Clonal hematopoiesis and differential outcomes after immune checkpoint blockade

Positive selection of hematopoietic stem cells harboring leukemogenic mutations in individuals without obvert hematologic malignancies, termed clonal hematopoiesis (CH), occurs at an increased frequency among older individuals without obvert signs of hematologic disease (Jaiswal and Libby, 2020Jaiswal S. Libby P. Clonal haematopoiesis: connecting ageing and inflammation in cardiovascular disease.Nat. Rev. Cardiol. 2020; 17: 137-144https://doi.org/10.1038/s41569-019-0247-5Crossref PubMed Scopus (166) Google Scholar). While CH is associated with a slight increased risk of hematologic cancers over time, CH has emerged as a significant risk factor for cardiovascular events via skewing innate immune cells toward a pro-inflammatory state (Jaiswal and Libby, 2020Jaiswal S. Libby P. Clonal haematopoiesis: connecting ageing and inflammation in cardiovascular disease.Nat. Rev. Cardiol. 2020; 17: 137-144https://doi.org/10.1038/s41569-019-0247-5Crossref PubMed Scopus (166) Google Scholar). Animal models of CH demonstrate that CH results in the production of altered myeloid cells with enhanced inflammatory responses and subsequent increased inflammasome activation (Jaiswal and Libby, 2020Jaiswal S. Libby P. Clonal haematopoiesis: connecting ageing and inflammation in cardiovascular disease.Nat. Rev. Cardiol. 2020; 17: 137-144https://doi.org/10.1038/s41569-019-0247-5Crossref PubMed Scopus (166) Google Scholar). The prevalence of CH is age dependent, with individuals in the fourth to fifth decades of life having a 1%–5% prevalence and individuals in the sixth to eighth decades of life having a 10%–15% prevalence (Genovese et al., 2014Genovese G. Kahler A.K. Handsaker R.E. Lindberg J. Rose S.A. Bakhoum S.F. Chambert K. Mick E. Neale B.M. Fromer M. et al.Clonal hematopoiesis and blood-cancer risk inferred from blood DNA sequence.N. Engl. J. Med. 2014; 371: 2477-2487https://doi.org/10.1056/nejmoa1409405Crossref PubMed Scopus (0) Google Scholar; Jaiswal et al., 2014Jaiswal S. Fontanillas P. Flannick J. Manning A. Grauman P.V. Mar B.G. Lindsley R.C. Mermel C.H. Burtt N. Chavez A. et al.Age-related clonal hematopoiesis associated with adverse outcomes.N. Engl. J. Med. 2014; 371: 2488-2498https://doi.org/10.1056/nejmoa1408617Crossref PubMed Scopus (0) Google Scholar). Given that most cancers are diagnosed between the sixth and seventh decades of life, a substantial proportion of patients with cancers harbor CH. Immune checkpoint inhibitors (ICIs) that target inhibitory immune cell checkpoints such as PD-1 are now used across many cancer types for multiple indications, albeit with limited benefit in most circumstances. Whether CH may impact immune modulating cancer therapies is unknown, but the altered immunity associated with CH could foreseeably enhance or limit the benefit of ICIs. For instance, inflammatory properties of CH may enhance innate immune cell activation, which has been implicated in the efficacy of ICIs, or, oppositely, induce immunosuppressive tumor microenvironments or produce dysfunctional lymphocytes rendering ICIs ineffective (Long et al., 2022Long H. Jia Q. Wang L. Fang W. Wang Z. Jiang T. Zhou F. Jin Z. Huang J. Zhou L. et al.Tumor-induced erythroid precursor-differentiated myeloid cells mediate immunosuppression and curtail anti-PD-1/PD-L1 treatment efficacy.Cancer Cell. 2022; 40: 674-693.e7https://doi.org/10.1016/j.ccell.2022.04.018Abstract Full Text Full Text PDF PubMed Scopus (31) Google Scholar; Veatch et al., 2022Veatch J.R. Lee S.M. Shasha C. Singhi N. Szeto J.L. Moshiri A.S. Kim T.S. Smythe K. Kong P. Fitzgibbon M. et al.Neoantigen-specific CD4(+) T cells in human melanoma have diverse differentiation states and correlate with CD8(+) T cell, macrophage, and B cell function.Cancer Cell. 2022; 40: 393-409.e9https://doi.org/10.1016/j.ccell.2022.03.006Abstract Full Text Full Text PDF PubMed Scopus (35) Google Scholar). By examining a large cohort of patients with solid cancers treated with ICIs who also had peripheral blood sequencing to evaluate the presence of CH, we sought to determine whether CH may affect the outcomes of patients treated with ICIs. Patients at a single institution with bladder, breast, colorectal, esophagogastric, glioma, head and neck, melanoma, non-small cell lung, renal cell, and unknown primary cancers who underwent peripheral blood-cell-targeted next-generation sequencing using the MSK-IMPACT assay in a prospective IRB-approved institutional research protocol (NCT01775072) were identified as previously described (Bolton et al., 2020Bolton K.L. Ptashkin R.N. Gao T. Braunstein L. Devlin S.M. Kelly D. Patel M. Berthon A. Syed A. Yabe M. et al.Cancer therapy shapes the fitness landscape of clonal hematopoiesis.Nat. Genet. 2020; 52: 1219-1226https://doi.org/10.1038/s41588-020-00710-0Crossref PubMed Scopus (252) Google Scholar). CH was diagnosed if a variant in leukemia-associated genes was detected in blood cells with a minimum variant allele fraction of 2% (Bolton et al., 2020Bolton K.L. Ptashkin R.N. Gao T. Braunstein L. Devlin S.M. Kelly D. Patel M. Berthon A. Syed A. Yabe M. et al.Cancer therapy shapes the fitness landscape of clonal hematopoiesis.Nat. Genet. 2020; 52: 1219-1226https://doi.org/10.1038/s41588-020-00710-0Crossref PubMed Scopus (252) Google Scholar; Mencia-Trinchant et al., 2020Mencia-Trinchant N. MacKay M.J. Chin C. Afshinnekoo E. Foox J. Meydan C. Butler D. Mozsary C. Vernice N.A. Darby C. et al.Clonal hematopoiesis before, during, and after human spaceflight.Cell Rep. 2020; 33108458https://doi.org/10.1016/j.celrep.2020.108458Abstract Full Text Full Text PDF PubMed Scopus (18) Google Scholar). The predictive value of CH was assessed in an ICI-treated cohort (MSK-ICI) comprised of 1,677 patients who were treated with anti-PD-1/L1 antibodies as any line of therapy where overall survival (OS) was calculated from the time of ICI initiation (Samstein et al., 2019Samstein R.M. Lee C.H. Shoushtari A.N. Hellmann M.D. Shen R. Janjigian Y.Y. Barron D.A. Zehir A. Jordan E.J. Omuro A. et al.Tumor mutational load predicts survival after immunotherapy across multiple cancer types.Nat. Genet. 2019; 51: 202-206https://doi.org/10.1038/s41588-018-0312-8Crossref PubMed Scopus (2129) Google Scholar). The prognostic value of CH was assessed in a treatment-agnostic cohort (MSK-prognostic) comprised of 13,729 patients (including those in the ICI cohort) treated with any systemic therapy where OS was calculated from the time of cancer diagnosis. The prevalence of CH in the MSK-ICI and MSK-prognostic cohorts was 20.5% and 19.8%, respectively. Multivariable Cox regression analysis showed that CH was associated with worse OS in the MSK-ICI cohort (HR = 1.28; 95% CI, 1.07–1.53), but not the MSK-prognostic cohort (HR = 1.06; 95% CI, 0.97–1.17) after adjusting for age, sex, race, smoking status, time from cancer diagnosis to CH testing, cancer types, and exposure to different classes of chemotherapy (Table S1). To determine whether CH was associated with OS across individual cancer types, we assessed the prevalence and OS association of CH within each histology. The prevalence of CH was highest in bladder and non-small cell lung cancers and highly correlated between the MSK-ICI and MSK-prognostic cohorts (Spearman r = 0.90, p < 0.001) (Figure S1A and B). In the MSK-ICI cohort, CH in nearly all cancer types was associated with numerically worse survival except in colorectal cancer (HR = 0.24; 95% CI, 0.08–0.69), where there was a significant decreased risk of death (Figure S1A). In the MSK-prognostic cohort, there was no consistent association between CH and survival across cancer types (Figure S1B). Among patients with colorectal cancers in the MSK-ICI cohort, 20.5% had mismatch repair deficiency (MMRd), including two patients with CH. However, even after excluding all patients with MMRd, OS among patients with CH was significantly prolonged (median OS 51 versus 14 months, log rank p value = 0.02) (Figure S1C). Nearly all patients with colorectal cancer and CH in the MSK-ICI cohort had DNMT3A mutations (52.6%) or TET2 mutations (26.3%), which was not statistically different from the proportions observed in patients with colorectal cancer and CH in the MSK-prognostic cohort. A subset of patients had clinical imaging accessible for radiographic measurements, which allowed us to determine if CH was also associated with anti-tumor responses. However, we note that the small sample size of this patient cohort limits statistical power to discern significant associations. Among MMR-proficient (MMRp) colorectal cancers, 25% of patients with CH (n = 2) had an objective response (complete or partial response according to RECIST), while 7% of patients without CH (n = 4) had an objective response (Figure S1D). Among MMRd colorectal cancers, 100% of patients with CH (n = 2) had an objective response, while 52% of patients without CH (n = 13) had an objective response (Figure S1D). Our analysis of a large cohort of ICI-treated patients indicates that CH is a negative predictive factor of ICI outcomes across multiple cancer types, with the exception of colorectal cancer. While colorectal cancers with MMRd have significantly improved outcomes after ICIs, there was no clear association between CH and MMRd in our cohort, and CH was still associated with a prolonged OS among MMRp colorectal cancers. This finding is notable, as colorectal cancers that are MMRp typically have poor responses and outcomes to ICIs. Thus, CH may define a subpopulation of patients with colorectal cancer where ICIs may be considered for treatment. The underlying mechanism of the histology-specific impact of CH on ICI outcomes remains to be elucidated, but the inflammatory properties of CH may preferentially enhance innate immune cell activation in ICI-resistant cancers such as colorectal cancers. CH may also impact T cell function, as loss of inactivation of DNMT3A and TET2 has been reported to enhance anti-tumor activity by lymphocytes (Lee et al., 2021Lee M. Li J. Li J. Fang S. Zhang J. Vo A.T.T. Han W. Zeng H. Isgandarova S. Martinez-Moczygemba M. et al.Tet2 Inactivation enhances the antitumor activity of tumor-infiltrating lymphocytes.Cancer Res. 2021; 81: 1965-1976https://doi.org/10.1158/0008-5472.can-20-3213Crossref PubMed Scopus (0) Google Scholar; Prinzing et al., 2021Prinzing B. Zebley C.C. Petersen C.T. Fan Y. Anido A.A. Yi Z. Nguyen P. Houke H. Bell M. Haydar D. et al.Deleting DNMT3A in CAR T cells prevents exhaustion and enhances antitumor activity.Sci. Transl. Med. 2021; 13eabh0272https://doi.org/10.1126/scitranslmed.abh0272Crossref PubMed Scopus (78) Google Scholar). Oppositely, CH may produce dysfunctional lymphocytes with limited ICI sensitivity in other contexts. Dissecting how CH differentially affects ICI outcomes may reveal new insights into the intersection between age-related immune function, cancer immunity, and ICI biology. Our results also highlight CH as a putative predictive biomarker for ICIs, which may broaden the utility of clinical testing for CH. However, whether biological features of CH such as clone size are associated with ICI outcomes remains to be investigated in larger studies, given the relatively small sample size of patients with colorectal cancer and CH in the MSK-ICI cohort. Both patients with CH and colorectal cancers in the MSK-ICI cohort that had partial responses harbored DNMT3A mutations, but given that DNMT3A was the most common CH gene mutated in the MSK-ICI cohort, it remains to be determined whether specific CH genes are associated with improved ICI outcomes. Limitations of this analysis include the relatively small sample of colorectal cancers in the MSK-ICI cohort and the lack of inclusion of other cancer types. Whether CH may govern outcomes to ICIs that target other inhibitory checkpoints or other cancer treatments such as targeted therapies remains to be examined. Further investigations are also warranted to identify clinical and biological factors that govern the association between CH and ICI outcomes. D.H. is supported in part by the Cancer Prevention and Research Institute of Texas and Josephine Sterling Foundation. L.G.T.M. is supported by grant R01 DE027738 from the National Institutes of Health, the Sebastian Nativo Fund, the Jayme and Peter Flowers Fund. The funding agencies had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication. L.G.T.M. reports intellectual property owned by Memorial Sloan Kettering and, on which he is listed as an inventor, licensed to PGDx, unrelated to this work. No other disclosures are reported. Download .pdf (.18 MB) Help with pdf files Document S1. Table S1 and Figure S1