Gene regulatory therapeutics for metabolic dysfunction-associated liver disease
Comprehensive profiling of human livers identifies novel therapeutic targets and diagnostic biomarkers for MASLD by modulating transcription factor gene regulatory networks, addressing the unmet need for effective MASLD treatment and prevention.
Patent Information
- Application Number
- PCT/US2026/026969
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2025-05-07
- Filing Date
- 2026-05-06
- Publication Date
- 2026-11-12
AI Technical Summary
There is a critical unmet need for identifying and therapeutically exploiting cell type-specific gene regulatory programs and epigenomic targets that drive metabolic dysfunction-associated steatotic liver disease (MASLD) progression, particularly in hepatocytes, and for methods of diagnosing and treating MASLD based on these targets.
Comprehensive single cell transcriptomic, epigenomic, and spatial profiling of human livers spanning the spectrum of MASLD progression to identify novel therapeutic targets and diagnostic biomarkers, involving the modulation of transcription factor gene regulatory networks (GRNs) such as HNF4A, RORA, CUX2, and others, and administering agents that target hepatocyte cCREs and genetic risk loci.
Enables the identification of novel therapeutic targets and diagnostic biomarkers for MASLD, allowing for effective treatment and prevention strategies by modulating specific GRNs and targeting altered gene activity in hepatocytes.
Smart Images

Figure US2026026969_12112026_PF_FP_ABST
Abstract
Description
Aty. Dkt. No.: 114198-7260 GENE REGULATORY THERAPEUTICS FOR METABOLIC DYSFUNCTION- ASSOCIATED LIVER DISEASE CROSS-REFERENCE TO RELATED APPLICATION
[0001] This application claims the benefit under 35 U.S.C § 119(e) of U.S. Serial No.:63 / 801,527, filed May 7, 2025, the content of which is incorporated herein by reference in its entirety.BACKGROUND
[0002] Metabolic dysfunction-associated steatotic liver disease (MASLD) and alcohol-associated liver disease (MetALD) are the major causes of liver fibrosis in patients with metabolic syndrome, which consists of obesity, insulin resistance, and hyperlipidemia (1, 2). MASLD is a spectrum of liver disease ranging from steatosis (metabolic dysfunction-associated steatotic liver [MASL]) to metabolic dysfunction-associated steatohepatitis ([MASH]), characterized by the development of steatosis, inflammation and fibrosis (2). Excessive alcohol consumption often exacerbates liver injury in MASH patients. (3, 4). Over a third of the world population is affected by MASLD (5, 6), of which an estimated 20% have MASH, and although MASH causes increased morbidity and mortality, there are few therapeutic options (2, 7, 8).
[0003] There remains a critical unmet need for the identification and therapeutic exploitation of cell type-specific gene regulatory programs, transcription factor networks, and epigenomic targets that drive MASLD progression, particularly in hepatocytes, and for methods of diagnosing, stratifying, and treating MASLD based on such targets. This disclosure satisfies this need and provide related advantages as well.SUMMARY OF THE DISCLOSURE
[0004] Studies in humans and experimental models have proposed a two-hit model for MASH and MetALD progression (2, 9). Chronic metabolic injury of hepatocytes causes steatosis, hepatocyte damage, apoptosis (10), and subsequent recruitment of inflammatory cells (11-14), which drive activation of hepatic stellate cells (aHSCs / myofibroblasts) that secrete Collagen Type I and extracellular matrix proteins and make the liver fibrotic (2, 7). The crosstalk between hepatocytes (15, 16), myeloid cells (17, 18), HSCs (19-23), HSCs, and liver sinusoidal endothelial cells (LSECs) plays a critical role in the pathogenesis of MASH (21, 24-26). Single cell transcriptomic studies have provided initial insights into gene expression changes in these cell types during MASLD (16, 27-30). However, characterization of the gene regulatory programs and signaling responses within liver cell types that drive MASH progression, areAty. Dkt. No.: 114198-7260 critical for understanding of how to modulate cellular function to prevent or reverse metabolic liver disease.
[0005] The epigenome describes covalent modifications to DNA and histone proteins throughout chromosomes, which collectively encodes multiple layers of gene regulation that define cell typespecific gene activity, signaling responses, and function (31, 32). Previous single cell chromatin accessibility studies have defined reference maps of cell type-specific candidate cis-regulatory elements (cCREs) in the liver but lack information on functional context of cCRE activity, diversity of cCRE activity across individuals and MASLD states, and annotation of the noncoding genome outside of cCREs defined from accessible chromatin. Recently developed single cell assays enable profiling additional modalities describing the epigenome in specific human cell types (33- 36). In this study Applicant leveraged these advances to comprehensively profile single cell transcription, accessible chromatin regions, two histone modifications associated with active (H3K27ac) and repressive (H3K27me3) states, and 3D chromatin conformation in 86 human livers, along with spatial transcriptomics in seven livers, spanning the spectrum of MASLD progression. Applicant defined cell type-specific gene regulatory programs and chromatin state in the liver, determined transcriptional drivers of progression to MASH and annotated genetic risk loci for MASLD in liver cell types, and characterized heterogeneity including MASH-associated sub-clusters in hepatocytes. Overall, Applicant identified novel genes, pathways and transcriptional networks that could serve as targets for therapeutic intervention in MASH.
[0006] In one aspect, the present disclosure addresses the limitations of the prior art by providing comprehensive single cell transcriptomic, epigenomic, and spatial profiling of human livers spanning the spectrum of MASLD progression, enabling identification of novel therapeutic targets, diagnostic biomarkers, and methods for treating or preventing MASH.
[0007] In another aspect, the disclosure provides a method of treating or preventing MASLD progression in a subject in need thereof, the method comprising, or consisting essentially or, or consisting of administering to the subject an agent that modulates the activity of a transcription factor gene regulatory network (GRN) selected from the group consisting of HNF4A, RORA, CUX2, CREB5, NFKB1, RELB, KLF6, TEAD1, ETS1, NFAT5, FOXO1, NR4A1, FOS, HNF4G, PPARG, STAT1, PPARA, and ATF3, wherein the GRN is associated with MASLD progression in hepatocytes.
[0008] In another aspect, the disclosure provides a method of diagnosing or stratifying MASLD in a subject, the method comprising, or consisting essentially or, or consisting of: obtaining a biological sample from the subject’s liver; determining the activity of one or more cell type-Aty. Dkt. No.: 114198-7260 specific gene regulatory programs in the sample; and classifying the subject into a MASLD stage based on the determined activity.
[0009] In another aspect, the disclosure provides a method of identifying a therapeutic agent for the treatment of MASLD, the method comprising, or consisting essentially or, or consisting of contacting a liver cell with a candidate agent; determining the effect of the candidate agent on the activity of one or more transcription factor gene regulatory networks (GRNs) associated with MASLD progression; and selecting the candidate agent as a therapeutic agent if the candidate agent modulates the activity of the one or more GRNs toward a non-disease state.
[0010] In a further aspect, the disclosure provides compositions comprising agents that target one or more of (a) hepatocyte cCREs with altered activity in MASH, including cCREs enriched for HNF4A, NR4A1, CEBPA / D, FOS / JUN, TEAD, and NFKB transcription factor motifs; (b) genes within transcription factor GRNs associated with MASH progression, including DTNA, AKAP12, ITGA2, LAMA3, MY07B, CD44, TNFAIP2, TMEM150B, and FGF1; and (c) genetic risk loci linked to hepatocyte regulatory programs, including loci at 12q 13 (KRT8 / KRT18), 8p23 (PPP1R3B), 19ql3 (CEBPA), 22ql2 (XBP1), and 2q37 (EFHD1).BRIEF DESCRIPTION OF THE DRAWINGS
[0011] FIGS. 1A- II: Integrated cell type-specific map of gene regulation in the human liver. (FIG. 1A) Study design consisting of profiling normal, MASL, MASH and MetALD livers from 86 human donors. Within MASH and MetALD livers were further classified by low fibrosis (Fib-) and high fibrosis (Fib+). Each donor was assayed using single cell multiome (paired ATAC-seq and RNA-seq), Paired tag for H3K27ac and H3K27me3 marks (paired Cut&Tag and RNA-seq), droplet Hi-C and genome-wide genotyping. A subset of donors was also assayed using spatial transcriptomics (Visium) (FIG. IB) UMAP plot showing clustering of 678,748 nuclei from multiome, Paired-Tag and droplet Hi-C. Clusters are labeled based on cell type identity which is defined using gene expression of key marker genes. (FIG. 1C) Activity of selected cell type marker genes across liver cell types in gene expression, ATAC-seq, H3K27ac and H3K27me3 profiles. (FIG. ID) Genome browser locus plots of selected cell type marker genes across liver cell types in ATAC-seq, H3K27ac and H3K27me3 profiles. (FIG. IE) Chromatin contacts from Hi-C at the ADAMTS2 locus in HSCs (blue) and endothelial (red) cells. Bottom shows the ATAC-seq, H3K27ac and H3K27me3 profiles in HSCs and endothelial cells for the same genomic region. (FIG. IF) Liver cell type proportions (top), number of cells from each assay (middle), and MASLD phenotypes (bottom) across all 86 donors profiled in the study. (FIG. 1G) Proportions of each liver cell type across fibrosis stages (top) and MASLDAty. Dkt. No.: 114198-7260 state (bottom). (FIG. 1H) Cell type annotations in spatial imaging of representative normal, MASL, Fib- MASH and Fib+ MASH livers. Applicant note that the colors denoting each cell type differ from the rest of the figure for the purposes of visualization of the spatial profiles. (FIG. II) Expression of selected marker genes of liver cell types in spatial profiles.
[0012] FIGS. 2A - 2H: Cell type-specific gene regulatory programs in the human liver. (FIG. 2A) Percentage of liver cell type cCREs in this study identified in the liver in ENCODE and HEA (top), and distribution of liver cell type cCREs across various genomic features (FIG.2B) Emission probabilities of ATAC-seq, H3K27ac and H3K27me3 for each chromatin state (active, open, weak, repressive, and unknown) and proportion of the genome annotated with each state in each cell type (top), and distribution of chromatin states in hepatocytes genome-wide (bottom left) and at cCREs (bottom right). (FIG. 2C) Examples of gene regulatory networks (GRNs) with hepatocyte-specific and myeloid-specific cCRE-target gene links and active chromatin state at the CRP and CD163 loci, respectively. Chromatin contacts in hepatocytes and myeloid cells support cell type-specific activity at each locus. Hepatocyte cCRE-target gene links from GRNs show enrichment of chromatin contacts compared to proximal regions, while myeloid target genes links show no enrichment. (FIG. 2D) Clustering of liver cell type cCREs into modules (MO-8) based on accessibility profiles across cell types (left) and chromatin state of cCREs in each module across cell types (right). (FIG. 2E) Enrichment of chronic alanine aminotransferase (cALT) level associated variants in cCREs and H3K27ac sites in each liver cell type (left), enrichment of cALT associated variants in cCREs annotated in each chromatin state in hepatocytes and cholangiocyte (top right), and number of fine-mapped cALT loci and variants overlapping hepatocyte cCREs and active chromatin states. Colors represent the posterior probability (PPA) of variants. (FIG. 2F) Liver cell type cCREs with accessibility profiles highly specific to a cell type. Plotted values represent log scaled counts-per-million (CPM) in the cCRE. (FIG. 2G) Transcription factor sequence motifs enriched in cCREs with cell typespecific accessibility for each cell type. (FIG. 2H) Expression level of genes in each cell type linked to cCREs with cell type-specific accessibility (left), and biological pathways enriched in genes linked to specific cCREs in each cell type (right).
[0013] FIGS. 3A - 3J: Cell type-specific genomic and epigenomic changes in MASLD. (FIG. 3A) Number of genomic features in each cell type for each modality with significant changes in activity in different MASLD stages compared to normal livers. (FIG. 3B) Pearson correlation in fold-change in hepatocyte activity in high fibrosis (Fib+) MASH across modalities (left), and genome browser plot of the DTNA locus where, in Fib+ MASH, several cCREs had increased accessibility and H3K27ac signal, DTNA had increased expression, and the region hadAty. Dkt. No.: 114198-7260 broadly reduced H3K27me3 signal. (FIG. 3C) Biological pathways in hepatocytes with altered expression in MASLD stages compared to normal livers. Values shown in the heatmap are normalized enrichment scores (NES) and cells contain a * if the NES were significant at FDR<.10 (top), and correlation in the fold-change of gene expression levels in hepatocytes in MASL compared to Fib- and Fib+ MASH (bottom). (FIG. 3D) Clusters of hepatocyte cCREs with significant change in Fib+ MASH across different MASLD stages in ATAC-seq, H3K27ac and H3K27me3 profiles (top), and transcription factor sequence motifs enrichments for cCREs in each cluster (bottom). (FIG. 3E) Chromatin state in each MASLD stage for hepatocyte cCREs with decreased activity in MASH (top) and increased activity in MASL / MASH (bottom).(FIG. 3F) Genome browser tracks showing example loci with altered activity in hepatocytes in MASH. A hepatocyte cCRE in the GOLGA1 gene (highlighted in black rectangle) with decreased activity in higher fibrosis and in an active hepatocyte chromatin state shows physical interactions via chromatin looping with the promoter region of the SC Al gene, which has decreased expression higher fibrosis (top); A hepatocyte cCRE in the SNX27 gene (highlighted in black rectangle) with decreased activity in higher fibrosis and in an active hepatocyte chromatin state shows physical interactions via chromatin looping with the promoter region of the RORC gene, which has decreased expression higher fibrosis (bottom). (FIG. 3G) Transcription factor gene regulatory networks (GRNs) in hepatocytes with increased or decreased activity in MASLD stages compared to normal livers. GRNs are grouped (GRN cl-6) based on change in activity across MASLD stages. Cells with significant (FDR< 10) change in GRN activity are highlighted with * (left), and biological pathways significantly (FDR< 10) enriched among genes in each GRN (right). (FIG. 3H) Connectivity of GRNs with altered activity in Fib- MASH, Fib+ MASH and MASL. Transcription factor genes as well as the genes in GRNs for each factor are colored based on the change in expression in the specified MASLD stage. (FIG. 31) Fold-enrichment of fine-mapped MASLD (cATL) associated variants for hepatocyte GRNs in each GRN group compared to background cCREs (top), and connectivity of hepatocyte GRNs based on fine-mapped MASLD variants overlapping a cCRE in the GRN. Colors are based on the clustering of GRN connectivity using igraph. *FDR<.10 (FIG. 3 J) Multiple clusters of TF GRNs altered in MASL and Fib- MASH based on overlapping MASLD variants.
[0014] FIGS. 4A - 4G: MASLD risk loci affect hepatocyte gene regulatory programs.(FIG. 4A) Number of features for each modality with a significant (FDR< 10) QTL in each liver cell type. (FIG. 4B) Transcription factor sequence motifs enriched in chromatin and H3K27ac QTLs for each cell type. *FDR< 10. (FIG. 4C) Proportion of QTLs in each cell type for eachAty. Dkt. No.: 114198-7260 modality which are colocalized with QTLs for another modality or not colocalized with any other QTLs (top), and proportion of QTLs in each modality that are shared across cell types or show cell type specificity. (FIG. 4D) Number of MASLD-associated loci that are significant QTLs for each modality and colocalize with QTLs for each modality in hepatocytes. (FIG. 4E) Annotation of specific MASLD loci with QTL association including loci with a fine-mapped variant overlapping a cCRE or active chromatin state, a fine-mapped variant with predicted effects on hepatocyte chromatin with chromBPnet, and a link to a predicted target gene. (FIG.4F) MASLD-associated variants at the 12q 13 locus affect activity of a cCRE in hepatocytes, where the risk allele A has increased chromatin accessibility and H3K27ac activity (top left) and has higher predicted accessibility and maps in a sequence motif for HNF1 A (bottom left). The MASLD association signal was strongly colocalized with chromatin accessibility and H3K27ac QTL association for the cCRE (highlighted in grey) in hepatocytes, and the cCRE reg33ion showed physical interactions in hepatocytes with the KRT8 and KRT18 promoter regions (right).(FIG. 4G) MASLD-associated variants at the 8p23 locus affect activity of a cCRE in hepatocytes, where the risk allele A has decreased chromatin accessibility and H3K27ac activity (top left) and has lower predicted accessibility and maps in a sequence motif for HNF1 A (bottom left). The MASLD association signal was strongly colocalized with chromatin accessibility QTL association for the cCRE (highlighted in grey) in hepatocytes, and the cCRE region showed physical interactions in hepatocytes with the PPP1R3B promoter regions (right).
[0015] FIGS. 5 A - 5M: Heterogeneity in hepatocytes and changes in MASLD. (FIG. 5 A) Subclusters of hepatocytes identified in single cell multi ome and Paired-Tag data including zone 1 (peri-portal), zone 2 (mid-lobular), zone 3 (peri-central) populations, a MASH-associated (MASH-a) population, and a Fibrosis-associated (Fib-a) population. (FIG. 5B) Genes with specific expression in each hepatocyte sub-cluster (left) and biological pathways enriched in subcluster-specific genes (right). (FIG. 5C) Annotation of hepatocyte sub-clusters in spatial transcriptome profiles from representative examples of normal, MASL, Fib- MASH and Fib+ MASH livers. (FIG. 5D) Proportions of cells in hepatocyte subclusters across different MASLD states and fibrosis score. (FIG. 5E) Hepatocyte cCREs with activity specific to each sub-cluster in accessible chromatin, H3K27ac and H3K27me3 profiles (left) and TF sequence motifs enriched in subcluster-specific cCREs (right). (FIG. 5F) TF GRNs with increased activity in each hepatocyte sub-cluster. (FIG. 5G) Features for each modality with changes in activity in MASLD states within each hepatocyte sub-cluster (top) and the proportion of bases within H3K27me3 bins altered in injured hepatocytes in Fib+ MASH in each chromatin state (bottom).(FIG. 5H) Biological pathways enriched in H3K27me3 bins with increased and decreasedAty. Dkt. No.: 114198-7260 activity in Fib-a hepatocytes in Fib+ MASH (top), and enrichment of cell structure / interaction pathways in Fib+ MASH in each hepatocyte sub-cluster (bottom). (FIG. 51) Sequence motifs enriched in cCREs with increased activity in Fib+ MASH in each hepatocyte sub-cluster. (FIG.5 J) Pseudo-time ordering of hepatocytes from the Fib-a sub-cluster (left), where Zone 3 hepatocytes are ordered closest to Fib-a cells compared to other hepatocyte zones (top right). Expression of the top genes up-regulated in Fib+ MASH in hepatocytes and proportion of each hepatocyte subcluster across pseudo time bins (bottom right). (FIG. 5K) Cells in spatial transcriptomics of a Fib+ MASH donor colored by the joint expression of the top genes up-regulated in Fib+ MASH. (FIG. 5L) Cellular neighborhood analysis showing enrichment of each liver cell type and hepatocyte sub-cluster for proximity to other cell types. (FIG. 5M) Strength of overall interactions between pairs of liver cell types and hepatocyte sub-clusters based on predicted cell-cell signaling in Fib+ MASH (left), strength of interaction between ligand expressed in HSCs (COL1 Al) and specific receptors expressed in liver cell types and subclusters in Fib+ MASH (right).
[0016] FIGS. 6A- 6C: Study design for single cell profiling of human livers. (FIG. 6A) Design of study to perform single cell assays of human livers in sample pools. Livers from normal and MASLD groups (MASL, MASH, MetALD) were randomly assigned to groups. Samples from livers in each group were pooled, and then nuclei were extracted from each pool. Multiple single cell libraries including lOx multi ome, H3K27ac and H3K27me3 Paired-Tag and Droplet Hi-C were generated on nuclei from each pool, and libraries were sequenced. Finally, the donor of origin was assigned to barcodes in each pool by demultiplexing by using genotyping array data for each donor. (FIG. 6B) Number of singlet and doublet barcodes obtained from demultiplexing of libraries for each single cell assay. (FIG. 6C) Proportion of cells from each liver donor obtained from libraries for each single cell assay.
[0017] FIGS. 7A - 7P: Characteristics of single cell profiles in liver. (FIG. 7A) Number of UMIs (left) and genes (right) per barcode for the RNA modalities for lOx Multi ome and Droplet Paired-Tag as counted by Seurat. (FIG. 7B) Fragment length (right) and fraction of reads in peaks (left) for AT AC and DNA modalities of Droplet Paired-Tag. (FIG. 7C) Enrichment of normalized ATAC, H3K27ac, and H3K27me3 signal around TSS sites (+-2kb) per single cell library. (FIG. 7D) Unique read pairs in Droplet Hi-C including cis long and trans contacts. (FIG. 7E) 10X Multi ome UMAP and cell type annotations. (FIG. 7F) H3K27ac Droplet Paired-Tag UMAP and cell type annotations. (FIG. 7G) H3K27me3 Droplet Paired-Tag UMAP and cell type annotations. (FIG. 7H) Droplet Hi-C UMAP and cell type annotations. (FIG. 71) Label transfer scored for Droplet Paired-Tag and Hi-C by cell type. (FIG. 7 J) Proportion ofAty. Dkt. No.: 114198-7260 disease, gender, and modality composition across cell types, visualized in UMAP space in (FIGS. 7K - 7N). (FIG. 70) Number of detected genes per donor after filtering in spatial transcriptomics profiling of human livers. (FIG. 7P) Number of counts per donor after filtering in spatial transcriptomics profiling of human livers.
[0018] FIGS. 8A - 8B: Identifying sub-types of liver cell types in single cell profiles. (FIG.8A) UMAP of combined 10X Multi ome, Droplet Paired-Tag and Droplet HiC data showing cell sub-type definitions in the liver. (FIG. 8B) Expression of selected marker genes for hepatocyte sub-types.
[0019] FIG. 9: Histological, immunohistochemical, and spatial transcriptomic profiling across MASLD disease stages. Liver sections from seven donors across MASLD, including normal, MASL, MASH (Fibrosis score 0-4), and MetALD — were profiled using histology (H&E), immunohistochemistry (HAL and GAS staining), and lOx Visium HD spatial transcriptomics. Each row corresponds to an individual liver sample (donor ID indicated at left with fibrosis stage noted). Left panels show H&E staining with annotated central vein (CV) and portal vein (PV) regions. Middle panels display HAL and GAS staining marking periportal and pericentral regions, respectively, and highlighting histological variation across conditions.Spatial transcriptomic maps (fourth column) show the spatial distribution of six major cell types (hepatocytes, cholangiocytes, endothelial cells, HSCs, myeloid cells, and B cells), inferred via Tangram mapping. Rightmost panels show violin plots of RNA counts per cell type across each sample. Notably, the severe fibrosis MASH sample (top row) shows architectural disruption and altered spatial composition compared to normal and MASL livers.
[0020] FIGS. 10A- 10E: Characteristics of liver cell type cCREs. (FIG. 10A) Variability (coefficient of variation) in chromatin accessibility signal (CPM) of novel cCREs identified in this study compared to known cCREs. (FIG. 10B) Distribution of known and novel cREs across genomic features. (FIG. 10C) Overlap of chromatin state annotations in hepatocytes and HSCs to chromatin states in bulk liver from the Epigenome Roadmap. Parenthesis show the percentage of each cell type among all cells in the single cell maps. (FIG. 10D) Proportion of all cCREs classified in each chromatin state for each liver cell type. (FIG. 10E) TF sequence motifs enriched in each chromatin state for each liver cell type using Homer.
[0021] FIGS. HA - 11B: Enrichment of MASLD variants for cell type-specific cCREs. Enrichment of fine-mapped variants for cALT / MASLD proxy GWAS in (FIG. 11 A) liver cell type-specific cREs and (FIG. 11B) Liver cell type cCREs within cCRE modules (MO-9).Aty. Dkt. No.: 114198-7260
[0022] FIGS. 12A - 12C: MASLD variants annotated in liver cell type cCREs. (FIGS. 12A-12B) Number of fine-mapped variants (FIG. 12A) and credible sets (FIG. 12B) cALT / MASLD proxy GWAS variants intersecting liver cell type cREs and active chromatin state regions, stratified by variant posterior inclusion probability (PIP) (FIG. 12C) Annotation of specific fine-mapped signals to liver cell types by intersection with cREs, colored by probability of variants annotated.
[0023] FIGS. 13A - 13B: Changes in gene expression in hepatocytes in MASLD. (FIG. 13A) Differentially expressed genes from DESeq using the 10X Multiome data for listed contrasts. Red dots highlight upregulated genes in disease, blue dots highlight down regulated genes, and grey represents genes not significant at 10% FDR. (FIG. 13B) Pathway enrichments for differentially expressed genes using fGSEA. The top 5 most significantly enriched and depleted pathways are plotted.
[0024] FIG. 14: Change in collagen family gene expression in MASLD progression. Log-transformed average expression of genes in the collagen family in liver cell types across MASLD stages.
[0025] FIGS. 15A- 15C: Changes in hepatocyte cCRE activity in MASLD. (FIG. 15 A) Differentially accessible cCREs (ATAC) across fibrosis stages clustered based on the epigenomic signal at different fibrosis stages. ATAC signal in navy, H3K27ac signal in blue, and H3K27me3 signal in brown is shown for each site. (FIG. 15B) Clustering of differentially accessible cCREs in MetALD versus Control and MASH versus Control, as in A. (FIG. 15C) Overlap of cCREs changing in MASH and MetALD.
[0026] FIGS. 16A- 16C: Changes in hepatocyte chromatin state in progression to MASH. (FIG. 16A) Chromatin state changes across fibrosis stage within clusters of hepatocyte cCREs associated with fibrosis stage (left) and average chromatin accessibility (CPM) across fibrosis stages for the same cCREs (right). (FIG. 16B) Chromatin state in normal and MASL livers across all hepatocyte cCREs (left) and only hepatocyte cCREs that change state between normal and MASL livers (right). (FIG. 16C) Sequence motif enrichment in cCREs that transition from open state in normal to active state in MASL, or from active state in normal to open state in MASL.
[0027] FIGS. 17A- 17C: Enrichment of MASLD variants in TF GRNs (FIGS. 17A-17B) Number of credible sets overlapping GRN cCREs and cumulative PIP of variants overlapping GRN cCREs. (FIG. 7C) Break down of variants annotating to GRN cCREs and the direction of accessibility relative to the risk allele in QTL signals. Red is increased accessibility with the riskAty. Dkt. No.: 114198-7260 allele, blue is decreased accessibility with the risk allele, and grey is no QTL association. GRNs are colored by cluster.
[0028] FIGS. 18A - 18D: Fine-mapping of liver cell type QTLs. (FIG. 18A) Summary of the number of fine-mapped QTL signals (credible sets) for each modality and the number of credible set variants per signal. (FIGS. 18B-18D) Locus zoom plot for the MASLD proxy GWAS signal (top), variant pips for GWAS fine-mapping (upper middle), caQTL locus zoom plot (lower middle), and variant pips from caQTL fine mapping (bottom) for the PPP1R3B (B), KRT8 (FIG.18C), and EFHD1 (FIG. 18D) loci.
[0029] FIGS. 19A - 19E: MASLD risk variants affect EFHD1 expression (FIG. 19A) eQTL (top), caQTL (middle) and H3K27ac hQTL (bottom) boxplots and mashr effect estimates for the EFHD1 locus. The vertical line represents the lead fine-mapped MASLD proxy SNP. (FIG. 19B) Mashr estimates for the lead fine-mapped MASLD variant including the estimated effect and standard error, where asterisks for local false sign rate (LFSR) p-values are:****0.0001, ***0.001, ** 0.01, and *0.05. (FIG. 19C) Locus zoom plots for MASLD proxy GWAS (top), EFHD1 eQTL (upper middle), caQTL (lower middle), and H3K27ac hQTL (bottom). (FIG. 19D, FIG. 19E) Interaction plot of effect between EFHD1 expression and fibrosis stage (FIG. 19E, interaction term p-value = 0.046) and MASLD stage.
[0030] FIGS. 20A - 2C: MASLD risk variants interact with distal target gene CEBPA at the 19ql3 locus. (FIG. 20A) Browser track of the 19q 13 locus showing ATAC, H3K27ac, and H3K27me3 signal in hepatocytes and HSCs, and chromatin states and chromatin loops in hepatocytes. The vertical black bar marks the MASLD-associated caQTL peak which interacts with CEBPA. (FIG. 20B) CEBPA eQTL boxplot (top) and caQTL (middle) and H3K27ac hQTL (bottom) for the linked cRE. The bottom plots show eQTL boxplots for PEPD (top) and CEBPG (bottom), the two other genes at the locus. Neither PEPD or CEBPG had significant eQTL association. (FIG. 20C) mashr estimates for the caQTL at this locus. Asterisks for local false sign rate (LFSR) p-values are: ****0.0001, ***0.001, **0.01, and *0.05.DETAILED DESCRIPTIONDefinitions
[0031] As it would be understood, the section or subsection headings as used herein is for organizational purposes only and are not to be construed as limiting and / or separating the subject matter described.Aty. Dkt. No.: 114198-7260
[0032] Unless defined otherwise, all technical and scientific terms used herein have the same meanings as commonly understood by one of ordinary skill in the art to which this disclosure belongs. Although any methods and materials similar or equivalent to those described herein can be used in the practice or testing of the present disclosure, the preferred methods, devices, and materials are now described. All technical and patent publications cited herein are incorporated herein by reference in their entirety. Nothing herein is to be construed as an admission that the disclosure is not entitled to antedate such disclosure by virtue of prior disclosure.
[0033] The practice of the present disclosure will employ, unless otherwise indicated, conventional techniques of tissue culture, immunology, molecular biology, microbiology, cell biology and recombinant DNA, which are within the skill of the art. See, e.g., Sambrook and Russell eds. (2001) Molecular Cloning: A Laboratory Manual, 3rd edition; the series Ausubel et al. eds. (2007) Current Protocols in Molecular Biology; the series Methods in Enzymology (Academic Press, Inc., N.Y.); MacPherson et al. (1991) PCR 1: A Practical Approach (IRL Press at Oxford University Press); MacPherson et al. (1995) PCR 2: A Practical Approach; Harlow and Lane eds. (1999) Antibodies, A Laboratory Manual; Freshney (2005) Culture of Animal Cells: A Manual of Basic Technique, 5th edition; Gait ed. (1984) Oligonucleotide Synthesis; U.S. Patent No. 4,683,195; Hames and Higgins eds. (1984) Nucleic Acid Hybridization; Anderson (1999) Nucleic Acid Hybridization; Hames and Higgins eds. (1984) Transcription and Translation; Immobilized Cells and Enzymes (IRL Press (1986)); Perbal (1984) A Practical Guide to Molecular Cloning; Miller and Calos eds. (1987) Gene Transfer Vectors for Mammalian Cells (Cold Spring Harbor Laboratory); Makrides ed. (2003) Gene Transfer and Expression in Mammalian Cells; Mayer and Walker eds. (1987) Immunochemical Methods in Cell and Molecular Biology (Academic Press, London); Herzenberg et al. eds (1996) Weir’s Handbook of Experimental Immunology; Manipulating the Mouse Embryo: A Laboratory Manual, 3rd edition (Cold Spring Harbor Laboratory Press (2002)); Sohail (ed.) (2004) Gene Silencing by RNA Interference: Technology and Application (CRC Press).
[0034] As used in the specification and claims, the singular form “a,” “an” and “the” include plural references unless the context clearly dictates otherwise. For example, the term “a cell” includes a plurality of cells, including mixtures thereof.
[0035] As used herein, the term “comprising” is intended to mean that the compounds, agents, compositions and methods include the recited elements, but not exclude others. “Consisting essentially of’ when used to define compounds, agents, compositions and methods, shall mean excluding other elements of any essential significance to the combination. Thus, a composition consisting essentially of the elements as defined herein would not exclude trace contaminants,Aty. Dkt. No.: 114198-7260 e.g., from the isolation and purification method and pharmaceutically acceptable carriers, preservatives, and the like. “Consisting of’ shall mean excluding more than trace elements of other ingredients. Embodiments defined by each of these transition terms are within the scope of this technology.
[0036] All numerical designations, e.g., pH, temperature, time, concentration, and molecular weight, including ranges, are approximations which are varied (+) or (-) by increments of 1, 5, or 10%. It is to be understood, although not always explicitly stated that all numerical designations are preceded by the term “about.” It also is to be understood, although not always explicitly stated, that the reagents described herein are merely exemplary and that equivalents of such are known in the art.
[0037] The term “about,” as used herein when referring to a measurable value such as an amount or concentration and the like, is meant to encompass variations of 20%, 10%, 5%, 1 %, 0.5%, or even 0.1 % of the specified amount.
[0038] As used herein, comparative terms as used herein, such as high, low, increase, decrease, reduce, or any grammatical variation thereof, can refer to certain variation from the reference. In some embodiments, such variation can refer to about 10%, or about 20%, or about 30%, or about 40%, or about 50%, or about 60%, or about 70%, or about 80%, or about 90%, or about 1 fold, or about 2 folds, or about 3 folds, or about 4 folds, or about 5 folds, or about 6 folds, or about 7 folds, or about 8 folds, or about 9 folds, or about 10 folds, or about 20 folds, or about 30 folds, or about 40 folds, or about 50 folds, or about 60 folds, or about 70 folds, or about 80 folds, or about 90 folds, or about 100 folds or more higher than the reference. In some embodiments, such variation can refer to about 1%, or about 2%, or about 3%, or about 4%, or about 5%, or about 6%, or about 7%, or about 8%, or about 0%, or about 10%, or about 20%, or about 30%, or about 40%, or about 50%, or about 60%, or about 70%, or about 75%, or about 80%, or about 85%, or about 90%, or about 95%, or about 96%, or about 97%, or about 98%, or about 99% of the reference.
[0039] As will be understood by one skilled in the art, for any and all purposes, all ranges disclosed herein also encompass any and all possible subranges and combinations of subranges thereof. Furthermore, as will be understood by one skilled in the art, a range includes each individual member.
[0040] “Optional” or “optionally” means that the subsequently described circumstance may or may not occur, so that the description includes instances where the circumstance occurs and instances where it does not.Aty. Dkt. No.: 114198-7260
[0041] As used herein, “and / or” refers to and encompasses any and all possible combinations of one or more of the associated listed items, as well as the lack of combinations when interpreted in the alternative (“or”).
[0042] “Substantially” or “essentially” means nearly totally or completely, for instance, 95% or greater of some given quantity. In some embodiments, “substantially” or “essentially” means 95%, 96%, 97%, 98%, 99%, 99.5%, or 99.9%.
[0043] The terms or “acceptable,” “effective,” or “sufficient” when used to describe the selection of any components, ranges, dose forms, etc. disclosed herein intend that said component, range, dose form, etc. is suitable for the disclosed purpose.
[0044] “Effective amount” of an agent refers to an amount of the agent that is an amount sufficient to obtain a pharmacological response; or alternatively, is an amount of the agent that, when administered to a patient with a specified disorder or disease, is sufficient to have the intended effect, e.g., treatment, alleviation, amelioration, palliation or elimination of one or more manifestations of the specified disorder or disease in the patient. An effect does not necessarily occur by administration of one dose, and may occur only after administration of a series of doses. Thus, an effective amount may be administered in one or more administrations.
[0045] As used herein, the phrase “derived from” means isolated from, purified from, or engineered from, or any combination thereof.
[0046] As used herein, “treating” or “treatment” of a disease in a subject refers to (1) preventing the symptoms or disease from occurring in a subject that is predisposed or does not yet display symptoms of the disease; (2) inhibiting the disease or arresting its development; or (3) ameliorating or causing regression of the disease or the symptoms of the disease. As understood in the art, “treatment” is an approach for obtaining beneficial or desired results, including clinical results. For the purposes of the present technology, beneficial or desired results can include one or more, but are not limited to, alleviation or amelioration of one or more symptoms, diminishment of extent of a condition (including a disease), stabilized (i.e., not worsening) state of a condition (including disease), delay or slowing of condition (including disease), progression, amelioration or palliation of the condition (including disease), states and remission (whether partial or total), whether detectable or undetectable. In one embodiment, treatment excludes prevention or prophylaxis.
[0047] As used herein, the term “subject” is used interchangeably with “patient,” and indicates a mammal, in particular a human, equine, bovine, porcine, feline, canine, murine, rat, or nonhuman primate. In some embodiments, the subject is a human.Aty. Dkt. No.: 114198-7260
[0048] The term “contacting” means direct or indirect binding or interaction between two or more. A particular example of direct interaction is binding. A particular example of an indirect interaction is where one entity acts upon an intermediary molecule, which in turn acts upon the second referenced entity. Contacting as used herein includes in solution, in solid phase, in vitro, ex vivo, in a cell and in vivo. Contacting in vivo can be referred to as administering, or administration.
[0049] “Administration” or “delivery” of a composition a therapeutic agent (e.g., an AAV vector) can be performed in one dose, continuously or intermittently throughout the course of treatment. Methods of determining the most effective means and dosage of administration are known to those of skill in the art and will vary with the composition used for therapy, the purpose of the therapy, the target cell being treated, and the subject being treated. Single or multiple administrations can be carried out with the dose level and pattern being selected by the treating physician or in the case of animals, by the treating veterinarian. Suitable dosage formulations and methods of administering the agents are known in the art. Route of administration can also be determined and method of determining the most effective route of administration are known to those of skill in the art and will vary with the composition used for treatment, the purpose of the treatment, the health condition or disease stage of the subject being treated, and target cell or tissue. Non-limiting examples of route of administration include oral administration, intraperitoneal, infusion, nasal administration, inhalation, injection, and topical application. In some embodiments, the administration is administration to a tumor microenvironment. In some embodiments, administering or a grammatical variation thereof also refers to more than one doses with certain interval. In some embodiments, the interval is 1 day, 2 days, 3 days, 4 days, 5 days, 6 days, 1 week, 10 days, 2 weeks, 3 weeks, 1 month, 2 months, 3 months, 4 months, 5 months, 6 months, 1 year or longer. In some embodiments, one dose is repeated for once, twice, three times, four times, five times, six times, seven times, eight times, nine times, ten times or more. Depending on the use, administration can be in vitro or in vivo.
[0050] The term administration shall include without limitation, administration by oral, parenteral (e.g., intramuscular, intraperitoneal, intravenous, intravascular, intraperitoneal, intracerebroventricular (ICV), intrathecal, intracistemal injection or infusion, intracranial, ocular, intradermally, percutaneously, subcutaneous injection, or implant), intratum orally, by inhalation spray nasal, intratracheal, vaginal, rectal, sublingual, urethral (e.g., urethral suppository) or topical routes of administration (e.g., gel, ointment, cream, aerosol, etc.) and can be formulated, alone or together, in suitable dosage unit formulations containing conventional non-toxic pharmaceutically acceptable carriers, adjuvants, excipients, and vehicles appropriate for eachAty. Dkt. No.: 114198-7260 route of administration. The disclosure is not limited by the route of administration, the formulation or dosing schedule.
[0051] An agent of the present disclosure can be administered for therapy by any suitable route of administration. It will also be appreciated that the optimal route will vary with the condition and age of the recipient, and the disease being treated.
[0052] Administration or treatment in “combination” refers to administering two agents such that their pharmacological effects are manifest at the same time. Combination does not require administration at the same time or substantially the same time, although combination can include such administrations.
[0053] The terms "oligonucleotide" or "polynucleotide" or "portion," or "segment" thereof refer to a stretch of polynucleotide residues which is long enough to use in PCR or various hybridization procedures to identify or amplify identical or related parts of mRNA or DNA molecules. The polynucleotide compositions of this invention include RNA, cDNA, genomic DNA, synthetic forms, and mixed polymers, both sense and antisense strands, and may be chemically or biochemically modified or may contain non-natural or derivatized nucleotide bases, as will be readily appreciated by those skilled in the art. Such modifications include, for example, labels, methylation, substitution of one or more of the naturally occurring nucleotides with an analog, intemucleotide modifications such as uncharged linkages (e.g., methyl phosphonates, phosphotriesters, phosphoamidates, carbamates, etc.), charged linkages (e.g., phosphorothioates, phosphorodithioates, etc.), pendent moieties (e.g., polypeptides), intercalators (e.g., acridine, psoralen, etc.), chelators, alkylators, and modified linkages (e.g., alpha anomeric nucleic acids, etc.). Also included are synthetic molecules that mimic polynucleotides in their ability to bind to a designated sequence via hydrogen bonding and other chemical interactions. Such molecules are known in the art and include, for example, those in which peptide linkages substitute for phosphate linkages in the backbone of the molecule.
[0054] As used herein, the term “purified” does not require absolute purity; rather, it is intended as a relative term. Thus, for example, a purified nucleic acid, peptide, protein, biological complexes, cell, virus or other active compound is one that is isolated in whole or in part from proteins or other contaminants. Generally, substantially purified peptides, proteins, biological complexes, cell, virus or other active compounds for use within the disclosure comprise more than 80% of all macromolecular species present in a preparation prior to admixture or formulation of the peptide, protein, biological complex, cell, virus or other active compound with a pharmaceutical carrier, excipient, buffer, absorption enhancing agent, stabilizer, preservative,Aty. Dkt. No.: 114198-7260 adjuvant or other co-ingredient in a complete pharmaceutical formulation for therapeutic administration. More typically, the peptide, protein, biological complex, cell, virus or other active compound is purified to represent greater than 90%, often greater than 95% of all macromolecular species present in a purified preparation prior to admixture with other formulation ingredients. In other cases, the purified preparation may be essentially homogeneous, wherein other macromolecular species are not detectable by conventional techniques.
[0055] In some embodiments, the term “non-natural,” “engineered” or “recombinant” refers to having at least one modification not normally found in a naturally occurring protein, polypeptide, polynucleotide, strain, wild-type strain or the parental host strain of the referenced species. In some embodiments, the term “non-natural,” “engineered” or “recombinant” refers to being synthetized by human intervention.
[0056] As used herein, “MASLD” or “metabolic dysfunction-associated steatotic liver disease” refers to the spectrum of liver disease ranging from simple steatosis (MASL) through metabolic dysfunction-associated steatohepatitis (MASH), characterized by the development of steatosis, inflammation, and fibrosis in the setting of metabolic syndrome.
[0057] As used herein, “MASH” or “metabolic dysfunction-associated steatohepatitis” refers to the progressive form of MASLD characterized by hepatocyte injury, inflammation, and fibrosis. MASH includes low-fibrosis MASH (Fib- MASH; fibrosis stages 0-2) and high-fibrosis MASH (Fib+ MASH; fibrosis stages 3-4).
[0058] As used herein, “candidate cis-regulatory element” or “cCRE” refers to a region of noncoding DNA that may regulate the transcription of nearby or distal genes. cCREs are identified based on chromatin accessibility (e.g., ATAC-seq profiles) and may be further annotated based on chromatin state (e.g., active, open, weak, repressive, or unknown) using additional epigenomic modalities including histone modification profiles (H3K27ac, H3K27me3). A cCRE may be “cell type-specific” when its accessibility profile is significantly enriched in one cell type relative to others as determined by specificity scoring methods.
[0059] As used herein, “chromatin state” refers to a functional classification of a genomic region based on the combined presence or absence of epigenomic features. Chromatin states include: (a) “active” state, characterized by positive chromatin accessibility (ATAC-seq) signal and positive H3K27ac signal; (b) “open” state, characterized by positive chromatin accessibility (ATAC-seq) signal without significant H3K27ac signal; (c) “weak” state, characterized by low-level epigenomic signals; (d) “repressive” state, characterized by positive H3K27me3 signal; and (e)Atty. Dkt. No.: 114198-7260 “unknown” state, where no significant signal is detected. Chromatin states may be identified using methods including chromHMM or equivalent chromatin state annotation algorithms.
[0060] As used herein, “gene regulatory network” or “GRN” refers to a set of candidate cis-regulatory elements (cCREs) associated with a transcription factor (TF), together with the target genes whose expression is regulated by the TF through said cCREs. A GRN is characterized by the TF whose binding motif is enriched in the associated cCREs and the target genes linked to those cCREs through computational analysis integrating chromatin accessibility and gene expression data. GRNs may be identified using methods including but not limited to SCENIC + or equivalent computational frameworks for inferring TF-cCRE-gene regulatory relationships from single-cell multi-omics data. The “activity” of a GRN refers to the aggregate expression level of the target genes within the network, as determined by enrichment analysis of GRN target genes among differentially expressed genes.
[0061] As used herein, “hepatocyte sub-cluster” refers to a distinct subpopulation of hepatocytes identified by single-cell profiling, including: (a) Zone 1 (peri-portal) hepatocytes characterized by expression of G6PC, GLS2, and PCK1 and localization around the portal vein; (b) Zone 2 (mid-lobular) hepatocytes; (c) Zone 3 (peri-central) hepatocytes characterized by expression of CYP2E1 and ADH4 and localization around the central vein; (d) MASH-associated (MASH-a) hepatocytes preferentially expressing PROXI, XBP1, and CBP2; and (e) fibrosis-associated (Fib-a) hepatocytes preferentially expressing NFKB1, STAT4, HIF1A, SOD2, TEAD1, and TGFBR2.
[0062] As used herein, “fibrosis-associated hepatocyte” or “Fib-a hepatocyte” refers to a subpopulation of hepatocytes that preferentially expresses inflammatory and stress-related genes including one or more of NFKB1, STAT4, HIF1 A, SOD2, TEAD1, and TGFBR2, and is present in increased proportions in subjects with high-fibrosis MASH (Fib+ MASH) and severe fibrosis compared to normal livers. The term “fibrosis-adjacent hepatocyte” or “Fib-adj hepatocyte” refers to the subset of Fib-a hepatocytes that are spatially localized adjacent to sites of fibrotic injury and express genes most upregulated in Fib+ MASH including CXCL8, CXCL6, HSPB8, ITGB6, CDKN2A, andKRT23.
[0063] As used herein, “quantitative trait locus” or “QTL” refers to a genomic locus where genetic variation is statistically associated with variation in a quantitative trait. In the context of this disclosure, QTLs include: (a) “caQTL” (chromatin accessibility QTL), where genetic variation is associated with variation in chromatin accessibility at a cCRE; (b) “hQTL” (histone modification QTL), where genetic variation is associated with variation in H3K27ac orAty. Dkt. No.: 114198-7260 H3K27me3 signal; and (c) “eQTL” (expression QTL), where genetic variation is associated with variation in gene expression levels.
[0064] As used herein, “modulating” or “modulation” of a GRN, cCRE, chromatin state, gene, or pathway refers to any alteration in the activity, expression, accessibility, or function thereof, whether positive (activating, enhancing, increasing, opening) or negative (inhibiting, suppressing, decreasing, repressing). Modulation may be direct (e.g., by an agent that binds a transcription factor or cCRE) or indirect (e.g., by an agent that acts on an upstream regulator, a signaling pathway, or an epigenomic modifier).
[0065] As used herein, “non-disease profile” refers to a pattern of GRN activity, gene expression, chromatin accessibility, or chromatin state in hepatocytes that more closely resembles the profile observed in normal (non-MASLD) liver tissue than the profile observed in MASH liver tissue. Specifically, a non-disease profile is characterized by: (a) reduced activity of GRNs enriched in MASH hepatocytes, including GRNs for NFKB1, RELB, TEAD1, CREB5, KLF6, ETS1, and NFAT5; (b) maintained or increased activity of GRNs associated with normal hepatocyte function, including GRNs for HNF4A, PPARA, and RORA; (c) reduced expression of genes in cytoskeletal remodeling, cell-ECM interaction, and inflammatory pathways; and (d) maintained expression of genes in lipid metabolism and cholesterol-related pathways.
[0066] As used herein, “MASLD stage” refers to the clinical and histological classification of a subject’s liver condition along the MASLD spectrum, including: (a) “normal,” referring to liver tissue without steatosis, inflammation, or fibrosis; (b) “MASL” or “simple steatosis,” characterized by accumulation of lipid droplets in hepatocytes without significant inflammation or fibrosis; (c) “Fib- MASH” or “low-fibrosis MASH,” referring to MASH with fibrosis stages 0-2; and (d) “Fib+ MASH” or “high-fibrosis MASH,” referring to MASH with fibrosis stages 3-4. MASLD stages may be determined by histological assessment, non-invasive biomarkers, imaging, or molecular profiling of liver tissue.
[0067] As used herein, “hepatocyte GRN associated with lipid metabolism and biosynthesis” refers to a transcription factor gene regulatory network in hepatocytes that shows increased activity in MASL and is associated with regulation of one or more genes involved in lipid biosynthesis, lipid transport, lipid metabolism, or lipid binding. In some embodiments, such a GRN is associated with one or more transcription factors selected from HNF4A, RORA, CUX2, and NR1I2. In some embodiments, genes regulated by such GRNs include one or more of SMPD3, LPIN3, SLC25A42, SLC22A1, CRYL1, and PSD4.Aty. Dkt. No.: 114198-7260
[0068] As used herein, “hepatocyte GRN associated with epithelial remodeling” refers to a transcription factor gene regulatory network in hepatocytes that shows increased activity in MASH and is associated with regulation of one or more genes involved in epithelial remodeling, cytoskeletal organization, cell adhesion, cell-extracellular matrix interaction, inflammation, autophagy, senescence, cytokine signaling, or immune signaling. In some embodiments, such a GRN is associated with one or more transcription factors selected from CREB5, NFKB1, RELB, KLF6, TEAD1, ETS1, NFAT5, STAT1, HNF4G, and PPARG. In some embodiments, genes regulated by such GRNs include one or more of DTNA, AKAP12, ITGA2, LAMA3, MY07B, CD44, TNFAIP2, TMEM150B, and FGF1.
[0069] As used herein, “motif-enriched hepatocyte cCRE” refers to a candidate cis-regulatory element in hepatocytes having a sequence motif enrichment for one or more transcription factor binding motifs identified in the present disclosure. In some embodiments, the motif-enriched hepatocyte cCRE is enriched for one or more motifs selected from HNF4A, NR4A1, CEBPA / D, FOS / JUN, TEAD, andNFKB motifs. In some embodiments, cCREs enriched for HNF4A, NR4A1, or CEBPA / D motifs are associated with MASL or early disease states, and cCREs enriched for FOS / JUN, TEAD, or NFKB motifs are associated with MASH progression.
[0070] As used herein, “Fib-a hepatocyte marker gene” refers to a gene preferentially expressed in fibrosis-associated hepatocytes relative to other hepatocyte sub-clusters. In some embodiments, a Fib-a hepatocyte marker gene is selected from NFKB1, STAT4, HIF1 A, SOD2, TEAD1, and TGFBR2.
[0071] As used herein, “Fib-adj hepatocyte marker gene” refers to a gene expressed in fibrosis-adjacent hepatocytes, which are a subset of fibrosis-associated hepatocytes spatially localized adjacent to fibrotic injury. In some embodiments, a Fib-adj hepatocyte marker gene is selected from CXCL8, CXCL6, HSPB8, ITGB6, CDKN2A, and KRT23.
[0072] As used herein, “reference liver sample” refers to a biological sample, dataset, or molecular profile obtained from liver tissue or liver-derived cells of a subject or population with a known liver condition, including normal liver, MASL, low-fibrosis MASH, or high-fibrosis MASH. A reference liver sample may be used for comparison of gene expression, GRN activity, cCRE activity, chromatin state, histone modification level, genotype, cell type proportion, or hepatocyte sub-cluster proportion.
[0073] As used herein, “reference profile” refers to a molecular profile derived from a reference liver sample and including one or more of gene expression, transcription factor gene regulatory network activity, candidate cis-regulatory element activity, chromatin state, histone modificationAty. Dkt. No.: 114198-7260 level, or chromatin conformation. In some embodiments, the reference profile is derived from normal liver tissue. In other embodiments, the reference profile is derived from MASL, low-fibrosis MASH, or high-fibrosis MASH liver tissue.
[0074] As used herein, “increased repression” refers to an increase in a repressive chromatin feature at a genomic region or gene, relative to a reference. In some embodiments, increased repression is characterized by increased H3K27me3 signal and / or transition of a genomic region to a repressive chromatin state. In some embodiments, increased repression is associated with reduced expression of a gene overlapping or linked to the genomic region.
[0075] As used herein, “decreased repression” refers to a decrease in a repressive chromatin feature at a genomic region or gene, relative to a reference. In some embodiments, decreased repression is characterized by decreased H3K27me3 signal and / or transition of a genomic region away from a repressive chromatin state. In some embodiments, decreased repression is associated with increased expression of a gene overlapping or linked to the genomic region.
[0076] As used herein, “genes involved in lipid metabolism, transport, or Wnt signaling” refers to genes participating in one or more biological pathways associated with lipid synthesis, lipid oxidation, fatty acid metabolism, cholesterol metabolism, lipid transport, or Wnt signaling. In some embodiments, such genes are among genes overlapping genomic regions with increased H3K27me3 activity in fibrosis-associated hepatocytes in Fib+ MASH.
[0077] As used herein, “genes involved in cell structure, adhesion, extracellular matrix interaction, fibrosis, or cytokine signaling” refers to genes participating in one or more biological pathways associated with cytoskeletal organization, epithelial remodeling, cell adhesion, cell-extracellular matrix interaction, fibrotic remodeling, inflammatory signaling, or cytokine signaling. In some embodiments, such genes are among genes overlapping genomic regions with decreased H3K27me3 activity in fibrosis-associated hepatocytes in Fib+ MASH.
[0078] As used herein, “proportion of fibrosis-associated hepatocytes” refers to the fraction or percentage of hepatocytes in a sample that are classified as fibrosis-associated hepatocytes relative to the total hepatocyte population in the sample. In some embodiments, an increased proportion of fibrosis-associated hepatocytes relative to a reference is indicative of high-fibrosis MASH.
[0079] As used herein, “reduced Zone 3 hepatocytes” refers to a lower proportion, abundance, or representation of Zone 3 hepatocytes in a liver sample relative to a reference liver sample. In some embodiments, reduced Zone 3 hepatocytes are identified together with an increased proportion of fibrosis-associated hepatocytes.Aty. Dkt. No.: 114198-7260
[0080] As used herein, “normal liver tissue” refers to liver tissue lacking steatosis, inflammation, or fibrosis as determined by histological, clinical, molecular, or combined assessment. In some embodiments, normal liver tissue provides a reference profile for comparison to MASLD liver tissue.
[0081] As used herein, “MASL reference profile” refers to a molecular profile characteristic of hepatocytes from simple steatosis livers, including increased activity of one or more lipid-associated GRNs and / or increased expression of one or more genes selected from SMPD3, LPIN3, SLC25A42, SLC22A1, CRYL1, and PSD4, relative to normal liver.
[0082] As used herein, “MASH reference profile” refers to a molecular profile characteristic of hepatocytes from MASH livers, including increased activity of one or more GRNs associated with epithelial remodeling, inflammation, or fibrosis, and / or increased expression of one or more genes selected from DTNA, AKAP12, ITGA2, LAMA3, MY07B, CD44, TNFAIP2, TMEM150B, and FGF1, relative to normal liver.
[0083] As used herein, “comparison to a reference profile” means evaluating one or more molecular features in a test sample relative to the same one or more molecular features in a reference profile. Such comparison may be qualitative or quantitative and may involve determining increase, decrease, enrichment, depletion, gain of chromatin state, loss of chromatin state, or other relative difference.
[0084] As used herein, “spatial transcriptomics” refers to a method that measures RNA expression in tissue while preserving spatial information regarding the location of cells or transcripts within the tissue. In some embodiments, spatial transcriptomics includes Visium or Visium HD-based profiling.
[0085] As used herein, “chromatin accessibility profiling” refers to a method of determining the accessibility of chromatin regions in cells or nuclei. In some embodiments, chromatin accessibility profiling includes ATAC-seq, including single-cell or single-nucleus ATAC-seq, and multiome assays comprising ATAC-seq.
[0086] As used herein, “histone modification profiling” refers to a method of determining the presence or level of one or more histone modifications at genomic regions in cells or nuclei. In some embodiments, histone modification profiling includes profiling of H3K27ac and / or H3K27me3, including single-cell or single-nucleus assays such as Paired-Tag, CUT&Tag, or related assays.Aty. Dkt. No.: 114198-7260
[0087] As used herein, “3D chromatin conformation profiling” refers to a method of determining chromatin contacts, chromatin interactions, or chromatin architecture in cells or nuclei. In some embodiments, 3D chromatin conformation profiling includes Hi-C, including droplet Hi-C and single-cell Hi-C.
[0088] As used herein, “single-cell multiome analysis” refers to a method of jointly profiling gene expression and chromatin accessibility in the same cell or nucleus. In some embodiments, single-cell multiome analysis comprises paired RNA-seq and ATAC-seq profiling.
[0089] As used herein, “gene expression indicative of MASH progression” refers to expression of one or more genes at a level that is increased in MASH relative to normal liver and / or MASL and is associated with progression along the MASLD spectrum. In some embodiments, such genes include one or more of DTNA, AKAP12, ITGA2, LAMA3, MY07B, CD44, TNFAIP2, TMEM150B, andFGFl.
[0090] As used herein, “gene expression indicative of MASL” refers to expression of one or more genes at a level that is increased in MASL relative to normal liver and associated with lipid metabolism, synthesis, transport, or binding. In some embodiments, such genes include one or more of SMPD3, LPIN3, SLC25A42, SLC22A1, CRYL1, andPSD4.
[0091] As used herein, “risk locus-associated gene” refers to a gene linked in the present disclosure to a MASLD-associated genetic locus through one or more of cCRE overlap, chromatin state annotation, QTL association, colocalization analysis, chromatin looping, or predicted functional variant effects. In some embodiments, a risk locus-associated gene is selected from KRT8, KRT18, PPP1R3B, CEBPA, XBP1, and EFHD1.
[0092] Modes For Carrying Out the Disclosure
[0093] The present disclosure provides comprehensive single cell transcriptomic, epigenomic, and spatial profiling of five modalities in 2.4 million cells from 86 human livers across MASLD stages, enabling the identification of cell type-specific regulatory networks driving MASLD progression and representing therapeutic avenues. The five modalities include single cell gene expression, accessible chromatin (ATAC-seq), two histone modifications associated with active (H3K27ac) and repressive (H3K27me3) states, and 3D chromatin conformation (Hi-C), along with spatial transcriptomics in seven livers.
[0094] The integrated single cell map of the human liver consisted of a total of 678,748 single nuclei combining data from all multiome, Paired-Tag, and Hi-C assays. The 86 donors were classified as normal (n=26), simple steatosis (MASL) (n=25), MASH (n=23) and MASH with1Aty. Dkt. No.: 114198-7260 history of alcohol consumption (MetALD) (n=12). The MASH / MetALD donors were further divided into livers with mild / low-fibrosis (Fib- MASH; stages 0-2, n=19) and severe / high-fibrosis (Fib+ MASH; stages 3-4, n=16).
[0095] In some aspects, the disclosure provides a method of treating or preventing metabolic dysfunction-associated steatohepatitis (MASH) in a subject in need thereof, the method comprising, or consisting essentially of, or consisting of administering to the subject an effective amount of an agent that modulates the activity of at least one transcription factor gene regulatory network (GRN) in hepatocytes, wherein the at least one GRN includes a transcription factor selected from the group consisting of HNF4A, RORA, CUX2, CREB5, NFKB1, RELB, KLF6, TEAD1, ETS1, NR1I2, NFAT5, FOXO1, NR4A1, HNF4G, PPARG, STAT1, PPARA, and ATF3, and wherein said modulating is sufficient to alter the expression of one or more target genes regulated by the at least one GRN.
[0096] In some aspects, the agent increases the activity of a GRN selected from HNF4A, RORA, CUX2, NR1I2, and PPARA in hepatocytes.
[0097] In other aspect, the agent decreases the activity of a GRN selected from NFKB1, RELB, TEAD1, CREB5, KLF6, ETS1, andNFAT5 in hepatocytes.
[0098] In further aspects, the one or more target genes are selected from the group consisting of DTNA, AKAP12, ITGA2, LAMA3, MY07B, CD44, TNFAIP2, TMEM150B, FGF1, SMPD3, LPIN3, SLC25A42, SLC22A1, CRYL1, and PSD4.
[0099] In some aspects, the agent modulates the activity of one or more candidate cis-regulatory elements (cCREs) within the GRN.
[0100] Also provided herein, is a method wherein the subject has a MASLD stage selected from MASL, low-fibrosis MASH (Fib- MASH; fibrosis stages 0-2), and high-fibrosis MASH (Fib+ MASH; fibrosis stages 3-4).
[0101] Further provided is a method, wherein the agent is selected from the group consisting of a small molecule, a nucleic acid, an antisense oligonucleotide, an siRNA, an shRNA, a CRISPR-based gene editing agent, a transcription factor modulator, a protein, a peptide, an antibody or antigen-binding fragment thereof, and a viral vector encoding a transcription factor or a modulator thereof.
[0102] In some aspects, the disclosure provides a method of diagnosing or stratifying metabolic dysfunction-associated steatotic liver disease (MASLD) progression in a subject, the method including: (a) obtaining a biological sample including liver cells from the subject; (b)Aty. Dkt. No.: 114198-7260 determining, in the liver cells, the activity of one or more gene regulatory features selected from: (i) transcription factor gene regulatory networks (GRNs); (ii) candidate cis-regulatory elements (cCREs); (iii) chromatin states at cCREs; and (iv) gene expression of GRN target genes; and (c) classifying the subject into a MASLD stage based on the determined activity, wherein the MASLD stage is selected from normal, MASL, low-fibrosis MASH, and high-fibrosis MASH.
[0103] In some respects step (b) includes determining the activity of GRNs with increased activity in MASL selected from HNF4A, RORA, and CUX2, and GRNs with increased activity in MASH selected from CREB5, NFKB1, RELB, KLF6, TEAD1, ETS1, and NFAT5.
[0104] In further aspects, step (b) includes determining the proportion of hepatocyte sub-clusters selected from Zone 1, Zone 2, Zone 3, MASH-associated (MASH-a), and fibrosis-associated (Fib-a) hepatocytes.
[0105] In some aspects, the disclosure provides a method, wherein an increased proportion of Fib-a hepatocytes and a decreased proportion of Zone 1 and Zone 3 hepatocytes is indicative of high-fibrosis MASH.
[0106] In some aspects, the disclosure provides a method, wherein step (b) includes determining the chromatin state of one or more hepatocyte cCREs, wherein a transition from repressive to active chromatin states at cCREs enriched for NFKB, FOS / JUN, and TEAL) motifs is indicative of MASH progression.
[0107] In some aspects, the method further comprises, or consists essentially of, or yet further consists of determining the genotype of the subject at one or more MASLD-associated risk loci selected from 12ql3 (KRT8 / KRT18), 8p23 (PPP1R3B), 19ql3 (CEBPA), 22ql2 (XBP1), and 2q37 (EFHD1).
[0108] Also provided is a method of identifying a candidate therapeutic agent for the treatment of metabolic dysfunction-associated steatohepatitis (MASH), the method including: (a) providing a population of hepatocytes; (b) contacting the hepatocytes with a test agent; (c) determining the effect of the test agent on the activity of one or more transcription factor gene regulatory networks (GRNs) associated with MASH progression, wherein the one or more GRNs are selected from the group consisting of CREB5, NFKB1, RELB, KLF6, TEAD1, ETS1, NFAT5, HNF4A, RORA, CUX2, FOXO1, NR4A1, HNF4G, PPARG, STAT1, PPARA, and ATF3; and (d) selecting the test agent as a candidate therapeutic agent if the test agent shifts the activity of the one or more GRNs toward a non-disease profile.Aty. Dkt. No.: 114198-7260
[0109] In some aspects, the determining the effect of the test agent includes measuring one or more of: (i) chromatin accessibility at cCREs within the GRN; (ii) histone modification levels (H3K27ac and / or H3K27me3) at cCREs within the GRN; (iii) gene expression of target genes within the GRN; and (iv) 3D chromatin contacts at loci associated with the GRN.
[0110] In some embodiments, a non-disease profile includes decreased activity of GRNs selected from NFKB1, RELB, TEAD1, CREB5, KLF6, ETS1, and NFAT5, and increased activity of GRNs selected from HNF4A, RORA, CUX2, and PPARA.[OHl] In other embodiments, the hepatocytes include primary human hepatocytes or hepatocytes derived from induced pluripotent stem cells (iPSCs).
[0112] In some aspects, the disclosure provides a method of treating liver fibrosis associated with MASH in a subject, the method including administering to the subject an effective amount of an agent that disrupts CD44-mediated signaling between hepatic stellate cells (HSCs) and fibrosis-associated hepatocytes, wherein the agent inhibits one or more of: (a) COL1 A1-CD44 interaction between HSCs and fibrosis-adjacent (Fib-adj) hepatocytes; and (b) CD44 expression or activity in hepatocytes. Non-limiting examples of the agent is an anti-CD44 antibody, a CD44-targeting siRNA, a CD44 small molecule inhibitor, or a combination thereof.
[0113] In further aspects, the disclosure provides a method that further comprises, or consists essentially of, or yet further consists of administering an agent that modulates the epigenomic profile of fibrosis-associated hepatocytes, wherein said modulating includes increasing H3K27me3 -mediated repression at epithelial remodeling and fibrosis-related genes, or decreasing H3K27me3 -mediated repression at lipid metabolism genes.
[0114] In one aspect of the above methods, the agent modulates the activity of one or more hepatocyte candidate cis-regulatory elements (cCREs) enriched for sequence motifs of one or more transcription factors selected from the group consisting of HNF4A, NR4A1, CEBPA / D, FOS / JUN, TEAD, and NFKB. In another aspect, the agent decreases activity of one or more hepatocyte cCREs enriched for FOS / JUN, TEAD, or NFKB motifs. In a further aspect, the agent increases activity of one or more hepatocyte cCREs enriched for HNF4A, NR4A1, or CEBPA / D motifs.
[0115] In one embodiment of the methods, step (b) comprises determining increased activity of one or more hepatocyte GRNs associated with lipid metabolism and biosynthesis, wherein the one or more GRNs are selected from the group consisting of HNF4A, RORA, CUX2, and NR1I2. In another embodiment, step (b) comprises determining increased activity of one or more hepatocyte GRNs associated with epithelial remodeling, cytoskeletal organization, cell-Aty. Dkt. No.: 114198-7260 extracellular matrix adhesion, inflammation, or immune signaling, wherein the one or more GRNs are selected from the group consisting of CREB5, NFKB1, RELB, KLF6, TEAD1, ETS1, NFAT5, STAT1, HNF4G, and PPARG. Alternatively, step (b) comprises determining expression of one or more genes selected from the group consisting of DTNA, AKAP12, ITGA2, LAMA3, MYO7B, CD44, TNFAIP2, TMEM150B, FGF1, SMPD3, LPIN3, SLC25A42, SLC22A1, CRYL1, and PSD4. Yet further, increased expression of one or more of DTNA, AKAP12, ITGA2, LAMA3, MY07B, CD44, TNFAIP2, TMEM150B, and FGF1 is indicative of MASH progression. In another aspect, increased expression of one or more of SMPD3, LPIN3, SLC25A42, SLC22A1, CRYL1, and PSD4 is indicative of MASL.
[0116] In one aspect, step (b) comprises determining whether hepatocytes in the sample comprise a fibrosis-associated hepatocyte population expressing one or more of NFKB1, STAT4, HIF1A, SOD2, TEAD1, and TGFBR2. Alternatively, the fibrosis-associated hepatocyte population is present in increased proportion in subjects with high-fibrosis MASH relative to normal liver. Yet further, step (b) comprises determining whether the sample comprises fibrosis-adjacent hepatocytes expressing one or more of CXCL8, CXCL6, HSPB8, ITGB6, CDKN2A, and KRT23. In addition, step (b) comprises determining whether Zone 3 hepatocytes are reduced and fibrosis-associated hepatocytes are increased relative to a reference liver sample.
[0117] In some aspects, step (b) comprises determining whether hepatocytes exhibit increased repression at genes involved in lipid metabolism, transport, or Wnt signaling and / or decreased repression at genes involved in cell structure, adhesion, extracellular matrix interaction, fibrosis, or cytokine signaling. In addition, the repression is determined by measuring H3K27me3 signal.
[0118] In one embodiment, the non-disease profile comprises decreased expression of one or more genes selected from the group consisting of DTNA, AKAP12, ITGA2, LAMA3, MY07B, CD44, TNFAIP2, TMEM150B, and FGF1. Alternatively, the non-disease profile comprises increased expression of one or more genes selected from the group consisting of SMPD3, LPIN3, SLC25A42, SLC22A1, CRYL1, and PSD4. Yet further, the non-disease profile comprises decreased activity of one or more candidate cis-regulatory elements enriched for NFKB, FOS / JUN, or TEAD sequence motifs. In one embodiment, the non-disease profile comprises increased activity of one or more candidate cis-regulatory elements enriched for HNF4A, NR4A1, or CEBPA / D sequence motifs.
[0119] In one aspect, the fibrosis-associated hepatocytes comprise hepatocytes expressing one or more of NFKB1, STAT4, HIF1A, SOD2, TEAD1, and TGFBR2.Aty. Dkt. No.: 114198-7260
[0120] In another embodiment, the fibrosis-adjacent hepatocytes comprise a subset of fibrosis-associated hepatocytes expressing one or more of CXCL8, CXCL6, HSPB8, ITGB6, CDKN2A, and KRT23.
[0121] In one aspect, the agent reduces signaling from hepatic stellate cells to fibrosis-adjacent hepatocytes involving COL 1 Al and CD44.
[0122] In one embodiment, the epithelial remodeling and fibrosis-related genes comprise genes involved in cell structure, adhesion, extracellular matrix interaction, or cytokine signaling.
[0123] In one aspect, the lipid metabolism genes comprise genes involved in lipid metabolism, lipid transport, or Wnt signaling.
[0124] In one aspect of the disclosed methods, the one or more MASLD-associated risk loci comprise a locus at 12ql 3 associated with KRT8 / KRT18, a locus at 8p23 associated with PPP1R3B, a locus at 19ql 3 associated with CEBPA, a locus at 22ql2 associated with XBP1, or a locus at 2q37 associated with EFHD1.
[0125] In a further aspect, the genotype determined at the one or more MASLD-associated risk loci is used together with the activity of one or more transcription factor gene regulatory networks (GRNs), candidate cis-regulatory elements (cCREs), chromatin states, or GRN target genes to classify the subject into a MASLD stage.
[0126] In another aspect, the biological sample is analyzed by one or more methods selected from the group consisting of single-cell RNA sequencing, single-cell multiome analysis, chromatin accessibility profiling, histone modification profiling, 3D chromatin conformation profiling, and spatial transcriptomics.
[0127] In one embodiment, the determining the effect of the test agent comprises comparing the activity of the one or more GRNs to a reference profile from normal liver tissue. Alternatively, determining the effect of the test agent comprises comparing the activity of the one or more GRNs to a reference profile from MASL, low-fibrosis MASH, or high-fibrosis MASH liver tissue.
[0128] It is to be understood that this disclosure provides the methods wherein the method is performed on a sample that has been obtained and ready for analysis such that the step of obtaining the sample is omitted from the claimed methods.
[0129] Cell Type-Specific Gene Regulatory Programs
[0130] A unified set of 611,576 candidate cis-regulatory elements (cCREs) was defined across cell types using accessible chromatin (ATAC-seq) profiles. Between 43.50-47.89% of cCREsAty. Dkt. No.: 114198-7260 annotated were novel compared to previous catalogs. Chromatin states were called in each cell type using ATAC-seq, H3K27ac, and H3K27me3 profiles, revealing active (ATAC-seq+H3K27ac), open (ATAC-seq), repressive (H3K27me3) and weakly active states. Chromatin states annotated 47.7% of the genome in hepatocytes, compared to 5.9% in cCREs based on chromatin accessibility only.
[0131] Transcription factor gene regulatory networks (GRNs) were defined in each cell type using SCENIC+. Cell types had GRNs for 186 TFs (105-137 per cell type), and on average, TF GRNs consisted of 123 cCREs linked to 82 genes. In hepatocytes, the STAT1 GRN was linked to interferon signaling and viral response and the ATF3 / 4 GRN was linked to endoplasmic reticulum stress and unfolded protein response.
[0132] Significant enrichment of MASLD-associated variants was identified in hepatocyte cCREs using LD score regression, with stronger enrichment for hepatocyte cCREs in active chromatin states. Fine-mapped variants at 68 / 77 (88%) of all MASLD loci were annotated to liver cell type cCREs or active states.
[0133] Transcriptional Drivers of MASLD Progression
[0134] Marked changes in gene expression, chromatin accessibility, and histone modification profiles were identified across MASLD stages compared to normal livers in hepatocytes, cholangiocytes, HSC, endothelial, and myeloid cells. Fibrosis was identified as a major cause of disease-associated changes, with more extensive changes in high fibrosis (Fib+) MASH compared to low fibrosis (Fib-) MASH.
[0135] Hepatocytes in MASH had increased expression of pathways related to epithelial remodeling, such as cytoskeletal structure and cell-ECM interactions, as well as pathways related to glycolysis, hypoxia, senescence, cytokine signaling, cell death, and signaling responses.MASH hepatocytes exhibited reduced expression of genes in faty acid and cholesterol related pathways. MASL and MASH were driven by divergent gene expression programs in hepatocytes.
[0136] TF GRNs with increased activity in MASL included HNF4A, RORA, and CUX2, which were connected to genes with increased expression in MASL involved in lipid biosynthesis (SMPD3, LPIN3, SLC25A42), binding (PSD4), transport (SLC22A1), and metabolism (CRYL1). GRNs with increased activity in MASH included CREB5, NFKB1 / RELB, KLF6, TEAD1, ETS1, and NFAT5, which regulated genes involved in cytoskeletal organization, adhesion and ECM interactions (DTNA, AKAP12, ITGA2, LAMA3, MY07B, CD44), inflammation (TNFAIP2), autophagy (TMEM150B), and lipid metabolism (FGF1).Aty. Dkt. No.: 114198-7260
[0137] Stratifying MASH into Fib+ and Fib- groups revealed that MASL hepatocytes uniquely induced HNF4A, NR1I2, CUX2, and RORA GRNs, while Fib- MASH hepatocytes shared some of these and uniquely upregulated FOXO1, NR4A1, and FOS. Fib+ MASH hepatocytes uniquely upregulated HNF4G, PPARG, NFKB, NFAT5 and STAT1 and downregulated PPARA and HNF4A GRNs, highlighting fibrosis as a primary driver of regulatory changes in MASH.
[0138] MASLD Risk Loci and Hepatocyte Regulatory Programs
[0139] QTL mapping of chromatin accessibility (caQTLs), histone modifications (hQTLs), and gene expression (eQTLs) across all 86 donors identified 10,591 caQTLs, 1,644 eQTLs, 418 H3K27ac hQTLs, and 391 H3K27me3 hQTLs across all cell types. Variants at over a third of all known MASLD loci (28 / 77) were QTLs in hepatocytes, of which 18 were colocalized with QTL associated variants.
[0140] Specific MASLD loci linked to hepatocyte regulatory programs include: (a) the 12q 13 locus, where hepatocyte cCREs affected by MASLD variants interact with the promoter region of KRT8 / KRT18 through chromatin looping; (b) the 8p23 locus, where the hepatocyte cCRE affected by MASLD variants interacted with the PPP1R3B promoter over 200kb distal; (c) the 19q 13 locus where cCREs interacted with the CEBPA promoter; and (d) the 22ql2 locus where cCREs interacted with the XBP1 promoter.
[0141] At the 2q37 locus, risk alleles increased cCRE accessibility, H3K27ac signal, and EFHD1 expression in hepatocytes. Sequence-based models predicted functional effects of variants at almost all (24 / 26) MASLD loci with hepatocyte caQTLs.
[0142] Hepatocyte Heterogeneity and Fibrosis-Associated Sub-Clusters
[0143] Five distinct sub-clusters of hepatocytes were identified: Zone 1 (peri -portal), Zone 2 (mid-lobular), Zone 3 (peri-central), MASH-associated (MASH-a), and fibrosis-associated (Fib-a) populations.
[0144] MASH-a hepatocytes preferentially expressed genes implicated in hepatocyte proliferation, identity and function such as PROXI, XBP1, and CBP2, and had significantly increased proportion in MASH compared to normal livers. Fib-a hepatocytes preferentially expressed inflammatory and stress-related genes (NFKB1, STAT4, HIF1A, SOD2, TEAD1, TGFBR2) and had significantly increased proportion in Fib+ MASH and severe fibrosis.
[0145] Fib-a hepatocytes had dramatically altered repressive domains marked by H3K27me3, including increased repression at lipid metabolism genes and loss of repression at epithelial remodeling and fibrosis-related genes. Zone 3 hepatocytes were ordered most closely to Fib-aAty. Dkt. No.: 114198-7260 cells on pseudotime analysis, supporting that epigenetic remodeling drives a transcriptional switch from Zone 3 cells toward the fibrosis-associated state.
[0146] Fibrosis-adjacent (Fib-adj) hepatocytes were specifically localized to cell types in fibrotic areas including HSCs and immune cells. Cell-cell interaction analysis revealed signaling from HSCs specifically to Fib-adj hepatocytes involving collagen (COL1 Al) and CD44, a cell surface adhesion molecule implicated in fibrogenic injury response and hepatocellular carcinoma.
[0147] Therapeutic Applications
[0148] Based on the identified cell type-specific gene regulatory programs and their association with MASLD progression, the present disclosure provides multiple therapeutic strategies.
[0149] First, agents that modulate MASL-associated GRNs (e.g., enhancing HNF4A, RORA, CUX2 GRN activity) may promote a beneficial hepatocyte response to lipids and prevent progression to fibrosis. Lipid-related networks increased in MASL and Fib- MASH were most enriched for MASLD-associated variants, linking genetic risk to processes upregulated in early stages of disease.
[0150] Second, agents that inhibit MASH-associated GRNs (e.g., inhibiting NFKB1 / RELB, TEAD1, KLF6, CREB5, ETS1, NFAT5 activity) may prevent or reverse inflammatory and fibrotic pathways in hepatocytes.
[0151] Third, modulation of hepatocyte sub-cluster identity — particularly preventing the transition from Zone 3 hepatocytes to fibrosis-associated states — represents a therapeutic strategy to prevent fibrosis progression.
[0152] Fourth, disrupting the CD44-mediated signaling between HSCs and fibrosis-adjacent hepatocytes may reduce fibrosis-associated hepatocyte remodeling.
[0153] Fifth, targeting specific MASLD risk loci and their candidate genes (PPP1R3B, CEBPA, KRT8 / KRT18, XBP1, EFHD1) provides opportunities for genotype-informed therapeutic strategies.
[0154] EXAMPLES
[0155] Experiment No.l:
[0156] Single cell and spatial map of the human liver from non-disease and MASLD donors
[0157] Applicant performed single cell assays to profile gene expression, accessible chromatin, two histone modifications, and 3D genome architecture in human livers from 86 donors classified as normal (n=26), simple steatosis (MASL) (n=25), metabolic dysfunction associatedAty. Dkt. No.: 114198-7260 steatohepatitis (MASH) (n=23) and MASH with history of alcohol consumption (MetALD) (n=12) (data not shown). The MASH / MetALD donors were further divided into livers with mild / low-fibrosis (Fib-MASH; stages 0-2, n=19) and severe / high-fibrosis (Fib+ MASH; stages 3-4, n=16). An overview of the study design and assays performed for each donor is shown in FIG. 1A and FIG. 6
[0158] For all 86 liver donors, Applicant first generated joint single nucleus accessible chromatin and gene expression (lOx multi ome) profiles using a pooled sample design, along with genome-wide genotyping (FIG. 6). After sample demultiplexing and removal of low-quality cells and doublets, Applicant clustered 408,178 profiles across all 86 donors using single cell transcriptomes in Seurat (Y. Hao, et al., Integrated analysis of multimodal single-cell data. Cell 184, 3573-3587. e29 (2021).) (FIG. 7). Applicant annotated the cell type identity of multiome clusters using marker gene expression which identified hepatocytes, cholangiocytes, endothelial, myeloid, hepatic stellate cell (HSC), lymphoid, and other cell types (FIG. IB, FIG.1C). Next, Applicant generated joint single nucleus profiles of gene expression and histone modification for the active chromatin mark H3K27ac and repressive mark H3K27me3 using droplet-based Paired-Tag (Y. Xie, et al., Droplet-based single-cell joint profiling of histone modifications and transcriptomes. Nat Struct Mol Biol 30, 1428-1433 (2023).) on pooled samples from the same 86 donors (FIG. 6). Following demultiplexing and removal of low-quality cells and doublets, Applicant reference mapped 94,125 H3K27ac and 86,554 H3K27me3 cells to the multiome atlas (FIG. 7). Finally, Applicant performed droplet-based single cell Hi-C (L. Chang, et al., Droplet Hi-C enables scalable, single-cell profiling of chromatin architecture in heterogeneous tissues. Nat Biotechnol, doi: 10.1038 / s41587-024-02447-l (2024).) on pooled samples from 85 of the 86 liver donors (FIG. 6). After demultiplexing and doublet removal from each pool, Applicant obtained a total of 89,891 nuclei with high quality Hi-C profiles based on the number of total, cis, and trans chromatin contacts (FIG. 7). Applicant integrated singlecell Hi-C profiles with the multiome and Paired-Tag map by using label transfer in Seurat (T. Stuart, et al., Comprehensive Integration of Single-Cell Data. Cell 177, 1888-1902. e21 (2019).) (see Methods, FIG. 7).
[0159] Overall, the resulting integrated map of the human liver consisted of a total of 678,748 single nuclei combining data from all multiome, Paired-Tag and Hi-C assays (FIG. IB, FIG. 7).Epigenomic profiles at cell type marker genes revealed accessibility (ATAC-seq) and active signal (H3K27ac) for the corresponding cell type and repressed signal (H3K27me3) for nonactive cell types (FIG. 1C, FIG. ID). Applicant also observed cell type-specific patterns of 3D chromatin architecture at known cell type marker genes, such as increased chromatin contacts atAty. Dkt. No.: 114198-7260 the ADAMTS2 locus in HSCs compared to other liver cell types such as endothelial cells (FIG. IE). Applicant further annotated single cell profiles at an additional layer of resolution to identify sub-types of cell types in the liver. This revealed known cell sub-types such as quiescent and activated HSCs (aHSCs), meschenchymal cells, sinusoidal endothelial cells (LSECs), liver resident Kupffer cells (KCs), myeloid cells (including lipid-associated macrophages) and other inflammatory cells (e.g. T / B / NK cells), and zone-specific hepatocytes, among other sub-types (FIG. 8).
[0160] Applicant next examined the distribution of cell types and cell sub-types within each of the 86 liver donors. Each donor contributed consistently to the single-cell profiles across both multiome and Paired-Tag data, with balanced representation across phenotypes such as age, sex, and BMI (FIG. IF). Similarly, single cell Hi-C profiles showed consistent representation across donors (FIG. IF). Applicant observed significant changes (FDR<.10) in the abundance of liver cell types across MASLD states, most prominently an increased abundance of HSCs in MASH, particularly in high-fibrosis (Fib+) MASH, reflecting their activation in MASLD (FIG. IF, FIG.1G). In addition, cholangiocytes, B cells and T cells were all increased, and hepatocytes decreased, in MASH progression, consistent with increased inflammation and hepatocyte loss (FIG. IF, FIG. 1G) There were similarly significant changes (FDR< 10) in the abundance of cell types in higher fibrosis stage including increased HSCs, cholangiocytes, and B cells, as well as decreased hepatocytes (FIG. IF, FIG. 1G). Applicant also observed significant changes (FDR< 10) in the abundance of liver cell sub-types in MASLD; for example, aHSCs, vascular and lymphatic endothelial cells, and lipid-associated macrophages were all more abundant across MASH progression and in higher fibrosis stage (data not shown).
[0161] Finally, Applicant used spatial transcriptomics (Visium HD) to profile the organization of liver cell types in seven livers representing normal (n=2), MASL (n=2), MASH (n=2) and MetALD (n=l) (see Methods, FIG. 9). After cell segmentation and filtering of low quality cells, Applicant annotated 1,729,432 cells based on reference mapping from the single cell multiome map using Tangram (T. Biancalani, et al., Deep learning and alignment of spatially resolved single-cell transcriptomes with Tangram. Nat Methods 18, 1352-1362 (2021).). Cell type assignments were confirmed based on expression of key marker genes (FIG. 1H, FIG. II).Normal livers were composed of liver lobules with central and portal veins and hepatocytes organized in cords, while quiescent HSCs resided in the space of Disse and myeloid cells were found throughout liver parenchyma. In MASL, normal liver architecture was preserved but with steatosis, characterized by accumulation of lipid droplets in hepatocytes. MASH livers had injured hepatocytes around the central vein, inflammation, and fibrosis. Low fibrosis (Fib-)Aty. Dkt. No.: 114198-7260 MASH was characterized by aHSCs with migration towards liver injury and inflammation. High fibrosis (Fib+) MASH or cirrhosis had severe inflammation with bridging fibrosis, and aHSCs, myeloid, T and B cells near the fibrous scar surrounding hepatic regenerative nodules.
[0162] Together, these results provide an integrated single cell map and spatial profiles from normal and MASLD livers. To provide these genomic data to the community as a resource, Applicant developed a browser available at epigenome.wustl.edu / MASLD / .
[0163] Cell type-specific gene regulatory programs in the liver
[0164] Applicant next leveraged the integrated single cell map of multiple modalities to characterize the gene regulatory programs of each liver cell type. Applicant first defined a unified set of 611,576 candidate c / .s-regulatory elements (cCREs) across cell types using accessible chromatin (ATAC-seq) profiles (data not shown). Compared to previous catalogs of cCREs in the adult and fetal liver (The ENCODE Project Consortium, F. Abascal, et al., Expanded encyclopaedias of DNA elements in the human and mouse genomes. Nature 583, 699-710 (2020)., K. Zhang, et al., A single-cell atlas of chromatin accessibility in the human genome. Cell 184, 5985-6001. el9 (2021).), between 43.50-47.89% of cCREs Applicant annotated were novel (FIG. 2A, FIG. 10). Liver cell type cCREs overall showed expected distribution across genomic features, although novel cCREs were more enriched in retro-transposon and other classes of elements relative to known cCREs (FIG. 2A, FIG. 10). The novel cCREs also exhibited greater inter-donor variability than known cCREs, consistent with the expanded number and diversity of profiled livers uncovering a larger portfolio of regulatory elements (FIG. 10).
[0165] Next, Applicant called chromatin states in each cell type using ATAC-seq, H3K27ac, and H3K27me3 profiles with chromHMM (J. Ernst, M. Kellis, Chromatin-state discovery and genome annotation with ChromHMM. NatProtoc 12, 2478-2492 (2017).), which revealed active (ATAC-seq+H3K27ac), open (ATAC-seq), repressive (H3K27me3) and weakly active chromatin states (FIG. 2B). Chromatin states annotated a substantially larger proportion of the non-coding genome in liver cell types compared to cCRE definitions. For example, in hepatocytes, 47.7% of the genome was annotated as an active, weak, open or repressed chromatin state, compared to 5.9% of the genome in cCREs based on chromatin accessibility only (FIG. 2B). Comparing chromatin states in liver cell types to states in bulk liver from NIH Epigenome Roadmap (Roadmap Epigenomics Consortium, A. Kundaje, et al., Integrative analysis of 111 reference human epigenomes. Nature 518, 317-330 (2015).) revealed strong concordance in state identity across the genome for hepatocytes with lower concordance amongAty. Dkt. No.: 114198-7260 less common cell types (e.g. HSCs) (FIG. 10), highlighting that bulk chromatin state profiles more closely reflect common cell types. As the functional roles of most cCREs are undefined, Applicant used chromatin state information to annotate cCREs in each cell type. In hepatocytes, for example, 30.9% of all cCREs were in an active state while 10% and 22.7% were weak and open states, respectively, and 36.4% were repressed or unknown (FIG. 2B). The cCREs annotated in different chromatin states had distinct properties; for example, in hepatocytes, cCREs in active states were enriched for HNF1 / 4, FOXA and PPAR TF motifs while cCREs in open states were enriched for CTCF, CUX1 / 2 and ONECUT motifs and cCREs in repressed states were enriched for REST motifs (FIG. 10).
[0166] Applicant defined TF gene regulatory networks (GRNs) in each cell type using SCENIC+ (C. Bravo Gonzalez-Blas, et al., SCENIC+: single-cell multiomic inference of enhancers and gene regulatory networks. Nat Methods 20, 1355-1367 (2023).). Cell types had GRNs for 186 TFs (105-137 per cell type) and, on average, TF GRNs consisted of 123 cCREs linked to 82 genes (data not shown). Biological pathways enriched in target genes of GRNs revealed known transcriptional regulators of cellular processes - for example, in hepatocytes, the STAT1 GRN was linked to interferon signaling and viral response and the ATF3 / 4 GRN was linked to endoplasmic reticulum stress and unfolded protein response - as well as many putative relationships (data not shown). Applicant utilized 3D chromatin contacts in each cell type to validate target gene predictions of cCREs in GRNs. For example, at the CRP locus, multiple hepatocyte-specific cCRE links were supported by chromatin contacts in hepatocytes and not in myeloid cells (FIG. 2C). Similarly, at the CD163 locus, myeloid-specific cCRE links were supported by chromatin contacts in myeloid cells (FIG. 2C). More broadly, cCRE-target gene links collectively had increased 3D chromatin contacts for the corresponding cell type compared to other cell types (FIG. 2C). Applicant also identified physical interactions between distal regions by calling chromatin loops using the 3D chromatin contacts in liver cell types with sufficient coverage (hepatocytes, endothelial, HSC, myeloid cells). Hepatocytes had the largest number of detected loops with fewer loops in less abundant cell types (data not shown).
[0167] Applicant next examined the chromatin states of cCREs across liver cell types to determine how their activity varies across cell types. Applicant grouped all 611,576 cCREs into modules based on accessible chromatin signal across cell types, revealing nine distinct modules (M0-M8) (FIG. 2D). One module (M0) had broadly shared accessibility across cell types while other modules had accessibility profiles that were more cell type restricted (FIG. 2D). Applicant next determined the chromatin state of cCREs in each module. The shared (M0) module was broadly enriched for active chromatin states across cell types, and a contained higher proportionAty. Dkt. No.: 114198-7260 of cCREs in the MO module mapping to gene promoters relative to other modules (13.4% vs. 2.5%). By comparison, the other cell-type restricted modules were enriched for active and open chromatin states in the corresponding cell type(s) where accessibility was observed (FIG. 2D).
[0168] The liver cell types and regulatory programs through which genetic variants influence MASLD risk remain poorly defined. Applicant utilized genome-wide association study (GWAS) data for circulating alanine aminotransferase (cALT) level (M. Vujkovic, et al., A multiancestry genome-wide association study of unexplained chronic ALT elevation as a proxy for nonalcoholic fatty liver disease with histological and radiological validation. Nat Genet 54, 761-771 (2022).), a proxy for liver damage and MASLD, to identify genetic determinants of MASLD in liver cell types. Applicant identified significant enrichment (FDR<.10) of MASLD-associated variants genome-wide in hepatocyte cCREs using LD score regression (B. K. Bulik-Sullivan, et al., LD Score regression distinguishes confounding from polygenicity in genomewide association studies. Nature Genetics 47, 291-295 (2015)., H. K. Finucane, et al., Partitioning heritability by functional annotation using genome-wide association summary statistics. Nature Genetics 47, 1228-1235 (2015).), as well as a similar magnitude of enrichment in hepatocyte H3K27ac sites (FIG.2E). When further classifying cCREs by chromatin state, Applicant observed stronger enrichment of MASLD variants for hepatocyte cCREs in active states and no corresponding enrichment for other states (FIG. 2E). Applicant also observed significant enrichment of MASLD variants among cCREs in the hepatocyte-specific module (M7) and no corresponding enrichment for other modules (FIG. 11).
[0169] Applicant next annotated specific MASLD-associated loci where, overall, 348 out of 659 fine-mapped variants with PPA>.01, representing 68 / 77 (88%) all MASLD loci, were annotated to liver cell type cCREs or active states (FIG. 12). When considering just hepatocytes, 274 fine-mapped variants with PPA>.01 at 58 loci were annotated to hepatocyte cCREs or active states, where active states annotated a larger number of variants compared to cCREs (FIG. 2E).Among variants in hepatocyte cCREs or active chromatin states were several with high causal probabilities (PPA>.99) and unannotated in previous catalogs of liver cell type cCREs (K.Zhang, et al., A single-cell atlas of chromatin accessibility in the human genome. Cell 184, 5985-6001. el9 (2021).), such as rs4841132 at the PPP1R3B locus and rsl 1621792 at the NFATC4 locus. Outside of hepatocytes, several loci were annotated specifically in other cell types such as TRPS1 (indexed by rs2737217) and HLA-DQA1 (rs686250) in myeloid cells, FCGR2B (rs74816838) in endothelial cells, and Fl CC 1 (rsl0201587) in HSCs, supporting that these cell types also play a role in MASLD risk (FIG. 12).Aty. Dkt. No.: 114198-7260
[0170] Finally, Applicant identified gene regulatory programs that specify the identity of each liver cell type. From the set of 611,576 cCREs, Applicant identified 32,663 (5.3%) cCREs with cell type-specific accessibility profiles, of which 12,410 of these cCREs were linked to a putative target gene (FIG. 2F). Sequence motifs enriched in cell type-specific cCREs revealed key regulators of cell type identity such as HNF4A, RXRA and NR2F1 TF motifs for hepatocytes, SPIB and SPIC motifs for myeloid cells, SOX and GATA motifs for endothelial cells, and FOS, JUN and TEAD motifs for HSCs (FIG. 2G). Cell type-specific cCREs in hepatocytes were significantly and specifically enriched for cALT level associated variants, confirming the prominent role of hepatocytes in MASLD risk (FIG. 11). Cell type-specific cCREs were linked to 7,422 target genes through GRNs, and these target genes overall showed corresponding cell type-specific expression (FIG. 2H). Furthermore, cell type-specific target genes were enriched for key process related to cellular function and identity, such as cholesterol transport and glycine metabolism in hepatocytes (FIG. 2H).
[0171] Together, these results provide an expanded annotation of gene regulatory programs of cell types in the human liver which facilitates annotation of variants involved in MASLD risk.
[0172] Changes in liver cell type gene regulation associated with MASLD
[0173] Given a comprehensive definition of gene regulatory programs in liver cell types, Applicant next determined how cell type regulatory programs change across progression to MASH. Applicant observed marked changes (FDR<.10) in gene expression, chromatin accessibility, and histone modification profiles across MASLD stages compared to normal livers in hepatocytes, cholangiocytes, HSC, endothelial, and myeloid cells (FIG. 3A, FIG. 13,). In each cell type, the number of gene or epigenomic features with significant changes in activity compared to normal livers was substantially more pronounced in MASH and MetALD than for MASL (FIG. 3A). Moreover, within MASH / MetALD livers, significant changes were more extensive in high fibrosis (Fib+) MASH compared to low fibrosis (Fib-) MASH (FIG. 3A, FIG.13), highlighting fibrosis as a major cause of disease-associated changes.
[0174] Notably, Applicant observed a larger number of significant H3K27ac and H3K27me3 features with altered activity in hepatocytes of Fib+ MASH livers compared to other cell types (FIG. 3A). Applicant determined the relationship between changes in genomic modalities in hepatocytes in Fib+ MASH livers, which revealed overall positive correlation in changes in gene expression, accessible chromatin and H3K27ac activity and a negative correlation with the repressive mark H3K27me3 (FIG. 3B). For example, across the DTNA locus there was a loss of repression in hepatocytes in Fib+ MASH, which coincided with increased cCRE activity and theAty. Dkt. No.: 114198-7260 expression of DTLA (FIG. 3B). For example, across the DTNA locus there was a loss of repression in hepatocytes in Fib+ MASH, which coincided with increased hepatocyte cCRE activity and the expression of DTNA, a component of the dystrophin-associated protein complex (DPC) that binds cytoskeleton and extracellular matrix (FIG. 3B).
[0175] Applicant identified pathways enriched among genes with altered expression in MASH compared to normal livers (FIG. 3C). Hepatocytes in MASH had increased expression of pathways related to epithelial remodeling, such as cytoskeletal structure and cell-ECM interactions, although they did not contribute to deposition of fibrillar collagens or ECM (FIG.14), as well as pathways related to glycolysis, hypoxia, senescence, cytokine signaling, cell death, and signaling responses (FIG. 3C). In turn, MASH hepatocytes exhibited reduced expression of genes in fatty acid and cholesterol related pathways. In MASH livers, HSCs upregulated Hedgehog signaling, Ras signaling, fiber formation, collagen degradation, and matrix metallothionein activity, while endothelial cells had increased ECM interactions and decreased inflammatory signaling (data not shown). When further characterizing changes in hepatocytes in Fib+ and Fib-MASH livers, the strong enrichment in cytoskeletal and cell-ECM interaction pathways (FIG. 3C) observed in MASH livers could be attributed to Fib+ MASH, while these pathways had either less pronounced induction or no effect in Fib- MASH hepatocytes (FIG. 3C). In turn, Fib- MASH livers showed distinct up-regulation of fatty acid metabolism, synthesis, and oxidation pathways compared to normal livers.
[0176] As both MASH and MetALD are caused by metabolic liver injury with exacerbated hepatic inflammation and fibrosis due to chronic alcohol consumption in the latter, Applicant next assessed their molecular differences and similarities. Despite profiling fewer MetALD livers, Applicant identified more genes and cCREs overall with significant changes in activity in MetALD compared to MASH (FIG. 3A). Changes in individual gene and pathway expression in MASH and MetALD livers were highly correlated across cell types (all ^>.5, P<2xl0'16), although gene effects were overall more pronounced in MetALD (hepatocytes binomial p=2xl0‘16). Among specific pathways, MetALD hepatocytes had more pronounced activity of cell-ECM interaction pathways while MASH hepatocytes showed increased interleukin signaling, glycolysis, oxidative phosphorylation, and respiration. In MetALD HSCs, there was increased cell-ECM interaction pathways, while MASH HSCs had increased oxidative phosphorylation, renin-angiotensin system (RAS) processing, translation and metabolism pathways (data not shown).
[0177] While MASL showed few significant changes in expression of individual genes compared to normal livers, numerous pathways were broadly altered in MASL in hepatocytesAty. Dkt. No.: 114198-7260 and other cell types (FDR< 10) (FIG. 3C, FIG. 13). There was significant correlation in effects on gene expression in MASL and MASH in non-parenchymal cells (including HSCs, endothelial cells, and cholangiocytes) (rho>.3), yet hepatocytes showed little to no correlation (rho=.05), highlighting the distinct gene-regulatory programs of MASL and MASH hepatocytes. When separating MASH into Fib+ and Fib- livers, there was a modest correlation in effects between MASL and Fib- MASH (rho=.19) and very low correlation with Fib+ MASH (rho=.03) (FIG.3D) Further, MASL and MASH exhibited negative correlation (rho=-.30) in pathway enrichments compared to normal livers. In contrast to MASH, hepatocytes in MASL up-regulated faty acid metabolism and biosynthesis and did not show increased epithelial remodeling, inflammation, and fibrotic signaling (FIG. 3C). Overall, this supports the concept that MASL and MASH are driven by divergent gene expression programs in hepatocytes. In Fib- MASH livers, hepatocytes retained characteristics of MASL hepatocytes while also sharing some features with Fib+ MASH hepatocytes (FIG. 3C), suggesting it may represent a transitional state between MASL and advanced fibrosis MASH.
[0178] Next, Applicant examined epigenomic changes in hepatocytes associated with progression to MASH. There were 18,057 and 49,422 cCREs with significant (FDR< 10) changes in activity in Fib+ MASH and fibrosis score, respectively, where these cCREs had concordant changes in both chromatin accessibility and H3K27ac signal (FIG. 3E, FIG. 15). Clustering of MASH-associated cCREs revealed multiple sets of up-regulated cCREs, including those altered starting in MASL and others altered only in MASH, as well as down-regulated cCREs (FIG. 3E, FIG. 15). To further understand cCRE dynamics, Applicant defined chromatin states separately in MASLD stages and categorized cCREs by state across progression to MASH. Hepatocyte cCREs with increased and decreased accessibility in higher fibrosis score livers exhibited a progressive gain and loss of active chromatin states, respectively, as well as, in the former, loss of repressive states (FIG. 3F, FIG. 16). Applicant also observed chromatin state changes in hepatocytes from normal to MASL livers, including subsets of cCREs in open states in normal livers that transitioned to active states in MASL and vice versa (FIG. 11).Applicant next identified the gene targets of MASH-associated cCREs using chromatin loops and TF GRNs in hepatocytes (FIG. 3G). For example, distal cCREs at the 9q33 and 1 q21 loci with decreased hepatocyte accessibility and H3K27ac activity in Fib+ MASH interacted with the promoter region of SCAI and RORC, respectively, both of which had reduced expression in MASH (FIG. 3G), and where the latter is involved in insulin signaling and lipid metabolism.
[0179] To identify upstream regulators of cCRE activity in MASLD, Applicant identified TF sequence motifs enriched in cCREs with altered activity in MASLD. In hepatocytes, there wereAty. Dkt. No.: 114198-7260 distinct motif enrichments among the different clusters of cCREs with altered activity in Fib+ MASH and high fibrosis score (FIG. 3E). For example, cCREs with up-regulated activity starting in MASL were enriched for HNF4A, NR4A1, CEBPA / D TF motifs. Similar motifs including NR4A1 / 2 and CEBP were also enriched in cCREs that transitioned to active chromatin states in MASL from normal livers while, conversely, MAF and CTCF motifs were enriched in cCREs that transitioned from active to open states in MASL (FIG. 16). In comparison, cCREs up-regulated only in MASH and MetALD were enriched in FOS / JUN, TEAL) and NFKB motifs, and cCREs down-regulated in MASH were enriched for CTCF, KLF, SP, ELK and ETV motifs (FIG. 3E). Together these findings further support that MASL and MASH in hepatocytes are characterized by distinct TFs and regulatory programs. Applicant also observed TF motifs enriched in cCREs altered in other cell types. In MASH HSCs, cCREs with increased accessibility were enriched for TEAL), RUNX, FOS / JUN, ATF3 and hormone receptor motifs (GRE, ARE), previously implicated in HSC activation (R. Isaac, et al., TM7SF3 controls TEAD1 splicing to prevent MASH-induced liver fibrosis. Cell Metab 36, 1030-1043. e7 (2024)., H. Y. Kim, et al., Multi-modal analysis of human hepatic stellate cells identifies novel therapeutic targets for metabolic dysfunction-associated steatotic liver disease. J Hepatol 82, 882-897 (2025).), while cCREs with decreased accessibility were enriched for farnesoid X receptor (FXR) and ETS / ETV motifs, associated with maintenance of quiescence in HSCs (S. Fiorucci, et al., The nuclear receptor SHP mediates inhibition of hepatic stellate cells by FXR and protects against liver fibrosis. Gastroenterology 127, 1497-1512 (2004)., X. Liu, et al., Identification of Lineage-Specific Transcription Factors That Prevent Activation of Hepatic Stellate Cells and Promote Fibrosis Resolution. Gastroenterology 158, 1728-1744. el4 (2020).) (data not shown). In MASH endothelial cells, cCREs with increased activity were enriched for FOS / JUN, TCF, and MEF motifs, while cCREs with decreased activity were enriched for IRF motifs (data not shown).
[0180] Applicant next sought to further define transcriptional drivers of progression to MASH in hepatocytes by identifying TF GRNs with altered activity across MASLD states. TF GRNs with increased activity in MASL included HNF4A, RORA, and CUX2 (FIG. 3H, FIG. 31). These MASL-enriched GRNs were highly connected to genes with increased expression in MASL and involved in lipid biosynthesis (SMPD3, LPIN3, SLC25A42), binding (PSD4), transport (SLC22AP), and metabolism (CRYLP) (FIG. 31). Applicant also identified GRNs with increased activity in MASH, which were highly distinct from those in MASL, including CREB5, NFKB1 / RELB, KLF6, TEAD1, ETS1, and NFAT5 (data not shown). MASH GRNs each regulated genes with increased expression in MASH including those involved in cytoskeletalAty. Dkt. No.: 114198-7260 organization, adhesion and ECM interactions (DTNA, AKAP12, ITGA2, LAMA3, MY07B, CD44), inflammation (TNFAIP2), autophagy (TMEM150B), and lipid metabolism (FGF1) (FIG.3H, FIG. 31). The overall activity of TF GRNs was similar in MASH and MetALD, although MetALD showed relatively higher activity of HNF4G, STAT1, and PPARG GRNs (FIG. 3H,).
[0181] Stratifying MASH livers into Fib+ and Fib- groups revealed additional dynamics of TF activity and their associated gene networks (FIG. 3H, FIG. 31). For example, MASL hepatocytes uniquely induced HNF4A, NR1I2, CUX2, and RORA GRNs, while Fib- MASH hepatocytes shared some of these GRNs, uniquely upregulated FOXO1, NR4A1, and FOS, and activated GRNs characteristic of Fib+ MASH hepatocytes such as CREB5, ETS1, RUNX1, CUX1, RELB, KLF6, TEAD1, and ATF3. In turn, Fib+ MASH hepatocytes uniquely upregulated HNF4G, PPARG, NFKB, NFAT5 and STAT1 and downregulated PPARA and HNF4A GRNs, highlighting fibrosis as a primary driver of regulatory changes in MASH.Applicant further grouped GRNs by activity across MASLD stages which revealed distinct clusters (cl-c5) of GRNs with increased activity in MASL and Fib-MASH, Fib- / Fib+ MASH, Fib- MASH only, and Fib+ MASH only (FIG. 3H). Genes in GRNs with altered activity in MASL and MASH were enriched (FDR< 10) in pathways with corresponding changes in MASL and MASH expression, highlighting drivers of altered pathway activity in MASLD (FIG. 3H).For example, GRNs with increased activity in MASL and Fib- MASH, such as HNF4A and RORA, were associated with regulation of lipid metabolism / synthesis; GRNs active in both Fib-and Fib+ MASH, including KLF6, ETS1, and CREB5, regulated epithelial remodeling processes such as cytoskeletal organization and cell-ECM adhesion; and GRNs with highest activity in Fib+ MASH, such as NFKB and RELB, regulated inflammation and immune signaling.
[0182] Finally, to understand genetic risk in GRNs associated with progression to MASH, Applicant tested for enrichment of MASLD (cALT) associated variants in hepatocyte GRNs. In hepatocytes, variants showed enrichment for GRN clusters with increased activity in MASL and Fib- MASH (GRN c4) and with increased activity in both Fib- and Fib+ MASH (GRN cl), with limited enrichment in other GRN clusters (FIG. 3J, FIG. 17). Specific TF GRNs annotated with MASLD variants included those increased in MASL and Fib- MASH (e.g. HNF4A, RORA, CUX2), increased in Fib- and Fib+ MASH (ETS, KLF6, ATF3), and decreased across all MASLD stages (THRB, ONECUT1). Applicant identified multiple clusters of TF GRNs altered in MASL and Fib- MASH based on overlapping MASLD variants (FIG. 3J), and these GRNs were linked to multiple genes (MYCLD, MTTP.:APOL2, MLXIP) in pathways such as lipid metabolism and transport.Aty. Dkt. No.: 114198-7260
[0183] Overall, these results identify transcriptional drivers of progression to MASH in liver cell types and show that hepatocytes exhibit distinct gene regulatory programs between MASL and MASH with additional genomic changes associated with severe fibrosis in MASH.
[0184] MASLD-associated variants affecting hepatocyte regulatory programs
[0185] Applicant next sought to identify specific loci and genes affecting MASLD risk in hepatocytes by combining quantitative trait locus (QTL) mapping, variant functional predictions, and 3D chromatin architecture.
[0186] Applicant performed QTL mapping of chromatin accessibility (caQTLs), histone modifications (hQTLs), and gene expression (eQTLs) levels in liver cell types across all 86 donors using TensorQTL (A. Taylor-Weiner, et al., Scaling computational genomics to millions of individuals with GPUs. Genome Biol 20, 228 (2019).) (data not shown). In total, Applicant identified significant QTLs (FDR< 10) for each modality, including 10,591 caQTLs, 1,644 eQTLs, 418 H3K27ac hQTLs, and 391 H3K27me3 hQTLs across all cell types (FIG. 4A, FIG.18). Across modalities, most QTLs were identified in hepatocytes, with fewer in less common cell types (FIG. 4A). Cell type caQTLs and hQTLs disrupted distinct TF sequence motifs, including HNF, FOX, PPAR, CUX, RXR, and CEBP motifs in hepatocytes, CEBP, STAT, IRF, SPIC, and ETV TFs in myeloid cells, FEV, FLU, ETS and KLF motifs in endothelial cells, and ETS and FOS motifs in HSCs (FIG. 4B). Overall, caQTLs and hQTLs had highly cell typespecific effects compared to eQTLs which were more often shared across cell types (FIG. 4C).Furthermore, caQTLs / hQTLs and eQTLs at the same locus were largely distinct based on statistical colocalization (C. Giambartolomei, et al., Bayesian Test for Colocalisation between Pairs of Genetic Association Studies Using Summary Statistics. PLOS Genetics 10, el004383 (2014).), suggesting they captured broadly different classes of variant regulatory effects (FIG.4C).
[0187] Applicant next annotated MASLD-associated loci that affect hepatocyte regulatory programs. For these analyses Applicant again utilized genetic association data for cALT levels, a proxy for liver damage and MASLD, due to the large number of established loci. Variants at over a third of all known cALT (MASLD) loci (28 / 77) were QTLs in hepatocytes, of which 18 were colocalized with QTL associated variants (FIG. 4D). The majority of MASLD loci with QTLs (26 / 28) were caQTLs, including 11 and 15 loci where risk alleles increased and decreased cCRE activity, respectively (FIG. 4D). Next, Applicant prioritized target genes of loci affecting hepatocyte cCRE activity using eQTLs. Several MASLD loci colocalized with eQTLs, forAty. Dkt. No.: 114198-7260 example at the 2q37 locus where risk alleles increased cCRE accessibility, H3K27ac signal, and EFHD1 expression in hepatocytes (FIG. 4D, FIG. 4E, FIG. 19).
[0188] Most loci did not have colocalized eQTLs, however, and we utilized chromatin loops in hepatocytes to annotate more distal target genes of MASLD loci. Overall, Applicant linked 12 MASLD loci with caQTLs to putative target genes in hepatocytes using chromatin loops (FIG.4E), including those not previously prioritized as candidate genes (M. Vujkovic, et al., A multiancestry genome-wide association study of unexplained chronic ALT elevation as a proxy for nonalcoholic fatty liver disease with histological and radiological validation. Nat Genet 54, 761-771 (2022).). For example, at the 12q 13 locus, the hepatocyte cCRE affected by MASLD variants interacted with the promoter region of KRT8 KRT18. which are keratins that form the intermediate filament cytoskeleton in hepatocytes (FIG. 4F). Hepatocyte chromatin state at this locus further supports KRT8 / KRT18 as candidate genes, as other previous implicated genes such as KRT78 are in broad regions of repression (FIG. 4F). At the 8p23 locus, the hepatocyte cCRE affected by MASLD variants interacted with the PPP1R3B promoter over 200kb distal (FIG.4G). Additional interactions between associated cCREs and distal target genes included the 19q 13 locus where cCREs interacted with the CEBPA promoter (FIG. 20), and the 22ql2 locus where cCREs interacted with the XBP1 promoter.
[0189] Finally, Applicant prioritized candidate causal variants underlying genetic effects on hepatocyte cCRE activity at MASLD loci. Applicant created sequence-based models of hepatocyte chromatin accessibility using chromBPNet (A. Pampari, et al., ChromBPNet: bias factorized, base-resolution deep learning models of chromatin accessibility reveal cis-regulatory sequence syntax, transcription factor footprints and regulatory variants. [Preprint] (2024). doi.org / 10.1101 / 2024.12.25.630221.), and then used models to predict the functional effects variants at each locus with a caQTL. At almost all (24 / 26) MASLD loci with hepatocyte caQTLs, at least one fine-mapped variant was predicted to have functional effects on hepatocyte cCRE activity (FIG. 4E). For example, at the 12q 13 locus, the most likely causal variant rs9668670 (PPA=.6O) was predicted to disrupt hepatocyte cCRE activity, where the MASLD risk allele of this variant increased hepatocyte cCRE activity and had higher predicted accessibility (FIG. 4F). At the 8p23 locus, likely causal variant rs4841132 (PPA=.99) was predicted to disrupt hepatocyte cCRE activity where the MASLD risk allele decreased cCRE activity and had lower predicted accessibility (FIG. 4G). At both loci, candidate variants mapped within sequence motifs for HNF1A, a key regulator of hepatocytes, among other TFs (FIG. 4F, FIG. 4G)Aty. Dkt. No.: 114198-7260
[0190] Overall, these results provide detailed annotation of MASLD loci directly affecting hepatocyte regulation and candidate functional variants and target genes of variant activity at these loci.
[0191] Changes in hepatocyte sub-clusters in progression to MASH
[0192] Applicant finally sought to understand cellular heterogeneity within hepatocytes and the role of hepatocyte heterogeneity in progression to MASH. Applicant identified five distinct subclusters of hepatocytes, which Applicant annotated based on multiple lines of evidence including gene expression profiles, spatial localization, and abundance in MASLD (FIGS. 5A-5D). This included three hepatocyte sub-clusters based on zonation within the liver including Zone 1 (Periportal), Zone 2 (Mid-lobular), and Zone 3 (Central-lobular) cells. For example, Zone 1 hepatocytes had increased expression of marker genes (G6PC, GLS2, PCK!) and pathways related to insulin signaling and localized around the portal vein, while Zone 3 hepatocytes had increased expression of known markers (CYP2E1, ADH4) and lipid metabolism genes and pathways and localized around the central vein (FIG. 5B, FIG. 5C). Hepatocyte zonation was preserved in MASL and Fib- MASH hepatocytes, like that observed in normal livers, although there were increased Zone 1 and Zone 3 hepatocytes in MASL (FIG. 5D). Applicant also observed altered abundance of hepatocyte zones in MASH livers, most notably a decrease in Zone 1 and 3 hepatocytes in Fib+ MASH and higher fibrosis score (FIG. 5C, FIG. 5D).
[0193] Outside of zonation-based heterogeneity Applicant observed two additional sub-clusters of hepatocytes. The first sub-cluster preferentially expressed genes implicated in hepatocyte proliferation, identity and function such as BROXL XBBL CBP2, as well as pathways related to proliferation, metabolism, inflammation and lipid transport. This sub-cluster had generally low abundance (mean=1.2%), with several exceptions, and had significantly increased proportion in MASH compared to normal livers, and therefore Applicant termed this sub-type ‘MASH-associated’ (MASH-a) hepatocytes (FIGS. 5A-5D). The second sub-cluster preferentially expressed many inflammatory and stress-related genes (NFKB1, STAT4.:HII'1A.:SO 1)2.:TEAD1, TGSBR2) and inflammation, cell organization and structure, and lipid transport pathways (FIG.5B) This sub-cluster was present in all livers (mean=l 1.1%) and had significantly increased proportion in Fib+ MASH and severe fibrosis compared to normal livers (FIG. 5D). Applicant also observed localization of this sub-cluster near areas of liver fibrosis in MASH compared to normal and MASL livers (FIG. 5C). Therefore, Applicant termed this sub-type ‘fibrosis-associated’ (Fib-a) hepatocytes There was an inverse correlation between Fib-a hepatocyte proportion and Zone 3 (rho=-.47, p=7.2xl0'6), and to a lesser degree Zone 1 (rho=-.32, p=.0022) across livers, with no relationship with Zone 2 hepatocytes (rho=.05).Aty. Dkt. No.: 114198-7260
[0194] Next, we determined the transcriptional regulatory programs driving hepatocyte subcluster identity. Applicant identified 1,010 hepatocyte cCREs with accessible chromatin profiles highly specific to each hepatocyte sub-cluster, and these cCREs had concordant specificity in H3K27ac activity (FIG. 5E). Interestingly, Applicant also observed a decrease in repressive H3K27me3 activity in Zone 3 cells among the cCREs with Zone 3-specific activity. Zonespecific cREs in hepatocytes were enriched for specific TF motifs such as GATA and TCF7 motifs in Zone 3 hepatocytes, and FOXA motifs in Zone 2 hepatocytes. The MASH-a and Fib-a specific cCREs were also enriched for TF sequence motifs, where Fib-a hepatocytes showed enrichment for many motifs including FOS / JUN, NFKB, TEAD, ELF, ETS1, and ATF family TFs and MASH-a hepatocytes were enriched for FOXA, CEBP, PPAR and steroid hormone (GR, AR) motifs (FIG. 5E). Applicant further identified TF GRNs with increased activity in each hepatocyte sub-cluster which implicated similar TFs driving zonation identity including HNF4A and GATA4 in Zone 3, FOXA1, PPARA and RXRA in Zone 2, and FOXO1 and NR4A1 in Zone 1 hepatocytes. Numerous TF GRNs were also highly active in Fib-a hepatocytes compared to other hepatocyte sub-types including NFKB1 / REL, TEAD1, ETS1, ELK1, KLF6, CREB5, CUX1, and ATF3, among others (FIG. 5F).
[0195] Changes in hepatocyte heterogeneity in progression to MASH
[0196] Applicant sought to define genomic changes within each hepatocyte sub-cluster in progression to MASH. Overall, there were dramatic changes in each sub-cluster in MASH / MetALD as well as in Fib+ MASH, and few changes in MASL and Fib- MASH, compared to normal livers (FIG. 5G). Zone 3 and Fib-a hepatocytes showed the largest number of changes in gene expression and cCRE activity in Fib+ MASH compared to other sub-clusters (FIG. 5G). Notably, Applicant also observed a substantial number of changes in the repressive mark H3K27me3 across genomic regions (15kb bins) in Fib-a hepatocytes, suggesting broadly altered repression profiles within hepatocytes in Fib+ MASH. Defining the chromatin state of H3K27me3 bins altered in Fib+ MASH further confirmed an increase in repression in fibrosis-associated hepatocytes (FIG. 5G). Specific biological pathways were enriched in genes overlapping regions with altered H3K27me3 activity in Fib-a hepatocytes, with increased repression at lipid metabolism, transport, and Wnt signaling pathways leading to decreased expression, and decreased repression at cell structure, adhesion, ECM interaction and cytokine signaling genes leading to their activation (FIG. 5H). The Fib-a hepatocytes also had the largest increases in expression of genes within selected cell structure, adhesion and ECM pathways in Fib+ MASH relative to other sub-clusters (FIG. 5H).Aty. Dkt. No.: 114198-7260
[0197] Applicant next characterized changes in cCRE activity in MASH within each hepatocyte sub-cluster. Applicant identified TF sequence motifs enriched in cCREs with up- and down-regulated activity in Fib+ MASH and identified motifs with differential enrichment across hepatocyte sub-clusters. For example, among cCREs up-regulated in Fib+ MASH, there was stronger enrichment of NFKB -related motifs as well as motif for oxidative stress-related TFs such as NFE2L2, NF-E2 and NRF2 in Fib-a and Zone 3 hepatocytes (FIG. 51). By comparison, there was stronger enrichment of ATF motifs as well as several FOS / JUN-related motifs in Zone 1 hepatocytes in Fib+ MASH (FIG. 51). Applicant observed similar patterns among cCREs down-regulated in Fib+ MASH, with strong enrichment of SP and KLF TF family motifs in Fib-a and Zone 3 hepatocytes, and enrichment of STAT motifs in Fib-a hepatocytes, and few enriched motifs in other cell states.
[0198] The abundance of Fib-a hepatocytes inversely correlated with Zone 3 and Fib-a hepatocytes repressed Zone 3 -related processes in MASH, and Applicant hypothesized that Fib-a hepatocytes may preferentially arise from Zone 3 cells. Applicant ordered hepatocytes in Fib+ MASH livers on a trajectory starting from Fib-a cells using pseudo-time analyses (see Methods), which revealed that Zone 3 cells were generally ordered more closely to Fib-a cells compared to Zone 1 and 2 cells (FIG. 5J). Applicant next ordered the expression of genes up-regulated in hepatocytes in Fib+ MASH, including cytokines (CXCL8, CXCL6), heat shock proteins (HSPB8), integrins (ITGB6) senescence-related genes (CDKN2A), and keratins (KRT2S), along this trajectory. Overall, genes up-regulated in Fib+ MASH showed a gradient of expression that was highest in Fib-a and Zone 3 hepatocytes. Notably, these genes had particularly high expression within a subset of Fib-a hepatocytes. Furthermore, the hepatocytes expressing these genes most up-regulated in Fib+ MASH showed highly specific localization around the areas of fibrotic injury in Fib+ MASH livers (FIG. 5K). Applicant thus termed this subset of Fib-a hepatocytes adjacent to fibrotic injury in Fib+ MASH as ‘Fib-adj’ hepatocytes.
[0199] Finally, Applicant hypothesized that changes in hepatocytes in MASH may be driven through altered signaling from cells in fibrotic areas in the liver. Based on neighborhood analysis, Fib-adj hepatocytes were specifically localized to cell types in fibrotic areas of the liver such as HSCs and immune cells in Fib+ MASH, while other hepatocyte sub-clusters mapped to different areas of the liver (FIG. 5L). Applicant performed cell-cell interaction analyses to identify signals altered between cell types in Fib+ MASH livers using CellChat (S. Jin, et al., Inference and analysis of cell-cell communication using CellChat. Nat Commun 12, 1088 (2021)., S. Jin, et al., CellChat for systematic analysis of cell-cell communication from singlecell transcriptomics. NatProtoc 20, 180-219 (2025).). Overall, there was evidence for signalingAty. Dkt. No.: 114198-7260 from HSCs specifically to Fib-adj hepatocytes relative to other hepatocyte sub-clusters, while hepatocyte zones had increased signaling amongst each other as well as from endothelial cells (FIG. 5M). The strongest signal from HSCs to Fib-adj hepatocytes involved collagen (COLIAF) and CD44, which is a cell surface adhesion molecule implicated in fibrogenic injury response and hepatocellular carcinoma (P. Govindaraju, et al., CD44-dependent inflammation, fibrogenesis, and collagenolysis regulates extracellular matrix remodeling and tensile strength during cutaneous wound healing. Matrix Biol 75-76, 314-330 (2019)., J. Zhang, et al., CD44 promotes hepatocellular carcinoma progression via upregulation of YAP. Exp Hematol Oncol 10, 54 (2021).) and is up-regulated in hepatocytes in MASH, while other hepatocytes had signaling to other receptors.
[0200] Discussion
[0201] Single cell studies of normal and MASLD livers have largely consisted of gene expression profiling from relatively few donors (P. Ramachandran, et al., Resolving the fibrotic niche of human liver cirrhosis at single-cell level. Nature 575, 512-518 (2019)., Z.-Y. Wang, et al., Single-cell and bulk transcriptomics of the liver reveals potential targets of NASH with fibrosis. Sci Rep 11, 19396 (2021)., T. S. Andrews, et al., Single-Cell, Single-Nucleus, and Spatial RNA Sequencing of the Human Liver Identifies Cholangiocyte and Mesenchymal Heterogeneity. Hepatol Commun 6, 821-840 (2022).-Y. Xiao, et al., Hepatocytes demarcated by EphB2 contribute to the progression of nonalcoholic steatohepatitis. Sci Transl Med 15, eadc9653 (2023).). Combining multiple single cell epigenomic modalities, including accessible chromatin, active and repressive histone modifications, and 3D chromatin interactions, with single cell and spatial transcriptomics across many human livers revealed unprecedented insight into the regulatory programs of liver cell types as well as transcriptional drivers of progression to MASH. Compared to cCREs defined from accessible chromatin alone, incorporating histone modifications annotated chromatin state in liver cell types for a much larger proportion of the human genome. Furthermore, chromatin state and chromatin interactions provided functional context to cCRE activity and revealed distinct properties and disease risk enrichments of cCREs in different chromatin states within liver cell types. The additional epigenomic modalities also enabled an improved annotation of genetic loci associated with MASLD risk in liver cell types, including candidate variants not annotated in previous cCRE catalogs and distal target genes of risk variant activity. Collectively, multimodal profiling enabled a more comprehensive understanding of how liver cell types are regulated and the regulatory programs driving MASLD risk and pathogenesis.Aty. Dkt. No.: 114198-7260
[0202] Only -20% of patients with MASL ultimately develop MASH, and the factors driving progression from MASL to MASH are not well understood. Hepatocytes respond first to metabolic injury early in disease and are key triggers of changes in HSCs and other cell types that lead to MASH and fibrosis. Our results revealed significant, stepwise differences in the genomic profiles of hepatocytes in MASL, low fibrosis MASH, and high fibrosis MASH, and uncovered key transcriptional drivers of these distinct profiles. Hepatocytes in MASL were marked by increased lipid metabolism, synthesis and uptake as well as increased activity of TF GRNs regulating these processes such as HNF4A, RORA and CUX2. In low fibrosis MASH, lipid-related pathways and GRNs remained active, while new programs associated with epithelial remodeling and adhesion regulated by TFs such as TEAD1, KLF6, ATF3, CREB5, and RELB were also activated. Finally, the transition to high fibrosis MASH resulted in a reduction in lipid metabolism and associated TF GRNs with a marked increase in inflammatory and fibrotic pathways regulated by TFs such as NFKB and NFAT5. Among these changes, lipid-related networks increased in MASL and Fib-MASH were most enriched for MASLD-associated variants, therefore linking genetic risk to processes up-regulated in the early stages of disease. Applicant speculate that a more beneficial hepatocyte response to lipids may prevent further exacerbation of disease to fibrosis, which may be leveraged to predict progressors, identify biomarkers and inform therapeutic discovery.
[0203] Severe fibrosis in the liver was also marked by dramatic transitions in cellular sub-types of hepatocytes, including a loss of hepatocyte zonation and an increase in specific fibrosis-associated cells. Cellular heterogeneity in hepatocytes in MALSD has been previously reported, although largely at the transcriptional level (P. Ramachandran, et al., Resolving the fibrotic niche of human liver cirrhosis at single-cell level. Nature 575, 512-518 (2019)., T. Tsuchida, S. L. Friedman, Mechanisms of hepatic stellate cell activation. Nat Rev Gastroenterol Hepatol 14, 397-411 (2017)., Y. Xiao, et al., Hepatocytes demarcated by EphB2 contribute to the progression of nonalcoholic steatohepatitis. Sci Transl Med 15, eadc9653 (2023).-N. Aizarani, et al., A human liver cell atlas reveals heterogeneity and epithelial progenitors. Nature 572, 199-204 (2019).). Our results revealed transcription factors and regulatory networks underlying zonationbased identity and other cellular heterogeneity in hepatocytes as well as transcriptional drivers of changes within hepatocyte zones and sub-types in MASH. Notably, the fibrosis-associated hepatocyte population had dramatically altered repressive domains marked by H3K27me3, including increased repression at lipid metabolism genes and loss of repression at epithelial remodeling and fibrosis-related genes. As this population was most closely related to Zone 3 hepatocytes, which themselves exhibited large genomic changes in MASH, Applicant speculateAty. Dkt. No.: 114198-7260 that epigenetic remodeling drives a transcriptional switch from Zone 3 cells toward the fibrosis-associated state. The strong impact of fibrosis on transcriptional and epigenetic changes in hepatocytes was further illustrated by the localization of a specific subset of hepatocytes with dramatic changes in severe fibrosis directly at the sites of fibrosis. This hepatocyte subset also exhibited distinct cell-cell interactions with HSCs, notably involving CD44 which is also implicated in hepatocellular carcinoma (L. A. Rodriguez, et al., Hepatocellular Carcinoma in Metabolic Dysfunction- Associated Steatotic Liver Disease. JAMA Netw Open 7, e2421019 (2024).), and downstream signaling from these interactions may help drive transition to the fibrosis-associated state.
[0204] Combining cell type-specific gene regulatory programs in the liver with genetic association data revealed mechanisms of individual loci associated with MASLD. At a large fraction of MASLD loci, Applicant detailed the regulatory logic of risk variant activity in hepatocytes, implicating specific variants, cCREs, transcriptional regulators, and genes as causally involved in disease. Through both QTL mapping and sequence-based models, Applicant also uncovered the direction of effect of variants on hepatocyte regulatory activity, which may help inform strategies to modify element or gene activity to protect hepatocytes during disease. Chromatin loops highlighted candidate genes in hepatocytes not prioritized in a previous study (M. Vujkovic, et al., A multiancestry genome-wide association study of unexplained chronic ALT elevation as a proxy for nonalcoholic fatty liver disease with histological and radiological validation. Nat Genet 54, 761-771 (2022).), including those distal to risk variants, and revealed novel targets for MASLD. Candidate genes included PPP1R3B, where knockout in both mice and humans is associated with increased steatosis and hepatic fat (K. T. Creasy, et al., PPP1R3B is a metabolic switch that shifts hepatic energy storage from lipid to glycogen. [Preprint] (2023). doi.org / 10.1101 / 2023.03.04.529958.), and CEBPA, a key hepatic TF involved in glucose and lipid metabolism implicated in liver fibrosis (T. Yan, et al., Hepatocyte-specific CCAAT / enhancer binding protein > restricts liver fibrosis progression. J Clin Invest 134, el 66731 (2024).), as well as other genes such as KRT8 / 18 and XBP1 involved in hepatocyte function and
[0205] In summary, our multimodal study provided the most comprehensive annotation to date of gene regulatory programs of liver cell types and cellular heterogeneity which enabled new mechanistic insight into drivers of MASLD progression and uncovered novel targets for therapeutic discovery.
[0206] Experiment No. 2:Aty. Dkt. No.: 114198-7260
[0207] Materials and methods
[0208] Donor collection and information
[0209] Deidentified livers (IRB 171883XX, certified by HRPP Director and IRB Chair as not human subjects according to the Code of Federal Regulations, Title 45, part 46 and UCSD Standard Operating Policies and Procedures) from 87 donors were obtained via Lifesharing OPO, which provided the informed consent, laboratory tests (ALT, AST, biochemistry, cell counts, liver biopsy, serology, and others), as well as patient history (cause of death, age, BMI, gender, and underlying diseases, history of alcohol consumption). Livers were graded by a pathologist using a double-blinded method and identified as normal, MASL, and MASH. Based on the donor history, biochemical, and histological evaluation, livers were graded as normal, MASL or MASH. MASH Clinical Research Network (MASH CRN, 2005) methodology was used for evaluation of livers including the degree of steatosis, steatohepatitis, inflammation, and fibrosis. A NASC / CRN score of 5 was used as the cut off for diagnosis of MASH. Livers with scores of <3 were diagnosed as not having MASH. Representative images of normal, MASL, and MASH livers are shown in FIG. 9.
[0210] Single cell multiome assays
[0211] Snap frozen liver samples (5x5 cm2) from each donor were used for single cell assays. For each pool, Applicant selected between 15 to 30 donors to represent all MASLD groups and sexes. For each donor, approximately 20-30 mg of snap frozen tissue (tissue samples measuring approximately 3 mm X 3 mm X 3mm) was sectioned on dry ice using pre-chilled scalpels and collected in a pre-chilled 5ml LoBind tube on dry ice. Sectioned tissue pools were stored at -80°C freezer. The nuclei suspension was then filtered through 100-pm and 30-pm CellTrics™ Filters and centrifuged for 5 min at 500g at 4 C. Anti-Nucleus MicroBeads (Miltenyi Biotec) were used for debris removal and enrichment of nuclei. The nuclei suspension was centrifuged for 5 min at 500g at 4 C. After resuspension in sort buffer (IX PBS, IX protease inhibitor, 1.5 U pl- RNasin® Ribonuclease Inhibitor and 1% BSA), nuclei were stained with 2 pM 7-AAD in sort buffer for 10 min on ice and were sorted by fluorescence-activated nuclei sorting with an SH800 cell sorter for single nuclei. Nuclei were collected in collection buffer (IX PBS, 5X protease inhibitor, 5 U pE1RNaseOUT, 5 U pE1SUPERaseln inhibitor, and 5% BSA) and centrifuged for 10 min at 500g at 4C. The nuclei pellet was resuspended in permeabilization buffer (ImM DTT, 0.2% IGEPAL-CA630, IX cOmplete EDTA-free protease inhibitor, 1.5 U pE1RNasin and 5% BSA in PBS), incubated on ice for 2 min and centrifuged for 5 min at 500g and 4°C and resuspended in lx nuclei buffer. Suspension of 30,000 permeabilized nuclei wereAty. Dkt. No.: 114198-7260 each loaded onto 8 lanes and processed using of 10X Genomics Controller following manufacturers recommendations. lOx Genomics Epi Multi ome AT AC + Gene Expression assays were processed following manufacturers recommendations. Generated libraries were sequenced at the Institute for Genomic Medicine at UC San Diego using NovaSeq X Plus (Illumina) sequencer.
[0212] Single cell droplet paired-Tag assays
[0213] To obtain joint profiles of histone modifications and transcriptomes from single cells in liver samples, Droplet Paired-Tag was performed as previously described, with minor modifications. For this study, Applicant performed separate assays for antibodies targeting the histone modification marks H3K27ac (Abeam, ab4729, polyclonal) and H3K27me3 (Abeam, ab 192985, recombinant). Briefly, liver tissues from 16-24 different donors were pooled in equal weights, and nuclei were extracted following the same protocol used for our lOx Genomics single-cell multiome data generation. The nuclei were stained with 0.2mg / ml Hoescht and were sorted via fluorescence-activated nuclei sorting using an SH800 cell sorter (Sony) to isolate single nuclei. The isolated nuclei were collected and counted using a cell counter (RWD C100-Pro) with DAPI staining. Applicant aliquoted 500,000 nuclei to initiate reactions in individual tubes for each histone modification mark and replicate. Two replicates were included for each tissue pool and histone modification.
[0214] First, nuclei were permeabilized with OMNI buffer and incubated with 2 pL PA-Tn5 (0.4 mg / mL) and 2 pg antibody in MED#1 buffer overnight at 4 °C. PA-Tn5 and antibodies were then removed by washing with MED#2 buffer. Tagmentation was carried out in MED#2 buffer supplemented with 10 mM MgC12 (Invitrogen, AM9530G) at 550 r.p.m. and 37 °C for 60 minutes in a ThermoMixer (Eppendorf), and the reaction was terminated by adding 2* stop solution. The nuclei were washed in 1 * nuclei buffer and counted again. From each tube, 25,000-40,000 nuclei were aliquoted and used for a single droplet generation reaction with the Chromium Next GEM Single Cell Multiome kit (lOx Genomics, 1000283). Finally, DNA and RNA library amplification was performed according to the Chromium Single Cell ATAC Library kit manual, with the following modifications: the starting amount of preamplification product for DNA library amplification was doubled, 13 amplification cycles were used for the DNA library, and a different SPRI bead size selection were applied for the DNA library (50 pL + 105 pL). The libraries were sequenced at the UCSD IGM using the Illumina NovaSeq X.
[0215] Single cell droplet Hi-C assaysAty. Dkt. No.: 114198-7260
[0216] To obtain cell-type specific chromatin architecture, Applicant performed Droplet Hi-C on all the livers from different donors, as previously described. Liver tissue from 16-24 donors were pooled together, and nuclei were extracted within the same reaction. Of the 87 donors, one donor did not have enough tissue available for pooling and was excluded from Droplet Hi-C pools. Nuclei were fixed with formaldehyde (Sigma, 47608-250ML-F) at a final concentration of 1% or paraformaldehyde (Electron Microscopy Sciences, 15714) at a final concentration of 2%, and then quenched with glycine solution at a final concentration of 0.2 M. Nuclei were permeabilized with lysis buffer on ice. Then, the nuclei were treated with 0.5% SDS (Promega, V6551), incubated at 62 °C for 10 minutes on a ThermoMixer, and then quenched with Triton X-100 (Sigma, 93443). Chromatin in the nuclei was sheared using three restriction enzymes:DpnII (NEB, R0543L), Mbol (NEB, R0147M), and Nlalll (NEB, R0125L), at 37 °C for 90 minutes on a ThermoMixer at 550 r.p.m. The three enzymes were then deactivated at 65 °C for 20 minutes, and the mixture was cooled to room temperature. The sheared chromatin was further ligated using T4 DNA ligase (NEB, M0202L). All enzymes were removed from the nuclei, which were then suspended in 1% BSA in IxPBS with 7AAD (Invitrogen, A1310) and incubated on ice for 1 hour.
[0217] The nuclei were washed three times with 1 x nuclei buffer. Approximately 20,000 nuclei were aliquoted into each PCR tube, and one reaction was performed using the Chromium Next GEM Single Cell ATAC Reagent Kits v2 (lOx Genomics). The tagmentation incubation time was 60 minutes, and the index PCR elongation time was extended from 20 seconds to 1 minute. Double-sided size selection was adjusted to 1.14x SPRIselect to remove only small fragments. For each liver sample pool, Applicant performed at least two replicates, with each replicate involving 1-3 lOx Genomics reactions. The libraries were sequenced at UCSD IGM using the Illumina NovaSeq X.
[0218] Genotyping arrays and imputation
[0219] From all donors with liver single cell profiles, Applicant sectioned 10-15mg of frozen tissue and extracted genomic DNA from each donor using the Monarch Genomic DNA Purification Kit (NEB #T3010L). Samples were eluted in 70-100uL UltraPure Distilled Water (Invitrogen 10977-015), and speed vacuumed to a concentration of 50ng / uL when necessary. Samples were genotyped using the Illumina InfiniumCoreExome-24vl-4_Al array at the UCSD IGM Genomics Center.
[0220] Cluster positions and SNP statistics were extracted from the array data using GenomeStudio v2.0.5. Genotypes were converted to plink format using GenomeStudio v2.0.5Aty. Dkt. No.: 114198-7260 with the plink input report plugin using an hg38 cluster file. Samples from different arrays were merged using plink vl.9, variants were filtered based on a minimum missing call rate of 0.05, minor allele frequency of at least .01, and Hardy-Weinberg equilibrium p-value greater than le-5 before generating VCF and frequency files. Applicant then used the HRC- 1000G-check-bim.pl script from the McCarthy Group Tools (www.chg.ox.ac.uk / ~wrayner / tools / , v4.3.0) to filter and prepare variants for imputation.
[0221] Genotypes were imputed on using TOPMed r3 panel using the TOPMed Imputation Server(l). Imputation was performed using the Array Build Hg38 with r2filter off and phasing using Eagle (v2.4). Following imputation, variants were filtered by imputation quality (r2>0.9 and minor allele frequency (MAF) > 0.01.
[0222] Spatial transcriptomics (Visium HD) assays
[0223] To minimize potential RNAse contamination dissection surfaces were cleaned with 70% EtoH sterile water was used for floating sections and clean new blades were use for microtome sectioning. FFPE blocks were hydrated for 5-10min on top of ice chilled flat metal surface and sections were cut at 5um thickness mounted to a Superfrost™ Slides (Fisher Scientific). Slides were dried for 2 hours in the oven at 42C after sectioning and stored at room temperature with desiccators. All sections were used for experiments within 2 weeks of mounting. To assess the RNA integrity of the tissue, Applicant assayed DV200 fragments. Slices of 3-5 pieces at lOum thick ribbon slices were saved prior to mounting. RNA was isolated as recommended in PureLink™ FFPE Kit (Invitrogen). All samples with DV200 fragments greater than 60% are used for the Visium HD experiments.
[0224] Deparaffmization and H&E staining conditions were carried out as recommended by 10X Genomics Visium HDFFPETissuePrepHandbook RevA CG000684. Briefly, H&E images were acquired by Olympus VS200 Slide Scanner at 20X magnification. Immediately after imaging, the slides are decrosslink and hybridize overnight with Visium Human Transcriptome Probes v2 >18,000 genes, >54,000 probe pairs. After permeabilizing and barcoding, the hybridized RNA was transferred to a Visium Slide slide 6.5mmX6.5mm oligo capture area. The ligated and barcoded transcripts were then subjected to amplification and indexing for library construction. Paired-end, dual indexed sequencing conditions were Readl: 43 cycles, i7 Index: 10 cycles, i5 Index: 10 cycles and Read 2S: 50 cycles. The libraries were sequenced at the UCSD IGM using Illumina NovaSeq X.
[0225] Single cell multiome clustering and annotationAty. Dkt. No.: 114198-7260
[0226] The lOx Multiome fastqs were aligned and processed via Cell Ranger Arc v2.0.2. For each library Applicant generated a ATAC count matrix of 5kb windows tiling the human genome by removing duplicate reads from the atac_possorted_bam.bam file from Cell Ranger, converted to tagAlign format, and quantified counts with bedtools v2.27.1. For each library, sample pools were deconvoluted using demuxlet(H. M. Kang, et al., Multiplexed droplet singlecell RNA-sequencing using natural genetic variation. Nat Biotechnol 36, 89-94 (2018).) (v2) (github.com / statgen / popscle) using a VCF subset to reads in open chromatin sites(K. Zhang, et al., A single-cell atlas of chromatin accessibility in the human genome. Cell 184, 5985-6001. el9 (2021).) and samples in that pool, a merged and sorted bam made from the Cell Ranger gex_possorted_bam.bam and atac_possorted_bam.bam, and the list of barcodes used to make the 5kb window matrix.
[0227] The raw RNA count matrix was loaded into Seurat v5.0.3, subset to the list of barcodes used to make the 5kb windows matrix, and SoupX vl.6.2 (M. D. Young, S. Behjati, SoupX removes ambient RNA contamination from droplet-based single-cell RNA sequencing data. Gigascience 9, giaal51 (2020).) was used to remove ambient RNA. The 5kb windows matrix was added to the object as a ChromatinAssay using Signac vl.12.0 (T. Stuart, et al., Single-cell chromatin state analysis with Signac. Nat Methods 18, 1333-1341 (2021).) and donor assignments were added based on the demuxlet outputs, filtering out any barcode called as a doublet. The resulting object was clustered using the weighted nearest neighbor (WNN) approach to check library quality.
[0228] Prior to merging libraries, the ATAC assay was subset to the 50k most variable windows. Cells were further filtered after merging all libraries together for <2% mitochondrial gene and TSS enrichment >2 as calculated by Signac(T. Stuart, et al., Single-cell chromatin state analysis with Signac. Nat Methods 18, 1333-1341 (2021).). After clustering, Applicant identified residual populations that appeared to be doublets based on high expression of markers from two different cell types and higher gene expression. To remove these potential doublets, Applicant ran Amulet vl.l (A. Thibodeau, et al., AMULET: a novel read count-based method for effective multiplet detection from single nucleus ATAC-seq data. Genome Biol 22, 252 (2021).) and reclustered the data at high resolution and removed clusters with greater than 5% cells marked as doublets by Amulet.
[0229] The resulting object was clustered using the RNA modality and cell types were annotated based on the expression of canonical liver marker genes, such as Albumin (ALB) for Hepatocytes and collagen genes in stellate cells. A full list of genes used to annotate cell types is provided in Supplemental Table 2. One donor (HL180069) in the original sample pools, was excludedAty. Dkt. No.: 114198-7260 from downstream analyses due to concerns with sample quality based on consistently being an outlier in count data.
[0230] Peaks were called by cell type using MACS2 v2.2.7.1 (Y. Zhang, et al., Model-based analysis of ChlP-Seq (MACS). Genome Biol 9, R137 (2008).) through Signac. An initial peak set across cell types was generated by iteratively clustering overlapping peaks, selecting the peak with the highest pileup score, removing other peaks directly overlapping that peak, and repeating until no overlaps remain. This consensus peak set was used to generate a new matrix using the FeatureMatrix command in Signac. The resulting data was re-clustered and final cell type annotations were identified based on expression patterns of canonical liver genes and marker genes in Seurat. A final set of consensus peaks were identified for cell types and sub-types using the following pipeline. ATAC-seq bam files were converted to tagAlign format and merged, then split by cell type or subtype using a barcode list. Peaks were called on the resulting split tagAlign file and a consensus peak set was generated across cell types and subtypes using the same iterative clustering approach described above. Finally, a barcode level count matrix was generated using FeatureMatrix and added back into the object and used for analyses.
[0231] Preprocessing of Droplet Paired-Tag data
[0232] Fastq files from Droplet Paired-Tag experiments were demultiplexed using the mkfastq command in cellranger-arc v2.0.0. After demultiplexing, the histone modification and transcriptome modalities were aligned separately to the human reference genome (hg38) using the count command in cellranger-atac v2.0.0 for histone modifications or cellranger v6.1.2 for transcriptome. To ensure the inclusion of high-quality nuclei, Applicant e implemented a four-step filtering strategy: i) Histone Modification Data Filtering: At the sample level, histone modification data were aggregated, and peak calling was performed using MACS2 v2.1.2 (Y. Zhang, et al., Model-based analysis of ChlP-Seq (MACS). Genome Biol 9, R137 (2008).) to identify narrow peaks (for H3K27ac) and broad peaks (for H3K27me3). High-quality barcodes were selected based on the number of fragments per cell and the Fraction of Reads in Peaks (FRiP) score, ii) Transcriptome Data Filtering: For RNA data, barcodes were filtered based on the number of unique molecular identifiers (UMIs) per cell, iii) Nuclei Matching: Pass-filter barcodes from both modalities were paired to select high-quality nuclei, iv) Donor Assignment: Applicant used the genotype information from each donor to distinguish the barcodes associated with single or multiple donors as described above for multiome data. High-quality nuclei confidently assigned to a single donor were retained for downstream analyses. Finally, to address ambient RNA contamination, Applicant used SoupX (M. D. Young, S. Behjati, SoupX removes ambient RNA contamination from droplet-based single-cell RNA sequencing data.Aty. Dkt. No.: 114198-7260 Gigascience 9, giaal51 (2020). to estimate and remove contamination for each library independently prior to clustering.
[0233] Reference mapping of Droplet Paired-Tag data with lOXMultiome
[0234] Applicant used the RNA modality from 10X Multi ome assays as reference to annotate cell types from Droplet Paired-Tag data using Seurat v5.1.0 (Y. Hao, et al., Dictionary learning for integrative, multimodal and scalable single-cell analysis. Nat Biotechnol 42, 293-304 (2024).). In brief, Droplet Paired-Tag datasets were first normalized and scaled using SCTransform (S. Choudhary, R. Satija, Comparison and evaluation of statistical error models for scRNA-seq. Genome Biol 23, 27 (2022).). Integration anchor cells was identified with the ‘FindTransferAnchors’ function to project the reference (10X Multiome) PCA structure to query (Droplet Paired-Tag) data. The first 30 principal components were then used for UMAP visualization and cell identity annotation on major cell type or cell subtype level using ‘MapQuery’ function with default parameters. For downstream analysis at the major cell type and cell subtype level, cells with a prediction score <0.8 were removed.
[0235] Visualization of AT AC and histone modification tracks
[0236] Cell type- or condition-specific alignments for histone modification data were extracted from cellranger bam files using custom scripts. BigWig files were generated using the ‘bamCoverage’ function in deepTools v3.5.1 applying the RPGC normalization method (F. Ramirez, et al., deepTools: a flexible platform for exploring deep-sequencing data. Nucleic Acids Res 42, W187-191 (2014).). Chromosome Y (chrY) was excluded from the normalization process.
[0237] For chromatin accessibility data, the shifted and extended tagAlign files originally used for peaks calling were converted back to bam files for BigWig generation. The bamCoverage function was used to generate BigWig files, employing the same normalization method RPGC and exclusion of chrY as applied for histone modification data.
[0238] Preprocessing of Droplet Hi-C data
[0239] Droplet Hi-C data were processed following previously described protocols. Fastq files were demultiplexed using the mkfastq command in cellranger-atac v2.0.0. After demultiplexing, cellular barcode sequences were extracted and aligned to the whitelist using Bowtie vl.3.0 (B. Langmead, et al., Ultrafast and memory-efficient alignment of short DNA sequences to the human genome. Genome Biol 10, R25 (2009).), and appended to the beginning of each read name to record cellular identity. Sequencing adapters were trimmed using Trim-Galore v0.6.10Aty. Dkt. No.: 114198-7260 (M. Martin, Cutadapt removes adapter sequences from high-throughput sequencing reads.EMBnet j. 17, 10 (2011).), and reads were aligned to the human reference genome (hg38) with BWA-MEM vO.7.17 using the arguments -SP5M'( H. Li, Aligning sequence reads, clone sequences and assembly contigs with BWA-MEM. arXiv [Preprint] (2013).doi.org / 10.48550 / ARXIV.1303.3997.). Contact pairs were then parsed, sorted, and deduplicated using Pairtools v0.3.0 (Open2C, N. Abdennur, et al., Pairtools: from sequencing data to chromosome contacts. bioRxiv, 2023.02.13.528389 (2023).), with barcode information stored in separate columns within the pairs file.
[0240] High-quality nuclei in each library were selected based on the number of unique contact pairs per cell. Genotype information was then used for donor phasing, and nuclei confidently assigned to a single donor were retained. These nuclei were split into individual pairs files, each containing the complete contact information for a single nucleus. Imputation was conducted using scHiCluster (vl.3.5) at three resolutions: 100 kb for compartment analysis, 25 kb for domain analysis, and 10 kb for cell embedding and loop analysis (J. Zhou, et al., Robust singlecell Hi-C clustering by convolution- and random-walk-based imputation. Proc Natl Acad Set U SA 116, 14011-14018 (2019).).
[0241] Reference mapping of Droplet Hi-C data
[0242] Applicant applied a previously described method to reference-map and annotate Droplet Hi-C profiles using gene expression data generated from the same tissue and samples (S. Shen, et al., scGAD: single-cell gene associating domain scores for exploratory analysis of scHi-C data. Bioinformatics 38, 3642-3644 (2022).). Briefly, the single-cell gene associating domain (scGAD) score Rij was calculated as the total contact number across the gene body for gene i in cell j at 10 kb resolution. The chromatin interaction profiles were represented as a cell-by-gene scGAD score matrix, which has been shown to correlate with gene activity. This matrix served as the input for reference mapping using the snRNA-seq data.
[0243] For reference mapping, the RNA modality from 10X Multi ome was used. The PC A model derived from the reference dataset was then applied to transform the scGAD score matrix, with features normalized and scaled to match the reference. Integration of the scGAD score matrix with the reference dataset was performed using canonical correlation analysis (CCA), and the integrated embedding was transformed and visualized using the pre-calculated UMAP embedding model from the reference dataset with Seurat function ‘ProjectUMAP’.
[0244] To annotate cell identities based on scGAD scores, Applicant identified the 15 nearest neighbors in the reference RNA data for each Droplet Hi-C cell using PyNNDescent vO.5.6 withAty. Dkt. No.: 114198-7260 Euclidean distance as the metric (W. Dong, et al., “Efficient k-nearest neighbor graph construction for generic similarity measures” in Proceedings of the 20th International Conference on World Wide Web (ACM, Hyderabad India, 2011;dl.acm.org / doi / 10.1145 / 1963405.1963487), pp. 577-586.). Distances were scaled and standardized, with the nearest neighbor receiving the highest score. The cell type identity of each Droplet Hi-C cell was inferred by assigning the cell type label that received the highest standardized score among its nearest neighbors.
[0245] Identifying chromatin loops in liver cell types
[0246] Applicant used scHiCluster to identify loops at lOkb resolution from imputed contact maps in selected major cell types. scHiCluster implements a modified version of SnapHi-C and performs loops calling between genomic distances of 50kb and 5Mb. In addition, Applicant identified loop summits which represent loop anchors with the highest normalized signal against the global background within a 20kb window.
[0247] Genome annotation using ChromHMM
[0248] Applicant used ChromHMM to summarize all epigenetic modalities (ATAC, H3K27ac, H3K27me3) and annotate the genome (J. Ernst, M. Kellis, Chromatin-state discovery and genome annotation with ChromHMM. NatProtoc 12, 2478-2492 (2017).). To generate ChromHMM model, Applicant first split the fragments files from cellranger-atac output by cell type. The resulting bed files were binarized using the binarizeBed function of ChromHMM, with the ' -center' flag specified. A model with states number set to 5 were learned using the LearnModel function with default parameters. Chromatin states were manually annotated based on the likelihood of epigenetic mark associated with each state. It should be noted that the states identified here may differ from those in other studies such as the Epigenome Roadmap project due to the more limited number of histone marks available.
[0249] Annotation of cCREs using ChromHMM
[0250] To annotate candidate cis-regulatory elements (cCREs), Applicant determined the overlap between ChromHMM segments and the identified cCREs from 10X Multi ome data using “bedtools intersect” command (A. R. Quinlan, I. M. Hall, BEDTools: a flexible suite of utilities for comparing genomic features. Bioinformatics 26, 841-842 (2010).). A ChromHMM state was assigned to a cCRE if it overlapped with at least 50% of the cCRE region. In rare cases (<1%) where a cCRE exhibited equal proportions of multiple chromatin states, chromatin state was randomly assigned. Since ChromHMM states are called using 200 bp bins while the medianAty. Dkt. No.: 114198-7260 length of cCREs is 300 bp, the inclusion of more than two chromatin states for a single cCRE is very rare and was not considered in this study.
[0251] Changes in sample cell type proportions.
[0252] To compare the proportion of cell types and cell sub-types across donors with different disease stages, Applicant first extracted and square root scaled the proportion of each cell type per donor. Applicant arranged disease classifications along a continuous scale with control as 0, MASL as 1, MASH as 2 and MetALD as 3. Applicant also considered fibrosis stage and steatosis grade. Then, Applicant trained generalized linear models to predict the scaled proportion of each cell type using condition, fibrosis stage, or steatosis, along with scaled donor age and BMI. Applicant extracted the coefficient and p-value associated with each covariate from the trained model and performed Benjamini Hochberg multiple test correction to identify donor covariates significantly predictive of cell type proportions. Additionally, Applicant performed an ANOVA on the donor condition information to identify differences in cell type proportions between disease groups. Two donors (HL 170065, HL201024) were excluded from these analyses because they had outlying values for BMI, over 3 standard deviations from the mean of all donors. When comparing the proportions of hepatocyte sub-types across donors, Applicant calculated the proportion of each sub-type relative to the total number of hepatocytes.
[0253] Defining gene regulatory programs in each cell type
[0254] Cell type marker genes were identified by generating cell type pseudobulk profiles for each donor, calculating the average transcripts-per-million (TPM) for each gene across donors, and calculating entropy using the Entropy function from the DescTools package in R. Applicant then considered markers of a cell type as those with the lowest entropy values and at least TPM>1 in the cell type.
[0255] Defining cell type specific cCREs
[0256] To identify cell type- or subtype-specific cCREs, Applicant first performed a regressionbased differential accessibility test for each cluster using SnapATAC2 v2.7.0 (K. Zhang, et al., A fast, scalable and versatile tool for analysis of single-cell omics data. Nat Methods 21, 217-227 (2024).). Differential cCREs were then assigned to specific groups using a described method based on a specificity score derived from Jensen-Shannon divergence (JSD) (Y. E. Li, et al., A comparative atlas of single-cell chromatin accessibility in the human brain. Science 382, eadf7044 (2023).). Briefly, Applicant calculated and normalized the fraction of accessibility for each differential cCRE across clusters to account for differences in complexity between groups. The normalized accessibility values served as input for JSD calculation. JSD was computedAty. Dkt. No.: 114198-7260 between the accessibility distribution of each cCRE and its reference distribution using the R package “philentropy”. The reference distribution (d) for a given cCRE in given group ( 7) is defined as:d = \1,91(0, otherwise
[0257] The Jensen Shannon-based specificity score (JSS) was defined as 1 — / jSD. Applicant calculated JSS for every cCRE for each group. Applicant consider JSS from cCREs that are not identified as differential from likelihood ratio test as a background distribution, and JSS from differential cCREs with positive values to be true positives. Applicant set an empirical false discovery rate (FDR) cutoff at where type I error no more than 5%. Applicant defined those cCREs that can be assigned to only one cell type or subtype as cell type- or subtype- specific cCREs. Finally, cCREs uniquely assigned to a single cell type or subtype based on their JSS were defined as cell type- or subtype-specific cCREs.
[0258] Identification of cis-regulatory modules
[0259] Applicant employed non-negative matrix factorization (NMF) to classify cCREs into modules based on their relative accessibility across cell types (Y. E. Li, et al., Identification of high-confidence RNA regulatory elements by combinatorial classification of RNA-protein binding sites. Genome Biol 18, 169 (2017).). Using the NMF implementation in scikit-learn (F. Pedregosa, et al., Scikit-learn: Machine Learning in Python, doi: 10.48550 / ARXIV.1201.0490 (2012).), Applicant decomposed the cell-by-cCRE matrix V (N*M, where N= cCREs, M= cells) into a coefficient matrix H (R M, where R = number of modules) and a basis matrix W (N*R). The basis matrix represents module-associated cCREs, while the coefficient matrix H describes the cell type components and their weights within each module. To determine an appropriate rank A, Applicant evaluated multiple metrics including ‘Sparseness’ and ‘Entropy’ (P. O. Hoyer, Non-negative matrix factorization with sparseness constraints. arXiv [Preprint] (2004). doi. org / 10.48550 / ARXIV.CS / 0408058., H. Kim, H. Park, Sparse non-negative matrix factorizations via alternating non-negativity-constrained least squares for microarray data analysis. Bioinformatics 23, 1495-1502 (2007).). These metrics were used to identify a rank A that optimally balances interpretability and clustering quality. To ensure stability, Applicant performed 100 NMF runs at each tested rank using different random seeds and calculated average metric values across these runs.
[0260] Sequence motif enrichment was performed on cell type specific ATAC peaks using the fmdMotifsGenome.pl command from HOMER v4.11 (S. Heinz, et al., Simple combinations ofAty. Dkt. No.: 114198-7260 lineage-determining transcription factors prime cis-regulatory elements required for macrophage and B cell identities. Mol Cell 38, 576-589 (2010).) with the size flag set to 200 and the JASPAR 2022 database (J. A. Castro-Mondragon, et al., JASPAR 2022: the 9th release of the open-access database of transcription factor binding profiles. Nucleic Acids Res 50, D165-D173 (2022).) obtained from the JASPAR2022 vO.99.7 bioconductor package using the dumpJaspar function from the MonaLisa package. The full set of peaks in a cell type was used as the background set.
[0261] Defining changes in regulatory programs in liver cell types in MASLD
[0262] Applicant tested for changes in gene expression, chromatin accessibility, and H3K27ac and H3K27me levels using DESeq2 vl.42.0 (M. I. Love, et al., Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol 15, 550 (2014).). The cell level count matrix for each modality was split by cell type and counts were aggregated by donor. Applicant performed differential analyses separately comparing MASL, MASH, and MetALD to normal livers using scaled age, scaled BMI, gender, and donor pool as covariates. Applicant also performed differential analyses for fibrosis score as both a quantitative variable as well as a categorical variable to separate MASH / MetALD donors (low:0-2, high:3-4) including the same covariates, as well as steatosis grade in non-liver disease including fibrosis score as an additional covariate. For each analysis, Applicant excluded donors with <10 cells for the given cell type, features with fewer than 10 total counts across all donors, and categorical contrasts with fewer than three samples in each group. The fGSEA package vl .28.0 was then used to measure enrichment of biological pathways from KEGG (M. Kanehisa, S. Goto, KEGG: kyoto encyclopedia of genes and genomes. Nucleic Acids Res 28, 27-30 (2000).), REACOME (M. Milacic, et al., The Reactome Pathway Knowledgebase 2024. Nucleic Acids Research 52, D672-D678 (2024).) and GO (M. Ashburner, et al., Gene ontology: tool for the unification of biology. The Gene Ontology Consortium. Nat Genet 25, 25-29 (2000).) in the differential gene expression results ranked by the ‘stat’ column.
[0263] Applicant identified cCREs changing in chromatin state in disease states by classifying chromatin state for each AT AC peak for each disease and fibrosis stage, then performing kmeans clustering of cCREs and manually annotating sites changing in MASH, MASL, or MetALD.
[0264] Sequence motif enrichment was calculated in the differentially accessible chromatin sites from the differential cCRE results or within clusters of cCREs using Homer as described above. All tested sites for the given contrast were used as the background set.Aty. Dkt. No.: 114198-7260
[0265] Pathway enrichment was calculated for genes linked to cell type specific cCREs using fGSEA with the GO biological processes pathways.
[0266] To identify the relationship between gene expression changes in MASL, Fib- MASH, Fib+ MASH, and all MASH livers, Applicant calculated the correlation in effect sizes for each condition compared to normal livers using Spearman’s rho. For cross-modality correlations, Applicant intersected features in each modality in hepatocytes, and calculated Pearson correlation in effect size in Fib+ MASH compared to normal for overlapping features across modalities.
[0267] Gene regulatory network modeling
[0268] Applicant identified gene regulatory networks (GRNs) in hepatocytes using SCENIC+(C. Bravo Gonzalez-Blas, et al., SCENIC+: single-cell multiomic inference of enhancers and gene regulatory networks. Nat Methods 20, 1355-1367 (2023).). Seurat SketchData() was employed to down sample either the entire multi ome dataset or just hepatocytes in single cell multi ome data to 50k nuclei. Utilizing multiome ATAC data from down sampled hepatocytes, Applicant imputed chromatin accessibility and generated binarized topics using pycisTopic. Topics above OTSU calculated thresholds were then retained. Gene expression counts, imputed chromatin accessibility, topics, and differentially accessible peaks in MASLD were then used as input for SCENIC+ (vl.Oal) (C. Bravo Gonzalez -Blas, et al., SCENIC+: single-cell multiomic inference of enhancers and gene regulatory networks. Nat Methods 20, 1355-1367 (2023).). Subsequent downstream analysis was performed with default parameters.
[0269] Applicant identified GRNs with increased or decreased activity in MASLD states by performing Gene set enrichment analysis of GRN target genes using the GSEA function of clusterProfiler package, employing the fGSEA method (T. Wu, et al., clusterProfiler 4.0: A universal enrichment tool for interpreting omics data. The Innovation 2, 100141 (2021).). GRN information, obtained from SCENIC+ output, was utilized as the TERM2GENE parameter. Selected GRNs and their target genes, based on core enrichment results, were then visualized using Cytoscape(P. Shannon, et al., Cytoscape: a software environment for integrated models of biomolecular interaction networks. Genome Res 13, 2498-2504 (2003).).
[0270] Spatial transcriptomics analysis
[0271] Visium HD sequence data was processed using SpaceRanger (v3.0.0) to generate raw count expression matrices with the command “spaceranger count”, using “-loupe-alignment” to manually align the image and the gene expression grid. Bin2cell (v0.3.0) was used to segment the 2x2 micron Visium HD bins into cells, through the StarDist model. StarDist parameters onAty. Dkt. No.: 114198-7260 the H&E image were stardist_model=“2D_versatile_he”, and prob_thresh=0.01, while StarDist parameters used for the gene expression were “stardist_model=“2D_versatile_fluo”, “prob_thresh=0.05”, and “nms_thresh=0.05”. An mpp of 0.25 was used. The counts were then normalized with Scanpy (1.10) using the functions “sc.pp.normalize total” and “sc.pp.loglp”
[0272] Tangram (vl.0.04) was used to map cell type annotations and gene expression information to Visium HD data using cell type markers and gene expression from the RNA portion of the multiome data. Individual sets of reference cell type markers were generated for different disease groups, namely normal, MASL, Fib-MASH, and Fib+ MASH. To generate cell type markers in each of these groups, cells from the multiome data were subset by each group and then down sampled to 20,000 cells. The Scanpy function sc.tl.rank genes group with the parameter “groupby=celltype” was used to identify the top 50 gene markers for each cell type. Tangram models for each spatial dataset was then trained on the cell type markers and gene expression data matching each tissue donor’s respective disease group.
[0273] Mapping cell types and gene expression information to each VisiumHD dataset using Tangram first involved splitting the data into patches of 10,000 cells. For each patch, the Tangram function tg.map cells to space with parameters “density _prior=”uniform” and “device=”cuda:0”” was used to find the probability of a particular celltype in a cell. For each patch, the top and bottom 2% quantile probability scores were clipped to remove outliers for each cell type label. The remaining probability scores for each cell type were then min-max scaled, and initial cell type annotations were assigned based on the highest scaled probability score. The separately annotated patches were then merged. Finally, to remove low quality cells, those with a scaled Tangram probability score < 0.5 were removed. To confirm annotations and quantify marker gene expression in spatial profiles, imputed gene expression from Tangram for each Visium HD dataset was down sampled to 100,000 cells. Datasets were then concatenated together, where the average expression of each known marker gene in each cell type was examined.
[0274] Hepatocyte zones and sub-cell types were annotated in Visium HD data using Tangram in similar fashion as above. Specifically, Tangram was used to align cell labels and gene expression of hepatocyte subclusters to only those cells within each Visium HD dataset that were previously labeled hepatocytes in the broad cell type Tangram runs. Reference markers for hepatocyte subclusters were identified using the same approach as for broad cell types, using the parameter “groupby=subcelltype” for sc.tl.rank genes group to identify hepatocyte subcluster markers. As with the broad annotations, reference markers sets were generated for different disease groups (normal, MASL, Fib-MASH, and Fib+ MASH) with the exception that if aAty. Dkt. No.: 114198-7260 hepatocyte sub-cluster comprised less than 5% of hepatocytes in the multi ome dataset from a given disease group, its markers were excluded from the reference marker set for that group.
[0275] To plot the expression of specific gene sets or pathways in spatial data, gene module scores for each set were plotted.
[0276] Genome-wide association study enrichments
[0277] Applicant utilized genome-wide association (GWAS) data for chronic elevation of alanine aminotransferase (ALT) levels(M. Vujkovic, et al., A multiancestry genome-wide association study of unexplained chronic ALT elevation as a proxy for nonalcoholic fatty liver disease with histological and radiological validation. Nat Genet 54, 761-771 (2022).), a proxy for liver damage and MASLD. Applicant performed GWAS enrichment analyses for cell typespecific regulatory elements using partitioned heritability following LD Score Regression (B. K. Bulik- Sullivan, et al., LD Score regression distinguishes confounding from polygenicity in genome-wide association studies. Nature Genetics 47, 291-295 (2015)., H. K. Finucane, et al., Partitioning heritability by functional annotation using genome-wide association summary statistics. Nature Genetics 47 , 1228-1235 (2015).). The H3K27ac and ATAC-seq peaks for each liver cell type were lifted to hgl9, and annotations were created using the make annot.py command in LDSC and 1000G EUR reference variants. LD scores were calculated using the Idsc.py command with 1000G_EUR_Phase3_baseline_snps and a Id window of IcM. Finally, enrichment was calculated using the Idsc.py command with the — h2-cts flag, with the 1000G_EUR_Phase3_baseline and all peaks of a given modality as background sets in the analysis. The same analysis was run on ATAC-seq peak sites split by chromatin state using the total set of sites for that cell type in the background set. All enrichments were used for calculating FDR using Storey’s method in the package FDR v2.34.0, and tests were considered significant at FDR<.10.
[0278] Applicant performed enrichment analyses of subsets of sites with altered activity in MASLD for fine-mapped variants using FINRICH vO.3.2, which performs a permutation test to identify enrichment of credible set variants in genomic windows relative to a background set. For differential AT AC sites all features passing 5% FDR were used as the foreground and all tested sites the background against credible set variants from the cALT GWAS trans-ancestry analysis lifted over to hg38 using ‘SNPlocs.Hsapiens.dbSNP155.GRCh38’ vO.99.24. Foldenrichment was assessed for sites changing in chromatin state and for regions from GRNs changing in activity in MASL and MASH from SCENIC+. The significance of the foldenrichments was determined via permutations followed by FDR correction. Applicant identifiedAty. Dkt. No.: 114198-7260 credible set variants from the cALT trans-ancestry fine-mapping overlapping cCREs in GRNs for each TF and constructed a graph of TF GRNs based on credible set variant overlap using igraph v2.1.4.
[0279] Quantitative trait locus mapping
[0280] Applicant performed QTL mapping using the RNA, ATAC, H3K27ac, and H3K27me3 modalities. Barcode level count matrices were subset by cell type and counts were aggregated by donor. The resulting libraries were transcripts per million (TPM) normalized (RNA) or counts per million (CPM) normalized and features were filtered based on at least 20% of samples with >0.1 TPM or CPM and 6 counts. The counts were normalized using a TMM adjustment in edgeR v4.0.16 (M. D. Robinson, et al., edgeR: a Bioconductor package for differential expression analysis of digital gene expression data. Bioinformatics 26, 139-140 (2010).) and inverse-normal transformed. Count matrices were converted to bed format as described on the TensorQTL GitHub repository (github.com / broadinstitute / tensorqtl). For ATAC and DPT modalities the bed coordinates were the feature peak coordinates, and for RNA TSS sites were derived from the GTF used in CellRanger based on strand and start and end coordinates, with the bed coordinates being a Ibp region at the start of the gene.
[0281] Applicant generated 10 genotype PCs based on analysis in plink vl.9 to account for ancestry. Sample genotypes were merged with 1000 Genomes genotypes and pruned prior to PCA. To account for latent effects in genomic profiles Applicant ran PCA on the inverse-normal transformed count matrices and extracted the first 15 PCs. Applicant used gender, scaled age, scaled BMI, fibrosis stage, disease label, and sequencing pool as additional biological and technical covariates.
[0282] TensorQTL vl (A. Taylor-Weiner, et al., Scaling computational genomics to millions of individuals with GPUs. Genome Biol 20, 228 (2019).) was run on a HPC at the San Diego Supercomputer Center with an al 00 gpu. First, pairwise summary stats were generated using the — cis nominal mode, then gene wise permutations and FDR cutoffs were determined using the — cis mode with the — fdr 0.05 flag, which performs permutations for each gene using the lead SNP to determine an empirical p-value and performs Storey’s method for FDR correction on the permuted p-values. These empirical p-values are back mapped to determine the nominal p-value cutoff for other variants for the given feature, which were back mapped onto the pairwise summary stats.
[0283] Applicant followed the mashr vO.2.79 (S. M. Urbut, et al., Flexible statistical methods for estimating and testing effects in genomic studies with multiple conditions. Nat Genet 51, 187-Aty. Dkt. No.: 114198-7260 195 (2019).) eQTL analysis vignette for each modality independently. First, we constructed two mashr datasets, one comprising the beta and error terms of the summary stats for 200,000 randomly selected feature variant pairs, and the other representing every feature-variant pair significant in any cell type. The randomly selected dataset was used to estimate the canonical covariance matrix, and the true effects the data drive covariance matrix. These two matrices and the randomly selected dataset were used to train the mash model. Then the model was used to estimate the effects in the dataset of significant associations using the get fitted g parameter. QTLs determined as specific based on a Ifsr value below 0.05.
[0284] Summary stats for Vujkovic et al 2022 (M. Vujkovic, et al., A multiancestry genomewide association study of unexplained chronic ALT elevation as a proxy for nonalcoholic fatty liver disease with histological and radiological validation. Nat Genet 54, 761-771 (2022).) were accessed from dbGaP and converted to hg38 coordinates using SNPlocs.Hsapiens.dbSNP155.GRCh38 in Bioconductor. Any QTL feature passing 5% FDR was colocalized with GWAS summary statistics using the coloc.abf function from coloc v5.2.3 (C. Giambartolomei, et al., Bayesian Test for Colocalisation between Pairs of Genetic Association Studies Using Summary Statistics. PLOS Genetics 10, el004383 (2014).). Signals were determined to be colocalized based on PPH4 greater than 0.8, the GWAS signal passing genome wide significance (5x1 O'8), and the lead GWAS and QTL variants being in high LD (r^O.5).
[0285] Cross-modality QTL modules were created following the process outlined in Arthur et al (T. D. Arthur, et al., Multi-omic QTL mapping in early developmental tissues reveals phenotypic and temporal complexity of regulatory variants underlying GWAS loci. [Preprint] (2024). doi.org / 10.1101 / 2024.04.10.588874.). Applicant identified QTL features by cell type with overlapping cis QTL mapping windows and performing colocalization. Colocalization signals passing PPH4>0.8 were constructed into a connected graph using igraph v2 (G. Csardi, et al., igraph for R: R interface of the igraph library for graph theory and network analysis, version v2.1.2, Zenodo (2024); doi.org / 10.5281 / ZENODO.7682609.) and clustered using the Louvain algorithm. Clusters with modularity scores above 0.3 were divided into a subgraph and clustered recursively.
[0286] Sequence-based models of chromatin accessibility
[0287] A ChromBPNet model was trained using the default preprocessing parameters outlined in the ChromBPNet tutorial (github.com / kundajelab / chrombpnet / wiki / Preprocessing). To account for Tn5 bias, a separate model was trained using our own lOx multi ome pseudobulk data. Four bias-Aty. Dkt. No.: 114198-7260 factorized models were trained on hepatocytes under different conditions: Control, MASL, MASH, and MetALD. One model was trained using untreated hepatocytes, while the other three were trained using hepatocytes from MASH, MASL, and MetALD conditions. For training, Applicant used macs2 narrowPeak regions with a p-value threshold of <0.05, called on pseudobulk ATAC fragments from all hepatocyte samples. Chromosomes 1, 3, and 6 were assigned for training, while chromosomes 8 and 20 were reserved for validation. Following training, default parameters were applied. ChromBPNet models consist of two output heads: (1) the profile head, which predicts the shape of a profile, and (2) the counts head, which estimates the total counts within a profile. The chrombpnet_nobias.h5 models were used for this analysis, and Applicant computed sequence contribution scores for one or both heads. Additionally, Applicant utilized ChromBPNet for variant prediction to assess the impact of genetic variants on chromatin accessibility within our dataset. Applicant considered variants where the absolute value of log counts diff was greater than .025 as having a functional effect. To generate predicted counts and contribution bigwigs for both alleles of example variants Applicant used bcftools to insert the alternate allele into the reference bigwig and ran the pred bw and contri bw functions from ChromBPNet on both the standard and custom reference genome using the chrombpnet_nobias.h5 model.
[0288] Annotating MASLD associated loci in hepatocytes
[0289] Applicant identified loci associated with MASLD (cALT levels) from a large genetic association study (M. Vujkovic, et al., A multiancestry genome-wide association study of unexplained chronic ALT elevation as a proxy for nonalcoholic fatty liver disease with histological and radiological validation. Nat Genet 54, 761-771 (2022).). Applicant identified loci where the lead fine-mapped variant was a significant QTL in hepatocytes for chromatin accessibility, H3K27ac or H3K27me3 histone modifications, or gene expression levels, as well as loci that were statistically colocalized with hepatocyte QTL association (shared probability>.8). Applicant next identified loci where at least one fine-mapped variant overlapped a hepatocyte cCRE or active chromatin state, as well as loci where at least one fine-mapped variant had a predicted effect on hepatocyte chromatin accessibility from chromBPNet models. Finally, Applicant annotated target genes that were either within lOkb of a chromatin loop to a hepatocyte cCRE harboring fine-mapped variants or a colocalized eQTL
[0290] Identification of trajectories and pseudotime.
[0291] To infer potential state changes between hepatocyte subtypes during Applicant performed trajectory analysis using the Monocle3 package (vl.3.4). Briefly, Applicant selected allAty. Dkt. No.: 114198-7260 hepatocyte cells from high fibrosis donors (MASH / MetALD with fibrosis score >=3) and learned a trajectory with monocle’s learn_graph() function with the following parameter settings: maximum euclidean distance ratio of 1, minimal branch length of 10, 25 nearest neighbors, and the default method for identifying nearest neighbors. Next, to reconstruct pseudotime from the Hepatocytes. Injured cluster, Applicant chose the node furthest in that cluster as a starting point and ordered all cells from there. The top 50 genes with the largest fold change in Fib-a hepatocytes when comparing high fibrosis (Fib+) MASH against normal livers were plotted along the trajectory after aggregating cells into bins (n=50) based on pseudotime order.
[0292] Identification of Fibrosis-adjacent hepatocytes and neighborhood enrichment analysis
[0293] Applicant first generated a fibrosis-associated gene set by selecting the top 50 genes with the largest fold change in the fibrosis-associated hepatocytes when comparing high fibrosis (Fib+) MASH against normal livers. These genes also mark a subset of cells identified in multiome pseudotime analysis described below. Module scores for this fibrosis-associated gene set were then calculated for each hepatocyte in a Fib+ MASH Visium HD dataset (HL2201019). Hepatocytes in the top 1% of module scores (>0.2) were isolated and labeled “Fib-adj hepatocytes” and used this population was used for neighborhood enrichment analysis.
[0294] Neighborhood enrichment was calculated using the sq.gr.nhood enrichment function from Squidpy (cl.6.1) (sq.gr.nhood enrichment (adge comb, cluster_key=“subcluster_label_comb”, show_progress_bar=False)). Enrichment was evaluated based on combined broad cell type classifications for non-hepatocytes, hepatocyte sub-cluster annotations, and Fib-adj hepatocytes.
[0295] Cell-cell interaction analysesApplicant performed cell-cell communication analysis on Visium HD spatial transcriptomic data using CellChat v2.1.0. A CellChat object was created from count data and meta data including Tangram cell types using the createCellChat function where spatial. factors were calculated with a conversion factor of 0.5 and a spot size of 2. Over expressed genes and interactions were calculated using the identifyOverExpressedGenes and identifyOverExpressedlnteractions functions. Interaction probabilities of all L-R pairs in the cellchat database were calculated using the computeCommunProb function, employing the truncatedMean method with the following parameters: trim = 0.05, distance. use = TRUE, interaction. range = 50, scale. distance = 1, contact. dependent = TRUE, and contact.range = 15. The resulting CellChat object was filtered to retain interactions involving a minimum of 10 cells expressing either the ligand or receptor. These interactions were then converted into a data frame using the subset Communication function. To account for multiple comparisons, Applicant applied the Benjamini -HochbergAty. Dkt. No.: 114198-7260 method via the q-value package, considering predicted interactions with an FDR < 0.1 as significant. Applicant generated a heatmap of weighted interactions using thenetVisual heatmap function in CellChat and additional visualizations were constructed using the ggplot2 package.
[0296] Clauses
[0297] Clause 1. A method of treating or preventing metabolic dysfunction-associated steatohepatitis (MASH) in a subject in need thereof, the method comprising administering to the subject an effective amount of an agent that modulates the activity of at least one transcription factor gene regulatory network (GRN) in hepatocytes, wherein the at least one GRN comprises a transcription factor selected from the group consisting of HNF4A, RORA, CUX2, CREB5, NFKB1, NR1I2, RELB, KLF6, TEAD1, ETS1, NFAT5, FOXO1, NR4A1, HNF4G, PPARG, STAT1, PPARA, and ATF3, and wherein said modulating is sufficient to alter the expression of one or more target genes regulated by the at least one GRN.
[0298] Clause 2. The method of clause 1, wherein the agent increases the activity of a GRN selected from HNF4A, RORA, CUX2, NR1I2, and PPARA in hepatocytes.
[0299] Clause 3. The method of clause 1, wherein the agent decreases the activity of a GRN selected from NFKB1, RELB, TEAD1, CREB5, KLF6, ETS1, and NFAT5 in hepatocytes.
[0300] Clause 4. The method of clause 1, wherein the one or more target genes are selected from the group consisting of DTNA, AKAP12, ITGA2, LAMA3, MY07B, CD44, TNFAIP2, TMEM150B, FGF1, SMPD3, LPIN3, SLC25A42, SLC22A1, CRYL1, and PSD4.
[0301] Clause 5. The method of clause 1, wherein the agent modulates the activity of one or more candidate cis-regulatory elements (cCREs) within the GRN.
[0302] Clause 6. The method of clause 1, wherein the subject has a MASLD stage selected from MASL, low-fibrosis MASH (Fib- MASH; fibrosis stages 0-2), and high-fibrosis MASH (Fib+ MASH; fibrosis stages 3-4).
[0303] Clause 7. The method of clause 1, wherein the agent is selected from the group consisting of a small molecule, a nucleic acid, an antisense oligonucleotide, an siRNA, an shRNA, a CRISPR-based gene editing agent, a transcription factor modulator, a protein, a peptide, an antibody or antigen-binding fragment thereof, and a viral vector encoding a transcription factor or a modulator thereof.
[0304] Clause 8. A method of diagnosing or stratifying metabolic dysfunction-associated steatotic liver disease (MASLD) progression in a subject, the method comprising: (a) obtaining aAty. Dkt. No.: 114198-7260 biological sample comprising liver cells from the subject; (b) determining, in the liver cells, the activity of one or more gene regulatory features selected from: (i) transcription factor gene regulatory networks (GRNs); (ii) candidate cis-regulatory elements (cCREs); (iii) chromatin states at cCREs; and (iv) gene expression of GRN target genes; and (c) classifying the subject into a MASLD stage based on the determined activity, wherein the MASLD stage is selected from normal, MASL, low-fibrosis MASH, and high-fibrosis MASH.
[0305] Clause 9. The method of clause 8, wherein step (b) comprises determining the activity of GRNs with increased activity in MASL selected from HNF4A, RORA, and CUX2, and GRNs with increased activity in MASH selected from CREB5, NFKB1, RELB, KLF6, TEAD1, ETS1, and NFAT5.
[0306] Clause 10. The method of clause 8, wherein step (b) comprises determining the proportion of hepatocyte sub-clusters selected from Zone 1, Zone 2, Zone 3, MASH-associated (MASH-a), and fibrosis-associated (Fib-a) hepatocytes.
[0307] Clause 11. The method of clause 10, wherein an increased proportion of Fib-a hepatocytes and a decreased proportion of Zone 1 and Zone 3 hepatocytes is indicative of high-fibrosis MASH.
[0308] Clause 12. The method of clause 8, wherein step (b) comprises determining the chromatin state of one or more hepatocyte cCREs, wherein a transition from repressive to active chromatin states at cCREs enriched for NFKB, FOS / JUN, and TEAL) motifs is indicative of MASH progression.
[0309] Clause 13. The method of clause 8, further comprising determining the genotype of the subject at one or more MASLD-associated risk loci selected from 12ql 3 (KRT8 / KRT18), 8p23 (PPP1R3B), 19ql3 (CEBPA), 22ql2 (XBP1), and 2q37 (EFHD1).
[0310] Clause 14. A method of identifying a candidate therapeutic agent for the treatment of metabolic dysfunction-associated steatohepatitis (MASH), the method comprising: (a) providing a population of hepatocytes; (b) contacting the hepatocytes with a test agent; (c) determining the effect of the test agent on the activity of one or more transcription factor gene regulatory networks (GRNs) associated with MASH progression, wherein the one or more GRNs are selected from the group consisting of CREB5, NFKB1, RELB, KLF6, TEAD1, ETS1, NFAT5, HNF4A, RORA, CUX2, FOXO1, NR4A1, HNF4G, PPARG, STAT1, PPARA, and ATF3; and (d) selecting the test agent as a candidate therapeutic agent if the test agent shifts the activity of the one or more GRNs toward a non-disease profile.Aty. Dkt. No.: 114198-7260
[0311] Clause 15. The method of clause 14, wherein determining the effect of the test agent comprises measuring one or more of: (i) chromatin accessibility at cCREs within the GRN; (ii) histone modification levels (H3K27ac and / or H3K27me3) at cCREs within the GRN; (iii) gene expression of target genes within the GRN; and (iv) 3D chromatin contacts at loci associated with the GRN.
[0312] Clause 16. The method of clause 14, wherein a non-disease profile comprises decreased activity of GRNs selected from NFKB1, RELB, TEAD1, CREB5, KLF6, ETS1, and NFAT5, and increased activity of GRNs selected from HNF4A, RORA, CUX2, and PPARA.
[0313] Clause 17. The method of clause 14, wherein the hepatocytes comprise primary human hepatocytes or hepatocytes derived from induced pluripotent stem cells (iPSCs).
[0314] Clause 18. A method of treating liver fibrosis associated with MASH in a subject, the method comprising administering to the subject an effective amount of an agent that disrupts CD44-mediated signaling between hepatic stellate cells (HSCs) and fibrosis-associated hepatocytes, wherein the agent inhibits one or more of: (a) COL1 A1-CD44 interaction between HSCs and fibrosis-adjacent (Fib-adj) hepatocytes; and (b) CD44 expression or activity in hepatocytes.
[0315] Clause 19. The method of clause 18, wherein the agent is an anti-CD44 antibody, a CD44-targeting siRNA, a CD44 small molecule inhibitor, or a combination thereof.
[0316] Clause 20. The method of clause 18, further comprising administering an agent that modulates the epigenomic profile of fibrosis-associated hepatocytes, wherein said modulating comprises increasing H3K27me3 -mediated repression at epithelial remodeling and fibrosis-related genes, or decreasing H3K27me3 -mediated repression at lipid metabolism genes.
[0317] Clause 21. The method of clause 1 or claim 14, wherein the agent modulates the activity of one or more hepatocyte candidate cis-regulatory elements (cCREs) enriched for sequence motifs of one or more transcription factors selected from the group consisting of HNF4A, NR4A1, CEBPA / D, FOS / JUN, TEAD, andNFKB.
[0318] Clause 22. The method of clause 21, wherein the agent decreases activity of one or more hepatocyte cCREs enriched for FOS / JUN, TEAD, or NFKB motifs.
[0319] Clause 23. The method of clause 21, wherein the agent increases activity of one or more hepatocyte cCREs enriched for HNF4A, NR4A1, or CEBPA / D motifs.
[0320] Clause 24. The method of clause 8, wherein step (b) comprises determining increased activity of one or more hepatocyte GRNs associated with lipid metabolism and biosynthesis,Aty. Dkt. No.: 114198-7260 wherein the one or more GRNs are selected from the group consisting of HNF4A, RORA, CUX2, andNRH2.
[0321] Clause 25. The method of clause 8, wherein step (b) comprises determining increased activity of one or more hepatocyte GRNs associated with epithelial remodeling, cytoskeletal organization, cell-extracellular matrix adhesion, inflammation, or immune signaling, wherein the one or more GRNs are selected from the group consisting of CREB5, NFKB1, RELB, KLF6, TEAD1, ETS1, NFAT5, STAT1, HNF4G, and PPARG.
[0322] Clause 26. The method of clause 8, wherein step (b) comprises determining expression of one or more genes selected from the group consisting of DTNA, AKAP12, ITGA2, LAMA3, MY07B, CD44, TNFAIP2, TMEM150B, FGF1, SMPD3, LPIN3, SLC25A42, SLC22A1, CRYL1, andPSD4.
[0323] Clause 27. The method of clause 8 or claim 26, wherein increased expression of one or more of DTNA, AKAP12, ITGA2, LAMA3, MY07B, CD44, TNFAIP2, TMEM150B, and FGF1 is indicative of MASH progression.
[0324] Clause 28. The method of clause 8 or claim 26, wherein increased expression of one or more of SMPD3, LPIN3, SLC25A42, SLC22A1, CRYL1, and PSD4 is indicative of MASL.
[0325] Clause 29. The method of clause 8, wherein step (b) comprises determining whether hepatocytes in the sample comprise a fibrosis-associated hepatocyte population expressing one or more of NFKB1, STAT4, HIF1A, SOD2, TEAD1, and TGFBR2.
[0326] Clause 30. The method of clause 29, wherein the fibrosis-associated hepatocyte population is present in increased proportion in subjects with high-fibrosis MASH relative to normal liver.
[0327] Clause 31. The method of clause 8 or claim 29, wherein step (b) comprises determining whether the sample comprises fibrosis-adjacent hepatocytes expressing one or more of CXCL8, CXCL6, HSPB8, ITGB6, CDKN2A, and KRT23.
[0328] Clause 32. The method of clause 8, wherein step (b) comprises determining whether Zone 3 hepatocytes are reduced and fibrosis-associated hepatocytes are increased relative to a reference liver sample.
[0329] Clause 33. The method of clause 8 or claim 12, wherein step (b) comprises determining whether hepatocytes exhibit increased repression at genes involved in lipid metabolism, transport, or Wnt signaling and / or decreased repression at genes involved in cell structure, adhesion, extracellular matrix interaction, fibrosis, or cytokine signaling.Aty. Dkt. No.: 114198-7260
[0330] Clause 34. The method of clause 33, wherein the repression is determined by measuring H3K27me3 signal.
[0331] Clause 35. The method of clause 14 or claim 16, wherein the non-disease profile comprises decreased expression of one or more genes selected from the group consisting of DTNA, AKAP12, ITGA2, LAMA3, MY07B, CD44, TNFAIP2, TMEM150B, and FGF1.
[0332] Clause 36. The method of clause 14 or claim 16, wherein the non-disease profile comprises increased expression of one or more genes selected from the group consisting of SMPD3, LPIN3, SLC25A42, SLC22A1, CRYL1, and PSD4.
[0333] Clause 37. The method of clause 14 or claim 16, wherein the non-disease profile comprises decreased activity of one or more candidate cis-regulatory elements enriched for NFKB, FOS / JUN, or TEAL) sequence motifs.
[0334] Clause 38. The method of clause 14 or claim 16, wherein the non-disease profile comprises increased activity of one or more candidate cis-regulatory elements enriched for HNF4A, NR4A1, or CEBPA / D sequence motifs.
[0335] Clause 39. The method of clause 18, wherein the fibrosis-associated hepatocytes comprise hepatocytes expressing one or more of NFKB1, STAT4, HIF1 A, SOD2, TEAD1, and TGFBR2.
[0336] Clause 40. The method of clause 18 or claim 39, wherein the fibrosis-adjacent hepatocytes comprise a subset of fibrosis-associated hepatocytes expressing one or more of CXCL8, CXCL6, HSPB8, ITGB6, CDKN2A, and KRT23.
[0337] Clause 41. The method of clause 18, wherein the agent reduces signaling from hepatic stellate cells to fibrosis-adjacent hepatocytes involving COL1A1 and CD44.
[0338] Clause 42. The method of clause 20, wherein the epithelial remodeling and fibrosis-related genes comprise genes involved in cell structure, adhesion, extracellular matrix interaction, or cytokine signaling.
[0339] Clause 43. The method of clause 20, wherein the lipid metabolism genes comprise genes involved in lipid metabolism, lipid transport, or Wnt signaling.
[0340] Clause 44. The method of clause 13, wherein the one or more MASLD-associated risk loci comprise a locus at 12ql 3 associated with KRT8 / KRT18, a locus at 8p23 associated with PPP1R3B, a locus at 19ql 3 associated with CEBPA, a locus at 22ql2 associated with XBP1, or a locus at 2q37 associated with EFHD1.Aty. Dkt. No.: 114198-7260
[0341] Clause 45. The method of clause 8 or claim 13, wherein the genotype determined at the one or more MASLD-associated risk loci is used together with the activity of one or more transcription factor gene regulatory networks (GRNs), candidate cis-regulatory elements (cCREs), chromatin states, or GRN target genes to classify the subject into a MASLD stage.
[0342] Clause 46. The method of clause 8, wherein the biological sample is analyzed by one or more methods selected from the group consisting of single-cell RNA sequencing, single-cell multiome analysis, chromatin accessibility profiling, histone modification profiling, 3D chromatin conformation profiling, and spatial transcriptomics.
[0343] Clause 47. The method of clause 14, wherein determining the effect of the test agent comprises comparing the activity of the one or more GRNs to a reference profile from normal liver tissue.
[0344] Clause 48. The method of clause 14, wherein determining the effect of the test agent comprises comparing the activity of the one or more GRNs to a reference profile from MASL, low-fibrosis MASH, or high-fibrosis MASH liver tissue.Equivalents
[0345] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this technology belongs.
[0346] The present technology illustratively described herein may suitably be practiced in the absence of any element or elements, limitation or limitations, not specifically disclosed herein. Thus, for example, the terms “comprising,” “including,” “containing,” etc. shall be read expansively and without limitation. Additionally, the terms and expressions employed herein have been used as terms of description and not of limitation, and there is no intention in the use of such terms and expressions of excluding any equivalents of the features shown and described or portions thereof, but it is recognized that various modifications are possible within the scope of the present technology claimed.
[0347] Thus, it should be understood that the materials, methods, and examples provided here are representative of preferred aspects, are exemplary, and are not intended as limitations on the scope of the present technology.
[0348] It should be understood that although the present invention has been specifically disclosed by certain aspects, embodiments, and optional features, modification, improvement and variation of such aspects, embodiments, and optional features can be resorted to by thoseAty. Dkt. No.: 114198-7260 skilled in the art, and that such modifications, improvements and variations are considered to be within the scope of this disclosure.
[0349] The present technology has been described broadly and generically herein. Each of the narrower species and sub-generic groupings falling within the generic disclosure also form part of the present technology. This includes the generic description of the present technology with a proviso or negative limitation removing any subject matter from the genus, regardless of whether or not the excised material is specifically recited herein.
[0350] In addition, where features or aspects of the present technology are described in terms of Markush groups, those skilled in the art will recognize that the present technology is also thereby described in terms of any individual member or subgroup of members of the Markush group.
[0351] All publications, patent applications, patents, and other references mentioned herein are expressly incorporated by reference in their entirety, to the same extent as if each were incorporated by reference individually. In case of conflict, the present specification, including definitions, will control.
Claims
Atty. Dkt. No.: 114198-7260 What is claimed is:
1. A method of treating or preventing metabolic dysfunction-associated steatohepatitis (MASH) in a subject in need thereof, the method comprising administering to the subject an effective amount of an agent that modulates the activity of at least one transcription factor gene regulatory network (GRN) in hepatocytes, wherein the at least one GRN comprises a transcription factor selected from the group consisting of HNF4A, RORA, CUX2, CREB5, NFKB1, NR1I2, RELB, KLF6, TEAD1, ETS1, NFAT5, FOXO1, NR4A1, HNF4G, PPARG, STAT1, PPARA, and ATF3, and wherein said modulating is sufficient to alter the expression of one or more target genes regulated by the at least one GRN.
2. The method of claim 1, wherein the agent increases the activity of a GRN selected from HNF4A, RORA, CUX2, NR1I2, and PPARA in hepatocytes.
3. The method of claim 1, wherein the agent decreases the activity of a GRN selected from NFKB1, RELB, TEAD1, CREB5, KLF6, ETS1, andNFAT5 in hepatocytes.
4. The method of claim 1, wherein the one or more target genes are selected from the group consisting of DTNA, AKAP12, ITGA2, LAMA3, MY07B, CD44, TNFAIP2, TMEM150B, FGF1, SMPD3, LPIN3, SLC25A42, SLC22A1, CRYL1, and PSD4.
5. The method of claim 1, wherein the agent modulates the activity of one or more candidate cis-regulatory elements (cCREs) within the GRN.
6. The method of claim 1, wherein the subject has a MASLD stage selected from MASL, low-fibrosis MASH (Fib- MASH; fibrosis stages 0-2), and high-fibrosis MASH (Fib+ MASH; fibrosis stages 3-4).
7. The method of claim 1, wherein the agent is selected from the group consisting of a small molecule, a nucleic acid, an antisense oligonucleotide, an siRNA, an shRNA, a CRISPR-based gene editing agent, a transcription factor modulator, a protein, a peptide, an antibody or antigenbinding fragment thereof, and a viral vector encoding a transcription factor or a modulator thereof.
8. A method of diagnosing or stratifying metabolic dysfunction-associated steatotic liver disease (MASLD) progression in a subject, the method comprising: (a) obtaining a biological sample comprising liver cells from the subject; (b) determining, in the liver cells, the activity of one or more gene regulatory features selected from: (i) transcription factor gene regulatory networks (GRNs); (ii) candidate cis-regulatory elements (cCREs); (iii) chromatin states at cCREs; and (iv) gene expression of GRN target genes; and (c) classifying the subject into aAty. Dkt. No.: 114198-7260 MASLD stage based on the determined activity, wherein the MASLD stage is selected from normal, MASL, low-fibrosis MASH, and high-fibrosis MASH.
9. The method of claim 8, wherein step (b) comprises determining the activity of GRNs with increased activity in MASL selected from HNF4A, RORA, and CUX2, and GRNs with increased activity in MASH selected from CREB5, NFKB1, RELB, KLF6, TEAD1, ETS1, and NFAT5.
10. The method of claim 8, wherein step (b) comprises determining the proportion of hepatocyte sub-clusters selected from Zone 1, Zone 2, Zone 3, MASH-associated (MASH-a), and fibrosis-associated (Fib-a) hepatocytes.
11. The method of claim 10, wherein an increased proportion of Fib-a hepatocytes and a decreased proportion of Zone 1 and Zone 3 hepatocytes is indicative of high-fibrosis MASH.
12. The method of claim 8, wherein step (b) comprises determining the chromatin state of one or more hepatocyte cCREs, wherein a transition from repressive to active chromatin states at cCREs enriched for NFKB, FOS / JUN, and TEAL) motifs is indicative of MASH progression.
13. The method of claim 8, further comprising determining the genotype of the subject at one or more MASLD-associated risk loci selected from 12ql 3 (KRT8 / KRT18), 8p23 (PPP1R3B), 19ql3 (CEBPA), 22ql2 (XBP1), and 2q37 (EFHD1).
14. A method of identifying a candidate therapeutic agent for the treatment of metabolic dysfunction-associated steatohepatitis (MASH), the method comprising: (a) providing a population of hepatocytes; (b) contacting the hepatocytes with a test agent; (c) determining the effect of the test agent on the activity of one or more transcription factor gene regulatory networks (GRNs) associated with MASH progression, wherein the one or more GRNs are selected from the group consisting of CREB5, NFKB1, RELB, KLF6, TEAD1, ETS1, NFAT5, HNF4A, RORA, CUX2, FOXO1, NR4A1, HNF4G, PPARG, STAT1, PPARA, and ATF3; and (d) selecting the test agent as a candidate therapeutic agent if the test agent shifts the activity of the one or more GRNs toward a non-disease profile.
15. The method of claim 14, wherein determining the effect of the test agent comprises measuring one or more of: (i) chromatin accessibility at cCREs within the GRN; (ii) histone modification levels (H3K27ac and / or H3K27me3) at cCREs within the GRN; (iii) gene expression of target genes within the GRN; and (iv) 3D chromatin contacts at loci associated with the GRN.Aty. Dkt. No.: 114198-7260 16. The method of claim 14, wherein a non-disease profile comprises decreased activity of GRNs selected from NFKB1, RELB, TEAD1, CREB5, KLF6, ETS1, andNFAT5, and increased activity of GRNs selected from HNF4A, RORA, CUX2, and PPARA.
17. The method of claim 14, wherein the hepatocytes comprise primary human hepatocytes or hepatocytes derived from induced pluripotent stem cells (iPSCs).
18. A method of treating liver fibrosis associated with MASH in a subject, the method comprising administering to the subject an effective amount of an agent that disrupts CD44-mediated signaling between hepatic stellate cells (HSCs) and fibrosis-associated hepatocytes, wherein the agent inhibits one or more of: (a) COL1 A1-CD44 interaction between HSCs and fibrosis-adjacent (Fib-adj) hepatocytes; and (b) CD44 expression or activity in hepatocytes.
19. The method of claim 18, wherein the agent is an anti-CD44 antibody, a CD44-targeting siRNA, a CD44 small molecule inhibitor, or a combination thereof.
20. The method of claim 18, further comprising administering an agent that modulates the epigenomic profile of fibrosis-associated hepatocytes, wherein said modulating comprises increasing H3K27me3 -mediated repression at epithelial remodeling and fibrosis-related genes, or decreasing H3K27me3 -mediated repression at lipid metabolism genes.
21. The method of claim 1 or claim 14, wherein the agent modulates the activity of one or more hepatocyte candidate cis-regulatory elements (cCREs) enriched for sequence motifs of one or more transcription factors selected from the group consisting of HNF4A, NR4A1, CEBPA / D, FOS / JUN, TEAD, andNFKB.
22. The method of claim 21, wherein the agent decreases activity of one or more hepatocyte cCREs enriched for FOS / JUN, TEAD, orNFKB motifs.
23. The method of claim 21, wherein the agent increases activity of one or more hepatocyte cCREs enriched for HNF4A, NR4A1, or CEBPA / D motifs.
24. The method of claim 8, wherein step (b) comprises determining increased activity of one or more hepatocyte GRNs associated with lipid metabolism and biosynthesis, wherein the one or more GRNs are selected from the group consisting of HNF4A, RORA, CUX2, and NR1I2.
25. The method of claim 8, wherein step (b) comprises determining increased activity of one or more hepatocyte GRNs associated with epithelial remodeling, cytoskeletal organization, cell-extracellular matrix adhesion, inflammation, or immune signaling, wherein the one or more GRNs are selected from the group consisting of CREB5, NFKB1, RELB, KLF6, TEAD1, ETS1, NFAT5, STAT1, HNF4G, and PPARG.Aty. Dkt. No.: 114198-7260 26. The method of claim 8, wherein step (b) comprises determining expression of one or more genes selected from the group consisting of DTNA, AKAP12, ITGA2, LAMA3, MY07B, CD44, TNFAIP2, TMEM150B, FGF1, SMPD3, LPIN3, SLC25A42, SLC22A1, CRYL1, and PSD4.
27. The method of claim 8 or claim 26, wherein increased expression of one or more of DTNA, AKAP12, ITGA2, LAMA3, MY07B, CD44, TNFAIP2, TMEM150B, and FGF1 is indicative of MASH progression.
28. The method of claim 8 or claim 26, wherein increased expression of one or more of SMPD3, LPIN3, SLC25A42, SLC22A1, CRYL1, and PSD4 is indicative of MASL.
29. The method of claim 8, wherein step (b) comprises determining whether hepatocytes in the sample comprise a fibrosis-associated hepatocyte population expressing one or more of NFKB1, STAT4, HIF1A, SOD2, TEAD1, and TGFBR2.
30. The method of claim 29, wherein the fibrosis-associated hepatocyte population is present in increased proportion in subjects with high-fibrosis MASH relative to normal liver.
31. The method of claim 8 or claim 29, wherein step (b) comprises determining whether the sample comprises fibrosis-adjacent hepatocytes expressing one or more of CXCL8, CXCL6, HSPB8, ITGB6, CDKN2A, and KRT23.
32. The method of claim 8, wherein step (b) comprises determining whether Zone 3 hepatocytes are reduced and fibrosis-associated hepatocytes are increased relative to a reference liver sample.
33. The method of claim 8 or claim 12, wherein step (b) comprises determining whether hepatocytes exhibit increased repression at genes involved in lipid metabolism, transport, or Wnt signaling and / or decreased repression at genes involved in cell structure, adhesion, extracellular matrix interaction, fibrosis, or cytokine signaling.
34. The method of claim 33, wherein the repression is determined by measuring H3K27me3 signal.
35. The method of claim 14 or claim 16, wherein the non-disease profile comprises decreased expression of one or more genes selected from the group consisting of DTNA, AKAP12, ITGA2, LAMA3, MY07B, CD44, TNFAIP2, TMEM150B, andFGFl.
36. The method of claim 14 or claim 16, wherein the non-disease profile comprises increased expression of one or more genes selected from the group consisting of SMPD3, LPIN3, SLC25A42, SLC22A1, CRYL1, and PSD4.Aty. Dkt. No.: 114198-7260 37. The method of claim 14 or claim 16, wherein the non-disease profile comprises decreased activity of one or more candidate cis-regulatory elements enriched for NFKB, FOS / JUN, or TEAD sequence motifs.
38. The method of claim 14 or claim 16, wherein the non-disease profile comprises increased activity of one or more candidate cis-regulatory elements enriched for HNF4A, NR4A1, or CEBPA / D sequence motifs.
39. The method of claim 18, wherein the fibrosis-associated hepatocytes comprise hepatocytes expressing one or more of NFKB1, STAT4, HIF1A, SOD2, TEAD1, and TGFBR2.
40. The method of claim 18 or claim 39, wherein the fibrosis-adjacent hepatocytes comprise a subset of fibrosis-associated hepatocytes expressing one or more of CXCL8, CXCL6, HSPB8, ITGB6, CDKN2A, and KRT23.
41. The method of claim 18, wherein the agent reduces signaling from hepatic stellate cells to fibrosis-adjacent hepatocytes involving COL 1 Al and CD44.
42. The method of claim 20, wherein the epithelial remodeling and fibrosis-related genes comprise genes involved in cell structure, adhesion, extracellular matrix interaction, or cytokine signaling.
43. The method of claim 20, wherein the lipid metabolism genes comprise genes involved in lipid metabolism, lipid transport, or Wnt signaling.
44. The method of claim 13, wherein the one or more MASLD-associated risk loci comprise a locus at 12ql 3 associated with KRT8 / KRT18, a locus at 8p23 associated with PPP1R3B, a locus at 19q 13 associated with CEBPA, a locus at 22ql2 associated with XBP1, or a locus at 2q37 associated with EFHD1.
45. The method of claim 8 or claim 13, wherein the genotype determined at the one or more MASLD-associated risk loci is used together with the activity of one or more transcription factor gene regulatory networks (GRNs), candidate cis-regulatory elements (cCREs), chromatin states, or GRN target genes to classify the subject into a MASLD stage.
46. The method of claim 8, wherein the biological sample is analyzed by one or more methods selected from the group consisting of single-cell RNA sequencing, single-cell multiome analysis, chromatin accessibility profiling, histone modification profiling, 3D chromatin conformation profiling, and spatial transcriptomics.
47. The method of claim 14, wherein determining the effect of the test agent comprises comparing the activity of the one or more GRNs to a reference profile from normal liver tissue.Aty. Dkt. No.: 114198-7260 48. The method of claim 14, wherein determining the effect of the test agent comprises comparing the activity of the one or more GRNs to a reference profile from MASL, low-fibrosis MASH, or high-fibrosis MASH liver tissue.
