Dear Editor,

Human enteroviruses (HEVs) comprise polioviruses, coxsackieviruses, echoviruses, rhinoviruses, and the enterovirus subgroups and belong to the genus Enterovirus in the family Picornaviridae (Baggen et al. 2018). The HEV genome was thought to contain a single open reading frame (ORF) encoding one polyprotein, which was post-translationally processed into structural capsid proteins (VP1-4) and non-structural accessory proteins (2A, 2B, 2C, 3A, 3B, 3C, 3D) (Baggen et al. 2018). Recently, we and others, have identified an additional upstream ORF (Guo et al. 2019a; Lulla et al. 2019) (ORF2/uORF). The encoded protein, ORF2p/UP, facilitates virion release from gut epithelial cells. Given HEVs are globally ubiquitous and cause a diverse array of clinical diseases (Wang et al. 2003; Verboon-Maciolek et al. 2008; Muehlenbachs et al. 2015), there is an urgent need to develop broad-spectrum vaccines and therapeutic treatments for enterovirus infections.

Ubiquitin and ubiquitin-like post-translational modifications (PTMs) are essential for determining cell fate and proliferation, embryonic development, and carcinoma progression (Nandi et al. 2006). Moreover, accumulating evidence suggests that different viruses frequently target these PTM signaling pathways (Hofmann et al. 2012; Dawson and Mehle 2018; Kumar et al. 2020) and modulate the intracellular environment of the host to enhance virus replication. Recently, we found that the PTM neddylation pathway is critical for activating the binding of lentiviral accessory protein Vpx to CRL4 E3 ligase, which induces the degradation of antiviral protein SAMHD1 (Wei et al. 2014; Guo et al. 2019b). The ubiquitin-like protein NEDD8 has a dedicated E1-activating enzyme (NAE1) and two E2- conjugating enzymes (UBE2M and UBE2F) and is essential for the enzymatic activity of the Cullin-RING E3 ligase family (Rabut and Peter 2008; Enchev et al. 2015). The neddylation pathway is considered an optimal target for cancer therapeutics, and the NAE1 inhibitor, MLN4924, has demonstrated remarkable therapeutic benefit in Phase-Ⅰ clinical trials for cancer therapy (Sarantopoulos et al. 2016; Shah et al. 2016). Besides, MLN4924 has been found to be an optimal antiviral drug candidate against diverse human viruses including HBV (Xie et al. 2021) and influenza A virus (Sun et al. 2018).

Herein, we set out to investigate the effects of the neddylation pathway on enterovirus replication and explore the potential use of MLN4924 as an antiviral drug candidate against enterovirus infection. Human rhabdomyosarcoma RD cells were treated with MLN4924 at various concentrations (0.02, 0.05, 0.1, 0.5, 1, and 2 μmol/L) and challenged with EV-A71 (multiplicity of infection, MOI of 0.01). Viral titers in the supernatant were determined at 48 h post-infection (hpi), and the cytopathic effect (CPE) was noted. MLN4924 treatment markedly inhibited EVA71 CPE (Fig. 1A) and significantly reduced the titers of progeny virus in the supernatant (Fig. 1B). In addition, EVA71 VP1 protein expression decreased in a dose-dependent manner in MLN4924-treated cells (Fig. 1C). MLN4924 inhibited EV-A71 RNA replication in RD cells at an EC50 value of approximately 0.011 μmol/L (Fig. 1D). No significant cytotoxicity or decrease in cell proliferation was observed in MLN4924-treated cells, even at concentrations as high as 20 μmol/L, as determined using the MTS assay (Fig. 1E). MLN4924 treatment conferred resistance to EVA71 infection in human colon carcinoma HT-29 cells (Supplementary Fig. S1).

Figure 1.  Neddylation blockade inhibits viral replication of enteroviruses. A–D MLN4924- or DMSO-treated RD cells were infected with EV-A71. A Cytopathic effect (CPE) was observed at 48 h postinfection (hpi). Scale bar = 200 μm. B Quantification of infectious virions. Supernatants were collected 48 hpi, and viral titers were calculated as TCID₅₀. The concentration of MLN4924 used ranged from 0.02–0.1 μmol/L. C The protein expression levels of enterovirus VP1 protein were detected using immunoblotting. A range of MLN4924 concentrations were used (0.02, 0.05, 0.1, 0.5, 1 and 2 μmol/L). D Effect of MLN4924 on EV-A71 RNA replication (EC₅₀ = 0.011 μmol/L). The concentration of MLN4924 used ranged from 0.02–2 μmol/L. E Cell viability assay. The cytotoxicity of MLN4924 was determined in mock-infected RD cells. Cells were cultured in media containing a range of MLN4924 concentrations (0–20 μmol/L), and cell proliferation and cytotoxicity were evaluated using the MTS assay. F Virus attachment and entry assays. RD cells were incubated at 4 ℃ for 30 min before EV-A71 infection and then incubated at 4 ℃ or 37 ℃ for 2 h to facilitate virus attachment or entry, respectively. qRT-PCR was performed to quantify viral RNA. The concentration of MLN4924 was 0.5 μmol/L. (G and H) the concentrations of intracellular and supernatant viral RNA at specific time points post-infection were determined using qRT-PCR. The concentration of MLN4924 was 0.5 μmol/L. (I) The effect of MLN4924 treatment on EV-A71 VP1 protein expression at specific time points post-infection. The concentration of MLN4924 was 0.5 μmol/L. J–L Stable RD-shNAE, RD-shUBE2M, and RD-shUBE2F cells were challenged with EV-A71 (MOI of 0.01). J Quantification of progeny virions. K CPE observed in RD cells at 48 hpi. Scale bar = 200 μm. L The survival rates of infected cells were determined using trypan blue staining. M–O MLN4924- or DMSO-treated HEK293T cells were infected with EV-D68. M CPE observed in HEK293T cells 48 hpi. The concentration of MLN4924 was 0.5 μmol/L. Scale bar = 200 μm. N Protein expression levels of EV-D68 VP1 protein were detected using immunoblotting. The concentration of MLN4924 used ranged from 0.02–0.5 μmol/L. O Quantification of progeny virions. Supernatants were collected 48 hpi and viral titers calculated using TCID50 assay. The concentration of MLN4924 used ranged from 0.02–0.5 μmol/L. ^*P < 0.05, ^**P < 0.01, ^***P < 0.001.

Furthermore, we determined the effects of MLN4924 on EV-A71 entry into host cells. RD cells were treated with either MLN4924 or DMSO and then incubated with EVA71 for 2 h at 4 ℃ or 37 ℃ to facilitate virus attachment or entry, respectively. qRT-PCR was used to detect viral RNA (vRNA). We found that MLN4924 treatment did not significantly affect EV-A71 virus entry (Fig. 1F). Timedose analysis of EV-A71 RNA replication indicated a strong decrease in the accumulation of viral RNA in the cell lysate (Fig. 1G) and in the supernatant (Fig. 1H) of MLN4924-treated RD cells, when compared to that in DMSO-treated cells. We next analyzed VP1 protein expression in RD cells infected with EV-A71 either in the presence of 0.5 μmol/L MLN4924 or DMSO over a time course. VP1 protein expression was detectable at 8 hpi in both MLN4924- and DMSO-treated cells. At later time points, a dramatic reduction in VP1 protein expression was observed in both the cell-associated and supernatantassociated fractions of MLN4924-treated cells as compared to the DMSO-treated controls (Fig. 1I). Hence, MLN4924 suppresses EV-A71 replication at a post-entry step.

Then, we evaluated the role of neddylation in EV-A71 replication. There are two well-known pathways involved in neddylation: Rbx1-UBE2M and Rbx2-UBE2F (Huang et al. 2009). We generated stable knockdown RD cell lines using shRNA targeting endogenous E1 (NAE1) and E2 (UBE2M or UBE2F) enzymes (Supplementary Fig. S2). The cells were then challenged with EV-A71. These results showed that the downregulation of NAE1, UBE2M, and UBE2F dramatically reduced the production of infectious virions (Fig. 1J) and enterovirus CPE (Fig. 1K and 1L), as compared with those in control cells. Thus, the PTM neddylation pathway is a determinant factor for supporting EV-A71 replication.

These results led us to examine the antiviral activity of MLN4924 against other enteroviruses. Unlike EV-A71, which is transmitted via the fecal-oral route, EV-D68 is transmitted through coughing and sneezing. We treated HEK293T cells with MLN4924 concentrations ranging from 0.02 to 0.5 μmol/L, and infected the cells with EVD68 (MOI 0.1) for 48 h. MLN4924 treatment significantly blocked EV-D68 CPE (Fig. 1M), VP1 protein expression (Fig. 1N), and infectious virion production (Fig. 1O). MLN4924 treatment potently inhibited EV-D68 replication in human lung carcinoma A549 cells (Supplementary Fig. S3).

Protein neddylation is the process of transferring NEDD8 to a lysine residue on the substrate; it modulates the activity of its substrates. UBE2M and UBE2F are the two NEDD8-conjugating E2 enzymes, and they exhibit distinct biological functions. UBE2M is responsible for the neddylation of Cullin1, Cullin2, Cullin3, Cullin4A, and Cullin4B, whereas UBE2F facilitates the neddylation of Cullin5. Moreover, UBE2M and UBE2F have different non-Cullin substrates. Both UBE2M and UBE2F are required for enterovirus replication, implying that enteroviruses utilized a complicated neddylation-ubiquitination network for modulating the intracellular microenvironment conducive for viral infection. Identification of NEDD8-conjugated substrates, which are critical for viral replication, is essential to extend our understanding of the life cycle of enteroviruses.

Considering the remarkable therapeutic benefit of MLN4924 treatment in patients with diverse cancers, we explored its potential as an inhibitor of enterovirus replication. Here we provide evidence that MLN4924 is a potent inhibitor of enterovirus replication and suggest that neddylation is crucial for the life cycle of enteroviruses. Furthermore, we show that the components of the neddylation pathway can serve as potential targets for therapeutic intervention.