Inhibition of Mammalian Gq Protein Function by Local Anesthetics

The studies were performed in Xenopus  oocytes. These cells express endogenous LPA and trypsin receptors; other G-protein–coupled receptors can be expressed conveniently. Intracellular Ca release as a response to receptor stimulation is easily assessed as Ca-activated Cl currents, and the size of the cells makes intracellular injection straightforward. In addition, using oocytes allowed comparison with our previous results obtained in this model. The study protocol was approved by the Animal Research Committee at the University of Virginia (Charlottesville, Virginia). Oocyte harvesting, receptor expression, intracellular injections, drug administration, and electrophysiologic recording were performed as described previously. 1,2,6–8 

Phosphorothioate oligonucleotides were synthesized by the University of Virginia Research Facility (Charlottesville, Virginia). The antisense sequence is complementary to specific 20-base segments with less than 50% homology with other types of X. laevis  Gα proteins. 9Sense oligonucleotides were used as control. Uninjected oocytes (for experiments on the LPA receptor) or those injected 24 h prior with cRNA encoding the m3 receptor were injected with 50 nl sterile water containing 50 ng/cell antisense or sense oligonucleotides. Forty-eight hours after oligonucleotide injection, the cells were tested as described previously.

Lysophosphatidic acid and methylcholine, used as agonists for the LPA or m3 muscarinic receptor, were diluted in Tyrode's solution to the required concentration and superfused (3 ml/min) over the oocyte for 10 s. The oocyte was positioned close to the inflow tubing so that complete exposure to test solutions was obtained in 4.8 ± 0.4 s (n = 20). Responses were quantified by measuring peak current and are reported as μA.

For intracellular administration of QX314 or lidocaine, a third micropipette was inserted into the voltage-clamped oocyte. The micropipette was connected to an automated microinjector (Nanoject; Drummond Scientific, Broomall, PA). Under voltage clamp, 25 nl (approximately 5% of total oocyte volume) of a 300-mm KCl solution was injected for determination of the control response; in the treatment group, we injected 25 nl KCl solution containing various concentrations of QX314 or lidocaine. Injection was followed by superfusion with Tyrode's solution for 10 min, preventing an extracellular effect of any QX314 or lidocaine leaked from the puncture site or through the membrane. I^Cl(Ca)was then induced by superfusion of LPA or methylcholine, as described previously. Control and treatment responses were obtained from different oocytes to prevent the effects of receptor desensitization from obscuring the results.

The purification of Gq is based on the original method of Biddlecome et al. , 10described in detail in Lindorfer et al.  11Minor changes have been made to this protocol in an attempt to increase yields. Briefly, for the chromatographic separation of Gq using Ni²⁺-NTA resin (Qiagen), the Q chromatography buffer was modified to contain the following: 20 mm Tris (pH 8.0), 150 mm NaCl, 1 mm MgCl², 0.5% (v/v) Genapol, 5 mm imidazole, 1 mm β-mercaptoethanol, 10 μm GDP, 17 μg/ml phenyl-methylsulfonyl fluoride (PMSF), and 2 μg/ml pepstatin, leupeptin, and aprotinin. Furthermore, during the washing of the Ni²⁺-NTA column, the 1-m NaCl wash was removed, and the 0.2% cholate and 0.3% cholate–GTPγS washes were combined into one wash with 0.3% cholate and 3 μm GTPγS. Approximately 20–30 μg Gq was purified from a 20 g (wet wt.) Sf9 cell pellet. For experiments, 25 nl purified mouse Gq protein (final intracellular concentration 5 × 10⁻⁸m) was dissolved in detergent (final intracellular concentration cholate 0.001%). Cholate in a final intracellular concentration of 0.001% was used as control.

Results are reported as mean ± SD. Measurements of at least 22 oocytes were averaged to generate each data point. As variability between batches of oocytes is common, responses were at times normalized to control response. Statistically significant differences were assessed using one-way analysis of variance followed by Student-Newman-Keuls correction for multiple comparisons. P < 0.05 was considered significant. Concentration–response curves were fit to the following logistic function, derived from the Hill equation

where y^maxand y^minare the maximum and minimum response obtained, n is the Hill coefficient, and X⁵⁰is the half-maximal effect concentration (EC⁵⁰for agonist) or the half-maximal inhibitory effect concentration (IC⁵⁰for antagonist).

Molecular biology reagents were obtained from Promega (Madison, WI), and other chemicals were obtained from Sigma (St. Louis, MO). QX314 was a gift from Astra Pharmaceuticals, L.P. (Westborough, MA).

To provide baseline measurements and to confirm that our model functioned appropriately, we determined the concentration–response relation for LPA. LPA induced inward currents (I^Cl(Ca)) as described previously by our group 1,7,12and others 13,14(fig. 1A). As shown in figure 1B, the response to LPA was concentration dependent. EC⁵⁰was 288 ± 30 nm (n > 22 for each data point). Maximal responses of 1.2 ± 0.1 μA were obtained at an LPA concentration of 10 μm. Calculated E^maxwas 1.4 ± 0.2 μA, and the Hill coefficient was 0.42 ± 0.09. These findings are similar to those reported in our previous studies. 1,15 

Whereas uninjected oocytes were unresponsive to methylcholine, oocytes injected with m3 muscarinic receptor cRNA responded to application of methylcholine (10⁻⁵–10⁻⁹m) with a transient I^Cl(Ca)(fig. 1C). We have shown previously that this response is mediated by m3 muscarinic receptors as it is inhibited by atropine and the selective m3 antagonist 4-diphenylacetoxy-N-methylpiperidine (4-DAMP). 16 

We determined the concentration–response relation for the m3 response. As shown in figure 1D, this response was also concentration dependent. EC⁵⁰, calculated from the Hill equation, was 200 ± 70 nm. Maximal responses of 1.5 ± 0.2 μA were obtained at an methylcholine concentration of 10 μm. Calculated E^maxwas 1.6 ± 0.2 μA. These findings also compare closely with data reported in our previous studies. 2,16–18 

We then studied the effect of intracellularly injected QX314 on LPA signaling. LPA-induced responses (at EC⁵⁰, 290 nm) were inhibited in a concentration-dependent manner after 10 min exposure to various concentrations of QX314. IC⁵⁰for intracellular QX314 was 209 ± 164 μm (figs. 2A and B). Maximal inhibition was obtained with QX314 10 mm; at this concentration, LPA responses were inhibited by 79%. The IC⁵⁰determined for intracellularly injected QX314 on LPA signaling (209 ± 164 μm) is similar to that obtained previously for m3 signaling (444 ± 226 μm), 2which would be expected if the intracellular site of action is the same, namely the Gq protein.

Since we planned to use lidocaine in further experiments, we next studied the inhibitory potency of intracellularly injected lidocaine on responses elicited by administration of LPA at EC⁵⁰(290 nm). A sample trace of a LPA response after 10 min exposure to intracellular lidocaine (100 μm) is shown in figure 2C. Figure 2Dillustrates the concentration–response relation for the effect of intracellular lidocaine. Curve fitting to the Hill equation revealed an IC⁵⁰of 148 ± 105 μm, which is very close to that determined previously for the intracellular inhibitory potency of lidocaine on trypsin signaling (445 ± 147 μm). 1Maximal inhibition (84% of control response) occurred at an intracellular lidocaine concentration of 10 mm.

Depletion of endogenous frog Gq protein reduces response sizes and eliminates the inhibitory action of intracellular LA. To assure that this effect demonstrated in our previous article 1also holds true for the different experimental conditions employed in the current study, we first determined whether oocytes injected with sense oligonucleotides (in KCl and detergent carrier) maintain a normal sensitivity to LAs. In the same set of experiments, we also verified that under the same conditions, anti-Gq–injected oocytes loose sensitivity to LA. As shown in figure 3, mean control response for oocytes injected with Gq sense, KCl, and detergent, elicited by 29 nm LPA (one tenth of EC⁵⁰), was 0.73 ± 0.14 μA (n = 26). Injection of QX314 (at IC⁵⁰, 210 μm) reduced LPA-evoked responses to 38% of control response (0.28 ± 0.11 μA; n = 29). This inhibitory effect by QX314 was completely abolished when the oocytes were first injected with anti-Gq (0.84 ± 0.15 μA; n = 33). These results demonstrate that the changed experimental conditions did not affect our previous findings. We now investigated whether intracellular injection of mouse Gq protein would reverse these effects on LPA and muscarinic m3 signaling pathways. Only one tenth of EC⁵⁰for the agonists was chosen since we knew from previous experiments that, after KCl injection, these concentrations elicit responses similar in size to those of full EC⁵⁰in the absence of intracellular KCl.

We used Gq-sense oligonucleotide–injected oocytes as control to exclude the possibility that injection of DNA oligonucleotides per se  affects agonist responses. Mean control response elicited by 29 nm LPA (one tenth of EC⁵⁰) was 0.6 ± 0.1 μA (n = 42). As shown in figure 4A, injection of antisense oligonucleotides against the endogenous frog Gq protein reduced the mean response size to 33 ± 7% of the corresponding control response (n = 47). In contrast, responses were unchanged (n = 57; 131 ± 29% of control) in cells in addition injected with mouse Gq protein (5 × 10⁻⁸m). Injection of mouse Gq protein “rescued” the inhibitory effect of intracellularly injected QX314 (at IC⁵⁰, 210 μm). Whereas QX314 was without effect on Gq-depleted oocytes, responses to 29 nm LPA (one tenth of EC⁵⁰) after QX314 injection were inhibited to 44 ± 10% (n = 54) of the corresponding control response in cells in addition injected with mouse Gq protein (5 × 10⁻⁸m). Thus, mammalian Gq protein both reversed the decreased responses induced by endogenous Gq depletion and enabled the inhibitory effect of intracellular LA.

To assure that this effect is not restricted to the experimental LA QX314, we repeated these studies using intracellularly injected lidocaine (fig. 4B). In Gq-depleted oocytes, mean response size following stimulation with 29 nm LPA (one tenth of EC⁵⁰) was reduced to 42 ± 11% (n = 44) of control responses (obtained in cells injected with Gq-sense oligonucleotides). Mean control response was 0.8 ± 0.1 μA (n = 39). Additional injection of mouse Gq protein (5 × 10⁻⁸m) prevented this inhibition (response size 109 ± 12% of control; n = 49) and reestablished the intracellular inhibitory effect of lidocaine (at IC⁵⁰, 150 μm; 53 ± 10% of corresponding control response; n = 48).

To rule out that this effect is specific for the endogenous LPA signaling pathway, we also studied recombinantly expressed m3 muscarinic receptors (fig. 4C). Mean control response was 1.0 ± 0.2 μA (n = 31). Oocytes injected with antisense oligonucleotides directed against endogenous Gq protein showed significantly reduced responses (to 49 ± 10% of control; n = 56). In contrast, responses after additional injection of mouse Gq protein (5 × 10⁻⁸m) were unchanged (94 ± 18% of control; n = 51). These responses were inhibited to 45 ± 16% (n = 56) of control response after intracellular application of QX314 (at IC⁵⁰, 210 μm).

In the current study, we have shown that mammalian Gq protein is able to couple to, and mediate signal transduction through endogenous and recombinantly expressed G-protein–coupled receptors in Xenopus  oocytes after depletion of the endogenous frog Gq protein. In addition, where knockdown of the endogenous frog Gq protein eliminates sensitivity to intracellular LA of several GPCR pathways, the presence of mammalian Gq protein can prevent this effect.

The Gq subunit is involved in a broad variety of signaling pathways, some of which are of interest to perioperative medicine. Examples of mediators signaling through this G protein are angiotensin, platelet-activating factor, and various cytokines. As a specific example, priming of PMNs, responsible for excessive stimulation of the inflammatory response, is mainly Gq protein mediated. 19LA inhibition of Gq protein function leads therefore to selective inhibition of PMN priming, explaining at least in part their well-known antiinflammatory actions (for review, see Hollman and Durieux 3). In contrast, “physiologic” free radical production by neutrophils is primarily mediated by Gi proteins and therefore not inhibited by LA.

As in our previous studies, we used the Xenopus  oocyte model. Potential problems with the technique have been discussed in our previous reports. 1,2,6–8Our results predict that every signaling pathway mediated by Gq would be at least partially inhibited by LAs. However, intravenous administration of LAs does not result in shutdown of major signaling pathways. Several explanations may exist for this apparent paradox. First, the concentrations required for intracellular block of GPCR signaling are significantly greater than those attained in blood after usual intravenous doses. However, additional LA binding sites on GPCRs may result in significantly greater sensitivity, 20and in some settings (e.g. , spinal anesthesia), concentrations able to block Gq signaling are likely to be attained. Second, GPCRs are only partially inhibited by LAs, even at high concentrations. Third, since most receptors couple to multiple G-protein subtypes, lack of activity of the Gq protein can in most instances be compensated for by other G-protein subunits. In particular, G¹¹, structurally very similar and functionally virtually identical to Gq, is likely to play a major role in this regard. Surprisingly, LAs have been shown to discriminate between those two subunits: whereas Gq protein function is inhibited by local anesthetics, G¹¹is not. 1Indeed, in this and previous studies, local anesthetics interfere selectively with Gq protein function, but not with the function of other Ca-signaling G proteins. LA inhibition of G-protein–gated inwardly rectifying K (GIRK) channels, for example, is not mediated by block of the coupling G^oor Gi protein, but rather by interaction with phosphatidylinositol 4,5-bisphosphate (PIP²). 21 

Inhibition of mammalian Gq protein function by LAs, as shown in this study, provides a potential explanation for several clinical effects of these compounds 4,22that cannot be attributed to sodium channel blockade. 3The antiinflammatory and probably the closely interwoven antithrombotic effects of LA 3might be explained at least in part by functional inhibition of a common Gq protein. Platelets in particular might be a target as these cells do not contain G¹¹proteins. Therefore, in these cells, some signaling pathways depend on functioning Gq proteins. Offermanns et al.  23reported that platelets from mice deficient in the α subunit of Gq are unresponsive to a variety of physiologic platelet activators. As a result, these mice are protected from collagen and adrenaline-induced thromboembolism. The authors concluded that the Gq protein may thus be a new target for drugs designed to block the activation of platelets. Together with our data, this suggests that the antithrombotic actions of LAs might result in part from inhibition of Gq in platelets.

To the best of our knowledge, LAs are the first compounds shown to be selective G-protein inhibitors. If the results of this study are confirmed in other models and structural modification of the LA molecule can increase their G-protein–blocking activity while reducing affinity for the sodium channel, new therapeutic indications for these drugs might be feasible.

Summarized, our study has shown that LAs inhibit mammalian Gq protein function. Therefore, next to the Na channel, the Gq protein must be seen as an intracellular target site for LAs, possibly explaining several of their clinical effects. In addition, LAs might be used as lead compounds for the design of novel therapeutics.

The authors thank Prof. Dr. med. Eike Martin (Department of Anesthesiology, Ruprecht-Karls-Universität Heidelberg, Heidelberg, Germany) for his support.