Association of high-dose radioactive iodine therapy with PPM1D-mutated clonal hematopoiesis in older individuals

Abstract

While radioactive iodine therapy (RAIT) has been an effective treatment for thyroid cancer, its link to clonal hematopoiesis (CH) has been yet underexplored. In this study, error-corrected sequencing (median depth: 1926×) of 93 CH-related genes was performed from the blood samples of 358 thyroid cancer patients, including 110 controls (no RAIT) and 248 RAIT recipients. RAIT recipients were stratified into low- and high-dose groups using a 7.4 GBq cutoff. Multivariable logistic regression revealed that the high-dose group had a higher CH prevalence with variant allele frequency (VAF) higher than 2% compared to controls, especially in patients aged ≥50 (OR = 2.44, CI = 1.04-6.00, P = 0.04). Thirteen genes had mutations with VAF >2%, with DNMT3A, TET2, and PPM1D being the most common. Notably, only the PPM1D mutations were significantly linked to RAIT, occurring more frequently in the high-dose group (13%) compared to the low-dose group (5%) or controls (2%) at a VAF cutoff of 0.5%. In silico analyses indicated that truncating PPM1D mutations confer a selective advantage under high-dose RAIT and with older age. Although the prognostic implications of PPM1D-mutated CH remain to be further elucidated, these findings offer valuable insights for optimizing RAIT dosing in thyroid cancer patients.

Keywords: PPM1D mutations; error‐corrected next‐generation sequencing; radioactive iodine therapy; therapy‐related clonal hematopoiesis; thyroid cancer.

Fig. 1

Overview of the study subjects and experimental design. A comprehensive overview of the study. A total of 382 blood samples were collected from 358 patients with thyroid cancer. Among these patients, 110 had no prior history of radioactive iodine therapy (RAIT), while 248 had undergone RAIT, including 24 individuals whose longitudinal samples were collected before and after receiving RAIT. We extracted DNA from the peripheral blood of the patients and conducted error‐corrected next‐generation sequencing (ecNGS) targeting the entire exon regions of 93‐CH‐related genes. We detected mutations with a variant allele frequency (VAF) as low as 0.5%. Consequently, this study aimed to reveal the relationship between the RAIT and CH.

Fig. 2

Associations between age, radioactive iodine therapy dose, and clonal hematopoiesis. (A) The prevalence of clonal hematopoiesis (CH), defined by a variant allele frequency (VAF) exceeding 2%, was notably greater in older patients (age at CH test; Age_CH ≥ 50) across all three groups: control (n = 110), low‐dose (n = 115), high‐dose (N = 133). (B) Distribution of the maximum VAF in each individual according to Age_CH and radioactive iodine therapy (RAIT) status/dose group. Mutations with high‐VAF levels were more prevalent at older ages (Age_CH ≥ 50 years), particularly within the high‐dose RAIT group. (C) In multivariable logistic regression analyses for the presence of CH with varying VAF cutoffs adjusting for Age_CH, sex, and smoking status, the high‐dose RAIT group displayed a significant association compared to the control group for 2% and 5% VAF cutoffs, but only in older patients (Age_CH ≥ 50 years). Error bars indicate confidence interval. (D) A plot of the marginal effect model with an interaction term between Age_CH and RAIT status/dose for maximum VAF. Although there was no significant difference between the RAIT‐received groups and the control group (colored lines and text), a significant interaction emerged when a continuous RAIT dose was included as an independent variable (black text) (Table S3). The maximum VAF is defined as the VAF of the mutation with the highest frequency in an individual. In regression analyses, the VAF was log‐transformed. Firth logistic regression was used to test for the presence of CH because of the sparse data. The units of the Age_CH and RAIT doses are per 10 years and per 1.11 GBq increase, respectively. *P value <0.05.

Fig. 3

Association of high‐dose radioactive iodine therapy with PPM1D‐mutated clonal hematopoiesis. (A) Analysis of mutations in the top 13 genes, chosen for having variant allele frequency (VAF) exceeding 2% in at least one sample, concerning age at clonal hematopoiesis test (Age_CH) and radioactive iodine therapy (RAIT) status/dose. High‐VAF mutations were notably more prevalent in older patients, and certain genes (PPM1D, TP53, ASXL1, RAD21) showed enrichment specifically in the high‐dose RAIT group. Sample sizes: Control (n = 110), low‐dose RAIT (n = 115), and high‐dose RAIT (n = 133). (B) Lollipop plots highlighting mutation positions and frequencies in PPM1D with the VAF of corresponding positions and their frequency densities (gray background). High‐VAF truncating mutations in PPM1D clustered near the C terminus, setting them apart from mutations in other genes. (C) Multivariable logistic regression analysis of the associations between the presence of CH mutations in the top 13 genes and multiple clinical variables. Analyses were conducted using mutation presence (VAF >0.5%) in each gene as the dependent variable. The independent variables included Age_CH (adjusted for sex, smoking, prior RAIT), age at first RAIT (Age_RAIT; adjusted for sex, smoking, RAIT dose, elapsed time from first RAIT), prior RAIT (adjusted for Age_CH, sex, smoking), and RAIT dose (adjusted for Age_CH, sex, smoking, elapsed time from first RAIT). Nonsignificant results (P value >0.05) are presented as gray dots. Error bars indicate confidence interval.

Fig. 4

Association between functional alterations in clonal hematopoiesis mutations and positive clonal selection. (A) Clonal hematopoiesis (CH) mutations were categorized according to the tertiles of CADD scores, and the proportion of clone sizes, determined by variant allele frequency (VAF) in each tertile, was analyzed. A trend was observed indicating a greater VAF with higher CADD scores. (B) Analysis of the nonsynonymous to synonymous substitution (dN/dS) ratio of each gene in all individuals and within each radioactive iodine therapy (RAIT) status/dose group revealed positive selection (dN/dS >1) for truncating PPM1D and missense TP53 mutations, particularly in the high‐dose RAIT group. The vertical and horizontal gray dashed lines indicate a dN/dS ratio of 1.

Fig. 5

Longitudinal assessment of clonal hematopoiesis before and after the radioactive iodine therapy. (A) The figure presents the number of CH mutations that changed the variant allele frequency (VAF) before and after radioactive iodine therapy (RAIT) for 24 patients with paired blood samples. Mutations undetectable after RAIT or those with decreased VAF were classified as ‘vanished/decreased’, while new mutations or those with increased VAF were classified as ‘emerged/increased’. (B) VAF changes for paired samples were examined based on the time between venipunctures. The figure exclusively displays existing mutations with elevated VAF or newly acquired mutations after RAIT. All these samples were obtained from individuals who received low‐dose RAIT (<7.4 GBq). The gray vertical lines connect all mutations detected in a sample, and the gray dashed lines connect mutations detected both before and after RAIT. Gene names are selectively labeled for mutations both detected before and after RAIT or for mutations with a maximum VAF after RAIT.