Smoking and chronic kidney disease are associated with expansion of TET2 clonal hematopoiesis of indeterminate potential (CHIP)

Abstract

Abstract Background CHIP is linked to blood cancer risk and adverse non-haematological health outcomes, with variant allele frequency (VAF) being an important determinant of risk. Longitudinal studies with serial CHIP measurements offer an opportunity to identify modifiable risk factors for CHIP and determine the clinical utility of monitoring CHIP growth. Studies to date have been limited by small numbers or a long interval between sequencing, resulting in limited power and attrition bias, and conflicting results. Utilising a prospective, extensively phenotyped cohort of older adults who were participants of the Aspirin in Preventing Events in the Elderly (ASPREE) trial, we aimed to 1) identify risk factors associated with clone growth and 2) determine if clone growth is related to cancer and/or mortality outcomes independent of clone size. Methods ASPREE enrolled community-dwelling Australians ≥70 years without cardiovascular disease or 5-year life-limiting illness. Sequencing was performed at baseline and after 3 years using an amplicon-based panel targeting 19 of the most common CHIP genes. Participants were followed for a median of 8.4 years and outcomes (haematologic malignancy, cancer, all-cause mortality and cancer mortality) were independently adjudicated. To determine risk factors and consequences of clone growth, we focused on variants that were quantifiable (VAF ≥1%) at both timepoints. Annual growth rate (AGR) was calculating assuming constant exponential growth between over the 3 years between sequencing (AGR = ((VAFY3/VAFBaseline)1/3-1)*100). Linear mixed-effects models were used to examine risk factors for change in VAF, including age, sex, >1 CHIP variant, smoking, alcohol, body mass index, hypertension, diabetes, history of cancer, chronic kidney disease (CKD; defined as eGFR≤60ml/min/1.73m2 or urine albumin-creatinine ratio >3mg/mmol), aspirin, haemoglobin, total cholesterol, high density lipoprotein, and statin use. We also explored whether clone growth, measured by fitness score of the CHIP clone [FS = log2(VAFY3/VAFBaseline)], was associated with clinical outcomes after adjusting for gene and year 3 VAF in landmark Cox proportional hazards regression models. Results There were 7246 participants sequenced at both timepoints. At year 3, 2073 CHIP (VAF ≥2%) variants were detected in 1745/7246 (24.1%) participants. The median VAF was 4.4% and 78% were in DNMT3A, TET2 or ASXL1. Consistent with other studies, we observed that growth rates differed by gene, but with considerable inter-individual variation in growth rate, even among variants in the same gene. The highest average growth rates were observed for KRAS, SF3B1 and JAK2 with median AGR of 25%, 13% and 12% respectively. By contrast, the median AGR was 3% for DNMT3A clones and 5% for TP53. Extrinsic risk factors for growth: Overall, smoking (p=0.04), CKD (p=0.03), and lower total cholesterol (p=0.01) at baseline were associated with greater change in VAF of CHIP clones. Aspirin and other candidate factors had no effect on growth. In analyses restricted by gene, smoking and CKD were associated with growth of TET2 clones, but not DNMT3A or ASXL1. Association between growth and disease outcomes: In those with CHIP at year 3, and after adjusting for gene and size of clone (VAF), fitness score was independently associated with risk of subsequent haematological malignancy (HR 1.44 [95% CI 1.02 to 2.04]). By contrast, fitness score was not associated with all-cause mortality, cancer-related mortality or incident cancer. Conclusion This is the largest cohort to date with serial sequencing to examine risk factors and consequences of measured growth of CHIP clones. We found that smoking and CKD were associated with faster growth of CHIP clones, specifically TET2 clones. Further work is required to verify these findings in different cohorts, and determine if smoking cessation or other modification of the pro-inflammatory state associated with smoking or CKD could potentially prevent CHIP growth and adverse outcomes. Our findings that growth of CHIP was independently associated with risk of blood cancer suggest that monitoring of CHIP could be considered in this context.