Previous Article in Journal
Application of Plasma Metabolomic Biomarker Panels in Early Diagnosis and Disease Staging of Alzheimer’s Disease
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Metabolites
Volume 16
Issue 6
10.3390/metabo16060378
metabolites-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Genetically Informed Single-Cell Analysis Reveals PLXND1 as a Cell-Type-Specific Molecular Switch in MASLD

by
Xianyi Ma
†,
Junbo Song
†,
Xin Hong
and
Zhibin Lin
*
Department of Hepatobiliary Surgery, Xijing Hospital, Fourth Military Medical University, Xi’an 710032, China
*
Author to whom correspondence should be addressed.
†
These authors contributed equally to this work.
Metabolites 2026, 16(6), 378; https://doi.org/10.3390/metabo16060378 (registering DOI)
Submission received: 26 April 2026 / Revised: 25 May 2026 / Accepted: 27 May 2026 / Published: 30 May 2026
(This article belongs to the Section Endocrinology and Clinical Metabolic Research)
Browse Figures
Versions Notes

Abstract

Background/Objectives: Metabolic dysfunction-associated steatotic liver disease (MASLD) is a systemic disorder driven by genetic predisposition, epigenetic programming, metabolic rewiring, and immune dysregulation. Although population genetics and single-cell transcriptomics have advanced our understanding, the multi-omic causal architecture of MASLD at cellular resolution remains poorly defined. This study aimed to establish an integrative framework linking genetic causality to cell-type-specific tissue dysfunction. Methods: Multi-layered Mendelian randomization (MR) and summary-data-based MR (SMR) across large-scale eQTL and pQTL datasets were applied to prioritize causal genes. Single-cell eQTL-based MR across 14 immune lineages generated cell-type-specific causal hypotheses, which were validated using human hepatic single-cell RNA-sequencing data (GSE136103). Two-step mediation MR quantified upstream epigenetic and downstream metabolic mechanisms. A high-fat diet (HFD)-induced murine model provided organismal validation. Results: Multi-layered MR nominated PLXND1 as a robust causal driver of MASLD. Single-cell eQTL-based MR revealed a functional dichotomy: PLXND1 upregulation in CD8+ effector memory T-cells decreased MASLD risk (OR = 0.486, 95% CI: 0.290–0.813, p = 0.006), whereas upregulation in natural killer cells (OR = 1.567, 95% CI: 1.337–1.837, p < 0.001), non-classical monocytes, and dendritic cells increased risk. Human hepatic single-cell transcriptomics confirmed that PLXND1 marks an anti-fibrotic, IFNG-high CD8+ T subset and a pro-inflammatory lipid-associated macrophage (LAM) population. Mediation MR identified DNA methylation at cg26767922 and cg08471739 as protective mediators acting predominantly via PLXND1 downregulation (92.39% and 64.50% mediation, respectively), and linked PLXND1 to six circulating metabolites. HFD mice showed significant hepatic PLXND1 upregulation. Conclusions: PLXND1 functions as a lineage-dependent molecular switch in MASLD, validated across genetic, epigenetic, metabolic, and single-cell dimensions. These findings caution against systemic PLXND1 blockade and support precision therapeutic strategies targeting hepatic innate immune cells.

1. Introduction

Metabolic dysfunction-associated steatotic liver disease (MASLD) has emerged as the defining chronic liver disorder of the metabolic era, with a rapidly rising global prevalence that shows no sign of geographical containment [1,2,3,4]. Although the earliest stage presents as seemingly benign hepatic steatosis, the disease can progress relentlessly through fibrosis and cirrhosis to hepatocellular carcinoma, inflicting substantial burdens on healthcare systems, economic productivity, and social welfare [5,6,7,8,9,10]. Despite major strides in understanding the physiological underpinnings of MASLD, effective pharmacological interventions remain scarce, and most approved or late-stage therapies offer only modest benefit [11,12,13,14]. This therapeutic gap reflects the fundamental complexity of MASLD pathogenesis, which cannot be reduced to a single insult; rather, it arises from a dynamic interplay among genetic predisposition, epigenetic programming, metabolic reorganization, and cross-talk between the liver and peripheral immune system [8,15,16,17].
Mendelian randomization (MR) provides a rigorous, quasi-experimental framework to interrogate causal associations in complex diseases [18,19,20]. Driven by increasingly powered genome-wide association studies and the integration of multi-omics datasets—including expression (eQTL), protein (pQTL), and epigenetic quantitative trait loci—MR is uniquely positioned to prioritize mechanistic biomarkers and therapeutic targets [21,22]. However, current MR investigations into MASLD remain highly fragmented, typically analyzing individual molecular layers—such as gene expressions, metabolites, or immune profiles—in isolation. This siloed approach obscures how disease risk propagates across multi-omic strata [23,24,25] and fails to identify the interconnected regulatory hubs that orchestrate MASLD pathogenesis [26,27].
To address these critical limitations, we developed a genetically informed, multi-scale analytical framework that sequentially links population-level causality with single-cell phenotypic validation in human MASLD tissue. By integrating multi-layered two-sample MR, summary-data-based MR (SMR), and cross-omics mediation analyses with high-resolution single-cell eQTL data and human hepatic single-cell RNA-sequencing, we sought to map the causal architecture of MASLD from epigenetic regulation through metabolic remodeling to immune microenvironmental dysfunction [18,28,29]. This approach enabled us to nominate PLXND1—a semaphorin receptor with lineage-restricted immune expression and established roles in immune cell trafficking and endothelial remodeling [30]—as a causal driver and to predict and directly validate its striking, cell-type-specific functional dichotomy within the steatotic liver.
Our study moves beyond descriptive association to establish a coherent, cross-omics causative model of MASLD pathogenesis. By closing the loop between genetic prediction and cellular phenotype, we provide a mechanistic blueprint for precision therapeutic strategies that target specific hepatic immune compartments rather than pursuing systemic, blunt inhibition, thereby offering a more rational path toward clinical translation.

2. Materials and Methods

2.1. eQTL Dataset

The information about local expression quantitative trait loci was based on whole-blood samples of the eQTLGen repository (https://www.eqtlgen.org/cis-eqtls.html, accessed on 10 October 2025) [31]. In order to identify strong genetic proxies, we used tight inclusion criteria, namely, genome-wide significance cut-offs (p < 5 × 10−8) and linkage disequilibrium pruning (r2 < 0.1) within a 10,000-kb range. Such an intensive screening finally resulted in finding 15,695 unique loci with adequate local regulatory variants.

2.2. pQTL Dataset

The unadjusted summary-level data on pQTL were downloaded from the deCODE repository (https://www.decode.com/summarydata/, accessed on 10 October 2025) [32] and involved 4907 proteomic traits measured among an Icelandic cohort. Cis-acting variants were selected through imposing tight limits in terms of geography: the candidate single-nucleotide polymorphisms (SNPs) had to be located on the same chromosome and had to lie wholly inside a 1-megabase region that enclosed the target gene sequence. Such a geographical limitation identified 4706 cis-oriented assemblies of interest. In order to obtain stringent genetic tools, independent variants were obtained using a strict genome-wide significance threshold (p < 5 × 10−8) and linkage disequilibrium clumping (distance cutoff = 10,000 kb; r2 < 0.1). After such quality control steps, there were 1615 strong groups of genetic proxies still accessible for further analysis.

2.3. Outcome Dataset

The summary statistics of MASLD were released by the FinnGen consortium (release R12; https://www.finngen.fi/en, accessed on 10 October 2025), consisting of 3504 MASLD patients and 496,844 controls of European ancestry. The large amount of data combines genetic and clinical information, enabling a detailed study of the genetic relationships in MASLD.

2.4. Single-Cell eQTL Dataset (sc-eQTL)

The transcriptomic profiles of individual cells were collected using the OneK1K resource (https://onek1k.org/, accessed on 11 December 2025) [33] which is a large library containing about 1.27 million peripheral blood mononuclear cells (PBMCs) sampled in 982 different individuals. We used publicly available, annotated transcriptomic profiles of 14 different immune cell subsets to map the expression patterns of the target gene. This fine-grained method made it possible to identify its distribution in certain immune populations and examine its role in the development of MASLD pathogenesis at the level of a single cell.

2.5. Single-Cell RNA-Sequencing Analysis of Human Hepatic Tissues

Publicly available single-cell RNA-sequencing data from human normal and MASLD liver samples were retrieved from the Gene Expression Omnibus (GEO) database under accession number GSE136103. The Seurat package in R was employed for quality control, normalization, dimensionality reduction, clustering, and visualization. For T-cell subclustering, all T-cells were extracted and re-clustered, followed by annotation into CD4+ and CD8+ subsets using lineage-specific markers. Differential gene expression analysis was performed using the Wilcoxon rank-sum test with Bonferroni correction. Gene Ontology (GO) biological process enrichment and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis were conducted using the clusterProfiler package (version 4.10.0). For macrophage analysis, cells were stratified into PLXND1-positive and PLXND1-negative populations, and differential expression and functional enrichment analyses were performed accordingly.

2.6. Two-Sample Mendelian Randomization Analysis

We have applied MR models through the TwoSampleMR computational package (v.0.6.13) to interrogate the causal mechanisms underlying MASLD. Our predictive factors were genetic architectures that defined cis-eQTL and cis-pQTL expression, which we compared with the disease phenotype. The protection against weak-instrument distortion required eliminating polymorphisms with F-values less than 10. The main statistics model used in our case was inverse variance-weighted (IVW) protocols, which led to the exclusion of all the traits with low variant counts to be able to apply strict IVW protocols. Further validation measures included MR-Egger intercepts to identify pleiotropic confounders, Cochran’s Q equations to measure variant dispersion and systematic leave-one-out permutations to ensure overall inferential resilience.

2.7. Summary-Data-Based Mendelian Randomization (SMR) Analysis

In order to explore possible causal structures between particular transcriptomic or proteomic profiles and MASLD etiology, we have implemented summary-data-based Mendelian randomisation frameworks based on genome-wide and quantitative-trait-locus summary statistics. The Heterogeneity in Dependent Instruments (HEIDI) protocol was incorporated in our pipeline in order to ensure that true pleiotropy could be distinguished in a definitive manner as opposed to simple genetic confounding. The statistical analysis is implemented with a null hypothesis of no horizontal pleiotropic effect; therefore, maintaining the null implies that every observed molecular association is due to a single common polymorphism rather than separate but tightly linked loci that are in linkage disequilibrium. Computational implementation was based upon the SMR suite release 1.3.1 (https://yanglab.westlake.edu.cn/software/smr/#Overview, accessed on 10 October 2025), and standard operating parameters were used.

2.8. Mediation Analysis

2.8.1. Upstream Analysis

We used a two-stage MR mediation analysis in an attempt to establish whether epigenetic modifications can regulate transcriptional outputs which eventually lead to the development of MASLD. mQTL statistics, available through the GoDMC consortium registry (http://www.GoDMC.org.uk/, accessed on 11 December 2025) [34], were taken as basic exposures. Instrumental variants were filtered through rigorous criteria, including keeping only those with genome-wide significance (p < 5 × 10−8) and rigid independence conditions (r2 < 0.1 in a 10,000-kb genome region). Our analysis pipeline measured three different causal measures: the overall effect of methylation state on the liver condition (βall), the epigenetic effect that determines particular transcript values (β1), and the downstream effect of these RNA transcripts that influence the disease predisposition (β2). We then calculated the indirect mediated pathway by multiplying these intermediate pathways (β12 = β1 × β2) and defined the partial epigenetic effect using the fraction (β12/βall) × 100%.

2.8.2. Downstream Analysis

In order to decode other physiological processes, we have used a two-step MR approach that determines if systemic circulating intermediates can affect the effects of target transcripts on hepatic steatosis. The summary statistics involving 1400 different blood-borne biochemicals were obtained through the official GWAS Catalog database (accession numbers GCST90199621-GCST90201020; https://www.ebi.ac.uk/gwas/, accessed on 11 December 2025). Instrumental variable extraction parameters were exactly based on our previous set filtration criteria. Three important causal vectors were measured: the overall clinical effect which connects transcript amounts with disease phenotype (βall), the initial vector that relates RNA abundance with circulating metabolic substances (β1), and final vector that relates these systemic biomarkers with the development of pathology (β2). The indirect mediated pathways (β12 = β1 × β2) and their respective fractional effects were mathematically calculated by using the same formulaic pipeline described above.

2.9. MASLD Mouse Model

Wild-type C57BL/6 male mice (8 weeks old; 20–22 g) were obtained through Shaanxi Yike Medical Biotechnology Co., Ltd (Xi’an, Shaanxi, China). The mice were kept on standard laboratory settings (20–25 °C, 30–50 percent relative humidity, and 12 h of light and darkness). After a habituation period (of seven days), we randomly allocated the murine models on standard laboratory chow (ND) or high-fat diet (HFD; containing 60% of calories as lipids). Each in vivo protocol strictly complied with the ethical requirements set by the Air Force Medical University regulation committee on animal welfare.

2.10. Histological Staining of Murine Liver Tissues

Following 12 weeks of dietary intervention, liver tissues from both normal chow and HFD groups were harvested, fixed in 4% paraformaldehyde, and embedded in paraffin. For hematoxylin and eosin (H&E) staining, 5-μm sections were deparaffinized, rehydrated, and stained with hematoxylin and eosin to evaluate general hepatic architecture, steatosis, and hepatocyte ballooning degeneration. For Oil Red O staining, frozen liver sections (8 μm) were fixed in 4% paraformaldehyde, rinsed with distilled water, and stained with 0.5% Oil Red O working solution for 15 min to visualize neutral lipid accumulation. Sections were then counterstained with hematoxylin and examined under a light microscope. For Masson’s trichrome staining, paraffin-embedded liver sections were deparaffinized and rehydrated, followed by staining with Weigert’s iron hematoxylin, Biebrich scarlet-acid fuchsin, and aniline blue to assess collagen deposition and early fibrotic changes. All histological images were captured under a light microscope (Leica Microsystems, Wetzlar, Germany) and analyzed by investigators blinded to the experimental groups.

2.11. Quantitative Real-Time PCR (qRT-PCR)

Total RNA was isolated with TRIzol reagent (Invitrogen, Carlsbad, CA, USA) as per the recommendations of the manufacturer. Reverse transcription of RNA into cDNA was performed with Evo M-MLV RT Premix (Accurate Biology, Changsha, China).
SYBR Green Premix Pro Taq HS (Accurate Biology) was used as the reagent in qRT-PCR and the reaction was conducted on a Bio-Rad CFX Maestro 2.2 system (Bio-Rad, Hercules, CA, USA). The internal control was β-actin.
Primer sequences were as follows:
PLXND1 forward: 5′-CGCAACCGTAGCCTAGAAGAC-3′
PLXND1 reverse: 5′-GGTTAAGGTCGAAGGTGAAGAG-3′
β-actin forward: 5′-GTGACGTTGACATCCGTAAAGA-3′
β-actin reverse: 5′-GCCGGACTCATCGTACTCC-3′

2.12. Western Blot

The liver samples were homogenized with RIPA extraction reagent (Thermo Fisher Scientific, Waltham, MA, USA, Cat# 89900). Following this, total peptide yields were measured by bicinchoninic acid colorimetry. Electrophoretic resolution of polypeptide fractions was performed on sodium dodecyl sulfate–polyacrylamide gel and electroblotted onto hydrophobic PVDF membranes purchased at Millipore (Billerica, MA, USA). The particular blots were probed overnight with goat-specific anti-PLXND1-recognizing immunoglobulins (Abcam, Cambridge, UK; Cat# ab28762) and finally incubated with horseradish peroxidase-conjugated anti-goat conjugates obtained from donkeys (Abcam, Cambridge, UK; Cat# ab97110).

2.13. Statistical Analysis

The R environment (v4.4.2; R Core Team) and GraphPad Prism (v10.0) were used to perform computational processing and statistical analyses. We have chosen a bilateral p-value of less than 0.05 as the threshold of statistical significance in all inferential tests.

3. Results

3.1. Multi-Layered MR Identifies PLXND1 as a Causal Driver of MASLD

We have conducted a two-sample MR analysis to explore the possible causal relationships between gene expression and MASLD in the context of transcriptomic and proteomic levels. The exposure variables were cis-acting quantitative trait loci, such as cis-expression quantitative trait loci (cis-eQTL) and cis-protein quantitative trait loci (cis-pQTL), and the outcome was summary-level data of a MASLD genome-wide association study (GWAS). All SNPs included in the instrumental variable selection had statistics of over 10, so it is highly unlikely that weak-instrument bias would occur. The inverse-variance weighted (IVW) was implemented as the main analytic strategy, and the value of p less than 0.05 was considered statistically significant. We found 1475 genes at the transcriptional (Supplementary Table S1) and 392 genes at the protein level (Supplementary Table S2) that are significantly and causally related to MASLD. Integration of these findings across layers generated 27 candidate genes with potential causal relevance to MASLD across both regulatory levels (Figure 1A). Additional evaluation of the effect direction showed that 18 of these genes had concordant causal patterns (10 of them were consistently correlated with higher risk of MASLD, and the other 8 with lower risk) (Figure 1B,C).
Evaluation of SMR of the 18 candidate genes indicated that the expression of PLXND1 (P_SMR = 0.0256), GSTZ1 (P_SMR = 0.0316) and RPIA (P_SMR = 0.0422) was associated with MASLD risk. In order to exclude false linkage disequilibrium, we have used HEIDI testing. All three loci passed this threshold (P_HEIDI > 0.05), confirming that direct causality or pleiotropy is the underlying cause of these associations (Table 1). Because PLXND1 is predominantly expressed in specific leukocyte subsets and has been implicated in immune cell–endothelial crosstalk, its causal effect on MASLD likely operates through cell-type-specific mechanisms that are invisible to bulk-tissue analyses. We therefore decomposed its genetic effect across individual immune lineages using single-cell eQTL-based MR.

3.2. Single-Cell eQTL MR Reveals a Functional Dichotomy of PLXND1 Across Immune Lineages

The MASLD immune landscape is also well documented to be patchy, and therefore we need to move beyond tissue analysis to achieve the real picture of the molecular dance taking place. We have measured the genetically estimated expression levels of PLXND1 in 14 peripheral blood immune subpopulations using single-cell eQTL data with the OneK1K cohort. The findings of the MR revealed the existence of a new reality. Rather than a nonspecific effect, the influence of PLXND1 on MASLD vulnerability is highly cell specific, working as a polarized rheostat among different immune elements (Figure 1D).
The gene seems to be used by the adaptive immune system as a defensive tool. Genetic fixity of increased PLXND1 expression exclusively in CD8+ effector T-cells (CD8+ Tem) means an unexpectedly strong decrease in the risk of MASLD (OR = 0.486, 95% CI = 0.290–0.813, p = 0.0059). However, entering the domain of the innate immune system reverses this situation completely, converting the gene into a particular liability. The maximum risk amplification impact occurred between natural killer (NK) cells (OR = 1.567, 95% CI = 1.337–1.837, p < 0.001), and an analogous pathogenic signature was also observed in non-classical monocytes and dendritic cells (OR = 1.511, 95% CI = 1.094–2.085, p = 0.0122). This dramatic functional dichotomy, protective in the adaptive arm but feeding pathogenesis actively along the innate lineages, places PLXND1 as a very context-specific, core node controlling the disease architecture.

3.3. Human Hepatic Single-Cell Transcriptomics Validate the Cell-Type-Specific Roles of PLXND1

While the sc-eQTL MR predictions are genetically grounded, they derive from peripheral blood and do not reveal the transcriptional programs or disease-associated states that these cells adopt within the steatotic liver itself. To bridge this gap and test whether the predicted functional dichotomy is reflected in human MASLD tissue, we analyzed publicly available single-cell RNA-sequencing data from normal and MASLD liver samples (GSE136103). Uniform manifold approximation and projection (UMAP) resolved nine major cell populations, including T-cells, B cells, macrophages, endothelial cells, and epithelial cells (Figure 2A,B). Comparison of cellular proportions between control and MASLD livers revealed a restructured immune microenvironment in the diseased state, with notable shifts in immune cell representation. Strikingly, the proportion of PLXND1-expressing cells was markedly elevated in MASLD livers compared to controls (11.3% versus 4.7%; Figure 2C). Feature plots mapping PLXND1 expression onto the UMAP space showed that the PLXND1 signal was most concentrated in endothelial cells and specific immune subsets, with a clear expansion of the PLXND1-positive population in MASLD (Figure 2D). Quantification across cell types confirmed that endothelial cells harbored the highest PLXND1 expression levels, while macrophages and T-cells also displayed detectable and disease-enriched expression (Figure 2E), establishing that PLXND1-upregulated cells constitute a genuine cellular component of the human MASLD liver.
We next dissected the T-cell compartment to interrogate the protective role of PLXND1 in adaptive immunity. Subclustering of all T-cells distinguished CD4+ and CD8+ lineages (Figure 2F). Among the CD8+ T-cell subsets, Cluster 4 exhibited the highest PLXND1 expression and was proportionally expanded in MASLD livers. Differential expression analysis comparing Cluster 4 against all other CD8+ T-cell clusters revealed a distinct transcriptional signature dominated by cytotoxic effector molecules, including killer lectin-like receptors (KLRC1, KLRD1), the transmembrane receptor CD160, the adaptor TYROBP, and the cytokine IFNG, alongside potent chemoattractants XCL1 and XCL2 (Figure 3A). Gene Ontology biological process enrichment demonstrated significant enrichment in leukocyte-mediated immunity, positive regulation of cell killing, cellular response to type II interferon, and lymphocyte chemotaxis (Figure 3B). These transcriptional features—characterized by high IFNG and cytotoxic molecule expression—suggest that PLXND1 marks an effector CD8+ T-cell subset with anti-fibrotic potential, aligning with the protective causal effect of genetically elevated PLXND1 in CD8+ effector memory T-cells.
In mirror image to the protective adaptive signature, the sc-eQTL MR had predicted that PLXND1 amplifies MASLD risk along innate immune lineages. To identify the cellular phenotype underlying this pathogenic effect, we compared PLXND1-positive versus PLXND1-negative macrophages within the same human MASLD dataset. The volcano plot of differentially expressed genes revealed that PLXND1-positive macrophages were robustly characterized by the upregulation of established lipid-associated macrophage (LAM) markers, including TREM2, APOC1, APOA1, PLA2G7, PLTP, and FABP1 (Figure 3C). These cells also showed elevated expression of pro-inflammatory mediators such as FCGR1A and CCR5. Gene Ontology enrichment analysis underscored a strong metabolic reprogramming phenotype, with significant enrichment in lipid transport, receptor-mediated endocytosis, lipid catabolic process, cholesterol efflux, and plasma lipoprotein particle remodeling (Figure 3D). Fisher’s exact test indicated that PLXND1-positive macrophages were significantly overrepresented in MASLD livers compared to controls (19.1% versus 5.9%, p < 0.001). Collectively, these findings establish that PLXND1 in macrophages marks a metabolically active, pro-inflammatory LAM subset that accumulates in the MASLD liver, providing a plausible cellular mechanism for the risk-enhancing causal effect of PLXND1 in innate immune lineages.
Taken together, these human hepatic single-cell transcriptomic data validate the functional dichotomy predicted by genetic MR: PLXND1 is associated with protective cytotoxic effector programs in CD8+ T-cells, whereas it defines a pathogenic lipid-associated macrophage phenotype in the innate immune compartment. This context-dependent behavior positions PLXND1 not as a uniform disease driver or suppressor, but as a cell-type-specific molecular switch in the MASLD immune microenvironment.

3.4. Cross-Omics Mediation Links PLXND1 to Epigenetic and Metabolic Remodeling

In addition to its cell-type-specific immunological roles, the causal impact of PLXND1 on MASLD is shaped by upstream regulatory mechanisms and downstream systemic effectors. We therefore turned to cross-omics mediation analyses to quantify how epigenetic modifications and circulating metabolites propagate disease risk through PLXND1. The data revealed a signal that was surprisingly clear: methylation of two separate CpG sites, cg08471739 and cg26767922 lead to a highly significant causal switch in the disease susceptibility (Figure 4A). And, more specifically, how do these epigenetic switches work? It was quite obvious that we should ask ourselves if they can directly control the outcome of MASLD by promoting or inhibiting the expression of targeted genes. In accordance with this causal chain (β1), as expected, the two loci are significantly influencing the expression of PLXND1 (Table 2). To examine the cascade wiring in more detail, we implemented a mediation MR model. The readouts were really shocking. The shielding of MASLD of cg26767922 is not just an indirect by-product of some other process—no less than 92.39 percent of it is caused by structural processes, in particular, by the specific downregulation of PLXND1 (Figure 4B). The subsequent locus, cg08471739, performs in the same way though perhaps somewhat weaker as inhibition of PLXND1 accounts of about 64.50 percent of its total protective impact (Figure 4B).
The massive PLXND1 effect on MASLD risk means that a larger downstream metabolic signature is present. The physiological cascade caused by this genetic switch necessitated expanding our mediation MR model to actively seek through the circulating effector molecules. Our first step involved identifying 69 different blood metabolites which may be associated with the presence of MASLD due to their demonstrable relationship (Supplementary Table S3). The initial challenge was to locate PLXND1 in the correct position within this metabolic hierarchy. With PLXND1 eQTL as our primary exposure, we utilized it to identify whether the gene is actually an upstream regulator that causes these specific alterations. This readout was exceptionally narrow. It turned out that the expression of PLXND1 directly controls the serum levels of just six different metabolites (Table 3). Equipped with this surprising clear molecular bridge, we performed a final mediation MR experiment measuring the indirect effects passing via this novel pathway (Figure 4C).

3.5. Hepatic PLXND1 Is Upregulated in HFD-Induced MASLD Mice

While the multi-omics computational analyses above establish a coherent causal chain from epigenetic perturbation to metabolic remodeling, they remain inferential; experimental validation is required to confirm that PLXND1 is genuinely upregulated in the steatotic liver. To establish a relevant in vivo disease context for the computational and human tissue findings, we utilized a HFD-induced murine model of MASLD. Male C57BL/6 mice were fed either a normal chow diet or an HFD (60% kcal from fat) for 12 weeks. Histological evaluation of liver sections confirmed successful metabolic injury induction in the HFD group. Hematoxylin and eosin (H&E) staining revealed pronounced hepatocyte ballooning degeneration and steatosis in HFD-fed mice compared to the normal hepatic architecture observed in NC controls (Figure 5A). Oil Red O staining demonstrated markedly increased lipid accumulation within hepatocytes of the HFD group (Figure 5A). Furthermore, Masson’s trichrome staining indicated elevated collagen deposition, signifying the presence of early fibrotic changes in steatotic livers (Figure 5A).
Having confirmed the MASLD phenotype at the tissue level, we next assessed hepatic PLXND1 expression. Quantitative real-time PCR showed a significant increase in PLXND1 transcript abundance in the livers of HFD-fed mice relative to NC-fed controls (Figure 5B). Immunoblot analysis corroborated this finding at the protein level, with densitometric quantification revealing a robust elevation of hepatic PLXND1 in the MASLD cohort (Figure 5C). Collectively, these multi-level data demonstrate that PLXND1 is significantly upregulated in the steatotic liver, providing experimental validation that the gene is not merely a statistical artifact of genetic inference but is actively engaged in the in vivo pathogenesis of MASLD.

4. Discussion

Current MR studies of MASLD have made important strides in identifying causal risk factors and prioritizing therapeutic targets [18,35,36]. Large-scale genome-wide association studies coupled with expression and protein quantitative trait loci have nominated numerous genes and pathways that influence disease susceptibility [26,37]. However, the field remains largely fragmented as most investigations analyze transcriptomic, proteomic, or metabolomic layers in isolation while the causal propagation of risk across these strata is rarely dissected in an integrated manner [38,39]. Consequently, while individual molecular associations have accumulated, the interconnected regulatory hubs that truly orchestrate MASLD pathogenesis have remained elusive [27,40,41,42].
In the present study, we focused on PLXND1 as a central node that bridges upstream epigenetic regulation, downstream metabolic reprogramming, and cell-type-specific immune dysfunction. Our multi-layered MR framework first established PLXND1 as a robust causal driver of MASLD at both the transcriptomic and proteomic levels. Mediation analyses then demonstrated that DNA methylation at cg26767922 and cg08471739 exerts protective effects primarily through the transcriptional downregulation of PLXND1, and that PLXND1 propagates disease risk via six circulating blood metabolites. Most importantly, single-cell eQTL-based MR revealed a striking functional dichotomy: genetically elevated PLXND1 expression was protective in CD8+ effector memory T-cells but pathogenic in natural killer cells, non-classical monocytes, and dendritic cells. Human hepatic single-cell transcriptomics directly validated these predictions by showing that PLXND1 marks a cytotoxic, IFNG-high CD8+ T subset with anti-fibrotic potential, while in macrophages it defines a disease-enriched, pro-inflammatory LAM subset. Finally, in vivo validation in high-fat diet-induced MASLD mice confirmed significant hepatic upregulation of PLXND1 in steatotic livers. Together, these findings trace a continuous causal chain from epigenetic perturbation through metabolic remodeling to immunometabolic dysfunction.
The central advance of our work lies in its deliberate bridging of population-level genetic causality with tissue-level cellular phenotyping. Previous multi-omics studies in MASLD have successfully mapped genetic, transcriptomic, and proteomic associations in parallel [26,43,44,45], yet they have largely remained descriptive or correlative, without establishing how a single molecular hub drives distinct cellular behaviors in the diseased liver [41,46]. Similarly, single-cell transcriptomic atlases of MASLD have richly documented hepatic immune heterogeneity [47], but because they lack genetic instrumentation, they cannot distinguish whether observed gene-expression patterns are causally implicated in disease or merely represent downstream consequences of metabolic injury [48,49]. The integration of single-cell eQTL data with MR has recently emerged as a powerful strategy to resolve cell-type-specific causal effects in autoimmune and infectious diseases [50], but its application to MASLD—particularly with direct validation in human steatotic livers—has, to our knowledge, not been attempted. By genetically predicting a cell-type-specific functional dichotomy and then corroborating it in human hepatic single-cell data, we close a critical gap between statistical causality and biological mechanism.
Several specific innovations distinguish this study from existing studies. First, rather than treating epigenetics, transcriptomics, metabolomics, and immunology as parallel but disconnected stories, we modeled them as sequential links in a unified causal chain. This enabled us to quantify, for example, that 92.39% of the protective effect of cg26767922 operates through PLXND1 downregulation, and to trace how PLXND1 upregulation reshapes circulating metabolites that in turn influence MASLD risk. Second, the identification of PLXND1 as a lineage-dependent molecular switch—protective in adaptive cytotoxic T-cells yet pathogenic in innate LAMs—challenges binary classifications of genes as uniformly beneficial or harmful. This context-dependent behavior explains why bulk-tissue transcriptomic studies of MASLD have often reported contradictory immune signatures: opposing effects across cell populations become attenuated or erased when tissue is homogenized. Third, our findings carry direct translational implications. Systemic pharmacological blockade of PLXND1 would simultaneously eliminate the protective cytotoxic CD8+ T-cell response and suppress the pathogenic LAM phenotype. Our data therefore support a shift toward cell-type-specific precision strategies—such as macrophage-targeted nanocarriers or liver-homing biologics—that selectively modulate PLXND1 in hepatic innate immune cells rather than bluntly inhibiting it throughout the organism.
The precise downstream mechanisms through which PLXND1 exerts its lineage-specific effects remain to be defined. As a direct receptor for class 3 semaphorins—most notably Sema3E—PLXND1 canonically signals through its intracellular GAP domain to suppress R-Ras/M-Ras activity, thereby regulating cytoskeletal remodeling, integrin-dependent adhesion, directional migration, and vascular morphogenesis [51,52]. Whether analogous semaphorin-guided mechanisms operate in hepatic innate immunity—for instance, directing PLXND1-expressing LAMs toward lipid-laden hepatocytes—remains speculative, and the downstream effector cascades operative in this context are unknown. The present single-cell transcriptomic analysis independently revealed enrichment of PLXND1+ macrophages in lipid transport, PPAR signaling, and receptor-mediated endocytosis gene sets, providing transcriptional evidence for a functional link between PLXND1 expression and the lipid-sensing and efferocytosis programs characteristic of LAMs in steatotic liver disease [53,54,55]. In CD8+ effector memory T-cells, PLXND1 has been implicated in modulating cell motility [56], and it is conceivable that semaphorin-guided positioning may contribute to cytotoxic T-cell trafficking within fibrotic liver tissue; however, the proposed coordination of XCL1/XCL2 secretion through PLXND1 signaling remains entirely hypothetical and requires experimental validation [57,58]. Collectively, these observations suggest that semaphorin-receptor signaling, immune cell positioning, and transcriptional lipid-metabolic programming may represent candidate mechanisms linking PLXND1 to MASLD-associated immune responses. Defining the precise context-dependent cascades will require receptor–ligand interaction profiling, phosphoproteomic characterization of PLXND1 signaling complexes in primary hepatic immune cells, and cell-type-resolved functional assays.
Several limitations warrant consideration. First, although MR-Egger intercept tests, Cochran’s Q assessments, and leave-one-out permutations were performed, residual horizontal pleiotropy cannot be entirely excluded. The concordant causal directionality across both eQTL- and pQTL-derived instruments, together with SMR-HEIDI co-localization confirming a shared causal variant rather than linkage disequilibrium artifacts, nonetheless supports the robustness of our principal estimates. Second, the underlying GWAS statistics derive predominantly from European-ancestry cohorts—FinnGen R12, eQTLGen, and deCODE—limiting generalizability to non-European populations, including East Asian groups in whom MASLD genetic architecture and environmental determinants may differ substantially. Third, single-cell analyses relied on a single publicly available dataset (GSE136103); the PLXND1-expressing LAM and CD8+ T-cell phenotypes described herein require corroboration in independent datasets with matched clinical and histological annotation. Fourth, while the HFD murine model confirms PLXND1 upregulation in steatotic liver at both transcript and protein levels, the absence of cell-type-specific conditional knockout experiments leaves the directional causal contributions of PLXND1 within individual hepatic immune compartments formally unestablished.
These limitations define a clear path forward. Replication in large-scale multi-ethnic resources—including Biobank Japan and the China Kadoorie Biobank—will establish cross-population generalizability, while independent scRNA-seq profiling of geographically diverse MASLD biopsies will validate the immune phenotypes identified here. Myeloid-specific and cytotoxic T-cell-specific conditional knockout models will permit direct in vivo interrogation of lineage-dependent contributions to steatohepatitis and fibrosis; parallel dissection of Sema3E–PLXND1–RhoA/PI3K signaling in sorted hepatic immune populations will identify the molecular nodes most amenable to intervention. Macrophage-targeted delivery platforms capable of selectively silencing Plxnd1 in hepatic myeloid cells while preserving CD8+ T-cell function will then provide the translational bridge between these mechanistic insights and precision therapeutic strategies for MASLD.
In summary, this integrative multi-omics framework establishes a transformative paradigm for understanding MASLD pathogenesis by demonstrating how epigenetic regulation, metabolic rewiring, and immune hyperresponsiveness converge through a single molecular hub. The cell type-dependent duality of PLXND1 underscores that therapeutic modulation of such axes must account for cellular context, highlighting the necessity of precision medicine approaches in the MASLD field.

5. Conclusions

This study identifies PLXND1 as a robust, cell-type-specific causal driver of MASLD. By integrating multi-layered MR with human hepatic single-cell transcriptomics and in vivo validation, we trace a continuous causal chain from DNA methylation through gene expression and metabolic remodeling to divergent adaptive and innate immune phenotypes in the steatotic liver. These findings establish PLXND1 as a lineage-dependent molecular switch and provide a mechanistic foundation for precision therapeutic strategies that selectively target hepatic innate immune cells.

Supplementary Materials

The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/metabo16060378/s1, File S1: PLXND1-Mediation Analysis; Table S1: Transcriptome-wide Mendelian randomization results for the causal effects of gene expression on MASLD risk; Table S2: Proteome-wide Mendelian randomization results for the causal effects of circulating protein levels on MASLD risk; Table S3: Mendelian randomization screening of blood metabolites associated with MASLD risk.

Author Contributions

X.M.: Conceptualization, Data curation, Methodology, and Writing—Original Draft. J.S.: Investigation, Validation, Visualization, and Writing—Review & Editing. X.H.: Formal analysis, Data curation, and Writing—Review & Editing. Z.L.: Supervision, Project administration, Writing—Review & Editing, and Funding acquisition. All authors have read and agreed to the published version of the manuscript.

Funding

This research received funding from the National Natural Science Foundation of China (No. 82400667).

Institutional Review Board Statement

The analyses of human data in this study were based exclusively on publicly available, de-identified datasets, including summary-level genetic data and publicly accessible single-cell RNA-sequencing data from the Gene Expression Omnibus (GEO, accession number: GSE136103). Therefore, no additional ethical approval or informed consent was required for the secondary analysis of these public datasets. All animal experiments were conducted in accordance with the institutional guidelines for the care and use of laboratory animals and were approved by the Ethics Committee of Air Force Medical University (Ethics Committee Name: Experimental Animal Ethics Committee of the First Affiliated Hospital of the Air Force Medical University; Approval Code: KY-20254134-1; Approval Date: 16 July 2025). All efforts were made to minimize animal suffering.

Informed Consent Statement

Not applicable.

Data Availability Statement

All data are available in the main text or the Supplementary Materials. The datasets analyzed during the current study are available from the corresponding author upon reasonable request.

Acknowledgments

We extend our sincere appreciation to the FinnGen, eQTLGen, deCODE, GoDMC, GWAS Catalog, GEO, and OneK1K consortia, as well as to all contributors who have made their datasets publicly available.

Conflicts of Interest

The authors declare no conflicts of interest.

Abbreviations

The following abbreviations are used in this manuscript:
CIConfidence interval
DEGDifferentially expressed gene
FCFold change
GEOGene Expression Omnibus
GOGene Ontology
GWASGenome-wide association study
H&EHematoxylin and eosin
HFDHigh-fat diet
IVWInverse-variance weighted
KEGGKyoto Encyclopedia of Genes and Genomes
LAMLipid-associated macrophage
MASLDMetabolic dysfunction-associated steatotic liver disease
mQTLMethylation quantitative trait locus
MRMendelian randomization
NCNormal chow
NKNatural killer
OROdds ratio
pQTLProtein quantitative trait locus
qRT-PCRQuantitative real-time polymerase chain reaction
SMRSummary-data-based Mendelian randomization
sc-eQTLSingle-cell expression quantitative trait locus
scRNA-seqSingle-cell RNA sequencing
UMAPUniform manifold approximation and projection

References

  1. Lazarus, J.V.; Mark, H.E.; Anstee, Q.M.; Arab, J.P.; Batterham, R.L.; Castera, L.; Cortez-Pinto, H.; Crespo, J.; Cusi, K.; Dirac, M.A.; et al. Advancing the global public health agenda for NAFLD: A consensus statement. Nat. Rev. Gastroenterol. Hepatol. 2022, 19, 60–78. [Google Scholar] [CrossRef]
  2. Riazi, K.; Azhari, H.; Charette, J.H.; Underwood, F.E.; King, J.A.; Afshar, E.E.; Swain, M.G.; Congly, S.E.; Kaplan, G.G.; Shaheen, A.A. The prevalence and incidence of NAFLD worldwide: A systematic review and meta-analysis. Lancet Gastroenterol. Hepatol. 2022, 7, 851–861. [Google Scholar] [CrossRef] [PubMed]
  3. Rinella, M.E.; Lazarus, J.V.; Ratziu, V.; Francque, S.M.; Sanyal, A.J.; Kanwal, F.; Romero, D.; Abdelmalek, M.F.; Anstee, Q.M.; Arab, J.P.; et al. A multisociety Delphi consensus statement on new fatty liver disease nomenclature. J. Hepatol. 2023, 79, 1542–1556. [Google Scholar] [CrossRef] [PubMed]
  4. Allen, A.M.; Arab, J.P.; Wong, V.W. MASLD: A disease in flux. Nat. Rev. Gastroenterol. Hepatol. 2024, 21, 747–750. [Google Scholar] [CrossRef] [PubMed]
  5. Miao, L.; Targher, G.; Byrne, C.D.; Cao, Y.Y.; Zheng, M.H. Current status and future trends of the global burden of MASLD. Trends Endocrinol. Metab. 2024, 35, 697–707. [Google Scholar] [CrossRef]
  6. Allen, A.M.; Lazarus, J.V.; Younossi, Z.M. Healthcare and socioeconomic costs of NAFLD: A global framework to navigate the uncertainties. J. Hepatol. 2023, 79, 209–217. [Google Scholar] [CrossRef]
  7. Singal, A.G.; Kanwal, F.; Llovet, J.M. Global trends in hepatocellular carcinoma epidemiology: Implications for screening, prevention and therapy. Nat. Rev. Clin. Oncol. 2023, 20, 864–884. [Google Scholar] [CrossRef]
  8. Llovet, J.M.; Willoughby, C.E.; Singal, A.G.; Greten, T.F.; Heikenwälder, M.; El-Serag, H.B.; Finn, R.S.; Friedman, S.L. Nonalcoholic steatohepatitis-related hepatocellular carcinoma: Pathogenesis and treatment. Nat. Rev. Gastroenterol. Hepatol. 2023, 20, 487–503. [Google Scholar] [CrossRef]
  9. Sookoian, S.; Pirola, C.J.; Sanyal, A.J. MASLD as a non-communicable disease. Nat. Rev. Gastroenterol. Hepatol. 2025, 22, 148–149. [Google Scholar] [CrossRef]
  10. Byrne, C.D.; Armandi, A.; Pellegrinelli, V.; Vidal-Puig, A.; Bugianesi, E. Μetabolic dysfunction-associated steatotic liver disease: A condition of heterogeneous metabolic risk factors, mechanisms and comorbidities requiring holistic treatment. Nat. Rev. Gastroenterol. Hepatol. 2025, 22, 314–328. [Google Scholar] [CrossRef]
  11. Loomba, R.; Friedman, S.L.; Shulman, G.I. Mechanisms and disease consequences of nonalcoholic fatty liver disease. Cell 2021, 184, 2537–2564. [Google Scholar] [CrossRef]
  12. Tilg, H.; Byrne, C.D.; Targher, G. NASH drug treatment development: Challenges and lessons. Lancet Gastroenterol. Hepatol. 2023, 8, 943–954. [Google Scholar] [CrossRef]
  13. Harrison, S.A.; Allen, A.M.; Dubourg, J.; Noureddin, M.; Alkhouri, N. Challenges and opportunities in NASH drug development. Nat. Med. 2023, 29, 562–573. [Google Scholar] [CrossRef]
  14. Harrison, S.A.; Bedossa, P.; Guy, C.D.; Schattenberg, J.M.; Loomba, R.; Taub, R.; Labriola, D.; Moussa, S.E.; Neff, G.W.; Rinella, M.E.; et al. A Phase 3, Randomized, Controlled Trial of Resmetirom in NASH with Liver Fibrosis. N. Engl. J. Med. 2024, 390, 497–509. [Google Scholar] [CrossRef] [PubMed]
  15. Yahoo, N.; Dudek, M.; Knolle, P.; Heikenwälder, M. Role of immune responses in the development of NAFLD-associated liver cancer and prospects for therapeutic modulation. J. Hepatol. 2023, 79, 538–551. [Google Scholar] [CrossRef]
  16. Grander, C.; Grabherr, F.; Tilg, H. Non-alcoholic fatty liver disease: Pathophysiological concepts and treatment options. Cardiovasc. Res. 2023, 119, 1787–1798. [Google Scholar] [CrossRef] [PubMed]
  17. Wu, Y.L.; Lin, Z.J.; Li, C.C.; Lin, X.; Shan, S.K.; Guo, B.; Zheng, M.H.; Li, F.; Yuan, L.Q.; Li, Z.H. Epigenetic regulation in metabolic diseases: Mechanisms and advances in clinical study. Signal Transduct. Target. Ther. 2023, 8, 98. [Google Scholar] [CrossRef] [PubMed]
  18. Sanderson, E.; Glymour, M.M.; Holmes, M.V.; Kang, H.; Morrison, J.; Munafò, M.R.; Palmer, T.; Schooling, C.M.; Wallace, C.; Zhao, Q.; et al. Mendelian randomization. Nat. Rev. Methods Primers 2022, 2, 6. [Google Scholar] [CrossRef]
  19. Skrivankova, V.W.; Richmond, R.C.; Woolf, B.A.R.; Yarmolinsky, J.; Davies, N.M.; Swanson, S.A.; VanderWeele, T.J.; Higgins, J.P.T.; Timpson, N.J.; Dimou, N.; et al. Strengthening the Reporting of Observational Studies in Epidemiology Using Mendelian Randomization: The STROBE-MR Statement. JAMA 2021, 326, 1614–1621. [Google Scholar] [CrossRef]
  20. Ren, Z.; Simons, P.; Wesselius, A.; Stehouwer, C.D.A.; Brouwers, M. Relationship between NAFLD and coronary artery disease: A Mendelian randomization study. Hepatology 2023, 77, 230–238. [Google Scholar] [CrossRef]
  21. Gaziano, L.; Giambartolomei, C.; Pereira, A.C.; Gaulton, A.; Posner, D.C.; Swanson, S.A.; Ho, Y.L.; Iyengar, S.K.; Kosik, N.M.; Vujkovic, M.; et al. Actionable druggable genome-wide Mendelian randomization identifies repurposing opportunities for COVID-19. Nat. Med. 2021, 27, 668–676. [Google Scholar] [CrossRef] [PubMed]
  22. Center, B.S. The GTEx Consortium atlas of genetic regulatory effects across human tissues. Science 2020, 369, 1318–1330. [Google Scholar] [CrossRef] [PubMed]
  23. Rosoff, D.B.; Wagner, J.; Bell, A.S.; Mavromatis, L.A.; Jung, J.; Lohoff, F.W. A multi-omics Mendelian randomization study identifies new therapeutic targets for alcohol use disorder and problem drinking. Nat. Hum. Behav. 2025, 9, 188–207. [Google Scholar] [CrossRef]
  24. Ha, S.; Wong, V.W.; Zhang, X.; Yu, J. Interplay between gut microbiome, host genetic and epigenetic modifications in MASLD and MASLD-related hepatocellular carcinoma. Gut 2024, 74, 141–152. [Google Scholar] [CrossRef]
  25. Gaunt, T.R.; Shihab, H.A.; Hemani, G.; Min, J.L.; Woodward, G.; Lyttleton, O.; Zheng, J.; Duggirala, A.; McArdle, W.L.; Ho, K.; et al. Systematic identification of genetic influences on methylation across the human life course. Genome Biol. 2016, 17, 61. [Google Scholar] [CrossRef] [PubMed]
  26. Sveinbjornsson, G.; Ulfarsson, M.O.; Thorolfsdottir, R.B.; Jonsson, B.A.; Einarsson, E.; Gunnlaugsson, G.; Rognvaldsson, S.; Arnar, D.O.; Baldvinsson, M.; Bjarnason, R.G.; et al. Multiomics study of nonalcoholic fatty liver disease. Nat. Genet. 2022, 54, 1652–1663. [Google Scholar] [CrossRef]
  27. Peiseler, M.; Schwabe, R.; Hampe, J.; Kubes, P.; Heikenwälder, M.; Tacke, F. Immune mechanisms linking metabolic injury to inflammation and fibrosis in fatty liver disease—Novel insights into cellular communication circuits. J. Hepatol. 2022, 77, 1136–1160. [Google Scholar] [CrossRef]
  28. Sadler, M.C.; Auwerx, C.; Deelen, P.; Kutalik, Z. Multi-layered genetic approaches to identify approved drug targets. Cell Genom. 2023, 3, 100341. [Google Scholar] [CrossRef]
  29. Yazar, S.; Alquicira-Hernandez, J.; Wing, K.; Senabouth, A.; Gordon, M.G.; Andersen, S.; Lu, Q.; Rowson, A.; Taylor, T.R.P.; Clarke, L.; et al. Single-cell eQTL mapping identifies cell type-specific genetic control of autoimmune disease. Science 2022, 376, eabf3041. [Google Scholar] [CrossRef]
  30. Mehta, V.; Pang, K.L.; Rozbesky, D.; Nather, K.; Keen, A.; Lachowski, D.; Kong, Y.; Karia, D.; Ameismeier, M.; Huang, J.; et al. The guidance receptor plexin D1 is a mechanosensor in endothelial cells. Nature 2020, 578, 290–295. [Google Scholar] [CrossRef]
  31. Võsa, U.; Claringbould, A.; Westra, H.J.; Bonder, M.J.; Deelen, P.; Zeng, B.; Kirsten, H.; Saha, A.; Kreuzhuber, R.; Yazar, S.; et al. Large-scale cis- and trans-eQTL analyses identify thousands of genetic loci and polygenic scores that regulate blood gene expression. Nat. Genet. 2021, 53, 1300–1310. [Google Scholar] [CrossRef]
  32. Ferkingstad, E.; Sulem, P.; Atlason, B.A.; Sveinbjornsson, G.; Magnusson, M.I.; Styrmisdottir, E.L.; Gunnarsdottir, K.; Helgason, A.; Oddsson, A.; Halldorsson, B.V.; et al. Large-scale integration of the plasma proteome with genetics and disease. Nat. Genet. 2021, 53, 1712–1721. [Google Scholar] [CrossRef]
  33. Ying, H.; Wu, X.; Jia, X.; Yang, Q.; Liu, H.; Zhao, H.; Chen, Z.; Xu, M.; Wang, T.; Li, M.; et al. Single-cell transcriptome-wide Mendelian randomization and colocalization reveals immune-mediated regulatory mechanisms and drug targets for COVID-19. EBioMedicine 2025, 113, 105596. [Google Scholar] [CrossRef]
  34. Min, J.L.; Hemani, G.; Hannon, E.; Dekkers, K.F.; Castillo-Fernandez, J.; Luijk, R.; Carnero-Montoro, E.; Lawson, D.J.; Burrows, K.; Suderman, M.; et al. Genomic and phenotypic insights from an atlas of genetic effects on DNA methylation. Nat. Genet. 2021, 53, 1311–1321. [Google Scholar] [CrossRef]
  35. Xie, J.; Huang, H.; Liu, Z.; Li, Y.; Yu, C.; Xu, L.; Xu, C. The associations between modifiable risk factors and nonalcoholic fatty liver disease: A comprehensive Mendelian randomization study. Hepatology 2023, 77, 949–964. [Google Scholar] [CrossRef]
  36. Ran, S.; Zhang, J.; Tian, F.; Qian, Z.M.; Wei, S.; Wang, Y.; Chen, G.; Zhang, J.; Arnold, L.D.; McMillin, S.E.; et al. Association of metabolic signatures of air pollution with MASLD: Observational and Mendelian randomization study. J. Hepatol. 2025, 82, 560–570. [Google Scholar] [CrossRef]
  37. Eldjarn, G.H.; Ferkingstad, E.; Lund, S.H.; Helgason, H.; Magnusson, O.T.; Gunnarsdottir, K.; Olafsdottir, T.A.; Halldorsson, B.V.; Olason, P.I.; Zink, F.; et al. Large-scale plasma proteomics comparisons through genetics and disease associations. Nature 2023, 622, 348–358. [Google Scholar] [CrossRef]
  38. Wei, S.; He, L.; Zhang, Y.; Li, X.; Zhong, S.; Xiao, L.; Shen, R.; Lu, X.; Shu, Z.; Quan, Y.; et al. Decoding the triglyceride-glucose index in metabolic dysfunction-associated steatotic liver disease: Integrative insights from Mendelian randomization, cross-tissue transcriptomics, and spatial multi-omics. Int. J. Surg. 2026, 112, 94–109. [Google Scholar] [CrossRef] [PubMed]
  39. King, A.; Wu, C. Integrative Multi-Omics Approach for Improving Causal Gene Identification. Genet. Epidemiol. 2025, 49, e22601. [Google Scholar] [CrossRef] [PubMed]
  40. Bourganou, M.V.; Chondrogianni, M.E.; Kyrou, I.; Flessa, C.M.; Chatzigeorgiou, A.; Oikonomou, E.; Lambadiari, V.; Randeva, H.S.; Kassi, E. Unraveling Metabolic Dysfunction-Associated Steatotic Liver Disease Through the Use of Omics Technologies. Int. J. Mol. Sci. 2025, 26, 1589. [Google Scholar] [CrossRef] [PubMed]
  41. Li, Z.; Luo, G.; Gan, C.; Zhang, H.; Li, L.; Zhang, X.; Xing, X.; Hu, S.; Tan, X.; Ding, J.; et al. Spatially resolved multi-omics of human metabolic dysfunction-associated steatotic liver disease. Nat. Genet. 2025, 57, 3112–3125. [Google Scholar] [CrossRef]
  42. Wen, W.; Liu, Z.; Tan, W.; Tan, Y.; Li, W.; Wan, J.; Hu, H.; Jiang, Z.; Tang, X.; Yang, J.; et al. Integrating multi-omics and machine learning systematically deciphers cellular heterogeneity and fibrotic regulatory networks in the progression from MASLD to MASH. NPJ Digit. Med. 2026, 9, 167. [Google Scholar] [CrossRef]
  43. Barchetta, I.; Zampieri, M.; Cimini, F.A.; Dule, S.; Sentinelli, F.; Passarella, G.; Oldani, A.; Karpach, K.; Bacalini, M.G.; Baroni, M.G.; et al. Association Between Active DNA Demethylation and Liver Fibrosis in Individuals with Metabolic-Associated Steatotic Liver Disease (MASLD). Int. J. Mol. Sci. 2025, 26, 1271. [Google Scholar] [CrossRef] [PubMed]
  44. Govaere, O.; Cockell, S.; Tiniakos, D.; Queen, R.; Younes, R.; Vacca, M.; Alexander, L.; Ravaioli, F.; Palmer, J.; Petta, S.; et al. Transcriptomic profiling across the nonalcoholic fatty liver disease spectrum reveals gene signatures for steatohepatitis and fibrosis. Sci. Transl. Med. 2020, 12, eaba4448. [Google Scholar] [CrossRef] [PubMed]
  45. Du, M.; Yuan, H.; Wu, T.; Jiang, Y.; Suo, C.; Jin, L.; Zhang, T.; Liu, Z.; Chen, X. Cross-trait genomic modeling reveals the polygenic architecture and systemic impact of MASLD. Sci. Adv. 2026, 12, eaeb5665. [Google Scholar] [CrossRef]
  46. Vujkovic, M.; Ramdas, S.; Lorenz, K.M.; Guo, X.; Darlay, R.; Cordell, H.J.; He, J.; Gindin, Y.; Chung, C.; Myers, R.P.; 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. 2022, 54, 761–771. [Google Scholar] [CrossRef] [PubMed]
  47. Colella, F.; Henderson, N.C.; Ramachandran, P. Dissecting the mechanisms of MASLD fibrosis in the era of single-cell and spatial omics. J. Clin. Invest. 2025, 135, 18. [Google Scholar] [CrossRef]
  48. Martin, O.P.; Wallace, M.S.; Oetheimer, C.; Patel, H.B.; Butler, M.D.; Wong, L.P.; Huang, P.; Elbaz, J.; Costentin, C.; Salloum, S.; et al. Single-cell atlas of human liver and blood immune cells across fatty liver disease stages reveals distinct signatures linked to liver dysfunction and fibrogenesis. Nat. Immunol. 2025, 26, 1596–1611. [Google Scholar] [CrossRef]
  49. Lee, S.H.T.; Garske, K.M.; Arasu, U.T.; Kar, A.; Miao, Z.; Alvarez, M.; Koka, A.; Darci-Maher, N.; Benhammou, J.N.; Pan, D.Z.; et al. Single nucleus RNA-sequencing integrated into risk variant colocalization discovers 17 cell-type-specific abdominal obesity genes for metabolic dysfunction-associated steatotic liver disease. EBioMedicine 2024, 106, 105232. [Google Scholar] [CrossRef]
  50. Perez, R.K.; Gordon, M.G.; Subramaniam, M.; Kim, M.C.; Hartoularos, G.C.; Targ, S.; Sun, Y.; Ogorodnikov, A.; Bueno, R.; Lu, A.; et al. Single-cell RNA-seq reveals cell type-specific molecular and genetic associations to lupus. Science 2022, 376, eabf1970. [Google Scholar] [CrossRef]
  51. Gu, C.; Yoshida, Y.; Livet, J.; Reimert, D.V.; Mann, F.; Merte, J.; Henderson, C.E.; Jessell, T.M.; Kolodkin, A.L.; Ginty, D.D. Semaphorin 3E and plexin-D1 control vascular pattern independently of neuropilins. Science 2005, 307, 265–268. [Google Scholar] [CrossRef]
  52. Shimizu, I.; Yoshida, Y.; Moriya, J.; Nojima, A.; Uemura, A.; Kobayashi, Y.; Minamino, T. Semaphorin3E-induced inflammation contributes to insulin resistance in dietary obesity. Cell Metab. 2013, 18, 491–504. [Google Scholar] [CrossRef] [PubMed]
  53. Seidman, J.S.; Troutman, T.D.; Sakai, M.; Gola, A.; Spann, N.J.; Bennett, H.; Bruni, C.M.; Ouyang, Z.; Li, R.Z.; Sun, X.; et al. Niche-Specific Reprogramming of Epigenetic Landscapes Drives Myeloid Cell Diversity in Nonalcoholic Steatohepatitis. Immunity 2020, 52, 1057–1074.e1057. [Google Scholar] [CrossRef]
  54. Tran, S.; Baba, I.; Poupel, L.; Dussaud, S.; Moreau, M.; Gélineau, A.; Marcelin, G.; Magréau-Davy, E.; Ouhachi, M.; Lesnik, P.; et al. Impaired Kupffer Cell Self-Renewal Alters the Liver Response to Lipid Overload during Non-alcoholic Steatohepatitis. Immunity 2020, 53, 627–640.e625. [Google Scholar] [CrossRef]
  55. Remmerie, A.; Martens, L.; Thoné, T.; Castoldi, A.; Seurinck, R.; Pavie, B.; Roels, J.; Vanneste, B.; De Prijck, S.; Vanhockerhout, M.; et al. Osteopontin Expression Identifies a Subset of Recruited Macrophages Distinct from Kupffer Cells in the Fatty Liver. Immunity 2020, 53, 641–657.e614. [Google Scholar] [CrossRef] [PubMed]
  56. Choi, Y.I.; Duke-Cohan, J.S.; Ahmed, W.B.; Handley, M.A.; Mann, F.; Epstein, J.A.; Clayton, L.K.; Reinherz, E.L. PlexinD1 glycoprotein controls migration of positively selected thymocytes into the medulla. Immunity 2008, 29, 888–898. [Google Scholar] [CrossRef] [PubMed]
  57. De Angelis Rigotti, F.; Wiedmann, L.; Hubert, M.O.; Vacca, M.; Hasan, S.S.; Moll, I.; Carvajal, S.; Jiménez, W.; Starostecka, M.; Billeter, A.T.; et al. Semaphorin 3C exacerbates liver fibrosis. Hepatology 2023, 78, 1092–1105. [Google Scholar] [CrossRef]
  58. Dudek, M.; Pfister, D.; Donakonda, S.; Filpe, P.; Schneider, A.; Laschinger, M.; Hartmann, D.; Hüser, N.; Meiser, P.; Bayerl, F.; et al. Auto-aggressive CXCR6+ CD8 T cells cause liver immune pathology in NASH. Nature 2021, 592, 444–449. [Google Scholar] [CrossRef]
Metabolites 16 00378 g001
Figure 1. Multi-layered MR identifies PLXND1 as a causal driver of MASLD. (A–C) Venn diagram showing the overlap of genes significantly associated with MASLD at the transcriptomic (eQTL) and proteomic (pQTL) levels. (D) Forest plot of single-cell eQTL-based MR estimates showing the cell-type-specific causal effects of genetically predicted PLXND1 expression on MASLD susceptibility across peripheral blood immune subpopulations. Displayed are odds ratios (ORs) and their 95% confidence intervals (CIs); arrows indicate CIs exceeding the x-axis limits.
Figure 1. Multi-layered MR identifies PLXND1 as a causal driver of MASLD. (A–C) Venn diagram showing the overlap of genes significantly associated with MASLD at the transcriptomic (eQTL) and proteomic (pQTL) levels. (D) Forest plot of single-cell eQTL-based MR estimates showing the cell-type-specific causal effects of genetically predicted PLXND1 expression on MASLD susceptibility across peripheral blood immune subpopulations. Displayed are odds ratios (ORs) and their 95% confidence intervals (CIs); arrows indicate CIs exceeding the x-axis limits.
Metabolites 16 00378 g001
Metabolites 16 00378 g002
Figure 2. Landscape of PLXND1 expression in the human MASLD liver at single-cell resolution. (A) UMAP plot of the GSE136103 dataset showing nine major annotated cell populations. (B) Dot plot of canonical marker genes used for cell-type annotation. (C) Bar plot quantifying the proportion of PLXND1-expressing cells in control versus MASLD samples. (D) Feature plots displaying the spatial distribution of PLXND1 expression across the UMAP in control and MASLD livers. (E) Dot plot summarizing PLXND1 expression levels and the percentage of expressing cells across all identified cell types. (F) UMAP plot of subclustered T-cells colored by CD4+ and CD8+ identity and proportional composition in control and MASLD livers; subsets were annotated using canonical lineage markers (CD4, CD8A, CD8B, and FOXP3).
Figure 2. Landscape of PLXND1 expression in the human MASLD liver at single-cell resolution. (A) UMAP plot of the GSE136103 dataset showing nine major annotated cell populations. (B) Dot plot of canonical marker genes used for cell-type annotation. (C) Bar plot quantifying the proportion of PLXND1-expressing cells in control versus MASLD samples. (D) Feature plots displaying the spatial distribution of PLXND1 expression across the UMAP in control and MASLD livers. (E) Dot plot summarizing PLXND1 expression levels and the percentage of expressing cells across all identified cell types. (F) UMAP plot of subclustered T-cells colored by CD4+ and CD8+ identity and proportional composition in control and MASLD livers; subsets were annotated using canonical lineage markers (CD4, CD8A, CD8B, and FOXP3).
Metabolites 16 00378 g002
Metabolites 16 00378 g003
Figure 3. PLXND1 marks functionally dichotomous immune subsets in human MASLD livers. (A) Volcano plot of differentially expressed genes comparing PLXND1-high Cluster 4 against all other CD8+ T-cell clusters. (B) Gene Ontology (GO) biological process enrichment of upregulated genes in Cluster 4. (C) Volcano plot of differentially expressed genes between PLXND1-positive and PLXND1-negative macrophages. (D) GO biological process enrichment of upregulated genes in PLXND1-positive macrophages. In volcano plots, dashed vertical lines indicate the log2 fold-change cutoff, and the dashed horizontal line indicates the adjusted p-value significance threshold, such as |log2FC| = 0.5 and adjusted p = 0.05.
Figure 3. PLXND1 marks functionally dichotomous immune subsets in human MASLD livers. (A) Volcano plot of differentially expressed genes comparing PLXND1-high Cluster 4 against all other CD8+ T-cell clusters. (B) Gene Ontology (GO) biological process enrichment of upregulated genes in Cluster 4. (C) Volcano plot of differentially expressed genes between PLXND1-positive and PLXND1-negative macrophages. (D) GO biological process enrichment of upregulated genes in PLXND1-positive macrophages. In volcano plots, dashed vertical lines indicate the log2 fold-change cutoff, and the dashed horizontal line indicates the adjusted p-value significance threshold, such as |log2FC| = 0.5 and adjusted p = 0.05.
Metabolites 16 00378 g003
Metabolites 16 00378 g004
Figure 4. Cross-omics mediation links PLXND1 to upstream epigenetic programming and downstream metabolic remodeling. (A) Forest plot depicting the causal effects of DNA methylation quantitative trait loci (mQTLs) at CpG sites cg26767922 and cg08471739 on MASLD risk. (B) Path diagrams illustrating the two-step mediation MR framework for cg26767922 and cg08471739, quantifying the proportion of the total protective effect mediated through PLXND1 downregulation. (C) Directed acyclic graphs showing the mediation of PLXND1 expression on MASLD risk through six distinct circulating blood metabolites.
Figure 4. Cross-omics mediation links PLXND1 to upstream epigenetic programming and downstream metabolic remodeling. (A) Forest plot depicting the causal effects of DNA methylation quantitative trait loci (mQTLs) at CpG sites cg26767922 and cg08471739 on MASLD risk. (B) Path diagrams illustrating the two-step mediation MR framework for cg26767922 and cg08471739, quantifying the proportion of the total protective effect mediated through PLXND1 downregulation. (C) Directed acyclic graphs showing the mediation of PLXND1 expression on MASLD risk through six distinct circulating blood metabolites.
Metabolites 16 00378 g004
Metabolites 16 00378 g005
Figure 5. Hepatic PLXND1 is upregulated in HFD-induced MASLD mice. (A) Representative histological images of liver sections from normal chow (NC) and high-fat diet (HFD) mice, including hematoxylin and eosin (H&E) staining, Oil Red O staining, and Masson’s trichrome staining. (B) Quantitative real-time PCR analysis of hepatic PLXND1 transcript levels. (C) Representative immunoblot (left) and corresponding densitometric quantification (right) of hepatic PLXND1 protein expression normalized to β-Actin. Statistical significance is indicated by asterisks: ** p < 0.01; **** p < 0.0001.
Figure 5. Hepatic PLXND1 is upregulated in HFD-induced MASLD mice. (A) Representative histological images of liver sections from normal chow (NC) and high-fat diet (HFD) mice, including hematoxylin and eosin (H&E) staining, Oil Red O staining, and Masson’s trichrome staining. (B) Quantitative real-time PCR analysis of hepatic PLXND1 transcript levels. (C) Representative immunoblot (left) and corresponding densitometric quantification (right) of hepatic PLXND1 protein expression normalized to β-Actin. Statistical significance is indicated by asterisks: ** p < 0.01; **** p < 0.0001.
Metabolites 16 00378 g005
Table 1. Proteomic Discoveries Derived via Summary-Data-Based Mendelian Randomization in MASLD.
Table 1. Proteomic Discoveries Derived via Summary-Data-Based Mendelian Randomization in MASLD.
GeneExposureOutcometopSNPP_SMRP_HEIDInsnp_HEIDI
ENSG00000004399PLXND1MASLDrs98666530.02560.77377
ENSG00000100577GSTZ1MASLDrs23636420.03160.431520
ENSG00000153574RPIAMASLDrs101850490.04220.66634
ENSG00000135919SERPINE2MASLDrs134125350.07760.622615
ENSG00000104899AMHMASLDrs48072160.11630.791420
ENSG00000197766CFDMASLDrs713352760.24900.277820
ENSG00000113249HAVCR1MASLDrs20334770.46130.713120
Analytical abbreviations employed throughout this dataset comprise the following: SMR, denoting summary-data-based Mendelian randomization methodologies; HEIDI, representing diagnostic evaluations for heterogeneity in dependent instruments; alongside SNP, which signifies individual single-nucleotide polymorphism markers.
Table 2. Mendelian randomization results for key CpG sites significantly associated with PLXND1 expression.
Table 2. Mendelian randomization results for key CpG sites significantly associated with PLXND1 expression.
IDnSNPβsep-ValueEgger_InterceptEgger_Intercept_seEgger_Intercept_pval
cg267679229−0.26730.0537<0.0010.02630.03200.4378
cg084717396−0.21440.0532<0.001−0.01910.02730.5238
Table 3. Causal effects of PLXND1 expression on six blood metabolites.
Table 3. Causal effects of PLXND1 expression on six blood metabolites.
IDReported TraitnSNPβsep-Value
GCST90199817Indoleacetylglutamine levels60.30750.10990.0051
GCST901999082-o-methylascorbic acid levels6−0.20010.09260.0307
GCST90200452Plasma free asparagine levels6−0.27700.09780.0046
GCST90200787Glutamine to asparagine ratio60.27020.09690.0053
GCST90200866Phosphate to asparagine ratio60.23870.09620.0131
GCST90200989Histidine to asparagine ratio60.34740.09820.0004
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Ma, X.; Song, J.; Hong, X.; Lin, Z. Genetically Informed Single-Cell Analysis Reveals PLXND1 as a Cell-Type-Specific Molecular Switch in MASLD. Metabolites 2026, 16, 378. https://doi.org/10.3390/metabo16060378

AMA Style

Ma X, Song J, Hong X, Lin Z. Genetically Informed Single-Cell Analysis Reveals PLXND1 as a Cell-Type-Specific Molecular Switch in MASLD. Metabolites. 2026; 16(6):378. https://doi.org/10.3390/metabo16060378

Chicago/Turabian Style

Ma, Xianyi, Junbo Song, Xin Hong, and Zhibin Lin. 2026. "Genetically Informed Single-Cell Analysis Reveals PLXND1 as a Cell-Type-Specific Molecular Switch in MASLD" Metabolites 16, no. 6: 378. https://doi.org/10.3390/metabo16060378

APA Style

Ma, X., Song, J., Hong, X., & Lin, Z. (2026). Genetically Informed Single-Cell Analysis Reveals PLXND1 as a Cell-Type-Specific Molecular Switch in MASLD. Metabolites, 16(6), 378. https://doi.org/10.3390/metabo16060378

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Article metric data becomes available approximately 24 hours after publication online.
Back to TopTop