Background/Purpose: The field of autoimmunity may benefit from the knowledge gained by studying immune checkpoint inhibitors (ICIs). These agents have shown enormous efficacy in oncology by (re)activating T-cells to assault cancer cells, but come at the cost of immune-related adverse events (irAEs). Since ICI’s mode of action is not antigen-specific, we hypothesized that tolerance may be broken not only to tumor antigens but also to autoantigens, leading to the formation of autoantibodies. Therefore, we investigated whether patients treated with ICIs develop autoantibodies, and whether this trait is associated irAEs and ICI efficacy.

Methods: In pre- and post-treatment sera of 133 ipilimumab (anti-CTLA4)-treated melanoma patients, we determined 23 common clinical autoantibodies associated with autoimmune diseases (Figure 1). The association between autoantibody development and irAEs (under ipilimumab or subsequent anti-PD-1 therapy), best overall response, and overall survival was investigated.

Results: Autoantibodies developed in 19.2% (19/99) of pre-treatment autoantibody-negative patients (p<0.0001; Figure 1). A non-significant association was observed between development of any autoantibodies and any irAEs: 5/19 (78.9%) patients that developed any autoantibodies had irAEs, versus 46/80 (57.5%) patients that did not develop autoantibodies (OR: 2.92 [95% CI: 0.85 to 10.01]). Predominantly anti-TPO (4.8%, 6/125) and anti-TG antibodies (6.0% , 8/132) developed (p=0.03 and p=0.008, respectively). Patients with anti-thyroid antibodies after ipilimumab had significantly more thyroid dysfunction under subsequent anti-PD-1 therapy: 7/11 (54.6%) patients with anti-thyroid antibodies after ipilimumab developed thyroid dysfunction under anti-PD1, versus 7/49 (14.3%) patients without antibodies (OR: 9.96 [95% CI: 1.94 to 51.1]). For most other autoantibodies, including RA-associated antibodies, post-treatment positivity increased only marginally and was not associated with occurrence of irAEs in the organ system related to the autoantibody specificity. Becoming autoantibody positive showed a trend towards better overall survival (HR for all-cause death: 0.66 [95% CI: 0.34 to 1.26]; Figure 2) and therapy response (OR: 2.64 [95% CI: 0.85 to 8.16]).

Conclusion: Breaking of humoral tolerance as measured by development of autoantibodies is relatively common under treatment with ipilimumab and holds promise as a marker of ICI toxicity and efficacy. The nature of the autoantigens towards which tolerance is broken is not reflected in the phenotype of the irAEs.