RHO GTPASE SIGNALING-BASED GENE SIGNATURE ENABLES EVALUATION OF HEPATOCELLULAR CARCINOMA IMMUNE INFILTRATION, THERAPEUTIC BENEFIT, AND PAN-CANCER PROGNOSIS

Background: Alcohol-associated hepatitis (AH) is the active inflammatory state of alcohol-associated liver disease (ALD). AH varies in severity from mild to severe with mortality rates reaching 30-50% in severe AH. Prednisolone, which is a glucocorticoid receptor agonist, is the standard treatment for severe AH. The duration of treatment is based on the clinical response to prednisolone that is calculated by Lille score, where a Lille Score <0.45 at day 7 indicates that a prednisolone “responder” (AH-R) should continue treatment for 28 days. A Lille Score >0.45 indicates that a patient is a “non-responder” (AH-NR) to prednisolone and treatment should be terminated at day 7. Utilizing a proteomic approach, the current study aimed to identify hepatic proteins/pathways associated with non-responsiveness to prednisolone treatment and to characterize proteomic changes that negatively impact mortality in AH-NR.
Methods: LC/MS proteomic analysis was performed on liver biopsy samples obtained from non-ALD controls (n=10) and AH patients (n=21). Six-month survival within the AH-R was 93% (n=13 out of 14 patients) and 43% (n= 3 out of 7 patients) within the AH-NR. Significant differences in hepatic protein expression between groups were identified by a one-way ANOVA.
Results: Analysis of the hepatic proteome identified 7,944 total proteins and revealed 222 proteins that were decreased and 294 that were increased in AH-NR vs AH-R. Notably, among the significantly increased proteins in AH-NR was DHI1, an enzyme involved in metabolism of both prednisolone and prednisone. Significantly downregulated proteins in AH-NR included members of the glucocorticoid signaling pathway, GCR and GMEB2. Further, when comparing to controls, AH-R and AH-NR shared multiple protein changes but there were various proteins and pathways explicitly changed in AH-R or AH-NR. For example, the proteotoxic stress response pathway including HSF1, PSD10, RAF1, and HPBP1 was uniquely downregulated in AH-NR but not in AH-R. Next, when comparing non-survivors vs survivors within the AH-NR 113 significant proteins (57 decreased and 56 increased) were identified. Proteins involved in blood coagulation, defense response, and response to stress were significantly reduced in AH-NR non-survivors. In addition, acute inflammatory response and the expression of several acute phase proteins were downregulated in AH-NR non-survivors. For example, A1AG2, a steroid transport protein, AACT, a protease inhibitor, and AMBP, a heme homeostasis protein, among others.
Conclusion: The current study identified the specific changes in the hepatic proteins associated with non-responsiveness to prednisolone treatment and mortality in AH patients. Further studies are needed to investigate the mechanisms as to how these changes may contribute to AH patient response to prednisolone therapy and mortality.

Background: Heavy alcohol consumption leads to an increased risk of alcohol-associated liver disease (ALD), a condition associated with significant morbidity and mortality, for which there are limited treatments. ALD pathogenesis is influenced by multiple factors including dietary fatty acids. Preclinical studies in mice show that n-3 PUFA supplementation or endogenous increase attenuate experimental ALD. The goal of the current study was to examine the effects of n-3 PUFA supplementation on liver injury, systemic inflammation, and intestinal permeability in a human population of heavy drinking individuals.
Methods: Heavy drinking individuals were admitted to an NIH inpatient treatment program for 21 days. At discharge, subjects were randomized to either 2 g/day of n-3 PUFAs (EPA+DHA, n=46) or placebo (n=48) for 12 weeks. Plasma markers of liver injury (ALT, AST, and GGT) were measured at randomization (baseline) and at the conclusion of the study (final). In a subset of patients, an additional marker of liver injury (K18), systemic inflammation (TNFα, IL-8, IL-6), and intestinal permeability (flagellin, CD14, and endotoxin) were evaluated. Significant differences were determined by paired Student’s t test (p<0.05).
Results: 12-week n-3 PUFA supplementation was associated with a substantial reduction in ALT (55.15±9.52 vs35.21±4.14 U/L, p<0.05), and GGT (55.03±10.58 vs 33.25±4.89 U/L, p<0.05) levels among heavy drinkers (baseline vs. final). ALT and GGT were unchanged among the placebo group. No significant changes in AST were observed for either group. The AST:ALT ratio was unchanged for the n-3 PUFA group, but was significantly increased among the placebo group (0.85±0.05 vs 1.10±0.05, p<0.05, baseline vs final. Compared to baseline, K18 levels (specifically M65, a marker of necrosis) had a non-significant downward trend in the n-3 PUFA group (-18.94±9.14%), whereas a slight upward trend was observed for the placebo group. As an indirect marker of intestinal permeability, endotoxin levels were unchanged in the n-3 PUFA supplemented group over the course of the study, but there was a significant increase in the placebo arm (0.79±0.10 vs 1.23±0.17 EU/mL, p<0.05). CD14 and flagellin were unchanged for both groups. Lastly, levels of inflammatory cytokines IL-6, TNFα, and IL-8 had a non-significant downward trend in n-3 PUFA supplemented individuals (-32.85±13.29%; -15.35±16.60%, -8.46±20.50%, respectively). However, there was a non-significant increase for these cytokines in the placebo group (9.21%±15.69, 9.12%±11.80, 46.00%±18.17, respectively).
Conclusion: Our data suggested that n-3 PUFA supplementation in heavy drinking individuals had beneficial effects on liver injury, endotoxemia, and systemic inflammation. Further investigations are needed on the mechanistic role of n3 PUFAs in the pathogenesis of ALD.

Background: Intrahepatic cholangiocarcinoma (ICC) and hepatocellular carcinoma (HCC) comprise more than 90% of primary liver cancers, and rank as the third leading cause of cancer-related deaths. Because of their high incidence and differences in treatments and prognoses, the diagnostic specificity of ICC and HCC are significant for clinical decision-making. However, currently available blood-based diagnostic markers are inadequate to avoid a biopsy, with a risk of severe complications. Exosomal cargo, especially miRNAs (exo-miRNAs), are emerging as attractive targets for biomarker development as these are stable and offer robust tumor-specificity. Hence, in this study, we hypothesized that a blood-based exo-miRNA signature might lead to developing a clinically significant tool for the differential diagnosis of patients with ICC and HCC.
Methods: For biomarker discovery, we analyzed genomewide miRNA expression profiling data tissue-based data from publicly available datasets and our own expression profiling data generated through blood-based exosomal small RNA sequencing in patients with ICC and HCC. Bioinformatically prioritized miRNAs were trained and validated to establish an exo-miRNA signature in serum specimens from 412 patients enrolled within the training and validation cohorts. The established panel was finally integrated by combining it with key clinical features specific for patients with ICC and HCC, respectively, for the differential diagnosis of both liver cancers.
Results: Following rigorous bioinformatic and biostatistical analyses, a panel of 7 miRNAs was identified that exhibited significantly differential expression between ICC and HCC in the public datasets (|Log2Foldchange| >1.0 and Adjusted P<0.05), as well as demonstrated a similar trend in the exosomal small RNA sequencing data. Analysis of these candidate biomarkers using RT-qPCR assays in 412 patients with ICC and HCC allowed us to narrow down this panel to 5 exo-miRNAs, which yielded a discriminatory area under the curve (AUC) value of 0.95 in the training cohort and an AUC of 0.77 in the independent validation cohort. When we combined our exo-miRNA panel with key clinicopathological features that were significant in the univariate analysis, including CA19-9, AFP levels, and patient gender, the final model exhibited a significantly superior diagnostic performance for discriminating patients with ICC and HCC (AUC = 0.93).
Conclusions: We identified a panel of exo-miRNAs that could successfully differentiate between patients with ICC and HCC, and when combined with key clinical features, allowed us to establish a novel diagnostic signature for differential diagnosis of these two types of liver cancers in clinical practice.

Background: Advanced fibrosis is the most important measurement of chronic fatty liver disease therefore analyzing the gut microbiome of patients with non-alcoholic fatty liver disease (NAFLD) is crucial to understand how the gut microbiome is associated with advanced fibrosis.
Methods and Results: Stool samples obtained from patients with NAFLD (n=29) underwent FibroScan (ultrasound elastography) at the VA Greater Los Angeles Healthcare System. 29 individuals with NAFLD were compared to a group of 25 healthy control patients. Fecal samples underwent 16S ribosomal RNA sequencing. Patients without advanced fibrosis and NAFLD had a significantly lower alpha diversity by Chao1 index (p < 0.05). Patients with advanced fibrosis and NAFLD had higher levels of alpha diversity according to Chao1 as compared to patients without advanced fibrosis and NAFLD (p < 0.05). There was a significant beta diversity difference between the three groups (healthy control, NAFLD + advanced fibrosis, NALFD - advanced fibrosis) (p = 0.003). Differential abundance testing showed patients with advanced fibrosis and NALFD had high levels of Prevotella coprias (P. Copri) compared to healthy controls and patients with NALFD without advanced fibrosis (q < 0.05). To show causality of P. Copri, we utilized a diet induced obesity model for NAFLD in mice. Antibiotic treated (ertapenem ampicillin, cefoperazone) C57/Bl6 mice (n=8 per group) on either a standard control diet or a NASH diet (40% fat, 2% cholesterol, 20% fructose) were each gavaged orally with 3x10⁸ CFUs of P. Copri or a vehicle control for 16 weeks. Mice were gavaged twice a week for the first four weeks and once a week afterwards. Liver tissue samples were obtained from each mouse at the completion of the study and stained with hematoxylin and eosin. The fat content visible in the histological slides were quantified as a percentage of the captured image using ImageJ. The livers were then also processed by trichrome stain and total collagen. On a standard diet, there was no statistical difference in the fat content between mice gavaged with P. Copri compared to the control (p = 0.45). However, on the NASH diet, mice gavaged with P. Copri had significantly higher fat content compared to those in the control (p = 0.02). Utilizing the trichrome stain and total collagenase there was no statistical significance between P. Copri and control in either of the dietary groups.

Conclusion: These results support the hypothesis that there is a causal relationship between P. Copri colonization and the progression of fatty liver disease. Even though there were not any significant differences in fibrosis, it is possible that a longer-term exposure of P. Copri may lead to further progression of fibrosis. This raises the possibility of altering the microbiome to halt or delay the progression of fatty liver disease.

Introduction: Emerging studies have linked Rho GTPases, the crucial signal transducers in cell regulation, to tumorigenesis and metastasis, whereas their role in the tumor immune microenvironment (TME) and prognosis of hepatocellular carcinoma (HCC) remains poorly characterized.
Methods: Univariate Cox and the least absolute shrinkage and selection operator (LASSO) are used to develop a tumor prognosis prediction system named Rho GTPases-related gene score (RGPRG score). The survival difference, clinicopathological features, tumor immune microenvironment, response to immunotherapy, and chemotherapy, pan-cancer prognosis are then analyzed between RGPRG score groups.
Results: HCC patients with High RGPRG score show significantly worse survival than low RGPRG score patients (P<0.0001). GSEA demonstrates that High RGPRG score is correlated with cell cycle and primary immunodeficiency activation, while the increased immunosuppressive cell fractions estimated by CIBERSORT could be observed in the samples with high RGPRG score. Single-cell cohort of HCC patients implies an immune-active tumor microenvironment in the patients with low RGPRG score, and communication from T/NK cells to other cells are strengthened through MIF signaling networks, which may contribute to Rho GTPase phenotypes. Targeting these alterations in TME, the patients with high RGPRG score have worse immunotherapeutic outcomes and decreased survival time in the immunotherapy cohort (IMvigor210 and GSE78220). Besides, Genomics of Drug Sensitivity in Cancer (GDSC) analysis reveals 91 potential anti-cancer drugs associated with Rho GTPase phenotypes, which indicates that RGPRG score is a valuable signature to predict patients’ response to immunotherapy and chemotherapeutic agents. More importantly, RGPRG score is found to be correlated with 27 cancers survival, suggesting that RGPRG score could be extended to predict prognosis for pan-cancer. Further in vitro experiment confirms that knockdown of key Rho GTPase-signaling gene SFN could significantly inhibit the proliferation of HCC cells.
Conclusions: Taken together, we proposed a high-sensitivity prognostic model named RGPRG score system, and revealed that a high RGPRG score is an independent prognostic factor of HCC that is associated with poor prognosis, increased immunosuppressive cell infiltration, and a low response to immunotherapy, which may provide additional clinical insights underlying biological features of Rho GTPase in HCC.