G proteins couple membrane receptors (GPCRs) to intracellular effectors and exist in various forms. Several different intracellular signaling pathways are affected by G proteins, some stimulated, some inhibited, and often converged upon by actions of plural receptors. For example, G proteins released from some GPCRs activate membrane-bound phospholipase C (PLC) that subsequently hydrolyzes the sugar moiety from the lipid phosphatidyl inositol, liberating inositol trisphosphate (IP3) and diacylglycerol (DAG). In turn, IP3 acts on “release channels” of the storage compartments inside cells to subsequently
release Ca2+ into the intracellular domains. Ionized calcium, in turn, enhances or inhibits a number of cytoplasmic enzymes to acutely regulate cellular activ-
ity, and can regulate the long-term expression of cellular enzymes by binding to proteins that enter the nucleus and change cellular transcription. The other
product of PLC activity, DAG, has a stimulating effect on protein kinase C (PKC) enzymes, many of which are also stimulated by the elevated cytoplasmic Ca2+. Another major direct target of G protein activation is the enzyme that forms cyclic AMP (cAMP), adenylate cyclase (AC). Different G protein α-subunits will activate (Gs ) or inhibit (Gi ) adenylate cyclase, respectively increasing or decreasing the intracellular cAMP concentration (Hollmann et al. 2001).
In this complex scheme of multiple steps and interacting pathways local anesthetics may act at several molecular locations. There are some contradic- tory results of direct local anesthetic actions on speciﬁc GPCRs. Some caution should be applied, however, since many of these studies have been conducted on heterologous expression systems, such as Xenopus oocytes, where it is as- sumed that the enzymes that couple the GPCR to the measured endpoint are the same as those in mammalian cells expressing that particular GPCR. Local anesthetics inhibit certain GPCRs (e.g., LPA, TXA2, PAF, and m1 muscarinic receptors) expressed in Xenopus oocytes (Hollmann and Durieux 2000; Holl- mann et al. 2000a, 2004; Honemann et al. 1999) but are ineffective on others, e.g., angiotensin receptor signaling (Nietgen et al. 1997). Although certain PLC- coupled GPCR activities are inhibited by local anesthetics, the actual release of
Ca2+, triggered by IP3 that is liberated by phospholipase C (see above), is not affected by local anesthetics (Sullivan et al. 1999), indicating that the site(s) of
LA action is located upstream of this segment of the pathway, at the receptor itself, or on the G protein, or at the interaction site in the complex of the two (Xiong et al. 1999).
For almost all the local anesthetic-sensitive GPCRs, the site of action can be located at the G protein:receptor interface. [In a minority of cases there appears to be an action at the extracellular surface of some receptors (Holl- mann et al. 2000a), but these are very low-afﬁnity interactions and we will not
discuss them further here.] Durieux and colleagues have proposed that GPCRs linked to Gq α-subunits are the ones distinctly sensitive to local anesthetics,
based on the known G protein coupling of the susceptible pathways (Hollmann et al. 2001b, c, 2002). This hypothesis was conﬁrmed by experiments where Gq was “knocked out” of a system with the accompanying loss of local anes- thetic sensitivity of those coupled responses, while the pathways not coupled to Gq were unaffected (Hollmann et al. 2001b). In complementary experiments, introduction of Gq to cells, thus permitting this protein to couple to extant GPCRs, induced a sensitivity to local anesthetics that was not present before that manipulation (Hollmann et al. 2002). It is worth noting that the G protein coupling a particular GPCR to a speciﬁc pathway may differ among cells, and there may be multiple G proteins that can interact with one receptor, allowing for a diversity of sensitivities to local anesthetics. For example, substance P’s
neurokinin (NK)-1 receptors may be coupled to Gq or to G11 α-subunits (Mac- donald et al. 1996). Cells with the ﬁrst G protein will have substance P responses
attenuated by local anesthetics (Li et al. 1995), cells with the second will be rela- tively insensitive. Interestingly, in one case Gαi protein function was enhanced by local anesthetics, apparently not by interaction with the coupled adenosine receptor but by a direct effect on the αi G protein subunit (Benkwitz et al.
2003). The extent to which such activation occurs in general, and contributes
to the overall inhibitory effect of local anesthetics, remains unexplored.
Acute pain is often accompanied by inﬂammation, and certain inﬂammatory processes are profoundly sensitive to local anesthetics via actions on GPCRs (Hollmann and Durieux 2000). In particular, the priming of neutrophils that is often critical for their rapid and vigorous response during inﬂammation
(Condliffe et al. 1998) is suppressed by bupivacaine concentrations of ca. 10−8 M (Hollmann et al. 2001d; Fisher et al. 2001), far lower than the IC50s for blocking ion channels (see Sects. 2.2 and 2.3 of this chapter, above). A second noteworthy
behavior of this phenomenon is the very slow time course, taking hours to develop (Hollmann et al. 2004) compared to the several minutes required for inhibiting other GPCRs or the seconds for blocking ion channels (Oyama et al.
1998). One explanation for this very slow time-course is that the suppressed
receptors are actually being removed from the plasma membrane—perhaps by some slow endocytotic process that is triggered or enhanced by the local anesthetic per se, or by some metabolic consequence, such as elevation of intracellular Ca2+, that results from other actions of local anesthetics (see
below, Sect. 3.3).
Inhibition of Protein Kinase A and Protein Kinase C
Many activated GPCRs cause acute cellular changes mediated by protein phos- phorylation via speciﬁc protein kinases, a set of phospholipid-dependent en- zymes, some of which are also dependent on calcium. Prominent among these are protein kinase A (PKA) and PKC, which are activated by separate path- ways and are directed toward different protein substrates, although they may
phosphorylate the same protein in interdependent ways (Cantrell et al. 2002; Cantrell and Catterall 2001). Other kinases, such as calmodulin-dependent ki- nase II, have also been identiﬁed as modulators of ion channels, but we will focus on PKA and PKC in this review as there are some data regarding their sensitivity to local anesthetics.
The PKC family is divided into three subgroups based on sequence homol- ogy and cofactor requirements: classic-conventional PKC isozymes (PKC-α,
-βΙ, -βΙΙ, and -γ), which are Ca2+-dependent and diacylglycerol (DAG)-stimulat- ed kinases; novel isozymes (PKC-δ, -ε, -θ, and -η), which are Ca2+-independent and DAG-stimulated kinases; and atypical PKC isozymes (PKC-ξ and λ), which
are Ca2+- and DAG-independent kinases (Newton 2003). Before stimulation, PKC is located almost exclusively in the cytosol, whereas its hydrophobic acti- vators are present in the membrane. On binding a soluble G protein α-subunit,
cytoplasmic PKC then translates to the plasma membrane where it associates with a lipid cofactor, i.e., phosphatidyl serine, and may subsequently modulate neuronal signal transduction by phosphorylation of several types of membrane
proteins, including voltage-dependent channels.
There are many different isoenzymes of neuronal PKC, expressed in the brain and in the distal peripheral tissues and the dorsal horn of the spinal cord, that may be involved in pain transmission and in the modulation of
nociceptor stimulation. In primary nociceptive neurons, the isozymes PKC-γ and -ε appear to be most important (Aley et al. 2000). They are implicated in the phosphorylation of a class of slow gating Na+ channels (channels often re-
sistant to inhibition by the classic blocker tetrodotoxin, thus termed “TTX-R,” and comprising the channels Nav1.8 and Nav1.9; England et al. 1996; Cardenas et al. 1997; Gold et al. 1996, 1998; Wood et al. 2002; Rush and Waxman 2004; Baker et al. 2003; Baker 2005). Such TTX-R channels are selectively expressed in primary nociceptors (Elliott and Elliott 1993; Rush et al. 1998; Cummins et al. 1999; DibHajj et al. 1999) and contribute critically to increased nocicep- tive ﬁring after inﬂammation or injury (Akopian et al. 1999; Porreca et al. 1999;
Abdulla and Smith 2002; Roza et al. 2003; Black et al. 2004). Excitability changes after injury may have both rapid and slower components, however, resulting form different underlying processes, such as phosphorylation of existing chan- nels, for the fast responses, and transcriptional regulation or redistribution of existing channels for the slower ones (Devor et al. 1993; Novakovic et al. 1998; Coward et al. 2001).
Several PKC-dependent pathways are suppressed by local anesthetics. One in vitro study assessed the effect of local anesthetics on a PKC-linked cascade of reactions and on the enzymatic activity of PKC itself, indicating that tetracaine and mepivacaine inhibited phosphatidylinositol hydrolysis, i.e., PLC action, an essential step in the activation of PKC (Irvine et al. 1978). Moreover, local anesthetics also inhibited puriﬁed PKC activity, possibly by competing with DAG or membrane phospholipid, and the lipid solubility of bupivacaine and mepivacaine correlated with their potency to inhibit PKC subtypes in vitro
(Uratsuji et al. 1985; Mikawa et al. 1990). Although some investigators have suggested that PKC-dependent pathways might have a role in the biochemi- cal mechanism producing spinal anesthesia, there is no correlation between changes in PKC levels and either potency or lipid solubility of the anesthet- ics (Nivarthi et al. 1996). In immunohistochemical experiments, activation of the MAP kinase extracellular receptor activated kinase (ERK, see Sect. 3.1) in dorsal horn neurons of spinal cord by the PKC activator phorbol myristal acetate was insensitive to bupivacaine, although bupivacaine did inhibit the activation of ERK stimulated by several inotropic receptors (Yanagidate and Strichartz 2006). Such ﬁndings indicate that bupivacaine’s action occurred nei- ther at PKC itself nor at sites downstream toward ERK. The discrepancy in the results on puriﬁed PKC and the enzyme’s effects in situ could be explained by a differential sensitivity of different PKC isozymes to local anesthetics or to drug sensitivities that depend on the cofactors for enzyme activation. Thus, particular PKC isozymes may be direct biochemical targets for local anesthetic action, but other sites of local anesthetics may be located along the upstream pathway that activates PKC.
cAMP-dependent protein kinase (PKA) is a major cellular participant in many neuronal functions, including modulation of ion channels (Gold et al.
1998a; Lopshire and Nicol 1998; Evans et al. 1999; Cantrell and Catterall 2001;
Vijayaragavan et al. 2004; Yang and Gereau 2004; Matsumoto et al. 2005), and also in the maintenance of inﬂammatory pain (Aley et al. 1999). These actions of PKA usually increase neuronal excitability, sometimes through the modulation of subthreshold oscillations of membrane potential (Xing et al.
2003; see Amir et al. 1997), in keeping with the general activating role of cAMP
in cellular functions. As a speciﬁc example in sensory systems, PKA contributes to the activation of ERK in dorsal horn neurons, where its effects are additive with those of PKC, indicating independent pathways (Kawasaki et al. 2004). Since ERK activation in spinal dorsal horn neurons is induced by nociceptive activity (Ji et al. 1999), the activation of spinal PKA is associated with pain signaling; other observations also implicate PKA in pain-related signaling in the peripheral and central nervous systems.
The cAMP–PKA pathway is a typical step in a neuronal second messenger pathway; binding of transmitter to receptor leads to the activation of a stimula-
tory G protein, Gs , which activates the enzyme adenylyl cyclase. The cyclase in turn catalyzes the conversion of ATP to cAMP. Four cAMP molecules bind to the two regulatory subunits of the cAMP-dependent protein kinase, liberating the two catalytic subunits, which are then free to phosphorylate speciﬁc substrate proteins that regulate several cellular response, often interacting in a physio- logically synergistic manner (Schwartz and Kandel 2000). As with PKC-related
activities, the steps leading to kinase activation as well as the enzyme’s activity per se may be affected by local anesthetics.
There are, however, few studies of these effects. Local anesthetics have been reported to exert multiple actions on the catecholamine-sensitive adenylate cy-
clase system of frogs, thereby reducing its overall responsiveness to stimulation (Voeikov and Lefkowitz 1980). In contrast, bupivacaine had no effect on ERK activation induced by 8-Br-cAMP, a direct activator of PKA, in dorsal horn neu- rons in slices of spinal cord (Yanagidate and Strichartz 2006). These ﬁndings should be cautiously weighted, however, because the cAMP–PKA pathway is a complex nociceptive pathway, and the effects of LAs on different components of the overall pathway have not been established in general.