Study of the receptors at which CGRP acts is complicated by their multi-subunit complexity, and until comparatively recently attempts to isolate or clone the re- ceptor were hampered by the fact that it was not a single protein gene product. Some information on the localisation of the CGRP receptor was obtained from experiments with an antibody mixture raised against puriﬁed CGRP binding sites isolated from pig cerebellum (e.g. Ye et al. 1999), but as we know now that the receptor is made up from three different proteins that can individually also have roles in other receptors of the family, this earlier work should be treated with caution (see Brain et al. 2002; Poyner et al. 2002). The CGRP receptor consists of a seven-transmembrane G protein-coupled calcitonin-like receptor (CRLR) unit linked to receptor activity modifying protein 1 (RAMP1) and to receptor component protein (RCP), which is important in the coupling of re- ceptor activation to stimulation of adenylyl cyclase (Brain et al. 2002). RAMPs and RCP are widely distributed in the body, suggesting that they interact with other receptors also (e.g see Oliver et al. 2001: Hay et al. 2006). It is important to note that CRLR can have different selectivity according to which RAMP it is associated with; with RAMP2, for example, it forms a receptor more selective for adrenomedullin and with RAMP3 a receptor sensitive to both CGRP and adrenomedullin (Hay et al. 2006). This makes it very difﬁcult to use mapping of the localisation of an individual protein of the triad to identify where CGRP receptors are functionally located, although it has been claimed that RCP maps well with the location of CGRP immunoreactivity (Ma et al. 2003) and RCP expression is up-regulated in DRG and DHSC following peripheral inﬂamma- tion. In addition to a widespread distribution in the CNS, the components of the CGRP receptor are also expressed in the human cerebral vasculature (Oliver et al. 2002). The receptor localisation issue is further complicated by the observation that some structures such as the cerebellum will bind CGRP, yet there is a relative lack of RAMP1 (Oliver et al. 2001).
Those receptors containing CRLR and RAMPs appear to signal through cyclic AMP (cAMP) and intracellular calcium with coupling by way of Gq or Gα (Hay et al. 2006). Intrathecal injection of CGRP in rats produced hyper- algesia that could be reduced with a protein kinase (PKA) or PKC blocker, suggesting the involvement of both these second messenger systems (Sun et al. 2004). In neonatal or adult rat DRG neurons, CGRP increased cAMP and produced phosphorylation of cAMP-response element binding protein (CREB), suggesting that CGRP can modulate gene expression (Anderson and Seybold 2004).
The complicated nature of the CGRP receptor has so far made it impossible to produce a genetically modiﬁed mouse lacking functional CGRP receptors. Thus, further elucidation of the function of CGRP in nociception has been cru- cially dependent on the availability of selective pharmacological antagonists.
Studies with CGRP Receptor Antagonists
The ﬁrst useful antagonists of the effects of peptides of the CGRP family were truncated analogues of the peptides themselves such as CGRP 8–37 or amylin
22–52 (see Brain and Grant 2004; Hay et al. 2004; Van Rossum et al. 1997). In vascular and other smooth muscle tissues, especially in vitro, these peptide antagonists have proved generally useful in spite of some issues around lack of biological stability (see Brain and Grant 2004). In the CNS results have not been as consistent and CGRP 8–37 sometimes shows agonist properties (e.g. Riediger et al. 1999).
Low-intensity rat spinal cord electrical stimulation induces cutaneous va- sodilatation that is blocked by CGRP 8–37 but not by hexamethonium (Tanaka et al. 2001) as is vasodilatation in the dental pulp and lip following stim- ulation of the rat inferior alveolar nerve (Kerezoudis et al. 1994) and the cerebro-dilator response following nasociliary nerve stimulation in the cat (Goadsby 1993). Using intravital microscopy in anaesthetised rats to visualise dural vessel dilatation, Williamson and colleagues (1997a, b) showed that the dilatation produced either by CGRP given systemically or neurogenic dilata- tion produced by perivascular electrical nerve stimulation were blocked by i.v. CGRP 8–37. This is suggestive of an anti-migraine proﬁle for CGRP re- ceptor antagonists which is also supported by the observation that triptans reduce the neurogenic dilatation of dural vessels (Williamson et al. 1997b) and that the terminals which release CGRP express the 5-HT1B/D receptors at which the triptans act to relieve migraine headache (Ma et al. 2001). The excitation of neurons within the dorsal horn of the rat spinal cord by nox- ious stimulation of the limbs in the presence or absence of inﬂammation and by ionophoretic application of exogenous CGRP was blocked by ionophoretic CGRP 8–37 (Ebersberger et al. 2000; Neugebauer et al. 1996). Excitatory re- sponses evoked by ionophoretic application of substance P, neurokinin A or
AMPA (S-α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid) were un- affected by CGRP 8–37 co-application (Neugebauer et al. 1996). Deep tissue
inﬂammation up-regulated CGRP expression within the trigeminal ganglion and lowered the somatic withdrawal thresholds to noxious stimulation, and this was prevented by infusing CGRP 8–37 prior to the inﬂammatory stimulus
(Ambalavanar et al. 2006). Thermal injury to one hind paw of a rat produced swelling, inﬂammation and a lowered threshold for paw withdrawal which was reversed by intrathecal CGRP 8–37 (Lofgren et al. 1997) as was the ther-
mal and mechanical allodynia in a chronic central pain model (Bennett et al.
2000). A colonic pain model where thresholds to distention were reduced by either CGRP instillation or by nerve growth factor (NGF) or brain-derived neurotrophic factor (BDNF) administration was also found to be sensitive to blockade by CGRP 8–37 (Delafoy et al. 2006). In contrast, exogenous CGRP microinjected into the periaqueductal grey matter can raise the nociceptive
threshold in rats, but this effect too can be blocked by co-administration of
CGRP 8–37 (Yu et al. 2003).
Non-peptide antagonists of the CGRP receptor have been sought for many years, but it is only recently that highly potent and selective agents that are active in vivo have been discovered. The discovery that SK-N-MC (neuroblastoma– neuroepithelioma) cells were a facile source of membranes expressing human CGRP receptors enabled a high-throughput screening campaign leading to quinuclidine compounds with micromolar afﬁnity and antagonist properties as shown by blockade of the ability of CGRP to raise cAMP levels in SK-N-MC cells (Daines et al. 1997). SB-273779 is a more potent non-peptide antagonist (IC50 310 nM) that showed some in vivo activity against CGRP-induced falls in systemic blood pressure but its low potency, short half-life and poor in vivo tolerability at higher doses limit its utility (Aiyar et al. 2001). Rudolph and colleagues (2005) were the ﬁrst to report highly potent antagonists of the action of CGRP suitable for clinical evaluation. A high-throughput screening effort led them to a series of (R)-Tyr-(S)-Lys dipeptide-like compounds, and lead optimisation produced BIBN4096, the ﬁrst picomolar antagonist that is active both in vitro and in vivo (Rudolph et al. 2005). The major drawback of this compound is low oral bioavailability (<1%) such that clinical proof of concept studies had to be performed with i.v. administration.
BIBN4096 was found to reduce both spontaneous and evoked activity of sin- gle neurons in the spinal trigeminal nucleus with receptive ﬁelds in the dura mater (Fischer et al. 2005). Supraspinal neurons within the central nucleus of the amygdala of anaesthetised rats showed an increase in plasticity after peripheral inﬂammation that was reversed after intra-amygdala injection of BIBN4096 or CGRP 8–37 (Han et al. 2005). Species selectivity has to be consid- ered when using BIBN4096 as a tool in rodents, and doses tenfold higher than those effective at human or primate receptors may be needed (see Hershey et al. 2005). Dilation of human cerebral arteries in vitro by CGRP was potently reversed by BIBN4096 (Moreno et al. 2002). In the ﬁrst reported randomised, double-blind, phase II clinical trial in migraine patients (Olesen et al. 2004) BIBN4096 at a dose of 2.5 mg i.v. was effective in relieving the headache and clearly differentiated from placebo. In a volunteer study (Petersen et al. 2005) BIBN4096 2.5 mg i.v. prevented the headache produced by an infusion of CGRP
(1.5 μg/min for 20 min).
Indirect evidence on the putative role of CGRP in pain is provided from
studies on the experimental analgesic drug cizolirtine as it has been shown to inhibit the spinal release of both substance P and CGRP in rats (Ballet et al. 2001) and is effective in rat models of neuropathic pain (Aubel et al.
2004; Kayser et al. 2003). Block of the effects of substance P does not produce analgesia in humans (see previous section of this chapter), thus any efﬁcacy in cizolirtine is likely to be due to an effect on CGRP signalling. In a double-blind cross-over study in neuropathic pain patients, cizolirtine 200 mg p.o. b.i.d did not reduce overall pain experience but did reduce primary allodynia in the
subgroup with this symptom (Shembalkar et al. 2001). At doses up to 150 mg it was ineffective against the pain following third molar extraction (Matthew et al. 2000); against the pain of renal colic, 350 mg of cizolirtine was also ineffective (Pavlik et al. 2004).
There is thus convincing preclinical and clinical evidence to show that blockade of CGRP receptors is likely to be an effective treatment for migraine
headache. Some preclinical evidence suggests that blockade of CGRP receptors might have a more general utility in the treatment of painful conditions, but clinical studies with the release inhibitor cizolirtine (which must be treated cautiously given that a mechanism involving CGRP has not been conﬁrmed
in humans) only support a use against allodynia in patients with neuropathic pain.
Kinins as Mediators of Pain and Inflammation
The topic of kinins as mediators of pain and inﬂammation has recently been reviewed in detail (Rupniak et al. 2000; Calixto et al. 2000; 2001; Couture et al.
2001) so it will only be dealt with brieﬂy here. When kinins are produced by tissue damage the nonapeptide bradykinin appears rapidly followed by its metabolite des-[Arg9] bradykinin and by des-[Arg10] kallidin. All three peptides are pharmacologically active but with differing receptor selectiv-
ity (Rupniak et al. 2000). The production of bradykinin and kallidin from largely inactive precursors depends on cleavage by serine proteases known as kallikreins (although tissue and plasma kallikreins belong to different enzyme families). Bradykinin and its decapeptide analogue kallidin are then further cleaved by carboxypeptidase to produce the des-[Arg] analogues (Couture et al. 2001; Rupniak et al. 2000).
The kinins are directly algogenic and also release other mediators such as prostaglandins and cytokines which amplify the local inﬂammatory and noci- ceptive response. They also release transmitters from the terminals of primary afferents, including substance P and CGRP and may exert a pro-inﬂammatory effect by way of sympathetic ﬁbres. Bradykinin (as a partially puriﬁed biolog- ical extract) was ﬁrst shown to produce pain in humans by Armstrong and his colleagues in 1951 when they applied it to an exposed blister base in volunteers (cited by Rupniak et al. 2000). Bradykinin is released from muscles during con- traction as part of normal physiology but when injected into a muscle it can evoke pain (Boix et al. 2005). Exercise studies show an association between per- ceived pain and levels of bradykinin and kallidin produced in the muscle (Boix et al. 2004). Bradykinin is able to increase the gain of peripheral nociceptive mechanisms by disinhibiting the TRPV1 receptor (Chuang et al. 2001) on pri-
mary afferents and also potentiates glutamate operated synaptic transmission in the spinal dorsal horn (Wang et al. 2005). It appears that roughly half of all primary afferent nociceptor ﬁbres are sensitive to kinins under baseline con- ditions but that this proportion rises to 80% in the presence of inﬂammation (Rupniak et al. 2000).
The traditional idea of the kinin system being exclusively peripheral has now passed and it is accepted that kinins are found within the CNS, for example in the spinal cord (Lopes and Couture 1997). The increased sensitivity to opioids that has been reported in animals following the induction of peripheral inﬂammation may be due to a central action of kinins to increase opioid receptor trafﬁcking (Patwardhan et al. 2005).
It is important to remember that kinins also have a role in smooth muscle contraction and relaxation, control of blood pressure, vasodilation and vascu-
lar permeability such that this may constrain the way in which their effects can be modulated to produced pain relief (Calixto et al. 2001).
Kinin Receptors and Pathophysiology
Kinins exert their effects through two receptors called the bradykinin 1 (B1) and bradykinin 2 (B2) receptors on the basis of a distinct pharmacological classiﬁcation on isolated tissues obtained by a study of the agonist peptides themselves and of a number of peptide analogues with a variety of agonist, partial agonist and antagonist properties (Couture et al. 2001). The existence of the two distinct receptors has been conﬁrmed by molecular cloning performed by Hess and his colleagues (cited in Rupniak et al. 2000). Both receptors are members of the GPCR family (see Oliver et al. 2000) and there is only some 38% structural homology between them (Couture et al. 2001). It has recently been shown that a proteolytic B1 and B2 receptor complex with enhanced signalling capacity can form and become inserted in the plasma membrane (Kang et al.
2004). Early pharmacological studies suggested clear species differences in the selectivity of kinin receptors for agonist and antagonist ligands, and this has now been conﬁrmed by studies on cloned receptors from a number of species. For the B1 receptor in particular, there are striking cross species differences,
with des-[Arg10] kallidin having much higher afﬁnity for the human receptor
than does des-[Arg9] bradykinin whereas on rat and mouse B1 receptors the situation is reversed (Jones et al. 1999; Rupniak et al. 2000). In most species, the preferred agonist for the B2 receptor is bradykinin itself, whereas the des-
[Arg] peptides are the preferred ligands for the B1 receptor (Rupniak et al.
2000). Under most circumstances the B2 receptor is constitutive, although it
can be up-regulated by inﬂammatory mediators (Seabrook et al. 1997; Calixto
et al. 2001) whereas the B1 receptor is inducible and normally only expressed by peripheral tissues when they are exposed to an inﬂammatory or damaging stimulus (Couture et al. 2001; Marceau et al. 1998; Rupniak et al. 2000). Consti-
tutive B1 receptors may be found in the CNS, B1 and B2 receptors were found in the human brain by Raidoo and Bhoola (1997) and B1 receptors have been found in spinal cord of rat, mouse, monkey and human (see Calixto et al. 2004). It appears that under some circumstances the expression of B2 receptors can be down-regulated (Baptista et al. 2002; Kang et al. 2004). It is noteworthy that the human B1 receptor has ligand-independent constitutive activity as it lacks critical epitopes that regulate activity (Leeb-Lundberg et al. 2001) although the physiological signiﬁcance of this is unclear and homo-oligomerisation of this receptor may be needed for cell surface expression (Kang et al. 2005).
The role of B1 receptors in pain and inﬂammation has recently been re- viewed in depth (Calixto et al. 2004) and there is an extensive literature on the B2 receptor (reviewed in Rupniak et al. 2000). Evidence from human studies, however, is limited at the present time. Tissue kallikrein, B1 and B2 recep- tors are expressed in mast and giant cells inﬁltrating oesophageal squamous cell carcinomas (Dlamini and Bhoola 2005) and receptors have been found in numerous other tumours (Calixto et al. 2004). It is likely that kinins act as mitogens to promote cell proliferation in addition to exacerbating any inﬂam- matory response in the tumour. Studies in human volunteers have shown that UV-B irradiation of skin enhanced pain signalling through both B1 and B2 receptors activated by skin microdialysis of des-[Arg] kallidin or bradykinin respectively (Eisenbarth et al. 2004). Arthritic rats showed an enhanced B1- mediated extravasation after 5 days of joint inﬂammation (Cruwys et al. 1994) and lipopolysaccharide (LPS)-induced paw oedema was potentiated by the B1
agonist des-[Arg] bradykinin (Ferreira et al. 2000). Interleukin-1β seems to be
an especially potent inducer of B1 receptor expression and activation of nuclear factor (NF)-κB via B2 receptors also stimulates B1 expression (Couture et al.
2001). Platelet-activating factor (PAF) also activates NF-κB and up-regulates
B1 receptors and this is believed to be via TNF-α and subsequent interleukin (IL)-1β release (Fernandes et al. 2005). In adrenalectomised rats, B1 receptor expression was increased and this could be reversed by administering dex-
amethasone or by inhibiting NF-κB with PDTC (pyrrolidine dithiocarbamate) (Cabrini et al. 2001). In sciatic nerve injured mice there was a switch from B2 to B1 receptor signalling with increased time after lesioning (Rashid et al. 2004).
Functional B1 receptors are expressed by nociceptive sensory neurons and expression is up-regulated by glia-derived neurotrophic factor (GDNF ;Vellani et al. 2004). There is also evidence for the involvement of p38 in the develop- ment of inﬂammation in the rat and the consequent increase in B1 receptor expression (Ganju et al. 2001). Acute chronic cystitis in rat bladder induces ex- pression of B1 receptors whereas only B2 receptors are seen in the non-inﬂamed state (Chopra et al. 2005). Intrathecal administration of a B2 agonist to mice produces thermal hyperalgesia that peaks 10 min after injection whereas in- trathecal injection of a B1 agonist produces persistent thermal hyperalgesia lasting up to 1 h (Calixto et al. 2001), suggesting the receptors localised within the spinal cord are functional. In the cardiovascular system B1 receptors are
involved in inﬂammatory pathologies such as endotoxic shock, atheromatous disease and myocardial ischaemia, although it is not yet clear whether acti- vation of B1 receptors is pathological or protective (McLean et al. 2000). In rats with kindled epilepsy, induction of B1 receptors resulted in potentiation of glutamate release in cortex and hippocampus (Bregola et al. 1994; Mazzuferi et al. 2005).
Both B1 and B2 receptors appear to be involved in nociception and inﬂam- matory processes, and experiments with transgenic animals and/or selective receptor antagonists were needed to elucidate the respective role of each re- ceptor (Calixto et al. 2004).