Cholecystokinin type B receptor-mediated inhibition of A-type K+ channels enhances sensory neuronal excitability through the phosphatidylinositol 3-kinase and c-Src-dependent JNK pathway

Background Cholecystokinin (CCK) is implicated in the regulation of nociceptive sensitivity of primary afferent neurons. Nevertheless, the underlying cellular and molecular mechanisms remain unknown. Methods Using patch clamp recording, western blot analysis, immunofluorescent labelling, enzyme-linked immunosorbent assays, adenovirus-mediated shRNA knockdown and animal behaviour tests, we studied the effects of CCK-8 on the sensory neuronal excitability and peripheral pain sensitivity mediated by A-type K+ channels. Results CCK-8 reversibly and concentration-dependently decreased A-type K+ channel (IA) in small-sized dorsal root ganglion (DRG) neurons through the activation of CCK type B receptor (CCK-BR), while the sustained delayed rectifier K+ current was unaffected. The intracellular subunit of CCK-BR coimmunoprecipitated with Gαo. Blocking G-protein signaling with pertussis toxin or by the intracellular application of anti-Gβ antibody reversed the inhibitory effects of CCK-8. Antagonism of phosphatidylinositol 3-kinase (PI3K) but not of its common downstream target Akts abolished the CCK-BR-mediated IA response. CCK-8 application significantly activated JNK mitogen-activated protein kinase. Antagonism of either JNK or c-Src prevented the CCK-BR-mediated IA decrease, whereas c-Src inhibition attenuated the CCK-8-induced p-JNK activation. Application of CCK-8 enhanced the action potential firing rate of DRG neurons and elicited mechanical and thermal pain hypersensitivity in mice. These effects were mediated by CCK-BR and were occluded by IA blockade. Conclusion Our findings indicate that CCK-8 attenuated IA through CCK-BR that is coupled to the Gβγ-dependent PI3K and c-Src-mediated JNK pathways, thereby enhancing the sensory neuronal excitability in DRG neurons and peripheral pain sensitivity in mice.


Background
Cholecystokinin (CCK), a gastrointestinal polypeptide hormone existing in a variety of amino acid chain lengths, has been isolated from the central nervous system and peripheral tissues [1]. Two types of functional membrane receptors, cholecystokinin A receptor (CCK-AR), located mainly on pancreatic acinar cells, and CCK-BR, mostly in the stomach and nervous system tissues, have been identified as the endogenous receptors of CCK [2]. CCK, acting through its receptors, is found to be involved in the regulation of a variety of physiological functions in the nervous system, including metabolic, neurotrophic and modulatory actions [3]. Additionally, in vitro experiments have suggested that CCK might regulate the sensitivity of nociceptive sensory neurons [4], where CCK-BRs were abundantly expressed [5,6]. It has been established that the elevated level of CCK mRNA in the dorsal root ganglia (DRG) induced by peripheral nerve injury sensitizes and excites primary afferent sensory neurons, leading to pain hypersensitivity [7], with the application of CCK inducing pronociceptive effects [8]. Evidence also suggests that CCK is a potential trigger for increased visceral sensitivity in healthy subjects [9] as well as in irritable bowel syndrome patients [10,11]. Moreover, antagonism of CCK receptors can effectively reverse burn-induced mechanical allodynia [5], and the deletion of the CCK-BR gene attenuates the symptoms of mechanical allodynia in neuropathic pain [12]. So far, however, the mechanisms underlying the CCK-mediated hyperalgesia still remain unclear.
Changes in neuronal excitability of peripheral sensory neurons might directly regulate symptoms of pain, such as allodynia, hyperalgesia and spontaneous pain [13]. Voltagegated K + channels (Kv) are one of the major classes of ion channels responsible for driving neuronal excitability in both the central and peripheral nervous system [14] and in whole-cell patch clamp recordings are separated into two major categories: a large transient component characteristic for fast-inactivating A-type channels and a sustained delayed-rectifying channels, which respectively mediate I A and I DR currents [15,16]. A-type channels are sensitive to millimolar concentrations of 4-aminopyridine (4-AP) [17], and they play pivotal roles in the control of the electrical properties and excitability of nociceptive neurons [14,15]. Recent evidence has suggested that A-type channels have been implicated in pain plasticity and neuropathic conditions [17], which begin with the aberrant firing of action potential bursts in damaged neuronal tissue. For example, peripheral nerve injury results in the reduction of channel expression, thereby decreasing I A , and enhancing the sensory neuronal excitability and pain sensitivity [18,19]. Manipulation of I A is, therefore, predicted to affect neuronal excitability and useful for pain treatment.
In the current study, we examined the regulation of CCK-8 on I A in small-sized (< 30 μm in soma diameter) DRG neurons to determine whether A-type K + channels mediate the nociceptive actions of CCK-BR.

Animals
Adult ICR mice (male, 6-8 weeks of age) were purchased from the Experimental Animal Center of Soochow University. Mice were maintained in specificpathogen-free facilities on a 12-h light-dark cycle at a room temperature of 22 ± 1°C, and housed in cages with access to food and water ad libitum. All animal studies were conducted in accordance with the National Institutes of Health's Guidelines for Animal Care and Use and approved by the Animal Care and Use Committee of Soochow University.

Co-immunoprecipitation (co-IP)
Total cellular proteins were extracted using homogenization buffer (Tris 20 mM, pH 7.4, NaCl 150 mM, EDTA 1 mM, 0.5% Triton, and DTT 1 mM) supplemented with a cocktail of protease inhibitors. After centrifugation at 21,000 g for 30 min at 4°C, the supernatant was saved and protein concentration was measured by BCA protein assay (Beyotime, Shanghai, China). Extract containing 500 μg of protein was incubated at 4°C for 3 h with 3 μg of goat polyclonal antibody against CCK-BR (1: 500, Abcam). Protein A-Sepharose beads (Amersham Biosciences) were added to the samples and gently shaken for 4 h at 4°C. Beads were then rinsed and removed in lysis buffer. The pellet was boiled with 4 × Laemmli sample buffer and separated by SDS-PAGE. Immunoreactive proteins on membranes were developed as described above.

Immunofluorescence staining
Immunohistochemistry was performed as previously described [20,23]. Tissue samples were sectioned into 15 μm thin slices using a cryostat (CM 1950; Leica, San Jose, USA). The sections were blocked with 5% normal goat serum in PBS, plus 0.2% Triton X-100 for 1 h at room temperature then incubated overnight at 4°C with primary antibody against CCK-BR (goat, 1: 500, Abcam), antibody against NF-200 (mouse, 1:1000, Abcam), or antibody against CGRP (mouse, 1:500, Abcam). Sections were washed three times with PBS at room temperature, followed by Cy3-conjugated donkey anti-goat IgG (1: 500, BBI Life Science), FITC-conjugated donkey antimouse IgG (1:200, BBI Life Science) or FITC-IB4 (1:200, Sigma) in PBS at room temperature for 2 h. After sections were washed three times with PBS at room temperature, images were captured with a fluorescence microscope (Nikon 104c, Japan). Negative controls, omitting each primary antibody, were used in each case, and no significant staining was observed in these samples (data not shown).

PKA activity assay
PKA activity in homogenates was determined by enzyme-linked immuno sorbent assay (ELISA, Promega), according to the manufacture's instructions. Briefly, the cells were pretreated with either vehicle or KT-5720 for 30 min, followed by treatment with either vehicle (0.1% DMSO), or forskolin for 15 min. The cells were washed with ice-cold phosphate-buffered saline (PBS), placed on ice, and incubated with 200 μl lysis buffer. After a 10-min incubation on ice, the cells were transferred to microcentrifuge tubes. Cell lysates were centrifuged for 15 min, and aliquots of the supernatants containing 0.2 μg of protein were assayed for PKA activity. The activity is expressed as RLU − 1 (relative light units)/amount of protein.

PI3K activity assay
Cells were stimulated with or without CCK-8 (100 nM) for 15 min. After stimulation, PI3K activity in homogenates was determined with a PI3-Kinase HTRF™ Assay kit (Millipore Corporation, Bedford, MA), using 20 μg of protein for each sample, as stated in the manufacturer's protocol. HTRF was then measured with an excitation wavelength of 335 nm and emission wave length of 620 and 665 nm with a spectrofluorometer (Tecan, Infinite M1000, Salzburg, Austria).

Behavioral test
Behavioral testing was conducted in an appropriately lighted, quiet room, always during the light cycle between 9:00 AM and 4:00 PM in a series and by the same experimenter. The operator who performed the behavioral tests was blinded to all treatments. Animals were allowed to acclimate to a testing room for at least 30 min before performing the assessments. Mechanical sensitivity was determined on paw withdrawal to manual application of graded von Frey hairs (0.02-2.56 g; Stoelting) to the plantar surface as described previously. Thermal sensitivity was tested using a commercially available paw thermal stimulation system (IITC Life Sciences), and are expressed as paw-withdrawal latency (PWL) and tail-flick latency. Animals were gently dropped into an acrylic box with a metal floor that was preheated to a certain temperature. The values of PWL were calculated using a timer that was started when the animal is released onto the preheated plate and stopped at the moment of withdrawal, shaking, or licking of either hind paw. The cutoff latency was set to prevent tissue damage. All animals were tested once for each temperature per session in a random sequence. All drugs or vehicle were injected subcutaneously into the plantar surface of one hind paw in a volume of 10 μl. The pH of the solutions was adjusted at 7.4 to prevent skin irritation.

Data analysis
In electrophysiological experiments, data acquisition and analysis were performed with Clampfit 10.2 (Axon Instruments) and/or GraphPad Prism 5.0 software (Prism Software). The amplitude of I A was measured at the peak. The data plots were fit by the Boltzmann equation for the activation and inactivation curves as described previously [24]. All data are presented as means ± SEM. Statistical significance between two groups was determined using a paired or unpaired Student's t two-tailed test. Comparisons of multiple groups against a pooled control were tested using one-way analysis of variance (ANOVA) followed by a Bonferroni's post-test. Differences in values over time among groups were done using two-way ANOVA. The criterion for significance in all analyses was considered as p < 0.05.

CCK-8 selectively decreased I A in DRG neurons
The studies in vitro of nociceptive processing usually examined different subtypes of peripheral sensory neurons [25,26]. In the present study, we limited patch-clamp recordings to small-sized DRG neurons (< 30 μm in soma diameter) as these neurons are primarily involved in nociceptive signaling [19,26]. Two main types of outward voltage-gated K + channel (Kv) currents have been characterized in these nociceptive neuronsthe transient A-type K + channel currents (I A ) and the sustained and delayed-rectifier K + channel currents (I DR ) [15,16]. We first isolated these two kinetically different wholecell currents. A total outward current exhibiting a rapidly inactivating and a more sustained component was elicited by a depolarizing pulse from the holding potential of − 80 mV to + 40 mV (Fig. 1a). Biophysical separation of a delayed-rectifier current (I DR ) was obtained by a depolarizing prepulse to − 10 mV, which inactivated the transient channels. I A was then isolated by subtracting I DR from the total current (Fig. 1a). Addition of 5 mM 4-aminopyridine (4-AP) inhibited the remaining outward current by 87.1 ± 5.3% (n = 6, Fig. 1b), further confirming the effective isolation of I A .
Application of 100 nM CCK-8 to small-sized DRG neurons significantly decreased I A by 30.9 ± 3.7% (n = 8, Figs. 1c and e), while I DR was not effectively affected (decreased by 1.2 ± 0.9%, n = 10, Figs. 1d and e). The amplitude of I A partially recovered after CCK-8 washout (Fig. 1c). The CCK-8 effect on I A was concentration-dependent (Fig. 1f). The half-maximal inhibitory concentration (IC 50 ) calculated from a sigmoidal Hill equation [23,24] observed at 47.3 nM (Fig. 1f). Further, we examined whether CCK-8 would alter the biophysical properties of I A . While no significant changes were observed in the activation properties of I A (V half from 5.8 ± 1.6 mV to 6.5 ± 2.5 mV, n = 9, Figs. 1h and i), CCK-8 shifted the steady-state inactivation curve to the hyperpolarized level by 7.8 mV (V half from − 55.9 ± 3.9 mV to − 63.7 ± 2.8 mV, n = 12, Figs. 1g and i). These findings reveal that the CCK-8-induced reduction in I A is mainly contributed by an increased proportion of channels retained in the inactivated state.

The CCK-BR mediated the CCK-8-induced I A decrease
The CCK-AR and CCK-BR have been identified as the endogenous receptors for CCK-8 [27]. To determine which one is involved in the CCK-8-induced I A reduction, we first examined the protein profile and subcellular expression of these receptors in mouse DRGs. Immunoblot analysis revealed that only CCK-BR (predicted size of 80 kDa), but not CCK-AR (predicted size of 95 kDa), were endogenously expressed (Fig. 2a). Protein lysates prepared from the gallbladder of the same mice were used as a positive control. Small unmyelinated sensory neurons have been classified into isolectin B4 (IB 4 )-positive (non-peptidergic) subset and peptidergic (IB 4 negative) subset expressing calcitonin gene-related peptide (CGRP), while large neurons in myelinated A-fibers express neurofilament 200 (NF200). We analyzed the CCK-BR expression in DRGs subsets by coimmunostaining of CCK-BR with the mentioned markers. CCK-BRs were found to be heavily colocalized with IB 4 and CGRP, and less with neurofilament-200 (NF-200), a marker for myelinated A-fibers (Fig. 2b). Next, we determined the participation of CCK-BR in the effect of CCK-8 on I A . While the CCK-8 mediated reduction of currents was not affected by the presence of 1 μM of the CCK-AR antagonist devazepide (decreased by 29 , where PD max is the maximal percent decrease of peak I A , EC 50 is the concentration that produces half-maximum effect occurs and n is the Hill coefficient. Cell numbers at each concentration were shown in round brackets. g, h, CCK-8 did not significantly alter the steady-state activation curve of I A (n = 9, g), but shifted the steady-state inactivation curve of I A leftward (n = 12, f). i, summary data showing the effects of 100 nM CCK-8 on V half of the activation and inactivation curves. Voltage-dependent activation was measured with voltage commands ranging from − 70 to + 70 mV (400 ms, in 10 mV increment). Steady-state voltage-dependent inactivation of I A was determined by varying a 150-ms conditioning prepulse from − 120 to + 20 mV followed by a 500-ms voltage step pulse to + 40 mV. *p < 0.05, **p < 0.01 and ***p < 0.001 vs. control application of 0.1 μM BC-264, a selective CCK-BR agonist, significantly decreased I A by 30.9 ± 2.7% (n = 7, Fig. 2f), while the selective CCK-AR agonist GW5823 (5 μM) elicited no such effect (decreased by 3.1 ± 1.2%, n = 9, Fig. 2f). Since it is known that both CCK-4 and CCK-8 are active forms of CCKs in the nervous system [28,29], we also test whether application of another selective CCK-BR agonist CCK-4 affects I A in small-sized DRG neurons. Indeed, CCK-4 at 300 nM significantly decreased I A by 32.2 ± 4.9% (n = 7). These findings further support the conclusion that the CCK agonist driven I A decrease was mediated specifically by CCK-BR in small-sized DRG neurons.
The CCK-BR-mediated I A decrease requires the βγ subunits (G βγ ) of go-protein CCK-BR coupled to heterotrimeric G-proteins, which are key transducers to control numerous cellular processes [3]. We next examined actions of different subtypes of Gproteins in the CCK-BR-mediated I A modulation. Inactivation of G s by pre-treating DRG neurons with cholera toxin (CTX, 500 ng/ml) had no significant effects on the CCK-8-induced I A decrease (decreased by 32.3 ± 3.6%, n = 11, Fig. 3a and c). Contrastingly, pre-treating cells with pertussis toxin (PTX, 200 ng/ml) to inactivate G i/o abrogated the CCK-8-induced response (decreased by 3.9 ± 2.2%, n = 7, Fig. 3b and c). This CCK-BR-induced PTXsensitive, but CTX-insensitive decrease in I A , indicated the involvement of G i/o , but not G s in the signaling cascade. Further, dialysis of cells with an antibody specifically against Gα o (2 μg/ml) blocked the effect of CCK-8 on I A reduction (decreased by 1.5 ± 2.9%, n = 8, Fig. 3d), whereas a Gα i -specific antibody (2 μg/ml) had no such effect (decreased by 27.6 ± 4.7%, n = 9, Fig. 3d). Together, these findings suggest that Gα o -protein mediates the response to CCK-8. Moreover, we found that endogenous Gα o (Fig. 3e), but not Gα i (Fig. 3f), was co-immunoprecipitated with an antibody against CCK-BR from DRG tissues, indicating that the CCK-BR and the Gα o subunit form a complex in situ. Further, intracellular application of a G β -specific antibody abrogated the CCK-8-induced I A reduction (decreased by 3.3 ± 3.5%, n = 9; Figs. 3g and h), while its denatured form did not elicit such effects (decreased by 32.5 ± 1.3%, n = 8; Fig. 3h). Similar results were obtained with a specific G βγ inhibitor, gallein. Pretreatment of cells with gallein (10 μM) completely abolished the CCK-8-induced I A reduction (decreased by 2.7 ± 0.9%, n = 9; Fig. 3h). Thus, the G βγ subunit of G o -protein is also required for the CCK-BRmediated I A reductions.
The CCK-BR-mediated I A decrease requires PI3K, but independently of Akt In view of the fact that protein kinase C (PKC) has been shown to act downstream of G βγ [30] and regulate I A superficial dorsal horn neurons [31], we pre-incubated cells with PKC inhibitors and found that pretreating with GF109203X (1 μM) did not affect the CCK-8induced I A response (decreased by 31.8 ± 4.3%, n = 9, Fig. 4a) while pre-incubation of cells with GF109203X substantially blocked the PKC activator PMA (phorbol 12-myristate 13-acetate)-induced I A reduction (5 μM, n = 9, Fig. 4b). Previous studies have highlighted the critical role of PI3K/Akt cascades in G βγ -mediated responses [32]. Thus, we investigated whether the inhibitory effect of CCK-8 on I A was PI3K/Akt-dependent. We found that CCK-8 application significantly induced PI3K activation and that pretreating cells with the PI3K inhibitor LY294002 at 20 μM abolished this effect (Fig. 4c). Consistently, pre-treatment with the PI3K inhibitor LY294002 (20 μM) (decreased by 1.9 ± 2.7%, n = 8, Fig. 4d  mediated I A decrease. Further, we examined whether the CCK-8 action is also mediated by Akt, a major downstream target of the PI3Ks [33]. We measured the Akt activity in DRG cells and found that 100 nM CCK-8 significantly increased the phosphorylated Akt (p-Akt) level, while the total Akt (t-Akt) remained unchanged (Fig. 4g). This effect was abrogated by the Akt inhibitor III (10 μM, Fig. 4g). To further determine the involvement of Akt in the modulation of I A by CCK-8, cells were pretreated with Akt inhibitor III prior to CCK-8 application. Interestingly, in the presence of 10 μM Akt inhibitor III, CCK-8 at 100 nM still induced a significant decrease in I A (decreased by 28.5 ± 3.8%, n = 9, Fig. 4h), revealing that the CCK-8-induced I A decrease was mediated by PI3K, but independently of Akt.

Activation of CCK-BRs induces DRG neuronal hyperexcitability
Kv exerts pivotal effects in modulating neuronal excitability in peripheral sensory neurons [38]. To determine the functional roles of the CCK-BR-mediated I A response, we determined whether the membrane excitability of DRG neurons would be affected by CCK-8. Bath application of CCK-8 (100 nM) had no significant effects on the whole-cell currents of Nav (n = 11, Fig. 6a) and the high voltage-activated (HVA) calcium channel currents (n = 8, Fig. 6b) in small DRG neurons, whereas CCK-8 increased LVA (T-type) channel currents by 8.3% (n = 9, Fig. 6c). Using an external solution including Z941 (10 μM) to block T-type channels, we found that 100 nM CCK-8 significantly increased action potential (AP) firing in response to 1-s current injection (by 56.6 ± 2.9% compared to control, n = 17, Fig. 6d and e). After washout, the firing rate was partially restored (Fig. 6e). Additionally, CCK-8 (100 nM) significantly shortened the first spike latency (Fig. 6f) and decreased AP threshold (n = 17, Fig. 6g). The other membrane properties of neuronal excitability, including resting membrane potential, were not significantly changed by 100 nM CCK-8 (not shown). Pretreating neurons with LY225910 (1 μM) abrogated the CCK-8-induced increase in AP firing rate, indicating the CCK-BR involvement (n = 12, Fig. 6h). To further verify the CCK-BR-induced neuronal hyperexcitability through I A decrease, 4-AP was applied prior to CCK-8. Pre-treatment of DRG neurons with 5 mM 4-AP abrogated the neuronal hyperexcitability induced by 100 nM CCK-8 (n = 12, Fig. 6j and k), indicating that the CCK-BR-mediated I A decrease subsequently induced neuronal hyperexcitability in small DRG neurons.

Involvement of A-type channels encoding I A in CCKinduced pain hypersensitivity
Further, we determined whether CCK-8 would affect in vivo pain sensitivity in animals. Intraplantar injection of CCK-8 (50 ng) markedly increased pain sensitivity to both mechanical and heating stimuli (Figs. 7a and b).
The CCK-8-induced pain hypersensitivity to mechanical or heating stimulation was completely abrogated by intraplantar pretreatment of the CCK-BR antagonist LY225910 (0.5 μg, Figs. 7c and d), but not by the CCK-AR antagonist devazepide (1 μg, Figs. 7c and d). Moreover, intraplantar pretreatment with 4-AP (25 nmol) induced a significant increase in mechanical and heat sensitivity as compared with animals received a saline injection (Figs. 7e and f ). Sensitivity assessed after intraplantar injection of CCK-8 showed that CCK-8 did not induce any additive effects to that of 4-AP on mechanical (Fig. 7e) and thermal (Fig. 7f ) pain sensitivity, strongly suggesting that CCK-8 and 4-AP likely target molecules in the same cellular signaling pathway in vivo.
Collectively, these findings reveal that A-type channels encoding I A contribute to the CCK-BR-mediated acute pain hypersensitivity.

Discussion
The present study provides mechanistic data describing a novel functional role of CCK-8 in modulating transient I A in small-sized DRG neurons, without any concurrent effect on I DR . Based on our findings, we propose a signaling cascade model in which CCK-8-stimulated PI3K recruits the Src-dependent JNK to suppress I A . This attenuation of I A induced by CCK-8 application is mediated by the stimulation of CCK-BR and leads to sensory neuronal hyperexcitability and pain hypersensitivity in mice (see Fig. 8).
The PKC family of isozymes mediates I A responses in a cell-type and tissue-specific manner. For instance, activation of group I metabotropic glutamate receptors led to an inhibition of I A through a PKC-dependent mechanism in striatal cholinergic interneurons, while in large aspiny neurons activation of PKCα increases I A . Interestingly, in murine proximal colonic myocytes, the PKCindependent regulation of I A has also been reported [39]. In this study, the CCK-8-induced I A decrease was independent of PKC and was mediated by PI3K, through the JNK-dependent signaling. These results are supported by previous studies that Kv currents including I A recorded from trigeminal ganglion neurons and pancreatic β cells decreased in response to PI3K pathway activation [21,40]. Interestingly, the activation of PI3K has also been reported to increase I A in cultured rat cerebellar granule cells [41]. In addition, PI3K-induced activation of Kv4.3 channels through glucocorticoid-inducible kinase-1 (SGK1) was also reported [42]. Although these discrepancies require further clarification, the regulatory effects of PI3K would be variable in tissues/cell types expressing different A-type channel subunits. Another appropriate alternative hypothesis is that the stimulatory PI3K can also phosphorylate an intermediate protein that in turn down-regulates I A in small-sized DRG neurons. Furthermore, different splice variants of KChIP auxiliary subunits of I A channels can engender different, even opposing, modulation of Kv4 channel currents [43].
A known target of G βγ is PI3K [33]. In contrast to many other common G βγ -dependent PI3K signal transduction events, the CCK-8-induced PI3K dependent I A attenuation isnot mediated by Akt, as demonstrated by the specific inhibition of Akt with pharmacological agents. Interestingly, previous studies have shown that Akt both negatively and positively regulates Kv4 [44,45], which forms one of the major components mediating I A . For example, Akt down-regulates the activity of Kv4 channels in cultured cerebellar granule cells of rats [45]; in the same neurons, a different study demonstrates that enhanced Akt activity is required for I A amplification and Kv4.2 induction [44]. This Akt-dependent stimulation of Kv4 also occurs in the arcuate nucleus [46]. Thus, it appears that Akt differentially regulates the activity of Kv4 channels in a tissue-specific manner. In our study, the CCK-BR-mediated I A response was found to be independent of Akt; therefore, we went on to investigate what mediating PI3K signals to suppress I A in DRG neurons. Considerable in vivo and in vitro studies indicate that ERK plays pivotal roles in neuropathic pain Representative traces were recorded when smallsized DRG neurons were subjected to 130 pA current injections. j, k exemplary current traces (j) and summary of results (k) indicating that application of 4-AP at 5 mM abrogated the 100 nM CCK-8-induced neuronal hyperexcitability (n = 12). Representative traces were recorded when small-sized DRG neurons were subjected to 80 pA current injections. *p < 0.05 and **p < 0.01 vs. control [34,47]. Phosphorylated ERK is elevated in DRG cells following peripheral nerve injury [48]. Intrathecal application of ERK inhibitors reduces the pain behavior associated with nerve injury [49]. Moreover, one of the most convincing evidence comes from the direct phosphorylation of the pore-forming channel subunit of Kv4.2 by ERK [50] that determine a downregulation of I A in superficial dorsal horn neurons [31]. Contrastingly, antagonism of ERK completely abrogated I A increase induced by dopamine in lateral pyloric neurons [51]. However, we found that the CCK-8-induced decrease of I A was unlikely induced by ERK phosphorylation, because the CCK-8 application did not change the ERK activity in DRG cells, whereas the levels of p-JNK were significantly increased. Moreover, the MAPK/ERK inhibitor did not affect the CCK-8-induced I A response. Our findings suggested that PI3K stimulated JNK in DRG neurons and that this signaling is essential for the CCK-BR-mediated I A response. Our results showed that 1) application of the JNK inhibitor SP600125, but not the p38 MAPK inhibitor SB203580 or the MAPK/ERK (MEK) inhibitor U016, blocked the CCK-8-induced I A decrease and 2) antagonism of PI3K blocked CCK-BRmediated JNK activation. Consistent with these findings, the increased activity of JNK in ventricular myocytes markedly decreased the amplitude of transient outward K + current density [52]. These observations are in line with an earlier study showing a C-reactive protein (CRP)-induced modulation of intracellular JNK and interactions with voltage-activated K + channels [53].
Up till now, it is still relatively unclear how PI3K activates JNK. It has been established that that PI3K may stimulate PKA, subsequently activating the downstream MAPK pathway [35]. In the present study, activation of CCK-BR did not influence the PKA activity in DRG cells, indicating some other mechanisms, but not of PKA, mediate the crosstalk between PI3K and JNK signaling. Src kinases are downstream of PI3K and can facilitate JNK activity [37], suggesting possible crosstalk between PI3K and JNK signaling. In support of this observation, the current study demonstrated that the Src kinase inhibitor PP2 blocked the CCK-8-induced JNK activation. The blockade of PI3K also abolished the CCK-BR-mediated increase in Src activity, indicating that PI3K may modulate the JNK pathway through c-Src. Therefore, it is likely that CCK-8-activated PI3K recruits Src to up-regulate JNK activity, and thereby regulating CCK-BR-mediated I A response in DRG neurons. I A , encoded by A-type K + channels, is important determinants of both the delay of spike onset (first spike latency) and the decrease in the firing frequency [17]. Acute decreases in I A in sensory neurons cause robust increases in neuronal excitability, [38] and may increase the responsiveness to nociceptive stimulation and contribute to mechanical hypersensitivity and thermal hyperalgesia [54]. Genetic studies have firmly established a prominent role for A-type channels in amplifying nociceptive signals in the periphery and in contributing to central sensitization in the spinal dorsal horn [17,19,55]. Further, recent evidence has suggested that modulation of peripheral A-type channels influences somatic and visceral nociceptive inputs and thus an increase of A-type channel currents results in significant anti-nociception in a variety of animal neuropathic pain models [19]. In the current study, consistently with the CCK-8-induced I A decrease, activation of CCK-BR led to increased excitability in DRG neurons with increased spike frequency and shortened first-spike latency, both of which are major parameters determining the timing of neurotransmitter release, and hence pain transmission [56]. In addition, acute mechanical hypersensitivity and thermal hyperalgesia mediated by CCK-BR can be occluded by the A-type K + channel blockade. As such, our findings are supportive of the reasonable assumption that nociceptive actions of CCK-BR are mediated, at least in part, through the JNK-dependent reduction of I A . Our present results are, indeed, in accordance with previous studies that CCK-8 might induce pro-nociceptive actions [8,9]. Intrathecal inhibition of JNK, a key modulator I A in the present study, has been found to attenuate the CCI-induced mechanical allodynia and thermal hyperalgesia in rats [57]. Following spinal nerve ligation (SNL), phosphorylated JNK in smallsized DRG neurons have been found to be greatly increased [58] and the intrathecal infusion of JNK inhibitor can reverse mechanical but not thermal hypersensitivity [59]. Fig. 8 Schematic shows the regulation of CCK-BR on I A and the involvement of CCK-8/CCK-BR in pain sensitivity. CCK-8 stimulates the G o -protein coupled CCK-BR and thereafter releases the βγ subunits (G βγ ). The released G βγ subsequently activates PI3K, which decreases the I A and induces neuronal hyperexcitability and pain hypersensitivity. PI3K catalyzes the conversion of PtdIns(4,5)P2 (PIP2) to PtdIns(3,4,5)P3 (PIP3), which serves as a second messenger that helps to activate Akt. However, neither PKA/PKC/Akt nor the direct binding of G βγ with A-type channels contributes to the CCK-BR-mediated I A response. In mouse DRG neurons, PI3K signaling may activate Src, which then phosphorylates JNK to modulate I A . Whether the activated p-JNK would phosphorylates Kv channels encoding I A or in turn stimulated intermediate molecules still needs further examined