Dysregulation of the Splicing Machinery Is Associated to the Development of Nonalcoholic Fatty Liver Disease

Abstract

Context: Nonalcoholic fatty liver disease (NAFLD) is a common obesity-associated pathology characterized by hepatic fat accumulation, which can progress to fibrosis, cirrhosis, and hepatocellular carcinoma. Obesity is associated with profound changes in gene-expression patterns of the liver, which could contribute to the onset of comorbidities.

Objective: As these alterations might be linked to a dysregulation of the splicing process, we aimed to determine whether the dysregulation in the expression of splicing machinery components could be associated with NAFLD.

Participants: We collected 41 liver biopsies from nonalcoholic individuals with obesity, with or without hepatic steatosis, who underwent bariatric surgery.

Interventions: The expression pattern of splicing machinery components was determined using a microfluidic quantitative PCR-based array. An in vitro approximation to determine lipid accumulation using HepG2 cells was also implemented.

Results: The liver of patients with obesity and steatosis exhibited a severe dysregulation of certain splicing machinery components compared with patients with obesity without steatosis. Nonsupervised clustering analysis allowed the identification of three molecular phenotypes of NAFLD with a unique fingerprint of alterations in splicing machinery components, which also presented distinctive hepatic and clinical-metabolic alterations and a differential response to bariatric surgery after 1 year. In addition, in vitro silencing of certain splicing machinery components (i.e., PTBP1, RBM45, SND1) reduced fat accumulation and modulated the expression of key de novo lipogenesis enzymes, whereas conversely, fat accumulation did not alter spliceosome components expression.

Conclusion: There is a close relationship between splicing machinery dysregulation and NAFLD development, which should be further investigated to identify alternative therapeutic targets.

Figure 1.

Differential expression of splicing machinery components in the liver of patients with and without steatosis with obesity. (A) Study design and pattern of dysregulation of spliceosome components and splicing factors in the liver of patients with and without steatosis with obesity. (Left) A graphical summary of the study design is shown. (Right) A schematic representation of fold-change levels of spliceosome components and splicing factors between the liver of patients with and without steatosis with obesity, represented in red (increase) or blue (decrease), is depicted. (B) Expression levels and receiver operating characteristic (ROC) curves of significantly altered spliceosome components and splicing factors in the liver of patients with obesity and steatosis compared with patients with obesity without steatosis. mRNA expression levels (adjusted by an NF, calculated from the expression level of HPRT and ACTB) of the different spliceosome components (first and second rows) and splicing factors (third and fourth rows) significantly altered in the liver of women with obesity with steatosis (ST) and without steatosis (NON ST). Values represent the means ± SEM. Asterisks indicate values that significantly differ from patients without steatosis (t test: *P < 0.05, **P < 0.01, ***P < 0.001). ROC curve analyses, to determine the accuracy of the components of the splicing machinery and splicing factors to discriminate between patients, with or without liver steatosis, are included below each graph. (C) ROC curves of subsets of spliceosome components and splicing factors generated by Random Forest computational algorithm, followed by crossvalidation analysis, considering the expression of a selection of the most relevant and discriminatory splicing machinery components. Specifically, three subset of specific splicing machinery components are included: PTBP1, RBM45, SRRM1, RNU4, and RNU6ATAC; PTBP1, RBM22, SRSF1, SRRM1, SNRNP70, and RNU6ATAC; and CELF1, PTBP1, RBM22, RBM3, SRRM1, and RNU6. AUC, area under curve.

Figure 2.

Pattern of dysregulation of spliceosome components and splicing factors in the liver of patients with and without steatosis with obesity, according to the grade of lipid accumulation. (Top) Fold-change expression levels among patients with different levels of steatosis compared with nonsteatotic livers, represented in red (increase) or blue (decrease). (Bottom) mRNA expression levels (adjusted by an NF calculated from the expression levels of HPRT and ACTB) of the different spliceosome components and splicing factors in the liver of women with obesity without steatosis (NON ST) and with three different levels of steatosis (mild, moderate, and severe). Values represent the means ± SEM. Asterisks indicate values that significantly differ from patients without steatosis (t test: *P < 0.05, **P < 0.01, ***P < 0.001).

Figure 3.

The expression of splicing machinery components is differentially altered in the liver of individuals with steatosis. (A) Unsupervised clustering analysis of the expression levels of the splicing machinery in patients with steatosis. This bioinformatic approach identified three molecularly defined populations of patients with steatosis (Clusters A, B, and C). (B) Specific changes of certain components of the splicing machinery defined each cluster of patients with steatosis. The three molecularly defined clusters of patients with steatosis were associated with the alteration in the expression of certain spliceosome components and splicing factors compared with patients without steatosis or included in the other clusters. The alteration of selected factors (within the frame) was able to classify patients in the three clusters with an AUC of 1, using the classification algorithm Random Forest. Data indicate mRNA expression levels (adjusted by an NF calculated from the expression level of HPRT and ACTB) in each cluster (A, B, and C) compared with the rest of patients, with and without steatosis (NON ST). Values represent the means ± SEM. Asterisks indicate values that significantly differ between groups (t test: *P < 0.05, **P < 0.01, ***P < 0.001).

Figure 4.

Each molecularly defined population of patients with steatosis [Clusters (CL) A, B, and C] was characterized by certain hepatic and clinical-metabolic alterations. Relevant clinical parameters associated to each cluster of patients were grouped according to their expression levels of the spliceosome components and splicing factors. Values represent the means ± SEM. Asterisks indicate values that significantly differ between groups (t test: *P < 0.05, **P < 0.01). AST, aspartate transaminase; GGT, gamma-glutamyltransferase; NON ST, without steatosis.

Figure 5.

Cluster C presented a worst response to bariatric surgery after 1 year of follow-up. Recovery from hepatic steatosis and evolution of BMI, waist/hip ratio (WHR), GGT, glucose, triglycerides, alkaline phosphatase, and high-density lipoprotein (HDL) levels, 1 year after bariatric surgery in patients from Cluster C vs Clusters A + B. The data are expressed as percentage of the value before surgery (normalized to 100%). Values represent the means ± SEM. Asterisks indicate values that significantly differ between Cluster C and Cluster A + B (t test: *P < 0.05, **P < 0.01).

Figure 6.

Modulation of splicing machinery components influenced lipid accumulation in HepG2 cells. (A) Validation of lipid accumulation in HepG2 cell lines by Oil Red O absorbance at 520 nm. Data are expressed as a percentage of the control (Ctrl; normalized to 100%). Asterisks indicate values that significantly differ from control cells (t test: ***P < 0.001). (B) Effect of lipid accumulation on the expression of certain spliceosome components and splicing factors at 10, 24, and 48 hours. mRNA expression levels (adjusted by an NF calculated from the expression level of HPRT and GAPDH) of the different spliceosome components and splicing factors in HepG2 cells treated with 500 µM OA. The data are expressed as a percentage of the control (normalized to 100%). Values represent the means ± SEM (n = 5). (C) qPCR validation of the silencing with specific siRNAs. mRNA expression levels adjusted by the expression level of ACTB. Data are expressed as a percentage of control random siRNAs (Scramble; set at 100%). (D) Effect of silencing of certain splicing factors in HepG2 cells on lipid accumulation determined by Oil Red O absorbance at 520 nm. Data are expressed as percentage of the control (normalized to 100%). Values represent the means ± SEM (n = 5; t test: *P < 0.05, **P < 0.01, ***P < 0.001).