The Proteasome Inhibitor Velcade Enhances rather than Reduces Disease in Mouse Hepatitis Coronavirus-Infected Mice

The cellular proteasome is a central actor in protein degradation in eukaryotes (48). This barrel-shaped cytosolic complex has at least three peptidase activities (i.e., caspaselike, trypsinlike, and chymotrypsinlike activities), which are mediated by different subunits (11, 14, 15). Proteasomal degradation is an important mechanism to control protein homeostasis, thereby regulating various basic cellular processes, including cell cycle regulation, cell adhesion, gene transcription, and apoptosis (12, 19, 28, 29, 37).

Because proteasome inhibitors induce apoptosis preferentially in tumor cells, they represent a novel class of drugs in anticancer therapy (10, 16, 30, 42). One of these drugs, Velcade (Bortezomib; formerly known as PS-341, LDP-341, and MLM341), selectively blocks the chymotrypsinlike activity of the catalytic core of the proteasome 20S subunit (1, 2). Velcade has been clinically approved for the treatment of multiple myelomas and mantle cell lymphoma (17, 18), while it also displays anticancer activity against a range of different human malignancies in cell culture and in animal models, including myeloma, pancreatic cancer, lung cancer, prostate cancer, chronic lymphocytic leukemia, and colon cancer (7, 31-33, 40, 41).

Viruses are obligatory intracellular parasites that exploit the host cell for generating their progeny virions. A functional cellular proteasomal system has been shown to be critical for many viruses, including coronaviruses (CoVs), at different stages of their life cycle (8, 20, 23, 24, 39, 43, 50). Thus, in addition to their potential as anticancer drugs, proteasome inhibitors also appear to be promising antiviral agents (39, 45). Supporting this idea, proteasome inhibitors have been shown to protect against coxsackievirus-induced myocarditis and to prolong the survival of mice inoculated with Epstein-Barr virus-transformed B cells (13, 51). However, proteasome inhibitors have also been shown to exhibit immunosuppressive properties. They can, for example, alter TLR4-induced dendritic cell activation and interfere with the immunological functions of T cells (4, 27). Furthermore, cancer patients treated with Velcade demonstrate increased reactivation of varicella herpes zoster virus infections (6, 21, 46). Thus, while playing essential roles in the replication of several viruses, the proteasome is also critically involved in generating an effective antiviral immune response.

In the present study we evaluated the ability of Velcade (obtained from Millennium Pharmaceuticals, Inc.) to inhibit mouse hepatitis virus (MHV) infection in living animals. Previously, we showed that different proteasome inhibitors can dramatically affect the replication of different coronaviruses (CoVs; i.e., MHV, FIPV, and SARS-CoV), as determined by measuring reporter gene expression at different time points postinfection (33a). We confirmed and extended these observations here by evaluating the antiviral effect of Velcade in cell culture by performing a one-step growth curve with MHV-A59 in LR7 mouse fibroblast cells (22). To this end, cells were infected in the absence or presence of 10 μM Velcade (Fig. 1). At the indicated time points, the virus titers in the supernatant were determined by using a quantal assay. In the presence of Velcade, virus production was dramatically affected (Fig. 1), a result that is in agreement with the decrease in virus replication observed previously (33a), whereas cell viability was not affected, as determined by a Wst-1 assay (47).

To investigate whether the proteasome inhibitor Velcade could also inhibit infection in vivo, we next used our recently developed bioluminescence imaging (BLI) model to monitor the spatial and temporal progression of MHV infection in living mice (35). To this end, C57BL/6 mice were inoculated intraperitoneally with 10⁶ 50% tissue culture infective doses (TCID₅₀) of MHV-EFLM, which is a recombinant MHV expressing the firefly luciferase (FL) reporter gene. The mice were either pretreated with Velcade in phosphate-buffered saline (PBS) at 1 mg/kg or with an equal volume of PBS. The drug was applied intraperitoneally on days −1 and 2 relative to the time of inoculation with MHV-EFLM. Replication, as measured by determining the amount of in vivo FL activity, was assessed by BLI as described previously (35). All mice were imaged for exactly 10 min on their ventral sides. As shown in Fig. 2, treatment of MHV-infected mice with Velcade resulted in a dramatic increase in FL expression. This increase was already significant at day 2 postinfection but developed rapidly further during the next 2 days. In sharp contrast, the mock-treated, MHV-infected animals showed a reduction in FL levels from days 2 to 4. Furthermore, the Velcade-treated mice, but not the mock-treated animals, showed severe clinical signs (i.e., weight loss and rough fur), with two mice already succumbing to the infection before day 4 (data not shown). These data strongly suggest that treatment of MHV-infected mice with the proteasome inhibitor Velcade does not inhibit infection but rather enhances it.

To corroborate these results, we subsequently performed an additional experiment in which we used more conventional methods to measure MHV replication in mice. Therefore, C57BL/6 mice were inoculated intraperitoneally with 10⁶ TCID₅₀ of wild-type MHV-A59. Treatment with Velcade was performed as described above. As an additional control, a group of mice (n = 4) was treated with Velcade only. Body weight measurements were performed daily; at day 4 postinfection all of the mice were sacrificed, and the livers were collected to determine the viral loads. The mice infected with MHV but not treated with Velcade and the uninfected Velcade-treated control animals did not show any clinical symptoms. In contrast, the animals infected with MHV and treated with Velcade showed severe clinical signs with a body weight drop at day 4 of ca. 15% relative to their weight at the start of the experiment (Fig. 3 A). One mouse in this group already succumbed to the infection at day 3. The viral loads, as measured by determining the PFU and the amounts of viral RNA in the liver homogenates (Fig. 3B and C), were also higher in the animals treated with Velcade. However, the viral loads were not elevated as high as might have been expected on the basis of the BLI data.

Next, we analyzed the expression of viral proteins in liver sections of the mice by staining with a polyclonal antiserum (k135) raised against the structural proteins of MHV (38). Although we could detect only moderate staining of viral antigen in the animals infected with MHV, the lesions in MHV-infected mice treated with Velcade appeared to be filled with viral protein (Fig. 4 A). In retrospect, these data fit with the much higher expression of the FL reporter gene observed in the treated mice in the BLI experiment (Fig. 2). Histopathologic evaluation of hematoxylin-and-eosin-stained liver sections of the MHV-infected mice showed randomly distributed coagulation necrosis of hepatocytes in mice treated with Velcade and in the control group. However, light microscopic morphometric analysis (using software from Olympus Soft Imaging Solutions Gmbh) of the liver showed a significant (P < 0.0001) 3-fold increase in the surface area with necrosis in the MHV-infected mice treated with Velcade compared to the control group (1.87 × 10⁵ μm² ± 4.1 × 10⁴ μm² versus 5.89 × 10⁴ μm² ± 8.3 × 10³ μm², respectively). Moreover, semiquantitative analysis of the inflammatory response showed a markedly reduced presence of leukocytes in the necrotic areas of the liver in the MHV-infected mice treated with Velcade compared to the control animals (Fig. 4B). Immunohistochemical evaluation of the affected areas indicated a reduction in T lymphocytes (Fig. 4C). In agreement with previous results (34), we observed an upregulation of cytokine gene expression (i.e., tumor necrosis factor alpha [TNF-α], CXCL-1, CXCL-2, and beta interferon [IFN-β]) in the livers of MHV-infected mice as determined by quantitative reverse transcription-PCR (RT-PCR). Interestingly, a similar induction of cytokine gene expression was also observed in the MHV-infected animals treated with Velcade, with the exception of the IFN-γ induction (Fig. 4D). This latter observation probably reflects the absence of activated immune cells in the virus-induced lesions. Altogether, we observed that proteasome inhibition during MHV infection results in an accumulation of viral antigen accompanied by the reduced presence of inflammatory cells at the sites of infection, which results in a fatal coagulation necrosis of hepatocytes.

The reduced inflammatory response (most notably, the reduced recruitment of T-lymphocytes) at the sites of MHV infection in the liver indicates that proteasome inhibition abrogates the induction of a protective immune response against MHV. Likewise, it has been shown recently that mice treated with Velcade were more susceptible to lymphocytic choriomeningitis virus infection (3). These observations correspond with the findings of recent studies, which demonstrate that proteasome inhibition promotes apoptosis in primary natural killer cells and dendritic cells (DCs) and inhibits DC maturation (26, 44, 49). In addition, exposure of mice to bortezomib radically impaired murine T-lymphocyte development in the thymus (25) and induced selective apoptosis and decreased Th1 responses among alloreactive T lymphocytes, while leaving unstimulated T cells unaffected (5). The present study shows for the first time that the immunosuppressive effects of the proteasome inhibitor Velcade may have a fatal outcome in mice infected with MHV, indicating that proteasome inhibitors are probably not useful as anti(corona)viral agents. Our results warrant a more in-depth investigation of the use of proteasome inhibitors in a clinical setting since they may result in adverse side effects in virus-infected individuals.