When inflammation is not the driver: Pauci-immune synovitis and the stromal basis of refractory arthritis

Abstract

Approximately 5% to 20% of patients with rheumatoid arthritis (RA) fail to respond to sequential biologic and targeted synthetic disease-modifying antirheumatic drugs, representing a major unmet clinical need. Increasing evidence indicates that a substantial proportion of these patients exhibit a pauci-immune (fibroid) synovial pathotype, characterised by sparse immune-cell infiltration and enrichment of fibroblast, extracellular matrix, and cell adhesion pathways rather than inflammatory immune signatures. Notably, this pathotype is present in up to one quarter of treatment-naïve patients with early RA. Clinically, it is associated with persistent disease activity and pain despite relatively low systemic inflammation and predicts poor therapeutic responses across multiple drug classes, including tumour necrosis factor inhibitors, rituximab, and tocilizumab. Recent advances in spatial transcriptomics and single-cell technologies have begun to elucidate the biological basis of this phenotype, identifying an endothelial–fibroblast signalling network driven by Notch and transforming growth factor beta pathways that spatially orchestrates fibrogenic responses. These studies have also defined distinct sublining fibroblast subsets, including DKK3⁺ fibroblasts associated with treatment refractoriness and CD200⁺ fibroblasts linked to resolution of inflammation. Together, these findings shift the pathogenic paradigm from immune-cell–driven inflammation to stromal cell dysfunction in a biologically distinct subset of RA, positioning stromal cells as promising therapeutic targets. This review synthesises the clinical, molecular, and mechanistic evidence supporting the pauci-immune synovial pathotype, and discusses its implications for precision medicine and treatment stratification.

Introduction

The treatment of inflammatory arthritis has been transformed over the past 2 decades by the advent of biologic and targeted synthetic disease-modifying antirheumatic drugs (DMARDs). However, approximately 5% to 20% of patients with rheumatoid arthritis (RA) fail to achieve adequate disease control despite sequential treatment with multiple biologic therapies [1], highlighting the existence of pathogenic mechanisms that are not sufficiently addressed by current immunomodulatory approaches. Increasing evidence suggests that, in a subset of patients, persistent disease is not primarily sustained by ongoing immune-cell activation but rather by intrinsic alterations of the synovium itself. This concept has led to the recognition of the pauci-immune (fibroid) synovial pathotype, a distinct tissue endotype characterised by limited immune-cell infiltration and a dominant stromal signature involving fibroblast activation, extracellular matrix remodelling, and altered tissue architecture. The pauci-immune pathotype is associated with reduced responsiveness to conventional immunosuppressive therapies, suggesting that immune-independent mechanisms contribute to disease persistence. This review summarises current evidence defining the pauci-immune synovial pathotype, explores its molecular and cellular determinants, discusses its clinical implications, and examines how improved understanding of this distinct disease endotype may contribute to precision medicine approaches and the development of novel stromal-targeted therapies in inflammatory arthritis. A graphical summary is presented in Figure 1.

Figure 1.

Graphical summary. COMP, cartilage oligomeric matrix protein; CRP, C-reactive protein.

Defining the Pauci-Immune Synovial Pathotype

Recent years have revolutionised the understanding of the pathogenesis of RA by deciphering the heterogeneity of the synovium at the cellular and molecular levels [[2], [3], [4], [5]]. In particular, 3 main synovial pathotypes have been described: the diffuse-myeloid pathotype with myeloid lineage predominance but poor in B cells, the lympho-myeloid dominated by the presence of B cells in addition to myeloid cells and the pauci-immune (fibroid) pathotype, which contains few or no inflammatory immune cells and is instead enriched in stromal components, including tissue-resident fibroblasts, endothelial cells, and extracellular matrix proteins [3]. At the molecular level, the pauci-immune pathotype is characterised by activation of fibroblast-associated transcriptional programmes, extracellular matrix remodelling, and cell adhesion pathways rather than classical inflammatory cytokine-driven responses [3,6,7]. These findings suggest that in this subset of patients, stromal cells are not merely passive structural components but represent active regulators of synovial pathology and potential drivers of persistent disease. Importantly, the pauci-immune pathotype does not simply reflect inactive, residual, or long-standing disease. It is already detectable in approximately 20% to 25% of treatment-naïve patients with early RA, demonstrating that this phenotype represents an intrinsic biological subtype rather than a consequence of prolonged inflammation or therapeutic exposure [3,8]. Moreover, patients exhibiting a pauci-immune pathotype frequently present with clinically active disease despite limited immune-cell infiltration within the synovium, emphasising the dissociation between clinical disease activity and tissue inflammatory burden [8,9].

Although initially described in RA, stromal-dominant synovial phenotypes have also been identified in other arthritic conditions, including psoriatic arthritis, juvenile idiopathic arthritis, and osteoarthritis. This suggests that activation of fibrostromal programmes may represent a conserved mechanism of chronic arthritis [10].

Clinical Implications

The recognition of synovial pathotypes has challenged the traditional assumption that clinical disease activity directly reflects the extent of immune-cell–mediated inflammation. In patients with a pauci-immune synovial phenotype, active disease can persist despite limited infiltration by inflammatory immune cells, highlighting the possibility of a dissociation between clinical inflammation and tissue pathology [8,9]. This observation has important implications for both disease assessment and therapeutic decision-making.

Compared with immune-rich synovial pathotypes, patients with pauci-immune synovitis frequently exhibit higher levels of pain, fatigue, and tender joint counts despite comparable composite disease activity scores. These clinical features are accompanied by lower systemic inflammatory markers and less structural progression [3,8,9]. Together, these observations suggest that symptoms and disease persistence may be driven by mechanisms beyond classical inflammatory pathways.

Consistent with this concept, the pauci-immune pathotype has been associated with therapeutic resistance across multiple drug classes. Nerviani et al [8] showed that patients with baseline pauci-immune synovitis had a poor response to the tumour necrosis factor (TNF) inhibitor certolizumab pegol, with response rates of only 28.6% compared with 83.3% in both the lympho-myeloid and diffuse-myeloid pathotypes. Similar observations were subsequently reported for conventional synthetic DMARDs as well as for B cell depletion [3,11]. The biopsy-driven phase 4 randomised controlled trial (rituximab vs tocilizumab in anti-TNF inadequate responder patients with rheumatoid arthritis (R4RA)) further extended these findings by demonstrating that the pauci-immune pathotype predicted poor response to both rituximab and tocilizumab, despite these agents targeting different immunological pathways [9]. Importantly, patients enrolled in R4RA had already failed conventional synthetic DMARDs and TNF inhibitors, indicating resistance across multiple therapeutic mechanisms.

Consistently, transcriptomic analyses of the synovium of patients refractory to all medications have identified enrichment of stromal and fibroblast-associated signatures characteristic of the pauci-immune/fibroid phenotype [9]. Similarly, cohorts of patients with noninflammatory, multirefractory RA demonstrated a predominance of pauci-immune synovitis, with markedly reduced immune-cell infiltration compared with nonrefractory disease [12]. Together, these findings suggest that persistent disease in this subgroup is maintained by mechanisms that are insufficiently targeted by current immunosuppressive strategies. When immune cells are largely absent from the tissue, further escalation of immune-directed therapy may provide limited benefit, whereas targeting stromal activation pathways may represent a more rational therapeutic strategy.

Beyond its mechanistic relevance, synovial pathotyping has demonstrated potential clinical utility for treatment stratification. In the Pathobiology of Early Arthritis Cohort, incorporation of synovial pathotype information into predictive models improved the accuracy of treatment response prediction from 78.8% to approximately 89% to 90%, supporting the concept that tissue-based biomarkers may complement clinical and serological parameters in guiding therapeutic decisions [13].

Mechanisms Underlying Pauci-immune Synovitis

Although the clinical relevance of the pauci-immune pathotype is increasingly recognised, the underlying biological mechanisms remain incompletely understood. Emerging evidence indicates that pauci-immune synovitis is not characterised by passive loss of inflammation, but rather by active reprogramming of the stromal compartment, involving fibroblast activation, extracellular matrix remodelling, and altered endothelial–mesenchymal communication.

Stromal activation and fibrogenic programming

A growing body of evidence links treatment-refractory, noninflammatory RA to activation of profibrotic pathways, including transforming growth factor beta (TGF-β) signalling, extracellular matrix remodelling, and expansion of fibroblast populations expressing cartilage oligomeric matrix protein (COMP) [14]. Spatial transcriptomic analyses have provided further insight into the organisation of these fibrogenic programmes within the synovial microenvironment. In patients who subsequently develop persistent nonremitting disease, fibrogenic signatures are already detectable at baseline within perivascular niches, characterised by increased COMP expression in adjacent fibroblasts. These observations suggest that the anatomical organisation of stromal cells within specific tissue niches may determine the trajectory towards treatment resistance.

Endothelial–fibroblast communication

Interestingly, one spatial study has identified an endothelial–fibroblast regulatory axis involving Notch and TGF-β signalling, which creates localised gradients of fibroblast activation [14]. Through Notch signalling, endothelial cells exert dual, opposing control over neighbouring fibroblasts by promoting TGF-β ligand expression while simultaneously modulating TGF-β receptor availability on fibroblasts. This creates a spatial gradient of TGF-β responsiveness that is highest close to vascular structures and progressively decreases with distance. Such organisation provides a mechanistic explanation for why fibroid lesions appear as spatially restricted niches rather than diffuse tissue changes: perivascular regions become highly active fibrogenic hubs, characterised by COMP-high fibroblasts and extracellular matrix deposition, whereas more distant regions may retain inflammatory characteristics.

This endothelial–stromal interaction may also carry therapeutic implications. In patient-derived synovial organoids, combined inhibition of Notch and TGF-β signalling has been shown to reverse the fibrogenic programme, supporting the concept that stromal communication pathways may be a targetable pathway [14].

Fibroblast heterogeneity

Single-cell and spatial transcriptomic studies have further revealed that synovial fibroblasts are not a homogeneous population but comprise multiple functionally distinct subsets with different biological roles [7,15]. Beyond their classical role as matrix-producing cells, fibroblasts actively shape immune-cell localisation through at least 2 distinct programmes [16]. Interferon gamma–responsive fibroblasts form lymphocyte-permissive niches that retain lymphocytes and sustain chronic inflammation. In contrast, TGF-β–responsive, matrix-synthesising fibroblasts build a reparative niche, composed of a collagen-rich perivascular barrier that restricts leukocyte migration and promotes resolution of tissue inflammation or profibrotic pathways in excess.

Single-cell and spatial transcriptomics have further resolved this reparative programme into distinct sublining fibroblast subsets. In the R4RA trial, a DKK3⁺ fibroblast population was enriched in patients with refractory RA who failed anti-TNF therapy and subsequently failed both rituximab and tocilizumab, suggesting an association between specific DKK3⁺ fibroblast subpopulations and therapeutic resistance [9]. Conversely, CD200⁺DKK3⁺ fibroblasts have been identified as a pro-resolving population that emerges during the resolution phase of inflammation. Through interaction with type 2 innate lymphoid cells via the CD200–CD200R1 axis, these fibroblasts promote the establishment of a reparative mesenchymal niche [17]. These findings support the concept that the biological function of DKK3⁺ fibroblasts is determined by their molecular context rather than DKK3 expression itself and that failure of inflammatory MMP3⁺/IL6⁺ fibroblasts to transition towards a pro-resolving CD200⁺DKK3⁺ state may represent one mechanism underlying treatment refractoriness [17].

Conclusion

The recognition of the pauci-immune synovial pathotype represents a conceptual shift in our understanding of RA, moving from an exclusively immune-centred model towards one in which dysregulated stromal cells actively maintain chronic disease in a subset of patients. The pauci-immune synovial pathotype therefore represents an actively maintained tissue state characterised by fibroblast reprogramming, endothelial–stromal crosstalk and altered extracellular matrix biology. This mechanistic framework provides a rationale for developing therapies that move beyond immunosuppression towards stromal reprogramming, including strategies targeting fibroblast activation states, endothelial–mesenchymal communication pathways, and mechanisms controlling tissue resolution.

Future studies should focus on defining the temporal stability of this pathotype, identifying reliable noninvasive surrogate biomarkers, and determining whether treatment algorithms guided by synovial biology can improve outcomes for patients with refractory disease (Table 1).

Table 1.

Research agenda

•Elucidate the biological mechanisms that drive the pauci-immune synovial pathotype, with a particular focus on endothelial–stromal interactions and the role of Notch signalling.

•Define the origins, plasticity, and functional roles of distinct fibroblast populations, including COMP⁺, MMP3⁺/IL6⁺, and CD200⁺/DKK3⁺ fibroblasts.

•Determine whether failure of tissue-resolution programmes contributes to treatment refractoriness.

•Establish whether the pauci-immune pathotype represents a stable endotype or a dynamic state.

•Identify robust noninvasive biomarkers capable of detecting and monitoring the pauci-immune phenotype.

•Evaluate the therapeutic potential of stromal-targeted interventions.

•Assess whether synovial pathotype-guided treatment strategies can improve outcomes in refractory rheumatoid arthritis.

COMP, cartilage oligomeric matrix protein.

Competing interests

Grant/research support from Vontobel Stiftung, Pfizer, Novartis Foundation for Bio-Medical Research; Iten Kohaut Foundation, Kurt und Senta Herrmann Foundation, Foundation for Research in Rheumatology (FOREUM), University Zurich; Congress Support from AstraZeneca and Janssen. Speaker and consultancy fees from Boehringer Ingelheim.