The anti-nociceptive effects of anti-depressants require intact descending in- hibitory bulbospinal pathways. A study by Ardid (1995) shows that the anti- nociceptive effects of clomipramine are suppressed only in the hindpaw ipsi- lateral to a unilateral lesion of the dorsolateral funiculus of the rat (Ardid et al.
1995). The effects of clomipramine, amitriptyline and desipramine were tested on a neuropathic pain model in rats induced by loosely tied ligatures around
the common sciatic nerve. Acute injections of clomipramine, amitriptyline and desipramine caused pain relief. Chronic injections of these three TCAs
resulted in pain relief as measured by a signiﬁcant and progressive increase in vocalization threshold.
Sierralta et al. (1995) use p-chlorophenylalanine and α-methyltyrosine to
examine the anti-depressant drugs clomipramine, zimelidine, imipramine and
maprotiline in the acetic acid writhing test in mice. Each anti-depressant demonstrated an anti-nociceptive effect in this study. This study shows that critical levels of both 5-HT and NA are responsible for mediating the anti- nociceptive effects of anti-depressants on the writhing test in mice (Sierralta et al. 1995).
The streptozocin-induced diabetic rat is a model of chronic pain with signs of hyperalgesia and allodynia. The TCAs clomipramine, amitriptyline and desipramine, as well as clonidine, an α-adrenergic stimulating agent, were studied in this pain model to show that noradrenergic drugs seem to be the
most active of these drugs that act on monoaminergic transmission to cause pain relief (Courteix et al. 1994).
Fishbain et al. (2000) conducted a review of 22 controlled animal studies and 5 double-blind placebo-controlled studies that examined anti-depressants in various pain models. This group found that anti-depressants that atten- uate both serotonin and norepinephrine levels have greater anti-nociceptive
activity than anti-depressants that solely act to modulate the NA levels. Fur- thermore, anti-depressants that act solely to modulate serotonin levels have weaker anti-nociceptive properties than the two previously mentioned classes of anti-depressants (Fishbain et al. 2000). Bomholt (2005) in a study of various anti-depressants in the chronic constriction injury rat model of neuropathic pain suggests that anti-depressant drugs that act on both serotonin and NA levels have greater anti-nociceptive effects than SSRIs (Bomholt et al. 2005). The anti-depressants amitriptyline, duloxetine, mirtazapine and citalopram were able to attenuate thermal hyperalgesia in the chronic constriction injury rat model of neuropathic pain. Interestingly, only amitriptyline, a TCA, and duloxetine, a dual serotonin-norepinephrine reuptake inhibitor, fully reversed thermal hypersensitivity. Amitriptyline, duloxetine and mirtazapine, a nora- drenergic and speciﬁc serotoninergic anti-depressant (NaSSA) caused a sig- niﬁcant reduction of mechanical hyperalgesia, whereas citalopram, an SSRI, was ineffective in attenuating mechanical hyperalgesia. Mechanical allodynia was not affected by any of these four anti-depressants (Bomholt et al. 2005).
Which receptors underlie the effects of the increased levels of monoamines in synapses as a result of their primary action on uptake mechanisms? Mico (1997) provides supporting evidence that β-adrenoceptors play a role in the analgesic effect of desipramine and nortriptyline. However, only the β1 adren- ergic receptor is involved when the painful stimulus is chemical, as tested by the non-neuropathic acetic acid and formalin tests (Mico et al. 1997). However, these are not neuropathic states, and further studies that show α-adrenoceptors play a role in the physiology of pain transmission. Anti-depressant anti- nociception is mediated by α2-adrenoceptors, and not α1-adrenoceptors, and not only by drugs that act by re-uptake inhibition of NA, but also of serotonin (Gray et al. 1999).
As seen in the treatment of diabetic neuropathy, the TCA imipramine has an NNT of 1.4 in a study with optimal doses. Other studies of tricyclics show an NNT of 2.4 in this pain disorder. Furthermore, SSRIs have an NNT of 6.7. The NNT is 3.3 for CBZ, 10.0 for mexiletine, 3.7 for GBP, 1.9 for dextromethorphan,
3.4 for both tramadol and levodopa, and 5.9 for capsaicin (Sindrup et al. 1999). In diabetic neuropathy, the NNT for anti-depressants was 3.4 and the NNT
for anti-convulsants was 2.7 (Collins et al. 2000). Anti-depressants and anti- convulsants had the same efﬁcacy and incidence of minor adverse effects in for diabetic neuropathy and PHN (Collins et al. 2000) In PHN, TCAs have an
NNT of 2.3. The NNT for PHN is 3.2 for GBP, 2.5 for oxycodone and 5.3 for capsaicin. Dextromethorphan was inactive in PHN (Sindrup et al. 1999). In PHN, the NNT for anti-depressants was 2.1 and the NNT for anti-convulsants was 3.2 (Collins et al. 2000). The NNT was 2.5 for tricyclics and 3.5 for capsaicin
in peripheral nerve injury (Sindrup et al. 1999).
In the treatment of pain in polyneuropathy, TCAs and anti-convulsants have become the conventional pharmacotherapy. According to the study by Sindrup and Jensen (2000) TCAs are the drugs of choice in the pharmacological treat-
ment of pain in polyneuropathy. GBP, CBZ, and tramadol are good alternatives to TCAs if contraindications or tolerability problems are encountered with this class of drugs. Sindrup et al. (1999) found no obvious relationship be- tween the mechanism of action of these drugs and their effect in distinct pain conditions or for speciﬁc drug classes and various pain conditions (Sindrup et al. 1999). McQuay et al. (1996) reviewed 21 placebo-controlled treatments including data on 10 different anti-depressants in 17 randomized controlled trials. An NNT of 3 for anti-depressants when compared with placebo shows signiﬁcant pain relief in the 6 of 13 studies of diabetic neuropathy. PHN studies have an NNT of 2.3. The odds ratio for anti-depressants from two atypical facial pain studies is 4.1 and the NNT is 2.8. From the only central pain study with analysable dichotomous data, the NNT for anti-depressants was estimated to be 1.7 (McQuay et al. 1996).
However, since the early anti-depressants have a number of potential phar- macological targets, there may be non-monoamine actions of these com- pounds, although it is unclear that they contribute to the clinical effects of these compounds. Skolnick et al. (1996) examined the effect of chronic anti- depressant treatment on NMDA receptors. Adaptive changes in radioligand binding to NMDA receptors resulted from chronic (14 days), but not acute (1 day), anti-depressant administration to mice. The TCA imipramine, the SSRI citalopram, and electroconvulsive shock slowly produce these adaptive changes that persist for some time after treatment has been stopped. How- ever, the ability of SSRIs, weakly effective in patients, to produce these actions, seemingly limited to the cerebral cortex, argue against a contribution to the clinical actions of anti-depressants in pain.
Valverde et al. (1994) have shown that anti-nociception associated with TCAs acts partly via the endogenous opioid system and partly by additional activation of noradrenergic and serotonergic pathways, although this study did not address nerve injury. De Felipe et al. (1985) found that rats treated chron- ically for 21 consecutive days with the typical anti-depressants clomipramine, desipramine and amitriptyline, as well as with the atypical anti-depressants iprindole and nomifensine, had increased levels of [Met5]enkephalin in stria- tum and nucleus accumbens.
Hamon et al. (1987) examined the central mechanisms involved in anti- depressant potentiation of morphine-induced analgesia. Levels of Leu-enke- phalin, the opioid peptide, are markedly increased in the spinal cord, hypotha- lamus and cerebral cortex after a 14-day chronic treatment of amoxapine or amitriptyline. Met-enkephalin levels were increased after amitriptyline treat- ment in the spinal cord and hypothalamus. Chronic treatment with amoxap-
ine or amitriptyline caused an increase of δ- and μ-opioid binding sites in the spinal cord and decreased in the hypothalamus; opioid receptor levels re-
mained unchanged in the cerebral cortex (Hamon et al. 1987). However, due to the close relations between monoamine systems and opioids in the brainstem
and midbrain, these secondary effects on opioid function are not unexpected.
Indeed, the anti-nociceptive properties of clomipramine and amitriptyline, as well as their ability to potentiate morphine-induced analgesia, seem to be linked to the activation of a serotonin-mediated endogenous opioid system. Sacerdote et al. (1987) have shown that acute administration of clomipramine and amitriptyline (acting on NA and 5HT), induce analgesia. Nortriptyline, a TCA that acts predominantly via the noradrenergic system, does not induce analgesia when administered acutely. However, acute dosing of nortriptyline, amitriptyline and clomipramine all potentiate the anti-nociceptive effects of morphine, which further exempliﬁes a relationship between the serotoninergic and the endogenous opioid systems (Sacerdote et al. 1987). Here as in many studies, the relation of effects seen with acute doses to the clinical proﬁle of drugs used for chronic treatments is unclear.
Using a rat model of neuropathic pain, caffeine blocks the thermal anti- hyperalgesic effect of acute amitriptyline in a rat model of neuropathic pain. Concurrent systemic administration of both caffeine and amitriptyline blocked amitriptyline thermal anti-hyperalgesic effect. No observable effects inherent to caffeine were found at this dose. Spinally, the mild anti-hyperalgesic effect of amitriptyline was unchanged by pretreatment with intrathecal caffeine. Using brief anaesthesia, peripherally administered amitriptyline into the neuropathic paw resulted in anti-hyperalgesia. Furthermore, anti-hyperalgesia due to both doses of amitriptyline were partially antagonized with co-administration of caffeine. Ultimately, this study suggests that acute amitriptyline’s thermal anti- hyperalgesic effects in this model may involve increase of a peripheral endoge- nous adenosine tone (Esser and Sawynok 2000). In a randomized, double- blind, placebo-controlled study of 200 adult patients, topically administered
3.3% doxepin hydrochloride, 0.025% capsaicin and a combination of 3.3% dox- epin/0.025% capsaicin were found to provide similar levels of anti-nociception
in human chronic neuropathic pain. However, a more rapid onset of analgesia was produced by the combination of doxepin/capsaicin (McCleane 2000).
In line with peripheral actions under some conditions, Abdi et al. (1998)
examined the acute effects of amitriptyline and GBP on behavioural signs of mechanical allodynia using the Chung rat model of neuropathic pain. In a second experiment, continuous discharges in fascicles of injured afferent ﬁbres were recorded. Amitriptyline and GBP increased the mechanical allody- nia threshold and amitriptyline depressed the rate of continuing discharges of injured afferent ﬁbres, whereas GBP had no effect on these discharges. Neuro- pathic pain behaviour in rats is clearly modulated by both amitriptyline and GBP. The site of action of GBP is exclusively central; amitriptyline appears to act both peripherally and centrally (Abdi et al. 1998), but with regard to the former, little is known whether this potential action acts on evoked responses or contributes to the effects seen in patients after systemic doses.
It is known that amitriptyline, as well as other anti-depressants, has a high binding afﬁnity for NMDA receptors in vitro. Intrathecal amitriptyline com-
pletely antagonized thermal hyperalgesia induced by NMDA. Some authors
have suggested that TCAs intrathecally administered may provide greater pain relief than systemically administered TCAs, which are known to provide modest activity in the treatment of clinical neuropathic pain. Inﬂammation- induced thermal hyperalgesia was reversed by intrathecal amitriptyline. The anti-nociceptive effect of amitriptyline was unaffected, except at the lowest dose, by block of NA and 5HT receptors. Hyperalgesia is reversed by amitripty- line in rats by a mechanism not linked to monoamine reuptake inhibition, and possibly due to NMDA receptor antagonism (Eisenach and Gebhart 1995).
However, these animal studies, often using thermal tests and acute doses, need to be expanded to include studies on allodynias and ongoing pain after
nerve injury to better replicate the clinical situation. They also fail to ex- plain why, if the effects of the TCAs are peripheral and/or central and not mediated by monoamines, sedation and cardiovascular effects, hallmarks of
noradrenergic activity, are so common in patients. A signiﬁcantly higher rate of serious adverse cardiac events occurs during nortriptyline treatment than during treatment with paroxetine (Roose et al. 1998).
Venlafaxine, a serotonin and noradrenergic reuptake inhibitor (SnaRI), is effective in multiple pain disorders and has an improved side-effect proﬁle when compared to TCAs (Mattia et al. 2002).
Schreiber et al. (1999) have shown that the κ- and δ-opioid receptor subtypes
strongly inﬂuenced the anti-nociceptive effect of venlafaxine, with the α2- adrenergic receptor also contributing to the anti-nociceptive effect of this drug (Schreiber et al. 1999). α2- and a minor α1-adrenergic mechanisms of
anti-nociception are implicated by tests where adrenergic and serotoninergic antagonists were used; venlafaxine-induced anti-nociception was decreased by yohimbine but not phentolamine or metergoline. Furthermore, clonidine,
an α2-adrenergic agonist, signiﬁcantly potentiated venlafaxine-mediated anti- nociception (Schreiber et al. 1999).
As has been pointed out in this account, there are multiple mechanisms of neu- ropathic and other pains, and in the case of the many peripheral neuropathies, changes can be charted from the periphery, through the spinal cord and into the brain. Thus, the fact that both anti-depressants and anti-convulsants, with very different mechanisms, can be effective is not surprising since they target key but parallel systems involved in this pain state. A better understanding of the multiple mechanisms of neuropathic pain should lead to a more effective use of existing drugs, possibly allowing a greater number of patients to beneﬁt, and provide a basis for the development of potential new therapies.