Binding of IFNλs to their receptor complex leads to activation of downstream signalling cascades that ultimately lead to the establishment of an antiviral state (6). The JAK/STAT pathway is one of the most critical and well-characterised pathways activated following IFNλ binding. An emerging view is that the kinetics of such a downstream response is a crucial determinant of the antiviral potential of IFNλs (11, 20). To probe the temporal basis of IFNλ signalling in greater detail, we first measured phosphorylation of STAT1 over time at Y701 ( Figure 1 ). Human hepatocyte HepaRG monolayers were incubated with conditioned media estimated to contain equivalent amounts of IFNλs (IFNλ3, IFNλ4 WT, P70S, L79F, and K154E) for 15, 30, 60, 120, and 360 min, and 24 h. Following stimulation, protein lysates were harvested, and STAT1 phosphorylation was assayed by immunoblot analysis ( Figure 1A ). Conditioned media generated following transfection of an enhanced green fluorescent protein (EGFP)-expressing plasmid served as a negative control. Results showed that IFNλ3 and a number of IFNλ4 variants induced detectable levels of pSTAT1 ( Figure 1A , quantified in Supplementary Figure 1A ). L79F gave extremely low levels of pSTAT1 (data not shown), which likely correlates with its very limited activity as described previously (17). Interestingly, IFNλ3 and IFNλ4 variants induced distinct kinetics of pSTAT1 activation ( Figure 1A and Supplementary Figure 1A ). Whilst all IFNλs peaked around similar times (30 min to 1 h), IFNλ4 WT and P70S showed clear transient activation, whilst IFNλ3 and K154E displayed persistent activation of pSTAT1. Importantly, levels of pSTAT1 correlated with previously measured antiviral potential for three IFNλ4 variants K154E > WT > P70S (17). As IFNλs can also signal in other tissues apart from the human liver (8, 21), and there is an emerging role for IFNλ4 in extra-hepatic environments, we assayed whether human colon carcinoma cells (T84) were capable of inducing pSTAT1 in response to IFNλ4 and its variants. Intestinal T84 cells were treated and incubated with conditioned media containing equivalent amounts of IFNλs (IFNλ3, IFNλ4 WT, P70S, and K154E), and their induction of pSTAT1 was assayed over time by immunoblot analysis ( Figure 1B and Supplementary Figure 1B ). We observed similar trends as to HepaRG, although differences in amplitude of pSTAT1 induction were noted, especially for IFNλ4 K154E in T84 cells. Together, these results show that both hepatic and intestinal cell lines can respond to both IFNλ3 and IFNλ4 and display variant-specific inductions of the JAK/STAT pathway.

Phosphorylation of STAT1 following receptor complex engagement by IFNλs results in STAT1/2 dimer formation and translocation to the nucleus to induce ISG transcription, which ultimately leads to the production of antiviral proteins and the establishment of an antiviral state (6). Our previous work showed that IFNλ variants induced different levels of ISG expression when measured at 24 h (17). To ascertain whether this ISG expression varied at earlier times after incubation in concert with the kinetics of pSTAT1 activation, we measured the relative induction of a panel of core ISGs (IFIT1, MX1, ISG15, and RSAD2/VIPERIN) compared to EGFP-treated conditioned media in HepaRG cells ( Figures 2A–D ) and T84 cells ( Figures 2E–H ). Compared to EGFP-conditioned media stimulated cells, all IFNλs induced measurable increases in ISG mRNA in HepaRG cells but with discernible differences in magnitude. T84 cells also showed ISG induction for four of the supernatants tested ( Figures 2E–H ). Additionally, looking at relative fold change, T84 cells gave a lower induction of all ISGs as compared to HepaRG cells ( Figure 2 ). The magnitudes of ISG induction for both cell lines mirrored the pSTAT1 induction that was observed in Figure 1 (IFNλ3/K154E > WT > P70S > L79F). IFNλ4 K154E induced a similar pattern of ISG induction as IFNλ3 in both cell lines. Interestingly the kinetics of ISG induction was distinct to each cell line. In HepaRG cells, all IFNλs induced an early peak induction of ISGs, which subsequently declined over time. Moreover, IFNλ4 K154E demonstrated a slightly faster induction and peaked by 2 h whilst all other IFNλs tested peaked at 6 h. By contrast, IFNλ3 and the IFNλ4 K154E showed no or little decline in ISG induction after induction at either 2 or 6 h in T84 cells ( Figures 2E–H ). Additionally, T84 cells yielded low to almost undetectable induction of ISGs following IFNλ4 WT and P70S treatment. Together, these results show that K154E provides similar stimulatory activity to IFNλ3 and that this is far greater than for either IFNλ4 WT or P70S, which are the most common IFNλ4 variants in the human population.

Induction of an antiviral state is the major downstream consequence of IFN signalling. To determine how STAT1 phosphorylation and ISG expression correlate with antiviral activity, we infected the hepatic and intestinal cell models with two different viruses, EMCV and VSV. Both EMCV and VSV are highly cytopathic, replicate very fast, and are sensitive to IFN, which makes them suitable for assessing the kinetics of antiviral activity. EMCV infectivity and replication were assayed by determining the cytopathic effects of the virus, whilst a VSV encoding luciferase (VSV-luc) was deployed and its infectivity was measured by luciferase assay. HepaRG and T84 cells were treated with increasing concentrations of EGFP or IFNλ3 or IFNλ4-containing supernatants at 24 h prior to virus infection. Following IFNλ pretreatment, cells were infected with EMCV or VSV [multiplicity of infection (MOI) of 0.3 and 1, respectively] in the continuous presence of IFNλs, and infection was assayed at 24 h post-infection for EMCV ( Figure 3A ) and 8 h post-infection for VSV ( Figures 3B, C ). Different assay times for VSV versus EMCV were due to differences in replication kinetics and cytopathic effects of either virus. Consistent with our previous work (17, 21), results show that VSV infection was inhibited by all IFNs in both cell lines ( Figures 3B, C ). IFNλ3 was the most potent IFN, as it reduced VSV infection with 10% of the maximum concentration in both HepRG and T84 cells. IFNλ4 WT and K154E showed similar antiviral activity; however, a much higher concentration of these two IFNs was required to reach a similar potency as IFNλ3. Consistent with previous low pSTAT1 and ISG inductions, P70S was only able to slightly reduce virus infection even at the highest concentrations in both cell lines. T84 cells were poorly infected with EMCV and highly resistant to the cytopathic effects of the virus, and therefore, antiviral activity was not assayed in this cell line, but similar patterns of antiviral activity were seen for HepaRG cells infected with EMCV ( Figure 3A ).

Having established antiviral assays in both liver- and intestinal-derived cell lines, we wished to determine whether IFNλ activity was time dependent and whether the continuous presence of IFNλs was required to maintain their antiviral activity. Therefore, we performed infections and antiviral assays over time, both in the continuous presence of IFNλs but also in cells that had been pretreated with IFNλs for varying lengths of time, yet the cytokines were then removed, monolayers washed, and fresh media provided (“non-washed” and “washed”, respectively, Figure 4A ) prior to infection. Initially, we conducted experiments in T84 cells that were infected with VSV following IFNλ pretreatment ( Figures 4B–E ). For clarity, it should be noted that the following data are presented differently than those in Figure 3 . In agreement with the data presented in Figure 3 , all IFNλs demonstrated antiviral activity with IFNλ3 and IFNλ4 P70S showing the greatest and least potency, respectively. The peak of IFNλ3 activity was delayed relative to all IFNλ4s. Moreover, we found that shorter incubation times with IFNλ3 followed by its removal before infection reduced its antiviral activity to a greater extent compared to the three IFNλ4 variants used in the experiment (compare early time points in Figure 4B with Figures 4C–E ); comparison of activity between washed and unwashed was significant (t-test) at all time points for IFNλ3, whilst for IFNλ4s, it was only significant for the first or second time points. In HepaRG cells infected with EMCV, we observed a similar pattern, i.e., removal of IFNλ3 after relatively short incubation (2 h) with cells gave a greater reduction in antiviral activity compared to the same timepoint for the IFNλ4 variants ( Figures 4F–I ); when compared with the difference between washed and unwashed for IFNλ4, that of IFNλ3 reached statistical significance (t-test). In addition, we observed that all IFNλs generally gave less antiviral activity after removal at the time of infection compared to activities in the continuous presence of the proteins. From these experiments, we suggest that IFNλ4 proteins may be more tightly bound to the heteromeric cell receptor as compared to IFNλ3. Alternatively, signalling with IFNλ4 is maintained for a longer period as compared to IFNλ3. To further assess the contribution of IFN-cell contact time compared to signalling time, we repeated the wash experiments in HepaRG cells with EMCV, but, on this occasion, cells were incubated with the IFNλs at 24 h prior to infection but then removed by washing at differing times before virus addition ( Supplementary Figure S2 ). The results show a greater decline in antiviral activity (~15-fold) from 24 h incubation to 6 and 2 h incubation for IFNλ3 compared to IFNλ4 WT and IFNλ4 K154E, which showed reductions in activity by ~0.75-4 fold.

Our data suggest that in human cells, human IFNλ4 and its variants induce a distinct antiviral response compared with human IFNλ3. As previous work has demonstrated that IFNλ4 from different primate species have varying levels of antiviral activity (17, 22), we next explored whether the distinct signalling kinetics that we observed were also species specific. We first analysed the amino acid homology between IFNλ3, IFNλ4, IFNλR1, and IL10R2 in humans, chimpanzees, and rhesus macaques ( Figure 5A ). Results showed that although the various orthologues shared a high degree of homology (92-97%), there were differences that could affect activity given that even a single amino acid change can alter signalling as in IFNλ4 variants P70S and K154E. Given these genetic differences, we next tested the antiviral kinetics of non-human IFNλs. First, we treated human HepaRG cells with human and non-human IFNλs as described in Supplementary Figure 2 , by treating cells for 2, 6, and 24 h, and then removing the cytokines prior to infection with EMCV at 24 h after initial stimulation ( Figure 5B ). For these experiments, we utilised non-human primate IFNλ3 or IFNλ4 proteins containing a C-terminal FLAG tag, which we characterised previously (17). In these experiments, we utilised IFNλ4 K154E as a model human IFNλ4, since it gave robust levels of detectable antiviral activity, with kinetics broadly similar to IFNλ4 WT. All IFNλs had antiviral activity against EMCV with chimpanzee IFNλ4 having greater activity than human and macaque IFNλ4 ( Figure 5C ), whilst human IFNλ3 had greater activity than macaque IFNλ3 ( Figure 5D ). Similar to human variants, the peak of IFNλ3 activity was delayed relative to all IFNλ4s. IFNλ3 washing experiments demonstrated that like human IFNλ4, non-human primate IFNλ4 were more refractory to early removal than human or macaque IFNλ3 ( Figures 5C, D ). To determine if these characteristics also occurred in non-human cells, we repeated these experiments in the rhesus macaque respiratory epithelial cell line LLCMK2 ( Figures 5E, F ). Results showed that all IFNλ4s had similar kinetics of antiviral activity but different levels of potencies as found in HepaRG cells. Washing following immediate infection supported the initial washing experiments with IFNλ3 antiviral activity being more sensitive to early removal of cytokine ( Figures 5G, H ).

Complex and dynamic interactions between cytokine ligands and their cognate receptors dictate the signalling output (23). To probe further the molecular genetic basis of IFNλ kinetics, we sought to mutate and disrupt the receptor binding faces of IFNλ, hypothesising that these residues were most likely to be responsible for IFN kinetics. IFNλ4 is highly divergent when compared with IFNλ1-3 with ~30% similarity detected, suggesting that there are likely to be distinct molecular determinants of differential signalling contained within IFNλ4 compared to the other human IFNλs (13, 15). To begin to identify those determinants, we constructed chimeric IFNλs between IFNλ4 and human IFNλ1. IFNλ1 was chosen, as it is known to have similar kinetics to IFNλ3 (10) but, like IFNλ4, is N-linked glycosylated (6). Initially, comparison of differentially conserved amino acids in IFNλ4 (human and non-human primate) with IFNλ1-3 (human and macaque) identified a divergent receptor binding interface between these groups of IFNλs suggestive of distinct receptor interactions ( Figure 6A ). We focused on divergent, likely surface-exposed residues near relevant helices (A, D, and F) and designed two chimeric IFNλs based on IFNλ1 containing candidate IFNλ4 residues from the IFNλR1-binding helix F (F) and the IL10R2-binding helices A and D (AD). An additional chimera with all three IFNλ4 binding helices was generated, termed ADF.

We first confirmed that IFNλ1 and its chimeras were produced and released into the supernatant using a split-luciferase assay; this showed that chimeras incorporating helices A and D yielded reduced production ( Figure 6B ). To test their antiviral activity, HepaRG cells were pretreated for 2, 6, or 24 h prior to EMCV infection with the WT IFNs and each of the indicated chimeras. Results showed that IFNλ1 had higher antiviral activity than IFNλ4 yet similar to IFNλ3. Additionally, IFNλ1 with IFNλ4 substitutions had reduced antiviral activity (IFNλ1 > F > AD > ADF) ( Figures 6C, D ). To determine if the chimeras impacted IFN kinetics, HepaRG cells were pretreated for 2, 6, or 24 h prior to EMCV with the WT IFNs and each of the indicated chimeras. The IFNs were either left for the duration of the infection or removed at the time of infection, and infection was commenced either at time of cytokine removal or at 24 h after initial incubation. Importantly, IFNλ1 kinetics were similar to IFNλ3 in HepaRG cells, with increasing activity over time and a delayed peak relative to IFNλ4 ( Figures 6C, D compared with Figure 2F and Supplementary Figure 2 ). IFNλ1/4 chimeras had similar kinetic profiles as IFNλ1. Results revealed that washing reduced the antiviral potency of all IFNs and IFNλ1, and all the chimeras were more greatly affected than IFNλ4. Taken with our antiviral activity results suggested that chimera F had reduced potency compared to IFNλ1, whilst the reduced activity of AD is likely due to reduced protein, and thus, the impact on ADF is due to reduced amount and potency ( Figure 6B ). However, despite alteration of the receptor interaction surfaces, the kinetics remain conserved similar to IFNλ1 (and IFNλ3), suggesting that these residues only modify the magnitude of the antiviral response and are not sufficient to alter the antiviral kinetics.

Altogether, our work described here demonstrates the distinct yet conserved antiviral kinetics of human and non-human primate IFNλ4 compared to other IFNλs.

HEK-293T (human embryonic kidney) and LLC-MK2 (rhesus macaque respiratory epithelial cell line) were cultured in high glucose Dulbecco’s modified Eagle’s medium (DMEM) with 10% foetal calf serum (FCS) and pen/strep. HepaRG.ISG15-EGFP [human hepatocyte-like cell line modified to express EGFP under the control of the endogenous ISG15 promoter (17)] were cultured in complete William’s media with FCS (10%), human insulin (4 µg/ml), hydrocortisone hemisuccinate (50 µM), and pen/strep (1%) (HepaRG cells). T84 (ATCC CCL-248) colon carcinoma cells were cultured in a 50:50 mix of DMEM:F12 with 10% FCS and 1% pen/strep on collagen-coated cell culture dishes. All cell lines were passaged routinely following PBS washing and trypsin-mediated detachment. Cell lines were routinely screened for Mycoplasma contamination and discarded if signs of contamination were detected.

Two viruses were used in this study: Ruckart strain of encephalomyocarditis virus (EMCV) and VSV. EMCV was produced in Vero cells following low MOI infection (MOI = 0.0001) and harvested between 1 and 2 days when extensive cytopathic effect was observed. EMCV infectivity was quantified by TCID₅₀ and typically grew to titres of ~10⁸/ml. VSV-luc was a kind gift from Sean Whelan (Washington University, St. Louis) and was produced and titrated as described in (26, 27).

Commercially available primary antibodies were mouse monoclonal antibodies recognizing β-actin (Sigma #A5441), pSTAT1 (BD Transductions #612233), mouse anti-STAT1 antibody (3987, Abcam), or mouse anti-phospho-STAT1 antibody (29025, Abcam) and used at a 1:1,000 dilution. Additionally, rabbit anti-beta-tubulin antibody (6046, Abcam) was also used (1:1,000). For secondary antibodies, antimouse (GE Healthcare #NA934V), coupled with horseradish peroxidase (HRP), was used at a 1:5,000 dilution (T84) or horseradish peroxidase-conjugated goat antirabbit IgG secondary antibody (A0545, Sigma-Aldrich) at 1:2,000 dilution or horseradish peroxidase-conjugated goat anti-mouse IgG secondary antibody (A4416, Sigma-Aldrich) at 1:2,000 (HepaRG).

Recombinant DNA technology was utilised to generate the IFNs for functional testing in this study, as previously described (15, 17). The mammalian expression plasmids expressing HiBiT variants and human IFNλ3, and chimpanzee (Pan troglodytes) and rhesus macaque (Macaca mulatta) IFNλ4 with a carboxy-terminal FLAG tag were described previously (17). Rhesus macaque IFNλ3-FLAG was generated synthetically (GeneArt) with sequence corresponding to XP_001086865.3 alongside WT IFNλ1-HiBiT, or IFNλ1/λ4-HiBiT chimeras were constructed synthetically (GeneArt) with sequences from helices A, D, and F as shown ( Figure 5 ) and cloned into expression vector pC1 and sequenced confirmed by Sanger sequencing. Correct plasmids were purified by midiprep or maxiprep and quality and quantity determined by NanoDrop prior to transfection. An EGFP expression plasmid prepared in identical conditions was used as a negative control throughout.

IFNλs were produced using the protocol described previously, which is capable of generating functional IFNλs (17). Briefly, IFN expression plasmids were transfected into sub-confluent HEK-293T cell monolayers, which are hyporesponsive to IFNλ signalling due to very low expression of IFNλR1 (15). Lipofectamine 2000 was used to transfect IFNλ plasmids per manufacturer’s instructions. IFNλs were routinely generated in six-well plates or 10 cm dishes, and 2 and 14 µg of plasmids were used, respectively. Lipofectamine 2000 (2 µl) was used per microgram of plasmid. Plasmids were transfected into cells in Optimem for 16–18 hours, before changing media to growth media (10% FCS) until 2 days posttransfection was reached. The conditioned media were harvested, clarified by centrifugation, aliquoted, and immediately frozen at −80 in. Relative levels of IFNλs were estimated using the extracellular HiBiT split luciferase assay by virtue of their C-terminal HiBiT tag by incubating IFN preparations with assay reagents and measured by manufacturer’s instructions (Nano-Glo HiBiT Extracellular Detection system, Promega) using a luminometer.

For quantitative real-time PCR (qRT-PCR) or immunoblotting experiments, IFN stimulation was achieved by incubating cell monolayers with IFNλ-containing conditioned media at a defined concentration to have equivalent HiBiT signal for each sample. Cells treated with IFNs were incubated for the indicated period of time before either being processed. A previous titration analysis indicated that a ~1:2–1:4 dilution of IFNλ4-WT is enough to give a robust induction of ISGs for all variants (17) and antiviral response whilst limiting the amount of conditioned media added to cells (<50% of total volume). Therefore, WT IFNλ4 was used at the standard, and the levels of other IFNs were normalised to this by virtue of the HiBiT tag. Based on HiBiT assay measurements, the relative ratios of supernatant were IFNλ4(WT):P70S:L79F:K154E:IFNλ3, ~1:2:2:0.2:0.01. For the analysis of pSTAT1 levels by immunoblotting, 100,000 cells were seeded in 500 µl of growth media, into sterile rat-tail-collagen-coated (T84) or untreated (HepaRG) 24-well plates. To analyse ISG expression levels by qRT-PCR, 50,000 T84 cells were seeded in 500 µl DMEM/F12 into sterile rat-tail-collagen-coated 48-well plates, or 1,000,000 HepaRG cells were seeded into 2 ml of growth media into 6-well plates. Cells were treated either with HEK293T cell supernatants containing either IFNλ3 or different IFNλ4 variants [λ4 wild type (WT), K154E, P70S] or GFP-conditioned medium (Mock). Prior to treatment, media were removed; cells were rinsed once in PBS and then treated with each IFN diluted in their corresponding growth media to achieve an equal concentration (as determined by HiBiT) and added to the cells in 500−1,000 µl/well. Cells were then incubated at 37°C and 5% CO₂ until harvest. Cells were harvested at 15 min, 30 min, 1 h, 2 h, 4 h, 24 h, and 48 h posttreatment for the analysis of pSTAT1 protein levels by immunoblotting, whereas total RNA was isolated from T84 cells at 2, 4, 8, 12, and 24 h posttreatment for the analysis of ISG expression levels by qRT-PCR. To control for batch-to-batch variability in protein production, at least two independent protein preps were used.

For the EMCV antiviral assays, 5,000 cells were seeded per well in a 96-well plate 24 h prior to treatment. At the day of treatment, IFNs were added in twofold serial dilutions to the cells 24 h prior to infection. Following IFN treatment, EMCV (MOI = 0.3) was added to the cells, and infection was scored by CPE 24hpi visually or by crystal violet staining. EMCV is highly cytopathic in certain cell lines and very sensitive to IFN. The reciprocal of the dilution giving ~50% protection was used as a semiquantitative measure of IFNλ-conditioned media activity.

For VSV infection, T84 or HepaRG cells were seeded in a white bottom 96-well plate. Cells were pretreated prior to infection as indicated time points and concentrations of IFN-λ3 and IFN-λ4, and its variants K154E and P70S. VSV-luc (MOI = 1) was added to the wells, and the infection was allowed to proceed for 8 h. At the end and the infection, media was removed, cells were washed 1× with PBS and lysed with cell lysis buffer (Promega) at room temperature (RT) for 20 min. The same volume of Steady Glo (Promega) was added to the cells and incubated for 15 min. Luminescence was read using Tecan Infinite M200 Pro.

At the time of harvest, cells were rinsed once with PBS and then lysed with 1× radioimmunoprecipitation assay (RIPA) buffer [150 mM sodium chloride, 1.0% Triton X-100, 0.5% sodium deoxycholate, 0.1% sodium dodecyl sulphate (SDS), 50 mM Tris at pH 8.0 supplemented with phosphatase and protease inhibitors (Sigma-Aldrich or Thermo Fisher) for 5–10 min at RT (T84) or ice (HepaRG)]. Cell lysates were collected, and roughly equal amounts of protein were then separated by SDS-PAGE in a 10% (HepaRG) or 12% (T84) polyacrylamide gel, following boiling and reducing. Lysates were then blotted onto a nitrocellulose membrane (T84) or PVDF (HepaRG) by wet blotting. Membranes were blocked with blocking buffer [5% BSA in TBS containing 0.1% Tween-20 (TBS-T)] for 1 h at RT whilst shaking. Primary antibodies (1:1,000 dilution) were diluted in blocking buffer and incubated overnight shaking at 4°C. The membranes were washed four times in TBS-T for 10 min at RT. Then, secondary antibodies were diluted in blocking buffer and incubated for 1 h shaking at RT. Membranes were again washed four times in TBS-T for 10 min at RT. HRP detection reagent (GE Healthcare) was mixed 1:1 and incubated at RT for 2–3 min, or ECL substrate is added (Immobilon crescendo western HRP substrate, WBLUR0100, Merck). Membranes were then exposed to film and developed or visualised by chemiluminescence using the G:BOX Chemi gel doc Imaging System Instrument (Syngene). The detection of β-actin (T84) or β-tubulin (HepaRG) was used as loading controls. For quantitative analysis, pSTAT1 intensities of each immunoblot were quantified for each timepoint using ImageJ or Image Studio Lite Version 5.2. For quantification with ImageJ, the background value (Mock) was manually subtracted from the calculated values. pSTAT1 levels were then determined relative to control.

The total RNA was purified from lysed cells using the Nucleo Spin^® RNA extraction kit (T84) by Marchery-Nagel (Catalog number 740955.50) according to the manufacturer’s instructions or (HepaRG) RNeasy Mini Kit (74106, Qiagen). RNA concentration was measured using the NanoDrop Lite spectrophotometer (Thermo Scientific). For T84 cells, 250 ng of total RNA was reverse transcribed into cDNA using the iScript™ cDNA Synthesis kit (BioRad Laboratories, Catalog number 1708891). The reaction contained a mixture of 1 μl Reverse Transcriptase, 4 μl Reaction Mix, and 15 μl of RNA template in nuclease-free water. The newly synthesised cDNA was diluted 1:2 in RNase/DNase free water. The following qRT-PCR was performed using a Bio-Rad CFX96 Real-Time PCR Detection System. Per reaction 7.5 µl of SsoAdvanced Universal SYBR Green Supermix, 2 µl of 1:2 diluted cDNA, 1.7 µl of nuclease free water, and 1.9 µl of either forward or reverse primers (2 µM) for the amplification of IFIT1 (fw: 5′-AAAAGCCCACATTTGAGGTG-3′; rev: 5′-GAAATTCCTGAAACCGACCA-3′), ISG15 (fw: 5′-CCTCTGAGCATCCTGGT-3′; rev: 5′-AGGCCGTACTCCCCCAG-3′), Viperin (fw: 5′-GAGAGCCATTTCTTCAAGACC-3′ and rev: 5′-CTATAATCCCTACACCACCTCC-3′), and Mx1 (fw: 5′-GGTCTATACCACACGCACAGA-3′; rev: 5′- ACTGGTTTCCTTTGCCTCGT-3′) were used. Data analysis was performed using the Bio-Rad CFX Manager 3.0. The expression of the targeted genes was then normalised to the housekeeping gene HPRT1 (fw: 5′-CCTGGCGTCGTGATTAGTGAT-3′; rev: 5′-AGACGTTCAGTCCTGTCCATAA-3′). For HepaRG cells, 1 μg of total RNA was reverse transcribed into cDNA using the High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems, UK). The reaction contained a mixture of 1 μl Reverse Transcriptase, 9 μl Reaction Mix, and 10 μl of RNA template in nuclease-free water. The newly synthesised cDNA was diluted 1:25 in RNase/DNase free water. The following qRT-PCR was performed using a Real-Time Ready PCR Kit (Roche) and TaqMan primer–primer–probe mixes. Each reaction mixture consisted of 10 μl of 2× LightCycler 480 Probes Master, 1 μl of 20× Real-Time Ready Assay with 4 μl PCR-grade H₂O (total volume, 15 μl). Template DNA, defrosted on ice, was first diluted 1:25 (v/v) with PCR-grade H₂O and then 5 μl diluted template added per reaction tube to the probes MasterMix to give a final volume of 20 μl. TaqMan assays (Catalogue number 4331182) for IFIT1 (Assay ID: Hs03027069_s1), ISG15 (Assay ID: Hs01921425_s1), MX1 (Assay ID: Hs00895608_m1), and RSAD2/VIPERIN (Assay ID: Hs00369813_m1) were used. The expression of the targeted genes was then normalised to the housekeeping gene GAPDH (Assay ID: Hs02786624_g1). Cells treated with conditioned media from EGFP-plasmid transfected HEK-293T cells were used as a mock control, and all values were normalised against this value at each time as “fold change to EGFP.”