The CysLT₂R receptor mediates leukotriene C₄-driven acute and chronic itch

Significance

Interactions between the nervous system and immune system are central regulators of chronic itch, a key feature of pathologies like atopic dermatitis and allergic contact dermatitis. Cysteinyl leukotrienes (LTC₄, LTD₄, and LTE₄) are eicosanoid lipids known for mediating inflammation, bronchoconstriction, and vascular leakage. We demonstrate here that CysLTs are potent itch inducers and that this effect depends on the specific coupling of LTC₄ with its receptor CysLT₂R, which is expressed in a population of peripheral sensory neurons in the mouse and in human. We show that the LTC₄/CysLT₂R pathway contributes to a model of chronic itch, suggesting that CysLT₂R could be a new therapeutic target for intractable chronic itch.

Abstract

Acute and chronic itch are burdensome manifestations of skin pathologies including allergic skin diseases and atopic dermatitis, but the underlying molecular mechanisms are not well understood. Cysteinyl leukotrienes (CysLTs), comprising LTC₄, LTD₄, and LTE₄, are produced by immune cells during type 2 inflammation. Here, we uncover a role for LTC₄ and its signaling through the CysLT receptor 2 (CysLT₂R) in itch. Cysltr2 transcript is highly expressed in dorsal root ganglia (DRG) neurons linked to itch in mice. We also detected CYSLTR2 in a broad population of human DRG neurons. Injection of leukotriene C₄ (LTC₄) or its nonhydrolyzable form NMLTC₄, but neither LTD₄ nor LTE₄, induced dose-dependent itch but not pain behaviors in mice. LTC₄-mediated itch differed in bout duration and kinetics from pruritogens histamine, compound 48/80, and chloroquine. NMLTC₄-induced itch was abrogated in mice deficient for Cysltr2 or when deficiency was restricted to radioresistant cells. Itch was unaffected in mice deficient for Cysltr1, Trpv1, or mast cells (W^Sh mice). CysLT₂R played a role in itch in the MC903 mouse model of chronic itch and dermatitis, but not in models of dry skin or compound 48/80- or Alternaria-induced itch. In MC903-treated mice, CysLT levels increased in skin over time, and Cysltr2^−/− mice showed decreased itch in the chronic phase of inflammation. Collectively, our study reveals that LTC₄ acts through CysLT₂R as its physiological receptor to induce itch, and CysLT₂R contributes to itch in a model of dermatitis. Therefore, targeting CysLT signaling may be a promising approach to treat inflammatory itch.

Itch, or pruriception, is defined as an uncomfortable sensation that triggers the desire to scratch, and is mediated by peripheral sensory neurons termed pruriceptors (1, 2). Increasing evidence indicates that inflammatory mediators are released into the skin by immune cells and other cell types that directly activate or sensitize pruriceptors to produce itch. Chronic itch is a debilitating symptom of many skin pathologies, including atopic dermatitis (AD) and allergic contact dermatitis (ACD), and the roles of individual molecular pathways in chronic itch are not clearly defined. While certain classes of mediators such as histamine, proteases, and cytokines have been investigated more recently, less is known about the roles of lipid mediators in itch. There is a great need for better understanding of the molecular mechanisms leading to itch and to develop novel therapeutic modalities to treat itch.

Leukotrienes (LTs) are eicosanoid lipid mediators generated upon activation of both immune and structural cells. LTs are comprised of LTB₄ and the cysteinyl LTs (CysLTs; LTC₄, LTD₄, and LTE₄; Fig. 1A). They were named “leukotrienes” to highlight their originally defined source: leukocytes, including mast cells (MCs), eosinophils, basophils, and macrophages (3). More recently, platelet–neutrophil aggregates (4) and tuft cells (5), which are specialized epithelial cells, were identified as potent producers of CysLTs, highlighting the ubiquitous and versatile sources of CysLTs during inflammation. The biosynthesis of LTs begins when arachidonic acid is liberated from membrane phospholipids and is converted into LTA₄ by the enzyme 5-lipoxygenase (5-LO) in the presence of the 5-LO–associated protein (FLAP; Fig. 1A). LTA₄ hydrolase processes LTA₄ into LTB₄, which binds to the LTB₄ receptors. LTC₄ synthase (LTC₄S), at the outer nuclear membrane, conjugates LTA₄ with reduced glutathione to produce LTC₄, the first and only intracellular CysLT (6). LTC₄ is rapidly transported extracellularly (within minutes) and converted sequentially by membrane-bound γ-glutamyl transferases and dipeptidases to LTD₄ and LTE₄. CysLTs exert their effects through three G protein-coupled receptors (GPCRs)—CysLT₁R, CysLT₂R, and CysLT₃R—with different affinities. CysLT₁R and CysLT₂R, the receptors for the short-lived LTC₄ and LTD₄, are widely expressed in hematopoietic and structural cells. The stable end metabolite LTE₄ binds to the epithelial CysLT₃R and mediates mucin release in response to the airborne fungus Alternaria (7, 8). Pharmacologic studies using heterologous transfectants indicated that CysLT₁R is the high-affinity receptor for LTD₄ and binds LTC₄ with lesser affinity. By contrast, CysLT₂R binds LTC₄ and LTD₄ at equimolar concentrations (9, 10). However, in vivo selectivity for LTC₄ and LTD₄ differ depending on the tissue distribution, frequency of CysLT receptor-expressing cells, and proximity to the ligand source (11–14). The brief half-life of LTC₄ within the tissue likely requires close proximity between the LTC₄ source and the target cell. CysLTs are potent inducers of airway smooth muscle constriction, vascular permeability, leukocyte recruitment, and chemokine production (3). Classically, the proinflammatory effects of CysLTs are attributed to the CysLT₁R receptor, which participates in the recruitment of eosinophils, the activation of ILC2s, and Th2 response during airway inflammation (12, 15, 16). The CysLT₁R-specific antagonist montelukast is widely used to treat bronchoconstriction and inflammation in asthmatic patients (17, 18). CysLT₂R is resistant to montelukast, and its role is not understood as well as CysLT₁R. CysLT₂R signaling in platelets is required for type 2 lung inflammation and MC activation (19). Additionally, CysLT₂R has a newly defined role in lung metastasis through an effect on angiogenesis (20). Eosinophil-derived LTC₄ also mediates skin fibrosis and inflammation through CysLT₂R in an ovalbumin-induced mouse model of AD (14). Several groups have worked to elucidate the role of LTB₄ and its receptors in pain and itch (21–24). However, the role of CysLTs and their associated receptors in itch has not been well studied. Recent transcriptome studies suggest Cysltr2 to be expressed in sensory neurons, but how CysLT₂R regulates neural responses and itch behavior in vivo has not been clarified.

Fig. 1.

Cysltr2 is expressed on a subset of DRG sensory neurons. (A) Diagram of the LT pathway. (B and C) Expression of selected transcripts (B, related to itch; C, LT receptors) in mouse DRG neuron populations clustered into functional subsets based on single-cell RNA-seq data. Full dataset and methods are available in a previous study (32). (D) Triple/double-label ISH done with RNAscope in DRG (Left, Cysltr2, red; Nppb, blue; Hrh1, green; Middle, Cysltr2, red; MrgprA3, green; Mrgprd, blue; Right, Cysltr2, red; Trpv1, green). (Scale bar: 50 µm.) (E and F) ISH quantification: (E) Venn diagram representing overlap between markers and (F) percentages of Cysltr2⁺ neurons within subsets defined by other markers (n = 3 mice). (G) Representative images of human DRG labeled with RNAscope ISH for CYSLTR2 (red), NPPB (green), and TRPV1 (blue) and costained with DAPI (cyan). Lipofuscin (globular structures) that autofluoresced in all three channels and appear white in the overlay image were not analyzed, as this is background signal that is present in all human nervous tissue. (Scale bar: 50 μm.) (H) Pie chart displaying the distribution of CYSLTR2 neuronal subpopulations in human DRG. (I) Percentage of all human sensory neurons expressing CYSLTR2 transcript (Top) and of TRPV1 neurons and NPPB neurons coexpressing CYSLTR2 (Bottom). Values presented as mean ± SEM.

Pruriceptors are primary sensory neurons that mediate itch, whose cell bodies reside in the dorsal root ganglia (DRG) and trigeminal ganglia (25). Pruriceptors express molecular receptors at their peripheral nerve terminals in epidermal layers of the skin that respond to pruritogens including histamine, cytokines, and proteases. HRH1 and HRH4 are the major histamine receptors linked to pruriceptor activation and itch (26, 27). IL-4, IL-13, and IL-31 are cytokines that mediate itch via their cognate receptors expressed by pruriceptors (28, 29). Chloroquine (CQ) and BAM8-22, an endogenous peptide derived from proenkephalin, act via the Mas-related G protein receptors MrgprA3 and MrgprC11, respectively, to produce itch (30). Recent transcriptomic analyses have revealed distinct subsets of DRG neurons that may correspond to distinct pruriceptor subtypes and function, known as NP1, NP2, and NP3 neurons (31, 32). NP1 neurons express Mrgprd that responds to β-alanine to produce itch (33); NP2 neurons express Mrgpra3 and Mrgprc11 (30); NP3 neurons express Il31ra and Osmr (Oncostatin M receptor), which bind the pruritogenic cytokine IL-31 (28). NP3 neurons are also characterized by expression of neuropeptides Nppb and Sst, which have both been functionally linked to neurotransmission in itch (34–36).

In our molecular analysis as well in published datasets from other groups, we find that Cysltr2 is highly enriched in the NP3 subset of DRG neurons (31, 32). However, the functional role of CysLT₂R in pruriception, and the role of specific CysLTs (LTC₄, LTD₄, LTE₄) in itch, are unknown. Given that CysLTs are characteristic of type 2 immune cell activation in tissues during allergies, neuronal CysLT₂R could allow the immediate response and induction of itch coupled to allergic-type inflammation.

In this study, our goal was to determine the functional role of specific CysLTs and CysLT₂R in itch induction and skin inflammation. We find that Cysltr2 is enriched in pruriceptor-lineage DRG sensory neurons in mouse and is also expressed in human DRG neurons. LTC₄, but not LTD₄ or LTE₄, induces acute itch behaviors in mice that differ in duration and quality from other pruritogens. We demonstrate that LTC₄-induced itch is dependent on the CysLT₂R receptor. Bone-marrow chimeras show that radioresistant cells expressing CysLT₂R are necessary for LTC₄-induced itch. Levels of LTC₄ are elevated in the late stage of a mouse model of chronic dermatitis and itch, and Cysltr2^−/− mice have a decreased scratching phenotype at this time point. By contrast, CysLT₂R did not mediate itch in other inflammatory skin models that we tested. Overall, our findings show that CysLTs play a role in itch in acute and chronic situations by acting through the CysLT₂R receptor.

Results

Expression of Cysltr2 Receptor in a Pruriceptive Subset of DRG Neurons.

Molecular and genetic analysis of DRG neurons has recently identified distinct neurons linked to itch and other somatosensory functions (31, 37, 38). We previously performed single-cell profiling of FACS-sorted Nav1.8-lineage and Parvalbumin-lineage mouse DRG neurons (39). We observed a population cluster (VI) from Nav1.8-lineage neurons with enriched levels of itch-associated transcripts including Il31ra, the receptor for the cytokine IL-31 that drives pruritus (28, 40), and Nppb, a neuropeptide that mediates neurotransmission of itch signaling from DRG to the spinal cord (36). We found that Cysltr2, the receptor for LTC₄, was highly expressed in the same neuronal subset (SI Appendix, Fig. S1A). Other recently published single-cell RNA sequencing datasets of mouse DRG neurons revealed similar expression of Cysltr2 in a population of neurons expressing Il31ra, Nppb, Hrh1, and Sst, termed NP3 neurons (31, 32, 38, 41). The Cysltr2⁺ population is distinct from neurons expressing Mrgpra3 or Mrgprd (Fig. 1B), markers of NP2 and NP1 neurons, respectively (42).

CysLTs are synthesized as part of arachidonic acid metabolism, where LTA₄ is processed by LTC₄S into LTC₄, which is subsequently metabolized to LTD₄ and LTE₄ (Fig. 1A). These ligands bind with different affinities to the receptors CysLT₁R, CysLT₂R, or CysLT₃R (Oxgr1). Our analysis of the publicly available mouse RNA-sequencing (RNA-seq) dataset showed that Cysltr2 was the only CysLT receptor expressed in sensory neurons, as transcripts for Cysltr1 and Oxgr1 were not detected (Fig. 1C). Ltb4r1 and Ltb4r2, the receptors for LTB₄, were absent from pruriceptors but highly expressed in tyrosine hydroxylase (TH) neurons, a population of unmyelinated low-threshold mechanoreceptors (C-LTMRs) characterized by the expression of TH and associated with pleasant touch; Ltb4r1 was additionally expressed in peptidergic subsets (PEP1.3 and PEP1.4; Fig. 1C). These data led us to focus on analysis of the role of CysLT₂R in itch signaling.

To confirm the presence of Cysltr2 transcript in mouse DRG neurons, we performed RNAscope in situ hybridization analysis. Cysltr2 was expressed in 9.4 ± 1.9% of neurons marked by neuronal markers Tubb3 (β-3 tubulin) and in 11.6 ± 2.2% of neurons marked by Scn10a (Nav1.8; SI Appendix, Fig. S1 C and D). Cysltr2 overlapped extensively with Nppb: 95.7 ± 0.9% of Nppb⁺ neurons were Cysltr2⁺, whereas 86.0 ± 3.1% of Cysltr2⁺ neurons are also Nppb⁺ (Fig. 1 D–F). While still overlapping, 76.0 ± 3.2% of Cysltr2⁺ neurons were Hrh1⁺ and 74.7 ± 3.3% of Hrh1⁺ neurons were Cysltr2⁺, which confirms that the histamine receptor H1 is not completely restricted to the NP3 population (41). RNAscope analysis confirmed that Mrgpra3, which marks NP2 neurons, and Mrgprd, which marks NP1 neurons, had very little overlap with Cysltr2⁺ neurons (Fig. 1 D–F), with 10.3 ± 3.2% and 7.0 ± 1.6% of Cysltr2⁺ neurons being, respectively, Mrgpra3⁺ and Mrgprd⁺, and, inversely, 20.7 ± 6.4% of Mrgpra3⁺ neurons and 2.0 ± 0.6% of Mrgprd⁺ neurons being Cysltr2⁺. The transient receptor potential (TRP) channel TrpV1 has been shown to play a role in histamine-dependent itch (26). RNA-seq data show that this ion channel is expressed in NP3 neurons (SI Appendix, Fig. S1B). Using RNAscope, we confirmed that the majority of Cysltr2⁺ neurons (87.3 ± 2.2%) were Trpv1⁺, while 43.0 ± 2.1% of Trpv1⁺ neurons are Cysltr2⁺ (Fig. 1 D–F).

A recent study performed developmental analysis of DRG neurons at the single-cell level to examine the evolution of transcript expression in sensory neurons at different stages of mouse development (SI Appendix, Fig. S1 E and F) (38). Our analysis of this database showed that DRG neurons started expressing Cysltr2 after postnatal day 0 (SI Appendix, Fig. S1 E and F), and that it remained restricted to the same lineage of Sst⁺ neurons from the moment it was expressed into adulthood (SI Appendix, Fig. S1F).

We next characterized the expression of CYSLTR2 in human DRGs using RNAscope analysis (Fig. 1G and SI Appendix, Fig. S1G). We found that 63.6 ± 2.2% of human DRG neurons expressed CYSLTR2 (Fig. 1 G–I), which is broader in expression than in mouse, and these neurons ranged in size from 30 to 122 µm (SI Appendix, Fig. S1H). As in mouse, the majority of NPPB⁺ neurons (94.5 ± 2.8%) coexpressed CYSLTR2, as well as a large proportion of TRPV1⁺ neurons (76.3 ± 1.8%; Fig. 1 H and I). Overall, we found that Cysltr2 was expressed in sensory neurons from both mice and humans that overlapped with NPPB, with a broader CYSLTR2 expression in humans.

LTC₄ Specifically Induces Dose-Dependent Acute Itch.

We next determined whether specific CysLTs induced itch when injected in vivo. Cheek injection of ligands into mice allows the distinguishing of pruritogens vs. algogens based on whether they trigger hind-paw scratching vs. nocifensive forepaw wiping behaviors, respectively (43) (Fig. 2A). For our behavioral analysis, we utilized an infrared behavior observation box (iBOB) in order to analyze scratching and wiping behaviors in the dark using infrared LEDs. We first injected the known ligands of CysLT₂R, LTC₄, or LTD₄. Because LTC₄ is rapidly converted to LTD₄ at the membrane, we used, in addition, a version of LTC4 conjugated with an N-methyl group (NMLTC₄), which is a nonhydrolyzable form resistant to conversion to LTD₄. We found that, while LTC₄ and NMLTC₄ cheek injections induced robust scratching behaviors, neither vehicle nor LTD₄ injections induced scratching (Fig. 2B). This was the case when quantified as total scratching bouts or duration of scratching (Fig. 2B and SI Appendix, Fig. S2A). LTC₄ and NMLTC₄ cheek injections did not induce wiping with the forepaws, indicative of pain (Fig. 2C). We next tested whether LTE₄, which is the terminal metabolite of the CysLT pathway, also had the ability to induce itch. We found that, while NMLTC₄ induced robust scratching behaviors, LTE₄ injections did not induce itch (Fig. 2D). Neither LTD₄ nor LTE₄ induced significant scratching or wiping behaviors over vehicle controls (Fig. 2 B and D). These data indicate that injections of LTC₄ and NMLTC₄, but not LTD₄ nor LTE₄, induced robust itch but not pain behaviors.

Fig. 2.

LTC₄ but not LTD₄ or LTE₄ induces dose-dependent acute itch behaviors. (A) Diagram of experimental design of acute itch/pain induction by intradermal injection of ligands in the cheek of the mice. (B) Scratching bouts in response to vehicle, LTC₄, NMLTC₄, and LTD₄ at 0.6 nmol (n = 8 to 11). (C) Wiping responses, indicative of pain, to vehicle, LTC₄, NMLTC₄, and LTD₄ at 0.6 nmol (n = 8 to 11). (D) Scratching bouts in response to vehicle, NMLTC₄, and LTE₄ at 0.6 nmol (n = 4 to 8). (E) Wiping responses, indicative of pain, to vehicle, LTE₄, and NMLTC₄ at 0.6 nmol (n = 4 to 8). (F) Scratching bouts’ dose responses to NMTLC₄ and LTD₄ were recorded for 30 min at different concentrations and scored live (n = 3 to 12). (G) Scratching bouts’ dose responses to LTC₄ and NMLTC₄ were recorded for 30 min at different concentrations (n = 6 to 8). (H and I) Analysis of kinetics and scratching bout duration differences in response to various pruritogens: LTC₄ (0.6 nmol), IL-31 (0.02 nmol), histamine (100 μg), compound 48/80 (100 μg), CQ (200 μg), and vehicles (phosphate-buffered saline [PBS] and dimethyl sulfoxide [DMSO]). (H) Absolute numbers of bouts shorter than 0.3 s, bouts between 0.3 and 1 s, and bouts longer than 1 s. (I) Distribution of bouts according to their duration. Values presented as mean ± SEM. One-way ANOVA with Dunnett’s posttest. ns, nonsignificant (*P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001).

We next determined whether NMLTC₄ and LTC₄ induced a dose-dependent scratching response. LTC₄ and NMLTC₄ scratching curves had a bell shape, peaking at 0.2 to 0.6 nmol, whereas higher doses of LTC₄ and NMLTC₄ did not induce itch (Fig. 2 F and G). As pain can inhibit itch, we checked whether NMLTC₄ at higher doses could have a nociceptive effect; however, we detected no wiping indicative of pain (SI Appendix, Fig. S2B). LTD₄ did not induce scratching at any of the doses tested (Fig. 2F). The scratching induced by LTC₄ and NMLTC₄ occurred mainly during the first 30 min and is gone by 45 min (SI Appendix, Fig. S2C). Our phenotype was not affected by the infrared setup: NMLTC₄ induced itch that was similar when scored traditionally by observers in the light (Fig. 2F) as when recorded in iBOB (Fig. 2G). Therefore, LTC₄, but not LTD₄, gives a classic bell-shaped curve in dose dependency, which is similar to some other GPCR ligands, for acute itch induction.

LTC₄ Induces Itch that Is Distinct in Quality Compared with Other Itch Ligands.

Most pruritogens have been classified based on their overall ability to induce itch, but the quality of the itch responses has not been well characterized. We next questioned whether scratching bouts induced by injections of different pruritogens could vary in kinetics or duration. We first looked at when itch occurred following intradermal cheek injections of several pruritogenic compounds: LTC₄ (0.6 nmol), IL-31 (0.02 nmol), histamine (100 µg), compound 48/80 (100 µg), and CQ (200 µg; Fig. 2 H and I and SI Appendix, Fig. S2D). IL-31, histamine, and CQ induce itch by acting directly on sensory neurons, while compound 48/80 triggers itch by activating MC through Mrgprb2 receptors (44).

We first measured the length of individual scratching bouts induced by pruritogens, and we empirically divided bouts in three categories: short bouts (<0.3 s), medium bouts (0.3 to 1 s), and long bouts (>1 s). By this analysis, LTC₄ induced a significant increase in medium and long bouts (Fig. 2H). Overall, 45% of bouts induced by LTC₄ were medium to long bouts (>0.3 s; Fig. 2I). By contrast, CQ and compound 48/80 induced a majority of shorter bouts, with less than 10% of longer bouts (>0.3 s; Fig. 2I). IL-31 showed a similar profile of scratching bouts as LTC₄, with a significant increase of long bouts and around 40% of medium-long bouts. Histamine produced a significant proportion of medium bouts but did not produce any long bouts (Fig. 2H). We observed that LTC₄-, CQ-, and histamine-induced itch started within the first 5 min, beginning at 4.1 min, 4.7 min, and 4.6 min on average, respectively (SI Appendix, Fig. S2E), while IL-31–induced itch started at 8.7 min and 48/80-induced itch started at 11.1 min (SI Appendix, Fig. S2E). LTC₄-induced itch is at the highest between 5 and 10 min (SI Appendix, Fig. S2C), while histamine-induced itch started and peaked at 10 min, CQ- and compound 48/80-induced itches started at 10 min and seemingly peaked at 15 to 20 min, whereas IL-31–induced itch was strongest at 25 to 30 min (SI Appendix, Fig. S2F).

Alloknesis is a form of itch that occurs when the skin gets sensitized by inflammatory mediators and responds to innocuous touch stimuli. For example, histamine injection induces strong alloknesis (45). The types of neurons involved and molecular mechanisms of alloknesis are different from acute ligand-induced itch (46). We asked whether LTC₄ was able to induce alloknesis following injection. The nape of the neck was injected by ligands, followed by stimulation with a thin Von Frey filament (SI Appendix, Fig. S2G). Histamine injection was able to increase the number of responses to that filament within 20 min (SI Appendix, Fig. S2H), and this alloknesis lasted until 3 h after injection (SI Appendix, Fig. S2J). However, LTC₄ did not induce sustained alloknesis during the first 60 min, nor at the 3 h time point (SI Appendix, Fig. S2 H–J). This shows that LTC₄ can induce acute scratching and itch but does not mediate alloknesis in naïve animals. These data, taken together, show that LTC₄ induces differences in kinetics and quality of itch following injection compared with other pruritogens.

LTC₄-Mediated Itch Is Dependent on CysLT₂R.

We next determined the functional role of the CysLT₂R receptor in CysLT-induced itch. In studies in the lung and also in heterologous systems, LTC₄ has been found to bind to both CysLT₂R and CysLT₁R with different affinities; therefore, it is important to clarify the roles of these receptors in vivo in itch. LTC₄-induced itch was significantly decreased in Cysltr2^−/− mice compared with Cysltr2^+/+ control littermates (SI Appendix, Fig. S3 A and B). Levels of LTC₄-induced scratching in Cysltr2^−/− mice did not appear to be completely gone (SI Appendix, Fig. S3A), so, to elucidate whether some residual itch was present, we repeated this experiment including a vehicle condition in wild-type (WT) mice as a comparison for baseline itch. We confirmed with this experiment that LTC₄-induced scratching was significantly decreased in Cysltr2^−/− mice to levels compared with the vehicle condition (Fig. 3 A and B). NMLTC₄-induced itch was eliminated in Cysltr2^−/− mice compared to Cysltr2^+/+ control littermates, indicating a major role for Cysltr2 in mediating this itch (Fig. 3 C and D).

Fig. 3.

LTC₄-induced itch is dependent on CysLT₂R. (A and B) Scratching bouts induced by intradermal cheek injection of vehicle or LTC₄ 0.6 nmol in B6 mice or by injection of LTC₄ 0.6 nmol in Cysltr2^−/− mice (7 to 44 wk old; A) and detailed scratching bout kinetics over 30 min (^#B6 + LTC₄ vs. B6 + vehicle; *B6 + LTC₄ vs. Cysltr2^−/− + LTC₄; B). (C and D) Scratching bouts induced by intradermal cheek injection in Cysltr2^+/+ or Cysltr2^−/− mice of NMLTC₄ (0.6 nmol; C) and detailed scratching bout kinetics over 45 min (D). (E and F) Generation of BM chimeras with WT or Cysltr2^−/− mice as donors and WT or Cysltr2^−/− mice as recipients. (E) Diagram of BM transplant procedure. (F) Scratching bouts induced by intradermal cheek injection of NMLTC₄ 0.6 nmol in naïve mice and in BM chimera. Behavior recorded in daylight settings (not in iBOB). (G–I) Scratching bouts induced by intradermal cheek injection of NMLTC₄ (0.6 nmol) in (G) Kit^Wsh/Wsh mice, (H) Cysltr1^−/− mice, and (I) Trpv1^−/− mice. (J) Scratching bouts induced by the coinjection of NMLTC₄ (0.6 nmol) and LTD₄ (2 nmol) in B6 mice and in Cysltr1^−/− mice. Values presented as mean ± SEM. Unpaired t test (C, G, and I), repeated-measures two-way ANOVA with Šidák’s posttest (B and D), one-way ANOVA with Dunnett’s posttest (F), and one-way ANOVA with Tukey’s posttest (A and J). ns, nonsignificant (*P < 0.05; **P < 0.01; ***P < 0.001).

We next generated bone marrow (BM) chimeric mice to determine which cells expressing CysLT₂R were involved in CysLT-induced itch, as the receptor is expressed on numerous immune cells in addition to sensory neurons. WT or Cysltr2^−/− mice were lethally irradiated to eliminate radiosensitive hematopoietic cells but not somatic cells (including neurons), followed by transplantation with WT or Cysltr2^−/− BM, thus reconstituting radiosensitive hematopoietic cells with WT or Cysltr2^−/− genotypes (Fig. 3E). Flow cytometry of skin cells from BM-chimera mice showed that nearly 100% of eosinophils were from donor mice (SI Appendix, Fig. S4B), macrophages showed fractions from both donor and recipient origins (SI Appendix, Fig. S4C), and skin T cells were mostly from recipient mice (SI Appendix, Fig. S4D). These differences likely reflect the effect of radiation on proliferating vs. tissue-resident nondividing immune cells. Mice were injected with NMLTC₄ and scored for itch behavior. Irradiated Cysltr2^−/− mice reconstituted with WT donor BM had lower levels of NMLTC₄-induced scratching compared with levels in naïve WT mice (Fig. 3F), whereas irradiated WT mice reconstituted from WT or Cysltr2^−/− BM showed levels unchanged from naïve WT mice, indicating that Cysltr2-expressing radioresistant cells mediate NMLTC₄ itch (Fig. 3F).

We also investigated whether CysLTs acted through MCs to produce itch. MCs express Cysltr2 (47) and generate LTC₄ during inflammation. MCs are major sources of histamine, serotonin, and other pruritogenic mediators that induce itch (41, 44). MCs are radioresistant and would not be affected in BM chimeras (48). We found that NMLTC₄ induced equivalent itch in Kit^Wsh/Wsh mice deficient for MCs as their WT littermates (Fig. 3G).

We next asked whether CysLT₁R was involved in CysLT-induced itch. When injected with NMLTC₄, Cysltr1^−/− mice had the same levels of scratching as their Cysltr1^+/+ littermates (Fig. 3H). These data indicate that CysLT₂R but not CysLT₁R is involved in CysLT-induced itch. TRP channels are known to be activated and induce calcium influxes downstream of GPCRs in itch pathways. TrpV1 mediates histamine-dependent neuronal signaling and itch (26), while TrpA1 mediates CQ or BAM8-11 scratching through MrgprA1 and MrgprC11, respectively (49). As most Cysltr2⁺ neurons expressed Trpv1 (Fig. 1E and SI Appendix, Fig. S1 A and B) and Trpa1 (SI Appendix, Fig. S1B), we next determined whether CysLT responses could also be mediated through those TRP channels. We first found that NMLTC₄-induced and LTC₄-induced itch was not altered in Trpv1^−/− mice (Fig. 3I and SI Appendix, Fig. S3C). We then found that preinjecting the TrpA1 antagonist HC-030031 in the cheek failed to inhibit NMLTC₄-induced scratching (SI Appendix, Fig. S3D). These data indicate that targeting TrpV1 or TrpA1 alone is unable to impact LTC₄-induced itch. Taken together, these results show that LTC₄ can induce itch in mice by acting through CysLT₂R in nonhematopoietic cells and in a manner independent from MC, CysLT₁R, and TrpV1.

We next asked if there could be interactions between the CysLTs in itch. LTD₄, which we found does not induce itch when injected acutely (Fig. 2 B and F), inhibited NMLTC₄-induced itch when coinjected with NMLTC₄ (Fig. 3J). LTD₄ is able to act through both CysLT₁R and CysLT₂R, so we tested whether this inhibition was dependent on CysLT₁R expression. Interestingly, LTD₄ inhibition of NMLTC₄-induced itch was intact in Cysltr1^−/− mice (Fig. 3J), showing that this inhibition is independent of CysLT₁R and suggesting a potential competitive antagonism at CysLT₂R.

CysLT₂R Does Not Mediate Dry Skin or Alternaria-Induced Itch.

We next wished to determine whether CysLT₂R plays an endogenous role in physiological models of skin pathology and itch. Cysltr2^−/− mice did not have a defect in histamine-induced itch (SI Appendix, Fig. S5A). Similarly, the MC degranulator compound 48/80 induced robust scratching that was not reduced in Cysltr2^−/− mice (SI Appendix, Fig. S5B). Alternaria alternata is an environmental airborne fungus involved in allergic diseases like asthma and potentially AD. It has recently been shown to be able to induce acute itch and scratching when injected intradermally (50). We show here that Alternaria-induced itch was CysLT₂R-independent (SI Appendix, Fig. S5 C–E). Dry skin is a major cause of itch (51), and dry skin-induced itch can be modeled in mice by repeated application of a mixture of acetone and ether followed by water (AEW) for 5 d (SI Appendix, Fig. S5F). We found that AEW-induced itch and scratching behaviors were equivalent in Cysltr2^−/− mice as compared with littermate controls (SI Appendix, Fig. S5H). Taken together, these data show that CysLT₂R does not mediate dry skin or Alternaria-induced itch.

CysLT Pathway and CysLT₂R in a Model of Chronic Itch.

We next investigated the role of the endogenous CysLT pathway in chronic itch and dermatitis and utilized the MC903 model of skin inflammation, which has some characteristics of AD and ACD (52). The vitamin D analog MC903 induces skin thickening, immune cell influx, inflammation, and itch (29, 53). When we applied MC903 to the ears of male and female mice over 12 d (Fig. 4A), we found that this treatment caused spontaneous scratching behaviors (Fig. 4B), thickening of the ear (Fig. 4C), and severe acanthosis (thickening of epidermis), hyperkeratosis (thickening of the stratum corneum), and inflammation (Fig. 4D and SI Appendix, Fig. S6A). Flow cytometry analysis of the ear skin revealed a major influx of immune cells, including eosinophils, neutrophils, dendritic cells, macrophages, and T cells detectable at day 6 and with a peak at day 12 (SI Appendix, Fig. S6 B and C). Staining with toluidine blue at day 12 revealed an increase in the MC number present in the ear (SI Appendix, Fig. S7 A and B). However, we found that MCs were unlikely to be involved in driving chronic itch or skin inflammation in this model: WT controls and Kit^Wsh/Wsh mice deficient in MCs showed similar levels of itch behavior and ear thickening (SI Appendix, Fig. S7 C and D). We next determined the role of Trpv1⁺ neurons in MC903-induced itch, given that Trpv1 expression encompasses Cysltr2 expression in the DRG. Using resiniferatoxin (RTX) to ablate Trpv1⁺ neurons, we observed that RTX-treated mice failed to develop any itch/scratching behaviors in response to MC903 treatment (SI Appendix, Fig. S8A) while showing the same levels of ear thickening as control mice (SI Appendix, Fig. S8B), indicating that these neurons do mediate itch but not overall inflammation in this model.

Fig. 4.

CysLT levels are increased in the skin of the MC903 model of dermatitis and itch. (A) Diagram of procedure: daily application of MC903 on the mouse ear for 12 d. (B) Scratching bouts recorded on days 3, 10, and 12 for 60 min before daily application of vehicle (ethanol) or MC903. (C) Percentage of ear thickness change following daily application of vehicle or MC903. (D) H&E staining performed on ear section at day 12 of MC903 model. (Scale bar: 50 µm.) (E) CysLTs (LTC₄, LTD₄, and LTE₄) levels measured by ELISA at days 3, 6, 10, and 12 (n = 5 to 6). (F and G) Liquid chromatography–mass spectrometry. (F) Representative chromatograms for LTC₄ (Left) and LTE₄ (Right) from ear homogenates at day 12 after daily vehicle treatment (Top) or MC903 treatment (Bottom). (G) LTC₄ and LTE₄ quantification per milligram of ear collected at day 12 after vehicle or MC903 treatment. Values presented as mean ± SEM. Repeated-measures two-way ANOVA, Šidák’s posttest (B and C), two-way ANOVA with Šidák’s posttest (E), and unpaired t test (G). ns, nonsignificant (*P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001).

We next investigated whether CysLTs and CysLT₂R played a role in the MC903 model. Given that many immune cells recruited into the skin, including eosinophils, MCs, and dendritic cells, can generate CysLTs, we hypothesized that we could detect an increase in these key lipid mediators. Using enzyme-linked immunoassays (ELISAs), we first measured the overall CysLT levels from skin homogenates. CysLTs steadily increased over time, with significant differences at day 10 and day 12 (Fig. 4E). We next used mass spectrometry to detect individual CysLTs (LTC₄, LTD₄, and LTE4) in ear homogenates at day 12. While MC903-inflammed ears had significant levels of LTC₄ and LTE₄, none were detectable in the vehicle conditions. LTC₄ in particular was highly elevated in MC903 samples, with LTE₄ detected at lower levels and levels of LTD₄ too low to be reliably measured (Fig. 4 F and G). Therefore, CysLTs are significantly induced in MC903-inflamed ears, with the highest level of LTC₄ at the later stage of inflammation.

We next ascertained the role for CysLT₂R in this model of skin inflammation and itch. We found that MC903-treated Cysltr2^−/− mice had similar scratching levels as wild-type control littermates at early time points of the model; however, at day 12, Cysltr2^−/− mice showed a significant decrease in scratching compared with WT controls (Fig. 5A). Our findings do indicate, therefore, a specific role for CysLT₂R in later phases of itch in this model. Previous work had shown contributions of other major mediators to earlier phases of MC903-induced itch, such as neutrophils and CXCL10, which were more involved before day 10 (54). Analysis of bout duration at the day 12 timepoint further showed that the decrease in scratching in Cysltr2^−/− mice was observed in both short bouts (<0.3 s) and longer bouts, with a significant decrease in long bouts (>1 s) in the Cysltr2^−/− mice (Fig. 5B).

Fig. 5.

CysLT₂R is involved in chronic itch but not in inflammation in MC903. (A) Scratching bouts recorded on days 3, 6, 10, and 12 for 60 min before daily application of MC903 in Cysltr2^+/+ or Cysltr2^−/− mice. (B) Bout duration analysis for Cysltr2^+/+ and Cysltr2^−/− mice at day 12 of MC903 (number of bouts: <0.3 s, between 0.3 and 1 s, and >1 s). (C) Percentage of ear thickness change following daily application of MC903 in Cysltr2^+/+ or Cysltr2^−/− mice. (D and E) H&E staining performed on ear section at day 12 of the MC903 model in Cysltr2^+/+ or Cysltr2^−/− mice. (D) Representative images of H&E staining. (E) Quantification of epidermal thickness (n = 8 to 9 with 10 to 15 fields quantified per animal). (F) Quantification of immune cell populations by flow cytometry from ear homogenates collected at day 12 from Cysltr2^+/+ or Cysltr2^−/− mice (n = 3). Values presented as mean ± SEM. Repeated-measures two-way ANOVA, Šidák’s posttest (A and C), and unpaired t test (B, E, and F). ns, nonsignificant (*P < 0.05; **P < 0.01; ****P < 0.0001).

We next determined whether CysLT₂R mediates skin inflammation in MC903-treated mice, given its known role in ovalbumin-induced skin thickening and collagen deposition in a different model of dermatitis (14). Interestingly, we found no difference between Cysltr2^−/− and littermate controls in ear swelling over time (Fig. 5C) or in epidermal thickening as analyzed by hematoxylin and eosin (H&E) staining (Fig. 5 D and E). However, we did find, by flow cytometry analysis, significant decreases in some specific immune cells, including eosinophils and macrophages, in the skin of Cysltr2^−/− mice at day 12, while T cells, dendritic cells, and neutrophils remained unchanged (Fig. 5F).

As a comparison with Cysltr2^−/− mice, we next determined that the increase in ear thickness and itch behaviors were not impaired in Cysltr1^−/− mice in the MC903 model (SI Appendix, Fig. S9), indicating that CysLT₁R is not involved in mediating chronic itch. These data, taken together, indicate that CysLT₂R specifically contributes to itch in the late stage of inflammation in MC903-treated mice.

Discussion

Chronic itch negatively impacts the quality of life in patients with AD and ACD. Recent advances in the field have highlighted the importance of molecular receptors for immune mediators and neuroimmune cross-talk in those processes (29, 54, 55). CysLTs have long been suspected to be an important mediator of allergic skin reactions. CysLTs produce wheal and flare reactions when injected into human skin (56). LTC₄ is released in great quantities during in vivo allergic cutaneous reactions to ragweed or grass pollen antigen (57), and LTC₄ levels are increased in the skin of patients with AD (58). The presence of Cysltr2 expressed by the NP3 subset of pruriceptors strongly hinted at a role for the CysLTs/CysLT₂R pathway in itch, but its functional relevance had not been explored prior to this study. Our study definitively shows that LTC₄, but no other CysLTs, can specifically and functionally induce itch in vivo through the CysLT₂R receptor, and that CysLT₂R contributes to itch in a physiological model of dermatitis.

Prior to our work, two previous studies have suggested a potential role of NMLTC₄ and LTD₄ in inducing itch (31, 41). In a first study, LTD₄ was injected intradermally at one dose and found to induce itch similarly to serotonin and IL-31 (31). By contrast, we found that LTD₄ does not produce itch in a detailed dose–response analysis. A more recent study found that NMLTC₄ induced calcium influx into mouse DRG neurons and induced itch following acute injections in mice at one dose, together with sphingosine 1 phosphate and serotonin (41). While this did show initial phenotypes induced by NMLTC₄ with scratching behaviors, without a comparison with endogenous LTC₄ or other CysLTs, the physiological relevance of this response in inflammatory itch and the role of CysLT₂R in both acute and chronic itch remained unknown.

Here, we show that LTC₄, in particular, is a critical mediator of itch, and that CysLT₂R mediates this functional response in vivo in acute LTC₄-induced itch, and also is relevant in a chronic model of itch. Even though CysLT₂R expressed in heterologous cell systems binds to both LTC₄ and LTD₄ in vitro (10), in vivo datasets show a more context-dependent role of each ligand in how it interacts with the receptor. CysLT₂R is preferentially activated by LTC₄ in vivo in platelets (13) and during type 2 airway immunopathology (19). Our data on itch fits with LTC₄ and not LTD₄ in being the major driver of itch in vivo. We found that LTD₄ was able to inhibit NMLTC₄ itch when it was coinjected with NMLTC₄. Previous work has shown that LTC₄/CysLT₂R and LTD₄/CysLT₁R pathways can have antagonistic effects: CysLT₂R negatively regulates CysLT₁R-dependent Th2 pulmonary inflammation to dust mites in mice sensitized and challenged with Dermatophagoides farinae (11) and CysLT₁R-induced mitogenic signaling responses of MC (59). However, in our study, LTD₄ inhibition of NMLTC₄-induced scratching was still present in Cysltr1^−/− mice, showing that the mechanism was CysLT₁R-independent. A recent study in platelet biology indicates interesting parallels with our data (60). In this study, it was shown that LTD₄ inhibits LTC₄-driven platelet activation in vitro in a manner independent of CysLT₁R, most likely by the ligand competing with LTC₄ on the CysLT₂R receptor (60).

We found that, while LTC₄ and NMLTC₄ induce increasing itch at lower doses, this response no longer occurred at high doses. The bell-shaped curve observed with LTC₄ and NMLTC₄ may be similar to other pruritogen responses, such as the inverted U-shaped itch response observed with CQ injections (61). Given that many of the pruritogen receptors are membrane-bound GPCRs, it could be related to receptor internalization and desensitization. CysLT₁R receptors are known to undergo rapid agonist-dependent internalization (62, 63). A recent study found that a high dose of LTC₄ induced internalization of CysLT₂R in MCs, while a low dose induced its expression at the membrane (64). The downstream signaling of CysLT₂R in neurons is yet unknown. Our results show that TrpV1, which mediates histamine-dependent neuronal signaling and itch (26), was unnecessary for LTC₄-induced itch. Our data also do not indicate a role for TrpA1, which is involved in itch dependent on MrgprA3 and MrgprC11 (49). Nonetheless, it is possible that double deficiency in both ion channels could show a phenotype, and this remains to be determined with further studies.

Other than sensory neurons, many other cell types express CysLT₂R, including myeloid immune cells (3). By using BM chimeras from Cysltr2^−/− mice, our results showed that receptor expression in radiosensitive hematopoietic cells such as eosinophils was unnecessary for NMLTC₄-induced itch, while Cysltr2 in radioresistant cells was necessary. These data fit with a role for Cysltr2 in sensory neurons, which are radioresistant nonhematopoietic cells. Future studies using conditional knockout mice for Cysltr2 are necessary to determine its exact role in neurons.

We also investigated the kinetics and the quality of itching induced by distinct ligands, finding that LTC₄-induced scratching differed from histamine, CQ, and compound 48/80 by having longer bouts. Potentially, the responsive neuronal subset could determine the subsequent itch responses. LTC₄ and IL-31, which are ligands for CysLT₂R and IL-31ra coexpressed by NP3 neurons, induce the same type of longer scratching bouts. CQ, which activates MrgprA3 expressed on NP2 neurons, produces almost exclusively short bouts. Histamine receptors are known to be spread across NP3 and NP2 subsets (32), which could explain the intermediary phenotype observed with histamine. Furthermore, LTC₄ did not induce alloknesis (touch-induced itch). While chemical ligands can be coupled to itch by direct gating of pruriceptors, alloknesis is processed by other sensory neurons, including TLR5-expressing Aβ low-threshold mechanoreceptors (reviewed in ref. 46).

Transcript types and expression levels in DRG neurons can differ significantly between different species (65). We showed here that the expression of the Cysltr2 transcript was broader in human DRG neurons (63%) than in mouse DRG neurons (10%). Most neurons expressing NPPB coexpressed CYLTR2 in human neurons; however, CYLTR2 expression was not limited to NPPB⁺ neurons. The broader CYLTR2 expression in human DRG neurons suggests that the receptor is relevant to human sensory physiology and might have roles beyond those we have highlighted in the mouse. A recent study highlighted the differences between mouse and human sensory neurons, including Trpv1 expression (∼30% in mice vs. ∼70% in humans) and TSLP receptor expression (∼70% in mice vs. ∼5% in humans) (66). In rats, a previous study found that Cysltr2 was expressed in 36% of DRG neurons and that CysLT₂R could be involved in mediating pain in rats (67). Cysltr2⁺ rat neurons were mainly isolectin B4⁺ neurons coexpressing ATP receptor P2rx3, and LTC₄ injected into the footpad of rats potentiates αβ-me-ATP–induced thermal hyperalgesia (67). This contrasts with mouse DRG neurons, where Cysltr2 is expressed in ∼10% of sensory neurons and P2rx3 is in a much broader population of neurons (31, 32, 41). In our study, we did not find nocifensive phenotypes with LTC₄ cheek injections, which does not rule out sensitization of pain, but shows no induction of acute pain phenotypes. Another study found that sensory neurons from the trigeminal ganglion innervating the nasal mucosa of guinea pigs expressed Cysltr1 rather than Cysltr2, and LTD₄ could increase the excitability of those neurons (68). By contrast, we do not detect Cysltr1 transcript in mouse DRG neurons and do not find a functional role for CysLT₁R in itch. These distinct findings in human, mouse, rat, and guinea pigs highlight a need for species-specific analysis of sensory transcripts, and may have implications for translatability of results from mouse to humans in itch biology.

We demonstrated the involvement of the CysLT pathway and CysLT₂R in the MC903 mouse model of chronic itch. In this mouse model of dermatitis, several immune and inflammatory mediators have been found to contribute to chronic itch. The cytokines TSLP, IL-4, and IL-13, as well as serotonin and its receptor 5HTR7, have been found to critically drive itch in this model (29, 52–54). Recent work has shown that neutrophils and the chemokine CXCL10 mediate distinct phases of inflammation and itch in MC903 mice (54). Interestingly, this study showed that Cysltr2, Nppb, and Il31ra are transcriptionally up-regulated in trigeminal neurons following cheek application of MC903 (days 5 to 8). We found that CysLT levels went up in the skin especially in the later stages of the MC903 model. The detection of LTC₄ at day 12 indicates that the production of CysLTs is actively ongoing at that stage, since LTC₄ is quickly converted to LTD₄ (69), and it is at day 12 that Cysltr2^−/− mice showed a significant decrease in itch, suggesting that active LTC₄ production at that stage is able to induce scratching. Future studies are needed to determine how CysLT₂R signaling synergizes with the other cytokine- and immune cell-driven pathways in this model to drive itch signaling.

We found that several immune cell types increased in MC903-treated mice, which could be sources of CysLTs. In the skin and other tissues, CysLTs are produced by eosinophils, MC, macrophages, and monocytes (70, 71). MCs generate CysLTs upon activation (72) and communicate with neurons bidirectionally in inflammation and itch (73). However, we observed that mice lacking MC (Kit^Wsh/Wsh) showed intact itch in MC903-treated mice. Eosinophils are other possible candidates. Eosinophil-derived LTC₄ acts on fibroblasts through CysLT₂R to induce collagen deposition and skin thickening in ovalbumin-induced allergic sensitization (14). Macrophages and dendritic cells are also high expressers of LTC₄S, which is the enzyme that produces LTC₄ (Immgen database; https://www.immgen.org/). Of note, we found that both eosinophils and macrophages decreased in Cysltr2^−/− mice after MC903 challenge. Therefore, LTC₄ might be acting through both immune cells and sensory neurons to drive itch in vivo, and the relative roles of sensory neuron-expressed CysLT₂R compared with nonneuronal CysLT₂R in the MC903 model remain to be determined. It is possible that there is a positive neuroimmune feedback loop between CysLT-induced itch and immune cell recruitment through CysLT₂R.

By contrast with the MC903 model, we did not find a role for CysLT₂R in the Alternaria model or compound 48/80-induced itch. CysLT₂R also did not mediate AEW model of dry skin itch. One major difference between MC903 and the AEW models is that AEW does not cause significant infiltration of inflammatory cells in the dermis (74). Alternaria and compound 48/80 are both acute inflammatory models that could drive other immune pathways and cellular recruitment distinct from MC903-driven skin immune responses.

Our study may have therapeutic implications for treatment of dermatitis. The 5-LO inhibitor zileuton, which targets upstream conversion of arachidonic acid to LTA₄, has demonstrated efficacy in treating pruritus in small clinical trials (75). The CysLT₁R-specific antagonist montelukast (Singulair), which is successful in treating bronchoconstriction in asthmatic patients (9, 17), was tested in the treatment of AD patients but showed mixed results (76–79). Our work indicates that montelukast might have failed to show positive results because it targets CysLT₁R instead of CysLT₂R.

The role of the CysLT pathway and its receptors in skin conditions has been suspected for a long time but has not been previously studied in itch. Here, we have shown that LTC₄, acting through CysLT₂R, is able to induce scratching in mice and participated in chronic itch in a mouse model of AD. This suggests that drugs targeting CysLT₂R could be useful to treat recalcitrant chronic itch.

Materials and Methods

Detailed descriptions of methods and materials are provided in SI Appendix, Materials and Methods.

Mice.

All animal experiments were approved by the Harvard Medical School Institutional Animal Care and Use Committee.

BM Chimera Generation.

BM was collected from tibia, femur, and hips of both sides of either WT (C57BL/6) or Cysltr2^−/− mice 8 to 12 wk of age, dissected and cleaned from all soft tissue. Recipient animals (WT or Cysltr2^−/−) 6 to 12 wk of age underwent lethal irradiation on the day of the transplantation. At 3 to 5 h after lethal irradiation, mice were anesthetized with isoflurane for retrobulbar injection of 3 × 10⁶ BM donor cells. After 6 to 8 wk, behavioral experiments were performed.

Mouse RNAscope In Situ Hybridization Analysis.

DRGs were dissected from mice and embedded in optimal cutting temperature compound, and cryosections of 16 µm were cut. Multilabeling in situ hybridization (ISH) was performed using the RNAscope technology (ACD) according to the manufacturer’s instructions. Probes against mouse Cysltr2, Trpv1, Nppb, Mrgpra3, Mrgprd, Hrh1, Tubb3, and Scn10a in conjunction with the RNAscope multiplex fluorescent development kit were used.

Human DRG RNAscope Analysis.

Procurement procedures for all human tissue were approved by the institutional review board at the University of Texas at Dallas, and samples were deidentified prior to use in the study. Human lumbar dorsal root ganglions collection and RNAscope in situ hybridization were performed as described previously (66).

Behavioral Analysis.

For all behavior experiments, experimenters were blinded to experimental groups and/or genotypes. Recording of behaviors was performed with an experimental setup that enables us to record the mice in the dark in an experimenter-free environment: iBOB (Crimson Scientific) unless otherwise specified. Acute itch and pain behavior experiments were performed as described previously (43).

Bout duration quantification.

Duration of individual bouts was measured, and bouts were then classified in three categories according to their length: <0.3 s, 0.3 to 1 s, and 1 s.

Touch-induced itch (alloknesis).

For touch-induced itch, the nape of neck was mechanically stimulated using a 0.07-g Von Frey filament for 1 s three times in a row, with this sequence repeated three times, and the scratching responses was recorded out of a total of nine.

Chronic itch models.

The MC903 model of chronic itch was performed as described previously (29, 53), and dry skin-evoked itch behaviors assessment was carried out as previously described (80).

Cysteinyl LT Detection.

Whole ears from mice were collected, and CysLT generation was measured in acetone-precipitated homogenates by a commercially available ELISA according to the manufacturer’s protocol (Cayman).

Mass Spectrometry.

Samples were analyzed on an ultimate 3000 LC coupled with a Q Exactive plus mass spectrometer (Thermo Fisher), with a method based on previous studies (81, 82).

Flow Cytometry.

Ears were mechanically separated and minced, then digested. The preparation was stained with antibodies. Flow cytometry was conducted on an LSRII flow cytometer.

Statistics Analysis.

Data in figures represent mean ± SEM. All significance tests were chosen considering the experimental design, and we assumed normal distribution and variance of data. No data were excluded from statistical analyses unless due to technical errors. Statistical significance was determined by unpaired Student’s t test for two-group comparisons, one-way ANOVA, or ANOVA for multivariate linear models. Statistical analyses were performed using Prism 7 (GraphPad Software).

Data Availability

All study data are included in the article and/or supporting information.