Clinically Important Alterations in Pharmacogene Expression in Histologically Severe Nonalcoholic Fatty Liver Disease

Abstract

Background Polypharmacy is common in patients with nonalcoholic fatty liver disease (NAFLD) and previous reports suggest that NAFLD is associated with altered drug disposition. This study aimed to determine if patients with histologically severe NAFLD are at risk for altered drug response by characterizing changes in hepatic mRNA expression of genes mediating drug disposition (pharmacogenes) across the spectrum of NAFLD severity. Methods We utilized hepatic RNA-seq from 93 patients with histologically staged NAFLD to test the relationship between pharmacogene expression and histological NAFLD Activity Score (NAS), fibrosis stage, and steatohepatitis (NASH). We then validated our results by meta-analysis of 16 studies, including ours. Results We identified 37 pharmacogene-NAFLD severity associations that were statistically significant after Bonferroni correction. Among all of the significant associations were 4 CYP enzymes (CYP2C19, CYP1B1, CYP2C8, CYP27B1), 3 phase II metabolic enzymes (GSTP1, GSTT1, GSTZ1), 6 ABC transporters (ABCB1, ABCB4, ABCB8, ABCC1, ABCC3, ABCC4), and 7 SLC transporters (SLC22A12, SLC16A1, SLCO3A1, SLC28A3, SLC2A4, SLC22A17, SLC6A6). We chose to validate CYP2C19 due to its actionability in clopidogrel prescribing and found that, compared to controls, it is significantly downregulated to 46% in NASH, to 58% in high NAS, and to 43% in severe fibrosis in our meta-analysis. In the regression analyses, CYP2C19 transcript abundance decreases to 69% with every 1 unit increase in fibrosis stage; reducing CYP2C19 mRNA expression levels by 77% in individuals with stage 4 fibrosis as compared to stage 0. With every 1 unit increase in NAS, CYP2C19 transcript abundance decreases to 83%; reducing CYP2C19 mRNA expression levels by 73% in individuals with a NAS of 7 as compared to 0. Hepatic CYP2C19 transcript abundance is 63% lower in NASH compared to those without. Conclusions Our data demonstrate that mRNA expression levels of several pharmacogenes are altered in livers of patients with NAFLD. Of these, the marked down-regulation of CYP2C19 presents a clear opportunity to further develop individualized treatment modifications for drugs that are sensitive substrates of the CYP2C19 enzyme (e.g., clopidogrel).

miR-27b targets MAIP1 to mediate lipid accumulation in cultured human and mouse hepatic cells

“…Gene expression analysis of MAIP1, B4GALT3 and PHLPP2 in liver biopsies form NASH patients. We also performed a separate analysis using a public database of RNA-seq data from the livers of NASH patients (GEO, GE225740) 64,65 , in which we searched for the expressions of MAIP1, B4GALT3, and PHLPP2 (Supplementary Fig. 4).…”

“…4). The RNA samples were categorized into three groups (0, no steatohepatitis; 1, borderline steatohepatitis; 2, definite steatohepatitis) according to Powell et al 65 . The expression of MAIP1 tended to decline with increasing level of steatohepatitis, suggesting that MAIP1 was involved in the manifestation.…”

Intermittent fasting regimens for metabolic dysfunction-associated steatotic liver disease: a systematic review and network meta-analysis of randomized controlled trials

“…Similar to other cardiometabolic risk factors, hypercholesterolemia in MASLD patients is usually treated with oral drugs as lifetime medications. This contributes to the issue of polypharmacy, which is common among patients with MASLD [35]. In these patients, IF can offer the benefit of adjusting lipid profile with reducing polypharmacy, ultimately, decreasing the risks of drug complications, interactions, and nonadherence.…”

Enformer: An encoder-decoder generative model to enhance prediction of disease outcomes using electronic health records

“…However, their pretraining objectives were limited in predicting a fraction of ICD codes within each visit. In reality, most patients have multiple diseases or outcomes at once, 16 many of which are highly correlated (such as obesity, diabetes, and hypertension [17][18][19] ) and thus collectively contribute to the disease or outcome trajectories. Therefore, a novel pretraining strategy, which predicts the complete set of diseases and outcomes within a visit, might improve clinical predictive modeling.…”