Association of germline TYK2 variation with lung cancer and non‐Hodgkin lymphoma risk

Abstract Deucravacitinib, a novel, selective inhibitor of TYK2 is currently under review at the FDA and EMA for treatment of moderate‐to‐severe plaque psoriasis. It is unclear whether recent safety concerns (ie, elevated rates of lung cancer and lymphoma) related to similar medications (ie, other JAK inhibitors) are shared with this novel TYK2 inhibitor. We used a partial loss‐of‐function variant in TYK2 (rs34536443), previously shown to protect against psoriasis and other autoimmune diseases, to evaluate the potential effect of therapeutic TYK2 inhibition on risk of lung cancer and non‐Hodgkin lymphoma. Summary genetic association data on lung cancer risk were obtained from a GWAS meta‐analysis of 29 266 cases and 56 450 controls in the Integrative Analysis of Lung Cancer Risk and Aetiology (INTEGRAL) consortium. Summary genetic association data on non‐Hodgkin lymphoma risk were obtained from a GWAS meta‐analysis of 8489 cases and 374 506 controls in the UK Biobank and InterLymph consortium. In the primary analysis, each copy of the minor allele of rs34536443, representing partial TYK2 inhibition, was associated with an increased risk of lung cancer (OR 1.15, 95% CI 1.09‐1.23, P = 2.29 × 10−6) and non‐Hodgkin lymphoma (OR 1.18, 95% CI 1.05‐1.33, P = 5.25 × 10−3). Our analyses using an established partial loss‐of‐function mutation to mimic TYK2 inhibition provide genetic evidence that therapeutic TYK2 inhibition may increase risk of lung cancer and non‐Hodgkin lymphoma. These findings, consistent with recent reports from postmarketing trials of similar JAK inhibitors, could have important implications for future safety assessment of deucravacitinib and other TYK2 inhibitors in development.

Summary genetic association data on non-Hodgkin lymphoma risk were obtained from a GWAS meta-analysis of 8489 cases and 374 506 controls in the UK Biobank and InterLymph consortium. In the primary analysis, each copy of the minor allele of rs34536443, representing partial TYK2 inhibition, was associated with an increased risk of lung cancer (OR 1.15, 95% CI 1.09-1.23, P = 2.29 Â 10 À6 ) and non-Hodgkin lymphoma (OR 1.18, 95% CI 1.05-1.33, P = 5.25 Â 10 À3 ). Our analyses using an established partial loss-of-function mutation to mimic TYK2 inhibition provide genetic evidence that therapeutic TYK2 inhibition may increase risk of lung cancer and non-Hodgkin lymphoma. These findings, consistent with recent reports from postmarketing trials of similar JAK inhibitors, could have important implications for future safety assessment of deucravacitinib and other TYK2 inhibitors in development.

What's new?
Increased rates of lymphoma and lung cancer associated with Janus kinase (JAK) inhibitors used in the treatment of chronic inflammatory conditions have raised significant concern. A promising alternative, particularly for the treatment of plaque psoriasis, is deucravacitinib, a selective inhibitor of JAK family member TYK2. Here, the authors explored possible carcinogenic effects of TYK2 inhibition by genetic proxy based on a partial loss-of-function variant in TYK2 that provides protection against psoriasis. Analyses show that genetically proxied TYK2 inhibition increases lung cancer and non-Hodgkin lymphoma risk. The findings could impact safety assessments of deucravacitinib and future novel TYK2 inhibitors.

| INTRODUCTION
On September 1, 2021 the U.S. Food and Drug Administration (FDA) announced that three Janus kinase (JAK) inhibitors approved to treat chronic inflammatory conditions would require safety warnings over increased rates of serious heart-related events, lung cancer and lymphoma. 1 Recently, oral deucravacitinib, a selective inhibitor of tyrosine kinase 2 (TYK2, a member of the JAK family), was shown to lead to larger improvements in symptom severity for patients with moderate-to-severe plaque psoriasis than oral Apremilast, the current standard of care. 2,3 Deucravacitinib, therefore, has the potential to become an important treatment option for patients with psoriasis requiring systemic treatment and is currently under review for approval at the FDA and European Medicines Agency. 2 It is unclear, however, whether the elevated cancer risk associated with some JAK inhibitors is shared with this novel TYK2 inhibitor.
In the absence of long-term clinical trial data, naturally occurring genetic variation can be leveraged to validate therapeutic targets and predict their adverse effects. 4 Specifically, germline mutations causing partial or complete loss-of-function (LOF) of genes encoding drug targets can be employed to mimic pharmacological inhibition of these targets and have been used to correctly predict adverse effects of new medications. 5 Here, we used an established partial LOF mutation in TYK2 (rs34536443), previously shown to protect against psoriasis and other autoimmune diseases, 6,7 to evaluate the potential effect of therapeutic TYK2 inhibition on risk of lung cancer and non-Hodgkin lymphoma.

Minor allele homozygosity of rs34536443 (observed in~1 in 600
Europeans) causes near-complete (~80%) loss of TYK2 function, while heterozygotes have a more modest reduction in function (<40%), suggesting nonadditive effects of this variant. 6,8,9 To validate this variant as a surrogate for therapeutic TYK2 inhibition, we evaluated the effect of each copy of the minor allele of rs34536443, representing partial TYK2 inhibition, on risk of psoriasis, inflammatory bowel disease, Crohn's disease and multiple sclerosis. These analyses were performed using summary genetic association data on up to 78 334 cases and 150 030 controls from genome-wide association studies (GWAS) of these autoimmune diseases. [10][11][12][13] Effect estimates were generated using the Wald ratio and standard errors were approximated using the delta method. 14 To evaluate the effect of this variant on risk of overall and histological subtype-specific lung cancer we obtained summary genetic association data on up to 29 266 cases and 56 450 controls from a GWAS meta-analysis of the Integrative Analysis of Lung Cancer Risk and Aetiology (INTEGRAL). 15 Summary genetic association data for non-Hodgkin lymphoma (NHL) were generated by meta-analysing genome-wide association (GWAS) data for rs34536443 from the UK Biobank and InterLymph consortium in METAL. 16 In secondary analyses, we explored whether rs34536443  [18][19][20] All analyses were restricted to participants of European ancestry.
Further information on statistical analysis, imputation and quality control measures for these studies is provided in Appendix S1 and in the original publications.

| RESULTS
To validate rs34536443 as a surrogate for therapeutic TYK2 inhibition, we confirmed that each copy of the minor allele, representing partial TYK2 inhibition, was associated with lower risk of psoriasis (OR 0.48, 95% CI 0.39-0.58, P = 9.01 Â 10 À14 ) and other autoimmune diseases (Table 1).
In analyses of 29 266 lung cancer cases and 56 540 controls, each copy of the minor allele of rs34536443 was associated with an increased risk of lung cancer (OR 1.15, 95% CI 1.09-1.23, P = 2.29 Â 10 À6 ). This association was stronger for current and former smokers (OR 1.15, 95% CI 1.07-1.24, P = 1.72 Â 10 À4 ) compared to never smokers (OR 1.09, 95% CI 0.91-1.32, P = .34). The magnitude of effect was similar across histology types. In analyses of 8489 cases and 374 506 controls, each copy of the minor allele of rs34536443 was associated with an increased risk of non-Hodgkin lymphoma (OR 1.18, 95% CI 1.05-1.33, P = 5.25 Â 10 À3 ). In secondary analyses, rs34536443 was weakly associated with risk of Note: OR represents the exponential change in odds of cancer per each copy of the minor allele of rs34536443. Advanced prostate cancer defined as Gleason score ≥8, prostate-specific antigen >100 ng/mL, metastatic disease (M1) or death from prostate cancer.
advanced prostate cancer, but not associated with other cancers examined. Exploratory analyses examining the association of rs34536443 with smoking initiation found little evidence to support an association (OR of ever smoking regularly: 1.00, 95% CI 0.99-1.02, P = .67; Appendix S1).

| DISCUSSION
Limitations to this analysis include the restriction of the majority of our analyses to summary genetic association data which precluded comprehensive assessment of effect modification across non-lung cancer analyses. Second, the primary objective of these analyses was to evaluate safety profiles of TYK2 inhibition and, therefore, we did not include follow-up assessment of potential pathophysiological mechanisms underpinning an effect of TYK2 inhibition on cancer risk. Finally, we were unable to compare the effect of geneticallyproxied TYK2 inhibition on cancer risk to that of genetically-proxied inhibition of other JAK targets which was outside of the scope of this analysis.
The efficacy of deucravactinib in treating plaque psoriasis is attributed to the selective inhibition of TYK2, a downstream mediator of proinflammatory signalling pathways critical to psoriasis. 3 Using an established partial loss-of-function variant to mimic therapeutic TYK2 inhibition, we show that potential protection from autoimmunity mediated by TYK2 inhibition may be counteracted by an increased risk of lung cancer and non-Hodgkin lymphoma. This finding, including the restriction of an increased lung cancer risk to current and former smokers, is consistent with recent reports of higher rates of these two cancers in users of the JAK inhibitor Tofacitinib in the ORAL Surveillance safety trial. 1