DNMT3A and TET2 mutant Clonal Hematopoiesis May Drive a Proinflammatory State and Predict Enhanced Response to Immune Checkpoint Inhibitors

Background. DNA methylation is a key epigenetic process involved in development, aging, and cancer. Mutations in DNMT3A and TET2 in the hematopoietic stem cell compartment lead to increased self-renewal. In addition to mutations in ASXL1, collectively, these DTA mutations are recognized as an aging phenomenon, known as the most common Clonal hematopoiesis of Indeterminate Potential (CHIP) mutations and alone are not predictive of increased risk for hematopoietic malignancy. Recently, DNMT3A mutations in donor hematopoietic cells were suggested to be associated with enhanced T-cell activity in allografted patients. Additionally, role of DNMT3A mutations in creating a proinflammatory state in cardiovascular disease setting and associated elevation of T-cell markers in the myocardium have been recently explored (Sano S et al. Circ Res. 2018). Since an inflamed tumor microenvironment is associated with improved immune checkpoint inhibitors (CPI) activity, we sought to determine the impact of CHIP (a proinflammatory state) on response to CPI and CPI's effects on clonal dynamics. Additionally, while classical chemotherapy (CTX) can create selective external pressure providing survival advantage to mutant stem cells, the selective pressure of T-cell activating therapies on hematopoietic stem cells is unclear.

Methods. To study the relationship between CHIP and CPI, we used paired peripheral-blood samples taken before and after treatment with CPI therapy in patients (pts) with melanoma (MEL; n= 32) and non-small cell lung cancer (NSCLC; n=109). Serial samples (or post CPI samples) were evaluable in 5 MEL pts and 6 NSCLC pts. Error-corrected sequencing of a targeted panel of genes recurrently mutated in clonal hematopoiesis (CH) was performed on peripheral blood genomic DNA. Statistical comparisons between baseline and serial sample VAFs were performed using two-sided fisher's exact test, with a p < 0.05 considered significant.

Results. In both the MEL and NSCLC cohort, baseline samples were collected before extensive therapy exposure. 90% (29/32) of the MEL cohort had no CTX or targeted therapy prior to the baseline sample; 28% (9/32) had prior radiotherapy (RT). 10% (11/109) of the NSCLC cohort samples had prior CTX, but only 2 of these were treated for more than 1 month before sample collection. CH was frequent in these minimally pre-treated patient samples; 28.1% (9/32) and 37.6% (41/109) of the baseline MEL and NSCLC samples, respectively. As expected, DTA mutations were the most common events in these cohorts. Samples with CH were from patients of older age, but had normal hematological parameters with exception of increased RDW (p=0.022). Primary tumor responses in this cohort were defined as durable (receipt of ≥12 CPI cycles) or not durable (<12 cycles). DNMT3A^mut patients (VAF ≥1%, n=5) had more durable responses, i.e. higher median number of CPI cycles (21 cycles, range:10-40) compared to non-DNMT3A^mut pts (7 cycles, range:1-13; p= NS). Additionally, pts with larger DNMT3A^mut clones (figure 1- MEL cohort) tended to receive higher numbers of CPI cycles. In the serial sample analysis, we observed that mutations in DNMT3A and TET2 increased in size with longer CPI exposures (Figure 2, MEL cohort); pts 2, 3 and 5 received 13, 15 and 18 CPI cycles respectively, while pt 4 with the most notable clonal expansion in DNMT3A received 40 CPI cycles. All serial samples in MEL cohort showed a statistically significant change in VAF from baseline. In the serial sample analysis of NSCLC pts, we observed that those with ≥ 3 months of CPI exposure demonstrated decreases in clone size for non-DTA gene mutations such as SRCAP, STK11 and TPM1 (Table 1), but increases or stability in DNMT3A and TET2 mutations (Table 1). However, this VAF increase in DNMT3A and TET2 mutations in NSCLC cohort was not statistically significant.

Conclusions. In this small cohort of pts with MEL and NSCLC, the presence of DNMT3A/TET2 CH was associated with longer checkpoint inhibitor exposure and increased allelic frequency over time. These findings need further validation in larger cohorts and delineation of the relationship between DTA mutations such as DNMT3A and enhanced immune activity.

Acknowledgement: Data and samples for this study were provided by the Data Bank and BioRepository (DBBR), which is funded by the National Cancer Institute (P30 CA016056) and is a Roswell Park Cancer Institute Cancer Center Support Grant shared resource.

Figure 1

View largeDownload slide

Figure 1

View largeDownload slide

Close modal