Proinflammatory chemokine CXCL14 activates MAS-related G protein-coupled receptor MRGPRX2 and its putative mouse ortholog MRGPRB2

Abstract

Patients with idiopathic pulmonary fibrosis show a strongly upregulated expression of chemokine CXCL14, whose target is still unknown. Screening of CXCL14 in a panel of human G protein-coupled receptors (GPCRs) revealed its potent and selective activation of the orphan MAS-related GPCR X2 (MRGPRX2). This receptor is expressed on mast cells and - like CXCL14 - upregulated in bronchial inflammation. CXCL14 induces robust activation of MRGPRX2 and its putative mouse ortholog MRGPRB2 in G protein-dependent and β-arrestin recruitment assays that is blocked by a selective MRGPRX2/B2 antagonist. Truncation combined with mutagenesis and computational studies identified the pharmacophoric sequence of CXCL14 and its presumed interaction with the receptor. Intriguingly, C-terminal domain sequences of CXCL14 consisting of 4 to 11 amino acids display similar or increased potency and efficacy compared to the full CXCL14 sequence (77 amino acids). These results provide a rational basis for the future development of potential idiopathic pulmonary fibrosis therapies.

Fig. 1. Identification of a target receptor for CXCL14.

a A volcano plot of gene expression (8912 genes) of healthy (11 lung samples) versus IPF (13 lung samples) groups displaying statistical significance (-log₁₀ p-value) versus magnitude of change (-log₂ fold change), using the online Gene Expression Omnibus database (GEO) and the Graphpad Prism software (accession number: GSE2052)^–. Genes in the blue box are significantly upregulated in IPF patients with an adjusted p-value cut-off of 0.005 and a more than 4-fold change in expression, see Supplementary Data 1 for more details. Genes coding for insulin-like growth factor 1 (IGF1), CXCL14, secreted phosphoprotein 1 (SPP1), twist family transcription factor 1 (TWIST1), matrix metallopeptidase 7 (MMP7), uridine 5’-diphosphoglucuronosyltransferase family 1 member A9 (UGT1A9), G protein-coupled receptor 87 (GPR87), cartilage oligomeric matrix protein (COMP), and interleukin 13 receptor subunit alpha 2 (IL13RA2); see Supplementary Data 1 for more details. b Screening of GPCRs using the β-arrestin recruitment assay: CXCL14 (1 µM) was screened at 160 well-known GPCRs and 78 orphan GPCRs (oGPCRs). For known GPCRs, The effect of CXCL14 was normalized to the maximal effect of the standard agonist of each investigated receptor (for further details see Supplementary Data 2). c Concentration-response curves of CXCL14 at MRGPRX2 in β-arrestin recruitment assays using β-arrestin-Chinese hamster ovary cells (β-arrestin-CHO) recombinantly expressing MRGPRX2 (CXCL14, EC₅₀ 0.905 ± 0.548 µM), the data are means ± SD of n = 4 biological replicates. d Effect of CXCL14 (10 µM) in β-arrestin recruitment assays at all MRGPRX subtypes (MRGPRX1, MRGPRX2, MRGPRX3, and MRGPRX4), each of which was recombinantly expressed in β-arrestin-CHO cells, the data are means ± SD of n = 3 or 4 biological replicates. e Effect of CXCL14 (1 µM) in β-arrestin recruitment assays at all chemokine receptors (CCR1, CCR2, CCR3, CCR4, CCR5, CCR6, CCR7, CCR8, CCR9, CCR10, CXCR1, CXCR2, CXCR3, CXCR4, CXCR5, CXCR6, and CX3CR1). The chemokine receptor XCR1, coupled to Gα_i, was tested in cAMP assays. The effect of CXCL14 was normalized to the maximal effect of the standard agonist of each chemokine receptor (for further details see Supplementary Data 2).

Fig. 2. Characterization of CXCL14 as MRGPRX2 agonist.

a Concentration-response curve of CXCL14 in calcium mobilization assays. An EC₅₀ value of 0.889 ± 0.163 µM was determined in CHEM-1 cells recombinantly expressing MRGPRX2, but not the parental CHEM1 cell line; data are means ± SD of 4 biological replicates. b Concentration-response curve of CXCL14 at MRGPRX2 in calcium mobilization assays using LN229 cells recombinantly expressing MRGPRX2 (CXCL14, EC₅₀ 0.504 ± 0.202 µM); data are means ± SD of 4 biological replicates. Inhibition of calcium mobilization induced by CXCL14 (1 µM) and PAMP-20 (1 µM) by the MRGPRX2 antagonist PSB-172656 (100 nM) (c) in calcium mobilization assays using LN229 cells recombinantly expressing MRGPRX2 and d in β-arrestin recruitment assays using β-arrestin-CHO cells recombinantly expressing MRGPRX2. Data represent means ± SD of 3 biological replicates.; ***p < 0.001 multiple t test. e Gα_i1 protein dissociation due to CXCL14-induced MRGPRX2 activation, determined by a BRET-based assay (TRUPATH). CXCL14 (1 µM) and PAMP-20 (1 µM) were added to LN229 cells recombinantly expressing MRGPRX2 and Gα_i1RLuc8, Gβ_3, and Gγ₉, and to LN229 cells recombinantly expressing only Gα_i1RLuc8, Gβ_3, and Gγ₉, but not MRGPRX2 (data are means ± SD of 3 biological replicates; ***p < 0.001 multiple t-test.

Fig. 3. Identification of the pharmacophore of CXCL14 (large fragments).

a 3D structure of CXCL14 (PDBid: 2HDL) (green, backbone and C-terminal α-helix; blue, β-strands; red, 3₁₀ helix; magenta, unique loop of CXCL14). b Effects of large CXCL14 fragments (1 µM) on MRGPRX2 activation measured in calcium mobilization assays using CHEM1 cells recombinantly expressing MRGPRX2, the effect was normalized to 1 µM CXCL14 (100%). c Sequences of the investigated CXCL14 fragments. d pEC₅₀ values of active CXCL14 fragments in comparison to the known MRGPRX2 agonist PAMP-20. Data are means ± SD of 3 to 4 biological replicates, or means of biological duplicates, respectively. Sigmoidal dose-response parameters (variable slope) were used to analyze the data. e 2D- and 3D-structure of CXCL14: the 7 amino acid stretch, STKRFIK, that was removed in the largest inactive fragment CXCL14(1-56/64-77), is highlighted by a rectangle.

Fig. 4. Identification of the pharmacophore of CXCL14 (small fragments).

a Investigated partial peptide sequences of CXCL14. b pEC₅₀ values of CXCL14 fragments determined in calcium mobilization assays. c Maximal effect of small CXCL14 fragments (10 µM) in calcium mobilization assays normalized to the effect of CXCL14 (1 µM). d EC₅₀ and E_max values determined in calcium mobilization assays in comparison to values determined in β-arrestin recruitment assays. e Concentration-response curves of CXCL14(57-65) (PSB-231409A; STKRFIKWY) at MRGPRX2 in calcium mobilization assays (orange, EC₅₀ 0.301 ± 0.112 µM), and β-arrestin recruitment assays (blue, EC₅₀ 1.96 ± 1.97 µM). Data represent means ± SD from three to four independent experiments performed in duplicates. The effect was normalized to the maximal effect of CXCL14 (100%). Sigmoidal dose-response parameters (variable slope) were used to calculate EC₅₀ and E_max values.

Fig. 5. Investigation of the proposed peptide-receptor interaction site.

a Cryo-EM structure of MRGPRX2 in complex Cortistatin-14 (PCKNFFWKTFSSCK with disulfide bridge: 2-13) in green color (PDB: 7S8L) superimposed by CXCL14(57-65) (STKRFIKWY) colored orange in its predicted binding pose, showing subpocket 1 and subpocket 2. The peptides that appear to be similar, FIKW and WKTF, are underlined in both sequences consisting of aromatic amino acid residues (F and W) framing two amino acid residues including a lysine (K). b Predicted binding pose of the peptide CXCL14(57-65). c Side-view of MRGPRX2 showing the transmembrane helices (TM1-TM7) and the suggested binding pocket of CXCL14(57-65), which is located near the surface of MRGPRX2 (shown in the snake plot of MRGPRX2). The amino acid residues glutamic acid (Glu164^4.60), cysteine (Cys168^4.64), phenylalanine (Phe170^ELC2), cysteine (Cys180^5.34), aspartic acid (Asp184^5.38), tryptophan (Trp243^6.55), leucine (Leu247^6.59), and tryptophan (Trp248^6.60) are shown; the amino acids in purple are polar, acidic amino acids are circled in red, and lipophilic amino acids are in green. d Two-dimensional view of CXCL14(57-65) docked into the predicted binding site of MRGPRX2 using the Molecular Operating Environment (MOE) software. A snake plot of MRGPRX2 is depicted and key amino acid residues important for interactions are highlighted. pEC₅₀ values of wild type (WT) CXCL14(57-65) compared to mutant sequences K59R, K59A, R60A, K63A, and W64F, at MRGPRX2 measured (e) in calcium mobilization assays and f in β-arrestin recruitment assays. g EC₅₀ and E_max values of the mutant peptides in both assay systems. All data are means ± SD of n = 3 to 5 biological replicates. The effects were normalized to the maximal effect of CXCL14 (1 µM, 100%). Sigmoidal dose-response parameters (variable slope) were used to analyze the data for calculating EC₅₀ and E_max values. ^ns > 0.05; * ≤ 0.05; ** ≤ 0.01; *** ≤ 0.001 p-value (one-way ANOVA with Bonferroni’s multiple comparisons test). The amino acid residues of the receptor are labeled using the three-letter code and numbered according to the Ballesteros-Weinstein system, while the amino acid residues of the tested fragments are designated using the one-letter code.

Fig. 6. Activation of MRGPRB2 by CXCL14 and its fragments.

a Concentration-response curves of CXCL14 at MRGPRB2 in calcium mobilization assays using 1321N1 astrocytoma cells stably expressing MRGPRB2 (CXCL14, EC₅₀, 0.972 ± 0.105 µM) as compared to non-transfected 1321N1 astrocytoma cells. b CXCL14-induced MRGPRB2 activation (100 nM) is blocked by the antagonist PSB-172656 (1 µM) in calcium mobilization assays (**** ≤ 0.0005 p value). c Comparison of pEC₅₀ values of CXCL14 fragments and mutants at MRGPRX2 and MRGPRB2 determined in calcium mobilization assays (multiple t-test). d EC₅₀ and E_max values of CXCL14 fragments (compared to CXCL14) and mutants (compared to the peptide CXCL14(57-65)), not significant ^ns > 0.05; *≤0.05; **≤ 0.01 p-value (one-way ANOVA with Bonferroni’s multiple comparisons test). All data are means ± SD of n  =  3 to 4 biological replicates. e Overall sequence alignment between human MRGPRX2 and mouse MRGPRB2. Key amino acid residues forming interactions with the peptide CXCL14(57-65) are indicated in red. Vertical lines indicate conserved amino acids, one dot indicates non-conserved amino acids, and two dots indicate similar amino acids. Generated using the Clustal Omega1 sequence alignment tool.