Sensory-selective local anesthesia, which inhibits nociceptors without affecting motor or autonomic function, has important clinical significance for preventing adverse effects. Although frequently sought, this discriminative block is only occasionally achieved [1] . Transient receptor potential vanilloid-1 (TRPV1) channels are highly and selectively expressed, both peripherally and centrally, by many neurons in the nociceptive pathways. Peripherally, TRPV1 channels are located on the cell bodies and axons of nociceptor Aδ and C fibers [2] . Centrally, in the spinal cord dorsal horn, they are present at presynaptic and postsynaptic sites of the junctions between primary and secondary neurons and interneurons, as well as on astrocytes and microglia [3] . In the brain, they are distributed in the amygdala, cerebellum, cerebral cortex, hypothalamus, hippocampus, locus coeruleus, mesencephalon, olfactory bulb, periaqueductal gray, striatum, and thalamus. Because TRPV1 channels are selectively expressed by nociceptive neurons, researchers have been interested in the combination of local anesthetics (LA) and capsaicin, a TRPV1 agonist [4] -[7] . The pore of an activated TRPV1 channel is large enough to admit molecules like the membrane-impermeant local anesthetic derivative QX-314, which only blocks sodium channels, and thereby suppresses impulse firing, from the inside of cells [8] [9] . A long-lasting differential blockade favoring nociceptive block was produced by a 10-minute pre-injection of QX-314 prior to injection of the TRPV1-specific agonist capsaicin near the sciatic nerve [4] . Capsaicin combined with membrane-permeant local anesthetics like bupivacaine, lidocaine, amitryptiline, and N-methyl-amitriptyline also preferentially prolongs sensory/nociceptive block in the rat sciatic nerve [5] . However, central intrathecal application of the QX-314/capsaicin combination administrated by a continuous infusion pump did not selectively inhibit nociceptors in naïve and neuropathic rats [6] . It is not known whether this failed inhibition was due to the interactions of co-administered drugs or changed environmental factors, e.g. vascular effects.

When combining capsaicin with a local anesthetic, the drugs’ pharmacokinetics in the injection environment should be considered. The sciatic nerve is surrounded by a periuneurial sheath that resists LA passage, and is often located in the midst of much more massive skeletal muscles, while the spinal cord is immersed in cerebrospinal fluid. During peripheral nerve blockade, LA must penetrate through the sheath to reach the nerve fibers, and then enter the nerve plasmalemma before direct binding to sodium channels leads to failure in the generation or propagation of action potentials. The mechanisms underlying the neuraxial block, however, are more complex, although the barriers to LA penetration from the intrathecal space may be fewer [8] . Additionally, neuraxial actions may involve pre- and postsynaptic terminals in the release and binding of neurotransmitters and neuropeptides to their respective receptors, followed by the downstream signaling processes [9] .

We performed the studies reported here based on the speculation that the addition of capsaicin to local anesthetics for intrathecal injection would potentiate the local anesthesia. To our surprise, not only was there potentiation of LAs’ anti-nociceptive actions but there was also a reduction of motor blocking activity.

After intrathecal injection, all local anesthetic plus capsaicin groups (LA + CAPS) demonstrated shorter and less intense motor blocks compared to LA alone. The differences in nociceptive block durations were smaller and usually insignificant (Figure 1, Table 1).

The results show that for spinal administration the addition of capsaicin to each local anesthetic significantly lessened the duration of motor block in comparison to that of local anesthetic alone. In contrast, nociceptive block was not shortened by capsaicin, and was actually lengthened, significantly for articaine, but insignificantly, albeit to a greater degree, for lidocaine (Table 1). The overall result is a nociceptive block that outlasts the motor block, by a ratio of 3 - 5, a desirable clinical condition allowing suppression of acute post-operative pain with much shorter motor deficits.

Rearing and “defensive posture” after injection of capsaicin alone is probably due discomfort or pain from that drug on motor or nociceptive pathways in the spinal roots or spinal cord. None of these behaviors was observed when LAs were co-injected with capsaicin, suggesting that the cellular signals induced by capsaicin were either not present or could not be conducted beyond the spinal cord, e.g., to the brain. The full reversibility of all behavioral changes, from capsaicin alone or in combination with LAs, implies that no pathology accompanied these treatments.

The reasons underlying these differential changes in block duration are not known. In situations where quaternary LAs are combined with capsaicin for Na channel blockade, the activated TRPV1 channels open large enough pores to allow local anesthetic molecules to pass through [4] . The cationic quaternary LA derivatives must be present in the cytoplasm to reach their blocking site on the Na channel [12] [13] , but their passage through the bilayer region of biological membranes is very slow [14] and diffusion is greatly facilitated through the hydrophilic pathway afforded by the open TRPV1 channels. Since TRPV1 is expressed only in nociceptive nerves, those Aδ and C “pain” fibers are selectively blocked by the quaternary compounds in combination with capsaicin.

When traditional LAs are used the possibilities for interactions are more complex, for three reasons. First, being tertiary amines these drugs exist in rapid equilibrium between a charged species and its neutral base form, with the charged form being proportionally higher, since the pK_as are 0.5 - 1.5 pH units above physiological Ph [15] . The neutral LA species diffuse relatively easily through a bilayer membrane [14] [16] and all fibers will be blocked by a solution of LA. If anything, the observed differential block susceptibility favors small myelinated Aδ- and Aγ-fibers, with the smallest axons, non-myelinated C-fibers, being the least susceptible to block [17] [18] . Since the Aγ-fibers are critical for normal motor neuron tone (through regulation of discharge from muscle spindle afferents), their blockade will result in flaccid paralysis of the innervated muscle Therefore, a differential blockade of Aγ-fibers over C-fibers would be expected to lead to a selective motor blockade, as has been observed for lidocaine’s block of the rat sciatic nerve [10] . For this situation, the addition of capsaicin and activation of TRPV1 channels will allow local anesthetics an additional route of influx into nociceptive fibers and thus, to the degree that entry is facilitated by this route, favor a selective increase of nociceptive blockade.

Secondly, tertiary amine LAs themselves are weak agonists of the TRPV1 receptor [19] and so will tend to favor their own entry into TRPV1-expressing nerves via the activated channel. Based on this action of LAs, a combination of a tertiary amine LA and a quaternary derivative has been shown to produce a strongly nociception-selective functional block in vivo [20] .

Thirdly, activation of the TRPV1 channel generates an inward current that depolarizes the membrane. Membrane depolarization per se increases the potency for LAs in inhibiting Na channels [9] , through an increase in the proportion of channels in the inactivated state, with its higher affinity for tertiary amine LAs [14] [16] . All three mechanisms provide an explanation for an increased duration of nociceptive blockade, but cannot account for the reduction in motor blockade reported here.

Chemicals injected into the intrathecal space will penetrate into the spinal cord and also diffuse along the roots and interact with cells in the dorsal root ganglion [21] . So some of the actions of this drug combination could result from conduction block of impulses in the dorsal and ventral roots. Although the presence of hydrophobic and amphipathic molecules in the vehicle used for dissolving capsaicin, Tween 80 and ethanol, might alter the anatomic distribution of the LA, allowing deeper penetration and facilitating a longer-lasting nerve block, as has been proposed for a prolonged sensory-selective block of peripheral nerve, this action cannot account for the shortening of motor block observed here. Furthermore, the vehicle was present in both the LA alone and LA + CAPS formulations, so its presence cannot account for the differential blockade.

One possible explanation might be found in the sensory-motor interactions, most processed without consciousness, in the central nervous system. Recent findings show that nociceptive input modulates motor pathways at both spinal and cortical levels [22] . A different report revealed that activated TRPV1 receptors excited the spinal ventral root after-discharge after nerve injury [23] . It is not apparent from this neurophysiological study, however, if the TRPV1 receptors were located on the motoneurons or if the after-discharge was facilitated by interneurons that are targets of primary affererent fibers expressing TRPV1 receptors. At this time we cannot explain the shortening of motor block by the mechanisms thought to account for the differential block after peripheral nerve application.

Synaptic transmission in the spinal cord involves, at the minimum, stimulated release of neurotransmitters and neuromodulators, e.g., Substance-P, from the pre-synaptic endings and binding and activation of respective receptors on the post-synaptic cells. Complicating any simple unimodal “pass through” circuit is the complex role of interneurons which, depending on the condition, may participate in facilitation or inhibition of nociceptive processing [24] . On top of all this, LA molecules themselves have actions at multiple targets involved in synaptic transmission, from an inhibition of voltage-gated Ca⁺² channels [25] [26] , a direct blockade of post-synaptic receptors, including inotropic receptors for acetylcholine [27] and glutamate [28] , and metabotropic GPCRs for, among others [29] , Substance P [30] . Capsaicin acting at TRPV1 receptors will increase intracellular Ca⁺², which, depending on the level of free Ca⁺² reached, could have opposing actions. Low “resting” levels of Ca⁺² inside the pre-synaptic ending, as might result from TRPV1 opening, can enhance or reduce the amount of transmitter released by the stimulation of an invading impulse [31] , and an elevated Ca⁺² level in the post-synaptic cell can lead to desensitization of certain receptors [32] , but also activate certain Ca⁺²-dependent enzymes, e.g. many PKCs, that enhance cellular excitability [33] . In addition, LAs themselves have actions on some PKCs [34] . Given the many possibilities for the actions of either drug, any explanation of their interactions to effect a differential block must be firmly rooted in speculation.

In summary, combined intrathecal injection of capsaicin and membrane-permeant LAs produces a differential block by shortening motor deficits and, in some cases, prolonging nociceptive deficits. These effects probably involve both peripheral and central mechanisms, although the spinal circuitry is too complex and the drugs’ actions too diverse for any simple explanation. Nociceptive block with minimum motor block has promise for clinical use in peri-operative spinal anesthesia and, possibly, pain management.