Review
`
`Molecular studies of CGRP and the
`CGRP family of peptides in the central
`nervous system
`
`Cephalalgia
`2019, Vol. 39(3) 403–419
`! International Headache Society 2018
`Article reuse guidelines:
`sagepub.com/journals-permissions
`DOI: 10.1177/0333102418765787
`journals.sagepub.com/home/cep
`
`Erica R Hendrikse1, Rebekah L Bower1,þ
`Christopher S Walker1
`
`, Debbie L Hay1,2
`
`and
`
`Abstract
`Background: Calcitonin gene-related peptide is an important target for migraine and other painful neurovascular
`conditions. Understanding the normal biological functions of calcitonin gene-related peptide is critical to understand
`the mechanisms of calcitonin gene-related peptide-blocking therapies as well as engineering improvements to these
`medications. Calcitonin gene-related peptide is closely related to other peptides in the calcitonin gene-related peptide
`family of peptides, including amylin. Relatedness in peptide sequence and in receptor biology makes it difficult to tease
`apart the contributions that each peptide and receptor makes to physiological processes and to disorders.
`Summary: The focus of this review is the expression of calcitonin gene-related peptide, related peptides and their
`receptors in the central nervous system. Calcitonin gene-related peptide is expressed throughout the nervous system,
`whereas amylin and adrenomedullin have only limited expression at discrete sites in the brain. The components of two
`receptors that respond to calcitonin gene-related peptide, the calcitonin gene-related peptide receptor (calcitonin
`receptor-like receptor with receptor activity-modifying protein 1) and the AMY1 receptor (calcitonin receptor with
`receptor activity-modifying protein 1), are expressed throughout the nervous system. Understanding expression of the
`peptides and their receptors lays the foundation for more deeply understanding their physiology, pathophysiology and
`therapeutic use.
`
`Keywords
`Amylin, adrenomedullin, calcitonin, CGRP, central nervous system, migraine
`
`Date received: 18 November 2017; revised: 19 February 2018; accepted: 26 February 2018
`
`Introduction
`
`Calcitonin gene-related peptide (CGRP) is an import-
`ant sensory neuropeptide that is involved in pain modu-
`lation (1,2). CGRP has attracted particular interest in
`regard to its role in migraine and likely plays a role in
`other primary headache disorders and painful condi-
`tions (2,3). Presently, three investigational classes of
`drug aim to reduce CGRP activity to prevent and
`treat migraine (3). These are small molecule antagonists
`against a CGRP receptor and antibodies that either
`block receptor activity or bind directly to the CGRP
`peptide. This ‘first generation’ of CGRP-based treat-
`ments will
`likely lead to further drugs over time,
`which will result from a deeper knowledge of the
`CGRP system at a cellular and molecular level and a
`greater understanding of the molecular neuroanatomy
`of CGRP and its receptors. This information guides
`
`understanding of the relative role of central and periph-
`eral processes in which CGRP participates in the patho-
`physiology of migraine, and other disorders.
`Understanding CGRP biology is complex due to the
`existence of similar peptides in the same family, and
`shared receptor complexes. The aim of this review is
`
`1School of Biological Sciences, University of Auckland, Auckland,
`New Zealand
`2Centre for Brain Research, University of Auckland, Auckland,
`New Zealand
`þ
`Current address: Department of Pharmacology and Toxicology,
`The University of Otago, Dunedin, New Zealand
`
`Corresponding author:
`Christopher S Walker, School of Biological Sciences, University of
`Auckland, Private Bag 92 019, Auckland 1142, New Zealand.
`Email: cs.walker@auckland.ac.nz
`
`EX2073
`Eli Lilly & Co. v. Teva Pharms. Int'l GMBH
`IPR2018-01427
`
`1
`
`

`

`404
`
`Cephalalgia 39(3)
`
`to summarize information about the expression of
`CGRP and to compare this to its binding sites and
`molecularly defined receptors. We will detail what is
`known about the expression of CGRP and related pep-
`tides, where binding sites for labelled peptides are
`found, and what
`information is available for the
`molecular correlate(s) to that binding. We focus our
`attention on the central nervous system (CNS) and per-
`ipheral ganglia, with emphasis on brain regions that are
`of particular relevance to migraine. The sensory cir-
`cumventricular organs (CVOs) will also be considered,
`as circulating factors can act directly on these sites
`within the brain.
`
`Migraine and the nervous system
`
`Historically, there has been considerable debate around
`the importance of central and peripheral mechanisms of
`migraine. This initially took the form of the competing
`vascular and neuronal hypotheses for migraine. It is
`now clear, based on recent clinical trials, that migraine
`can be treated through peripheral blockade of CGRP
`action (3,4). This does not preclude a central origin of
`migraine. Although the peripheral aspects of the trige-
`minovascular system are important in migraine, it is
`still worthwhile to consider central contributions (5).
`Numerous regions of the CNS are activated during a
`migraine attack and central phenomena are associated
`with migraine symptoms, such as the link between cor-
`tical spreading depression and migraine aura (3). It is
`unclear whether blocking central CGRP action may be
`beneficial or perhaps a hindrance in treating migraine.
`However, central CGRP receptors
`should be an
`important consideration for developing a ‘second gen-
`eration’ of CGRP system-based treatments with poten-
`tially greater effectiveness or fewer side effects.
`This review will focus on the craniofacial pain path-
`way and other brain regions associated with migraine.
`The craniofacial pain pathway begins with Ad and C
`
`fibres of the trigeminal nerve, whose cell bodies are
`located peripherally in the trigeminal ganglia and pro-
`ject centrally primarily into the spinal
`trigeminal
`nucleus (STN) of the brainstem and into the C1/C2
`levels of the spinal cord. The STN is proposed as a
`possible
`site of migraine
`initiation and displays
`increased activity immediately prior to a migraine
`attack (6). Higher order processing of painful signals
`primarily involves the thalamus,
`insular cortex and
`somatosensory cortex.
`Interconnected with these
`regions are other sites, including the amygdala, raphe
`nuclei, periaqueductal gray (PAG), parabrachial area,
`gracile nucleus and locus coeruleus, which play roles in
`pain processing, proprioception, stress or aversion.
`Interestingly, in migraine patients, altered connectivity
`between these regions is reported, such as from the thal-
`amus to the insular cortex and somatosensory cortex or
`from the PAG to the cortex and amygdala (7,8). Other
`brain regions involved in migraine may include the
`hypothalamus,
`implicated in attack initiation;
`the
`hippocampus, which displays greater pain-induced
`activity in migraine patients; and the cortex, cerebellum
`and visual network, which may be involved in cortical
`spreading depression and symptoms of migraine aura
`(9–11).
`
`The calcitonin gene-related peptide family
`of peptides
`
`CGRP belongs to a small family of structurally related
`peptides (Figure 1). There are two forms of CGRP, a
`and b. The broad term ‘‘CGRP’’ is used for either a or
`b, unless specifically noted. The other major members
`of this family are amylin, adrenomedullin (AM) and
`adrenomedullin 2 (AM2). Another member of this
`family,
`the calcitonin receptor-stimulating peptide
`(CRSP) was reported in some mammals (12); however,
`CRSP is not expressed in primates or rodents. Each of
`these peptides contains a conserved pair of cysteine
`
`aCGRP
`bCGRP
`CT
`AM
`AM2
`Amylin
`
`AC-DTATCVTHRLAGLLSRSGGVVKNN-FVPTN-VGSKAF–NH2
`AC-NTATCVTHRLAGLLSRSGGMVKSN-FVPTN-VGSKAF–NH2
`CGNLSTCMLGTYTQDFNKFHTF-------PQTAIGVGAP–NH2
`GC-RFGTCTVQKLAHQIYQFTDKD-KDNVAPRSKISPQGY–NH2
`GC-VLGTCQVQNLSHRLWQLMGPAGRQDSAPVDPSSPHSY–NH2
`KC-NTATCATQRLANFLVHSSNNFGAI-LSSTN-VGSNTY–NH2
`
`Figure 1. Amino acid sequences of calcitonin gene-related peptide family members. Alignment of the amino acid sequences for
`human aCGRP, bCGRP, calcitonin (CT), AM, AM2 and amylin using the single letter code. AM and AM2 have been truncated at the
`N-terminal. All the peptides have a C-terminal amide (-NH2). A conserved disulfide bond between the two N-terminal cysteine
`residues is indicated by a solid line.
`
`2
`
`

`

`Hendrikse et al.
`
`405
`
`form a disulfide bond, creating a
`that
`residues
`loop. They also all contain a C-terminal amide.
`Amylin and CGRP are the most closely related pep-
`tides in terms of amino acid sequence, resulting in over-
`lapping actions
`in pain modulation and nutrient
`balance (13,14). The similarities between the peptides
`within the CGRP family causes significant overlap in
`their ability to activate each other’s receptors. This
`‘‘blurred’’ receptor pharmacology is particularly evi-
`dent at rodent receptors, where there is noticeable
`cross-reactivity. This makes working with this peptide
`family challenging because it is often very difficult to
`determine which of the many possible molecularly
`defined receptors actually mediates an effect for a
`given peptide (15).
`
`Receptor composition and pharmacology
`
`The molecular composition of the CGRP family of
`receptors is illustrated in Figure 2, which highlights
`the overlapping activity of the peptides at the different
`receptors. Overlap occurs, not only because the pep-
`tides have shared features but also because the recep-
`tors
`themselves have shared and closely related
`components. All of the receptors within this family
`are G protein-coupled receptors (GPCRs). Two specific
`GPCRs form high affinity receptors for the different
`peptides due to their association with accessory pro-
`teins. There are three of these accessory proteins, recep-
`tor activity-modifying protein (RAMP) 1, 2 and 3. The
`CGRP, AM1 and AM2 receptors consist of the calci-
`tonin receptor-like receptor
`(CLR) and RAMP1,
`
`CGRP Receptors
`
`Adrenomedullin receptors
`
`CGRP
`
`CGRP > AM
`
`AM1
`AM > AM2
`
`Amylin receptors
`
`AM2
`AM = AM2
`
`AMY1
`CGRP = Amylin
`
`Calcitonin receptor
`
`AMY2
`Amylin
`
`AMY3
`Amylin
`
`Legend
`
`CTR
`
`CT > Amylin
`
`CLR
`
`CTR
`
`RAMP1 RAMP2 RAMP3
`
`Figure 2. The calcitonin gene-related peptide (CGRP) receptor family. The relative potency of CGRP, calcitonin (CT), AM, AM2 or
`amylin is shown below a schematic representation of the appropriate receptor complex (17). The potential overlap between the
`CGRP and amylin receptors is highlighted in purple. The receptor components are described in the legend.
`
`3
`
`

`

`406
`
`Cephalalgia 39(3)
`
`RAMP2 or RAMP3, respectively (16,17). The AMY1,
`AMY2 and AMY3 receptors consist of the calcitonin
`receptor (CTR) with each RAMP (17,18). CLR alone
`does not appear to act as a functional receptor, whereas
`CTR is the receptor for calcitonin (16,17). Figure 2
`shows that CGRP can activate both CLR/RAMP1
`and CTR/RAMP1 equally. CGRP is weaker at other
`human receptor complexes, but there are clear devia-
`tions in other species (15). At rat CLR and CTR-based
`receptors, CGRP is potent at both RAMP3-based
`receptors (19). Amylin can potently activate CTR/
`RAMP1, CTR/RAMP2 and CTR/RAMP3 but not
`the CLR-based receptors. AM activates all
`three
`CLR/RAMP complexes but has a preference for
`CLR/RAMP2 and 3 (17).
`One pharmacological method for teasing apart clo-
`sely-related receptors with overlapping peptide binding
`parameters is to use specific antagonists. Although sev-
`eral antagonists have been reported that are capable of
`antagonizing members of the CGRP receptor family,
`they generally lack the required specificity. For exam-
`ple, although CGRP8-37, an N-terminally truncated
`form of CGRP, is often cited as a specific CGRP recep-
`tor antagonist, it does not display sufficient specificity
`to be practically useful. CGRP8-37 is less than 10-fold
`selective for CGRP receptors over AMY1 and is cap-
`receptors
`able of antagonizing AM2 and AMY3
`(19–22). Similarly, CTR and amylin receptor antagon-
`ists including truncated salmon calcitonin (sCT8-32) and
`AC187 are potent antagonists, but do not distinguish
`between amylin receptor subtypes (19,21). The develop-
`ment of the small molecule ‘‘gepant’’ class of drugs has
`yielded some useful tools for separating the CGRP
`receptor from AM receptors and,
`if used carefully,
`receptors
`can tell apart
`the CGRP and AMY1
`(Table 1). For instance, olcegepant displays greater
`than 10,000-fold selectivity for the CGRP over AM
`receptors and 100–200 fold selectivity for the CGRP
`receptor over the AMY1 receptor (23–25). However,
`other gepants including telcagepant and MK-3207 are
`less selective and the degree of selectivity measured may
`
`depend on other variables, including the signaling path-
`way measured (23,26). With the careful selection of sev-
`eral different agonist and antagonist concentrations, it
`may be possible to pharmacologically characterize
`members of the CGRP receptor family present in a
`particular cell
`line or tissue. However, for practical
`use in in vivo systems more specific pharmacological
`tools are required.
`
`Challenges associated with determining the molecular target for
`a given peptide. To understand the biology of the CGRP
`peptide family, it is crucial to define the molecular iden-
`tity of the peptide and receptor responsible for a bio-
`logical effect in physiological systems. The overlapping
`pharmacology described above represents a major chal-
`lenge associated with studying this important family of
`receptors. Therefore, pharmacological
`approaches
`should be complemented with other methods. The
`measurement of mRNA can be a useful guide.
`However, the detection of mRNA does not necessarily
`mean that protein is present, especially in neurons
`where proteins can be transported to projections dis-
`tant from the cell body (27). Stable membrane proteins
`including CLR, CTR and RAMPs may not need high
`levels of mRNA expression to maintain protein expres-
`sion in steady-state cells. This has been illustrated
`where moderate to high levels of immunoreactivity
`were observed in regions where the corresponding
`mRNA was low or not detected (28,29).
`The development of techniques including high-reso-
`lution confocal
`imaging has allowed researchers to
`detect the precise cellular localization of a receptor
`using antibodies. However, when used to directly
`detect protein expression, antibodies also have limita-
`tions. They require comprehensive validation to con-
`firm specificity and selectivity for their targets (30). In
`many cases, antibodies are used without sufficient evi-
`dence of validation and it may not be possible to draw
`the conclusions that the authors suggest. For example,
`studies may simply lack the required information
`regarding the antibodies used to allow evaluation of
`
`Table 1. Summary of small molecule CGRP receptor antagonists (gepants) for migraine treatment.
`
`Small molecule
`
`Olcegepant
`
`Telcagepant
`
`MK-3207
`
`BHV-5000
`
`Atogepant
`
`Ubrogepant
`
`Rimegepant
`
`Blocks CGRP
`receptor
`Blocks AMY1
`receptor1
`Clinical
`development
`
`Yes
`
`Yes
`
`Yes
`
`Yes
`
`Yes
`
`Yes
`
`Yes
`
`Yes (100–200
`fold lower)
`
`Development
`stopped
`
`Yes (50–100
`fold lower)
`
`Development
`stopped
`
`Yes (50
`fold lower)
`
`Development
`stopped
`
`No data
`
`No data
`
`No data
`
`No data
`
`Preclinical
`
`Phase 2
`
`Phase 3
`
`Phase 3
`
`No data, no published data is available.
`1The relative affinity or potency for the AMY1 receptor compared to the CGRP receptor is reported in parenthesis (20,23,26,154).
`
`4
`
`

`

`Hendrikse et al.
`
`407
`
`the data (31) or the antibodies used have been shown to
`recognize multiple proteins. This prevents observed
`immunoreactivity from being conclusively associated
`with a single protein (30). The latter situation has
`been a significant issue for antibodies designed to
`detect RAMPs (32).
`Even if a receptor subunit has been identified with a
`well-validated antibody, this alone is not meaningful
`for this receptor family, except when considering calci-
`tonin activity at CTR alone. Expression of CLR or
`CTR by themselves says little about
`the receptor
`phenotype without
`further examining potential co-
`localization with a RAMP. RAMPs are also known
`to modify the signaling of other receptors, forming
`alternative complexes that complicate interpretation
`(33). However, valuable information has been obtained
`from studies showing co-localization between CLR and
`receptor-component protein or between CLR and
`RAMP1, indicating the likely presence of the CGRP
`receptor (34–36). The ideal solution maybe to develop
`antibodies that specifically target the heteromeric com-
`plex. This strategy was employed to detect the CLR/
`RAMP1 complex in the trigeminal ganglia, dura mater
`and the spinal trigeminal nucleus (37). However, anti-
`bodies that recognize heteromeric complexes are diffi-
`cult
`to generate and characterize. Thus, protein
`expression data should be used alongside other experi-
`mental observations. For example, mRNA and peptide
`binding correlations were informative in early studies
`validating CLR/RAMP complexes as receptors for
`CGRP and AM (38). Another consideration relates to
`the level of expression reported. For many receptors,
`there is considerable amplification in signal following
`ligand-receptor binding. Hence, relatively low expres-
`sion does not automatically translate
`into little
`function.
`Regardless of method, much research relies on quali-
`tative description of the biological target within ana-
`tomical locations. This is a highly subjective process,
`and is open to individual interpretation of intensity
`and location. Collating data, improving its accessibility,
`and standardisation all help to address these issues. For
`example, the Human Protein Atlas, the Allen Brain
`Atlas and other web-based tools provide excellent
`resources for expression data, assisting identification
`of patterns (39–41).
`
`Peptide accessibility to the CNS and relevance of receptors in
`different locations. The blood-brain barrier (BBB) is an
`important consideration when looking at the origin,
`expression and activity of neuropeptides in the CNS.
`Peptides present in the blood may cross the BBB either
`through active transport or passive diffusion (42,43).
`Alternatively, circulating peptides can interact directly
`with CVOs. CVOs are highly vascularized, with
`
`fenestrated capillaries, allowing peptides access to
`these discrete parts of the CNS. The CGRP family of
`peptides do not freely cross the BBB (44). The propor-
`tion of peripheral amylin that can cross the BBB is low
`and is unlikely to be sufficient to activate receptors
`inside
`the BBB at physiological
`concentrations
`(45,46). Interestingly, the highest amount of labelled
`amylin that was reported to penetrate the brain was
`observed in the hypothalamus and medulla (46),
`which are associated with sensory CVOs and thus are
`permeable to circulating hormones. Hence, relatively
`high levels of amylin experimentally reaching these
`regions is not surprising. The reported actions of
`exogenous amylin at sites inside the BBB, such as the
`ventral tegmental area, are unlikely to be explained by
`amylin crossing the BBB (47–49). Alternative explan-
`ations could include local production of amylin or
`another member of the CGRP peptide family triggering
`receptor activation in these regions. The BBB pene-
`trance of CGRP and AM have been examined primar-
`ily in vascular models, which suggest that neither
`peptide can cross the BBB when the vascular endothe-
`lium is intact (50–52). A more definitive study has been
`performed for AM, which, under normal conditions,
`does not appear to cross the BBB and penetrate the
`brain (53). Curiously, definitive experiments have not
`been performed for CGRP. However, given the lack of
`brain penetrance displayed by calcitonin, amylin and
`AM and data from vascular models, significant cross-
`ing of the BBB by CGRP seems unlikely. Overall, the
`low BBB permeability for CGRP family peptides sug-
`gests that CNS expression of these peptides is required
`for them to have activity inside the BBB.
`
`AM and AM2
`
`AM and AM2 expression in the CNS. AM is best known as
`a regulator of the cardiovascular and lymphatic sys-
`tems. It is a potent vasodilator and is highly expressed
`in the vascular endothelium (54,55). Vascular AM may
`be involved in maintaining the BBB, cerebral circula-
`tion and the volume of cerebrovascular fluid (56–58).
`Expression in the vasculature may complicate expres-
`sion analysis
`that
`relies on homogenized tissue,
`including mRNA analysis and membrane binding.
`mRNA and immunoreactivity for AM has been
`reported in the human and rat brain, localized to the
`vasculature, the choroid plexus and in neurons and glia
`of the hypothalamus, cerebellum and medulla (59–63).
`However, the cerebellum, which has been implicated in
`AM-regulated blood pressure control, was the only
`brain region where mature AM peptide was detected
`(59,64). Interestingly, in a genetic mouse model that
`was modified to lack AM expression in the CNS,
`altered responses to pain, anxiety and stress were
`
`5
`
`

`

`408
`
`Cephalalgia 39(3)
`
`observed (65,66). Similarly, AM has been reported to
`induce pain, and AM-like immunoreactivity was
`detected in dorsal root ganglia neurons and axon ter-
`minals in the dorsal horn from rats (67,68). These stu-
`dies suggest that AM may play a physiological role in
`the CNS despite limited expression.
`The physiology of AM2 is poorly understood; how-
`ever, central or peripheral administration can modulate
`blood pressure (69). AM2 mRNA and immunoreactiv-
`ity has been found in the pituitary and hypothalamus of
`humans and rats as well as the spinal cord and dorsal
`root ganglia of rats (70–74). AM2 may be involved in
`the hypothalamo-pituitary axis and the central control
`of blood pressure in the hypothalamus (69,75–77).
`
`AM and AM2 binding sites in the CNS. AM binding sites
`have been measured in rat and human brain homogen-
`ates, with high levels of 125I-AM binding in the hypo-
`thalamus, thalamus and spinal cord (72,78–80). AM2
`binding sites in the CNS have not been specifically stu-
`died. More work is required to determine the preva-
`lence and location of AM and AM2 binding sites
`within the brain.
`
`Molecular composition of AM and AM2 binding sites in the
`CNS. CLR, RAMP2 and RAMP3 are the most relevant
`to the AM peptides (Figure 2; Hay et al., 2017). The
`expression of CLR mRNA and protein has been stu-
`died extensively within the nervous system, where it is
`widespread (Table 2 (28,70,81–83)). Thus, the presence
`of functional AM1 and AM2 receptors is dependent
`upon the expression of RAMP2 or RAMP3, respect-
`ively. The expression of mRNA for these receptor com-
`ponents has been well documented in the vasculature
`(50,84,85). However, there is only limited data describ-
`ing RAMP2 or RAMP3 in discrete rat nervous system
`regions (Table 2). In rats, abundant mRNA for
`RAMP2 was detected in the hypothalamus, and
`RAMP3 was detected in nuclei of
`the thalamus
`(86,87). Relatively high RAMP2 expression was also
`detected in the hippocampus and spinal cord (70,86).
`RAMP2 and RAMP3 have not been examined in a
`human brain (Table 3). Unfortunately, there are cur-
`rently no antibodies for the RAMP2 or RAMP3 pro-
`teins that have been sufficiently characterized for use in
`immunohistochemical mapping (33). Characterized
`antibodies that have been used for the detection and
`mapping of CLR, CTR and RAMP1 are discussed
`later.
`
`Calcitonin
`
`Calcitonin expression in the CNS. Canonically, calcitonin is
`expressed by thyroid C cells and regulates calcium
`(88–90). The CALCA gene
`homeostasis
`encodes
`
`calcitonin and CGRP mRNA (91). Tissue-specific alter-
`native splicing of CALCA mRNA results in the prefer-
`ential expression of CGRP mRNA in the nervous
`system; calcitonin mRNA is not detectable in these
`regions (1,88,92,93).
`
`Calcitonin binding sites in the CNS. Although calcitonin
`does not appear to be expressed in the nervous
`system, binding sites for calcitonin are found in many
`brain structures (94–96). Salmon calcitonin has fre-
`quently been used as a tool in calcitonin studies because
`it has higher affinity than human or rat calcitonin and
`can bind CTR alone, AMY1, AMY2 and AMY3 recep-
`tors with high affinity (18). Thus, depending upon the
`presence or lack of a RAMP, salmon calcitonin binding
`may indicate the presence of a receptor for amylin,
`CGRP or calcitonin. Salmon calcitonin binding sites
`have been examined in rat and primate nervous systems
`using autoradiography (Tables 2 and 3). The data are
`generally consistent between these studies with only
`minor differences. Overall, autoradiography shows
`high levels of salmon calcitonin binding in the hypo-
`thalamus, sensory CVOs and the nucleus accumbens,
`along with moderate binding in the bed nucleus of the
`stria terminalis (BNST), PAG and locus coeruelus
`(95,97–100). The spinal cord has minimal salmon calci-
`tonin binding in the dorsal horn (95,97). There also
`seems to be very little binding in the cerebral cortex
`and in the cerebellum. Data from the human brain is
`more limited. In a single study, salmon calcitonin bind-
`ing was examined by autoradiography in the human
`medulla (Table 3). Discrete binding sites were observed
`in several nuclei, including high densities of binding in
`the NTS and raphe nuclei. Very little binding was
`observed in the fibre tracts (96). In another study,
`human hypothalamic membranes showed high levels
`of salmon calcitonin binding (101). The variability in
`displacement of salmon calcitonin by human calcitonin
`between brain regions suggests that AMY receptors
`account for a significant proportion of salmon calci-
`tonin binding sites (102,103). However, more work is
`required to determine the prevalence and location of
`calcitonin binding sites within the human brain, and
`their direct relevance to calcitonin, amylin and/or
`CGRP activity.
`
`in the
`Molecular composition of calcitonin binding sites
`CNS. Consistent with widespread calcitonin binding
`sites, CTR mRNA and protein expression has been
`observed throughout the nervous system of rodents
`(Table 2). Several splice variants of the CTR have
`been reported. These variants are not necessarily con-
`served across species and can display altered functions
`(104). This review does not distinguish between these,
`because there are no tools to determine the expression
`
`6
`
`

`

`Hendrikse et al.
`
`409
`
`(81,86,87,97,98,115,145,146,150,151,158,170)
`
`(29,81,86,87,97,99,105,115,139,145,162,170)
`
`(29,81,86,87,115,139,145,147,151,158,170)
`
`(29,86,87,92,97-99,105,115,139,145,146,148,
`
`151,152,162-164,170,171)
`
`(86,87,92,97-99,105,115,139,145,146,148,151,
`
`(81,86,87,92,99,115,145,146,150,151,162,169)
`
`152,162-164,170)
`
`148,150,158,162-164,168)
`
`(29,81,86,87,92,97,98,105,114,115,139,145-
`
`MP
`
`(29,86,92,97,98,105,115,139,145,153,162)
`
`164,167)
`
`(29,47,81,92,97-99,105,115,139,145-148,162-
`
`(29,81,87,92,97,105,115,145-147,162,163,166)
`
`(81,86,87,97-99,105,115,143,145-148,151)
`
`145-148,150,151,158,162-164)
`
`(29,81,82,86,87,92,97-99,105,115,132,139,
`
`146,158,161,165)
`
`(29,67,70,82,86,87,92,97,101,131,139,140,
`
`P
`
`(115,151)
`
`(115)
`
`(87,99,115,139,164)
`
`(99,115,139,145,146,162)
`
`(87,99,115,145,146,150,151)
`
`(99,105,115,139,145,146,150,151,162,163)
`
`(29,45,70,107,128,131,158-161)
`
`(20,29,35,92,93,132,133,138,139,155-158)
`
`MP
`
`P
`
`P
`
`P
`
`MP
`
`MP
`
`P
`
`MP
`
`MP
`
`P
`
`MP
`
`MP
`
`P
`
`MP
`
`MP
`
`P
`
`P
`
`P1
`P1
`
`MP
`
`MP
`
`References
`
`Peptide
`
`Y
`
`Y
`
`Y
`
`Y
`
`Y
`
`Y
`
`N
`
`Y
`
`Y
`
`Y
`
`Y
`
`Y
`
`Y
`
`Y
`
`Y
`
`N
`
`Y
`
`N
`
`N
`
`N
`
`Y
`
`Y1
`
`Y
`
`N
`
`Y
`
`Y
`
`N
`
`Y
`
`Y
`
`Y
`
`Y
`
`N
`
`N
`
`N
`
`Y
`
`Y
`
`Y
`
`Y
`
`Y
`
`Y
`
`Y
`
`Y
`
`Y
`
`Y
`
`Y
`
`Y
`
`Y
`
`Y
`
`Y
`
`Y
`
`M
`
`125I-sCTCGRPAmylin
`
`125I-Amy
`
`125I-CGRP
`
`Bindingsites
`
`1Verylowexpression.
`ordatanotavailable;CVOs:circumventricularorgans;OVLT:organumvasculosumoflaminaterminalis.
`M:mRNAdetectedbypolymerasechainreactionorinsituhybridisation;P:protein/immunoreactivityintissues;Y:specificradioligandbindingdetected;N:nobindingdetected;Emptyspace:noexpression
`
`M
`
`MP
`
`MP1
`
`P
`
`M
`
`M
`
`P
`
`P
`
`Cerebralcortex
`
`Septalarea
`
`Hippocampus
`
`Basalganglia
`
`Amygdala
`
`Thalamus
`
`MP
`
`MP
`
`Hypothalamus
`
`MP
`
`MP
`
`MP
`
`P
`
`grey
`
`P
`
`Periaqueductal
`
`P
`
`MP
`
`Midbrain
`
`Pons
`
`Cerebellum
`
`MP
`
`MP
`
`Medulla
`
`MP
`
`MP
`
`Spinalcord
`
`Pinealgland
`
`organ
`
`Subcommisural
`
`Medianeminence
`
`Pituitary
`
`OVLT
`
`Subfornicalorgan
`
`Areapostrema
`
`CVOs
`
`P
`
`MP
`
`MP
`
`Dorsalrootganglia
`
`Trigeminalganglia
`
`CLRCTR
`
`Region
`
`Receptorsubunit
`
`Table2.ExpressionofCGRPfamilyreceptorcomponentsandpeptidesandtheirbindingsitesinrodents.
`
`M
`
`M
`
`M
`
`M
`
`M
`
`M
`
`M
`
`M
`
`M
`
`M
`
`M
`
`M
`
`M
`
`M
`
`M
`
`M
`
`M
`
`M
`
`M
`
`M
`
`M
`
`M
`
`M
`
`M
`
`M
`
`M
`
`M
`
`MP
`
`P
`
`MP
`
`M
`
`M
`
`MP
`MP1
`
`M1P
`
`M
`
`M1
`
`M
`
`MP
`
`P
`
`MP
`
`MP
`
`MP
`
`M
`
`M
`
`M
`
`MP1
`MP
`
`M
`
`M
`
`M
`
`MP
`
`P
`
`RAMP3
`
`RAMP2
`
`RAMP1
`
`7
`
`

`

`410
`
`Cephalalgia 39(3)
`
`Table 3. Expression of CGRP family receptor components and peptides and their binding sites in primates including humans.
`
`Receptor subunit
`
`Binding sites
`
`Peptide
`
`Region
`
`CLR CTR RAMP1 RAMP2 RAMP3
`
`125I-CGRP
`
`125I-Amy
`
`125I-sCT CGRP Amylin References
`
`Trigeminal ganglia
`
`MP
`
`P
`
`P
`
`Dorsal root ganglia MP
`
`CVOs
`
`Area postrema
`
`MP
`
`P
`
`MP
`
`Subfornical organ
`
`OVLT
`
`Pituitary
`
`MP
`
`Median eminence MP
`
`Subcommisural
`organ
`
`Pineal gland
`
`Spinal cord
`
`Medulla
`
`MP
`
`MP
`
`MP
`
`P
`
`MP
`
`MP
`
`MP
`
`MP
`
`MP
`
`Y
`
`Y
`
`Y
`
`N
`
`Y
`
`Y
`
`Y
`
`Y
`
`Y
`
`Y
`
`Y
`
`N
`
`MP
`
`MP
`
`(20, 35, 37, 83, 133, 172)
`
`(140)
`
`(28, 95, 96, 106, 163)
`
`(95)
`
`(95)
`
`(28, 101)
`
`(28)
`
`(28)
`
`P
`
`P
`
`P
`
`(28, 82, 95, 101, 140)
`
`(20, 28, 37, 82, 95, 96,
`106, 147, 163, 173)
`
`(34, 95, 101, 147)
`
`MP
`
`MP
`
`MP
`
`MP
`
`MP
`
`Cerebellum
`
`Pons
`
`Midbrain
`
`MP
`
`MP
`
`MP
`
`Periaqueductal grey MP
`
`MP
`
`P
`
`Hypothalamus
`
`Thalamus
`
`Amygdala
`
`Basal ganglia
`
`Hippocampus
`
`Septal area
`
`Cerebral cortex
`
`Y
`
`Y
`
`Y
`
`Y
`
`Y
`
`Y
`
`Y
`
`Y
`
`P
`
`P
`
`NS
`
`Y
`
`Y
`
`N
`
`Y
`
`Y
`
`Y
`
`Y
`
`NS
`
`Y
`
`Y
`
`Y
`
`Y
`
`Y
`
`Y
`
`Y
`
`Y
`
`Y
`
`Y
`
`Y
`
`(28, 95, 147, 163, 173)
`
`(28, 95, 101, 147, 163)
`
`(28, 95, 147, 173)
`
`(28, 95, 101, 147, 163)
`
`(95, 147)
`
`(95, 147, 163)
`
`(95, 147, 163)
`
`(95)
`
`(95)
`
`(95)
`
`M: mRNA detected by polymerase chain reaction or in situ hybridisation; P: protein/immunoreactivity in tissues; Y: specific radioligand binding detected;
`N: no binding detected; Blank space: no expression or data not available; NS: non-specific binding; CVOs: circumventricular organs; OVLT: organum
`vasculosum of lamina terminalis.
`1Very low expression or binding.
`
`profile of specific splice variants at the protein level. We
`discuss CTR only in general due to limited data, even at
`the mRNA level. In the peripheral nervous system,
`CTR staining was observed in neuronal cell bodies in
`the
`trigeminal and dorsal
`root ganglia (20,29).
`Centrally, dense protein expression was reported in
`fibres and cell bodies of brain regions including the
`nucleus accumbens, the substantia nigra, the NTS and
`the area postrema (105). Moderate staining of cell
`bodies and fibres was observed in the spinal cord,
`amygdala, PAG and dorsal
`raphe nucleus
`(105).
`Limited but discrete staining has also been observed
`in the thalamus, BNST and regions of the somatosen-
`sory cortex (29,105). Interestingly, CTR protein has
`also been reported in several regions involved in the
`perception and higher-order processing of craniofacial
`pain (Table 2). The pattern of CTR expression in
`human neural tissues is similar to that of rodents,
`although data are limited to the trigeminal ganglia
`and the medulla (Table 3;
`(20,106). Widespread
`
`expression of CTR was observed in the human medulla,
`including the NTS, STN, cuneate, gracile and hypo-
`glossal nuclei (106). Considering the lack of calcitonin
`expression in central and peripheral nervous tissues and
`its limited ability to cross the BBB, it is surprising that
`CTR is so widely expressed within the nervous system.
`This suggests that the CTR probably functions in con-
`cert with one or more RAMPs inside the nervous
`system to bind amylin, CGRP, or another peptide.
`
`Amylin
`
`Amylin expression in the CNS. The pancreatic b-cell is a
`major site of amylin expression and releases amylin into
`the circulation to induce meal-ending satiation (13).
`Early reports show a lack of amylin expression in the
`nervous
`system (107,108). However,
`later work
`reported limited amylin expression within the brain
`and in sensory neurons (Table 2; (109–111). It is
`likely that some amylin antibodies can also detect
`
`8
`
`

`

`Hendrikse et al.
`
`411
`
`CGRP and some CGRP antibodies can detect amylin.
`Potential cross-reactivity between amylin and CGRP
`antibodies may account for some discrepancies between
`studies (112). Hence, care should be taken to confirm
`that antibodies detect only the desired target. More
`recently, amylin mRNA expression was reported in
`the preoptic area and BNST neurons of lactating
`dams (113). Amylin mRNA and immunoreactivity
`has also been reported in hypothalamic neurons in
`mice (114). Although amylin expression has not been
`examined in the human brain (Table 3), the limited
`amylin expression in the CNS suggests that its central
`actions likely occur predominantly outside the BBB
`through the CVOs, although discrete expression in a
`limited number of CNS regions is possible.
`
`Amylin binding sites in the CNS. Despite the reportedly low
`expression of amylin in the CNS, 125I-amylin binds with
`widespread distribution in rat brain wh