The U-Rich Untranslated Region of the Hepatitis E Virus Induces Differential Type I and Type III Interferon Responses in a Host Cell-Dependent Manner

Abstract

Hepatitis E virus (HEV), a single-strand positive-sense RNA virus, is an understudied but important human pathogen. The virus can establish infection at a number of host tissues, including the small intestine and liver, causing acute and chronic hepatitis E as well as certain neurological disorders. The retinoic acid-inducible gene I (RIG-I) pathway is essential to induce the interferon (IFN) response during HEV infection. However, the pathogen-associated motif patterns (PAMPs) in the HEV genome that are recognized by RIG-I remain unknown. In this study, we first identified that HEV RNA PAMPs derived from the 3' untranslated region (UTR) of the HEV genome induced higher levels of IFN mRNA, interferon regulatory factor-3 (IRF3) phosphorylation, and nuclear translocation than the 5' UTR of HEV. We revealed that the U-rich region in the 3' UTR of the HEV genome acts as a potent RIG-I PAMP, while the presence of poly(A) tail in the 3' UTR further increases the potency. We further demonstrated that HEV UTR PAMPs induce differential type I and type III IFN responses in a cell type-dependent fashion. Predominant type III IFN response was observed in the liver tissues of pigs experimentally infected with HEV as well as in HEV RNA PAMP-induced human hepatocytes in vitro In contrast, HEV RNA PAMPs induced a predominant type I IFN response in swine enterocytes. Taken together, the results from this study indicated that the IFN response during HEV infection depends both on viral RNA motifs and host target cell types. The results have important implications in understanding the mechanism of HEV pathogenesis.IMPORTANCE Hepatitis E virus (HEV) is an important human pathogen causing both acute and chronic viral hepatitis E infection. Currently, the mechanisms of HEV replication and pathogenesis remain poorly understood. The innate immune response acts as the first line of defense during viral infection. The retinoic acid-inducible gene I (RIG-I)-mediated interferon (IFN) response has been implicated in establishing antiviral response during HEV infection, although the HEV RNA motifs that are recognized by RIG-I are unknown. This study identified that the U-rich region in the 3' untranslated region (UTR) of the HEV genome acts as a potent RIG-I agonist compared to the HEV 5' UTR. We further revealed that the HEV RNA pathogen-associated motif patterns (PAMPs) induced a differential IFN response in a cell type-dependent manner: a predominantly type III IFN response in hepatocytes, and a predominantly type I IFN response in enterocytes. These data demonstrate the complexity by which both host and viral factors influence the IFN response during HEV infection.

Keywords: U-rich region RNA PAMPs; hepatitis E virus (HEV); retinoic acid-inducible gene I (RIG-I); type I interferon (IFN); type-III IFN.

FIG 1

Differential induction of type I (α) and type III (λ1/3) IFN mRNA expression levels in liver tissues of conventional and gnotobiotic pigs experimentally infected with a genotype 3 human HEV. Type I (IFN-α) and type III (IFN-λ1/3) IFN mRNA expression levels in uninfected or HEV-infected conventional (N = 2 uninfected, 4 infected) (a) or gnotobiotic (N = 5 uninfected, 6 infected) (b) pigs at 4 weeks postinfection (wpi) were quantified using gene-specific RT-qPCR. Swine RSP32 was used as a housekeeping control. Fold change was calculated using the threshold cycle (2^−ΔΔCT) method. *, P < 0.05 versus uninfected pigs. The data represent means ± standard errors of the means (SEMs).

FIG 2

HEV 3′ UTR induced higher levels of IFN mRNAs in Huh7-S10-3 liver cells. Type III IFN (hIFN-λ1) (a) and type I IFN (hIFN-β) (b) mRNA levels in plain Huh7-S10-3 cells stimulated by different HEV RNA PAMPs (SL1cap, SL1, SL2, or SL3) at 18 h poststimulation were quantified using gene-specific qPCR. The fold change was calculated compared to unstimulated cells (mock), using the 2^−ΔΔCT method. RPS18 was used as a housekeeping control. (c) IFN-β promoter activity in HEV RNA PAMP-stimulated Huh7-S10-3 cells and unstimulated (mock) cells. Human IFN-β promoter firefly luciferase was used as a reporter plasmid, and TK-Renilla luciferase was used as a control vector. The IFN-β promoter activity was calculated by determining the ratio of firefly luciferase (FLuc)/Renilla luciferase (RLuc) levels as measured by Dual-Glo kit. **, P ≤ 0.01 versus HEV SL3 using paired Student t test. The data represent means ± SEMs of results from three independent experiments.

FIG 3

Establishment of Tet-On RIG-I Huh7-S10-3 liver cells. The Tet-On RIG-I Huh7-S10-3 cells were established by transducing human Huh7-S10-3 with pLIX402 lentivirus containing hRIG-I CDS under a doxycycline-inducible promoter. RIG-I mRNA levels (a) and RIG-I protein levels by Western blotting (b) in plain Huh7-S10-3 cells and Tet-On RIG-I cells in the presence or absence of doxycycline. The Tet-On RIG-I cells were stimulated with RIG-I agonist 5′ppp ligand (5ppp) or left unstimulated (mock), and cultured in the presence (Doxy+ve) or absence (Doxy-ve) of doxycycline (5 μg/ml). (c) IFN-β promoter activity at 12 h poststimulation of Tet-On RIG-I cells with 5′ppp or left unstimulated. IFN-β promoter activity was calculated by determining the ratio of firefly luciferase (FLuc)/Renilla luciferase (RLuc) levels as measured by Dual-Glo kit. (d) Type III IFN (IFN-λ1) mRNA expression levels in 5′ppp-stimulated Tet-On RIG-I cells at 6 h and 18 h poststimulation were quantified using gene-specific RT-qPCR. The fold change was calculated compared to unstimulated cells (mock) using the 2^−ΔΔCT method. RPS18 was used as a housekeeping control. *, P < 0.05; **, P < 0.01 versus mock. Data represent means ± SEMs of results from three independent experiments.

FIG 4

Type I and type III IFN mRNA expression levels in HEV RNA PAMP-stimulated Tet-On RIG-I Huh7-S10-3 liver cells. Type I IFN (hIFN-β) (a) and type III IFN (hIFN-λ1) (b) mRNA expression levels in HEV RNA PAMP (SL1cap, SL1, SL2, or SL3)-stimulated Tet-On RIG-I cells at 6 h and 18 h poststimulation were quantified using gene-specific RT-qPCR. (c) RIG-I mRNA levels at 18 h poststimulation. The fold change was calculated compared to the unstimulated cells (mock) using the 2^−ΔΔCT method. RPS18 was used as a housekeeping control. ^a, P < 0.05; ^aa, P < 0.01 versus HEV SL3-stimulated cells. Data represent means ± SEMs of results from two independent experiments.

FIG 5

IFN-λ1 mRNA expression levels in SL3- and SL2-stimulated Huh7-S10-3 cells under RIG-I knockdown conditions. Huh7-S10-3 cells were transfected with control siRNA (siCnt) or RIG-I siRNA (siRIG). At 24 h posttransfection, the cells were transfected with SL3 (a and c) or SL2 (b and d). The mRNA expression levels of IFN-λ1 and RIG-I were estimated using gene-specific RT-qPCR. The fold change was calculated compared to unstimulated cells (mock), using the 2^−ΔΔCT method. RPS18 was used as a housekeeping control. (e) RIG-I protein levels in siCnt- and siRIG-transfected cells at 24 h posttransfection. **, P < 0.01 versus SL3/SL2-stimulated cells; ^a, P < 0.05; ^aa, P < 0.01 versus siCnt+SL3/SL2-stimulated cells; NS, not significant. Data represent means ± standard deviations (SDs) of results from two independent experiments.

FIG 6

Phosphorylated IRF3 levels in HEV RNA PAMP-stimulated Tet-On RIG-I Huh7-S10-3 liver cells. Tet-On RIG-I Huh7-S10-3 cells were stimulated with various HEV RNA PAMPs (SL1cap, SL1, SL3), 5′ppp (RIG-I agonist), or left unstimulated (mock) in the presence (Doxy+) or absence of doxycycline. The cells were harvested at the indicated time points and probed for phospho-IRF3 (pIRF3) and total-IRF3 (IRF3) by Western blotting. β-Actin or GAPDH was used as a loading control. (a and c) Representative Western blots of p-IRF3 and IRF3. (b and d) Fold changes in pIRF3/IRF3 as estimated by densitometric analyses of the Western blots. Data in panel d represent means ± SDs of results from two independent experiments.

FIG 7

IRF3 nuclear translocation in HEV RNA PAMP-stimulated Huh7-S10-3 cells. (a) Secondary antibody control. Plain Huh7-S10-3 cells were left unstimulated (b) or stimulated with 200 ng of 5′ppp (RIG-I agonist, as a positive control) (c), SL1cap (d), SL1 (e), SL2 (f), or SL3 (g). At 18 h poststimulation, the cells were stained by IFA for IRF3 (red), and the nuclei were counterstained using DAPI (blue). Representative IRF3 nuclear localization in panels is indicated by white arrows.

FIG 8

The U-rich region is essential for HEV UTR-induced IFN response. (a) Schematic representation of various lengths of SL1 and SL3 HEV RNA PAMP constructs. Capped-SL1 (SL1cap; 250 nt), SL1 (250 nt), SL1-169 (169 nt), SL1-85 (85 nt), SL3 without poly(A) tail (SL3w/oA; 85 nt), SL3 [contains poly(A) tail; 169 nt]. Black-filled triangle represents 5′ cap in SL1cap. (b and c) IFN-λ1 mRNA expression levels in HEV RNA PAMP-stimulated cells at 18 h poststimulation were estimated using gene-specific RT-qPCR. The fold change was calculated compared to unstimulated cells (CC) using the 2^−ΔΔCT method. RPS18 was used as a housekeeping control. **, P < 0.01 versus SL3-stimulated cells; ^aa, P < 0.01 versus SL3w/oA-stimulated cells. Data represent means ± SDs of results from two independent experiments.

FIG 9

U-rich region variability and secondary structure stability of HEV UTRs. Genetic variability in the U-rich regions among genotypes 1 to 8 HEV genomic sequences was determined by a multiple-sequence alignment of SL1-85 (85 nt) (a) and SL3 without poly(A) tail (SL3w/oA; 85 nt) (b) using the MEGA6 software program (36). The U residues are marked in red, and the C residues are marked in blue. The initiation codon AUG of HEV ORF1 in SL1-85 is underlined. The secondary structures of SL1-85 (c) and SL3w/oA (d) were predicted using the mFOLD server.

FIG 10

The HEV 3′ UTR induced predominantly type I IFN responses in enterocytes IPEC-J2 cells. Swine type III IFN (swIFN-λ1) (a) and type I IFN (swIFN-β) (b) mRNA expression levels in HEV RNA PAMP-stimulated IPEC-J2 cells at 18 h poststimulation were estimated using gene-specific RT-qPCR. Fold change was calculated using the 2^−ΔΔCT method. Swine RPS18 was used as a housekeeping control. *, P ≤ 0.05 versus SL3 using paired Student t test. Data represent means ± SEMs from two independent experiments.