|Home | About | Journals | Submit | Contact Us | Français|
Hand, foot, and mouth disease (HFMD) is a reemerging illness caused by a variety of enteroviruses. The main causative agents are enterovirus 71 (EV71), coxsackievirus A16 (CVA16), and, most recently, coxsackievirus A6 (CVA6). Enterovirus infections can vary from asymptomatic infections to those with a mild fever and blisters on infected individuals' hands, feet, and throats to infections with severe neurological complications. Viral persistence for weeks postinfection (wpi) has also been documented by the demonstration of virus in children's stools. However, little is known about disease progression, viral spread, and tissue tropism of these viruses. These types of studies are limited because many recently developed mouse models mimic the severe neurological complications that occur in a small percentage of enterovirus infections. In the present study, we documented real-time EV71 infection in two different mouse strains by the use of in vivo imaging. Infection of BALB/c mice with a bioluminescent mouse-adapted EV71 construct (mEV71-NLuc) resulted in a lack of clinical signs of disease but in relatively high viral replication, as visualized by luminescence, for 2 wpi. In contrast, mEV71-NLuc infection of AG129 mice (alpha/beta and gamma interferon receptor deficient) showed rapid spread and long-term persistence of the virus in the brain. Interestingly, AG129 mice that survived infection maintained luminescence in the brain for up to 8 wpi. The results we present here will allow future studies on EV71 antiviral drug susceptibility, vaccine efficacy, transmissibility, and pathogenesis.
IMPORTANCE We report here that a stable full-length enterovirus 71 (EV71) reporter construct was used to visualize real-time viral spread in AG129 and BALB/c mice. To our knowledge, this is the first report of in vivo imaging of infection with any member of the Picornaviridae family. The nanoluciferase (NLuc) gene, one of the smallest luciferase genes currently available, was shown to be stable in the EV71 genome for eight passages on rhabdomyosarcoma cells. Real-time visualization of EV71 infection in mice identified areas of tropism that would have been missed by traditional methods, including full characterization of EV71 replication in BALB/c mice. Additionally, the bioluminescent construct allowed for increased speed and sensitivity of cell culture assays and will allow future studies involving various degrees of enterovirus infection in mice, not just severe infections. Our data suggest that interferon plays an important role in controlling EV71 infection in the central nervous system of mice.
Over the past 2 decades, millions of hand, foot, and mouth disease (HFMD) cases have been reported in the Asian-Pacific region (1, 2). Sporadic outbreaks in day care centers across the United States and Europe have also been documented (3,–5). HFMD is caused by members of the genus Enterovirus (family Picornaviridae), with the predominant causative agents being enterovirus 71 (EV71), coxsackievirus A16 (CVA16), and coxsackievirus A6 (CVA6) (1, 2). Symptomatic cases caused by these viruses are indistinguishable, with symptoms including fever as well as rashes and blisters on the hands, feet, and mouths of infected individuals, usually under the age of 5 years (1, 2). However, EV71 infection poses a higher risk because it has been associated with neurological complications, including brainstem encephalitis and aseptic meningitis. Additionally, the presence of viral RNA in stool weeks after initial infection suggests viral persistence (6). Asymptomatic infections have also been reported and hypothesized to play a role in the spread of these viruses (7, 8). Because there is no cross-reactive, broadly protecting vaccine or antiviral for HFMD, viruses that cause HFMD have become the enteroviruses with the highest global public health concern since the poliovirus epidemic in the early to mid-1900s.
Recent studies have highlighted the fact that humans are the only known natural host for enteroviruses, which has led to a lack of a small animal model for HFMD (9). This has limited in vivo studies on viral pathogenesis, vaccine efficacy, viral dynamics, and spread. The creation of small animal models of EV71 infection has been attempted with a variety of rodent species and adaptation protocols and with the development of transgenic mice containing a human EV71 receptor (scavenger receptor class B, member 2 [SCARB2]) (10,–14). These models have been used successfully for various in vivo studies, but they are unable to monitor real-time patterns of infection. One of the drawbacks of these models is that they require survival and/or clinical signs of infection as a measure of susceptibility (9). They also require euthanasia in order to collect tissues for the assessment of viral infection or therapeutic efficacy. Lastly, many of the models mimic only severe EV71 infection and do not allow for acute, persistent, or asymptomatic infections to be studied.
Bioluminescent reporter systems offer the advantage of visualizing viral dynamics in real time. This strategy has been utilized for many viruses, including monkeypox virus, dengue virus, influenza virus, and herpes simplex virus 1 (15,–18). However, constraints with genetically unstable viruses have halted the development of any type of reporter construct for picornaviruses (19,–21). For example, a recombinant poliovirus containing the Renilla luciferase (RLuc) gene was genetically unstable, losing the reporter gene after only one passage in cell culture (20). Here we report that we have overcome these limitations by using the very small nanoluciferase (NLuc) gene to develop a replication-competent reporter virus suitable for in vivo imaging. NLuc is a 19-kDa engineered luciferase that possesses 150-fold more light output than both RLuc and firefly luciferase (FLuc) (22). This approach allowed us to visualize real-time viral replication in the AG129 (alpha/beta and gamma interferon receptor deficient) and BALB/c mouse strains and to develop faster methods for traditional cell culture assays (e.g., 50% tissue culture infective dose [TCID50] and neutralization assays). These models will allow for further assessment of nonlethal EV71 infections, including persistent infections in immunocompromised individuals. These models could also be used for antiviral drug susceptibility, vaccine efficacy, transmissibility, and pathogenicity studies.
In order to visualize real-time EV71 replication and dynamics in vivo, as well as to develop in vitro reporter systems, replication-competent NLuc constructs were generated using infectious cDNAs for the parent EV71 strain (vEV71) and mouse-adapted EV71 (mEV71) (23). mEV71-NLuc contained two synonymous and two nonsynonymous mutations allowing the virus to infect adult AG129 mice (23). We successfully rescued the NLuc reporter viruses through electroporation of RNA transcripts into RD cells. vEV71-NLuc and mEV71-NLuc passaged 8 times on RD cells did not result in loss of the NLuc gene, suggesting reporter construct stability (Fig. 1B). The growth kinetics of the wild-type and bioluminescent viruses were compared by use of a multistep growth curve. vEV71 reached its peak titer at 24 h postinfection (hpi), whereas mEV71 replicated at a lower rate, requiring 48 h to reach a similar titer (Fig. 1C). mEV71-NLuc and vEV71-NLuc were both attenuated in RD cells and were not capable of replicating to the same peak titers as those of their parent viruses (Fig. 1C). The growth trend for both bioluminescent viruses was similar to that for mEV71. Next, the titers of mEV71-NLuc and vEV71-NLuc samples from the multistep growth curve were measured with a Nano-Glo TCID50 assay. To begin, it was determined if luminescence could serve as a direct measure of viral titer. At 48 hpi, luminescence readings for mEV71-NLuc and vEV71-NLuc stocks paralleled virus titers determined with a standard TCID50 assay (read at 120 hpi) (Fig. 1D). This time point was used to read the luminescence of all the multistep growth curve samples. Titers from the Nano-Glo and traditional TCID50 assays paralleled each other, but the Nano-Glo assay could be used to determine the virus titer in half the time needed for a traditional TCID50 assay (Fig. 1C). Finally, we determined if the insertion of the NLuc gene affected the number of viral particles that completed infectious cycles by calculating the genome/TCID50 ratio. Bioluminescent viruses had higher genome/TCID50 ratios than those of their parent viruses, suggesting that more viral particles were either defective or not completing whole life cycles when the NLuc gene was placed into the viral genome (Fig. 1E). Overall, these initial experiments demonstrated that the NLuc gene was stable in the EV71 genome but that viruses harboring the NLuc gene were slightly attenuated in RD cells. We also demonstrated that luciferase expression was an accurate measure of viral titer at 48 hpi.
Initial studies with mice were done in order to compare mEV71 and mEV71-NLuc infections of AG129 mice. Three-week-old AG129 mice infected with 106 TCID50 units of mEV71 developed neurological clinical signs and succumbed to infection between 9 and 19 days postinfection (dpi) (Fig. 2A). However, only one AG129 mouse succumbed, at 52 dpi, when mice were injected with an equal amount of mEV71-NLuc (Fig. 2A). We also tested the virulence of mEV71 in 3-week-old BALB/c mice. BALB/c mice infected with mEV71 showed no clinical signs of disease, and none of the mice succumbed to infection (Fig. 2A). To determine if viral loads in the sera of infected mice positively correlated with clinical signs of disease, real-time reverse transcription-PCR (RT-PCR) was done on RNA extracts of sera taken from mice at 1, 3, and 5 dpi. AG129 mice infected with mEV71-NLuc had a significant difference in viral loads on all 3 days compared to AG129 mice infected with mEV71 (Fig. 2B). BALB/c mice infected with mEV71 and AG129 mice infected with mEV71 had significant differences in viral loads only at 3 and 5 dpi (Fig. 2B). Taken together, these data showed that mEV71-NLuc was attenuated compared to mEV71 in 3-week-old AG129 mice. Also, BALB/c mice showed no clinical signs of disease or weight loss when they were infected with mEV71, but they had detectable viral loads in serum at 1 dpi.
To better understand the dynamics of viral replication and patterns of spread, both BALB/c and AG129 mice were infected with 2.4 × 106 TCID50 units of mEV71-NLuc and imaged in vivo. Initial studies established that mEV71-NLuc was attenuated in mice; therefore, we increased the amount of virus given for in vivo imaging studies. BALB/c mice were used because of their white fur and previous use in development of mouse models for EV71 (24). AG129 mice were chosen because they have been used successfully in the past as an EV71 animal model for vaccine efficacy studies (13). Initial attempts at imaging determined that visualization of the luminescence in AG129 mice was hindered because of their dark fur. Hair remover (Nair; Church & Dwight Co., Ewing, NJ) was used to remove the stomach fur of the animals. The heads of the mice were shaved to keep the hair remover away from the animals' eyes. A significantly larger amount of luminescence was detected after the fur was removed (Fig. 3). A rapid systemic spread of luminescence from the site of injection (intraperitoneal [i.p.]) was observed in AG129 mice at 1 dpi (Fig. 4 and and5).5). Luminescence was seen in the footpads of infected animals early during experimentation, but no blisters were observed (Fig. 5). This mimicked the tropism of EV71 for human hands and feet. While luminescence on the ventral side of infected AG129 animals continued to diminish throughout the course of infection (Fig. 4), luminescence was observed in surviving mice until the end of the study (56 days) (Fig. 6). A slight increase in luminescence was observed on the ventral side of mice before they met the criteria for euthanasia (e.g., AG129 mice 5 and 6 at 35 dpi) (Fig. 6). This suggested that the mice could no longer control the spread of the virus, succumbing to the infection. In contrast, luminescence in BALB/c mice was localized around the site of injection (i.p.) and was cleared by 3 weeks postinfection (wpi) (Fig. 4 and and5).5). The peak average radiance for both animal strains was obtained at 3 dpi for the ventral and dorsal sides (Fig. 7). Radiance decreased over the course of infection for all mice for the dorsal and ventral views, but as visualized, radiance in the brain region of AG129 mice peaked on day 3 and was maintained throughout the course of infection (Fig. 7).
Surprisingly, 50% of the AG129 mice survived infection and continued to gain weight, but they did show slight clinical signs of disease (e.g., ruffled fur) (Fig. 8A and andB).B). At the end of the study (day 56), the brains of surviving AG129 mice were collected for isolation and genome sequencing. No mutations were found in the genome of any virus from the brain samples, and the NLuc gene was fully intact. For both mouse strains, viral loads in serum were relatively high at 1 dpi and decreased at 3 dpi (Fig. 8C). The numbers of viral copies in fecal samples from both animal strains were also determined weekly by real-time RT-PCR, starting at 1 wpi. No viral RNA was detected in any of the tested fecal samples. These data demonstrate that in vivo imaging can readily be used to visualize viral spread, even in the absence of clinical signs of disease and weight loss, and therefore is a useful tool for studying the various disease outcomes of EV71 infection.
To determine if antibodies were playing a role in the clearance of viral infection from tissues, neutralizing antibody (nAb) titers were measured in AG129 and BALB/c mice at 35 and 56 dpi. nAbs are known to be a major correlate of protection against picornaviruses. We developed a Nano-Glo luminescence neutralization assay to decrease the amount of time needed to determine antibody titers. As found previously with an influenza virus Nano-Glo assay, our luminescence assay was more sensitive than traditional assays that measure cytopathology. This increased sensitivity may make final nAb titers lower than those revealed by traditional methodology, since wells that might be negative based on cytopathic effect (CPE) may still express high levels of luciferase, suggesting active viral replication. All mice seroconverted at 35 dpi, with the presence of nAbs in both animal strains (Fig. 8D). nAb titers were significantly higher in AG129 mice than in BALB/c mice at 35 and 56 dpi. nAb titers also increased over time in AG129 mice, while nAb titers in BALB/c mice remained constant.
Tissue samples collected at the end of the study from AG129 and BALB/c mice were examined for histological changes. Slight lesions were observed in the brains of AG129 mice, with mild brainstem encephalitis represented by multifocal areas of perivascular cuffing and mild gliosis and vacuolation of the neuropil in the thalamic area (Fig. 9). Mononuclear meningitis was also observed. No histological changes were observed in the brain or any tissue sample for BALB/c mice (Fig. 9). This shows that even though AG129 mice did not lose weight or succumb to infection after 56 dpi, neurological abnormalities were still present. Other tissues examined included heart, lung, spleen, liver, intestine, kidney, and hind-leg muscle tissues for both AG129 and BALB/c mice. No histological lesions were observed in any of these tissues at 56 dpi.
In vivo imaging provided useful data for tracking viral spread without having to sacrifice animals. However, it was critical to determine whether the luminescence signal could be used as a proxy measure for virus presence and therefore be used for comparison to previous studies investigating tissue tropism and viral spread. To confirm this, 3 mice each were sacrificed at 3, 6, 12, and 42 dpi to measure viral loads in various tissues. Viral loads in all tissues closely paralleled the average radiances and images from in vivo imaging (Fig. 10). High viral loads were detected in all organs at 3 dpi for both animal strains (Fig. 10). The data also followed viral load trends similar to those seen in previous work looking at viral loads in tissues of 6-week-old AG129 mice infected with mEV71 (23). Viral loads in the brains of AG129 mice could also be detected at 42 dpi. Interestingly, BALB/c mice and AG129 mice had similar viral tissue distributions, but BALB/c mice displayed no clinical signs of disease. We hypothesized that an intact interferon system aided in the clearance of the virus from BALB/c animals, keeping viral loads low in the brains of infected animals. We concluded that the luminescence signal is an adequate proxy for measuring viral tissue distribution.
HFMD continues to be a major public health concern, with millions of cases reported globally every year. Asymptomatic, acute, severe, and even persistent human infections have been documented. However, mouse models for HFMD are limited, mimicking only the rare neurological complications seen in humans. Here we analyzed the dynamics of viral spread in BALB/c and AG129 mice by using a novel bioluminescent EV71 construct. Our work showed that EV71 infections in these animal models paralleled many different types of HFMD manifestations in humans. AG129 mice displayed rapid viral spread and long-term virus persistence in the brains of surviving animals. Virus replication was also visualized in BALB/c mice for 2 wpi, in the absence of clinical signs of disease. These studies provide significant knowledge about enterovirus infections in animal models, with the potential for use in future therapeutic and pathogenesis studies.
The development of reporter constructs for picornaviruses, including enteroviruses, has proven difficult. The large size of many reporter genes (including the RLuc, green fluorescent protein [GFP], and FLuc genes) contributes to their lack of stability in many picornavirus genomes (19, 20). Replicon particles have been created, but they allow for only one round of infection and are therefore not suitable for tracking the dynamics of viral spread (20). In our studies, we used the NLuc reporter gene, which has been used successfully for the generation of influenza virus reporter constructs (17, 25). The NLuc gene was placed after the 5′ untranslated region (UTR), a location recently found to support the insertion of Gaussia luciferase (GLuc) (26), allowing for stable yet slightly attenuated EV71 reporter constructs to be developed.
Using the mEV71-NLuc construct, viral spread was tracked in AG129 and BALB/c mice. Our data revealed a range of EV71 infections in mice, from asymptomatic to severe. Before our studies, adult BALB/c mice were considered resistant to EV71 infection (9). However, we demonstrated that these mice had relatively high viral loads in many tissues and serum and expressed luciferase for several weeks after being infected with mEV71-NLuc. Alternatively, mEV71-NLuc spread to the brains of AG129 mice at 1 dpi and was maintained throughout the course of infection. AG129 mice mimicked the severe complications reported for humans, including brainstem encephalitis, polio-like paralysis, and death (1). We hypothesized that the high serum viral loads at 1 dpi contributed to the quick viral spread to many organs in both animal models. However, the lack of alpha/beta and gamma interferon receptors in AG129 mice contributed to the virus being maintained in the brain, causing the onset of neurological clinical signs of disease. Additionally, in vivo imaging allowed for the visualization of mEV71-NLuc-infected footpad tissues of BALB/c and AG129 mice, even though blisters were not present. Viral spread to the hands, feet, and mouth is a classical symptom in humans (1). Other tissues with small amounts of viral replication were also identified due to the extreme sensitivity of NLuc. The presence of virus in tissues was confirmed by real-time RT-PCR. These results provide evidence that in vivo imaging provides a good representation of EV71 infection and can be related directly to human infection. Also, AG129 mice provide the opportunity to study persistent infection in an immunocompromised model.
Persistent infections have been well documented for many enteroviruses. Postpolio syndrome has affected many poliomyelitis survivors decades after they achieved clinical stability. Poliovirus genomic sequences were detected in 11 spinal fluid samples from 20 patients with postpolio syndrome (27). Poliovirus has also been show to persist in neuroblastoma cells, primary cell cultures of human fetal brains, and mouse spinal cords 12 months after infection (28,–30). In our studies, EV71 was recovered from the brains of AG129 mice at 56 dpi, showing persistence in the central nervous system (CNS) as with poliovirus. Another enterovirus, coxsackievirus B3, persists in mouse heart tissue due to terminal deletions in its 5′ UTR (31). However, unlike the coxsackievirus B3 example, upon sequencing of mEV71-NLuc from persistently infected animals, there were no signs of terminal deletions or mutations.
We hypothesize that the persistent virus expression and high viral loads in the AG129 brains were due to the lack of alpha/beta and gamma interferon receptors. The innate response is critical for EV71 clearance in humans. EV71 infection was shown to boost IP-10 expression, leading to increases in gamma interferon and CD8 T cells that aided in clearing the virus from many tissues (32). C57BL/6J-derived IP-10−/− mice had a 45% increase in mortality when infected with EV71 (32). Interferon is also important in controlling infections by other neurotropic viruses. Alpha/beta interferon helps to control the tissue tropism and pathogenicity of poliovirus (33). Upon infection with West Nile virus, increases in alpha/beta interferon expression in endothelial cells and astrocytes enhanced tight junction integrity at the blood-brain barrier (34). Gamma interferon signaling also aided in closing the blood-brain barrier, preventing entry of virus (34). Together, these studies show the importance of interferon in controlling viral infection. However, further studies will be needed to determine if the lack of interferon contributes to EV71 persistence in AG129 mice. This work may add to our understanding of why immunocompromised patients are at greatest risk for developing severe neurological complications from EV71 infections (35).
Another major factor in controlling picornavirus infections are nAbs, but their role in helping to prevent viral spread to the CNS remains unknown. nAbs cannot cross the blood-brain barrier; however, alteration of blood-brain barrier permeability as a result of proinflammatory cytokines, such as tumor necrosis factor, interleukin-1β (IL-1β), and IL-6, may allow nAbs to enter the CNS parenchyma (36). Recent findings also showed that treatment with antibodies after infection with EV71 led to decreases in clinical signs of disease and viral loads in the brains of infected mice (37). We detected nAbs in AG129 mice at 35 and 56 dpi. However, their role in controlling EV71 infection in the brains of infected mice remains unknown.
We conclude that in vivo imaging provides a rapid means to assess viral spread and allows for better animal models of acute and asymptomatic EV71 infections. It also allows for smaller groups of study animals and enhances the sensitivity for determining protection from viral infections. We also provided data suggesting that interferon is needed to control EV71 infection in the CNS. These results greatly increase the types of future studies that can be done both in vitro and in vivo with EV71 and also expanded to other enteroviruses involved with HFMD.
All animal studies were carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health (38). The IACUC protocol (protocol V5391) was approved by the Institutional Animal Care and Use Committees at the University of Wisconsin (UW)-Madison. AG129 mice were obtained from B&K Universal Limited (Hull, England) and bred in a pathogen-free animal facility at the UW-Madison School of Veterinary Medicine. BALB/c mice were received from Jackson Laboratories.
Human rhabdomyosarcoma (RD; ATCC CCL-136) cells were grown in Dulbecco's modified minimal essential medium (DMEM; Gibco, Carlsbad, CA) supplemented with 2% fetal bovine serum (FBS), 100 U/ml of penicillin, and 100 μg/ml of streptomycin (DMEM growth medium). Plated cells were incubated at 37°C under 5% CO2. The parent EV71-B2 isolate (vEV71), MS/7423/87 (accession no. U22522.1), was obtained from Inviragen Inc. Viral stocks of vEV71 were prepared following previously described procedures (14). Mouse-adapted EV71 (mEV71) was created by serial passage of vEV71 in young, interferon receptor-deficient mice and extensively characterized (14, 23). There were two synonymous mutations (T108A in VP4 and C51T in 2B) and two nonsynonymous mutations (H37R and K244E, both in VP1) between mEV71 and vEV71. The K244E mutation in VP1 was determined to be the critical mutation for mouse adaptation (23).
EV71-NLuc fusions were created by circular polymerase extension cloning (CPEC) (39). Briefly, NLuc sequences were amplified from a full-length PA-2A-NLuc fusion construct, obtained from A. Mehle (UW-Madison) (17). NLuc amplicons were placed between the 5′ UTR and the VP4 gene of previously characterized infectious cDNAs for vEV71 and mEV71 (23) to create a continuous open reading frame (ORF) (Fig. 1A). An EV71 2A cleavage sequence (AITTL) was placed immediately after the NLuc coding sequence. All infectious cDNAs were Sanger sequenced upon completion. Primer and cDNA sequences are available upon request.
Virus recovery from full-length infectious cDNAs was completed for NLuc constructs as previously described (23). Briefly, cDNAs were linearized with SbfI-HF (New England BioLabs, Ipswich, MA), and full-length RNA transcripts were synthesized using a MEGAscript T7 transcription kit (Thermo Fisher Scientific, Waltham, MA) following the manufacturer's instructions. RNA transcripts (18 μg) were electroporated into 1 × 107 RD cells. Once 90% cytopathic effect (CPE) was observed, flasks were freeze-thawed three times, cellular debris was removed, and virus was titrated using a traditional TCID50 assay. The recovered virus was Sanger sequenced to ensure that the virus and NLuc sequences were correct.
To test the stability of bioluminescent EV71 constructs, the reporter viruses were continually passaged eight times on RD cells. For passage 1, 80 to 90% confluent T25 flasks of RD cells were infected with vEV71-NLuc or mEV71-NLuc at a multiplicity of infection (MOI) of 0.1. When CPE was observed, virus was harvested as described above, and 1 μl was blindly passaged into another T25 flask of RD cells. RNA extractions were performed on the viruses harvested from the culture fluids by use of ZR viral RNA kits (Zymo Research, Irvine, CA). The stability of the NLuc gene was tested using a Qiagen One-Step RT-PCR kit (Qiagen, Valencia, CA) with primers that flanked either side of the NLuc gene following the manufacturer's instructions (primer sequences are available upon request). vEV71 was used as a control. A negative control without template was also done.
A multistep growth curve was completed for viruses with and without the NLuc gene. When 48-well plates of RD cells reached 80 to 90% confluence, triplicate wells were infected at an MOI of 0.1 for each virus. After a 1-h absorption period, cells were washed three times with 1× phosphate-buffered saline (PBS), followed by the addition of DMEM growth medium. This was considered time point zero. At each desired time point, one plate (triplicate wells for each virus) was placed at −80°C and the virus harvested as described above. Infectious viral titers were tested for all viruses by using a traditional TCID50 assay. Bioluminescent virus samples were also tested using a Nano-Glo TCID50 assay (described below).
A Nano-Glo TCID50 assay was developed as an alternate measure to determine virus titers. To begin, the optimal time point to test for luminescence was determined. Confluent clear 96-well plates of RD cells were infected with 100-μl aliquots of 10-fold serial dilutions of mEV71-NLuc or vEV71-NLuc. At predetermined time points postinfection, 75 μl of sample was removed from each well, and 25 μl of prepared Nano-Glo substrate solution (Promega, Madison, WI) was added following the manufacturer's instructions. Samples were transferred from the clear plate to a white 96-well plate and read on a Veritas microplate luminometer (Turner Biosystems, Sunnyvale, CA). A comparison was done with a control plate read by traditional TCID50 methods (CPE) after 5 days postinfection (dpi). Following optimization, luminescence-based TCID50 assays were performed on bioluminescent samples from the multistep growth curve analysis (described above).
Real-time RT-PCR was used to determine the ratio of viral genomes to TCID50 titers in RD cells. Triplicate wells in a confluent 48-well plate of RD cells were infected at an MOI of 0.1 for each virus. Virus was harvested at 24 hpi as described above. Supernatants were treated with 50 U/ml of Benzonase nuclease (Sigma-Aldrich Corp., St. Louis, MO) for 1 h at 37°C and 5% CO2 to remove nonencapsidated viral RNA, followed by the addition of TRIzol and extraction of RNA by use of a Zymo Direct-zol RNA miniprep kit (Zymo Research, Irvine, CA). Viral copies were determined using a QuantiTect SYBR green RT-PCR kit (Qiagen, Valencia, CA) following the manufacturer's instructions for a Bio-Rad iCycler iQ5 thermal cycler (Bio-Rad, Hercules, CA). A previously described primer set that targeted a 100-bp region of the VP1 gene was used (23). A standard curve was generated using in vitro-transcribed RNA from an mEV71 infectious cDNA and used to quantify the total amount of viral RNA in each sample. Threshold cycle (CT) values for virus samples were plotted against the standard curve to obtain viral copy numbers. Viral copy numbers were used to determine the number of genomes per milliliter for each sample and then divided by the TCID50 titer per milliliter to get the genome/TCID50 ratio.
Initial experiments used groups (n = 8) of 3-week-old AG129 (alpha/beta and gamma interferon receptor deficient) and BALB/c mice that received intraperitoneal (i.p.) injections of 106 TCID50 units of mEV71-NLuc or mEV71 in 200 μl. Control groups of both AG129 and BALB/c mice received 1× PBS via the same route and in the same volume. All mice were bled at 1, 3, and 5 dpi to test for viral loads in serum by use of the real-time RT-PCR assay described above. RNA extractions from serum were done with a ZR viral RNA kit, and samples were not treated with Benzonase. Survival was also recorded for all mice.
In vivo imaging was done on groups (n = 8) of 3-week-old AG129 and BALB/c mice. Mice received i.p. injections of 2.4 × 106 TCID50 units of mEV71-NLuc in 200 μl (maximum dose possible). Control groups of both AG129 and BALB/c mice received 1× PBS via the same route and in the same volume. Viral loads were measured as described above. Weight loss and survival were also recorded. In vivo imaging was done on 6 animals per group every other day for 3 weeks and weekly after that, for an additional 5 weeks. A control animal from each mock-infected group was also imaged. For imaging, Nano-Glo reagent was diluted 1:20 in 1× PBS and used immediately for imaging. One hundred microliters of the diluted Nano-Glo reagent was injected retro-orbitally into anesthetized mice. In vivo imaging was performed on an IVIS 200 imaging system, and images were analyzed using Living Image software (PerkinElmer). The limit of detection was an average radiance of 2.85 × 103 photons (p)/s/cm2/sr. Fecal samples were collected at each week postinfection (wpi) to test for viral loads by real-time RT-PCR. Samples were prepared as 10% (wt/vol) homogenates with DMEM growth medium, and RNA was extracted using a Direct-zol RNA miniprep kit (no Benzonase treatment). Mice were bled at 35 and 56 dpi to test for neutralizing antibodies. At the end of the study, mice were euthanized and tissues dissected and fixed in 2% paraformaldehyde. Brains from AG129 mice that survived infection with mEV71-NLuc were cut in half and either placed in tissue homogenization tubes for viral sequencing or fixed for histology. Fixed tissues were sent to the histology laboratory at the UW-Madison School of Veterinary Medicine, where they were paraffin embedded, sectioned, and stained with hematoxylin and eosin (H&E).
Groups (n = 12) of 3-week-old AG129 and BALB/c mice were injected i.p. with 2.4 × 106 TCID50 units of mEV71-NLuc. At 3, 6, 9, and 12 dpi, 3 mice were sacrificed and tissue samples aseptically removed to test for the presence of mEV71-NLuc viral RNA. Tissues were placed in tissue homogenization tubes (Biospec Products Inc., Bartlesville, OK) with a mixture of homogenization beads and 300 μl of DMEM growth medium. Tubes were homogenized using a Mini-BeadBeater 24 machine (Biospec Products Inc., Bartlesville, OK) for 3.5 oscillations/min (×1,000) in 1-min intervals. Between intervals, tubes were placed on ice for 1 min. Once they were homogenized, tissues were freeze-thawed and made into 10% (wt/vol) homogenates by use of DMEM growth medium. Homogenates were centrifuged at 2,500 × g for 10 min to remove tissue and cell debris. Viral loads were determined using the real-time RT-PCR assay described above. RNA was extracted using a Zymo Direct-zol RNA miniprep kit. Control samples from mock-infected animals were also included for each mouse tissue.
A Nano-Glo neutralization assay was developed to measure the antibody response against EV71, based on a previously described assay that uses CPE as a readout (14, 25). Individual serum samples were heat inactivated at 56°C for 30 min. Twofold serial dilutions of each sample were mixed with equal volumes of vEV71-NLuc suspension at 2,000 TCID50/ml and incubated at 37°C for 1.5 h. One hundred microliters of serum-virus mixture per well (triplicate wells per sample) was added to a clear 96-well plate of RD cells (final virus titer, 100 TCID50 per well). At 48 hpi, 75 μl of medium was removed from each well, and 25 μl of prepared Nano-Glo substrate solution was added following the manufacturer's instructions. Samples were transferred to a white 96-well plate and read on a luminometer. Neutralization was considered for any well that generated a luminescence reading below the luminescence generated by a well infected with 10 TCID50 units of vEV71-NLuc. The endpoint neutralizing titer was defined as the highest serum dilution for which at least two of the three replicates were below the luminescence limit. A control was done by repeating four samples by use of a traditional neutralization method (with CPE as a readout) after 5 dpi to ensure the accuracy of the Nano-Glo assay.
RNA extractions were performed on the brains of AG129 mice infected with 2.4 × 106 TCID50 units of mEV71-NLuc and euthanized at 8 wpi, using a Direct-zol RNA miniprep kit. Genome-length cDNA synthesis was achieved for each, using Superscript III First Strand synthesis systems (Invitrogen, Carlsbad, CA) with the provided oligo(dT) primer following the manufacturer's instructions. Sequence-specific primers were used to amplify each full-length genome in 9 overlapping fragments by using Q5 high-fidelity DNA polymerase (New England BioLabs, Ipswich, MA) (primer sequences are available upon request). Amplicons were visualized by electrophoresis (1% agarose) and Sanger sequenced using amplifying primers. Sequences were assembled and aligned using Vector NTI (version 11.5; Invitrogen).
Data from animal studies were compiled in Microsoft Excel and analyzed using Prism 6 software (GraphPad, La Jolla, CA). Statistical analyses of growth curves, viral loads, and neutralizing antibody titers were performed using Student's t test (40).
We thank Matthew T. Aliota for his critical review of the manuscript and Monica D. Ronderos for her expertise in pathology and for providing useful histological analysis for the manuscript. We also thank Laura Knoll for providing her IVIS imaging system and animal space to allow us to complete these studies.
The research reported in this publication was supported by the National Institute of Allergy and Infectious Diseases of the National Institutes of Health under award number T32AI055397.
The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.