Poor association between dihydropyrimidine dehydrogenase ( DPYD ) genotype and fluoropyrimidine‐induced toxicity in an Asian population

Abstract Objective Dihydropyrimidine dehydrogenase (DPYD) genotype is closely associated with fluoropyrimidine (FP)‐induced toxicities in Caucasian population and European Medicines Agency now recommends DPYD genotype‐based FP dosing strategy. Patients and Methods The current study aimed to investigate their impact on FP‐related toxicities in an Asian population using genome‐wide association study (GWAS) data set from 1364 patients with colon cancer. Results Among 82 variants registered in the Clinical Pharmacogenetics Implementation Consortium, 74 DPYD variants were directly genotyped in GWAS cohort; however, only 7 nonsynonymous DPYD variants (CPIC variants) were identified and none of the four recurrent DPYD variants (DPYD*2A, c.2846A>T, c.1679T>G, c.1236G>A) were included. Seven CPIC variants were investigated for their association with the incidence of FP‐related toxicities; however, none of these variants revealed a significant correlation with FP‐related toxicities. Conclusion These data suggested that the DPYD genotype registered in CPIC plays a minor role in FP‐related toxicities in an Asian population.


| INTRODUCTION
Fluoropyrimidine (FP) is widely used for a variety of cancers; however, some patients suffer from severe toxicities such as diarrhea, mucositis, hand-foot syndrome, and neutropenia. Dihydropyrimidine dehydrogenase (DPYD) is a key enzyme in FP degradation and approximately 8% of the Caucasian population have reduced DPYD genotype and 0.1% having no DPYD enzyme activity. 1,2 Since impaired DPYD variants delay FP degradation, patients harboring these variants have the increased risk of developing severe FP-related toxicities. A recent prospective study demonstrated that FP dose adjustment based on the four recurrent DPYD variants (DPYD*2A, c.2846A>T, c.1679T>G, c.1236G>A) could reduce the risk of FP-related toxicities in the Caucasian population. 2 A cost analysis revealed that upfront DPYD genotyping did not increase the burden of medical expenses and that expected total costs per patient were comparable regardless of whether or not they underwent universal screening. 3 Based on the aforementioned data, the European Medicines Agency (EMA) has recommended universal DPYD genotyping before initiating FP-based chemotherapy in 2020 4 and Clinical Pharmacogenetics Implementation Consortium (CPIC) has published 82 DPYD variants including 27 variants with impaired enzyme activity 5 ; however, it remains unclear whether universal DPYD screening and genotype-based FP dose adjustment are feasible for the Asian population. Therefore, the current study aimed to investigate the prevalence of DPYD variants registered in CPIC and its association with FP-related toxicities using GWAS data from 1364 patients with colon cancer who underwent adjuvant chemotherapy under clinical trials.

| Genotyping and clinical data
obtained from a genome-wide association study targeting colon cancer who underwent adjuvant chemotherapy Genotyping and clinical data obtained from a genomewide association study (GWAS) targeting the 1364 patients with colon cancer who underwent adjuvant chemotherapy under clinical trials (JOIN or ACIEVE) were utilized. Among 82 DPYD variants registered in the CPIC 5 (CPIC variants), 74 variants were directly genotyped in GWAS cohort. JOIN was a phase II trial to test the safety of 6 months of modified FOLFOX6 (mFOL-FOX6) in Japanese patients with colon cancer. 6 In contrast, ACHIEVE was a phase III trial planned as a part of phase III trials of the International Duration Evaluation of Adjuvant Chemotherapy (IDEA), 7 and tested 3 versus 6 months of mFOLFOX6 or capecitabine plus oxaliplatin (CAPOX) adjuvant chemotherapy in patients with colon cancer. 8,9 Diarrhea/mucositis/hand-foot syndrome/neutropenia were selected as representative FP-related toxicities and were graded according to the CTCAE version 3.0 (JOIN) and 4.0 (ACHIVE), respectively. Given those essential elements of diarrhea, mucositis, hand-foot syndrome, and neutropenia of CTCAE versions 3.0 were equivalent to those of version 4.0, data from both clinical Patients and Methods: The current study aimed to investigate their impact on FP-related toxicities in an Asian population using genome-wide association study (GWAS) data set from 1364 patients with colon cancer.

Results: Among 82 variants registered in the Clinical Pharmacogenetics
Implementation Consortium, 74 DPYD variants were directly genotyped in GWAS cohort; however, only 7 nonsynonymous DPYD variants (CPIC variants) were identified and none of the four recurrent DPYD variants (DPYD*2A, c.2846A>T, c.1679T>G, c.1236G>A) were included. Seven CPIC variants were investigated for their association with the incidence of FP-related toxicities; however, none of these variants revealed a significant correlation with FP-related toxicities.
Conclusion: These data suggested that the DPYD genotype registered in CPIC plays a minor role in FP-related toxicities in an Asian population.

K E Y W O R D S
Asian population, DPYD, fluoropyrimidine, genotyping, toxicity trials were comparable. In both clinical trials, clinical data were collected using an electronic data capture system with regular central monitoring. The primary outcomes of the GWAS were published elsewhere. 10

| Comprehensive search for DPYD variants using the Japanese Haplotype Reference Panel database
Nonsynonymous variants of DPYD were comprehensively searched using the registry database of the Japanese Haplotype Reference Panel (JHRP, n = 3135. https:// www.hgvd.genome.med.kyoto -u.ac.jp/).

| Statistical analysis
Associations between the presence of FP-related toxicities during adjuvant chemotherapy and genotype frequencies of nonsynonymous DPYD variants were investigated. p values and odds ratios (OR) with 95% confidence intervals (95% CI) were calculated using logistic regression analysis according to the additive genetic model. The significance level was set according to Bonferroni's correction for multiple testing with nine variants (p < 0.005). All statistical analyses were performed using R (version 3.2.0; http://www.r-proje ct.org/).

| Ethics
The protocols of GWAS study were approved by the institutional review boards of all participating institutions, and all patients provided written informed consent for the use of genomic and clinical data for research purposes.

| Prevalence of DPYD variants registered in CPIC guidelines in GWAS cohort
Among 82 DPYD variants registered in CPIC, 74 variants were directly genotyped in GWAS cohort; however, only seven nonsynonymous variants (CPIC variants) were identified and none of the four most clinically relevant DPYD variants (DPYD*2A, c.2846A>T, c.1679T>G, c.1236G>A) in the Caucasian population were included. Furthermore, these seven CPIC variants were presumed to have normal function according to the CPIC guidelines (Table 1). 42 variants not registered in CPIC guidelines, two variants were directly genotyped in GWAS cohort and were identified (non-CPIC variants). One variant (rs188052243) had been reported to have decreased enzymatic activity in vitro (Table 1). 11

| Association between seven CPIC and two non-CPIC DPYD variants and FP-related toxicities
The prevalence of the seven CPIC and two non-CPIC DPYD variants were compared between those who had grade 0 and grade ≥2 or 3 diarrhea, mucositis, hand-foot syndrome, and neutropenia; however, no significant associations were observed ( Table 2).

| DISCUSSION
Since 2020, EMA has recommended universal DPYD genotyping before initiating FP-based chemotherapy. 4 In this study, we aimed to investigate the feasibility of DPYD genotype-based FP dosing strategy in an Asian population. The four most clinically relevant DPYD variants (DPYD*2A, c.2846A>T, c.1679T>G, c.1236G>A) in the Caucasian population were not found in both GWAS cohort (n = 1364) and the JHRP database (n = 3135), which were in line with the previous study searching for DPYD variants in 341 Japanese subjects. 12 Comprehensive search for nonsynonymous DPYD variants using the JHRP database identified 42 DPYD variants which were not published in the CPIC guideline; however, most of them were rare and median minor allele frequency (MAF) was 0.0006 (range 0.0001-0.28, Table S1). Only two variants were directly genotyped and were identified in GWAS cohort. Notably, no significant association had been found between seven CPIC and two non-CPIC DPYD variants identified in the GWAS cohort and FP-related toxicities. Two patients harboring presumed impaired DPYD enzymatic activity (rs188052243) did not develop severe FPrelated toxicities. Thus, the prevalence of DPYD genotype registered in CPIC guideline is quite low and plays a minor role in FP-related toxicities in an Asian population. In contrast, incidence of FP-related toxicities in Japanese patients were comparable or higher than those in Caucasian patients (Table S2). For example, the incidence of grade 3 or higher neutropenia observed in this cohort (23.2%) was significantly higher compared to that in the IDEA trial excluding ACHIEVE cohort (8.2%) (Table S2). This suggests that the development of FP-related toxicities in the Asian population could be attributed to factors other than the DPYD variants registered in CPIC guideline.
We also evaluated the association between other hematological toxicities (leucopenia, anemia, thrombocytopenia, and febrile neutropenia) and nine nonsynonymous DPYD variants in our GWAS cohort; however, no significant associations were found (data not shown).
The current study has some limitations. First, although data regarding FP-related toxicities were collected in large prospective clinical trials, their actual relationship with FP was unclear given that patients received a combination chemotherapy of FP and oxaliplatin. Second, most of rare DPYD variants identified in Japanese subjects with presumed impaired enzymatic activity 11,13 were not directly genotyped in GWAS cohort. Therefore, we could not exclude the possibility of rare DPYD variants were involved in FP-related toxicities in the Asian population.
In summary, the current study suggested that DPYD genotypes registered in CPIC guidelines play a minor role in FP-related toxicities in the Asian population and that universal DPYD screening does not appear to be feasible.