The voltage-gated sodium channel NaV1.7 underlies endometriosis-associated chronic pelvic pain

Chronic pelvic pain (CPP) is the primary symptom of endometriosis patients, but adequate treatments are lacking. Modulation of ion channels expressed by sensory nerves innervating the viscera have shown promise for the treatment of irritable bowel syndrome and overactive bladder. However, similar therapies have not been explored for endometriosis-associated CPP. Here we examined the role of the voltage-gated sodium (NaV) channel NaV1.7 in the sensitivity of vagina-innervating sensory afferents and investigated whether NaV1.7 inhibition reduces nociceptive signals from the vagina and ameliorates endometriosis-associated CPP. The mechanical responsiveness of vagina-innervating sensory afferents was assessed with ex vivo single unit recording preparations. Pain evoked by vaginal distension (VD) was quantified by the visceromotor response (VMR) in vivo. In control mice, pharmacological activation of NaV1.7 with OD1 sensitised vagina-innervating pelvic afferents to mechanical stimuli. Using a syngeneic mouse model of endometriosis, we established that endometriosis sensitized vagina-innervating pelvic afferents to mechanical stimuli. The highly selective NaV1.7 inhibitor Tsp1a revealed that this afferent hypersensitivity occurred in a NaV1.7-dependent manner. Moreover, in vivo intra-vaginal treatment with Tsp1a reduced the exaggerated VMRs to VD that is characteristic of mice with endometriosis. Conversely, Tsp1a did not alter ex vivo afferent mechanosensitivity or in vivo VMRs to VD in Sham control mice. Collectively, these findings suggest that NaV1.7 plays a crucial role in endometriosis-induced vaginal hyperalgesia. Importantly, NaV1.7 inhibition selectively alleviated endometriosis-associated CPP without the loss of normal sensation, suggesting that selective targeting of NaV1.7 could improve the quality of life of women with endometriosis.


Introduction
Endometriosis is a debilitating condition characterised by infertility and chronic pelvic pain (CPP) that affects approximately 11% of women worldwide [4; 25]. Vaginal hyperalgesia, also known as dyspareunia, or painful intercourse, is one of the most debilitating symptoms affecting women with endometriosis. Current treatments aimed at relieving the CPP associated with endometriosis include hormonal suppression of ovarian function or surgery. Unfortunately, both interventions have variable success and negatively impact fertility [61], reflecting a clear need for alternative treatment options.
Physiological and painful stimuli are detected by ion channels and receptors expressed within the terminals of sensory afferent nerve fibres projecting from the periphery to the central nervous system (CNS). Modulation of ion channels expressed within sensory neurons innervating the colon and the bladder has proved be a promising approach for treatment of chronic pain associated with irritable bowel syndrome (IBS) and overactive bladder syndrome (OAB) [9; 32; 40; 41; 51; 55; 56]. Remarkably, a similar approach directed to relieve endometriosis-associated CPP has yet to be described. This is partially due to the lack of knowledge of specific ion channels and receptors expressed within sensory neurons projecting to the female reproductive tract. In fact, only the transient receptor potential vanilloid 1 channel TRPV1 [20], purinergic receptor P2X3 [59], voltage-gated sodium (NaV) channels [16; 45], and the voltage-gated potassium channels (Kv) KV6.4 and KV2.1 [45], have been reported to be expressed in these sensory afferents. While only tetrodotoxin (TTX)-sensitive NaV channels have been reported to have a functional role in pain sensation from the female reproductive tract of healthy mice [16], the identity of the NaV channel(s) contributing to this pain signalling pathway, and whether they contribute to endometriosis-associated CPP, remains unknown.
In this study we used a highly selective inhibitor of NaV1.7 [41] to investigate the role of this channel in vaginal sensation and evoked pain. Additionally, we examined whether NaV1.7 contributes to the allodynia and hyperalgesia experienced by mice with endometriosis. We show that NaV1.7 contributes to endometriosis-associated vaginal hyperalgesia, which can be ameliorated by treatment with a Nav1.7 inhibitor. Our data suggest that inhibition of NaV1.7 may be a viable option to treat endometriosis-associated CPP.

Activation of NaV1.7 increases vaginal afferent sensitivity to mechanical stimuli
Little is known about how mechanical stimuli is sensed by female reproductive organs. We recently showed that the sensitivity of pelvic vaginal afferents to mechanical stimuli can be altered by nonspecific modulators of NaV channels [16]. This includes enhancing afferent responses to mechanical stimuli with veratridine and conversely inhibiting responses with TTX [16]. In this study, we dissected these mechanisms further by investigating the specific contribution of NaV1.7, which is expressed by 100% of vagina-innervating neurons [16].
Compared to baseline responses, application of the α-scorpion toxin OD1, a selective agonist of NaV1.7 [23], significantly increased pelvic vaginal afferent responses to vaginal stroking (Figure 1Ai-iii), focal compression (Figure 1Bi-ii), and circular stretch (Figure 1Ci-ii). In addition to an increased number of action potentials (AP) generated by circular stretch, OD1 significatively reduced the latency of the response (time for the first AP to be generated) to circular stretch (Figure 1Ciii). Interestingly, greater than 50 % of the vaginal afferents studied continued to fire APs after cessation of the mechanical stimuli ( Figure 1Ci). Overall, these results indicate that the mechanical responsiveness of vaginal afferents can be augmented by activation of NaV1.7 expressed within these fibres.  (Figure 2Ai-ii), focal compression (Figure 2Bi-ii), and circular stretch (Figure 2Ci-iii).   We then investigated whether Tsp1a could modulate pain sensitivity evoked by vaginal distension (VD) in conscious Sham control mice. In vivo vaginal pain sensitivity was monitored by measuring the VMR to increasing VD pressures using EMG electrodes surgically implanted into the abdominal muscles as previously described [16; 17; 48]. In Sham control mice, we found that VD evoked an increase in the VMR, with the degree of VMR related to the amount of pressure applied to the vagina ( Figure 3A-B). The in vivo sensitivity to pain evoked by VD in Sham control mice administered Tsp1a intravaginally was equal in magnitude to that displayed by mice treated with intravaginal saline (Figure   3Ci). This effect was observed across both non-noxious and noxious distension pressures (Figure 3Cii-iii). Collectively, these results suggest that while activation of NaV1.7 can enhance afferent hypersensitivity in control states, inhibition of NaV1.7 does not alter the capacity of vaginal afferents to sense non-noxious and noxious stimuli.

Vaginal afferents from mice with endometriosis display hypersensitivity to mechanical stimuli
We next investigated whether NaV1.7 plays a functional role in nociceptive signalling associated with endometriosis. We previously showed that, in a mouse model of IBS, the sensory afferents innervating the colon develop mechanical hypersensitivity. These hypersensitive colonic afferents send enhanced nociceptive signals to the CNS, ultimately causing increased pain sensitivity to colorectal distension in vivo [13; 15; 31; 33; 35]. Although similar neuroplasticity of sensory afferents has been suggested to occur in endometriosis [2; 8; 47], this has not been demonstrated experimentally.
Therefore, we investigated whether sensory afferents innervating the vagina of mice with endometriosis become hypersensitive to mechanical stimuli.
Using our ex vivo vaginal afferent recordings, we characterised the response of vaginainnervating afferents to three different mechanical stimuli, in both Sham control mice and mice with fully developed endometriosis (Endo) (Figure 4). Compared to Sham control mice, vaginal afferents from Endo mice fired significantly more APs in response to vaginal surface stroking and circular stretch ( Figure 4A and C). Moreover, AP firing evoked by circular stretch commenced more rapidly within afferents from Endo mice (Figure 4Civ). Overall, these data indicate that the development of endometriosis enhances the sensitivity of pelvic vaginal afferents to mechanical stimuli.

Inhibition of NaV1.7 reduces endometriosis-associated vaginal afferent hypersensitivity
We then examined whether selective inhibition of NaV1.7 with Tsp1a [41] was able to reverse endometriosis-associated vaginal afferent hypersensitivity. We found that exposure of vaginal afferent endings to Tsp1a significantly reduced vaginal afferent firing to vagina surface stroking (Figure 5Ai-iii), focal compression (Figure 5Bi-iii), and circular stretch (Figure 5Ci-iii) in Endo mice. Moreover, the time taken for these afferents to fire the first AP was significantly longer after incubation with Tsp1a ( Figure   5Civ). Overall, these data indicate that Na V 1.7 activity contributes to the enhanced sensitivity developed in vaginal afferents from Endo mice.

Inhibition of NaV1.7 reverses endometriosis-associated chronic pelvic pain to vaginal distension (VD)
We then investigated whether the observed alteration in sensory signalling from the vagina to the CNS translates to changes in pain sensitivity evoked by vaginal distension in conscious mice with endometriosis. We first confirmed our previous findings and demonstrated that Endo mice display enhanced pain sensitivity to VD [17]. Here we found that, compared to Sham control mice, Endo mice had elevated VMR responses to VD distension at all pressures tested (20-70 mmHg) (Figure 6A-C).
Further analysis revealed the VMR to non-noxious (20-30 mmHg) (Figure 6Cii) and noxious (40-70 mmHg) (Figure 6Ciii) Figure 7A-C). This effect was observed at all non-noxious and noxious VD pressures that we tested (Figure 7Cii-iii). Interestingly, Tsp1a reduced the VMRs in Endo mice to levels observed in Sham controls, suggesting that pharmacological inhibition of NaV1.7 on vaginal afferents is an effective strategy to reverse the vaginal allodynia and hyperalgesia associated with endometriosis.

Discussion
Endometriosis is a chronic disorder characterised by infertility and CPP. Currently there is a lack of effective analgesic treatments for endometriosis-associated CPP mainly due to our lack of knowledge about its aetiology and pathogenesis [43; 61]. How sensory neurons innervating the female reproductive tract detect and transmit pain and how they are altered in endometriosis remains poorly understood. We recently showed that all nine members of the NaV channel family are expressed in pelvic vaginal afferents, although their relative expression in these neurons varies greatly. The pan-NaV channel agonist, veratridine, can drive vaginal afferent hypersensitivity, whilst the pan-NaV channel blocker TTX reduces vaginal afferent sensitivity. These findings demonstrate that NaV channels play an important role in regulating vaginal sensation in healthy animals and point towards a key role of TTXsensitive NaV channels [16]. However, the specific identity of the TTX-sensitive NaV channels contributing to vaginal nociceptive signalling and their contribution to endometriosis-associated CPP has remained unclear.
First, we demonstrated that activation of NaV1.7 with OD1, a α-highly selective NaV1.7 agonist [23], dramatically enhanced the responsiveness of vaginal afferents from Sham control mice to three different types of mechanical stimuli. Interestingly, we observed that whilst OD1 failed to elicit AP discharge in the absence of mechanical stimuli (direct activation of afferents), half of the afferents incubated with OD1 continued to fire APs after cessation of the mechanical stimulus. This observation could be explained by the mechanism by which OD1 alters NaV1.7 channel function. Kinetically, NaV1.7 is slow to close and subsequently inactivates following its activation [34; 42]. OD1 inhibits the fast inactivation of Na V 1.7, which ultimately disrupts the resting threshold equilibrium, maintaining hyperexcitable sensory neurons in the absence of further stimuli [23; 42; 49]. This is particularly relevant in genetic gain-of-function mutations, where the enhanced role of NaV1.7 in sensory neurons maintains them in a hyperexcitable state, increasing AP firing and ultimately driving sensory signalling in the absence of stimuli [22; 42]. Overall, our findings indicate that vaginal afferent responses to mechanical stimuli in control conditions can be enhanced by pharmacological activation of NaV1.7.
We next examined whether selective inhibition of NaV1.7 could reduce the ability of vaginal afferents to sense baseline mechanical stimuli in Sham control mice. For this we used the peptide Tsp1a, a selective inhibitor of Nav1.7 derived from venom of the Peruvian tarantula Thrixopelma spec. [41]. We found that Tsp1a does not alter the ability of vaginal afferents to respond to mechanical stimuli ex vivo.
These findings were further supported by in vivo studies, whereby intravaginal application of Tsp1a failed to alter vaginal sensitivity to distension in conscious Sham control mice. Overall, our results are consistent with previous reports indicating that NaV1.7 does not appear to contribute to baseline visceral nociception in healthy conditions [36; 41].
However, a possible role for NaV1.7 in endometriosis-associated CPP has previously been suggested.
There is evidence of SCN9A (the gene encoding NaV1.7) overexpression in endometriosis lesions collected from women with pain, compared to women without endometriosis/without pain [30].
Furthermore, macrophages activated with peritoneal fluid from women with endometriosis can also enhance the expression of SCN9A in cultured human sensory neurons in vitro, via a macrophagederived insulin-like growth factor 1 (IGF-1) mechanism [27]. While changes in the expression profile of NaV1.7 have been associated with endometriosis in these studies, the functional role of NaV1.7 in sensing pain from pelvic organs affected by endometriosis has not yet been explored. Therefore, we investigated the role of NaV1.7 in our established syngeneic mouse model of endometriosis [48]. We show here, for the first time, that pelvic afferents innervating the mouse vagina become hypersensitive to mechanical stimuli in endometriosis. This hypersensitivity was also displayed in vivo, with conscious Endo mice displaying both allodynia and hyperalgesia to vaginal distension compared to their Sham control counterparts. These findings demonstrate that increased nociceptive signalling from the vagina results in increased pain sensitivity in vivo, which closely corresponds with nociceptive mechanisms observed in other visceral organs [13; 15; 31; 33; 35].
Importantly, inhibition of NaV1.7 with Tsp1a was effective in reducing ex vivo vaginal afferent hypersensitivity in mice with endometriosis. This attenuation was also apparent in vivo; the VMR evoked by vaginal distension in conscious Endo mice was reversed with a single intravaginal dose of Tsp1a, with sensitivity restored to the levels seen in Sham control animals. Notably, this effect was apparent at non-noxious and noxious distension pressures, suggesting that Tsp1a is able alleviate vaginal allodynia and hyperalgesia in mice with endometriosis. Overall, our findings are consistent with the notion that enhanced activity of NaV1.7 is able to drive, and contribute to, heightened pain and nociception in chronic disease states [34; 41; 57]. However, NaV1.7 does not appear to contribute to visceral nociception in healthy conditions [36; 41]. This is an important concept, because an ideal analgesic should target the key mechanisms responsible for the pathological pain underlying the condition and provide pain relief without the loss of baseline sensory functions, which are essential for host responses to the external and internal environments.
Although NaV1.7 is a widely accepted as pain target, the translation to clinical significance as a treatment for chronic pain has been lacking [24]. This is likely due to a multiplicity of factors including: (i) genetic mutations impacting pain sensitivity in humans; (ii) the lack of selectivity and tissue penetration of some NaV1.7 inhibitors, and (iii) the underestimated impact of the effect of the auxiliary β subunits on NaV1.7 pharmacology [24]. Our findings showing different roles for NaV1.7 in nociceptive signalling in health and endometriosis states are in keeping with recent studies. For example, inhibition of NaV1.7 does not alter colonic afferent signalling in healthy states [36; 41], whereas NaV1.7 inhibition in a mouse model of IBS reverses mechanical hypersensitivity ex vivo and in vivo [41]. Other studies show that indirect targeting of NaV1.7 activity in sensory neurons by specifically inhibiting NaV1.7 trafficking and surface expression with compound 194, is effective at reversing thermal and mechanical hypersensitivity in a rodent model of neuropathic pain [28; 46]. Overall, the role of NaV1.7 in pain is complex, with somatosensory studies suggesting that NaV1.7 blockers alone may not replicate the analgesic phenotype of NaV1.7 -/mutant mice and humans with congenital insensitivity to pain due to loss-of-function mutations in SCN9A. However, their effects may be potentiated with exogenous opioids [52] and G-protein coupled receptors (GPCRs) [39; 53]. For example, some studies have shown that NaV1.7 regulates the efficacy and balance of GPCR-mediated pro-and anti-nociceptive intracellular signalling, so that without NaV1.7 the balance of these processes is shifted toward anti-nociception [39].
Other animal studies show that post-surgical pain can be successfully treated with NaV1.7 inhibitors alone or at subtherapeutic doses in combination with baclofen or opioids, suggesting super-additive antinociceptive effects when NaV1.7 inhibitors are applied in combination with baclofen or opioids [53].
Whether opioidergic and GPCR mechanisms contribute to NaV1.7's role in endometriosis-associated CPP will be the subject of future studies. However, it is interesting to note that several GPCR and opioidergic mechanisms are upregulated in colonic nociceptors [10-12; 14; 21; 37; 54; 55] in IBS models and that inhibition of NaV1.7 also reverses pathological pain in these IBS models [41].
In conclusion, our study demonstrates that NaV1.7 contributes to the vaginal hyperalgesia associated with endometriosis. Moreover, it reveals that pharmacological inhibition NaV1.7 with Tsp1a could be a viable therapeutic option to treat endometriosis-associated CPP. Together with our recently published findings showing that Tsp1a could reverse colonic hypersensitivity in a mouse model of IBS [41], these results suggest that pharmacological targeting of NaV1.7 is a valid approach to treat CPP associated with visceral pain disorders including endometriosis.

Animals
All experiments involving animals were approved by the Animal Ethics Committee of the South Female mice were group housed in IVC cages and the littermate male mice were separated at weaning.
All female mice use in this study were unmated virgins.

Syngeneic inoculation mouse model of endometriosis
In this study we used a syngeneic mouse model of endometriosis, previously established and characterised by our group [48]. Briefly, female mice were ovariectomised under isoflurane and received estrogen (100 μg/kg estradiol benzoate) intraperitoneally (IP). Following ovariectomy recovery (minimum of 5 days), endometriosis (Endo) or Sham inductions were performed as previously described [48]. For this, recipient mice were anesthetized under isoflurane, and a small (0.5 cm) incision was made into the peritoneal space just below the umbilicus. 250 μL of minced endometrial tissue suspension obtained from donor mice (in PBS plus penicillin: 100 U/ml and streptomycin: 100 μg/ml) was then inoculated into the peritoneal cavity using a 1-ml pipette, as described previously [48]. A ratio of 1 donor mouse per 2 recipient mice was used. The incision was then closed using 6.0 Prolene sutures, and a gentle massaging of the abdominal cavity was performed to help disperse the inoculated endometrial fragments. Sham surgeries were performed by inoculating 250 μL of sterile PBS plus penicillin (100 U/ml) and streptomycin (100 μg/ml) in the absence of any tissue. All recipient mice received a low dose (0.05 mg/kg) of analgesic buprenorphine prior to the commencement of surgery.
Throughout the surgery and during the recovery period, animals were kept on a heating pad to maintain body temperature and monitored daily (for 5 consecutive days) for postsurgical complications. To maintain steady levels of circulating estrogen and minimize any difference related to the stage of the oestrous cycle, both Endo mice and Sham control mice were given IP injections of estrogen (100 μg/kg estradiol benzoate) immediately after surgery. All mice continued to receive estrogen IP once a week until full development of endometriosis (up to 8-10 weeks).

Ex vivo afferent recording from pelvic nerves innervating the female reproductive tract
Single-unit afferents recordings from the pelvic nerve innervating the vagina of Sham control mice and mice with endometriosis were performed using an ex vivo afferent recording preparation as previously described [16]. One of the bundles was placed onto a platinum recording electrode. A separate platinum reference electrode rested on the mirror in a small pool of Krebs solution adjacent to the recording electrode.
Action potentials, generated by mechanical stimuli applied to the afferent's receptive field, were recorded using a differential amplifier, and filtered and sampled (20 kHz) using a 1401 interface (Cambridge Electronic Design, Cambridge, UK).

Mechanosensory profile of pelvic afferents innervating the vagina
Receptive fields tested in this study were limited to the vagina (above vaginal opening and below cervix) and were identified by systematically stroking the mucosal surface of the vagina with a stiff brush to activate all subtypes of vaginal mechanoreceptors. Mechanosensory properties of the pelvic afferents innervating a particular receptive field within the vagina were assessed by three distinct mechanical stimuli as previously described [16]. These included: (i) static probing with calibrated von Frey hairs (vfh; 2 g force; applied 3 times for a period of 3 s); (ii) mucosal stroking of the vaginal surface with calibrated vfh (10-1000 mg force; applied 10 times each); and (iii) circular stretch (5 g; applied for a period of 30 s). Once baseline mechanosensitivity was tested, a small chamber was then placed onto the mucosal surface of the vagina, surrounding the afferent receptive field. Residual Krebs solution within the chamber was aspirated and the NaV channel modulators OD1 (100 nM) and Tsp1a (200 nM) were applied in separate experimental preparations for 5 min each. The afferent receptive field was then re-tested using the same three mechanical stimuli.

Statistical analysis of afferent recording data
Action potentials were analysed off-line using Spike 2 (version 5.21) software (Cambridge Electronic Design, Cambridge, UK) and discriminated as single units based on distinguishable waveforms, amplitudes, and durations. Data are expressed as mean ± SEM. n = the number of afferents recorded, N = the number of animals used for those specific experiments. Data were statistically compared using Prism 7 software (GraphPad Software, San Diego, CA, USA) and, where appropriate, analysed using paired or un-paired Student's t-test and one-or two-way analysis of variance (ANOVA) with Bonferroni post hoc tests. Differences were considered statistically significant at P < 0.05.

Assessment of vaginal pain sensitivity in vivo
We quantified the visceromotor response (VMR) to vaginal distension (VD) as an objective measurement of vaginal sensitivity to pain in fully awake mice as previously described [16; 17; 48].
After surgery, animals were single housed to protect the EMG electrodes. Animals were allowed to recover from surgery for at least 3 days before VMR assessment. On the day of VMR assessment, animals were briefly sedated with isoflurane and Tsp1a (200 nM

Statistical analysis of VMR data
To quantify the magnitude of the VMR at each distension pressure, the area under the curve (AUC) during the distension (30 s) was corrected for the baseline activity (AUC pre-distension, 30 s).
Total area under the curve was quantified by adding the individual AUC at each distension pressure.
VMR data are presented as mean ± SEM. N represents the number of animals. VMR data were statistically analysed by generalised estimating equations followed by Fisher's least significant difference (LSD) post-hoc test when appropriate using SPSS 23.0. Analysis and figures were prepared using Prism 7 (GraphPad Software, San Diego, CA, USA). Differences were considered statistically significant at P < 0.05.