Early Steps in Herpes Simplex Virus Infection Blocked by a Proteasome Inhibitor

Viruses usurp host cell functions to advance their replicative agenda. HSV relies on cellular proteasome activity for successful infection. Proteasome inhibitors, such as MG132, block HSV infection at multiple stages of the infectious cycle. Targeting host cell processes for antiviral intervention is an unconventional approach that might limit antiviral resistance. Here we demonstrated that the proteasome inhibitor bortezomib, which is a clinically effective cancer drug, has the in vitro features of a promising anti-HSV therapeutic. Bortezomib inhibited HSV infection during the first hours of infection at nanomolar concentrations that were minimally cytotoxic. The mechanism of bortezomib’s inhibition of early HSV infection was to halt nucleocapsid transport to the nucleus and to stabilize the ND10 cellular defense complex. Bortezomib and acyclovir acted synergistically to inhibit HSV infection. Overall, we present evidence for the repurposing of bortezomib as a novel antiherpesviral agent and describe specific mechanisms of action.

a guanosine analogue that targets HSV DNA replication for termination (11). Acyclovirresistant strains can lead to severe disease, including disseminated infection of immune-dysregulated individuals (12,13). New, effective therapeutics with different mechanisms of action are needed.
26S proteasomes are ϳ2.5-MDa, ATP-dependent multisubunit proteolytic machines present in both the nucleus and the cytoplasm of all eukaryotic cells. They are composed of a barrel-shaped, central 20S core that contains the proteolytic activities. The proteasome executes the controlled degradation of functional proteins as well as the hydrolysis of aberrantly folded polypeptides (14). Proteasome-dependent degradation plays a key role in many cellular processes, such as cell cycle control, proliferation, and apoptosis (15). Viruses can commandeer proteasome activity to promote a diversity of functions critical for their replicative cycles (16). MG132 is a peptide aldehyde that competitively inhibits the degradative activity of the proteasome. Functional proteasomes facilitate HSV entry at a postpenetration stage. MG132 impairs incoming HSV capsid transport to the nuclear periphery (17). Proteasome inhibitors also block HSV infection by preventing degradation of promyelocytic leukemia (PML) isoforms, stabilizing nuclear domain 10 (ND10), and, in turn, preventing lytic replication (18)(19)(20). Thus, the proteasome is important for early events in HSV infection. Many antiviral drugs are designed to target viral proteins to ensure specificity and avoid toxicity, but such antivirals select for drug-resistant virus mutants. In contrast, antivirals that target cellular proteins required for viral replication, such as proteasomal components, avert the development of resistance.
Here, we report anti-HSV activity of the proteasome inhibitor bortezomib. Overall, we provide evidence that bortezomib inhibits infection of multiple HSV strains early in the infectious cycle, exhibits minimal cytotoxicity, mechanistically halts viral capsid transport to the nucleus and stabilizes ND10 structure, and exhibits synergy with acyclovir.
contain mutations in the HSV thymidine kinase or DNA polymerase genes that render them resistant to acyclovir (26)(27)(28). Bortezomib was similarly effective at blocking infection by HSV-1 clinical isolate H129, derived from the brain of a herpes encephalitis patient (29). The concentrations of bortezomib that inhibited 50% of HSV infection (EC 50 ) are depicted in Fig. 2 The EC 50 values ranged from 3.7 to 50.6 nM, indicating that bortezomib, at low nanomolar concentrations, inhibits infection by wild-type, acyclovirresistant, and clinically isolated HSV. In this assay, the EC 50 value for acyclovir was 380 nM (see Fig. S1 in the supplemental material), which falls within the broad range of EC 50 s reported for acyclovir. Experiments with bortezomib and primary human foreskin fibroblasts (HFFs) yielded an EC 50 of 6.0 nM (Fig. 2B).
Bortezomib exhibits low cytotoxicity at concentrations effective against HSV infection. Proteasome inhibition ultimately leads to cell death (30). Thus, it was important to determine that bortezomib cytotoxicity did not explain the loss in HSV infectivity under our experimental conditions. We exposed Vero cells to doses of bortezomib similar to those utilized in the experiments whose results are presented Fig. 2. The cytotoxicity of bortezomib was quantitated via measurement of the levels of extracellular lactate dehydrogenase (LDH), an enzyme released upon cell death. Bortezomib-induced cytotoxicity peaked at ϳ30% at concentrations of Ͼ10 M but did not increase further, even at 1 mM (Fig. 3). Thus, the concentration of bortezomib that is cytotoxic to 50% of cells (CC 50 ) is Ͼ1 mM. At bortezomib concentrations near the EC 50 for HSV infection, cytotoxicity levels ranged from 0% to 8% (Fig. 3). Similar results were obtained with human foreskin fibroblasts (Fig. S2). Thus, bortezomib exhibits low cytotoxicity at concentrations that are effective against HSV infection.
Bortezomib is effective when added prior to 3 h postinfection (p.i.). To begin to determine the mechanism by which bortezomib blocks HSV infection, a time-ofaddition study was implemented. HSV-1 KOS (multiplicity of infection [MOI] of 0.001) was added to Vero cells. At time points from 0 to 6 h p.i., 200 nM bortezomib was added to the infected cells. HSV-1 titers were determined at 24 h p.i. The later bortezomib was added, the less of an effect it had on HSV infection (Fig. 4). Bortezomib added at time zero almost completely inhibited HSV infection but consistently lost effectiveness over time, inhibiting only minimally after 3 h p.i. At least two proteasome-dependent processes occur during the first hours of HSV-1 infection. These results suggest bortezomib acts at an early step of HSV infection that requires the proteasome.
Bortezomib is not virucidal against HSV. We determined whether the inhibitory effect of bortezomib on HSV infection was due to a direct, inactivating effect on viral particles. HSV-1 KOS was treated with 100 nM bortezomib for 1 h. The virus-drug mixture was diluted to noninhibitory concentrations of bortezomib, and titers were determined on Vero cells. The infectivity of the bortezomib-treated samples was similar to that seen with the vehicle control (Fig. 5). This result suggests that bortezomib does not have a direct, virucidal effect on the infectivity of HSV particles.
HSV attachment to cells is unaltered by bortezomib. To rule out the possibility that the bortezomib inhibition was due to an effect on viral attachment to the cell surface, HSV-1 KOS (40 genomes/cell) was added to Vero cells on ice for 1 h at 4°C in the presence of 200 nM bortezomib, a vehicle control, or a heparin treatment control. Cell-attached HSV-1 was quantitated by quantitative PCR (qPCR). Bortezomib-treated HSV-1 attached to cells in a manner similar to that seen with mock-treated HSV-1  ( Fig. 6A). Control soluble heparin inhibited HSV-1 attachment to cells by >84%. These results suggest that the inhibitory activity of bortezomib is not due to a defect in HSV-1 attachment to cells.
Transport of the HSV capsid to the nucleus is halted by bortezomib. Following fusion with a cell membrane, entering HSV nucleocapsids are transported in a proteasome-dependent manner to the nucleus, the site of herpesviral genome replication (17). To determine whether this step in the viral life cycle is affected by bortezomib, HSV-1 K26GFP was added to Vero cells in the presence of 100 or 500 nM bortezomib at 37°C. Herpesviruses utilize multiple entry pathways in a cell-specific manner (31). HSV-1 entry into Vero cells proceeds via direct penetration with the host cell plasma membrane (32,33). By 2.5 h p.i., in untreated cells, capsids were detected  at the nuclear periphery (Fig. 6B). In contrast, in cells treated with bortezomib, HSV-1 capsids were halted at the cell periphery ( Fig. 6C and D). Thus, bortezomib impacts HSV-1 infection at an early step, prior to capsid arrival at the nucleus.
HSV-induced ND10 disruption is prevented in the presence of bortezomib. Host cell ND10 nuclear bodies contain many proteins responsible for normal functions, including cell cycle regulation, apoptosis, gene transcription, and antiviral defense mechanisms (34)(35)(36). A hallmark of infection by viruses, including HSV, is disruption of ND10 (35). HSV-induced disruption of ND10 is proteasome mediated and dependent on viral immediate early protein infected cell protein 0 (ICP0) (20). MG132 halts virusmediated ND10 dispersal (20). To determine if bortezomib blocks the HSV-triggered dissolution of host ND10, we infected Vero cells with HSV-1 KOS and visualized PML, a major protein component of ND10, via immunofluorescence microscopy. In uninfected, untreated control samples, PML was present in punctate nuclear dots (Fig. S3). Upon infection, confirmed by the presence of HSV-1 ICP4, vehicle control-treated samples lacked punctate dots. Instead, there was diffuse PML throughout the nucleus ( Fig. 7A to C), indicating ND10 disruption triggered by HSV infection. The staining in Fig. 7 is consistent with antibody detection of ICP4 expressed by the infected cell (37). Tegument ICP4 is present at ϳ150 copies per virion (38) and has not been detected in infected cells. When 200 or 500 nM bortezomib was added prior to infection, the punctate PML staining was sustained during infection ( Fig. 7G and K, open triangles). Uninfected (ICP4-negative) cells were also detected when bortezomib was present ( Fig. 7G and K, closed triangles), which might have been due to effects of bortezomib on incoming capsid transport. Interestingly, ICP4 staining was still detected in a subset of nuclei, suggesting that viral entry and ICP4 gene expression had occurred while the ND10 structure remained intact (open triangles). Transport of incoming HSV-1 capsids to the nucleus is largely proteasome dependent. However, proteasome-independent transport also occurs and is particularly notable when tegument ICP0 is missing from incoming particles (17,39). The results represented in Fig. 6 and 7 are consistent with the notion that bortezomib shuts down HSV-1 infection at two steps. If the drug is ineffective at the earlier capsid transport step, it has a second opportunity to inhibit infection by preserving ND10. Overall, these results suggest that bortezomib prevents the disruption of the ND10 host cell defense complex, which is coincident with successful lytic replication of HSV.
Acyclovir and bortezomib work synergistically to inhibit HSV infection. Combination therapies with two or more drugs have the potential to successfully inhibit viral infection more effectively than either drug alone (40). To determine if this is the case for acyclovir and bortezomib, HSV-1 KOS (MOI of 0.1) was added to Vero cells in the presence of acyclovir and bortezomib simultaneously at various concentrations. At 24 h p.i., the cells and supernatant were collected, and the titers of each drug combination sample were determined. Increasing concentrations of either bortezomib alone (topmost bars) or acyclovir alone (blue bars) reduced HSV infectivity (Fig. 8A). Selected combinations of the two drugs also inhibited infectivity. These combinations resulted in minimal cytotoxicity (Fig. S4). Results were analyzed with CompuSyn software, which provided a combination index (CI) value that suggested whether the effect of two drugs was additive, antagonistic, or synergistic. At software-determined concentrations that resulted in 50%, 75%, or 90% inhibition of HSV infection, the CI values were all less than 1, indicating synergy ( Fig. 8B and C). CI values for most of the individual concentrations tested were also indicative of synergy, with a minority indicating additivity or antagonism (see Table S1 in the supplemental material). These results suggest that the combination of bortezomib and acyclovir inhibits HSV infection more effectively than treatment with either drug alone, indicative of a synergistic relationship.

DISCUSSION
The proteasome is a host cell component required for successful infection by HSV. It is needed at multiple steps in the viral life cycle and thus is an attractive target for therapeutic intervention. Targeting a host cell component is expected to decrease the development of viral resistance. Here, we demonstrate that proteasome inhibition by bortezomib effectively inhibits HSV infection in cell culture and identify a mechanism of action. Bortezomib inhibits wild-type, acyclovir-resistant, and clinically isolated HSV infection. It has low cytotoxicity at effective concentrations. Mechanistically, bort- Bortezomib Inhibition of HSV Infection ® ezomib acts during the capsid transport and ND10 disruption steps of the HSV life cycle. Bortezomib did not affect HSV attachment to cells and was not virucidal. Bortezomib exhibited synergy with acyclovir. Overall, we present evidence that proteasome inhibition represents an attractive intervention for HSV infection and that the anticancer drug bortezomib might be able to be repurposed as a novel anti-HSV therapeutic.
Incoming HSV-1 nucleocapsid transport to the nucleus is blocked by the proteasome inhibitors lactacystin, epoxomicin, MG132, and bortezomib (17) (Fig. 6C and D). Nuclear delivery of HSV capsids is facilitated by ICP0 present in the virion tegument but is independent of E1 activation of ubiquitin (17,39,41). The substrate targeted by the proteasome for HSV capsid transport is not currently known. Nuclear transport of capsids from incoming virions that lack tegument ICP0 is not blocked by proteasome inhibitors; these virions remain infectious in a proteasome-independent manner (39).
Our results indicate that while bortezomib blocks the majority of capsid transport to the nucleus, a subset of virions might still enter. HSV-1 ICP4 is detected in nuclei even after bortezomib treatment (Fig. 7). This is consistent with the notion that HSV capsids, while largely dependent on proteasomal activity, can also reach the nucleus via a proteasome-independent mechanism.
HSV infection is also facilitated by the proteasome's degradative activity at later steps in the viral life cycle. Following HSV genome entry into the nucleus, the viral DNA relies partly on nuclear factor kappa-␤ (NF-␤) for transcriptional activation (59). Induction of NF-␤ depends on proteasomal degradation of the inhibitor of kappa-␤ kinase (IK). Proteasome inhibitors downregulate NF-␤ and, consequently, the HSV transcripts that depend on NF-␤ induction (60). Future efforts will test directly the effect of bortezomib on the expression of HSV-1 immediate early, early, and late genes and the localization of their gene products. HSV ICP0-mediated proteasomal degradation of DNA-dependent protein kinase (61) and centromeric proteins CENP-A (62) and CENP-C (63) has also been described. Proteasome activity is also required during HSV reactivation from latency (64). Here we demonstrate that bortezomib blocks infection by inhibiting two key, proteasome-dependent processes that occur during the first hours of infection: HSV capsid transport to the nucleus (Fig. 6B to D) and ND10 disruption (Fig. 7).
The primary treatment for HSV infection uses the acyclovir family of drugs. Acyclovir is a guanosine analogue that interferes with viral DNA replication. Direct pressure on the viral kinase results in the development of HSV strains resistant to acyclovir (26).
Immunocompromised patients are particularly susceptible to infection by acyclovirresistant strains, which can result in disseminated disease and death (12,13). Secondline treatments for these severe infections include the use of foscarnet, a pyrophosphate analogue, and of cidofovir, a cytosine analogue, both of which target viral DNA replication (65). Thus, treating acyclovir-resistant strains with these second-line agents might result in the development of multiresistant strains, as has been reported previously for foscarnet (66).
The CC 50 values for foscarnet and cidofovir on Vero cells are 50 mM and 560 M, respectively (67,68). We were unable to reach 50% Vero cell cytotoxicity with concentrations as high as 1 mM. The foscarnet and cidofovir EC 50 values for HSV-1 on Vero cells are 32.6 M and 6.4 M, respectively (69,70). Bortezomib yields ϳ1,000-times-lower EC 50 values. Moreover, foscarnet and cidofovir both exhibit toxicity in vivo, particularly in the form of nephrotoxicity (71,72). Other side effects of cidofovir include neutropenia and metabolic acidosis (73). The therapeutic dose of foscarnet is 40 to 90 mg/kg of body weight, and its level of toxicity in mice is 500 mg/kg (74). The therapeutic dose of cidofovir is 5 mg/kg, but its level of toxicity in animals is 0.25 to 1 mg/kg (75). The therapeutic dose of bortezomib in cancer treatment is 1.3 mg/m 2 . The level of bortezomib toxicity in animals is 0.6 to 0.9 g/m 2 (76). Of course, toxicity in animal models is not always predictive of toxicity in humans. Bortezomib, unlike acyclovir, foscarnet, and cidofovir, targets a host process; thus, antiviral resistance might be less likely.
The EC 50 values for bortezomib against HSV infection ranged from 3.7 to 50.6 nM (Fig. 2). Similar concentrations inhibited proteasome activity and kill cancer cells in culture models of B cell lymphoma and mantle cell lymphoma (85)(86)(87). Clinical side effects of bortezomib treatment include peripheral neuropathy (PN) and thrombocytopenia (88,89). PN might be due to serine protease inhibition by bortezomib in mitochondria, which has a role in the survivability of neurons (90). At a dose of 1.45 to 2.0 mg/m 2 , bortezomib has a half-life of 23 days or 8.7 to 14.8 h in whole blood or plasma, respectively (91,92). Bortezomib is given twice weekly, and the mean maximum plasma concentration is 200 to 300 nM (92,93). Thus, the EC 50 s that we obtained were well below the clinically achievable plasma concentration. Therefore, it is possible that plasma concentrations of bortezomib lower than those used for multiple myeloma might be effective against HSV, potentially ameliorating side effects. Bortezomib is clinically effective in patients with immune deficiency resulting from plasma cancers. Thus, bortezomib might be appropriate for administration to the immunocompromised, an important target population for new anti-HSV therapeutics. Bortezomib delivery by subcutaneous administration might be safer than delivery by intravenous administration as suggested by a previous report from a clinical trial in multiple myeloma patients (94). Herpes zoster (shingles), caused by varicella zoster virus, is an adverse event associated with bortezomib use (95). For this reason, acyclovir prophylaxis is recommended for multiple myeloma patients under treatment (96). For the HSV-infected immunocompromised population, acyclovir and low-concentration bortezomib therapy could be imagined and would also reduce the risk of shingles.
Repurposing of existing drugs has been successful in treating other medical conditions. For example, eflornithine is an ornithine decarboxylase inhibitor used to treat African sleeping sickness that was repurposed to treat female facial hirsutism (97). Thalidomide, an anti-inflammatory medication for treatment of leprosy, has recently been approved for treatment of multiple myeloma (98). Bortezomib, like these examples, has already been FDA approved for treatment, which would aid in streamlining the redevelopment process. Altogether, we provide evidence that bortezomib is a novel potential therapeutic for HSV with a defined mechanism of action. The results warrant preclinical testing of bortezomib efficacy in an animal model of HSV infection.
Chemicals. Stocks of 50 mM bortezomib (Selleckchem, Houston, TX, or Sigma, St. Louis, MO) and 20 mM acyclovir (Sigma) were prepared in dimethyl sulfoxide (DMSO; Fisher Scientific, Fair Lawn, NJ) and stored at Ϫ80°C and Ϫ20°C, respectively. Stocks of 0.5 mg/ml heparin (Sigma) were prepared in water and stored at Ϫ20°C.
Cytotoxicity of bortezomib. Bortezomib or a DMSO vehicle control was added to confluent cell monolayers in 96-well plates. At 24 h, a sample of supernatant was assayed for lactate dehydrogenase (LDH) activity using a Pierce LDH cytotoxicity assay kit (Thermo Scientific, Rockford, IL) according to the manufacturer's instructions. All concentrations were tested in triplicate. As a positive-control sample, cells were lysed with 1% SDS for 30 min. Cytotoxicity is reported as LDH activity as a percentage of the detergent-lysed sample.
Time of addition of bortezomib. HSV-1 strain KOS (MOI of 0.001) was added to Vero cells. Bortezomib or vehicle control was added at a concentration of 200 nM at various times postinfection (p.i.) from 0 to 6 h. Cells were incubated at 37°C and 5% CO 2 for 18 to 24 h in total and were subjected to HSV plaque assay. Plaques were quantitated, and data are presented as percent inhibition of vehicle control infectivity.
HSV attachment to cells. HSV-1 KOS was treated with 2 g/ml DNase (Turbo DNAFree; Thermo Fisher Scientific) according to the manufacturer's instructions. This treatment removes any free HSV-1 DNA that is not protected inside viral capsids. Virus was diluted in ice-cold binding medium (carbonatefree, serum-free DMEM supplemented with 20 mM HEPES and 0.2% bovine serum albumin [BSA]). Prechilled Vero cells were simultaneously exposed to 200 or 500 nM bortezomib or 2 g/ml heparin control and virus (40 genome copies/cell). Virus was subjected to spinoculation onto the cells at 200 ϫ g for 1 h at 4°C. Cells were washed twice with ice-cold phosphate-buffered saline (PBS; Thermo Fisher Scientific), and cell-associated HSV-1 DNA was isolated with a QIAamp DNA blood minikit (Qiagen, Germany). The HSV-1 ICP22 copy number was determined via qPCR. HSV-1 was quantitated using a CFX96 real-time PCR detection system (Bio-Rad). Primers (Integrated DNA Technologies [IDT], Coralville, IA) were based on KOS ICP22 sequences, both forward (5= GAG TTT GGG GAG TTT G 3=) and reverse (5= GGC AGG CGG TGG AGA A 3=) (101,102). A standard curve was generated for the assay using known quantities of a plasmid containing the HSV-1 ICP22 coding region diluted in glycogen.
Direct effect of bortezomib on viral particles. HSV-1 KOS virions (ϳ3 ϫ 10 7 PFU) were directly treated with 100 nM bortezomib-culture medium for 1 h at 37°C. Control samples were treated with DMSO vehicle. Samples were diluted 10-fold in culture medium, and HSV-1 titers were determined on Vero cells. The concentration of residual bortezomib in diluted virus preparations (0.01 nM) does not inhibit HSV-1 infection.

Synergy of bortezomib and acyclovir.
Vero cells were infected with HSV-1 KOS (MOI of 0.1) and exposed simultaneously to various combinations of acyclovir and bortezomib. Cells were incubated at 37°C and 5% CO 2 for 24 h. A sample of supernatant was taken for LDH cytotoxicity analysis as described above. Cells were collected and the titers determined via HSV plaque assay as described above. Data are presented as percent inhibition compared to an untreated control. The three-dimensional (3D) graph was constructed using Microsoft Excel. Combination indices and isobolograms were generated with Com-puSyn software.
Ethics statement. Low-passage-number HSV-1 isolate H129 was part of an already-existing, publicly available collection (26). Institutional review board (IRB) approval was not sought. Samples were anonymized in the previous study.