F. Yanagidate• G. R. Strichartz (✉)
Pain Research Center, BWH/MRB611, 75 Francis Street, Boston MA, 02115-6110, USA firstname.lastname@example.org
Abstract Local anesthetics are used broadly to prevent or reverse acute pain and treat symptoms of chronic pain. This chapter, on the analgesic aspects of local anesthetics, reviews their broad actions that affect many different molecular targets and disrupt their functions in pain processing. Application of local anesthetics to peripheral nerve primarily results in the blockade of propagating action potentials, through their inhibition of voltage- gated sodium channels. Such inhibition results from drug binding at a site in the channel’s inner pore, accessible from the cytoplasmic opening. Binding of drug molecules to these channels depends on their conformation, with the drugs generally having a higher afﬁnity for the open and inactivated channel states that are induced by membrane depolarization. As a result, the effective potency of these drugs for blocking impulses increases during high-frequency repetitive ﬁring and also under slow depolarization, such as occurs at a region of nerve injury, which is often the locus for generation of abnormal, pain-related ectopic impulses. At distal and central terminals the inhibition of voltage-gated calcium channels by local anesthetics will suppress neurogenic inﬂammation and the release of neurotransmitters. Actions on receptors that contribute to nociceptive transduction, such as TRPV1 and the bradykinin B2 receptor, provide an independent mode of analgesia. In the spinal cord, where local anesthetics are present during epidural or intrathecal anesthesia, inhibition of inotropic receptors, such as those for glutamate, by local anesthetics further
interferes with neuronal transmission. Activation of spinal cord mitogen-activated protein (MAP) kinases, which are essential for the hyperalgesia following injury or incision and occur in both neurons and glia, is inhibited by spinal local anesthetics. Many G protein- coupled receptors are susceptible to local anesthetics, with particular sensitivity of those
coupled via the Gq α-subunit. Local anesthetics are also infused intravenously to yield
plasma concentrations far below those that block normal action potentials, yet that are
frequently effective at reversing neuropathic pain. Thus, local anesthetics modify a variety of neuronal membrane channels and receptors, leading to what is probably a synergistic mixture of analgesic mechanisms to achieve effective clinical analgesia.
Keywords Ion channels · Action potentials · Nociception · G protein-coupled receptors · MAP kinases
Local anesthetics are widely used for the prevention and relief of both acute and chronic pain (Strichartz and Berde 2005). The reduction or abolition of acute pain from accidental or intentional trauma (surgery) is accomplished by delivery of local anesthetics to the skin by topical application or subcutaneous inﬁltration, to peripheral nerve by percutaneous injection, or to the neuraxis by administration into the epidural or intrathecal spaces (Gokin and Strichartz
1999). Chronic pain symptoms are also relieved, albeit mostly temporarily, by the “nerve blocking” procedures just described. In contrast, systemic local
anesthetics, administered intravenously and to much lower plasma concen- trations than those used for direct nerve block, also relieve many forms of neuropathic pain in humans and in animal models, and with a therapeutic
beneﬁt that often endures for weeks, months, or longer, far outlasting the presence of the drug in vivo (Boas et al. 1982; Mao and Chen 2000).
Whereas mechanisms for the nerve blocking actions of local anesthetics are relatively well understood, those that underlie the neuraxial block are proba- bly more complex. The inhibition of neuronal voltage-gated sodium channels (VGSC) by direct binding of local anesthetics (Nau and Wang 2004) leads to failure in the generation or propagation of action potentials, the primary mechanisms for functional deﬁcits during peripheral nerve blockade. After the delivery of local anesthetics to the epidural or intrathecal compartments, these drugs diffuse into the spinal cord where they can interact with a variety of other ion channels involved (1) in excitation/depolarization of presynaptic terminals, (2) in regulating release of neurotransmitters, and (3) with both pre- and postsynaptic receptor proteins for small neurotransmitters and neu- ropeptides (Gokin and Strichartz 1999). These receptors include both inotropic
and metabotropic receptors, membrane macromolecules that are coupled to the activation of various intracellular signaling pathways using enzymes that themselves may be direct targets of local anesthetic action. Since clinically used local anesthetics are relatively impotent drugs, with inhibitory actions dependent on millimolar concentrations at their target molecule, and often have restricted access to their desired target site, they are used clinically at high enough concentrations that many different molecular targets are likely to be modiﬁed by them.
Like many “amphipathic” molecules, which contain both hydrophobic and hydrophilic—even ionized—regions, local anesthetics distribute into and have
effects upon the dynamic properties of membranes (Lissi et al. 1990). Almost all of the recognized targets of local anesthetics are membrane-intrinsic or membrane-associated macromolecules, and the partitioning within and per-
meation through biomembranes is a feature of local anesthetics that is essential for their fundamental actions and also governs their clinical effectiveness (de Paula and Schreier 1995).
In this chapter we will describe the various mechanisms and molecular and cellular targets for local anesthetics during neural blockade, as well present the phenomenon of and speculate on mechanisms for long-term relief of chronic pain by intravenous local anesthetics. Unlike the acute nerve-blocking actions
of local anesthetics, the mechanisms underlying the long-term effects of intra- venous drugs to relieve chronic pain symptoms remain a mystery. We will also consider the long-held possibility of using certain naturally occurring toxins to
effect analgesia by potent and speciﬁc binding to single classes of ion channels.
Finally, we will examine recently described nontraditional actions of local anesthetics, particularly their antiinﬂammatory effects, and relate the known mechanisms in neuronal tissues to those on immune cell functions.
Molecular Targets of Local Anesthetics
Blockade of Action Potentials: The Classic Local Anesthetic Action
Action potentials are the hallmark of excitable membranes, and their occur- rence is absolutely dependent on ion channels that open in response to depo- larization and allow current to enter the cell. In most excitable membranes of nerve, skeletal, and cardiac cells, the major inward current that drives the fast depolarizing phase of the action potential is carried through VGSCs. In order to achieve the conditions for an action potential to occur, i.e., “threshold,” the cell membrane must be adequately stimulated, depolarized sufﬁciently to open enough VGSCs to produce inward current that will overcome the
outward current that ﬂows through the K+ and Cl− channels that coexist in
the membrane. The initial sources of these impulse-generating stimuli are physiological transducing events, such as occur at distal sensory endings, or postsynaptic “excitatory” receptors, such as occur in the CNS (spinal cord). Local circuit current from a depolarized membrane under an action potential ﬂows to an unexcited adjacent region to stimulate it to threshold, accounting for the propagation of the depolarizing wave of the action potential along an axon. This mode of propagating stimulation has a higher “margin of safety” than the other stimulation modes, i.e., the current supplied by the adjacent excited membrane is about 5–10 times greater than that required to reach threshold. At any one of these locations, impulse-generating processes can be suppressed by local anesthetics to prevent or abolish action potentials. Cor- responding to its higher margin of safety for conduction, propagating axonal action potentials are more resistant to treatments that block inward current than are, for example, the initial generation of action potentials at sensory end- ings. It therefore takes lower concentrations of local anesthetics to suppress the initiation of an impulse than those necessary to abolish that impulse once it has started propagating (Raymond 1992). Furthermore, the margin of safety for conduction is apparently greatest in nonmyelinated C-ﬁbers (and in the smallest, i.e., slowest conducting, among these ﬁbers) and is least in the small myelinated, A-delta and A-gamma ﬁbers (Huang et al. 1997; Gokin et al. 2001). This accounts for the differential order of susceptibility to local anesthetics of axons documented directly, in vitro and in vivo (small myelinated>large myelinated>non-myelinated). These observations have been made by many experimentalists but are still challenged by the historical “size principle” that states that smaller ﬁbers, regardless of micro-anatomy, are always more suscep- tible to local anesthetics than larger ﬁbers, a principal extrapolated from very early observations on compound action potentials at a time before C-ﬁbers had been identiﬁed (Raymond and Gissen 1987).
Binding of Local Anesthetics to Voltage-Gated Na+ Channels
Blockade of action potentials relies on the inhibition of VGSCs. The regions of the VGSC that interact directly to bind local anesthetics have been identiﬁed, principally through site-directed mutation (Ragsdale et al. 1994; Wang et al.
2000, 2001). The picture that emerges is of a dynamically altered drug binding site that varies with the state of the channel. Resting closed channels have the lowest afﬁnity. Inactivated closed channels are much higher in afﬁnity, pri- marily because of a much slower dissociation rate (Chernoff 1990; Hille 1977), perhaps because the drug must escape from a sterically occluded channel, blocked by an “inactivation moiety” that closes the pore’s cytoplasmic ending (Catterall 2000; Vedantham and Cannon 1999; Wang et al. 2004; Courtney and Strichartz 1987). Open channels also have a higher afﬁnity, but bind the drugs much faster than inactivated channels do (Hille 1977; Chernoff 1990). One
thermodynamic consequence of the differential binding afﬁnity to different channel states is that the transitions between these states, the so-called “gat- ing” of the channel, must be modiﬁed when local anesthetics are bound (Hille
1977; Balser et al. 1996).
Many of the amino acid side chains that appear to interact with the drug molecules are located around the cytoplasmic vestibule that forms a portal to the channel’s narrow pore (Kondratiev and Tomaselli 2003; Ragsdale et al.
1994), although parts of the “inner pore domain,” located deeper in the channel, also seem to inﬂuence binding (Sunami et al. 1997). Of the residues lining the vestibule, some are more inﬂuential in determining drug afﬁnity for the
resting state; others are more important for determining inactivation state binding (Li et al. 1999; Nau et al. 1999, 2003). From this we conclude that local anesthetic binding does not occur at a constant locus in the channel and, given
the conformational ﬂexibility of most local anesthetics, it is likely that they adjust their shape to ﬁt the channel’s altered conformations. The name that has been applied to this reciprocal adjustment of drug and channel during
state-dependent binding is the “modulated receptor” (Hille 1977).
All VGSCs are composed of one large (ca. 180 kDa) α-subunit and 1 to 2 auxiliary β-subunits (Catterall 2000). The ion pore of the channel and the major gating machinery, including the loci for local anesthetic binding, are
located on the α piece, although combination with β-subunits can inﬂuence the overall gating pattern and thus, through this modulation of state transitions, inﬂuence the state-dependent binding of these drugs (Wang et al. 1996; Wright
et al. 1997, 1999). Each α-subunit carries four homologous domains, each domain contributing one segment of the channel’s ion conducting pore and one part of the overall gating apparatus that couples membrane voltage to
channel states. Each domain also contains regions that interact with local anesthetics, although it is unclear whether a tightly bound local anesthetic molecule simultaneously interacts with all four regions. Judging from their relatively low afﬁnity, rarely having KD values below 10 μM, and their weak stereoselectivitytency ratios rarely exceeding 5 (Lee-Son et al. 1992; Valenzuela
et al. 1995; Brau et al. 2000; see review by Nau and Strichartz 2002), the ﬁt of local anesthetics to VGSCs is apparently not very tight (Courtney and Strichartz 1987).
Local Anesthetic Binding to K+ and Ca2+ Channels
There is much structural similarity among the voltage-gated, cation-selective ion channels (Yu et al. 2005). The charged transmembrane regions that act as “voltage-sensors,” the hairpin segments that line the narrowest part of the pore to form the ion “selectivity ﬁlter,” and several of the other transmem- brane helices that support these critical functional regions are homologous structures among many of these ion channels. Not surprisingly, then, they
also possess a similar pharmacology for drugs that interact with these regions, among which are the local anesthetics, although, in general, their afﬁnity for these targets is less than that for the Na+ channels.
Both the transient “A-type” and the more persistent “delayed rectiﬁer” types of K+ channels, both of which are activated by membrane depolariza- tions, are blocked by local anesthetics (Courtney and Kendig 1988; Castle 1990; Josephson 1988; Olschewski et al. 1998; Komai and McDowell 2001). The ki-
netics of the block are most consistent with high-afﬁnity binding to the open state, and the mechanistic picture that accompanies these kinetics portrays one local anesthetic molecule entering an open channel’s pore and occlud- ing it, analogous to the mode for blocking open Na+ channels (Castle 1990; Valenzuela et al. 1995). Classical “inward rectiﬁer” K+ channels are much less susceptible to local anesthetics (Castle 1990; Carmeliet et al. 1986), perhaps because they are not activated by voltage but instead may have a metal cation binding site in their pore that has few of the physicochemical properties that bind a local anesthetic, e.g., hydrophobicity. However, other K+ channels that are not voltage-gated but are open at the resting potential are potently in- hibited by local anesthetics (Brau et al. 1995; Olschewski et al. 1996; Kindler
et al. 1999).
Voltage-gated Ca2+ channels also are blocked by local anesthetics (Palade and Almers 1985; Oyama et al. 1988; Guo et al. 1991; Sugiyama and Muteki
1994; Xiong and Strichartz 1998; Liu et al. 2001). Interestingly, in the case of the slowly inactivating L-type Ca2+ channels, the primary mode of block seems to be a promotion of the inactivated state (Carmeliet et al. 1986; Xiong and Strichartz 1998), like one of the actions of local anesthetics (LAs) on VGSCs
and also like the dihydropyridine drugs, e.g., nifedipine, that are classical inhibitors of L-type channels. The LAs do not, however, bind at the dihy- dropyridine binding site on these channels (Hirota et al. 1997; Xiong and Strichartz 1998).
Pacemaker currents in neurons (and in cardiac muscle), Ih , carried by the hyperpolarization-activated, cyclic nucleotide-gated (HCN) class of channels which conduct Na+ and K+ about equally well, are LA sensitive too (Bischoff et al. 2003). Concentrations of bupivacaine and lidocaine to half block these channels, ca. 50 and 100 μM, respectively, are actually lower than those re-
ported to block Na+ channels in the same sensory neurons (Bischoff et al.
2003). Membrane-impermeant derivatives of local anesthetic—such as the permanently charged quaternary homolog of lidocaine, QX314—which block most cation channels from the intracellular direction, do not attenuate pace- maker currents when applied to the cell’s exterior, suggesting that there is also an intracellular route to the local anesthetic binding site of these channels. Inasmuch as pacemaker channels are found in many neurons (Doan et al.
2004) and have been implicated as sources that drive the repetitive ﬁring often associated with abnormal pain (Chaplan et al. 2003; Yao et al. 2003)—and
because selective inhibitors of Ih delivered systemically are known to relieve
certain neuropathic pain symptoms (Chaplan et al. 2003; Lee et al. 2005; cf. Raes et al. 1998)—their relative susceptibility to blockade by local anesthetics nominates them as likely targets for some of the clinical analgesic actions of these drugs.
The physiological, functional consequences of blocking these different channels will obviously depend on that channel’s role in neural activity. Potas- sium channel block is often synergistic with the Na+ channel blocking activity for effecting impulse blockade by local anesthetics (Drachman and Strichartz
1991). Blocking K+ channels can depolarize the resting membrane and usually slows the repolarization of the action potential. In the absence of an accom-
panying Na+ channel blockade by the drug, these actions will often enhance excitability, bringing the resting membrane closer to threshold and prolong- ing the duration of the action potential, an action that raises the margin of
safety for impulse propagation and increases the likelihood of repetitive ﬁring after a single stimulus (Raymond et al. 1990). When conjoined with the Na+ channel blocking properties of local anesthetics, however, K+ channel block- ade has a remarkably synergistic action on action potentials. Both the steady depolarization of the resting potential and the prolonged depolarization of the action potential potentiate impulse blockade by increasing the presence of the open and inactivated Na+ channels that bind LAs with high afﬁnity. This over- all effect illustrates an important pharmacological principle: the integrated actions of an agent that acts on several, physiologically coupled targets may differ from the action that would be predicted from the actions of that agent on the separate targets. Since the nervous system almost always utilizes plural ion
channels and receptors in neural conduction and synaptic transmission (Hille
2001), the net effect of local anesthetics cannot always be predicted directly from their actions on only one of these targets.
Inhibition of the Ca2+ channels at nerve terminals by local anesthetics has dramatic physiological effects, disproportionally greater than the inhibition of action potentials through Na+ channel blockade. Transmitter release from
nerve terminals depends on the free cytosolic concentration of Ca2+ in the presynaptic ending, raised to the third power or higher. Therefore, if half the Ca2+ channels are blocked, the release of transmitter might be reduced
to about 10% of the control value in drug-free conditions. By comparison, the margin of safety for impulse propagation is so large for most peripheral nerve axons that about 80% of the channels must be blocked before impulse
conduction fails. These calculations are not intended to indicate the functional potency relationships for local anesthetics acting at different ion channels, but rather to emphasize the different forms of nonlinear relations between blockade of ion channels and the inhibition of the physiological functions
to which they are coupled. Ultimately, it is these ﬁnal physiological actions that are most directly related to analgesia, and the toxic side effects of local anesthetics.