Phospholipase C-related but catalytically inactive protein modulates pain behavior in a neuropathic pain model in mice

Background An inositol 1,4,5-trisphosphate binding protein, comprising 2 isoforms termed PRIP-1 and PRIP-2, was identified as a novel modulator for GABAA receptor trafficking. It has been reported that naive PRIP-1 knockout mice have hyperalgesic responses. Findings To determine the involvement of PRIP in pain sensation, a hind paw withdrawal test was performed before and after partial sciatic nerve ligation (PSNL) in PRIP-1 and PRIP-2 double knockout (DKO) mice. We found that naive DKO mice exhibited normal pain sensitivity. However, DKO mice that underwent PSNL surgery showed increased ipsilateral paw withdrawal threshold. To further investigate the inverse phenotype in PRIP-1 KO and DKO mice, we produced mice with specific siRNA-mediated knockdown of PRIPs in the spinal cord. Consistent with the phenotypes of KO mice, PRIP-1 knockdown mice showed allodynia, while PRIP double knockdown (DKD) mice with PSNL showed decreased pain-related behavior. This indicates that reduced expression of both PRIPs in the spinal cord induces resistance towards a painful sensation. GABAA receptor subunit expression pattern was similar between PRIP-1 KO and DKO spinal cord, while expression of K+-Cl--cotransporter-2 (KCC2), which controls the balance of neuronal excitation and inhibition, was significantly upregulated in DKO mice. Furthermore, in the DKD PSNL model, an inhibitor-induced KCC2 inhibition exhibited an altered phenotype from painless to painful sensations. Conclusions Suppressed expression of PRIPs induces an elevated expression of KCC2 in the spinal cord, resulting in inhibition of nociception and amelioration of neuropathic pain in DKO mice.

Naive PRIP-1 knockout (KO) mice demonstrate a marked decrease in the withdrawal threshold in the von Frey hair test because of altered expression of GABA A receptor subunit in their central nervous system [8]. In the present study, we investigated the role of PRIP-1 and PRIP-2 in pain sensation using PRIP-1 and PRIP-2 double knockout (DKO) mice, and PRIP-1 and/or PRIP-2 knockdown (KD) mice.

Materials and methods Animals
Ten-to fourteen-week-old male PRIP-1 KO [5,8] and DKO [7,9] mice, in a C57BL/6J mouse background, and ddY mice were used. All procedures and handling of animals were performed with permission according to the guidelines of Hiroshima University.

Seltzer model and paw withdrawal threshold test
Partial sciatic nerve ligation (PSNL) was performed according to the procedure described by Seltzer et al. [10]. A paw withdrawal threshold in response to probing with von Frey hair (gram weight to buckling) was measured.

Statistical analyses
The density of each band was analyzed using NIH ImageJ software, and the densitometric units were corrected for tubulin. The data were expressed as the mean ± S.E.M. Statistical analyses are described in the figure legends.

Results and discussion
To examine pain-related behavior in DKO mice, PSNL was performed, and the withdrawal threshold of the hind paw was measured by applying von Frey filaments. Naive DKO mice had normal sensation levels in terms of withdrawal threshold ( Figure 1A). This differed greatly from the significant reduction in the withdrawal threshold observed in PRIP-1 KO mice [8]. After PSNL, the withdrawal threshold in the contralateral hind paw of DKO mice was not significantly different from presurgical baselines Table 1 The sequences used in siRNA knockdown methods ( Figure 1A and B). The significant reduction of the withdrawal threshold of wild-type (WT) ipsilateral hind paw was dramatically ameliorated in the DKO mice ( Figure 1B), suggesting that DKO mice exhibit a neuropathic painresistant phenotype. Since PRIP expression in WT mice was similar to that in the PSNL and sham-operated mice ( Figure 1C), the onset of neuropathic pain was not induced by the change of PRIP expression. To better understand the involvement of PRIP in nociceptive signaling, we produced the spinal cord-specific PRIP-1 knockdown (PRIP-1 KD), PRIP-2 knockdown (PRIP-2 KD), and PRIP-1 and PRIP-2 double knockdown (DKD) mice by using molecular specific siRNAs (Table 1) in the ddY mouse strain. We reported that a peak of gene suppression following intrathecal injection of a siRNA occurs at 2-3 days postinjection, and this recovers to original levels approximately 8 days after injection [14]. The significantly reduced expression of PRIP-1 in PRIP-1 KD and DKD mice, or of PRIP-2 in PRIP-2 KD and DKD mice, was observed 3 days after the siRNA injection (Figure 2A and B). We then examined mechanical allodynia by using the von Frey hair test in animals 3 days after siRNA injection. Allodynia was observed in the PRIP-1 KD mice, but not in the other mice ( Figure 2C), indicating that the PRIP-1 KD mice mimicked the phenotypes of pain sensitivity observed in PRIP-1 KO mice.
Next, we observed the influence of suppression of the PRIP gene on pain sensation by using a PSNL model. PSNL was performed on ddY mice 10 days before intrathecal siRNA injection, after which an allodynia score of contralateral and ipsilateral sides was analyzed during the 8 days after the injection of PRIP siRNA. In the contralateral paw, PRIP-1 KD mice showed an allodynia in accordance with PRIP-1 fluctuation 2-5 days after siRNA injection (initial score, 1.17 ± 0.09 at day 0; peak score, 0.46 ± 0.117 at day 2; and recovered score, 1.22 ± 0.08 at day 8) ( Figure 2D). The PRIP-1 protein expression was analyzed by immunoblotting (data not shown). The allodynia observed in PRIP-1 KD mice was not seen in PRIP-2 KD, DKD, and other control mice ( Figure 2D). However, the withdrawal threshold for the ipsilateral paw was dramatically increased in DKD mice (initial score, 0.104 ± 0.02 at day 0 and peak score, 0.628 ± 0.068 at day 3), but not other experimental mice, including PRIP-1 KD and PRIP-2 KD mice; the relief gradually reverted to painful levels within 7 days ( Figure 2E). This suggested that suppression of both PRIP genes, but not either, induces resistance for pain sensation associated with allodynia.
Neuropathic pain in a model animal induces an altered expression of GABA A receptors, including the downregulation of γ2 subunit-containing receptors [15,16]. PRIP is a modulator for GABA A receptor intracellular trafficking [7,9,17]. The β 2/3 subunit is upregulated, and the γ2 subunit is downregulated in the spinal cord of PRIP-1 KO or DKO mice [8,18]. Therefore, we examined the expression levels of GABA A receptor subunits by immunoblotting using commercially available subunitspecific antibodies. The examined expressions were similar between the genotypes, with the exception of α5 expression, which was increased in DKO mice ( Figure 3A). Knabel et al. reported that α2 and α3 contribute to diazepam-induced antihyperalgesia actions, but that α1 and α5 subunits do not [19], suggesting less involvement of spinal α5 subunit-containing GABA A receptors in nociception [16]. Therefore, the different pain sensation between PRIP-1 KO and DKO mice is probably not due to the alteration of GABA A receptor expression in the spinal cord.
Inhibitory signaling is regulated by the intracellular chloride ion concentration, which is established in part via KCC2. Therefore, a high level of KCC2 expression drives chloride extrusion from neurons and maintains a low intracellular chloride ion concentration, i.e., GABAergic input may even acquire a net cell inhibitory response [20]. We next examined KCC2 expression by immunoblotting. The expression in PRIP-1 KO spinal cord was similar to that in WT; however, the expression in DKO mice was significantly increased compared with WT and PRIP-1 KO mice ( Figure 3B). We then investigated the influence of PSNL surgery on expression of NKCC1 and KCC2, both of which are required for maintaining a fine balance between chloride ion influx and efflux, respectively ( Figure 3C). KCC2 expression was higher in DKO sham-operated mice than in corresponding WT mice. PSNL surgery induced decreased KCC2 expressions in WT and DKO mice compared with sham-operated mice. Despite the decrease, KCC2 levels in DKO PSNL mice were similar to those in WT sham-operated mice. On the other hand, expression levels of NKCC1 were similar, and PSNL surgery did not affect the expressions ( Figure 3C). We also examined expression of GlyR, contributing as a chloride ion channel dominantly expressing in the spinal cord, whose activation is known to ameliorates neuropathic pain [14]. Expressions of GlyRα1, a most prevalent subunit in central nervous system [21], were similar in the genotypes and were not changed by PSNL operation ( Figure 3C). These data suggested that PRIP deficiency affects the expression of KCC2 at basal and after PSNL surgery. Since PRIP-2 KO mice are not currently available, we were unable to define changes of GABA A receptor, KCC2, NKCC2, and GlyRα1 expressions as a result of PRIP-2 KO alone.
In immature neurons, a low expression of KCC2 results in a physiologically high concentration of intracellular chloride ions, which leads to the depolarization of cells [22]. Similarly, when PSNL was performed in mice, KCC2 expression was decreased in the spinal cord, resulting in a high concentration of intracellular chloride ion and reduced nociceptive threshold [23]. In addition, upregulation of KCC2 induced inhibitory postsynaptic potentials [24], suggesting that high expression level of KCC2 observed with DKO spinal cord enhances inhibitory synaptic transmission. We then tested if inactivation of KCC2 by R-(+)-[(dihydroindenyl)oxy] alkanoic acid (R-DIOA), an inhibitor of KCC2, affects pain sensitivity. The paw withdrawal threshold was decreased dose-dependently by intrathecal administration of R-DIOA ( Figure 3D), indicating the importance of KCC2 activity. Therefore, similar KCC2 expression in the spinal cord of naive WT and DKO PSNL mice ( Figure 3C) may cause the allodyniaresistant phenotype observed in the DKO PSNL model ( Figure 1B).  To further confirm the involvement of KCC2 in neuropathic pain regulated by PRIP, we performed a hind paw withdrawal test by using WT, PRIP-1 and PRIP-2 KD, and DKD mice with R-DIOA. Withdrawal thresholds were not changed in PSNL-operated WT and a single gene KD mice, and R-DIOA administration did not affect the pain threshold ( Figure 3E). However, relief from pain (as indicated by the increase of the threshold) in the DKD PSNL model was significantly inhibited by the administration of R-DIOA 3 days postinjection ( Figure 3E).

Conclusions
We demonstrated that the regular expression of KCC2 in DKO mice even after PSNL surgery induces the  inhibition of nociceptive transmission and ameliorates PSNL-mediated neuropathic pain, even though the alteration of GABA A receptor subunits in PRIP-1 KO mice causes allodynia [8]. The current findings led us to hypothesize that regulation of KCC2 expression is a critical modulator of pain sensation.

Competing interests
The authors declare that they have no competing interests.
Authors' contributions TK carried out the paw withdrawal threshold test, immunoblotting, and data analyses. KM performed PSNL surgery and intrathecal siRNA injection. SR, NK, and KM participated in the data analyses. SU and MH provided the knockout mice and participated in the design of the study. TK conceived of the study, participated in its design and coordination of the experiments, and wrote the manuscript. All authors read and approved the final manuscript.