Rheumatic Diseases Clinics of North America Volume 28 • Number 2 • May 2002


Srinivas G. Rao, MD, PhD

The neuropharmacology of centrally-acting analgesic medications in fibromyalgia


Srinivas G. Rao, MD, PhD

Cypress Bioscience
4350 Executive Drive
Suite 325
San Diego, CA 92131, USA

E-mail address:  srao@cypressbio.com

 

PII S0889-857X(01)00004-7

Chronic, widespread pain represents the sine qua non of the fibromyalgia syndrome (FMS), a fact reflected in the requirements of the American College of Rheumatology's 1990 diagnostic criteria for FMS [1] [2] [3] . Patients with FMS display abnormalities in pain perception in the form of both allodynia (pain with innocuous stimulation) and hyperalgesia (increased sensitivity to painful stimuli) [4] [5] . Such abnormalities, which are also found in other forms of chronic pain, imply that the “gain” of nociceptive processing in these patients is increased [1] .

Some early theories of FMS pathophysiology posited peripheral abnormalities (particularly alterations in skeletal muscle) as underlying the pathophysiology of FMS pain [6] . More recent studies, however, have generally failed to confirm the presence of such alterations [6] [7] [8] . The lack of peripheral abnormalities, coupled with the widespread nature of the pain, has shifted the focus away from the periphery and towards the central nervous system (CNS) [1] [9] [10] . In particular, it is currently thought that “central sensitization” may underlie the abnormal sensitivity to pain in FMS patients [1] [4] [9] . In this context, central sensitization has been operationally defined as a generalized heightened pain sensitivity due to pathological nociceptive processing within the central nervous system [1] [11] . An important caveat to bear in mind when considering theories of central sensitization, however, is that FMS patients report a number of other symptoms—sleep abnormalities, fatigue, perceived swelling of their extremities, and irritable bowel syndrome—in addition to pain [1] [3] . How the pathophysiology of such symptoms is related, causally, to central sensitization is still an area of active research [1] .

The focus of this article is on the neuropharmacology of the central nervous system pain pathways, emphasizing the spinal to midbrain sites-of-action of medications commonly used in FMS. The review begins with an overview of the ascending and descending pain pathways, with a particular emphasis on their respective neuropharmacology.

Central pain pathways

It has been recognized for some time that nociception is not a passive, one-directional process [12] . Rather, a complex interaction between ascending and descending pathways (Fig. 1) exists with the ability to dramatically alter the relationship between stimulus and response [13] . In addition, over the past decade, it has become apparent that chronic pain is quite different, clinically and pharmacologically, from acute pain [14] . As described later, a variety of alterations occurring both within the central nervous system and at the periphery may contribute to the chronic pain state.


Fig. 1.   An overview of both the ascending and descending pain pathways. The ascending and descending pathways are respectively shown by arrows. The periaqueductal gray (PAG) represents a central structure in this system, linking the cortex and other higher structures with the dorsal horn and processing both ascending and descending nociceptive information. Nociceptive input from the periphery is relayed via Aδ and C peripheral afferent fibers (PAF) to the dorsal horn of the spinal cord. After significant local processing, signals are relayed to higher centers via the spinothalamic tract (STT) to the nuclei gracilis and cuneiformis (NG/NC) and then on to the thalamus. The PAG has excitatory connections to the rostral ventromedial medulla (RVM) and dorsolateral pontine catecholamine cell groups (DLP). The former includes the nucleus magnus raphe and the reticular formation and is the primary source of spinal 5-HT, which, in turn, is primarily inhibitory at the level of the dorsal horn (dashed line). The DLP consists of the locus ceruleus, the subceruleus, and the Kölliker-Fuse nucleus, and this system represents the major source of descending inhibitory NE spinal innervation. The 5-HT and NE pathways descend in the spinal cord primarily within the dorsolateral funiculus (DLF).

 

Several excellent reviews summarize the current knowledge of pathways that underlie nociceptive processing [15] [16] . Rather than duplicate these efforts, a brief review of the pathways and their pharmacology specifically as it relates to drug therapy is presented here.

Ascending systems

Nociceptive information from peripheral nociceptors is relayed to the CNS via primary afferent fibers (PAF, either unmyelinated C fibers or myelinated Aδ fibers) that terminate within specific laminae of the dorsal horn (Fig. 2) . These fibers synapse onto several classes of dorsal horn interneurons and projection neurons, including nociceptive-specific and wide-dynamic range neurons [16] . The latter class of neurons receives input from both nociceptive and nonnociceptive afferents. A subset of dorsal horn neurons project supraspinally (via the spinothalamic tract and other pathways), and such connectivity forms the basis of pain perception [16] . Significant local interconnectivity is present, however, within the spinal cord, and such connections underlie aspects of a number of phenomena, including spinal motor reflexes, wind-up, and diffuse noxious inhibitory controls (DNIC). The spinal motor reflexes (i.e., tail-flick reflex) are mediated by connections between nociceptive dorsal horn neurons and anterior horn motor neurons. “Wind-up” refers to a very specific augmentation in the response of a dorsal horn neuron that results from tonic, peripheral nociceptive input [17] [18] [19] . Wind-up has been demonstrated to occur on both nociceptive-specific and wide-dynamic range neurons, and wind-up of the latter cells may be critically important to the development of allodynia [20] [21] . Finally, the DNIC serves, in some ways, the opposite role of wind-up: its function is manifest in the observation that nociceptive stimuli applied to one area of the body can actually suppress the activity of nociceptive neurons corresponding to other body areas [22] [23] [24] .


Fig. 2.   An expanded view of the dorsal horn. The critical component at this level is the dorsal horn neuron, which can be either of the nociceptive-specific or wide-dynamic range (WDR) variety. Such neurons project both locally within the spinal cord and to high centers via the spinothalamic tract (STT). The local connections may occur either within the dorsal horn (mediating DNIC function, for example) or in the anterior horn (mediating spinal motor reflexes). The main excitatory transmitter from the PAF to the dorsal horn neuron is glutamate (Glu) acting at postsynaptic AMPA, kainate, and NMDA receptors. In addition, neurokinins including substance P (SP), calcitonin gene-related peptide (CGRP), and neurokinin A (NKA) are co-released with Glu, and SP and NKA act via postsynaptic NK1 receptors. The PAF terminal itself receives modulatory input from a number of receptors, including α2 (inhibitory, black box), μ-opioid, and 5-HT3 (excitatory, white box). Activation of the inhibitory inputs causes a reduction in Glu and SP release, whereas 5-HT3–receptor activation causes enhanced release. The dorsal horn neuron is also subject to GABA, NE, 5-HT, and opiate inhibitory input, acting respectively at GABAA/B, α2, 5-HT 1A, and μ-opioid receptors. In particular, note that whereas 5-HT has an excitatory effect on the PAF terminals, it can inhibit the dorsal horn both directly (via 5-HT1A activation) or indirectly (through activation of a GABAergic interneuron).

 

The neuropharmacology of the dorsal horn neuron is complicated, as a staggering array of excitatory and inhibitory peptidergic and amino acid neurotransmitters are present [13] ; a subset of these neurotransmitters is presented in Fig. 2 . The primary excitatory neurotransmitter released by the PAF is glutamate. The latter acts on both ion-channel (i.e., AMPA [α-amino-3-hydroxy-5-methylisoxazole-4-propionate], kainate, and NMDA [N-methyl-Daspartate]) and G-protein–linked receptors located on the dorsal horn neurons. Because of blockade by a magnesium ion, the NMDA receptors become functional only at relatively high levels of neuronal activation [25] . This point is important, as a number of studies suggest NMDA activation represents a critical step in initiating wind-up and related phenomena in dorsal horn neurons [17] [18] [26] [27] [28] [29] . The downstream mediators of NMDA activation include nitric oxide (NO) and phosphokinase C (PKC), both of which are activated by increased intracellular levels of calcium that results from activation of the NMDA receptor [30] [31] [32] .

As shown in Fig. 2 , a number of neuropeptides—including substance P (SP), neurokinin A, calcitonin gene-related peptide (CGRP), and somatostatin—are colocalized with glutamate at the PAF-dorsal horn neuron synapse, and their effects on dorsal horn neurons appear to be cooperative [13] [14] [33] . The effects of SP, the best characterized of these neuropeptides [34] [35] , are primarily mediated via postsynaptic NK1 receptors [36] [37] [38] [39] . As described below, SP-mediated neurotransmission has been found to play a role in animal pain models, particularly those replicating chronic pain [40] . Note, however, that the role of SP-mediated neurotransmission in human nociceptive processing is still controversial [41] .

At the level of the dorsal horn, noxious stimulation results in the release of SP [33] , and the application of SP to dorsal horn neurons has been shown to augment NMDA-induced wind-up [42] [43] . Elevated levels of SP are found in the CSF of FMS patients [44] [45] , and such elevations are also found in some other chronic pain conditions, including chronic headache [46] , trigeminal neuralgia [47] and painful osteoarthritis [48] , although not in others, including painful diabetic polyneuropathy [49] . Interestingly, intrathecally administering SP to experimental animals also results in a state of hyperalgesia and allodynia [50] [51] [52] , and such pain sensitivity can be relieved by intrathecal administration of selective NK1 antagonists [53] or NMDA antagonists [54] . Further evidence of the cooperative effects of NMDA and SP in the generation of pain states can be found in a study looking at transgenic mice that overexpress nerve growth factor [55] . Such animals display spontaneous hyperalgesia and allodynia and show evidence of increased SP production. Once again, intrathecal treatment with either NMDA or NK1 antagonists was shown to normalize pain responses.

Inhibition of the PAF terminals and dorsal horn neurons is a result of glycine, γ-amino butyric acid (GABA), and opioid-mediated neurotransmission [13] [16] [56] . Glycine receptors and GABAA receptors are both ligand-gated anion channels, and their activation results in a rapid hyperpolarization of the postsynaptic neuron [57] [58] . The GABAB receptor, on the other hand, is G-protein linked, and the activation of these receptors results in a slower hyperpolarization mediated through increases in potassium conductance [58] . At the level of the spinal cord, the μ-opioid receptor is predominant, and it is mainly located presynaptically on the PAF terminal [59] . Activation of this receptor tends to hyperpolarize the PAF terminal, thus causing a reduction in glutamate and SP release [60] and a subsequent reduction in wind-up [61] .


Descending systems

The major components of the descending nociceptive system are shown in Fig. 1 . The periaquductal gray (PAG) represents a central structure in this system, linking the cortex and other higher structures with the dorsal horn and processing both ascending and descending nociceptive information [62] . Stimulation of the PAG results in the inhibition of dorsal horn neurons [63] [64] . These effects of the PAG on the dorsal horn are primarily mediated via connections from the PAG to components of the rostral ventromedial medulla (RVM) [65] and to cells within the dorsolateral pontine (DLP) catecholamine cell groups [66] . As described below, the monoamines serotonin (5-HT) and norepinephrine (NE, also referred to as noradrenaline) play a central role in descending pain modulation. The RVM (which includes nucleus magnus raphe and the reticular formation) represents the major source of spinal 5-HT, whereas the dorsolateral pons is the primary source of CNS NE.

Three functional categories of neurons are found in the RVM: on cells, off cells, and neutral cells [67] (Fig. 3) . In acute pain models, activation of on cells is generally pronociceptive, whereas off-cell activation is antinociceptive. Neutral cells, as their name implies, show no activity modulation with pain testing. Subsets of both the on-and off-cell populations have been shown to project directly to the dorsal horn [68] [69] . A component of the off-cell population is likely responsible for the serotonergic innervation of the dorsal horn [67] [70] . In contrast, a subset of on cells may be GABAergic interneurons that provide tonic inhibitory input to nearby off cells [67] [71] . This theory is supported by the observation that the application of GABAA agonists and antagonists at the level of the RVM is pro- and antinociceptive, respectively [72] . Presumably, these effects are the result of off cells being further inhibited by the GABAA agonist and, conversely, being disinhibited by the antagonist.


Fig. 3.   An expanded view of the RVM. Three functional categories of neurons are found in the RVM: on cells, off cells, and neutral cells. In acute pain models, activation of on cells is generally pronociceptive, whereas off-cell activation is antinociceptive. The activity of these two cell classes is generally reciprocal, and it is possible that at least some on cells are inhibitory GABAergic neurons that synapse onto off cells. Off cells also appear to receive excitatory glutamatergic input from the PAG, and may, in part, be responsible for descending 5-HT projections. On cells receive excitatory input from dorsal horn neurons and from CCK-B–mediated input. The effects of NE on on cells are complex, having both stimulatory and excitatory effects depending on whether the neurotransmission is being mediated by postsynaptic α1 or α2 receptors. Finally, μ-opioid–receptor agonists are profoundly inhibitory to on cells, presumably as a result of the presence of μ receptors on these neurons.

 

Data suggest that the disinhibition of off cells may be important for the analgesic effects of opiates. It has been shown that the application of opioid receptor agonists directly into the RVM causes antinociception via suppression of on-cell activity and a reciprocal increase in off-cell activity [73] [74] [75] . As activation of opioid receptors is typically inhibitory, a parsimonious explanation for this phenomenon is that off cells are being disinhibited from tonic on-cell inhibitory input [73] . Data also supports the hypothesis that opioid-induced analgesia is, in part, mediated via descending 5-HT pathways [76] . It has been shown, for example, that direct injection of opiates into the RVM causes an increased release of 5-HT in the spinal cord, and that the analgesia thus induced can be augmented by increasing local 5-HT concentrations by selectively blocking its reuptake [70] .

Disinhibition may also play an important role in the hyperpolarization seen in dorsal horn neurons as a result of high-level electrical stimulation of the RVM [77] . At the level of the dorsal horn, such hyperpolarization can be blocked by the local application of 5-HT3 antagonists [78] . Further, the selective destruction of 5-HT3 receptors reduces the analgesic effectiveness of intrathecal 5-HT [79] . As the 5-HT3 receptor is a ligand-gated cation channel, application of 5-HT has an excitatory effect on neurons expressing this receptor [80] . Thus, the hyperpolarizing effects of PAG-stimulation may be explained partly by the presence of the 5-HT3 receptors on GABAergic interneurons within the dorsal horn (Fig. 2) [81] . Activation of interneurons bearing these receptors by local 5-HT release (which, in turn, is induced by PAG stimulation) will cause hyperpolarization of dorsal horn neurons postsynaptic to these interneurons. In keeping with this theory, GABAA antagonists can also block PAG stimulation induced hyperpolarization, confirming the presence of an indirect, GABA-mediated effect [82] [83] . Note, however, that studies suggest that 5-HT3 receptors are also found on PAF terminals [84] , and the activation of such receptors has been shown to be pronociceptive due to depolarization of the terminal [85] [86] [87] . Finally, in addition to the indirect, GABA-mediated pathway described above, 5-HT likely has direct hyperpolarizing effects on dorsal horn neurons, perhaps as a result of 5-HT1A receptor activation [88] [89] [90] .

The dorsolateral pontine catecholamine cell groups (DLP, Fig. 2 ) from which the major NE descending fibers originate include the locus ceruleus, the subceruleus, and the Kölliker-Fuse nucleus [16] [91] [92] . Chemical or electrical stimulation of the dorsolateral pons results in an α2-mediated analgesia and direct inhibition of dorsal horn neurons [16] . Microionotophoretic application of NE at the level of the dorsal horn also results in the inhibition of local neurons [63] , and intrathecal administration of NE or of an α2 agonist results in inhibition of dorsal horn neurons and a pronounced behavioral analgesia [93] [94] [95] . Data suggest that such analgesia may be particularly relevant against mechanical allodynia [96] . In addition to its direct effect on dorsal horn neurons, spinal NE has also been shown to reduce SP release from PAFs [97] [98] and the dorsal horn in general [99] and may thus help prevent or reduce wind-up–like phenomena (Fig. 3) . Both α1- and α2-class receptors may play a role in mediating such effects [100] [101] . The DLP also sends projections to the RVM [102] [103] , and data suggest that NE may serve to modulate the activity of the RVM [104] [105] [106] . In particular, application of the α2 agonist clonidine into the RVM results in an inhibition of on-cell firing [67] . Conversely, direct injection of NE is actually pronociceptive and associated with a transient increase in on-cell activity [67] . Such effects are thought to be α1-mediated.

In summary, the 5-HT and NE systems originating in the RVM and dorsolateral pons, respectively, are thought to represent the primary mediators of descending nociceptive modulation. At the level of the spinal cord, the projections from the RVM appear to have both pro-and antinociceptive components, whereas the pontine projections appear to be mainly antinociceptive. These two systems are tightly interconnected [107] ; in fact, the analgesic effects of spinal 5-HT are partially dependent on NE [81] , although their respective antinociceptive effects appear to be additive under some circumstances [108] .

The role of the descending systems in chronic pain is an area of active research. One may hypothesize that either disabling one or more antinociceptive pathways or activating a pronociceptive one can lead to a behavioral state of central sensitization. Indeed, recent studies have demonstrated that reduced spinal NE outflow results in a chronic hyperalgesic state in laboratory animals (L Jasmin, personal communication, 2001). Likewise, a significant body of research has implicated the RVM in maintaining the hyperalgesia state. For example, inactivation of the RVM (by lesion, injection of lidocaine, or spinal transaction) can reverse the allodynia and hyperalgesia seen in animal models of chronic pain [31] . Recent data suggest that such effects are mediated specifically by the on cells of the RVM. Selective ablation of μ-opioid–receptor-positive cells in the RVM (presumed on cells [60] [67] ) reverses hyperalgesia caused by experimental nerve injury [109] . Further, the application of CCK-B receptor antagonists (which selectively block the activation of on cells) reverses the mechanical allodynia seen in animal models of chronic, neuropathic pain [110] .

Studies have demonstrated that NMDA receptors at the level of the RVM may also play a role in maintaining a hyperalgesic state. Direct injection of NMDA antagonists into the RVM is effective in reversing hyperalgesia caused by several chronic pain paradigms [111] [112] [113] . Presumably, the on cells are the recipients of the NMDA-mediated input, possibly directly from the dorsal horn, in these hyperalgesic states (Fig. 3) .

FMS: classes of therapeutic agents

As reviewed by Andre Barkhuizen in this issue, a wide variety of medications are used in clinical practice to treat the symptoms of FMS [114] [115] [116] [117] . The classes of agents that are used for their analgesic effects include the antidepressants, opiates, antiepileptic drugs, and antispasticity agents. Other agents, such as sedatives and/or hypnotics, have not been shown to be effective in treating the pain of FMS, although they may have a role in treating other associated symptoms (see later). Finally, whereas NSAIDs may be used in some clinical settings to treat FMS, their effectiveness in as analgesics in FMS has not been demonstrated [118] [119] . The pharmacology of the agents and their respective drug classes is summarized in Table 1 and detailed below.

 

Table 1.   Drug classes: mechanism of potential analgesic actions

Drug class

Specific agents

Analgesic mechanisms

Tricyclic antidepressants

Amitriptyline

NE-reuptake inhibition

Doxepin

5-HT–reuptake inhibition

Cyclobenzaprine

NMDA antagonist?

Cation channel blockade?

SSRI antidepressants

Fluoxetine

5-HT–reuptake inhibition

Sertraline

Citalopram

SNRI antidepressants

Venlafaxine

NE-reuptake inhibition

Milnacipran

5-HT–reuptake inhibition

Duloxetine

RIMA antidepressants

Moclobemide

Reversible inhibition of MAO-A

Pirlindole

NARI antidepressants

Reboxetine

NE-reuptake inhibition

Other antidepressants

Nefazadone

5-HT2 antagonist

NE-reuptake inhibition (weak)

5-HT–reuptake inhibition (weak)

Mirtazipine

2A (autoreceptor) antagonist

5-HT2 antagonist

5-HT3 antagonist

Buproprion

Dopamine reuptake inhibition

5-HT–reuptake inhibition

NE reuptake inhibiton

Opiates

Morphine

μ-opioid agonist

Tramadol

μ-opioid agonist

5-HT–reuptake inhibition

NE-reuptake inhibition

Antiepileptics

Gabapentin

Cation channel blockade

Lamotrigine

Enhanced GABA neurotransmission

Topiramate

Tiagabine

Carbamazepine

Antispasticity agents

Tizandine

2 agonist

Baclofen

GABAB agonist

Diazepam

Enhanced GABAA neurotransmission

Lorazepam

Other drugs

Ketamine

NMDA antagonists

Dextromethorphan

Tropisetron

5-HT3 antagonists

Ondansetron



Antidepressants

Antidepressants of all varieties represent a common form of therapy for many chronic pain conditions, including FMS [114] [115] [116] [117] [120] . All of the antidepressants described here increase 5-HT-and/or NE-mediated neurotransmission, either directly or indirectly, within the CNS. As discussed in the previous section, increasing the spinal concentrations of either 5-HT or NE has been shown to be antinociceptive in a number of animal models. In particular, increasing 5-HT–mediated neurotransmission has the effect of hyperpolarizing dorsal horn neurons, both by direct effects possibly mediated by 5-HT1A and by indirect, 5-HT3 effects (see Fig. 2 ). Increasing NE α1-and α2-mediated neurotransmission has also been shown to hyperpolarize both dorsal horn neurons and PAF terminals. This latter activity may counteract the potential pronociceptive effects of 5-HT on PAF terminals noted previously. Conversely, whereas increasing levels of NE in the RVM appears to be pronociceptive [67] , the effects of 5-HT–reuptake inhibition in the RVM may serve to counteract such effects [70] [76] . Such complementary actions may explain, in part, why increasing both 5-HT and NE levels simultaneously have additive effects on analgesia.

Antidepressants drugs can be classified on both historical and pharmacological grounds as tricyclic, selective serotonin reuptake inhibitors (SSRIs), and atypical agents (see Table 1 ) [121] . The specific pharmacology for each of these antidepressant classes are discussed in turn.

Classes of antidepressants

The tricyclic antidepressants (TCA) represent the oldest class of mood elevating agents. The use of these drugs in the treatment of FMS is well established, and specific agents in common use within the United States today for this indication include amitryptyline, doxepin, and cyclobenzaprine [114] [115] [116] [117] . Note that whereas the latter agent is commonly classified as a muscle relaxant rather than an antidepressant, it is tricyclic in structure and has effects on both the NE and 5-HT systems [122] [123] .

Members of the TCA drug class may reduce pain by increasing CNS concentrations of 5-HT and/or NE by blocking their respective reuptake; however, they also have prominent antagonist effects on histaminergic and cholinergic neurotransmission [124] . Other effects include NMDA antagonist action and ion-channel blocking activity (like antiepileptic drugs; see later) [125] [126] [127] [128] [129] [130] [131] [132] . Such effects may play a role in augmenting the analgesic efficacy of TCAs; however, these myriad effects also undoubtedly contribute to this class's relatively poor side effect profile and poor patient tolerance [123] .

The SSRIs have revolutionized the treatment of major depressive disorder and several other psychiatric conditions, including social phobia and anxiety [133] [134] . Much of their success is attributable to the fact that such drugs display a much improved side-effect profile compared to TCAs, which, in turn, is a result of their much higher degree of pharmacological specificity [124] . As implied by their name, SSRIs conceptually inhibit the reuptake of only 5-HT, although the actual selectivity of these agents for the monoamines is not absolute and varies by agent [124] . Citalopram is generally considered the most selective SSRI currently on the market. On the other hand, recent evidence suggests that paroxetine may also block the reuptake of NE at typical doses (CB Nemeroff, personal communication, 2001).

SSRIs that have been studied in FMS include fluoxetine, sertraline, and citalopram [135] [136] [137] [138] ; however, their relative efficacy, particularly compared to TCAs, is the subject of some debate [135] [139] [140] . Of note, the most selective SSRI—citalopram—also appears to be the least efficacious [137] [141] . In other chronic pain paradigms, SSRIs are generally considered to be inferior to TCAs [120] [142] [143] [144] . The simplest explanation for this phenomenon is that SSRIs only augment one of the two descending inhibitory systems.

The atypical class of antidepressants covers a great deal of pharmacologic variety, including 5-HT-NE dual reuptake inhibitors (SNRIs); reversible, enzyme-specific monamine-oxidase inhibitors (RIMAs); NE-specific reuptake inhibitors (NARIs); and other agents [121] . SNRIs are quite similar to some TCAs (e.g., amitriptyline) in increasing the levels of both NE and 5-HT by inhibiting their respective reuptake [124] . Unlike TCAs, however, SNRIs are generally devoid of significant activity at other receptor systems, thus greatly improving the side effect profile and general tolerability of TCAs. Currently, only one SNRI—venlafaxine—is on the market within the United States (although see nefazadone later), although several others are under development. Interestingly, data suggests that venlafaxine primarily affects the 5-HT system at lower doses; only at high doses are NE effects apparent [145] [146] [147] . In light of its pharmacology, it is perhaps no great surprise that venlafaxine has been shown to be efficacious in FMS and other pain paradigms [148] [149] .

The synaptic and extrasynaptic breakdown of the monoamines 5-HT and NE is a result of the activity of monoamine oxidase (MAO) enzymes. Two versions of the MAO enzyme are present in mammals—the A and B types [150] [151] . While significant functional overlap exists, the main substrates for MAO-A include NE, 5-HT, and dopamine, whereas those of MAO-B include dopamine, tyramine, and phenylethylamine [152] . Blocking either enzyme will increase the concentration of its respective substrates. Irreversible, enzyme-nonspecific monoamine oxidase inhibitors—including phenelzine and tranylcypromine—have been on the US market for over 20 years; however, concerns about potentially fatal interactions with other medications and with certain foods containing tyramine have limited their widespread usage [151] [152] . Newer agents that reversibly inhibit MAO-A—so-called RIMAs—have a much improved safety profile compared with older drugs [151] . Currently, no RIMAs are available in the United States, but at least two such agents are available in parts of Europe—moclobemide and pirlindole. Pharmacologically, these agents have effects that resemble those of SNRIs, and, thus, one would expect reasonable efficacy in chronic pain; however, early data with moclobemide has been unimpressive, with the agent demonstrating poor analgesic efficacy in cases of neuropathic pain [153] and poor efficacy compared to amitriptyline in FMS [154] . The data for pirlindole in FMS, however, shows more promise. In a recent 4-week, randomized, double-blind controlled trial, Ginsberg et al. found that pirlindole may be beneficial for certain symptoms of FMS, including pain [155] .

As stated above, NARIs specifically inhibit the reuptake of only norepinephrine [121] . While no NARIs are currently sold within the US marketplace, one such agent, reboxetine, is marketed as an antidepressant in other parts of the world [156] . The results of research into reboxetine's efficacy in chronic pain have yet to be published. Theoretically, one may expect analgesic efficacy in chronic pain for this class to be perhaps slightly superior to that of SSRIs and below that of SNRIs and TCAs. As is the case for SSRIs, only one descending antinociceptive system is being activated. Unlike the projections from the RVM, however, the pontine NE projections are thought to be entirely antinociceptive at the level of the dorsal horn. Note, however, that the effects on unopposed NE reuptake inhibition in RVM may actually be pronociceptive, as discussed previously (see Fig. 3 ) [67] .

Other atypical agents include nefazodone, mirtazipine, and buproprion. Nefazodone is a potent 5-HT2 antagonist, although it also weakly blocks the reuptake of both 5-HT and NE like an SNRI [157] [158] . The 5-HT2 antagonist actions appear important for increasing 5-HT1A–mediated neurotransmission in animal models [158] , and such activity may be useful in this agent's potential role as an analgesic [90] . There are no data on the efficacy of this agent in pain syndromes at the present time; however, trazadone, an agent related to nefazadone, has been relatively ineffective in the treatment of various pain syndromes, including FMS [14] . Mirtazipine blocks α2 autoreceptors (mainly α2A ) and 5-HT2 and 5-HT3 receptors [159] . This agent has shown some potential in some clinical pain conditions [160] , and such analgesic activity may be mediated by this agent's ability to increase NE levels (by α2A blockade) and increase 5-HT 1A neurotransmission. As discussed later, the blockade of 5-HT 3 receptors has both pro- and antinociceptive actions. Finally, buproprion is thought to be a nonspecific monoamine reuptake inhibitor, preferentially blocking the reuptake of dopamine, with lesser effects on 5-HT and NE [142] . A recent trial suggests that this agent may be effective in certain neuropathic pain states [161] .


Opiates

Three different opioid receptors have been isolated within the CNS—the μ, κ, and δ receptors—and all three appear to play a role in analgesia [60] . As discussed previously, opiates act on both the ascending and descending pain pathways. For example, it has been shown that μ agonists such as morphine both reduce transmitter release from the PAF terminals, and activate off cells within the RVM [59] [60] [61] [73] .

In general, concerns about side effects and addiction have limited the chronic use of opiates in FMS, particularly as the latter is not a life-threatening condition [114] ; however, one particularly interesting agent with modest opiate activity, in widespread use, and with demonstrated efficacy in FMS is tramadol [162] . This agent is unique in that it combines μ-opiate–receptor agonist activity with 5-HT and NE reuptake inhibition [163] [164] . This combination of activities allows tramadol to act at both ascending and descending sites [165] , including those mentioned previously for both μ agonists and for antidepressants. Further, new research suggests that the 5-HT1A receptor may also be involved in tramadol's analgesic effects [166] . Interestingly, tramadol may also be effective in psychiatric conditions including depression and obsessive compulsive disorder [167] [168] [169] , a fact consistent with its ability to block monoamine reuptake.


Antiepileptic drugs

A number of antiepileptic drugs (AEDs) have seen substantial use outside of their primary indication, including in chronic pain and as mood stabilizers [120] [170] . Specific examples of agents within this class include gabapentin, lamotrigine, topiramate, tiagabine, phenytoin, benzodiazepines (such as diazepam), valproic acid, and carbamazepine. Pharmacologically, many of these agents—including gabapentin, lamotrigine, topiramate, carbamazepine, and valproic acid—are cation channel (mainly sodium and calcium) blockers [170] . In addition, many of these agents also have enhancing effects on GABAergic neurotransmission; such agents include benzodiazepines, tiagabine, topiramate, and valproic acid [171] . While the details vary, AEDs as a class have the potential for relatively broad pharmacological effects across many components of the peripheral and central nervous systems, generally decreasing excitability, reducing ectopic discharge, and reducing neurotransmitter release [171] . In particular, the effects of AEDs on the ascending pathways may include reducing glutamate/SP release from PAF terminals, directly decreasing the activation of dorsal horn neurons, and increasing GABAergic input onto these neurons (Fig. 2) . The pharmacology of AEDs may be particularly suited for chronic pain due to nerve injury, as such injury appears to lead to the expression of particular cation channels that may play a role in ectopic discharge [172] .

In some neuropathic pain paradigms, such as trigeminal neuralgia, AEDs represent the first line of treatment [170] . However, only anecdotal data supports the use of most of these agents in FMS, although two exceptions do exist. Pregabalin—a molecule related to neurotonin—has recently been tested and found to be efficacious in a number of chronic pain conditions [173] [174] . Bryans and Wustrow provide an excellent review of both neurontin and pregabalin [175] . Note, however, that pregabalin is currently still in clinical development and is thus not, as yet, on the market.

The other class of antiepileptic agents that have been studied in FMS is that of the benzodizepines. As alluded to previously, benzodiazepines have been shown to enhance GABAergic inhibitory neurotransmission within the dorsal horn [176] . However, studies demonstrate that these agents appear to have only modest effects on FMS pain, although they do appear to exert more robust effects on sleep [177] [178] [179] [180] [181] . Like opiates, however, concerns about their side effects tend to discourage their long-term use in FMS and in other chronic pain syndromes [182] .


Antispasticity agents

Antispasticity agents are indicated for the treatment of skeletal muscle spasticity resulting from various CNS insults, including multiple sclerosis and stroke. Agents of this class with a demonstrated ability to reduce muscle tone include tizanidine, baclofen, and diazepam [183] . Benzodiazepines, including diazepam, are discussed above in the antiepileptic drug section. Tizanidine is an α2 agonist, similar in many ways to clonidine [184] . Compared to clonidine, however, tizanidine has less pronounced effects on blood pressure, possibly as a result of its lower affinity for the imidazoline 1 and 2 receptors [185] . While studies specifically targeting FMS have yet to be performed, both tizanidine and clonidine have demonstrated analgesic efficacy in a variety of clinical and animal pain paradigms, although the pronounced sedation caused by these agents can be problematic in some patients [186] [187] [188] [189] [190] . The analgesic efficacy of these agents is not surprising, as α 2 agonists can affect both the ascending and descending pain pathways at a number of points. As described in the previous section, α2 agonists reduce activation of PAF terminals, thus reducing glutamate/SP release. They may also directly inhibit dorsal horn projection neurons. Finally, increased α2 agonist activity within the RVM may increase off-cell activation, thus further decreasing pain by activation of the 5-HT descending system.

Baclofen is a GABAB agonist that is structurally related to GABA [58] . While not tested specifically in FMS, baclofen is widely used in a number of chronic pain conditions [191] , and it has been shown to be efficacious trigeminal neuralgia [192] [193] . The analgesic effects may be due to suppression of dorsal horn neuron activity [191] .


Other agents

Current thoughts on the use in FMS of three other classes of agents—NMDA antagonists, NK1 antagonists, and 5-HT3 antagonists—are detailed elsewhere in this journal. However, a few words about the respective site- and mechanism-of-action of these agents are in order. As discussed above in the pain pathways section, NMDA-mediated neurotransmission may play an important role in mediating wind-up and related phenomena in at least two sites in the pain pathways: at the PAF-dorsal horn neuron synapse and at the glutamatergic synapses onto on cells within the RVM. In addition, NMDA antagonist may help normalize SP-mediated neurotransmission, a feature that may be particularly relevant to FMS (see below). In fact, three recent studies have demonstrated that NMDA antagonists improve pain symptoms in FMS patients [194] [195] [196] . A poor side-effect profile, however, represents a significant problem for this class of agents [197] .

One group has extensively studied the use of tropisetron, a 5-HT3 antagonist, in the treatment of fibromyalgia [80] [198] [199] [200] [201] . Overall, this agent was found to be modestly effective only within certain range of doses, with a loss of efficacy at both lower and high levels [198] . A possible explanation of this phenomenon lies in the fact that the blockade of 5-HT3 receptors has both pro-and antinociceptive effects due to the presence of these receptors on both PAF terminals and inhibitory dorsal horn interneurons (see Fig. 2 ). Thus, the balance of pro- and antinociceptive effects may be highly dose-dependent, a fact that may lead to unpredictable results in clinical practice.

The rationale for the use of NK1 antagonists in FMS is linked, in part, to the observation that SP levels within the CSF of FMS patients are routinely elevated [44] [45] . As discussed previously, SP-mediated neurotransmission from the PAF to the dorsal horn neuron has been shown to be important in the generation of wind-up, although its role in the maintenance of such phenomena is unclear [17] [18] [28] . NK1 antagonists have demonstrated analgesic efficacy in a number of preclinical pain paradigms, particularly those modeling chronic pain. To date, there have been no published reports of the use of NK1 antagonists in FMS; however, the track record of this class of agents in human acute-and chronic-pain studies has been extremely poor [41] .

Summary

As demonstrated above, the anatomy and neuropharmacology of the pain pathways within the CNS, even to the level of the midbrain, are extraordinarily complex. Indeed, discussions of the effects of these agents on the neuropharmacology of the thalamus, hypothalamus, and cortex were excluded from this review owing to their adding further to this complexity. Also, the dearth of data regarding FMS pain pathophysiology necessitated a relatively generic analysis of the pain pathways. As mentioned in the introduction, the current thought is that central sensitization plays an important role in FMS. However, we see in this chapter that the behavioral state of central sensitization may be a result of alterations in either the ascending systems or in one or more descending systems. Studies to assess the presence or relative importance of such changes in FMS are difficult to perform in humans, and to date there are no animal models of FMS.

Accepting these limitations, it is apparent that many drugs considered to date for the treatment of FMS do target a number of appropriate sites within both the ascending and descending pain pathways. The data regarding clinical efficacy on some good candidate agents, however, is extremely preliminary. For example, it is evident from the present analysis that SNRIs, α2 agonists, and NK1 antagonists may be particularly well suited to FMS, although current data supporting their use is either anecdotal or from open-label trials [114] [149] . Other sites within the pain pathways have not yet been targeted. Examples of these include the use of CCKB antagonists to block on-cell activation or of nitric oxide synthetase antagonists to block the downstream mediators of NMDA activation. Efficacy of such agents may give considerable insight into the pathophysiology of FMS.

Finally, as indicated previously, FMS consists of more than just chronic pain, and the question of how sleep abnormalities, depression, fatigues, and so forth tie into disordered pain processing is being researched actively. Future research focusing on how the various manifestations of FMS relate to one another undoubtedly will lead to a more rational targeting of drugs in this complex disorder.

 

 

Use of this content is subject to the Terms and Conditions of the MD Consult web site.

Rheumatic Diseases Clinics of North America
Volume 28 • Number 2 • May 2002
Copyright © 2002 W. B. Saunders Company