Longitudinal genomic profiling of chemotherapy-related CHIP variants in patients with ovarian cancer

Abstract

Introduction: Clonal hematopoiesis (CH) is characterized by the presence of hematopoietic stem cells (HSCs) with the potential of clonally expanding and giving rise to hematological malignancies. Clonal hematopoiesis of indeterminate potential (CHIP) is the outgrowth of a single HSC clone with an acquired somatic mutation in the absence of hematological abnormalities. CHIP variants occur with a variant allele frequency (VAF) of at least 2% in peripheral blood. This definition does not account for less frequent mutations that give rise to hematopoietic clones. Previous studies indicate an association between CH and secondary hematologic malignancies in cancer patients who receive chemotherapy.

Methods: To discover novel candidate CHIP mutations, including those with extremely low VAFs, we performed an in-depth characterization of low-frequency CHIP variants in a highly selected group of patients with high-grade serous ovarian cancer (HGSC) before and after neoadjuvant chemotherapy (NACT). We performed comprehensive ultra-high-depth whole-exome sequencing of circulating free DNA (cfDNA) and matched white blood cell (WBC) DNA from pre- (n=9) and post-NACT (n=9) samples from HGSC patients who had excellent response (ER; n=4) or poor response (PR; n=5) to NACT.

Results: Variants present in both the WBC DNA and cfDNA from a patient were considered candidate CHIP variants. We identified 93,088 candidate CHIP variants in 13,780 genes. Compared with pre-NACT samples, post-NACT samples tended to have fewer CHIP mutations with VAFs of less than 5%, which may reflect the negative selective pressure of chemotherapy on rare hematopoietic clones. Finally, we identified CHIP variants in tumor samples matched to the liquid biopsies.

Discussion: Our innovative sequencing approach enabled the discovery of a large number of novel low-frequency candidate CHIP mutations, whose frequency and composition are affected by chemotherapy, in the cfDNA of patients with HGSC. The CHIP variants that were enriched after chemotherapy, if validated, might become essential predictive markers for therapy-related myeloid neoplasia.

Keywords: biomarkers; clonal hematopoiesis of indeterminate potential (CHIP); oncology; ovarian cancer; t-MDS/AML.

Figure 1

Graphical representation of the study design.

Figure 2

(A) Numbers of CHIP variants in each pre-chemotherapy sample and each post-chemotherapy sample. (B) Numbers of CHIP variants in all pre-chemotherapy samples versus all post-chemotherapy samples. (C-F) Numbers of CHIP variants with allele frequencies of 0.5-1% (C), 1-2% (D), 2-5% (E), and more than 5% (F) in pre-chemotherapy samples versus post-chemotherapy samples. (G) Heatmap showing the distribution of CHIP variants in POST nast-NACT and in pre-NACT. pre, pre-chemotherapy; post, post-chemotherapy; var, variant; AF, allele frequency; ER, excellent response; PR, poor response; cf, circulating free DNA.

Figure 3

(A) Heatmap showing the distribution of the top 50 candidate CHIP mutations whose VAFs were higher in post-NACT samples than in PR pre-NACT samples in the PR group. (B) Heatmap showing the distribution of the top 50 candidate CHIP mutations whose VAFS were higher in post-NACT samples than in pre-NACT samples in the ER group. (C) Heatmap showing the distribution of the top 50 candidate CHIP mutations whose VAFs in pre-NACT samples were higher in the PR group than in the ER group. Asteriscs indicate stop codons.

Figure 4

(A) Numbers of CHIP variants in cfDNA and tumor DNA in each pre-chemotherapy sample and each post-chemotherapy sample. (B) Heatmap showing the distribution of the top 50 candidate CHIP mutations whose VAFs were higher in pre-NACT samples than in post-NACT tumor samples.