**Preface to "Recent Advances in the Diagnosis and Treatment of Chronic Liver Diseases"**

Dear readers,

It is my great pleasure and honor to publish a book entitled Advances in the Diagnosis and Treatment of Chronic Liver Diseases. Chronic liver diseases develop from a wide range of causes, including hepatitis B virus (HBV) infection, hepatitis C virus (HCV) infection, alcoholic-related liver disease, non-alcoholic fatty liver disease (NAFLD), and autoimmune liver diseases. Recent advances in molecular and cellular techniques have succeeded in providing new aspects in the diagnosis and treatment of chronic liver diseases. This book includes seven state-of-the-art studies on chronic liver diseases.

I believe that the present special collection would be beneficial for a wide range of readers. Sincerely,

> **Hirayuki Enomoto** *Editor*

## *Article* **Inhibitory E**ff**ect of a Human MicroRNA, miR-6133-5p, on the Fibrotic Activity of Hepatic Stellate Cells in Culture**

**Susumu Hamada-Tsutsumi <sup>1</sup> , Masaya Onishi <sup>1</sup> , Kentaro Matsuura <sup>2</sup> , Masanori Isogawa <sup>1</sup> , Keigo Kawashima <sup>1</sup> , Yusuke Sato <sup>3</sup> and Yasuhito Tanaka 1,4,\***


Received: 30 July 2020; Accepted: 28 September 2020; Published: 1 October 2020

**Abstract:** Background: We recently identified 39 human microRNAs, which effectively suppress hepatitis B virus (HBV) replication in hepatocytes. Chronic HBV infection often results in active, hepatitis-related liver fibrosis; hence, we assessed whether any of these microRNAs have anti-fibrotic potential and predicted that miR-6133-5p may target several fibrosis-related genes. Methods: The hepatic stellate cell line LX-2 was transfected with an miR-6133-5p mimic and subsequently treated with Transforming growth factor (TGF)-β. The mRNA and protein products of the *COL1A1* gene, encoding collagen, and the *ACTA2* gene, an activation marker of hepatic stellate cells, were quantified. Results: The expression of *COL1A1* and *ACTA2* was markedly reduced in LX-2 cells treated with miR-6133-5p. Interestingly, phosphorylation of c-Jun N-terminal kinase (JNK) was also significantly decreased by miR-6133-5p treatment. The expression of several predicted target genes of miR-6133-5p, including *TGFBR2* (which encodes Transforming Growth Factor Beta Receptor 2) and *FGFR1* (which encodes Fibroblast Growth Factor Receptor 1), was also reduced in miR-6133-5p-treated cells. The knockdown of *TGFBR2* by the corresponding small interfering RNA greatly suppressed the expression of *COL1A1* and *ACTA2*. Treatment with the JNK inhibitor, SP600125, also suppressed *COL1A1* and *ACTA2* expression, indicating that TGFBR2 and JNK mediate the anti-fibrotic effect of miR-6133-5p. The downregulation of *FGFR1* may result in a decrease of phosphorylated Akt, ERK (extracellular signal-regulated kinase), and JNK. Conclusion: miR-6133-5p has a strong anti-fibrotic effect, mediated by inactivation of TGFBR2, Akt, and JNK.

**Keywords:** liver fibrosis; hepatic stellate cells; TGF-β; JNK signaling pathway

#### **1. Introduction**

The progression of liver fibrosis often leads to fatal outcomes, such as the development of cirrhosis and hepatocellular carcinoma. Infections with viruses such as the hepatitis B virus (HBV) and hepatitis C virus are the major causes of liver fibrosis and contributed to around 50% of cirrhosis and hepatocellular carcinoma cases in 2017 [1]. Approximately 250 million people worldwide were infected with hepatitis viruses, resulting in nearly 1.4 million deaths in 2016, which is more than those caused by human immunodeficiency virus infection or tuberculosis [2]. Therefore, the development of novel treatment options to prevent the progression of liver fibrosis is important for reducing risks to health.

Hepatic stellate cells (HSCs) are major producers of extracellular matrix proteins, such as collagen fibers, during the development of fibrosis. HSCs are activated by fibrogenic cytokines, such as Transforming growth factor (TGF)-β, angiotensin II, and leptin, induced by liver injury [3]. Once activated, HSCs proliferate and differentiate into myofibroblasts and start to produce α-smooth muscle actin (α-SMA). Although extensive efforts have revealed that various signaling molecules such as Akt and c-Jun N-terminal kinase (JNK) control the activation and fibrogenesis of HSCs [4–8], the molecular processes involved in HSC activation are not entirely understood [9].

MicroRNAs (miRNAs) are small non-coding RNAs, 21–25 nucleotides in length, encoded in the human genome. Each miRNA targets hundreds of mRNAs and downregulates them post-transcriptionally by base pairing with their 3′ -untranslated regions (3′ -UTRs) [10,11]. Extensive studies have revealed that miRNAs regulate various biological and cellular processes, including proliferation, differentiation, cell behavior, and cancer development [12–17]. The involvement of miRNAs in fibrosis of the liver and other organs also has been reported [18–22].

Recently, by screening a human miRNA mimic library, we identified 39 miRNAs that effectively suppress HBV replication [23]. A significant portion of chronically HBV-infected patients suffer from progression of liver fibrosis toward cirrhosis [24]. Hence, we investigated whether any of these miRNAs have additional effects related to fibrosis and found that one of them, miR-6133-5p, potentially targets several fibrosis-related genes. Interestingly, miR-6133 greatly suppressed not only the *COL1A1* gene that encodes the α-chain of collagen type I (collagen Iα1), the major component of fibrous tissue in the liver, but also the *ACTA2* gene, which encodes α-SMA, indicating its anti-fibrotic potential. In the present study, we explored the molecular mechanisms by which miR-6133-5p suppresses the fibrotic activity of hepatic stellate cells.

#### **2. Results**

## *2.1. MiR-6133-5p Suppresses the Synthesis of* α*-Chain of Collagen Type I and* α*-Smooth Muscle Actin in LX-2 Cells*

To explore the role of miR-6133-5p in HSC functions, we transfected an miR-6133-5p mimic or a negative control miRNA mimic (hereafter referred to as miControl) into a human HSC line, LX-2, 24 h before treatment with 5 ng/mL of recombinant human transforming growth factor β1 (rhTGF-β1), a strong inducer of fibrogenesis. RNA and protein were collected at each time point, as indicated in Figure 1A. As shown in Figure 1B, in the miControl-treated cells, rhTGF-β1 treatment dramatically increased the expression of *COL1A1* and *ACTA2*. Interestingly, the levels of *COL1A1* and *ACTA2* were significantly decreased in the miR-6133-5p-treated cells, irrespective of rhTGF-β1 treatment, indicating that miR-6133 has strong anti-fibrotic property. Western blot analyses showed that the amounts of collagen Iα1 and α-SMA were also decreased in the miR-6133-5p-transfected cells, with or without rhTGF-β1 treatment (Figure 1C). It was noted that the amount of α-SMA protein increased only 72 h after rhTGF-β1 treatment.

β1 ( β1 encoding collagen Iα1, encoding α α collagen Iα1 α α **Figure 1.** The impact of miR-6133 on the anti-fibrotic activity of LX-2 cells. (**A**) LX-2 cells were transfected with a miR-6133-5p mimic or a negative control microRNA mimic (miControl) at a final concentration of 20 nM, 24 h before treatment with recombinant human transforming growth factor β1 (rhTGF-β1, 5 ng/mL). Total RNA and proteins were extracted from the cells collected at the time points indicated. (**B**) The expression of the *COL1A1* gene encoding collagen Iα1, and the *ACTA2* gene encoding α-smooth muscle actin was determined by RT-qPCR. (**C**) The amounts of α-chain of collagen type I (collagen Iα1) and α-smooth muscle actin (α-SMA) were determined by Western blot analysis. Error bars represent means ± standard deviations (*n* = 3). \* *p* < 0.05; \*\* *p* < 0.01., n.s. not significant.

To identify the molecular mechanisms by which miR-6133-5p suppressed *COL1A1* and *ACTA2*, we performed an RNAseq analysis to compare gene expression patterns of the cells transfected with miR-6133-5p with those transfected with miControl, 24 h after rhTGF-β treatment. Several genes (*n* = 373) were downregulated by more than 50%, with statistical significance (Figure 2A). Among them, 36 genes were also found among the putative target genes of miR-6133-5p predicted by TargetScanHuman v7.2 (total 438 genes, http://www.targetscan.org/vert\_72/) [11]. We then examined the role of each gene on the expression of *COL1A1* and *ACTA2* by transfecting with the corresponding small interfering RNAs (siRNA). As shown in Figure 2B, only knockdown of the *TGFBR2* gene, which encodes a component of the human TGF-β receptor, downregulated *COL1A1* and *ACTA2* more than 20%, compared with cells treated with a non-targeting control siRNA (siControl), with statistical significance. However, suppression of *COL1A1* and *ACTA2* by miR-6133-5p was also observed in the absence of rhTGF-β. Moreover, miR-6133 did not alter rhTGF-β-induced phosphorylation of Smad2/3, indicating the presence of TGFBR2-Smad independent pathways (Figure 3A).

By analyzing the RNAseq data in terms of gene ontology, we found that several genes encoding extracellular matrix proteins and genes involved in epithelial-to-mesenchymal transition, such as *CTGF* (connective tissue growth factor), *COL1A2* (collagen Iα2), *COL5A3* (collagen Vα3), *LOX* (Lysyl oxidase), *SNAI2* (Snail 2), and *CDH2* (Cadherin 2), were also significantly downregulated in the miR-6133-5p-treated group (Supplementary Tables S1 and S2). These results indicated that miR-6133-5p partially affects the activation and fibrotic function of LX-2 cells.

#### *2.2. MiR-6133 Decreased Phosphorylation of Akt, ERK, and JNK*

We then examined the impact of miR-6133 on the major cellular signaling pathways mediated by the serine/threonine kinases, Akt, ERK (extracellular signal-regulated kinase), JNK, and p38. Surprisingly, the amounts of phosphorylated forms of Akt, ERK, and JNK, but not p38, were smaller in the miR-6133-treated LX-2 cells than the miControl-treated cells, 24 h after rhTGF-β treatment (Figure 3A). Moreover, these differences were also observed in the mock-treated groups (Figure 3A). To determine whether the inhibition of phosphorylation of any of these kinases affected the expression of *COL1A1* and *ACTA2*, we used siRNAs targeting the *SMAD2*, *SMAD3*, and *SMAD4* genes, which are the central mediators of canonical TGF-β signaling and those targeting the *AKT1*, *AKT2*, and *AKT3* genes; and chemical inhibitors of MEK (MAPK/ERK kinase), JNK, and p38.

As shown in Figure 3B, treatment with SMAD2/3/4 siRNAs slightly decreased *COL1A1* and *ACTA2*, but without statistical significance, indicating that *COL1A1* and *ACTA2* are not solely regulated by the canonical TGF-β-Smad2/3/4 pathway. The knockdown of *AKT1*/*2*/*3* decreased the level of *COL1A1*, but not *ACTA2*. Treatment with the MEK inhibitor, U0126, which inhibits phosphorylation of ERK by the upstream kinase MEK, slightly increased the levels of *COL1A1* and *ACTA2*. In contrast, inhibition of JNK and p38 by their corresponding inhibitors (SP600125 and SB203580, respectively) significantly suppressed *COL1A1* and *ACTA2* expression (Figure 3B). Collectively, these results suggested that Akt may partially account for the suppression of *COL1A1* by miR-6133-5p and, similarly, JNK may partially account for the suppression of both *COL1A1* and *ACTA2* by miR-6133-5p.

We next investigated the relationship between the suppression of the *TGFBR2* gene and that of Akt, ERK, and JNK phosphorylation using an siRNA targeting the *TGFBR2* gene (siTGFBR2). As shown in Figure 4A, the amount of collagen Iα1 was greatly decreased and that of α-SMA was slightly decreased in the LX-2 cells treated with siTGFBR2 compared with the siControl-treated cells. While the amount of phosphorylated forms of Smad2/3 and Akt was also significantly decreased by siTGFBR2, it had no impact on the amounts of phosphorylated ERK, JNK, and p38 (Figure 4A). These results indicated that the suppression of JNK by miR-6133 is independent of TGFBR2.

**(A)**

(collagen Iα2) (collagen Vα3)

**Figure 2.** The impact of miR-6133 target genes on the anti-fibrotic activity of LX-2 cells. (**A**) RNAseq analysis revealed that 373 genes were downregulated in miR-6133-treated LX-2 cells compared with those treated with miControl, with statistical significance (*p* < 0.05). Thirty-six of these genes were also among 438 putative miR-6133-5p target genes predicted by TargetScanHuman v7.2. The levels of *TGFBR2* (which encodes Transforming Growth Factor Beta Receptor 2) in the miR-6133-treated and miControl-treated LX-2 cells (*n* = 3) were determined by RT-qPCR. (**B**) LX-2 cells were transfected with small interfering RNAs (siRNAs) corresponding to each of the 36 genes or a non-targeting control siRNA (siControl) at a final concentration of 20 nM, 24 h before treatment with rhTGF-β1 (5 ng/mL). The expression of *COL1A1* and *ACTA2* was determined by RT-qPCR. Fold change was calculated as the ratio over the expression in the siControl-treated sample (indicated by a dotted line). Error bars represent means ± standard deviations (*n* = 3). \* *p* < 0.05; \*\* *p* < 0.01.

β signaling

β1 (5

β treatment

**Figure 3.** The effect of miR-6133 on the activation of intracellular signaling pathways. (**A**) The amounts of phosphorylated-Smad2/3 (p-Smad2/3), total Smad2/3, p-Akt, Akt, p-ERK (extracellular signal-regulated kinase), ERK, p-JNK (c-Jun N-terminal kinase), JNK, p-p38, p38, and GAPDH (glyceraldehyde-3-phosphate dehydrogenase) were determined by western blot analysis, using the LX-2 extracts collected 24 h after rhTGF-β1 treatment. (**B**) LX-2 cells were treated with a mixture of siRNAs corresponding to the *SMAD2*, *SMAD3*, and *SMAD4* genes (siSMAD2/3/4), a mixture of siRNAs corresponding to the *AKT1*, *AKT2*, and *AKT3* genes (siAKT1/2/3), or siControl, at a final concentration of 30 nM (10 nM for individual siRNA), 24 h before treatment with rhTGF-β1 (5 ng/mL). LX-2 cells were also treated side by side with an MEK (MAPK/ERK kinase) inhibitor (U0126), a JNK inhibitor (SP600125), a p38 inhibitor (SB203580), or DMSO (dimethylsulfoxide), 24 h before rhTGF-β1 treatment. Total RNA was extracted from the samples collected 24 h after rhTGF-β1 treatment and the expression of COL1A1 and ACTA2 was determined by RT-qPCR analysis. Error bars represent means ± standard deviations (*n* = 3). \* *p* < 0.05; \*\* *p* < 0.01.

β1 β1 treatment β1. The **Figure 4.** The effect of knockdown of *TGFBR2* and *FGFR1* (which encodes Fibroblast Growth Factor Receptor 1), on the activation of intracellular signaling pathways. (**A**) LX-2 cells were treated with an TGFBR2 siRNA or siControl at a final concentration of 20 nM, 24 h before treatment with rhTGF-β1 (5 ng/mL). Protein extracts were prepared from the samples collected 24 h after rhTGF-β1 treatment and subjected to Western blot analysis. (**B**) LX-2 cells were treated with an FGFR1 siRNA, TGFBR2 siRNA, or siControl at a final concentration of 20 nM, 24 h before treatment with rhTGF-β1. The lysates were analyzed by Western blot analysis. (**C**) Values of fragments per kilobase of exon per million reads mapped (FPKM) for *FGF1*, *FGF2*, *FGF5*, *FGFR1*, *HBEGF*, *VEGFA*, *IRS1*, *PIK3CD*, and *PIK3R2* genes, deduced from the RNAseq analysis comparing samples treated with miR-6133-5p and miControl. Error bars represent means ± standard deviations (*n* = 3). \* *p* < 0.05; \*\* *p* < 0.01.

#### *2.3. Possible Involvement of the Fibroblast Growth Factor Receptor 1 (FGFR1) Gene in the Suppression of JNK Phosphorylation by MiR-6133-5p*

Next, to further elucidate the mechanism by which miR-6133-5p suppressed the phosphorylation of Akt, ERK, and JNK, we performed gene set enrichment analysis (GSEA) and constructed miRNA-mRNA networks from RNAseq data. By analyzing GSEA of gene ontology (GO) gene sets in the miR-6133-5p-treated LX-2 cells, we found that a group of genes annotated as 'extracellular matrix organization' were significantly downregulated in the miR-6133-5p-treated cells (Supplementary Figure S1A,B). From the putative miR-6133-target genes predicted by TargetScanHuman v7.2 and experimentally validated miR-6133-5p target genes listed in miRTarBase v8.0 (http://mirtarbase.cuhk. edu.cn/php/index.php) [25], we selected 98 genes whose expression was significantly lower in the miR-6133-5p-treated LX-2 cells (fold change in log2 ratio > 0.8, *p* < 0.05, fold discovery rate < 0.01; Supplementary Figure S1C). Among them, we found the *FGFR1* gene annotated in 'extracellular matrix organization', which is known to transmit a signal to the PI3K-Akt and ERK/JNK/p38 signaling pathways upon activation by its ligands, fibroblast growth factors [26]. As shown in Figure 4B, knockdown of *FGFR1* by the corresponding siRNA significantly decreased the phosphorylated form of JNK. It also decreased phosphorylated forms of Akt and ERK to some extent. These results suggested that the anti-fibrotic function of miR-6133-5p may partially be mediated by the FGFR-Akt/ERK/JNK axis. In contrast, knockdown of *FGFR1* increased the amount of phosphorylated form of Smad2/3 (Figure 4B), indicating that FGFR1 acts inhibitory to the TGF-β pathway, including Smad2/3, as reported by Li et al. [27]. Interestingly, the levels of expression of genes encoding FGFR ligands (*FGF1*, *FGF2*, and *FGF5*) were also reduced in the miR-6133-5p-treated cells. The level of genes involved in other growth factor signaling pathways, such as *HBEGF* (the epidermal growth factor signaling pathway), *VEGFA* (the vascular endothelial growth factor signaling pathway), and *IRS1*, *PIK3CD*, *PIK3R2* (the IGF-PI3K signaling pathway), were also significantly decreased (Figure 4C).

#### **3. Discussion**

Recently, we reported that the human miRNA, miR-6133-5p, has strong antiviral activity against HBV replication [23]. In this study, we found that miR-6133-5p effectively suppressed *COL1A1* and *ACTA2*—the main component of fibrous tissue in the liver and a representative marker of HSC activation, respectively. An RNAseq analysis also revealed that several genes encoding other extracellular matrix proteins and those involved in epithelial-to-mesenchymal transition were significantly downregulated in the miR-6133-5p-treated cells, suggesting that miR-6133-5p has strong, but partial, anti-fibrotic property when introduced in HSCs. Functional analyses revealed that siRNA-mediated knockdown of *TGFBR2* and *AKT1*/*2*/*3*, and inhibition of JNK by an appropriate chemical, suppressed *COL1A1* and *ACTA2* expression, suggesting that the anti-fibrotic effects of miR-6133-5p may be mediated by TGFBR2, Akt, and JNK (Figure 5).

β signaling mediated by the **Figure 5.** Schema of anti-fibrotic mechanisms via miR-6133-5p. TGF-β signaling mediated by the receptor TGFBR2 and its receivers, Smad2/3, control, *COL1A1*, and *ACTA2*. miR-6133-5p directly suppresses the expression of *TGFBR2*. miR-6133-5p decreases phosphorylation (activation) of Akt, ERK, and JNK, possibly by targeting *FGFR1* which results in the suppression of FGF/FGFR axis. Akt and JNK regulate *COL1A1* and *ACTA2* expression.

The decrease of phosphorylated Akt, ERK, and JNK in the miR-6133-treated LX-2 cells could be due to the downregulation of *FGFR1*, a target gene of miR-6133-5p. Several reports have shown that Akt is involved in collagen synthesis and the activation of HSCs [4,5]. JNK has also been reported to regulate collagen synthesis and the activation of HSCs [6–8]. A chemical compound, GS-444217, that specifically inhibit ASK1 (Apoptosis signal-regulating kinase 1), a protein kinase upstream of JNK and p38, has been reported to reduce liver fibrosis in a mouse model with a *Nlrp3* (NLR family pyrin domain containing 3) loss-of-function mutation [28]. These findings, together with our present study, indicate that signaling pathways including Akt and JNK are therapeutic targets for the control of liver fibrosis.

2 cells were cultured in Dulbecco's modified Eagle's medium (Thermo Fisher Scientific, The roles of endogenous miR-6133-5p in humans are largely unknown. miR-6133-5p is expressed in almost all tissues, including the liver. One report revealed that the amount of urinary exosomal miR-6133-5p was increased in type II diabetic nephropathy patients [29]. On the other hand, we found that the level of miR-6133-5p was not altered in the rhTGF-β1-treated LX-2 cells, compared with the control (data not shown), indicating that endogenous level of miR-6133 have no impact on the fibrogenic function of HSCs.

miR-6133-5p is found in the genome of several primates. Although it is important to determine whether miR-6133-5p effectively ameliorates liver fibrosis in vivo, at present, we could not employ physiologically relevant small animal models for evaluating the effect of primate-restricted miRNA such as miR-6133-5p on liver fibrosis in vivo. Chimeric mice, with liver repopulated with human hepatocytes, were frequently used to study HBV replication in vivo [30]. The development of similar experimental animal models such as small animals harboring human HSCs in the liver would help to examine the effect of reagents on liver fibrosis in future studies.

Lipofectamine RNAiMAX Reagent (Thermo Fisher Scientific), according to the manufacturer's β1. The cells were then harvested and subjected to RNA and protein Some issues remain to be addressed. While the expression of *ACTA2* was induced by rhTGF-β1 and peaked 24 h after treatment, the increase of α-SMA protein became obvious only 72 h after rhTGF-β1 treatment (Figure 1B,C). Similarly, while knockdown of *TGFBR2* greatly suppressed *ACTA2* expression 24 h after rhTGF-β1 treatment, the amount of α-SMA protein was not decreased so much (Figures 2B and 4A). This could be due to the balance between the efficiency of translation and the degradation of α-SMA protein, which may cause a delay in the outcome of the increase/decrease of *ACTA2* mRNA and changes in the amount of its protein product.

On the other hand, knockdown of *TGFBR2* by the corresponding siRNA greatly decreased the phosphorylated form of Smad2/3. However, although miR-6133-5p treatment also suppressed *TGFBR2* effectively, it had no impact on the amount of phosphorylated Smad2/3. The amount of phosphorylated Smad2/3 was increased by the knockdown of *FGFR1* (Figure 4B); hence, it is possible that the downregulation of *FGFR1* by miR-6133 may partially cancel the suppressive effect by the downregulation of *TGFBR2* in terms of the level of phosphorylation of Smad2/3.

Our results showed that the inhibition of p38 by its inhibitor, SB203580, also effectively suppressed *COL1A1* and *ACTA2* expression (Figure 3B). On the other hand, miR-6133-5p had no impact on the phosphorylation of p38 in LX-2. JNK and p38 are differently regulated by upstream kinases (MKK4 (mitogen-activated protein kinase kinase 4)/MKK7 and MKK3/MKK6, respectively); hence, miR-6133-5p could selectively inhibit JNK without affecting p38 and it is sufficient for the suppression of *COL1A1* and *ACTA2* in LX-2 cells.

A comprehensive transcriptome analysis covering many time points is needed in a future study to dissect the effect of miR-6133-5p from the immediate-early suppression of direct target genes, middle-stage changes of signaling pathways, and late-stage changes, such as the downregulation of *COL1A1* and *ACTA2*. The role of miR-6133-5p in the fibrosis of other organs or tissues has also not been documented; further studies will be required to examine the effect of miR-6133-5p in various fibroblast cells of different organ origins.

In conclusion, we found that miR-6133-5p has strong anti-fibrotic effect which could be mediated by inactivation of TGFBR2, Akt, and JNK.

#### **4. Materials and Methods**

#### *4.1. Cell Culture and Transfection*

LX-2 cells were cultured in Dulbecco's modified Eagle's medium (Thermo Fisher Scientific, Waltham, MA, USA) supplemented with 1% antibiotic antimycotic solution and 2% heat-inactivated fetal bovine serum (Thermo Fisher Scientific). MiRIDIAN MicroRNA miR-6133-5p Mimic and MiRIDIAN MicroRNA Mimic Negative Control #1 (miControl) were purchased from Horizon Discovery Group plc. (Cambridge, UK). ON-TARGETplus siRNA SMARTpools targeting *SMAD2*, *SMAD3*, *SMAD4*, *AKT1*, *AKT2*, *AKT3*, and *FGFR1*; and ON-TARGETplus Non-targeting Pool (siControl) were purchased from Horizon Discovery Group plc. Each SMARTpool contains four independent siRNA molecules to ensure efficient gene knockdown. Silencer Select siRNA reagents corresponding to 36 predicted target genes of miR-6133-5p and Silencer Select negative control siRNA were purchased from Thermo Fisher Scientific. Two independent Silencer Select siRNA molecules for each gene were mixed equally and used in the following assay to ensure efficient gene knockdown. The sequence information of the siRNAs used in this study was shown in Supplementary Table S4.

miRNAs or siRNAs were introduced into LX-2 cells at a final concentration of 20 nmol/L using Lipofectamine RNAiMAX Reagent (Thermo Fisher Scientific), according to the manufacturer's instructions. Twenty-four hours after transfection, the cells were further cultured with medium containing 5 ng/mL of rhTGF-β1. The cells were then harvested and subjected to RNA and protein extraction at the time points indicated in Figure 1A.

#### *4.2. Gene Expression Analysis by RT-qPCR*

Total RNA was extracted using ISOGEN (Nippon Gene, Toyama, Japan). Gene expression was determined by RT-qPCR using StepOne Plus (Thermo Fisher Scientific). The primer–probe sets for RT-qPCR analysis of human *COL1A1*, *ACTA2*, *TGFBR2*, *FGFR1*, and *GAPDH* genes were purchased from Thermo Fisher Scientific.

#### *4.3. Quantification of Protein*

Cell lysates were prepared using a lysis buffer containing 1% Nonidet P40, 150 mM sodium chloride, and 50 mM Tris-Cl buffer (pH 7.4). A cOmplete Mini EDTA-free tablet and a PhosSTOP tablet (Roche diagnostics, Basel, Switzerland) were added to each 10 mL of the lysis buffer immediately before use. Western blot analyses were performed by a routine procedure using the primary antibodies listed below with the species, target, company, catalogue number, and dilution: mouse anti-collagen Iα1 (sc-293182, Santa Cruz, Dallas, TX, USA, 1:1000); rabbit anti-α-SMA (GTX100034, GeneTex, Irvine, CA, USA, 1:1000); rabbit anti-phospho-Smad2/3 (#8828, Cell Signaling Technology, Danvers, MA, USA, 1:2000); rabbit anti-Smad2/3 (#8685, Cell Signaling Technology, 1:2000); rabbit anti-phospho-Akt (#9271, Cell Signaling Technology, 1:2000); rabbit anti-Akt (#9272, Cell Signaling Technology, 1:2000); mouse anti-phospho-ERK (#9106, Cell Signaling Technology, 1:2000); rabbit anti-ERK (#9102, Cell Signaling Technology, 1:2000); rabbit anti-phospho-JNK (#9251, Cell Signaling Technology, 1:2000); rabbit anti-JNK (#9252, Cell Signaling Technology, 1:2000); rabbit anti-phospho-p38 (#9211, Cell Signaling Technology, 1:2000); rabbit anti-p38 (#9212, Cell Signaling Technology, 1:2000); mouse anti-GAPDH (ab8245, Abcam, Cambridge, UK, 1:10,000). The intensity of the bands was calculated using ImageJ v1.8.0 (https://imagej.nih.gov/ij/index.html).

#### *4.4. RNAseq Analysis*

Total RNA samples collected from miR-6133-5p- and miControl-treated LX-2 cells (*n* = 2) 24 h after rhTGF-β1 treatment were subjected to an RNAseq analysis. PolyA + RNA was extracted, fragmented, and reverse-transcribed to yield a single-stranded cDNA mixture. Double-stranded DNA was then synthesized using the cDNA mixture as a template. The ends of the product were blunted, phosphorylated, followed by addition of 3′ -deoxyadenosine, and ligated with adapter DNA fragments containing an index sequence unique to each sample. After amplification by PCR, the resultant sequencing libraries were subjected to pair-end sequencing (sequence length = 150 bases) using NovaSeq 6000, NovaSeq 6000 S4 Reagent Kit, and NovaSeq Xp 4-Lane Kit (Illumina Inc., San Diego, CA, USA). The reads were mapped and annotated using GeneData Profiler Genome v11.0.4a (GeneData, Basel, Switzerland) and STAR v2.5.3a (https://github.com/alexdobin/STAR). *Homo sapiens* genome assembly GRCh37 (hg19) was used as the reference. The number of reads and the percentage of mapped reads for each sample are shown in Supplementary Table S3. Read counts underwent the trimmed mean of M values (TMM) normalization and log2 computes counts per million (CPM) transformation using the edgeR software v3.30.3 [31]. Differences in gene expression between miR-6133-5p and miControl were tested by a quasi-likelihood test function (glmQLFit). We set a false-discovery rate (FDR) threshold of 0.01 to correct for multiple testing and set a log-fold change (Log2FC) threshold of 0.8. The RNAseq data were deposited in the Gene Expression Omnibus database (accession number: GSE158478, https://www.ncbi.nlm.nih.gov/geo/).

To functionally characterize miR-6133-5p, we performed a pathway analysis using the GSVA R package (https://www.bioconductor.org/packages/release/bioc/html/GSVA.html) [32]. The gene sets used were the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway and all GO gene sets from the Broad Institute's Molecular Signatures Database (MSigDB) v7.1. The top differentially enriched pathways were yielded along with *p*-values adjusted for multiple testing correction using the Benjamini–Hochberg FDR controlling procedure. Cytoscape software v3.6.2 [33] was employed to construct the miRNA–mRNA gene network. All data were analyzed in R (http://www.r-project.org/).

#### *4.5. Statistical Analysis*

The student's *t*-test was performed using Microsoft Excel. Data are depicted as the mean ± standard deviation, and *p*-values < 0.05 were considered significant: \* *p* < 0.05, \*\* *p* < 0.01.

**Supplementary Materials:** Supplementary materials can be found at http://www.mdpi.com/1422-0067/21/19/ 7251/s1. Table S1: Downregulated genes listed in the gene set 'Extracellular matrix'; Table S2: Downregulated

genes listed in the gene set 'Hallmark epithelial mesenchymal transition'; Table S3: The RNAseq Results; Table S4: Sequence information of the siRNAs used in this study; Figure S1: Characterization of genes controlled by miR-6133-5p.

**Author Contributions:** Conceptualization, Y.T.; investigation, S.H.-T., M.O., K.M., M.I., and K.K.; resources, Y.S.; writing—original draft preparation, S.H.-T.; writing—review and editing, Y.T.; project administration, Y.T.; funding acquisition, Y.T. All authors have read and agreed to the published version of the manuscript.

**Funding:** This study was partly supported by the Research Program on Hepatitis from the Japan Agency for Medical Research and Development (AMED), Grant Numbers JP20fk0210048 and JP20fk0310101.

**Acknowledgments:** We are grateful to Kyoko Ito for gene expression analyses and Mayumi Hojo for Western blot analyses.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **Abbreviations**


#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## *Review* **Co-Occurrence of Hepatitis A Infection and Chronic Liver Disease**

## **Tatsuo Kanda \* , Reina Sasaki, Ryota Masuzaki, Hiroshi Takahashi, Taku Mizutani, Naoki Matsumoto , Kazushige Nirei and Mitsuhiko Moriyama**

Division of Gastroenterology and Hepatology, Department of Medicine, Nihon University School of Medicine, 30-1 Oyaguchi-kamicho, Itabashi-ku, Tokyo 173-8610, Japan; sasaki.reina@nihon-u.ac.jp (R.S.); masuzaki.ryota@nihon-u.ac.jp (R.M.); hiroshi.t.215@gmail.com (H.T.); mattakunotaku1981@yahoo.co.jp (T.M.); matsumoto.naoki@nihon-u.ac.jp (N.M.); nirei.kazushige@nihon-u.ac.jp (K.N.); mizutani.taku@nihon-u.ac.jp (M.M.)

**\*** Correspondence: kanda.tatsuo@nihon-u.ac.jp; Tel.: +81-3-3972-8111

Received: 10 August 2020; Accepted: 1 September 2020; Published: 2 September 2020

**Abstract:** Hepatitis A virus (HAV) infection occasionally leads to a critical condition in patients with or without chronic liver diseases. Acute-on-chronic liver disease includes acute-on-chronic liver failure (ACLF) and non-ACLF. In this review, we searched the literature concerning the association between HAV infection and chronic liver diseases in PubMed. Chronic liver diseases, such as metabolic associated fatty liver disease and alcoholic liver disease, coinfection with other viruses, and host genetic factors may be associated with severe hepatitis A. It is important to understand these conditions and mechanisms. There may be no etiological correlation between liver failure and HAV infection, but there is an association between the level of chronic liver damage and the severity of acute-on-chronic liver disease. While the application of an HAV vaccination is important for preventing HAV infection, the development of antivirals against HAV may be important for preventing the development of ACLF with HAV infection as an acute insult. The latter is all the more urgent given that the lives of patients with HAV infection and a chronic liver disease of another etiology may be at immediate risk.

**Keywords:** HBV; HCV; HIV; acute liver failure; nonalcoholic fatty liver diseases; NASH; GRP78

#### **1. Introduction**

Liver failure is a common disease with high mortality, and its incidence is increasing with the use of alcohol and the prevalence of obesity and diabetes [1–3]. It has also been reported that the prognosis of acute hepatitis or acute liver injury was affected by the preexistence of chronic liver diseases and cirrhosis [1,2], extrahepatic diseases, such as metabolic, malignant, and psychiatric diseases [4], and host factors, such as older age and obesity [3,5,6], although the etiology of acute insults is one of the most important risk factors for the development of severe liver diseases [1,7].

Hepatitis A virus (HAV) infection is still one of the major causes of acute hepatitis worldwide. HAV infection occasionally causes acute liver failure [4,8]. It has been reported that a superinfection of HAV in patients with a chronic hepatitis C virus (HCV) infection is associated with fulminant hepatitis [9], although much research denies this association [5,10]. HAV infection rarely causes acute liver failure in patients without underlying chronic liver diseases [9].

There are excellent, safe, and effective HAV vaccines to prevent HAV infection. However, HAV vaccination costs a lot. As no universal vaccination program against HAV infection exists in certain countries, such as Japan, it may be important to develop potential drugs against HAV infection [11].

In this review, we searched the recent literature concerning the association between HAV infection and chronic liver diseases, including metabolic associated fatty liver disease (MAFLD), in PubMed. We also discussed the mechanism of severe acute hepatitis A.

#### **2. Acute-On-Chronic Liver Failure with HAV Infection as an Acute Insult**

Acute-on-chronic liver diseases include acute-on-chronic liver failure (ACLF) and non-ACLF [12]. ACLF, which presents acutely with multiple organ failure and is precipitated by an acute insult, has high short-term mortality [2,13]. In general, the prognosis of ACLF is worse than that of acute liver failure. ACLF is a distinct concept, where acute hepatic decompensation occurs in patients with chronic liver disease or cirrhosis in encountering an acute insult, leading to high short-term mortality [2]. In Asian countries, hepatitis viruses are important factors of acute insults, unlike in European countries and the United States [2], and HAV is one of the acute insults of ACLF [1,12,14–17].

HAV superinfection was found to be the most common etiology (42%) of acute deterioration in children with ACLF in India [15]. ACLF in adults was found to be due to HEV, HAV, or both in 61%, 27%, and 6% of cases [1], respectively, although HAV infections occur in childhood, and HAV infection as an acute insult in adult ACLF is relatively uncommon in India [17]. Agrawal et al. reported an adult patient with ACLF and HAV as an acute insult who had an underlying cirrhotic liver due to nonalcoholic steatohepatitis (NASH) [17]. Among the children and adults with ACLF, acute insults caused by both HAV and HEV are important. It may be important to consider them in order to improve the prognosis of ACLF by developing a treatment for HAV infection.

#### **3. HAV Infection and Metabolic Associated Fatty Liver Disease (MAFLD)**

ACLF may occur among patients with chronic liver diseases or cirrhosis due to nonalcoholic fatty liver diseases (NAFLD), including NASH and alcoholic liver diseases (ALD), in eastern and western countries [2,13]. NASH is the most rapidly increasing etiology for ACLF [18]. Agrawal et al. reported a nonobese 34-year-old man presenting ACLF with acute HAV infection superimposed on NASH without cirrhosis [17] (Table 1). Kahraman et al. also reported a human immunodeficiency virus (HIV)-positive case presenting ACLF with acute HAV infection superimposed on cirrhosis due to NASH [19]. NASH is also observed among people less than 40 years old, and acute-on-chronic liver diseases may have an atypical course among these patients [20].



CLD, chronic liver diseases; HIV, human immunodeficiency virus; ALD, alcoholic liver disease; HCV, hepatitis C virus.

Fatty liver diseases associated with metabolic dysfunction are common and have a heterogeneous genetic predisposition, metabolic syndrome, and environmental factors [23]. Recently, experts suggested "MAFLD" should replace NAFLD/NASH [23]. The diagnosis of MAFLD is based on the detection of liver steatosis in the presence of overweight or obesity, diabetes mellitus, and/or clinical evidence of metabolic abnormalities, such as hypertension, dyslipidemia, and hyperglycemia.

A Japanese nationwide survey of ALF and late-onset hepatic failure (LOHF) caused by HAV infection suggested that diabetic mellitus was more common among deceased patients than among rescued patients (29% vs. 8%; *p* < 0.05), excluding patients with liver transplantations, and that diabetic

mellitus was independently associated with the outcome [24]. Patients with diabetes are at risk of developing severe hepatitis [25].

We observed that HAV HA11-1299 genotype IIIA strain replication is enhanced by the accumulation of lipids or high-concentration glucose in the human hepatoma cell line, Huh7 [26–28]. Hyperglycemia or the accumulation of lipids induces an endoplasmic reticulum (ER) stress response in human hepatocytes. HAV replicates in the ER of human hepatocytes and induces an ER stress response. The ER stress response is mediated by the sensor molecules, inositol-requiring enzyme 1α (IRE1α), PKR-like ER kinase (PERK), and activating transcription factor 6 (ATF6), which are usually associated with molecular chaperone glucose-regulated protein 78 (GRP78) [27]. GRP78 is a negative regulator of ER stress response. We also observed that the overexpression of GRP78 could inhibit HAV replication, while the knockdown or knockout of GRP78 enhanced HAV replication [26,28]. In sum, GRP78 is an antiviral protein against HAV replication [28].

#### **4. HAV Infection and Alcoholic Liver Diseases (ALD)**

There are several factors affecting the severity of HAV infection and the rates of fulminant hepatic failure [29]. These important factors include older age, concomitant virus infection, chronic liver disease, sexual orientation, intravenous drug use, and alcohol abuse [6,29]. Feller et al. reported that 12 patients developed hepatic encephalopathy, ascites, or both, among 20 patients with alcoholic cirrhosis and a superimposed episode of acute viral hepatitis [30]. HAV infection was excluded in only three of these patients [30].

Lefilliatre et al. reported that three patients with fulminant hepatitis A had preexisting liver diseases, and one of the three had biopsy-proven alcoholic cirrhosis [21] (Table 1). Spada et al. reported that two individuals were HCV-coinfected alcohol abusers, had underlying liver cirrhosis, and died of acute liver failure due to HAV infection [22] (Table 1).

While the direct effects of alcohol on HAV replication is unknown, excess alcohol intake (binge drinking) could induce hepatic fibrosis. As only alcohol intake is responsible for worsening ACLF with alcoholic chronic liver diseases and alcoholic cirrhosis [31], HAV may have an additive responsibility for worsening ACLF with ALD.

#### **5. Coinfection of HAV with HIV**

In Japan, where no universal vaccination programs against HAV infection exist, 10–20% of those with HIV infection tested positive for immunoglobulin G (IgG) anti-hepatitis A (HA) antibodies [32–35]. This prevalence is similar to that of IgG anti-HA in those without HIV infection [36,37], although a higher prevalence area can also be observed in Japan [38]. In general, individuals of high-risk groups, such as healthcare workers, sewage workers, and drug addicts, have ~60% of IgG anti-HA worldwide [39,40]. The seroprevalence of IgG anti-HA is relatively higher in people living with HIV worldwide [41–43].

HIV infection has also been reported as a cause of liver damage in patients infected with HIV [40]. Thus, it is as important to consider patients with HIV infection as those with chronic liver disease. Not only chronic viral hepatitis B or C but also drug-induced liver injury induced by the antiretroviral drugs, NAFLD and ALD, has also been observed in people with HIV [40].

HAV infection in patients with chronic liver diseases and coinfected with HIV are shown in Table 2 [21,22,44,45]. Prolonged HAV infection was also reported in an HIV-seropositive patient [44]. It was reported that the recovery of immunity through recently developed anti-HIV therapies may lead to more severe hepatocellular damage in patients with HAV infection [45].

HAV infects humans through fecal–oral routes, when HAV-contaminated water and food are consumed. Among men who have sex with men (MSM), HAV is sexually transmitted [46], and HAV outbreaks have been observed [47–55]. It is noteworthy that acute hepatitis A among MSM is one of the male-dominant diseases, although, in general, no gender difference exists in patients with an HAV infection caused by HAV-contaminated water and food. While HAV may cause severe hepatitis in

people living with HIV, two doses of an HAV vaccine are more effective for them to achieve a sustained HAV seroresponse than a single dose of an HAV vaccine [56].


**Table 2.** Coinfection with hepatitis A virus (HAV) and human immunodeficiency virus (HIV).

CLD, chronic liver diseases; HBV, hepatitis B virus; HCV, hepatitis C virus.

#### **6. Coinfection of HAV with HBV**

Several cases of ACLF with HAV as an acute insult and chronic hepatitis or cirrhosis due to HBV, as well as cases with a superinfection of HAV in patients with HBV, have been reported (Table 3) [9,21,57–62]. A superinfection of HBsAg carriers with HAV seems not to cause more severe conditions [57]. Patients with HBV plus HAV infection had a less advanced baseline liver disease and a better prognosis than those with HBV plus hepatitis E virus infection [60].

**Table 3.** Acute-on-chronic liver failure and/or superinfection of hepatitis virus (HAV) in patients with hepatitis B virus (HBV).


CLD, chronic liver diseases; HCV, hepatitis C virus; HIV, human immunodeficiency virus; HBeAg, hepatitis B virus e antigen.

Vento et al. reported that, among 10 patients with an acquired HAV superinfection and chronic HBV infection, one (10%), who had cirrhosis, had marked cholestasis [9]. Pramoolsinsap et al. evaluated acute superinfection with HAV in 20 HBV asymptomatic carriers and fulminant hepatitis or submassive hepatitis in 11 (55%) of 20 HBsAg carriers [63]. A superinfection of HAV in patients with HBV occasionally leads to critical conditions in HBV carriers with or without cirrhosis, although patients with advanced fibrosis or cirrhosis are more susceptible to severe conditions [9,21,57–63].

A total of 310,746 cases with acute hepatitis A were observed during the Shanghai hepatitis A epidemic [58]. A total of 47 fatal cases (0.015%) were reported. Fatality rates were 0.05% (15/27,346) and 0.009% (25/283,400) in patients with or without HBV infection, respectively. It is worth noting that there were 5.6-fold greater fatality rates in patients with HBV infection than in those without [58]. Cooksley et al. reported that patients infected with HBV who have raised ALT levels and high HBV levels have a higher risk of liver failure following HAV superinfection [58]. HAV vaccination seems to be effective in preventing liver failure associated with HAV in patients with or without HBV infection [64–67]. However, HAV vaccination may not be necessary in the case of countries in which HAV is endemic, such as India [68,69].

It has been reported that the transient suppression of HBV replication and the disappearance of HBV DNA with the seroconversion of HBeAg were observed in several cases of double infections with HAV and HBV carries [57]. Beisel et al. also reported that an HBsAg carrier case with HAV superinfection presented the seroconversion of HBsAg, suggesting that unspecific immunological responses to HAV could lead to a functional cure of HBV [62]. It was reported that the sharp peak in interferon-gamma production induced by a superinfection of HAV may lead to the suppression of HBV replication in patients with chronic hepatitis B [70]. This peak in interferon-gamma production occurred just before the rise in serum transaminase activity, resulting in a decrease in HBV DNA and HBeAg.

Berthillon et al. infected the human hepatoma cell line, PLC/PRF/5 [71], which integrates HBV DNA and produces HBsAg, with the HAV CF53 strain [72]. The inhibition of HBsAg production in PLC/PRF/5 cells infected with HAV was observed, compared with those without HAV infection, demonstrating that HAV interferes with the expression of HBsAg from hepatocytes harboring integrated HBV DNA sequences [71]. We also infected HepG2.2.15, which produces HBV virion or HepG2 cells, with the HAV HA11-1299 strain. We demonstrated that the HAV replication is similar between HepG2.2.15 and HepG2, 96 h after HAV infection. However, HBV replication is inhibited in HAV-infected HepG2.2.15, compared to HepG2.2.15 without HAV infection [73].

We also observed that the replication of both HAV and HBV is suppressed in human hepatocyte PXB cells superinfected with HAV and HBV, compared to those mono-infected with HAV or HBV [73]. Thus, HAV infection seems to inhibit HBV replication. Further studies are required to support this point, although it indicates that the existence of cirrhosis or advanced liver fibrosis should cause severe hepatitis in the superinfection of HAV in patients with HBV.

#### **7. Coinfection of HAV with HCV**

In general, HCV is a rare cause of fulminant hepatitis or acute liver failure [74,75]. We did not identify any cases of fulminant hepatitis with HCV RNA in 82 cases of fulminant hepatitis and late-onset hepatic failure from 1986 to 2001, which were examined at Chiba University School of Medicine, Japan [74]. There were several reports that HAV infection in patients with chronic hepatitis C is associated with increased mortality [9,21,22,76], although several contrary opinions exist [59,77] (Table 4).



CLD, chronic liver diseases; HBV, hepatitis B virus; ALD, alcoholic liver disease; HIV, human immunodeficiency virus.

Vento et al. reported that, among 17 patients with an acquired HAV superinfection with chronic hepatitis C, seven patients (41.2%) possessed fulminant hepatic failure, and six (85.7%) of those seven patients died [9]. It is interesting to note that antinuclear antibodies, anti-smooth-muscle antibodies, and/or anti-asialoglycoprotein receptor antibodies were detected in five of seven patients with fulminant hepatitis (71.4%) [9]. Moreover, six of these seven patients possessed chronic active hepatitis, and one patient recovered from fulminant hepatitis and was treated with methylprednisolone [9]. There are some reports indicating a higher fatality rate of HAV superinfection in patients with chronic HCV infection, not considering those with or without cirrhosis [21]. However, it is unclear whether the high fatality rates were due to severe underlying liver damage or not [21,22].

It was reported that the superinfection of HAV is associated with decreased HCV replication, which may lead to a clearance of HCV [77,78]. Esser-Nobis et al. found that Huh7-Lunet cells supported HAV and HCV replication with similar efficacy and limited interference with each other [79].

In fact, as several severe hepatitis A cases have been observed in patients with chronic HCV infection, clinicians should pay attention to HAV infection in HCV-infected individuals [80]. At present, although direct-acting antivirals against HCV can lead to a higher sustained virological response with less adverse events, no effective HCV vaccines are available. Thus, HAV vaccination should be considered for HCV-infected patients, especially those with cirrhosis or advanced fibrosis [81–88].

#### **8. HAV and Other Chronic Liver Diseases**

It was reported that a prospective study of 31 children in the age group of 1–16 years, who fulfilled the criteria for ACLF of the Asian Pacific Association for the Study of the Liver (APASL) 2008 consensus, found 13 ACLF cases of HAV as an acute insult and autoimmune hepatitis or Wilson disease as causes of chronic liver disease [15]. In children, acute-on-chronic liver diseases, HEV, and HAV are more frequently causes of acute insults and Wilson disease, while autoimmune liver disease and primary sclerosing cholangitis are more frequently causes of chronic liver disease [12]. It is possible that HAV infection, as an acute insult, could result in ACLF in patients with any chronic liver disease, especially cirrhosis. Careful attention should also be paid to HAV infection in adults and children who have certain chronic liver diseases.

#### **9. Host Genetic Factors in HAV Infection**

Acute insults in ACLF are different, depending on the country in which they are found [2]. In Asian countries, European countries, and the United States, hepatic, hepatic, and extrahepatic or infection (extrahepatic) causes, respectively, are representative acute insults in the definition of the APASL, EASL, and NACSELD ACLF guidelines [2,89–91]. Of course, not only a sanitary environment but also host genetic factors are different in these different regions. Among Mexican Americans, transforming growth factor beta 1 (TGFB1) rs1800469 (adjusted odds ratio (OR), 1.38; 95% confidence interval (CI), 1.14–1.68; P value adjusted for false discovery rate (FDR-P) = 0.017) and X-ray repair cross complementing 1 (XRCC1) rs1799782 (OR, 1.57; 95% CI, 1.27–1.94; FDR-P = 0.0007) were associated with an increased risk of HAV infection [92]. ATP-binding cassette subfamily B member 1 (ABCB1) rs1045642 (OR, 0.79; 95% CI, 0.71–0.89; FDR-P = 0.0007) was associated with a decreased risk [92]. Host genetic factors may also play an important role in determining the differential susceptibility to HAV infection [92–94].

#### **10. Prevention of HAV Infection in Patients with Chronic Liver Diseases**

#### *10.1. HAV Vaccination*

HAV vaccination may be important for patients with chronic liver diseases, especially those with cirrhosis [81–88]. While a universal vaccination program against HAV seems to be the most effective solution for the prevention of HAV infection, it may be difficult to carry out this program worldwide due to the high costs of HAV vaccine production and its low effectiveness in certain countries in which the infection is endemic [88,95]. HAV vaccination targeting certain populations may also be effective and important in this regard [96]. Antivirals against HAV infection may also be needed (Figure 1). The unknown causes for chronic injury constitute only 5–15% of cases of ACLF [2].

**Figure 1.** Effects of hepatitis A infection (HAV) on the prognosis of chronic liver disease. Possible acceleration and inhibition of the disease progression of hepatitis A are indicated by red and blue arrows, respectively. MAFLD, metabolic associated fatty liver disease; ALD, alcoholic liver disease; HIV, human immunodeficiency virus; HBV, hepatitis B virus; HCV, hepatitis C virus.

#### *10.2. Japanese Rice-Koji Miso Extracts and Zinc Sulfates Could Inhibit HAV Replication with the Enhancement of GRP78 Expression*

Japanese rice-koji miso extracts enhanced GRP78 expression and inhibited HAV HA11-1299 genotype IIIA strain replication in the human hepatocytes, Huh7 and PXB cells [97]. We investigated the effect of miso extracts on virus replication in HepG2.2.15 cells infected with the HAV HA11-1299 strain [73]. It is noteworthy that miso extracts have an inhibitory effect on HAV replication but no inhibitory effect on HBV replication. Japanese rice-koji miso extracts may have an inhibitory effect on HAV replication in patients superinfected with HAV and HBV.

The zinc homeostasis pathway was identified as a key pathway of the antiviral activity of Japanese rice-koji miso against HAV infection using transcriptome-sequencing analysis [98]. We also demonstrated that zinc sulfate has an inhibitory effect on HAV HA11-1299 replication in human hepatocytes with the enhancement of GRP78 expression [98]. As Japanese miso soup and zinc sulfate are traditional foods and drugs, respectively, they induce GRP78 expression and are useful and safe antiviral compounds against HAV, with fewer adverse events. Gut dysbiosis and increased permeability cause pathological bacterial translocation and endotoxemia, which play an important role in the development of ACLF [2]. HAV infects the liver by the gut-portal vein–liver axis through fecal–oral routes. The digestion and absorption of Japanese rice-koji miso extracts and zinc sulfate may be used through similar routes.

#### *10.3. Candidates of Antivirals against HAV in Chronic Liver Diseases*

β The inhibitory effects of interferon-alpha, interferon-gamma, interferon-lambda, ribavirin, amantadine, sirtinol, and AZD1480 as host-targeting drugs and HAV 3C cysteine protease inhibitors, as well as small interfering RNAs against HAV, as antivirals that directly act on HAV replication, have been reported [11,46,99]. Interferon has antiviral potential against HAV [100,101], but it is difficult to use interferon in patients with ACLF, as interferon generally has cytotoxicity. Peginterferon-lambda has fewer side effects than peginterferon-alpha and may be useful in some patients with HAV infection. Amantadine is a broad-spectrum antiviral and has an inhibitory effect on HAV replication through the targeting of HAV internal entry site (IRES) activity [100,102,103]. The sirtuin inhibitor, sirtinol, also inhibits HAV replication by inhibiting HAV IRES activity [104]. Further studies on the mechanism of the sirtuin inhibitor and JAK pathways in HAV replication are needed [104,105]. In patients with chronic liver diseases or ACLF, these drugs should be improved, and more safe drugs are needed and should be explored. It has been reported that HCV receptor candidates, such as HAV cellular receptor 1 (HAVcr-1), integrin β1, and gangliosides, are the entry receptor candidates for HAV. Further studies in this vein are needed [106–108]. Gangliosides seem to function as endosome receptors for infection using both naked and quasi-enveloped HAV virions [108]. Blocking the cellular entry of HAV is also an attractive drug target for combating HAV infection.

#### *10.4. HAV Infection Is Associated with the Activation of the Host Immune System and Severe Systemic Inflammation*

Acute hepatitis A usually exhibits more severe inflammation, such as a higher fever and higher C-reactive protein levels, compared to acute hepatitis due to other hepatitis viruses [109–111]. Some cases of acute HAV infection present acute renal failure [112–114]. These results suggest that HAV infection activates human immune systems and induces cytokines [115–119]. Innate immunity also seems to be involved in the pathogenesis of hepatitis A [120,121]. Hypergammaglobulinemia and a high occurrence of autoantibodies are observed in HAV infection [122,123]. This may support the immunological basis of its pathogenesis. Moreover, the higher gammaglobulinemia in fulminant HAV suggests the existence of a more aggressive immunological reaction in severe hepatitis A [123].

Severe systemic inflammation can affect the functions of somatic cells in tissue and modify the clinical manifestation of cirrhosis and ACLF [124,125]. Patients with acute liver failure or ACLF are susceptible to infection, and early transplant-free survival is poor [126–129]. In liver transplantation for patients with ACLF, the role of the timing, bridging, and management of liver transplantation is important [130,131].

#### *10.5. Recent Outbreak of HAV Infection in MSM*

It has recently been reported that HAV susceptibility parallels the high COVID-19 mortality [132]. The 2019 coronavirus disease (COVID-19) has been observed in Japan, where the HAV susceptibility of the general population is high [34,35]. An HAV vaccination program is urgently required for individuals with or without HIV infection in this area. HAV infection is an imported infection, like novel severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection [133]. In the era of COVID-19, attention should also be paid to dual infection with HAV and SARS-CoV-2.

An outbreak of HAV infection in MSM has been observed worldwide. An outbreak of acute HAV infection among HIV-coinfected MSM in Taiwan was observed from June 2015 to September 2017 [50,134,135]. Between July 2016 and February 2017, 48 male cases of HAV infection were found in the Netherlands [48]. A total of 17 of them were MSM. This strain is identical to a strain causing a large outbreak among MSM in Taiwan [48]. In the United States, HAV infections also increased among MSM from 2016 to 2018 [54,136,137]. Since 2017, HAV infection has increased among MSM in Japan [34,37,52]. RIVM-HAV16-090-like hepatitis A virus strains, which were >99.6% identical to the 66 reported strains isolated from Taiwan and European countries from 2015 to 2017, were also recovered from Japanese MSM [52]. A recent outbreak of HAV infection was also reported in various countries, such as Brazil, Spain, and Italy [53,138–140].

#### **11. Possible Molecular Mechanism of the Development of ACLF in Patients with HAV Infection**

The molecular mechanism of the development of ACLF in patients with HAV infection is not fully understood. The possible mechanisms of the development of liver failure in the presence of coinfection with HCV and HAV are as follows. HAV is a virus that is generally sensitive to interferon [100–102]. In comparison with HCV, HAV induces a limited production of type I interferon when HAV infects chimpanzees [141]. Compared with HBV and HCV, HAV weakly induced the activation of NF-κB signaling pathways in human hepatocytes [142,143]. While HAV VP3 activates cell growth signaling [143], HAV VP1/2A reduces cell viabilities in HCV sub-genomic replicon cells [144].

HAV is usually a non-cytopathic virus, and HAV inhibits double-stranded (dsRNA)-induced interferon-beta gene expression by influencing the interferon-beta enhanceosome, as well as dsRNA-induced apoptosis [145]. Compared with HBV and HCV, HAV could evade mitochondrial antiviral signaling protein (MAVS)-mediated type I interferon responses [146]. HAV 3ABC is capable of MAVS cleavage, like HCV NS3/4A, which cleavesMAVS and disrupts interferon signaling [147]. HAV 3C inhibits HAV IRES-dependent translation and cleaves the polypyrimidine tract-binding protein [148]. HCV induces interferon-beta signaling pathways in human hepatocytes [149]. Controlling the effects of interferon signaling may determine the prognosis of patients coinfected with HCV and HAV (Figure 2).

(**a**)

(**b**)

κ **Figure 2.** Possible molecular mechanism of the development of acute-on-chronic liver failure (ACLF) in patients coinfected with hepatitis A virus (HAV) and HCV. (**a**) Only HAV infection; (**b**) coinfection HAV and HCV. RIG-I, retinoic acid-inducible gene-I; MDA-5, melanoma differentiation associated gene 5; TLR3, toll-like receptor 3; MAVS, mitochondrial antiviral signaling protein; IRF3, interferon regulatory factor 3; IFN, interferon; ISG, interferon-stimulated gene; NF-κB, nuclear factor kappa B subunit 1.

HBV is a stealth virus which efficiently infects humans without alerting the innate immune system, although HCV strongly induces but cunningly evades the innate immune response [150]. The high glucose and fat deposition of hepatocytes seem to induce a chaperon-mediated autophagy (CMA) [151]. CMA targets interferon-alpha receptor chain-1 for degradation, dampens hepatic innate immunity, and disrupts interferon signaling pathways [151]. CMA is also observed in patients with ALD or MAFLD [152,153]. Altering interferon signaling may contribute to ALF-associated acute HAV infection. However, further studies are needed. Among HIV-positive patients with acute HAV infection, lower peaks in total bilirubin, AST, and ALT levels were observed in comparison with HIV-negative patients with acute HAV infection [154], suggesting that weaker immune responses occur in HIV-positive patients. These immune responses could enhance HAV replication and modify the pathogenesis in HIV-positive patients with acute hepatitis A [155].

#### **12. Conclusions**

We reviewed the literature concerning HAV infection in patients with chronic liver diseases. In patients with chronic liver diseases, HAV infection can occasionally lead to a critical condition, such as acute liver failure. There seems to be no etiological association between liver failure and HAV infection, but there is a significant correlation between the severity of liver disease and the degree to which the liver has already been damaged. While there are effective HAV vaccines currently in existence, antivirals against HAV should be further explored. The latter is urgent given that the lives of patients with HAV infection and a chronic liver disease of another etiology may be at immediate risk.

**Author Contributions:** Conceptualization, T.K., R.S. and R.M.; formal analysis, T.K.; investigation, T.K.; resources, T.K.; writing—original draft preparation, T.K.; writing—review and editing, T.K., R.S. and R.M.; supervision, H.T., T.M., N.M., K.N. and M.M.; funding acquisition, T.K. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was supported by the Japan Agency for Medical Research and Development (AMED), under grant number JP20fk0210075.

**Conflicts of Interest:** The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

## **Abbreviations**


#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## *Article* **Clusters of Circulating let-7 Family Tumor Suppressors Are Associated with Clinical Characteristics of Chronic Hepatitis C**

**Yi-Shan Tsai <sup>1</sup> , Ming-Lun Yeh 1,2, Pei-Chien Tsai 1,3, Ching-I Huang <sup>1</sup> , Chung-Feng Huang 1,2,3,4 , Meng-Hsuan Hsieh 1,2,3,4, Ta-Wei Liu <sup>1</sup> , Yi-Hung Lin <sup>1</sup> , Po-Cheng Liang <sup>1</sup> , Zu-Yau Lin 1,2 , Shinn-Cherng Chen 1,2, Jee-Fu Huang 1,2 , Wan-Long Chuang 1,2,5, Chia-Yen Dai 1,2,3,4,5,6,7,\* ,**† **and Ming-Lung Yu 1,2,3,5,6,**†


Received: 10 June 2020; Accepted: 10 July 2020; Published: 13 July 2020

**Abstract:** Hepatitis C virus (HCV) infections can cause permanent liver-related diseases, including hepatocellular carcinoma (HCC). Low mortality and incidence of HCC have been observed in patients with chronic hepatitis C undergoing direct-acting antiviral therapy. Tumor suppressive let-7 family members are down-regulated in HCC. The present study, therefore, aimed to investigate whether expression levels for the full spectrum of let-7 family members (let-7a, 7b, 7c, 7d, 7e, 7f, 7g, 7i, and miR-98) in the circulatory system are useful as surveillance biomarkers for liver-related diseases to monitor treatment efficacy during HCV infection. To this end, we measured the levels of mature circulating let-7 family members using quantitative reverse transcription-PCR in 236 patients with HCV infection, and 147 age- and sex-matched controls. Using hierarchical cluster analysis and principal component analysis, three clusters were obtained after measuring expression levels of let-7 family members in the patients and controls. Cluster 1 included let-7a/d/e/g, Cluster 2 comprised let-7b and let-7i, and Cluster 3 comprised let-7c/f/miR-98. Let-7b/c/g represented the three clusters and showed the best survival response to liver cancer when analyzed with respect to patient data. Therefore, considering the circulating levels of let7 b/c/g as representatives of the let-7 family may facilitate effective monitoring of liver-related disease.

**Keywords:** chronic hepatitis C; let-7 family; hepatitis C virus; miRNA; biomarker; hepatocellular carcinoma

#### **1. Introduction**

Chronic hepatitis C (CHC) virus infection can cause serious liver disorders. It can also increase the risk of hepatocellular carcinoma (HCC) and the progression of severe hepatic and extrahepatic diseases [1]. Compared to non-virological response (NVR), the sustained virological response (SVR) resulting from treatment with pegylated interferon-α (PegIFNα) and ribavirin (RBV) is associated with a lower risk of HCC development [2]. However, patients with significant hepatic fibrosis remain at high risk for HCC, even when they achieve SVR with antiviral therapy [3,4]. Since 2013, the development of direct-acting antivirals (DAAs) has increased SVR rates by 95% above those of interferon (IFN)-based treatments in patients with chronic HCV infection regardless of viral genotype [5]. However, there is no evidence that HCC occurrence or recurrence differs between patients receiving DAA or IFN therapy [6]. Furthermore, the annual post-SVR HCC incidence (approximately 1%) remains higher than that for cancers of other organs. Therefore, it is necessary to establish a clinical strategy for monitoring cancer risk in post-SVR patients [7,8].

Mature microRNAs (miRNAs) are short, single-stranded RNA molecules (approximately 22 nucleotides in length) that post-transcriptionally silence gene expression and play important roles in a broad variety of biological processes, including intrinsic antiviral immunity. Chen et al. systematically characterized serum/plasma miRNAs and found that they were stable, reproducible, and consistent among individuals of the same species. These miRNAs also represent promising, non-invasive biomarkers for diagnosing cancer and other diseases [9]. Furthermore, it has been suggested that miRNAs can be used as a prediction model for the treatment outcome of HCV virus genotype 1 infection [10].

Previous studies have shown that the miRNA let-7b exhibits a significant anti-HCV effect [11], and that IFNα rapidly modulates the expression of let-7s with anti-HCV activity by targeting *IGF2BP* [12]. In our previous study, we demonstrated the effects of let-7g on HCV infection in vitro in clinical tissue and serum samples. We found that IFN/RBV treatment induces let-7g expression. Furthermore, overexpression of let-7g reduces the expression of the HCV gene and core protein level, thereby inhibiting viral replication. Let-7g and IFN/RBV treatment also synergistically inhibits HCV replication and represses Lin28A/B [13].

The let-7 family, comprised of ten members (let-7a, 7b, 7c, 7d, 7e, 7f, 7g, 7i, miR-98, and miR-202), target the 3′ untranslated regions (UTRs) of genes essential for development and are conserved from *Caenorhabditis elegans* to humans [14]. Next-generation sequencing and microarray studies have revealed that various HCC-specific miRNA signatures in the liver tissue showed lower let-7 (a/b/c/d/e/f/g) expression levels compared to healthy liver tissues [15]. Matsuura et al. performed a longitudinal miRNA microarray study on plasma and extracellular vesicles (EVs) in patients with CHC and found that the plasma levels of circulating let-7(a/c/d) were higher than those in EVs, and were inversely correlated with the severity of hepatic fibrosis [16].

Despite these studies, the roles of mature let-7 family members (let-7a, 7b, 7c, 7d, 7e, 7f, 7g, 7i, and miR-98) in the circulating plasma of patients with CHC and their clinical relevance remain unclear. Moreover, due to the large number of let-7 family members, and the difficulty associated with obtaining liver tissues from CHC patients and the control group, detection of all family members for the purposes of monitoring liver-related diseases is not feasible. Nevertheless, this study aimed to evaluate the association between circulating let-7 family members and the clinical characteristics of CHC patients. We also examined the potential for use of nine mature let-7 family members as non-invasive biomarkers for CHC patients, using cluster analysis and principal component analysis (PCA). An independent dataset was then employed to evaluate the clusters of the let-7 family in normal tissue (N), and tumors (T) from the Cancer Genome Atlas (TCGA) liver cancer samples. Here we demonstrate that assessing the levels of the circulating let-7 family members could represent a promising new method to monitor liver-related diseases.

## **2. Results**

#### *2.1. Clinical Characteristics of Patients*

The characteristics of the subjects enrolled in this study are listed in Table 1. Compared with the healthy control group, CHC patients had significantly higher aspartate transaminase (AST, *p* < 0.0001), alanine aminotransferase (ALT, *p* < 0.0001), gamma-glutamyl transferase (γGT, *p* < 0.0001), and fasting plasma glucose (*p* < 0.0001) levels. Alternatively, creatinine (Cr, *p* < 0.0001), white blood cell counts (WBC, *p* < 0.0001), platelet counts (PLT, *p* < 0.0001), hemoglobin (*p* < 0.0001), total triglycerides (TG, *p* = 0.0002), total cholesterol (CHLO, *p* < 0.0001), HDL cholesterol (HDL-C, *p* < 0.0001), and LDL cholesterol (LDL-C, *p* < 0.0001) were significantly lower in CHC subjects.



<sup>1</sup> BMI, body mass index; HCC, hepatocellular carcinoma; LC, liver cirrhosis; Cr, creatinine; AST, aspartate aminotransferase; ALT, alanine aminotransferase; γGT, gamma-glutamyl transferase; WBC, white blood cell count; PLT, platelet count; TG, triglyceride; CHLO, cholesterol; HDL-C, HDL cholesterol; LDL-C, LDL cholesterol; AC\_Sugar, fasting plasma glucose; HbA1c, hemoglobin A1c. All values are expressed as the mean ± standard deviation (SD). The *p* value was calculated for the continuous variables using the Student's *t*-test or Mann–Whitney test, and the χ 2 test was used for the categorical variables, with α = 0.05. \* = *p* < 0.05. # The HCV virus loads were determined by log-transformation.

#### *2.2. Circulating Let-7 Family Member Profiling*

To determine the expression of let-7 family members in the blood, a qRT-PCR was performed. The −△*C*t (cycle threshold) value of each miRNA was measured and normalized to that of cel-39, as it showed the most consistent Ct value among all donors (27.10 ± 0.82 and 27.18 ± 1.06 for healthy controls and CHC patients, respectively) (Table S1). After statistical analyses, the expression levels of the let-7 family members were significantly lower in patients with CHC compared to the control group (Table 2 and Figure 1). The distribution of individuals is provided in Figure S2.


**Table 2.** Circulating let-7 family expression at baseline (Log102 −△*C*t ) 1 .

<sup>1</sup> Data is presented as the mean <sup>±</sup> SD. △*C*<sup>t</sup> <sup>=</sup> <sup>C</sup>target – CTcel39; \* <sup>=</sup> *<sup>p</sup>* <sup>&</sup>lt; 0.05 following statistical analysis using an ANOVA with Bonferroni correction (α = 0.0056). α

**Figure 1.** Hierarchical clustering plot (heatmap) of circulating let-7 family in the control group without hepatitis C virus (HCV) infection (*n* = 147) and HCV-infected patients (*n* = 236).

#### *2.3. Three Clusters Distinguish the Let-7 Family Members*

Due to let-7 family members (let-7a, 7b, 7c, 7d, 7e, 7f, 7g, 7i, and miR-98) with highly correlated (Table S2). To further investigate associations between the expression levels of let-7 family members in individuals from the control and HCV-infected groups, hierarchical clustering was performed using Ward's method. Three distinct clusters of let-7 family members were identified, in both the control and HCV-infected groups as represented in the hierarchical clustering plot (heatmap, Figure 1). Cluster 1 included let-7a, let-7d, let-7e, and let-7g. Cluster 2 was comprised of let-7b and let-7i, while Cluster 3 was made of let-7c, let-7f, and miR-98. We also performed a PCA to investigate the similarities between the expression levels of let-7 family members among the study participants and found that Cluster 3 (let-7c, f, and miR-98) was clearly distinct from clusters 1 and 2 in both patient groups (Figure S3).

#### *2.4. Circulating Let-7 Family Expression Levels Correlate with Baseline Clinical Parameters*

Correlation analyses between differentially expressed let-7 family members and clinical parameters were performed, including characteristics such as age, indicators of liver damage or injury (AST and ALT), platelet count, and HCV viral load (Table 3). The expression levels of Cluster 1 members (let-7a and let-7g) were significantly negatively correlated with AST (*r*=−0.1898 and−0.2038; *p* = 0.0037 and 0.0018, respectively). Meanwhile, the expression of Cluster 1 members (let-7a/d/e/g) were significantly

−

positively correlated with PLT (*r* = 0.2433, 0.2209, 0.2113, and 0.1911; *p* = 0.0002, 0.0007, 0.0012, and 0.0034, respectively). However, no correlation was observed between any of the three let-7 clusters and ALT or HCV load. These results demonstrate that Cluster 1 (let-7a/d/e/g) was a better indicator of clinical characteristics than the other two let-7 clusters. Additionally, the circulating let-7 levels were not correlated with the HCV genotype (Table S3) or HCV viral load (Table S4).

**Table 3.** Correlation between circulating expression of let-7 family members and various clinical parameters (*n* = 236) <sup>1</sup> .


<sup>1</sup> Data represents correlation between miRNA expression and the values of each clinical parameter, as reported in Table 1. Correlation was determined using Pearson's test. AST, aspartate aminotransferase; ALT, alanine aminotransferase; PLT, platelet count; HCV, hepatitis C viral load. \* = *p* < 0.05 following the Bonferroni correction (α = 0.0056).

#### *2.5. Let-7b*/*c*/*g Levels Are Associated with Clinical Progression*

An independent dataset was used to further evaluate the let-7 clusters in normal tissue (N) and tumors (T) from TCGA liver cancer samples. A hierarchical clustering plot revealed three clusters, grouped into let7a-1/a-2/a-3/f-1/f-2/g/e, let-7b/c, and let7d/i (Figure 2). The expression levels of let-7a/b/c/g/i were significantly lower in tumors than in healthy tissue. Furthermore, the greatest reduction in let-7b/c/g expression was observed in tumors (red line, Figure 3). This observation was supported by OncomiR (www.oncomir.org.), which was used to explore the associations between TCGA-LIHC (liver hepatocellular carcinoma) survival data and the let-7 family expression levels. Kaplan–Meier survival analysis demonstrated that patients with high expression of let-7b/c/g had significantly increased overall survival compared to patients with low expression (*p* = 0.03162, Figure 4). Taken together, these data indicate that reduced expression of let-7b/c/g may be associated with liver tumor progression.

**Figure 2.** Hierarchical clustering plot (heatmap) of the let-7 miRNA family expression in healthy (*n* = 50) and tumor (*n* = 366) liver tissue samples obtained from TCGA-LIHC.

**Figure 3.** Dot plot showing the relationship between the expression of let-7 family members in normal (*n* = 50) and tumor (*n* = 366) tissues from liver cancer patients in the TCGA dataset. Log<sup>2</sup> (RPM +1) transformed values for let-7 family members are shown as the mean ± SD. Prominent declines in let-7b/c/g are indicated by red lines. Statistical significance was assessed using the Mann–Whitney test. The *p* values are represented as follows: \*\* = *p* < 0.01, \*\*\* = *p* < 0.001, \*\*\*\* = *p* < 0.0001.

α **Figure 4.** Kaplan–Meier (KM) survival analysis curve for let-7 family expression in TCGA-LIHC patients. The KM survival curves were examined according to (**A**) let-7 a/b/c/d/e/f/g/i expression level and (**B**) let-7 a/b/c/g/i expression levels. No differences in overall survival were observed. (**C**) let-7 b/c/g expression levels and (**D**) let-7 b/c expression levels show clear differences in overall survival. Statistical significance was obtained by log-rank tests, with α = 0.05.

#### **3. Discussion**

′ Circulating mature miRNAs are small RNAs measuring approximately 22 nucleotides, and are known to be stable in the serum/plasma [9]. These nucleic acids represent novel non-invasive biomarkers for liver inflammation, liver fibrosis, liver cancer [17], and non-alcoholic fatty liver disease (NAFLD), [18]; they have also been shown to be useful for cancer detection [19]. Next-generation sequencing or microarray methods may be useful for identifying let-7 (a/b/c/d/e/f/g) members that are down-regulated in plasma or liver tissues to diagnose patients with hepatic fibrosis or hepatocellular carcinoma [15,16]. The ten mature let-7 family members are derived from 13 precursors located in nine different chromosomes with similar seed regions [20]. These let-7 family members target 3′ UTRs of genes that are essential for development and have been conserved from *C. elegans* to humans [14]. However, the precise roles of circulating let-7 family members in humans remain uncharacterized. Here, we studied nine mature let-7 family members, including let-7a, 7b, 7c, 7d, 7e, 7f, 7g, 7i, and miR-98 (*Homo sapiens* [has]-miR-202 was excluded), in the circulating plasma of patients with CHC and healthy controls, using TaqMan quantitative RT-PCR assays. To the best of our knowledge, this is the first study to demonstrate that circulating let-7 family members can be classified into three similar clusters in control and CHC groups. The highest expression levels were found for Cluster 2 (let-7b and let-7i), while the lowest expression levels were found in Cluster 3 (let-7c/let-7f/miR-98). Moreover, the expression levels of circulating let-7 family members in patients with CHC were lower than those in the healthy population. Here we also identified the circulating let-7 b/c/g as representatives from each of the three let-7 clusters as the most effective markers for detecting liver-related diseases based on their strong association with clinical characteristics.

Our previous study showed that the expression level of let-7g in liver tissues was significantly lower in NVR patients than in SVR patients and that antiviral treatment with IFN/RBV could induce let-7g expression [13]. Moreover, the ectopic overexpression of let-7 family members was found to

repress HCV core protein, and HCV loads in a cell model system [12,13]. However, in this study, the HCV load and HCV genotype were not found to be dependent factors and were negatively correlated with the expression levels of let-7 family members in the circulatory system. Therefore, let-7 family members might specifically target naked HCV RNA in liver tissue or cell models, but not HCV RNA within the envelope coat.

Previously Kirschner et al. reported specific miRNAs for which the profiling suggested an influence of hemolysis (miR-16, -451, -92a, let-7b, -103, -106a, 17, -21, -210, -27a, -31,-625-3p, -92a) as well as miRNAs (let-7a and let-7d) that appeared to be unaffected by hemolysis [21]. We, therefore, also measured the absorbance peak at 414 nm to detect the level of hemolysis and observed values of 0.192 ± 0.1478 (mean ± SD) for the control group, and 0.3143 ± 0.1723 (mean ± SD) for the CHC group (*p* < 0.05; Figure S4). Hemolysis was higher in CHC group. Therefore, the circulating let-7 (a/d/e/g) Cluster1 for the control group and CHC group might be unaffected by hemolysis.

Down-regulated let-7 miRNA expression in the circulatory system might result in an increase in interleukin-10 (IL-10) from CD4+ T-cells, providing the virus with an important survival advantage by manipulating the host immune response [22]. Early IL-10 elevation has been shown to strongly suppress the priming of naïve HCV–specific CD8+ T-cells, causing T-cell failure and viral persistence [23]. In addition, HCV induces the expression of toll-like receptor 4 (TLR4), enhancing the production of IFNβ and IL-6 [24]. Vespasiani-Gentilucci et al. confirmed that dominant TLR4 hyperexpression in patients with CHC was significantly correlated with the inflammatory score and degree of fibrosis, indicating that TLR4 plays an important role in the pathogenesis of HCV-related chronic liver diseases [25]. It has also been reported that let-7b can target TLR4 through 3′ UTR post-transcriptional regulation and attenuates NF-κB activity [26]. Furthermore, expression levels of circulating let-7 (a/c/d) are inversely correlated with the severity of hepatic fibrosis [16], indicating that decreases in expression levels of let-7 family members might be involved in HCV infection and may cause inflammation, which is clinically and epidemiologically linked to cancer. NF-κB has been shown to have a causative role in inflammation. Iliopoulos et al. previously showed that transient activation of Src oncoproteins could trigger an NF-κB-mediated inflammatory response that directly activates Lin28 transcription and rapidly reduces the miRNA levels of let-7 family members [27].

The HCV core protein has been shown to have oncogenic potential [28]. Ali et al. demonstrated that the expression of an HCV sub-genomic replicon in cultured cells could cause them to acquire cancer stem cell-like signatures, including the enhanced expression of Lin28 and other proteins [29]. More importantly, Lin28b has been shown as sufficient to drive liver cancer and necessary for cancer maintenance in murine models. Many human cancers exhibit deregulated let-7 expression [14,30], the specificity of which is inhibited by Lin28A/B [31].

Interestingly, we also observed that let-7 (a/d/e/g) expression levels were positively correlated with PLT. Platelets are known therapeutic targets for preventing ischemic damage, enhancing liver regeneration, and inhibiting hepatitis progression [32]. These miRNAs are also inversely correlated with AST and ALT, which are markers of liver damage. Together, these results indicate that decreases in let-7 b/c/g levels might be involved in HCV infection and damage; moreover, the literature suggests these could promote inflammatory responses and tumorigenesis associated with HCV-related fibrosis or HCC.

In human genomic loci clusters, miRNA genes, including let-7 genes, are frequently located at fragile sites. Inflammatory status promoters increase the production of reactive oxygen species, leading to oxidative DNA damage by reducing DNA repair and increasing genomic instability of these fragile sites [33]. Genomic regions involved in cancers as tumor suppressor genes, such as *let-7g*/*miR-135*−*1*, are located in fragile sites of the ARP-DRR1 region in 3p21.1-21.2. [34]. Therefore, decreasing levels of let-7 family members may not only be an important marker for disease progression, but may also suggest fragile site instability and promotion of oncogenes.

Certain limitations were noted in this study. First, plasma samples are easily contaminated with peripheral blood mononuclear cells. Although previous studies have reported that let-7b/c/g levels in PBMCs decline rapidly in HIV infection [22,35], HIV patients were excluded in the current study. Second, the relationship between expression levels of mature let-7 family members and inflammatory factors was unclear. Detailed analysis of other inflammatory mediators as intermediate variables is required. Such an analysis may provide additional information regarding the roles of other inflammatory mediators in the progression of HCV infection. Third, a large population-based follow-up study that includes additional circulating miRNAs is needed. This would allow for the investigation of baseline vs. post-antiviral treatment expression of the tumor suppressor let-7 b/c/g genes as potential early monitoring targets for patients who are at high risk for fibrosis, cirrhosis, and HCC.

#### **4. Materials and Methods**

#### *4.1. Ethics Approval and Consent to Participate*

This study was reviewed and approved by the IRB ethics committee (KMUHIRB-980176, KMUHIRB-20120097, and KMUHIRB-20140054 with the approval dates of 16 July 2009, 11 April 2012 and 7 November 2014, respectively). All protocols were approved by the ethical committee of the Kaohsiung Medical University Hospital, based on the International Conference on Harmonization for Good Clinical Practice. All participants provided written informed consent before enrollment.

#### *4.2. Patient Cohort*

We analyzed stored plasma extracted from the HCV patient database collected from one medical center and two regional core hospitals of the Kaohsiung Medical University between October 2009 and December 2016. The members of the age- and sex-matched control groups were enrolled from the same geographic communities. A total of 236 patients with HCV infection were enrolled in the HCV group. The following inclusion criteria were established before selection: i) Positive for anti-HCV antibodies for more than 6 months, and positive for HCV RNA by PCR assay; ii) negative for hepatitis B surface antigen (HBsAg) and no concomitant HIV; iii) negative for other types of hepatitis, including autoimmune hepatitis, primary biliary cirrhosis, sclerosing cholangitis, Wilson's disease, and α1-antitrypsin deficiency; iv) daily ethanol consumption of <20 g (both females and males), as confirmed through an interview with the patient and a family member; and v) a high serum ALT level for 6 months preceding the study entry. The control group comprised 147 age- and sex-matched subjects without viral hepatitis who were selected based on the following inclusion criteria: i) Normal liver echogenicity, as determined by ultrasound sonography and ii) normal liver function test. The exclusion criteria for the control group included a current or past history of alcohol abuse (≥20 g ethanol per day), being pregnant, and seropositivity for HBsAg or anti-HCV antibody. All participants were advised to fast for 12 h overnight before the standard biochemistry tests, which included tests for Cr, AST, ALT, γGT, WBC, PLT, hemoglobin, TG, CHLO, HDL-C, LDL-C, AC\_sugar, and HbA1c. Anthropometric data, including body weight and height, were obtained using standardized techniques. HBsAg was detected with commercially available enzyme-linked immunosorbent assay kits (Abbott Laboratories, North Chicago, IL, USA). HCV RNA and HCV genotype were assayed using a real-time PCR assay (RealTime HCV; Abbott Molecular, Des Plaines, IL, USA; detection limit, 12 IU/mL) [36].

#### *4.3. MicroRNA Extraction from Plasma*

A fixed volume of 200 µL of plasma was extracted using 1 mL Trizol LS reagent (Thermo Scientific, Wilmington, DE, USA), according to the manufacturer's protocol. Molecular biology-grade 1–bromo–3–chloropropane (BCP; 300 µL/1 mL TRIzol LS) was then added. After centrifuging for 15 min at 12,000× *g* and 4 ◦C, a fixed volume of the aqueous phase was transferred into a new tube, which was treated with 2 µL of 100-pmol/L synthetic *C. elegans* Cel-39 as a spike-in control, and 10 µg of RNase-free glycogen as a co-precipitant carrier. The aqueous sample was mixed thoroughly with 100% molecular-grade isopropanol and incubated on ice for 1 h. A gel-like pellet was obtained after centrifugation at 12,000× *g* for 15 min. The flow-through was discarded and the pellet was washed

with 80% ethanol. The pellet was again centrifuged at 7500× *g* for 5 min at 4 ◦C, and the supernatant was discarded. The RNA (including the miRNA) was air dried, eluted in 15 µL of RNase-free H2O and stored at −80 ◦C until further analysis. A total of 2 µL of each sample was used to measure the absorbance peak at 414 nm for the level of hemolysis on a NanoDrop™ 2000 (Thermo Fisher Scientific, Inc., Waltham, MA, USA) [37].

#### *4.4. Quantification of Circulating miRNAs*

The total RNA (20 ng) was used as a template for reverse transcription, which was performed using the TaqMan MicroRNA Reverse Transcription kit and the associated miRNA-specific stem-loop primers to convert miRNA to cDNA as step 1 (Thermo Scientific, Wilmington, DE, USA), according to the manufacturer's instructions. The *C. elegans* synthetic cel-miR-39-3p, a suitable and reproducible normalizer, was used as the spiked-in control [38]. The relative levels of individual miRNAs were amplified using the TaqMan® Universal PCR Master Mix II, without uracil N-glycosylase (UNG), on a 7900HT Sequence detection system. Specific amplification was performed using the following program: 95 ◦C for 10 min, and amplification followed by 40 cycles of 95 ◦C for 15 s and 60 ◦C for 1 min as step 2. Specific primers of mature miRNA sequences were used for the TaqMan R microRNA assays (assay ID): hsa-let-7a (000377), hsa-let-7b (002619), hsa-let-7c (000379), hsa-let-7d (002283), hsa-let-7e (002406), hsa-let-7f (000382), hsa-let-7g (002282), hsa-let-7i (002221), hsa-miR-98 (000577), and an internal control, cel-miR-39-3p (000200). The hsa-miR-202 (002362) primer was excluded as it was under-determined in 99% of the samples with different nucleotides of the mature let-7 family members (Figure S1). The relative expression level of each let-7 member was determined using the comparative C<sup>t</sup> method, which was defined as 2−∆*C*<sup>t</sup> , where <sup>∆</sup>*C*<sup>t</sup> <sup>=</sup> *<sup>C</sup>*<sup>t</sup> of the let-7 member <sup>−</sup> *<sup>C</sup>*<sup>t</sup> of cel-39.

#### *4.5. miRNA-seq and Clinical Data from UCSC Xena Platform*

LIHC samples, including healthy (*n* = 50) and primary tumor tissues (*n* = 366) were obtained from the UCSC Xena platform. This platform provides interactive online visualization of cancer genomics datasets, such as TCGA, a public data resource [39]. Expression of the let-7 miRNA family mature strand was transformed according to RNA sequencing guidelines (Illumina Hiseq 2000) and presented as log<sup>2</sup> (RPM +1). Kaplan–Meier survival curves and log-rank methods for the let-7 family clusters in LIHC were performed using OncomiR to evaluate overall survival (OS) rate [40].

#### *4.6. Statistical Analysis*

Statistically significant differences between the expression levels of let-7 family members in the different groups were determined using the Mann–Whitney test with Bonferroni correction for multiple comparisons. Pearson's correlation analysis was used to assess the relationships between the let-7 family members and the different clinical parameters. Hierarchical clustering and PCA were performed to identify the distinguishable let-7 family members. All analyses were performed using JMP 12.0 (SAS Institute, Cary, NC, USA). Graphs were generated using the GraphPad Prism 5.0 software (San Diego, CA, USA).

**Supplementary Materials:** Supplementary materials can be found at http://www.mdpi.com/1422-0067/21/14/ 4945/s1.

**Author Contributions:** Investigation, Y.-S.T.; Formal analysis, Y.-S.T. and P.-C.T.; Patient sample collection, M.-L.Y. (Ming-Lun Yeh), C.-I.H., C.-F.H., M.-H.H., T.-W.L., Y.-H.L., P.-C.L., Z.-Y.L., S.-C.C., J.-F.H., C.-Y.D., M.-L.Y. (Ming-Lung Yu), and W.-L.C.; Writing—Original Draft Preparation; Y.-S.T., C.-Y.D., and M.-L.Y.; Writing—Review and Editing, C.-Y.D. and M.-L.Y. All authors have read and agree to the published version of the manuscript.

**Funding:** This work was supported by grants from the National Science Council of Taiwan (grant number MOST104-2314-B-037-075-MY3) and Kaohsiung Medical University Hospital (grant number KMUH106-6R06 and KMUH107-7R06). Taiwan Liver Research Foundation (TLRF) provides free liver disease surveillance and health educational lectures in local community.

**Conflicts of Interest:** The authors declare no conflict of interest. The sponsors had no role in the design, execution, interpretation, or writing of the study.

### **Abbreviations**


#### **References**


development in non-cirrhotic patients with successful hepatitis C virus eradication. *J. Hepatol.* **2014**, *61*, 67–74. [CrossRef] [PubMed]


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

*Article*
