Sublethal enteroviral infection exacerbates disease progression in an ALS mouse model

Background Amyotrophic lateral sclerosis (ALS) is a fatal neurodegenerative disease of the motor neuron system associated with both genetic and environmental risk factors. Infection with enteroviruses, including poliovirus and coxsackievirus, such as coxsackievirus B3 (CVB3), has been proposed as a possible causal/risk factor for ALS due to the evidence that enteroviruses can target motor neurons and establish a persistent infection in the central nervous system (CNS), and recent findings that enteroviral infection-induced molecular and pathological phenotypes closely resemble ALS. However, a causal relationship has not yet been affirmed. Methods Wild-type C57BL/6J and G85R mutant superoxide dismutase 1 (SOD1G85R) ALS mice were intracerebroventricularly infected with a sublethal dose of CVB3 or sham-infected. For a subset of mice, ribavirin (a broad-spectrum anti-RNA viral drug) was given subcutaneously during the acute or chronic stage of infection. Following viral infection, general activity and survival were monitored daily for up to week 60. Starting at week 20 post-infection (PI), motor functions were measured weekly. Mouse brains and/or spinal cords were harvested at day 10, week 20 and week 60 PI for histopathological evaluation of neurotoxicity, immunohistochemical staining of viral protein, neuroinflammatory/immune and ALS pathology markers, and NanoString and RT-qPCR analysis of inflammatory gene expression. Results We found that sublethal infection (mimicking chronic infection) of SOD1G85R ALS mice with CVB3 resulted in early onset and progressive motor dysfunction, and shortened lifespan, while similar viral infection in C57BL/6J, the background strain of SOD1G85R mice, did not significantly affect motor function and mortality as compared to mock infection within the timeframe of the current study (60 weeks PI). Furthermore, we showed that CVB3 infection led to a significant increase in proinflammatory gene expression and immune cell infiltration and induced ALS-related pathologies (i.e., TAR DNA-binding protein 43 (TDP-43) pathology and neuronal damage) in the CNS of both SOD1G85R and C57BL/6J mice. Finally, we discovered that early (day 1) but not late (day 15) administration of ribavirin could rescue ALS-like neuropathology and symptoms induced by CVB3 infection. Conclusions Our study identifies a new risk factor that contributes to early onset and accelerated progression of ALS and offers opportunities for the development of novel targeted therapies. Supplementary Information The online version contains supplementary material available at 10.1186/s12974-022-02380-7.

Similar to carcinogenesis, the development of ALS was proposed to be a multistep process involving interactions between genetic mutations and environmental risk factors [1,2]. This hypothesis places a high interest in exploring the role of environmental factors including viral infection in the development of ALS. Over the past 40 years, efforts have been made to study the association of ALS with different neurotropic viruses, including enteroviruses (EVs) and retroviruses [8,9]. Among them, human endogenous retroviruses (HERVs) have been widely investigated using clinical diagnostic data, and through in vitro and in vivo experimentations [8, [10][11][12]. There is evidence that HERV-K gene expression and reverse transcriptase activity can be detected in the blood and brain tissues of some ALS patients.
Transgenic mice expressing HERV-K in their neurons display ALS-like motor dysfunction, suggesting a possible role for HERV in ALS pathogenesis [12]. Studies are underway to examine the therapeutic potential of anti-retroviral drugs for ALS (ClinicalTrials.gov Identi er: NCT02437110) [13,14].
Strong evidence supporting the link between ALS and EV infection is still lacking [9]. EVs are a group of single, positive-stranded RNA viruses of the Picornaviridae family that include poliovirus and nonpolioviruses such as coxsackievirus B3 (CVB3), echovirus, EV-A71, and EV-D68 [15,16]. Although EVs commonly cause asymptomatic infections in adults, infection with EVs can lead to severe neurological complications such as poliomyelitis, aseptic meningitis, encephalitis, and non-polio accid paralysis in children [15,16]. It has been reported that EVs can persistently exist in various tissues, including the central nervous system (CNS), and reactivate either spontaneously or in response to external stimulations [17][18][19]. As EVs can target motor neurons, multiple clinical studies have been conducted to explore the possible role of EV infection in ALS [20][21][22][23][24][25][26][27]. However, all these previous studies have been limited to viral genome detection in human ALS blood and tissues, and the available data are controversial and correlative in nature [9]. There is a lack of careful investigation of the potential causal relationship between EV infection and ALS.
Given the possible limitations/challenges faced as discussed in detail previously [9], human viral interrogation study may be unable to provide de nitive answers beyond the uncertainty thus far known.
Therefore, in the current study, we utilized animal models to determine whether EV infection induces ALSlike phenotype in normal mice and/or promotes early onset and progression of ALS in genetically susceptible mice. We also investigated if administration of anti-EV drugs can alleviate virus-mediated ALS-like symptoms. We demonstrated that sublethal CVB3 infection in an ALS mouse model expressing a human mutant SOD1, SOD1 G85R , accelerates disease progression and decreases mouse survival. This observation was accompanied by signi cantly increased immune cell in ltration and proin ammatory cytokine/chemokine gene expression and detection of tissue damage and transactive response DNA binding protein-43 (TDP-43) pathology in the CNS. Moreover, we discovered that application of anti-viral drug ribavirin during the acute phase of viral infection rescues virus-mediated ALS-like pathology and phenotype in mice.
Ribavirin (Sigma-Aldrich, St. Louis, MO; R9644) treatment in CVB3-infected SOD1 G85R mice was performed by subcutaneously injecting either ribavirin (100 mg/kg mouse body weight) diluted in DMEM or an equal volume of DMEM in the neck of neonatal mice. The injections were done every 3 days starting at either day 1 or day 15 post-infection (PI) for a total of 5 injections.
Hindlimb re ex was performed by lifting the mice from the tip of the tail to assess and score their hind leg re ex on a scale of 4-0 (4 = healthy animal with full extension of both hind legs, and 0 = both limbs are fully clasped). Reduction in leg extension is an early de cit observed in mutant SOD1 transgenic mice.
An inverted grid test was performed by placing the mice on an inverted metal grid (10 cm above the surface) and measuring the time (in seconds, with the maximum monitoring time as 120 seconds) for the mice to fall. The test is used to assess the arm/leg strength of the mice [30].
Gait analysis is a test that quanti es the mouse movement based on its gait distance. In the assessment, the mouse feet were rst marked by non-toxic paints and then placed on a long paper (60 cm long, 20 cm wide) that was covered by cardboard, allowing the mice to walk forward in one direction. After the paints were dried, the stride distance (the distance of forward movement between each stride) was measured and recorded. In addition to the gait measurements, the time in seconds (footprint time) needed for each mouse to walk over the track was recorded as an additional measurement of the animal's motor function [31].

Tissue Preparation And Immunohistochemistry Staining
Mouse brains harvested were xed with 4% formaldehyde and para nized. Para n-embedded sections (4 µm thickness) were depara nized through xylene and a series of isopropanol solutions (100%, 90%, and 70%). Hematoxylin and eosin (H&E) staining was conducted to evaluate tissue damages, while immunohistochemical (IHC) staining was used for detecting the presence and localization of proteins of interest. For IHC, antigen retrieval was done by heating tissue sections in pH 6.0 citrate buffer (Life Technologies Carlsbad, CA; 005000) for 25 minutes at 121 o C, and then peroxidase blocked using 30 mg/ml hydrogen peroxide solution. Multiple washes with 1⋅ TBS pH 7.6 (Tris-buffered Saline, 0.

Rna Extraction And Gene Expression Assays
Mouse brain tissues were homogenized using Qiagen Tissuelyzer LT, and RNAs were extracted using the Monarch total RNA miniprep kit (New England Biolabs, T2010) following the manufacturer's instructions. Gene expression was measured using NanoString Mouse Immunology Panel (561 targets with 15 internal reference targets) that was ran on a NanoString nCounter® Pro ler (NanoString Technologies Inc.

Quanti cation And Statistical Analysis
Quanti cation of immunohistochemistry images was performed using ImageJ (version 1.0) with the combination of Colour Deconvolution Plugin (version 1.5) (http://www.mecourse.com/landinig/software/cdeconv/cdeconv.html -accessed in November 2017) to generate the optical density value based on the intensity of the staining as described [32]. Statistical analysis and the corresponding graphs were done using GraphPad Prism 8 V8.4. Further details in statistical analysis are described in Figure Legends.

Sublethal CVB3 infection exacerbates ALS-like phenotypes and decreases survival of ALS SOD1 G85R mice
To determine the impact of EV infection on the development of ALS, the neonates (2-3 days old) of transgenic mice carrying human mutant SOD1 (SOD1 G85R ) and non-transgenic C57BL/6J (background strain used as experimental control) mice were intracerebroventricularly inoculated with a sublethal dose of CVB3 (500 pfu) or an equal volume of DMEM (mock infection) for 10 days, 20 weeks or 60 weeks (Fig.  1A). The viral dosage was chosen based on our previous experience in non-transgenic mice, which does not cause virus-related mortality despite the virus persisting in the CNS for months. Weekly motor function measurements were performed by hindlimb re ex score, inverted grid test, and gait analysis starting at week 20 post-infection (PI). We selected the SOD1 G85R model because these mice develop a slowly progressive and late adult-onset paralysis [33], unlike other models of ALS such as SOD1 G93A mice, which rapidly develop paralysis [34], obscuring the role of EV infection in ALS. In addition, the SOD1 G85R mice express levels of SOD1 G85R equivalent to those of endogenous SOD1, thus closely mimicking human cases of ALS.
Decreased body weight was also observed in CVB3-infected SOD1 G85R mice starting at around week 19 PI compared with mock-infected SOD1 G85R mice (Fig. 1C, right). While CVB3 did not cause mortality in C57BL/6J mice (Fig. 1B), the body weight in CVB3-infected C57BL/6J mice was lower than that in mockinfected counterparts (Fig. 1C, left). Motor function measurements revealed a signi cant decrease in hindlimb re ex score, hanging time, and stride distance, as well as a drastic increase in footprint time in SOD1 G85R mice after CVB3 infection starting around week 40 PI compared with mock infection (Fig. 1D). It was observed that CVB3 infection of C57BL/6J mice did not trigger motor dysfunction (Fig. 1D). Altogether, these results suggest that sublethal CVB3 infection accelerates disease onset, exacerbates motor dysfunction, and decreases survival in the SOD1 G85R mice but does not seem to cause similar phenotypes in C57BL/6J mice.

Sublethal CVB3 infection leads to ALS-related protein pathologies in vivo
Having shown that CVB3 infection leads to exacerbated disease phenotypes and shortened survival in SOD1 G85R mice, we next examined the neuropathological changes in CVB3-infected mice. Similar to our previous observations in Balb/c mice [35], we were able to detect CVB3 capsid protein VP1 in the hippocampus regions of mouse brain at day 10 PI (Fig. 2A, top panel). VP1 signals were not detectable in the brain at week 60 PI as expected due to immune-mediated viral clearance (data not shown). H&E staining showed signi cant tissue damages in the same regions of the mouse brain as VP1 was detected ( Fig. 2A, middle panel).
We also examined ALS-related pathologies through IHC staining of TDP-43 (an RNA-binding protein), SQSTM1/p62 (a ubiquitin-binding autophagy adaptor protein), and ubiquitin. Similar to our previous ndings [35], we observed that CVB3 infection led to cytoplasmic mislocalization of TDP-43 in both SOD1 G85R and C57BL/6J mice at day 10 and week 60 PI (Fig. 2B). In mock-infected mice, TDP-43 remained in the nucleus. Furthermore, SQSTM1/p62-and ubiquitin-positive inclusions were detected in the brain of CVB3-infected but not mock-infected mice at both day 10 and week 60 PI (Fig. 2B). Collectively, these results indicate that CVB3 infection induces ALS-related pathologies (e.g., TDP-43 mislocalization, the pathological hallmark of ALS), which may contribute to the accelerated ALS phenotypes observed in CVB3-infected SOD1 G85R mice.

Sublethal CVB3 infection promotes immune cell in ltration and proin ammatory gene expression
The role of neuroin ammation in the development of neurodegenerative diseases, especially in ALS, has been widely studied and emphasized [36][37][38]. Therefore, we next sought to investigate the in ammatory response in both SOD1 G85R and C57BL/6J mice after CVB3 infection. IHC staining was conducted to assess the astrocyte reactivity with anti-GFAP (glial brillary acidic protein) antibody, microglial activation using anti-Iba1 (ionized calcium-binding adaptor molecule 1) and anti-CD68 (cluster of differentiation 68) antibodies, the presence of B cells with anti-CD19 and anti-CD79A antibodies, the presence of T cells using anti-CD4 and anti-CD8 antibodies, and the presence of natural killer (NK) cells with anti-NK1.1 (an NK cell-speci c antigen) antibody.
We further investigated the alteration of proin ammatory gene expression in both the brain and spinal cord from mock-and CVB3-infected SOD1 G85R mice at week 20 PI by NanoString using nCounter mouse immunology panel. Gene Ontology (GO) analysis showed upregulated genes in the brain (105 genes) and spinal cord (141 genes) of CVB3-infected relative to mock-infected SOD1 G85R mice, which can be categorized into six different functional groups related to immune responses as indicated (Fig. 4A). A closer examination of these signi cantly upregulated genes revealed that many critical proin ammatory cytokine/chemokine genes were present in both the brain and spinal cord samples (Fig. 4B). RT-qPCR analysis con rmed the upregulation of four genes (TNFA, NFKB2, CXCL10, and CCL2), which are shared between the brain and spinal cord samples, in the brain of week 20 CVB3-infected compared to mockinfected SOD1 G85R mice (Fig. 4C). Similarly, we found signi cantly increased expression in multiple proin ammatory genes (TNFA, CCL2, CCL5, IL6, and IL10) in the brain of week 60 CVB3-infected compared to mock-infected SOD1 G85R mice (Fig. 4D). Taken together, our results indicate that CVB3 infection further enhances immune cell in ltration and proin ammatory gene expression in SOD1 G85R mice.

CVB3-accelerated disease progression is mitigated by early viral intervention
CVB3-accelerated disease progression is mitigated by early viral intervention Finally, we examined whether administration of ribavirin, a broad-spectrum anti-RNA viral drug, during the acute or chronic stage of infection can rescue virus-mediated ALS-like neuropathology and symptoms. Ribavirin is a nucleoside analog and acts as a mutagen via incorporation into the viral RNA genome. It can cross the blood-brain barrier and inhibit neurotropic replication of a variety of EVs [16,42]. Ribavirin (100 mg/kg body weight), diluted in DMEM, was subcutaneously injected to CVB3-infected SOD1 G85R mice every 3 days starting at either day 1 (early intervention) or day 15 (late intervention) PI for a total of 5 injections (Fig. 5A). This amount of ribavirin has been previously shown to inhibit CVB3 infection in the CNS [43]. After the last injection, all mice underwent similar experimental procedures as illustrated in Fig.  1A. Kaplan-Meier plots of mouse survival showed that early intervention with ribavirin signi cantly extended mouse survival compared with CVB3-infected SOD1 G85R mice receiving no treatment (average lifespan of 54.4 weeks vs 47.5 weeks, p = 0.0037) (Fig. 5B). However, late intervention with ribavirin did not cause improvement of mouse survival (average lifespan of 48.0 weeks vs 47.5 weeks).
A signi cant improvement in the motor function was observed in CVB3-infected SOD1 G85R mice after early intervention but not late intervention with ribavirin (Fig. 5C). Immunologically, there was a signi cantly lower expression of proin ammatory genes, including TNFA, CXCL10, CCL2, CCL5, IL6, and IL10, in the early intervention group (Day 1) as compared to the non-treated group (Fig. 5D). No difference in the expression levels of proin ammatory genes between the late intervention group (Day 15) and no intervention group was observed. Finally, we were able to show TDP-43 mislocalization in the late, but not in the early intervention group (Fig. 5E). Jointly, the results suggest that application of anti-EV drugs at the early stage of infection attenuates ALS-like phenotype and improves animal survival.

Discussion
The rationale for the current study stems from early evidence that EVs can target motor neurons and from the recent exciting discovery that EV infection produces the hallmark molecular phenotypes of ALS [9], including neuroin ammation, RNA-processing defects, compromised protein quality control and protein aggregation, impaired nucleocytoplasmic transport, and most intriguingly, cytoplasmic mislocalization, aggregation, and cleavage of TDP-43, termed TDP-43 pathology [35,44]. TDP-43 pathology, found in more than 95% of all deceased ALS patients, is not only a pathological hallmark of ALS but also a key disease mechanism for ALS [45]. Despite these exciting observations, a compelling causal relationship between EV infection and ALS development has not yet been established.
Utilizing mouse models, in the present study we demonstrated that chronic, postnatal CVB3 infection hastens disease onset, accelerates motor dysfunction, and shortens the lifespan of SOD1 G85R mice, while similar sublethal viral infection is unable to elicit an ALS-like phenotype in normal non-transgenic mice at least within the timeframe of this study (60 weeks PI). Overall, our results suggest that EV infection serves as a risk/susceptibility factor, rather than a cause for ALS. We propose that at least three mechanisms contribute to the exacerbating effects of EV infection on ALS onset and progression (Fig. 6).
First, in line with previous reports [17,35,43], we showed here that CVB3 infection of the CNS causes focal damages in multiple regions of the brain, which can lead to potential behaviour and motor function changes.
Second, we observed that immunological markers and proin ammatory gene expression are upregulated in the CNS of mice expressing mutant SOD1 and CVB3 infection further enhances immune cell in ltration/activation and cytokine/chemokine gene expression. Neuroin ammation has been identi ed as a key player in the pathogenesis of ALS [36-38]. The evidence presented here suggested an important mechanism by which EV infection worsens ALS-linked phenotypes through promoting aberrant immune responses. We found the presence of several immune cell types such as macrophages, T cells, B cells, and NK cells. While there is evidence supporting the involvement of macrophage, T and NK cells in the development of ALS [46][47][48][49], the role of B cells appears to be insigni cant when comparing the disease progression of SOD1 G93A mice with that of SOD1 G93A mice de cient of B cells [50].
Third, similar to our previous report [35], we observed ALS-related pathologies (TDP-43 mislocalization, SQSTM1/p62-and ubiquitin-positive inclusions) within CVB3-infected mouse brains. TDP-43 pathology is not generally detectable in SOD1-related ALS animal models and human tissues with SOD1 mutations [51,52]. Therefore, CVB3-induced TDP-43 mislocalization likely serves as another mechanism contributing to accelerated disease progression, It is noted that, despite shared pathological and molecular characteristics as that in CVB3-infected SOD1 G85R mice, such as virus-induced tissue damages, increase immune cell in ltration, and positive ALS-related pathologies, CVB3 infection of the C57BL/6J background strain mice is not su cient in inducing motor dysfunction and decreasing survival. We speculate that virus-induced tissue damages, molecular pathologies, and immune responses, under a healthy genetic background, could be tolerated by the host and eventually surmounted (i.e., not reaching the threshold for disease phenotype). However, similar damages/immune responses in combination under a detrimental genetic mutation background such as SOD1 G85R , could contribute to the acceleration of disease progression (Fig. 6).
Finally, as a proof-of-concept, we tested whether the application of an anti-EV drug can attenuate ALS phenotype and improve animal survival. To further understand how CVB3 infection affects the disease progression in SOD1 G85R mice, we treated CVB3-infected mice with ribavirin at either day 1 or day 15 PI. By comparing the results of early versus late treatment, we gain a better understanding of the question whether neurodegeneration will continue if viral infection is stopped and the possible role of "prion-like mechanism" in ALS [53,54]. For example, less effectiveness of late treatment compared with early treatment may suggest a "prion-like mechanism" independent of persistent viral infection for disease progression. In other words, persistent/active infection would not necessarily be required for disease progression. In this study, we demonstrated that early intervention with ribavirin mitigates disease progression and improves mouse survival, whereas late intervention fails to provide a protective effect, suggesting a potential involvement of a "prion-like mechanism" in EV-related ALS. Future studies are warranted to examine the role of a combination of both anti-viral and anti-propagation drugs in this model.

Conclusions
In this study, we identi ed EV as a novel risk factor for the initiation and progression of ALS, which offers potential for future therapeutic interventions. Knowledge gathered from this project will also have broad implications for the study of host-pathogen interactions in other neurodegenerative diseases, in particular frontotemporal dementia, which share many common neuropathological hallmarks and disease mechanisms with ALS [55,56].

Declarations
Ethics approval and consent to participate All procedures involved mouse study were approved by the University of British Columbia Research Ethics Board (Animal Care #A20-0156 and A17-0227).

Consent for publication
Not applicable Availability of data and materials The data supporting the ndings of this study are available from the corresponding author upon request.

Figure 6
Schematic summary of mechanisms that enteroviral infection exacerbates disease phenotype in an ALS mouse model. Sublethal infection of enteroviruses, such as CVB3 modelled in this study, promotes early onset and progression of ALS-like phenotype, and decreases the lifespan of mice in SOD1G85R ALS mouse model. This exacerbation is associated with at least threemolecular and pathological phenotypes induced by CVB3infection:1) direct tissue damages or neuronal cell death, 2)dysregulated immune responses, and 3) TDP-43 pathologies.It appears that sublethal viral infection alone is not su cient to provokeALS-like phenotypesin C57BL/6J mice. However, together with SOD1G85R-mediated toxicities (e.g., increase proin ammatory response, oxidative stress, protein aggregates and neuronal cell death), viral infection accelerates disease progression and reduces mouse lifespanin SOD1G85R ALS mice.