroxybutyrate-co-3-hydroxyvalerate) (PHB/VA). J Biomed Mater Res 1993;27:1135e48. [30] Siparsky GL. Degradation kinetics of poly(hydroxy) acids: PLA and PCL. ACS Symp Ser 2001;764:230e51. [31] Hare GM, Evans PJ, Mackinnon SE, Best TJ, Bain JR, Szalai JP, et al. Walking track analysis: a long-term assessment of peripheral nerve recovery. Plast Reconstr Surg 1992;89:251e8. [32] Bendszus M, Wessig C, Solymosi L, Reiners K, Koltzenburg M. MRI of peripheral nerve degeneration and regeneration: correlation with electrophysiology and histology. Exp Neurol 2004;188:171e7.
Contents lists available at ScienceDirect
Biomaterials journal homepage: www.elsevier.com/locate/biomaterials
Peripheral nerve regeneration using a microporous polylactic acid asymmetric conduit in a rabbit long-gap sciatic nerve transection model Shan-hui Hsu a, b, *, Shan-Ho Chan c, d, Chih-Ming Chiang d, Clayton Chi-Chang Chen d, Ching-Fen Jiang e a
Institute of Polymer Science and Engineering, National Taiwan University, Taipei, Taiwan Rehabilitation Engineering Research Center, National Taiwan University, Taipei, Taiwan c Department of Chemical Engineering, National Chung Hsing University, Taichung, Taiwan d Department of Radiology, Taichung Veterans General Hospital, Taichung, Taiwan e Department of Biomedical Engineering, I-Shou University, Kaohsiung, Taiwan b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 30 December 2010 Accepted 26 January 2011 Available online 10 March 2011
The performance of an asymmetric conduit made of microporous polylactic acid (PLA) in promoting the long-term peripheral nerve regeneration across a 20-mm-long sciatic nerve gap was evaluated by a rabbit sciatic nerve transection model. Magnetic resonance imaging (MRI) was employed to monitor the nerve regeneration process. The extents of nerve regeneration and conduit degradation were quantified by image analysis. Functional and histological analyses were followed to assess nerve reinnervation. MR images showed that the transected nerve was connected at about 4 months. The diameter of the regenerated nerve continued to increase while the conduit was gradually degraded. The conduit was completely degraded in 18 months. The degradation kinetics in vivo was estimated based on MR images. The functional recovery after 18 months was w82% based on electrophysiology. The extension range of the operated limb was slowly recuperated to w81% at 18 months. Histology showed that nerve bundles were self-assembled after 16e18 months, but the morphologies were still different from those of normal sciatic nerve. This was the first work on the long-term evaluation of peripheral nerve regeneration in a rabbit model, and the first to report the use of MRI to obtain the real-time images of regenerated nerve in a biomaterial conduit as well as to define the degradation rate of the conduit in vivo. The platform established in this study serves to evaluate the regeneration of larger-diameter (>3-mm) nerve across a long-gap bridged by a conduit. Ó 2011 Elsevier Ltd. All rights reserved.
Keywords: Nerve regeneration Asymmetric nerve conduit Magnetic resonance imaging (MRI) Functional recovery
1. Introduction Peripheral nerve injuries are frequently caused by trauma and may lead to a significant loss of sensory or motor functions. Wallerian degeneration occurs at the distal stump of the axonal injury site within 24 h of injury. The proximal end of the nerve fiber sends out sprouts toward the distal end [1e3]. If the gap defect is too wide, surgical treatments such as transplantation of an autologous nerve graft (“autograft”) can help to guide the direction of axon from the proximal to the distal ends [4,5]. Mackinnon and Dellon proposed the use of synthetic tube made of biodegradable polymers as the nerve guide (“nerve conduit”) [6]. The nerve conduit can provide an unperturbed environment for tissue repair. A fibrin
* Corresponding author. No. 1, Sec. 4 Roosevelt Road, Institute of Polymer Science and Engineering, National Taiwan University, Taipei 10617, Taiwan. Tel.: þ886 2 33665313; fax: þ886 4 22854734. E-mail address: [email protected] (S.-h. Hsu). 0142-9612/$ e see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.biomaterials.2011.01.065
bridge connecting the two stumps usually forms in one or two weeks, along which axons can migrate from one to the other end. The Schwann cells simultaneously participate in the remodeling of the endoneurial basal lamina [7]. Conduits made of different polymers have been developed to regenerate a nerve gap wider than 10 mm [8,9]. Especially, microporous polymer conduits with novel asymmetric membrane structure and permeability were found to enhance nerve regeneration significantly [10,11]. There are a few challenges encountered in studying the process of peripheral nerve reconstruction. The most popular animal model is the rat sciatic nerve transection, which provides a gap no more than 15 mm. Animal models for nerve regeneration across a large gap have been quite limited in literature [12]. In addition, few observations have been made on long-term recovery after an extensive period of time (e.g. over a year) [13e15]. Skepticism thus remains regarding if peripheral nerve regeneration observed in the rat sciatic nerve model is adaptable to that in human. Besides, the degradation rate is an important character of the nerve conduit and essential information for the product to be approved by the
S.-h. Hsu et al. / Biomaterials 32 (2011) 3764e3775
regulatory authorities, but little has been known with respect to the degradation rate of the conduit in vivo. Magnetic resonance imaging (MRI) is a method of providing detailed pictures of internal organs of human body without X-ray or other irradiation. In particular, MRI can visualize the brain, the spinal cord and nerves much more clearly than those by computed tomography [16,17]. MRI has been broadly applied in the diagnosis of neurological, musculoskeletal, cardiovascular and oncological diseases [18]. With the benefit from MRI, nerve regeneration in experimental animals can be monitored for a long period of time. We consider that MRI may be a potential tool for evaluation of the performance of tissue engineering scaffolds/conduits, as well as their degradation in vivo. However, to the best of our knowledge, there has been no report regarding the application of MRI in monitoring the peripheral nerve regeneration process in a biodegradable polymer conduit. A rabbit model of sciatic nerve transection was employed in this study. Regeneration of the larger-diameter (w3.4 mm) nerve across a 20 mm gap in an asymmetric poly(D,L-lactic acid) (PLA) conduit was evaluated. MRI was used to visualize the nerve regeneration during an 18-month period. The quantitative features of MRI could provide information regarding the dimension of the regenerated nerve and the degradation rate of the polymer conduit in vivo. Functional assessment and histological evaluation were used to verify the consistency of the results. 2. Materials and methods 2.1. Preparation of nerve conduits Microporous PLA conduits with novel asymmetric membrane structure and permeability were fabricated as previously described [10]. Briefly, 10% solution of PLA weight average molecular weight about 180 kDa (8300D, Cargill, USA) in 1,4dioxane was prepared and cast on a polydimethylsiloxane mold fit into glass dishes. The dishes were then placed in 40% alcohol that served as the nonsolvent. The substrates formed as a result of solvent-nonsolvent phase conversion. They were washed in water for 8 h and then dried in a 40 C oven for 24 h. The substrates were rolled to form nerve conduits and cut in 24 mm length. The average inner diameter of the conduits was 3.4 0.10 mm and the average wall thickness was 0.3 0.06 mm (Fig. 1A). The conduits were sterilized by 70% alcohol for 15 min and washed by phosphate buffered saline (PBS) for 5 min before surgery.
3765
2.2. Animal implantation 2.2.1. Animals and microsurgical procedures All animal procedures followed the ethical guidelines of the experimental animals and were approved by the Institutional Animal Care and Use Committee. A total number of twenty-six male adult New Zealand rabbits weighing 3000e3500 g were used. The rabbits were housed under temperature-controlled conditions at 22 3 C, with a cycle of 12-h light and darkness, and with free access to food and water. The animals were anesthetized and maintained with isoflurane (Halocarbon, USA) throughout the microsurgical procedures. The skin from clipped lateral thigh was scrubbed in a routine fashion with antiseptic solution. The incision extended from the greater trochanter to the mid-calf distally. A muscle splitting incision exposed the sciatic and posterior femur nerves. The sciatic nerve was originated from sacrum spine nerve creating an adequate distal segment. The conduit (w24 mm long) was placed into the nerve gap by 7-0 Nylon sutures. Nerve stumps at both ends were sutured into both ends of the conduit to a length of approximately 2 mm [19]. The nerve gap was thus 20 mm in length. Muscle and skin were closed with 4-0 Dexon sutures. The outcomes were evaluated by imaging techniques, functional assessment and histological analysis. Twenty-four animals were sacrificed before 18 months. Two spare animals were sacrificed after 2 years. 2.3. Imaging technique The MRI equipment used was a sonata 1.5 T system from Siemens (Germany). The image acquisition employed a 5-inch surface coil (small loop coil). The MRI protocol included the T1 weighted image (T1WI), T2 weighted image (T2WI), and T2 weighted short time of inversion recovery (STIR) [20]. T1WI for the anatomy of sciatic nerve (bright signal) was acquired inside the nerve conduit. The optimized parameters for T1WI in the current study were the repeat time (TR)/echo time (TE) 300/12 ms, field of view (FOV) 50e80 mm, and matrix pixel 256 256. T2WI was acquired to demonstrate the tissues (bright signals) near the region of the implanted conduit [21]. The optimized parameters for T2WI in the current study were TR/TE 3000e4000/100 ms, FOV 50e80 mm, and matrix pixel 256 256 [22]. T1WI and T2WI had different brightness and contrast. They were complementary to each other so better justification regarding soft tissues and non-soft tissues (e.g. nerve conduit) could be made. STIR did not provide images of better resolution in this case, and therefore was not presented in this study. MR images could be taken from the sagittal, axial, and coronal directions to show different views in vivo. The nerve and muscle signal ratios on each sequence were measured and processed with computers. The imaging parameters of MRI should be refined according to the different animals and experimental conditions. In addition to the qualitative feature of MRI mentioned previously, a computer program (MATLAB 7.6.0) was employed to convert the biomedical images into data regarding the volume of regenerated nerve and that of nerve conduits as described below. MRI images based on analog transmission and data traffic were converted into intelligent data. By the use of MATLAB in operation and multiplication of MR images, the DICOM file was changed into data. In accordance with the software function, the images of the interested regions were pricked. The whole construct was divided into a few regions where the volume of nerve conduits and that of regenerated nerve were calculated and then combined. The calculation of volume was based on T2WI. During the later period, the newly repaired regions of the regenerated nerve were expanded to occupy the areas of the degraded conduit in MR images. The pixels and matrix from interested regions were converted into computational data and functional values. Array 2D was used as the calculation method to compute the measurement of the area. Array 2D data were assembled to form 3D combinations. Finally, 3D images were assembled to form the relations between nerve volume (regenerated) and conduit volume (degraded). The volume of regenerated nerve obtained by this method was then normalized to the original volume of the initial sciatic nerve and represented by the percentage value. The volume of the remaining conduit was also normalized to the original volume of the initial nerve conduit and represented as percentage. 2.4. Functional assessment
Fig. 1. (A) The dimensions of the asymmetric PLA nerve conduits used in this study; and (B) the SEM images showing the microporosity in the cross-section of the asymmetric conduits. Scar bar ¼ 100 mm.
In the initial stage, rabbits limped with the operated legs. After five to eight days, the muscle of thigh and that of calf had obvious dystrophy and the animals acted on healthy extremity. No walking track analysis was available from literature for animals other than rats. For the reason, the differences in the muscle periphery of thigh and calf between the control (right) and the experimental (left) legs could be measured as a way to evaluate the functional recovery. The muscle periphery measurement was performed on rabbits before surgery and every two weeks after the implantation until no difference of muscle periphery between both legs was identify. Ten male rabbits were selected to measure the muscle periphery initially. Because some rabbits were sacrificed for the electrophysiological analysis and histology, after 16 months only six rabbits were used for data collection. Electrophysiological measurements were performed on the experimental animals before they were sacrificed for histological analysis. All animals were anesthetized. The amplifier used in the study had two channels and a ground line. The ground line was connected on the ear of the rabbit. Nerves were stimulated by a pair of needle electrodes. The needle electrode which was connected with a DC electrical
3766
S.-h. Hsu et al. / Biomaterials 32 (2011) 3764e3775
stimulator (PowerLab ML866, AD Instrument, Australia) was placed on the proximal end of the nerve. The action potential recorded from sciatic nerve in this study actually represented the aggregate electrical effect of individual action potentials generated by the electrical response of nerves to stimulation. The stimulation voltage
was generally from 1 to 10 mV and in 2 Hz pulse. The recording electrode had 10-s interval to avoid nerve fatigue and was avoided interference with radio wave. The pulses of muscle and nerve action potential were recorded by computer software (Scope for Windows, AD Instrument, Australia). Based on the nerve-to-nerve
Fig. 2. The exposed rabbit sciatic nerve (A) was transected and replaced by a nerve conduit (B). The abbreviation: P, the proximal end; D, the distal end. The anatomical position of the rabbit could be shown by the localizer image of MRI (C) to confirm the relative locations.Sciatic nerve was at 2-cm depth parallel to the side of the femur bone (D). After surgery, the nerve conduit was shown to be parallel to the side of the femur bone (E). Based on images of the early regeneration stage (<4 months, F, G and H), the regenerated nerve had either sprouted from the proximal end (F, 12 weeks) or sprouted from both ends (G, 10 weeks; H, 11 weeks; the sprouting rate from the proximal being faster). C, D, G, H: T1WI images; E, F: T2WI images.
S.-h. Hsu et al. / Biomaterials 32 (2011) 3764e3775 distances and time (from stimulation point to the maximum pulse amplitude), the nerve conduction velocity (NCV) for each group of conduits was evaluated. The peak amplitude was obtained from the muscle action potential (MAP). The dynamic motion behavior (jumping) of rabbits after nerve repair was continuously recorded as video and analyzed by a semi-automation program developed by us using MATLAB 7.4.0. In this analysis, the video with the mpg/mpeg format was converted into a sequence of static images to record the data (with the * frame rate of 40 frames per second). The vector a was defined by manually selecting two points horizontal to the ground in the first frame. If the video was taken steadily, * then the direction of vector a would be consistent in the rest of the frames. Another * vector b was defined as the central axis of the lower limb. The direction of * the vector b varied in accordance to the swing of the lower limb. The angle q * * between the two vectors a and b could be calculated according to equation
3767
** * * q ¼ cos1 ð a: bÞ=ðj aj:j bjÞ; 0+ q 180+ . A map indicating the variation of q versus time in terms of frame numbers could be used to evaluate the function of the lower limb. A complete swing cycle was given to show the changes of q corresponding to the movement of the lower limb. The extension range of the lower limb was denoted as Q and defined as the absolute value of the gap between the maximum and the minimum of the q during a swing cycle. Rabbit activities quantified by Q may be used to evaluate the degree of recovery of the limb. L Q values are supposed to indicate better and refined recovery. 2.5. Histological analysis After a period of time (4, 6, 8, 12, 16 and 18 months), four animals were euthanatized by CO2 overdose treatment for histological observation. The implanted
Fig. 3. The harvested conduits from sacrificed animals after 4 months (A), 12 months (C) and 18 months (E) and the corresponding sagittal MR images (B, D, F) before sacrifice are shown. The long black arrows in (B, D, F) indicated the regenerated nerve. The short black arrows in (B, D, F) indicated the conduit. The cross-section of the regenerated nerve became larger while the conduit was slowly degraded. The conduit in (F) was almost completely degraded. All MR images were T1WI images.
3768
S.-h. Hsu et al. / Biomaterials 32 (2011) 3764e3775
conduits were harvested and fixed in cold-buffered 4% glutaraldehyde solution immediately. After 48 h, the nerve conduits were cut open longitudinally. If the nerve gap was successfully connected, the regenerated nerve was in the form of a white string-like substance that connected the two ends of the initial gap. The specimens were washed in PBS several times and then transected into the proximal, medium and distal segments. The central portion of the medium segments (midconduit) was a completely regenerated region and was often used for comparison. All samples were postfixed in 1% osmium tetroxide (5 ml, Polysciences, USA), dehydrated in a graded series of ethanol solutions, and then embedded. The embedded samples were cut with 5 mm thickness, and stained with H&E, which did not stain the conduit (PLA). All nerve sections were observed under an optical microscope (Nikon Labophot, Japan), and photographs were taken using a digital camera (Nikon H666L, Japan).
2.6. Statistical analysis Twenty-six rabbits underwent the nerve conduit implantation.Five rabbits were evaluated by the imaging technique to trace the nerve regeneration every two weeks from 0 to 16 weeks, every four weeks after 16 weeks and every six to eight weeks after 64 weeks. Typical images were shown. The average diameters of the regenerated nerve based on the axial MR images at the midconduit as well as the computed volumes of the regenerated nerve and the degraded conduits were the mean values from five rabbits. Dynamic motion behavior was estimated at 4, 8, 12, 16 and 18 months (n ¼ 5 for each group) and the normal side was the control group. Electrophysiological measurement and histological analysis were analyzed at 8, 12 and 18 months (n ¼ 4). The results from multiple samples were expressed as mean standard deviation. Statistical differences were analyzed by one-way analysis
Fig. 4. The coronal views (A, C, E) and the axial views (B, D, F) of the rabbits receiving conduits for 4 months (A, B), 6 months (C, D) and 8 months (E, F). The long white arrows in images indicated the regenerated nerve and the short white arrows in images were nerve conduits (A, C, E). These images were clearer than those in Fig. 3 and could be used for quantification purposes. After 8 months, the regenerated nerve and the surrounding tissue became next to each other and less distinguishable in the image of axial view as a result of conduit degradation. All images were T1WI images.
S.-h. Hsu et al. / Biomaterials 32 (2011) 3764e3775 of variance (ANOVA). p < 0.05 was considered as statistically significant. After sacrificing (at 4, 6, 8, 12, 16 and 18 months, n ¼ 4 for each group), the nerve sections were retrieved for histological process. Typical histological data were shown.
3. Results The dimension of the PLA conduits used in the study is shown in Fig. 1A The cross-section of each conduit showed the typical asymmetric microporous structure, as illustrated in Fig. 1B During the surgery, the rabbit sciatic nerve (Fig. 2A) was replaced by the nerve conduit (Fig. 2B). The anatomical position shown in Fig. 2C was used to locate the sciatic nerve (2.0 cm below femur bone). Normal sciatic nerve in rabbits had a smooth linear structure with isointense MR images (Fig. 2D). Right after surgery, the nerve conduit was shown to be parallel to the side of the femur bone (Fig. 2E). The T2WI image (Fig. 2E) showed the black lumen of the conduit. Images taken at the early regeneration stage (<4 months) showed that in some rabbits the regenerated nerve sprouted from the proximal end (Fig. 2F, as commonly perceived. In other rabbits (two out of five, Fig. 2G,H)), the regenerated nerve sprouted from both ends. In the latter case, the rate of nerve sprouting from the proximal was much faster than that from the distal. All conduits maintained a stable supporting structure during the first four months. The newly regenerated nerve in the conduit was completely connected after about four months (Fig. 3, sagittal view at 4, 12, and 18 months). All subjects had successful connection. The gross appearance of the harvested nerve conduits showed that the regenerated nerve inside the nerve conduits gradually became thicker and angiogenesis was obvious as red color in the surrounding tissue (Fig. 3A,C,E). T1WI images obtained after 4 months showed the regenerated sciatic nerve (Fig. 3B,D,F). The bright signal of the regenerated area inside the tube was fluid component and those outside the tube were the neighboring blood vessels as indicated by the thin white arrows in Fig. 3D,F. During this period, degradation of the nerve conduits was visible from the MR images. The dark images around the conduits (indicated by the short black arrows in Fig. 3B,D,F) were the signals of nerve conduits.
3769
The regenerated nerve and repaired tissue replaced the spaces of the degraded conduits and had isointense images with soft tissue. The dark image lines of the nerve conduit in the early stage (as indicated by the short arrows in Fig. 3B) were blocked and replaced by the new regenerated nerve in the later stage. The size (diameter) of the regenerated nerve increased with time (from Fig. 3B to F). Images of the coronal and the axial views (4, 6, and 8 months) are shown in Fig. 4. During this period, the conduit and the regenerated nerve could be distinguished by the axial view of both T1WI images and T2WI images taken at the middle section of the conduit. The dark signals inside the ring area were the images of regenerated nerve, which became thicker gradually. The white signals of the ring were the fluid components inside the degraded conduits, which became narrower during this period. The diameter of the regenerated nerve at the midconduit quantified by the axial MR images was about 0.7 0.08 mm at 4 months (Fig. 4B), 1.8 0.13 mm at 6 months (Fig. 4D), and 2.8 0.17 mm at 8 months (Fig. 4F). The three-dimensional volume of the regenerated nerve and that of the nerve conduit was obtained from computation of the MR images. The tendency of changes for the volumes of the nerve conduit and the regenerated nerve along the implantation period is shown in Fig. 5. The changes of volumes for the nerve conduit and for the regenerated nerve were in the opposite trend. The regenerated nerve increased in diameter when the degradation of polymer continued in vivo. The two tendency curves crossed over each other at about 16e20 weeks, which was approximately when the nerve first connected. The nerve conduits were mostly (w95%) degraded after 52 weeks, and were almost completely (w99%) degraded after 18 months. Based on MR image analysis, the initial three-dimensional volume (V0) of the PLA nerve conduit in situ right after implantation was about 83.7 0.84 mm3. This value was in good agreement with that calculated based on the geometry and dimension shown in Fig.1A. The volume of the nerve conduit in vivo at each time point (V) and the original volume V0 were used to estimate the degradation kinetics by assuming that the volume was proportional to the amount of the reactant. The relationship of V0 vs. time appeared to be an exponential decay, as shown in Fig. 6A. The natural log of V0/V
Fig. 5. The volume change of the nerve conduit and that of the regenerated nerve during the period of 18 months. These data were obtained by MATLAB software based on MR images. The percent volume of nerve conduit was calculated based on the original volume of the nerve conduit. The percentage of nerve regeneration (including sprouted nerve before connection) was calculated based on the original volume of normal sciatic nerve. Nerve regeneration was accompanied by the conduit degradation and after 18 months the conduit was almost completely degraded.
3770
S.-h. Hsu et al. / Biomaterials 32 (2011) 3764e3775
Methods. The map of the variation of q versus time in terms of frame numbers as well as the maximum and minimum angles were indicated from the video image frames (Fig. 8B). The extension range of the lower limb (denoted as Q) was defined as the difference between the maximum and the minimum q during a swing cycle. Based on the map, Q was 29 2 at 4 months (Fig. 8C) and 68 3 for the control limb (Fig. 8D). The values of Q obtained at different periods after implantation are shown in Fig. 8E. The values of Q increased steadily from 4 to 18 months. This indicated that the motion function was slowly recovered from the nerve injury. The functional recovery was about 81% of that of the control after 18 months, based on the motion behavior analysis.
Fig. 6. (A) The exponential decay of the volume (V) of the nerve conduit. (B) The degradation kinetics of the conduit, showing the approximately linear relationship between natural log of V0/V and time.
vs. time was plotted in Fig. 6B, and was approximately linear. This suggested the pseudo first-order kinetics for the degradation of the PLA conduit. The slope of the linear curve reflected the rate constant, which was 0.0706 week1 or 0.0101 day1. The muscle peripheries of thigh and calf were measured and are shown in Fig. 7A. In the initial stage of injury, rabbits limped with the operated legs and had obvious muscular dystrophy. The decrease of muscle peripheries was an indirect proof of the early nerve injury in the rabbits. Both thigh muscle and calf muscle had smaller periphery that the original values before 36 weeks. The muscle peripheries were the lowest at 16e18 months and were slowly recovered to the original values. After 36 weeks, the muscular peripheries remained similar. The electrophysiological analysis was conducted at 8, 12, and 18 months before the animals were sacrificed. The nerve conduction velocity (NCV) was calculated based on NAP and the peak amplitude was obtained from MAP. The values of NCV at different time periods are shown in Fig. 7B. The NCV of the normal sciatic nerve was about 64.3 3 m/s, which was significantly higher (p < 0.05) than that of the regenerated nerve. The percent recovery of the rabbits based on NCV was about 62% at 8 months, 64% at 12 months, and 82% at 18 months. The peak amplitude for the normal side muscle was 0.0036 0.0001 mV, which was significantly higher (p < 0.05) than that of the experimental side (Fig. 7C). The recovery of the rabbits based on peak amplitude was about 58%, 69%, and 81% at 8 months,12 months, and 18 months respectively. Based on the electrophysiology, the percent functional recovery after 18 months had reached 80%. The jumping motion of the experimental rabbits was evaluated by a novel method involving the measurement of motion angle and extension range for the lower limb, as shown in Fig. 8. The motion * * angle q was defined from two vectors a and b obtained from the lower limb of each rabbit (Fig. 8A), as detailed in the Materials and
Fig. 7. (A) The changes in muscle periphery of lower limbs in experimental rabbits receiving conduits. At early implantation (before 36 weeks), the peripheries of the thigh and calf were obviously smaller (muscular dystrophy). After 36 weeks later, the peripheries of the thigh and calf were recovered to nearly normal so the measurement was done less frequently. (B, C) Electrophysiological data in the experimental rabbits receiving conduits recorded after 8, 12 and 18 months. The NCV values in (B) were obtained from the compound nerve action potential (NAP). The peak amplitude values in (C) were obtained from the compound muscle action potential (MAP).
S.-h. Hsu et al. / Biomaterials 32 (2011) 3764e3775
Histology of the nerve sections at the midconduit is shown in Fig. 9. All 24 subjects had successful connection (i.e. success rate ¼ 100%). Normal sciatic nerve of the rabbit (Fig. 9Aa) was compared with that of the regenerated nerve at 4 months (Fig. 9Ab), 6 months (Fig. 9Ac), 8 months (Fig. 9Ad), 12 months (Fig. 9Ae), 16 months (Fig. 9Af) and 18 months (Fig. 9Ag, h). The
3771
presence of myelins (M), blood vessels (V), regenerated nerves (N), connective tissues (CT), or fats (F) was indicated in the repaired tissues. At 4 months, the newly regenerated nerves showed two nerve bundles (Fig. 9Ab). At 6 months, the newly regenerated nerves had formed several nerve bundles (Fig. 9Ac). From 6 to 12 months, the bundles continued to expand their size (Fig. 9Acee).
.
.
Fig. 8. The dynamic behavior analysis for rabbits showing the extension range of the lower limbs (Q). The motion angle (q) was defined as the angle between the vectors a and b (panel A). During a swing (panel B; from (1), (2) to (3)), the motion angle was constantly recorded. The data curves of motion angle (q) after 4 months (C) and the control (normal side, D). The calculated data during 4e18 months were compared with that of the normal side (E).
3772
S.-h. Hsu et al. / Biomaterials 32 (2011) 3764e3775
Inside each bundle, the number of myelinated axons also increased. Before 12 months, the nerve bundles were in irregular shape (Fig. 9Abee). The structures of blood vessels, nerve fibers, myelins, and their distributions in newly regenerated nerves were different from those in the normal sciatic nerve. Blood vessel formation was found both inside the nerve bundles and outside the nerves near the connective tissues. After 16 months, the independent nerve bundles had self-organized into large nerve bundles (Fig. 9Af, h). At 18 months, the nerve bundles were merged or combined with each other. This self-assembly was accompanied by a decrease in the number of blood vessels. At this time, the morphology of the regenerated nerve tissue finally became more homogeneous (Fig. 9Ah), which was more like that of the normal sciatic nerve (Fig. 9Aa). The diameter of the regenerated nerve at 18 months was slightly larger than that of the normal nerve. Histology at larger magnification (400) is shown in Fig. 9B. Nerve sections at different periods were compared with that of the normal sciatic nerve (Fig. 9Ba). Thinner myelins were observed in 4e12 months (Fig. 9Bjem). The self-assembly of myelins into large bundles was clearly visualized after 16e18 months and the margins were indicated by the long black arrows in Fig. 9Bnep. 4. Discussion The rat model of sciatic nerve injury was the most common animal model for studying peripheral nerve regeneration. However, the longest gap that could be created in rats was 15 mm [9]. Rabbits are larger in size than rats and are more cost-efficient compared to larger animals such as dogs [12,23] and cats [24]. Sciatic nerve regeneration in rabbits across a 10 mm gap bridged by an autograft or an epineural flap has been examined for a period of 3 months [25,26]. So far there is no literature regarding large-gap sciatic nerve regeneration bridged by a conduit in rabbits. There was a study on bridging the gap of rabbit common peroneal nerve (thinner than sciatic nerve) with poly-3-hydroxybutyrate conduits, but nerve regeneration was only evaluated by histology after 4 months [27]. Our present study described nerve regeneration across a 20 mm sciatic nerve gap in rabbits during a period of 18 months, which could provide better translations to long-term clinical applications. The adequate nerve regeneration and functional recovery may be attributed to the asymmetric structure of our PLA conduits. The commercial Neurotube made of polyglycolic acid did not supply tubes longer than 20 mm with diameter similar to our conduits. Therefore, a comparison could not be made. However, we discovered that asymmetric PLA conduits with additional surface microgrooves (width/spacing/depth ¼ 20 mm/20 mm/3 mm) [19] could advance the nerve connection time from four months to three months in a parallel study, suggesting that the rabbit sciatic nerve model and MRI monitoring presented here may be a good platform for studying peripheral nerve regeneration and may help correlating the biomaterials design and device performance. MRI successfully monitored the regeneration progresses of transected peripheral nerve in the biomaterial conduit in this study. In a recent report, the acute crush injury (created by clamps) in rabbit sciatic nerve was diagnosed by MRI; but nerve repair was not followed [20]. Based on the MR images in the present study, we identified that some of the transected nerve sprouted from both the proximal and distal ends. The proximal stump of the transected nerve regenerated faster than the distal stump and the two sprouts finally connected not in the center but about 3/4 distance from the proximal end after about 4 months. This was completely surprising because it was different from the textbook concept that the transected nerve sprouts from the proximal end and that the distal end undergoes Wallerian degradation. It could be argued that the sprouted “nerve” from the distal end we observed may have been
the image of the organized fibrin bridge and not the nerve itself. Nevertheless, if it were the fibrin bridge, it should have linked both ends. Another argument could be that this phenomenon may only occur in acute nerve injury, which we have not had any clue yet. We suggest that this observation may deserve further confirmation and in-depth study for better understanding of the nerve regeneration mechanism inside a polymer conduit. Material degradation is essential information required by all tissue engineering medical products; however, few studies have fully characterized the biodegradation on site in vivo. It was indicated that the degradation rate of PLA differed according to the site of implantation [28]. Based on the present study, our nerve conduits were gradually degraded during the 18-month period (50% degraded within the first 20 weeks and 95% degraded within the first year). PLA was reported to degrade partially (56% or more based on molecular weight change) after 6 months in vivo (subcutaneous mice) [29]. The degradation rate of our PLA conduits in the first 6 months based on a different definition (volume reduction) roughly agreed with that reported. The proper degradation rate allowed the conduit to provide a stable structural support for the newly regenerated nerve during the first 4e6 months. This may account for the good success rate and functional outcome of our conduits after 18 months. Due to the heterogeneous nature of the PLA degradation, it is usually difficult to obtain the information regarding the degradation in vivo. Most studies measured the decrease in molecular weight [28], but it was less likely to obtain the degradation kinetics by measuring the average molecular weight of the remaining device. The advantage of MRI was to quantify the intensity of the conduit. The degradation kinetics estimated from the MR images indicated a pseudo first-order reaction. PLA was reported to undergo selfcatalysis in vitro [30]. The pseudo first-order kinetics of PLA was normally observed when the concentration of catalyst HCl was kept constant [30]. Therefore, the degradation information obtained from MR images suggested that in vivo the acid concentration near the conduit was relatively constant. PLA is not soluble in water unless dissociation of the acid end-group, which is expected to result in an acidic environment and a significant contribution to the hydrolysis. The degradation kinetics found in this study suggested a fast drainage of the acid generated near the conduit, which could be attributed to the special asymmetric design in the structure of the conduit. Results from the dynamic behavior analysis showed the rehabilitation of sciatic nerve in rabbits (81% after 18 months). Few papers have mentioned the analytical method for evaluation of the functional recovery in rabbits. The sciatic functional index (SFI) or dynamic cat walks (DCW) for rats are not applicable for rabbits because of the unique rabbit motivation (jumping, not walking) and size (much larger than rats) [31].The analysis of extension range is similar to SFI or DCW in a way that it analyzes the spontaneous coordinate activity. The percent functional recovery based on the extension range was very consistent with that based on the peak amplitude of compound muscle action potential obtained from electrophysiology. Therefore, the behavior analysis developed in this study was a novel and effective method to analyze the functional recovery during the period of nerve regeneration in the experimental animals. In sciatic nerve crushed rabbits, the myelin formation was immature and different from that of the normal sciatic nerve [20]. This was also observed during the early stage of nerve regeneration in our study. In the current model, nerves were connected at about 4 months, but the process of nerve bundle reorganization took much longer (more than a year). In the rat model of 10 mm sciatic nerve transection, these phenomena (including increase and decrease in the vessel number) were observed in 4e6 weeks [19]. The ligated sciatic nerves in rats recovered even faster and had the
S.-h. Hsu et al. / Biomaterials 32 (2011) 3764e3775
3773
Fig. 9. (A) Histology of the nerve sections at the midconduit after 4 months (b), 6 months (c), 8 months (d), 12 months (e), 16 months (f) and 18 months (g, h) of implantation, compared with that of the normal sciatic nerve (a). The abbreviation M: myelin, V: blood vessel, N: regenerated nerve, F: fat tissue and C: conduit. Scar bar ¼ 200 mm. Small nerve bundles inside the conduits were observed after 4e12 months (bee). Small bundles were self-assembled into large bundles after 16e18 months (with the margins indicated by the long black arrows in feh). (B) Images were displayed at larger magnification (400) for normal sciatic nerve (i), and for regenerated nerve after 4 months (j), 6 months (k), 8 months (l), 12 months (m), 16 months (n) and 18 months (o, p) of implantation.Scar bar ¼ 50 mm. Thinner myelins were observed within a year (jem). The self-assembly occurred after 16e18 months (nep). The margin of the self-assembled area was marked by the long black arrows (nep).
same histology as normal nerves after 12 weeks [32]. In our study, the evolutions of morphology for the regenerated nerves were consistent with the change in functional recovery demonstrated by the electrophysiology and behavior analysis. For minor injury in literature, the muscular function began to recover in some rabbits at six weeks following the crush of sciatic nerve [20]. Obviously it
took much more time for the large-gap sciatic nerve defected rabbits to recover from the injury, even in the presence of a protecting and guiding biomaterial conduit. On the other hand, since none of the rabbits died during the experimental period, and the average data collected from the two spare rabbits before their sacrifice at 24 months showed the promising NCV value at 61.5 m/s
3774
S.-h. Hsu et al. / Biomaterials 32 (2011) 3764e3775
Fig. 9. (continued).
(98% recovery) and extension range (Q) at 62.0 (91% recovery), we considered that the asymmetric PLA conduit may provide satisfactory long-term recovery for the injured peripheral nerve. 5. Conclusions The performance of an asymmetric PLA conduit could be evaluated by the rabbit sciatic nerve transection model during a much longer period compared with that observed in the rat model. The
PLA conduit was slowly degraded while the nerve regeneration proceeded. Some of the regenerated nerve sprouted from the proximal end, while others sprouted from both the proximal and distal stumps. In both cases, the nerve was connected after four months. After connection, the nerve continued to increase its diameter. The degradation of the conduit quantified by MR images showed pseudo first-order kinetics with a rate constant w0.01 day1. The functional recovery after 18 months was about 82% based on electrophysiology and was about 81% based on behavior
S.-h. Hsu et al. / Biomaterials 32 (2011) 3764e3775
analysis. Histology showed that the regenerated nerve inside the conduit was in the form of several small nerve bundles after six months and merged into a single large bundle after 18 months, which was close to the histology of normal sciatic nerve. By adjusting the parameters to display the nerve conduit and regenerated nerve, MRI could provide fine resolution to quantify the performance and the degradation of the nerve conduits in vivo. Acknowledgments The authors thank Dr. Su Jing Yang and the MRI center, Department of Radiology, Taichung Veterans General Hospital, for their group support. This work was supported by grants from the Taichung Veterans General. Hospital/National Chung Hsing University (grant number TCVGH-NCHU 977608) and the National Science Council (grant number NSC 99-2321-B-002-043). Appendix Figures with essential color discrimination. Certain figures in this article, particularly Figs. 2,3,8 and 9 are difficult to interpret in black and white. The full color images can be found in the online version, at doi:10.1016/j.biomaterials.2011.01.065. References [1] Faweett JW, Keynes RJ. Peripheral nerve regeneration. Annu Rev Neurosci 1990;13:43e60. [2] Vargas ME, Barres BA. Why is Wallerian degeneration in the CNS so slow? Annu Rev Neurosci 2007;30:153e79. [3] Coleman MP, Freeman MF. Wallerian degeneration, WldS, and Nmnat. Annu Rev Neurosci 2010;33:245e67. [4] Schmidt CE, Leach JB. Neural tissue engineering: strategies for repair and regeneration. Annu Rev Biomed Eng 2003;5:293e347. [5] Guertin AD, Zhang DP, Mak KS, Alberta JA, Kim HA. Microanatomy of axon/ glial signaling during Wallerian degeneration. J Neurosci 2005;25:3478e87. [6] Mackinnon SE, Dellon AL. A study of nerve regeneration across synthetic (Maxon) and biologic (collagen) nerve conduits for nerve gaps up to 5 cm in the primate. J Reconstr Microsurg 1990;6:117e21. [7] Cornbrooks CJ, Carey DJ, McDonald JA, Timpl R, Bunge RP. In vivo and in vitro observations on laminin production by Schwann cells. Proc Natl Acad Sci USA 1983;80:3850e4. [8] Chang JY, Lin JH, Yao CH, Chen JH, Lai TY, Chen YS. In vivo evaluation of a biodegradable EDC/NHS-cross-linked gelatin peripheral nerve guide conduit material. Macromol Biosci 2007;7:500e7. [9] Deumens R, Bozkurt A, Meek MF, Marcus MA, Joosten EA, Weis J, et al. Repairing injured peripheral nerves: bridging the gap. Prog Neurobiol 2010;92:245e76. [10] Chang CJ, Hsu SH. The effect of high outflow permeability in asymmetric poly (dl-lactic acid-co-glycolic acid) conduits for peripheral nerve regeneration. Biomaterials 2006;27:1035e42. [11] Oh SH, Kim JH, Song C, Jeon BH, Yoon JH, Seo TB, et al. Peripheral nerve regeneration within an asymmetrically porous PLGA/Pluronic F127 nerve guide conduit. Biomaterials 2008;29:1601e9.
3775
[12] Ichihara S, Inada Y, Nakada A, Endo K, Azuma T, Nakai R, et al. Development of new nerve guide tube for repair of long nerve defects. Tissue Eng Part C Methods 2009;15:387e402. [13] Meek MF, Den Dunnen WF, Schakenraad JM, Robinson PH. Long-term evaluation of functional nerve recovery after reconstruction with a thin-walled biodegradable poly (dl-lactide-epsilon-caprolactone) nerve guide, using walking track analysis and electrostimulation tests. Microsurgery 1999;19:247e53. [14] Ansselin AD, Fink T, Davey DF. Peripheral nerve regeneration through nerve guides seeded with adult Schwann cells. Neuropathol Appl Neurobiol 1997;23:387e98. [15] Kobayashi J, Mackinnon SE, Watanabe O, Ball DJ, Gu XM, Hunter DA, et al. The effect of duration of muscle denervation on functional recovery in the rat model. Muscle Nerve 1997;20:858e66. [16] Veraa S, Dijkman R, Meij BP, Voorhout G. Comparative imaging of spinal extradural lymphoma in a Bordeaux dog. Can Vet J 2010;51:519e21. [17] Bendszus M, Koltzenburg M, Wessig C, Solymosi L. Sequential MR imaging of denervated muscle: experimental study. AJNR Am J Neuroradiol 2002;23:1427e31. [18] Bendszus M, Wessig C, Reiners K, Bartsch AJ, Solymosi L, Koltzenberg M. MR imaging in the differential diagnosis of neurogenic foot drop. AJNR Am J Neuroradiol 2003;24:1273e4. [19] Hsu SH, Ni HC. Fabrication of the microgrooved/microporous polylactide substrates as peripheral nerve conduits and in vivo evaluation. Tissue Eng Part A 2009;15:1381e90. [20] Li X, Shen J, Chen J, Wang X, Liu Q, Liang B. Magnetic resonance imaging evaluation of acute crush injury of rabbit sciatic nerve: correlation with histology. Can Assoc Radiol J 2008;59:123e30. [21] Valentine HL, Does MD, Marshall V, Tonkin EG, Valentine WM. Multicomponent T2 analysis of dithiocarbamate-mediated peripheral nerve demyelination. Neurotoxicology 2007;28:645e54. [22] Stroman PW, Dorvil JC, Marois Y, Poddevin N, Guidoin R. In vivo time course studies of the tissue responses to resorbable polylactic acid implants by means of MRI. Magn Reson Med 1999;42:210e4. [23] Wang XD, Hu W, Cao Y, Yao J, Wu J, Gu XS. Dog sciatic nerve regeneration across a 30-mm defect bridged by a chitosan/PGA artificial nerve graft. Brain 2005;128:1897e910. [24] Suzuki Y, Tanihara M, Ohnishi K, Suzuki K, Endo K, Nishimura Y. Cat peripheral nerve regeneration across 50 mm gap repaired with a novel nerve guide composed of freeze-dried alginate gel. Neurosci Lett 1999;259:75e8. [25] Ignatiadis IA, Yiannakopoulos CK, Barbitsioti AD, Avram AM, Patralexis HG, Tsolakis CK, et al. Diverse types of epineural conduits for bridging short nerve defects. An experimental study in the rabbit. Microsurgery 2007;27:98e104. [26] Ignatiadis IA, Tsiampa VA, Yiannakopoulos CK, Xeinis SF, Papalois AE, Xenakis TH, et al. A new technique of autogenous conduits for bridging short nerve defects. An experimental study in the rabbit. Acta Neurochir Suppl 2007;100:73e6. [27] Young YC, Wiberg M, Terenghi G. Poly-3-hydroxybutyrate (PHB): a resorbable conduit for long-gap repair in peripheral nerves. Br J Plast Surg 2002;55:235e40. [28] Tschakaloff A, Losken HW, Oepen RV, Michaeli W, Moritz O, Mooney MP, et al. Degradation kinetics of biodegradable dl-polylactic acid biodegradable implants depending on the site of implantation. Int J Oral Maxillofac Surg 1994;23:443e5. [29] Gogolewski S, Jovanovic M, Perren SM, Dillon JG, Hughes MK. Tissue response and in vivo degradation of selected polyhydroxyacids: polylactides (PLA), poly (3-hydroxybutyrate) (PHB), and poly(3-hydroxybutyrate-co-3-hydroxyvalerate) (PHB/VA). J Biomed Mater Res 1993;27:1135e48. [30] Siparsky GL. Degradation kinetics of poly(hydroxy) acids: PLA and PCL. ACS Symp Ser 2001;764:230e51. [31] Hare GM, Evans PJ, Mackinnon SE, Best TJ, Bain JR, Szalai JP, et al. Walking track analysis: a long-term assessment of peripheral nerve recovery. Plast Reconstr Surg 1992;89:251e8. [32] Bendszus M, Wessig C, Solymosi L, Reiners K, Koltzenburg M. MRI of peripheral nerve degeneration and regeneration: correlation with electrophysiology and histology. Exp Neurol 2004;188:171e7.