Acta Neurochir DOI 10.1007/s00701-014-2009-9

EXPERIMENTAL RESEARCH - NEUROSURGICAL TECHNIQUES

Functional recovery after repair of peroneal nerve gap using different collagen conduits Vincent Pertici & Jérôme Laurin & François Féron & Tanguy Marqueste & Patrick Decherchi

Received: 22 November 2013 / Accepted: 18 January 2014 # Springer-Verlag Wien 2014

Abstract Background Currently, autologous nerve implantation to bridge a long nerve gap presents the greatest regenerative performance in spite of substantial drawbacks. In this study, we evaluate the effect of two different collagen conduits bridging a peroneal nerve gap. Methods Rats were divided into four groups: (1) the gold standard group, in which a 10-mm-long nerve segment was cut, reversed, and reimplanted between the nerve stumps; (2) the CG-I/III group, in which a type I/III collagen conduit bridged the gap; (3) the CG-I, in which a type I collagen conduit was grafted; and (4) the sham group, in which a surgery was performed without injuring the nerve. Peroneal Functional Index and kinematics analysis of locomotion were performed weekly during the 12 weeks post-surgery. At the end of the protocol, additional electrophysiological tests, muscular weight measurements, axon counting, and g-ratio analysis were carried out. Results Functional loss followed by incomplete recovery was observed in animals grafted with collagen conduits. At 12 weeks post-surgery, the ventilatory rate of the CG-I group in V. Pertici : J. Laurin : T. Marqueste : P. Decherchi (*) Institut des Sciences du Mouvement: Etienne-Jules MAREY, Equipe, Plasticité des Systèmes Nerveux et Musculaire, Parc Scientifique et Technologique de Luminy, Faculté des Sciences du Sport de Marseille, Aix-Marseille Université (AMU) et Centre National de la Recherche Scientifique (CNRS), UMR 7287, CC910 - 163 Avenue de Luminy, 13288 Marseille Cedex 09, France e-mail: [email protected] URL: www.ism.univmed.fr F. Féron Neurobiologie des Interactions Cellulaires et Neurophysiopathologie, Equipe, Plasticité Olfactive et Réparation du Système Nerveux, Faculté de Médecine Nord, Institut Fédératif de Recherche Jean Roche (IFR11), Aix-Marseille Université (AMU) et Centre National de la Recherche Scientifique (CNRS), UMR 7259, 51, boulevard Pierre Dramard, 13916 Marseille cedex 20, France

response to exercise was similar to the sham group, contrary to the CG-I/III group. After KCl injections, an increase in metabosensitive afferent-fiber activity was recorded, but the response stayed incomplete for the collagen groups compared to the sham group. Furthermore, the CG-I group presented a higher number of axons and seemed to induce a greater axonal maturity compared to the CG-I/III group. Conclusions Our results suggest that the grafting of a type I collagen conduit may present slight better prospects than a type I/III collagen conduit. Key words Kinematic . Movement . Ventilatory rate . g-ratio

Introduction Peripheral nerve transections are often clinically reconnected by epineurial sutures. However, peripheral nerve injuries resulting in long gaps cannot be simply end-to-end sutured without generating a tension that impedes axonal regeneration. As previously shown in sciatic nerve rat model, a gap anastomosis as short as 2 mm resulted in mild tension [25]. Bridging strategies are therefore required to repair long nerve gaps [9]. Currently, the autograft is recognized as the “gold standard” due to the fact that it presents the greatest regenerative performance. It consists in removing a segment of nerve from another part of the body (usually the sural nerve, which purely drives sensory information) in order to implant it in between the sectioned nerve stumps. However, a limited amount of nerve-donor material and subsequent loss of function at the nerve donor-site are the major drawbacks. Potential mismatches in size and fasciculation patterns between the autograft and the sectioned nerve stumps cannot be excluded. Additionally, only 40-50 % of autografted patients show functional benefits [16].

Acta Neurochir

Promising alternative strategies are the use of natural, material-based guides (collagen, fibrin, alginate, etc.). These guides, also known as channels or conduits, promote axonal regeneration from the proximal to the distal nerve stump and prevent fibrous scar tissue formation [24]. Previous investigations have demonstrated the potential recovery induced by such grafts [23]. Among them, guides made of collagen offer several advantages, including biocompatibility, non-toxic degradation products, and the induction of minimal foreign body responses [9]. Collagen guides display adhesive properties for different cell types that allow long-time survival, proliferation in the inner surface of the guide, and angiogenesis [18]. Those characteristics promote the needed alignment of Schwann cells, and therefore, the formation of guiding columns of Büngner [15]. Those columns, which release extracellular matrix proteins and specific adhesion molecules, serve as the physical framework for regenerating axons [12]. The interstitial structure of many tissues includes two distinct collagens, termed types I and III [11]. The implantation of pure I and mixed I/III collagen conduits exhibits functional recovery close to the autograft technique. Indeed, a highly purified type I bovine collagen conduit bridging a 4-mm-long nerve gap was demonstrated to be as effective as the autograft based on electrophysiological study in rodents and nonhuman primates [2]. Porcine-derived type I/III collagen conduits have also proven to support axonal regeneration across a 10-mmlong peroneal nerve gap in rats [1]. To date, the comparison of guides’ recovery effectiveness between studies is problematic due to disparate protocols. Nerve regeneration can be studied on a large variety of models (rats, mice, dogs) and on several nerves (sciatic, peroneal, tibial, and others). Moreover, the use of different gap lengths and multiple biomaterials might influence the axonal regrowth and the subsequent functional outcomes [23]. Therefore, the present study, based on the repair of a 10mm peroneal gap in the rat model, is designed to compare the recovery effectiveness induced by the implantation of two different types of collagen guides: a purified porcine tendon type I collagen guide and a porcine skin type I/III collagen. Functional measurements were performed during the 12 weeks post-surgery. Improvement of the locomotor function was assessed by: (1) the Peroneal Functional Index (PFI), a method to evaluate the functional condition based on footprint measurements of walking rats [3], and (2) the analysis of the ankle angles during the swing phase of locomotion [22]. Complementary electrophysiological techniques were also used to measure the response of metabosensitive afferent fibers to KCl injections and the ventilatory rate changes to prolonged isometric contractions. Furthermore, morphometric measurements were used to estimate the number of axons, the mean axon diameter and the myelination at the proximal and distal nerve levels.

Materials and methods Animals Sixty adult male Sprague Dawley rats, weighing 300-400 g (Centre d’Elevage Roger Janvier®, Le Genest Saint Isle, France), were housed in smooth-bottomed plastic cages at 22 °C with a 12-h light/dark cycle. Food (Safe®, Augy, France) and water were available ad libitum. An acclimation period of 1 week was allowed before the initiation of the experiment. All animals were weighed before each experiment step. Animals were randomized into four groups: (1) the gold standard group (GS) (n=15), in which the nerve segment was cut, reversed, and reimplanted between the nerve stumps before being sutured using three epineurial sutures; (2) the type I/III collagen group (CG-I/III) (n=15), in which a type I/III collagen conduit was grafted; (3) the type I collagen group (CG-I) (n=15), in which a type I collagen conduit bridged the 10-mm-long gap, and (4) the sham group (n= 15), in which a surgery was performed without injuring the peroneal nerve. Ethical approval Anesthesia and surgical procedures were performed according to the French law on animal care guidelines, and the Animal Care Committees of Aix-Marseille Université (AMU) and the Centre National de la Recherche Scientifique (CNRS) approved our protocols. Individuals conducting research were listed in the authorized personnel section of the animal research protocol or added to a previously approved protocol. Furthermore, experiments were performed following the recommendations provided in the Guide for Care and Use of Laboratory Animals (U.S. Department of Health and Human Services, National Institutes of Health) and in accordance with the European Community’s council directive of 24 November 1986 (86/609/ EEC). No clinical sign of pain or unpleasant sensation (i.e., screech, prostration, hyperactivity, anorexia) and no paweating behavior were observed during the study. At the end of the experiments, animals were sacrificed by an intra-arterial overdose (3 ml) of sodium pentobarbital solution (0.6 g.kg -1, Nembutal®, Sanofi Santé Animale, Libourne, France). Surgical protocol For the initial surgery, rats were deeply anesthetized with an intra-peritoneal injection of chloral hydrate (0.5 g.kg-1, Sigma, St. Louis, MO, USA) and an additional 0.1 ml of anesthesia was given approximately every 20 min (0.5 g/10 ml). The central temperature was maintained at a constant (around

Acta Neurochir

37 °C) with a homeothermic blanket (Harvard Apparatus®, Holliston, USA) driven by a rectal thermal probe. Atropine (1 ml.kg-1) was subcutaneously administered to reduce secretions. Surgical procedures were aseptically performed with the aid of a dissecting microscope. Rats were positioned in the ventral decubitus position, and the peroneal nerve from the left hindlimb was dissected free from the surrounding tissues at a length of 3–4 cm. In the lesioned groups (GS, CG-I/III and CG-I), the peroneal nerve was cut 5 mm away from its inserting point into the tibialis anterior muscle. Then, a 10-mm-long nerve segment was removed. In the GS group, the removed nerve segment was reversed and reimplanted in between the nerve stumps using three epineural sutures in each extremity (Ethilon® 10– 0, Ethicon Inc., Johnson & Johnson Company, Auneau, France). In the CG-I/III and CG-I groups, the nerve gap was bridged by a 10-mm-long type I/III collagen guide or a type I collagen guide, respectively. Each extremity was sutured onto the nerve stumps by three epineurial sutures. An additional sham group was also operated on with a mock surgery, without injuring the peroneal nerve. Muscles and skin were stitched (Vicryl® 6–0 and 4–0 respectively, Ethicon Inc., Johnson & Johnson Company, Auneau, France) and the animals were replaced in individual cages. Collagen guides: CG-I/III and CG-I Two nerve guides called CG-I/III and CG-I were fabricated (Orthomed S.A. Saint-Jeannet, and Biom’Up® S.A., Lyon, France) from highly purified porcine collagens following the respective patented process (respectively, FR2902661 and EP2424579). Briefly, for CG-I/III (ID Ø: 600 μm, wall: 100 μm), a smooth mold was coated with a homogeneous solution of collagen, dried, and coated again to reach the final targeted thickness. The guide was then chemically crosslinked and dried again. The collagen tube wall constituted a succession of continuous cylindrical and coaxial collagen films. This innovative fabrication method resulted in a particular microscopic structure in which collagen molecule spacing was very narrow. Type I collagen guides (CG-I) were manufactured (ID Ø: 600 μm, wall: 100 μm) using type I collagen purified from porcine tendons. Briefly, a concentrated aqueous collagen dispersion with an oxidized polysaccharide (crosslinker) was realized and put in a cylindrical mold. The crosslinking reaction was allowed using an alkaline atmosphere. The gel obtained was then removed from the mold and dried using solvent and air to form a tube. The collagen tube was comprised of a single layer of collagen. As a consequence of the manufacturing process, the macroscopic structure was smooth and the nerve guides both

exhibited (1) limited swelling, (2) cell-tight properties, and (3) a good mechanical resistance and flexibility that made it easy to handle and suture. The structural stability and the mechanical strength of the nerve guides were increased by the crosslinking processes. The aqueous aldehyde crosslinking for CG-I/III and the oxidized polysaccharide crosslinking for CG-I allowed us to control the rate of in vivo resorption. Final resorption was adjusted to meet the technical requirements for nerve regeneration. Moreover, the production process of each collagen contains a viral inactivation step that allows a full viral safety [10]. Collagen conduit biocompatibility was evaluated positively using the International Organization for Standardization (ISO) document 10993 (ISO 10993 Technical Committee, 2007), and reached agreement for UE marking. Biocompatibility was assessed by quantifying tissue inflammation and Schwann cell colonization, survival, and migration (unpublished data from Biom’Up). While the macroscopic structure was similar in the two collagen guides tested in this study, their collagen compositions were different. The collagen guide used in the CG-I/III group was made of highly purified type I/III collagens derived from porcine skin, while the guide of the CG-I group consisted of a purified porcine type I fibrous collagen derived from porcine tendons. Type I fibrous collagen provided strength to the collagen guide and consequently allowed easier suturability (data from Biom’Up). Functional assessment of hind limb recovery All the animals were familiarized to go through a walking track apparatus (150-cm-long, 9-cm-wide, and 40-cm-high) to perform two functional tests during the pre-surgical week and every even week during the 12 weeks post-surgery [1,5]. Heavy lighting was provided with two 500-W spots, and a dark box was positioned at the chamber’s end to promote walking. Peroneal Functional Index (PFI) The Peroneal Functional Index (PFI) is a formula (PFI=174.9[(ePL-nPL)/nPL] + 80.3[(eTS-nTS)/nTS]-13.4) to evaluate the functional condition of the peroneal nerve based on footprint measurements of walking rats [3]. Footprint length (PL, or longitudinal distance between the tip of the longest toe and the heel) and total toe spreading (TS, or cross-sectional distance between the first and fifth toes) are the main two factors modified by the peroneal nerve transection due to motor loss of the toe extensors, foot dorsiflexors, and everters [3]. In order to do so, the animal hind feet were dipped into Chinese ink (Swop-Pads®, Trodat, France) and the footprints were recorded weekly on paper track and manually analyzed. The normal/unoperated (N) footprints and the contralateral experimental (E) footprints were compared.

Acta Neurochir

Kinematic analysis of movement Before recording, both the lower back and the left hind limb of the rat were shaved. The knee joint (condylus femoralis), the ankle joint (lateralis malleolus) and the fifth metatarsal head were marked with a black permanent marker. A numerical camcorder (Basler A602fc High Speed Camera, Basler Vision Technologies, Basler AG, Ahrensburg, Germany) with a full resolution of 656 (horizontal) × 490 (vertical) was positioned perpendicular to the vertical plane of the chamber to get a sagittal view of the gait cycles. The rats were placed in the chamber and several walking cycles were recorded until they arrived in the dark box. Each camcorder field allowed visualizing an 80-cm-long zone in the middle of the glass chamber. Kinematic data were recorded with a 100-Hz acquisition frequency using a 2D movement and behavior system analysis software (Simi Motion®, SIMI Reality Motion Systems GmbH, Unterschleissheim, Germany). For each rat, gait trials were repeated until two or three satisfactory gait cycles without pause were obtained in a single run. During the off-line analysis, the two-dimensional positions of the anatomical markers were tracked manually on each frame (Simi Motion®). Kinematic data were smoothed using a cubic smoothing spline procedure (Simi Motion®). The gait cycles were divided into a stance and a swing phase based on the position of the toe marker. The ankle-joint angle, defined as the intersection between the lines extended from the knee joint to the ankle joint and from the ankle joint to the metatarsal head, was analyzed through the swing phase. The minimum and the maximum ankle angles were then calculated. Gait trials were performed once pre-surgery and weekly during 12 weeks post-surgery. Processing of the gait parameters was performed using Matlab® software (Mathworks Inc., Massachusetts, USA).

Electrophysiological assessment Twelve weeks post-surgery, rats were re-anesthetized by an intra-peritoneal injection of urethane (120 mg/kg, Sigma, St. Louis, MO). A tracheotomy was performed in order to artificially ventilate the rats (Harvard volumetric pump: rate 40– 60 min, tidal volume 2–4 ml; Southmatick, MA USA) and to evaluate the modification in ventilation due to electrically induced-fatigue (EIF). A catheter was inserted into the right femoral artery and pushed up to the fork of the abdominal aorta. This catheter, which did not obstruct the blood flow to the left lower limb muscles, allowed potassium chloride (KCl) injections. Animals were placed on a stereotaxic apparatus (Model 902, David Kopf Instrument, Tujunga, California, USA). The left peroneal nerve was dissected free from surrounding tissues and tendons were left intact. Core temperature was maintained at 36 °C using a heating pad controlled by a thermostat driven by a rectal probe. To prevent dehydratation, subcutaneous injections of 0.5 ml of isotonic

glucose-saline, pH balanced to 7.4, were administered at 30min intervals. The animal’s forelimbs and hindlimbs were secured with tape while knee and ankle joints were firmly held by clamps on a horizontal support to prevent any movement during nerve and muscle stimulations. Special care was taken not to apply any unnecessary pressure or stretch to any part of the hindlimbs. Ventilatory rate According to our previous study [7], changes in ventilatory rate were recorded after tibialis anterior muscle stimulation. The experiments were performed after regional circulatory occlusion that isolated and maintained the neural drive, but abolished humoral communication. To elicit electrically-induced fatigue (EIF), rhythmic muscle contractions were produced by the neurostimulator (Grass S88K®) that delivered pulse trains through a pair of intramuscular electrodes (pulse duration: 0.1 ms; frequency: 10 Hz, i.e., five shocks in each 500 ms train; duty cycle: 500/ 1500 ms, voltage range: 5–8 volts). The voltage was supramaximal, i.e., 20 % higher than the one used to elicit a maximal twitch. Fatigue was assessed from the decay of force throughout the 3-min EIF period by the use of a strain gauge attached to the tibialis anterior muscle’s distal tendon (Micromanometer 7001®, Ugo Basile Srl., Biological Research Apparatus, Comerio VA, Italy). The muscle contraction force was measured during muscle electrical stimulation. As previously shown, low frequency stimulation of the tibialis anterior muscle was used in order to strongly activate metabosensitive afferent fibers [4]. Based on this assumption, a previous study demonstrated that EIF led to respiratory adjustments [7]. Ventilatory rate was recorded from 2 minutes before EIF (pre-EIF) to 5 minutes after EIF (post-EIF) by a thermocouple inserted into the tracheal canula. The ventilatory rate was expressed in cycle/min. The change in post-EIF ventilatory rate was expressed in percentage [Δcycle/min (%)] of the pre-EIF rate. Metabosensitive activity The proximal portion of the peroneal nerve was cut to exclude efferent discharges during nerve recording. To analyze the cumulative response of the metabosensitive afferent fibers, the distal segment of the nerve was positioned on a bipolar tungsten hook electrode and immersed in paraffin oil. The neural signals were referred to a ground electrode implanted in a nearby muscle, amplified (50–100 K), and filtered (30 Hz to 10 kHz) with a differential amplifier (P2MP® SARL, Marseille, France). The afferent discharge was recorded (Biopac MP150® and AcqKnowledge software, Biopac Systems, Inc., Goleta, USA) and fed into pulse window discriminators (P2MP® SARL, Marseille, France), which simultaneously analyzed afferent populations. The output of these discriminators provided noise-free tracings (discriminated units), which were counted using a data analysis system (Biopac AcqKnowledge® software) at 1-sec

Acta Neurochir

intervals (in Hz) and then displayed on a computer. Distinct concentrations of KCl (1, 5, and 10 mM) in 0.5 ml were randomly injected into the contralateral artery while nerve activity was continuously recorded. The dose–response curve to KCl from peroneal nerve was determined in the present experiment. Metabosensitive activity was expressed in afferent discharge frequency (impulse.s-1). Change in postinjection metabosensitive activity was expressed in percent [Δimpulse.s-1 (%)] of the pre-injection afferent discharge frequency. Muscle weight measurements After the electrophysiological study, rats were then sacrificed by a 1-ml intra-arterial overdose of pentobarbital sodium solution (Pentobarbital Sodique®, 60 mg.kg−1 Sanofi Santé Animal, Sanofi France, Paris France). Right and left tibialis anterior muscles were harvested and immediately weighed on a precision scale (Navigator™ N30330 model, OHAUS Corp., Parsippany, NJ, USA). Comparisons of muscle mass were performed using a left denervated muscle weight/right healthy muscle weight ratio [26]. Histology For axon counting, peroneal nerves (n=8 per group) were harvested free from surrounding tissues, rapidly washed in phosphate buffer (PBS)(Gibco®, Life Technologies Corp. CA, USA) and immersed in a 4 % paraformaldehydecontaining PBS solution during 24 hours at 4 °C. Peroneal nerves were sectioned in three parts (proximal end, middle segment, and distal end) before being immunostained. After paraffin embedding, sections (5-μm thickness±0.5 μm) were cut on a microtome (RM2155, Leica®, Solms, Germany) and collected on coated slides (SuperFrost Plus®, Gerhard Menzel-Glaser, GmbH, Germany). Sections were immunostained with a mouse monoclonal antibody to the light chain of the neurofilament (NF-L) protein (Neurofilament Protein Clone 2 F11, MO762, Dako®, Glostrup, Denmark)(dilution: 1/100) using a robot (Benchmark® XT, Ventana Medical Systems, Inc., Arizona, USA). After washing, an appropriate biotinylated-conjugated secondary antibody was applied to the sections. The final staining step was performed using diaminobenzidine (Ventana® iview DAB 760 091, Ventana Medical Systems, Inc., Arizona, USA). For the g-ratio assessment, peroneal nerves (n=8 per group) were harvested free from surrounding tissues, washed in phosphate buffer (PBS; Gibco®, Life Technologies Corp., Saint Aubin, France) and immersed in a 4 % glutaraldehydecontaining PBS solution for 24 hours. Samples were stained with p-phenylenediamine (PPD). After inclusion, semithin sections (0.8 μm) were cut using an ultramicrotome (Ultracut® R, Leica, Solms, Germany) and collected on

coated slides (SuperFrost Plus®). After being dried for 12 hours on a hot plate, sections were stained with a PPDethanol solution (70°), washed in distilled water, dried for 5 hours on a hot plate, and mounted with glycerol-containing medium (Glycergel®, DakoCytomation, Glostrup, Denmark). Slides were examined using an optical microscope (Eclipse® E800, Nikon, Champigny-sur-Marne, France) that was associated with high-resolution, color digital camera (DXM 1200, Nikon). The slides were digitized and analyzed by Histolab software (Alphelys®, Plaisir, France). The axon counting was performed by the robot and therefore, observerassociated biases were avoided. For the g-ratio assessment, slides were coded, five regions of interest in each section were randomly chosen, and data analysis was blindly performed. Statistical analysis Data processing was performed using a software program (SigmaStat® 2.03, Statistical software, San Jose, CA, USA). Data were expressed as mean±SEM. PFI scores and minimum and maximum ankle angles were averaged and compared in pre-surgical and post-surgical sessions. Ventilatory rate (cycle/min) and metabosensitive afferent-fiber activity in response to EIF and KCl injections were averaged and compared before and after respective stimulus. In order to determine differences between groups over time, a two-way repeated measures analysis of variance (ANOVA, group x time) was calculated. Post-hoc multiple group comparisons were performed with a Student-NewmanKeuls Method. Differences between groups in (1) the ventilatory rate adjustment in response to EIF (in % of pre-EIF ventilatory rate), (2) the metabosensitive afferent-fiber activity adjustment in response to EIF and KCl injections, and (3) the muscle atrophy were evaluated by a one-way ANOVA (group). Posthoc multiple group comparisons were performed with a Student-Newman-Keuls Method. Differences in (1) axon number, (2) axonal diameter, and (3) g-ratio were determined by a two-way ANOVA (group x position). Post-hoc multiple group comparisons were performed with a Student-Newman-Keuls Method. Results were considered significant if the p-value fell below 0.05.

Results Functional assessment of hind limb recovery The Peroneal Functional Index (PFI) PFI results before surgery were similar in all groups. In the sham group, the mean PFI was −12.26±7.43 and remained constant during the

Acta Neurochir

protocol period. In all operated groups, PFI values significantly decreased from the second week to the end of the protocol compared to the sham group (p

Functional recovery after repair of peroneal nerve gap using different collagen conduits.

Currently, autologous nerve implantation to bridge a long nerve gap presents the greatest regenerative performance in spite of substantial drawbacks. ...
433KB Sizes 0 Downloads 0 Views