Predicting In Vivo Compound Brain Penetration Using Multi-task Graph Neural Networks.

Assessing whether compounds penetrate the brain can become critical in drug discovery, either to prevent adverse events or to reach the biological target. Generally, pre-clinical in vivo studies measuring the ratio of brain and blood concentrations (Kp) are required to estimate the brain penetration potential of a new drug entity. In this work, we developed machine learning models to predict in vivo compound brain penetration (as LogKp) from chemical structure. Our results show the benefit of including in vitro experimental data as auxiliary tasks in multi-task graph neural network (MT-GNN) models. MT-GNNs outperformed single-task (ST) models solely trained on in vivo brain penetration data. The best-performing MT-GNN regression model achieved a coefficient of determination of 0.42 and a mean absolute error of 0.39 (2.5-fold) on a prospective validation set and outperformed all tested ST models. To facilitate decision-making, compounds were classified into brain-penetrant or non-penetrant, achieving a Matthew's correlation coefficient of 0.66. Taken together, our findings indicate that the inclusion of in vitro assay data as MT-GNN auxiliary tasks improves in vivo brain penetration predictions and prospective compound prioritization.

Haplotype structure defines effects of common DPYD variants c.85T > C (rs1801265) and c.496A > G (rs2297595) on dihydropyrimidine dehydrogenase activity: Implication for 5-fluorouracil toxicity.

AIMS: The aim of this study was to identify risk variants and haplotypes that impair dihydropyrimidine dehydrogenase (DPD) activity and are, therefore, candidate risk variants for severe toxicity to 5-fluorouracil (5-FU) chemotherapy. METHODS: Plasma dihydrouracil/uracil (UH2 /U) ratios were measured as a population marker for DPD activity in a total of 1382 subjects from 4 independent studies. Genotype and haplotype correlations with UH2 /U ratios were assessed. RESULTS: Significantly lower UH2 /U ratios (panova < 2 × 10-16 ) were observed in carriers of the 4 well-studied 5-FU toxicity risk variants with mean differences (MD) of -43.7% for DPYD c.1905 + 1G > A (rs3918290), -46.0% for DPYD c.1679T > G (rs55886062), -37.1%, for DPYD c.2846A > T (rs67376798), and -13.2% for DPYD c.1129-5923C > G (rs75017182). An additional variant, DPYD c.496A > G (rs2297595), was also associated with lower UH2 /U ratios (P < .0001, MD: -12.6%). A haplotype analysis was performed for variants in linkage disequilibrium with c.496A > G, which consisted of the common variant c.85T > C (rs1801265) and the risk variant c.1129-5923C > G. Both haplotypes carrying c.496A > G were associated with decreased UH2 /U ratios (H3, P = .003, MD: -9.6%; H5, P = .002, MD: -16.9%). A haplotype carrying only the variant c.85T > C (H2) was associated with elevated ratios (P = .004, MD: +8.6%). CONCLUSIONS: Based on our data, DPYD-c.496A > G is a strong candidate risk allele for 5-FU toxicity. Our data suggest that DPYD-c.85T > C might be protective; however, the deleterious impacts of the linked alleles c.496A > G and c.1129-5923C > G likely limit this effect in patients. The possible protective effect of c.85T > C and linkage disequilibrium with c.496A > G and c.1129-5923C > G may have hampered prior association studies and should be considered in future clinical studies.

Fluoropyrimidine chemotherapy: recommendations for DPYD genotyping and therapeutic drug monitoring of the Swiss Group of Pharmacogenomics and Personalised Therapy.

Fluoropyrimidines (FPs), mainly 5-fluorouracil (5-FU) and its oral prodrug capecitabine (Cap), remain the backbone of the treatment of many different solid tumors. Despite their broad use in clinical routine, 10–40% of patients experience severe, and in rare cases (0.2–0.5%) even lethal, FP-related toxicity in early chemotherapy cycles. Today, there is a plethora of evidence that genetic variants in the gene encoding for the 5-FU catabolising enzyme dihydropyrimidine dehydrogenase (DPD, encoded by DPYD) are predictive of severe FP-related toxicities, and international clinical practice recommendations for DPYD genotype-guided FP dosing and therapeutic drug monitoring (TDM) are available. In spite of this strong evidence and DPYD genotyping becoming standard practice in other countries, it is has not been widely adopted in Switzerland to date. Here, we discuss current guidelines on genotype-guided FP dosing and TDM, and propose recommendations tailored to the situation in Switzerland to facilitate their clinical uptake for the further individualisation of FP chemotherapy. We recommend preemptive testing of four DPYD variants (c.1905+1G>A (rs3918290), c.1679T>G (rs55886062), c.2846A>T (rs67376798) and c.1129-5923C>G (rs75017182, c.1236G>A/HapB3)) in patients with an indication for FP-based chemotherapy, with the costs reimbursed through the compulsory health insurance in Switzerland. Carriers of these variants (6.5% in the Swiss population) have a 40–50% risk of developing severe early-onset toxicity when treated with standard FP doses. In these patients, we therefore recommend the use of a reduced starting dose, based on a dose adjustment scheme provided herein. Furthermore, we recommend the use of infusional 5-FU in patients with a DPYD risk genotype in order to enable TDM-based dose escalation. Only if the use of an infusional 5-FU regimen is not feasible should a slow titration of Cap, starting with the recommended reduced dose and basing further doses on monitoring of toxicity, be considered. Given that several studies have shown that TDM in 5-FU treatment improves not only the therapy’s safety, but potentially also its efficacy, we also include detailed TDM-based dosing guidelines and discuss the pre-analytical aspects of 5-FU TDM.

A Novel Nomenclature for Repeat Motifs in the Thymidylate Synthase Enhancer Region and Its Relevance for Pharmacogenetic Studies.

Inhibition of thymidylate synthase (TS) is the primary mode of action for 5-fluorouracil (5FU) chemotherapy. TS expression is modulated by a variable number of tandem repeats in the TS enhancer region (TSER) located upstream of the TS gene (TYMS). Variability in the TSER has been suggested to contribute to 5FU-induced adverse events. However, the precise genetic associations remain largely undefined due to high polymorphism and ambiguity in defining genotypes. To assess toxicity associations, we sequenced the TSER in 629 cancer patients treated with 5FU. Of the 13 alleles identified, few could be unambiguously named using current TSER-nomenclature. We devised a concise and unambiguous systematic naming approach for TSER-alleles that encompasses all known variants. After applying this comprehensive naming system to our data, we demonstrated that the number of upstream stimulatory factor (USF1-)binding sites in the TSER was significantly associated with gastrointestinal toxicity in 5FU treatment.

Evaluating the role of ENOSF1 and TYMS variants as predictors in fluoropyrimidine-related toxicities: An IPD meta-analysis.

To assess the proposed associations of the c.742-227G>A (rs2612091) polymorphism within the Enolase Superfamily Member 1 gene (ENOSF1) and two variants in the adjacent Thymidylate Synthase gene (TYMS): the 5'VNTR 28bp-repeat (rs45445694) and 3'UTR 6bp-indel (rs11280056) with severe toxicity in fluoropyrimidine-treated cancer patients, we performed an individual patient data meta-analysis. Only studies investigating all three-abovementioned variants with fluoropyrimidine-related toxicities were considered for meta-analysis. Associations were tested individually for each study using multivariate regression. Meta-analysis was performed using a random-effects model. One-stage multivariate regressions including tests for independent SNP effects were applied to investigate individual effects of the variants. Multivariate haplotype regression analyses were performed on a pooled dataset to test multi-SNP effects. Of four studies including 2'067 patients, 1'912 were eligible for meta-analysis. All variants were exclusively associated with severe hand-foot-syndrome (HFS) (TYMS 2R: OR = 1.50, p = 0.0002; TYMS 6bp-ins: OR = 1.42 p = 0.0036; ENOSF1 c.742-227G: OR = 1.64 p < 0.0001, per allele). We observed independent effects for ENOSF1 c.742-227G>A and the TYMS 28bp-repeat: each toxicity-associated allele increased the risk for severe HFS (OR = 1.32 per allele, p < 0.0001). Patients homozygous for both variants were at the 3-fold higher risk for severe HFS compared to wild-type patients. Our results confirm an essential role for ENOSF1 c.742-227G and TYMS 2R-alleles in the development of fluoropyrimidine-related HFS. This suggests an important function of these genes in the development of severe HFS. Furthermore, these variants might help stratify patients in studies investigating measures of HFS prevention.