Abstract
Patients with liver diseases often suffer from chronic itch or pruritus, yet the itch-causing pruritogen(s) and their cognate receptor(s) remain largely elusive. Using transcriptomics and GPCR activation assays, we found that an orphan, primate specific MRGPRX4 is expressed in human dorsal root ganglia (hDRG) and selectively activated by bile acids. In situ hybridization and immunohistochemistry revealed that MRGPRX4 is expressed in ∼7% of hDRG neurons and co-localizes with HRH1, a known itch-inducing GPCR. Bile acids elicited a robust Ca2+ response in a subset of cultured hDRG neurons, and intradermal injection of bile acids and an MRGPRX4 specific agonist induced significant itch in healthy human subjects. Surprisingly, application of agonist for TGR5, a known sequence conserved bile acid receptor previously implicated in cholestatic itch, failed to elicit Ca2+ response in cultured hDRG neurons, nor did it induce pruritus in human subjects. In situ hybridization and immunostaining results revealed that hTGR5 is selectively expressed in satellite glial cells, unlike mTGR5 (in mouse DRG neurons), likely accounting for the inter-species difference functionally. Finally, we found that patients with cholestatic itch have significantly higher plasma bile acid levels compared to non-itchy patients and the bile acid levels significantly decreased after itch relief. This elevated bile acid level in itchy patients is sufficient to activate MRGPRX4. Taken together, our data strongly suggest that MRGPRX4 is a novel bile acid receptor that likely underlies cholestatic itch, providing a promising new drug target for anti-itch therapies.
INTRODUCTION
Chronic itch, or pruritus, is a severe and potentially debilitating clinical feature associated with many dermatological and systemic conditions1, severely affecting quality of life and potentially leading to lassitude, fatigue, and even depression and suicidal tendencies2. The most well-characterized itch receptors are the H1 and H4 histamine receptors (HRH1 and HRH4)3. Although antihistamines, which act by inhibiting histamine receptors, are generally effective at relieving itch symptoms induced by inflammation and allergens, these compounds are usually ineffective at treating chronic itch caused by systemic diseases and most skin disorders. To date, no effective treatment is available for treating histamine-resistant itch2.
A high percentage of patients with systemic liver failure develop itch with cholestatic symptoms4. For example, the prevalence of itch is as high as 69% among patients with primary biliary cirrhosis, and severe itch is an indication for liver transplantation5. Moreover, itch occurs in more than half of pregnant woman with intrahepatic cholestasis of pregnancy, a condition that has been associated with an increased risk of preterm delivery, perinatal mortality, and fetal distress6.
Several medications have been tested for treating cholestatic itch, including ursodeoxycholic acid (UDCA), cholestyramine, and rifampicin; however, these compounds either are ineffective or induce severe side effects5. Therefore, safe and effective treatments for cholestatic itch are urgently needed, and identifying the underlying molecular mechanisms—particularly the receptor and ligand—is the essential first step.
Although the link between cholestasis and itch was first described more than 2000 years ago7, the detailed mechanisms underlying cholestatic itch remain unidentified. To date, a handful of molecules have been proposed as the pruritogens that mediate cholestatic itch, including bile acids, bilirubin, lysophosphatidic acid, autotaxin, and endogenous opioids4. With respect to the cognate receptor for the pruritogen, a few receptors have been proposed, albeit based primarily on rodent models. For example, the membrane-bound bile acid receptor TGR5 has been reported to mediate bile acid□induced itch in mice8,9. However, a recent study found that administering TGR5-selective agonists failed to elicit an itch response in mouse models of cholestasis10, raising doubts regarding whether TGR5 is indeed the principal mediator for cholestatic itch. Recently, Meixiong et al. reported that Mrgpra1 and MRGPRX4 (in mice and humans, respectively) can be activated by bilirubin, a compound that serves as one of the pruritogens in cholestatic itch in mice11. Nevertheless, the precise molecular mechanism that underlie cholestatic itch in humans remains to be determined.
We specifically focused our search on genes that are selectively expressed in the human dorsal root ganglia (DRG), where the cell bodies of primary itch-sensing neurons are located. Our search revealed a novel ligand-receptor pair comprised of bile acids (the ligand) and the receptor MRGPRX4. Moreover, we found that MRGPRX4 is expressed selectively in a small subset of neurons in the human DRG, and bile acids directly trigger intracellular Ca2+ increase in these neurons. In addition, intradermal injection of both bile acids and the MRGPRX4-specific agonist nateglinide induce detectable itch in human subjects, and this bile acid-induced itch is histamine-independent. Surprisingly, application of agonist for TGR5, a known sequence conserved bile acid receptor previously implicated in cholestatic itch, failed to elicit Ca2+ response in cultured hDRG neurons, nor did it induce pruritus in human subjects. In situ hybridization and immunostaining results revealed that unlike mTGR5 expressing in mouse DRG neurons, hTGR5 is selectively expressed in satellite glial cells, likely accounting for the inter-species difference functionally. Finally, we found that plasma bile acid levels are well correlated with itch sensation in cholestatic patients and that this elevated bile acid level is sufficient to activate MRGPRX4. Taken together, our results provide compelling evidence that the ligand-receptor pair of bile acids and MRGPRX4 is likely to be one of the critical mediators for human cholestatic itch.
RESULTS
MRGPRX4 is activated by bile extract
DRG neurons are primary somatosensory neurons that express a variety of receptors and ion channels for detecting both extrinsic and intrinsic stimuli12. To identify a receptor in mediating cholestatic itch in human, we reason that this candidate receptor could be expressed in human DRG neurons and activated by bile extracts. Since the majority of itch receptors identified to date belong to the G protein□coupled receptor (GPCR) superfamily13, we analyzed two published transcriptomics datasets compiled from a variety of human tissues14,15, specifically focusing on GPCRs. Among the 332 transcripts that are enriched in the human DRG (Table S1), we identified the following seven highest-enriched orphan GPCRs: GPR149, MRGPRX4, GPR139, GPR83, MRGPRE, MRGPRX1, and MRGPRD16,17 (Fig. 1a and Table S2). Next, we cloned and expressed these candidate receptors in HEK293T cells (Supplementary Fig. 1a, b), finding that all seven receptors were expressed at the plasma membrane (Supplementary Fig. 1b). We measured the activation of each receptor by bovine bile extract using two reporter assays, the Gs-dependent luciferase assay18 and the Gq-dependent TGFα shedding assay19 (Fig. 1b,c). No signal was detected with the Gs-dependent luciferase assay (Fig. 1b). Interestingly, bile extract elicited a significant increase in reporter activity in cells expressing MRGPRX4 measured using the TGFα shedding assay, but had no effect on cells expressing the other six GPCRs (Fig. 1c). These results suggest that MRGPRX4 is activated by one or more compounds present in bile extract, and that MRGPRX4 likely signals through the Gq but not the Gs pathway. Further experiments revealed that bovine, porcine, and human bile extract activate MRGPRX4 to a similar extent in a dose-dependent manner (Fig. 1d); in contrast, extracts obtained from bovine brain, spleen, heart, kidney, and liver tissues induced no detectable signal on MRGPRX4-expressing cells (Fig. 1e). Taken together, these results suggest that MRGPRX4 is potently activated by bile extract and active compound(s) is/are highly enriched in bile extract.
Identifying which in bile extract activate MRGPRX4
Next, to identify the component(s) in bovine bile extract that activate(s) MRGPRX4, we separated the extract into six fractions using silica gel column chromatography (Fig. 2a). Each fraction was then applied to MRGPRX4-expressing HEK293T cells, and MRGPRX4 activation was measured using the TGFα shedding assay. Among the six fractions tested, fraction 4 caused the strongest activation of MRGPRX4, whereas fractions 1 and 6 caused the weakest activity (Fig. 2b), indicating that the active component(s) are mainly present in fraction 4. Mass spectrometry of fractions 4 and 6 revealed a peak enriched specifically in fraction 4 (Fig. 2c); this peak corresponded to ions with an m/z value of 410.3265 in the positive ion mode and was annotated to prostaglandin F2α diethyl amide and/or dihydroxy bile acids. Further experiments using 1H-NMR revealed that two pure dihydroxy bile acids—deoxycholic acid (DCA) and chenodeoxycholic acid (CDCA)—produced peaks that were identical to the peaks in fraction 4 (Fig. 2d); other fractions that only weakly activated MRGPRX4 also contained characteristic peaks of bile acids as shown by 1H-NMR (Supplementary Fig. 2). These results suggest that DCA and/or CDCA are enriched in the active fraction of bile extract and may be the key compounds that activate MRGPRX4.
Characterization of bile acids: MRGPRX4 activation and the downstream signaling
To further characterize the efficacy and potency of DCA, CDCA, other bile acids, and their derivatives in activating MRGPRX4, we systematically measured their ability to activate MRGPRX4 in HEK293T cells, using TGFα shedding assay and FLIPR (fluorescent imaging plate reader) Ca2+ assay. All of the bile acids tested activated MRGPRX4 to some extent; DCA had the highest potency measured with both assays, with an EC50 value of 2.7 μM and 2.6 μM in the TGFα shedding and FLIPR assays, respectively; cholic acid (CA), CDCA, and lithocholic acid (LCA)—three close analogs of DCA—were less potent (Fig. 3a-c). Based on the structural differences between DCA and the less potent bile acids, we reasoned that hydroxylation at position of R1 and/or R2, as well as taurine/glycine conjugation at position R3, is important for specific bile acids to activate MRGPRX4 (Fig. 3c).
Next, we examined the potential signaling events downstream of bile acid-induced MRGPRX4 activation by measuring intracellular Ca2+ concentration ([Ca2+]i) in MRGPRX4-expressing HEK293T cells loaded with Fluo-8 AM, a fluorescent Ca2+ indicator. We found that DCA, CA, CDCA, and LCA induced a robust fluorescence response in these cells (Fig. 3d-f), and pretreating the cells with the phospholipase C inhibitor U73122 significantly reduced the DCA-evoked Ca2+ signals; in contrast, the Gβγ inhibitor gallein had no effect on DCA-evoked signaling (Fig. 3g-h). Taken together, these results indicate that a Gq-dependent signaling pathway involving phospholipase C is downstream to MRGPRX4 activation by bile acids.
Interestingly, even though MRGPRX1, MRGPRX2, and MRGPRX3 are close analogs of MRGPRX4, none of these receptors was activated by bile acids, even at 100 μM concentration (Supplementary Fig. 3a-e). We therefore investigated the putative ligand-binding sites in MRGPRX4 by comparing the primary amino acid sequence of MRGPRX4 with these three analogs (Fig. 3i). We identified amino acid residues that are conserved in MRGPRX1, MRGPRX2, and MRGPRX3 but not in MRGPRX4 and mutated these residues, once per time, to an alanine residue in MRGPRX4. We found that mutating amino acids 159, 180, and 235 reduced the receptor’s affinity for DCA (Fig. 3j), without affecting trafficking to the cell membrane (Fig. 3k); thus, these three sites may play a critical role in the binding of bile acids to MRGPRX4. In addition, we examined whether mouse and/or rat Mrgpr family members also respond to bile acids. Intriguingly, bile acids failed to activate any mouse or rat Mrgpr members tested (Supplementary Fig. 3g-h), suggesting that the ability of MRGPRX4 to sense bile acids may be a new functional addition during evolution.
A subset of human itch-related DRG neurons express MRGPRX4 and respond to bile acids
Next, we examined endogenous expression pattern of MRGPRX4 in hDRGs. We performed in situ hybridization using a digoxigenin-labeled riboprobe against MRGPRX4 mRNA, and found that MRGPRX4 mRNA is expressed in only ∼6-8% of hDRG neurons (Fig. 4a, c); similar results were obtained with immunofluorescence using an MRGPRX4-specific antibody (Fig. 4b, c and Supplementary Fig. 4). Morphologically, these MRGPRX4-expressing neurons are small-diameter neurons, with a diameter of approximately 50 μm, which is similar to small-diameter neurons that express the neurotrophic tyrosine kinase receptor type 1 (TrkA) (Fig. 4d), suggesting a function in nociception and/or pruriception20.
To further characterize the molecular profile of these MRGPRX4-positive hDRG neurons, we performed triple-labeling of MRGPRX4 and two additional molecular markers using RNAscope in situ hybridization (Fig. 4e). Our analysis revealed that >90% of MRGPRX4-positive neurons also express the histamine receptor HRH1, a well-characterized itch receptor in humans21, and TRPV1 (transient receptor potential cation channel subfamily V member 1) (Fig. 4f-g), which functions downstream of Mrgprs and histamine receptors22,23. Interestingly, the majority of MRGPRX4-expressing neurons also co-express Nav1.7 voltage-gated sodium channel, the peptidergic marker CGRP (calcitonin gene-related peptide), and TrkA24,25 (Fig. 4f-g). These results suggest that MRGPRX4 is specifically expressed in a subset of small diameter peptidergic hDRG neurons.
Next, we tested whether MRGPRX4 in DRG neurons can be activated by bile acids. Because bile acids failed to induce a detectable Ca2+ signal in cultured rat DRG neurons (Supplementary Fig. 5), we expressed the human MRGPRX4 in cultured rat DRG neurons. Bile acids triggered a robust Ca2+ response in MRGPRX4-expressing rat DRG neurons (Supplementary Fig. 5), indicating that MRGPRX4 expressed in rat DRG neurons mediates the bile acid□induced activation. Consistent with our finding that DCA is a more potent agonist of MRGPRX4 than CA, DCA induced a significantly larger Ca2+ response and activated a larger number of MRGPRX4-expressing rat DRG neurons than CA (Supplementary Fig. 5).
Next, we asked whether hDRG neurons can also be activated by bile acids. Application of DCA induced a robust fluorescence increase in a subset (∼6%) of these hDRG neurons loaded with Fluo-8 AM; this percentage of DCA-responsive cells is similar to the percentage of MRGPRX4-expressing cells measured with in situ hybridization (Fig. 4a, c). Moreover, the less potent MRGPRX4 agonist CA also induced a response, albeit much weaker than DCA (Fig. 4h and Supplementary Fig. 6). In addition, nearly all (∼90%) of DCA-responsive hDRG neurons were capsaicin-sensitive, and approximately one-third of DCA-responsive neurons also responded to histamine (Fig. 4j). Together, our results indicate that expression of MRGPRX4 is sufficient to render bile acid sensitivity of primary somatosensory neurons.
Pharmacological activation of MRGPRX4 triggers itch sensation in human subjects
Given the specific expression pattern of MRGPRX4 in a subset of hDRG neurons, and the known role of Mrgpr family members in mediating itch sensation, we next asked whether pharmacologically activating MRGPRX4 could trigger itch sensation in human subjects. We recruited healthy volunteers and performed a double-blind skin itch test, in which each subject received a 25-μl intradermal injection of the test compounds or vehicle in four separate sites on both forearms (Fig. 5a1, inset), after which the subject was asked to rank the itch sensation at each injection site using a generalized labeled magnitude scale (LMS)26. Interestingly, the pharmacological MRGPRX4 specific agonist nateglinide, a previously reported MRGPRX4 agonist27, (Supplementary Fig. 3f) —but not vehicle—induced a robust itch sensation in healthy subjects (Fig. 5a1, a2). These results show that activation of MRGPRX4 is sufficient to trigger itch sensation in humans, suggesting that MRGPRX4 is a human itch receptor.
Bile acid□induced itch in humans is both histamine- and TGR5-independent
Previous studies have implicated that bile acids could induce itch in human28,29. Here, we systematically test pruritic effect of bile acids on human and whether bile aicd-induced itch shows some features similar to that of cholestatic itch17. We found that 500 μg (25 μl) of DCA induced a significant itch sensation that peaked within 5 min and declined slowly over time; in contrast, control injections with vehicle did not induce an itch response (Fig. 5a1, a2). Moreover, itch intensity induced by DCA was in a dose-dependent manner (Fig. 5b1, b2). We also found that less potent MRGPRX4 agonists, including CA, CDCA, taurochenodeoxycholic acid (TCDCA), and LCA, also induced a weaker—albeit still significant—itch sensation (Fig. 5c1, c2). Given that antihistamines are largely ineffective for treating cholestatic itch4, we tested whether itch induced by bile acids can be blocked by antihistamines. We found that pretreating subjects with an antihistamine prevented histamine-induced itch but had no effect on DCA-induced itch (Fig. 5d1, d2), suggesting that itch induced by bile acids does not involve histamine signaling. Taken together, these results indicate that bile acids trigger an itch sensation with features similar to cholestatic itch.
In mice, the membrane bile acid receptor TGR5 has been reported to mediate bile acid□induced itch8,9. To test whether bile acid-induced itch in human is also mediated by TGR5, we chose a non-bile acid TGR5 agonist compound 1530, which is nearly 70-fold more potent than DCA in activating human TGR5 and does not activate human MRGPRX4 (Fig. 6b, c). Intradermal injections of 10 μg (25μl) of compound 15 did not induce detectable itch in humans, whereas DCA, as the positive control, induced significant itch (Fig. 6a1, a2, d). These results suggest that TRG5 is not the receptor mediating bile acid-induced itch. Furthermore, we examined the expression of TGR5 in the human, monkey, and mouse DRG tissues. Very surprisingly, although the amino acid sequence of TGR5 is relatively conserved between rodents and primates (Supplementary Fig. 7a), we found the different expression pattern of TGR5 in DRG tissues. In human and monkey, both in situ hybridization and immunostaining revealed that TGR5 is highly expressed in satellite glial cells surrounding DRG neurons but not the primary sensory neurons (Fig. 6e-i and Supplementary Fig. 7), while in mouse, the same in situ probe and antibody detected the expression of TGR5 in mouse DRG neurons (Fig. 6f, h and Supplementary Fig. 7c), similar to the previous publication8,9. These results revealed an interesting species difference in TGR5 expression and function between mouse and primate. Taken together, our results demonstrate that the function of TGR5 in human somatosensory system is different from that in mouse, and TGR5 is not the receptor for mediating bile acid-induced itch in human.
The elevated levels of bile acids in cholestatic itchy patients are sufficient to activate MRGPRX4
Lastly, to investigate whether bile acids are the pruritogens under pathological conditions, we collected plasma samples from patients with liver or skin diseases and measured the concentration of 12 major bile acids using HPLC-MS/MS (Fig. 7a and Supplementary Fig. 8a). We found that glycine- and taurine-conjugated primary bile acids, including glycocholic acid (GCA), taurocholic acid (TCA), glycochenodeoxycholic acid (GCDCA), and TCDCA are the major bile acids present in cholestatic patients (Fig. 7a), consistent with previously published results31-33. Compared to itchy patients with liver diseases, non-itchy patients had significantly higher levels of total bile acids (defined here as the sum of the 12 bile acids shown in Fig. 7a) (Fig. 7a, b). The level of total plasma bile acids in the itchy patients with skin diseases was barely detectable and significantly lower than the itchy patients with liver diseases. Among the 12 bile acids measured, the ones with the largest differences between the patients with itch and those without itch were for GCA, GCDCA, TCA, and TCDCA (Fig. 7a, b), suggesting that these four bile acids play key roles in mediating chronic itch under pathological conditions. Indeed, intradermal injections of TCDCA caused significant itch in healthy subjects (Fig. 5c1, c2). For DCA, the most potent ligand for MRGPRX4 among all tested bile acids, we did not see the significant difference between itchy and non-itchy patients with liver diseases (Fig. 7a), suggesting it is not the major contributor for cholestatic itch under pathological conditions. More importantly, although bile acid levels vary among itchy patients with liver diseases both from our data (Fig. 7a, b) and previously reported results32-34, we found that the total plasma bile acids, as well as the individual levels of GCDCA, TCDCA, TCA, and GCA, significantly decreased in 11 out of 13 patients following itch relief (Fig. 7c, d and Supplementary Fig. 8c). Taken together, these results suggest that high levels of bile acids are well correlated with itchy symptom in patients with liver diseases and that bile acids—particularly GCDCA, TCDCA, TCA, and GCA— could be main metabolites triggering cholestatic itch.
Next, we examined whether combinations of bile acids at pathologically relevant levels are sufficient to activate MRGPRX4. We prepared mixtures of bile acids similar to the plasma/serum levels in healthy subjects (“healthy mix”) or in patients with liver diseases and itch (“liver itch mix”), which are estimated based on previously published data31,35 and our quantification results (Fig. 7a). These mixtures were then applied to MRGPRX4-expressing HEK293T cells while performing Ca2+ imaging. We found that the “liver itch mix” but not “healthy mix” induced a significant Ca2+ signal (Fig. 7e, f), suggesting that pathological relevant level of bile acids is sufficient to activate MRGPRX4.
Recently, Meixiong et al. reported that MRGPRX4 can also be activated by bilirubin, which is another potential pruritogen for triggering cholestatic itch11. We therefore compared bilirubin and DCA with respect to binding and activating MRGPRX4. We found that compared to bile acids, bilirubin is a less potent, partial agonist of MRGPRX4 (Supplementary Fig. 9a). Given the structural differences between bilirubin and DCA, we then tested whether bilirubin is an allosteric modulator of MRGPRX4. Indeed, we found that bilirubin can potentiate the activation of MRGPRX4 by DCA (Supplementary Fig. 9b), and—conversely—DCA potentiate the activation of MRGPRX4 by bilirubin (Supplementary Fig. 9c). Moreover, we found that both total bilirubin and conjugated bilirubin levels were significantly higher in itchy patients with liver diseases compared to non-itchy patients (Supplementary Fig. 9d) and plasma bilirubin levels decreased significantly after itch relief (Supplementary Fig. 9e). Compare to total bilirubin, total bile acids show better correlation with itch intensity (measured using a self-report numerical rating scale36) (Supplementary Fig. 9f). Taken together, these results suggest that bile acids are the major pruritogens in MRGPRX4-mediated cholestatic itch and bilirubin facilitates the activation of MRGPRX4 by bile acids and may also contribute to cholestatic itch in pathological conditions.
DISCUSSION
Here, we report that MRGPRX4 is a novel GPCR that fits with the criteria we set for identifying putative receptor in mediating cholestatic itch. MRGPRX4 is selectively expressed in a small subset of human DRG neurons. Bile acids triggered a robust Ca2+ response in a subset of hDRG neurons as well as rat DRG neurons expressing MRGPRX4 exogenously. Both bile acids and an MRGPRX4-specific agonist induce itch in human. Bile acid-induced itch in human is histamine independent, which is consistent with antihistamines are largely ineffective for treating cholestatic itch. Surprisingly, application of agonist for TGR5 failed to elicit Ca2+ response in cultured hDRG neurons, nor did it induce pruritus in human subjects. The expression pattern of TGR5 is different between mouse and human. hTGR5 is selectively expressed in satellite glial cells, while mTGR5 is expressed in DRG neurons, likely accounting for the inter-species difference functionally. We also found that plasma levels of bile acids were well correlated with itchy patients with liver diseases. Importantly, a mixture of bile acids with components and concentrations similar to that of cholestatic itchy patients—but not healthy volunteers—was sufficient to activate MRGPRX4. Our data indicate bile acids are the major pruritogens in MRGPRX4-mediated cholestatic itch and bilirubin facilitates the activation of MRGPRX4 by bile acids and may also contribute to cholestatic itch in pathological conditions. Based on our results, we propose a new working model for cholestatic itch (Fig. 7g): patients with cholestasis usually display increased plasma levels of bile acids and bilirubin, which are precipitated in the skin and activate MRGPRX4 receptors in itch-related primary fibers, thereby triggering itch in these patients. Our results exclude TGR5 as a primary itch receptor in human, and the broad expression of TGR5 in satellite glial cells implies a more general function which remains to be determined in the future.
Here, we provide important evidence that MRGPRX4 is sufficient for mediating bile acid□induced itch, and thus should play an important role in cholestatic itch. Since specific antagonist for MRGPRX4 is currently unavailable, we could not determine whether MRGPRX4 is necessary for bile acids induced itch in human. Future studies will be designed to further examine the role of MRGPRX4 in cholestatic itch using to-be-developed pharmacological and/or human genetic approaches. For example, several single-nucleotide polymorphisms (SNPs) have been identified in the human MRGPRX4 gene37, and it would be interesting to screen for loss-of-function and gain-of-function MRGPRX4 variants. Characterizing the relationship between these variants and itch intensity in cholestatic patients and healthy subjects with bile acid□induced itch could help to further delineate the relationship between MRGPRX4 activity and cholestatic itch. These experiments will also help to determine whether MRGPRX4 is the main molecular receptor for mediating cholestatic itch, or whether other GPCRs4, such as lysophosphatidic acid receptors and serotonin receptors also play roles in cholestatic itch.
Our current understandings about mechanisms underlying somatosensation in the mammalian system are mainly derived from studies of rodents. Despite the great value and insights we gained using rodent models, notable failures have happened in translating results obtained in rodents into effective and safe clinical treatments in human38-41. The bile acid receptors we study here is a great example demonstrating the species differences between rodent and human somatosensory systems. Although TGR5, a bile acid membrane receptor, was previously reported to be expressed in mouse DRG neurons and mediate bile acid□induced itch in mice8,9, our expressing characterizations as well as functional assays revealed that TGR5 is not expressed in human DRG neurons and doesn’t directly mediate itch sensation in human. Instead, primate MRGPRX4 gains the novel function of bile acid sensitivity during evolution. Therefore, it is crucial to study and validate the mechanism of cholestatic chronic itch and develop the correspondent treatment within the context of human physiology.
Recently, Meixiong et al. reported that mouse Mrgpra1 and human MRGPRX4 can be activated by bilirubin, suggesting that bilirubin may serve as a pruritogen in cholestatic itch11. Bilirubin, a yellow compound that causes the yellow discoloration in jaundice, has not been considered a likely candidate pruritogen though, because the clinical observation that itch often precedes the appearance of jaundice, particularly in patients with intrahepatic cholestasis of pregnancy (ICP)42 and patients with primary biliary cirrhosis7. Our results suggest that bilirubin is a partial agonist of MRGPRX4 and may potentiate the activation of MRGPRX4 by bile acids. This notion is consistent with our finding that the correlation between bile acid levels and itch intensity is stronger than the correlation between bilirubin levels and itch intensity. Based on these findings, we propose that bile acid is the major contributor to cholestatic itch, and bilirubin serves to increase bile acid□induced cholestatic itch under pathological conditions.
In summary, we found that the membrane-bound GPCR MRGPRX4 is a novel bile acid receptor and may serve as an important molecular mediator of chronic itch in patients with systemic liver diseases. Our results suggest that MRGPRX4 is a promising molecular target for developing new treatments to alleviate devastating chronic itch in these patients.
Data availability statement
The data that support the findings of this study are available from the corresponding author upon request. All figures have associated raw data. There is no restriction regarding data availability.
Conflict of interest
The authors declare no competing interests.
MATERIALS AND METHODS
Analysis of GPCRs expressed in human DRG neurons
The expression profile of all genes in hDRG neurons was compared to human reference tissues, including trigeminal ganglia, brain, colon, liver, lung, muscle, and testis14,15. To identify DRG-enriched GPCRs, we using the following formula: [(the expression level of a given gene in the DRG)/(the total expression level of that gene in all tissues)]; a value ≥0.5 was used to define DRG-enriched genes. The expression level of a gene refers to the number of fragments per kilobase of exon per million fragments mapped (FPKM) in the tissue transcriptome.
Bovine tissue extracts
Fresh bovine heart, brain, kidney, spleen, and liver tissues (40 g each) were dissected and then boiled for 5□min in 200 ml water. Acetic acid and HCl were then added to a final concentration of 1 M and 20 mM, respectively, and the mixture was homogenized thoroughly and then centrifuged at 11,000 rpm for 30 min. The supernatant was collected and concentrated to a volume of 40 ml using a rotary evaporator. Acetone (80 ml) was then added to the concentrated solution, and the new solution was again centrifuged at 11,000 rpm for 30 minutes. The supernatant was collected using a rotary evaporator and freeze-dried in a vacuum. The final product was weighed, and equal amounts of each extract were used to test for activity.
Generation of stable GPCR-expressing cell lines
Stable cell lines expressing orphan GPCRs were generated using the PiggyBac Transposon System. In brief, each orphan GPCR was subcloned into the PiggyBac Transposon vector and co-transfected with the hyperactive PiggyBac transposase43 into the HEK293T-based TGFα shedding reporter cell line19 using polyethylenimine (PEI). Receptor-expressing cells were selected and maintained in DMEM containing 10% fetal bovine serum (FBS), 1 μg/ml puromycin, 100 U penicillin, and 100 μg/ml streptomycin in a humidified atmosphere at 37°C containing 5% CO2.
TGFα shedding assay
Cultured cells expressing orphan GPCRs were rinsed once with Mg2+-free and Ca2+-free phosphate-buffered saline (PBS) and then detached with 0.05% (w/v) trypsin. The cell suspension was transferred to a 15-ml tube and centrifuged at 190xg for 5 min. The supernatant was discarded, and the cell pellet was suspended in 10 ml PBS and incubated for 15 min at room temperature (RT). The cells were re-centrifuged and suspended in 4 ml HBSS (Hanks’ balanced salt solution) containing 5 mM HEPES (pH 7.4). The suspended cells were then seeded in a 96-well plate at 40,000-50,000 cells per well and placed in a 37°C incubator in 5% CO2 for 30 min. A 10x stock solution of each drug was prepared in assay buffer (HBSS containing 5 mM HEPES, pH 7.4), and 10 μl of 10x stock solution was added to each well. The plate was then placed in the incubator for 2 hr, after which alkaline phosphatase (AP) activity was measured in the conditioned media and cells.
FLIPR assay
HEK293T cells stably expressing human MRGPRX4 were seeded in 96-well plates at a density of ∼50,000 cells per well. The following day, the cells were loaded with Fluo-8 (Screen Quest Fluo-8 No-Wash Calcium Assay Kit, AAT Bioquest, Cat. No. 36316) for 2 hr, and test compounds were added to the wells. The Fluo-8 signal was measured using the FLIPR TETRA system (PerkinElmer).
Luciferase assay
We generated a luciferase reporter plasmid that encodes secreted NanoLuc under the control of a cAMP response element (CRE) and a minimal promoter. The hygromycin-resistance gene and EBFP driven by the SV40 promoter in the reporter plasmid were used to generate stable cell lines. HEK293T cells were transfected with this plasmid, and a stable cell line was generated by selecting with hygromycin.
This stable reporter cell line was then transfected with various GPCRs and used to monitor the activation of these receptors. In brief, the cells were seeded in 96-well plates; the next day, the culture medium was replaced, and compounds were added to the wells; forskolin (10 μM final concentration) and 0.01% DMSO (v/v) were used as positive and negative controls, respectively. The plates were incubated at 37°C in 5% CO2 for 24 hr, after which a 10-μl aliquot of cell culture medium was removed from each well and combined with 40 μl culture medium plus 50 μl assay buffer (containing 20 μM of the luciferase substrate coelenterazine); after 5 min incubation, luminescence was measured using an EnVision plate reader (PerkinElmer).
Fractionation of bile acid components
A commercially available bovine bile acid powder (126.6 mg) was loaded in a silica gel column (DCM:MeOH = 10:1). The smaller fractions were combined to form six larger fractions (F1 through F6) based on analytical thin-layer chromatography performed using 0.25-mm silica gel 60-F plates. Flash chromatography was performed using 200−400 mesh silica gel.
MS and NMR
High-resolution mass spectrometry was performed at the Peking University Mass Spectrometry Laboratory using a Bruker Fourier Transform Ion Cyclotron Resonance Mass Spectrometer Solarix XR. 1H-NMR spectra were recorded on a Bruker 400-MHz spectrometer at ambient temperature with CDCl3 as the solvent.
Immunostaining and flow cytometry analysis
Suspended live HEK 293 cells stably expressing the point-mutated MRGPRX4 were washed in washing buffer (1X PBS solution, mixed with 5% fatal bovine serum (FBS)) for 3 times. Then cells were incubated with primary antibody (Sigmal-Aldrich Cat. No. C3956, 1:25 dilution) for 30 minutes, and secondary antibody (AAT Bioquest iFluroTM Alexa 488 goat antirabbit IgG Cat. No. 1060423, 1:50 dilution) for 1 hour. Cells were washed for two times after each antibody treatment. Next, cells were resuspended with 300 uL to 500 uL FACS buffer, and fluorescence-activated cell-sorting analysis was performed, using the BD FACS Calibor Flow cytometer (BD Biosciences), and the data were analyzed using FlowJo software (Ver. 7.6.1).
Cultured human DRG neurons
Collection of DRG tissue from adult humans was approved by the Committee for Medical Science Research Ethics, Peking University Third Hospital (IRB00006761-2015238), and collection from human embryos was approved by the Reproductive Study Ethics Committee of Peking University Third Hospital (2012SZ-013 and 2017SZ-043) and Beijing Anzhen Hospital (2014012x). DRG tissues were obtained from adult patients undergoing surgical excision of a schwannoma; the tissues were placed immediately in ice-cold DMEM/F12 medium. The tissues were then cut into pieces <1 mm in size and treated with an enzyme solution containing 5 mg/ml dispase and 1 mg/ml collagenase at 37°C for 1 hr. After trituration and centrifugation, the cells were washed in 15% (w/v) bovine serum albumin (BSA) resuspended in DMEM/F12 containing 10% FBS, plated on glass coverslips coated with poly-D-lysine and laminin, cultured in an incubator at 37°C, and used within 24 hr of plating.
Culture and electroporation of rodent DRG neurons
Rat DRG tissues were obtained from the thoracic and lumbar vertebrae and placed in ice-cold DMEM/F12 medium. The tissues were cut into pieces <1 mm in size and then treated with an enzyme solution containing 5 mg/ml dispase and 1 mg/ml collagenase at 37°C for 1 hr. After trituration and centrifugation, the cells were washed in 15% BSA, resuspended in DMEM/F12 containing 10% FBS, plated on glass coverslips coated with poly-D-lysine and laminin, cultured in an incubator at 37°C, and used within 24 hr of plating.
Rat DRG neurons were electroporated as follows. After washing the neurons with 15% BSA, the neurons were resuspended in DMEM/F12 and electroporated using a P3 Primary Cell 4D-Nucleofector X Kit L (cat. no. V4XP-3012, Lonza) in accordance with the manufacturer’s instructions. After electroporation, the neurons were cultured for 72 hr before use in order to allow the transgenes to express.
Ca2+ imaging
For Ca2+ imaging experiments, cells were loaded at 37°C for 1 hr with 10 μg/ml Fluo-8 AM (AAT Bioquest, Inc.) supplemented with 0.01% Pluronic F-127 (w/v; Invitrogen). Bile acids, bio-mimicked bile acid mixes, and/or various drugs to be tested were added to the cells in a chamber containing a custom-made 8-channel perfusion valve control system. Fluorescence images were acquired using a Nikon A1 confocal microscope.
In situ hybridization and immunostaining
Single colorimetric in situ hybridization in hDRG sections was performed as follows. The sections were fixed in freshly prepared 4% paraformaldehyde (PFA) in PBS for 20 min at RT, and then washed in fresh-DEPC PBS (1:1000 DEPC was added to 1x PBS immediately before use) and DEPC-pretreated PBS (1:1000 DEPC in PBS overnight, followed by autoclaving) for 10 min each. The sections were then immersed in a DEPC-containing antigen-retrieval solution containing 10 mM citric acid, 0.05% Tween-20 (pH 6.0) in a 95°C water bath for 20 min, and then cooled at RT for 30 min. After washing in DEPC-pretreated PBS for 10 min, the sections were incubated in a Proteinase K solution (25 μg/mL in DEPC-pretreated water) for 20 min and then washed in fresh-DEPC PBS and DEPC-pretreated PBS (10 min each). The sections were incubated in freshly prepared acetylation solution containing 0.1 M TEA and 0.25% acetic anhydride in DEPC-pretreated water for 10 min at RT, followed by a 10-min wash in DEPC-pretreated PBS. The prehybridization step was performed in probe-free hybridization buffer consisting of 50% formamide, 5x SSC, 0.3 mg/ml yeast tRNA, 100 μg/ml heparin, 1x Denhardt’s solution, 0.1% Tween-20, 0.1% CHAPS, and 5 mM EDTA in RNase-free water at 62°C for 30 min in a humidified chamber, followed by an overnight hybridization step in hybridization buffer containing 5 ng/μl DIG-labeled riboprobes at 62°C in a humidified chamber (under a Parafilm coverslip). After the hybridization step, the sections were washed in 0.2x SSC at 68°C (once for 15 min and twice for 30 min each), followed by blocking in PBS containing 0.1% Triton X-100 and 20% horse serum for 1 hr at RT. The sections were then stained overnight at 4°C with pre-absorbed AP-conjugated sheep anti-DIG antibody (1:1000, Roche, cat. 11093274910) in PBS containing 0.1% Triton X-100 and 20% horse serum. The sections were washed 3 times for 10 min each in PBS containing 0.1% Triton X-100, followed by overnight incubation in the dark in AP buffer containing 100 mM Tris (pH 9.5), 50 mM MgCl2, 100 mM NaCl, 0.1% Tween-20, 5 mM levamisole, 0.34 mg/ml NBT (Roche cat. no. 11383213001), and 0.17 mg/ml BCIP (Roche, cat. no. 1138221001) to allow the color reaction to develop. The sections were washed 3 times for 10 min each in PBS, and then fixed for 30 min in 4% PFA in PBS. The sections were quickly rinsed 5 times in ddH2O, dried at 37°C for 1 hr, and dehydrated in xylene (3 times for 2 min each). Finally, the sections were mounted under a glass coverslip using Permount (Fisher).
Immunostaining was performed using a rabbit anti-hMRGPRX4 antibody obtained (Abcam, cat. no. ab120808). The sections were fixed in freshly prepared 4% PFA in PBS for 20 min at RT and then washed in PBS containing 0.1% Triton X-100 3 times for 10 min each, followed by block in PBS containing 0.1% Triton X-100 and 20% horse serum for 1 hour at RT. The sections were then incubated overnight in primary antibody at 4°C, washed with PBS containing 0.1% Triton X-100 3 times for 15 min each, and incubated with secondary antibody for 1 hour at RT. After washing with PBS 3 times for 15 min each, the sections were mounted under glass coverslips and Fluoromount-G (Invitrogen).
RNAscope in situ hybridization
RNAscope in situ hybridization was performed in accordance with the manufacturer’s instructions (Advanced Cell Diagnostics). In brief, human DRG sections were fixed, dehydrated, and treated with protease. The sections were then hybridized with the respective target probe for 2 hours at 40°C, followed by four-round signal amplification. The sections were then mounted under coverslips, sealed with nail polish, and stored in the dark at 4°C until imaged.
Human itch test
The human itch test studies were approved by the Committee for Protecting Human and Animal Subjects at the Department of Psychology, Peking University (#2018-05-02). Volunteers were students and faculty members recruited from Peking University. All subjects provided written informed consent and were provided with the experimental protocol. All injections were performed using an INJEX 30 needle-free injection system (INJEX Pharma GmbH, Berlin, Germany). We performed two studies as described below.
In the first study (to measure bile acid□induced itch sensation), each tested compound was dissolved in physiological saline containing 7% Tween-80 (Sigma-Aldrich). The injection sites were cleaned with rubbing alcohol, and 25 μl of each solution was injected intradermally on the volar surface of each arm. The same volume of vehicle (saline containing 7% Tween-80) served as the negative control. Itch was defined as the desire to initiate scratching during the experiment, and the subjects rate the perceived intensity according the generalized labeled magnitude scale (LMS) described by Green et al.26
In the second study (to measure the effect of antihistamines on DCA-induced itch), two experimental sessions were performed, separated by 2 weeks, with 14 and 12 subjects participating in the first and second sessions, respectively. Approximately 1.5 g of topical antihistamine cream (doxepin hydrochloride cream, Chongqing Huapont Pharm. Co., China) or a placebo cream (cold cream, Eau Thermale Avène, Paris, France) was applied 2.5 hr before injection of DCA or histamine (Sigma-Aldrich); any unabsorbed cream was removed with alcohol. A 500 μg/25 μl solution of DCA was prepared as described above, and a 2.5 μg/25 μl solution of histamine was dissolved in saline; 25 μl of the DCA or histamine solution was injected into the volar surface of the arm as described above. In the first session, each subject received two intradermal injections of DCA (one at the antihistamine-treated site and one at the placebo-treated site). In the second session, each subject received two intradermal injections of histamine (one at the antihistamine-treated site and one at the placebo-treated site). The subjects then rate the itch sensation as described above.
Quantification of plasma bile acids and bilirubin
These experiments were approved by the Committee for Biomedical Ethics, Peking University First Hospital (2017-R-94). Itch intensity was measured using a self-report numerical rating scale (NRS)36, and whole blood samples were collected from patients with skin diseases and patients with liver diseases. Plasma was obtained by centrifuging 2 ml of whole blood at 4°C, 11,000 g for 10 min; 100 μl of each plasma sample was then mixed with 400 μl acetonitrile and left to sit at 4°C for 20 min. The mixture was centrifuged, and the supernatant was dried in a rotatory evaporator (45°C under vacuum), and the dried residue was retrieved and dissolved in 60% methanol for further analysis.
The bile acid level in plasma samples was measured using HPLC-MS/MS (Agilent model LC1260 QQQ 6495). Chromatographic separation was performed in an ACQUITY UPLC HSS T3 column (2.1 mm × 100 mm, 1.8 μm; Waters Corp.). The mobile phase consisted of solution A (water) and solution B (acetonitrile). The total running time was 23 min, and a linear gradient (0.3 ml/min) was applied as follows: 0-2 min: 10% B - 40% B; 2-18 min: 40% B - 50% B; 18-19 min: 50-100% B; 19-20 min: 100% B; 20-21 min: 100-10% B; 21-23 min: 10% B. The injection volume was 5 μl, and the mobile phase flow rate was 3 ml/min. Deoxycholic-2,2,4,4,11,11-d6 acid (Sigma, cat. no. 809675) was used as an internal standard.
Total bilirubin and direct bilirubin values were obtained from the patients’ hospital blood chemistry reports.
Statistical analysis
Summary data are presented as the mean ± SEM. Human subjects were randomly assigned to control and experimental groups, and the subjects and investigators were double-blinded with respect to the experiment treatments. Data were analyzed using the Student’s t-test, two-proportion z-test, Chi-square test or One-way ANOVA and differences with a P-value of < 0.05 were considered significant.
Acknowledgments
We thank Dr. Y. Rao for sharing the tissue culture room, Dr. J.H. Zhao for collecting clinical blood samples. We are grateful to Dr. L.Q. Luo and Dr. Y. Song for critical reading of the manuscript. We also thank Dr. X.Z. Dong for sharing unpublished data. This work was supported by the Junior Thousand Talents Program of China to Y.L.