The Human Sodium Iodide Symporter as a Reporter Gene for Studying Middle East Respiratory Syndrome Coronavirus Pathogenesis

Advanced medical imaging such as single photon emission computed tomography with computed tomography (SPECT/CT) enhances fields such as oncology and cardiology. Application of SPECT/CT, magnetic resonance imaging, and positron emission tomography to infectious disease may enhance pathogenesis studies and provide alternate biomarkers of disease progression. The experiments described in this article focus on insertion of a SPECT/CT-compatible reporter gene into MERS-CoV to demonstrate that a functional SPECT/CT reporter gene can be inserted into a virus.

iments are frequently hampered by sacrificing the subject to identify the source of the reporter gene signal (5)(6)(7). Applying medical imaging technology, such as single photon emission computed tomography (SPECT) or positron emission tomography (PET), to animal models of human infectious disease provides the capability to serially monitor anatomical and physiological responses to infection in the same subject that can be clinically translated. Generating a virus that carries a SPECT/PET-compatible reporter gene furthers that capability by serial, real-time evaluation of virus kinetics, identification of tissue tropism, and determination of pathogenic mechanisms. The human sodium iodide symporter (hNIS) gene has emerged as one of the most promising reporter genes in preclinical and translational research for oncology and gene therapy (8,9).
The hNIS symporter is an intrinsic plasma membrane protein belonging to the sodium/solute symporter family, which drives negatively charged solutes into the cytoplasm using a sodium ion electrochemical gradient (10). The advantages of hNIS as an imaging reporter gene include its relatively small size (ϳ2 kb), wide availability of substrates, such as radioiodines, tetrafluoroborate, and 99m Tc-pertechnetate, and wellunderstood metabolism and clearance mechanisms of these substrates (11). Oncolytic viruses such as measles virus and replication-deficient adenovirus that contain hNIS have demonstrated value as theranostics, as hNIS is used both as a therapeutic platform and to track the therapeutic effect (8,9). In addition, hNIS is unlikely to perturb the underlying cell biochemistry, and no negative effects of resultant sodium influx have been observed (12). Finally, once incorporated into the viral genome, the relatively small size of the reporter gene is less likely than larger reporter genes to alter viral pathogenic properties (13).
Middle East respiratory syndrome-CoV (MERS-CoV) recently emerged and is associated with Middle East respiratory syndrome (MERS), a severe, frequently lethal pneumonia in humans (14)(15)(16). Viral pathogenesis is not well understood, in part, because of limited autopsy information and a lack of animal models that fully recapitulate human disease. As with most lethal infectious diseases, animal models are the cornerstone for preclinical countermeasure evaluation and understanding pathogenesis. MERS-CoV provides a unique opportunity to incorporate reporter gene technology to better understand viral pathogenesis because its larger genome size may be more amenable to reporter gene insertion than other viruses.
Animal models for MERS are under development with no single model identified as the standard. New World and Old World nonhuman primates infected with MERS-CoV develop transient respiratory disease with little or no virus replication and varying disease outcome (17)(18)(19). MERS-CoV-exposed New Zealand White rabbits develop limited lung pathology with evidence of viral replication but did not show overt clinical signs of disease (20,21). Transgenic mice globally expressing the human CD26/ dipeptidyl peptidase 4 (DPP4) receptor (22), expressing the human receptor under the murine promoter (23) or transduced with DPP4 receptor (24) become permissive to the virus but do not develop fulminant, lethal respiratory disease. Therefore, changes in reporter gene signal may serve as a biomarker for countermeasure evaluation.
The objective of this study was to incorporate hNIS into MERS-CoV to improve the MERS animal models. Incorporation of a SPECT/PET-compatible reporter gene with an emerging virus such as MERS-CoV requires functional evaluation of the recombinant virus to ensure similar fitness to the parental pathogen. We hypothesized that insertion of hNIS would result in stable expression of a SPECT/PET-compatible reporter gene. A recombinant MERS-CoV carrying hNIS (rMERS-CoV/hNIS) was generated. We assessed the stability, fitness, and functionality of this recombinant pathogen in vitro and in CRISPR-generated transgenic mice that support replication of wild-type MERS-CoV (25).

RESULTS
rMERS-CoV/hNIS genetic stability, kinetics, and fitness. Recombinant virus was evaluated by one-step and multistep kinetics and by serial passaging of the virus. rMERS-CoV/hNIS replicated similarly to rMERS-CoV in Vero E6 cells infected at a multi-plicity of infection (MOI) of 0.01 with a peak in virus yield of 4 log 10 plaque-forming units (PFU)/ml at 24 h. At an MOI of 3, viral yields peaked at 7 log 10 PFU/ml at 48 h and plateaued at 72 h postinfection ( Fig. 1a and b). The correlations between the multistep growth curves for cells infected at an MOI of 0.01 g (r ϭ 0.96 Pearson correlation, P ϭ 0.0082) (Fig. 1a) and one-step growth curves of cells infected at an MOI of 3 (r ϭ 0.90, P ϭ 0.03) (Fig. 1b) for both viruses were high. Both viruses had comparable cytopathic effects ( Fig. 1c and d).
Expression of the hNIS transgene was evaluated by reverse transcriptase PCR (RT-PCR) using primers to specifically detect hNIS expression by MERS-CoV. The resulting 635-bp PCR product includes the MERS-CoV leader sequence and a portion of the hNIS transgene. RT-PCR confirmed hNIS expression in rMERS-CoV/hNIS-infected Vero E6 cells up to 96 h (Fig. 2a). The PCR product was not detected in parental rMERS-CoVinfected cells (Fig. 2b). The RT-PCR assay was also performed to determine the stability of the hNIS transgene following serial passage. The hNIS transgene was stable in rMERS-CoV/hNIS for up to six cell culture passages (Fig. 2c).
hNIS functionality in MERS-CoV/hNIS-infected cells. To demonstrate in vitro hNIS functionality, a series of assays were performed to characterize the kinetics of 99m Tcpertechnetate cellular uptake, the relationship between virus concentration and probe cellular uptake, and between probe dose and its detectability by the gamma camera. The in vitro hNIS functionality assay is outlined in Fig. 3a and b. The 99m Tcpertechnetate uptake by rMERS-CoV/hNIS-infected cells in six-well plates was visible on the tissue culture plate images and distinguishable from the background radioactivity in the wells with uninfected and rMERS-CoV-infected cells (Fig. 3c). In rMERS-CoV/hNISinfected cells, uptake increased with time after infection and reached a maximum, 20% of the dose applied to the well, at the final 96-h postinfection time point (Fig. 4a). The mean radioactivity values measured in the wells were 89.8 Ϯ 5.6, 0.77 Ϯ 0.15, and 0.57 Ϯ 0.04 Ci for rMERS-CoV/hNIS-infected, rMERS-CoV-infected, and uninfected cells, respectively. 99m Tc-pertechnetate uptake was dependent on the virus MOI (Fig. 4b). Higher MOI was associated with greater accumulation of 99m Tc-pertechnetate by rMERS-CoV/hNISinfected cells.
Uptake by rMERS-CoV/hNIS-infected cells was also dependent on the dose of 99m Tc-pertechnetate (Fig. 4c). The relationship between radioactivity measured by the gamma camera on plate images and the dose of 99m Tc-pertechnetate added to each well was linear with high correlation (R 2 ϭ 0.99, linear regression). At 96 hours postinfection with an MOI of 0.01, uptake was slightly above the background level when the 99m Tc-pertechnetate dose of 0.02 mCi per well was applied, and no uptake was detected when 0.004 mCi of 99m Tc-pertechnetate was added to each well (Fig. 4c). Therefore, the 99m Tc-pertechnetate dose should be considered a limiting factor for detection of hNIS expression in rMERS-CoV/hNIS-infected cells. Importantly, no changes  The specificity of 99m Tc-pertechnetate uptake by rMERS-CoV/hNIS-infected cells was confirmed by adding sodium perchlorate, a specific and competitive inhibitor of hNIS protein function, and assessing uptake at 24 h postinfection ( Fig. 5a). At a concentration of 0.1 mM, sodium perchlorate blocked 99.5% of the cellular uptake of 99m Tcpertechnetate in rMERS-CoV/hNIS-infected cells at an MOI of 0.01. Sodium perchlorate concentrations below 0.1 mM reduced the inhibitory effect. For example, radiotracer uptake was inhibited by 50% at sodium perchlorate concentrations of 0.001 mM at 24 h postinfection (Fig. 5a). The effective inhibition of 99m Tc-pertechnetate uptake with 0.1 mM sodium perchlorate declined between 48 and 96 h postinfection (Fig. 5b), suggesting that viral transcription can overcome the inhibitory effect. By 96 h postinfec-  rMERS-CoV Expressing the Sodium Iodide Symporter tion, 99m Tc-pertechnetate uptake in rMERS-CoV/hNIS-infected cells was nearly double the uptake at 24 h postinfection in the presence of sodium perchlorate (Fig. 5b).
In vivo evaluation of rMERS-CoV/hNIS. CRISPR-generated 288/330 ϩ/ϩ DPP4 humanized mice were challenged with parental rMERS-CoV or rMERS-CoV/hNIS by the intranasal route and monitored by SPECT imaging. One group each of mice infected with rMERS-CoV/hNIS and parental rMERS-CoV were euthanized on day 3 postexposure (pe), while another set was monitored to day 7 pe and euthanized. SPECT/CT imaging was performed to determine whether sites of virus replication were evident in these animals. Unfortunately, the low resolution of the clinical scanner used in our biocontainment facility prevented clear discrimination of hNIS-expressing tissues and limited the utility of the imaging. Details of the SPECT/CT data can be found in Fig. S1 and the methodology is described in Text S1 in the supplemental material.
Macroscopic evaluation of the lungs from virus-and sham-exposed animals did not reveal any significant pathological changes except in one of the five mice in group 2 (rMERS-CoV). The right caudodorsal lung lobe of this animal was congested. Microscopically, minimal-to-mild perivascular and peribronchiolar inflammation was noted multifocally with variable congestion in rMERS-CoV-and rMERS-CoV/hNIS-infected mice (Fig. 6a to e). Kidney congestion was also noticeable in 40 and 50% of mice in groups 2 and 3, respectively, on day 3 pe and 33 and 50% on day 7 pe in mice in groups 4 and 5, respectively, infected with rMERS-CoV or rMERS-CoV/hNIS ( Table 1). The percentage of mice with alterations in the lung was higher on day 7 pe compared to that observed on day 3 pe (100% vs Ͻ83%, respectively). Extensive focal congestion in the lungs and multifocal congestion in the kidneys were observed in one of five mice from group 1 receiving sham infection, but inflammation was not manifested in this group (Table 1).
Infectious virus could be detected only in the lungs from 3/6 mice from group 3 (5.01 Ϯ 0.45 log 10 PFU/mg) but was detected in the lung tissue by plaque assay in 4 of 5 mice from group 2 (6.7 Ϯ 0.43 log 10 [mean Ϯ SD] PFU/mg) (Fig. 6f). The differences in viral loads between groups 2 and 3 were not statistically significant (two-tailed t test, P ϭ 0.214). RT-PCR indicated maintenance of the hNIS transgene in 3 of 6 mice in group 3 (Fig. 6g). The inability to detect virus in 3 of 6 mice and the nearly 2 log 10 decrease in PFU/mg observed for rMERS-CoV/hNIS at day 3 supports a loss of viral fitness in vivo. Virus could not be detected by plaque assay or RT-PCR at 7 days postinfection.

DISCUSSION
This is the first report describing the application of a recombinant RNA virus expressing a SPECT/PET reporter gene to study viral pathogenesis. Our data demonstrate the feasibility of using hNIS as a reporter protein in animal models of human disease, provided a higher-resolution SPECT/CT or PET/CT is available. Despite the presence of a robust signal, the clinical grade SPECT/CT in our facility did not provide the necessary resolution to demonstrate in vivo functionality of the rMERS-CoV/hNIS in mice. If available, a micro-PET/CT and the use of tetrafluorborate, an alternate hNIS ligand, could be used to increase sensitivity and improve signal to noise (26). The recombinant MERS-CoV/hNIS has similar fitness to the parental virus, rMERS-CoV, as evidenced by similar kinetics, fitness, and cytopathic effect in vitro. The results of in vivo evaluation also support similar fitness; however, further development would require insertion of the hNIS transgene into other locations within the virus and insertion of the transgene into the MERS-15 virus, which can cause lethal respiratory disease in the 288/300 ϩϩ mice.
Genetic stability of the hNIS transgene cloned into MERS-CoV was confirmed over six in vitro passages. hNIS maintained its function in rMERS-CoV/hNIS-infected cells upon incubation with 99m Tc-pertechnetate. The probe uptake (percentage of radioactive dose applied to each well) positively correlated with virus concentration and time postinfection. rMERS-CoV/hNIS expression of the hNIS protein in infected Vero E6 cells resulted in specific uptake of 99m Tc-pertechnetate and kinetics that correlated with viral replication. Similarly, in a previous study using hNIS-expressing adenoviral vector in carcinoma cells, radionuclide accumulation of 99m Tc-pertechnetate correlated with the amount of adenovirus delivered and the activity of the added radioisotope (27).
Consistent with other studies (28), sodium perchlorate greatly decreased rMERS-CoV/hNIS uptake in cell culture. However, the sodium perchlorate dose that blocked hNIS protein function by 99.5% at 24 h postinfection became less effective at ϩ96 h postinfection, corresponding with increased viral replication, and presumably symporter expression from the virus.
These experiments utilized the 288/330 ϩ/ϩ transgenic humanized mice (25). Results from previous work indicated that these mice, when challenged with the mouseadapted MERS-15 virus, developed severe acute respiratory distress syndrome, includ- (d) Representative mouse infected with rMERS-CoV (group 4) with multifocal, minimal-to-mild perivascular and peribronchiolar inflammation with congestion at day 7 pe. (e) Representative mouse infected with rMERS-CoV/hNIS (group 5) with multifocal, minimal-to-mild perivascular and peribronchiolar inflammation with congestion at day 7 pe. (f) Viral load in lung tissue on day 3 pe in mice infected with rMERS-CoV or rMERS-CoV/hNIS as determined by plaque assay. (g) RT-PCR analysis of the lungs recovered from rMERS-CoV/hNIS-or rMERS-CoV-infected mice. The leftmost lane contains molecular size markers (in kilobases). The next three lanes contain stock controls, rMERS-CoV/hNIS (stock control, without RT step), rMERS-CoV stock, and rMERS-CoV/hNIS stock. Cϩ indicates positive-control virus, and CϪ indictes uninfected negative-control cells. Samples from groups 1 to 5 are shown in the five sets of lanes as follows: group 1, uninfected control (CϪ) group, lanes 1 to 5; group 2, exposed to rMERS-CoV, lanes 1 to 5; group 3, exposed to rMERS-CoV/hNIS, lanes 1 to 6; group 4, exposed to rMERS-CoV, lanes 1 to 5; group 5, exposed to rMERS-CoV/hNIS, lanes 1 to 6. The asterisk in lane 5* of the rMERS-CoV/hNIS-treated group on day 3 pe indicates altered PCR conditions (48°C annealing temperature, 3% dimethyl sulfoxide) to improve the sensitivity to detect rMERS-CoV/hNIS. rMERS-CoV Expressing the Sodium Iodide Symporter ing decreased survival, extreme weight loss, decreased pulmonary function, pulmonary hemorrhage, and pathological signs indicative of end-stage lung disease (25). Similar to the previously published work, the parental (MERS-CoV) and recombinant virus (MERS-CoV/hNIS) grew to high titers in the 288/330 ϩ/ϩ mice, but the mice did not develop disease (25). Although the SPECT signal was dependent on the viral load in tissue culture, the SPECT signal from infected mice as determined by comparing lung-to-heart ratios was not sufficiently different between infected and uninfected mice. Unfortunately, rMERS-CoV/hNIS was detected by PCR in only 50% of infected mice on day 3 pe, suggesting limited in vivo genetic stability.
This inconsistent detection of rMERS-CoV/hNIS could be due to hNIS transgene insertion into the ORF5 gene, which antagonizes innate immunity and thus may impact viral pathogenesis in vivo (29). Cockrell et al. demonstrated that a clone of the MERS-15 virus with an ORF5 deletion was attenuated compared to a clone in which the ORF5 gene was not deleted (25). This attenuation may also explain the lower titer in the rMERS-CoV/hNIS mice compared to that observed with the parental virus (Fig. 6f). Further utilization of hNIS as a reporter for virus growth and localization would include developing the mouse-adapted version (rMERS-15/hNIS-CoV) for evaluation and model improvement and insertion of the hNIS into other locations of the viral genome.
Imaging with a small-animal PET scanner and PET radiotracer 18 F-tetrafluoroborate would improve image quality and sensitivity for visualizing rMERS-CoV/hNIS replication. An alternative to the SPECT is use of a gamma counter to measure the radioactivity in the tissues; unfortunately, a suitable gamma counter is not available in our biosafety level 4 (BSL-4) laboratory. Alternatively, utilization of rMERS-CoV/hNIS in larger animals, i.e., nonhuman primates, should help overcome the SPECT clinical scanner's limited spatial resolution. This technology also aids identification of tissues that are directly infected, which would improve virological analysis, histological analysis, and characterization of tissue-specific response to infection.
Although the animal models of MERS require further improvement, the model chosen for this experiment was useful in supporting the relative in vivo stability of the hNIS reporter in MERS-CoV. The results from the mouse experiments agree with published studies (20,21,25) and corroborate previous data that indicate that wild-type virus grows to a high titer in the 288/300 ϩϩ transgenic mice. Realizing the potential and limitations of the existing MERS-CoV animal models, further incorporation of real-time imaging technology will expand MERS-CoV research. Development of recombinant viruses with imaging reporter genes and its application with suitable animal models will be instrumental in furthering our understanding of viral pathogenesis, potentially leading to improved animal models of human disease and more efficient countermeasure evaluation.

Cells, viruses, and animals. Recombinant parental virus (rMERS-CoV) and MERS-CoV expressing hNIS
(rMERS-CoV/hNIS) were generated as previously described (30). Briefly, the hNIS open reading frame (ORF) was inserted into the NS5 ORF of MERS-CoV. The nine bases, UCCUUCAUA, between the M gene TRS and its start codon were included at the start of the hNIS sequence and cloned using the SanDI restriction enzyme site upstream of ORF 6. Transfections for recovery of the recombinant viruses were performed at biosafety level 3 (BSL-3). Except for generation of the recombinant viruses, all experimental  (3) procedures were conducted at BSL-4 biocontainment laboratories due to the location of the SPECT clinical scanner. MERS-CoV isolate Hu/Jordan-N3/2012 was propagated in MRC-5 cells (ATCC) at a multiplicity of infection (MOI) of 0.1 for 5 days following published procedures (19,31). Virus was recovered by removal of the tissue culture media followed by centrifugation, titers of the virus were determined by plaque assay, and the cultures were evaluated for mycoplasma and endotoxin contamination (19,31). Vero E6 cells (ATCC CRL-1586) were maintained in Dulbecco's modified Eagle medium (DMEM) (Lonza) supplemented with 5% fetal bovine serum (FBS) and incubated at 37°C and 5% CO 2 without antibiotics or antimycotics. CRISPR-Cas9-engineered mice (288/330 ϩ/ϩ ; n ϭ 30) encoding two amino acids (positions 288 and 330) that match the human sequence in the DPP4 receptor on both chromosomes were obtained from the University of North Carolina (25).
rMERS-CoV and rMERS-CoV/hNIS plaque assays. Plaque assays were performed in Vero E6 cells as previously described (32). For titration of virus in tissues, the collected tissues were homogenized using a bead-based tissue homogenizer to make a 10% homogenate in DMEM plus 2% FBS (Sigma).
One-step and multistep growth curves and serial passage. Kinetic studies with rMERS-CoV and rMERS-CoV/hNIS were performed in parallel experiments in triplicate. Vero E6 cells were infected with rMERS-CoV or rMERS-CoV/hNIS at an MOI of 0.01 or 3. Samples were collected at 0, 24, 36, 48, 72, and 96 h postinfection, and virus titers were determined as described above. Serial passage was performed to demonstrate stability of the hNIS transgene. Virus was passaged every 72 h for six passages. An aliquot was collected and assayed by reverse transcriptase PCR (RT-PCR) to detect the hNIS transgene.
RT-PCR of viral RNA. For evaluation of stability of the hNIS gene, total RNA was isolated from Vero E6 cells collected 24, 36, 48, 72, and 96 h after rMERS-CoV and rMERS-CoV/hNIS infection at MOIs of 0.01 and 3. The RNA was reverse transcribed into cDNA using deoxythymidine oligonucleotide primers and SuperScript reverse transcriptase II (Thermo Fisher Scientific) according to the manufacturer's protocol. Two custom primers, hNIS Leader (forward primer sequence, 5=CTATCTCACTTCCCCTCGTTCTC) with hNIS6 (reverse primer sequence, 5=GAAGCCACTTAGCATCAC) were used to create PCR products. PCR products were separated and identified on a SYBR safe DNA-stained agarose gel (Thermo Fisher Scientific). Stock rMERS-CoV/hNIS was used as a positive control, and uninfected-cell lysates were used as a negative control for RT-PCR.
Cell culture radio-uptake assay. Functionality of the hNIS gene was assessed by hNIS-mediated accumulation of 99m Tc-pertechnetate by rMERS-CoV/hNIS-infected cells. Vero E6 cells were infected with rMERS-CoV or rMERS-CoV/hNIS at two different MOIs, 0.01 and 0.004, and incubated for 0, 24, 36, 48, 72, and 96 h. 99m Tc-pertechnetate, 0.004 to 0.6 mCi in 0.2 ml, was added to each well and incubated for 60 min. 99m Tc-pertechnetate was removed from the wells, and the cells were washed with room temperature media. Planar images of the plates were acquired over 30 min with a single gamma camera head using the Precedence SPECT/computed tomography (CT) clinical system (Philips Medical Systems), that was modified to operate in a biocontainment environment (33). The in vitro samples were quantified using the SPECT/CT scanner. Uninfected cells were treated with 99m Tc-pertechnetate to determine the background level. Specificity of 99m Tc-pertechnetate uptake by rMERS-CoV/hNIS-infected cells was assessed by adding sodium perchlorate, NaClO 4 , 1 to 100 M, to the wells prior to 99m Tc-pertechnetate introduction. 99m Tc-pertechnetate counts from each well were converted into radioactivity units and decay-corrected relative to the time of probe introduction to the wells. Uptake of 99m Tc-pertechnetate by infected cells was expressed as a percentage of the radioactive dose applied to the well. The radio-uptake assays were conducted in triplicate.
Transgenic mouse experiments. Thirty 288/330 ϩ/ϩ mice, 15 males and 15 females, were split into 5 groups of 6 animals. Two mice, one from group 1 and one from group 4, did not recover from anesthesia after a preexposure scan and were not replaced. Mice were lightly anesthetized with isoflurane and were exposed intranasally with phosphate-buffered saline (PBS) without virus (group 1) or with rMERS-CoV (groups 2 and 4) or rMERS-CoV/hNIS (groups 3 and 5). The virus/PBS was instilled into both nostrils, 25 l per nostril. The mice were monitored daily for clinical signs of disease. The animals from groups 2 and 3 were euthanized on day 3 postexposure (pe), and the animals from groups 1, 4, and 5 were euthanized on day 7 pe. The lungs were collected for gross pathological and microscopic evaluation and virus detection via plaque assay.
Ethics statement. All work with animals was performed in an AAALAC-accredited research facility and approved by the National Institute for Allergy and Infectious Disease Division of Clinical Research Animal Care and Use Committee. The experiments complied with the Animal Welfare Act regulations, Public Health Service policy, and the Guide for the Care and Use of Laboratory Animals (34). Animals were monitored daily. Clinical health, body weight, and food consumption were recorded.
Statistical analysis. Statistical analysis was performed using Graphpad Prism 7.03 (Graphpad Software Inc.) as indicated in the text.