Tissue tropisms, infection kinetics, histologic lesions, and antibody response of the MR766 strain of Zika virus in a murine model

All experiments involving mice were approved by the Louisiana State University (LSU) Institutional Animal Care and Use Committee (IACUC protocol 15-078) in adherence with policies of the American Veterinary Medical Association and in compliance with the guidelines laid out by the National Institutes of Health’s Guide for Care and Use of Laboratory Animals, 2011.

The IRF3/7 double knockout (DKO) mice were originally provided by Dr. Michael Diamond and are deficient in IRF 3 and 7. These mice have a blunted but not abrogated type I interferon response [18]. The prototypical ZIKV strain MR766 was used. Virus was originally obtained from Dr. Robert Tesh at the World Reference Center for Emerging Viruses and Arboviruses at the University of Texas Medical Branch as lyophilized stock and was passaged it once in C6/36 cells and once in Vero cells prior to use in this study. MR766 was originally isolated from the Zika forest in Uganda in the late 1940s and belongs to the African genotype. For post-infection challenge, the PRVABC59 (Asian lineage) was originally received from the CDC (Ft. Collins, CO) and was passaged once on Vero cells prior to our additional passage on Vero cells. Both strains of ZIKV were determined to have a titer of ~10⁷ plaque forming units (PFU)/mL via plaque assay prior to beginning the experiments. Mice were infected by injection subcutaneously in the back with 10⁶ PFU in 100 μl.

Blood samples were allowed to clot at room temperature for 30 min and were then centrifuged for 4 min at 4 °C and 6000 rcf. Serum was separated from the clot and stored in a sterile microcentrifuge tube at −80 °C until further processing. Eye-swab samples were processed by adding 250 μl of M199E media to the swab cotton in a centrifuge tube, vortexing and centrifuging for 2 min at 4 °C and 6000 rcf. The supernatant was extracted and stored in a separate microcentrifuge tube at −80 °C. RNA extraction was performed using the Kingfisher^TM (Thermo-Fisher) automated extraction platform and the Ambion MagMax^TM viral isolation kit (Thermo-Fisher), as per manufacturer’s instructions. Viral RNA was detected by qRT-PCR using the SuperScript™ III One-Step RT-PCR System with Platinum® Taq DNA Polymerase (Life Technologies) on a Roche Lightcycler® 480 (Roche). ZIKV-specific primers and probe were previously designed by Faye et al. [20].

PRNTs were performed following the WHO PRNT protocol for dengue (DENV) which we have previously shown to be applicable to ZIKV [21, 22] on serum collected from surviving ZIKV-infected mice. Briefly, virus was standardized to 25 PFU of ZIKV in 50 μl per well. Mouse serum was complement inactivated for 30 min at 56 °C and then serially diluted 1:2 in M199E media from 1:10 to 1:1280 and mixed with 50 μl of the standardized virus solution and allowed to incubate at 37 °C for 1 h. The mixture was then inoculated onto ~80% confluent Vero cell monolayers in 12-well plates. Also included were a negative control well (media only) and a positive control (virus only). The first overlay was added immediately after incubation, and the second overlay was added 3 dpi for ZIKV. Plaques were visible and counted the following day (4 dpi). The percent reduction in plaques per dilution was calculated. When a negative plaque reduction was observed, the value was set at a baseline of zero. Results are expressed as the reciprocal of the dilution in which the desired percentage of plaque reduction was achieved; we report both PRNT50 and PRNT80.

Statistics were performed using SAS 9.4 (Cary, NC) or RStudio with R version 3.2.2. The changes in weight were reported as the percent lost compared to initial weight. Viremia was logarithmically transformed and reported as log10 titer. We evaluated the differences in weight and viremia between sexes and day post infection with a repeated measures ANOVA analysis, using a mixed model to evaluate the effect of sex, day post-infection, and the interaction of each on the percent weight reduction or the log viremia titer (SAS, PROC MIXED). As our sampling regimen was temporally unevenly spaced, we specified a spatial power covariance structure to account for the uneven intervals. For the heterotypic challenge experiment, PRNT50 and PRNT80 titers were expressed as the reciprocal of the highest dilution at which the desired percentage of plaque reduction was achieved. Differences in percent reduction of neutralization were analyzed using the Kruskal-Wallis rank sum test and the 95% confidence interval calculated from a t-distribution (df = 5). Figures were created with R version 3.2.2 or histologic images compiled in Microsoft PowerPoint (Seattle, WA).

To test the susceptibility of IRF 3/7 DKO mice to ZIKV, we inoculated 10⁶ PFU/mouse of MR766 ZIKV Uganda strain subcutaneously in mice between 6 and 10 weeks of age, one group of males and one group of females. Viral RNA was detected in the serum of all mice, peaking at 2 dpi for both groups. The female mice had on average higher serum viral titers each day, with a peak average log titer of 6.2 compared to 4.49 in males, although this difference was not found to be significant (Fig. 1a, p > .05). Both groups of mice began to lose weight within 2 days of infection. The mice had a maximum average weight loss of 7.6% at 9 dpi for the males and 18.4% at 8 dpi for the females. Weight loss was not significantly different between sexes (Fig. 1b, p < .05). One male died at 7 dpi and two females had to be euthanized on 8 dpi because they lost over 21% of their initial body weight. This translates to an approximate 72% survival rate. The male mouse that died had a higher log titer than the average of the surviving male mice (log titer of 6.1 versus 4.39 of survivors). The two euthanized females had the two highest peak viral titers of the group (log titer of 7.24 and 6.58 versus the average of 5.84 of the other 4 females), although another female of the group had log titer of 6.53 and a peak weight loss of 16.3% (the second highest weight loss), but survived infection.

Signs of overt disease observed in all infected mice included reduced activity, hunched posture, and ruffled fur. These signs began around 5 dpi and increased in severity concurrently with weight lost, reaching a maximum at 8–9 dpi, at which point they began to regain weight and recover. In addition to the non-specific signs, two mice developed ocular disease in the form of crusty discharge in the eye: one male mouse in the right eye at 11 dpi, and one female mouse in the left eye at 6 dpi. The male mouse fully recovered as of 7 months post infection. The female mouse was euthanized at 8 dpi due to excessive weight loss. The ocular discharge was collected and tested for presence of ZIKV RNA, and found in the case of the female mouse to contain 2.280x10³ genome copies/mL of ZIKV NS5 as per the qPCR assay. Attempts to isolate infectious virus from the supernatant were unsuccessful.

To explore the lesions produced by acute infection with ZIKV in IRF 3/7 DKO mice, we infected a second group of IRF3/7 DKO females with the purpose of analyzing their tissues histologically. The viremia in this second group was not statistically different from the other groups (Additional file 1: Figure S1, p > .05). We observed multifocal mild to moderate histologic lesions in the brain, spinal cord and eye of the sacrificed females. Two out of five female mice had encephalomyelitis with lymphocytic perivascular cuffing, gliosis and neuronal necrosis (Fig. 2a). Using IHC, we were able to visualize viral antigen in the hippocampus (Fig. 3a-b) of one of these females. Another female mouse from this group had bilateral epiphora and crusting at 7 dpi; at the time of euthanasia at 12 dpi, it had retinal ganglion cell necrosis and vitreitis, presenting inflammatory cell infiltration of the vitreous humor where in normal circumstances there is none (Fig. 2b).

In addition to these females, we also evaluated histologically the tissues of two mice from the infection kinetics groups: the one male mouse that died and the eyes of the female with ocular discharge that had to be euthanized. In spite of finding ZIKV RNA in the ocular discharge from this female, no histological lesions were found in either the sclera or the retina. The male mouse presented viral antigen visualized by IHC in the in the cerebral cortex (Fig. 3c–d), as well as in the male reproductive organs.

The reproductive organs of the male mouse presented severe necrosuppurative epididymitis associated with abundant viral antigen within the epididymal lining and sloughed off epithelial cells in the lumen (Fig. 4a and b). The affected portion of the epididymis was in stark contrast to other portions of the epididymis with normal architecture and no viral antigen (Fig. 5). Viral antigen was also visualized in germ cells in the seminiferous tubules of the testicles (Fig. 6). In addition, we observed abundant ZIKV antigen staining in the seminal fluid inside the lumen of the ductus deferens, mostly concentrated in what were interpreted as sloughed off epithelial cells (Fig. 7).

No lesions or viral antigen were found in the reproductive organs of the females, nor in skull bones, vertebrae, bone marrow, nerves, spinal root and trigeminal ganglia, upper respiratory tract, lungs, salivary glands, tonsils, spleen, lymph nodes, kidneys, adrenal glands, oral cavity, esophagus, stomach, intestines, skeletal muscle, skin, and pituitary gland.

We show that the IRF3/7 DKO model has similar ZIKV susceptibility characteristics as the other recent models, such as viremia and signs of disease, without being 100% lethal or necessitating humane euthanasia in the majority of cases [1, 2, 5‐7]. Since this is a predominantly non-lethal model, we explored the utility of this model for studies in which the antibody responses would be an important endpoint.

A group of mice was infected with ZIKV MR766 presented with similar infection kinetics and signs of disease as the previous experimental groups (Additional file 2: Figure S2), and one mouse was euthanized at 8 dpi due to excessive weight loss. After the initial infection with the MR766 strain (time point 1), convalescent serum demonstrated a high capacity for neutralization to both the MR766 strain and the PRVABC59 strain from the current outbreak and of the Asian lineage. There was no significant difference in the neutralizing titers at both the 50 and 80% levels between virus strains (p > .05) (Fig. 8), indicating that in vitro the antibody response was cross-neutralizing between the two genotypes. When the neutralization curves of secondarily (PRVABC59) challenged mice against both ZIKV MR766 and PRVABC59 strains were compared, there was no significant difference in the percent reduction of plaques between strains or pre- and post-challenge (Fig. 9) (p > .05), indicating that secondary exposure to PRVABC59 did not result in a change in the neutralizing capacity of the antibodies for either strain. This recapitulates what others have demonstrated [23].

At three-weeks post-secondary challenge with PRVABC59, the average PRNT50 titer was approximately 59733 and 78720 for MR799 and PRVABC59, respectively. The PRNT80 titers were approximately 72747 and 65293 for MR799 and PRVABC59, respectively (Fig. 8). There was no significant difference at either neutralization level for the two strains (p > .05). In addition, there was no significant difference in PRNT50 or PRNT80 levels for either virus strain between time points, suggesting no altering effect of heterotypic challenge as demonstrated by this in vitro assay.