The involvement of B-cells in atherosclerosis has gained attention in recent years, although mechanistic underpinnings are yet to be clarified. The results reported in this study demonstrate an altered expression of GPR55 in the early phase of RA, also in relation to inflammatory circuits, B-cell activation and remodelling responses. These results were independently validated in transcriptomic datasets and in vitro assays. To the best of our knowledge, this is the first description of GPR55 expression in patients with rheumatic and musculoskeletal diseases (RMDs).

Our findings demonstrate not only that GPR55 is expressed in leucocyte subsets from human populations, but also that altered GPR55 expression could be linked to human diseases, namely RA. These results were validated at the gene expression levels using public datasets, also highlighting differences between tissues (stromal vs haematopoietic). Differences in the cellular composition between tissues may account, at least in part, for these differences. As the endocannabinoid system is known to regulate several pathways,^{23 24} it may be feasible that different tissues and organs may be hallmarked by differential expression of its receptors, including GPR55. Furthermore, cell-specific differences within leucocyte subsets, as well as differences in their modulation on stimuli in vitro, were also observed in our study. Synovial-infiltrating leucocyte composition may differ from that of the systemic compartment, hence adding to this discrepancy. Moreover, important differences were found within the B-cell compartment, with a dominance of memory B-cells. These results are relevant in a twofold manner. First, memory subsets are instrumental to understand the chronicity and perpetuation of autoimmunity in RA.^{25 26} Second, these differences may help to understand the differences between RA and SLE in the datasets, as distinct B-cell subpopulations hallmark these conditions.27,30 Differences in B-cell activation have also been reported. Since GPR55 has been demonstrated to be differentially expressed among B-cell subsets, disturbances within B-cell pools may thus explain disease-specific patterns.

GPR55 expression in patients with RA was unrelated to clinical features, including patient-reported outcomes related to nociception, such as HAQ or pain scales. This finding emphasised the differences observed between GPR55 and other cannabinoid receptors.⁷ Interestingly, evidence from cytokine analyses, proteomic signatures, metabolomic features and in vitro experiments revealed associations between GPR55 expression, especially in B-cells, and inflammatory circuits. First, correlation analyses confirmed that GPR55 expression was associated with a number of proinflammatory mediators. Moreover, positive regulation of cytokine production was also retrieved after pathway annotation. The analysis of GlycA and GlycB levels, which are linked to systemic inflammation, reinforced these findings. Interestingly, distinct correlation profiles were noted between monocyte and B-cell subsets, thus pointing to cell context-specific roles or regulation. In fact, divergent patterns were also observed with lymphocytes and neutrophils counts. Furthermore, LPS exposure in vitro led to a dose-dependent GPR55 downregulation in B-cells, whereas no effect was observed on monocytes, in line with existing evidence.¹⁰ Equivalent results were obtained in neurons in an animal model, where the LPS challenge was also associated with a proinflammatory response.³¹ LPS is a well-known stimulant of B-cells, hence prompting activation, proliferation and differentiation.³² This may suggest a link between GPR55 expression and B-cell activation. Our findings support this idea, as LPS-induced GPR55 downregulation overlapped with increasing B-cell activation in vitro, measured as CD86 increased expression. Furthermore, the analysis of functional proteomic pathways also reinforced a potential connection with B-cell activation programmes. Furthermore, data from immunoglobulin levels and disease-related autoantibodies strengthen these observations. Interestingly, despite the associations found with IgA and IgG, no associations were observed with IgM. This divergence may relate to the differential GPR55 expression between memory and naïve subpopulations. These findings confirm previous evidence from GPR55-deficient mouse models.⁴

Recent evidence has reported that GPR55 deficiency may be connected with an altered IL-4 signalling,³³ also related to B-cell differentiation. Transcriptomic and mitochondrial content analyses emphasised the involvement of GPR55 signalling in regulating B-cell activation and differentiation through a pleiotropic effect.^{4 34} On the other hand, GPR55 signalling has also been reported to inhibit Nuclear Factor kappa B (NFkB) (reviewed in a study by Trigo et al³⁵), a master regulator of B-cell activation. Thus, it is tempting to speculate that reduced GPR55 signalling due to diminished GPR55 expression could in turn lead to an enhanced B-cell activation, thus creating a positive feedback loop. This may be of particular relevance under chronic stimulation, such as in autoimmunity. The negative association between GPR55 expression in B-cells and B-cell frequency in patients with RA support this idea. Moreover, the positive association between GPR55 expression and neutrophil-to-lymphocyte ratio also aligns with a context of chronic stimulation. In fact, prolonged GPR55 stimulation has been reported to downregulate expression.³⁶ Collectively, our findings seem to suggest that GPR55 on monocytes and B-cells may not be related to pain, but rather to inflammation and B-cell activation. Decreased GPR55 may thus account for the B-cell overactivation observed in RA, and potentially to altered B-cell compartment in these patients, hence pointing to a potential role as an actionable target to guide therapeutic interventions. Further studies elucidating signalling pathways linked to GPR55 downregulation specifically in B-cells are warranted.

In addition to immune circuits, GPR55 expression was also associated with several matrix metalloproteinases and other mediators such as IL-18, C-X-C motif Ligand 1 (CXCL1) or Thrombospondin-2 (THBS2), which are involved in vascular and extracellular matrix turnover. Pathway analysis from our proteomic platform validated this finding. These results are in line with the existing literature, as altered or blocked GPR55 signalling has been related to impaired remodelling on experimental myocardial infarction,³⁷ ventricular dysfunction³³ or cardiomyocyte hypertrophy.³⁸ Moreover, blocked GPR55 signalling has been linked to increased Matrix Metalloproteinase-9 (MMP-9) expression under inflammation.^{39 40} Furthermore, GPR55 knockdown blunted angiogenesis and endothelial functions, including growth factors production, in vitro.⁴¹ The findings presented in this study are relevant in a twofold manner. First, we set the role of B-cells in the centre of this altered remodelling response. Whereas most of the existing evidence came from broad GPR55 deletion in animal models, our results are suggestive of a specific effect of B-cells. This may explain why mice deficient in GPR55 in the haematopoietic system exhibited an aberrant remodelling response on myocardial infarction, to some extent recapitulating that of cardiac-specific knockout mice,³⁷ despite the stromal compartment being unaffected in the latter. Second, our findings expand previous evidence on the role of B-cell specific GPR55 deficient mice in atherosclerosis development. Lack of GPR55 had been reported to trigger B-cell hyperactivation,⁴ which is also in line with our findings. However, effects beyond the B-cell compartment to account for vascular outcomes were largely unclear. Our study points to the involvement of remodelling and reparative processes in this scenario. GPR55 deficiency has been observed to modify the activity of a number of kinases, which relate to several cellular processes such as CREBP and mTORC2 signalling, response to oxidative stress, or even lipid metabolism.³³ Whether these kinases could also directly activate vascular remodelling, or those events may indirectly trigger this phenomenon, is conceivable. Furthermore, altered tissue repair and matrix remodelling may also be related to limited GPR55 signalling, in a Peroxisome Proliferator–Activated Receptors-gamma (PPARγ)-mediated effect (reviewed in a study by Trigo et al³⁵). Translational studies are needed to identify the presumed mechanisms underlying the connection between B-cell activation and remodelling pathways, either within B-cells or through cell-to-cell communication. Such an identification would be instrumental in elucidating potential therapeutic targets to dampen this phenomenon.

Experimental evidence has highlighted an association between GPR55 and vascular outcomes in animal models, including atherosclerosis. The association between GPR55 downregulation and B-cell activation may be in line with this protective effect. Surprisingly, our study failed to show an association between diminished GPR55 expression and atherosclerosis burden in RA. However, a number of remarks should be considered. Current evidence came from preclinical research using GPR55 knockout models or strong pharmacological inhibition, whereas our study assessed GPR55 expression at specific cellular subset level in real-world patient populations. Additionally, the nature of the association with vascular outcomes in the literature is intriguing. GPR55 deficiency has been linked to plaque size and phenotype, but not to atherosclerosis occurrence itself.⁴ Similarly, infarct occurrence and size were unaffected by GPR55 status,³⁷ although it led to an altered subsequent remodelling response. Equivalent results were observed for cardiomyocytes or left ventricular hypertrophy.^{33 38} Furthermore, GPR55 antagonisation did not affect plaque size or composition under different atherogenic conditions, but augmented plaque stability and infiltration.³⁹ Taken together, these studies seem to rule out a direct, causative effect of decreased GPR55 on the occurrence of vascular outcomes. Alternatively, evidence seems to suggest that diminished GPR55 may create a microenvironment more prone to vascular outcomes via altered remodelling, rather than a vascular insult itself. Experimental evidence also demonstrated that GPR55 led to decreased insulin resistance⁴² or physical activity,⁴³ which are also known to enhance, but not necessarily trigger, vascular outcomes. Additional hits may explain the occurrence of vascular events in certain individuals but not in others (figure 4). This conceptual model aligns with the associations with reparative processes and/or maladaptive responses elsewhere described,^{33 37} also intertwining inflammation. It also supports the hypothesis that GPR55 signalling is required for the maintenance of physiological immune homeostasis under stress conditions.³⁷ This model also aligns with the fact that GPR55 was profoundly decreased overall in the patient group, hence pointing to a ‘common hallmark’ rather than to disease biomarker for stratification. Furthermore, this model also reconciles our findings with the existing literature, since associations with GPR55 were dependent on atherosclerosis status. GPR55 was correlated with remodelling, B-cell proteomic signatures, immunoglobulins, proinflammatory cytokines and metabolomic features, all being related to cardiometabolic risk, in patients with atherosclerosis. On the contrary, GPR55 was unrelated to these features in their atherosclerosis-free counterparts. Interestingly, a negative association was found with BAFF levels, which aligns with the atheroprotective effects recently described for this factor,^{44 45} also in the context of GPR55-deficient models.⁴ Furthermore, differential GPR55 expression among B-cell subpopulations and differences in their BAFF/APRIL dependency, especially for memory cells and plasmablasts, may account for the different picture observed for APRIL and BAFF in our study.46,49 This is also supported by the strong association observed with Transmembrane activator and CAML interactor (TACI, TNFRSF13B) in our proteomic platform. Similarly, response to LPS proteomic signature predominated in patients with atherosclerosis, but not in those without atherosclerosis. This could be in line with the association between systemic LPS exposure and atherosclerosis occurrence in humans.⁵⁰ Moreover, GPCRs have been linked to intestinal barrier permeability,⁵¹ and GPR55 has been linked to intestinal inflammation,⁵² both being major sources of systemic LPS. However, the involvement of the B-cell subset in this regard remains unexplored. Taken together, all these lines of evidence may shed some light to understand how the atheroprotective role of B-cells depends on GPR55. These findings could inform future studies identifying the exact mechanisms by which disturbed B-cells integrate with additional hits for paving the ground for atherosclerosis development.

However, a number of limitations must be considered for this study. Our study is based on treatment-naïve patients, and whether GPR55 expression is normalised on DMARD initiation is unknown. Prospective studies are needed to elucidate this phenomenon. Larger trials including patients in all disease stages beyond early arthritis may be useful in elucidating its potential role as a biomarker for disease stratification. Moreover, our study was based on real-world patient samples, and most of our findings were correlative in nature. Then, mechanistic data linking GPR55 expression to immune cell behaviour should inspire future studies. Transcriptomic studies are warranted to better understand the role of GPR55 on B-cell activation. Moreover, our functional assays represent a proof-of-concept study to link GPR55 downregulation and B-cell activation. Additional experiments are needed to characterise the process of B-cell activation, including B-cell proliferation rates, changes in cellular metabolism and cell differentiation trajectories. Finally, we hypothesise that GPR55 signalling may be diminished in RA due to GPR55 downregulation. In fact, no differences have been reported in Lysophosphatidylinositol (LPI) levels in patients with RA.⁵³ Nevertheless, a systematic characterisation of the GPR55:ligand axis is needed, also capturing the heterogeneity of LPI composition,⁵⁴ as well as the wide diversity of ligands described for GPR55 beyond LPI.⁸

In conclusion, reduced GPR55 expression hallmarked monocytes and B-cell subsets, mainly memory subpopulations, in early arthritis, whereas no differences were found in the preclinical stage. GPR55 expression was linked to B-cell activation-related pathways, especially in patients with atherosclerosis, and presumably via T-cell-independent mechanisms. Associations with inflammatory mediators and immunoglobulin levels were also atherosclerosis status dependent. Functional studies are warranted to validate and ensure clinical translation. These findings open new avenues for treatment aimed at reverting diminished GPR55 signalling and may also help to understand the therapeutic outcomes on certain drugs, as decreased GPR55 expression in certain conditions may limit beneficial effects. Of note, targeted GPR55 stimulation has been proposed to beneficially impact clinical outcomes on infarction occurrence.³⁷ Furthermore, our findings strengthen the crosstalk between neuroendocrine, namely the endocannabinoid system, and immune circuits. Although targeting the endocannabinoid system holds promise to block pain, inflammation and joint destruction in RMDs, data is scarce and mostly limited to conventional CB receptors.⁵⁵ Emerging data from animal models also suggests a positive effect on atherosclerosis prevention,⁵⁶ which may also support this notion. Moreover, our findings provide novel understanding towards the role of B-cells in atherosclerosis, especially in autoimmunity. Finally, the results herein presented expand the knowledge of non-conventional lipids in RA and atherosclerosis, as GPR55 signalling mostly relies on complex, arachidonic-derived lipid species. Collectively, GPR55 may be a novel hub to understand the crosstalk between immune responses and maladaptive remodelling in atherosclerosis.