Cocaine- and amphetamine-regulated transcript peptide (CART) is present in peptidergic C primary afferents and axons of excitatory interneurons with a possible role in nociception in the superficial laminae of the rat spinal cord

CART antisera

Monoclonal CART antibody was raised in mouse against purified recombinant CART(54–102) peptide (Thim et al., 1998), as described by Vrang et al. (1999). The antibody reacted equally well with CART(54–102), CART(61–102) and CART(62–102) peptides (Singru et al., 2007). Polyclonal CART antibody was raised in rabbit against the CART(79–102) fragment, as reported by Koylu et al. (1997). This antibody reacted with the immunizing peptide but not by other CART peptide fragments, such as the CART(2–27), CART (31–52) or CART(55–76). No immunostaining was detected when primary antibodies were preabsorbed with their immunizing peptides (Koylu et al., 1997; Vrang et al., 1999).

In our study, further controls were carried out by omitting either the primary or secondary antibodies: the immunostaining did not differ from the background. Immunoperoxidase staining was also abolished by preabsorption of the monoclonal CART antibody with the naturally occurring, biologically active CART(55–102) fragment (American Peptide Company; 0.01 mm). Perfect matching of the staining patterns of the mouse and rabbit CART antibodies was also demonstrated by dual immunofluorescent labelling (see Supplementary material, Fig. S1).

Distribution of CART peptide in the rat spinal cord

The distribution of CART-immunoreactivity was examined along the length of the rat spinal cord. As reported earlier (Koylu et al., 1998; Dun et al., 2000; Ohsawa et al., 2000), CART staining was present from cervical to sacral segments, and CART-stained fibres and terminals were found in each lamina and in the dorsolateral fasciculus including the lateral spinal nucleus (Fig. 1). The densest network of CART-positive fibres was observed in the superficial laminae, especially in lamina I, and the CART-positive axon density dramatically decreased away from the superficial laminae. CART-immunoreactive cell bodies were found around the central canal, and a few pale neurons appeared in the dorsal horn. Strongly stained neurons and fibres were also located in the intermediolateral cell column of the thoracic cord, as reported previously (Koylu et al., 1998; Ohsawa et al., 2000).

Colocalization of CART with various immunocytochemical markers in the superficial laminae

The majority (72.6%) of CART-positive axons in lamina I and a smaller population (34.8%) of those in lamina II were found to be immunoreactive for CGRP (Fig. 2A–P). The proportion of CART terminals in laminae I and II that contained both CGRP and SP was 62.70 ± 0.90% and 33.90 ± 1.30%, respectively (Fig. 2E–H). A few CART terminals were immunoreactive for both CGRP and SOM (9.89 ± 1.17% in lamina I and 0.94 ± 0.34% in lamina II; Fig. 2I–L). CART immunoreactivity was not found in non-peptidergic primary afferent terminals, which were detected by IB4 and occupied mainly the inner part of lamina II (Fig. 2A–D). CTb injected into the sciatic nerve labelled the myelinated afferents and gave a similar staining pattern to that reported earlier (LaMotte et al., 1991; Rivero-Melian & Grant, 1991; Woolf et al., 1995). CTb-labelled Aδ fibres and terminals were found in lamina I, and in a band extending ventrally from the ventral half of lamina II to lamina V, but hardly any were present in the dorsal half of lamina II. There was no colocalization between CTb-stained axons and the CART peptide (Fig. 2M–P).

Many of the CART-immunoreactive axons showed VGLUT2 immunoreactivity (57.22 ± 4.29% in lamina I and 48.77 ± 8.36% in lamina II; Fig. 2Q–T).

Examination of the extent of colocalization of CART peptide-containing axons with interneuronal markers revealed that only a few of the former were immunoreactive for VGAT (2.42 ± 0.42% and 6.57 ± 1.71% in lamina I and II, respectively), and no colocalization was found with GLYT2 (Fig. 2U–X). In order to identify the subpopulations of excitatory interneurons that coexpress CART and VGLUT2, we used SP, NT and SOM antibodies in triple- and quadruple-labelling studies. These experiments showed that many of the CART-immunopositive boutons lacking CGRP were immunoreactive for SP (17.10 ± 0.10% and 33.9 ± 2.70% in lamina I and II, respectively; Fig. 2E–H) and/or for SOM (15.33 ± 1.83% and 25.5 ± 1.39% in lamina I and II, respectively; Fig. 2I–L), while the number of NT-labelled CART-ergic terminals was negligible (3.23 ± 0.55% and 3.38 ± 0.15% in lamina I and II, respectively; Fig. 2Q–T).

CART immunoreactivity was never colocalized with 5-HTT or DβH (see Supplementary material Fig. S2), which are markers of descending serotoninergic and noradrenergic pathways, respectively.

For the details of the quantitative analysis, see Table 2.

CART immunoreactivity in the DRG

A relatively small number (less then 10%) of CART-immunoreactive neuronal perikarya and axons were detected in sections from DRG. The stained cells belonged to the ‘small’ cell group. Most of them were also CGRP- and SP-positive, but there were a few cells that contained neither of these two peptides (Fig. 3).

Fixation and antibody specificity

The CART antibodies used in this study have been used frequently and are well characterized (Vrang et al., 1999; Koylu et al., 1997, 1998; Smith et al., 1999; Singru et al., 2007). Because the best CART staining appeared when the fixative contained 4% paraformaldehyde and 4% acrolein, we had to test the other markers with this fixative. All antibodies used showed a staining pattern in the spinal cord similar to that reported previously using paraformaldehyde-based fixatives without acrolein (Todd & Spike, 1993; Koylu et al., 1998; Ohsawa et al., 2000).

In our studies, we did not characterize the CART-positive cell bodies as these were rare and very weakly immunoreactive, which would have made neurochemical characterization unreliable.

Subpopulations of CART-immunoreactive terminals in lamina I and II

The neurochemical diversity of CART-containing fibres and cell bodies has been shown in various areas of the central and peripheral nervous systems. Coexistence of CART peptide with nitric oxide synthase or calbindin in intracardiac neurons (Richardson et al., 2006), with SP in neurons of the nucleus accumbens (Hubert & Kuhar, 2005), and with choline acetyltransferase in sympathetic preganglionic neurons of the thoracolumbar spinal cord (Dun et al., 2000) have all been demonstrated.

We have shown that the majority of CART-containing fibres in lamina I, together with a smaller population in lamina II, express CGRP. CGRP is a marker for peptidergic primary afferents, most of which are unmyelinated, and is never found in spinal interneurons (Carlton et al., 1987; Ju et al., 1987). In agreement with these findings, CART-positive small neurons expressing CGRP (and SP) were also found in DRG. However, the small number of CART-immunoreactive cells in the DRG seems to contradict the large number of CART/CGRP/SP axon terminals in the superficial laminae. As in the case of galanin (Zhang et al., 1993), a possible explanation could be that many DRG neurons synthesize CART peptide, but because axonal transport of the peptide to the terminals is very fast the peptide is undetectable in the cell bodies of most of them. A similar situation in the spinal cord probably explains the low number of CART-immunoreactive cell bodies in the dorsal horn.

Previous studies have reported that although most of the peptidergic afferents do not express VGLUT2 at detectable levels, some of the peptidergic fibres show a low level of immunoreactivity (Li et al., 2003; Oliveira et al., 2003; Todd et al., 2003; Alvarez et al., 2004; Landry et al., 2004). Because in our experiments we did not distinguish between the weak and strong VGLUT2 immunoreactivity, this could result in the relatively high proportion of VGLUT2-positive CART terminals in lamina I.

While the majority of CART axons in lamina I originate from primary afferents, colocalization of CART peptide with interneuronal markers becomes dominant in lamina II. Approximately half of the CART-immunoreactive terminals in lamina II contained VGLUT2, and more than 60% of the CART axons were immunoreactive for SP (34%), SOM (25%) or NT (3%) without CGRP (Table 2). It has been reported that the SOM- and NT-immunoreactive cells form two different sets of excitatory interneuronal subpopulations in the dorsal horn (Proudlock et al., 1993; Todd & Koerber, 2005), and almost all of the SP-, SOM- and NT-positive interneuronal terminals express VGLUT2 (Todd et al., 2003). In contrast to Tuchscherer & Seybold (1989), who reported no colocalization of SP and SOM in varicosities of the spinal dorsal horn, we found previously that a small proportion of CART terminals coexpress SP and SOM without CGRP (Kozsurek & Puskár, unpublished observation). The cross-reactivity between SP and SOM antibodies was excluded with radioimmunoassay applications performed by the suppliers. Differences in the fixatives and the resolution of the microscopes in the z-axis may yield these divergent results. This suggests, however, that while SP and SOM are present in two different populations of primary afferents (Hökfelt et al., 1976), they may be colocalized in a small population of interneurons, and this can explain the discrepancy in our quantitative data.

Functional implications

Physiological studies suggest that all SP-containing primary afferents function as nociceptors (Hylden & Wilcox, 1983; Kellstein et al., 1990; Lawson et al., 1997). Our results demonstrate that most of the CART axons in the superficial laminae belong to this type of nociceptor. It has also been demonstrated that one population of SP-containing primary afferents express mu opioid receptor on which morphine acts (Li et al., 1998). A possible mechanism of the potentiation of morphine action by CART peptide, reported by Damaj et al. (2004), could be the facilitation of the presynaptic morphine effect on CART/SP-containing primary afferents. To test this hypothesis, it will be necessary to test for the existence of mu-receptors on CART-positive afferents.

As the receptor for the CART peptide has not yet been identified, and because no appropriate antagonists are available, it is difficult, at present, to determine the exact role of the peptide in the spinal cord. However, it has been reported in behavioural studies that intrathecal application of CART(55–102) causes hyperalgesia in acute pain states (Ohsawa et al., 2000), and attenuates hyperalgesia and allodynia in neuropathic but not in inflammatory pain (Ohsawa et al., 2000; Damaj et al., 2004, 2006). In in vivo experiments, when the effect of CART peptide on thermal hyperalgesia induced by N-methyl-d-aspartate (NMDA) and α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) was studied in rats, the peptide alone had no effect on the tail flick latency. However, CART significantly enhanced NMDA- but not AMPA-induced nociceptive effects (Hsun Lin et al., 2005). In in vitro studies, using whole-cell patch-clamp measurements on substantia gelatinosa neurons in spinal cord slices, CART peptide alone caused no significant change of resting membrane potential, but increased the amplitude of NMDA-induced depolarization in substantia gelatinosa neurons and had little effect on AMPA-induced responses (Hsun Lin et al., 2005). These studies demonstrate that exogenously applied CART peptide selectively facilitates NMDA receptor-mediated glutamatergic sensory and nociceptive transmission. Although the exact roles of the excitatory interneuron populations containing SP, SOM or NT have not been determined, the subpopulations that express CART peptide, together with the CART-containing primary afferents, could be involved in the NMDA-mediated physiological effects including nociception.

The fact that both the C fibres and the interneurons that are CART-immunoreactive express SP and/or SOM raises the possibility that CART can affect not only glutamatergic neurotransmission but also the release and effects of SP or SOM.