The Quaternary Lidocaine Derivative, QX-314, Produces Long-lasting Local Anesthesia in Animal Models In Vivo 

The experimental protocol comprising the three animal models was approved by The University of British Columbia Committee on Animal Care (Vancouver, British Columbia, Canada). We used a randomized, blinded, and controlled experimental design. The experimenter was blinded to all treatment allocations and drug doses. Animal allocation to treatment groups was randomized by computer with the use of Research Randomizer, version 3.0.∥Each experiment was conducted with both negative (placebo: normal saline) and positive (70 mm lidocaine) controls.

A slightly modified version of the method by Bulbrig and Wadja24was used. In brief, the backs of guinea pigs (weight, 800–1,200 g; n = 8) were shaved 24 h before experimentation. Each animal received four treatments intradermally on the lower flank using 27-gauge, ½-inch hypodermic needles. Volumes of 0.5 ml were injected, producing wheals with a diameter of 2 cm. Six pinpricks were applied along the perimeter of a 1-cm-diameter circle drawn with the origin at the center of the wheal. This was done to standardize the reversal of local anesthesia, because preliminary observations indicated that the duration of blockade was inversely related to the distance from the center of the wheal. Pinpricks were standardized using a weighted 20-g, 30-gauge needle. The number of pinpricks out of six producing a cutaneous trunci reflex was recorded at 1, 5, 15, 35, 50, 80, 110, 140, and 180 min after injection and every hour afterward. The number of responses out of six was converted to a nociceptive blockade score out of 100%. At 110 min after injection, wheals to which the animal responded six out of six times were no longer monitored.

A modification of the method of Grant et al.  25was used. In brief, mice (total, n = 30) were placed in restraining tubes, and the 2-cm tips of their tails were submerged into a 50°C water bath. The tail-flick latency was measured, and animals with latencies less than 3 s were selected.

Two 20-μl injections were performed bilaterally 7 cm proximal to the tip of the tail with 29-gauge hypodermic needles. Needles were inserted with the bevel facing medially until their tips contacted the caudal vertebrae, and were then withdrawn by 1 mm before injection. After injection, the tail-flick latency was tested for at 1, 10, 20, 30, 45, 60, 90, and 120 min after injection and every hour afterward until the time to offset was determined. Mice were monitored for a minimum of 60 min. At each sampling time, the maximum exposure to the hot water bath was 4 s, and a tail-flick latency greater than 4 s was defined as sensory blockade. Because of the tendency for the mice to struggle when placed in restraining tubes, the onset of sensory blockade was defined as two consecutive readings with a tail-flick latency greater than 4 s, and, likewise, the offset of sensory blockade was defined as two consecutive readings with a tail-flick latency less than 4 s.

The method of Leszczynska and Kau26was used. In brief, all mice (total, n = 45) were placed in the middle of a 20 × 25-cm inverted mesh and were taught to climb to the outside and over the edge of the mesh. All mice were able to climb on the inverted mesh with all four limbs before treatments. Mice were placed into restraining tubes for injection, and injections were performed in the area of the popliteal fossa of the left hind limb using 30-gauge, 1-inch hypodermic needles. The bevel tips of the needles faced proximally. After injection, mice were removed from the restraining tubes and positioned onto the inverted mesh. The primary endpoint was the time to loss of ability to hang on to the inverted mesh with the injected hind limb, which was tested for at 1, 5, 10, 20, 30, 45, and 60 min after injection and every hour afterward. The animals were tested on the inverted mesh for a minimum of 60 min. After this time, only mice that were unable to hang on to the inverted mesh with the injected hind limb were monitored.

NaCl, N -(2,6-dimethylphenyl carbamoylmethyl)triethylammonium chloride (QX-314), and lidocaine HCl were purchased from Sigma-Aldrich Canada Ltd. (Oakville, Ontario, Canada). Normal saline (NaCl) solution was 0.9% wt/vol. Lidocaine and QX-314 were dissolved in saline, with the NaCl concentration decreased by equivalent moles of lidocaine or QX-314 in solution to maintain equiosmolality. The pH of all solutions was titrated to pH 6.9 ± 0.1 with NaOH and HCl before freezing at −15°C in 1-ml aliquots. Before the start of each experiment, solutions were thawed and placed within syringes up to 24 h before use. While in syringes, solutions were kept refrigerated at 10°C. Before commencement of the individual experiments, drugs were equilibrated at room temperature for 30 min or more.

Statistical analyses of continuous data were conducted with the use of one-way and repeated-measures analysis of variance for multisample comparisons as appropriate, with the Tukey multiple comparison test for post hoc  testing. We used the Kolmogorov-Smirnov test to test if values came from a gaussian distribution. The Pearson correlation was used to test for a linear relation between two gaussian variables. The Student t  test was used for comparisons of two groups of normally distributed data and testing for differences from a theoretical mean. Nonnormally distributed data were analyzed with the Kruskal-Wallis test. The Grubbs test was used to test for statistical outliers. Categorical data were analyzed with Fisher exact test. Survival curves of categorical time-to-event data were compared with the log-rank test; survival fractions were calculated using the product limit (Kaplan-Meier) method. Differences were considered significant at P < 0.05. Statistical tests were two-tailed, and data are expressed as mean ± SD, n = sample size, unless mentioned otherwise. The data were analyzed using Prism version 4 (GraphPad, San Diego, CA) and Microsoft Excel version 2003 software (Microsoft Corporation, Redmond, WA).

In the guinea pig intradermal assay, QX-314 reversibly and dose-dependently inhibited the cutaneous trunci muscle reflex (fig. 1; F^4,55= 5.01; P < 0.01). At 10 mm, QX-314 produced a partial reflex blockade, which peaked at 34 ± 11 min (n = 6). At 30 mm, complete nociceptive blockade (> 90% effect) was observed at 12 ± 5 min (n = 6), and at 70 mm, complete blockade occurred at 18 ± 13 min (n = 6). In contrast, the onset of 70 mm lidocaine was within 1 min in all animals tested (n = 6). Normal saline produced no inhibition of the cutaneous trunci muscle reflex in any animal tested (n = 5). The duration of complete inhibition of the cutaneous trunci muscle reflex was 62 ± 20 min for 70 mm lidocaine. For 70 mm QX-314, this value was 250 ± 132 min, a greater than fourfold increase in duration when compared with the equimolar concentration of lidocaine (Tukey multiple comparison test, q = 5.58; P < 0.01). The time to complete cutaneous trunci muscle reflex recovery after QX-314 injection also was dose dependent (fig. 2; F^3,20= 26.83; P < 0.0001). At 10 mm QX-314, blockade reversed completely at 175 ± 103 min after injection. At 30 mm, reversal occurred at 323 ± 111 min, and at 70 mm, reversal occurred at 650 ± 171 min. By comparison, in animals injected with 70 mm lidocaine, the cutaneous trunci muscle reflex completely recovered at 100 ± 24 min, representing a sixfold difference in offset time at an equivalent concentration (Tukey multiple comparison test, q = 11.69; P < 0.001).

In the mouse tail-flick test, 70 mm lidocaine inhibited the tail-flick reflex response to a thermal stimulus in all animals tested (n = 6). QX-314 blocked the tail-flick response in a concentration-dependent fashion: two of six mice in the 10 mm QX-314 group did not produce a tail flick to a thermal stimulus, five of six mice did not respond in the 30 mm QX-314 group, and six of six mice showed no tail flick with 70 mm QX-314 (all groups, P > 0.05 compared with 70 mm lidocaine). The offset of nerve block due to QX-314 also was concentration dependent, with higher concentrations producing longer times to offset (cf.  table 1; F^3,15= 26.33; P < 0.0001; Pearson correlation coefficient, 0.775). Onset times were fast when 70 mm lidocaine was injected, blocking the response to a thermal stimulus within 1 min in all animals tested, whereas the onset of block in the QX-314 groups was considerably slower (table 1; F^3,15= 7.76; P < 0.01). In this assay, the mean sensory conduction block produced by 70 mm QX-314 was more than 10 times longer than that produced by 70 mm lidocaine (Tukey multiple comparison test, q = 7.32; P < 0.001; table 1). Figure 3illustrates the onset and offset of sensory blockade due to lidocaine and QX-314 in the form of survival curves. In all animals, the actions of both lidocaine and QX-314 were reversible. No animal injected with saline exhibited evidence of sensory blockade (n = 6).

In the mouse sciatic nerve blockade model, injection of 50 μl lidocaine, 70 mm, into the popliteal space prevented all mice (n = 9) from using the injected hind limb to hang on to an inverted mesh. On the other hand, only 1 mouse of 10 that received 10 mm QX-314 (Fisher exact test, P < 0.001 compared with 70 mm lidocaine), and 5 of 10 mice injected with 30 mm QX-314 (P < 0.05 compared with lidocaine) were unable to hang on to the inverted mesh. All 10 mice that received 70 mm QX-314 were unable to hang on to the inverted mesh (P > 0.05 compared with lidocaine). Mice that received QX-314 and showed a positive response usually did not respond immediately to the QX-314, whereas each mouse that received lidocaine responded within 1 min. In this assay, 70 mm QX-314 produced nerve blockade for 282 ± 113 min, more than 12 times longer than 70 mm lidocaine (23 ± 10 min; Tukey multiple comparison test, q = 9.93; P < 0.0001). In each group, there was a single mouse that was unable to hang on to the inverted mesh with the injected hind limb for an extended period of time. The time to recovery of these four animals' ability to hang on to the mesh was 120 min (70 mm lidocaine), 7 days (10 mm QX-314), 15 h (30 mm QX-314), and 6 days (70 mm QX-314), respectively. These data met the criteria of statistical outliers using the Grubbs test (P < 0.01).27Furthermore, each of these four mice was unable to hold on to the mesh with the injected hind limb immediately upon injection (within 1 min), raising the possibility of direct nerve trauma. For these reasons, these animals were excluded from statistical analysis. None of the six control animals injected with placebo (normal saline) exhibited signs of motor blockade. The effects of lidocaine and QX-314 injection into the popliteal space are summarized in table 2. Figure 4depicts the onset and offset of sciatic nerve motor blockade due to lidocaine and QX-314 in the form of survival curves.

Of the six animals injected with 70 mm lidocaine in the mouse tail-flick test, three exhibited visible erythema which appeared 24 h after injection. This also was observed in two of six animals injected with 10 mm QX-314, one of six injected with 30 mm QX-314, and five of six injected with 70 mm QX-314 (all groups, P > 0.05 compared with 70 mm lidocaine). In all animals, the erythema disappeared within 6 days. Saline did not produce any erythema. We observed no signs of local toxicity, such as tissue induration or necrosis, in any animal. Inspection of the tails of mice receiving tail injections 24 h after experimentation also revealed no signs of local tissue toxicity.

In this study, we have demonstrated that the quaternary lidocaine derivative, QX-314, concentration-dependently and reversibly produced robust local anesthesia of long duration in three different in vivo  animal models, at concentrations equivalent to those clinically relevant for lidocaine. Compared with 70 mm lidocaine (i.e. , approximately 2%), QX-314 produced local anesthesia up to 6 times longer in the guinea pig intradermal wheal assay, up to 10 times longer in the mouse tail sensory conduction blockade test, and up to 12 times longer in the mouse sciatic nerve motor blockade model. Hence, QX-314 was effective in producing impulse blockade along free nerve endings, C fibers, and motor (Aα) fibers. In each animal model, QX-314 produced local anesthesia with a slower onset compared with lidocaine, but with substantially longer duration. Collectively, these results raise the possibility that quaternary ammonium compounds may be clinically useful as long-acting local anesthetics in humans and challenge the conventional notion that these agents are ineffective when applied extracellularly.28 

Our findings correspond well to previous observations by others that extracellularly applied QX-314 does affect action potentials. For example, in experiments on squid axons, external QX-314 (10 mm) reduced the action potential rate of rise by 9%.8In single myelinated frog fibers, 5 mm external QX-314 reduced evoked current amplitudes by 7%.5In frog sciatic nerve, a 40-min exposure to QX-314 (20 mm) reduced the compound action potential amplitude by 63%.22Whereas these results per se  were not suggestive of clinically useful local anesthesia, it is noteworthy that these experiments were performed at relatively low temperatures (22°, 6°, and 4°C, respectively). Should the activity of extracellular QX-314 be dependent on diffusion through the nerve membrane, the low temperatures would have greatly reduced its ability to do so, possibly explaining the relatively modest activity observed in vitro . It is noteworthy, though, that hypothermia potentiated the blocking action of extracellular lidocaine in a study on desheathed rat sciatic nerves.29 

The precise mechanism of action that underlies the observed local anesthetic effects of QX-314 in the current study remains uncertain. In experiments on frog sciatic nerve, the effect of external QX-314 was noted to last more than 1 h after drug washout, and marked frequency-dependent inhibition of the compound action potential was observed.22Similar results showing marked frequency-dependent blockade also were obtained in rat vagus nerve.23Because local anesthetic frequency-dependent blockade is mediated by the cationic form of the drug, which becomes “trapped” by the inactivation gate and ion selectivity filter of voltage-gated Na⁺channels,5,30,31the observation that use-dependence occurs even when QX-314 is applied extracellularly22suggests that the local anesthetic action of QX-314 may be due to intracellular blockade of Na⁺channels through the hydrophilic pathway. It is hence possible that external QX-314 produces local anesthesia by traversing the lipid membrane to bind to its conventional intracellular receptor on the Na⁺channel. The slow onset and long duration of QX-314–mediated local anesthesia provides additional support for the hypothesis that external QX-314 penetrates the nerve membrane (albeit slowly) and then interacts with an intracellular binding site. Because QX-314 is permanently charged, to allow penetration across lipid membranes, an appropriate combination of a large concentration gradient and time would be required. Our observed large differences in local anesthetic duration with increasing QX-314 concentrations further supports the notion that QX-314 penetrates the nerve membrane(s) to act intracellularly, in which case higher QX-314 concentrations would be associated with a higher rate of intracellular penetration and, consequently, a longer duration of blockade.

In the three animal models, 10 and 30 mm QX-314 did not produce local anesthesia with the same consistency as 70 mm. To be effective for conduction blockade, local anesthetics must prevent depolarization at a minimum of three successive nodes of Ranvier so as to prevent axonal saltatory conduction, and the nerve must be exposed to a minimum blocking concentration of local anesthetic.32Because QX-314 does not penetrate lipid membranes easily, this agent would be expected to possess a relatively high (extracellular) minimum blocking concentration. Hence, it is likely that the minimal distance from the nerve fiber that a local anesthetic must be applied to achieve effective conduction blockade is lower for QX-314 than for a conventional tertiary agent such as lidocaine. Consequently, the duration and magnitude of blockade due to QX-314 would be expected to vary greatly with concentration, injected volume, and the proximity of the injection to the nerve. Consistent with this hypothesis is our observation in each of the three animal models of a high variability in both local anesthetic efficacy and duration of blockade at the lower QX-314 concentrations.

Collectively, our results suggest that long-acting local anesthetics can be created through quaternization of the molecule. For example, whereas quaternization decreases the potency of lidocaine, it markedly increased the duration of action, an effect that is not exclusive to lidocaine derivatives: In previous experiments, a 2% solution of the quaternary methyl derivative of cocaine produced local anesthesia with an onset within 30 min and a duration of 4–5 h.21Similar results were obtained with the quaternary benzyl derivative of diphenyl-piperidinoethyl-acetamide, whereas the tertiary derivatives of these compounds produced local anesthesia within 5 min and possessed a much shorter duration of action, again highly consistent with the results of our study.

In addition to prolonged duration, quaternary local anesthetics have several other potential advantages over conventional tertiary compounds. For example, quaternary agents do not readily penetrate the blood–brain barrier, raising the possibility that these agents are associated with decreased central nervous system toxicity. In fact, one study with QX-314 provides support for this hypothesis.33It also is conceivable that quaternary agents produce fewer cardiotoxic effects, because negative inotropism of local anesthetics correlates with lipid solubility.34Furthermore, systemic plasma concentrations of quaternary compounds after peripheral nerve blockade are expected to be low, because the permanent charge should impede systemic absorption through capillaries and facilitate rapid renal excretion. Moreover, the use of quaternary local anesthetics may obviate the use of added vasoconstrictors such as epinephrine to prolong the duration of blockade. There are many clinical situations where this would be advantageous, such penile blockade, regional anesthesia in patients with cardiovascular disease, or narrow-angle glaucoma. Finally, one may speculate that quaternary agents also could provide enhanced sensory–motor separation. A major determinant of a local anesthetic's magnitude of sensory–motor separation is its use-dependent blocking activity.35The capability for producing use-dependent blockade inversely correlates with lipid solubility, the latter of which is decreased by quaternization.36However, whereas the current study was not designed to investigate differences between tertiary and quaternary agents in terms of sensory–motor separation, we observed no apparent differences in effective blocking doses of QX-314 between the sensory and motor blockade assays.

In summary, we have demonstrated here that the quaternary lidocaine derivative, QX-314, concentration-dependently produces long-lasting, reversible local anesthesia in three standard in vivo  animal models, at concentrations equivalent to those clinically relevant for lidocaine. Our findings raise the possibility that quaternary local anesthetics may produce clinically useful, long-lasting local anesthesia in humans and challenge the conventional notion that quaternization of tertiary amines obliterates extracellular local anesthetic activity.

The authors thank Ernest Puil, Ph.D., Michael J. A. Walker, Ph.D., and Richard A. Wall, Ph.D. (all Professors Emeriti, Department of Anesthesiology, Pharmacology and Therapeutics, The University of British Columbia, Vancouver, British Columbia, Canada), for insightful comments, and Jimmy T. C. Wang, B.Sc. (Graduate Student, Department of Anesthesiology, Pharmacology and Therapeutics, The University of British Columbia, Vancouver, British Columbia), for technical assistance.