Neuropeptide Y-expressing dorsal horn inhibitory interneurons gate spinal pain and itch signalling

Somatosensory information is processed by a complex network of interneurons in the spinal dorsal horn. It has been reported that inhibitory interneurons that express neuropeptide Y (NPY), either permanently or during development, suppress mechanical itch, with no effect on pain. Here we investigate the role of interneurons that continue to express NPY (NPY-INs) in adulthood. We find that chemogenetic activation of NPY-INs reduces behaviours associated with acute pain and pruritogen-evoked itch, whereas silencing them causes exaggerated itch responses that depend on cells expressing the gastrin-releasing peptide receptor. As predicted by our previous studies, silencing of another population of inhibitory interneurons (those expressing dynorphin) also increases itch, but to a lesser extent. Importantly, NPY- IN activation also reduces behavioural signs of inflammatory and neuropathic pain. These results demonstrate that NPY-INs gate pain and itch transmission at the spinal level, and therefore represent a potential treatment target for pathological pain and itch.

INTRODUCTION Click here to enter text. hypersensitivity of the ipsilateral paw, compared to pre-surgery thresholds. Both the 291 mechanical and heat hypersensitivity were blocked in CNO-treated mice (Figures 4E 292 and 4F). Because de novo expression of NPY is known to occur in injured A-fibre 293 afferents following nerve injury 33,[38][39][40]  would contribute to the blockade of neuropathic pain that we observed. We therefore 305 conclude that this effect is due to activation of spinal inhibitory NPY-INs. We also 306 assessed mCherry expression in the L4 and L5 DRG of 5 CFA-treated 307 AAV.flex.hM3Dq-mCherry-injected NPY::Cre mice, 3 days following CFA injection. In 308 contrast to nerve injury, neuropeptide upregulation is not observed in rodent DRG 309 under inflammatory conditions 33,40 . As expected, we observed no mCherry-labelled 310 cells in the contra-or ipsilateral L4 or L5 DRG of these mice (data not shown).

311
Spinal NPY signalling has been implicated in the suppression of neuropathic pain 312 through inhibition of NPY Y1 receptor (Y1R)-expressing excitatory interneurons in 313 the dorsal horn 29,43,44 . Therefore the suppression of neuropathic hypersensitivity that 314 we observed during chemogenetic activation of NPY-INs could be due to GABAergic 315 transmission, NPY signalling, or a combination of both. To assess the potential role 316 of Y1R signalling, we systemically co-administered CNO and the Y1R-selective 317 antagonist BMS 193885 21 prior to behavioural testing in AAV.flex.hM3Dq-mCherry-318 injected NPY::Cre mice that had undergone SNI surgery. Administration of the Y1R 319 antagonist had no effect on the CNO-mediated suppression of tactile and heat 320 hypersensitivity in these mice (Figures 4E and 4F), suggesting that action of NPY on 321 Y1 receptors is not required for this effect.
In addition to evoked hypersensitivity, peripheral nerve injury induces ongoing 323 neuropathic pain in rodents, as well as engaging affective-emotional responses to 324 pain 45 . To determine the contribution of NPY-INs to ongoing pain we tested whether 325 CNO induced conditioned place preference (CPP) in a separate cohort of 326 AAV.flex.hM3Dq-mCherry-injected NPY::Cre mice following SNI surgery ( Figure 4G).

327
A wildtype control group that had undergone SNI was also included to test for any 328 possible preference of (or aversion to) the effects of CNO that could have resulted 329 from off-target effects independent of DREADD activation. CNO did not induce  Toxin-mediated silencing of NPY interneurons causes spontaneous itch and 352 enhances pruritogen-evoked itch but does not alter nocifensive reflexes. 353 We then tested whether tetanus toxin light chain (TeLC)-mediated silencing of   Figure 5A). AAV.flex.eGFP-injected mice of 392 the same genotype were again used as a control group. NPY::Cre;GRPR CreERT2 mice 393 that received injections of AAV.flex.TeLC.eGFP showed no significant difference in 394 CQ-induced itch, compared to AAV.flex.eGFP-injected controls (P=0.34, 2-way 395 ANOVA with Tukey's post-test, Figure 5B). However, when comparing NPY::Cre and 396 NPY::Cre;GRPR CreERT2 mice that had received injections of AAV.flex.TeLC.eGFP, 397 we found that the NPY::Cre;GRPR CreERT2 mice showed significantly less CQ-induced 398 itch behaviour than NPY::Cre mice (P<0.0001, 2-way ANOVA with Tukey's post-test, 399 Figure 5B). Furthermore, AAV.flex.TeLC.eGFP-injected NPY::Cre;GRPR CreERT2 mice 400 did not display a significant increase in spontaneous biting prior to CQ administration 401 (compared to AAV.flex.eGFP-injected controls; P=0.82, 2-way ANOVA with Tukey's 402 post-test, Figure 5C) and never developed skin lesions ( Figures 5D and 5E). These 403 data demonstrate that both the spontaneous itch and the increased pruritogen-    coding for Cre-dependent constructs 7,16 . While this approach failed to capture a 517 minority of NPY-expressing neurons, it enabled us to target a large number of these 518 cells. Importantly, expression was restricted to those cells that continue to express Click here to enter text.
NPY. This was confirmed by our finding that up to 85% of the virally transfected cells 520 contained detectable levels of NPY.

521
The main differences in interpreting the roles of NPY cells are likely to depend on 522 whether the cells were inactivated (through ablation or synaptic silencing) or 523 chemogenetically activated. In agreement with Bourane et al 18 , we found that 524 silencing NPY cells had no effect on acute nociceptive thresholds. However, 525 chemogenetically activating these cells increased thresholds for both thermal and

554
Here we show that activating NPY cells also strongly suppresses CQ-evoked itch.

555
This is at odds with findings of Acton et al 21 , who reported that chemogenetic 556 activation of NPY-lineage neurons failed to alter scratching in response to CQ. There 557 are technical differences between these studies, since Acton et al used a reporter 558 mouse line to express hM3Dq, and injected CQ intradermally behind the ear. The 559 discrepancy between the results of these studies is most likely to result from higher 560 levels of DREADD expression following viral transfection, and therefore more 561 effective neuronal activation. However, there may also have been a contribution from 562 regional differences in the itch tests used (hindlimb versus head), as well as in the 563 neuronal populations targeted (as noted above). Although Bourane et al 18 reported 564 that ablating ~70% of NPY-lineage neurons had no effect on itch evoked by CQ, we 565 found that synaptic silencing of the NPY cells with TeLC increased CQ-evoked itch, 566 and often resulted in development of skin lesions, presumably secondary to the 567 spontaneous itch-related biting that was also observed. In fact, the antipruritic action  Figure 7A). This inhibitory input to GRPR cells appears to be even more 582 powerful than that originating from the dynorphin/galanin cells, since NPY-583 immunoreactive boutons accounted for 45% of the inhibitory synapses on the GRPR 584 cells, compared to the 21% from dynorphin-immunoreactive boutons. Consistent with this we found that optogenetic activation of NPY cells elicited oIPSCs in all of the 586 GRPR cells tested. Interestingly, these were of much higher mean amplitude (~250 587 pA), than the ~80 pA oIPSCs reported by Liu et al 22 in GRPR cells when galanin 588 cells were optogenetically activated using a very similar experimental approach. The 589 inhibition of GRPR cells by NPY-INs is likely to be predominantly GABAergic, since 590 oIPSCs were reduced by gabazine in all cells (with one also sensitive to strychnine).

591
Also, consistent with previous evidence showing that the majority of GRPR cells lack 592 Y1 receptors 21,23 , we did not detect outward currents in any of the GRPR cells that 593 were tested with a Y1 agonist.

594
Although GRPR-expressing excitatory interneurons have been strongly implicated 595 in itch, we have recently shown that these cells respond to noxious as well as pruritic 596 stimuli, that they correspond morphologically to a class of SDH excitatory Activating NPY cells suppresses hypersensitivity in persistent pain states 610 Importantly, in addition to its effect on acute nocifensive reflexes and itch, 611 activating NPY cells also blocked thermal and mechanical hypersensitivity in both 612 inflammatory and neuropathic pain states. In the SNI model, we found that 613 administration of a Y1 antagonist had no effect on the reversal of mechanical and 614 heat hypersensitivity when NPY cells were activated. NPY acting on Y1 receptors 615 expressed by spinal neurons is known to reduce signs of neuropathic pain 29,43,59 ; 616 however, it appears that chemogenetic activation of NPY cells generated GABAergic 617 inhibition that was sufficiently powerful to reverse the hypersensitivity independently Click here to enter text.  and spinal cord tissue was processed for imaging and analysis as described below.              were injected into the L3 segments of wild-type mice (n = 8 for both groups), and