Metagenomic and transcriptomic investigation of pediatric acute liver failure cases reveals a common pathway predominated by monocytes

Abstract

In 2022, a cluster of severe childhood hepatitis was detected primarily in Europe and North America, leading to a global alert by the World Health Organization. An association with adeno-associated virus 2 (AAV2) in conjunction with human adenoviruses was found. Five percent of the cases progressed to acute liver failure, necessitating transplantation. The mechanism of disease that accounts for fulminant liver failure in these patients remains incompletely described. An upsurge was observed of in the five total cases of acute liver failure that presented to the Dutch national referral center for pediatric liver transplantation in the spring of 2022. An in-depth molecular analysis of the mechanism of pediatric acute liver failure was performed using targeted transcriptomics and metagenomics to identify any virus present in the cases, immune profile haplotypes, and differentially expressed gene groups. Explanted liver tissue and plasma samples (n = 15) were subjected to viral metagenomic and human transcriptomic profiling, targeting >600 inflammatory genes. Liver transcriptomic signatures of transplanted cases were compared with those of pediatric controls from a liver biobank (n = 6). AAV2, adenoviruses, and herpesviruses were detected in liver explant tissue and plasma samples of the cases. Epstein-Barr virus and varicella zoster virus infection with pathognomonic clinical symptomatology preceded liver failure in two respective cases. AAV2 was detected in one-third of control livers. Excessive activation of monocyte pathways was detected in liver explants from cases compared with controls. Remarkably, this signature was comparable for AAV2, adenoviruses, and/or herpesviruses-positive transplant cases. Our multi-omic findings suggest a common transcriptomic profile, with an upregulation of monocyte pathways in the presented transplanted cases, which had similar severe clinical outcomes. In the cohort presented, AAV2 was not exclusively associated with acute liver failure, suggesting that other processes may have contributed to a uniform cascade of irreversible pathology.

Importance: Since the appearance of the cluster of pediatric hepatitis of unknown origin in 2022, several groups have reported an association of adenoviruses and AAV2 in a high number of cases in contrast to controls. The adenoviruses detected were heterogeneous in both species-adenovirus C and F-and sequences. The mechanisms of disease that accounts for fulminant liver failure, occurring in 5% of pediatric hepatitis cases, remain incompletely described. The current study adds to previous data by including pediatric acute liver failure cases during the upsurge, enabling the analyses of inflammation expression profiles in cases with different viruses in relation to pediatric controls. This led to the discovery of transcriptome upregulation of monocyte pathways in liver explants from the cases. This inflammatory transcriptomic signature was comparable for AAV2, adenoviruses, and/or herpesviruses-positive transplant cases.

Keywords: adeno-associated virus; childhood hepatitis; metagenomics; transcriptomics.

Fig 1

(a) Study design. Cases of non-A-E hepatitis (n = 5) and non-hepatitis controls (n = 6). Explanted liver biopsies and plasma samples from cases were subjected to metagenomic sequencing analyses. Controls were tested by PCR for metagenomic findings in the cases. Liver tissue samples from both cases and controls were subjected to transcriptomics. HLA DRB1*04:01 positivity is indicated. *Not transplanted, no liver explant available. Hib, Haemophilus influenzae b; GAS, group A streptococcus; VZV, varicella zoster virus; B19, human parvovirus B19. (b) Viral metagenomic detection of viruses in pediatric cases of acute liver failure. Numbers correspond to Ct values of confirmatory PCRs. Circle sizes are representative of the number of sequence reads normalized per virus genome size (RPKM, reads per kilobase genome per million). − , not detected; HHV, human herpes virus; VZV, varicella zoster virus; TTV, torque teno virus; HPyV, human polyomavirus. Created using Biorender.com.

Fig 2

Transcriptomics results of liver samples from transplanted patients and controls. (a) Heatmap of the expressions of inflammation gene transcripts targeted, in cases (n = 4) and controls (n = 6), as log₂ counts per million (Table S2). (b) Principal component analyses of gene transcripts in cases and controls. (c) Volcano plot showing differentially expressed genes, with all genes included. (d–f) Single sample gene set enrichment analysis (ssGSEA), comparing cases against the average of controls. Reactome and blood transcriptional modules (14) were included as pathways. Only pathways with statistically significant enrichment (P-adjusted < 0.05) in at least two cases are shown. Size and color indicate direction and magnitude per pathway and case. Enrichment scores are shown for two example pathways per case. Colored lines below depict the rank, based on fold change relative to controls, of all genes present in the pathway. Highest-ranking genes indicate an increase relative to controls and are shown on the left, and lowest-ranking genes indicate the greatest decrease relative to controls and are shown on the right. (g) Monocyte/macrophage-derived pro-inflammatory core (leading edge) genes in cases. (h) Decreased expression of complement pathways leading edge genes in cases compared with controls. Boxplots center line, median; box limits, upper and lower quartiles; whiskers, 1.5× interquartile range; points, outliers.

Fig 3

Transcriptomics integration of liver samples with published cohorts. (a) Principal component analyses of z-scale normalized gene expression levels in cases and controls (PC1/2 on left and PC3/4 on right). This includes 17 controls from study GSE96851 (blue), six controls from this study (green), four HBV-associated acute liver failure cases from study GSE96851 (red), four pediatric acute liver failure cases from Morfopoulou et al. (purple) and four pediatric acute liver failure cases from this study (olive). (b) Single sample gene set enrichment analysis (ssGSEA), comparing GSE96851 cases against the average of GSE96851 controls, and cases from this study and Morfopoulou et al. against our cases. Reactome and blood transcriptional modules (14) were included as pathways. Only pathways with statistically significant enrichment (P-adjusted < 0.05) in at least two cases are shown. Size and color indicate direction and magnitude per pathway and case. (c and d) Boxplots showing leading edge genes expression levels (either normalized expression levels from microarray or log count per million from RNA-Sequencing) from (c) the “enriched in monocytes” pathways and (d) the “complement cascade.” Boxplots center line, median; box limits, upper and lower quartiles; whiskers, 1.5× interquartile range; points, outliers.