Differential Skewing of Circulating MR1-Restricted and γδ T Cells in Human Psoriasis Vulgaris

Psoriasis vulgaris (PV) is a chronic, recurrent inflammatory dermatosis mediated by aberrantly activated immune cells. The role of the innate-like T cells, particularly gammadelta T (γδT) cells and MR1-restricted T lymphocytes, is incompletely explored, mainly through animal models, or by use of surrogate lineage markers, respectively. Here, we used case-control settings, multiparameter flow cytometry, 5-OP-RU-loaded MR1-tetramers, Luminex technology and targeted qRT-PCR to dissect the cellular and transcriptional landscape of γδ and MR1-restricted blood T cells in untreated PV cases (n=21, 22 matched controls). High interpersonal differences in cell composition were observed, fueling transcriptional variability at healthy baseline. A minor subset of canonical CD4+CD8+MR1-tet+TCRVα7.2+ and CD4+CD8-MR1-tet+TCRVα7.2+ T cells was the most significantly underrepresented community in male PV individuals, whereas Vδ2+ γδ T cells expressing high levels of TCR and Vδ1-δ2- γδ T cells expressing intermediate levels of TCR were selectively enriched in affected males, partly reflecting disease severity. Our findings highlight a formerly unappreciated skewing of human circulating MAIT and γδ cytomes during PV, and reveal their compositional changes in relation to sex, CMV exposure, serum cytokine content, BMI, and inflammatory burden. Complementing numerical alterations, we finally show that flow-sorted, MAIT and γδ populations exhibit divergent transcriptional changes in mild type I psoriasis, consisting of differential bulk expression for signatures of cytotoxicity/type-1 immunity (EOMES, RUNX3, IL18R), type-3 immunity (RORC, CCR6), and T cell innateness (ZBTB16).


INTRODUCTION
Psoriasis is a common and diverse, but poorly understood autoinflammatory dermatosis affecting up to 3% of the Caucasian population. Plaque-type (vulgar) psoriasis (PV) comprises most cases, but other forms have been also described (1). Once manifest, it is typically a relapsing disease often associated with systemic manifestations and comorbidities. The etiology is not clear, but genetic predisposition, in addition to microbial dysbiosis, dietary factors, and immune response, can trigger the disease (2).
On the molecular level, a range of resident and recirculating TNFa-and IL-17A-producing cells instructs the development of aberrant skin inflammation (3), but understanding how this diversity fits into psoriasiform inflammation is still limited.
Consequently, the precise composition of blood and lesional cells in PV remains unclear and almost certainly varies with different pathologic settings.
In humans, a variety of IL-17-producing CD4 + and CD8 + , conventional (4-6) and innate-like (gd (7-10)), mucosal associated invariant T (MAIT) (11) and invariant natural killer (iNKT) (12)) T cells, is enriched in psoriatic skin. The latter ones, particularly semi-invariant gd T cells, are central to dermal integrity and repair (13), and represent the principal IL-17 source in several animal models of skin inflammation and PV (8,(14)(15)(16)(17)(18). Their human counterparts differ in number, distribution, and T-cell receptor (TCR) repertoire (19), and are often delineated into Vd2 + and Vd2subsets (20). Vd2 + cells largely dominate peripheral adult blood, often co-express Vg9 chain and mount prototypic anti-microbial innate immune responses. The Vg9Vd2 population, in particular, delineates an important proinflammatory, skin-homing gdT cell compartment in PV (10). Conversely, Vd2 -T cells, particularly Vd1 + T cell compartment, are mostly confined to epithelial layers and mucosal surfaces, exhibiting clonally expanded TCR repertoires (21). In addition, different subsets of circulating gd T cells can be discriminated in healthy individuals based on CD3 and/or TCR expression levels: a larger subset of gd T cells expressing intermediate levels of TCR (hereafter CD3 + gdTCR int ), and a smaller fraction of gd T cells expressing high levels of TCR (CD3 + gdTCR high ), the latter containing IL-17 + d1/d3 + effectors (22,23), but no data regarding the variance between these gd T cell subsets currently exists for PV. Complementing these observations, distinct gene co-expression networks have been associated with the functional heterogeneity, TCRd usage, and cell-type specification of gdT cells, but the transcriptional landscape of the human gd T cell lineage in PV remains mostly uncharacterized.
Similarly, even less is known, about the MAIT population, another major innate-like T cell subset in humans. Evidence suggests that these cells traffic extensively (24) and may contribute significantly to IL-17 production in a highly inflammatory environment (25) by exerting rapid and direct effector responses prior to and independently of the TCR signaling (26). Under homeostatic conditions, the majority of the CD8 + IL-17 + T cell population in blood belongs to MAIT cell (27), but how these cells are distributed in PV remains unclear. Much alike to gd T cells, MAIT cells express a high level of surrogate markers, such as CD161 and IL-18Ra, and rely on semi-invariant TCR (Va7.2) which recognizes microbial riboflavin and folate metabolites bound on major histocompatibility complex class Irelated protein-1 (MR1) (26,(28)(29)(30). Consistent with this, MAIT and gd T cells also exhibit close similarity in their transcriptional nature, and share analogous effector subsets (31). At present, insights into the regulation of human gd and MAIT cells in PV remain scarce (32,33) and are mainly fueled by findings obtained in mouse models (8,14,16,18) or by use of CD161 and TCRVa7.2 as surrogate markers of MAIT cells (11). A single study is available on dermal MAIT cells (11), describing similar ratios of CD8 + CD161 + TCRVa7.2 + cells in healthy and psoriatic human skin; however, CD4 -CD8 -, CD4 + , and CD4 + CD8 + MAIT cells also exist but have not been probed yet. In addition, the prevalence and phenotype of circulating MR1-restricted T cells, encompassing not only the canonical TRAV1-2 + MAIT set but also a broader class of atypical TRAV1-2 -MR1-reactive T cells, are even less constrained (34,35).
Here, we exploit the advantages of MR1-Ag tetramers to provide unbiased estimates of MR1-restricted T cells and their peripheral blood numbers in PV across a range of cell subsets, thus obviating the need for surrogate markers. Next, we refine the data on blood gd T cells and their various fractions in relation to disease severity, circulating signature cytokines, and trafficking mediators by using flow cytometry in a cohort of therapeuticallynaïve, well-defined PV cases, and their matched healthy controls. In the last step, we perform a targeted gene expression analysis of purified MAIT and gd blood T cells, and generate their respective transcriptomic profiles in relation to health and PV.

Study Design and Subject Selection
Twenty-two healthy controls (16 males, 6 females; 23-54 years of age, median age 32 years) and 21 clinically active, well-characterized psoriatic patients (14 males, 7 females; 19-49 years of age, median age 33 years) were recruited at the Department of Dermatology and Venereology of the Osijek University Hospital, Croatia. All participants were unrelated and had no history of impaired hepatic or renal function. Psoriasis vulgaris was defined according to the pathohistological findings of a skin biopsy, and disease severity estimated by the treating dermatologist using Psoriasis Area and Severity Index (PASI) and the Dermatological Life Quality Index (DLQI) questionnaires (36,37). A complete blood count (CBC) encompassing red blood cell (RBC), white blood cell (WBC), platelet count, WBC differential, hemoglobin, and hematocrit measurements, together with C-reactive protein (CRP) serum levels and erythrocyte sedimentation rate (ESR) were performed in the hospital central laboratory. The body mass index (BMI) and markers of bacterial and viral burden including Mycobacterium tuberculosis (QuantiFERON-TB Gold test), cytomegalovirus (anti-CMV IgG, anti-CMV IgM), hepatitis B (anti-HBsAg) and hepatitis C (anti-HCV) antibody titre, were assessed on the same day, respectively. Patients undergoing systemic immunomodulatory, PUVA (psoralen and ultraviolet A) or nbUVB (narrow band UVB) phototherapy, with autoimmune, malignant and infectious comorbidities or allergic reactions within 6 weeks prior to the testing, were excluded from the study. Written informed consent was collected from all participants prior to the testing, and the study protocol was reviewed and approved by the ethical committee of the Osijek University Hospital (number: R2-9042/2018) and the Faculty of Medicine in Osijek (number: 2158-61-07-18-135).

Peripheral Blood Mononuclear Cell Isolation, Storage, and Thawing
Peripheral blood mononuclear cells (PBMCs) were isolated from 10 mL of freshly collected, heparinized blood samples fractionated during 25 min at 800g gradient density centrifugation on Lymphoprep medium (STEMCELL Technologies; Germany). Harvested mononuclear cell layer was transferred to a sterile conical tube, topped with PBS buffer up to the 14 mL mark and pelleted by centrifugation at 550g for 10 min. Washing step was repeated once more. Collected PBMCs were gently resuspended in 5 ml of 1x PBS and their cell count and viability were determined with the use of the Countess II automated cell counter (Thermo Fisher Scientific, USA). Following final 5 min centrifugation at 550g, 1x10e6 cells were resuspended in 1 ml of 1xPBS buffer and used immediately in downstream applications. The remaining PBMC collection was counted, pelleted and cryopreserved in 3x10e6 cell aliquots suspended in 0.5 ml of cold FBS (Biosera, France), and an equal volume of pre-chilled (4°C) freezing medium [FBS + 20% of DMSO (AppliChem)] added dropwise. Cryovials were placed in a styrofoam container and stored shortly (24-72h) at −80°C, before being transferred into liquid nitrogen tank.
For staining, cryopreserved PBMCs were thawed rapidly (60s) in a 37°C water bath, transferred into sterile 15 mL tube and dropwise diluted with 10 mL of pre-warmed, supplemented RPMI-1640 culture medium [10% FBS, 1% Na-pyruvate, 0.01M HEPES, Sigma-Aldrich]. Thawed cells were pelleted at 350g for 10 min, suspended in 5 mL of RPMI-1640 buffer, and shortly rested (30 min at 37°C in a CO 2 water jacketed incubator) before being washed and resuspended in cell staining buffer for the respective experiments. Initial gating strategy for peripheral gd T cell frequency analysis included evaluation of CD3 and gdTCR cell surface expression ( Figures 1A, B). Thereafter, we assessed gd cell lineages in more detail, by considering TCRd chain usage in a subset of cryopreserved PBMC samples ( Figure 1C). For this purpose, a cross-validation was performed by using a panel of CD3, gdTCR, TCRVd1 and TCRVd2 antibodies to show that relative cell populations, as measured by FACS, were equivalent for paired, fresh and cryopreserved samples (Supplementary Figure 1). The results demonstrated that our fresh and thawed PBMCs were comparable, showing consistent cellular proportions were recovered with different sample preparation methods. Only data that passed viability control metrics (>70%) from the FACS instruments were included (n=34).
Compensation parameters were set according to the single stained samples, while fluorescence-minus-one (FMO) and isotype controls were used for gate adjustments (Supplementary Figure 2). Flow cytometry data were collected on BD FACS Canto II cytometer and processed with FlowLogic v7.2.1. software (Inivai Technologies, Australia). Simultaneous 2-way cell sorting was performed on a 4color S3e cell sorter (Bio-Rad Laboratories, SAD) and was restricted to CD3 + gdTCR + and CD3 + TCRVa7.2 + MR1-5-OP-RU tetramer + expressing T cells. A minimum of 3,000 MR1-reactive T cells and 15,000 gd T cells from each, freshly collected PBMC sample were sortpurified directly into miRVana ™ miRNA Lysis/Binding buffer (Thermo Fisher Scientific) and used immediately for RNA extraction according to the manufacturer's instructions.
cDNA Synthesis and RT-qPCR Experiments cDNA synthesis was accomplished with the PrimeScript ™ RT Reagent kit (Takara Bio, USA) using 100 ng of total RNA input in a 40 ml final mastermix reaction, as previously described (39) Quantity and purity of RNA samples were initially verified by OD 260 /OD 280 ratio > 1.8. using IMPLEN NanoPhotometer P-Class P-330 (IMPLEN GmbH, Germany). Obtained cDNA samples were diluted 6-fold and used as a template for subsequent quantitative real-time PCR analysis of eight selected targets (PLZF/ZBTB16, RUNX3, RORC, TBX21, EOMES, IL-18R, CCR6, CCR10) and two reference genes (ACTB and TBP) using QuantStudio 5 instrument (Thermo Fisher Scientific, USA). All qPCR experiments were performed in triplicate 15 ml reactions containing 7.5 ml of TaqMan Universal PCR Master Mix II and 0.75 µl of predesigned individual TaqMan gene expression assay (Applied Biosystems, USA). The cycling conditions were set according to the manufacturer's guidelines and the list of TaqMan expression assays and amplicon sizes is given in Supplementary Table 1.
Ct values were determined with the use of QuantStudio Design & Analysis Software v 1.5.1. qPCR precision (R 2 range 0.990-0.999) and amplification efficiency (80.4-99.2%) of all investigated targets were validated by analysis of 5-point fourfold serial dilutions of arbitrary standards that were run in parallel to samples during each experiment. Intra-assay variability was less than 1.57% and less than 2.93% in-between different qPCR experiments. Expression levels of investigated transcripts were normalized relative to the ACTB reference gene, validated via NormFinder algorithm as the best endogenous control for both MR1-reactive (M=0.359) and gd T cell (M=0.515) sample cohort. Fold difference in mRNA expression was finally calculated according to the efficiency corrected model of 2 -DDCt method as described by Pfaffl 2001 (40).

Luminex Analysis of Cytokine and Chemokine Serum Levels
Peripheral blood samples for serum analysis were collected in anticoagulant-free vacutainers and centrifuged for 10 min at 1,000 g to obtain, aliquot and store (−80°C) serum samples until use. The Human Custom Procarta Plex 5-plex kit (eBioscience, Affymetrix) and the Luminex 200 platform were used for the multiplex quantitative analysis of IL-17A/F, IL-18, IL-23, CCL20, and CCL27 serum levels according to instructions in the manufacturer's leaflet. Briefly, all samples were tested in duplicate reactions comprising 50 ml of prewashed Antibody Magnetic Bead Mixture, 25 ml of Universal Assay Buffer and 25 ml of serum. Following 2 h incubation at 500 rpm, antibody captured analytes were mixed with 25 ml of biotinylated Detection Antibody Mixture, and the 96-well plate was incubated for 30 min with constant shaking at 500 rpm. Streptavidin conjugated phycoerythrin (50 ml) was added in the next step, mixed with 120 ml of Reading buffer and incubated for 5 min at 500 rpm before reading the plate in the Luminex instrument.

Statistical Analysis
Normality of distributions was assessed by the Shapiro-Wilk test and the homogeneity of variances by Levene's test. Subsequently, a nonparametric approach was adopted. Continuous data are presented as median with interquartile range (IQR). Before analysis, serum CMV IgG levels were winsorized at the upper level of detection range. The Mann-Whitney U-test was used for

Demographic and Biochemical Data
Subjects' characteristics are outlined in Table 1 In line with previous reports (22), two circulating gd T cell populations could be distinguished in most individuals, giving rise to CD3 + gdTCR int cells, and a smaller, CD3 + gdTCR high fraction (Figures 1B I, II, Table 2). For some donors, however, more-diffuse staining patterns were observed ( Figure 1B III).
Overall, CMV seropositive ( Figure 2A) and lean subjects ( Figure  2B) shared an expansion of CD3 + gdTCR high cells at systemic level, supporting a prominent position of gd T cell communities in CMV defence (42) and adipose tissue biology (23). By contrast, the counts of CD3 + gdTCR high cells declined with higher acute-phase inflammatory burden ( Figure 2C). With regard to TCRd chain usage, gdTCR int cells were mostly Vd2 + , outnumbering Vd1 + and Vd1 -Vd2subsets ( Figure 2D). Conversely, Vd1 + cells dominated the CD3 + gdTCR high subset ( Figure 2E), more so in CMVexperienced than CMV-naive individuals ( Figure 2F), with only a paucity of gdTCR high cells expressing Vd2 chain. For the Vd1 -Vd2subset of gdTCR high cells, a weak enrichment in highly antiCMVIgG-positive subjects ( Figure 2F) confirmed that the human CMV response is not restricted to Vd1 population (43).
Overall, CMV status and BMI emerged as major covariates underlying TCRd profiles in circulating CD3 + gdTCR high cells. Serum cytokine content further modified these proportions, principally affecting gdTCR int compartment and its TCRd composition. Specifically, the peripheral abundance of total CD3 + gdTCR + and CD3 + gdTCR int cells diminished with increasing serum CCL27, largely in response to declining Vd2 + cell numbers ( Figure 2G), and this association appeared restricted to subjects with PV (Supplementary Figure 3). In a similar vein, the frequency of Vd1 + gdTCR int T cells positively correlated with IL-18 serum levels ( Figures 2H, I, Supplementary Figure 4).  Table 3; the number of female participants was too low for a meaningful comparison). As judged by PASI, Vd1 -d2 -gdTCR int T cell numbers increased with the severity of disease ( Figure 2J), leading to a marginally higher proportion of Vd1 -d2 -gdTCR int T cells in affected male donors ( Figure 2K).   3D I, II), and Vd2 + gdTCR int cells (Figures 3E I, II) in gd T cell mixture, but no evidence of exquisite restriction to a single compartment was observed. RUNX3 expression, which reportedly promotes the maturation of DN TCRgd + thymocytes (44,45), was broader, and apparently not constrained to any major cell subset. In healthy controls, bulk RORC expression co-ordinately increased with the relative size of the Vd1 + gdTCR int subset ( Figure 3F), suggesting that these TFs might operate in different cellular compartments. Altogether, these results indicate that PV might promote numerical and transcriptomic reorganization of the gd cytome early in disease course, at least in type I PV.

Atypical MR1-Reactive T Cells Are Not Numerically Altered, but Inversely Correlate With IL-18 Serum Levels in PV Patients
In the next step, the atypical MR1-reactive, but TCRVa7.2-negative peripheral T cell pool was enumerated as well. In line with previous   reports on human atypical MR1-restricted ab T cell compartment [40], majority of circulating MR1-tet + TCRVa7.2 -T cells of healthy controls ( Table 4) were either CD8 + or CD4 + , whereas minority was DN or DP. In our PV dataset, the total size of the atypical CD3 + MR1-restricted, TCRVa7.2compartment was inversely related to serum IL-18 levels ( Figure 5A), whereby peripheral DP ( Figure 5B) and CD4 + ( Figure 5C) MR1-tet + TCRVa7.2 -T cell subsets were largely responsible for the observed effect. No association was observed with the case-control status, CMV seropositivity, sex and age (data not shown) for any atypical MR1-restricted T cell class.

RORC, CCRC6, and EOMES Transcripts Are Differentially Expressed in Circulating MR1-tet + TCRVa7.2 + T Cells of PV Patients
The transcriptional profile of purified blood MR1-tet + TCRVa7.2 + T cells was assayed next. In view of developmental similarities between MAIT and gd T cells (48), we probed the same selection of target genes ex vivo via RT-qPCR ( Figures 6A, B). As a result, an overlapping correlation pattern, which was significantly stronger for markers of innateness as compared to the other genes, emerged by comparing MAIT and gd T cells. Among the differentially expressed genes ( Figures 6C, D), transcripts whose products are involved in MAIT17 response (RORC, CCR6), tissue residency (CCR6), memory-like differentiation, and cytotoxicity (EOMES) were observed. While RORC expression was significantly up-regulated in conventional MAIT cells of PV patients, EOMES was down-regulated compared to controls, evoking differential, mutually exclusive requirements for lineage decisions in Tc17 and metabolite-specific T cells (48,49). Interestingly, CCR6, which is involved in thymic egress and guides tissue localization of other innate-like T cells, was downregulated in PV, but the significance of this finding for MR1restricted T cells and PV remains to be addressed. No significant associations were observed for age, sex, BMI, anti-CMV IgG, serum cytokine levels, disease severity, and CD4/CD8 cell composition (data not shown; note, however, that this kind of deconvolution is inherently inefficient for rare cell populations, such as CD4 + MAIT cells). Overall, these data suggest that the transcriptomic response of circulating innate-like T cells evolves parallely and early in disease course of PV, in a manner that differs fundamentally between gd and MAIT cell populations.

DISCUSSION
Despite numerous animal studies, the composition of innate-like T cells, and their contribution to human PV remain ambiguous. Here, we dissect the diversity of gd and MR1-restricted blood T cells in untreated, mildly affected PV patients, and show that the largest effect on their compositional profile was exerted by CMV exposure, BMI status, and sex, respectively, which largely, but not completely overpowered the disease effect. As a result, we propose that multiple, circulating innate-like T cell subsets undergo a range of distinct, previously unrecognized compositional alterations in PV, by identifying novel subcommunities with significant numerical changes in male PV patients. Of these, a minor subset of canonical CD4 + (SP and DP) MR1-tet + TCRVa7.2 + T cells was the most significantly underrepresented community during type I disease, which was absent or low in affected male individuals, irrespectively of PASI/DLQI measures. Likewise, the proportion of circulating, very sparsely represented Vd2 + gdTCR high and Vd1 -d2 -gdTCR int was increased in male patients, the latter mirroring disease activity, while no association was seen for the entire, sexmixed PV cohort. No evidence of case-control association was observed for canonical CD4 -CD8 + and CD4 -CD8 -(DN) MAIT cells, or for any other subset of atypical MR1-restricted TCRVa7.2blood T cells, at least within the spectrum of mildly to moderately affected young adults. In line, no alterations in measured serum chemokine levels (IL-17A/F, IL-23, IL-18, CCL20, and CCL27) were observed either, confirming a weak relationship between serum cytokines and skin changes in all but most severely affected individuals (50)(51)(52)(53). Instead, a relative depletion of circulating Vd2 + gdTCR int , and atypical CD4 + CD8 + (DP) and CD4 + CD8 -MR1restricted T cells was observed in the face of increasing CCL27 and IL-18 levels in PV sera, respectively, possibly reflecting their different sensitivity to activation-induced cell death (54), or homeostatic trafficking and redistribution of two cell populations competitively best poised to respond to CCR10 and IL-18R ligation (55). CCL27, together with IL-18, is prominently expressed in keratinocytes (56)(57)(58), acting as a chemoattractant for a subset of skin-specific CCR10 + gd cells in mice (59) and humans (10). For murine MAIT and Treg cells, IL-18 may play a similar role in lungs and thymus, respectively (60,61). How IL-18 regulates these migratory events is at present elusive; however, IL-18-dependent induction of the key homing chemokine receptor -CCR6 on thymic IL-18R + Tregs may provide some cues (61). Echoing these findings, a robust decline in human blood CCR10 + and CCR6 + Vg9Vd2 cells has been observed in advanced PV, mirroring their cutaneous accumulation in psoriatic lesions (10). Together, these observations revealed that affected individuals had different analyte:cytome co-associations compared to those in healthy donors, indicating that innate T cell subsets might be coordinated differently in PV and healthy participants. A similar, significant increase in the degree of correlation has been recently demonstrated in prediabetes (62) and in cotwins (63) showing signs of early subclinical neuroinflammation, suggesting that very early disease stages may indeed be associated with changes in blood components when using multiple, orthogonal "omic" signatures. Nevertheless, the molecular mechanisms underlying these associations remain elusive, precluding a  Plužarić et al.

Innate-Like T Cells in Psoriasis
Frontiers in Immunology | www.frontiersin.org December 2020 | Volume 11 | Article 572924 distinction between the cause and effect: bystander activation (64), microbial dysbiosis (65,66), and confounding by unknown modifiers could all play a role. Furthermore, the generalizability of these findings to female patients remains an issue, because we could not efficiently test or control for many non-heritable and heritable modifiers (such as HLA composition) in our small cohort of women, raising the need for independent replication in a well-powered, longitudinal study. In addition, sampling variations, incomplete data sets, and batch effects may create analytic difficulties. Meanwhile, more details on potential sexrelated differences in human immune responses have become available elsewhere (67). Human MR1-restricted ab T cells display a profound functional and compositional heterogeneity (26,27,46), presenting with a unique opportunity to shape immune responses. In the context of human PV (11) and psoriatic arthritis (33), MAIT cells have been evaluated in skin samples and synovial fluid, respectively, using TRAV1-2, CD161 and IL-18Ra as surrogate markers that relatively accurately estimate CD8 + and DN, but not CD4 + and DP MAIT cells. No difference in skin composition has been reported regarding CD8 + CD161 + TCRVa7.2 + T cell frequency in PV and healthy controls (11), but the exact number, as well as the actual contribution of other MR1-restricted, skin or blood T cell subsets to PV have remained unaddressed. We complement and expand these findings by reporting an inverse association of canonical (TCRVa7.2 + ) circulating MR1-tet + CD4 + CD8and CD4 + CD8 + , but not CD4 -CD8 + and CD4 -CD8 -, or MR1-restricted TCRVa7.2 -T cells with PV, adding to a growing body of literature on immune cell aberrations in PV. These cells exhibit sex-based differences in the prevalence of TRAV1-2 + sets (47), and their canonical CD4 -CD8 ± fractions commonly decline among adult PBMC in response to various (auto)inflammatory processes (68)(69)(70), and aging (47,71). In this context, it is significant that stimulation of MAIT cells may itself result in TCRVa7.2 downregulation, potentially affecting subsequent detection by flow cytometry (72,73), but this has yet to be demonstrated in vivo. Conversely, a minor CD4 + and TRAV1-2 -cell subset predominate in neonates (74), and show differential cytokine production, TCR pairing and antigen reactivity (27), but have not been studied in the context of human pathology yet. The number of the latter in the blood, however, is invariably low (27,46), limiting the conclusiveness of our results; thus, their functional role, and differentiation status in PV should be further investigated. Furthermore, gd T cell lineage also contains a minor subset of MR1reactive Vd1/d3 cells (75); consequently, a modified gating strategy would be necessary to disentangle these cells from their TCRb + MR1-tet + TCRVa7.2counterparts, highlighting the limits of our work. Accordingly, in situ demonstration of MR1-restricted T cells will be required to establish their translational potential.
The diversity of MAIT cell phenotype is also reflected at transcriptional level, varying according to their developmental stage, tissue localization, activation status, and CD4/CD8 census (49). Here, MAIT blood cells demonstrated several transcriptional differences in PV, surpassing their numerical variations. Within this module, RORC and CCR6 are involved in type-3 (MAIT17) ontogeny, mucosal residence, and early TCR activation of MAIT cells (48,76). By contrast, EOMES, which marks CD8 + and early TCR-activated CD4 -CD8 -MAIT cells (76), controls key checkpoints of cytotoxic maturation and exhaustion, suggesting a coordinated, multifaceted transcriptional reprogramming of MAIT cytome emerges early in PV. mRNA, however, is a poor proxy for protein expression, thus, the functional relevance of these findings is currently unknown. Consequently, there is a need for studying the cell types and states within the tissue, paving the way for potential multiomic, and single cell genomic efforts. Additionally, the patients with a more severe PV presentation should also be examined.
Next to MAIT cells, gdT lymphocytes are major innate IL-17 producers that richly populate dermal layers of lesional skin (8). These dermal populations are locally maintained (77), receive input from circulating precursor (15), and may disseminate to aggravate inflammation at distant sites (17). The full repertoire of human skin and blood gd T cells, however, has yet to be determined. Recently, several distinct communities (gdTCR int and gdTCR hi ), which differ in TCRd chain composition, IL-17 production, and transcriptional drivers (PLZF hi vs PLZF lo GATA3 + T-bet lo ), have been described in human gd blood cells (22). However, very little is known about how this heterogeneity in human gd cells relates to PV. Building on this census, we first show that CMV exposure and BMI status reciprocally shape gdTCR int :gdTCR hi ratio, mostly through the accumulation of predominantly Vd2 lo and gdTCR hi blood cells in CMV experienced, and lean subjects having low acute-phase inflammatory burden. As expected, the baseline, bulk gd transcriptome aligned with interpersonal differences in cell composition, broadly mirroring the findings from Venken et al (22). In the next step, we demonstrated that circulating Vd2 + gdTCR high and Vd1 -d2 -gdTCR int T cells are relatively enriched in mildly affected, therapeutically-naïve males with type I psoriasis compared to age, CMV, BMI, and sex-matched baseline, partly in relation to disease severity. The picture that emerges is distinct and complementary to the earlier study, whereby a decline in circulating Vg9Vd2 T cells was observed in more severely affected, heavily pretreated and older patients (10), for whom CMV status and BMI remain unknown. From scRNA-seq data, it is also evident that  Vg9Vd2 T cells are actually a heterogeneous population, comprising Th1-like gd and Th17-like gd cells (78). Consequently, we still lack the clear understanding of the earliest events in blood gd cytome, at both the cellular and molecular levels. Beyond these compositional differences, we also uncover formerly unappreciated relationships between gd transcriptional phenotype and PV, by showing a loss of innateness-associated transcription markers in bulk gd blood transcriptome. Of those, PLZF, a TF central to the lineage commitment of innate-like T cells (79), has been associated with type-2 and type-3 cell fates (78), Vd2 + and gdTCR int sets (22), cytokine receptor activity (80), and cell survival (81). Clearly, deep, unbiased characterization of human gd cells is necessary to better pinpoint the subsets underlying PV associations; meanwhile, these results hint that circulating gd T set may be disproportionately altered already in mildly affected male patients. In support, the lesional and non-lesional skin in PV shows many shared features across the epithelial and immune compartments (82)(83)(84), including the increased presence of certain gd subsets (85).
In summary, we found that mildly-to-moderately affected male patients display distinct numerical and transcriptional profiles of association between PV and certain understudied innate-like T cell subsets in peripheral blood. We also show here that identification of culprit cell subpopulations in human datasets is beset by extensive confounding from multiple sources, motivating new work to make the currently unresolved issues more tractable. Using this observation, we inform the ongoing discussion by dissecting the factors that drive the complexity of gd, and proper MR1-restricted blood T cells, in human PV.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material. Further inquiries can be directed to the corresponding authors.

ETHICS STATEMENT
The studies involving human participants were reviewed and approved by Osijek University Hospital (number: R2-9042/2018) and the Faculty of Medicine in Osijek (number: 2158-61-07-18-135). The patients/participants provided their written informed consent to participate in this study.