Impact of IL-27 on hepatocyte antiviral gene expression and function

Background: Interleukin (IL)-27 is a member of the IL-6/IL-12 family of cytokines. It is a potent cytokine, with potential antiviral impact, and has been shown to play a role in modulating functions of diverse cell types, including Th1, Th2, and NK and B cells, demonstrating both pro- and anti-inflammatory roles. In hepatocytes, it is capable of inducing signal transducer and activator of transcription (STAT)1, STAT3 and interferon-stimulated genes. Methods: To address its role in viral hepatitis, the antiviral activity of IL-27 against hepatitis C virus (HCV) and hepatitis B virus (HBV) was tested in vitro using cell-culture-derived infectious HCV (HCVcc) cell culture system and the HepaRG HBV cell culture model. To further investigate the impact of IL-27 on hepatocytes, Huh7.5 cells were treated with IL-27 to analyse the differentially expressed genes by microarray analysis. Furthermore, by quantitative PCR, we analyzed the up-regulation of chemokine (CXCL)-10 in response to IL-27. Results: In both HCV and HBV infection models, we observed only a modest direct antiviral effect. Microarray analysis showed that the up-regulated genes mostly belonged to antigen presentation and DNA replication pathways, and involved strong up-regulation of CXCL-10, a gene associated with liver inflammation. Overall, gene set enrichment analysis showed a striking correlation of these genes with those up-regulated in response to related cytokines in diverse cell populations. Conclusion: Our data indicate that IL-27 can have a significant pro-inflammatory impact in vitro, although the direct antiviral effect is modest. It may have a potential impact on hepatocyte function, especially chemokine expression and antigen presentation.


Introduction
Interleukin (IL)-27 is a member of the IL-6/IL-12 family of cytokines that includes IL-12 and IL-23. It is secreted by antigen presenting cells and has diverse impacts on host immune responses, including inducing the formation of antiviral Th1 cells 1,2 . It also inhibits Th2 humoral immune responses 2 and inhibits Th17 cells by blocking IL-17A production, and preventing up-regulation of RORγT 3 . Recent data suggest it may also act on CD4+ T cells to negatively regulate function in tuberculosis 4 .
IL-27 has also been described to possess antiviral functions. It inhibits HIV replication in CD4 cells and has been reported to control hepatitis C virus (HCV) in the cell-culture-derived infectious HCV (HCVcc) model, while in hepatitis B virus (HBV) infection, the virus activates IL-27 and interferon (IFN)λ1 and these co-ordinate to inhibit HBV replication 5,6 . In hepatoma cell lines, human and primary rat hepatocytes, the antiviral activity is thought to be induced by the induction of signal transducer and activator of transcription (STAT)1 and STAT3, leading to the induction of IFN regulated proteins, such as interferon response factor (IRF)-1, IRF-9, guanylate binding protein 2 and myxovirus resistance A 6,7 .
In this study, we used the HCVcc model and the HepaRG cell line (an HBV infection model) to address the antiviral impact of IL-27 on hepatocytes, and used a genome wide microarray to analyse the interplay of different genes that were regulated after stimulation with IL-27.
Huh7.5 cells were infected with genotype 2a chimeric HCV (MOI 0.05; J6/JFH1) (kindly provided by Apath) 8 . Infected hepatocytes were treated with IL-27 at 100ng/ml and controls cultured with PBS as a control for up to 20 days. Cells were stimulated with IFNa (R&D systems Europe) at 1000u/ml as a positive control for antiviral activity. Immuno-fluorescence assays (IFA) were performed over this timecourse; cells were probed with 9E10 mouse monoclonal anti-non-structural protein 5 (NS5A) antibody (a gift from Dr. Charles Rice, Rockefeller University, USA) conjugated to Alexa 488 for the presence of HCV by fluorescence microscopy. Visualisation was performed using a Nikon Eclipse TE2000-U inverted microscope at x20 magnification using a Nikon DXM1200F camera. The software used for taking pictures was ACT-1 v2.63 and Adobe photoshop CS4 was used to count the infected cells. To obtain viral load genome equivalents, HBV DNA was isolated from the inoculum using Invitrogen PureLink Viral DNA Mini kit (cat# 12280050), following the manufacturer's instructions. The viral DNA was then quantified using forward: 5'-GGT CTC TTT CGG AGT GTG GA-3'; reverse: 5'-ATA GGG GCA TTT GGT GGT CT-3' primers. qPCR was performed on a Roche 480 light cycler machine and using the Roche LC480 SYBR green master mix (cat# 04707516001) as below.

HBV infection
HepaRG cells were infected with the above inoculum with a genome equivalent of 20GE/cell overnight at 5% CO 2 at 37°C. The infected HepaRG cells were treated with IL-27 (100ng/ml) or IFN-α (1000IU/ml) (R&D systems Europe).

HBV ELISA
At day 7 post infection hepatitis B surface (HBs) antigen levels were assayed using MONOLISA HBs Ag Ultra Kit (Bio-Rad), according to the manufacturer's instructions.
RNA extraction and relative reverse transcription quantitative (q)PCR analysis Total RNA was prepared using RNeasy kits (Qiagen). In-column DNAse treatment was performed. The quality of RNA was checked using Agilent Technologies 2100 Bioanalyser.
Two-step reverse transcription was performed using Superscript III Reverse Transcriptase (Invitrogen) and qPCR was performed using Roche Light Cycler 480 to detect CXCL10, low molecular mass peptide 7 (LMP7) and transporter associated with antigen processing 1 (TAP1). Primers were designed using the Roche Universal Probe Library system as follows: Relative gene expression was calculated using the comparative cycle threshold method, as described previously 9 .
Gene expression and statistical analysis Gene expression profiles were obtained by hybridising the samples to GeneChip Human Gene 1 STU Arrays (HuGene-1_0-st-v1; Affymetrix), according to the manufacturer's instructions.
Statistical testing was performed using Linear Models for Microarray Analysis (LIMMA) package (http://bioconductor. org/packages/release/bioc/html/limma.html) 10 . Raw p-values were corrected using the false discovery rate controlling procedure of Benjamini and Hochberg 11 . Following this correction, adjusted p-values <0.01 were considered significant. Gene annotation was added to the final probe list from the NetAffx™ Analysis Center (https://www.affymetrix.com/analysis/index.affx).

Antiviral studies
In the HCVcc model, it was previously demonstrated 6 that IL-27 had antiviral properties. Using the same model, we have demonstrated that there is no significant reduction in viral infectivity at 10 days post treatment with IL-27 at 100ng/ml, as observed by an immuno-fluorescence assay (IFA) of infected hepatocytes ( Figure 1A). However, at 16 days a modest impact of IL-27 on infection was observed with 73.8% of cells infected compared to controls (normalized to 100%; p=0.0116 t test; Figure1B). Analysis of viral RNA at day 10 revealed a small difference between treated and untreated cells ( Figure 1C); however, the effect is overall limited compared to the IFNγ positive control ( Figure 1B) and was not significantly different when measured at a later time point (day 13).
Similarly, at day 7 post-infection, HepaRG cells infected with HBV showed modest reductions in HBsAg levels after treatment with IL-27 ( Figure 1D).

Gene expression studies
To further understand the impact of IL-27 on hepatocytes, we next addressed the gene expression changes occurring during IL-27 treatment. Microarray analysis showed that IL-27 significantly induced (>two-fold) a total of 446 genes, while 129 genes were down-regulated in Huh 7.5 cells, 72 hours post-stimulation (Supplementary Table 1). Table 1 shows the top 20 differentially regulated genes obtained.
Amongst the top 20 hits in our data were a number of signaling pathway genes notably Rab-like protein 2A (RABL2a) (a GTPase that mediates signal transduction), neurotensin (NTS) and signal peptidase complex subunit 2 homolog (SPCS2). Cell cycle proteins, e.g. the calmodulin binding ASPM gene, CDC6, which regulates DNA replication, along with DNA replication checkpoint gene, CDC45L, were also up-regulated. Other relevant up-regulated genes were those involved in inflammation, including CXCL-10, which is known to be IFN responsive, orosomucoid (ORM)-1/2 and lysozyme (LYZ) ( Table 1).
The DAVID program was used to analyse the data set to identify pathways that are differentially regulated in response to stimulation by IL-27. Kegg functional annotation analysis within this program (Supplementary Table 2A) of the up-regulated genes, showed high fold-enrichment and significance in genes involved in DNA replication, the cell cycle and homologous recombination. Gene functional classification classified genes into 14 clusters, the largest, with a high enrichment score, included genes involved in the cell cycle, cell division, DNA replication, the spliceosome, and nucleic acid metabolism, and a strong signal from the proteasomal signaling pathway (Supplementary Table 2B).
Analysing the down-regulated genes, gene functional classification showed only one gene cluster with a significant score, which included transcriptional regulators v-maf musculoaponeurotic fibrosarcoma oncogene homolog MAF, early growth response 1 (EGR1) and Jun-D proto-oncogene (JUND). NR1D1, a nuclear receptor and transcription repressor involved in circadian regulation, which is normally highly expressed in the liver, was also reduced in response to IL-27. Target genes of NR1D1 are ApoA1 and ApoCIII and anti-fibrinolytic factor PAI-1, while NR1D2 is involved in the control of lipid and energy homeostasis in skeletal muscle 13 .

Gene set enrichment analysis
Gene set enrichment analysis (GSEA) was used to understand the significance of IL-27 induced gene expression, in comparison with other published experimental datasets. Genes differentially regulated in response to IL-27 were compared against a list of immunological signatures from the Broad Institute GSEA database. We found linked gene expression sets in experiments using related cytokines on distinct cell types. Figure 2 shows representative enrichment plots of relevant gene signatures from datasets with significant correlation to our experiments on Huh7. Table 3 give a more comprehensive list (top 20 most significant enriched gene sets) of experiments that have similar differentially up (S-2A) or down (S-2B) regulated genes, respectively. Overall, this shows clear overlapping gene sets between genes induced in our experiments and cellular activation by lymphocytes in response to related cytokines.

cells (CD8+ T cells and NK cells following treatment with IL-15, a related cytokine in vitro). Supplementary
Quantitation of CXCL-10, TAP-1 and LMP7 One of the genes that was up-regulated by IL-27 was CXCL-10 -an IFNγ-responsive gene. Therefore, as a test of principle and to reconfirm the results obtained by microarray analysis, we tested the level of gene expression of CXCL10 in hepatocyte lines in response to IL-27. qPCR analysis showed that the levels of CXCL-10 increased with time, reaching a peak at about 24 hrs post-stimulation (p=0.0001; Figure 3A).
Since proteasomal signaling emerged as an enriched pathway, TAP1 and LMP7, important proteasomal genes involved in MHC class I presentation, were also examined by qPCR 14 . We addressed this by analyzing the amount of TAP1 in cells stimulated with IL-27 over time. qPCR analysis showed that the expression levels of TAP1 increased from 6 hrs post stimulation (p=0.0001) and LMP7 significantly increased increased from 24 hours (p=0.0132) post-stimulation with IL-27 ( Figure 3B).

Discussion
IL-27 functions to activate diverse intracellular pathways in hepatocytes, but only a modest impact on viral replication was demonstrated in this study with HCVcc, and similarly for HBV. Consistent with this result, the gene expression dataset did not show a marked up-regulation of classical antiviral genes. This is different to previous findings using different cell models 7 , where IL-27 has been shown to post-translationally modulate phosphorylation of STAT1/3 in HepG2 cells. Other studies have also proposed that IL-27 might possess antiviral functions similar to IFNα 5,7,15 . The fact that we did not observe a similar up-regulation of gene expression in our study could be due to differences in the cell lines used in the two studies. Consistent with this, in the above report another cell line within the same study did not show similar upregulation of STAT phosphorylation 7 . However, as suggested by others, IL-27 may synergize with other antiviral treatments, such as IFNα, for therapy for HBV 16 . One further possible difference between studies was the time-point assessed -at 72hrs we may have missed some early antiviral gene expression, although overall the data is consistent with the lack of substantial impact of IL-27 in virus culture experiments over 1-2 weeks. We chose 72 hrs based on experience with other antiviral gene expression studies responses (e.g. IFN-lambda and alpha), and RNAseq studies in the HCVcc system, where we have used time-points up to day 10 17 .
Although we did not observe a clear antiviral gene set activated, we did observe some specific responses within the hepatocytes, such as up-regulation of CXCL10, TAP1 and LMP7. These responses were confirmed as relevant by comparison with other datasets using GSEA, which revealed consistent patterns of response in diverse cell types, in response to related cytokines. CXCL-10 is associated with the expression of IFNγ, but can be induced by other cytokines, and it has been shown that IL-27 is able to induce CXCL-10 in bronchial epithelial cells 18  The overall impact of IL-27 in vivo is still unclear. IL-27 has been observed to increase in patients with chronic HBV 20 , and has also been shown to be able to modulate immune responses to prevent hepatic injury 21 . Our in vitro data indicate it may have a potential impact on hepatocyte chemokine secretion and MHC class I antigen presentation, and thus in vivo studies of the role of IL-27 in modulating hepatocyte interaction with host CD8+ T cell responses may be of value in future.

Data availability
IL-27 control dataset available from NCBI GEO (accession number, GSE89610; https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE89610) Author contributions NR and PK planned the study and wrote the manuscript. NR and SB carried out the experimental protocols. NR and RA performed qPCR in Figure 3. DB performed the microarray, and NS and HL analysed the initial bioinformatics from the microarray data. EM helped with the final analysis of the bioinformatics data and generated the GSEA data.

Competing interests
No competing interest were disclosed.

Grant information
This work was supported by the Wellcome Trust [WT091663MA, WT109965MA, and 090532/Z/09/Z].
The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. The choice of IL-27 dosing as well as the treatment duration is unclear; rationales need to be given here.
The immunofluorescent staining in Figure 1 needs a DAPI signal to indicate HCV negative cells.
There are discrepancies between the text and the Figure. Page 4: "… however, the effect is overall limited compared to the IFNγ positive control ( Figure 1B) and was not significantly different when measured at a later time point (day 13)." There is no IFNγ mentioned in Figure 1.
The HBV data given in Figure 1D are insufficient to draw any conclusions on IL-27 treatment. These data are of minor quality and need to be complemented (HBeAg, HBV DNA) or discarded.
The number of replicates and repetitions within diverse experiments is not clear, especially for the gene array. Furthermore, the validation was performed with a single gene (CXCL-10) at a diverse time point.
The links in the legend of Figure 2 seemed to be wrong. No IL-15 treatment is mentioned there.
Drawing conclusions on antiviral responses of hepatocytes, while using Huh7 hepatoma cell line, is false. Here only suggestions can be made.
I have read this submission. I believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard, however I have significant reservations, as outlined above.
No competing interests were disclosed. Competing Interests: 05 December 2016 Referee Report PCR and IF and demonstrates no effect on HCV replication. The reasons for such discrepancy are not clear. Cell lines could have acquired or lost characteristics over time and could behave differently. Do we know the genotype of this cell line? Is it IL28B CC or TT?
Gene expression results shown in Figure 2 shows overlapping patterns with IL-15 induced gene expression in CD8 and NK cells. I am not sure what this means since gene expression in different cell types are likely to be controlled by cell specific transcriptional patterns and not comparable to effects of different cytokine on completely different cell types. This would benefit form addition of a description that will explain the results better.
Also, what was the rationale to choose 72 hours post treatment for gene expression analysis? Earlier studies have shown immediate induction of ISGs with Interferons and other antiviral cytokines.
Cell type specific expression of ISGs have been associated with favorable responses to interferon-alpha in patients. For example, higher expression of ISGs in hepatocyte is associated with a poor response to IFN-alpha, while a higher expression of ISGs in monocytes is associated with SVR. Hence, a lack of ISG expression in this hepatocyte cell line may be associated with an overall favorable host ISG response. Effects of IL-27 on other cell types may provide clues on the net effect of this cytokine on host innate response to HCV.
I have read this submission. I believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard, however I have significant reservations, as outlined above.
No competing interests were disclosed. Competing Interests: