Skip to main content
Download PDF

Genome-wide association study of abnormal elevation of ALT in patients exposed to atabecestat

BMC Genomics volume 24, Article number: 513 (2023) Cite this article

Abstract

Background

Atabecestat, a potent brain penetrable BACE1 inhibitor that reduces CSF amyloid beta (Aβ), was developed as an oral treatment for Alzheimer’s disease (AD). Elevated liver enzyme adverse events were reported in three studies although only one case met Hy’s law criteria to predict serious hepatotoxicity.

Method

We performed a case-control genome-wide association study (GWAS) to identify genetic risk variants associated with liver enzyme elevation using 42 cases with alanine transaminase (ALT) above three times the upper limit of normal (ULN) and 141 controls below ULN. Additionally, we performed a GWAS using continuous maximal ALT/ULN (expressed as times the ULN) upon exposure to atabecestat as the outcome measure (n = 285).

Results

No variant passed the genome-wide significance threshold (p = 5 × 10− 8) in the case-control GWAS. We identified suggestive association signals in genes (NLRP1, SCIMP, and C1QBP) implicated in the inflammatory processes. Among the genes implicated by position mapping using variants suggestively associated (p < 1 × 10− 5) with ALT elevation case-control status, gene sets involved in innate immune response (adjusted p-value = 0.05) and regulation of cytokine production (adjusted p-value = 0.04) were enriched. One genomic region in the intronic region of GABRG3 passed the genome-wide significance threshold in the continuous max(ALT/ULN) GWAS, and this variant was nominally associated with ALT elevation case status (p = 0.009).

Conclusion

The suggestive GWAS signals in the case-control GWAS analysis suggest the potential role of inflammation in atabecestat-induced liver enzyme elevation.

Peer Review reports

Background

Idiosyncratic drug-induced liver injury (DILI) is a challenge in drug development. It is both a challenge and an opportunity in the field of pharmacogenomics as genetic variants are often discovered in smaller sample sizes, compared to disease genetics. It is a challenge as the clinical development of compounds may be stopped for safety concerns before the sample size required for genetic discovery is reached. Even if causal genetic variants are identified, they may not have the necessary positive predictive value needed for a clinical application, as DILI is typically a rare event [1].

The discovered genetic risk factors for idiosyncratic DILI are primarily concentrated on polymorphisms in the human leukocyte antigen (HLA) region including HLA-DQB1*06:02 and HLA-A*02:01 associated with amoxicillin-clavulanate induced hepatotoxicity [2], HLA-DRB1*16:01-DQB1*05:02 for flupirtine-induced hepatotoxicity [3], HLA-B*57:01 for flucloxacillin-induced hepatotoxicity [4], HLA-DRB1*1501-HLA-DQB1*0602-HLA-DRB5*0101-HLA-DQA1*0102 for lumiracoxib-induced hepatotoxicity [5], HLA-DRB1*07:01 for lapatinib-induced hepatotoxicity [6], and HLA-DRB1*07 and HLA-DQA1*02 for ximelagatran-induced hepatotoxicity [7]. Additionally, variants in ST6GAL1, which plays a role in the systemic inflammatory response, and variants in the intron of FAM65B, which play roles in liver inflammation, have been associated with flucloxacillin [4] and antituberculosis drug-induced hepatotoxicity [8], respectively. Polymorphisms from transporters and metabolizing enzymes may also play a role [1].

A guidance for the management of suspected DILI has recommended consideration be given to stopping the suspected drug in any of the following situations: when ALT exceeding 8X the upper limit of normal (ULN); ALT exceeding 5XULN for more than 2 weeks; ALT exceeding 3XULN with total bilirubin > 2XULN or international normalized ratio (INR) > 1.5 (Hy’s Law criteria [9,10,11,12]]; ALT exceeding 3XULN with symptoms including fatigue, nausea, vomiting, right upper quadrant pain or tenderness, fever, rash, or eosinophilia [13]. The Roussel UCLAF Causality Assessment Method (RUCAM) may also be applied to assess causality taking age, gender, and medical history into account. In the context of drug development, Hy’s law is more widely used for predicting serious hepatotoxicity [9,10,11,12, 14]. The mechanisms of DILI have been reviewed previously and mitochondrial dysfunction, drug-induced free radical production, depletion of glutathione, peroxidation, perturbation of bile flow and activation of innate immune response [1, 15, 16] were implicated.

Atabecestat, a potent brain penetrable BACE1 inhibitor that reduces CSF amyloid beta (Aβ), was developed by Janssen Research & Development LLC and Shionogi and Co Ltd as an oral treatment for Alzheimer’s disease (AD). Elevated liver enzyme (> 3x ULN) adverse events were reported in approximately 10% of participants in three studies [17, 18] although only one case met the Hy’s law criteria to predict serious hepatotoxicity and was described earlier in a case report [19]. Previous challenging/cloning experiments using atabecestat and its metabolites (diaminothiazine [DIAT] and N-acetyl DIAT etc.) implicated the involvement of adaptive immunity in the observed liver enzyme elevation [20]. We herein performed genome wide association study (GWAS) to identify genetic risk variants associated with liver enzyme elevation in cohorts of patients from three phase II clinical studies exposed to atabecestat [17, 18]. Two outcome definitions [dichotomous trait, ALT elevation case control status and a continuous trait [i.e. maximal ALT (max(ALT/ULN)] level any time after drug exposure)] were used in the GWAS.

Results

The demographic characteristics of patients in the case control cohort is described in Supplemental Table 1. The top variant rs3865350 (Table 1, p = 4.63 × 10− 7, Manhattan and QQ plots in Supplemental Figs. 1 and 2, and genotype cluster plot in Supplemental Fig. 3) associated with ALT elevation was from a genomic region on chromosome 17p13.2, annotated to genes DERL2, MIS12, and NLRP1. The NLRP1 intronic variant kgp8586964 (also known as rs8067359), that was in linkage disequilibrium (LD) with rs3865350 (chr17: 5,381,867) and an eQTL variant for NLRP1 in fibroblast cells, was nominally associated with ALT elevation case control status (Table 1; Fig. 1B, p = 3.02 × 10− 6). The G allele of kgp8586964 is associated with greater risk of having ALT elevation (OR = 3.4, 95% confidence interval [2.0, 5.8]), and is also associated with greater level of NLRP1 expression in fibroblast cell (p = 2.2 × 10− 12, Supplemental Fig. 4, Data Source: GTEx Analysis Release V8).

Table 1 Variants associated with ALT elevation case-control analysis (p-value < 1 × 10− 5)
Full size table
Fig. 1
figure 1

Top variants associated with maximal ALT level. (A) kgp5495470/rs114579373 (B) rs3865350

Full size image

The genomic region (chr17: 5,320,001–5,360,000) forms a long-range 3D chromatin interaction loop with a genomic region (chr17: 5,120,001–5,160,000) containing SCIMP (chr17: 5,112,215-5,138,155) encoding SLP adaptor and CSK interacting membrane protein in liver (Source: HiC dataset GSE87112 [21], Fig. 2, Supplemental Table 2). Both NLRP1 and SCIMP are involved in inflammatory processes. rs3865350 was also suggestively associated with max(ALT/ULN) upon exposure to atabecestat (p = 5.87 × 10− 5). A list of variants associated with ALT elevation case status with a Fisher’s exact test association p-value less than 1 × 10− 5 is available in Supplemental Table 3 A together with association test results from continuous trait max(ALT/ULN) and sensitivity analyses including a logistic regression model for the ALT elevation case-control status (p = 1.55 × 10− 5 for rs3865350) and the similar GWAS analyses in the subsets of samples of European ancestry (p = 1.84 × 10− 5 for case-control analysis using 39 cases and 127 controls; p = 5.60 × 10− 5 for the continuous trait analysis in a cohort of 258 samples).

Fig. 2
figure 2

Circos plot for rs3865350 highlighting eQTLs and/or chromatin interactions. The circos plot was generated via FUMA [43] v1.3.4b (https://fuma.ctglab.nl/), the outer most layer is Manhattan plot and the middle layer highlights genomic risk loci (as defined by FUMA [43] using minimum P-value of lead SNPs of 1 × 10− 5 and default values for other parameters) in blue, while the inner most layer highlights eQTLs and/or chromatin interactions. Only SNPs with p < 0.05 are displayed in the outer ring. SNPs in genomic risk loci are color-coded as a function of their maximum r [2] to the one of the independent significant SNPs in the locus. The rsID of the top SNPs in each risk locus are displayed in the most outer layer. For the inner most layer, if the gene is mapped only by chromatin interactions or only by eQTLs, it is colored orange or green, respectively. It is colored red when the gene is mapped by both

Full size image

Two variants in the intronic region of GABRG3 on chromosome 15 passed the genome-wide significance threshold in the max(ALT/ULN) single-variant analyses (rs188010617- p = 3.04 × 10− 9 and rs193168063 - p = 1.18 × 10− 8). Rs188010617 was also nominally associated with ALT elevation case-control status (p = 0.01) in the logistic regression sensitivity analysis (this variant was not tested in Fisher’s exact test as it was not directly genotyped).

A list of variants associated with max(ALT/ULN) with an association p-value less than 1 × 10− 6 is available in Supplemental Table 3B together with association test results from Fisher’s exact and other sensitivity analyses. One of the top variants associated in the continuous maximal ALT trait analysis was kgp5495470 (also known as rs114579373) on chromosome 5q34 [p = 2.19 × 10− 7 for max(ALT/ULN), Manhattan and QQ plots in Supplemental Figs. 1 and 2, and genotype cluster plot in Supplemental Fig. 3]. Using the dichotomized ALT elevation trait, the same variant was associated with the case/control status (p = 3.23 × 10− 5, Fig. 1A). Eleven out of 84 (13%) chromosomes in cases vs. 4 out of 282 (1%) chromosomes in controls carried the T allele.

No individual gene passed the genome-wide significance threshold in any of the gene-based analyses (multi-marker analysis of genomic annotation (MAGMA) gene-based analysis). GSTM4 was nominally associated with case control ALT elevation status (p = 0.003). A full list of genes with nominal association p-value less than 0.005 is available in Supplemental Table 4.

No gene set passed the study-wide significance threshold in any of the gene set-based analyses. Gene sets involved in immune signaling pathways such as regulation of IFNa signaling (p = 1.25 × 10− 5), type I interferon pathway (p = 1.66 × 10− 5), IL27 pathway (p = 3.22 × 10− 5), IL10 pathway (p = 0.0002), alpha/beta T-cell activation (p = 0.0003) were nominally associated with ALT elevation case control status (Table 2, a full list of gene sets with nominal association p-value less than 0.005 is available in Supplemental Table 5). In addition, genes involved in the glutathione metabolism pathway were nominally associated with ALT elevation (p = 0.002). Gene sets involved in lectin pathway of complement activation was nominally associated with max(ALT/ULN) (p = 1.23 × 10− 5) upon exposure to atabecestat.

Table 2 Pathways enriched in ALT elevation case control status GWAS in the MAGMA analysis
Full size table

Among the genes implicated by position mapping, using variants suggestively associated with ALT elevation case control status (p < 1 × 10− 5), gene sets involved in innate immune response (adjusted p-value = 0.05) and regulation of cytokine production (adjusted p-value = 0.04) were enriched, among those in the GO [22] biological processes (from MsigDB [23] c5; Supplemental Table 6). NLRP1 and C1QBP were the genes responsible for these pathway enrichments.

Discussion

We performed GWAS using two phenotype definitions to identify genetic variants that might explain ALT elevation observed in a subset of patients exposed to atabecestat. Even though we did not identify any variant or gene passing the genome wide significance threshold associated with case-control status, several genes and pathways implicated in the immune response and glutathione metabolism pathway surfaced as candidate signals explaining the drug-induced ALT elevation. There was a genome wide significant finding in the continuous trait max(ALT/ULN) GWAS, with nominal support (p = 0.01) from the case-control status analysis.

Among the top variants associated with ALT elevation case control status, three genes implicated in the inflammatory process were implicated. NLRP1 (NLR family pyrin domain containing 1) was implicated by position mapping and encodes a member of the Ced-4 family of apoptosis proteins. Ced-family members contain a caspase recruitment domain (CARD) and are known to be key mediators of programmed cell death. NLRP1 contains a distinct N-terminal pyrin domain (PYD), which may mediate protein-protein interactions. NLRP1 and other leucine-rich repeat (LRR)-containing protein (NLR) family members can initiate the formation of inflammasomes which are cytosolic multi-protein complexes serving as sensors for infection and damage to illicit inflammatory response [24, 25]. It has been reported that inflammasome is activated in liver of cholestatic patients and aggravates hepatic injury in bile duct-ligated mouse [26]. SCIMP was implicated by long range chromatin interaction. SCIMP encodes a transmembrane adaptor protein that is expressed in antigen-presenting cells and localized in the immunologic synapse and is involved in major histocompatibility complex class II signal transduction and immune synapse formation, and mediates TLR-4 mediated cytokine responses in macrophages upon LPS stimulation [27].

In addition, pathways enrichment analysis using MAGMA uses the entire p-value distribution and over-representation analysis implemented in GENE2FUNC of FUMA uses only the genes implicated by position mapping among the top associated variants (p < 1 × 10− 5). Many pathways in the inflammatory processes (Supplemental Tables 4 and 5 and Fig. 3) were implicated. In particular, innate immune response and regulation of cytokine production (implicated by C1QBP and NLRP1) were enriched (adjusted p-value < = 0.05) among genes implicated by these top associated variants. rs3865350 is an sQTL variant for C1QBP in fibroblast and lymphocytes (Supplemental Fig. 5), which encodes complement C1q binding protein. It regulates the transcriptional activity of hypoxanthine catabolic enzyme xanthine dehydrogenase (XDH) in cancer cells presumably by regulating the mRNA level of XDH transcriptional stimulators IL-6, TNF-α, and IFN-γ, and promotes the catabolism of hypoxanthine [28]. XDH also mediates reactive oxygen species (ROS) generation and apoptosis. Both innate immune response and oxidative stress are hypothesized mechanisms for DILI. It is worth to mention in the in vitro challenging/cloning experiment also implicated the role of adaptive response in atabecestat liver injury where CD4 + T-cell clones activated by the atabecestat metabolite DIAT were detected in five out of eight patients, and CD4 + and CD8 + clones activated by atabecestat were also detected although with a lower cloning efficiency [20]. These clones proliferated and secreted IFN-γ and IL-13 following atabecestat or DIAT stimulation [20].

Fig. 3
figure 3

Pathway enrichment among genes implicated by position mapping for variants associated with ALT elevation case control status (p < 1 × 10− 5)

Full size image

It is also of interest that several eQTL/sQTL relationships between chr17 variants (rs3865350 and kgp8586964) and NLRP1 and C1QBP were detected in fibroblast cells. Hepatic stellate cells (HSCs) account for 15% of total resident cells in the normal human liver and are resident mesenchymal cells with features of resident fibroblasts and pericytes. HSCs play key roles in liver fibrosis. HSCs are quiescent in the normal liver and become activated during liver injury, resulting in inflammation and activation of the innate immune system leading to liver fibrosis [29]. On the other hand, transient fibrosis could be a protective mechanism by maintaining cell attachment and the architecture of liver tissue via production of extracellular matrix proteins and a regenerative mechanism by repairing liver damage by producing growth factors [30, 31].

In the continuous trait analysis, we identified intronic variants in GABRA3 associated with max(ALT/ULN) reaching study wide significance threshold. GABRA3 in liver plays a protective role for DILI as GABRA3 is involved in altered signaling for driving hepatocyte proliferation to restore damaged hepatoctyes [32].

With the limited sample size, the suggestive GWAS signals implicate the potential role of inflammation in the atabecestat induced liver enzyme elevation. There are several limitations for this study besides the sample size. First, samples were genotyped in 4 batches (in particular, the last two samples were singleton samples, and there is no convenient way of assessing genotyping quality); second, due to the sample size limitation, the case control analysis used all samples and Fisher’s Exact test was used for association analysis (to handle the risk of potential cell count less than five in the 2 × 2 table used for statistical test), although the quantitative trait max(ALT/ULN) corrected for 5 principal components (PCs). We however cross-checked the top association results to provide additional supporting evidence and performed sensitivity analyses using logistic regression correcting for 2 PCs (running the risk of small cell count) and performing similar analysis in the population of European ancestry alone (with even smaller sample size); third, the phenotypes used is a risk factor for liver hepatotoxicity and not DILI, which is a more extreme phenotype and could help to uncover the genetic cause better; fourth, given the small sample size, the model did not explicitly correct for age and gender to use up another 2 degrees of freedom, even though the extent of ALT elevation was expressed as ALT/ULN and ULN is age and gender dependent. Clinical trials on several compounds with similar mechanism of action including lanabecestat and atabecestat (but not verubecestat) had revealed similar issues of liver enzyme elevation [33, 34]. It will be of interest to learn if there are overlapping mechanisms underlying the observed liver enzyme elevation. Future endeavors of this line of work may also encompass longitudinal repeat measurement modeling and/or use a different aggregation method other than maximum.

Conclusions

We performed GWAS to identify genetic risk factors for atabecestat induced ALT elevation. The suggestive GWAS signals implicate the potential role of inflammation in the atabecestat induced liver enzyme elevation.

Methods

Clinical cohorts

The samples included in the analysis were pooled from three studies sponsored by Janssen Research & Development, LLC including 54861911ALZ2002 (NCT02260674), a Phase 2a study to evaluate the safety and tolerability of atabecestat (JNJ-54,861,911), an oral BACE1 inhibitor in participants with early Alzheimer’s Disease [17]; 54861911ALZ2004 (NCT02406027, terminated), the long-term extension study of (JNJ-54,861,911) in early Alzheimer’s disease spectrum patients [17]; and 54861911AZ2003(NCT02569398, terminated), Phase 2b efficacy and safety study of Atabecestat in participants who are asymptomatic at risk for developing Alzheimer’s Dementia (EARLY) [18]. In 54861911ALZ2002 and 54861911ALZ2004, 12 of 104 subjects (11.5%) experienced liver enzyme elevation, resulting in dosage modification and increased frequency of safety monitoring, but no case met the Hy’s law criteria (ALT or aspartate aminotransferase (AST) > 3× ULN and total bilirubin > 2× ULN) [17] to predict serious hepatotoxicity. Treatment discontinuation normalized ALT or AST in all except one with pretreatment elevation [17]. In 54861911ALZ2003, 40 of 371 subjects (10.8%) receiving atabecestat and 1 subject in the placebo from 54861911ALZ2003 experienced ALT or AST elevation (> 3x ULN), among which one subject met the Hy’s law criteria [18]. As a result, 54861911ALZ2003 and 54861911ALZ2004 were terminated early. Samples included in this genetic analysis were from subjects treated with atabecestat before the decision for trial termination and therefore not all cases with liver enzyme elevation were included. The clinical studies were carried out in accordance with the ethical principles outlined in the Declaration of Helsinki, Good Clinical Practices guidelines, and applicable regulatory requirements.

Phenotype definitions

Two phenotype definitions for ALT elevation were used. First, a dichotomized ALT elevation case control status phenotype was used where cases were defined as having ALT level > = 3x ULN at any time during the study after atabecestat exposure, while controls were defined as having ALT < = 1x ULN at all times; Second, a continuous phenotype was used using log transformation of maximal ALT value (expressed as times the ULN) at any time during the study to approximate normal distribution. In total 42 cases and 141 controls were included in the GWAS using the dichotomized phenotype, while 285 samples were included in the GWAS using the continuous phenotype. A full list of case by ULN was provided in Supplemental Table 7.

Genotyping

A total of 380 samples were genotyped using Omni2.5MExome or Omni2.5 M chips over 4 batches (2 of them were singleton presumably with samples from other studies by the genotyping vendor) at three facilities (Supplemental Table 8).

Genotype data quality control

The genotype data were merged and QC’ed where samples exceeding 5% missingness rate were excluded, whereas SNP exceeding 5% missingness rate, deviating from Hardy-Weinberg equilibrium (p < 1 × 10− 6) or having minor allele frequency (MAF) below 1% were excluded. The total genotyping rate was 99.86%. Among the remaining participants, relatedness was assessed using pairwise identity by descent (IBD) estimation in PLINK [35]. Participants were excluded as needed to ensure that the estimated proportion of IBD (PI HAT) between any two remaining individuals was less than 0.3. Additionally, subjects (if any) with genetically inferred gender discrepancies from the gender provided in the clinical database were excluded. No subject was excluded in this study due to this reason. No outlier removal was applied to remove subjects 6 sigma away using EIGENSTRAT [36, 37] to maximize the power of the study. After genotyping QC, 374 samples were retained removing samples with low call rates and related samples (Supplemental Table 9).

The genotype data were further QC’ed using RICOPILI pipeline [38] with standard parameters [SNP-wise missingness rate 2%, Hardy-Weinberg equilibrium (p < 1 × 10− 6 in controls, or p < 1 × 10− 10 in cases)], except when sample-wise missingness rate was set to 4%.

Genotype data imputation

To allow the exploration of a greater density of markers than what was genotyped directly, the imputation was performed using RICOPILI (minimac3 [39]/Eagle [40] v2.3.5) pipeline. Imputation of unobserved genotypes was based on the reference haplotypes from the 1,000 Genomes haplotypes; Phase 3 integrated variant set release in NCBI build 37 (hg19) coordinates from the 1000 Genome Project [41].

Single marker GWAS analysis

Fisher exact test was used in the case-control association analysis without correcting for PCA using the directly genotyped markers passing SNP QC (n = 1,526,079). ~11 M imputed and directly genotyped markers were used in the max(ALT/ULN) genome wide association analysis. The continuous trait max(ALT/ULN) was analyzed in a linear regression model correcting for five principal components to account for population substructure and using the imputed genotype dosages. All association tests were performed using PLINK [35, 42]. Genotype quality was spot-checked post-GWAS to ensure the genotype quality where Genome Studio (Illumina, Inc.) Project files were available (Supplemental Fig. 3). A conventional genome-wide significance threshold of 5 × 10− 8 was used to declare study-wide significance. A list of variants with unadjusted p-value less than 1 × 10− 5 (Fisher’s exact test) or 1 × 10− 6 (linear regression) was also reported.

GWAS sensitivity analysis

Fisher’s exact test cannot account for covariates. We also performed case-control logistic regression GWAS as a sensitivity analysis and additionally correct for two principal components. These analyses were also performed using RICOPILI pipeline [38] and the 1,000 Genomes haplotypes phase 3 reference panel. Additionally, the subset of homogenous samples of European ancestry were used as another sensitivity analysis for both case-control and continuous max(ALT/ULN) outcomes. These results were only used for cross-reference purpose only to ensure that population stratification was not the reason for the observed association signals.

Multi-marker analysis of genomic annotation (MAGMA) gene-based analysis, gene set-based analysis, and variant annotations

Variant clumping to identify independent genomic locus and annotation was performed using FUMA [43]. In addition to single-marker-based GWAS, gene-based analyses followed by pathway enrichment analysis were computed using MAGMA (v1.07) [44] based on GWAS summary statistics. SNPs were mapped to 18,035 protein-coding genes (Ensembl v92). Genome-wide significance was defined at p = 0.05/18,035 = 2.77 × 10− 6. The MAGMA analyses were performed using FUMA [43].

In addition to MAGMA gene set based enrichment analysis, which uses the entire p-value distribution, gene set enrichment analysis was performed using variants with association p-value less than 1 × 10− 5 using GENE2FUNC feature of FUMA. Variants were annotated using FUMA using position-based, eQTL-based, and 3D chromatin interaction information. The Manhattan plots, QQ plots, and circos plots were generated using FUMA (v1.3.4b) [43] available from https://fuma.ctglab.nl/.

Data Availability

Due to the proprietary nature of the data and privacy obligation, the request for the data will be a considered on a case-by-case basis and a Data Transfer Agreement is required to ensure the research is consistent with the scope of informed consent and necessary precaution is taken to protect patient privacy. Data request shall be directed to QSL.

Abbreviations

AD:

Alzheimer’s disease

ALT:

Alanine transaminase

AST:

Aspartate aminotransferase

DILI:

Drug-induced liver injury

GWAS:

Genome wide association study

ULN:

Upper limit of normal

References

  1. Shao Q, Mao X, Zhou Z, Huai C, Li Z. Research Progress of Pharmacogenomics in Drug-Induced Liver Injury. Front Pharmacol. 2021;12:735260.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  2. Lucena MI, et al. Susceptibility to amoxicillin-clavulanate-induced liver injury is influenced by multiple HLA class I and II alleles. Gastroenterology. 2011;141:338–47.

    Article  CAS  PubMed  Google Scholar 

  3. Nicoletti P, et al. HLA-DRB1*16: 01-DQB1*05: 02 is a novel genetic risk factor for flupirtine-induced liver injury. Pharmacogenet Genomics. 2016;26:218–24.

    Article  CAS  PubMed  Google Scholar 

  4. Daly AK, et al. HLA-B*5701 genotype is a major determinant of drug-induced liver injury due to flucloxacillin. Nat Genet. 2009;41:816–9.

    Article  CAS  PubMed  Google Scholar 

  5. Singer JB, et al. A genome-wide study identifies HLA alleles associated with lumiracoxib-related liver injury. Nat Genet. 2010;42:711–4.

    Article  CAS  PubMed  Google Scholar 

  6. Spraggs CF, et al. Characterisation of the HLA-DRB1*07:01 biomarker for lapatinib-induced liver toxicity during treatment of early-stage breast cancer patients with lapatinib in combination with trastuzumab and/or taxanes. Pharmacogenomics J. 2018;18:480–6.

    Article  CAS  PubMed  Google Scholar 

  7. Kindmark A, et al. Genome-wide pharmacogenetic investigation of a hepatic adverse event without clinical signs of immunopathology suggests an underlying immune pathogenesis. Pharmacogenomics J. 2008;8:186–95.

    Article  CAS  PubMed  Google Scholar 

  8. Petros Z, et al. Genome-wide association and replication study of anti-tuberculosis drugs-induced liver toxicity. BMC Genomics. 2016;17:755.

    Article  PubMed  PubMed Central  Google Scholar 

  9. Reuben A. Hy’s law. Hepatology. 2004;39:574–8.

    Article  PubMed  Google Scholar 

  10. Lewis JH. Hy’s law,‘ the ‘Rezulin Rule,‘ and other predictors of severe drug-induced hepatotoxicity: putting risk-benefit into perspective. Pharmacoepidemiol Drug Saf. 2006;15:221–9.

    Article  CAS  PubMed  Google Scholar 

  11. Senior JR. How can ‘Hy’s law’ help the clinician? Pharmacoepidemiol Drug Saf. 2006;15:235–9.

    Article  PubMed  Google Scholar 

  12. Temple R. Hy’s law: predicting serious hepatotoxicity. Pharmacoepidemiol Drug Saf. 2006;15:241–3.

    Article  PubMed  Google Scholar 

  13. Services USDoHaH, Administration FaD, CDER). C.f.D.E.a.R. & (CBER), C.f.B.E.a.R. Guidance for Industry Drug-Induced Liver Injury: Premarketing Clinical Evaluation. Vol. 2023 (2009).

  14. Regev A, Bjornsson ES. Drug-induced liver injury: morbidity, mortality, and Hy’s law. Gastroenterology. 2014;147:20–4.

    Article  PubMed  Google Scholar 

  15. Petros Z, Makonnen E, Aklillu E. Genome-Wide Association Studies for Idiosyncratic Drug-Induced Hepatotoxicity: Looking Back-Looking Forward to Next-Generation Innovation. OMICS 21, 123–131 (2017).

  16. Stephens C, Lucena MI, Andrade RJ. Genetic risk factors in the development of idiosyncratic drug-induced liver injury. Expert Opin Drug Metab Toxicol. 2021;17:153–69.

    Article  CAS  PubMed  Google Scholar 

  17. Novak G, et al. Long-term safety and tolerability of atabecestat (JNJ-54861911), an oral BACE1 inhibitor, in early Alzheimer’s disease spectrum patients: a randomized, double-blind, placebo-controlled study and a two-period extension study. Alzheimers Res Ther. 2020;12:58.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  18. Sperling R, et al. Findings of Efficacy, Safety, and Biomarker Outcomes of Atabecestat in Preclinical Alzheimer Disease: a truncated randomized phase 2b/3 clinical trial. JAMA Neurol. 2021;78:293–301.

    Article  PubMed  Google Scholar 

  19. De Jonghe S, Weinstock D, Aligo J, Washington K, Naisbitt D. Biopsy Pathology and immunohistochemistry of a case of Immune-Mediated Drug-Induced Liver Injury with Atabecestat. Hepatology. 2021;73:452–5.

    Article  PubMed  Google Scholar 

  20. Thomson PJ, et al. Drug-specific T-cell responses in patients with liver injury following treatment with the BACE inhibitor atabecestat. Allergy. 2021;76:1825–35.

    Article  CAS  PubMed  Google Scholar 

  21. Schmitt AD, et al. A compendium of chromatin contact maps reveals spatially active regions in the Human Genome. Cell Rep. 2016;17:2042–59.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  22. Boyle EI, et al. GO::TermFinder–open source software for accessing Gene Ontology information and finding significantly enriched gene ontology terms associated with a list of genes. Bioinformatics. 2004;20:3710–5.

    Article  CAS  PubMed  Google Scholar 

  23. Liberzon A, et al. Molecular signatures database (MSigDB) 3.0. Bioinformatics. 2011;27:1739–40.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  24. Jenster LM et al. P38 kinases mediate NLRP1 inflammasome activation after ribotoxic stress response and virus infection. J Exp Med 220(2023).

  25. Broz P, Dixit VM. Inflammasomes: mechanism of assembly, regulation and signalling. Nat Rev Immunol. 2016;16:407–20.

    Article  CAS  PubMed  Google Scholar 

  26. Cai SY, et al. Inflammasome is activated in the liver of Cholestatic Patients and aggravates hepatic Injury in bile Duct-Ligated mouse. Cell Mol Gastroenterol Hepatol. 2020;9:679–88.

    Article  PubMed  Google Scholar 

  27. Curson JEB et al. TLR4 phosphorylation at tyrosine 672 activates the ERK/c-FOS signaling module for LPS-induced cytokine responses in macrophages. Eur J Immunol, e2250056 (2023).

  28. Wang Y, et al. C1QBP regulates apoptosis of renal cell carcinoma via modulating xanthine dehydrogenase (XDH) mediated ROS generation. Int J Med Sci. 2022;19:842–57.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  29. Wen YA, et al. Biliary intervention aggravates cholestatic liver injury, and induces hepatic inflammation, proliferation and fibrogenesis in BDL mice. Exp Toxicol Pathol. 2011;63:277–84.

    Article  CAS  PubMed  Google Scholar 

  30. Dechene A, et al. Acute liver failure is associated with elevated liver stiffness and hepatic stellate cell activation. Hepatology. 2010;52:1008–16.

    Article  CAS  PubMed  Google Scholar 

  31. Li J, Zhao YR, Tian Z. Roles of hepatic stellate cells in acute liver failure: from the perspective of inflammation and fibrosis. World J Hepatol. 2019;11:412–20.

    Article  PubMed  PubMed Central  Google Scholar 

  32. Kim M, et al. Expression levels of GABA-A receptor subunit alpha 3, Gabra3 and lipoprotein lipase, Lpl are Associated with the susceptibility to Acetaminophen-Induced Hepatotoxicity. Biomol Ther (Seoul). 2017;25:112–21.

    Article  CAS  PubMed  Google Scholar 

  33. Egan MF, et al. Further analyses of the safety of verubecestat in the phase 3 EPOCH trial of mild-to-moderate Alzheimer’s disease. Alzheimers Res Ther. 2019;11:68.

    Article  PubMed  PubMed Central  Google Scholar 

  34. May PC, et al. The potent BACE1 inhibitor LY2886721 elicits robust central abeta pharmacodynamic responses in mice, dogs, and humans. J Neurosci. 2015;35:1199–210.

    Article  PubMed  PubMed Central  Google Scholar 

  35. Purcell S, et al. PLINK: a tool set for whole-genome association and population-based linkage analyses. Am J Hum Genet. 2007;81:559–75.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  36. Patterson N, Price AL, Reich D. Population structure and eigenanalysis. PLoS Genet. 2006;2:e190.

    Article  PubMed  PubMed Central  Google Scholar 

  37. Price AL, et al. Principal components analysis corrects for stratification in genome-wide association studies. Nat Genet. 2006;38:904–9.

    Article  CAS  PubMed  Google Scholar 

  38. Lam M, et al. Rapid Imputation for COnsortias PIpeLIne. Bioinformatics. 2020;36:930–3.

    Article  CAS  PubMed  Google Scholar 

  39. Das S, et al. Next-generation genotype imputation service and methods. Nat Genet. 2016;48:1284–7.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  40. Loh PR, et al. Reference-based phasing using the Haplotype Reference Consortium panel. Nat Genet. 2016;48:1443–8.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  41. Genomes Project C, et al. A map of human genome variation from population-scale sequencing. Nature. 2010;467:1061–73.

    Article  Google Scholar 

  42. Chang CC, et al. Second-generation PLINK: rising to the challenge of larger and richer datasets. Gigascience. 2015;4:7.

    Article  PubMed  PubMed Central  Google Scholar 

  43. Watanabe K, Taskesen E, van Bochoven A, Posthuma D. Functional mapping and annotation of genetic associations with FUMA. Nat Commun. 2017;8:1826.

    Article  PubMed  PubMed Central  Google Scholar 

  44. de Leeuw CA, Mooij JM, Heskes T, Posthuma D. MAGMA: generalized gene-set analysis of GWAS data. PLoS Comput Biol. 2015;11:e1004219.

    Article  PubMed  PubMed Central  Google Scholar 

Download references

Acknowledgements

We are grateful to the study volunteers for participating in the clinical studies and to the clinicians and support staff for enabling patient recruitment and blood sample collection. Informed consent was obtained from all participants. We thank the staff at Covance and the Neuroscience Biomarkers team at Janssen for managing the EDTA blood samples and managing genotyping of batch 2–4 samples, the staff at Gentris, Eurofins, and CGL for DNA extraction and plating, and the staff at Illumina, Eurofins, and CGL for genotyping the batch 1, 2, and 3/4 Janssen DNA samples, respectively.

Funding

The work was funded by Janssen Research & Development, LLC.

Author information

Author notes

  1. Stephan Francke

    Present address: Pharmacogenomics & Biomarker in Clinical Development, Cary, NC, USA

  2. Gary Romano

    Present address: Passage Bio, Philadelphia, PA, USA

Authors and Affiliations

  1. Neuroscience, Janssen Research & Development, LLC, Titusville, NJ, 08560, USA

    Qingqin S. Li, John Thipphawong, Gary Romano & Gerald P. Novak

  2. JRD Data Science, Janssen Research & Development, LLC, Titusville, NJ, 08560, USA

    Qingqin S. Li

  3. Computational Science Translational Platforms, Janssen Research & Development, LLC, Spring House, PA, 19477, USA

    Stephan Francke

  4. Translational Pharmacokinetics Pharmacodynamics and Investigative Toxicology, Janssen Research & Development, Beerse, Belgium

    Jan Snoeys

Authors
  1. Qingqin S. Li
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Stephan Francke
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Jan Snoeys
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. John Thipphawong
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Gary Romano
    View author publications

    You can also search for this author in PubMed Google Scholar

  6. Gerald P. Novak
    View author publications

    You can also search for this author in PubMed Google Scholar

Contributions

QSL, SF, and GN conceived and initiated the project, QSL performed the analysis, and wrote the first draft of the manuscript. All authors contributed to the data analysis review, discussions, and contributed to the final manuscript.

Corresponding author

Correspondence to Qingqin S. Li.

Ethics declarations

Competing interests

All authors were employees of Janssen Research & Development, LLC and equity shareholders in Johnson & Johnson, the parent company of the Janssen companies when this work was conducted. Drs. Li, Francke, Snoeys, Thipphawong, Romano, and Novak declared that, except for income received from their primary employer, no financial support or compensation has been received from any individual or corporate entity over the past three years for research or professional service, and there were no other personal financial holdings that could be perceived as constituting a potential conflict of interest.

Ethics approval and consent to participate

The study protocols were approved by the respective central, regional, or site-specific Institution Review Board or Ethic Committee. All subjects participating in the study provided Informed consent for participating in the clinical study and for genetic testing.

Consent for publication

N/A.

Additional information

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Supplementary Material 1

Supplementary Material 2

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Li, Q.S., Francke, S., Snoeys, J. et al. Genome-wide association study of abnormal elevation of ALT in patients exposed to atabecestat. BMC Genomics 24, 513 (2023). https://doi.org/10.1186/s12864-023-09625-6

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1186/s12864-023-09625-6

Keywords

BMC Genomics

ISSN: 1471-2164

Contact us