Global molecular landscape of early MASLD progression in obesity

  1. Qing Zhao
  2. William De Nardo
  3. Ruoyu Wang
  4. Yi Zhong
  5. Umur Keles
  6. Gabrielé Sakalauskaite
  7. Li Na Zhao
  8. Huiyi Tay
  9. Sonia Youhanna
  10. Mengchao Yan
  11. Ye Xie
  12. Youngrae Kim
  13. Sungdong Lee
  14. Rachel Liyu Lim
  15. Guoshou Teo
  16. Pradeep Narayanaswamy
  17. Paul R. Burton
  18. Volker M. Lauschke
  19. Hyungwon Choi
  20. Matthew J. Watt
  21. Philipp Kaldis

Reviewed by

Read the full article See related articles

Discuss this preprint

Start a discussion What are Sciety discussions?

Listed in

Log in to save this article

Abstract

Metabolic dysfunction-associated steatotic liver disease (MASLD) is often asymptomatic early on but can progress to irreversible conditions like cirrhosis. Due to limited access to human liver biopsies, systematic and integrative molecular resources remain scarce. In this study, we performed transcriptomic analyses on liver and metabolomic analyses on liver and plasma samples from morbidly obese individuals without liver pathology or at early-stage MASLD. While, the plasma metabolomic profile did not fully mirror liver histological features, dual-omics integration of liver samples revealed significantly remodeled lipid and amino acid metabolism pathways. Integrative network analysis uncoupled metabolic remodeling and gene expression as independent features of hepatic steatosis and fibrosis progression, respectively. Notably, GTPases and their regulators emerged as a novel class of genes linked to early liver fibrosis. This study offers a detailed molecular landscape of early MASLD in obesity and highlights potential targets of obesity-linked liver fibrosis.

Article activity feed

  1. Review Commons

    Note: This response was posted by the corresponding author to Review Commons. The content has not been altered except for formatting.

    Learn more at Review Commons

    Reply to the reviewers

    We thank the editor and the reviewers for their positive and constructive comments. Below is our point-by-point responses.

    Reviewer #1 (Evidence, reproducibility and clarity (Required)):

    Metabolic dysfunction-associated steatotic liver disease (MASLD) ranges from simple steatosis, steatohepatitis, fibrosis/cirrhosis, and hepatocellular carcinoma. In the current study, the authors aimed to determine the early molecular signatures differentiating patients with MASLD associated fibrosis from those patients with early MASLD but no symptoms. The authors recruited 109 obese individuals before bariatric surgery. They separated the cohorts as no MASLD (without histological abnormalities) and MASLD. The liver samples were then subjected to transcriptomic and metabolomic analysis. The serum samples were subjected to metabolomic analysis. The authors identified dysregulated lipid metabolism, including glyceride lipids, in the liver samples of MASLD patients compared to the no MASLD ones. Circulating metabolomic changes in lipid profiles slightly correlated with MASLD, possibly due to the no MASLD samples derived from obese patients. Several genes involved in lipid droplet formation were also found elevated in MASLD patients. Besides, elevated levels of amino acids, which are possibly related to collagen synthesis, were observed in MASLD patients. Several antioxidant metabolites were increased in MASLD patients. Furthermore, dysregulated genes involved in mitochondrial function and autophagy were identified in MASLD patients, likely linking oxidative stress to MASLD progression. The authors then determined the representative gene signatures in the development of fibrosis by comparing this cohort with the other two published cohorts. Top enriched pathways in fibrotic patients included GTPase signaling and innate immune responses, suggesting the involvement of GTPase in MASLD progression to fibrosis. The authors then challenged human patient derived 3D spheroid system with a dual PPARa/d agonist and found that this treatment restored the expression levels of GTPase-related genes in MASLD 3D spheroids. In conclusion, the authors suggested the involvement of upregulated GTPase-related genes during fibrosis initiation. Overall, the current study might provide some resources regarding transcriptomic and metabolomic data derived from obese patients with and without MASLD. However, several concerns should be carefully addressed.

    A recent study, via proteomic and transcriptomic analysis, revealed that four proteins (ADAMTSL2, AKR1B10, CFHR4 and TREM2) could be used to identify MASLD patients at risk of steatohepatitis (PMID: 37037945). It is not clear why the authors did not include this study in their comparison. Thank you for the suggestion. The RNA sequencing dataset (GSE135251) from study PMID 37037945 is the same dataset we used as an external benchmark in our study, referred to as the EU cohort on page 4 in the manuscript. In addition to PMID 37037945, we have cited the original transcriptomic study (PMID 33268509) for the EU cohort. In the revised manuscript, we discussed this proteome-transcriptome paper in the Discussion section and highlighted the potential of AKR1B10 as a biomarker in early MASLD.

    The authors recruited 109 patients but only performed transcriptomic and metabolomic analysis in 94 liver samples. Why did the authors exclude other samples?

    We thank the reviewer for their question and we understand the confusion. The discrepancy in sample size between liver and plasma cohorts is due to the fact that, for certain cases, we were unable to get sufficient liver tissue slices (“Exclusion criteria included: age The authors mentioned clinical data in Table 1 but did not present the table in this manuscript.

    Table 1 (key patient characteristics) was included in the main document after the Methods section, and Table S1 (additional patient characteristics) was provided as a supplemental file in our original submission.

    The generated metabolomic data could be a very useful resource to the MASLD community. However, it is very confusing how the data was generated in those supplemental tables. There is no clear labeling of human clinical information in those tables. Also, what do those values mean in columns 47-154? This reviewer assumed that they are the raw data of metabolomic analysis in plasma samples. However, without clear clinical information in these patients, it is impossible that any scientist can use the data to reproduce the authors' findings.

    We appreciate this suggestion. To ensure accessibility of the data resources, we created a GitHub repository for both data and code, available at https://github.com/SLINGhub/MASLD_dual_omics____.

    The GitHub repository includes clinical data for all 109 participants with patient characteristics and histological gradings, as well as processed omics data (log₂-transformed). We have generated artificial IDs for each patient so that we can include all the requested data in an organized manner. A code template is also provided to replicate the main statistical results from this study. In addition, for readers interested in conducting analyses from the raw data, we have deposited the raw sequencing files and mass spectrometry data in GEO and Zenodo, as detailed in the ‘Data Availability’ section.

    In Fig. 5B, the authors excluded the steatosis and fibrosis overlapped genes. Steatosis and fibrosis specific genes could simply reflect the outcomes rather than causes. In this case, the obtained results might not identify the gene signatures related to fibrosis initiation.

    We appreciate this comment, but we do not fully understand the reviewer’s point since we did not exclude overlapped genes in our analysis, and it was unclear to us whether excluding overlapping genes has anything to do with causality of both processes.

    In Figure 5B, we identified the gene signatures associated with steatosis and fibrosis after adjusting for potential confounders such as age, sex, BMI and diabetes status. Our results showed that these signatures were relatively independent, sharing a limited number of genes. We then examined genes uniquely associated with each process by additional adjustment (e.g., adjusting steatosis models for fibrosis grades). To us this was not an unreasonable approach, given that steatosis precedes fibrosis in most cases, especially in morbid obesity.

    We nevertheless agree with the reviewer’s point that the gene expression changes we identified represent statistical associations without warranting causality. To specifically address fibrosis initiation mechanisms within the limitation of the current study design, we performed a separate comparative analysis between patients with fibrosis+steatosis versus those with steatosis alone (Table S11), which still identified GTPase regulation as a potential key mechanism in fibrosis initiation (Figure 6B).

    In Fig. 6D, the authors used 3D liver spheroid to validate their findings. However, there is no images showing the 3D liver spheroid formation before and after PPARa/d agonist treatment. It is not clear whether the 3D liver spheroid was successfully established.

    There is extensive literature (>40 papers) from the Lauschke lab on 3D liver spheroid culture, including but not limited to PMIDs 27143246, 28264975, 32775153, 37870288 and 39605182. Images of the spheroids can be seen in Figure 1c of Adv. Sci. 2024, 2407572 and elafibrinor treatment did not affect the morphology of the spheroids.

    The authors suggested that targeting LX-2 cells with Rac1 and Cdc42 inhibitors could reduce collagen production. Did the authors observe these two genes upregulated in mRNA and protein expression levels in their cohort when compared MASLD patients with and without fibrosis? Did the authors observe that the expression levels of Rac1 and Cdc42 are correlated with fibrosis progression in MASLD patients?

    Regarding comments 7 and 8, we targeted Rac1 and Cdc42 in the LX-2 cell experiment as they are common and major GTPases. Protein-level data are not available in our dataset, but we examined their transcript-level expression. RAC1 and CDC42 expression levels were positively associated with fibrosis progression, with coefficients of 0.362 (q = 0.027) and 0.342 (q = 0.031), respectively. These results are presented in Table S5, and the corresponding boxplots are shown here.

    Figure R1. RAC1 and CDC42 expression levels in individuals with different fibrosis *levels. *

    Other studies have revealed several metabolite changes related to MASLD progression (PMID: 35434590, PMID: 22364559). However, the authors did not discuss the discrepancies between their findings with the previous studies.

    Thank you for the suggestion. We have incorporated a discussion of the two studies into the Discussion section, highlighting the consistencies and discrepancies between our plasma metabolomic results and previous findings. The main differences may stem from variations in MASLD spectrum and the degree of obesity in the cohorts.

    Reviewer #1 (Significance (Required)):

    Overall, the current study might provide some new resources regarding transcriptomic and metabolomic data derived from obese patients with and without MASLD. The MASLD research community will be interested in the resource data.

    We thank this reviewer for the positive and constructive evaluation of our manuscript.

    Reviewer #2 (Evidence, reproducibility and clarity (Required)):

    Summary:

    In this paper, Kaldis and collaborators investigate the molecular heterogeneity of a 109 morbidly obese patient cohort, focusing on liver transcriptomics and metabolomics analysis from liver and serum. The main finding (i.e. upregulation of GTPase-coding genes) was validated in spheroids and a human HSC cell line. As these proteins are involved in critical cellular functions related to metabolism and cytoskeleton dynamics, these findings shed light on their involvement in human liver pathology which so far has been poorly (or even not) documented to date. This is an interesting addition to the current knowledge about chronic liver pathology. However the manuscript suffers from the lack of a clear-cut definition of patient subgroups and the seemingly indistinct use of generic (MASLD, NAS score) and more granular terms (MASH, fibrosis) across the various analysis they performed.

    We thank this reviewer of highlighting the novelty of our manuscript. We agree that mixing generic and granular terms can be confusing and we tried to use of terms consistently throughout, which has been further improved in the revised version.

    Figure 1 and Table 1 provide comprehensive information regarding histological phenotypes, NAS scores, and patient characteristics. From Figure 2 onward, we specifically focused on steatosis and fibrosis as distinct histological features, identifying molecular signatures associated with each process.

    The term ‘MASH’ was used only when referring to the ex vivo 3D spheroids derived from histologically confirmed MASH patients for validation purposes. As our primary cohort represents early disease stages, we did not characterize molecular features of MASH in that data set.

    In this cohort, the term 'NAS' was mentioned only in Section 1 to characterize the disease spectrum. Additionally, in Figures 3A and 6A, we illustrated the association between gene expression levels and NAS in two external cohorts. This was due to the absence of steatosis grades in the two datasets. NAS is an additive measure of multiple scores (steatosis, inflammation and ballooning), but does not account for fibrosis grades.

    Our study focuses on the molecular features of steatosis grades and fibrosis grades as the main histological processes, with all terminology aligned with this stated objective. This allows us to map the transcriptome and metabolome to pathologist-defined steatosis/fibrosis severity (i.e., 0,1,2,3) and identify genes/metabolites that are correlated with increasing steatosis/fibrosis score.

    Major comments:

    • Are the key conclusions convincing?

    The conclusions are generally consistent with findings from numerous previous studies, as many of the genes identified and their associations with disease states have been previously reported. However, I found it difficult to discern which specific disease stages the authors are referring to throughout the manuscript. Terms such as MASLD (Fig. 1F), steatosis (Fig. 4A), MASH, fibrosis (Fig. 6), and the composite NAS score (Fig. 1G) are used interchangeably, without clearly explaining whether or how the patient cohort was stratified to distinguish between isolated steatosis, MASH, and MASH with or without fibrosis. It is also unclear whether subgroups were propensity score-matched.

    As explained in our previous point, we believe that we did not carelessly use the terms interchangeably, but rather used them as they were available or pertinent to the comparisons in discussion. We have provided a comprehensive cohort description in the first section (Table 1, including all histological features and NAS scores), then focused specifically on steatosis and fibrosis in subsequent analyses. We identified distinct molecular processes underlying these two histological features and validated key fibrosis-related pathways.

    Regarding the comment of ‘propensity score-matched subgroups’, we would like to clarify that the only “sub”-group analysis performed in this paper is the transition from steatosis to steatosis with fibrosis. We have consistently used linear regression as the association analysis framework, without binarization of outcomes. We recall that this is a cross-sectional study with challenging recruitment situation from a bariatric surgery clinic that naturally represents the spectrum of MASLD in obesity. We acknowledge that the sampling can always be biased in such a study. However, given the invasiveness of liver resection, the study is also limited by the reality that not all patients would agree to the study, nor it is feasible to form a perfect subgroup meeting 1:1 ratio as in large-scale epidemiology studies based on plasma samples.

    In a related point, the authors mention that 76% of patients are non-fibrotic, introducing a marked imbalance between fibrotic (n=26) and non-fibrotic (n=83) samples. Given this disparity and potential inter-individual variability, it would be helpful to include observed fold changes or effect sizes to give readers a sense of the magnitude of the biological dysregulations being reported.

    As explained in our previous response, our study design examines associations between histological and molecular features rather than using a case-control approach. For effect size quantification, we report standardized linear regression coefficients, i.e. the change in gene expression Z-score per one-point increase in steatosis or fibrosis grade. We also provided fold changes in our comparative analysis of steatosis+fibrosis versus fibrosis-free steatosis. These effect sizes were fully documented in the Supplemental Tables.

    • Should the authors qualify some of their claims as preliminary or speculative, or remove them altogether?

    • The authors seem pretty enthusiastic about elafibranor, despite a failed phase 3 clinical trial. I would qualify elafibranor as a useful tool in preclinical model. We agree with the reviewer and indeed used elafibranor as a research tool for PPARa/d modulation rather than a clinically promising prospect. Discussion regarding elafibranor has been updated.

    • The authors should make clearly the pronounced sex bias in their study, which includes mostly women (and btw refer to sex and not gender in the manuscript). Thank you for this important point. We added "Notably, the cohort was predominantly female (76.1%)" to the 'Overview of the study' section in the manuscript. We also replaced all 'gender' with 'sex' throughout the manuscript. In this cohorts, individuals with previous gender reassignment were excluded (see Materials and Methods).

    • The "MASH" status of the spheroid model is overstated. As described in the text it is much closer to a lipotoxicity model (and even glucotoxicity as Glc concentration is 2g/L). The 3D cultures were established from cells isolated from patients with histologically confirmed MASH. Besides steatosis, we observe increased secretion of pro-inflammatory cytokines, activation of hepatic stellate cells and increased deposition of collagen, thus phenocopying the critical disease hallmarks. Additionally, unbiased omics profiling (transcriptomics, proteomics and lipidomics) reveals significant increases in collagen biosynthesis, inflammatory signaling and cholesterol biosynthesis in MASH patient-derived cultures compared to controls. These differences largely overlapped with the results from analyses of six MASH case-control cohort studies. All of these results have been published previously (PMID 39605182).

    This is confusing with panel D in which the authors establish a relationship between fibrotic patients (F2/F3 vs F0/S0, so I guess "no MASLD liver?) and this model. Is the relationship maintained for steatotic-only patients?

    In Figure 6D, we compared GTPase-related gene expression between patients with fibrosis grade 2/3 (n = 26) and those without fibrosis and steatosis (n = 24). Principal component regression resulted in a positive correlation (β = 9.97) between log2 fold changes in 3D spheroids and human fibrosis samples, indicating consistent directional changes in both systems.

    To answer the question from the reviewer, we compared the expression levels of GTPase-related genes in patients with steatosis but no fibrosis (n = 18) to those without fibrosis and steatosis (n = 24), we observed a negative correlation (β = -10.91). This indicates that GTPase-related gene changes in our 3D spheroids do not align with steatosis-related changes in humans.

    Therefore, under the assumption that fibrosis follows steatosis in the majority of the cases of MASLD progression, the result indicates that the alterations in GTPase-related gene expression in the 3D spheroid model specifically is reflective of fibrosis rather than steatosis.

    Figure R2. Comparison of expression level changes in GTPase-related genes between this human cohort and an independent 3D spheroid system: (A) positive correlation with fibrosis grade 2/3 patients versus controls (left), and (B) negative correlation with steatosis-only patients versus controls (right).

    • Would additional experiments be essential to support the claims of the paper? Request additional experiments only where necessary for the paper as it is, and do not ask authors to open new lines of experimentation.

    I am not convinced that HSC and LX2 cells express significant levels of PPARα. However, did the authors check for this parameter in their LX2 cell line and assessed whether PPARα/b activation by elafibranor (and/or pemafibrate as it is PPARα selective) alter GTPase expression? Whether negative or positive, this could give a clue about possible intercellular crosstalk in the spheroid model.

    We thank this reviewer to point this out. In response, we analysed the mRNA expression of all PPARs in LX-2 cells with and without Elafibranor treatment, respectively (see Figure R3, same as Figure S8G in the Supplemental Material). We confirmed PPARs are expressed in LX-2 cells at the mRNA level (Figure R3A). Elafibranor does not affect their mRNA levels, which is consistent with previous reports that its primary mechanism is through binding and altering the activity of PPAR proteins, not gene expression (PMID 33326461 and PMID 37627519).

    *Figure R3. Gene signatures in LX-2 cells with and without Elafibranor treatment (n = 3). *

    In addition, we assessed mRNA levels of selected GTPase-related genes in LX-2 cells with and without Elafibranor treatment (Figure R3B). Although statistical power was limited, we observed a consistent trend toward reduced RHOU, DOCK2, and RAC1 expression with Elafibranor. this preliminary signal suggests that Elafibranor may counter the elevated GTPase levels seen in MASH patient spheroids, potentially via crosstalk among hepatic cell types, including HSCs.

    To further investigate intercellular crosstalk in GTPase regulation among hepatic cell types, we evaluated signature GTPase-related genes in LX-2 cells, spheroid co-cultures (hepatocytes, HSCs, Kupffer cells), and hepatocyte monocultures. As shown in __Figure R4 __(same as Figure S10 in the supplemental material), TGFB1 served as a positive control, exhibiting the most pronounced induction upon TGF-β1 treatment in hepatocytes. Despite varied alterations across the selected GTPase-related genes, TGF-β1 treatment produced a trend toward increased VAV1 and DOCK2 expression in co-culture, hepatocytes, and LX-2 cells, and this was reversed by the TGF-β inhibitor in co-culture and hepatocytes. Other GTPase genes, including RAC1, RAB32, and RHOU, displayed cell type–specific responses to TGF-β1. These observations suggest that the regulation of GTPases is mediated by multiple hepatic cell types, supporting the importance of intercellular crosstalk.

    Figure R4. Expression of GTPase-related genes in spheroid co-culture, hepatocyte monoculture, and LX-2 cells (n = 3). Controls for each gene and experiment were normalized to 1 to enable comparison across treatment groups.

    • Are the suggested experiments realistic in terms of time and resources? It would help if you could add an estimated cost and time investment for substantial experiments.

    The experiment mentioned above is cheap (cell culture, RT-QPCR) and can be performed within a couple of weeks.

    • Are the data and the methods presented in such a way that they can be reproduced? Yes

    • Are the experiments adequately replicated and statistical analysis adequate? There is no indication of group size, number of replicates for in vitro experiments

    Thank you for this suggestion. We have added the sample sizes to all relevant sections: ‘n = 4’ in the figure legends for 3D spheroid experiments and ‘n = 8–10’ for the LX-2 experiments. This information has also been incorporated into the corresponding experimental descriptions in the Methods section.

    **Referees cross-commenting**

    I believe there is a general consensus on this potentially interesting contribution to the field, with three main points: (1) the need for a careful group-by-group comparison that accounts for potential confounders, (2) a more rigorous exploitation/characterization of the spheroid system, and (3) the need to benchmark the authors' findings against the available literature.

    Thank you for summarizing the main points. Our responses are as follows:

    • We adjusted for key confounders (sex, gender, age, BMI, diabetes) in all statistical analysis to minimize potential bias, mostly using linear regression (rather than group-to-group comparison). In response to Reviewer 3, comment 1, we also conducted additional statistical analyses exploring molecular changes in diabetic vs. non-diabetic individuals.
    • We provided detailed characterization of the spheroid model (response to Reviewer 3, comment 3) and we have done additional experiments in LX-2 cells.
    • We benchmarked our findings using external human cohorts, mouse models, and single cell spheroid systems:
    • Compared our liver transcriptomics data with two published liver RNA-seq datasets (EU cohort, PMID 31467298; VA cohort, PMID 33268509) as shown in Figure 1G. In Figures 3A and 6A, we also included sidebars indicating gene alterations in these cohorts, showing consistent trends. Moreover, we examined the expression alterations of GTPase-related genes in these datasets in response to Reviewer 3’s comment 2.
    • Assessed genes linked to fibrosis progression in hepatic stellate cells from a murine liver fibrosis model (PMID 34839349), confirming differential expression of GTPases and their regulators during fibrosis initiation (Figure S9A).
    • Examined GTPase-related genes in an independent single-cell human spheroid system (PMID 37962490). This enabled cell-type-specific information of GTPase regulation in response to TGF-β (Figure S9C). We also expanded the discussion section on both the consistencies and discrepancies between our findings and previously published studies.

    Reviewer #2 (Significance (Required)):

    The authors identified GTPases as players in the progression of MASLD. This is an interesting preliminary report warranting further molecular investigations (in which liver cell types, which GTPase pathway(s) are involved, which functions are controlled through this pathway...)

    • State what audience might be interested in and influenced by the reported findings.

    This paper will have an impact in the hepatology field

    • Define your field of expertise with a few keywords to help the authors contextualize your point of view. Indicate if there are any parts of the paper that you do not have sufficient expertise to evaluate.

    I have expertise in the analysis of "MASLD" human cohorts and in the molecular biology of chronic liver diseases.

    Reviewer #3 (Evidence, reproducibility and clarity (Required)):

    Summary:

    Metabolic dysfunction associated liver disease (MASLD) describes a spectrum of progressive liver pathologies linked to life style-associated metabolic alterations (such as increased body weight and elevated blood sugar levels), reaching from steatosis over steatohepatitis to fibrosis and finally end stage complications, such as liver failure and hepatocellular carcinoma. Treatment options for MASLD include diet adjustments, weight loss, and the receptor-β (THR-β) agonist resmetirom, but remain limited at this stage, motivating further studies to elucidate molecular disease mechanisms to identify novel therapeutic targets. In their present study, the authors aim to identify early molecular changes in MASLD linked to obesity. To this end, they study a cohort of 109 obese individuals with no or early-stage MASLD combining measurements from two anatomic sides: 1. bulk RNA-sequencing and metabolomics of liver biopsies, and 2. metabolomics from patient blood. Their major finding is that GTPase-related genes are transcriptionally altered in livers of individuals with steatosis with fibrosis compared to steatosis without fibrosis.

    Major comments:

    1. Confounders (such as (pre-)diabetes) The patient table shows significant differences in non-MASLD vs. MASLD individuals, with the latter suffering more often from diabetes or hypertriglyceridemia.

    Rather than just stating corrections, subgroup analyses should be performed (accompanied with designated statistical power analyses) to infer the degree to which these conditions contribute to the observations. I.e., major findings stating MASLD-associated changes should hold true in the subgroup of MASLD patients without diabetes/of female sex and so forth (testing for each of the significant differences between groups).

    Our original statistical analysis employed linear regression to examine associations between molecular variables (genes/metabolites) and histological progression (steatosis and fibrosis), with adjustment for potential confounders including diabetic status, age, sex, and BMI. We specifically focused on these two histological features to elucidate the disturbed molecular processes during their progression. Regression coefficients represent the expected change in abundance levels (in units of standard deviation of the corresponding molecule) per one-unit increase in histological grades.

    To address the reviewer's question, we conducted additional subgroup analyses to determine whether our major findings remain consistent in individuals with and without diabetes. We assessed linear associations between gene signatures and histological features separately in non-diabetic (n = 71) and diabetic individuals (n = 23). Statistical power was estimated by comparing the variance explained by the full regression model (y ~ x + a + b + c) against the reduced model (y ~ a + b + c), converting the incremental R² for x into Cohen's f², and applying pwr.f2.test with the corresponding degrees of freedom and sample size at α = 0.05.

    For both steatosis and fibrosis, the results in the non-diabetic subgroup (n = 71) showed high consistency with findings in our original analysis (n = 94, adjusted for diabetes), indicating that our originally reported gene signatures, after correction for diabetic status, remain valid in non-diabetic individuals.

    In contrast, for diabetic individuals (n = 23), associations between genes and histological features did not closely replicate our original findings. Notably, we observed larger estimate effects for fibrosis-associated genes in diabetic individuals, suggesting a potential interaction between diabetes and fibrosis progression.

    Figure R5. Subgroup analysis of the association between gene expressions and steatosis grades

    *Figure R6. *Subgroup analysis of the association between gene expressions and fibrosis grades

    On the comment "degree to which these conditions contribute to the observations," our original analysis adjusted for diabetes status to identify molecular signatures independently associated with fibrosis without the confounding of diabetes status. Consequently, the reported gene signatures in the original analysis more closely reflect patterns in the non-diabetes group, as demonstrated in our subgroup analysis plots. We also comment that, unfortunately, we did not adjust for the interaction of fibrosis and diabetes in the original analysis.

    Furthermore, our additional analyses revealed a close relationship between diabetes and liver fibrosis. Consistent with Figure 1C, hepatic fibrosis is significantly correlated with insulin resistance parameters in clinical assays, including blood insulin levels and HOMA2-IR. To explore this association further, we compared gene expression profiles between diabetic MASLD patients (n = 21) and non-diabetic MASLD patients (n = 43). Although few genes reached significance after multiple testing correction, 166 genes showed differential expression (p 0.32) between these groups.

    We identified 55 genes as potential "diabetic markers" that both showed differential expression between diabetic and non-diabetic MASLD patients and were significantly associated with steatosis or fibrosis progression. These genes are predominantly downregulated metabolic genes (e.g., BAAT, G6PC1, SULT2A1, MAT1A), suggesting that diabetes may exacerbate metabolic suppression as fibrosis advances. Given the high prevalence of diabetes in the MASLD population, our analysis supports the hypothesis that diabetes worsens MASLD outcomes, likely through impaired metabolic capability during fibrosis progression.

    Regarding the comment on the "subgroup of female sex," our original analysis also adjusted for sex as a potential confounder. Since our cohort is predominantly female (>76%), the majority of our findings likely holds true in the female sub-population, similar to what we observed in our diabetes subgroup analysis.

    External validation

    Additionally, to back up the major GTPase signature findings, it would be desirable to analyze an external dataset of (pre)diabetes patients (other biased groups) for alternations in these genes. It would be important to know if this signature also shows in non-MASLD diabetic patients vs. healthy patients or is a feature specific to MASLD. Also, could the matched metabolic data be used to validate metabolite alterations that would be expected under GTPase-associated protein dysregulation?

    We appreciate the comments regarding the validation GTPase as a unique MASLD signature by external datasets. As shown in our previous analysis, after adjusting for diabetes status, the gene signatures remained largely preserved in the non-diabetes subgroup. Before we respond further, we also preface that publicly available liver tissue data, with appropriate and full-scale clinical metadata and sufficient sample sizes, are extremely rare. To the best of our knowledge, the public data sets we brought into our paper were the most prominent data of reliable quality.

    In the paper, we benchmarked our RNAseq dataset against two datasets: the VA cohort and EU cohort (Figure 1). Our cohort focused primarily on early MASLD patients with obesity, which aligns more closely with the disease spectrum represented in the VA cohort (Figure 1G). Notably, in the published paper for the VA cohort, Hoang et al. highlighted Rho GTPase signaling as one of the top pathways in the fibrosis PPI network (Figure 1B from publication PMID 31467298).

    We interrogated GTPase-related genes in both the VA and EU cohorts. As shown in Figure R7 (below), GTPase-related genes demonstrated a strong association with fibrosis grades in the VA cohort, as expected. The EU cohort comprises more advanced MASLD cases with higher fibrosis grades, and our re-analysis in this cohort specifically focused on MASH patients (as designated by the authors). In those MASH patients, GTPase-related genes did not show significant positive associations with fibrosis progression. This finding is consistent with our hypothesis that GTPase regulation is triggered more prominent during the early progression of fibrosis rather than at later stages.

    Unfortunately, diabetes status was not available in the GEO repository for the VA cohort. Available liver tissue sequencing datasets with balanced representation of diabetic and nondiabetic patients are rare, especially those derived from obese individuals and reflecting the early-to-middle stages of MASLD. In our own cohort, for instance, only two diabetic patients without MASLD were recruited (Table 1). While we cannot rule out a role for insulin resistance in GTPase regulation, we will plan future experiments using mouse models to examine GTPase-mediated fibrosis under diabetic and nondiabetic conditions.

    Regarding the comment ‘validate metabolite alterations that would be expected under GTPase-associated protein dysregulation,’ we note that GTPases are primarily involved in cytoskeletal organization, vesicle trafficking, and other cellular processes, with few well-established links to specific metabolite signatures. Nevertheless, in our partial correlation network integrating hepatic genes and metabolites, we observed co-regulated metabolites associated with GTPase-related genes (Figure R8). These included palmitoleoyl ethanolamide (N-acylethanolamine, an anti-inflammatory metabolite and PPARα ligand), phenylacetic acid (a phenylalanine metabolite), biotin (a coenzyme), arginine, lysine, melatonin (a tryptophan metabolite), and several lipid species such as PC 32:0 and CAR 20:1. While causal relationships cannot be inferred from this dataset, our integrative network highlights potential connections related to the trafficking of these metabolites that warrant further investigation.

    *Figure R7. Associations between GTPase-related genes to fibrosis in this study and two external cohorts. Asterisks denote significant associations with q value Figure R8. Integrative subnetwork of GTPase-related genes. Blue squares represent GTPase-related genes, red circles indicate metabolites connected to these genes, and the purple diamond denotes fibrosis, which is connected to RHOU.

    3D liver spheroid MASH model, Fig. 6D/E

    This 3D experiment is technically not an external validation of GTPase-related genes being involved in MASLD, since patient-derived cells may only retain changes that have happened in vivo. To demonstrate that the GTPase expression signature is specifically invoked by fibrosis the LX-2 set up is more convincing, however, the up-regulation of the GTPase-related genes upon fibrosis induction with TGF-beta, in concordance with the patient data, needs to be shown first (qPCR or RNA-seq).

    We agree with the reviewer that experiments in LX-2 (HSC) cells are important and as we have described under ‘Reviewer #2’ we have done this (Figure R3 and Figure R4). Because HSCs only comprise a minor cell population of liver cells, the signals observed in patient bulk RNA data are likely driven primarily by hepatocytes. Nevertheless, we have highlighted the importance of hepatic cell crosstalk in Figure R4 and in our response to Reviewer #2. Additionally, in Supplementary Figure S9B, we identify the potential cell types of origin for the GTPase signals (predominantly hepatocytes and HSCs) using a single-cell dataset from an independent study (PMID: 37962490).

    Additionally, the description of the 3D model is too uncritical. The maintenance of functional human PHHs in 3D has only become available this year (PMID: 40240606) marking a break-through in the field. Since the authors did not use this system, I would strongly assume their findings are largely attributable to the mesenchymal cells in the 3D culture, and these limitations need to be stated.

    We humbly disagree with the reviewer on the 3D liver spheroids. The paper that the reviewer is referencing is related to the proliferation of hepatocytes in organoids, not – at least not directly – their functional maintenance. Here, we use a spheroid model of mature fully differentiated cells, which is conceptually different from the organoid approach. Maintenance of such functional human hepatocytes for multiple weeks in culture has been possible for close to a decade (PMID 27143246). Moreover, particularly for the modeling of chronic liver disease, such as MASH, it is important to use directly patient-derived cells as short induction cycles (typically 1-2 weeks) of disease phenotypes in organoid models do not faithfully reproduce the molecular signatures that stem from chronic exposures in vivo.

    The 3D liver spheroid model we used here is derived from livers from patients with a histologically confirmed diagnosis of MASH. The isolated cells are fully mature and thus do not require in vitro differentiation. There are no MSCs in the 3D cultures; rather the spheroids contain hepatocytes, stellate cells, Kupffer cells as well as various other immune cell types present in the liver at the time of isolation (T cells, B cells, NK cells). Furthermore, the model is extensively characterized at the transcriptomic, proteomic and lipidomic level (PMID 39605182).

    Novelty / references

    Similar studies that also combined liver and blood lipidomics/metabolomics in obese individuals with and without MASLD (e.g. PMID 39731853, 39653777) should be cited. Additionally, it would benefit the quality of the discussion to state how findings in this study add new insights over previous studies, if their findings/insights differ, and if so, why.

    Thank you for the suggestion. We added the two papers into the discussion section. Specifically, we discussed the consistent findings (such as AKR1B10 in PMID 37037945 and mitochondrial dysfunction in PMID 39731853) and discrepancies (such as limited plasma metabolomic changes and circulating sphingolipid alterations in multiple human and mouse models) in comparison with previously published omics studies in MASLD patients. Also, we thoroughly discussed our findings (e.g., lipid dysregulation, dysregulated tryptophan metabolism, GTPase regulation) and potential mechanisms with extensive literature supports from of human, animal, and cell studies.

    Minor comments:

    1. The quality of Supplementary Figures (e.g. S7) makes is impossible to read the labels Thank you for this feedback. The resolution of the figures was impaired in the initial upload. We will provide all supplementary figures with high resolution in our revised submission and ensure all labels are clearly readable.

    For Figure S7C, we presented the correlation matrix of more than 200 GTPase-related genes along with the TGF-β genes TGFB1 and TGFB3. This illustrates the overall co-expression patterns of GTPase-related genes rather than displaying individual gene labels, with arrows now included to highlight TGFB1 and TGFB3.

    Reviewer #3 (Significance (Required)):

    The authors provide an overall sound study on the hepatic transcriptomic and metabolomic signatures in an Australian cohort of 109 obese non-to-early stage MASLD patients. They perform thorough analyses of metabolome and transcriptome in liver biopsies and metabolome in blood, using standard technologies such as RNA sequencing and mass spectrometry. Their key finding is a GTPase-associated gene signature related to fibrosis onset. Limitations of the study include potential cohort confounders (raising the need for expanded control experiments), limited discussion of similar studies, and limits in cell-type resolution, the latter of which is related to the molecular read out, and has in parts been started to be addressed by in vitro experiments in an immortalized HSC lines. Taken together, given additional control analyses will be performed, the results could be of interest to an expert community in the field of molecular hepatology and, while still descriptive, hold the potential to prompt mechanistic follow-up studies.

    We thank this reviewer for a balanced, positive, and constructive evaluation of our manuscript.

  2. Review Commons

    Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Referee #3

    Evidence, reproducibility and clarity

    Summary:

    Metabolic dysfunction associated liver disease (MASLD) describes a spectrum of progressive liver pathologies linked to life style-associated metabolic alterations (such as increased body weight and elevated blood sugar levels), reaching from steatosis over steatohepatitis to fibrosis and finally end stage complications, such as liver failure and hepatocellular carcinoma. Treatment options for MASLD include diet adjustments, weight loss, and the receptor-β (THR-β) agonist resmetirom, but remain limited at this stage, motivating further studies to elucidate molecular disease mechanisms to identify novel therapeutic targets.

    In their present study, the authors aim to identify early molecular changes in MASLD linked to obesity. To this end, they study a cohort of 109 obese individuals with no or early-stage MASLD combining measurements from two anatomic sides: 1. bulk RNA-sequencing and metabolomics of liver biopsies, and 2. metabolomics from patient blood. Their major finding is that GTPase-related genes are transcriptionally altered in livers of individuals with steatosis with fibrosis compared to steatosis without fibrosis.

    Major comments:

    1. Confounders (such as (pre-)diabetes) The patient table shows significant differences in non-MASLD vs. MASLD individuals, with the latter suffering more often from diabetes or hypertriglyceridemia. Rather than just stating corrections, subgroup analyses should be performed (accompanied with designated statistical power analyses) to infer the degree to which these conditions contribute to the observations. I.e., major findings stating MASLD-associated changes should hold true in the subgroup of MASLD patients without diabetes/of female sex and so forth (testing for each of the significant differences between groups).
    2. External validation Additionally, to back up the major GTPase signature findings, it would be desirable to analyze an external dataset of (pre)diabetes patients (other biased groups) for alternations in these genes. It would be important to know if this signature also shows in non-MASLD diabetic patients vs. healthy patients or is a feature specific to MASLD. Also, could the matched metabolic data be used to validate metabolite alterations that would be expected under GTPase-associated protein dysregulation?
    3. 3D liver spheroid MASH model, Fig. 6D/E This 3D experiment is technically not an external validation of GTPase-related genes being involved in MASLD, since patient-derived cells may only retain changes that have happened in vivo. To demonstrate that the GTPase expression signature is specifically invoked by fibrosis the LX-2 set up is more convincing, however, the up-regulation of the GTPase-related genes upon fibrosis induction with TGF-beta, in concordance with the patient data, needs to be shown first (qPCR or RNA-seq). Additionally, the description of the 3D model is too uncritical. The maintenance of functional human PHHs in 3D has only become available this year (PMID: 40240606) marking a break-through in the field. Since the authors did not use this system, I would strongly assume their findings are largely attributable to the mesenchymal cells in the 3D culture, and these limitations need to be stated.
    4. Novelty / references Similar studies that also combined liver and blood lipidomics/metabolomics in obese individuals with and without MASLD (e.g. PMID 39731853, 39653777) should becited. Additionally, it would benefit the quality of the discussion to state how findings in this study add new insights over previous studies, if their findings/insights differ, and if so, why.

    Minor comments:

    1. The quality of Supplementary Figures (e.g. S7) makes is impossible to read the labels

    Significance

    The authors provide an overall sound study on the hepatic transcriptomic and metabolomic signatures in an Australian cohort of 109 obese non-to-early stage MASLD patients. They perform thorough analyses of metabolome and transcriptome in liver biopsies and metabolome in blood, using standard technologies such as RNA-sequencing and mass spectrometry. Their key finding is a GTPase-associated gene signature related to fibrosis onset. Limitations of the study include potential cohort confounders (raising the need for expanded control experiments), limited discussion of similar studies, and limits in cell-type resolution, the latter of which is related to the molecular read out, and has in parts been started to be addressed by in vitro experiments in an immortalized HSC lines. Taken together, given additional control analyses will be performed, the results could be of interest to an expert community in the field of molecular hepatology and, while still descriptive, hold the potential to prompt mechanistic follow-up studies.

  3. Review Commons

    Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Referee #2

    Evidence, reproducibility and clarity

    Summary:

    Provide a short summary of the findings and key conclusions (including methodology and model system(s) where appropriate).

    In this paper, Kaldis and collaborators investigate the molecular heterogeneity of a 109 morbidly obese patient cohort, focusing on liver transcriptomics and metabolomics analysis from liver and serum. The main finding (ie upregulation of GTPase-coding genes) was validated in spheroids and a human HSC cell line. As these proteins are involved in critical cellular functions related to metabolism and cytoskeleton dynamics, these findings shed light on their involvement in human liver pathology which so far has been poorly (or even not) documented to date. This is an interesting addition to the current knowledge about chronic liver pathology. However the manuscript suffers from the lack of a clear-cut definition of patient subgroups and the seemingly indistinct use of generic (MASLD, NAS score) and more granular terms (MASH, fibrosis) across the various analysis they performed.

    Major comments:

    • Are the key conclusions convincing?

    The conclusions are generally consistent with findings from numerous previous studies, as many of the genes identified and their associations with disease states have been previously reported. However, I found it difficult to discern which specific disease stages the authors are referring to throughout the manuscript. Terms such as MASLD (Fig. 1F), steatosis (Fig. 4A), MASH, fibrosis (Fig. 6), and the composite NAS score (Fig. 1G) are used interchangeably, without clearly explaining whether or how the patient cohort was stratified to distinguish between isolated steatosis, MASH, and MASH with or without fibrosis. It is also unclear whether subgroups were propensity score-matched.

    In a related point, the authors mention that 76% of patients are non-fibrotic, introducing a marked imbalance between fibrotic (n=26) and non-fibrotic (n=83) samples. Given this disparity and potential inter-individual variability, it would be helpful to include observed fold changes or effect sizes to give readers a sense of the magnitude of the biological dysregulations being reported.

    • Should the authors qualify some of their claims as preliminary or speculative, or remove them altogether?
      • The authors seem pretty enthusiastic about elafibranor, despite a failed phase 3 clinical trial. I would qualify elafibranor as a useful tool in preclinical model.
      • The authors should make clearly the pronounced sex bias in their study, which includes mostly women (and btw refer to sex and not gender in the manuscript).
      • The "MASH" status of the spheroid model is overstated. As described in the text it is much closer to a lipotoxicity model (and even glucotoxicity as Glc concentration is 2g/L). This is confusing with panel D in which the authors establish a relationship between fibrotic patients (F2/F3 vs F0/S0, so I guess "no MASLD liver?) and this model. Is the relationship maintained for steatotic-only patients?
    • Would additional experiments be essential to support the claims of the paper? Request additional experiments only where necessary for the paper as it is, and do not ask authors to open new lines of experimentation. I am not convinced that HSC and LX2 cells express significant levels of PPARα. However, did the authors check for this parameter in their LX2 cell line and assessed whether PPARα/b activation by elafibranor (and/or pemafibrate as it is PPARα selective) alter GTPase expression? Whether negative or positive, this could give a clue about possible intercellular crosstalk in the spheroid model.
    • Are the suggested experiments realistic in terms of time and resources? It would help if you could add an estimated cost and time investment for substantial experiments.

    The experiment mentioned above is cheap (cell culture, RT-QPCR) and can be performed within a couple of weeks.

    • Are the data and the methods presented in such a way that they can be reproduced?

    Yes

    • Are the experiments adequately replicated and statistical analysis adequate?

    There is no indication of group size, number of replicates for in vitro experiments

    Referees cross-commenting

    I believe there is a general consensus on this potentially interesting contribution to the field, with three main points: (1) the need for a careful group-by-group comparison that accounts for potential confounders, (2) a more rigorous exploitation/characterization of the spheroid system, and (3) the need to benchmark the authors' findings against the available literature.

    Significance

    • Describe the nature and significance of the advance (e.g. conceptual, technical, clinical) for the field. The authors identified GTPases as players in the progression of MASLD. This is an interesting preliminary report warranting further molecular investigations (in which liver cell types, which GTPase pathway(s) are involved, which functions are controlled through this pathway...)
    • State what audience might be interested in and influenced by the reported findings. This paper will have an impact in the hepatology field
    • Define your field of expertise with a few keywords to help the authors contextualize your point of view. Indicate if there are any parts of the paper that you do not have sufficient expertise to evaluate. I have expertise in the analysis of "MASLD" human cohorts and in the molecular biology of chronic liver diseases.
  4. Review Commons

    Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Referee #1

    Evidence, reproducibility and clarity

    Metabolic dysfunction-associated steatotic liver disease (MASLD) ranges from simple steatosis, steatohepatitis, fibrosis/cirrhosis, and hepatocellular carcinoma. In the current study, the authors aimed to determine the early molecular signatures differentiating patients with MASLD associated fibrosis from those patients with early MASLD but no symptoms. The authors recruited 109 obese individuals before bariatric surgery. They separated the cohorts as no MASLD (without histological abnormalities) and MASLD. The liver samples were then subjected to transcriptomic and metabolomic analysis. The serum samples were subjected to metabolomic analysis. The authors identified dysregulated lipid metabolism, including glyceride lipids, in the liver samples of MASLD patients compared to the no MASLD ones. Circulating metabolomic changes in lipid profiles slightly correlated with MASLD, possibly due to the no MASLD samples derived from obese patients. Several genes involved in lipid droplet formation were also found elevated in MASLD patients. Besides, elevated levels of amino acids, which are possibly related to collagen synthesis, were observed in MASLD patients. Several antioxidant metabolites were increased in MASLD patients. Furthermore, dysregulated genes involved in mitochondrial function and autophagy were identified in MASLD patients, likely linking oxidative stress to MASLD progression. The authors then determined the representative gene signatures in the development of fibrosis by comparing this cohort with the other two published cohorts. Top enriched pathways in fibrotic patients included GTPas signaling and innate immune responses, suggesting the involvement of GTPas in MASLD progression to fibrosis. The authors then challenged human patient derived 3D spheroid system with a dual PPARa/d agonist and found that this treatment restored the expression levels of GTPase-related genes in MASLD 3D spheroids. In conclusion, the authors suggested the involvement of upregulated GTPase-related genes during fibrosis initiation. Overall, the current study might provide some resources regarding transcriptomic and metabolomic data derived from obese patients with and without MASLD. However, several concerns should be carefully addressed.

    1. A recent study, via proteomic and transcriptomic analysis, revealed that four proteins (ADAMTSL2, AKR1B10, CFHR4 and TREM2) could be used to identify MASLD patients at risk of steatohepatitis (PMID: 37037945). It is not clear why the authors did not include this study in their comparison.
    2. The authors recruited 109 patients but only performed transcriptomic and metabolomic analysis in 94 liver samples. Why did the authors exclude other samples?
    3. The authors mentioned clinical data in Table 1 but did not present the table in this manuscript.
    4. The generated metabolomic data could be a very useful resource to the MASLD community. However, it is very confusing how the data was generated in those supplemental tables. There is no clear labeling of human clinical information in those tables. Also, what do those values mean in columns 47-154? This reviewer assumed that they are the raw data of metabolomic analysis in plasma samples. However, without clear clinical information in these patients, it is impossible that any scientist can use the data to reproduce the authors' findings.
    5. In Fig. 5B, the authors excluded the steatosis and fibrosis overlapped genes. Steatosis and fibrosis specific genes could simply reflect the outcomes rather than causes. In this case, the obtained results might not identify the gene signatures related to fibrosis initiation.
    6. In Fig. 6D, the authors used 3D liver spheroid to validate their findings. However, there is no images showing the 3D liver spheroid formation before and after PPARa/d agonist treatment. It is not clear whether the 3D liver spheroid was successfully established.
    7. The authors suggested that targeting LX-2 cells with Rac1 and Cdc42 inhibitors could reduce collagen production. Did the authors observe these two genes upregulated in mRNA and protein expression levels in their cohort when compared MASLD patients with and without fibrosis?
    8. Did the authors observe that the expression levels of Rac1 and Cdc42 are correlated with fibrosis progression in MASLD patients?
    9. Other studies have revealed several metabolite changes related to MASLD progression (PMID: 35434590, PMID: 22364559). However, the authors did not discuss the discrepancies between their findings with the previous studies.

    Significance

    Overall, the current study might provide some new resources regarding transcriptomic and metabolomic data derived from obese patients with and without MASLD. The MASLD research community will be interested in the resource data.

  5. Version published to 10.1101/2025.01.13.632747 on bioRxiv