We have previously reported in detail the distribution of CGRP and its receptor components in the human trigeminal ganglion . Small to medium sized neurons of the trigeminal ganglion store CGRP and have further connections with STN in the brainstem, and in related extensions down to the C1-2 level via un-myelinated C-fibers . CLR and RAMP1 are expressed in large neurons and thick fibers, which are co-expressed with a marker for Aδ-fibers. These fibers from the trigeminal ganglion extend further to the brainstem. Several studies point to the involvement of the brainstem in migraine pathophysiology [1, 19] and a region within the brainstem has been demonstrated to be active during attacks in migraine patients [14, 20].
The present study has revealed the detailed distribution and relations between CGRP and its receptor components in human and rat projection regions from the trigeminal ganglion. The study demonstrates for the first time the expression of CLR and RAMP1 within human STN in the brainstem and in the spinal cord at C1-level, using a series of novel well characterized antibodies. Thus, this indicates CGRP signaling in areas of the brainstem/spinal cord putatively involved in the migraine pathophysiology. The reason for choosing these regions in the present study is based on a series of detailed neuroanatomical mappings in rat showing pain pathways that may be involved in primary headaches .
Distribution of CGRP, CLR and RAMP1 in STN
We observed CGRP expression around nerve fiber bundles in the superficial laminae of human and rat STN. We found no CGRP expressing neurons in the STN neither in man nor in rat. However, in rat we observed a small group of CGRP positive neurons located close to the central canal, in the inferior olive and in the hypoglossal nucleus of the brainstem. Notably these neurons were absent in man. The results were confirmed using 2 different antibodies against CGRP. The functional role of CGRP in these regions remains to be disclosed. The presence of CGRP in the brainstem is supported by early autoradiography studies [22, 23]. In addition, CGRP expression in STN and spinal cord of different species has been studied [6, 24–26]
The expression of CGRP in the brainstem differs between species . It has been shown that the distribution of CGRP fibers is similar in rat and alpaca brainstem. However, CGRP containing neurons are more widespread in rat than in alpaca. In addition, the localization of CGRP positive neurons in the cat and alpaca brainstem differs . Hence, the distribution pattern observed between rat and man in our study is likely due to species differences.
CGRP and MBP double-staining showed no co-localization, indicating that CGRP is as expected expressed in un-myelinated fibers in STN and this is in agreement with a previous study on the trigeminal ganglion . Synaptophysin, used as a marker for synaptic vesicles, and CGRP were expressed in the same structures within the STN. This supports that CGRP is stored and released from nerve terminals as supported by a study on rat dorsal horn  and acts post-synaptic at CLR/RAMP1 expressing fibers.
In human STN, we found CLR and RAMP1 positive fibers mostly in the spinal trigeminal tract, spanning towards the superficial laminae. In rat STN, expression of CLR and RAMP1 were also found mostly in the spinal trigeminal tract, around fiber bundles and fibers in the spinal trigeminal tract spanning towards the superficial laminae. These results are in contrast to an earlier study by Lennerz et al., 2008, where CLR and RAMP1 expression was detected in the superficial laminae, partially co-localizing with CGRP. The reason for this difference could reside in that different antibodies were used.
There were no neuronal cell bodies immunopositive for CLR and RAMP1 in the human or rat STN. CLR and RAMP1 were observed to be co-expressed in the fibers, suggesting the presence of functional CGRP receptor in both types of species. Interestingly, we found that the receptor components were co-expressed on the walls of capillaries in rat STN
In rat STN, it was suggested that CGRP and its receptor components are localized in terminals from primary afferents . In contrast, CLR did not co-localize with neuropeptides of primary spinal afferents in the dorsal horn of rat . In the present work, we observed no co-localization between the receptor components and CGRP positive fibers, neither in man nor in rat. These results suggest that CGRP and the receptor components appear in nerve terminals, where C-fiber released CGRP may act post-synaptic at CGRP receptors on second-order neurons or modify the responses of trigeminal Aδ-fibers.
Distribution of CGRP, CLR and RAMP1 in C1
We examined the localization and expression of CGRP and its receptor components in the spinal cord at the C1-level, since the main part of the trigeminovascular projection occurs at this level . CGRP positive fibers were found in laminae I and II. Similar staining pattern has been demonstrated previously in the chick, quail dorsal horn of the spinal cord  and in the cat . In addition, we observed some CGRP positive neurons and fibers close to the central canal (laminae X). This finding is in agreement with a previous study in rat spinal cord with neurons being positive for CGRP .
To determine which fibers express CGRP, co-localization experiments were performed with CGRP and IB4. IB4 has previously been used as a marker for C-fibers . In the present study we found that IB4 and CGRP could be expressed in the same laminae, but most often in different types of fibers. This observation was confirmed by confocal microscopy. The IB4 staining was more prominent in the deeper laminae of the spinal cord, which is in agreement with a study of rat spinal cord . It has been shown that the degree of co-expression of CGRP and IB4 in neurons vary in the rat. More neurons expressing both markers are found in dorsal root ganglia compared to the trigeminal ganglion of rat . Electron microscopy showed that IB4 and CGRP expressing axons were distinct, but both could be present in the same bundle of un-myelinated fibers .
The MBP marker showed that in some areas within laminae I and II myelinated fibers are absent. Co-staining experiments with CGRP showed that CGRP is expressed in areas that are absent of MBP, suggesting that CGRP is indeed expressed in un-myelinated fibers.
To further scrutinize the CGRP positive fibers, double-staining of CGRP and synaptophysin was performed. In laminae I and II, where CGRP staining was found, co-expression of CGRP and synaptophysin was observed. With confocal microscopy, we obtained a detailed three-dimensional view of the staining pattern. This clearly showed that CGRP and synaptophysin were detected in the same fibers.
Reportedly, CLR and RAMP1 are expressed in fibers within laminae I and II in the dorsal horn of rat spinal cord . Similar observation but in different tissue was seen in our study; CLR and RAMP1 expressions were found within laminae I and II of human and rat C1. No CLR or RAMP1 positive neuronal cell bodies were observed at the C1-level. In rat, CGRP was expressed in the same laminae, but it did not co-localize with CLR or RAMP1. CLR and RAMP1 were co-expressed, suggesting expression of functional CGRP receptor in fibers within laminae I and II. The presence of receptor components in the spinal cord is supported by the mRNA expression of RAMP1 and RCP, detected with specific oligonucleotides for in situ hybridization .
CLR and RAMP1 were in addition detected in fibers close to the central canal. As described above, CGRP positive fibers were also detected in this area. Interestingly, tracing experiments in cat have revealed projections from the periaqueductal gray (PAG) region to the spinal cord. Horseradish peroxidase (HRP) injections into the PAG region resulted in labeled fibers close to the central canal, terminating in laminae X of C1 and C2-levels . The same authors found labeled fibers in segments C4 to T8 adjoining the ependymal layer of the central canal and next to the basal membrane of the nearby capillaries. Histochemical studies in different species have revealed neurons, axons and terminals within laminae X containing neuropeptides such as substance P . Thus, our results demonstrate fibers containing CGRP and its receptor components close to the central canal. The function of this is not known; one may speculate that these fibers can release CGRP directly into the cerebrospinal fluid or stimulate the ependymal cells of the central canal.
Methodology and technical considerations
The hematoxylin-eosin stained material revealed well-preserved human tissue adequate for immunofluorescence technique, even though the tissue samples were collected 24 to 36 hrs after death. Due to the relatively high age of the subjects, lipofuscin is accumulated in the tissue, causing auto-fluorescence. We have earlier examined, in rat trigeminal ganglion, if storage of the animals for 24 hrs at +4°C prior fixation would affect the immunohistochemistry; which was not the case . Similar results were obtained in the present study on the rat brainstem (data not shown). In rat we also performed a direct comparison of perfusion-fixed and immersion-fixed brainstems, and found no observable difference in antibody expression pattern.
The C1-level compared well in the Paxinos atlas for rat and man. The STN is a structure that is distributed for a considerable length in the brainstem and could therefore not be examined for its entire distribution. It is a limitation of the present study that we only studied a portion of the STN; the exact part is given in the method part.
In our previous study, antibodies against human and rat CLR and RAMP1 were generated, and the specificity of the antibodies was confirmed in HEK293 cells stably expressing the human CGRP receptor. The specificity of the raised antibodies was also confirmed by Western blotting . The same antibodies were used in this study.
The staining of the receptor components was weaker and a bit more diffuse in the human tissue compared to rat. This could be due of several factors: differences in antibodies recognizing the epitopes, tissue condition, or low level of CLR and RAMP1 in human tissue. In rat STN and C1, the RAMP1 antibody displayed a stronger staining pattern compared to the CLR antibody. If this was due to differences in antibody recognizing the epitopes or more expression of RAMP1 in these areas could not be verified.
Within the rat brainstem, we found expression of RAMP1 and CLR in the capillary walls. This was blocked with the specific blocking peptides (epitopes used in the production of the respective antibodies). The staining of the capillaries was similar in the endogenously activity blocked DAB-stained sections.