Variants of the MRGPRX2 Gene Found in Patients With Hypersensitivity to Quinolones and Vancomycin Show Amplified and Drug-Specific Activation Responses In Vitro

Abstract

The diagnosis of immediate hypersensitivity reactions (IHDRs) to fluoroquinolones and vancomycin is challenging. Though they may be partially mediated by the immunoglobulin E (IgE) high-affinity receptor in mast cells [1, 2], key features suggest that other mechanisms could be at play [2, 3]. They often occur in drug-naïve individuals [1, 4] receiving first doses of treatment [1, 4], and skin testing results are frequently unreliable [1]. “Vancomycin flushing syndrome” often resolves by adjusting infusion velocity [4-6]. Interestingly, both antibiotics induce mast-cell activation via the Mas-related G-protein coupled receptor-X2 (MRGPRX2) [3-5]. In this study, we sought to evaluate whether patients with confirmed hypersensitivity to these ligands presented variants of the MRGPRX2 gene, and to assess these polymorphisms' functional implications.

From 2018 to 2019, in 3 centres in Spain, we prospectively recruited patients who suffered reactions clinically compatible with IHDRs, occurring within 1 h of receiving quinolones or vancomycin. Four to 8 months (mean: 16.6 [SD ± 14.8] weeks) after the index reaction, skin tests with these drugs were performed using standardized dilutions. Patients with a history of chronic urticaria and/or NSAID hypersensitivity, and, –to avoid including potentially IgE-mediated cases– those showing positive skin tests, were excluded. Provocation tests with the suspected culprit and alternatives were conducted when risk assessment did not contraindicate them. All index reactions were confirmed by a clinician at either the time of occurrence or during provocation. Twelve patients (median age: 49.5 [IQR 44.0–61.5] years, 66.6% females) finally participated and signed their informed consent.

Whole exome sequencing was performed using blood samples from patients after genomic DNA isolation and processing, and variables located in exonic and intronic splice site-flanking regions of the MRGPRX2 gene were assessed and filtered. Two non-synonymous, exonic variants of the gene were found in patients with quinolone hypersensitivity: Asn16His (N16H) (20% of patients, 10% of total alleles affected) and Asn62Ser (N62S) (40% of patients, 30% of alleles affected). A Ser313Arg (S313R) variant was detected in heterozygosis in one of the vancomycin-reactive patients, and N62S was detected in the other, affecting both alleles.

To test their functional effects, wild-type (WT) MRGPRX2 and MRGPRX2 carrying these three SNPs were transiently transfected in HEK293LTV cells (Cell Biolabs Inc., San Diego, CA, USA). Activation responses after incubation with different stimuli (substance P (SP), cortistatin-14, quinolones, and vancomycin) were measured using calcium and inositol-phosphate 1 (IP1) read-outs. Variants N16H and N62S, found together in 20% of quinolone hypersensitivity cases, were tested as “double-mutants”, as well as separately (Figure 1A).

Levels of MRGPRX2 expression were comparable between WT and other variant-transfected cells. All transfectants led to a significantly higher calcium release than WT after activation with substance P (SP) (Figure 1B). Double variant (N16H/N62S)-transfected cells induced the highest calcium concentrations, followed by S313R and N62S. N16H/N62S exhibited increased IP1 production after incubation with substance P at 1 μM, whereas S313R induced a significantly higher accumulation of IP1 at 10 μM. The degree of MRGRPX2 receptor internalization was more pronounced for the N16H/N62S and S313R variants. Comparable IP1 release results were observed after stimulation with different concentrations of cortistatin-14.

Cells carrying N16H/N62S showed increased calcium and IP1 responses when stimulated with ciprofloxacin at all concentrations assayed (Figure 1C). N62S-transfected cells induced augmented IP1 accumulation, but only when stimulated with the highest dose. After stimulus with levofloxacin at 90 μg/mL, double variant and N62S-transfected cells accumulated more IP1 than those with N16H and WT. S313R did not increase reactivity after stimulation with ciprofloxacin and levofloxacin, acting similarly to WT (Figure 1C). Receptor internalization was more significant for N62S and N16H/N62S variants under ciprofloxacin.

When stimulated with vancomycin, on the other hand, S313R-transfected cells showed a significantly elevated calcium influx response, compared to WT (Figure 1D). There was an increased, slightly less significant response for N62S. Cells containing the double variant showed similar activation to MRGPRX2 WT cells when exposed to vancomycin. Using IP1 as a read-out, S313R and N62S-transfected cells also produced more IP1 than WT cells when stimulated with vancomycin (500 and 1000 μg/mL). Both N62S and S313R enhanced receptor internalization in the presence of vancomycin.

This study reveals three nonsynonymous, exonic SNPs of the MRGPRX2 gene in individuals with ligand hypersensitivity, inducing enhanced cell activation responses in different contexts, using a transfection model validated for MRGPRX2 variants [6] and two read-out techniques. Interestingly, patterns of enhancement seem to be drug-specific, dependent on the variant-ligand combination, cumulative (more than one polymorphism), and potentially influenced by dosing.

Identification of changes in HLA alleles has already led to rapid assays that predict delayed reactions to quinolones and vancomycin [7]. Applying similar translational applications to our research, however, requires comprehensively understanding the likely complex relationship between variant presence and hypersensitivity risk. In a cohort of 59 quinolone-tolerant individuals who donated samples to the University of Navarra Biobank as part of a separate genetic study, we found N16H in 10.2% of individuals (5.1% of alleles), N62S in 47.5% (27.1% alleles), and the N16H/N62S combination in 10.2%. A polygenic effect on receptor mechanisms could explain why variants present in general [8], and in tolerant populations, could augment the risk of reactions to common antibiotics. The conformational effects of each variant –N62S affects a cytoplasmic MRGPRX2 domain, while N16H affects an extracellular one [8]–along with zygosity, variability in receptor expression, and changes in genes involved in cell activation cascades or drug pharmacokinetics, are probably also relevant in predicting susceptibility. Further studies can address this need by including larger cohorts of reactive and tolerant individuals, by considering modifications to skin testing cut-off criteria [9] (including positive and negative skin test results), and by testing patients' mast cells to establish the role of zygosity in variant-influenced receptor activity.

G.G., M.M. M.S.-B. and P.L.Q.: conceptualized the study design. P.L.Q.: performed the assessment of clinical data from the patients. P.L.Q. and M.S.-B.: conducted the interpretation of the exome sequencing data processed by bioinformatics (E.G.). L.O.: performed the laboratory experiments and created result figures under M.M.'s supervision. P.L.Q., M.S.-B., L.O., M.M. and G.G.: wrote the manuscript text. G.G., P.L.Q., R.M.-C., J.J.L. and I.D.: recruited patients for the case cohort. J.L.P.-G.: provided exome sequencing and clinical data for quinolone-tolerant individuals. All authors reviewed the final version of the manuscript.

The authors have nothing to report.

The authors declare no conflicts of interest.