Dynamic Alterations to Hepatic MicroRNA-29a in Response to Long-Term High-Fat Diet and EtOH Feeding

Abstract

MicroRNA-29a (miR-29a) is a well characterized fibro-inflammatory molecule and its aberrant expression is linked to a variety of pathological liver conditions. The long-term effects of a high-fat diet (HFD) in combination with different levels of EtOH consumption on miR-29a expression and liver pathobiology are unknown. Mice at 8 weeks of age were divided into five groups (calorie-matched diet plus water (CMD) as a control group, HFD plus water (HFD) as a liver disease group, HFD plus 2% EtOH (HFD + 2% E), HFD + 10% E, and HFD + 20% E as intervention groups) and fed for 4, 13, 26, or 39 weeks. At each time point, analyses were performed for liver weight/body weight (BW) ratio, AST/ALT ratio, as well as liver histology assessments, which included inflammation, estimated fat deposition, lipid area, and fibrosis. Hepatic miR-29a was measured and correlations with phenotypic traits were determined. Four-week feeding produced no differences between the groups on all collected phenotypic traits or miR-29a expression, while significant effects were observed after 13 weeks, with EtOH concentration-specific induction of miR-29a. A turning point for most of the collected traits was apparent at 26 weeks, and miR-29a was significantly down-regulated with increasing liver injury. Overall, miR-29a up-regulation was associated with a lower liver/BW ratio, fat deposition, inflammation, and fibrosis, suggesting a protective role of miR-29a against liver disease progression. A HFD plus increasing concentrations of EtOH produces progressive adverse effects on the liver, with no evidence of beneficial effects of low-dose EtOH consumption. Moreover, miR-29a up-regulation is associated with less severe liver injury.

Keywords: EtOH; high-fat diet; liver; long-term feeding; microRNA.

Figure 1

Experimental design and body weight. (A) Overview of feeding groups, lengths of feeding, and sample collection points. Blood collections for serum analyses indicated with red arrows. Tissue collections for histology indicated with black arrows. Approximate human ages are presented at the bottom. (B) Average weekly BW of all mice in the study plotted over 1–39 weeks. Graph represents mean values +/− SEM for each week. After each tissue harvesting time point, available animal numbers decrease. From week 1–4, N = 42–43/group. From week 5–13, N = 30–31/group. From week 14–26, N = 20–21/group. From week 27–39, N = 10–11/group. Comparisons were made using a two-way ANOVA, followed by Dunnett’s multiple comparisons test. Comparisons made against the CMD group for a given time point. Statistical significance was labeled as follows: * as p < 0.05, ** as p < 0.01, *** as p < 0.001, and **** as p < 0.0001.

Figure 2

Hepatic miR-29a expression in liver. (A) Relative hepatic miR-29a expression in mice fed with a CMD, HFD, HFD + 2% E, HFD + 10% E, and HFD + 20% E for 4, 13, 26, and 39 weeks. Comparisons were made using a two-way ANOVA, followed by Dunnett’s multiple comparisons test. Comparisons made against the CMD group for a given time point. (B) Hepatic miR-29a expression relative to a CMD at 4, 13, 26, and 39 weeks. Significant correlation between miR-29a expression and (C) AST/ALT ratio at 13 weeks, (D) AST/ALT ratio and lipid area at 26 weeks, and (E) lipid area at 39 weeks of feeding. Correlations were performed, generating Pearson r values. This was followed by a simple linear regression. Statistical significance was labeled as follows: ** as p < 0.01, *** as p < 0.001, and **** as p < 0.0001. Each black dot represents data from an individual sample.

Figure 3

Average liver weight:BW ratios and epididymal fat weight:BW ratios over time. (A) Average liver weight:BW ratios plotted for 4, 13, 26, and 39-week groups. Graphs represent mean values +/− SEM, with N = 9−12/group for each time point. Comparisons were made using a two-way ANOVA, followed by Dunnett’s multiple comparisons test. Comparisons made against the CMD group for a given time point. (B) Comparisons between the 25 samples with the highest and 25 samples with the lowest miR-29a expression (i.e., top and bottom 25% of the samples) revealed differences in the liver weight:BW ratio. Comparisons were made using Student’s t-test, with the statistical significance labeled as follows: * as p < 0.05. (C) A negative correlation was found between miR-29a expression and the liver weight:BW ratio. (D) Average epididymal fat weight:BW ratios plotted for 4, 13, 26, and 39-week groups. The average epididymal fat weight:BW ratios increase over time and demonstrate an inverse correlation with ethanol concentration after 39 weeks of feeding. Graphs represent mean values +/− SEM, with N = 9–12/group for each time point. Comparisons were made using a two-way ANOVA, followed by Dunnett’s multiple comparisons test. Comparisons made against the CMD group for a given time point. (E) Comparisons between the 25 samples with the highest and 25 samples with the lowest miR-29a expression (i.e., top and bottom 25% of the samples) revealed no differences in the average epididymal fat weight and the epididymal fat weight:BW ratio. (F) No correlation was found between miR-29a expression and the epididymal fat weight:BW ratio. Correlations were performed, generating Pearson r values. This was followed by a simple linear regression. Statistical significance labeled as follows: * as p < 0.05, ** as p < 0.01, and **** as p < 0.0001. NS: not significant. Each black dot represents data from an individual sample.

Figure 4

Average AST, ALT, and AST:ALT ratios over time and their association with miR-29a expression. (A) Average AST and (B) ALT activity (mU/mL) and (C) AST:ALT ratios plotted for each collection time. With the exception of the 4-week feeding length, blood samples were collected at two time points for all mice. Line graphs reflect average measurements over time. Graphs represent mean values +/− SEM, with N = 9–12/group for each time point. Comparisons were made using a two-way ANOVA, followed by Dunnett’s multiple comparisons test. Comparisons made against the CMD group for a given time point. (D–F) Comparisons between the 25 samples with the highest and 25 samples with the lowest miR-29a expression (i.e., top and bottom 25% of the samples) revealed a negative correlation between miR-29a expression and ALT activity. Comparisons were made using Student’s t-test. Statistical significance labeled as follows: * as p < 0.05, ** as p < 0.01, *** as p < 0.001, and **** as p < 0.0001. NS: not significant.

Figure 5

Liver histology at 13 weeks. (A) Representative gross morphology of livers from each feeding group. (B) Oil Red O staining, with enlarged area indicated by box, from representative liver sections. (C) Tri-Chrome staining, with enlarged area indicated by box, from representative liver sections. (D) Hematoxylin and eosin staining from representative liver sections.

Figure 6

Liver histology at 26 weeks. (A) Representative gross morphology of livers from each feeding group. (B) Oil Red O staining, with enlarged area indicated by box, from representative liver sections. (C) Tri-Chrome staining, with enlarged area indicated by box, from representative liver sections. (D) Hematoxylin and eosin staining from representative liver sections.

Figure 7

Liver histology at 39 weeks. (A) Representative gross morphology of livers from each feeding group. (B) Oil Red O staining, with enlarged area indicated by box, from representative liver sections. (C) Tri-Chrome staining, with enlarged area indicated by box, from representative liver sections. (D) Hematoxylin and eosin staining from representative liver sections.

Figure 8

Total lipid area and droplet size analysis. (A) Average lipid area as determined by densitometric analysis of Oil Red O photomicrographs. Graphs represent mean values +/− SEM, with N = 7/group. Comparisons were made using a one-way ANOVA, followed by Dunnett’s multiple comparisons test. Comparisons made against the CMD group for a given time point. (B) Abundance of lipid droplets of particular sizes. Graphs represent the average number of droplets of a given size counted for each feeding group. (C) Change in the relative proportions of different sized lipid droplets over time. Graphs represent the percent abundance of particular sized lipid droplets relative to the CMD group at the same time point. Statistical significance labeled as follows: ** as p < 0.01, and **** as p < 0.0001.

Figure 9

Histological assessments and correlation with miR-29a at 26 and 39 weeks. (A) Average inflammation scores for the feeding groups at 26 and 39 weeks, as determined by a professional histopathologist. Graphs represent mean values +/− SEM, with N = 6–11/group. (B) Average fibrosis scores for the feeding groups at 26 and 39 weeks, as determined by a professional histopathologist. Graphs represent mean values +/− SEM, with N = 5–11/group. (C) Average microvesicular steatosis (MiS) scores for the feeding groups at 26 and 39 weeks, as determined by a professional histopathologist. Graphs represent mean values +/− SEM, with N = 4–10/group. Comparisons were made using a one-way ANOVA, followed by Dunnett’s multiple comparisons test. Comparisons made against the CMD group for a given time point. The 25 samples with the highest and 25 samples with the lowest miR-29a expression (i.e., top and bottom 25% of the samples) demonstrated a significant difference in (D) inflammation and (E) fibrosis, but not (F) the MiS score. Graphs represent average values +/− SEM for the given traits. Comparisons were made using Student’s t-test. Moreover, miR-29a expression is negatively correlated with (D) inflammation (N = 40) and (E) fibrosis (N = 41) scores, but not with (F) the MiS scores. Correlations were performed, generating Pearson r values. This was followed by a simple linear regression. Statistical significance for all graphs was labeled as follows: * as p < 0.05, ** as p < 0.01, *** as p < 0.001, and **** as p < 0.0001. Each dot represents data from an individual sample.