Analgesia linked to Nav1.7 loss of function requires μ and δ opioid receptors

Functional deletion of the SCN9A gene encoding sodium channel Nav1.7 makes humans and mice pain-free (1,2). Opioid signaling contributes to this analgesic state (3). Here we show that the pharmacological block or deletion of both μ and δ opioid receptors is required to abolish Nav1.7 null opioid-related analgesia.κ-opioid receptor antagonists were without effect. Enkephalins encoded by the Penk gene are upregulated in Nav1.7 nulls (3). Deleting Nfat5, a transcription factor with binding motifs upstream of Penk (4), induces the same level of enkephalin mRNA expression as found in Nav1.7 nulls, but without consequent analgesia. These data confirm that a combination of events linked to SCN9A gene loss is required for analgesia. Higher levels of endogenous enkephalins (3), potentiated opioid receptors (5), diminished electrical excitability (6,7) and loss of neurotransmitter release (2,1) together contribute to the analgesic phenotype found in Nav1.7 null mouse and human mutants. These observations help explain the failure of Nav1.7 channel blockers alone to produce analgesia and suggest new routes for analgesic drug development.

Functional deletion of the SCN9A gene encoding sodium channel Nav1.7 makes humans and mice painfree (1,2). Opioid signaling contributes to this analgesic state (3). Here we show that the pharmacological block or deletion of both and  opioid receptors is required to abolish Nav1.7 null opioid-related analgesia -opioid receptor antagonists were without effect. Enkephalins encoded by the Penk gene are upregulated in Nav1.7 nulls (3). Deleting Nfat5, a transcription factor with binding motifs upstream of Penk (4), induces the same level of enkephalin mRNA expression as found in Nav1.7 nulls, but without consequent analgesia. These data confirm that a combination of events linked to SCN9A gene loss is required for analgesia. Higher levels of endogenous enkephalins (3), potentiated opioid receptors (5), diminished electrical excitability (6,7) and loss of neurotransmitter release (2,1) together contribute to the analgesic phenotype found in Nav1.7 null mouse and human mutants. These observations help explain the failure of Nav1.7 channel blockers alone to produce analgesia and suggest new routes for analgesic drug development.
Pain is numerically the greatest clinical challenge of the age, affecting about half the population, whilst 7% of people have debilitating pain conditions (9). Finding new analgesic targets and drugs has proved challenging. One approach has been to identify the genes involved in human monogenic loss of pain conditions (10). The association of human gain-of-function mutations in Nav1.7 with enhanced pain phenotypes, and the pain-free state linked to loss of Nav1.7 expression focused considerable attention on this voltage gated sodium channel as a potential analgesic drug target (11). Nav1.7 is found in damage-sensing peripheral sensory neurons, sympathetic neurons and CNS structures like the hypothalamus, as well as in non-neuronal locations such as the pancreas. Deletion in all sensory neurons and sympathetic neurons abolishes acute, inflammatory and neuropathic pain, although some pain disorders such as oxaliplatin-evoked cold allodynia are retained (2,12).
As human and mouse Nav1.7 null mutants are effectively pain-free, this channel should be an excellent analgesic drug target. However, channel blockers are very weak analgesics (11,13). This is likely due to the fact that partial channel block cannot recapitulate the many physiological effects of gene deletion. This explanation is supported by experiments that show that only 100% channel block with very high dose tetrodotoxin can recapitulate some effects of gene deletion (3). In null mutants neurotransmitter release is diminished, and synaptic integration is also diminished. In addition, the opioid peptide enkephalins are upregulated in the absence of Nav1.7, and opioid receptor signaling is potentiated. Both of these latter events may be linked to loss of sodium ingress through Nav1.7 (3).
Consistent with an opioid component of analgesia, the opioid antagonist naloxone substantially reverses Nav1.7 loss-associated analgesia (3). We wondered which opioid receptors were involved in this process. Here, using pharmacological studies and opioid receptor knockout mice, we show that both  and  opioid receptors contribute to Nav1.7 null mutant analgesia and deleting both receptors mimics the effects of naloxone on Nav1.7 null analgesia in mice. In addition, we show that elevating enkephalin mRNA levels to those found in Nav1.7 nulls is not alone sufficient to cause measurable analgesia.
We first examined the role of -opioid receptors (MOR) in Nav1.7 null-associated analgesia ( Figure 1,A  B). Nav1.7 null mutant mice show dramatic thermal analgesia. Global deletion of MOR on a Nav1.7 null background had a small effect on acute heat pain behaviour ( Figure 1A). This effect did not match the effects of naloxone that substantially diminished analgesia ( Figure 1B). Consistent with this, naloxone further diminished the analgesic phenotype of Nav1.7 / MOR double mutant mice, demonstrating that MORs alone do not account for the opioid-mediated component of Nav1.7 null associated analgesia (Figure1B).
Next we tested the effect of deleting -opioid receptors (DOR) on Nav1.7 null pain behaviour (16). Once again there was a small diminution in analgesia compared to Nav1.7 null mice (Figure1C). Naloxone further diminished the analgesic phenotype of the Nav1.7/DOR double null mutants ( Figure 1D), demonstrating that DORs alone do not account for the opioid mediated component of Nav1.7 null associated analgesia. However, when the potent selective MOR antagonist CTOP was applied to DOR receptor null mice (17), the analgesia associated with Nav1.7 deletion was reduced by the same level as with naloxone ( Figure 1F). CTOP and the kappa-opioid (KOR) antagonist norBNI (18) together also had the same effect as naloxone when applied to a Nav1.7/DOR double null (Figure1E). However, norBNI on a Nav1.7/DOR null background was without effect (Figure1G). These data show that KORs do not mediate analgesia in Nav1.7 null mutants, but pharmacological block of MOR on a DOR null background can account for all opioid mediated analgesia (17).
To provide further evidence that both MOR and DOR contribute to opioid-mediated analgesia in Nav1.7 nulls, we generated double opioid receptor null mutant mice on a Nav1.7 null background. Double MOR/DOR knockouts on a Nav1.7 null background showed exactly the same loss of analgesia as that caused by naloxone in Nav1.7 knockout mice (Figure2A,B). Application of   and  antagonists (19) together did the same (Figure2D), although the KOR antagonist norBNI alone was inactive, confirming that KOR activation did not contribute to analgesia (Figure2C). These pharmacological and genetic studies demonstrate that MOR and DOR together account for opioid-mediated analgesia in Nav.7 null mutant mice.
Elevated levels of enkephalins are found in Nav1.7 null mutant mice (3). Interestingly, there are five consensus binding sites for the transcription factor NFAT5 upstream of the Penk coding region. NFAT5 recognizes DNA elements similar to those bound by NFATc proteins (4). As NFAT5 activity is regulated by hyperosmolarity and salt kinases (20), there is a potential link between sodium ingress through Nav1.7 and transcriptional regulation. We manipulated sodium levels in sensory neuron cultures using either monensin as a sodium ionophore ([Na + ] control 6.65 mM, SEM 0.27; [Na + ] Monensin 9.46 mM, SEM 0.44; n = 19;) or veratridine as an activator of voltage gated sodium channels (control [Na + ] 5.5 mM, SEM 0.25; [Na + ] Veratridine 7.6 mM, SEM 0.41; n = 9) to increase sodium levels, and very high doses of TTX (500nM) to block voltage gated sodium channel activity and potentially lower intracellular sodium (Supplementary data). Intriguingly, agents that alter intracellular sodium concentrations impact similarly on Nav1.7 and NFAT5 mRNA levels. Monensin ( Figure 3A) lowered both PENK and NFAT5 mRNA levels, whilst TTX elevated them (Figure3B). The TTX effect was apparent in wild type mice, but not in Nav1.7 nulls, implying that this channel is the locus of action for PENK mRNA control by TTX (Figure3C,D).
Veratridine lowered both PENK and NFAT5 mRNA levels in wild type but not Nav1.7 null mutant mice, again linking transcriptional events to Nav1.7 channel activity (Figure3E,F). We examined the role of NFAT5 using conditional NFAT5-Wnt1-Cre null mutants in sensory neurons of wild type and Nav1.7 null mutant mice. Expression levels of NFAT5 and Nav1.7 transcripts in single and double mutants were analysed to confirm Cre activity at the floxed loci (Supplementary data). NFAT5 conditional null mutant mice showed enhanced expression of PENK mRNA ( Figure 4A). When the NFAT5 null mice were crossed with Nav1.7 null mutants, PENK mRNA levels further increased ( Figure 4A). As NFAT5 null mice have the same levels of PENK mRNA as Nav1.7 null mutants, this allowed us to examine the contribution of enhanced opioid peptide expression to the analgesia seen in Nav1.7 null mutant mice. Opioid signalling in Nav1.7 null mutants is potentiated in at last two ways. First, there are enhanced levels of enkephalins, and second the opioid receptors have much enhanced activity, as measured indirectly through the quantitation of PKA signalling (5). There was, perhaps surprisingly, no analgesic effect of elevated enkephalin levels in the Nfat5 null sensory ganglia. By measuring noxious mechanosensation ( Figure 4B), thermal thresholds ( Figure 4C), and noxious heat-induced pain related behaviour ( Figure  4D), the NFAT5 null enkephalin-induced mice showed normal pain behaviour, compared to Nav1.7 null mice ( Figure 4). As opioids clearly play a role in Nav1.7 null analgesia, as demonstrated by the naloxone effects, this suggests that the enhanced activity of opioid receptors may make a major contribution to the Nav1.7 null opioid-mediated analgesia.
What are the implications of these findings for drug development? Firstly the complexity of physiological changes that occur in Nav1.7 nulls is striking. Receptors (e.g. 5HTr4) and transcription factors (e.g. Runx1) implicated in nociception are dysregulated (3) Opioid peptide expression is increased (3) and opioid signalling is potentiated (5), whilst electrical excitability (6) and integration of nociceptive stimuli is lost (7). There is evidence that these events require the complete loss of Nav1.7 function, as occurs in null mutants. For example, only complete channel block with very high doses of TTX can induce increased PENK mRNA expression (3). Should small molecule-specific Nav1.7 antagonists be able to replicate all these events then they would be excellent analgesics. All the evidence thus far demonstrates that this is not the case, and the necessarily partial blockade of Nav1.7 does not cause analgesia (13). Molecules with limited specificity like Biogen's BIIB074 are good analgesics but much of their activity likely results from blockade of sodium channels other than Nav1.7 (20).
The role of MOR and DOR and the lack of a role for KOR in Nav1.7 null analgesia fit with recent data. There is evidence for  receptor interactions in nociceptive sensory neurons (21), and primates express  heteromultimers as targets of opioid analgesia (22). As Nav1.7 deletion with peripheral nervous system-dependent Cre mice causes analgesia, then the actions on opioid receptors must occur either on primary sensory neurons, or on their synaptic targets within the spinal cord. Evidence that coadministration of opioids with Nav1.7 antagonists can have synergistic therapeutic effects has been demonstrated with a number of specific Nav1.7 antagonists. However, human proof of concept studies on synergistic analgesia with Nav1.7 antagonists and opioids have yet to be published. The evidence for potentiation of opioid receptor signalling in Nav1.7 null mice is significant (5). Although diminished electrical excitability may provide the necessary landscape for endogenous opioid effects, it is surprising that elevated enkephalin levels alone do not produce any detectable levels of analgesia in the NFAT5 null mice. Exogenous administration of enkephalins in humans delivered through gene therapy has useful analgesic effects (23). The focus is then upon potentiated opioid receptor signaling (5 ). There is some evidence linking the ingress of sodium through Nav1.7 to effects on GPCR activity. Pert and Snyder showed the influence of sodium on opioid receptor activity in 1974, demonstrating that increased sodium concentrations caused diminished agonist binding (24). Intracellular sodium levels may control this process (25) and the proximity of Nav1.7 channels to opioid receptors may influence sodium occupancy of these GPCRs (26).
In summary,  and opioid receptors are required for the opioid component of Nav1.7 null mutant analgesia. Co administration of / agonists with specific Nav1.7 antagonists may therefore have useful analgesic effects (29). If analgesia depends substantially upon both potentiated receptor activity , as well as increased enkephalin expression, analgesic drug development using small molecules to mimic Nav1.7 gene deletion will be problematic. Nociceptor silencing through CRISPR mediated gene deletion of Nav1.7 may prove a more tractable analgesic strategy for extreme chronic pain conditions (30).
respectively dissolved in ethanol, saline and DMSO and incubated for the time specified in the figure legend. For controls, the same volumes of vehicle were used.

DRG neuron cultures
DRG from all spinal levels were harvested and dissociated as described (28). Dissociated neurons were plated on poly-L-lysine and laminin coated 35-mm plastic dishes (Nunc, Denmark). Incubation with drugs was started at least 24h after dissociation. Monensin (Sigma, UK, in ethanol absolute), TTX (Sigma, UK, in extracellular solution) or Veratridine (Sigma, UK, DMSO) were used at concentrations described in the figure legends before RNA extraction and quantification. For each experiment, control DRG neurons were treated with the appropriate vehicle.

Quantitative PCR
For fresh DRG analysis, DRG from lumbar segments L4, L5 and L6 were harvested and pooled. For DRG cultures, cells were collected after incubation with the drug and concentrated by centrifugation. RNA was extracted using TRIzol® Reagent (Invitrogen) according to the manufacturer's instructions. Reverse transcription was performed using iScript™ Reverse Transcription Supermix for RT-qPCR following Bio-Rad supplied protocol. cDNA amplification was performed in triplicate, using SsoAdvanced™ Universal SYBR® Green Supermix (Bio-Rad). DNA amplification was quantified with a Bio-Rad CFX Connect™ Real-Time PCR Detection System thermocycler. The expression level of target genes was normalized to housekeeping gene mRNA (Gapdh). Fold changes were determined using the 2 −ΔΔCt equation in which wild-type littermate or vehicle-treated cultured DRG cDNA samples were designated as the calibrator. The data presented are given as the mean of the fold changes.

Statistical analysis
Data were analysed using GraphPad Prism 7. Statistical tests performed for a given experiment are described in figure legends.