Enterovirus D68 Infection Induces TDP-43 Cleavage, Aggregation, and Neurotoxicity

ABSTRACT

Enterovirus D68 (EV-D68), which causes severe respiratory diseases and irreversible central nervous system damage, has become a serious public health problem worldwide. However, the mechanisms by which EV-D68 exerts neurotoxicity remain unclear. Thus, we aimed to analyze the effects of EV-D68 infection on the cleavage, subcellular translocation, and pathogenic aggregation of TAR DNA-binding protein 43 kDa (TDP-43) in respiratory or neural cells. The results showed that EV-D68-encoded proteases 2A and 3C induced TDP-43 translocation and cleavage, respectively. Specifically, 3C cleaved residue 327Q of TDP-43. The 3C-mediated cleaved TDP-43 fragments had substantially decreased protein solubility compared with the wild-type TDP-43. Hence, 3C activity promoted TDP-43 aggregation, which exerted cytotoxicity to diverse human cells, including glioblastoma T98G cells. The effects of commercially available antiviral drugs on 3C-mediated TDP-43 cleavage were screened, and the results revealed lopinavir as a potent inhibitor of EV-D68 3C protease. Overall, these results suggested TDP-43 as a conserved host target of EV-D68 3C. This study is the first to provide evidence on the involvement of TDP-43 dysregulation in EV-D68 pathogenesis.

IMPORTANCE Over the past decade, the incidence of enterovirus D68 (EV-D68) infection has increased worldwide. EV-D68 infection can cause different respiratory symptoms and severe neurological complications, including acute flaccid myelitis. Thus, elucidating the mechanisms underlying EV-D68 toxicity is important to develop novel methods to prevent EV-D68 infection-associated diseases. This study shows that EV-D68 infection triggers the translocalization, cleavage, and aggregation of TDP-43, an intracellular protein closely related to degenerative neurological disorders. The viral protease 3C decreased TDP-43 solubility, thereby exerting cytotoxicity to host cells, including human glioblastoma cells. Thus, counteracting 3C activity is an effective strategy to relieve EV-D68-triggered cell death. Cytoplasmic aggregation of TDP-43 is a hallmark of degenerative diseases, contributing to neural cell damage and central nervous system (CNS) disorders. The findings of this study on EV-D68-induced TDP-43 formation extend our understanding of virus-mediated cytotoxicity and the potential risks of TDP-43 dysfunction-related cognitive impairment and neurological symptoms in infected patients.

INTRODUCTION

Enterovirus D68 (EV-D68), an emerging pathogen belonging to the Enterovirus genus of the Picornaviridae family, is a small, nonenveloped, icosahedral virus (1, 2). EV-D68 has a positive-sense single-stranded RNA genome of ~7.4 kb and consists of a 5′-untranslated region (UTR) with an internal ribosome entry site and a spacer region, an open reading frame (ORF) encoding a single precursor polypeptide, and a short 3′-UTR with a poly(A) tail. The ORF-encoded polypeptide undergoes post-translational proteolytic processing to yield four structural proteins (VP1, VP2, VP3, and VP4) and seven nonstructural proteins (2A, 2B, 2C, 3A, 3B, 3C, and 3D) (3). EV-D68 is biologically similar to rhinoviruses but distinct from traditional enteroviruses in that it exhibits acid sensitivity, prefers low growth temperatures, and proliferates in the respiratory tract rather than in the intestine (4). EV-D68 is mainly transmitted through the respiratory tract rather than the fecal-oral route (4). Host membrane proteins serving as EV-D68 receptors include sialic acids (e.g., α2,6-linked and α2,3-linked SAs) (5, 6), heparan sulfate proteoglycans (7), and intercellular adhesion molecule 5 (8), which facilitate virus entry into host cells.

The incidence of EV-D68 infection has recently increased worldwide, making it a serious public health threat. Although only 26 sporadic cases of EV-D68 infection have been reported since its first identification in 1962 (9), this number has been increasing in the United States, Asia, and Europe since 2005 (10), with an outbreak occurring in the United States in 2014 (9). EV-D68 infection causes mild-to-severe respiratory diseases, with lower respiratory tract infections leading to dyspnea, wheezing pneumonia, and hypoxia. Some patients even require ventilatory support and intensive care (11). In addition, EV-D68 infection has attracted global attention owing to its association with various central nervous system (CNS) complications, of which acute flaccid myelitis (AFM) is the most common, causing pain in the affected limb and diminished or absent reflexes (12, 13). However, effective antiviral drugs or vaccines for controlling EV-D68 transmission remain unavailable to date. Thus, understanding the pathogenesis of EV-D68 is urgently needed to design control strategies against the spread of this virus.

Host TAR DNA-binding protein 43 kDa (TDP-43) is a heterogeneous nuclear ribonucleoprotein composed of 414 amino acids and is encoded by the TARDBP gene on chromosome 1. TDP-43 is involved in RNA splicing, transcription, mRNA transport, and stabilization (14, 15). It also regulates microRNA biogenesis (16). TDP-43 is synthesized in the cytoplasm and shuttled between the nucleus and cytoplasm in a transcription-dependent manner (17). It is predominantly localized in the nucleus under normal physiological conditions (17) but under pathological conditions is translocated to the cytoplasm, where it undergoes multiple post-translational modifications, including hyperphosphorylation, ubiquitination, and truncation, leading to its accumulation (18). Abnormal TDP-43 expression has been frequently detected in patients with degenerative diseases. TDP-43 is the main protein component of intracellular insoluble aggregates in nerve samples from patients with frontotemporal lobar degeneration (FTLD) and amyotrophic lateral sclerosis (ALS) (19). Cytoplasmic mislocalization and aggregation of TDP-43 proteins are common features in ALS, FTLD, and other neurodegenerative diseases, including Alzheimer’s disease (20) and limb-dominant age-related TDP-43 encephalopathy (21). However, the effects of EV-D68 on TDP-43 expression in respiratory and neural cells have not been investigated.

This study aimed to investigate the influence of EV-D68 infection on TDP-43 distribution and expression and revealed that viral protease 3C triggers TDP-43 cleavage, aggregation, and cytotoxicity in host cells, including human neural cells. With the recent global transmission of EV-D68, the finding that EV-D68 3C can induce pathological TDP-43 formation widens our understanding of virus cytotoxicity and serves as a basis to monitor and control EV-D68-associated clinical symptoms.

RESULTS

EV-D68 infection triggers the cytoplasmic accumulation and cleavage of TDP-43.

This study investigated the influence of EV-D68 on TDP-43 localization. EV-D68 infection triggered the cytoplasmic accumulation and cleavage of TDP-43. Fluorescence imaging was performed by generating a construct expressing TDP-43 and fluorescent mCherry fusion protein. Fluorescence data suggested that TDP-43 translocated into the cytoplasm and formed inclusions in the EV-D68-infected cells, whereas TDP-43 accumulated in the nuclei of the uninfected cells (Fig. 1A). Considering that the cytoplasmic aggregation of TDP-43 is a key pathological feature of several neurodegenerative diseases, including amyotrophic lateral sclerosis (22), we conducted further analyses.

FIG 1.

EV-D68 infection triggers cytoplasmic translocation and cleavage of TDP-43. (A) Subcellular localization of mCherry-TDP-43 in HEK293T cells. The cell nuclei were stained with DAPI (4′,6-diamidino-2-phenylindole). White arrows indicate the inclusions in the cytoplasm. Bar = 10 μm. (B) Immunoblotting data of endogenous TDP-43 cleavage in the presence of EV-D68 Fermon. HEK293T cells were cultured in 12-well plates until they reached 80% to 90% confluence and infected with EV-D68 (multiplicity of infection [MOI] = 0.04) for 24 h. TDP-43 FL refers to full length TDP-43. TDP-43 Cl. refers to cleaved TDP-43. (C) Endogenous TDP-43 cleavage in HEK293T cells infected with an increasing amount of EV-D68 Fermon at MOIs of 0.01, 0.02, 0.04, and 0.16 for 24 h. (D) Cleavage of overexpressed hemagglutinin (HA)-TDP-43 in HEK293T cells infected with EV-D68 Fermon (MOI = 0.04) for 24 h. (E) Endogenous TDP-43 cleavage after infection of EV-D68 prototype Fermon, EV-D68 US/MO/14-18947 (MO), or EV-D68 US/KY/14-18953 (KY) at a MOI of 0.04. HEK293T cells were infected with EV-D68 for 24 h. The cell lysates were analyzed using immunoblotting.

Specifically, we investigated the effects of EV-D68 infection on the expression of endogenous TDP-43. Immunoblotting assay results revealed a TDP-43 truncation at ~35 kDa in the EV-D68-infected HEK293T cells but not in the uninfected cells (Fig. 1B), suggesting that TDP-43 proteins underwent cleavage during EV-D68 replication. This cleavage occurred in a dose-dependent manner in the cells challenged with increasing concentrations of EV-D68 (Fig. 1C). N-terminal hemagglutinin (HA)-tagged TDP-43 was ectopically expressed in HEK293T cells, which were subsequently challenged with EV-D68 at a multiplicity of infection (MOI) of 0.04. The treated cells were harvested 24 h postinfection and then prepared for immunoblotting. The results showed that TDP-43 was cleaved after EV-D68 infection (Fig. 1D). Considering the distinct features between the EV-D68 prototype strain Fermon and circulating isolated strains (5, 8), we compared the effects of different EV-D68 strains on TDP-43 cleavage. The results showed that all the tested EV-D68 strains efficiently induced TDP-43 cleavage (Fig. 1E).

EV-D68 can infect human neural and respiratory cells. Thus, we measured the influence of EV-D68 on TDP-43 proteins in human airway cells (A549 and BEAS-2B), glioblastoma cells (T98G), and neuroblastoma cells (SH-SY5Y). Despite having different replication capacities, EV-D68 could productively infect all these cell lines (Fig. 2A). Immunoblotting data demonstrated that EV-D68 infection induced TDP-43 cleavage in all the tested cells (Fig. 2B to D). The induction of TDP-43 cleavage was more efficient in respiratory cells (A549 and BEAS-2B) than in neural cells (T98G and SH-SY5Y) when the cells were challenged with EV-D68 at low virus doses (MOI = 0.008), consistent with the diverse replication capacities of EV-D68 in these cells. Hence, TDP-43 is a conserved host factor that is targeted by EV-D68 for cleavage. In addition, consistent with the previous results, we further observed that EV-D68 resulted in TDP-43 translocalization in T98G cells (Fig. 2E).

FIG 2.

EV-D68 induces TDP-43 cleavage in respiratory and neural cells. (A) Determination of infectious viral titers in the culture supernatants of EV-D68-infected A549, BEAS-2B, T98G, and SH-SY5Y cells. These test cells were infected with EV-D68 (MOI = 0.008) for 48 h. (B) Human lung carcinoma A549 and lung epithelial BEAS-2B cells were separately challenged with EV-D68 Fermon (MOI = 0.008) for 24 h. Expression of TDP-43 proteins was analyzed through immunoblotting. (C, D) Cleavage of endogenous TDP-43 by EV-D68 Fermon (MOI = 0.08) in glioblastoma T98G (C) and neuroblastoma SH-SY5Y (D) cells for 48 h. (E) Subcellular localization of TDP-43 in T98G cells. T98G cells were infected with EV-D68 (MOI = 0.08) for 48 h. The cell nuclei were stained with DAPI. White arrows indicate the inclusions in the cytoplasm. Bar = 10 μm.

EV-D68 3C protease is responsible for the cleavage of TDP-43.

To explore the mechanisms by which EV-D68 induced pathological TDP-43 formation, we performed unbiased screening to analyze the effects of EV-D68-encoded proteins on TDP-43 expression. The results of the mCherry-TDP-43 fluorescent reporter assay confirmed that EV-D68 2A triggered TDP-43 translocation. In the 2A-overexpressing HEK293T cells, TDP-43 was distributed in the cytoplasm and nuclei, whereas the wild-type TDP-43 was mainly located in the nuclei (Fig. 3A).

FIG 3.

EV-D68 3C is responsible for the cleavage of TDP-43. Experiments were performed in HEK293T cells. (A) Subcellular localization of mCherry-TDP-43 in the presence or absence of EV-D68 2A. The cell nuclei were stained with DAPI. Bar = 10 μm. (B, C) Screening of EV-D68 structural proteins (B) and nonstructural proteins (C) on TDP-43 cleavage. (D, E) HEK293T cells were cotransfected with HA-TDP-43 and increasing amounts of 3C expression vectors (0.05, 0.1, 0.2, 0.4, and 0.6 μg) for 48 h. TDP-43 cleavage was detected through immunoblotting and quantified using ImageJ. (F) Endogenous TDP-43 cleavage in HEK293T cells overexpressing 3C proteins for 48 h.

Immunoblotting data of TDP-43 expression in our screening further indicated that EV-D68 protease 3C induced HA-tagged TDP-43 cleavage, whereas the other structural and nonstructural proteins exerted no influence on TDP-43 expression in HEK293T cells transfected with indicated vectors (Fig. 3B and C). We also observed that HA-TDP-43 expression was decreased when cotransfected with 2A (Fig. 3C), which may be due to the cytotoxicity or inhibition of translation by 2A (23, 24). TDP-43 cleavage was dose-dependent in the 3C-expressing HEK293T cells (Fig. 3D and E). Endogenous TDP-43 was cleaved in the 3C-expressing cells (Fig. 3F).Hence, EV-D68-encoded proteases 2A and 3C contributed to TDP-43 translocation and cleavage, respectively.

TDP-43 cleavage depends on the enzymatic activity of 3C.

The cleavage and pathological aggregation of TDP-43 in neurodegenerative diseases are mainly attributed to the activation of caspases (25), and EV-D68 infection could activate caspase pathways (26). Thus, we inhibited caspase activity by using the pan-caspase inhibitor Z-VAD to confirm the role of caspase activation in EV-D68-mediated TDP-43 cleavage. As expected, Z-VAD treatment drastically blocked the EV-D68-mediated activation of caspase 3 (Fig. 4A). However, Z-VAD showed no detectable effects on EV-D68-mediated TDP-43 cleavage (Fig. 4B), nor did it affect TDP-43 cleavage in the presence of 3C (Fig. 4C). These data suggest that the cleavage of TDP-43 by EV-D68 is not associated with caspase activation.

FIG 4.

TDP-43 cleavage is dependent on the proteolytic activity of EV-D68 3C. (A) Z-VAD inhibits EV-D68-induced caspase 3 activation. HEK293T cells were infected with EV-D68 Fermon (MOI = 0.04) with the medium containing 20 μmol/liter Z-VAD for 24 h. Caspase 3 levels were analyzed through immunoblotting. (B) Z-VAD exerts no influence on EV-D68-mediated TDP-43 cleavage. HEK293T cells were infected or mock-infected with EV-D68 (MOI = 0.04) for 24 h in the medium containing 20 μmol/liter pan-caspase inhibitor Z-VAD-FMK. (C) Z-VAD exerts no suppressive effect on 3C-mediated TDP-43 cleavage. HEK293T cells were transfected with EV-D68 3C with the medium containing 20 μmol/liter Z-VAD for 48 h. (D) Enzyme-defective 3C mutants lose the ability to induce TDP-43 cleavage. The EV-D68 3C and its mutants were transfected into HEK293T cells. (E) 3CL inhibitor GC376 blocks TDP-43 cleavage by EV-D68 infection. HEK293T cells were infected with EV-D68 Fermon (MOI = 0.04) with the medium containing increasing concentrations of GC376 (0.01, 0.1, 0.5, and 1 μmol/liter) for 24 h. TDP-43 expression levels were analyzed through immunoblotting. (F) GC376 inhibited the cleavage of Gasdermin D (GSDMD) by 3CL protease Nsp5 of SARS-CoV-2. HEK293T cells were cotransfected with GSDMD and Nsp5 in the medium containing GC376 5 μmol/liter. GSDMD and Nsp5 expression levels were analyzed through immunoblotting. DMSO, dimethyl sulfoxide.

We investigated the proteolytic activity of 3C during TDP-43 cleavage. We generated expression constructs of 3C enzyme-defective mutants (H40G, E71A, and C147G). These 3C constructs were cotransfected into HEK293T cells by using HA-TDP-43 expression vectors. The samples were then prepared for immunoblotting, and the results indicated that all tested 3C enzyme-defective mutants failed to induce TDP-43 cleavage compared with the wild-type 3C (Fig. 4D). In addition, we demonstrated that the coronavirus 3C-like protease inhibitor GC376 suppressed EV-D68 3C activity, which consequently inhibited EV-D68-mediated TDP-43 cleavage (Fig. 4E). As a positive control, GC376 as an inhibitor targeting coronavirus 3C-like protease Nsp5 was also validated in our system (Fig. 4F) (27). Our results confirmed that EV-D68 3C cleaves TDP-43 proteins through its proteolytic activity.

EV-D68 3C cleaves TDP-43 protein at the residue Q327.

We performed a sequence logo analysis of the known cleavage sites of EV-D68 3C (Fig. 5A) and found a similar motif in the C terminus of TDP-43 (Fig. 5B). To confirm the importance of this motif in the sensitivity of TDP-43 to 3C, we generated a site mutant in the region of TDP-43 (Q327L and Q331A) and included TDP-43 D89E in our investigation. Immunoblotting data indicated that Q327L and Q331A dramatically impaired EV-D68-mediated cleavage, whereas D89E showed less sensitivity to EV-D68 compared with the wild-type TDP-43 (Fig. 5C). Similar results were observed in the 3C-mediated TDP-43 cleavage. TDP-43 Q327L completely abolished recognition by 3C, Q331A was dramatically resistant to 3C, and D89E showed minimal influence on 3C-mediated cleavage (Fig. 5D). Based on the above results, we confirmed that 3C can cleave TDP-43 at Q327.

FIG 5.

EV-D68 cleaves TDP-43 at amino acid Q327 residue. (A) Sequence logo analysis of the cleavage site prediction of EV-D68 3C. (B) Diagram of the functional domains and identified cleavage site of TDP-43. (C) TDP-43 mutants are resistant to EV-D68-mediated cleavage. HEK293T cells were transfected with HA-TDP-43 wild type (WT)/mutants for 24 h and then infected or mock-infected with EV-D68 Fermon (MOI = 0.04) for another 24 h. Cell lysates were subjected to immunoblotting using anti-HA and anti-histone antibodies. (D) TDP-43 mutants are resistant to 3C-mediated cleavage. HEK293T cells were cotransfected with indicated HA-TDP-43 wild type/mutants and EV-D68 3C for 48 h. HA-TDP-43 expression levels were detected through immunoblotting. NES, nuclear export signal; NLS, nuclear localization signal; RRM, RNA-recognition motif.

EV-D68 infection and 3C expression increase TDP-43 aggregation.

In patients with neurodegenerative diseases, abnormal expression or cleavage of TDP-43 is frequently associated with TDP-43 inclusion formation, which is a major cause of disease development. Therefore, we measured the effects of 3C-mediated cleavage on TDP-43 solubility using a cell fraction assay widely used in TDP-43 proteinopathy studies (28). Compared with TDP-43 in the soluble fraction of the uninfected control cells, the 3C-mediated cleaved TDP-43 fragments showed dramatically increased accumulation in the radioimmunoprecipitation assay (RIPA)-insoluble urea-soluble fraction (Fig. 6A), which implied that the 3C-mediated cleavage of TDP-43 decreased its solubility. Consistent with these results, immunofluorescence studies showed that the full-length TDP-43 proteins were uniformly distributed in the nucleus, while much more truncated proteins were translocated into the cytoplasm with observable puncts (Fig. 6B and C).

FIG 6.

3C-mediated cleavage decreases the solubility of TDP-43 proteins. (A) The cleaved TDP-43 fragment solubility decreases in HEK293T cells. The cells were transfected with TDP-43 FL or TDP-43 Cl for 48 h, and the cell fraction and proteins were sequentially extracted using radioimmunoprecipitation assay (RIPA) and urea buffers. Immunoblotting was performed using an anti-HA antibody to detect full-length and truncated TDP-43. (B) Subcellular localization of full-length TDP-43 and cleaved TDP-43 in HEK293T cells. HEK293T cells were transfected with TDP-43 FL and TDP-43 Cl plasmids for 48 h. The cell nuclei were stained with DAPI. Bar = 10 μm. (C) Proportion of TDP-43 distributed in the cytoplasm in cultures from panel B. (D) EV-D68 infection induces endogenous TDP-43 aggregation. The cells were infected with EV-D68 Fermon (MOI = 0.04) for 24 h, and the cell fractions were prepared for immunoblotting using an anti-TDP-43 antibody to detect full-length and truncated TDP-43. (E) Ectopic expression of EV-D68 3C induces endogenous TDP-43 aggregation.

To confirm this conclusion, we investigated the influence of EV-D68 infection on TDP-43 solubility and found that EV-D68 infection decreased TDP-43 solubility (Fig. 6D). In addition, the expression of cleaved fragments and full-length proteins of TDP-43 increased in the insoluble fraction. Similarly, the ectopic expression of 3C significantly increased TDP-43 aggregation (Fig. 6E).

TDP-43 cleaved fragments exert cytotoxicity to T98G glioblastoma cells.

TDP-43 aggregation is associated with cell death. Indeed, we noticed that the expression of 3C-generated TDP-43 truncated fragments obviously disrupted the normal proliferation and morphology of HEK293T cells. Thus, we examined the release of lactate dehydrogenase (LDH) from cells expressing full-length and truncated TDP-43. As expected, the 3C-mediated cleaved TDP-43 fragments showed much higher toxicity than the wild-type TDP-43 (Fig. 7A). Importantly, we further confirmed that the TDP-43 fragments were more cytotoxic to human glioblastoma T98G cells than the full-length TDP-43 (Fig. 7B and C).

FIG 7.

Cleaved TDP-43 fragments are cytotoxic. (A) Cleaved TDP-43 fragments induce cell death in HEK293T cells. The cells were transfected with TDP-43 FL or TDP-43 Cl for 72 h. The medium was harvested for lactate dehydrogenase (LDH) assay. (B) The cleaved TDP-43 fragments induce cell death in T98G cells. The cells were transfected with TDP-43 FL or TDP-43 Cl for 72 h. The medium was harvested for LDH assay. (C) Expression levels of wild-type and cleaved TDP-43 in T98G cells.

In addition, we confirmed the cytotoxicity of EV-D68 3C by measuring cell proliferation (Fig. 8A and B) and LDH release (Fig. 8C). Treatment with the small-molecule inhibitor GC376, which can block TDP-43 cleavage, significantly restored cell viability (Fig. 8A to C) without exerting additional effects on cell proliferation (Fig. 8D). These data imply that 3C protease can be used to treat 3C-induced neurotoxicity.

FIG 8.

GC376 treatment relieves the cytotoxicity of EV-D68 3C. (A) GC376 disrupts 3C-mediated suppression of cell proliferation. The cells were transfected with 3C expression vector with or without GC376 (10 μmol/liter) and then stained with crystal violet. The cell viability was determined with Cell Counting Kit-8 (CCK8) assay (B). (C) GC376 decreased the LDH release triggered by EV-D68 3C. (D) Viability of cells treated with indicated GC376 concentration.

Lopinavir inhibits 3C-mediated TDP-43 cleavage.

The cleavage of host TDP-43 proteins by EV-D68 3C can also serve as a reporter to screen novel drug candidates targeting the enzyme activity of 3C. We then investigated the anti-3C effects of two commercial protease inhibitors (lopinavir and nelfinavir) that have been used in the clinical treatment of HIV. Our results indicated that only lopinavir significantly inhibited EV-D68-mediated TDP-43 cleavage (Fig. 9A and B). The inhibition of 3C function by lopinavir was further confirmed by dosage-effect experiments (Fig. 9C and D). Consistent with these results, lopinavir treatment directly suppressed TDP-43 cleavage in the presence of 3C expression (Fig. 9E and F). As a positive control, the ability of lopinavir as an inhibitor targeting HIV protease was validated in our system (Fig. 9G). Hence, lopinavir has potential applications in controlling the emerging EV-D68 pandemic by directly inhibiting the enzymatic activity of 3C.

FIG 9.

Lopinavir is a potent inhibitor for 3C-mediated TDP-43 cleavage. (A, B) Screening of protease inhibitors against EV-D68-mediated TDP-43 cleavage. The cells were challenged with EV-D68 Fermon (MOI = 0.04) for 24 h with the medium containing 20 μmol/liter lopinavir or 10 μmol/liter nelfinavir. Cell lysates were subjected to immunoblotting. ****, P < 0.0001. DMSO and ethyl alcohol (EtOH) were used as control for lopinavir and nelfinavir, respectively. (C, D) Lopinavir inhibits TDP-43 cleavage by EV-D68 infection. The cells were infected or mock-infected with EV-D68 Fermon (MOI = 0.04) for 24 h in the medium containing increasing concentrations of lopinavir (4, 8, 16, and 20 μmol/liter). The cell lysates were subjected to immunoblotting. (E, F) Lopinavir inhibits 3C-mediated TDP-43 cleavage. The cells were transfected with EV-D68 3C for 48 h with the medium containing increasing concentrations of lopinavir (4, 8, and 16 μmol/liter). Endogenous TDP-43 expression levels were analyzed through immunoblotting and quantified using ImageJ. TDP-43 Cl % is the percentage of cleavage product compared to the control in the absence of the inhibitor. (G) Lopinavir inhibited the infection of HIV-1ΔEnv-GFP virus-like particles.

Silencing endogenous TDP-43 expression has no influence on EV-D68 replication.

To further clarify whether TDP-43 plays a role in EV-D68 infection, A549 cells and T98G cells were treated with small interfering RNA (siRNA) specific for TDP-43 mRNA, followed by EV-D68 infection. Fig. 10 (A and D) shows that TDP-43 expression was significantly downregulated in A549 and T98G cells transfected with related siRNA. Silencing of TDP-43 did not affect the expression of viral structural protein VP1 compared to that in the control group (Fig. 10B and E). Furthermore, there was no significant difference in the titer of the progeny virus compared to that in the control (Fig. 10C and F). Hence, these data indicate that TDP-43 has no influence on EV-D68 infection.

FIG 10.

TDP-43 knockdown has no influence on EV-D68 replication. A549 (A to C) cells and T98G cells (D to F) were transfected with an small interfering RNA (siRNA) targeting TDP-43 (siTDP-43) or a scrambled siRNA (siCON) for 24 h followed by EV-D68 infection for 48 h at an MOI of 0.008 and 0.08, respectively. Western blotting was conducted to examine protein expression of TDP-43 (A and D) and EV-D68 VP1 (B and E). Virus titers of EV-D68 (C and F) in the supernatant were determined.

DISCUSSION

Cytoplasmic translocation and aggregation of TDP-43 are the key pathological features of several neurodegenerative diseases. Abnormal TDP-43 expression has been observed in all cases of ALS except those carrying SOD1 and FUS mutations (29, 30), ~50% of FLTD cases (31), and many cases of Alzheimer’s disease and Parkinson’s disease (32). In the present study, we found that EV-D68 can induce similar pathological TDP-43 formation and cytotoxicity upon infection.

The results of this study help us understand the mechanisms by which EV-D68 exerts toxic effects to host cells during infection. Viral proteases 2A and 3C are responsible for the translocation and cleavage of TDP-43, respectively. Specifically, 2A induces the accumulation of TDP-43 from the nucleus to the cytoplasm. Meanwhile, 3C triggers TDP-43 cleavage and aggregation. In addition, we confirmed that the aggregated TDP-43 fragments are cytotoxic to different human cells, including neural T98G cells.

After entering the host cells, enteroviruses undergo viral RNA replication, protein synthesis, and newly formed virion assembly and release. Most of these processes occur in the cytoplasm, whereas a series of nuclear RNA-binding proteins are essential for viral replication. To induce the accumulation of nuclear cofactors in the cytoplasm, enterovirus-encoded 2A disrupts the host nuclear transport pathway and alters nuclear permeability by proteolyzing nuclear pore complex proteins (Nups), particularly Nup153, Nup98, and Nup62 (33, 34). TDP-43 typically resides in the nucleus but can shuttle between the cytoplasm and nucleus. Thus, blockage of the nuclear transport system by 2A would retain TDP-43 protein accumulation in the cytoplasm.

Emerging evidence suggests that TDP-43 proteins play important roles in modulating innate immune responses. Cytoplasmic TDP-43 facilitates the release of mitochondrial DNA and subsequently activates the cGAS-STING pathway to elicit type I interferon (IFN-I) production (35). Abnormal TDP-43 expression triggers a retinoic acid-inducible gene-I/mitochondrial antiviral signaling protein (RIG-I/MAVS)-dependent IFN response via intracellular double-stranded RNA (dsRNA) exposure (36). Hence, the EV-D68 2A-induced cytoplasmic translocation of TDP-43 increases the risk of initiating the host cGAS-STING innate immune cascade, which disrupts virus replication. Another protease encoded by EV-D68, 3C, specifically cleaves TDP-43 at Q327 (Fig. 4). The cleaved TDP-43 fragments retain the integrity of two RNA-recognition motifs, which can bind and protect dsRNA from RIG-I recognition but cannot target the mitochondria and release mitochondrial DNA (mtDNA). The detailed mechanism by which EV-D68 evades the TDP-43-dependent innate immune activation must be elucidated in the future to develop novel strategies against viral immune evasion.

TDP-43 was initially identified as a restriction factor that binds to the HIV-1 LTR region and suppresses HIV-1 transcription and gene expression (37). However, we did not detect significant changes in the viral replication capacity of EV-D68 in A549 and T98G cells with knocked down endogenous TDP-43 expression by RNA interference (RNAi) compared with control cells (Fig. 10). Considering that TDP-43 is ubiquitously expressed, we cannot eliminate the influence of TDP-43 expression on EV-D68 infection and replication in other types of human cells or in vivo. However, at the current stage, we need to pay attention to the virus-mediated dysfunction of host factors that modulate human disease development, including those factors that are not closely related to virus infectivity but may contribute to viral diseases.

Accumulating evidence has demonstrated that TDP-43 solubility interferes with infection by different viruses, including Theiler’s murine encephalomyelitis virus (38) and coxsackievirus CV-B3 (39). The results of the present study on the human respiratory virus EV-D68 support this hypothesis. Intracellular protein aggregation is a common cause of brain disease. Aside from the increasing TDP-43 inclusions observed in virus-infected cells, α-synuclein aggregation (40), an important hallmark of Parkinson’s disease, also occurs during H1N1 influenza infection. Other forms of virus-triggered protein aggregation are expected and urgently need to be identified in the future.

CNS disorders have been associated with acute infection or postinfection. After the EV-D68 outbreak in North America, EV-D68 has been analyzed for its potential relationship with AFM. The EV-D68-induced cytoplasmic aggregation of TDP-43 warrants further analyses to control the intracranial infection of EV-D68, which has been detected in vivo and in vitro. In addition, our results suggest that counteracting the enzymatic activity of viral proteases is effective for relieving EV-D68 cytotoxicity. Meanwhile, the cleavage of TDP-43 by 3C could serve as a reporter for screening novel drug candidates targeting EV-D68 3C. We have identified lopinavir as an inhibitor of 3C function (Fig. 9). Whether EV-D68 and other viruses that induce TDP-43 aggregation could cause cognitive disorders in immunocompetent animal models challenged with sublethal titers requires further investigation. The establishment of EV-D68-associated CNS disease models may also advance TDP-43-related neuroscience.

Taken together, our results demonstrate that the protease activity of EV-D68 3C can cleave TDP-43, leading to cytotoxicity in host cells, including neural cells. In view of the critical role of pathological TDP-43 formation in neural disorders, our studies further emphasize that young patients with EV-D68 infection should be monitored to control viral load in the CNS and prevent cognitive sequelae.

MATERIALS AND METHODS

Plasmids and reagents.

VR1012-HA-TDP-43 and pmC1-mCherry-TDP-43 plasmids were constructed by Generay Biotech (Shanghai) based on the sequence we provided (GenBank accession number 23435). HA-TDP-43-Cl was constructed using the Hieff Clone Plus Multi one-step cloning kit (YEASEN, catalog no. 10912), and HA-TDP-43 was used as a template. For PCR amplification, the following primers were used: forward 5′-TCT GCA GTC ACC GTC GTC GAC GCC ACC ATG-3′ and reverse 5′-AGG CAC AGC AGA TCT GGA TCC CTA GGC GGC AGC CAT CAT-3′. The N-terminal HA tag and amino acids 1 to 326 of HA-TDP-43 were retained in HA-TDP-43 Cl. The EV-D68 2A, 2B, 2C, 3A, 3C, and 3D expression vectors with an N-terminal HA tag were purchased from Generay Biotech (Shanghai). EV-D68 VP1, VP3, and VP2/4 strains were provided by Tao Wang. The empty vector VR1012 was provided by Vical (San Diego, CA, USA). HA-TDP-43 and EV-D68 3C mutants were generated through site-specific mutagenesis. Gasdermin D (GSDMD) plasmid was kindly provided by Dr, Xing Liu. 3C-like protease Nsp5 of SARS-Co-2 plasmid was provided by Peihui Wang. pCMV-VSV-G was purchased from Addgene (Cambridge, MA, USA). HIV-1ΔEnv-GFP plasmid was provided by Robert Siliciano (Johns Hopkins University).

An anti-TDP-43 antibody (dilution 1:5,000; catalog no. ab104223) was purchased from Abcam (Cambridge Biomedical Campus, UK), an anti-enterovirus D68 VP1 antibody (dilution 1:1,000; catalog no. GTX132313) was purchased from GeneTex (San Antonio, CA, USA), a monoclonal mouse anti-α-tubulin antibody (dilution 1:2,000; catalog no. A01410) was purchased from GenScript (Piscataway, NJ, USA), a polyclonal rabbit anti-HA antibody (dilution 1:2,000; catalog no. 71-5500) was purchased from Thermo Fisher Scientific (Waltham, MA, USA), an anti-histone H3 antibody (dilution 1:3,000; catalog no. ab176842) was purchased from Abcam (Cambridge Biomedical Campus, UK), an anti-Flag antibody (dilution 1:1,000; catalog no. F1804) was purchased from Sigma (Shanghai, China), and an anti-GSDMD antibody (dilution 1:2,000; catalog no. HPA044487) was purchased from Sigma (Shanghai, China). The reagents used in this study were lopinavir (catalog no. A8204; APExBIO Technology), nelfinavir (catalog no. A3653; APExBIO Technology), GC376 (catalog no. S0475; Selleck), and DAPI (4′,6-diamidino-2-phenylindole) (catalog no. F6057; Sigma).

Cells and viruses.

Human embryonic kidney 293T cells (American Type Culture Collection [ATCC], CRL-3216), human lung carcinoma A549 cells (CRM-CCL-185), human bronchial epithelial BEAS-2B cells (CRL-9609), human glioblastoma multiforme T98G cells (CRL-1690), and human rhabdomyosarcoma RD cells (ATCC, CCL-136) were cultured in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum (FBS) and penicillin/streptomycin solution. Human neuroblastoma SH-SY5Y cells (CRL-2266) were cultured in MEM/F12 supplemented with 20% FBS, 1% nonessential amino acid, and penicillin/streptomycin solution.

EV-D68 prototype Fermon (ATCC, VR-1826) and EV-D68 circulating strains from the 2014 U.S. outbreak, US/MO/14-18947 (ATCC, VR-1823D), and US/KY/14-18953 (ATCC, VR-1825D) were propagated in RD cells. EV-D68 viruses in the supernatants of infected cells were harvested and passed through a 0.45-μm filter. Viral particles were pelleted through a 20% sucrose cushion in an SW28 rotor at 28,000 rpm for 90 min. Purified virions were stored at −80°C.

Transfection and immunoblotting.

HEK293T and T98G cells were grown in 12-well plates for 24 h until 60% confluence. A transfection solution was prepared by mixing plasmids with Lipofectamine 2000 (Thermo Fisher) in serum-free Opti-MEM. The cells were incubated with the transfection mixture for 4 h and then grown in full culture medium. The cells were harvested 48 h after transfection for further analyses. Each transfection was performed in triplicate.

Cell samples were harvested and lysed in RIPA buffer (1 M Tris, pH 7.8, 1 M NaCl, 1% NP-40, and 0.5 M EDTA). The cell lysate was separated on 12% SDS-PAGE gels and then transferred to nitrocellulose membranes using a semidry apparatus (Bio-Rad).

RNAi.

RNAi against TDP-43 in A549 and T98G cells was conducted using siRNA targeting TDP-43 or control nontargeting siRNA (RIBO Biotechnology, Guangzhou). The cells were transfected using Lipofectamine 3000 (Invitrogen, Life Technologies), and 24 h post-transfection, the cells were infected with EV-D68. The cells were collected for immunoblotting 48 h postinfection, and culture medium supernatants were collected to determine viral titers.

Cell viability assay.

Cell viability assays were performed using the cell counting kit-8 (catalog number HY-K0301) assay in accordance with the manufacturer’s instructions.

Cytotoxicity assay.

LDH cytotoxicity assays were performed using an LDH cytotoxicity assay kit (YEASEN, catalog number 40209ES76) in accordance with the manufacturer’s instructions.

Cell fractionation.

The solubility analysis of TDP-43 was performed as previously described (28). Briefly, the cells were washed twice with phosphate-buffered saline (PBS) and lysed with cold RIPA buffer supplemented with protease and phosphatase inhibitors. The cell lysates were sonicated and cleared by centrifugation at 100,000 × g for 30 min at 4°C. The supernatants were collected as RIPA-soluble fractions. RIPA-insoluble protein pellets were washed with RIPA buffer and recentrifuged under the same conditions to prevent carry-over. Only the supernatants from the first extraction were used. RIPA-insoluble pellets were extracted using urea buffer (7 M urea, 2 M thiourea, 4% CHAPS [3-[(3-cholamidopropyl)-dimethylammonio]-1- propanesulfonate], 30 mM Tris, pH 8.5). Immunoblotting was performed for protein analysis.

Crystal violet staining.

Crystal violet staining was conducted as previously described by Feoktistova et al. (41). Briefly, the medium was discarded, and the cells were washed three times with PBS. The cells were incubated with 1 mL of methanol for fixation and rocked gently for 30 min. The methanol was removed, and the cells were washed three times with PBS. The cells were incubated with crystal violet staining solution (0.5% crystal violet in 80% water and 20% methanol) and rocked gently for 30 min. The staining solution was removed, and the cells were washed four times with PBS. The stained cells were left to dry for 4 h on a laboratory bench or in a chemical hood.

Sequence logo analysis.

Multiple sequence alignment was executed using MEGA7 (Molecular Evolutionary Genetic Analysis 7), and the alignment results were used to create the sequence logo on the WebLogo website (http://weblogo.threeplusone.com/create.cgi).

Virus titer assay.

Virus titers were determined by endpoint dilution assay. Briefly, RD cells were cultured under standard conditions in 96-well plates at a density of 10,000 cells/well. EV-D68 was serially diluted (10-fold) with DMEM containing 1% FBS and added to cells. Virus titers were determined by the appearance of cytopathic effects in RD cells using a microtitration analysis in accordance with the Reed-Muench method (42).

HIV-1 pseudovirus package.

For HIV-1 virus-like particle packaging assays, HEK 293T cells were cotransfected with HIV-1ΔEnv–GFP and pCMV encoding VSV-G plasmids at a ratio of 9:1. Dimethyl sulfoxide (DMSO) and lopinavir (20 and 100 nM) were added to the culture 4 h post-transfection, and the cell culture medium was harvested 48 h later. Cell debris was removed by centrifugation at 3,500 rpm for 10 min and filtration through a 0.45-μm filter. Viral particles were pelleted through a 20% sucrose cushion in an SW41Ti rotor at 28,000 rpm for 90 min. Purified virions were stored at −80°C.

Immunostaining and confocal fluorescence imaging.

The cells were fixed with 4% paraformaldehyde in PBS for 30 min, permeabilized in 0.1% Triton X-100 in PBS for 30 min, and blocked in 5% bovine serum albumin (BSA) solution for 1 h. Then, the T98G cells and HEK293T cells were incubated overnight with anti-TDP-43 or anti-HA antibodies at 4°C and incubated with Alexa Fluro 594 antibody (Life Technologies, A-11012) for 1 h at room temperature. The nuclei were counterstained with DAPI (4′,6-diamidino-2-phenylindole). Confocal images were captured using a Nikon laser scanning confocal microscope With a maximum magnification of ×100.

Flow cytometry.

The percentage of GFP-positive cells was measured using a BD LSRFortessa flow cytometer in the FITC channel. A total of 10000 single-cell events were gated using flow cytometry. The percentage of GFP-positive cells was analyzed using the FlowJo (v.10.0.7).

Statistical analysis.

The statistical data were analyzed using GraphPad Prism software (version 8.0; GraphPad Software Inc.). Differences among test groups were analyzed by an unpaired Student’s t test. Statistical significance was set at P < 0.05.

Data availability.

The data that support the findings of this study are available from the corresponding author upon reasonable request.