Rheumatoid arthritis (RA) is a chronic, systemic, inflammatory autoimmune disease that affects up to 1% of the population (1). Inflammation predominantly targets the synovial tissue (ST; the joint lining), leading to pain, swelling and irreversible joint deformities. Disease complications are not confined to joints, and include systemic involvement (pulmonary, cardiac, haematological, amongst others). Despite treatment advances, a majority of patients do not attain remission, and sustained remission occurs in as few as 15% of patients (2). Although the eventual inflammatory milieu in the RA ST is diverse, checkpoint, or co-stimulatory pathways controlling immune cell activation are critical in engendering, maintaining, and perpetuating the inflammatory response, driven by chronic autoimmune antigenic drive. In parallel with the rise of immune checkpoint inhibitors (ICIs) that have transformed oncologic therapeutics, has been the greater recognition of checkpoint pathways in RA pathogenesis.

In this review, we provide an overview of immune checkpoints at various stages of immune cell development and describe the involvement of these in the development and progression of RA. We discuss the involvement of immune checkpoints in RA pathobiology, the relationship between immune checkpoint therapy in cancer and autoimmune adverse events, and the potential to manipulate these pathways as a future treatment strategy.

T cell activation is a tightly regulated, intricate, and specific process requiring multiple initiating signals. The first is provided through the interaction between a peptide antigen presented by an antigen presenting cell (APC), and a T cell expressing a cognate T cell receptor (TCR). The T cell then requires a secondary signal, referred to as co-stimulation, and the proteins that provide this are called co-stimulatory, or checkpoint molecules. In helper T cells, this is provided through the B7/CD28 co-stimulation axis (3). CD28 expressed on the T cell binds to either CD80 (B7.1) or CD86 (B7.2) on the APC, and together with TCR signalling, provides survival and activation signalling to initiate T cell proliferation ( Figure 1 ). The extent of co-stimulation controls the extent of the initial T cell proliferative response (4), and further signalling provided through specific cytokines dictates cell lineage (5).

Checkpoint molecule expression is largely controlled by the tissue microenvironment, and is influenced by alterations in homeostasis (6). These molecules can exert stimulatory or inhibitory signals in either direction of the interaction (i.e., in the APC and in the T cell) and cumulatively dictate the threshold of T cell activation, allowing for finetuning of the immune response (7, 8). Inhibitory checkpoints, including cytotoxic T-lymphocyte-associated protein 4 (CTLA-4) and programmed death protein 1 (PD-1) are upregulated during cell activation, and are important for limiting the response duration, aiding resolution, and maintaining self-tolerance (7). The importance of immune checkpoints in restricting the immune response is exemplified by disturbances in these pathways. Patients with inherited CTLA-4 deficiencies present with a range of clinical features including recurrent respiratory tract infections and increased risk of autoimmunity (9), while inherited PD-1 deficiency has been linked with early-onset autoimmunity and severe tuberculosis (10). Mice deficient in CTLA-4 experience lethal lymphoproliferative disorders (11), and those deficient in PD-1 develop accelerated autoimmunity and lethal pathology in response to acute infection (12–14).

The importance of the T cell response in RA has long been appreciated (15–18) and evidence suggests T cell-driven autoimmunity may occur years prior to clinical onset (19, 20). Distinct changes have been observed in the proportions of T follicular helper (Tfh) and CD19⁺ B cells in the lymph nodes of patients at risk of developing RA (individuals with rheumatoid factor (RF) and anti-citrullinated protein antibodies (ACPA) but asymptomatic for arthritis) (21), and altered proportions of naïve T cells and T regulatory cells (Treg) have been observed in the peripheral blood (22). In lymph nodes, stromal cells from at-risk patients had distinct hypomethylated sites in genes associated with antigen processing and presentation compared to healthy controls, suggesting altered antigen presentation pathways prior to clinically identifiable disease (23), implicating stromal-lymphoid cell interactions in early RA. Finally, it is well established that RF and ACPA can often be detected in the serum for years in individuals who progress to develop inflammatory arthritis (20, 24).

In clinically identifiable early and established RA, T cells accumulate in the ST sublining layer, where they’re found in association with activated macrophages and B cells (16), and often cluster in aggregates consisting of T cells, B cells, and dendritic cells (DC) (17). While there is enormous variability in synovitis in RA (17, 25), evidence suggests a T helper 1 (Th1) bias in the ST (17), and the demonstration of T cell involvement across all stages of disease suggests that T cell targeting may be efficacious therapeutically broadly throughout disease progression.

Aligning with the importance of the T cell response in RA progression, CTLA-4 fusion protein (CTLA-4-Ig; Abatacept), a soluble protein that binds CD80/CD86 expressed by activated APCs and outcompetes CD28 expressed by T cells at the beginning of the immune response, was approved for use in RA in 2005. Abatacept therefore inhibits circulating and synovial T cell activation by inhibiting CTLA-4:CD28 interactions, and has shown remarkable efficacy in disease management (26), reducing progression of undifferentiated arthritis to RA, and reduction of radiological progression and levels of anti-CCP antibodies (27). Early outcomes from two ongoing trials further suggest Abatacept may also prevent RA development in at-risk healthy individuals (28–30).

In cancer, tumour cells can exploit checkpoint pathways controlling the adaptive immune response to neoantigens, inducing tolerance and immune evasion (31, 32). ICIs targeting co-inhibitory receptors in the tumour microenvironment have had remarkable success by unleashing the anti-tumour function of T cells. However, ICI usage is complicated by the occurrence of immune-related adverse events (irAEs) which can present with broad manifestations varying by ICI type (33). Additionally, while combinations of ICIs have shown improved oncologic therapeutic response compared to monotherapies (34), the proportion of patients who developed treatment-related severe irAEs were markedly higher with combination therapy (55%) compared with those in the nivolumab- (16.3%) or ipilimumab-only groups (27.3%), demonstrating that combination therapy may dramatically increase irAE risk (34).

Rheumatological complications occur in ~5-10% of patients treated with ICIs (35), and can manifest in the form of joint inflammation indistinguishable from RA (36), which can persist after therapy cessation (37) and exacerbate existing RA symptoms (38). Additionally, inflammatory arthritis is the most common irAE reported in trials of ICIs (39), an occurrence that has offered a serendipitous window into RA pathogenesis.

Research into other inhibitory immune checkpoints including TIGIT, TIM-3, and LAG-3 is currently underway in cancer (32, 95), and are tantalising future targets in RA. Increased TIGIT expression on CD4⁺ RA T cells correlated with autoantibody levels and DAS28 (96), and an agonistic anti-TIGIT antibody, capable of inhibiting Tfh and Tph cells and enhancing Treg function, has recently been described (97). Tregs expressing TIGIT supressed Th1 and Th17 cell responses (98) and signalling through TIGIT on functionally defective Tregs from patients with multiple sclerosis restored suppressive function (99), highlighting the potential of TIGIT-targeting. LAG-3 is constitutively expressed on Tregs; LAG-3 positive Tregs in peripheral blood were reduced in RA patients with higher disease activity, and increased following abatacept therapy (100). Additionally, soluble LAG-3 was increased in early and established RA, correlated with ACPA/RF status and erosive progression in early RA, and decreased inflammatory cytokine release in chronic RA (101). Finally, T cell expression of TIM-3, was increased in RA patients and correlated with decreased DAS28 (102).

Leukocyte-associated immunoglobulin receptors 1 and 2 (LAIR-1 and LAIR-2, respectively) are inhibitory immune receptors for collagens that are expressed broadly across the immune compartment (103, 104). LAIR-1 engagement by collagen or the complement component C1q induces inhibitory T cell signalling (105), and in collagen-induced arthritis, administration of anti-LAIR-1 antibodies significantly attenuated disease (106, 107). In RA, T cell expression of LAIR-1 is reduced compared to healthy and osteoarthritis controls and elevated on ST monocytes and macrophages (108). LAIR-2 is a secreted homolog capable of inhibiting LAIR-1-mediated inhibitory signalling, and is elevated in the RA synovial fluid (109), suggesting dysregulated LAIR-1 action in RA. Together, these suggest that manipulating inhibitory LAIR-1 signalling may offer a future therapeutic approach for RA.

VSIG4 (complement receptor immunoglobulin; CRIg) is a B7-family related protein capable of modulating T cells through interaction with a yet to be identified receptor and has additional roles in immunosuppression (110, 111). VSIG4 binds the complement fragment C3b, inhibiting the alternative complement pathway (112), and approaches utilising this function have shown efficacy in rodent models of RA (113, 114). Administration of soluble VSIG4 reduced disease severity and prevented bone erosion in murine arthritis models (114), and administration of the dual complement inhibitor CRIg-CD59 alleviated symptoms in an adjuvant-induced arthritis rat model (113). In humans, VSIG4 is expressed by tissue-resident regulatory macrophages in the healthy and RA ST from patients in remission (115), and importantly, polymorphisms in VSIG4 have been associated with RA severity (116).

Finally, the signalling lymphocytic activation molecule (SLAM) family has gained recent attention in RA. SLAMF6 was increased in osteoarthritis and early RA ST (117), and in circulating early RA PD-1⁺CD4⁺ cells (54) and inhibition of SLAMF6 on CD4⁺PD-1⁺ peripheral cells from established RA patients decreased IgG production and plasmablast differentiation (53).

With their undeniable role in the RA autoimmune milieu, the potential of targeting immune checkpoint pathways with novel approaches is an area of immense interest in RA research. These include modulation utilising small molecules including non-coding microRNAs, and their use in downregulating inhibitory checkpoint pathways are currently under investigation in cancer (118, 119). These are attractive because of their ability to target not only a single gene, but entire pathways (120). Similar strategies capable of inducing these pathways may therefore be efficacious in alleviating autoimmune inflammation. Indeed, MEG3, a long non-coding RNA can modulate TIGIT expression on CD4⁺ T cells from aplastic anaemia patients (121).

The co-expression of checkpoint molecules on individual cells suggests the approach of targeting multiple checkpoints simultaneously, and there has been significant recent interest in targeting dual targets by bi-specific antibodies. CTLA-4/OX40 (ATOR-1015) (122), and PD-L1/CTLA-4 (KN046) bi-specifics are currently in trial for use in cancer (123). In RA, bi-specific approaches have the potential advantage of targeting checkpoint molecules specifically in the inflammatory niche. For instance, a bi-specific targeting TNF and synovial-specific domain scFv-A7, shows remarkable tissue and disease specificity for the microvascular compartment of the human arthritic ST (124). Interestingly, fewer irAEs appear to be observed following the use of bi-specific antibodies in cancer therapy compared to combination therapy, likely as a result of their site-specific nature (125).

Finally, while the adverse events associated with the use of ICIs in cancer therapy are reasonably well documented, the current lack of agonistic antibodies targeted to checkpoint pathways in human trials makes it difficult to remark on the potential risks associated with the use of checkpoint agonists for the treatment of chronic diseases such as RA. In initial trials, severe adverse events were documented while testing an anti-CD28 agonist in healthy control patients (126), while early trials of Peresolimab in RA do not suggest the development of severe adverse events (61), and in cancer, agonistic antibodies targeting OX40 showed acceptable safety profiles (127). However longer, and larger trials are required to fully assess safety profiles and effect on oncologic risk (61).

The role of immune checkpoint molecules in RA is a growing area of research, and the elucidation of how these pathways influence the breaching of self-tolerance and drive RA pathogenesis will undoubtedly be critical for selecting the most appropriate therapy for individual patients. However, much remains to be explored, from deciphering the complex interactions between immune cells to fine-tuning the delicate equilibrium of immune activation and suppression. As this field progresses, agonistic immune checkpoint-based therapies may emerge as valuable therapeutic tools for treating RA and may offer new avenues for improved patient care and quality of life.

AS: Conceptualization, Writing – original draft, Writing – review & editing. KL: Writing – original draft, Writing – review & editing. MW: Conceptualization, Funding acquisition, Project administration, Resources, Supervision, Writing – review & editing.