Sensory TRP channels contribute differentially to skin inflammation and persistent itch

Although both persistent itch and inflammation are commonly associated with allergic contact dermatitis (ACD), it is not known if they are mediated by shared or distinct signaling pathways. Here we show that both TRPA1 and TRPV1 channels are required for generating spontaneous scratching in a mouse model of ACD induced by squaric acid dibutylester (SADBE), a small molecule hapten, through directly promoting the excitability of pruriceptors. TRPV1 but not TRPA1 channels protect the skin inflammation, as genetic ablation of TRPV1 function or pharmacological ablation of TRPV1-positive sensory nerves promotes cutaneous inflammation in the SADBE-induced ACD. Our results demonstrate that persistent itch and inflammation are mediated by distinct cellular and molecular mechanisms in a mouse model of ACD. Identification of distinct roles of TRPA1 and TRPV1 in regulating itch and inflammation may provide new insights into the pathophysiology and treatment of chronic itch and inflammation in ACD patients.

The manuscript by Feng and co-workers investigates the mechanisms underlying squaric acid dibutylester (SADBE) induced itch and inflammation. Previous studies have shown that SADBE elicits inflammation and itch-like behaviors in mice (Qu et al 2014; Qu et al 2015). The novel feature of the current manuscript is that the authors show that SADBE directly activates both TRPV1 and TRPA1 ion channels that are expressed by sensory neurons. The evidence for this mode of action is good, although there are clearly other important actions of SADBE that are not explored. The finding that an electrophilic compound like SADBE activates TRPA1 is not surprising; it would have been surprising if it didn't. The data suggesting an action via binding to the putative capsaicin binding site in TRPV1 adds a more interesting observation. The current study also points to a role of sensory nerves and TRPV1 in regulating SADBE-evoked inflammation. This is perhaps the most interesting observation in the study but it is not investigated in depth. I think that there are aspects of the study that need clarification either by provision of further data or by stronger arguments. 1. SADBE activates both TRPV1 and TRPA1, which makes it highly unlikely that the behavioral effect of acute exposure is to induce only itch. Such stimulation will activate all nociceptive neurons and evoke pain, which is consistent with the results of Qu et al. 2014, 2015, which showed as much or more pain-like responses than itch-like responses after SADBE administration. Were pain-like responses measured and what happened in the Trpa1<sup>-/-</sup> , Trpv1<sup>-/-</sup> and Trpa1<sup>-/-</sup>/Trpv1<sup>-/-</sup> mice? 2. The conclusions about the relative roles of TRPA1 and TRPV1 would be strengthened by the addition of pharmacological evidence that selective antagonists recapitulated the effects seen in the knockout mice. Deleting these channels (Trpa1<sup>-/-</sup>, Trpv1<sup>-/-</sup> and Trpa1<sup>-/-</sup>/Trpv1<sup>-/-</sup> mice) may have unexpected 'compensatory' or 'modifying' effects on physiological processes. 3. Scratching was not totally eliminated in the Trpa1<sup>-/-</sup>/Trpv1<sup>-/-</sup> mice (Lines 211 -218). There was still a substantial response. What is the mechanistic basis for this residual pruritic effect -is it related to other actions on sensory neurons, perhaps increased action potential firing due to enhanced voltage gated sodium currents (Qu et al 2014;) or reduced potassium currents? 4. The mechanism by which the absence of eithe r TRPV1 expressing sensory neurons (RTXtreatment) or functional TRPV1 channels (Trpv1<sup>-/-</sup> mice) increases SADBE-evoked inflammation is not investigated in depth. Other studies have shown opposite effects of RTX treatment on inflammation elicited by some haptens (e.g. Imiiquimod -Riol-Blanco et al 2014 Nature 510: 157) and similar effects of TRPV1 inhibition for other haptens (e.g. Oxazalone -Bánvölgyi A et al 2005 J Neuroimmunol 169: 86). There is a similar dichotomy between the conclusion that TRPA1 does not influence inflammation in this SADBE model but is important for oxazalone-evoked inflammation (Liu et al., 2013 FASAB J 27:3549). The authors provide some evidence for cytokine changes to explain their findings but the mechanisms by which s ensory neurons drive these changes and why this is linked to TRPV1 and not to TRPA1 are not addressed. I recommend that the authors add more details or amend wording to aid the reader. 5. Methods. Lines 439 -453, 459 -462 and e.g. Figure 1. A). The protocols for administration of SADBE need to be consistently explained in the methods and the figures. The diagrams in the figures show SADBE administered first to the abdomen and then to the ear. But the scratching behavior is measured after SADBE is administered to either the nape of the neck . B) The timing of SADBE administration when measuring ear edema is unclear. The method (line 461) states measurements made after the challenge at day 3, but the protocol shown in figure 1 shows measurement 3 days after the last SADBE treatment. Do you mean 'measurements were made 3 days after the last SADBE challenge?' 6. What vehicle controls were used in the in vitro experiments and in the intradermal injection experiments? Was it acetone? If so, what was the final acetone concentration and did this evoke scratching. Vehicle controls should be presented. 7. Line 399 onwards -Methods. There are no methods given for the current clamp electrophysiology experiments. These should be added. 8. Line 111. Was there any measurement to show that the FTY720 treatment was effective in blocking T cell migration? 9. Lines 155 -164. The results shown in Figure 3, panels a-d need to be expressed quantitatively. How many neurons were studied, how many neurons responded to SADBE , how many coverslips? What were the percentages -the term 'similar percentages' is vague? The 3mM SADBA concentration used in these DRG experiments is a sub-maximally effective concentration for TRPV1 activation in the heterologous expression experiments (see Figure 4g). From Fig 3b it looks as though there are many TRPV1 expressing neurons that did not respond to SADBE. Were there SADBE responsive neurons that were not AITC or capsaicin sensitive in DRG neurons from wildtype, Trpa1<sup>-/-</sup> or Trpv1<sup>-/-</sup> mice? What was the percentage of responding neurons in the double knockout (Trpa1<sup>-/-</sup>/Trpv1<sup>-/-</sup>) DRG neurons -was it really 0%? 10. Figure 3, panels b and c. The time courses of the responses in the Trpa1<sup> -/-</sup> and Trpv1<sup>-/-</sup> mouse DRG neurons looks very different (Trpa1<sup>-/-</sup> transient, Trpv1<sup>-/-</sup> more persistent). Was this a consistent finding? Two components of response time courses also appear to be present in the wild-type DRG neurons. A difference in time course would be consistent with the time courses shown for TRPV1 -and TRPA1-mediated calcium responses in Figure 4a and b. 11. Lines 191 -194 & Figure 4. The SADBE-evoked responses of hTRPA1 mutants are depressed, but is this due to a reduced expression level? For human TRPA1 3C mutants, responses to a nonelectrophilic agonist such as carvacrol would demonstrate that the channels were well expressed. 12. Lines 197 -198 and Figure 4. The concentration response curve for the TRPV 1 M548L mutant appears to be shifted leftwards compared to the wild-type TRPV1, yet the authors say there was no effect. Some quantification of results and evidence of reproducibility should be added to justify the conclusion. 13. Lines 209 -210. Presumably there are some words missing in this sentence (e.g. mice lacking functional TRPA1 and/or TRPV1).

Reviewer #2 (Remarks to the Author):
This study shows distinct contributions of TRP channels to skin inflammation and itch in the SADBE-induced contact dermatitis model. By using TRPA1 knockout (KO), TRPV1 KO and TRPA1/TRPV1 double KO mice, the authors demonstrate that persistent scratching caused by SADBE treatment is mediated through both TRPA1 and TRPV1 channels. C alcium imaging and electrophysiological experiments using acutely dissociated DRG cells and HEK cells transfected with TRPA1 or TRPV1 show that SADBE can activate directly both TRP channels. Unlike persistent scratching induced by repeated application of SADBE, acute scratching induced by i ntradermal SADBE injection is mediated mainly through TRPA1. In contrast to scratching behavior, deficiency or pharmacological inhibition of TRPV1 exacerbates skin inflammation probably through promoting Th1 cytokines. These findings strongly suggest that SADBE causes itch independent of skin inflammation. Although most experiments were appropriately conducted and the manuscript is well organized, there remain a number of questions that should be addressed.
1. The authors found an increase in scratching a t 3 days after the final SADBE challenge in sensitized mice (Fig. 1), thereby regarding this scratching as "persistent" itch. On the other hand, they show that intradermal injection of SADBE elicited "acute" scratching in naïve mice (Fig. 5). I strongly recommend that the authors should present the time course of scratching after SADBE application in both the "persistent" and "acute" model, and also should examine the effects of TRP channel deficiency on scratching at different time points. Perhaps, increas ed scratching will be observed immediately after the final application of SADBE in sensitized mice; does the deficiency of TRPA1 alone or both TRPA1 and TRPV1 inhibit this scratching response? Also, how long does the scratching induced by SADBE injection in naïve mice continue? The present results raise an interesting question whether persistent scratching is due to continuation of direct activation of TRPA1 and/or TRPV1 by SADBE existing in the skin, or whether other mediator(s) induced by repeated SADBE application eventually activate TRPA1 and TRPV1 channels.
2. As shown in Figs. 1e and i, SADBE challenge on the nape of the neck causes scratching, irrespective of systemic presensitization on the abdominal skin, although skin inflammation is significantly but very weakly promoted by the presensitization. Therefore, SADBE causing symptoms, especially itch (or scratching), should be considered as irritant contact dermatitis (IC D), but not allergic contact dermatitis (AC D).
3. The authors showed that intradermal SADBE injection in naïve mice exclusively caused scratching compared to wiping behaviors (Fig. 5); however, each behavior should be compared to when the vehicle was injected. Scratching and wiping behaviors should not be compared to each other. Furthermore, Qu et al. have reported that repeated SADBE application induces not only itchrelated scratching but also pain-like behaviors, such as wiping and licking (Qu L, Brain, 2014). As described, both TRPA1 and TRPV1 play a critical role in pain. Have yo u checked pain behaviors in your SADBE model? How does the TRP channel deficiency affect pain caused by SADBE treatment 6. C alcium imaging and electrophysiological experiments clearly showed mutual compensatory effect in the DRG neurons isolated from TRPA1 KO or TRPV1 KO mice (Fig. 3). On the other hand, SADBE-induced persistent scratching could be suppressed in either TRPA1 KO or TRPV1 KO mice (Fig. 2). If DRG neurons play a dominant role in SADBE-induced persistent itch as shown in Fig. 8, these results would be theoretically inconsistent. Is it possible that TRPA1 -and/or TRPV1expressed in other cells besides DRG neurons is involved in the persistent itch? The authors should fully address this question. 7. The responses to SADBE in TRPV1-expressing HEK293 cells were clearly weaker than those in TRPA1-expressing cells (Fig. 4). Indeed, the EC 50 value for TRPV1-expressing cells was 5.6 times higher than that for TRPA1 (1.30 mM for TRPA1 vs. 7.26 mM for TRPV1). Why is that? Also, what concentration of SADBE did you use? If it was 3 mM similar to other in vitro experiments, the authors should use a higher (or submaximal) concentration (e.g., 10 mM) to determine the SADBE activation of TRPV1.
8. The authors suggest that increased Th1 cytokines would contribute to aggravation of SADBEinduced skin inflammation by TRPV1 deficiency. On the other hand, SADBE -induced skin inflammation was not affected in Rag1-/-mice, which lack T cells (Fig. 1f), suggesting that T cells are hardly required for SADBE-induced skin inflammation. Also, since the authors have not examined the effect of NK cell depletion alone on skin inflammation, the role of NK cells and T cells in skin edema by SADBE remains unclear. They thus should further address this question and revise the working model ( Fig. 8) to fit their findings.

Reviewer #3 (Remarks to the Author):
Jing et al. showed that the SADBE-induced persistent itch was not depend on lymphocytes, but mediated by TRPA1 and TRPV1channels. They demonstrated SADBE can directly activate both TRPA1 and TRPV1 in vivo using freshly isolated DRG cells and TRPA/TRPV1 -expressing HEK293 cells. Further, their observation suggests that TRPV1 also affect SADBE -induced ear swelling via inhibiting the production of Th1 cytokines. In this paper, data were well-presented; however, I had concerns relating to the interpretation of some of the results, and lack of mechanistic insight. Followings are my specific comments: 1. Fig.1e showed that irritant response (innate response) by SADBE challenge was very strong with this experimental protocol. It makes hard to evaluate adaptive immune response in this condition. Author need to modify the protocol to reduce irritant response to SA DBE. 2. Authors need to explain why ear swelling was not attenuated in Rag1-deficient ( Fig.1f) and FTY720-treated ( Fig.1g) mice compared to WT controls. It seems that authors just failed to induce adaptive immune response in these experiments. 3. Authors need to explain the interpretation of the result shown in Fig.1f. Does this result suggest SADBE-induced C HS response is mediated by NK cells but not by T/B cells? If so, is there an antigen-specificity in this response? 4. In Fig.1e, ear thickness increment looks over than 100% in SADBE-sensitized and SADBEchallenged group; however, the increment in same group looks less than 80% in Fig.1f and Fig.1g. Moreover, it was less than 50% in Fig. 6b. What causes these discrepancies? 5. Authors demonstrated that SADBE can directly activate TRPA1/TRPV1 channels. However, they did not present any data evaluating the indirect effect of SADBE; for instance, keratinocytes, mast cells, ILC s, which can be activated by SADBE might subsequently activate TRP channels, as authors described. 6. TRPA1 and TRPV1 play roles in SADBE-induced scratching behavior and ear swelling via Th1 cytokine production, respectively; although in-vivo data suggest that their function in calcium influx is compensable in response to SADBE. Authors need to discuss the mechanistic insight of this discrepancy. 7. Fig7: How about the expression level of IFN-gamma, the most important Th1 cytokine. 8. Fig8: In this paper, there is no data that demonstrate the involvement of DDC in SADBEinduced C HS. Therefore, DDC function shown in Fig . The novel feature of the current manuscript is that the authors show that SADBE directly activates both TRPV1 and TRPA1 ion channels that are expressed by sensory neurons. The evidence for this mode of action is good, although there are clearly other important actions of SADBE that are not explored. The finding that an electrophilic compound like SADBE activates TRPA1 is not surprising; it would have been surprising if it didn't. The data suggesting an action via binding to the putative capsaicin binding site in TRPV1 adds a more interesting observation. The current study also points to a role of sensory nerves and TRPV1 in regulating SADBE-evoked inflammation. This is perhaps the most interesting observation in the study but it is not investigated in depth. I think that there are aspects of the study that need clarification either by provision of further data or by stronger arguments.
1. SADBE activates both TRPV1 and TRPA1, which makes it highly unlikely that the behavioral effect of acute exposure is to induce only itch. Such stimulation will activate all nociceptive neurons and evoke pain, which is consistent with the results of Qu et al. 2014, 2015, which showed as much or more pain-like responses than itch-like responses after SADBE administration. Were pain-like responses measured and what happened in the Trpa1 -/-, Trpv1 -/and Trpa1 -/-/Trpv1 -/mice?
Response: Thanks for bringing up this important point. According to your suggestions, we have performed additional experiments investigating SADBE-elicited pain-like behavior. Surprisingly, we didn't find evident wiping behavior elicited by injections of 30 mM SADBE using the cheek model as shown in the revised manuscript ( Figure 5a). Furthermore, we didn't observe any paw licking or flicking behavior within the first 10 mins after intraplantar injections of 30 mM SADBE (data not shown).
To further exclude the possibility that SADBE contributes to mechanical and thermal pain sensitivities, we have performed Von Frey and Hargreaves tests after paw injections of 30 mM SADBE. As shown in Figures 1a and 1b below, we did not find significant differences in paw withdrawal threshold to mechanical stimuli or paw withdrawal latency to heat stimuli compared with mice injected with vehicle only. These results suggest that SADBE acts primarily as an itchevoking compound when applied acutely at the given dosage.
We further tested pain-like behavior in the SADBE-induced allergic contact dermatitis model. As shown in Figures 1e and 1f below, no wiping behavior was observed while scratching behavior gradually increased during the first 3 days when compared with the vehicle control group. However, excessive scratching produced excessive skin lesions at day 4 ( Figure 1c and 1d) where scratching behavior was markedly reduced while wiping behavior started to show up, suggesting that tissue damage and injury promote pain response and suppress itch response, which is consistent with previous findings that pain sensation constitutively suppresses itch sensation [1,2].
We used 20 µl of 0.5% SADBE compared with 25 µl of 1% SADBE used by Qu et al. in their original papers. Our protocol yielded a much better itch response because the low concentration of SADBE we used has delayed the development of pain behavior caused by excessive scratching and prolonged the chronic phase of itch. Taken together, these results suggest that SADBE does not directly elicit acute pain or chronic pain sensation in wt mice at the concentrations used in our studies. 2. The conclusions about the relative roles of TRPA1 and TRPV1 would be strengthened by the addition of pharmacological evidence that selective antagonists recapitulated the effects seen in the knockout mice. Deleting these channels (Trpa1 -/-, Trpv1 -/and Trpa1 -/-/Trpv1 -/mice) may have unexpected 'compensatory' or 'modifying' effects on physiological processes.
Response: Thanks for this constructive comment. Per your suggestion, we have performed additional experiments and tested the effect of a selective TRPA1 antagonist A967079 and a selective TRPV1 antagonist AMG517 on SADBE-induced CHS.
As expected, the number of scratching bouts was significantly reduced in mice treated with A967079, AMG517 or a combination of A967079 and AMG517 (A967079/AMG517) (Supplementary Figure 2a). Moreover, AMG517-treated mice showed a markedly increased edema when compared with vehicle-or A967079-treated mice (Supplementary Figure 2b). These results are consistent with results from genetic ablation studies and have further confirmed the distinct roles of TRP channels in SADBE-induced skin inflammation and persistent itch.
3. Scratching was not totally eliminated in the Trpa1 -/-/Trpv1 -/mice (Lines 211 -218). There was still a substantial response. What is the mechanistic basis for this residual pruritic effect -is it related to other actions on sensory neurons, perhaps increased action potential firing due to enhanced voltage gated sodium currents (Qu et al 2014;) or reduced potassium currents?
Response: Thanks for brining this up. To address this important question we have performed additional experiments and compared the number of scratches of the Trpa1 -/-/Trpv1 -/-dKO mice injected with either vehicle or SADBE. Although the SADBE group tended to scratch more when compared with vehicle group, there was no significant difference between these two groups ( Figure 2). Moreover, even 10 mM SADBE could not induce action potential firings in DRG neurons dissociated from the Trpa1 -/-/Trpv1 -/-dKO mice (please also see our the response to your Comment 9), we thus conclude that both TRPA1 and TRPV1 are required for SADBEinduced cellular and behavioral responses.  . The authors provide some evidence for cytokine changes to explain their findings but the mechanisms by which sensory neurons drive these changes and why this is linked to TRPV1 and not to TRPA1 are not addressed. I recommend that the authors add more details or amend wording to aid the reader.
Response: Thanks for your constructive suggestions. We have performed additional experiments and provided evidence for the involvement of dermal macrophages in TRPV1-mediated modulation of the SADBE-induced skin inflammation in the revised manuscript.

5.
Methods. Lines 439 -453, 459 -462 and e.g. Figure 1. A). The protocols for administration of SADBE need to be consistently explained in the methods and the figures. The diagrams in the figures show SADBE administered first to the abdomen and then to the ear. But the scratching behavior is measured after SADBE is administered to either the nape of the neck. B) The timing of SADBE administration when measuring ear edema is unclear. The method (line 461) states measurements made after the challenge at day 3, but the protocol shown in figure 1 shows measurement 3 days after the last SADBE treatment. Do you mean 'measurements were made 3 days after the last SADBE challenge?' Response: A) Thanks for pointing out this important issue. We apologize for not making this clear in our original submission. Mice were all challenged to their ears in our experiments. We have corrected this description in the revised manuscript. B) Sorry for the confusion. As shown in the schematic experimental protocol, we did the measurements 3 days after the last SADBE challenge. We have refined our statement in the revised manuscript. Response: Thanks for pointing this out. We have added the methods of electrophysiology in the revised manuscript.
8. Line 111. Was there any measurement to show that the FTY720 treatment was effective in blocking T cell migration?
Response: Thanks for bringing this up. As an immunosuppressant drug, FTY720 has been widely used to inhibit lymphocyte emigration from lymphoid organs. By downregulating Sphingosine-1-phosphate receptor 1 (S1P 1 ), the mechanism of FTY720-induced lymphocyte sequestration was reported by a Nature paper in 2004. [1]. So far, FTY720 is commonly used in immunity study of effector T cells and even proved by FDA in the treatment of T cell-mediated autoimmune diseases. By searching references, we also found that the effect of FTY720 is consistent among different groups [2][3][4] and some of these studies did not further confirm the effect of FTY720 in blocking T cell migration [5]. 9. Lines 155 -164. The results shown in Figure 3, panels a-d need to be expressed quantitatively. How many neurons were studied, how many neurons responded to SADBE , how many coverslips? What were the percentages -the term 'similar percentages' is vague? The 3mM SADBA concentration used in these DRG experiments is a sub-maximally effective concentration for TRPV1 activation in the heterologous expression experiments (see Figure 4g). From Fig 3b it looks as though there are many TRPV1 expressing neurons that did not respond to SADBE. Were there SADBE responsive neurons that were not AITC or capsaicin sensitive in DRG neurons from wild-type, Trpa1 -/or Trpv1 -/mice? What was the percentage of responding neurons in the double knockout (Trpa1 -/-/Trpv1 -/-) DRG neurons -was it really 0%?
Response: Thanks for your suggestions. To address your questions, we have performed additional experiments and also added detailed description for Figure 3 in the revised manuscript.
As suggested, we tried higher concentration of SADBE (10 mM) in both patch-clamp recordings and calcium imaging assays. Robust depolarization and spontaneous action potential firings were elicited by 10 mM SADBE in current clamp recordings, which desensitized the DRG neurons isolated from wt mice and blunted the AITC and capsaicin responses in the same cells (Figure 3a and 3b). Similar results were also found in calcium imaging experiments (Figure 3e), i.e. neither AITC nor capsaicin evoked significant calcium response after application of 10 mM SADBE. Furthermore, we compared the percentage of wt DRG neurons responded to 3 mM and 10 mM SADBE. Although there was a trend of increase in the number of neurons activated by 10 mM SADBE, there was no significant difference between the 3 mM and 10 mM groups (Figure 3g). Interestingly, a subpopulation of capsaicin-sensitive DRG neurons showed tiny or no calcium influx in response to 3 mM SADBE in calcium imaging assays (see Figure 3b in the revised manuscript), which might be due to its weaker potency to activate TRPV1.
To further confirm no other receptors other than TRPA1 and TRPV1 were involved in SADBE activation of sensory neurons, we used 10 mM SADBE to fully activate DRG neurons. As shown in Figure 3c, 3d and 3f below, no membrane depolarization, action potential firing or calcium influx was found in both current-clamp recordings and calcium imaging assays in response to 10 mM SADBE in DRG neurons isolated from Trpa1 -/-/Trpv1 -/-dKO mice, suggesting that TRPA1 and TRPV1 are the only SADBE receptors in mouse DRG neurons. Response: Thanks for bringing up this interesting phenomenon. We did observe two different patterns of calcium responses mediated by TRPA1 or TRPV1 channels. TRPA1 channel is continuously activated while TRPV1 is quickly desensitized upon activation in calcium imaging experiments, which might contribute to the weaker potency of SADBE activation of TRPV1.
Unfortunately, based on these results we can't further interpret whether these two patterns lead to any behavior differences. Figure 4. The SADBE-evoked responses of hTRPA1 mutants are depressed, but is this due to a reduced expression level? For human TRPA1 3C mutants, responses to a non-electrophilic agonist such as carvacrol would demonstrate that the channels were well expressed.

Lines 191 -194 &
Response: Thanks for your suggestions. (1) As reported by other groups [1][2], both three cysteine and lysine residues are important for activation of TRPA1 by reactive compounds through covalent modification, which can profoundly affect channel gating and channel conformation. Thus, these mutants might show depressed or no response to different reactive compounds when compared with native TRPA1 channel. However, the expression patterns are similar among cysteine mutants and wildtype TRPA1 [1]. In our case, three cysteines play a major role in mediating sensitivity to SADBE while TRPA1-K also showed a 4.6-fold reduction in EC 50 . These factors may contribute to the depressed response to SADBE in TRPA1 mutants.
(2) As suggested, we used flufenamic acid (FFA) as a positive control in patch clamp recording as our previous studies showed that FFA is a non-electrophilic agonist of TRPA1 [3]. Consistent with our previous results, SADBE barely evoked inward-and outward-currents while FFA strongly activated TRPA1-3C channel (Figure 4), confirming the transfection efficiency.   Figure 4. The concentration response curve for the TRPV1 M548L mutant appears to be shifted leftwards compared to the wild-type TRPV1, yet the authors say there was no effect. Some quantification of results and evidence of reproducibility should be added to justify the conclusion.

Lines 197 -198 and
Response: Sorry for the confusion. We agree with this reviewer that the potency of SADBE on M548L was significantly increased when compared with wt (Supplementary Table 2 in revised manuscript), which may be due to the enhanced interaction between SADBE and Leucine. In the original manuscript, we meant that SADBE potency was not decreased by substitution of R115 and M548. We have corrected the description and added the quantification results in the revised manuscript (Supplementary Table 1 and Table 2 in the revised manuscript).
Response: Thanks. We have revised this sentence according to your suggestion.

Reviewer #2 (Remarks to the Author):
This study shows distinct contributions of TRP channels to skin inflammation and itch in the SADBE-induced contact dermatitis model. By using TRPA1 knockout (KO), TRPV1 KO and TRPA1/TRPV1 double KO mice, the authors demonstrate that persistent scratching caused by SADBE treatment is mediated through both TRPA1 and TRPV1 channels. Calcium imaging and electrophysiological experiments using acutely dissociated DRG cells and HEK cells transfected with TRPA1 or TRPV1 show that SADBE can activate directly both TRP channels. Unlike persistent scratching induced by repeated application of SADBE, acute scratching induced by intradermal SADBE injection is mediated mainly through TRPA1. In contrast to scratching behavior, deficiency or pharmacological inhibition of TRPV1 exacerbates skin inflammation probably through promoting Th1 cytokines. These findings strongly suggest that SADBE causes itch independent of skin inflammation. Although most experiments were appropriately conducted and the manuscript is well organized, there remain a number of questions that should be addressed.
1. The authors found an increase in scratching at 3 days after the final SADBE challenge in sensitized mice (Fig. 1), thereby regarding this scratching as "persistent" itch. On the other hand, they show that intradermal injection of SADBE elicited "acute" scratching in naïve mice (Fig. 5). I strongly recommend that the authors should present the time course of scratching after SADBE application in both the "persistent" and "acute" model, and also should examine the effects of TRP channel deficiency on scratching at different time points. Perhaps, increased scratching will be observed immediately after the final application of SADBE in sensitized mice; does the deficiency of TRPA1 alone or both TRPA1 and TRPV1 inhibit this scratching response? Also, how long does the scratching induced by SADBE injection in naïve mice continue? The present results raise an interesting question whether persistent scratching is due to continuation of direct activation of TRPA1 and/or TRPV1 by SADBE existing in the skin, or whether other mediator(s) induced by repeated SADBE application eventually activate TRPA1 and TRPV1 channels.
Response: Thanks for your valuable suggestions. As recommended, we have recorded the time course of SADBE-induced acute and persistent itch in wt and mice deficient in TRPA1, TRPV1, and both TRPA1 and TRPV1.
In the acute itch model, scratching behavior was immediately observed after intradermal injection of SADBE and gradually diminished 30 min later in wt mice. As expected, Trpa1 -/and dKO but not Trpv1 -/mice showed markedly reduced scratching behavior in response to SADBE injection, suggesting that TRPA1 is required in SADBE-induced acute itch (Figure 5a).
In the model of SADBE-induced persistent itch, as the reviewer envisioned, wt mice showed a robust scratching response within the first hour after SADBE application presumably due to direct activation of TRPA1/V1 channels by SADBE. Although the number of scratching bouts was significantly reduced by more than half, which may due to the potential TRP channel desensitization and hapten clearance by immune response, itch sensation remained steady during the first 24 hours and dramatically increased at day 2 and day3 ( Figure 5b). Interestingly, we also found that other mediators (such as 5-HT) released by activated innate immune cells contribute to the increased itch sensation in the later phase of SADBE-induced CHS (manuscript accepted by J. Allergy Clin. Immunol).
Taken together, SADBE could induce persistent itch sensation by directly activating TRP channels and/or indirectly inducing release of endogenous pruritogens from innate immune cells. Nevertheless, TRPA1 and TRPV1 are the dominant downstream targets in both pathways as genetic ablation of TRPA1/TRPV1 significantly reduced SADBE-induced persistent itch sensation.  Figs. 1e and i, SADBE challenge on the nape of the neck causes scratching, irrespective of systemic presensitization on the abdominal skin, although skin inflammation is significantly but very weakly promoted by the presensitization. Therefore, SADBE causing symptoms, especially itch (or scratching), should be considered as irritant contact dermatitis (ICD), but not allergic contact dermatitis (ACD).

As shown in
Response: We strongly agree with this reviewer and added new paragraphs in the discussion. In SADBE-induced CHS, both allergic and irritant contact dermatitis could occur but persistent itch is a typical irritant contact dermatitis response. On the other hand, we could not rule out the possibility that endogenous pruritogens including many cytokines released in the immunemediated inflammation process could also contribute to the persistent itch.
3. The authors showed that intradermal SADBE injection in naïve mice exclusively caused scratching compared to wiping behaviors (Fig. 5); however, each behavior should be compared to when the vehicle was injected. Scratching and wiping behaviors should not be compared to each other. Furthermore, Qu et al. have reported that repeated SADBE application induces not only itch-related scratching but also pain-like behaviors, such as wiping and licking (Qu L, Brain, 2014). As described, both TRPA1 and TRPV1 play a critical role in pain. Have you checked pain behaviors in your SADBE model? How does the TRP channel deficiency affect pain caused by SADBE treatment?
Response: Thanks for your suggestions. We have modified Figure 5 in the revised manuscript. For the pain-related behavior, please also see our response to Reviewer #1's Comment 1.
We have performed additional experiments investigating SADBE-elicited pain-like behavior. Surprisingly, we didn't find obvious wiping behavior by injection 30 mM SADBE with cheek model as shown in the revised manuscript (Figure 5a). Furthermore, we didn't observe any paw licking or flicking behavior within the first 10 mins after imtraplantar injection of 30 mM SADBE (data not shown). To further exclude the possibility that SADBE contributes to mechanical and thermal pain sensitivities, we have performed Von Frey and Hargreaves tests after paw injection of 30 mM SADBE. As shown in Fig 1a and 1b (response to Reviewer #1's Comment 1), we did not find significant difference in paw withdrawal threshold to mechanical stimuli or paw withdrawal latency to heat stimuli compared with mice injected with vehicle only. These results suggested that SADBE acts primarily as an itch-evoking compound when applied acutely at the given dosage.
We further tested pain-like behavior in SADBE-induced allergic contact dermatitis model. As shown in Figure 1e and 1f (response to Reviewer #1's Comment 1), no wiping behavior was observed while scratching behavior gradually increased during the first 3 days when compared with the vehicle control group. However, excessive scratching produced excessive skin lesions at day 4 (Figure 1c and 1d, response to Reviewer #1's Comment 1) where scratching behavior was markedly decreased while wiping behavior began to show up, suggesting that tissue damage and injury promotes pain response but suppresses itch response. This finding is consistent with previous findings that pain sensation constitutively suppresses itch sensation [1,2].
We used 20 µl of 0.5% SADBE compared with 25 µl of 1% SADBE used by Qu et al. in their original papers. Our protocol yielded a much better itch response because the low concentration of SADBE we used delayed the development of pain behavior caused by excessive scratching and prolonged the chronic itch phase. Taken these together, we suggest that SADBE elicits neither acute pain nor chronic pain sensation directly in wt mice at the concentrations we used in our studies. (1) 0.5% SADBE was used in the induction of SADBE-induced CHS. The molar concentration is about 23 mM. So we decided to use 30 mM SADBE for the intradermal injection. For in vitro experiments, we have tried different concentrations and found that 3 mM was the most appropriate concentration because it was sufficient to activate DRG neurons without producing a robust desensitization (please also see response to Reviewer #1's comment # 9).
(2) We have added the description in the revised method section (please also see our response to Reviewer #1's comment #6). 6. Calcium imaging and electrophysiological experiments clearly showed mutual compensatory effect in the DRG neurons isolated from TRPA1 KO or TRPV1 KO mice (Fig. 3). On the other hand, SADBE-induced persistent scratching could be suppressed in either TRPA1 KO or TRPV1 KO mice (Fig. 2). If DRG neurons play a dominant role in SADBE-induced persistent itch as shown in Fig. 8, these results would be theoretically inconsistent. Is it possible that TRPA1and/or TRPV1-expressed in other cells besides DRG neurons is involved in the persistent itch? The authors should fully address this question.
Response: Thanks for bringing up this question. Functional interaction between TRPV1 and TRPA1 occur in several ways [1] and accumulating evidence suggests TRPA1 and TRPV1 assemble channel complexes in heterologous expression systems and sensory neurons [2][3][4]. Weng et al also proposed a TRPA1-TRPV1 complex model and revealed how this complex contributed to persistent pain sensation under the control of Tmem100 [5]. In brief, TRPA1 activity could be potentiated by TRPV1 in the presence of Tmem100. However, TRPA1-or TRPV1-mediated currents were partially inhibited by Tmem100 in DRG neurons isolated from Trpv1 -/-or Trpa1 -/-mice, respectively. This theory could also be applied to our findings in TRPA1-and TRPV1-mediated persistent itch. Although TRPA1 or TRPV1 still mediated calcium influx in Trpv1 -/-or Trpa1 -/-mice, we did observe reduced amplitude in calcium responses in DRG neurons from Trpa1 -/-and Trpv1 -/-mice (Fig 3a  versus Fig 3b and 3c in the revised manuscript). The compromised channel function may lead to the reduced scratching in the TRP channel deficient mice.
It remains controversial whether TRPA1 and TRPV1 channels are expressed in nonneuronal cells [6][7][8][9][10]. To exclude the possibility that functional TRP channels expressed by keratinocytes and immune cells contribute to SADBE-induced itch, we performed calcium imaging experiments with skinresident cells dissociated from the ear preparations of wt mice. Neither AITC nor capsaicin evoked measurable calcium responses (Fig 6), suggesting it is unlikely that the SABDE-induced persistent itch is mediated by TRPA1 and TRPV1 channels that are expressed by nonneuronal cells. 7. The responses to SADBE in TRPV1-expressing HEK293 cells were clearly weaker than those in TRPA1-expressing cells (Fig. 4). Indeed, the EC50 value for TRPV1-expressing cells was 5.6 times higher than that for TRPA1 (1.30 mM for TRPA1 vs. 7.26 mM for TRPV1). Why is that? Also, what concentration of SADBE did you use? If it was 3 mM similar to other in vitro experiments, the authors should use a higher (or submaximal) concentration (e.g., 10 mM) to determine the SADBE activation of TRPV1.
Response: Although TRPV1 and TRPA1 channels are structurally related, non-selective cation channels; these two channels possess different primary sequences and properties. Therefore, it is not surprising that they show different affinities against different chemicals. For instance, capsaicin and AITC selectively activate TRPV1 and TRPA1, respectively. In our case, SADBE possesses different potencies to activate TRPA1 and TRPV1, which might result from distinct intermolecular force and/or interaction mode caused by special residues between these two channels, for instance, cysteine and lysine residues in TRPA1 and tyrosine, serine, and threonine residues in TRPV1.
For in vitro experiments, 3 mM SADBE was used to activate TRPV1 because a higher concentration of SADBE (10 mM) excessively activated and subsequently desensitized DRG neurons in both calcium imaging assays and patch clamp recordings, which blunts AITC and capsaicin responses in the same cells (please also see Fig 3 in response to Reviewer #1's comment # 9). Therefore, we used 3mM but not 10 mM SABDE to classify DRG neurons together with capsaicin and AITC. 8. The authors suggest that increased Th1 cytokines would contribute to aggravation of SADBEinduced skin inflammation by TRPV1 deficiency. On the other hand, SADBE-induced skin inflammation was not affected in Rag1-/-mice, which lack T cells (Fig. 1f), suggesting that T cells are hardly required for SADBE-induced skin inflammation. Also, since the authors have not examined the effect of NK cell depletion alone on skin inflammation, the role of NK cells and T cells in skin edema by SADBE remains unclear. They thus should further address this question and revise the working model (Fig. 8) to fit their findings.
Response: Thanks for your suggestions. We have tested the function of NK cells in SADBEinduced inflammation by depletion of NK cells using anti-NK1.1 antibodies and we found that depletion of NK cells alone didn't alleviate SADBE-induced skin edema (Fig 1g in the revised  manuscript), which is consistent with a previous report [1]. On the other hand, Rag-1 -/mice receiving anti-NK1.1 antibody displayed a significantly decreased skin edema in response to SADBE challenges (Fig 1h in the revised manuscript), suggesting neither T/B cell nor NK cell priming alone is sufficient to mediate SABDE-induced inflammation but deficiency in all lymphocytes is effective in reducing SADBE-induced skin inflammation. We have already modified our working model according to these new findings. Jing et al. showed that the SADBE-induced persistent itch was not depend on lymphocytes, but mediated by TRPA1 and TRPV1channels. They demonstrated SADBE can directly activate both TRPA1 and TRPV1 in vivo using freshly isolated DRG cells and TRPA/TRPV1-expressing HEK293 cells. Further, their observation suggests that TRPV1 also affect SADBE-induced ear swelling via inhibiting the production of Th1 cytokines. In this paper, data were well-presented; however, I had concerns relating to the interpretation of some of the results, and lack of mechanistic insight. Followings are my specific comments: 1. Fig.1e showed that irritant response (innate response) by SADBE challenge was very strong with this experimental protocol. It makes hard to evaluate adaptive immune response in this condition. Author need to modify the protocol to reduce irritant response to SADBE.
Response: Thanks for your suggestions. In this manuscript, we aimed to determine the distinct roles of TRP channels in persistent itch and inflammation. As we showed in Figure 1e and 1i in the manuscript, SADEB challenges led to a significantly increased inflammatory response that was mediated by cutaneous immune cells.
On the other hand, SADBE-induced persistent itch is mainly an irritant response that is mediated by TRP channels. Moreover, TRPV1 modulates skin inflammation. We have tried three different concentrations of SADBE (0.25%, 0.5%, and 1%) in order to find the best to generate a moderate contact dermatitis that we can use to investigate the mechanisms of skin inflammation (mediated by immune cells) and persistent itch (mediated by TRP channels).
We found that 0.25% SADBE was unable to induce an irritant contact dermatitis (Fig 7) while 1% SADBE-induced response was too strong to investigate the function of TRP channels (immune cells were excessively activated and TRP channels were desensitized). On the other hand, 0.5% SADBE produced a moderate skin inflammation associated with persistent itch which we used in our studies to effectively determine the roles of TRP channels in SADBE-induced CHS. 2. Authors need to explain why ear swelling was not attenuated in Rag1-deficient (Fig.1f) and FTY720-treated (Fig.1g) mice compared to WT controls. It seems that authors just failed to induce adaptive immune response in these experiments.
Response: Thanks for brining this important point. The induction of CHS was effective as these mice displayed significantly increased inflammation and spontaneous scratching (see Figure 1f and 1g in the revised manuscript). In a Nature Immunology paper [1], the authors also showed an intact cutaneous CHS in Rag2-deficient mice and suggested that NK cells are both necessary and sufficient to mediate a potent CHS response in the absence of other adaptive lymphocytes. We have also tried to induce CHS response in wt mice treated with anti-NK1.1 antibodies or control IgG antibodies. NK cell depletion alone did not abolish CHS response in wt mice, which can still develop a T-cell dependent CHS. On the other hand, Rag1-deficient mice treated with anti-NK1.1 antibodies had a reduced inflammation (Figure 1h in revised manuscript). Taken together, both T/B cells and NK cells are likely involved in mediating the antigen-specific adaptive recall response after SADBE challenge. [1] O'Leary, Jacqueline G., et al. "T cell-and B cell-independent adaptive immunity mediated by natural killer cells." Nature immunology 7.5 (2006): 507. 3. Authors need to explain the interpretation of the result shown in Fig.1f. Does this result suggest SADBE-induced CHS response is mediated by NK cells but not by T/B cells? If so, is there an antigen-specificity in this response?
Response: Sorry for the confusion. We have made changes in Fig 1 of the revised manuscript to emphasize the functions of T/ B cell and NK cell in the revised manuscript.
4. In Fig.1e, ear thickness increment looks over than 100% in SADBE-sensitized and SADBEchallenged group; however, the increment in same group looks less than 80% in Fig.1f and Fig.1g. Moreover, it was less than 50% in Fig. 6b. What causes these discrepancies?
Response: Thanks for bringing up this question. We used mice of 8-weeks of age with a body weight of 20-23g in our behavior testing. However, the Trpa1 -/-/Trpv1 -/-dKO mice were lighter at 8-weeks of age when compared with wt, Trpa1 -/-, and Trpv1 -/mice. We thus used the Trpa1 -/-/Trpv1 -/-dKO mice at 12 weeks of age when they had a comparable body weight. Similarly, the Rag1-deficient mice were used at about 10 weeks of age. Although we do not understand the mechanisms underlying the relationship between age and immunity, it seems the severity of inflammation decreased as the mice matured. On the other hand, these discrepancies of the inflammation at different ages shouldn't affect our conclusions as mice were strictly matched with their littermate control groups in our experiments. 5. Authors demonstrated that SADBE can directly activate TRPA1/TRPV1 channels. However, they did not present any data evaluating the indirect effect of SADBE; for instance, keratinocytes, mast cells, ILCs, which can be activated by SADBE might subsequently activate TRP channels, as authors described.
Response: Thanks for bringing up this important point. In this manuscript, we mainly focused on the distinct roles of TRP channels in SADBE-induced inflammation and persistent itch. The classic hapten-induced inflammatory response mediated by adaptive immune system has already been well studied. Moreover, we did not find functional expression of TRPA1/TRPV1 by keratinocytes and innate immune cells in the skin (see also Fig 6 in the response to Review #2's Comment 6). Therefore, only the functions of TRPA1/TRPV1 in primary sensory neurons were included in this manuscript. However, we do not exclude the possibility that other TRP channelexpressed resident skin cells might be involved in SADBE-induced inflammatory response and persistent itch.
6. TRPA1 and TRPV1 play roles in SADBE-induced scratching behavior and ear swelling via Th1 cytokine production, respectively; although in-vivo data suggest that their function in calcium influx is compensable in response to SADBE. Authors need to discuss the mechanistic insight of this discrepancy.
Response: Thanks for your suggestions. Functional interaction between TRPV1 and TRPA1 occur in several ways [1] and accumulating evidence suggests TRPA1 and TRPV1 assemble channel complexes in heterologous expression systems and sensory neurons [2][3][4]. Weng et al also proposed a TRPA1-TRPV1 complex model and revealed how this complex contributed to persistent pain sensation under the control of Tmem100 [5]. In brief, TRPA1 activity could be potentiated by TRPV1 in the presence of Tmem100. However, TRPA1-or TRPV1-mediated currents were partially inhibited by Tmem100 in DRG neurons isolated from Trpv1 -/-or Trpa1 -/-mice, respectively. This theory could also be applied to our findings in TRPA1-and TRPV1-mediated persistent itch. Although TRPA1 or TRPV1 still mediated calcium influx in Trpv1 -/-or Trpa1 -/-mice, we did observe reduced amplitude in calcium responses in DRG neurons from Trpa1 -/-and Trpv1 -/-mice (Fig 3a  versus Fig 3b and 3c in the revised manuscript). The compromised channel function may lead to the reduced scratching in the TRP channel deficient mice.
It remains controversial whether TRPA1 and TRPV1 channels are expressed in nonneuronal cells [6][7][8][9][10]. To exclude the possibility that functional TRP channels expressed by keratinocytes and immune cells contribute to SADBE-induced itch, we performed calcium imaging experiments with skinresident cells dissociated from the ear preparations of wt mice. Neither AITC nor capsaicin evoked measurable calcium responses (Fig 6), suggesting it is unlikely that the SABDE-induced persistent itch is mediated by TRPA1 and TRPV1 channels that are expressed by nonneuronal cells.