Mode of Action
Pain-transmitting neurons are organized at three levels of the neuraxis: the midbrain, medulla and spinal cord. Pain-suppressing actions of neurons may be mediated partially by the endogenous opiate-like compounds endorphins. Neurons in the midbrain periaqueductal grey matter excite neurons in the rostral medulla. The medullary neurons, some of which contain serotonin, project to and inhibit trigeminal and spinal pain-transmitting neurons. Pain acts in a negative-feedback loop to inhibit pain transmission via this pain- suppression pathway (Basbaum and Fields 1978) but also in positive loops (Suzuki et al. 2005). The actions of the ant-depressants used in neuropathic pain are common in that they all block the re-uptake of the monoamines NA and serotonin; their differential effects on the synaptic levels of these modulatory transmitters varies from actions on both through to the SSRIs that are selective for serotonin (Millan 2002).
The ﬁrst rational for the use of these agents came from early studies based on stimulation produced analgesia (SPA), largely mediated through midbrain and brainstem areas that projected to the spinal cord through descending pathways. Over the years, this was extended to show similar sites could support opioid analgesia and that monoamines, both serotonin and NA, were important in both SPA and opioid actions in the brainstem. Later it became clear that there are descending noradrenergic pathways from the brainstem that modulate spinal activity with a relatively universal inhibitory action, mediated through spinal α2 adrenoceptors (Millan 2002). By contrast, the multitude of 5HT receptors suggested that this monoamine may play roles other than a simple inhibition of function.
There is now a great deal of evidence to suggest that the maintenance of chronic pain states, whether the result of nerve trauma or inﬂammation, is dependant to a large degree on descending pathways from the brainstem (Ur- ban and Gebhart 1999; Monconduit et al. 2002; Porreca et al. 2002; Ren and Dubner 2002). While in the early stages the emphasis was on descending in- hibitory pathways, it is now clear that descending facilitation is required for the full expression of chronic pain. Much of this research has focussed on the rostroventral medulla (RVM), an area that contains subsets of neurochemically distinct projection neurons some of which are serotonergic (5HT). Electrical stimulation of RVM results in inhibition or facilitation of spinal nociception, depending upon the intensity of the stimulus. Low-intensity stimulation tends to produce facilitation, and high intensity stimulation inhibition (Urban and Gebhart 1999; Fields 2004). Neurons in the RVM are able to code well for the intensity of noxious stimulus and have large receptive ﬁelds, covering most areas of the body. Stimulation of the RVM can enhance nociceptive and non- nociceptive input, and local anaesthetic block of the same area attenuates the mechanical hypersensitivity accompanying nerve injury. Based on responses to tail ﬂick nocifensive behaviour, cells in the RVM have been classiﬁed into three groups: on, off and neutral. The off cells are thought to produce a tonic in- hibition that is turned off by pain, but the ﬁnding of an increased cell discharge following prolonged noxious thermal stimulation led to the speculation that these cells facilitate nociception through activation of a descending pathway to the spinal cord. Destruction of ON-like cells abolished mechanical allodynia in the spinal nerve ligation (SNL) model of neuropathy (Porreca et al. 2001). RVM is in many ways seen as a relay between higher brain areas and the spinal cord and can be inﬂuenced by many areas of the brain involved in pain process- ing such as the periaqueductal grey (PAG), amygdala, and anterior cingulate and insular cortex. This is thought to allow the brain to make appropriate ad- justments to nociceptive processing informed by environmental contingencies such as threat or recovery from injury (Fields 2004). Thus it is likely that the proﬁle of anti-depressants in terms of their actions on NA and 5HT will lead to different functional effects on pain.
Over the past 4 years a general hypothesis concerning the regulation of spinal excitability and the control of chronic pain that builds upon previous work on both ascending and descending pathways has been proposed (Hunt 2000; Suzuki et al. 2002, 2004a). Essentially, starting from earlier observations, a key role of ascending pathways derived from a small group of neurons has been described. These neurons sit almost exclusively within laminae I and III of the dorsal horn of the spinal cord and:
– Express the substance P receptor (NK1) (Todd et al. 2000, 2002; Cheunsuang et al. 2002; Todd 2002)
– Are needed for both wind-up and long-term potentiation (LTP) of spinal neurons (Ikeda et al. 2003) following activation of nociceptive sensory af- ferents
– Project upon the brainstem particularly within areas that subsequently send information to the limbic system and somatosensory cortex (Gauriau and Bernard 2004).
Both neuropathic and inﬂammatory pain states are attenuated when this path- way is lesioned with a saporin-substance P conjugate applied intrathecally (Nichols et al. 1999). The lamina I–III/NK1 pathway was essential for the gen- eration of both ’wind-up’ and LTP in deep dorsal horn neurons: both were lost following ablation of lamina I neurons, the coding of peripheral stimuli and properties of deep dorsal horn neurons in rats with neuropathic injury that increase considerably in terms of their increased response to innocuous stimuli and their enlarged receptive ﬁeld sizes (Suzuki et al. 2002, 2004b). The link to the descending serotonergic pathway came from the observation that many of the effects of lamina I–III/NK1 lesions could be reproduced in part by either ablation of the descending serotonergic pathways or antagonism of the excitatory 5HT3 receptor with the selective 5HT3 receptor antagonist ondansetron given intrathecally over the dorsal horn.
We were also able to show that full activation of RVM serotonergic neurons did not occur following lamina I–III/NK1 ablation. Recent work has also in-dicated that local ablation of descending serotonergic pathways in the lumber spinal cord with 5,7 DHT can also severely attenuate the maintenance phase of both neuropathic and inﬂammatory pain states. Since this research, other work has begun to show some efﬁcacy of ondansetron in clinical trials with chronic pain patients (McCleane et al. 2003) and so-called ’at level’ pain that develops close to the site of a spinal cord transection has been ameliorated by reducing serotonergic function (Oatway et al. 2004).
There is compelling evidence to suggest that the ionotropic 5HT3 receptor mediates many of the pro-nociceptive effects of 5HT, although there is also evidence to suggest a role for 5HT2 receptors. Both are expressed by primary afferents and dorsal horn neurons mainly pre-synaptically (Miquel et al. 2002; Zeitz et al. 2002; Maxwell et al. 2003; Okamoto et al. 2005). The efﬁcacy of the 5HT3 receptor antagonist ondansetron on at-level allodynia following cord damage in rat is thought to reﬂect sprouting of axotomized serotonergic axons at the site of lesion and increased release of 5HT (Oatway et al. 2004). Within the DRG, 5HT3R is found on a small proportion of CGRP-containing sensory neu- rons and in mouse expressed by an uncharacterized group of small-diameter myelinated and unmyelinated nociceptors (Miquel et al. 2002; Zeitz et al. 2002). The evidence strongly suggests a role for 5HT3 receptors and release of 5HT as essential for the maintenance of chronic pain states. However, consistent inhibitory effects of 5HT1 receptor activation have been reported in animals, and 5HT1B/D agonists have established roles in migraine. In the case of non- cranial pains, the 5HT1A receptor appears to be the key target for descending 5HT inhibitory actions (Millan 2002).
Noradrenergic inhibition mediated through α2 receptors would therefore seem a key issue for effectiveness of anti-depressants in pain, since 5HT mech- anisms can both enhance and inhibit pain.
Neuropathic pain is successfully treated with anti-depressant drugs. Of neuropathic pain patients treated with anti-depressants, 30% will achieve over 50% pain relief (McQuay et al. 1996). Anti-depressant anti-nociceptive action is independent of the presence of clinical depression (Clifford 1985). Although literature exists that shows that neither SSRIs nor TCAs seem to have superior anti-nociceptive effects to the other (see Mattia et al. 2002), older TCAs do ex- hibit superior anti-nociceptive effects to other anti-depressants when used in the treatment of chronic neuropathic pain (Mattia and Coluzzi 2003). Collins et al. (2000) found no data that would support the idea that SSRIs are bet- ter analgesics than older anti-depressants. According to Max, zimelidine and paroxetine, two SSRIs, have weak clinical effectiveness in the treatment of neu- ropathic pain (Max 1994). Clinically, there is no signiﬁcant difference between TCAs. However, TCAs are signiﬁcantly more effective than benzodiazepines according to three studies available for review. Imipramine, a TCA, is more ef- fective than paroxetine, an SSRI, and mianserin, a tetracyclic anti-depressant (McQuay et al. 1996).