Cyclophilin Inhibitors Block Arterivirus Replication by Interfering with Viral RNA Synthesis

Cell culture, infection, and virus titration.

BHK-21 (28), Vero E6 (29), and MARC-145 (30) cells were cultured as described previously. 293/ACE2 cells (31) were cultured in Dulbecco's modified Eagle's medium (DMEM; Lonza) supplemented with 8% fetal calf serum, 100 U/ml of penicillin, 100 μg/ml of streptomycin, 2 mM l-glutamine, and 12 μg/ml blasticidin (PAA Laboratories). A cell culture-adapted derivative of the EAV Bucyrus isolate (32) and green fluorescent protein (GFP)-expressing recombinant EAV (33) were used to infect monolayers of BHK-21, Vero E6, and 293/ACE2 cells at a multiplicity of infection (MOI) of 5, as described previously (28, 29). MARC-145 cells were infected with a GFP-expressing recombinant PRRSV (SD01-08-GFP) at an MOI of 0.1, as previously described (34). EAV titers in cell culture supernatants were determined by plaque assay on BHK-21 cells (28), whereas PRRSV titers were determined by fluorescent focus assay (FFA) on MARC-145 cells, as described previously (35). For half-maximal inhibitory concentration (IC₅₀) determinations, cells were grown in black 96-well plates (Greiner), infected with EAV-GFP or PRRSV-GFP, and treated with compounds in octuplet. GFP reporter expression was quantified by measuring the fluorescence in a 96-well plate reader, using an excitation wavelength of 485 nm and an emission wavelength of 535 nm.

Antibodies and drugs.

Rabbit polyclonal antibodies against CypA (Abcam), CypB (Abcam), and calnexin (BD), a goat polyclonal antiserum against GAPDH (glyceraldehyde-3-phosphate dehydrogenase; Santa Cruz), and a mouse monoclonal antibody (MAb) against β-actin (Sigma) were used. Rabbit antisera recognizing the EAV replicase subunits nsp3 (36) and nsp9 (27) and the EAV membrane (M) protein (29) and a MAb against the EAV nucleocapsid (N) protein (37) have been described previously. The cyclophilin inhibitors CsA (Sigma) and Debio-064 (Debiopharm, Switzerland) were dissolved in dimethyl sulfoxide (DMSO). CsA was stored as a 50-mg/ml stock at −20°C, and Debio-064 was stored as a 10 mM stock at 4°C in aliquots for single use. The IC₅₀ of inhibitors was calculated with GraphPad Prism (version 5) software using a nonlinear regression model.

Immunofluorescence microscopy.

EAV-infected or mock-infected BHK-21 cells, grown on coverslips at 39.5°C, were fixed with 3% paraformaldehyde in phosphate-buffered saline (PBS), permeabilized with 0.1% Triton X-100, and processed for immunofluorescence microscopy as described previously (26). Specimens were examined with a Zeiss Axioskop 2 fluorescence microscope with an Axiocam HRc camera and Zeiss Axiovision (version 4.4) software.

Western blot analysis.

After SDS-PAGE, proteins were transferred to Hybond-LFP membranes (GE Healthcare) by semidry blotting. Membranes were blocked with 1% casein in PBS containing 0.1% Tween 20 (PBST) and were incubated with anti-nsp3 (1:2,000), anti-nsp9 (1:2,000), anti-M (1:2,000), anti-N (1:10,000), anti-CypA (1:1,000), anti-CypB (1:2,000), or anti-β-actin (1:50,000) antiserum diluted in PBST with 0.5% casein. Biotin-conjugated swine anti-rabbit IgG (1:2,000) or goat anti-mouse IgG (1:1,000) antibodies (Dako) and Cy3-conjugated mouse anti-biotin (1:2,500) antibodies were used for detection. Blots were scanned with a Typhoon 9410 imager (GE Healthcare) and analyzed with ImageQuant TL software.

Isolation of EAV RTC-containing replication structures and in vitro RNA synthesis assays.

EAV replication structures were isolated from BHK-21 or Vero E6 cells, and in vitro RNA synthesis assays were performed essentially as described previously (27). In short, approximately 1 × 10⁸ EAV-infected BHK-21 or Vero E6 cells were harvested at 6 or 7 h postinfection (p.i.), and cells were lysed to obtain a postnuclear supernatant (PNS) (27). A standard in vitro RNA synthesis assay mixture contained 20 μl of PNS (the equivalent of 6 × 10⁴ cells) from EAV-infected BHK-21 cells, 5 μl of an inhibitor solution, or 5 μl of RTC dilution buffer (control). Following gel electrophoresis, ³²P-labeled reaction products were analyzed by denaturing agarose gel electrophoresis and by exposing a PhosphorImager screen directly to the dried gel, after which screens were scanned with a Typhoon 9410 imager (GE Healthcare), and incorporation of label was quantified using ImageQuant TL software.

Density gradient fractionation.

Subcellular fractionation of PNS was performed in continuous 0 to 30% OptiPrep density gradients in RTC dilution buffer. The gradients were prepared in 13.2 ml Ultra-Clear centrifugation tubes (Beckman Coulter) using a Gradient Master gradient former (Biocomp). One milliliter of PNS from Vero E6 cells was carefully loaded on top of the preformed gradient. After centrifugation for 17 h at 48,000 × g in an SW41 rotor at 4°C, the gradient was fractionated into 0.5-ml fractions. The density of each fraction was determined with a refractometer (GETI).

Metabolic labeling of viral RNA synthesis.

Labeling of viral RNA with [³H]uridine was performed essentially as described previously (38). Briefly, at 4.5 h p.i., 4 × 10⁵ EAV-infected BHK-21 cells in 4-cm² dishes were given medium containing 10 μg/ml actinomycin D (ActD; Sigma-Aldrich) and either 4 μM CsA or 0.01% DMSO as a solvent control. After 1 h, viral RNA synthesis was labeled by adding 100 μCi of [³H]uridine to the medium. The ³H-labeled RNAs were isolated, separated in denaturing agarose gels, and visualized by fluorography. To verify that equal amounts of total RNA were loaded, the gel was hybridized with a ³²P-labeled oligonucleotide probe (5′-TTCACGCCCTCTTGAACTCTCTCTTC-3′) recognizing 28S rRNA, as described previously (27).

RNA interference.

ON-TARGETplus smart-pool small interfering RNA (siRNA) duplexes (Dharmacon) against CypA (PPIA; catalog no. L-004979-04) and CypB (PPIB; catalog no. L-004606-00) were used to silence CypA and CypB expression in 293/ACE2 cells. A nontargeting siRNA (D-001810-10) was used as a control, and a GAPDH-targeting siRNA (D-001830-10) was used to monitor transfection and knockdown efficiency. Stock solutions of 2 μM were prepared by dissolving siRNAs in 1× siRNA buffer (Dharmacon). For transfection of cells in 96-well clusters, 1 × 10⁴ 293/ACE2 cells per well were transfected with a 100-μl mixture containing 100 nM siRNA, 0.2 μg DharmaFECT1 transfection reagent (Dharmacon), Opti-MEM reduced-serum medium (Invitrogen), and antibiotic-free culture medium, according to the manufacturers' instructions. For cells in 12-well clusters, 600-μl transfection mixtures were used. Medium was replaced at 24 h posttransfection (p.t.) by antibiotic-free culture medium, and at 48 h p.t., cells were infected with EAV-GFP or wild-type (wt) EAV. Duplicate cultures were used either to prepare lysates to analyze protein expression levels or to monitor cell viability using a CellTiter 96 AQ_ueous nonradioactive cell proliferation assay (Promega), according to the manufacturer's instructions.

EAV-GFP and PRRSV-GFP replication is inhibited by CsA.

The effect of CsA on arterivirus replication was investigated in cell culture using two representatives of the arterivirus family, EAV and PRRSV (European genotype). For the initial experiments, we employed GFP-expressing recombinant viruses, since quantification of GFP expression provides a rapid and reliable method to detect inhibition of virus replication. BHK-21 cells were grown in 96-well plates and infected at an MOI of 5 with GFP-expressing recombinant EAV (33). Upon removal of the inoculum (at 1 h p.i.), medium containing 0.03 to 4 μM CsA was given, and at 18 h p.i., cells were fixed and GFP expression was quantified. We observed a strong dose-dependent inhibition of EAV-GFP replication (Fig. 1A) in the absence of significant cytotoxic effects at the CsA concentrations used (Fig. 1B). The IC₅₀ of CsA for EAV-GFP replication in BHK-21 cells was determined to be 0.95 μM.

A similar experiment was performed with PRRSV-GFP in MARC-145 cells (Fig. 1C). Although less sensitive to CsA treatment than EAV-GFP, the replication of PRRSV-GFP was completely blocked at 16 μM CsA, and an IC₅₀ of 5.22 μM was determined. Cell viability was only slightly affected by CsA concentrations above 4 μM (Fig. 1D).

The cyclophilin inhibitor Debio-064 blocks EAV-GFP and PRRSV-GFP replication.

Although CsA has been found to effectively block the replication of various RNA viruses in cell culture (14), its use in antiviral therapy would be complicated by the immune suppression (39) that is a major side effect. Therefore, several alternative Cyp inhibitors that lack the immune-suppressive properties of CsA, like SCY-635, NIM811, and Debio-025, which all block HCV replication, have been developed (17, 18, 40).

In this study, we tested whether the nonimmunosuppressive Cyp inhibitor Debio-064 is able to block EAV-GFP and PRRSV-GFP replication (Fig. 2A). Debio-064 is a structurally modified cyclosporine exhibiting an approximately 5-fold higher affinity for CypA than CsA. Debio-064 is 300-fold less active than CsA at inhibiting mouse T-cell proliferation induced by concanavalin A, suggesting that the compound does not inhibit calcineurin (41). EAV-GFP-infected BHK-21 cells (Fig. 2B) or PRRSV-GFP-infected MARC-145 cells (Fig. 2D) were treated with various noncytotoxic concentrations of Debio-064, and viral replication was quantified as described for CsA treatment. Compared to CsA, Debio-064 had a stronger inhibitory effect on EAV-GFP replication. At a concentration of 0.5 μM Debio-064, the EAV-GFP signal was hardly detectable (Fig. 2A), and an IC₅₀ of 0.29 μM, which is about 3-fold lower than that of CsA, was determined. For PRRSV-GFP, an almost complete block in GFP expression was observed at 8 μM, and an IC₅₀ of 5.14 μM (Fig. 2C), which is comparable to the inhibitory effect of CsA on PRRSV-GFP replication, was determined.

CsA and Debio-064 prevent arterivirus protein expression.

In our initial experiments, we tested the effect of CsA on the replication of a GFP-expressing recombinant EAV. To verify that wt EAV replication could also be inhibited by the drug, we analyzed wt EAV-infected BHK-21 cells that were treated with 0.25 to 8 μM CsA. At 6 h p.i., cells were lysed and lysates were subjected to Western blot analysis. The expression of viral nonstructural proteins (the nsp5-8 precursor and nsp9) and the structural M and N proteins was hardly detectable after 4 μM CsA treatment, while a clear reduction in protein expression could already be observed at 2 μM CsA (Fig. 3A). As observed for EAV-GFP, Debio-064 had a stronger inhibitory effect than CsA on the replication of wt EAV. Viral protein expression was clearly reduced in the presence of 0.5 μM the drug and became almost undetectable at 1 μM Debio-064 (Fig. 3B).

The effect of CsA and Debio-064 was further characterized by immunofluorescence microscopy of infected cells. For wt EAV, double-stranded RNA (dsRNA) (data not shown) and viral proteins were undetectable after a dose of 4 μM CsA (Fig. 3C) or 1 μM Debio-064 (Fig. 3D). In the case of PRRSV-GFP-infected MARC-145 cells (data not shown), maximal inhibition was observed at a 16 μM CsA dose. However, as previously observed for coronavirus-infected Vero E6, 17Cl1, or Huh7 cells (9), a small fraction of the MARC-145 cells remained capable of supporting PRRSV-GFP replication, even at high CsA doses of up to 64 μM (data not shown). We did not observe such a small, residual population of apparently drug-insensitive cells in EAV-infected BHK-21 cultures treated with either CsA or Debio-064 (Fig. 3C and D).

CsA and Debio-064 block the production of arterivirus infectious progeny.

To assess the extent to which CsA and Debio-064 treatment affected infectious progeny titers, we performed plaque assays to measure EAV titers at 12 h p.i. using supernatants from infected (MOI, 5) BHK-21 cells that had been treated with CsA or Debio-064 (Fig. 4A). CsA strongly reduced EAV progeny titers, with an almost 4-log-unit reduction at 4 μM CsA. Treatment probably completely abolished virus production, as the titers observed after treatment with 4 μM CsA were similar to those measured at 1 h p.i., which likely reflected the remainder of the high-MOI inoculum used (data not shown). These data correlated well with the barely detectable expression of the nsp5-8, nsp9, and the M and N proteins and the lack of dsRNA after treatment with 4 μM CsA (Fig. 3A and C). Treatment with Debio-064 also resulted in an ∼4-log-unit reduction of infectious progeny at 2 μM, while a 2- to 3-log-unit reduction was already achieved by treatment with 1 μM the compound.

Using a fluorescent focus assay, we also analyzed the production of PRRSV-GFP infectious progeny in culture supernatants of CsA-treated MARC-145 cells at 24 h p.i. As observed for EAV, the production of PRRSV-GFP infectious progeny was affected by CsA treatment, although significantly higher concentrations were required. At 16 μM CsA, an ∼1.5-log-unit reduction in the yield of infectious progeny was observed, while an apparently complete block (2.5-log-unit reduction) required a dose of 32 μM (Fig. 4B, gray bars). Treatment with Debio-064 resulted in an ∼1.5-log-unit reduction of infectious progeny at 16 μM and an ∼2.5-log-unit reduction of infectious progeny at 32 μM (Fig. 4B, white bars), which are comparable to the reduction in PRRSV progeny observed upon CsA treatment.

Cyclophilin inhibitors affect EAV RNA synthesis both in vivo and in vitro.

The above-described experiments showed that CsA can effectively block both EAV and PRRSV replication in cell culture. To establish that this lack of viral protein synthesis was due to a block of viral RNA synthesis, we measured the effect of CsA treatment on EAV RNA synthesis in infected Vero E6 cells in vivo, by metabolic labeling with [³H]uridine (in the presence of actinomycin D). When 4 μM CsA was given at 4.5 h p.i., ³H incorporation into viral RNA during a pulse-labeling from 5.5 to 6.5 h p.i. was reduced to 8% of the incorporation measured for nontreated control cells (Fig. 5A). To obtain more insight into the mechanism by which CsA inhibits arterivirus replication, we tested its effect in a previously developed in vitro assay to study the RNA-synthesizing activity of semipurified EAV RTCs (27). These assays were performed with PNS from EAV-infected BHK-21 cells, and the reactions, during which ³²P-labeled CTP is incorporated into viral RNA, were conducted in the presence of various concentrations of CsA. In the absence of the drug, in vitro synthesis of EAV genomic and sg RNAs was observed (Fig. 5B, lane 1), as documented previously (27). RNA-synthesizing activity was completely abolished when the reaction was performed in the presence of 12 μM CsA (lane 4), while a >50% reduction of viral RNA synthesis was observed in the presence of 8 μM CsA (lane 3). Comparable results were obtained with Debio-064, which also caused a >50% reduction of EAV RTC activity at about 8 μM and a complete inhibition at 16 μM (Fig. 5C, lanes 3 and 5). These data strongly suggest that Cyp inhibitors can directly affect the RNA-synthesizing capacity of the membrane-associated EAV RTCs in PNS samples. We recognize that the concentrations needed to fully block EAV RTC activity in vitro are ∼3-fold higher than those required to block virus replication in cell culture. This might be due to differences in the experimental setup, as the PNS used for the in vitro reaction constituted a concentrated preparation of RTCs (and host factors), and reaction conditions might influence the interaction between Cyps and their inhibitor.

EAV replication depends on cyclophilin A.

CsA is known to inhibit the PPIase activity of several members of the cyclophilin family. In particular, CypA and CypB have been implicated in the replication of several viruses (reviewed in reference 14). We therefore analyzed the effect of siRNA-mediated knockdown of CypA and CypB expression levels on the replication of EAV-GFP. We made use of the same human 293/ACE2 cells that we previously used to study the role of Cyps in SARS-CoV replication (9). This cell line was also susceptible to EAV infection, although only ∼40% of the cells became GFP positive after infection with EAV-GFP at a high MOI, as judged by immunofluorescence microscopy of infected cells fixed at 8 h p.i. (data not shown).

Knockdown of CypA and CypB expression was monitored by Western blotting, and an ∼80% reduction of expression compared to the level in control cells transfected with a nontargeting control siRNA was typically observed (Fig. 6A). Depletion of CypA or CypB did not have a significant effect on cell viability during the 48 h of the knockdown experiment (Fig. 6B). Compared to control cells, knockdown of CypB expression did not influence GFP reporter expression when these cells were infected with EAV-GFP, in which GFP fluorescence was measured at 24 h p.i. (Fig. 6C). In contrast, knockdown of CypA resulted in an ∼60% reduction of the GFP signal in EAV-GFP-infected cells compared to that in the control cells (Fig. 6C). Furthermore, at 32 h p.i., wt EAV titers in the culture medium of infected 293/ACE2 cells (MOI, 0.01) that had been depleted for CypA showed an ∼4-fold decrease in virus progeny compared to control cells (Fig. 6D). These data strongly suggest that EAV replication and the production of virus progeny depend on the availability of the host factor CypA.

Cyclophilin A cosediments with EAV RTCs.

Since our RNAi experiments suggested that EAV RNA synthesis depends on the availability of CypA, we investigated whether CypA cosediments with EAV RTC-containing membranes. We therefore fractionated PNSs from EAV-infected and mock-infected Vero E6 cells in a 0 to 30% OptiPrep density gradient. Gradient fractions were analyzed by Western blotting using antisera against CypA, CypB, several EAV nsps, and various organelle marker proteins (Fig. 7). Densities in the gradient ranged from 1.04 g/ml to 1.18 g/ml (Fig. 7A), and the low-density fractions of both EAV-infected and mock-infected PNSs contained the cytosolic marker GAPDH, while several organelle markers, like the ER protein calnexin and the mitochondrial marker CoxIV (data not shown), were found in higher-density fractions (Fig. 7B and C). This confirmed the separation of membrane-containing fractions from the cytosol. The membrane-associated EAV RTCs, detected with an antiserum directed against nsp9 (RNA-dependent RNA polymerase [RdRp]), sedimented at densities of about 1.15 g/ml (Fig. 7C). The nsp9-containing fraction also contained significant amounts of the normally cytosolic CypA (Fig. 7C). Upon density gradient fractionation of PNSs from uninfected cells, CypA was found only in the low-density cytosolic fractions (Fig. 7B), but in PNSs from EAV-infected cells, a fraction of CypA was found to cosediment with the RTC-containing membranes (compare fractions 3 in Fig. 7B and C). CypB, being an ER-associated protein, was observed in the high-density gradient fractions of both mock- and EAV-infected cell lysates (Fig. 7B). Therefore, the protein was present in the fractions containing the EAV RTCs, in particular, in the nsp9-containing fraction (Fig. 7C, fractions 3 and 4), but was clearly more dispersed in the gradient with mock-infected PNSs.

To analyze whether CsA can prevent the cosedimentation of CypA with EAV RTCs, we pretreated the PNSs from EAV-infected cells with CsA at 12 μM—the concentration that completely inhibited EAV RTC activity in vitro—for 30 min on ice before separating the material in an OptiPrep density gradient. Subsequently, the high-density nsp9-containing membrane fractions were analyzed for the presence of CypA by Western blotting (Fig. 7D). In the absence of CsA, a clear cosedimentation of CypA and nsp9 was observed, while CypA was no longer detectable in the high-density nsp9-containing fraction of CsA-treated lysates. This suggests that CsA can prevent the cosedimentation of CypA with the membrane-associated EAV RTCs.