Uncovering Novel Roles of miR-122 in the Pathophysiology of the Liver: Potential Interaction with NRF1 and E2F4 Signaling

Abstract

MicroRNA miR-122 plays a pivotal role in liver function. Despite numerous studies investigating this miRNA, the global network of genes regulated by miR-122 and its contribution to the underlying pathophysiological mechanisms remain largely unknown. To gain a deeper understanding of miR-122 activity, we employed two complementary approaches. Firstly, through transcriptome analysis of polyribosome-bound RNAs, we discovered that miR-122 exhibits potential antagonistic effects on specific transcription factors known to be dysregulated in liver disease, including nuclear respiratory factor-1 (NRF1) and the E2F transcription factor 4 (E2F4). Secondly, through proteome analysis of hepatoma cells transfected with either miR-122 mimic or antagomir, we discovered changes in several proteins associated with increased malignancy. Interestingly, many of these proteins were reported to be transcriptionally regulated by NRF1 and E2F4, six of which we validated as miR-122 targets. Among these, a negative correlation was observed between miR-122 and glucose-6-phosphate dehydrogenase levels in the livers of patients with hepatitis B virus-associated hepatocellular carcinoma. This study provides novel insights into potential alterations of molecular pathway occurring at the early stages of liver disease, driven by the dysregulation of miR-122 and its associated genes.

Keywords: E2F4; G6PD; NRF1; liver cancer; miR-122; polyribosomes.

Figure 1

Direct identification of miR-122 targets by polyribosomal profiling on sucrose gradient. (A) Huh-7 cells were transfected with either miR-122 mimic (122-MIM) or miR-122 inhibitor (122-PM) and lysed 48 h after transfection. The cytosolic extracts were sedimented by ultracentrifugation on a 10–50% sucrose gradient. Upon fractionation, RNA isolation, quantification, and quality control (QC) were performed for the corresponding polyribosomal fractions. The UV absorbance profile of the nucleic acids associated with polyribosomes was measured at 254 nm from miR-122-overexpressing and miR-122-depleted cells. The collected fractions are depicted as (A1–A12). Analysis of miRNA and mRNA distribution in isolated polyribosomes of cells transfected with miR-122 mimics (blue) or cells treated with miR-122-antagomiR (red) by qPCR. (B) The analysis showed an increased level of miR-122 associated to polyribosomes in cells treated with miR-122 mimics (C), while the distribution of unrelated miRNAs as shown for miR-194 was unaffected by the transfection. (D) In polyribosomes isolated from antagomiRs treated cells, a significant enrichment toward the heavy transcribed fraction in the distribution of SLC7A1 mRNA, the best characterized target for miR-122, was observed. (E) In contrast, the transferrin receptor 2 (TFR2) mRNA, which is not an miR-122 target, did not show any significant changes in its polyribosomal distribution. qPCR data were median-normalized as described in [14]. Data are shown either as miRNA/mRNA expressions in individual fractions or as box plot using the mean ± SD for the pooled fractions (n = 4). Statistical analysis was performed with a two-tailed t-test; p ≤ 0.05 was considered significant. * p ≤ 0.05.

Figure 2

Changes in mRNA distribution on polyribosomes in response to miR-122 overexpression (122-MIM). Huh-7 cells were treated with miR-122 mimic and lysed 48 h later. Following sedimentation on 10–50% sucrose gradient, lysates were fractionated, and fractions were pooled as indicated (A) to create a heavy pool (P1), a middle pool (P2), and a light pool (P3). RNA isolated from each individual pool were hybridized to Affymetrix arrays, and data were analyzed by AltAnalyze [15] using a cutoff of 1.5 fold changes (significance level p ≤ 0.05, one-way ANOVA) The arrow indicates the expected target gene shift across polyribosomes in response to miR-122 overexpression. Microarray data are available via the GEO repository under the accession number GSE234690. (B) Hierarchical clustering representing the differential distribution of mRNAs associated with the polyribosomal pools (P2 vs. P1, P2 vs. P3, and P3 vs. P1). Cosine matrix was applied to generate a hierarchical tree of gene clusters (n = 2). (C) Cake chart illustrating the number of regulated transcripts between the different pools P2 vs. P1 (left), P2 vs. P3 (middle), and P3 vs. P1 (right). (D) Evaluation of microarray data by GO-Elite algorithm identified a link between miR-122-responsive genes and the nuclear respiratory factor 1 (NRF1), E2F transcription factor 4 (E2F4), transcription Forkhead box protein P3 (FOXP3), and Yin Yang 1 (YY1).

Figure 3

qPCR validation of potential miR-122 target genes in miR-122-overexpressing Huh-7 cells. (A) Human hepatoma cells Huh-7 were transfected with miR-122 mimic and RNA isolated 24 h and 48 h after transfection. The overexpression of cellular miR-122 was measured 24 h and 48 h after transfection by miQPCR [23], while miRNA expressions were normalized to miR-192. (B) Relative expression changes in the mRNA levels of NRF1, YY1, and E2F4 are depicted at 24 h and 48 h after miR-122 overexpression. (C) Relative expression changes for selected candidates are depicted at 24 h and 48 h post transfection. Death effector domain-containing protein (DEDD) mRNA was identified as the most stable transcript in the dataset using GeNorm algorithm for reference gene quality. The validated miR-122 target SLC7A1 mRNA served as a positive control. The values of control cells were set to 100%. Data are shown as the mean ± SD (n = 5). Statistical analyses were performed with a two-tailed t-test; p ≤ 0.05 was considered significant. ns = not significant; * p ≤ 0.05; ** p ≤ 0.01; *** p ≤ 0.001. Full gene names are listed in Supplementary Data S2.

Figure 4

Proteome analysis of miR-122-enriched and -depleted Huh-7 cells. (A) Quantitative proteome analysis of Huh-7 cell lines transfected with miR-122 mimics, antagomiRs, or scrambled oligo control (scr). Unsupervised hierarchical clustering illustrating the relative abundance (Z-score) of all quantified proteins in Huh-7 upon modulation of cellular miR-122 levels (n = 5). (B) Volcano plot illustrating the pairwise differences in protein abundances in Huh-7 treated with miR-122 mimic-, antagomiR-, and scrambled oligo-transfected cells. Expression changes between groups are plotted in log₂ scale on the abscissa (x-axis), while p-values measured by one-way ANOVA are plotted in log₂ scale on the ordinate (y-axis). Thresholds for fold changes (FC ≥ 1.5 or FC ≤ −1.5) are shown as perpendicular dashed lines, while the threshold for statistical relevance (p ≤ 0.05) is shown as a horizontal dashed line. Proteins highlighted in green were found in significant lower abundance, while significantly more abundant proteins are highlighted in red. (C) Functional analysis of regulated proteins was conducted using Gene Ontology enrichment tool GOrilla [27]. The schematic representation illustrates the number of proteins associated with the six most significantly (i.e., with lowest p-value) depleted (green) and enriched (red) GO terms regarding biological process (upper) and cellular components (lower). The number of proteins related to the GO terms is given in red (enriched) or green (depleted) next to the bars, while the p-value for each GO term is written next to the y-axis. (D) Relative expression changes for selected target proteins at mRNA levels in response to miR-122 overexpression (122-MIM). Expressions were normalized to hypoxanthine phosphoribosyltransferase 1 (HPRT1). The values of scrambled control (CTRL) transfected cells were set to 100%. Data are shown as mean ± SD (n = 5). Statistical analyses were performed with a two-tailed t-test; p ≤ 0.05 was considered significant. *** p ≤ 0.001.

Figure 5

Correlation between miR-122 and G6PD expression in tumor tissue of HCC patients with or without HBV infection. (A) The full-length 3′UTRs for the seven GOIS were cloned into the luciferase promoter vector in the sense orientation [pMir(+)] or in the antisense orientation [pMir(−)] as a negative control. The plasmids pMir(+)_122 containing the perfect miR-122 binding site in sense orientation, and pMir(−)_122 (miR-122 binding site in the antisense orientation) served as the positive and negative controls, respectively. Plasmids were transfected in the presence (gray) or absence (black) of miR-122 mimic into HEK293 cells, and luciferase activity was measured 24 h post transfection. Relative luciferase activities are shown as a percentage of plasmid transfection control (n = 4). Western blot analysis demonstrated differential (B) EPS15L1 and (C) G6PD expression in Huh-7 treated with miR-122 mimic (122-MIM) or miR-122 inhibitor (122-PM) for 48 h (n = 4). Quantification of protein expression was normalized to actin. Statistical analyses were performed with a two-tailed t-test; p ≤ 0.05 was considered significant. ns = not significant; * p ≤ 0.05; ** p ≤ 0.01; *** p ≤ 0.001; **** p < 0.0001. The uncropped bolts are available in the Supplementary Data.

Figure 6

G6PD protein levels are significantly increased in the tumor tissue isolated from the livers of HCC patients. (A) Total proteins were isolated from the livers tissue of HCC patients, and the Western blot analysis for EPSL15L1 and G6PD was carried out in the tumor (HCC, n = 7) and in the non-tumor (NTL, n = 7) cells. (B) Quantification of protein expression was normalized to GAPDH. Statistical analyses were performed with s two-tailed t-test, p ≤ 0.05 was considered significant. ns = not significant; *** p ≤ 0.001. The uncropped bolts are available in Supplementary Data.

Figure 7

Correlation between miR-122 and G6PD expression in tumor tissue of HCC patients with or without HBV infection. (A) Analysis of LIHC cohorts from TCGA. Expression of G6PD was significantly increased in normal vs. primary tumor tissue (Left panel), as well as in the livers of patients with cancer stages 1 to 3 (Right panel). (B) Total RNAs were isolated from the FFPE liver tissue of HCC patients with (HBV-infected; n = 7) or without (non-viral: n = 26) HBV infection. Expression profiling of miR-122 (Left panel, normalized to miR-192) and G6PD (Right panel, normalized to HPRT1) were measured by using qPCR. (C) Linear regression analysis showed a significant (p = 0.0383) inverse correlation between miR-122 and G6PD in the livers of HBV-associated HCC (HBV-HCC) patients, whereas no correlation was found in non-viral HCC patients. A two-tailed t-test was used to compare two groups, One-way ANOVA was used to compare three or more groups, whereas linear regression (R²) was used to assess correlation between samples; p ≤ 0.05 was considered significant. ns = not significant; * p ≤ 0.05.

Figure 8

Proposed model for miR-122 in health and HBV-associated HCC. (A) Physiological regulation of miR-122: Physiological regulation of miR-122: miR-122 is transiently downregulated in response to (yet) unknown factors during inflammation, leading to upregulation of its targets and fine-tuning of gene expression downstream of NRF1, E2F4, and YY1 TFs. This model suggests that the transient downregulation of miR-122 leads to the fine-tuning of immune response and regeneration-associated networks. (B) Contribution of miR-122 in HBV-associated HCC: We propose that the dysregulation of miR-122 expression is a central factor in HCC development among HBV-infected patients. We propose that miR-122 transcription is inhibited by inflammation (model A). In addition, in HBV-infected patients, the viral-synthesized HBx has a dual role: (i) inhibiting miR-122 [41], and (ii) activating G6PD expression [33]. The combined downregulation of miR-122 and the activation of HBx leads to a “super activation” of G6PD, promoting cell survival (i.e., Bcl2/Bcl-xL activation [42]), energy metabolism [33], and ROS detoxification [43]. We propose that these changes might promote the transformation of hepatocytes to a malignant phenotype in HBV-infected patients.