A single approach, multiple insights: Harnessing spatial transcriptomics and proteomics to identify novel therapeutic targets of alcohol-associated hepatitis

Introduction

Alcohol-associated hepatitis (AH) is one of the deadliest manifestations of alcohol-associated liver disease (ALD) with limited treatment options and a high mortality rate of approximately 30% within a month following the onset of acute disease (Ali et al., 2022, Yamazaki and Schnabl, 2023). Apart from abstinence and liver transplantation, current treatment modalities for AH are restricted to corticosteroid therapy, which only offers minimal survival benefit (Shah et al., 2023, Thursz et al., 2015). Therefore, it is of significant interest to investigate the molecular mechanisms of AH pathogenesis and identify novel therapeutic targets to mitigate this disease. AH has been shown to affect multiple cell populations including hepatic and non-parenchymal cells, which all contribute to the pathology of AH. Recently, with the advent of omics technology, particularly spatial transcriptomics, researchers are now able to visualize and map the varying cellular and gene expression profiles of tissue sections at a high cellular resolution (Chen et al., 2023). The use of Visium spatial transcriptomics and proteomics on AH patient samples by Gripshover and colleagues offered valuable insights into the cellular plasticity, mitochondrial dysfunction, inflammatory responses, and identification of premalignant hepatocytes in AH (Gripshover et al., 2024). This commentary will discuss the study’s findings, their application to AH research, the methodological strengths and challenges, and potential future directions.

Cellular Plasticity and Regeneration in AH

Gripshover et al. investigated whether there were any cellular differences between AH and non-ALD controls using Visium spatial transcriptomics and proteomics (Figure 1). In brief, they stained the liver sections with hematoxylin and eosin, followed by hybridizing paired mRNA probes and oligo-tagged antibodies to label target RNA and protein, respectively. The mRNA probes and antibodies were then transferred onto a Visium gene expression slide, thereby ligating to spatially preserved barcoded tags (attached on the Visium slide) for library construction. Subsequently, the samples underwent next-generation sequencing to reveal unique RNA and protein signatures among cells. After bioinformatic processing of all obtained data, they were visualized and mapped onto the tissue via Bioturing Lens software. Sixteen Seurat clusters were identified, but only a few of them were shared between the AH and non-ALD control samples. Notably, AH samples had increased amounts of cholangiocytes, endothelial cells, macrophages, and stellate cells, with a concurrent reduction in hepatocytes, which appeared in distinct clusters rather than uniformly distributed throughout the tissue as seen in control samples. This finding implies that either hepatocytes were lost due to increased cell death, or that other cell types were compensating for hepatocyte loss through expansion.

Figure 1: Spatial transcriptomics identifies novel mechanistic insights in Alcohol-Associated Hepatitis.

Human AH or non-ALD samples were prepared as formalin-fixed paraffin-embedded tissue blocks and sectioned onto slides. After de-paraffinization, hematoxylin and eosin (H&E) staining was performed. Following de-crosslinking, paired mRNA probes and oligo-tagged antibodies bound to RNA and protein, respectively. Ligated probes and antibodies were captured onto a Visium slide and labeled with a barcode tag assisted by Visium CytAssist. The tagged products were enzymatically or chemically detached, and the RNA underwent reverse transcription (RT) to generate cDNA. Adapter sequences were attached to the ends of cDNA and oligos. Although not explicitly stated, PCR is used to amplify the cDNA and create library. Using Illumina NextSeq 2000, samples were sequenced, and the output data was visualized via analysis software. AH samples demonstrated increased cellular plasticity (indicated by the PCNA⁺ hepatocytes exhibiting expression patterns similar to cholangiocytes), increased mitochondrial dysfunction by the downregulation of mitochondrial genes involved in oxidative phosphorylation (OXPHOS), increased macrophage-derived exosome activity, and identification of daHeps accompanied with CEACAM8⁺ neutrophils. Illustration was created using BioRender.com.

Cellular plasticity refers to the ability of cells to adapt or transform in response to various stress stimuli or injury. In AH, this process is altered, where the liver is unable to appropriately compensate for the loss of hepatocyte function. Previous studies have demonstrated the presence of cholangiocyte-hepatocyte hybrid cells (Kim et al., 2021, Li et al., 2019), and these hybrid cells are predicted to restore injuries through a trans-differentiation process when hepatocyte proliferation or function is impaired (Deng et al., 2018, Font-Burgada et al., 2015, Schaub et al., 2018). Through spatial transcriptomic analysis, hepatocytes expressing proliferating cell nuclear antigen (PCNA), indicative of liver regeneration, shared transcriptional profiles closer to cholangiocytes in AH samples compared to non-ALD controls, expressing cholangiocyte markers keratin 19 (KRT19), mucin 6 (MUC6), cystic fibrosis transmembrane conductance regulator (CFTR), and secretin receptor (SCTR). This shift in molecular features suggests that the unique pathogenic state of AH could induce a change in cellular identity, contributing to the cellular plasticity typically observed in AH. This observation aligns well with previous studies of ductular reactions, a phenomenon characterized by the proliferation or hyperplasia of biliary epithelial cells, which is well-documented in liver diseases such as ALD and hepatitis (Roskams et al., 2004).

The identification of these altered cellular phenotypes suggests that the emergence of such hybrid cell populations contribute to a failed attempt at liver regeneration, ultimately leading to the loss of liver functionality. These findings establish a foundation for other investigators to decipher the implications and cellular and molecular drivers behind these cells not only in AH but in other liver diseases such as metabolic dysfunction-associated steatohepatitis (MASH) and hepatocellular carcinoma (HCC), where the role or presence of these cells have not been deeply understood. If such hybrid cells could serve as indicators of disease progression across a spectrum of liver diseases, then it would be pivotal to elucidate the mechanisms that govern their formation, persistence, and function.

Mitochondrial Dysfunction in AH Hepatocytes

As a central processing hub for numerous metabolites, the liver highly depends on the mitochondrial function of hepatocytes to maintain metabolic homeostasis. Mitochondrial dysfunction is widely implicated in ALD and AH, contributing to increased oxidative stress, the release of proinflammatory molecules, and impaired energy production (Chen et al., 2023). In line with these previous reports, the use of spatial transcriptomics showed a downregulation of genes involved in oxidative phosphorylation of hepatocytes in AH samples such as ubiquinol-cytochrome C reductase core protein 1 (UQCRC1), ubiquinol-cytochrome C reductase, complex III subunit X (UQCR10), cytochrome C oxidase subunit 6C (COX6C), and cytochrome C oxidase subunit 7A2-like (COX7A2L). This proposes that mitochondrial dysfunction is a defining characteristic of AH and further underscores the advantage of the investigation method. Although mitochondrial dysfunction was previously shown in ALD, this study expands upon this by identifying unique markers and affected regions specific to AH. Therapies that could restore impaired mitochondrial gene expression and function in AH could offer an alternative method of treatment, potentially reversing the defective oxidative phosphorylation process or alleviating oxidative stress.

Inflammation and Macrophage-Derived Exosome Activity

Alcohol is a known inflammatory agent that activates both the innate and adaptive immune responses responsible for disease progression in ALD (Nagy, 2015). One of the players involved in the innate immune response are macrophages, amplifying disease severity by releasing inflammatory mediators and exosomes (Ju and Mandrekar, 2015, Xiang et al., 2023). Spatial transcriptomics from this study unveiled an increased gene expression signature associated with macrophage-derived exosome release and trafficking in AH samples. Representatively, flotillin-1 (FLOT1), ras-related protein rab-7a (RAB7A), and ras homolog family member C (RHOC) were upregulated, proteins associated with extracellular matrix (ECM) organization, cytoskeletal remodeling, and vesicle trafficking. The enrichment of this signature suggests that the secretion of macrophage-derived exosomes is promoted through structural remodeling, which may intensify detrimental cellular signaling pathways and further exacerbate the AH phenotype. Although many investigations showcase the predominant role of macrophage-derived exosomes in response to alcohol exposure, questions remain about the molecular content of these exosomes, the downstream signaling pathways they activate, and their target hepatic cell populations.

Disease-Associated Hepatocytes in AH

Another striking finding from Gripshover et al. is the discovery of disease-associated hepatocytes (daHeps), a subpopulation of hepatocytes with high mutational burden whose presence is suggestive of increased risk for HCC development (Carlessi et al., 2023). To our knowledge, this study is the first to characterize these cells in the setting of ALD, specifically in AH. In alignment with previous studies, daHep gene markers C-X-C motif chemokine ligand 1 (CXCL1) and annexin 2 (ANXA2) were highly expressed in AH samples compared to non-ALD controls. Interestingly, the authors found that the protein marker carcinoembryonic antigen-related cell adhesion molecule 8 (CEACAM8, more commonly known as CD66b) could partition these daHeps. CEACAM8 is a glycosylphosphatidylinositol (GPI)-anchored glycoprotein primarily expressed on mature neutrophils (Ribon et al., 2019). It has been previously documented that cancer-associated fibroblasts can upregulate CEACAM8 expression in neutrophils, promoting their activation and driving them toward an immunosuppressive phenotype in HCC (Arvanitakis et al., 2021, Cheng et al., 2018). Notably, CEACAM8⁺ neutrophils have been implicated in promoting MASH progression by releasing proteases, elastase, myeloperoxidase, and reactive oxygen species (Peiseler and Tacke, 2021). Taken together, this suggests that the co-occurrence of CEACAM8⁺ neutrophils and daHeps in AH samples may signify a synergistic relationship that contributes to exacerbating the inflammatory environment and severity of AH. As the authors note, this data, along with previous reports, may indicate an increased risk for HCC in AH patients. Future investigations to determine the temporal dynamics of the co-expression of CEACAM8⁺ neutrophils and daHeps during the distinct stages of AH development and how this relationship evolves over time may provide further insights into the progression of AH. More in-depth studies are also needed to examine if this joint expression is present in other liver diseases, which could broaden the applicability of these findings and disclose shared mechanisms.

Methodological Innovations, Challenges, and Applications in AH Research

The integration of next-generation sequencing-based spatial transcriptomics and proteomics offers a novel strategy to elucidate the complex pathology of AH. As evidenced by this study, this technique can provide information about the molecular interactions specific to distinct cell populations and regions of the liver. This study confirms the feasibility of the method using a small sample size for both AH and non-ALD controls (n=2 each). An immediate future direction for the work would be the expansion of the cohort both to confirm the findings presented in this original work and to characterize novel disease drivers in the greater population. The inclusion of diverse cohorts may allow for further stratification by unique spatial transcriptomic signatures based on demographic data, comorbidities, and disease severity. Additionally, the methodology presented here is still in its infancy and poses a significant technical challenge to execute and analyze with high rigor due to potential overlapping signals from neighboring cells and high threshold gene expression detection cutoffs. Here, the authors highlighted these challenges and executed the assays with precision, however future technological advances to improve the sensitivity and specificity of the combined next-generation sequencing-based spatial transcriptomics and proteomics approach will greatly enhance the impact of potential findings from this method.

Although other non-mesenchymal cell types were not deeply investigated in this study, their further exploration, through integration with spatial transcriptomics, holds potential to illuminate key targets of AH pathogenesis. For example, AH samples showed increased expression of endothelial cells, which could indicate enhanced vascularity in an attempt to compensate for tissue injury and elevated oxidative stress typically observed in ALD. Oval cells, progenitor cells that could differentiate into hepatocytes or biliary epithelial cells upon liver damage, may offer an alternative perspective for understanding the altered liver regeneration process when hepatocyte function is compromised. These cells could also be important in deciphering the origin and formation of the cholangiocyte-hepatocyte hybrid cells. Additionally, while the study focuses on macrophage-derived exosomes, distinguishing liver-resident Kupffer cells from infiltrating macrophages may shed light on how these different cell types uniquely affect the molecular cargo of exosomes and their downstream effects. Exploring these distinctions could also identify molecular drivers of inflammatory signaling pathways in AH. Lastly, lymphocytes, such as T cells, have been implicated in acute AH as exhibiting upregulated expression of negative immune regulators, programmed cell death protein 1 (PD-1) and T-cell immunoglobulin and mucin domain-containing protein 3 (TIM-3), contributing to an immunosuppressive environment (Markwick et al., 2014). Exploiting this knowledge, spatial transcriptomics could help identify specific immune cell subsets and signaling pathways which could pave the way for novel therapeutic opportunities in AH patients showing limited efficacy with the standard of care corticosteroid therapy.

Conclusion and Future Directions

The study by Gripshover et al. serves as a foundation for advancing AH research through the stellar utilization of spatial transcriptomics combined with proteomics. They report the presence of increased hepatocyte-cholangiocyte hybrid cell populations, mitochondrial dysfunction, inflammatory macrophage-derived exosome activity, and the presence of daHeps with CEACAM8⁺ neutrophils (Figure 1). The progressive advancement of spatial transcriptomics has the potential to uncover novel transcriptomic changes during AH progression and in response to conventional and novel therapies. Future applications of this study focused on identifying shared molecular vulnerabilities among diverse AH patients would potentiate the applicability of the reported findings, while also providing a window of opportunity for translating these findings into the clinic.

Abbreviations

AH

alcohol-associated hepatitis

ALD

alcohol-associated liver disease

ANXA2

annexin 2

CEACAM8

carcinoembryonic antigen-related cell adhesion molecule 8

CFTR

cystic fibrosis transmembrane conductance regulator

COX6C

cytochrome C oxidase subunit 6C

COX7A2L

cytochrome C oxidase subunit 7A2-like

CXCL1

C-X-C motif chemokine ligand 1

daHeps

disease-associated hepatocytes

ECM

extracellular matrix

FLOT1

flotillin-1

HCC

hepatocellular carcinoma

KRT19

keratin 19

MASH

metabolic dysfunction-associated steatohepatitis

MUC6

mucin 6

OXPHOS

oxidative phosphorylation

PCNA

proliferating cell nuclear antigen

PD-1

programmed cell death protein 1

RAB7A

ras-related protein rab-7a

RHOC

ras homolog family member C

SCTR

secretin receptor

TIM-3

T-cell immunoglobulin and mucin domain-containing protein 3

UQCRC1

ubiquinol-cytochrome C reductase core protein 1

UQCR10

ubiquinol-cytochrome C reductase, complex III subunit X