Offer, Steven M.

Clinical studies have identified specific genetic variants in dihydropyrimidine dehydrogenase (DPD; DPYD gene) as predictors of severe adverse toxicity to the commonly used chemotherapeutic 5‐fluorouracil (5‐FU); however, these studies have focused on European and European‐American populations. Our laboratory recently demonstrated that additional variants in non‐European haplotypes are predictive of 5‐FU toxicity. The objective of this study was to identify potential risk variants in an understudied East African population relevant to our institution's catchment area. The DPYD protein‐coding region was sequenced in 588 individuals of Somali or Kenyan ancestry living in central/southeast Minnesota. Twelve novel nonsynonymous variants were identified, seven of which significantly decreased DPD activity in vitro. The commonly reported toxicity‐associated variants, *2A, D949V, and I560S, were not detected in any individuals. Overall, this study demonstrates a critical limitation in our knowledge of pharmacogenetic predictors of 5‐FU toxicity, which has been based on clinical studies conducted in populations of limited diversity.

Abstract The antimetabolite 5-fluorouracil (5-FU) is one of the most widely used chemotherapy drugs. Dihydropyrimidine dehydrogenase (DPD) is a major determinant of 5-FU response and toxicity. Although DPYD variants may affect 5-FU metabolism, they do not completely explain the reported variability in DPD function or the resultant differences in treatment response. Here, we report that H3K27 trimethylation (H3K27me3) at the DPYD promoter regulated by Ezh2 and UTX suppresses DPYD expression by inhibiting transcription factor PU.1 binding, leading to increased resistance to 5-FU. Enrichment of H3K27me3 at the DPYD promoter was negatively correlated with both DPYD expression and DPD enzyme activity in peripheral blood specimens from healthy volunteers. Lastly, tumor expression data suggest that DPYD repression by Ezh2 predicts poor survival in 5-FU–treated cancers. Collectively, the findings of the present article suggest that a previously uncharacterized mechanism regulates DPD expression and may contribute to tumor resistance to 5-FU. Cancer Res; 76(21); 6362–73. ©2016 AACR.

Abstract Nearly one-third of cancer patients that receive the commonly prescribed chemotherapy drug 5-fluorouracil (5-FU) experience severe (grade 3+) adverse toxicity related to treatment. Approximately 1,300 deaths are directly attributed to 5-FU toxicity annually in the U.S. While three deleterious genetic variations in the dihydropyrimidine dehydrogenase (DPD) gene (DPYD) are well-established predictors of severe toxicity (i.e., *2A, p.I560S, p.D949V), they only explain 30-35% of adverse cases of 5-FU toxicity. Recently, an intronic variant (rs75017182) has been suggested to contribute to 5-FU toxicity risk by promoting the alternative splicing of DPYD. However, recent clinical studies are conflicting as to the degree to which rs75017182 affects DPD enzyme function, and it remains unclear as to the extent to which DPYD splicing is affected in carriers. The primary aim of this study was to directly determine correlations between rs75017182, DPD enzyme function, and DPYD splicing. Genotyping for rs75017182, *2A, p.I560S, and p.D949V was performed on DNA obtained from 3950 healthy participants in the Mayo Clinic Biobank. Individuals carrying one or more of the genotyped variants, as well as matched non-carriers, were asked to provide an additional blood sample that was used to measure DPYD expression, DPYD splicing relevant to rs75017182, and DPD enzyme activity (n = 204). A moderate, but significant, reduction was noted in the expression of canonically spliced DPYD in rs75017182 carriers compared to non-carriers. DPD enzyme activity in peripheral blood mononuclear cells was similarly reduced in rs75017182 carriers. The degree to which DPD activity was reduced in rs75017182 carriers was not as severe as was noted for carriers of the toxicity-associated variants *2A, p.I560S, or p.D949V. Collectively, our data support a model in which rs75017182 leads to alternative splicing of a fraction of primary transcripts that harbor the variant, which in turn causes a modest reduction in the levels of functional DPD enzyme. The modest phenotypes noted may partially explain some of the discrepant results noted between various clinical studies that have attempted to evaluate the variant. Citation Format: Shikshya Shrestha, Steven M. Offer, Robert B. Diasio. The intronic DPYD variant rs75017182 correlates with a modest reduction in DPD enzyme activity. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr LB-195.

Abstract Dihydropyrimidine dehydrogenase (DPD) is a major determinant of the efficacy and toxicity of 5-fluorouracil in various cancer therapies. Single nucleotide polymorphisms (SNPs) within DPYD have been studied extensively for years. However, known SNPs do not explain most cases of altered DPD activity and response to 5-FU. Furthermore, variations of DPYD expression in cancer patients have been reported; however, the underlying molecular mechanism is unclear. This suggests that regulation of DPYD expression may be an additional mechanism to control DPD activity. In this study, we focused on epigenetic regulation of DPD and specifically investigated the role of histone methylation on DPYD expression. Inhibition of the H3K27 methyl-transferase Ezh2 by either GSK-126 or a dominant-negative histone H3 mutant significantly increased DPYD expression in various cell lines. The expression of thymidylate synthetase, a major target of 5-fluorouracil toxicity, was not altered. Consistent with elevated DPYD expression, cells that lost H3K27me3 were more resistant to 5-fluorouracil. Furthermore, we demonstrated that Ezh2 and H3K27me3 were enriched on the promoter and the first exon of DPYD in low-DPD expressing cells, but not in cell lines with high DPD levels. Additionally, the enrichment of H3K27me3 on the DPYD promoter was negatively correlated with cellular DPD activity in peripheral mononuclear cells from healthy volunteers. Finally, CRISPR/Cas9-mediated targeting of Ezh2 to the DPYD promoter artificially increased H3K27me3 and suppressed DPYD expression. Collectively, these data suggest that histone methylation can regulate DPYD expression, providing a previously unrecognized mechanism of 5-fluorouracil sensitivity. The tri-methylation of H3K27 is a potential biomarker for guiding individualized 5-FU therapy. Citation Format: Rentian Wu, Steven Offer, Calvin Jerde, Garrett Dunlap, Robert Diasio. Histone H3K27 tri-methylation regulates DPYD expression and cellular sensitivity of 5-fluorouracil. [abstract]. In: Proceedings of the AACR Special Conference on Chromatin and Epigenetics in Cancer; Sep 24-27, 2015; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2016;76(2 Suppl):Abstract nr B49.

Abstract 5-Fluorouracil (5-FU) is a cornerstone drug for the treatment of several aggressive tumors; however, adverse 5-FU toxicity contributes to approximately 1,300 deaths annually in the U.S. While three genetic variations in dihydropyrimidine dehydrogenase (DPD; DPYD gene) are well-established predictors of toxicity (i.e., *2A, I560S, D949V), the contributions of additional reported missense DPYD variants (>130) have not been adequately evaluated in clinical studies. Our lab has previously developed methods to study DPYD variants for potential clinical relevance to 5-FU toxicity using in vitro and cellular approaches (see Offer et al. Cancer Research 2013 & 2014). In the current study, we measured the in vitro enzyme activity of 47 uncharacterized missense variants that have been previously reported in dbSNP and NHLBI ESP databases. Additionally, we assessed the impact of these variants on 5-FU sensitivity in multiple transgenic human cellular models using real-time cellular analysis. Approximately one-third of the studied variants showed significantly reduced enzyme activity compared to wild type. Of the DPD-deficient variants identified in this study, 8 were profoundly deficient (<12.5% of “wildtype” activity), with enzyme activities similar to *2A. Two deficient variants showed modest reductions in enzyme activity (12.5-25% activity), similar to I560S. The remaining 5 deleterious variants exhibited mild deficiency (>25% activity, but significantly lower than “wildtype”), similar to D949V. Cellular sensitivity to 5-FU for cells expressing these variants paralleled the in vitro enzyme activity results, with deleterious variants showing reduced tolerance to 5-FU relative to cells transgenic for wildtype DPD. These trends were observed in multiple cell lines derived from various tissues and tumor types. Overall, our results add to the growing database of deleterious DPYD variations. Many of these variants have been observed primarily in non-Caucasian racial groups, consistent with our previous reports. These data suggest that the deleterious variants classified in this study may be important predictors of 5-FU toxicity, particularly in individuals of non-Caucasian ancestry. Additionally, these data suggest that genetic testing for rare variants may improve the success of predictive tests for 5-FU toxicity. Finally, this study suggests that in vitro and cellular models may be important pharmacogenomic methods to rapidly characterize variants of potential relevance in additional drug-gene pathways. Citation Format: Shikshya Shrestha, Steve Offer, Robert B. Diasio. Functional analysis of rare dihydropyrimidine dehydrogenase variants relevant to 5-fluorouracil toxicity. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 5487. doi:10.1158/1538-7445.AM2015-5487

Abstract Purpose: The microRNA miR-27a was recently shown to directly regulate dihydropyrimidine dehydrogenase (DPD), the key enzyme in fluoropyrimidine catabolism. A common polymorphism (rs895819A>G) in the miR-27a genomic region (MIR27A) was associated with reduced DPD activity in healthy volunteers, but the clinical relevance of this effect is still unknown. Here, we assessed the association of MIR27A germline variants with early-onset fluoropyrimidine toxicity. Experimental Design: MIR27A was sequenced in 514 patients with cancer receiving fluoropyrimidine-based chemotherapy. Associations of MIR27A polymorphisms with early-onset (cycles 1–2) fluoropyrimidine toxicity were assessed in the context of known risk variants in the DPD gene (DPYD) and additional covariates associated with toxicity. Results: The association of rs895819A>G with early-onset fluoropyrimidine toxicity was strongly dependent on DPYD risk variant carrier status (Pinteraction = 0.0025). In patients carrying DPYD risk variants, rs895819G was associated with a strongly increased toxicity risk [OR, 7.6; 95% confidence interval (CI), 1.7–34.7; P = 0.0085]. Overall, 71% (12/17) of patients who carried both rs895819G and a DPYD risk variant experienced severe toxicity. In patients without DPYD risk variants, rs895819G was associated with a modest decrease in toxicity risk (OR, 0.62; 95% CI, 0.43–0.9; P = 0.012). Conclusions: These results indicate that miR-27a and rs895819A>G may be clinically relevant for further toxicity risk stratification in carriers of DPYD risk variants. Our data suggest that direct suppression of DPD by miR-27a is primarily relevant in the context of fluoropyrimidine toxicity in patients with reduced DPD activity. However, miR-27a regulation of additional targets may outweigh its effect on DPD in patients without DPYD risk variants. Clin Cancer Res; 21(9); 2038–44. ©2015 AACR.

Abstract 5-fluorouracil (5-FU) is a commonly administered chemotherapy drug used for the treatment of various aggressive cancers. Dihydropyrimidine dehydrogenase (DPD, encoded by the DPYD gene) is the rate limiting enzyme of the 5-flourouracil (5-FU) catabolic pathway, and as such limits the amount of drug available for conversion into active metabolites. Reduced expression of DPD has been linked to severe adverse toxicity to 5-FU, including mucositis, diarrhea, hematological toxicities, and, in extreme cases, death. DPD has been previously shown to follow circadian patterns of expression in rodent models and in humans. Many microRNAs also exhibit circadian variations in expression and have been suggested to control oscillations in target protein expression. Using in silico prediction, we identified two putative binding domains for the microRNAs mir-27a and mir-27b within the DPYD gene. One of these sites is located in the 3’ untranslated region (UTR) and the other is located within the coding region just upstream from the stop codon. Using a reporter system, we show that mir-27a and mir-27b are able to bind to both predicted binding sites. Furthermore, over-expression of either microRNA reduced endogenous DPD protein levels in the HT-29 colon cancer cell line. To investigate mir-27a/b as a potential circadian regulator of DPD expression, expression was measured in livers collected at 4-hour intervals from chrono-synchronized mice. As expected, the circadian pattern of DPD enzyme activity closely matched that of DPD expression. The expression patterns for mir-27a and mir-27b were inverse of those for DPD expression and activity. These data suggest that mir-27a and mir-27b may direct circadian variations in DPD enzyme activity. Citation Format: Gabriel L. Butterfield, Steven M. Offer, Robert B. Diasio. Possible circadian modulation of dihydropyrimidine dehydrogenase by mir-27a and mir-27b. [abstract]. In: Proceedings of the 104th Annual Meeting of the American Association for Cancer Research; 2013 Apr 6-10; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2013;73(8 Suppl):Abstract nr 2198. doi:10.1158/1538-7445.AM2013-2198

Abstract The pyrimidine analog 5-fluorouracil (5-FU) was developed in the 1950s and remains one of the most widely used chemotherapy drugs in use today. A significant portion of patients who receive 5-FU suffer severe adverse toxicities, and, in extreme cases, death. Dihydropyrimidine dehydrogenase (DPD, encoded by the DPYD gene) rapidly degrades 85% of administered 5-FU, and as such, limits the amount of drug available for conversion into active metabolites. Genetic variations in DPYD have been suggested to increase the risk of 5-FU toxicity. Large-scale sequencing efforts have identified nearly 100 missense and nonsense DPYD variations, however only a handful of alleles have been studied for their association with 5-FU toxicity. Most studies have focused on a single allele (DPYD*2A), which only accounts for a small fraction of DPD deficiencies. Further complicating the interpretation of DPYD alleles, many individuals carry multiple variations within the DPYD gene. The purpose of this study was to identify missense variants that affect the enzyme function of DPD using a mammalian isogenic system of expression. Expression vectors encoding missense DPYD variations were generated by site directed mutagenesis and the expressed protein assayed for enzyme activity. Cells expressing selected variants were also teasted for sensitivity to 5-FU using a novel implementation of electrical impedance based realtime cellular analysis. The results of functional studies are compared to in silico predictions used to classify variants as deleterious or benign. An appreciable fraction of DPD variants showed enzyme activity that was significantly lower than that of wildtype protein. Within the deficient group of variants, a gradient of activities was noted, prompting the grouping of variants as mildly, moderately, and severely impaired. In addition to loss-of-function variations, a few hyperactive DPYD alleles were identified. This novel class of DPYD variants may be protective against 5-FU toxicity at the cost of treatment efficacy. The further evaluation of hyperactive alleles in clinical studies will be necessary to determine their impact on patient outcomes. Overall these data will facilitate the development of predictive tests for 5-FU toxicity by allowing researchers to prioritize variations for association studies. Citation Format: Steven M. Offer, Natalie J. Wegner, Croix Fossum, Robert D. Diasio. Classification of DPYD variations using an isogenic system of expression. [abstract]. In: Proceedings of the 104th Annual Meeting of the American Association for Cancer Research; 2013 Apr 6-10; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2013;73(8 Suppl):Abstract nr 2202. doi:10.1158/1538-7445.AM2013-2202