Secretagogin is expressed in sensory CGRP neurons and in spinal cord of mouse and complements other calcium-binding proteins, with a note on rat and human

Background Secretagogin (Scgn), a member of the EF-hand calcium-binding protein (CaBP) superfamily, has recently been found in subsets of developing and adult neurons. Here, we have analyzed the expression of Scgn in dorsal root ganglia (DRGs) and trigeminal ganglia (TGs), and in spinal cord of mouse at the mRNA and protein levels, and in comparison to the well-known CaBPs, calbindin D-28k, parvalbumin and calretinin. Rat DRGs, TGs and spinal cord, as well as human DRGs and spinal cord were used to reveal phylogenetic variations. Results We found Scgn mRNA expressed in mouse and human DRGs and in mouse ventral spinal cord. Our immunohistochemical data showed a complementary distribution of Scgn and the three CaBPs in mouse DRG neurons and spinal cord. Scgn was expressed in ~7% of all mouse DRG neuron profiles, mainly small ones and almost exclusively co-localized with calcitonin gene-related peptide (CGRP). This co-localization was also seen in human, but not in rat DRGs. Scgn could be detected in the mouse sciatic nerve and accumulated proximal to its constriction. In mouse spinal cord, Scgn-positive neuronal cell bodies and fibers were found in gray matter, especially in the dorsal horn, with particularly high concentrations of fibers in the superficial laminae, as well as in cell bodies in inner lamina II and in some other laminae. A dense Scgn-positive fiber network and some small cell bodies were also found in the superficial dorsal horn of humans. In the ventral horn, a small number of neurons were Scgn-positive in mouse but not rat, confirming mRNA distribution. Both in mouse and rat, a subset of TG neurons contained Scgn. Dorsal rhizotomy strongly reduced Scgn fiber staining in the dorsal horn. Peripheral axotomy did not clearly affect Scgn expression in DRGs, dorsal horn or ventral horn neurons in mouse. Conclusions Scgn is a CaBP expressed in a subpopulation of nociceptive DRG neurons and their processes in the dorsal horn of mouse, human and rat, the former two co-expressing CGRP, as well as in dorsal horn neurons in all three species. Functional implications of these findings include the cellular refinement of sensory information, in particular during the processing of pain.


Background
Calcium-binding proteins (CaBPs) play a major role in neuronal functions, and their cellular distribution in the nervous system has in many cases been thoroughly mapped by immunohistochemistry [1,2]. In particular, parvalbumin (PV), calretinin (CR) and calbindin D-28k (CB) have received much attention due to their robust, developmentally regulated and cell type-specific expression in the nervous system, and have emerged as effective markers to identify subpopulations of neurons [2][3][4]. In general terms, these proteins act either as Ca 2+ sensors or buffers of Ca 2+ transients in neurons, defined by their molecular properties and the signaling context they participate in [5]. Chemically, CaBPs are characterized by a tandem repeat of the Ca 2+ -binding loop surrounded by two helices, the EF-hand binding site [1,6,7]. Secretagogin (Scgn) is a recently cloned member of the EF-hand CaBP superfamily, first identified from a human pancreatic cDNA library by immunoscreening with the murine monoclonal antibody D24 generated using human insulinoma as immunogen [8,9]. Structurally, Scgn's deduced amino acid sequence specifies a protein of 276 amino acids with a calculated molecular mass of 32 kDa that can bind up to four Ca 2+ ions simultaneously [8]. Using immunohistochemical methodology, Scgn has been detected in several tissues, such as the brain of various mammalian species including humans [10][11][12][13][14][15][16], where it may associate with SNAP-25 [17], a protein(s) participating in the vesicular exocytosis of neurotransmitters [18], possibly neurodegeneration [10][11][12], as well as development, including embryonic expression in dorsal root ganglia (DRGs) and trigeminal ganglia (TGs) [15].
In the present study we have therefore analyzed, with quantitative (real-time) PCR (qPCR), in situ hybridization and high-resolution immunohistochemistry, the localization of Scgn in mouse DRGs (mDRGs) and spinal cord, with emphasis on its possible co-localization with PV, CR, and CB, as well as with calcitonin gene-related peptide (CGRP) or isolectin B4 (IB4), classic markers of nociceptive neurons [51,52]. In addition, the presence of three further molecules known to be expressed in DRGs/dorsal horn was studied: transient receptor potential vanilloid subtype 1 (TRPV1) [53], gastrin releasing peptide (GRP) [54][55][56], and protein kinase C gamma (PKCgamma) [57]. Dorsal root transection and unilateral peripheral sciatic nerve injury were performed in mice. Finally, we have, in a preliminary way, studied the extent of phylogenetic conservation in Scgn's distribution by comparing mouse DRGs, TGs and spinal cord with rat DRGs (rDRGs), rat TGs (rTGs) and rat spinal cord, as well as with human DRGs (hDRGs) and spinal cord. Some of these results were presented in a preliminary form at the 13 th World Congress on Pain [58].

Scgn mRNA detection: methodological considerations and tissue distribution pattern
Recently, Scgn has been localized in the brain with immunohistochemistry using affinity-purified antibodies raised against distinct peptide domains ("epitopes") of this protein [15], producing results that correspond well with publicly-available mRNA distribution maps [59]. Nevertheless, correlative analysis of Scgn mRNA and protein has not been performed. Therefore, we first probed Scgn mRNA distribution in the olfactory bulb, containing highest Scgn protein and mRNA levels in the nervous system [15,16]. We visualized, using riboprobes, Scgn mRNA as a "band" in deep neurons populating the granular layer (Additional file 1: Figure S1A), as well as the inner sublayer of the external plexiform layer (Additional file 1: Figure S1A 1 ). In addition, Scgn mRNA, though at relatively low levels, was found in cells scattered around olfactory glomeruli Additional file 1: Figure S1A 2 ,B), likely periglomerular interneurons. Thus, corresponding Scgn mRNA (Additional file 1: Figure  S1A-A 2 ) and protein distribution patterns (Additional file 1: Figure S1C,C 1 ) support the specificity of the antibodies used in the present and previous [15,16,60] studies.
Next, we profiled relative Scgn mRNA levels between amongst mouse olfactory bulb, dorsal and ventral spinal cord, and DRGs lumbar 4 and 5 (L4-5) by means of qPCR. While Scgn mRNA levels in the olfactory bulb were exceptionally high (Figure 1a), they were under detection threshold and at very low levels in the dorsal and ventral spinal horns, respectively (Figure 1a). Moderate Scgn mRNA levels were detected in DRGs (Figure 1a, n = 3/region) suggesting that only a restricted population of DRG neurons might be Scgn immunoreactive (IR).
We also addressed Scgn expression in DRGs using in situ hybridization. We demonstrate, by using oligoprobes, that a small population of cells in the hDRG expresses notable levels of Scgn mRNA (Figure 1b,c). However, we could not reliably detect in situ hybridization signal in mDRGs, probably reflecting the sparse mRNA levels found in our qPCR analysis.
In the hDRGs, 13.3±0.4% of all NPs were Scgn-IR (Figure 3a), and there were also distinct processes (Figure 3e

Localization of Scgn protein in spinal cord and sciatic nerve
Scgn-LI was found both in the neuronal cell bodies and fibers/processes in the mouse spinal cord: Scgn-IR fibers formed a dense plexus in the superficial dorsal horn (Lissauer's tract) (Figures 5a,b; 6a). Scgn-IR neurons were mainly found in inner lamina II ( Figures. 5c; 6a), but some cells were also seen in layers II-V, including both small neurons ( Figure 5f) and large, multipolar neurons (Figure 5d,e). In the ventral horn, Scgn-positive neurons were sporadically seen (Figure 5g), and some of them (3.0%) co-expressed CGRP (Figure 5g-i), in this region a marker for motoneurons [61].
In mouse dorsal horn, using double staining, we detected most superficial Scgn-IR fibers were CGRP-IR ( Figure 5r and r´expressing Scgn and CB appeared PV-negative. There were also neurons co-expressing Scgn and CB but apparently not CR (Figure 5s,t).
Unilateral dorsal rhizotomy strongly reduced both Scgnand CGRP-LIs in the superficial region of the ipsilateral dorsal horn as compared to the contralateral side (cf. Figure 6a-c with d-f). However, there were still cell bodies and processes in inner lamina II (Figure 6d,f). In the sciatic nerve, a moderate number of Scgn-IR axons could be seen (Figure 6g), however fewer than the CGRP-positive ones (Figure 6h), partially overlapping ( Figure 6i). Ten hours after ligation of the sciatic nerve there was distinct accumulation of Scgn-LI ( Figure 6j) and CGRP-LI ( Figure 6k) proximal to the site of the injury. In contrast to the normal nerve, there seemed to be a more equal number of fibers immunoreactive for Scgn and CGRP with prominent overlap, further supporting their coexistence ( Figure 6l).
In the rat dorsal horn, Scgn-LI was less prominent and mainly observed in medial, inner lamina II (Figure 7a In the human spinal cord, a dense network of Scgn-IR fibers was observed in the superficial, especially lateral, dorsal horn, most of which were CGRP-IR (Figure 3h-j). As in mouse, the CGRP fibers extended ventrally beyond the Scgn zone. A few Scgn-positive cell bodies, surrounded by Scgn nerve endings were seen in inner lamina II, but they did not, as in mouse, from a distinct band (

Scgn protein expression after peripheral nerve injury
Transection of the sciatic nerve did not significantly affect the percentage of Scgn-IR NPs in ipsilateral mouse DRGs as compared to contralateral ones (6.6±1.0% vs. 6.3±0.9; P>0.05). In agreement, Western blotting showed no change of Scgn protein levels in ipsilateral vs. contralateral mDRGs ( Figure 8a). Finally, in mouse spinal cord, the total protein levels of Scgn did not change after peripheral nerve injury (Figure 8b).

Discussion
The present study shows that Scgn, a recently identified member of the CaBP superfamily [8], is expressed in distinct neuronal populations at the spinal level of several species. In mDRGs, subpopulations of these nociceptive neurons express CGRP (98%)-and TRPV1-(~20%) but are IB4-negative [51,52]. Thus, the TRPV1-IR population of these neurons may be sensitive to noxious heat [53].
The apparent lack of GRP in Scgn-IR DRG neurons indicates that they are not involved in itching [56], and Scgn-positive, PKCgamma-negative dorsal horn interneurons may not be excitatory [57]. We have, however, not been able to assign these neurons to any of the categories identified in extensive developmental studies [19,20]. They are distinctly different from those harboring the three most-studied CaBPs (PV, CB and CR). Scgn mRNA was detected in a subpopulation of hDRG neurons, and Scgn mRNA transcripts were found in mDRGs by means of qPCR. The specificity of the antiserum has further been supported by adsorption experiments and Western blot analysis, as also shown in previous studies on brain [15,16], where results were compared with those in the Allen brain atlas [59]. Finally, double-staining experiments with two different antisera raised against different epitopes and in two animal species stained the same cell bodies in mDRGs.

Scgn is present in all major compartments of the mDRG neuron
An interesting question is to what extent Scgn produced in the mDRGs is transported to the dorsal horn. Our findings with dorsal rhizotomy suggest that the staining in the superficial layers, but not in inner lamina II, originates in the DRGs. Moreover, Scgn is detected in the sciatic nerve and is transported peripherally from the cell body, as shown by the accumulation of Scgn-LI proximal to a compression of the sciatic nerve. Therefore, Scgn may have a function(s) not only in cell bodies but possibly also in central and peripheral nerve terminals.
A similar situation may exist for the other three CaBPs discussed here, since there is a loss of CaBPs in the ipsilateral dorsal column/column nucleus after unilateral, multiple dorsal root ganglionectomy [40] and dorsal rhizotomy [34]. The latter is the projection area of large DRG neurons [62], and this is the category of DRG neurons that to large extent harbor CR and PV.

Other CaBPs in DRGs
CaBPs have in mouse mainly been studied as markers for the diverse neuron populations, especially during development and in cultures (e.g. [19,20,28,[30][31][32]), but detailed in vivo quantitative and colocalization analyses of adult mDRGs are less common. Nevertheless, Ichikawa et al. [50] reported the presence of PV-LI in~5% of adult mDRG NPs.
In contrast, a large number of studies have dealt with this issue in rat (for refs. see Introduction). In the most recent study on rDRGs by Ichikawa and colleagues [44] and using triple-label immunostainings, CR and PV are both present in mostly large-sized NPs, and~10% of the NPs contain both CaBPs (most CR neurons contain PV and~40% of PV neurons contain CR). A bimodal size curve has been earlier described for CR by the same a b c d group [39]. Neither PV nor CR coexists with CB, the latter being expressed in neurons of various sizes. With regard to nociceptors, known to be small-sized DRG neurons [51,52], none of these three CaBP populations, mostly encompassing large neurons, seem to be extensively involved: 1% of the PV-positive NPs express CGRP [42], colocalization of CGRP with either PV or CB is "rare" [37,38]. Nevertheless, Honda [49] reported that 9% of CB neurons are CGRP-IR, and 7% of CR neurons are substance P-positive [39]. The present results suggest that Scgn is the major CaBP in a population of peptidergic nociceptors in mDRGs [51,52]. Similar to mDRGs, Scgn is also expressed in mTGs, and majority of them are CGRP-positive. and CB (r'; read), but not PV (r'; dark blue). s and t show neurons (filled arrowhead; light blue) containing Scgn (green) and CB (red), but not CR (dark blue). Empty arrowheads indicate a dorsal horn neuron only express Scgn-LI (j'',k',l',o,q,t) or a Scgn-IR fiber (j'; green) does not overlap with TRPV1-IR fiber (j'; red).. Scale bars indicate 200μm (a), 50 μm (b=m=n=p; c=d=e=f; g=h=i; j=k=l; j'=j"; k'=l'; m'=o=q; r=s; r'=t).

Phylogenetic differences
We detected both similarities and differences in expression of Scgn-LI when comparing DRGs, TGs and spinal cord of mouse with rat and human tissues, suggesting partially conserved protein expression. In rDRGs, even fewer NPs were Scgn-IR as compared to mouse and, surprisingly, none of them was CGRP-IR. In contrast, Scgn expression was quantitatively similar in mouse and human DRGs, most of them expressing CGRP and none seemed IB4-positive. In rTGs many Scgn-IR neurons coexpressed CGRP and many, unexpectedly, stained for IB4, whereas no Scgn-IB4 containing neurons were detected in mouse TGs.
With regard to spinal cord, Scgn in mouse is present mainly in cell bodies in inner lamina II and, albeit in low numbers, in several other layers (I, II-IV) of the dorsal horn, and in ventral horn neurons. This pattern was similar in human spinal cord. However, cell bodies were only detected in inner lamina II, and they did not form a distinct band as in mouse. In rat, Scgn staining was less pronounced with little fiber staining in the very superficial region, possibly reflecting the low numbers of Scgn-IR cell bodies in DRGs and with the staining in inner lamina II mainly located medially. In spinal ventral horn, no Scgn-IR neurons could be detected in rat. Scgn-LI in mouse was found together with CB-, PV-or CR-LI, albeit at very different proportions. In the rat spinal cord some Scgn-IR interneurons also expressed CB or PV, but not CR-LI.
Taken together, mouse is similar to human with respect to Scgn expression in DRGs, and to large extent in the spinal cord, while rats exhibit substantial differences. With regard to CaBPs, in rDRGs more than 40% of the Scgn-IR NPs co-expressed CB, but hardly any CR or PV, a species difference here is being the low CB-Scgn coexistence in mouse.

Scgn and other CaBPs in spinal cord
Several studies on CaBPs in the rat spinal cord have been published [21,[34][35][36]40,48,63] but only few on mouse [29], the latter focusing on a select neuron population, so called V1 neurons in the deeper layers. The  studies on rat have shown that in lamina I many neurons are CB-IR, fewer CR-IR and none PV-IR. In lamina II CB-and CR-IR neurons are densely packed, and PV is confined to a distinct band in inner lamina II. There is only limited coexistence of the three CaBPs in the superficial laminae, although examples of cell positive for both CR and CB have been observed [40]. Laminae III and IV have in general cell bodies expressing CaBPs [35,48]. In the present study Scgn was, in contrast to CB, CR and PV, also expressed in mouse ventral horn neurons.

Functional aspects on Scgn
In view of their function as 'gate keepers' of Ca 2+ homeostasis, CaBPs have been hypothesized a protective role for preventing abnormal, cytotoxic Ca 2+ levels, thus likely participating in neurodegenerative processes and disease [64][65][66][67][68][69][70][71][72]. Interestingly, involvement of Scgn in neuronal survival in Alzheimer's disease has also been reported [10][11][12]. There is functional evidence that Ca 2+ buffering is important also in sensory neurons [73] and its dysfunction facilitates sensory neuron degeneration [74]. An alternative function as Ca 2+ sensor may be considered for Scgn in DRG neurons, since an interaction between Scgn and SNAP-25 proteins has been reported by Rogstam et al. [17]. This finding suggests a role in the control of neurotransmitter release since N-ethylmaleimide-sensitive factor-attachment protein receptors (SNAREs)-associated proteins are part of the exocytotic machinery [18]. Interestingly, inhibition of exocytosis causes long-lasting attenuation of pain [75].

Conclusions
Scgn represents a novel CaBP, which here and earlier has been found expressed in subpopulations of neurons in the rodent and primate nervous system, complementary to several other well-known members of this proteins superfamily. However, Scgn-positive DRG neurons, presumably a subtype of nociceptors both in mouse and humans, do not seem to perfectly match any of the DRG neuron populations identified during development in mice [19,20]. Scgn is also present in cell bodies in various layers of the dorsal horn. Analysis of corresponding tissues in rat suggests species differences in Scgn expression. The similarity between mouse and human DRGs suggest that results from future experiments on Scgn, e. g. using gentically modified mice, may be relevant to decipher molecular pathomechanisms in humans.

Tissues and animal models
Experiments were performed on male C57BL/6J Bommince mice (A/S Bomholtgaard, Ry, Denmark) weighing 25-28 g, and on adult male Sprague-Dawley rats (200-250 g; B and K Universal, Stockholm, Sweden). All animals were kept under standard conditions on a 12-hour daylight cycle with free access to food and water. Unilateral sciatic nerve transection (axotomy) was made on n=10 mice as described earlier [76]. Surgical procedures were performed under anesthesia with isoflurane. Operated animals were allowed to survive for 2 weeks after surgery. Dorsal root rhizotomy was done on n=5 mice, and animals were allowed to survive for 10 days. Sciatic nerve ligation was done in n=5 mice, which were sacrificed after 10 hours. Human ganglia were harvested from children with obstetric brachial plexus lesions, and undergoing reconstructive nerve surgery. Human spinal cord was harvested from a 48-year-old women died from stroke. The studies have been approved by the local Ethical Committee for animal experiments (Norra Stockholms djursförsöksetiska nämnd), and experiments on hDRGs were approved by a local Ethical Committee with written consent from the next of kin.

mRNA detection in tissues
In situ hybridization analysis of Scgn mRNA in mouse olfactory bulbs using riboprobes was performed as previously described [77]. Briefly, adult brains were perfusion fixed followed by post-fixation in the same fixative overnight (4% paraformaldehyde, in 0.1M PB), cryoprotected (30% sucrose in 0.1M PB), embedded in Tissue-tek OCT compound (Miles Laboratories, Elkhart, IN) and sectioned at a thickness of 20 μm. The fragment of scgn cDNA used for riboprobe synthesis was amplified from an adult mouse olfactory bulb cDNA preparation by PCR using Pfu DNA polymerase (Promega). The primers used for scgn cDNA amplification were flanked at their 5' ends with T7 and SP6 polymerase acceptors PCR products onto 1% agarose gels. Digoxigenin-labeled antisense and sense riboprobes for mouse scgn were synthesized by in vitro transcription using SP6 and T7 RNA polymerases (Roche). After synthesis, probes were cleaned by using the RNeasy kit (Qiagen) and DNA digested with RNase-free DNase I (Qiagen). Alkaline phosphatase-conjugated anti-digoxigenin antibodies (Roche, 1:2,000) were used with their signal developed by 5-bromo-4-chloro-3-indolyl-phosphate/nitro blue tetrazolium as substrate (BCIP/NBT; Roche).

mRNA detection by quantitative real-time PCR
Quantitative PCR (qPCR) reactions were performed with custom designed primers according to published protocols [16]. RNA isolated from tissues microdissected from adult C57Bl/6N mice (n = 3/DRG/spinal cord/olfactory bulb) were subjected to Scgn expression analysis after validating RNA integrity (data not shown). Gapdh was used to normalize Scgn mRNA expression levels. Results from qPCR experiments were subsequently compared to Scgn mRNA distribution as determined by in situ hybridization (primers: 5'CCCAGAAGTGGATGGATTTG 3'; reverse: 5'GTTGGGGATCAGGGGTTTAT 3'.

Immunohistochemistry
All operated (n=20) and control mice (n=10), as well as rats (n=10) were deeply anesthetized with sodium pentobarbital (10 mg/kg for mouse and 50 mg/kg for rat, both i.p.) and transcardially perfused with 20 ml (50 ml) of warm saline (0.9%; 37°C), followed by 20 ml (50 ml) of a warm mixture of 4% paraformaldehyde (37°C) and 0.4% picric acid in 0.16 M phosphate buffer (pH 7.2), and then by 50 ml (250 ml) of the same, but ice-cold fixative [79,80]. The L 5 mDRGs, mTGs, rDRGs and rTGs, as well as the L4 and L5 segments of both mouse and rat spinal cord were dissected out and postfixed in the same fixative for 90 min at 4°C. Specimens were subsequently stored in 10% sucrose in phosphate buffered saline (PBS, 0.1 M, pH 7.4) containing 0.01% sodium azide (Sigma, St. Louis, MO) and 0.02% bacitracin (Sigma) as preservatives at 4°C for 2 days. The hDRGs and spinal cord were immersion-fixed for four hours in ice-cold fixative and rinsed as mentioned above. Tissues were embedded with OCT compound (Tissue Tek), frozen and cut in a cryostat (Microm, Heidelberg, Germany) at 10 μm (mDRGs), 14 μm (mTGs, rDRGs and hDRGs) or 20 μm (mouse, rat and human spinal cords) thickness and mounted onto chrome-alum-gelatin-coated slides. Thaw-mounted sections were dried at room temperature (RT) for 30 min and rinsed with PBS for 15 min. Sections were incubated for 18 hours at 4°C in a humid chamber with rabbit anti-Scgn antiserum (1:1,000); [15,16,60] diluted in PBS containing 0.2% (w/v) bovine serum albumin and 0.03% Triton X-100 (Sigma). In addition a monospecific polyclonal rabbit anti-human Scgn antibody was used [8]. Briefly, purified recombinant Scgn (540 μg) was emulsified in complete Freund's adjuvant and injected subcutaneously into a rabbit followed by two more injections of incomplete Freund's adjuvant. Serum collected one month after the third immunization contained high titer antibody activity against the recombinant protein when tested by ELISA [8].
The specificity of Scgn antiserum was tested by preabsorption of the antiserum with homologous antigen at a concentration of 1 and/or 10 μM for 24 hours at 4°C. After incubation with control serum, i.e. Scgn antiserum pre-absorbed with the excess of Scgn, no fluorescent neuronal cells could be observed (data not shown).

Image analysis and quantification
Specimens were analyzed on a Bio-Rad Radiance Plus confocal scanning microscope (Bio-Rad, Hemel, Hempstead, UK) installed on a Nikon Eclipse E 600 fluorescence microscope (Nikon, Tokyo, Japan) equipped with x10 (0.5 numerical aperture, NA), x20 (0.75 NA) and x60 oil (1.40 NA) objectives. Fluorescein labeling was excited using the 488-nm line of the argon ion laser and detected after passing a HQ 530/60 (Bio-Rad) emission filter. For the detection of lissamine rhodamine sulfonyl chloride and rhodamine, the 543-nm HeNe laser was used in combination with the HQ 570 (Bio-Rad) emission filter. For the detection of DAPI a 405-nm laser was used. All the slides were scanned in a series of 1μm-thick optical sections. Consequently, images were analyzed separately and merged to evaluate possible colocalization. In some cases a Zeiss laser scanning microscope 780 system with a planapochromat x 20 (0.8NA) M27 objective was used.
To determine the percentage of IR NPs, the counting was performed on 10 or 14 μm thick sections, and every 4th or 6th section was selected (Nike Microphot-FX microscope, 20x objectives). The total number of immuno-positive NPs was divided by the total number of propidium iodide-stained [85] NPs in the DRG sections, and the percentage of positive NPs was calculated. Five to ten sections of each DRG from five animals in each group were included in the analysis, and 1,200-3,000 NPs were counted in each ganglion. The size distribution of NPs with a visible nucleus was measured using the Nikon Eclipse E 600 fluorescence microscope with Wasabi Image Software. We divided the NPs into small (a somal area of 100-600 μm 2 ); medium-sized (600-1400 μm 2 ) and large (>1400 μm 2 ) according to earlier studies [76,86]. The percentages of DRG NPs in each of these categories were calculated.

Statistical analyses
Differences between the percentage of Scgn-IR NPs as well as the gray levels of Scgn in mDRG neurons in ipsilateral and contralateral samples were evaluated by Student's t test.