Inhibition of itch by neurokinin 1 receptor (Tacr1) -expressing ON cells in the rostral ventromedial medulla

The rostral ventromedial medulla (RVM) is important in descending modulation of spinal nociceptive transmission, but it is unclear if the RVM also modulates spinal pruriceptive transmission. RVM ON cells are activated by noxious algesic and pruritic stimuli and are pronociceptive. Many RVM-spinal projection neurons express the neurokinin-1 receptor (Tacr1), and ON-cells are excited by local administration of substance P (SP). We hypothesized that Tacr1-expressing RVM ON cells exert an inhibitory effect on itch opposite to their pronociceptive action. Intramedullary microinjection of SP significantly potentiated RVM ON cells and reduced pruritogen-evoked scratching while producing mild mechanical sensitization. Chemogenetic activation of RVM Tacr1-expressing RVM neurons also reduced acute pruritogen-evoked scratching. Optotagging experiments confirmed RVM Tacr1-expressing neurons to be ON cells. We conclude that Tacr1-expressing ON cells in RVM play a significant role in the modulation of pruriceptive transmission.


Introduction 30
The transmission of somatosensory information in the spinal cord is under top-down modulation 31 and has been extensively studied in the context of pain signaling. Descending modulation of 32 pain is reflected by phenomena such as expectation (placebo and nocebo), diffuse noxious 33 inhibitory control (DNIC) and conditioned pain modulation (CPM) ( In rats, responses of RVM ON cells were potentiated by intramedullary microinjection of the 82 Tacr1 agonist SP (Budai et al., 2007) and we wished to determine whether this is true in mouse. 83 We identified RVM ON cells using in vivo extracellular electrophysiological recordings, with the 84 criterion that noxious pinch elicited a > 30% increase in firing that preceded the onset of the 85 hindpaw withdrawal. Using a microinjection cannula attached to a recording microelectrode (Fig.  86 1A), SP or saline was microinjected while recording from an ON cell in anesthetized mice. 87 (E) An implanted intramedullary microinjection cannula allowed assessment of itch and pain behavior after injection of SP into RVM. (F, G) Graphs plot the number of scratch bouts elicited by intradermal injection of histamine (F) or chloroquine (G) for each mouse (red dots and lines), and mean scratch bouts (black line; error bars: SEM), following intramedullary microinjection of saline or SP. Experiments with saline and SP microinjections were conducted at least 7 days apart. Microinjection of SP significantly attenuated histamine-and chloroquine-evoked scratching (F, G). (H) Mechanical withdrawal thresholds were reduced by intramedullary SP. (I) Thermal withdrawal latency was not significantly affected by intramedullary SP. *p < 0.05, **p < 0.01, ***p < 0.001. n = 5-7 males, 3 females/ group. Microinjection of SP potentiated pinch-evoked responses (Fig.1B). Following SP microinjection, 88 normalized responses of ON cells to repeated pinch stimuli were significantly increased (Fig.  89   1C). The enhancement of responses lasted > 1 hr. Histologically recovered recording sites 90 were located within the RVM and adjacent regions of the medullary reticular formation (Fig. 1D). 91 These results show that, similar to rats, RVM ON cells in the mouse are potentiated following 92 localized injection of SP. 93

Intramedullary SP inhibits scratching 94
Since RVM ON cells are potentiated following injection of SP, we next tested if intramedullary 95 microinjection of SP affected itch and pain related behaviors. Mice were implanted with an 96 intramedullary microinjection cannula dorsal to the RVM to allow microinjection of SP or vehicle 97 (Fig. 1E). The number of scratch bouts elicited by intradermal injection of histamine was 98 significantly lower following intramedullary microinjection of SP compared to saline vehicle (Fig.  99 1F). Scratching elicited by intradermal injection of chloroquine was also significantly reduced 100 following intramedullary injection of SP compared to saline (Fig. 1G). 101 There was a significant decrease in mechanical force to elicit a hindlimb withdrawal following 102 intramedullary SP compared to vehicle injection (Fig. 1H), indicating mild mechanical 103 sensitization. There was no significant effect of intramedullary SP injection on thermal hindpaw 104 withdrawal latency (Fig. 1I). We observed that both male and female mice showed a reduced 105 hindlimb withdrawal, but that males had a significantly higher force for both saline and sp 106 injection when compared with females (Supplemental Fig. 6d). Thus, activation of RVM Tacr1-107 expressing neurons, through intramedullary injection of SP, resulted in facilitation of mechanical 108 nociceptive behavior and inhibition of pruritogen-evoked scratching behavior. 109 Targeted expression of DREADDs in RVM Tacr1 expressing neurons 110 We next wanted to determine whether activation of RVM Tacr1 expressing neurons, would 111 modulate itch related behaviors. To selectively target Tacr1-expressing neurons in the RVM we 112 employed a chemogenetic approach. AAV-DIO-hM3dq-mCherry was injected into the RVM of 113 Tacr1 cre mice, resulting in selective expression of hM3Dq in Tacr1-expressing neurons (84% 114 expression of DREADDs in Tacr1 positive cells, Fig. 2A-C). 115 To activate the DREADDs (hM3Dq), clozapine was injected at a dose 0.01 mg/kg (Gomez et 116 al., 2017). This dose of clozapine was sufficient to produce increased activity of DREADDs-117 expressing neurons measured by c-fos expression. When we tested DREADDs mice there was 118 a significant increase in the number of c-fos positive cells following ip administration of 119 clozapine (62.22% +/-11.61%) when compared with ip saline (15.66% +/-1.678%) (Fig. 2h).
In 120 addition, we tested if clozapine elicited a greater increase in c-fos expression in DREADDs 121 compared to control vector (mCherry) mice. Indeed, following clozapine administration, the 122 number of c-fos positive mCherry-expressing cells was significantly higher in DREADDs 123 (hM3Dq-mCherry) mice (68.98% +/-11.83%) compared with neurons from control vector 124  Expression of hM3Dq-mCherry was limited to the RVM. Gt: gigantocellularis; RMg: raphe magnus; RPa: raphe pallidus. (C) Administration of clozapine caused a significant reduction in histamine-evoked scratching (7 males, 7 females) in hM3Dq expressing mice, but not in control vector mice (6 males, 3 females). (D) Administration of caused a significant reduction in chloroquine-evoked scratching in hM3Dq mice (7 males, 8 females) but not in control vector mice (6 males, 4 females). (E) Clozapine administration did not significantly change mechanical withdrawal thresholds in hM3Dq (7 males, 8 females) or control vector mice (6 males, 4 females). (F) Clozapine administration did not significantly change thermal withdrawal thresholds in hM3Dq (7 males, 8 females) or control vector mice (6 males, 4 females). *p < 0.05, **p < 0.01, ***p < 0.001, Two-way ANOVA with Sidaks multiple comparison test.  Taro) once per day produced a significant increase in spontaneous scratching at day 3 for both DREADDs (dark blue) and control vector (dark green) mice (n = 4 males, 4 females/group). Following clozapine, there was a significant reduction in spontaneous scratching on day 5 for the DREADDs mice (dashed light blue) but no significant change in control vector mice (dashed light green). (B) Graph shows individual DREADDs animals' spontaneous scratch bouts pre-and post-clozapine. Following clozapine there was a significant reduction in scratching. Blue: individual counts; red: mean +/-SEM. (C) Graph as in B for mice in vector control group, in which clozapine had no significant effect. Green: individual counts; red: mean +/-SEM. (D) Imiquimod induced significant increases in alloknesis scores on day 1, 3 and 5 of treatment in DREADDs (dark blue) and vector control (dark green) mice. Following clozapine administration on days 3 and 5 there were significant reductions in alloknesis scores for the DREADDs mice (dashed light blue) but not control vector groups (dashed light green). (E) Clozapine resulted in a significant reduction in alloknesis scores (format as in B). (F) Clozapine had no effect on alloknesis scores in vector controls (format as in C). *p < 0.05, **p < 0.01, ***p < 0.001. a model of chronic itch in addition to acute pruritogen-evoked scratch bouts. 173

RVM ON cells express Tacr1 174
Previous reports had shown, in rats, that RVM ON cells were facilitated by local microinjection Occasionally Neutral cells were identified by their response to optic stimulation (n = 3), and 205 each was strongly entrained (Fig. 5F). Since Neutral cells are numerous, not affected by 206 noxious stimulation and their contribution, if any, to descending modulation is unknown, they 207 were not investigated further. We thus conclude that while most presently-recorded light-208 sensitive cells were ON cells, other cell types including Neutral cells also expresses Tacr1. 209 To validate our results from optotagging, we used several metrics to determine whether a light-210 sensitive neuron was truly entrained to the optic stimulus. We assumed that each optic stimulus 211 was likely to directly excite a neuron expressing ChR2 and tested this in different ways. An 212 efficiency index was calculated to determine whether a neuron was entrained to the optic 213 stimulus, by counting the number of evoked action potentials divided by the number of optic 214 stimuli (Supplemental Fig. 3). The 3 Neutral cells had the highest efficiency index (>1), which 215 was due to occasional firing of "doublet" action potentials in response to the optic stimulus 216 (Supplemental Fig. 3C,D). ON cells classified as light sensitive also had a high efficiency index, 217 which decreased with increasing optic stimulation frequency most likely due to desensitization of 218 ChR2. We additionally assumed that the neuronal firing rate would reflect the number of optic 219 stimuli. Indeed, the neuronal firing rate in light sensitive neurons increased with the stimulation 220 frequency, while non-light sensitive neurons showed no change (Supplemental Fig. 4). In   Experiments were performed used wild type mice (c57BL/6j, Jackson Labs, Bar Harbor ME), 317 In vivo single-unit recording 384 Adult mice (6 wk) mice were anesthetized with pentobarbital sodium (60 mg/kg, ip). The head 385 was secured in a stereotaxic frame and an opening was made in the occipital bone. The 386 animal's body temperature was maintained with a heating pad and external heating source. 387 Teflon coated silver wires were inserted into the biceps femoris to record electromyographic 388 (EMG) activity. A single-unit recording microelectrode (10 MOhm, Frederick Haer Inc., Bowdoin 389 ME) was coupled to a 33-gauge injection cannula such that the tip of the recording 390 microelectrode extended several µm beyond the tip of the injection cannula and was inserted 391 into the RVM. ON cells were identified by a hindpaw pinch-evoked increase in firing that 392 preceded the hindlimb withdraw reflex measured as EMG activity from fine wires inserted in the 393 biceps femoris. Once an ON cell was identified, either saline or SP (0.5 μl, 10 nmol) was 394 microinjected and changes in firing rate to repeated consistent pinch stimuli were recorded. 395 Responses to pinch stimuli were normalized to the initial response for each unit. Electrode 396 voltages were amplified and digitized (CED 1401, CED, Cambridge UK) and analyzed with 397 Spike2 (CED). At the end of the recording, a lesion was produced at the last recording site by 398 passing direct current through the microelectrode, the brain was harvested postmortem and 399 postfixed in 10% formalin. 400 Optotagging 401 At least 4 weeks following injection of AAV5: hSyn-DIO-ChR2-eYFP into RVM, single-unit 402 recordings were made from RVM as described above using a Tungsten microelectrode attached 403 to an optic fiber such that the microelectrode tip extended a few hundred microns beyond the 404 optic fiber. In most experiments ON cells were functionally characterized as described above, 405 and OFF cells were characterized by a pinch-evoked pause in ongoing activity that preceded 406 the withdrawal reflex. ON and OFF cells were then tested for entrainment to optic stimulation at 407 473 nm wavelength and 0.25 mW -5 mW light output (5-20 Hz, 10ms pulse duration) from a 408 laser (Laserglow R471003GX). In some experiments, optic stimulation was used as to isolate 409 light-sensitive neurons, which were then identified as ON, OFF or Neutral based on their 410 response to pinch. Efficiency indices were calculated as the number of action potentials firing 411 within a 20 ms window following the onset of the optic stimulus divided by the total number of 412 optic stimulus pulses, and were used to determine whether a neuron was entrained to the optic 413 stimulus. Both male and female mice were used for all electrophysiological experiments. 414

Imiquimod Treatment 415
Imiquimod cream was applied topically to the shaved area on the rostral back once per day for 5 416 consecutive days. Treatment groups consisted of age-matched male and female NK-1-cre mice 417 that had received intra-RVM injection of AAV5: DIO-hM3Dq-mCherry or the control vector 418 AAV5: hSyn-DIO-mCherry. Imiquimod treatment induced signs of skin pathology including skin 419 scaling and erythema. As measures of chronic itch, we assessed spontaneous scratching, and 420 alloknesis, 23 hours following imiquimod treatment. Mice were videotaped and tested between 421 10 AM and 5 PM, with each individual mouse tested at the same time each day. The mice were 422 habituated to glass cylinders for 3 successive days prior to recording. Animals were videotaped 423 from above for 30 min. Behavioral videos were analyzed by two blinded observers. Only 424 discrete bouts of spontaneous hindlimb scratches directed towards the application site were 425 counted, as described previously (Akiyama et al., 2016) and summed over the 30 min period. 426 Alloknesis was assessed as previously described (Akiyama et al., 2012). The mouse was 427 placed in an enclosed area and a 0.07g von Frey monofilament was applied to the perimeter of 428 the imiquimod application area 5 consecutive times. The alloknesis score consisted of the 429 number of immediately-occurring hindlimb scratch bouts directed to the stimulus site. 430 Immunofluorescence 431 Four weeks after intra-RVM injection of AAV5:hSyn-DIO-hM3Dq-mCherry or the control vector 432 (AAV5:hSyn-DIO-mCherry) in Tacr1 cre mice, clozapine (0.01 mg/kg, i.p.) was injected, 433 Dallas TX) at 1:50 overnight at 4°C. c-fos was stained with a primary c-fos antibody (ab190289, 438 Abcam, Cambridge UK) at 1:10,000 overnight at room temperature. Alexa Fluor 488 (ab15077, 439 Abcam) was applied at 1:2000 for 2 hours at room temperature. All slides were mounted with 440 vectashield and imaged with confocal microscopy. Staining intensity was measured relative to 441 the red fluorescence (from DREADDs) and was quantified using FIJI (Schindelin et al., 2012). 442

Statistical analysis 443
All statistical analyses were performed using GraphPad Prism. Values are presented as mean 444 +/-SEM. Statistical significance was assessed using students t-test or a two-way, repeated 445 measures ANOVA with Bonferroni's correction, unless otherwise specified. Significance was 446 indicated by p < 0.05. Sample sizes were based on pilot data and are similar to those typically 447 used in the field. 448 For analysis of effects of intracranial microinjections, a paired students t-test was used to 449 compare the effects of intracranial injection of saline versus SP on behavioral measures. Two-450 way ANOVA determined a lack of significant sex x SP microinjection interaction, so the data 451 from both sexes were pooled. For chemogenetic experiments, a paired t-test was used to 452 compare scratch counts and nociceptive measures following vehicle vs. clozapine or CNO 453 injection, or between vector controls and DREADDs mice following clozapine or CNO. For  Figure 1. Chemogenetic activation of RVM Tacr1 expressing neurons inhibits itch related behavior. Tacr1creER mice were injected with AAV2-DIO-hM3Dq-mCherry. (A) CNO administration significantly reduced chloroquine evoked scratch bouts in DREADDs but not control vector mice (n = 5-7 males, 5-7 females). (B) CNO reduced spontaneous scratching in hM3Dq-expressing but not control vector mice ( n = 5 males, 5 females). (C) CNO administration significantly reduced mechanical withdrawal thresholds in DREADDs but not control vector mice (n = 5-7 males, 5-7 females). (D) CNO did not affect thermal withdrawal latency in any group (n = 5-7 males, 5-7 females).
Supplemental Figure 2: Clozapine administration does not affect acute itch or pain behavior. Saline, low dose clozapine (0.01 mg/kg) or high dose clozapine (0.1 mg/kg) was administered systemically, followed by tests for acute itch and pain behaviors. (A, B) Clozapine did not significantly affect the number of scratch bouts elicited by intradermal injection of histamine (A) or chloroquine (B). (C, D) Clozapine also did not significantly affect the respective latency or threshold of hindlimb withdrawals elicited by acute thermal (C) or mechanical (D) stimuli. (A,C and D) n = 6 males, 6 females; (B) n = 5 males, 3 females. P>0.05, 1-way ANOVA followed by Bonferroni post hoc test.