Experimental Neurology 255 (2014) 1–11

Contents lists available at ScienceDirect

Experimental Neurology journal homepage: www.elsevier.com/locate/yexnr

Assessment of sensory thresholds and nociceptive fiber growth after sciatic nerve injury reveals the differential contribution of collateral reinnervation and nerve regeneration to neuropathic pain Stefano Cobianchi, Julia de Cruz, Xavier Navarro ⁎ Group of Neuroplasticity and Regeneration, Institute of Neurosciences and Department of Cell Biology, Physiology and Immunology, Universitat Autònoma de Barcelona, Bellaterra, and Centro de Investigación Biomédica en Red sobre Enfermedades Neurodegenerativas (CIBERNED), Spain

a r t i c l e

i n f o

Article history: Received 25 November 2013 Revised 31 January 2014 Accepted 10 February 2014 Available online 16 February 2014 Keywords: Collateral sprouting Nerve regeneration Skin innervation Neuropathic pain Hyperalgesia Sensory nerve fibers Algesimetry test

a b s t r a c t Following traumatic peripheral nerve injury reinnervation of denervated targets may be achieved by regeneration of injured axons and by collateral sprouting of neighbor undamaged axons. Experimental models commonly use sciatic nerve injuries to assess nerve regeneration and neuropathic pain, but behavioral tests for evaluating sensory recovery often disregard the pattern of hindpaw innervation. This may lead to confounding attribution of recovery of sensory responses to improvement in sciatic nerve regeneration instead of collateral reinnervation by the undamaged saphenous nerve. We used a standardized methodology to assess the separate contribution of collateral and regenerative skin reinnervation on sensory responses. Section and suture of the sciatic nerve induced loss of sensibility in the lateral and central areas of the injured paw, but nociceptive responses rapidly recovered by expansion of the intact saphenous innervation territory. We used electronic Von Frey and Plantar test devices to measure mechanical and thermal withdrawal thresholds in specific sites of the injured paw: lateral site innervated by the sciatic nerve, medial site that remained innervated by the saphenous nerve, and central site originally innervated by the sciatic nerve but affected by saphenous sprouting. After sciatic section, signs of early hyperalgesia developed in medial and central paw areas due to saphenous sprouting and expansion. The regenerating sciatic nerve fibers reached the paw at 3–4 weeks and a late mechanical hyperalgesia was observed at the lateral site. Immunohistochemical staining of sensory fibers innervating the medial and lateral areas revealed a different pattern of skin reinnervation. Hypersensitivity in the intact saphenous nerve area was paralleled by early fiber sprout growth in the subepidermal plexus, but not entering the epidermis. On the other side, late sciatic hyperalgesia was accompanied by gradual skin reinnervation after 4 weeks. The standardization of algesimetry testing in sciatic nerve injury models, as proposed in this study, provides a suitable model for studying in parallel neuropathic pain and sensory nerve regeneration processes. Our results also indicate that collateral sprouting and axonal regeneration contribute differently in the initiation and maintenance of neuropathic pain. © 2014 Elsevier Inc. All rights reserved.

Introduction Peripheral nerve injuries induce functional deficits caused by degeneration of injured axons, that can be recovered by two endogenous mechanisms: the reinnervation of denervated targets by regeneration of injured axons, and the reinnervation by collateral branching of undamaged axons (Navarro et al., 2007). Each of these mechanisms may also contribute to the appearance of secondary effects, such as variations of sensory thresholds, hyperreflexia and neuropathic pain. Despite numerous studies on potential pain targets and involvement of hundreds of genes (LaCroix-Fralish et al., 2007), few basic discoveries ⁎ Corresponding author at: Facultat de Medicina, Universitat Autònoma de Barcelona, E-08193 Bellaterra, Spain. Fax: +34 935812986. E-mail address: [email protected] (X. Navarro).

http://dx.doi.org/10.1016/j.expneurol.2014.02.008 0014-4886/© 2014 Elsevier Inc. All rights reserved.

have been translated so far from rodent models into effective therapy for neuropathic pain after nerve lesions. As recently observed (Mogil et al., 2010), one possibility is the non-predictability of preclinical animal pain models. Several animal models of peripheral nerve injury reflecting human neuropathic pain syndromes have been developed and characterized. These models mostly use injuries to the sciatic nerve, and behavioral testing used to evaluate the recovery of sciatic sensory function typically measures the hypersensitivity to touch stimuli by means of the Von Frey test, and/or withdrawal responses to a heat source (Wood et al., 2011). On the other hand, also studies investigating treatments for enhancing nerve regeneration most frequently use the sciatic nerve of rodents (Rodríguez et al., 2004) and algesimetry tests for assessment of sensory recovery. However, in these studies the compensatory mechanism of collateral reinnervation is often neglected, although it has been

2

S. Cobianchi et al. / Experimental Neurology 255 (2014) 1–11

extensively shown to strongly affect the recovery of sensibility (Devor et al., 1979; Inbal et al., 1987). We particularly argue that measures of sensibility obtained by algesimetry tests may be not properly conducted in studies employing sciatic nerve injury models. Very often assessment of sensory thresholds is made by grossly stimulating the injured paw without consideration of the anatomical innervation territories, and then concluding that recovery of sensory responses is only due to an improvement in nerve regeneration rather than to collateral reinnervation. After sciatic nerve injuries, compensatory reinnervation of the hindpaw denervated skin is mediated by collateral sprouting of the saphenous nerve (Devor and Govrin-Lippmann, 1979; Kingery and Vallin, 1989; Kingery et al., 1994), which normally provides innervation to the most medial side of the rat paw, and partly overlaps with the tibial nerve (Decosterd and Woolf, 2000). Indeed, the decrease of withdrawal thresholds to mechanical and heat stimulation of medial and central areas of the paw skin after sciatic nerve lesions has been related to saphenous sprouting (Kingery et al., 1994; Vallin and Kingery, 1991). This phenomenon causes early and enhanced responses to stimulation that can be misinterpreted as the result of sciatic nerve fiber regeneration. Several studies have attempted to explain the occurrence of neuropathic pain symptoms related to neuroplastic changes in the innervation of the skin. Following sciatic nerve lesion, there is a loss of dermal and epidermal sensory fibers in the denervated area, with a course of sciatic reinnervation dependent on the severity of the injury (Hsieh et al., 2000; Navarro et al., 1995, 1997). At the same time, the spared sensory fibers rapidly increase in density (Ma and Bisby, 2000) with peptidergic calcitonin gene-related peptide (CGRP) immunoreactive fibers hyperinnervating the footpad skin (Duraku et al., 2012; Grelik et al., 2005) compared with other types of fibers (Duraku et al., 2013). Since the sprouts produced by intact saphenous nerve fibers are induced by target-derived nerve growth factor (NGF) (Diamond et al., 1987; Mendell et al., 1999), collateral reinnervation is mostly expected by peptidergic nociceptors that express the TrkA receptor for NGF (Wolf and Ma, 2007). The maintenance of neuropathic symptoms has been attributed to this altered pattern of reinnervation (Duraku et al., 2013), although it is still unknown how collateral sprouts growing through the skin induce sensory changes before and after regeneration of the sciatic nerve. Since the return of nociceptive responses anticipates the reinnervation of the epidermis, a different fiber growth profile through dermis and epidermis has also been suggested (Navarro et al., 1997; Verdú and Navarro, 1997), so nociceptors reinnervating sub-epidermal layers of the skin may mediate early stimulus-dependent pain responses. In this study we made a standardized evaluation of nociceptive reinnervation and changes in sensory thresholds differently induced by saphenous nerve sprouting and sciatic nerve regeneration. We show that differentiation of the contribution of these two processes is essential for the assessment of neuropathic pain. We used the sciatic nerve section and suture repair model, which divides the rat paw into a saphenoussensitive and a sciatic-insensitive area, and stimulated specific skin test sites. We eliminated the saphenous nerve at different times before and after the return of sciatic reinnervation, and showed that stimulusinduced pain depends on the area of stimulation and has a time course otherwise attributable to the two distinct pathways. We also characterized the time course of skin reinnervation by immunohistochemical labeling, and found that neuropathic pain is related to early changes in the sub-epidermis differently induced by saphenous sprouting and sciatic regeneration. Materials and methods Animals and surgery Adult female Sprague–Dawley rats (240 ± 30 g) were housed in standard polyethylene cages (four per cage) and kept on standard laboratory food and water ad libitum with a light–dark cycle of 12 h. All

experimental procedures were approved by the Ethics Committee of our institution, and followed the European Communities Council Directive 86/609/EEC and the guidelines of the Committee for Research and Ethical Issues of IASP (Zimmermann, 1983). Rats were anesthetized with pentobarbital (40 mg/kg i.p.). The right sciatic nerve was exposed at the mid thigh, transected at 92 mm from the tip of the third toe, and repaired by epineurial sutures (10–0), maintaining the fascicular alignment of tibial, peroneal and sural branches. The wound was closed in two layers and disinfected. Rats were kept in a warm environment until complete recovery from anesthesia. They were observed daily for behavioral signs indicative of pain, including signs of autotomy of the injured limb. In this study we used a total number of 21 rats. Nine rats were used for sensory testing during the complete follow-up. Samples for immunohistochemical analysis were collected from the same animals euthanized the day after the last functional test. At two selected intervals (14 and 28 days) after sciatic lesion, under isoflurane anesthesia, the right saphenous nerve of other rats (6 at each time point) was exposed by a small skin incision at the thigh, and transected. A fragment of approx 10 mm was excised from the distal stump to ensure no regeneration. The wound was sutured and disinfected. The following day, the animals were tested to verify complete loss of responses. Assessment of skin nociceptive reinnervation The progression of nociceptive reinnervation of the hindpaw was assessed by means of the pinprick test (Navarro et al., 1994) performed at different days post injury. Awake animals were gently kept in a cloth with the injured paw facing upwards, and the plantar skin was stimulated with a blunt needle progressively from distal to proximal at distinct sites (see Fig. 1B). Each site was stimulated twice, and responses were recorded as positive only when clear pain reactions (as fast withdrawal and vocalization) were triggered to the stimulation. Positive responses were taken as sign of skin functional reinnervation. An index of sprouting, calculated as the rate of responsive areas over the total areas stimulated by pinpricking at different times, indicated the increasing contribution of saphenous sprouting to nociception. Statistical comparison was made with one-way ANOVA and Bonferroni post-hoc test. The criterion for statistical significance was p b 0.05. Algesimetry tests for sensory threshold measurement One week before surgery, all the animals were habituated to experimental devices for nociceptive baseline threshold measurement. The algesimetry tests for mechanical and thermal stimuli were performed on both hindpaws before and at 3, 7, 14, 21, 28, 40, 50 and 60 days post-injury (dpi). Sensibility to a normally non-noxious mechanical stimulus was measured by means of an electronic Von Frey algesimeter (Bioseb, Chaville, France). Rats were placed on a wire net platform in plastic chambers 10 min before the experiment for habituation. We stimulated three different regions of the hindpaw plantar surface: the lateral (normally innervated by tibial and sural branches of the sciatic nerve), the central (normally innervated by the tibial nerve) and the medial (covered by tibial and saphenous nerves) (Casals-Díaz et al., 2009) (Fig. 1A). For each region, we chose a specific test site on the skin at the outer edge of the most proximal medial and lateral footpads, or at the center of the foot (Fig. 2A). The medial and lateral test sites were chosen as selective saphenous and sciatic sensory territories respectively (Cobianchi et al., 2013); the central test site was chosen as a non-selective test site. The experimenter was blind regarding the objectives of the testing, but trained in the correct execution of algesimetry tests. By using the electronic Von Frey algesimeter, a 0.4 mm non-noxious pointed probe was gently applied to each test site, and then slowly increasing the pressure. The threshold was expressed as the force (in grams) at which rats withdrew the paw in response to the stimulus. A cutoff force was set to 40 g

S. Cobianchi et al. / Experimental Neurology 255 (2014) 1–11

3

Fig. 1. Reinnervation of rat plantar skin after sciatic nerve section and repair. (A) The rat hind paw plantar skin is innervated by saphenous, tibial and sural nerves (tibial and sural are branches of the sciatic nerve), covering distinct territories shown in the figure in different gray shadows, with overlapping territories in white. (B) Schema indicating the different areas of the plantar surface that were stimulated by pinpricking to assess functional innervation. Nomenclature of the considered areas according to Kennedy et al. (1988). (C) Mapping of responses to pinprick in foot skin areas at different days post-injury (dpi), represented in a gray scale ranking from 0% responsive rats (white) to 100% responsive rats (black), shows the progression of nociceptive reinnervation from medial to central and lateral areas and toes. (D) Mapping of responses to pinprick after cutting the saphenous nerve at 2 different times (16 and 29 dpi) reveals the contribution of saphenous collateral sprouting to hindpaw skin reinnervation before and after sciatic nerve regeneration. (E) An index of sprouting, calculated as the rate of responsive areas over the total areas stimulated by pinpricking at different times after sciatic nerve injury, indicates the increasing contribution of saphenous sprouting to nociception (one-way ANOVA, Bonferroni post-hoc comparison; ***p b 0.001, 3 dpi vs. 28, 50 and 60 dpi; ### p b 0.001, 15 and 29 dpi vs. 14 and 28 dpi respectively).

at which stimulus lifted the paw with no response. Mechanical nociceptive threshold was taken as the mean of three measurements per test site, with 5 min interval between each measurement. Thermal sensibility was assessed by means of a Plantar test algesimeter (Ugo Basile, Comerio, Italy). Rats were placed into a plastic box with an elevated plexiglass floor. The beam (6 mm diameter) of a lamp was focused on the hindpaw plantar surface, and pointed to the same test sites as above. We paid attention to avoid cross-stimulation of distinct skin territories; particularly, in the medial and lateral test sites, we pointed the stimulating beam to not affect beyond the skin borders of α and β footpads as showed in Fig. 2A. Intensity was set to low power (40 mW/cm2) with a heating rate of 1 °C/s, to induce activation of unmyelinated fibers (Le Bars et al., 2001). A cutoff time of stimuli was set at 20 s to prevent tissue damage. Heat pain threshold was

calculated as the mean of three trials per test site, with 10 min resting period between each trial, and expressed as the latency (in seconds) of paw withdrawal response. Particular attention was paid during the execution of tests to avoid stimulating the same paw consecutively, and when the animal was not in a balanced posture. Data are presented as mean ± SEM. Statistical comparisons for nociceptive thresholds were made with repeated measures ANOVA and Bonferroni post-hoc test. The criterion for statistical significance was p b 0.05. Immunohistochemistry of paw skin innervation Plantar pads corresponding to the medial and lateral algesimetry test sites (named as α and β in Fig. 1B) were removed from both

4

S. Cobianchi et al. / Experimental Neurology 255 (2014) 1–11

S. Cobianchi et al. / Experimental Neurology 255 (2014) 1–11

hindpaws of rats euthanized at 3, 8, 16, 29 and 60 days after sciatic nerve injury. Cryotome sections 60 μm thick were processed together as previously described (Navarro et al., 1997; Verdú and Navarro, 1997). We used antibodies against protein gene product 9.5 (PGP, rabbit, 1:1000, ABDserotec) and calcitonin gene-related peptide (CGRP, goat, 1:500, Abcam) to label all the nerve fibers and the peptidergic fibers, respectively. Immunolabeling against growth-associated protein 43 (GAP-43, rabbit, 1:1000, Millipore) was performed in order to visualize the growing axons. Four sections of four samples per group were used for quantification of skin innervation. Images were captured with a Zeiss LSM 700 confocal microscope, and analyzed using ImageJ software (NIH, Bethesda, USA). To avoid measuring degenerating fibers, that still show some fading immunoreactivity (Navarro et al., 1997), we selected only nerve fibers with immunoreactivity above the threshold of the completely denervated lateral footpad at 3 dpi. Intraepidermal nerve fibers (IENFs) were counted under an epifluorescence microscope in a footpad epidermis field of 500 μm adjacent to the algesimetry test site (Fig. 3A). The mean length of sub-epidermal nerve fibers (SENFs) was measured on reconstructed images of the entire footpad section. Data were analyzed using two-way ANOVA and Tukey post-hoc test. The criterion for statistical significance was p b 0.05. Results Mapping of plantar skin nociceptive reinnervation The rat plantar hindpaw skin is normally innervated by saphenous, tibial and sural nerves (tibial and sural are branches of the sciatic nerve), with overlapping territories between them (Fig. 1A). Transection of the sciatic nerve divides the rat paw into a saphenous-sensitive (medial) and a sciatic-insensitive (lateral) area. Suture repair permitted the sciatic nerve to regenerate and to reinnervate the paw skin after 3–4 weeks. Pinprick stimuli were applied at 17 different plantar areas (Fig. 1B). For each area, we plotted the mean percentage of rats with positive responses over time after sciatic nerve injury (Fig. 1C). At 3 dpi, all the rats had positive responses only in the medial side of the sole and toes 1 and 2, which are normally innervated by saphenous and tibial nerves. A few rats had also positive responses in α and A footpads, also partially innervated by the saphenous nerve. Stimulation of these medial footpads induced withdrawal response in most of the rats at 7 and 14 dpi, whereas new areas in the central medial sole and B footpad became gradually responsive. At 21 dpi almost all the rats had responses in the medial half of the plantar surface except toe 3. One week later, central lateral areas, D and β footpads and toe 3 became responsive in some animals, and at days 50–60 almost all the sciatic nerve territory was reinnervated, except for toe 5 and the most lateral area, corresponding to the sural nerve territory, in some rats. To analyze the separate contribution on nociceptive responses of collateral and regenerative reinnervation of the skin, we transected the saphenous nerve at two different times, and repeated the mapping test the following day. As shown in Fig. 1D, the entire plantar skin was unresponsive after cutting the saphenous nerve at 14 dpi, demonstrating that it exclusively mediated the early nociceptive responses after sciatic nerve injury. However, when cutting the saphenous nerve at 28 dpi, we found that ankle, central areas and lateral footpads remained responsive in most of the rats the day after (Fig. 1D). This can be attributed to the arrival of newly regenerated axons from the sciatic nerve

5

between 3 and 4 weeks after injury. Notably, at 4 weeks, positive responses in distal areas that are solely innervated by the tibial nerve in the normal condition, such as the B footpad and the toe 3, completely disappeared after saphenous cut. In order to assess the contribution of collateral sprouting from the saphenous nerve to nociception before and after reinnervation by the sciatic nerve, a composite score was calculated as the ratio of responding versus total stimulated areas (Fig. 1E). The proportion of responsive areas significantly increased from about 23% at 3 dpi to 53% at 28 dpi (p b 0.001), to 73% at 50 dpi (p b 0.001) and to 84% at 60 dpi (p b 0.001). Interestingly, the elimination of the saphenous nerve at 29 dpi still left 22% areas responding to pinprick (p b 0.001 vs. 28 dpi), implying that at 28 dpi the saphenous nerve was mediating innervation of about 30% of the skin area, a larger surface than that at 3 dpi. Nociceptive thresholds measured at different skin innervation territories Fig. 2 shows sensory thresholds measured at three different test sites (Fig. 2A) as mentioned above. Withdrawal responses to Von Frey mechanical stimulation in normal paws were found at 30–31 g. After sciatic nerve injury, no withdrawal responses were found at 3, 7 and 14 dpi in the sciatic lateral test site (Fig. 2B, top charts). Since muscles proximal to the knee in the injured leg have spared innervation and allow withdrawal of the paw when innervated skin areas are stimulated, the lack of withdrawal response is attributable to the absence of sensation to painful stimuli. At 21 dpi responses to stimuli reappeared in a few rats but at higher force than in the contralateral intact paw (p b 0.001). At later times, the recovery of sensitivity progressed, but with abnormal responses below the normal threshold (p b 0.001 at 28, 40, 50 and 60 dpi), reflecting a state of mechanical hyperalgesia similar to that reported after reinnervation in sciatic nerve crush and spared nerve injury models (Casals-Díaz et al., 2009; Vogelaar et al., 2004). Withdrawal responses to heat stimulation had a latency around 14 s in the intact hindpaw. As observed for mechanical stimuli, in the denervated paw there were no responses to thermal stimuli applied at the sciatic lateral region at 3, 7 and 14 dpi. At 21 dpi recovery of sensitivity started in some rats, but at longer than normal withdrawal latency (p b 0.001). Nociceptive thermal threshold of the injured paw progressively normalized by 4 weeks. Unlike the mechanical response, no thermal hyperalgesia developed until the end of follow-up at 60 dpi. The time course from hypoesthesia to hyperalgesia observed at the lateral side of the paw reflects the expected course of denervation and regeneration of the injured sciatic nerve. However, compared with the lateral site, responses to thermal and mechanical stimuli applied at the central test site (Fig. 2B, mid charts), used in the study as non-specific site, started to recover already at 3 (p b 0.001) and 7 dpi (p b 0.001) respectively, and they were similar to contralateral values after 2–3 weeks. Then, early responses recorded at central areas of the paw may be erroneously attributed to reinnervation by sciatic regeneration, whereas as demonstrated in the mapping tests, they are due to collateral reinnervation from the saphenous nerve. To confirm this point, we tested the rats at the central site after cutting the saphenous nerve. We observed at 14 dpi a complete loss of responses, and at 28 dpi a significant increase of mechanical and thermal thresholds (Fig. 2C), demonstrating that saphenous nerve mediated the changes in sensory thresholds even if stimuli were delivered to sciatic central areas.

Fig. 2. Sensory thresholds measured with algesimetry tests. (A) Assessment of sensory thresholds was conducted by applying mechanical (Von Frey test) and thermal (Plantar test) stimuli to a medial and a lateral test site, corresponding to selective saphenous and sciatic nerve territories, or to the center of the foot. (B) Mechanical and thermal thresholds measured on the lateral sciatic, medial saphenous and central test sites of the hindpaw (*p b 0.05; **p b 0.01; ***p b 0.001, injured vs. contralateral paw). Sciatic nerve injury induced loss of sensitivity at lateral site (top charts) until 21 dpi when rats started to respond due to nerve regeneration, and subsequently developed mechanical but not thermal hyperalgesia. However, responses to stimuli applied at the central test site (mid charts) started to recover already at 7 dpi, approaching normal values. Early responses recorded at central areas of the paw may be erroneously attributed to reinnervation by sciatic regeneration. Thresholds measured at the medial site in the saphenous territory (bottom charts) were significantly decreased after sciatic nerve injury, revealing marked mechanical and milder thermal hyperalgesia. (C) Mechanical and thermal thresholds in the central test site measured before and after resection of the saphenous nerve at 14 or at 28 dpi (arrows) demonstrated that the saphenous nerve mediated the changes in sensory thresholds even if stimuli were delivered to the central paw site.

6

S. Cobianchi et al. / Experimental Neurology 255 (2014) 1–11

Fig. 3. Sensory innervation of footpad epidermis. (A) Epidermis (E), sub-epidermal plexus (SEp), sweat glands (SG) and algesimetry test site (ts) are represented on a rat footpad section (scale bar = 100 μm). In microscope images, the test site can be easily identified as adjacent to a concavity that the footpad forms with the adjacent skin. (B) Representative images of PGPand CGRP-immunoreactive fibers in the epidermis of a normal footpad and of the medial and lateral test sites taken at 8, 29 and 60 days after sciatic nerve section and repair. Note extensive epidermal denervation, followed by clear reinnervation at 60 dpi. Scale bar = 50 μm. (C) Counts of PGP- and CGRP-immunoreactive intra-epidermal nerve fibers (IENFs) at 3, 8, 16, 29 and 60 dpi show that these fibers are almost completely lost in medial and lateral sites between 3 and 29 dpi. IENFs return to around normal values at 60 dpi, but CGRP-immunoreactive fibers were higher than normal in the medial site (*p b 0.05; **p b 0.01; ***p b 0.001, injured vs. contralateral paw).

S. Cobianchi et al. / Experimental Neurology 255 (2014) 1–11

7

On the other hand, the nociceptive thresholds measured in the saphenous territory were dramatically decreased after sciatic nerve injury (Fig. 2B, bottom charts). Withdrawal threshold to mechanical stimulation in the medial site progressively decreased to about 25% of contralateral values from 3 dpi and remained so during the two month follow-up (p b 0.001 at 3, 7, 14, 21, 28, 40, 50 and 60 dpi). The measurement of heat sensitivity revealed a similar but lower decrease of withdrawal latency in the medial region of the injured paw, by 35% of normal, at 21 dpi. Compared to the contralateral paw, heat threshold was significantly lower at 14 (p b 0.001), 21 (p b 0.01), 28 (p b 0.001) and 50 dpi (p b 0.001), and gradually recovered to near normal at 60 dpi. Taken together, the algesimetry results show that the first pain responses are mediated by the saphenous nerve, that rapidly induced mechanical and thermal hyperalgesia in the saphenous normal territory and then gradually sensitized contiguous sciatic areas, as observed at the central test site. Responses mediated by the regenerated sciatic nerve appeared 3–4 weeks after injury and produced a late mechanical hyperalgesia. Plantar skin reinnervation We showed that the section and suture of sciatic nerve induces opposite changes in nociceptive threshold, followed by a different sensory recovery, if thresholds are measured in different test sites where sciatic and saphenous nerves selectively cater for skin innervation. Then we wanted to analyze if the observed changes were paralleled by changes in the fibers innervating the specific medial and lateral footpad test sites. Fig. 3A shows representative footpad epidermal and subepidermal territories and test site area evaluated for fiber innervation. The test site can be easily identified as adjacent to a concavity that the footpad forms with the flat skin. Immunohistochemical images of PGP immunoreactive profiles showed important changes after sciatic nerve injury (Fig. 3B). The number of IENFs was markedly reduced in the epidermis of both lateral and medial skin samples until 29 dpi (p b 0.001 at 3, 8, 16 and 29 dpi). The epidermis appeared thinner and numerous PGP positive Langerhans cells were found in both medial and lateral sites, as previously reported (Duraku et al., 2013; Stankovic et al., 1999). The first regenerating axons were found from day 29, but only a few reached the epidermis and reformed less corpuscular endings than normal, as previously described (Navarro et al., 1997). At 60 dpi nerve fibers extensively reinnervated medial and lateral sites of the paw, but with a less organized pattern than that in the control skin. The subpopulation of CGRP immunoreactive profiles was also significantly reduced after injury (medial site: p b 0.01 at 3 dpi, p b 0.001 at 8 dpi, p b 0.05 at 16 and 29 dpi; lateral site: p b 0.01 at 3, 8 and 16 dpi, p b 0.05 at 29 dpi) (Fig. 3B). CGRP staining evidenced a more complex pattern of reinnervation at two months after injury in the medial site compared with the lateral site. Counts of IENFs (Fig. 3C) show that PGP- and CGRP- positive fibers almost disappeared from both medial and lateral epidermis during the first month. However, the skin was well reinnervated at 60 dpi, when the mean number of peptidergic IENFs was even significantly higher than normal in the medial site (p b 0.05). This result indicates that nociceptive epidermal hyper-reinnervation may be involved in the maintenance of hypersensitivity at long time from the injury (Duraku et al., 2013). Since the progression of IENF innervation did not reflect the early changes of sensory thresholds observed at the medial site with algesimetry tests, we hypothesized that sprouting fibers that initially grow in deep skin layers at the overlapping territory and reach the sub-epidermal plexus rather than entering the epidermis, may be able to respond to stimulation. We observed peptidergic SENFs elongating from medial test site to the medial footpad dermis and along the basal membrane during the first days following injury (Fig. 4A), whereas in the lateral footpad SENFs immunoreactivity almost disappeared during the first weeks, and was back to normal levels at 60 dpi (Fig. 4B). When

Fig. 4. Sensory innervation of footpad sub-epidermis. (A) Representative images of PGPand CGRP-immunoreactive fibers in the sub-epidermal skin layer (SENFs) of the medial test site, at 3, 8 and 29 days after sciatic nerve injury, show fibers elongating through the sub-epidermal plexus which are longer with time. Scale bar = 100 μm. The PGPand CGRP-positive fiber area in the subepidermal plexus of the medial footpad linearly increased with time, whereas in the lateral footpad SENFs immunoreactivity almost disappeared during the first weeks and reappeared near to normal after 29 dpi (*p b 0.05; **p b 0.01; ***p b 0.001, injured vs. contralateral paw). (B) Measurement of the length of PGP- and CGRP-positive profiles at 3, 8, 16, 29 and 60 dpi. Medial SENFs progressively elongated until 29 dpi, and CGRP positive SENFs even exceeded the normal length at 29 dpi (*p b 0.05; **p b 0.01; ***p b 0.001, injured vs. contralateral paw).

analyzing the immunoreactive profile length (Fig. 4C) in the subepidermal plexus of the medial pad, we found an increase with time; SENFs were significantly shorter at 3 dpi (p b 0.01, PGP; p b 0.05, CGRP) but rapidly elongated until 29 dpi (p b 0.05 CGRP). CGRP positive SENFs significantly exceeded the normal length at 29 dpi, indicating an abnormal

8

S. Cobianchi et al. / Experimental Neurology 255 (2014) 1–11

Fig. 5. Growth phenotype of peptidergic reinnervating fibers. (A) Representative images of double-labeled CGRP- and GAP43-positive fibers in the normal and in the injured skin of the medial test site at 8, 29 and 60 days post-injury show a growing fiber phenotype after injury. Scale bar = 100 μm. (B) Plots of the rate of CGRP/GAP43-immunoreactive SENFs in medial and lateral skin test sites at different times after injury demonstrate that active fiber growth is different in timing at the medial and lateral sites (*p b 0.05, **p b 0.01, ***p b 0.001, injured vs. contralateral paw).

S. Cobianchi et al. / Experimental Neurology 255 (2014) 1–11

prolongation and perhaps a difficulty to cross the basal membrane to enter the epidermis. At 60 dpi, both PGP and CGRP labeled SENFs were found with similar length to normal, suggesting retraction after reinnervation by sciatic nerve fibers. In contrast, in the lateral pad, the SENFs almost completely disappeared, indicating degeneration, and then reappeared with normal length in samples taken at 29 and 60 dpi due to the arrival of sciatic regenerating axons. This result demonstrates that, at least during the first 2 weeks after sciatic injury, SENFs elongating through the medial skin identify sprouting saphenous axons. Growing state of reinnervating fibers We examined the proportion of CGRP-positive fibers co-labeled with GAP43 in order to characterize their growing state in the skin (Fig. 5). Only a few CGRP nociceptive fibers expressed GAP43 in the normal skin (Fig. 5A). However, after sciatic nerve section most spared SENFs at the medial site showed positive GAP43 staining, with a significantly increased proportion of GAP43 positive fibers at 3 (p b 0.01), 8 (p b 0.001), 16 (p b 0.001) and 29 dpi (p b 0.05) than in the contralateral intact paw (Fig. 5B), suggesting a sprouting phenotype. At 60 dpi the proportion of fibers expressing GAP43 was lower, but still higher than normal. In contrast, in the lateral pad of the paw GAP43 expression was not found distinctly until 29 dpi. After 29 dpi, when CGRP positive axons newly entered the skin, lateral skin samples showed significant colabeling (p b 0.001 at 29 dpi; p b 0.05 at 60 dpi). These observations evidence that the injured paw skin underwent a differential pattern of reinnervation by early sprouting of saphenous and late regeneration of sciatic nerve fibers. Discussion After sciatic nerve injuries, algesimetry tests are used to assess sensitivity to mechanical and thermal stimuli, but usually the stimuli are indiscriminately applied to the plantar foot skin, and the improvement of sensory function is often related to recovery or changes of the sciatic nerve innervation. The results of this study show that sensory assessment can be misleading in the evaluation of sciatic nerve regeneration if saphenous and sciatic territories are not distinguished in the injured paw. Different sensory stimuli have been applied in different ways and on different areas of the foot for assessment of nerve regeneration and functional reinnervation (Casals-Díaz et al., 2009; Nichols et al., 2005; Vogelaar et al., 2004). In the most common experimental model used, i.e. the sciatic nerve in rodents, the generally accepted rule was that the medial aspect of the paw should be avoided, as the saphenous nerve partially supplies this region (De Koning et al., 1986; Devor et al., 1979). However, in this way the information on the development of collateral hyperalgesia may be missed. In this study we selectively detected the changes in response to stimulation applied to specific sites in the paw, and assessed in parallel sciatic nerve regeneration and saphenous nerve sprouting effects on pain thresholds. The medial side of the paw remained sensitive due to sparing of saphenous fibers after sciatic transection, and showed marked hyperalgesia from a few days postlesion, consistent with adjacent neuropathic pain (Kingery and Vallin, 1989). In contrast, the lateral side was anesthetic until reinnervation by regenerating sciatic axons, which started by the fourth week, and then it became quickly hyperalgesic (Cobianchi et al., 2013). Interestingly, stimuli applied to the center of the paw showed mild recovery of responses from the first week and a slow progression of threshold decrease to below normal levels, which were attributable first to sprouting extension of the saphenous nerve and later to the arrival of regenerating sciatic axons. The development and the resolution of hyperalgesia are dependent on the injury model and the reinnervation of end-organs. Neurotmesis of the sciatic nerve with distal stump excision causes that a large area of the hindpaw remains unresponsive, but the saphenous mediated hyperalgesia is sustained over weeks/months (Attal et al., 1994; Casals-Díaz et al., 2009; Kingery et al., 1994). In contrast, after nerve

9

crush or section repaired lesions, regeneration of sciatic axons allows for regaining skin innervation and initiates the resolution of hyperalgesia (Casals-Díaz et al., 2009; Kingery et al., 1994). However, areas of hyperalgesia and hypoalgesia may coexist in the affected hindpaw, due to variable extension of saphenous sensory fibers sprouting and incomplete sciatic nerve regeneration. For this reason the assessment of somatosensory function mediated by the sciatic nerve after injury should be made in areas of the paw that are not affected by sprouting of the intact saphenous nerve, i.e. the lateral aspects of plantar skin and the lateral toes (Wood et al., 2011). Our results clearly demonstrate that after sciatic nerve injury a primary source of early hyperalgesia is due to the saphenous nerve. Sprouting of the saphenous nerve following sciatic denervation has been well documented, allowing functional extension of the saphenous territory for nociceptive and autonomic fibers (Devor et al., 1979; Kennedy et al., 1988; Kingery and Vallin, 1989; Markus et al., 1984). Saphenous nerve mediated mechanical hyperalgesia was long-lasting and remained after return of sciatic regenerating axons, whereas saphenous heat hyperalgesia gradually returned to normal values concomitantly to reappearance of responses at the sciatic territory from 4 weeks. Other studies found that adjacent hyperalgesia was more pronounced to mechanical than to heat stimulation (Kauppila and Xu, 1996; Mansikka and Pertovaara, 1997), and resolution or reduction of saphenous-mediated hyperalgesia coincident with sciatic nerve reinnervation (Kingery et al., 1994; Kinnman et al., 1992). Nonetheless, the hyperexcitability of the saphenous neurons could reflect chronic sensitization of its pathway, including spinal cord circuits (LaMotte et al., 1989; Markus et al., 1984; Sotgiu and Biella, 1995), contributing to maintain the hyperalgesia affecting more mechanical than thermal sensitive neurons. Collateral sprouting is a characteristic of normal undamaged nerves in response to elimination of other axons within the same target tissues. It is well known that skin denervation induces growth of collateral sprouts from neighbor uninjured axons (Diamond and Foerster, 1992; Diamond et al., 1987; Mendell et al., 1999), which is promoted by the secretion of neurotrophins by the denervated tissues (Diamond et al., 1992; Gloster and Diamond, 1992). Langerhans cell activation may also participate in intact nerve fiber sprouting or sensitization. For instance, Langerhans cells are in close apposition with epidermal fibers, and when activated they release pro-inflammatory mediators, such as nitric oxide and cytokines tumor necrosis factor and interleukin-1 (Andersson et al., 2004; Hsieh et al., 1996; Qureshi et al., 1996). This compensatory reinnervation allows for partial restoration of autonomic and nociceptive functions after nerve damage (Ahcan et al., 1998; Inbal et al., 1987). Small myelinated and unmyelinated fibers show stronger capacity to collaterally sprout and enlarge their territory of innervation than that of large myelinated fibers (Nixon et al., 1984; Doucette and Diamond, 1987; Jackson and Diamond, 1984; Kennedy et al., 1988; Gloster and Diamond, 1992), thus underlying increased nociceptive sensitivity that may be a protective mechanism for the previously denervated skin. Collateral reinnervation by saphenous nociceptive fibers rapidly extends to overlapping and contiguous areas previously innervated by the sciatic nerve (Devor et al., 1979; Kingery and Vallin, 1989; Kinnman et el., 1992). This explains the early recovery of sensitivity in the central areas of the paw, that disappeared after sectioning the saphenous nerve at 14 dpi. These observations point out that the site of sensory testing is crucial for assessing neuropathic pain and sensory recovery after nerve lesions because of the different pattern of reinnervation. We also found that saphenous peptidergic fibers showed early sprouts in the sub-epidermal plexus expressing GAP43 much earlier in time than returning sciatic regenerative axons, demonstrating that signs of neuropathic pain are associated to remodeling of sensory innervation in dermal layers of the skin. Sciatic nerve repair is a model of injury clinically relevant for the study of peripheral nerve regeneration (Rodríguez et al., 2004), which allows reinnervation differently from models of partial sciatic lesion

10

S. Cobianchi et al. / Experimental Neurology 255 (2014) 1–11

(Grelik et al., 2005; Ma and Bisby, 2000) in which sprouting and regenerative responses may be mixed. The reinnervation of the skin by axons regenerating after sciatic nerve section and suture repair started after 4 weeks, progressing from proximal to distal areas of the paw and gradually reaching sub-epidermal and epidermal layers up to the hyperreinnervation of the skin at 2 months. Sciatic reinnervation induced mechanical hyperalgesia in the lateral side of the paw, that reverted slowly over time (Lancelotta et al., 2003; Vogelaar et al., 2004). However, we found in parallel a higher proportion of CGRP fibers only in the saphenous territory, suggesting that skin nociceptive hyperreinnervation is not necessary to induce neuropathic phenotype after regeneration of peripheral nerves. This observation raises the possibility that different models of injury may induce a dishomogeneous distribution of receptors and sensory fiber types due to different contributions of collateral and regenerative reinnervation. Changes occurring in the primary afferent neurons alter the processing of nociceptive information following nerve injury. Peptidergic and non-peptidergic fibers play different roles in the development and maintenance of skin sensitization (Taylor et al., 2012). Previous works (Duraku et al., 2012, 2013; Grelik et al., 2005) showed a prominent role of peptidergic but not non-peptidergic fibers in reinnervating the skin after peripheral nerve injury. Most peptidergic fibers terminate in the more superficial layers of the epidermis (Zylka et al., 2005), and convey sensory information to specific laminae of the spinal cord dorsal horn (Molliver et al., 1995). We observed that epidermal and subepidermal layers have different patterns of reinnervation in relation with neuropathic abnormal responses. Epidermal reinnervation was not correlated with the significant decrease of withdrawal thresholds (Duraku et al., 2013). However, we found that peptidergic fibers early growing through the dermis anticipated epidermal reinnervation to trigger hyperalgesic responses. As previously hypothesized (Verdú and Navarro, 1997), the fact that pain responses started earlier than nerve fibers entered the epidermis may be explained by the activation of afferent fibers in the subepidermal plexus. To prove that active fiber growth is concomitant to hyperalgesic responses, we used GAP43 as a marker of axonal growth (Oestreicher et al., 1997). GAP43 is expressed in unmyelinated fibers of dorsal root ganglion following nerve injury (Coggeshall et al., 1991; Jaken et al., 2011; Woolf et al., 1988). We found that saphenous peptidergic fibers rapidly elongated through the subepidermal plexus after sciatic nerve injury. After one month, sciatic regenerating axons enter the skin and highly express GAP43 until 2 months, showing that collateral sprouting gradually extends in overlapping territory and is later substituted by growth of regenerative axons. In conclusion, peripheral nerve injuries induce intact nerve collateral sprouting, that is the cause of early recovery of sensory responses and neuropathic pain symptoms. The outcome of sensory testing depends on the site of stimulation, and landmarks of denervated and adjacent innervated skin should be used when performing algesimetry tests, together with histological analysis, for parallel investigation of the sprouting and regenerative capacities of sensory neurons. These processes are complementary in functional recovery, but contribute differently in the initiation and maintenance of neuropathic pain. Acknowledgments This research was supported by a grant FP7-278612 (BIOHYBRID) from the EU and funds of RETIC TERCEL from the Instituto de Salud Carlos III of Spain. The authors are grateful for the technical help of Mónica Espejo, Jessica Jaramillo and Marta Morell. References Ahcan, U., Arnĕz, Z.M., Bajrović, F., Janko, M., 1998. Contribution of collateral sprouting to the sensory and sudomotor recovery in the human palm after peripheral nerve injury. Br. J. Plast. Surg. 51, 436–443.

Andersson, By U., Tani, E., Andersson, U., Henter, J.I., 2004. Tumor necrosis factor, interleukin 11, and leukemia inhibitory factor produced by Langerhans cells in Langerhans cell histiocytosis. J. Pediatr. Hematol. Oncol. 26, 706–711. Attal, N., Filliatreau, G., Perrot, S., Jazat, F., Di Giamberardino, L., Guilbaud, G., 1994. Behavioural pain-related disorders and contribution of the saphenous nerve in crush and chronic constriction injury of the rat sciatic nerve. Pain 59, 301–312. Casals-Díaz, L., Vivó, M., Navarro, X., 2009. Nociceptive responses and spinal plastic changes of afferent C-fibers in three neuropathic pain models induced by sciatic nerve injury in the rat. Exp. Neurol. 217, 84–95. Cobianchi, S., Casals-Díaz, L., Jaramillo, J., Navarro, X., 2013. Differential effects of activity dependent treatments on axonal regeneration and neuropathic pain after peripheral nerve injury. Exp. Neurol. 240, 157–167. Coggeshall, R.E., Reynolds, M.L., Woolf, C.J., 1991. Distribution of the growth associated protein GAP-43 in the central processes of axotomized primary afferents in the adult rat spinal cord; presence of growth cone-like structures. Neurosci. Lett. 131, 37–41. De Koning, P., Brakkee, J.H., Gispen, W.H., 1986. Methods for producing a reproducible crush in the sciatic and tibial nerve of the rat and rapid and precise testing of return of sensory function. Beneficial effects of melanocortins. J. Neurol. Sci. 74, 237–246. Decosterd, I., Woolf, C.J., 2000. Spared nerve injury: an animal model of persistent peripheral neuropathic pain. Pain 87, 149–158. Devor, M., Govrin-Lippmann, R., 1979. Selective regeneration of sensory fibers following nerve crush injury. Exp. Neurol. 65, 243–254. Devor, M., Schonfeld, D., Seltzer, Z., Wall, P.D., 1979. Two modes of cutaneous reinnervation following peripheral nerve injury. J. Comp. Neurol. 185, 211–220. Diamond, J., Foerster, A., 1992. Recovery of sensory function in skin deprived of its innervation by lesion of the peripheral nerve. Exp. Neurol. 115, 100–103. Diamond, J., Coughlin, M., Macintyre, L., Holmes, M., Visheau, B., 1987. Evidence that endogenous beta nerve growth factor is responsible for the collateral sprouting, but not the regeneration, of nociceptive axons in adult rats. Proc. Natl. Acad. Sci. U. S. A. 84, 6596–6600. Diamond, J., Holmes, M., Coughlin, M., 1992. Endogenous NGF and nerve impulses regulate the collateral sprouting of sensory axons in the skin of the adult rat. J. Neurosci. 12, 1454–1466. Doucette, R., Diamond, J., 1987. Normal and precocious sprouting of heat nociceptors in the skin of adult rats. J. Comp. Neurol. 261, 592–603. Duraku, L.S., Hossaini, M., Hoendervangers, S., Falke, L.L., Kambiz, S., Mudera, V.C., Holstege, J.C., Walbeehm, E.T., Ruigrok, T.J., 2012. Spatiotemporal dynamics of reinnervation and hyperinnervation patterns by uninjured CGRP fibers in the rat foot sole epidermis after nerve injury. Mol. Pain 8, 61. Duraku, L.S., Hossaini, M., Schüttenhelm, B.N., Holstege, J.C., Baas, M., Ruigrok, T.J., Walbeehm, E.T., 2013. Re-innervation patterns by peptidergic Substance-P, nonpeptidergic P2X3, and myelinated NF-200 nerve fibers in epidermis and dermis of rats with neuropathic pain. Exp. Neurol. 241, 13–24. Gloster, A., Diamond, J., 1992. Sympathetic nerves in adult rats regenerate normally and restore pilomotor function during an anti-NGF treatment that prevents their collateral sprouting. J. Comp. Neurol. 326, 363–374. Grelik, C., Allard, S., Ribeiro-da-Silva, A., 2005. Changes in nociceptive sensory innervation in the epidermis of the rat lower lip skin in a model of neuropathic pain. Neurosci. Lett. 389, 140–145. Hsieh, S.T., Choi, S., Lin, W.M., Chang, Y.C., McArthur, J.C., Griffin, J.W., 1996. Epidermal denervation and its effects on keratinocytes and Langerhans cells. J. Neurocytol. 25, 513–524. Hsieh, S.T., Chiang, H.Y., Lin, W.M., 2000. Pathology of nerve terminal degeneration in the skin. J. Neuropathol. Exp. Neurol. 59, 297–307. Inbal, R., Rousso, M., Ashur, H., Wall, P.D., Devor, M., 1987. Collateral sprouting in skin and sensory recovery after nerve injury in man. Pain 28, 141–154. Jackson, P.C., Diamond, J., 1984. Temporal and spatial constraints on the collateral sprouting of low-threshold mechanosensory nerves in the skin of rats. J. Comp. Neurol. 226, 336–345. Jaken, R.J., van Gorp, S., Joosten, E.A., Losen, M., Martínez-Martínez, P., De Baets, M., Marcus, M.A., Deumens, R., 2011. Neuropathy-induced spinal GAP-43 expression is not a main player in the onset of mechanical pain hypersensitivity. J. Neurotrauma 28, 2463–2473. Kauppila, T., Xu, X.J., 1996. Sciatic nerve section induces mechanical hyperalgesia in skin adjacent to the deafferented region in rats: lack of correlation with autotomy behavior. Neurosci. Lett. 211, 65–67. Kennedy, W.R., Navarro, X., Kamei, H., 1988. Reinnervation of sweat glands in the mouse: axonal regeneration versus collateral sprouting. Muscle Nerve 11, 603–609. Kingery, W.S., Vallin, J.A., 1989. The development of chronic mechanical hyperalgesia, autotomy and collateral sprouting following sciatic nerve section in rat. Pain 38, 321–332. Kingery, W.S., Lu, J.D., Roffers, J.A., Kell, D.R., 1994. The resolution of neuropathic hyperalgesia following motor and sensory functional recovery in sciatic axonotmetic mononeuropathies. Pain 58, 157–168. Kinnman, E., Aldskogius, H., Johansson, O., Wiesenfeld-Hallin, Z., 1992. Collateral reinnervation and expansive regenerative reinnervation by sensory axons into "foreign" denervated skin: an immunohistochemical study in the rat. Exp. Brain Res. 91, 61–72. LaCroix-Fralish, M.L., Ledoux, J.B., Mogil, J.S., 2007. The Pain Genes Database: an interactive web browser of pain-related transgenic knockout studies. Pain 131 (3), e1–e3 (e4). LaMotte, C.C., Kapadia, S.E., Kocol, C.M., 1989. Deafferentation-induced expansion of saphenous terminal field labelling in the adult rat dorsal horn following pronase injection of the sciatic nerve. J. Comp. Neurol. 288, 311–325.

S. Cobianchi et al. / Experimental Neurology 255 (2014) 1–11 Lancelotta, M.P., Sheth, R.N., Meyer, R.A., Belzberg, A.J., Griffin, J.W., Campbell, J.N., 2003. Severity and duration of hyperalgesia in rat varies with type of nerve lesion. Neurosurgery 53, 1200–1209. Le Bars, D., Gozariu, M., Cadden, S.W., 2001. Animal models of nociception. Pharmacol. Rev. 53, 597–652. Ma, W., Bisby, M.A., 2000. Calcitonin gene-related peptide, substance P and protein gene product 9.5 immunoreactive axonal fibers in the rat footpad skin following partial sciatic nerve injuries. J. Neurocytol. 29, 249–262. Mansikka, H., Pertovaara, A., 1997. Submodality-selective hyperalgesia adjacent to partially injured sciatic nerve in the rat is dependent on capsaicin-sensitive afferent fibers and independent of collateral sprouting or a dorsal root reflex. Brain Res. Bull. 44, 237–245. Markus, H., Pomeranz, B., Krushelnycky, D., 1984. Spread of saphenous somatotopic projection map in spinal cord and hypersensitivity of the foot after chronic sciatic denervation in adult rat. Brain Res. 296, 27–39. Mendell, L.M., Albers, K.M., Davis, B.M., 1999. Neurotrophins, nociceptors, and pain. Microsc. Res. Tech. 45, 252–261. Mogil, J.S., Davis, K.D., Derbyshire, S.W., 2010. The necessity of animal models in pain research. Pain 151, 12–17. Molliver, D.C., Radeke, M.J., Feinstein, S.C., Snider, W.D., 1995. Presence or absence of TrkA protein distinguishes subsets of small sensory neurons with unique cytochemical characteristics and dorsal horn projections. J. Comp. Neurol. 361, 404–416. Navarro, X., Verdú, E., Butí, M., 1994. Comparison of regenerative and reinnervating capabilities of different functional types of nerve fibers. Exp. Neurol. 129, 217–224. Navarro, X., Verdú, E., Wendelscafer-Crabb, G., Kennedy, W.R., 1995. Innervation of cutaneous structures in the mouse hind paw: a confocal microscopy immunohistochemical study. J. Neurosci. Res. 41, 111–120. Navarro, X., Verdú, E., Wendelschafer-Crabb, G., Kennedy, W.R., 1997. Immunohistochemical study of skin reinnervation by regenerative axons. J. Comp. Neurol. 380, 164–174. Navarro, X., Vivó, M., Valero-Cabré, A., 2007. Neural plasticity after peripheral nerve injury and regeneration. Prog. Neurobiol. 82, 163–201. Nichols, C.M., Myckatyn, T.M., Rickman, S.R., Fox, I.K., Hadlock, T., Mackinnon, S.E., 2005. Choosing the correct functional assay: a comprehensive assessment of functional tests in the rat. Behav. Brain Res. 163, 143–158.

11

Nixon, B.J., Doucette, R., Jackson, P.C., Diamond, J., 1984. Impulse activity evokes precocious sprouting of nociceptive nerves into denervated skin. Somatosens. Res. 2, 97–126. Oestreicher, A.B., De Graan, P.N., Gispen, W.H., Verhaagen, J., Schrama, L.H., 1997. B-50, the growth associated protein-43: modulation of cell morphology and communication in the nervous system. Prog. Neurobiol. 53, 627–686. Qureshi, A.A., Hosoi, J., Xu, S., Takashima, A., Granstein, R.D., Lerner, E.A., 1996. Langerhans cells express inducible nitric oxide synthase and produce nitric oxide. J. Invest. Dermatol. 107 (6), 815–821. Rodríguez, F.J., Valero-Cabré, A., Navarro, X., 2004. Regeneration and functional recovery following peripheral nerve injuries. Drug Discov. Today Dis. Models 1, 177–185. Sotgiu, M.L., Biella, G., 1995. Spinal expansion of saphenous afferents after sciatic nerve constriction in rats. Neuroreport 6, 2305–2308. Stankovic, N., Johansson, O., Hildebrand, C., 1999. Increased occurrence of PGP 9.5immunoreactive epidermal Langerhans cells in rat plantar skin after sciatic nerve injury. Cell Tissue Res. 298, 255–260. Taylor, A.M., Osikowicz, M., Ribeiro-da-Silva, A., 2012. Consequences of the ablation of nonpeptidergic afferents in an animal model of trigeminal neuropathic pain. Pain 153, 1311–1319. Vallin, J.A., Kingery, W.S., 1991. Adjacent neuropathic hyperalgesia in rats: a model for sympathetic independent pain. Neurosci. Lett. 133, 241–244. Verdú, E., Navarro, X., 1997. Comparison of immunohistochemical and functional reinnervation of skin and muscle after peripheral nerve injury. Exp. Neurol. 146, 187–198. Vogelaar, C.F., Vrinten, D.H., Hoekman, M.F., Brakkee, J.H., Burbach, J.P., Hamers, F.P., 2004. Sciatic nerve regeneration in mice and rats: recovery of sensory innervation is followed by a slowly retreating neuropathic pain-like syndrome. Brain Res. 1027, 67–72. Wood, M.D., Kemp, S.W., Weber, C., Borschel, G.H., Gordon, T., 2011. Outcome measures of peripheral nerve regeneration. Ann. Anat. 193, 321–333. Woolf, C.J., Ma, Q., 2007. Nociceptors — noxious stimulus detectors. Neuron 55, 353–364. Woolf, C.J., Thompson, S.W., King, A.E., 1988. Prolonged primary afferent induced alterations in dorsal horn neurones, an intracellular analysis in vivo and in vitro. J. Physiol. Paris 83, 255–266. Zimmermann, M., 1983. Ethical guidelines for investigations of experimental pain in conscious animals. Pain 16, 109–110. Zylka, M.J., Rice, F.L., Anderson, D.J., 2005. Topographically distinct epidermal nociceptive circuits revealed by axonal tracers targeted to Mrgprd. Neuron 45, 17–25.