Immune checkpoint inhibitors (ICIs) have revolutionized the treatment and prognosis of cancer (1). ICIs are monoclonal antibodies against key regulatory proteins such as programmed death 1 and its ligand (PD1-PDL1) or cytotoxic T-lymphocyte-associated protein 4 (CTLA4), enhancing the immune response against cancer cells. Preventing CTLA-4 from binding to its ligands can disrupt immune checkpoint signaling and reduce Treg-mediated immunosuppression. Anti-PD-1 or anti-PD-L1 antibodies reverse the inhibitory signals sent to T cells and activate them. The ICIs acts on their targets, which are the immune cells and, more precisely, on T cells, but the desired effect is achieved indirectly on tumor cells. Reversing the cancer inhibition by ICIs increases the survival rate but exposes the patients to side effects. ICI toxicity refers to a range of inflammatory manifestations collectively termed immune-related adverse events (irAEs). They vary from mild and self-limiting to severe and potentially life-threatening manifestations (2). Rheumatic irAEs are among the most common irAEs and include the development of ICI-induced arthritis (ICI-IA) (3) in 1–7% of the treated patients (4). From a clinical perspective, it usually appears in the first 6 months after ICI initiation, without a gender ratio, as a seronegative oligo-polyarticular arthritis in the absence of rheumatoid factor (RF) or anti-citrullinated protein antibodies (ACPA) (5). Its pathogenesis is largely unknown: the obvious connection with ICIs suggests an immune pro-inflammatory reaction potentially related to T-cell activation (6).

Synovial tissue analysis has been developed in recent years and can be easily obtained by knee mini-arthroscopy and, more recently, by ultrasound (US)-guided synovial biopsies (USGSB). In recent times, the latter procedure is a widespread technique allowing the collection and analysis of high-quality samples of synovial tissue (ST) in various infectious, inflammatory, and non-inflammatory conditions (7). Our group and others have developed studies to analyze the synovial tissue in this early phase of RA (ERA) (8, 9). Different types of synovial pathotypes are described based on the cell types of the immune infiltration. Despite the potential differences between ERA and ICI-IA, some clues link the two. Recent studies have shown that synovial infiltrating T cells in RA patients are PD1-positive and correlate with disease activity. Still, at the same time, PD-L1 is poorly expressed in synovial tissue (10, 11). Thus, an agonistic action on PD1 in RA patients is believed to inhibit T cells. A first phase two trial with peresolimab has recently suggested a potential effect of anti-PD1 in RA (12).

This study aimed to provide a histopathological assessment of ICI-IA tissue, with a further comparison with a control cohort of untreated ERA patients.

In this study, we have presented the first synovial tissue histological evaluation of an ICI-IA cohort, compared to ERA.

We showed that ICI-IA and ERA share similar clinical and histological characteristics.

From the clinical point of view, ICI-IA has been extensively deciphered (5). It is typically an oligo-polyarthritis involving small, medium, and large joints resembling RA. Our cohort is in line with these observations. Indeed, when comparing ICI-IA with ERA, we did not identify any difference in terms of disease activity (DAS28), joint involvement (TJC and SJC), and serum inflammation (CRP) levels. The only major discriminant was the absence of ACPA antibodies in ICI-IA, confirming the seronegative profile of this condition.

ANAs have been historically associated with cancer as an epiphenomenon, and their prevalence could reach 44.4% also in patients without any signs or symptoms of autoimmune diseases such as systemic lupus erythematosus, Sjogren’s Syndrome, or systemic sclerosis (17–19). More recently, they have been associated with a higher risk of developing irAEs, but more studies are necessary to confirm these observations (20, 21). We observed a prevalence of 36.4% in ICI-IA, which was not different from that found in RA (6/18, 33.3%).

ICI-IA has a favorable outcome and is controlled by corticosteroids and immunosuppressants. bDMARDs have also been positively evaluated in such conditions. Compared to RA, we do not know the long-term course or evolution of ICI-IA. In the synovial tissue analysis, ICI-IA demonstrated significant levels of synovial inflammation, including sublining hyperplasia, inflammatory infiltrate, fibrinoid necrosis, and hypervascularization, all found at similar levels in the ERA cohort. The immunohistochemical analysis further confirmed comparable levels of immune cell infiltration—specifically CD3 + T cells, CD20 + B cells, CD68 + macrophages, and CD138 + plasma cells—between the two groups.

However, it should be underlined that hypervascularization tends to be lower in ICI-IA than in ERA. Hypervascularization is a typical histological feature of different inflammatory arthritides, including RA and Psoriatic Arthritis (PsA) (22). In this study, we show relatively low vascularization in ICI-IA compared to ERA, which is the only feature that differentiates them. While the low sample size requires external confirmation, this finding suggests that a deeper investigation into mechanisms involved in angiogenesis, such as vascular cell adhesion molecule 1 (VCAM-1) and intercellular adhesion molecule 1 (ICAM-1), could unravel differences between the diseases (23).

These similarities regarding clinical appearance and pathology may suggest a common immune activation between the two entities. ICI-IA represents the prototype of inflammatory arthritis and, more generally, of autoimmune diseases: once triggered by an appropriate stimulus, such as ICIs for ir-AEs, the immune system breaks self-tolerance control and targets self-tissues, such as the synovium. On the other hand, RA shares the same pathogenic model, but the triggering stimulus has never been unequivocally demonstrated. However, in addition to the initial trigger, ICI-IA and RA share common ground, including biological pathways and immune dysregulations. PD-1, for instance, plays a central role in both conditions. It is a key regulator of the cellular cycle whose expression on CD4 + and CD8 + lymphocytes reflects T-cell activation, and its binding to PD-L1 and PD-L2 inhibits lymphocytes. Targeting PD1 with blocking antibodies is the rationale behind the development of cancer immunotherapy, leading to the development of several anti-PD1 drugs such as pembrolizumab, nivolumab, and cemiplimab. On the other hand, these same pathways have been shown to play a role in inflammation and autoimmune diseases. Indeed, PD-1 is highly expressed in RA-inflamed synovium, especially on infiltrating T lymphocytes (24, 25), and its targeting showed promising results in a phase two trial (10). The cell cycle regulator CTLA4 shows activity in both directions and has become a therapeutic target in oncology and rheumatology. Ipilimumab and abatacept manipulate the same immune pathway in opposite directions, demonstrating once again extensive overlap of immune mechanisms in cancer and autoimmunity (26).

Despite the growing interest in ICI-IA and ir-AEs in general, knowledge about them is still scarce. However, some studies suggest a prominent role of T lymphocytes over B cells, which may also explain the extremely low ICI-IA seropositivity for ACPA in ICI-IA. More specifically, shared CD8 T-cell clones may be found between the tumor, the lymph nodes, the blood, and the inflamed tissue in cases of irAEs (27, 28). We know now that “toxic” CD8 T cells can also arise directly on-site (29). The specificity of these pathogenic lymphocytes has recently been better characterized as a disease-specific T CD8 subpopulation (CD38high) present both in ST and blood.

On the other hand, RA is a complex disease with numerous determinants, including a key role for B cells and autoantibodies. However, evidence also supports a T-cell-driven mechanism in its pathogenesis: shared T-cell clones have been identified between the joints and lungs of RA patients with interstitial lung disease (30), as well as among inflamed joints within individual RA patients (31). This finding suggests that ICIs may activate this CD8 T-cell phenotype, which in turn contributes to the development of arthritis. In this case report, the histologic features of ST were identical to those observed in RA (32), as previously shown in another case report (33). Taken together, these observations support the hypothesis of a common immune dysregulation between ICI-induced arthritis and ERA.

Beyond its clinical resemblance to RA, ICI-IA represents a rare opportunity to directly study inflamed tissue in the context of immune-related adverse events (irAEs). Compared to other irAEs—such as pneumonitis, colitis, or myocarditis—arthritis can be studied more easily due to the safety, feasibility, and reproducibility of accessing the synovial membrane. In this setting, the study of synovial tissue in ICI-IA offers a valuable model for investigating local immune dysregulation triggered by ICIs. Ultrasound-guided synovial biopsy (USGSB), a minimally invasive and widely adopted procedure in major academic centers, enables direct sampling of affected tissue, allowing for in-depth immunopathological analyses that are rarely feasible in other irAEs.

We acknowledge several limitations in our study. First, the low numerosity of the cohort necessarily impacts the statistical analysis when compared with the ERA cohort; however, in terms of clinical presentation, our cohort is in line with several previous cohorts, assuring its good representativity.

A potential limitation directly linked to patient enrollment is the possibility of referral bias, as some patients may not have been evaluated due to prior corticosteroid treatment, comorbidities (e.g., anticoagulants), or delayed referral; moreover, glucocorticoids are sometimes administered together with chemotherapy and immunotherapy, further complicating patient inclusion.

Second, we compared ICI-IA with a sex- and age-matched ERA cohort. RA is a female-gender-unbalanced disease, while ICI-IA is slightly more common in men⁵. Moreover, ERA tends to be often seropositive for RF and ACPAs. However, our primary objective was to compare ST tissue and male/female histology differences that had never been demonstrated in RA. Third, our study has some limitations, such as the absence of ultrasound features of ICI-IA compared to RA, since we did not systematically perform power Doppler evaluation in all joints. However, by selecting at least a Grade 2 synovitis according to the synovial hypertrophy score, we ensure a representative tissue collection of the synovial inflammatory process. Finally, histology and IHC have limited power in decrypting underlying pathogenic mechanisms, as already demonstrated in RA, and we cannot speculate on similarities between ICI-IA and RA. Histology has low specificity, and our study is not sufficient to propose a biological overlap between the two diseases. More mechanistic studies, including transcriptomic ones, are necessary to confirm these hypotheses.

In conclusion, for the first time, we performed a histological analysis of ST from a cohort of ICI-IA patients. When comparing it with untreated ERA, we showed a similar histological profile.