Isozyme-specific Effects of Protein Kinase C in Pain Modulation

• Protein kinase C plays an important role in the generation and maintenance of hypersensitivity after inflammation and nerve injury, but the roles of specific protein kinase C isoenzymes have not been clearly defined

• In mice with deletion of specific protein kinase C isoenzymes, there was no effect on baseline sensory testing but differing reliance of isoenzymes on behavioral responses to inflammation and nerve injury

THE International Association for the Study of Pain defines pain as “an unpleasant sensory and emotional experience associated with actual or potential tissue damage, or described in terms of such damage.” Pain is critical for the survival of an organism because it enables the organism to detect and avoid injury. Unfortunately, sometimes pain can become a serious medical problem (e.g. , cancer pain, neuropathic pain, inflammatory pain, etc. ) that requires effective management.

Activation of nociceptors by noxious stimuli (e.g. , heat, cold, mechanical injury, etc .) causes the transmission of nociceptive input from the peripheral site of noxious stimuli to the central nervous system, which can result in the perception of physiologic pain. The transmission of nociceptive input through pain pathways is modulated by multiple mechanisms at different levels of the nervous system. For example, descending inhibitory pathways attenuate the perception of pain, whereas facilitatory pathways enhance it. Multiple neurotransmitters, receptor systems, and intracellular signal transduction pathways participate in nociceptive transduction and modulation.

Protein kinase C (PKC) is an umbrella term for a family of serine/threonine kinases. At least 10 PKC isoforms have been identified, and these have been subdivided into three groups based on calcium and diacylglycerol dependence: classic (α, βI, βII and γ; calcium and diacylglycerol-dependent), novel (δ, ε, η, and θ; calcium-independent and diacylglycerol-dependent), and atypical (ζ and ι/λ; both calcium and diacylglycerol-independent).1These different PKC isozymes play important roles in the regulation of a variety of cellular functions (e.g. , differentiation, proliferation, cell migration, apoptosis, neuron excitability, and neurotransmitter release), making them attractive therapeutic targets for a host of human diseases.

Interestingly, a number of PKC isozymes have been found in the anatomical areas that regulate pain and nociception (e.g. , amygdala, periaqueductal gray, dorsal root ganglia, spinal cord, etc. ).2,3As such, the role of PKC in pain and analgesia has been an area of intense research over the last two decades.3The large number of PKC isoforms and the overlapping substrate specificities of these isoforms in vitro  make the unequivocal identification of isozymes-specific roles for PKC a major challenge. Many studies that employed pharmacological methods have shown the important roles of the PKC family in the modulation of pain.3However, the relatively poor specificity of most PKC activators and inhibitors has not only prevented the identification of the PKC isozyme-specific effects in pain, but also has caused conflicting results on the effect of PKC on pain. For example, using different PKC inhibitors, Igwe and Chronwall found that thermal hyperalgesia induced by complete Freund's adjuvant (CFA) was mediated by protein kinase C βII,4but Yajima et al.  reported that thermal hyperalgesia induced by CFA did not depend on PKC.5To gain specific insight into isozyme-specific roles for the PKCs, we used targeted gene knockout techniques in mice to investigate the specific roles of four different PKC isozymes (α, β, γ, and δ) in the modulation of pain. We hypothesized that different PKC isozymes play distinct roles in various aspects of pain modulation and analgesia. Because of the existing of conflicting reports on the effect of PKCγ on the development of morphine tolerance,6,7we also investigated the effect of PKCγ on the development of morphine tolerance using PKC γ knockout mice.

All experiments were performed according to the guidelines of the National Institutes of Health and were approved by the Animal Care and Use Committee of Washington University School of Medicine, St. Louis, Missouri. Mice were housed with a 12 h light-dark cycle with ad libitum  access to food and water. Targeted deletions of PKC α,8β,9γ,10or δ11were achieved by homologous recombination in processes described in detail previously. These animals were obtained from various sources, and thus have a variety of genetic backgrounds, each involving a varying number of backcrosses onto the C57BL/6 strain. PKCα, β, and δ mice were generated in house by one of the authors, whereas PKCγ knockout mice were purchased from Jackson Labs, Bar Harbor, ME. All lines were maintained as het x het crosses to generate wild-type and knockout mice from the same crosses. Because of the varying genetic backgrounds of these mice, each knockout was compared only to its own wild-type littermates as controls. The experimenter was blind to the genotypes of mice in all behavior experiments.

An accelerating Rotarod (Ugo Basile, Comerio, Italy) was used to assess motor coordination. Latency to fall as the Rotarod accelerated from 4 to 40 rpm over 5 min was assessed. Five consecutive trials were performed, with at least 15 min between each trial.

The surface of the hot plate (IITC Life Science, Woodland Hills, CA) was heated to a constant temperature of 52°C. Mice were placed on top of the hot plate, which was enclosed with clear Plexiglas chambers (width, 10 cm; length, 10 cm; height, 15 cm). The latency to respond with hindpaw licking or jumping (whichever occurred first) was measured. The mouse was immediately removed from the hot plate and returned to its home cage. Each animal was tested only once.

Each mouse was held gently in a towel and the distal one-third of its tail was immersed in a temperature-controlled water bath (52°C). The latency between tail immersion and the tail-flick response was recorded. A cutoff of 10 s was applied to prevent tissue damage.

Paw-withdrawal latencies to radiant heat stimuli were recorded using a plantar thermal stimulator (IITC Life Sciences).12Mice were placed into Plexiglas chambers on a glass surface that was heated to 30°C. Following a 30-min acclimation period, a radiant heat stimulus was applied to each hindpaw, and the time to paw withdrawal was measured. Each hindpaw was tested three times with a minimum intertrial interval of 5 min. A 20-s exposure limit was imposed to prevent tissue damage. The paw withdrawal latency of each mouse was calculated as the mean of the three trials.

Mechanical paw withdrawal thresholds were measured using the up-down testing paradigm.13,14The von Frey filaments (0.02, 0.04, 0.08, 0.16, 0.32, 0.64, 1.28, and 2.56 g) were applied to the lateral part of the hindpaw (in spared nerve injury [SNI] neuropathic pain model) or the middle plantar surface of the hindpaw until paw withdrawal or a maximum duration of 1 s. The 0.32-g stimulus was applied first. Whenever a withdrawal response to a given filament occurred, the next smaller von Frey filament was applied. Whenever a negative response occurred, the next higher von Frey filament was applied. The test continued until either the responses of four additional stimuli after the first change in response had been obtained, or the upper/lower end of the von Frey filament was reached. Threshold values were derived according to the method described by Chaplan et al.  13 

Mice were placed into a transparent observation chamber (30 × 30 × 25 cm) for adaptation 30 min before the experiments. Mice were then gently restrained by hand and injected subcutaneously with 10 μl formalin solution, 2%, in the plantar surface of the right hindpaw with a 30-gauge needle. After formalin injection, mice were then placed into the original chamber and were observed for licking and flinching behavior. Time spent licking or flinching was recorded in 5-min sections, starting from the time of formalin administration through 60 min postinjection.

Intraplantar injection of CFA (10 μl, 1.0 mg/ml; Sigma-Aldrich) was used as a chronic inflammatory pain model. Following CFA injection, thermal and mechanical sensitivity were measured using the von Frey filament and Hargreaves tests, as described above.

All animals underwent the SNI operation as described previously.15,16Briefly, during anesthesia with pentobarbital (50 mg/kg, intraperitoneal), the left sciatic nerve and its three terminal branches (sural, common peroneal, and tibial) were exposed at the mid-thigh. The left tibial and common peroneal nerves were tightly ligated with 8–0 silk and transected distal to the ligation, and 1 mm of the distal nerve stump was removed. Care was taken to prevent damage to the sural nerve. Muscle layers were closed using 6–0 suture, and the skin incision was closed with wound staples.

To measure the analgesic effect of morphine, two cumulative morphine-dosing regimens were used: In naïve mice the doses of morphine were 0.5, 1.0, 2.0, 3.0, 4.0, 6.0, and 10.0 mg/kg; in morphine-tolerated mice the doses of morphine were 3, 6, 10, 15, 30, 60, 100, and 300 mg/kg. Morphine was administered intraperitoneally in a volume of 200 μl, which was diluted in normal saline. The interval between drug injections was 30 min. Tail-flick latencies were tested 30 min after each injection of a drug dose.

Morphine analgesia was calculated using the percent maximum possible effect using the following equation: percent maximum possible effect = ([Measured tail-flick latency]−[baseline tail-flick latency]) × 100/(10 −[baseline tail-flick latency]), wherein 10 is the number of seconds used as the cut-off time for tail-flick latency measurement.

Two paradigms were used to induce morphine tolerance: One 75 mg slow-release morphine pellet (provided by National Institute on Drug Abuse) was implanted subcutaneously; one 25 mg release-controlled morphine pellet was implanted subcutaneously, followed by another 25 mg morphine pellet implanted subcutaneously 4 days later. All implantations were done during anesthesia with pentobarbital.

Results are expressed as mean ± SEM. Statistical analysis was performed using STATISTICA software (version 6; StatSoft, Inc., Tulsa, OK). Comparisons of two means were made by unpaired two-tailed Student t  test. Comparisons of time sequence data between groups were made by repeated measures two-way ANOVA, followed by Tukey post hoc  test. Differences were considered statistically significant when P ⩽ 0.05.

The accelerating rotarod test was employed to measure the gross motor functions of mice. Compared with their respective wild-type littermates, PKCα, β, γ, or δ knockout mice exhibited no statistically significant difference in the latency to drop from the rotarod (fig. 1). In addition, all knockout mice demonstrated trial-to-trial improvement in performance on the rotarod that was indistinguishable from their respective wild-type littermates, consistent with a lack of impairment in simple motor learning. These results suggest that deletion of PKCα, β, γ, or δ does not change gross motor function or impair simple motor learning in mice.

Three different tests were used to determine if deletion of an individual PKC isozyme alters sensitivity to noxious heat: hot-plate test, tail-flick test, and Hargreaves test.

In the hot-plate test, sensitivity to noxious heat applied to the paws was evaluated by measuring the latency paw-licking or jumping after mice were placed on the surface of a 52°C hot plate. The response latencies for PKCα, β, γ, or δ knockout mice were not statistically significantly different from their respective wild-type littermate mice (fig. 2A).

The tail-flick test was used to evaluate sensitivity to noxious heat applied to the tail. The latency to respond with a tail flick, following immersion of the tail in a hot water bath (52°C), was measured. The tail-flick latencies of PKCα, β, γ, or δ knockout mice were not statistically significantly different from their respective wild-type littermate mice (fig. 2B).

The Hargreaves test was used to evaluate sensitivity to noxious heat applied to the hindpaws. The withdrawal latency to a radiant heat stimulus applied to the hindpaw was measured. The paw withdrawal latencies of PKCα, β, γ, or δ knockout mice were not significantly different from their respective wild-type littermate mice (fig. 2C).

Together, these data indicate that deletion of individual PKC isozymes α, β, γ, or δ does not change the baseline sensitivity of mice to noxious heat stimuli.

The mechanical sensitivity of mice was tested with von Frey filaments. The paw-withdrawal thresholds to mechanical stimuli were calculated using the up-down method.13,14The paw-withdrawal thresholds of PKCα, β, γ, or δ knockout mice were not statistically significantly different from their respective wild-type littermate mice (fig. 2D). These results indicate that deletion of an individual PKC isozyme α, β, γ, or δ does not change the baseline sensitivity of mice to mechanical stimuli.

In the formalin test, we recorded the time that mice spent licking or flicking their hindpaws after intraplantar injection of formalin solution. The time spent licking or flicking after formalin injection was not statistically significantly different in PKCα, β, γ, or δ knockout mice compared with their respective wild-type littermate mice (fig. 3). These results indicate that deletion of individual PKCα, β, γ, or δ isozymes has no effect on the nociceptive response induced by formalin.

Paw-withdrawal thresholds to application of a series of calibrated von Frey filaments were measured in the CFA-induced persistent inflammatory pain model. After intraplantar injection of CFA, paw-withdrawal thresholds for all groups of mice were statistically significantly decreased in the injected hindpaw compared to baseline, demonstrating the induction of mechanical hypersensitivity, or allodynia. No statistically significant changes in the extent of CFA-induced mechanical hypersensitivity were seen in PKCα, β, γ, and δ knockout mice relative to their respective wild-type littermate mice (fig. 4). These results indicate that deletion of individual PKCα, β, γ, or δ isozymes has no effect on the mechanical allodynia induced by intraplantar injection of CFA.

Paw-withdrawal latencies to radiant heat stimuli were measured in the CFA-induced chronic inflammatory pain model. After intraplantar injection of CFA, paw-withdrawal latencies of wild-type mice were statistically significantly decreased compared with their respective baseline, indicating the mice developed thermal hyperalgesia. CFA-induced thermal hyperalgesia was not significantly different in PKCα knockout mice compared with wild-type littermates (fig. 5A). In contrast, compared with their respective wild-type littermates, PKCβ, γ, or δ knockout mice developed statistically significantly less thermal hyperalgesia (group main effect; no statistically significant differences in group x time effects) following CFA injection (figs. 5B, C, D). These results indicate that deletion of PKCβ, γ, or δ, but not PKCα, attenuates the thermal hyperalgesia induce by CFA.

The SNI model of neuropathic pain was employed to investigate the effect of PKCα, β, γ, and δ on the development of neuropathic pain. Paw-withdrawal thresholds to mechanical stimuli were measured with von Frey filaments before and after SNI. After tibial and femoral nerve injury, the paw-withdrawal thresholds measured in the lateral sides of mouse hindpaws were statistically significantly decreased in the isplilateral paws of wild-type mice (fig. 6). This result demonstrates the development of mechanical allodynia induced by SNI. Nerve injury-induced mechanical allodynia was not significantly different in PKCβ or δ knockout mice compared with wild-type littermate control mice (fig. 6B and D). Nerve injury-induced mechanical allodynia was significantly increased in PKCα knockout mice compared with their respective wild-type littermates (fig. 6A). In contrast, nerve injury-induced mechanical allodynia was significantly decreased in PKCγ knockout mice compared with littermate control mice (fig. 6C). The above results suggest different roles for individual PKC isozymes in the development of nerve injury-induced pain. PKCγ attenuates nerve injury-induced mechanical allodynia, whereas deletion of PKCα has an opposite effect. Deletion of PKC β or δ does not alter nerve injury-induced mechanical allodynia.

There is evidence suggesting that PKC may be involved in the development of morphine tolerance,17,18but the effect of PKCγ on morphine tolerance is still controversial.6,7We investigated the effect of PKCγ on the development of morphine tolerance using PKC γ knockout mice.

We initially used subcutaneous implantation of 75 mg morphine pellets to induce morphine tolerance. As table 1shows, this dose resulted in a statistically significant increase in the mortality of PKCγ knockout mice compared with their wild-type littermate mice within the first day after implantation of morphine pellets.

We next tried a novel method to induce morphine tolerance by subcutaneously implanting one 25 mg morphine pellet followed by implantation of another 25 mg pellet 4 days later. The tail-flick latency was determined before implantation of the first morphine pellet (baseline) and again 4 days after the implantation of the second pellet. The tail-flick latency after the implantation of morphine pellets was not statistically significantly different compared with their respective baselines in PKCγ knockout mice and their wild-type littermates (data not shown). The dose-response curves of morphine analgesia 4 days after the implantation of the second morphine pellet in PKC γ knockout mice and their wild-type littermates were significantly shifted to the right compared with that of naïve PKCγ knockout mice and their wild-type littermates (fig. 7), which indicates the development of morphine tolerance in both PKC γ knockout mice and their wild-type littermates. There was no statistically significant difference in the dose-response curves of morphine analgesia between PKCγ knockout mice and their wild-type littermate mice either before or after the implantation of morphine pellets, indicating that deletion of PKCγ does not alter morphine analgesia or the development of morphine tolerance. However, the absence of PKCγ led to increased mortality at high doses of morphine, indicating that PKCγ may play a role in opposing the toxic effects of morphine that are not directly related to analgesia.

To investigate the effects of different PKC isoforms on pain modulation, we used a variety of pain models to study multiple aspects of pain. The behavior tests used to evaluate thermal and mechanical sensitivity depend upon the ability of the mouse to withdraw from a stimulus, requiring intact motor reflexes. Using the accelerating rotarod test, we evaluated gross motor function in PKC-mutant mice to ensure that deletion of individual PKC isoforms does not impair gross motor function or simple motor learning. Our results from the rotarod test show that deletion of PKCα, β, or δ did not result in any statistically significant impairment of gross motor function. It has been reported that PKCγ knockout mice displayed impaired motor coordination,19but no statistically significant difference of drop latencies was found between PKCγ knockout mice and their wild-type littermates in our experiments. The lack of effect of knockout of PKCα, β, γ, or δ on locomotor function supports the use of these animals in research on pain using motor-reflex pain models in these mice.

PKC is present at multiple sites of the nervous system that contribute to pain transmission and modulation. It has been reported that PKC is involved in multiple aspects of pain modulation.3We first explored whether knockout of PKCα, β, γ, or δ alters basal pain perception. Our results from hot-plate, Hargreaves and tail-flick tests indicate that PKCα, β, γ, and δ are not required for noxious heat perception; our results using the von Frey test indicate that PKCα, β, γ, and δ are also not required for mechanical sensation. Our results indicate that these individual PKC isoforms are not required for normal acute noxious heat or mechanical sensitivity. This agrees with the literature reporting that PKC inhibitors have no effect on normal acute noxious heat and mechanical sensitivity.3,20,21It is also possible that multiple PKC isoforms subserve similar functions in noxious heat and mechanical sensation, thus preventing observable phenotypes in mice where an individual isoform has been ablated.

The formalin test is a well-established model which has been used extensively to study the mechanisms of inflammation-induced pain. The intraplantar injection of formalin induces characteristic nociceptive behaviors, which appear in two distinct phases. The first phase is thought to result from direct activation of nociceptive afferent fibers. The second phase is believed to be mediated by the activation of central neurons that are sensitized due to peripheral inflammation, as well as ongoing activity of primary afferents.22,–,24Intrathecal administration of PKC inhibitors has been shown to decrease the second phase of nociceptive behavior in the formalin test.25,26Our results show that deletion of individual PKC isoforms α, β, γ, or δ in mice had no effect on formalin-induced nociceptive behavior, suggesting that none of the PKCα, β, γ, or δ isoforms is individually required for formalin-induced pain. This may suggest that other PKC isoforms that were not tested in this study might be candidate modifiers of the formalin second phase. An alternative possibility is that the individual PKC isoforms have redundant functions, and that simultaneous knockout of combinations of these would reveal dependence on multiple PKC isoforms. Yet another possibility is that the PKC inhibitors used in previous studies do not act in a PKC-specific manner at the doses used in vivo . Our observation that PKCγ is not required for formalin-induced pain behavior is in contrast with the report of Malmberg et al.  that the second phase of the formalin test was significantly reduced in PKCγ knockout mice compared with wild-type mice.27The reason for this difference is unclear.

Chronic inflammatory pain conditions, like rheumatoid arthritis, are among the most common health problems and are difficult to treat. CFA-induced inflammation in the mouse is widely used as a model to study the mechanisms and management of chronic inflammatory pain. Our results indicate that PKCβ, γ and δ, but not PKCα, are involved in the development of CFA-induced thermal hyperalgesia and that none of PKC isoforms are individually necessary in the development of CFA-induced mechanical allodynia.

Neuropathic pain is a devastating consequence of injury or diseases of the peripheral or central nervous system and is characterized by spontaneous pain, hyperalgesia, and allodynia. The mechanisms that underlie neuropathic pain are incompletely understood. It has been reported that PKCγ knockout mice do not develop nerve injury-induced mechanical allodynia in the partial sciatic nerve ligation model of neuropathic pain.27In addition, intrathecal lentiviral-mediated RNA interference targeting PKC γ attenuates chronic constriction injury-induced neuropathic pain in rats.28These studies suggest that PKCγ is required for the full expression of nerve injury-induced pain. The SNI model of neuropathic pain involves tightly ligating and transecting two of the three terminal branches of the sciatic nerve (tibial and common peroneal nerves), leaving the remaining sural nerve intact.15SNI induces very stable, robust, and long-lasting mechanical allodynia.29,30Using the SNI model, we found that PKCγ, but not PKCβ and δ, plays a critical role in the development of nerve injury-induced mechanical allodynia. Further, our results show that deletion of PKCα enhances the mechanical allodynia induced by SNI. This may be caused by potential adaptive changes in the nerve system, such as up-regulation of PKCγ expression, induced by PKCα knockout, or by a loss of an inhibitory effect of PKCα on neuropathic mechanical allodynia. All these possibilities need to be further investigated.

The effect of PKCγ on the development of morphine tolerance has been investigated intensively.6,7,17,18,31Zeitz et al.  reported that there was reduced development of morphine tolerance in PKCγ knockout mice,6but Yukhananov and Kissin reported that they did not find any difference in the development of morphine tolerance in PKCγ knockout mice compared with wild-type control mice. We tried to use the same 75 mg morphine pellets as reported by the above two studies to induce morphine tolerance. Interestingly, we found the mortality caused by morphine pellet implantation in PKCγ knockout mice was very high, which made our further research using the same morphine pellets unpractical. Our two 25 mg morphine pellets implantation paradigm did not cause death of either PKCγ knockout mice or in their wild-type littermates, and induced very robust morphine tolerance. Our results indicate that PKCγ is not individually required for morphine analgesia or in the development of morphine tolerance.

Other PKC isoforms may also be involved in different aspects of pain modulation. For example, PKCε in primary afferent nociceptors plays multiple crucial roles in hyperalgesic priming.32It is necessary to further investigate the effects of other PKC isoforms in the modulation of pain.

In conclusion, our study examines the specific roles of different PKC isoforms (α, β, γ, and δ) in different aspects of pain modulation. These studies suggest a limited and modality-specific role of these PKC isoforms in inflammatory pain. However, our findings extend earlier reports and substantially support the hypothesis that PKCγ might be effectively targeted for the management of multiple forms of neuropathic pain.

The authors thank Sherri Vogt, B.S., Lab Manager, and Changshen Qiu, M.D., Lab Manager, for assistance with the animal colonies, and Judith Golden, Ph.D., Assistant Professor; Benedict Kolber, Ph.D., Postdoctoral Associate; and Steve Davidson, Ph.D., Postdoctoral Associate, for assistance with the manuscript. All are from Washington University School of Medicine, St. Louis, Missouri.