Sofosbuvir protects Zika virus-infected mice from mortality, preventing short- and long-term sequela

Zika virus (ZIKV) caused significant public health concern, because of its association with congenital malformations, neurological disorders in adults and, more recently, with deaths. Considering the necessity to mitigate the cases ZIKV-associated diseases, antiviral interventions against this virus are an urgent necessity. Sofosbuvir, a drug in clinical use against Hepatitis C Virus (HCV), is among the FDA-approved substances endowed with anti-ZIKV activity. In this work, we further investigated the in vivo activity of sofosbuvir against ZIKV. Neonatal Swiss mice were infected with ZIKV (2 x 107 PFU) and treated with sofosbuvir at 20 mg/kg/day, a concentration compatible with pre-clinical development of this drug. We found that sofosbuvir reduced acute levels of ZIKV from 60 to 90 % in different anatomical compartments, such as in blood plasma, spleen, kidney and brain. Early treatment with sofosbuvir doubled the percentage and time of survival of ZIKV-infected animals, despite the aggressive virus challenge assayed and also prevented the acute neuromotor impairment triggered by the virus. On the long-term behavior analysis of ZIKV-associated sequelae, sofosbuvir prevented loss of hippocampal- and amygdala-dependent memory. Our results point out that sofosbuvir inhibits ZIKV replication in vivo, which is consistent with the prospective necessity of antiviral drugs to treat ZIKV-infected individuals.

INTRODUCTION with a 0.1 % solution of crystal violet in 70 % methanol, and the virus titers were calculated by scoring the plaque-forming units (PFU).

Molecular detection of virus RNA levels.
Total RNA from culture, extract containing organs in PBS or plasma was extracted using QIAamp Viral RNA or RNeasy Mini Kits (Qiagen®), according to manufacturer's instructions.
Quantitative RT-PCR was performed using QuantiTect or QuantiNova Probe RT-PCR Kit (Quiagen®) in an ABI PRISM 7300 Sequence Detection System (Applied Biosystems).
Amplifications were carried out in 25 µL reaction mixtures containing 2× reaction mix buffer, 50 µM of each primer, 10 µM of probe and 5 µL of RNA template. Primers, probes and cycling conditions recommended by the Centers for Disease Control and Prevention (CDC) protocol were used to detect the ZIKV 20 . The standard curve method was employed for virus quantification. For reference on the cell amounts used, the housekeeping gene RNAse P was amplified 20 . The Ct values for this target were compared to those obtained to different cell amounts, 10 7 to 10 2 , for calibration.

Animals.
Swiss albino mice (Mus musculus) (pathogen free) from the Oswaldo Cruz Foundation breeding unit (Instituto de Ciência e Tecnologia em Biomodelos (ICTB)/Fiocruz) were used for these studies. The animals were kept at a constant temperature (25 °C) with free access to chow and water and a 12-h light/dark cycle. The Animal Welfare Committee of the Oswaldo Cruz Foundation (CEUA/FIOCRUZ) approved and covered (license number L-016/2016) the experiments in this study. The procedures described in this study were in accordance with the local guidelines and guidelines published in the National Institutes of Health Guide for the Care and Use of Laboratory Animals. The study is reported in accordance with the ARRIVE guidelines for reporting experiments involving animals 21 . The experimental laboratory received pregnant mice (at approximately the 14th gestational day) from the breeding unit. Pregnant mice were observed daily until delivery, to accurately determine the postnatal day. We established a litter size of 10 animals for all experimental replicates.

Experimental infection and treatment.
3-days-old Swiss mice were infected intraperitoneally with 2 x 10 7 PFU of virus 22 .
Treatments with sofosbuvir were carried out with sofosbuvir at 20 mg/kg/day intraperitoneally.
Treatment started one day prior to infection (pre-treatment) or two days after infection (late treatment). Either way, treatment was conducted for 7 days. For comparisons, mock-infected and -treated groups of animals were used as controls. Animals were monitored daily for survival, weight gain and virus-induced short-term sequelae (righting in up to 60 seconds) (Supplementary Video). Animals were euthanized, when differences in weight gain between infected and control groups were > 50% and/or severe illness present. Blood was collected by cardiac puncture and placed on citrate-containing tubes for plasma separation. Tissues (spleen, kidney and brain) were collected. Initially, the tissues were analysed macroscopically, for the presence of pathological signs. Whenever possible, the pathological signs were quantified by counting per verified tissue/organs. Alternatively, tissues were lysed (RLT buffer; Qiagen) and homogenized with Potter-Elvehjem homogenizer (Teflon pestle and glass mortar). Homogenates were cleared by centrifugation and total RNA was extracted.

Behavioral tests.
To test the righting reflex, animals were tested daily during the course of acute infection.
Animals were held in a supine position with all four paws facing up in the air for 5 seconds. After that, animals were released and the time the animal took to flip over onto its stomach with all four paws touching the surface was measured. A maximum of 60 seconds was given for each trial and animals were tested twice a day with a 5 minutes minimum interval between trials. For each animal, the lowest time was plotted in the graph. Animals that failed the test were included in the graph with a time of 60 seconds. Please see the Supplementary Video. Animals at age of 6 to 8 weeks of old were assayed. The Morris water maze (MWM) is a behavioral task to evaluate hippocampal dependent learning and spatial memory. The water maze comprised a black circular pool (100 cm in diameter) that was conceptually divided into four equal, but imaginary, quadrants for the purpose of data analysis. The water temperature was 25 °C. A platform (10 cm 2 ), which was hidden from the mouse's view, was located 2 cm beneath the surface of the water, allowing the mouse to easily climb onto it once its presence was detected. The water maze was located in a well-lit white room with several posters and other distal visual stimuli of timing, it is more opportune to administer sofosbuvir rather than leave the animals without treatment, for the sake of survival.

Sofosbuvir decreases ZIKV loads during acute infection.
Since sofosbuvir inhibits ZIKV replication, enhanced survival due to this treatment was presumably associated with reduction in viral levels during acute infection. We evaluated this hypothesis and measured the magnitude of virus inhibition in vivo. To do so, sofosbuvir-treated ZIKV-infected animals were euthanized daily from the first to the fifth day after infection. Next, viral loads were measured in different tissues (Fig 3). We observed that sofosbuvir reduced the mice viremia over 90 %, especially during the first 3 days after infection, when an exponential increase in viral levels in the plasma was observed ( Fig 3A). Sofosbuvir reduced by up to 60 % the peak of virus detection between 2 and 4 days post-infection in the kidney ( Fig 3B). Virus detection in the spleen and brain was abundant at 4 to 5 days after infection and sofosbuvir treatment reduced up to 80 % virus levels in these tissues (Fig 3C and D). Remarkably, during this experiment, we observed that ZIKV-infected animals had focci of cerebral microhemorrhage.
Altogether, our results indicate that sofosbuvir effects on mice survival was indeed followed by a reduction in virus detection in different anatomical compartments.

Sofosbuvir prevents short-and long-term sequelae in ZIKV-infected mice.
The neonatal animal model may represent a relevant model to evaluate short, and especially, long-term behavioral sequelae after infections. Consistently, we observed an acute neuromotor impairment in ZIKV-infected mice (Supplementary Video). To determine the magnitude of this injury and the benefits of sofosbuvir use, we applied the righting test reflex, for up to 60 seconds. ZIKV-infected animals, untreated with sofosbuvir, took over 12-fold more time to get to the upright position than sofosbuvir-treated animals or controls (Fig 5). No statistically significant differences were observed between sofosbuvir-treated ZIKV-infected mice and the controls (Fig 5). Our date indicate that sofosbuvir protected the animals from ZIKV-associated neuromotor impairment.
Moreover, we had a few ZIKV-infected mice that did not succumb to the infection ( Fig   1A). We kept these mice for 6 to 8-week to further monitor behavioral sequelae. We applied the Morris water maze test to assess hippocampal learning and memory. On the learning tests, training to find a platform 2 cm beneath the surface of the water was carried for 4 days. Healthy control animals and survivors from ZIKV infection responded similarly to learning, independently on whether infected animals were treated or not (Fig 6A). On day 5, the platform was removed and memory was evaluated, by measuring the timing to stay in the platform's quadrant (latency).
Untreated ZIKV-infected mice did not stay on the quadrant where the platform had been previously located, in comparison to control (uninfected) and sofosbuvir-treated ZIKV-infected mice ( Fig   6B).
Subsequently, an amygdala-dependent aversive memory test was performed (freezing test).
This test consists of two foot-shocks on mice. On the next day, mice are exposed to the same environment, without shock, when latency is measured. Our data showed that untreated ZIKVinfected animal lost the aversive memory, whereas the sofosbuvir-treated ZIKV-infected mice and healthy controls behaved normally (Fig 6C). Our results indicate that besides the increase in survival lead by sofosbuvir, this drug also prevent from ZIKV-induced behavior sequelae, neuromotor impairments and memory loss.

In the last years, the risk perception on ZIKV infection increased substantially. Although
Zika fever is a mild and self-limited disease to most cases 24 , ZIKV-associated morbidities have been described 2,3 . Since 2013 ZIKV spread explosively across immunologically naïve populations throughout the world, and especially in the Americas 25 . For instance, in Brazil during 2015, it is estimated that over 4 million people were affected by this virus 26 . Major concerns were raised due to the association of ZIKV infection with neurological disorders during fetal development and adulthood 2,3 . More recently, a few Zika-associated deaths have also been reported 4,5 . We and others have shown that sofosbuvir, a clinically approved drug against HCV, show strong antiviral activity against ZIKV 12-14 . Particularly, we showed that sofosbuvir is functionally active against ZIKV in cells derived from peripheral organs and the CNS, by targeting the viral RNA polymerase 13 . To advance on the pre-clinical development of sofosbuvir as an anti-ZIKV drug, we further examined whether this uridine analog is active in vivo.
We understand that pharmacological studies on animal models, such as Swiss outbred mice, with representative genetic heterogeneity, may allow further exploration of the generated data to a broader population 16 . For this reason, we have chosen to use these WT and outbred animal model for these analysis. Immunocompetent mice is a more natural model and may have limited susceptibility to ZIKV infection 15 , allowing some mice to survive even after an aggressive virus challenge. Furthermore, immunocompetent neonatal mice represent an interesting model, because key processes of brain development in rodent occur postnatally (while in humans they occur during the third trimester of fetal development) 27 . Besides, with the subset of animals that survive, sequelae associated with ZIKV infection may be further examined 15 . Taking these information into account, we infected 3-day old Swiss mice with ZIKV (2 x 10 7 PFU). Indeed, this is a very high viral dose challenge, to the best of our knowledge the highest titers described to infected WT neonatal mice 22 . Treatments were carried out with sofosbuvir at 20 mg/kg/day, a dose administrated to mice during the preclinical development of this substance. The studies using this dosage supported, later on, the clinical dossier to approve the safe and effective use of sofosbuvir in humans at 400 mg/day to treat HCV infection 23 .
The virus challenge used was expected to cause significant mortality in untreated mice (indeed, less than 5 % of all the animals assayed survived). Therefore, we initially hypothesized that the expected outcomes with sofosbuvir would be an increase in the time of survival and/or reduction in mortality of infected animals. In fact, both expectations were achieved. Our results show that sofosbuvir treatment was associated with reduced mortality in infected mice, especially compared with animals that received no intervention. We identified an associated likelihood of lower mortality when comparing pre-treatment with late initiation of treatment (2 days after infection). This narrow and early time frame for antiviral intervention is common to other acute viral infections 28 . Mortality to pandemic influenza for example is reduced if neuraminidase inhibitors are administered early in the time course of infection, preferentially within 2.5 days after infection 28 . Sofosbuvir-treated animals not only had reduced mortality risk, but they also lived longer, even with our high dose viral challenge. Moreover, sofosbuvir prevented ZIKV mortality, suggesting its potential use for pre-exposure prophylaxis. This could be a desirable feature to treat pregnant women under in high risk areas of ZIKV infection.
ZIKV detection in kidney, urine, plasma, spleen before reaching out the brain 29-31 has been associated as was a hallmark of the natural history of ZIKV infection. Sofosbuvir effects on survival were associated with an inhibition of acute virus infection and spread through different anatomical compartments. Since sofosbuvir reduced acute virus loads, sofosbuvir-treated ZIKVinfected animals displayed less foci of brain microhemorrhage; when compared to untreated counterparts. Moreover, sofosbuvir-treated mice responded properly to neuromotor reflex (righting); whereas untreated ZIKV-infected animals had severe impairment of this parameter.
On the long-term analysis of animals that survived, we noticed that ZIKV-infected animals had behavioral sequelae compatible with memory impairment. This is consistent with virusinduced cell death and cerebral inflammation in memory-forming areas 14,19,22,32,33 . Sofosbuvirtreated ZIKV-infected mice survived longer and in more numbers, than untreated animals. The surviving animals were also healthier-responding to memory testing behavior consistently with uninfected control mice.        3-days old Swiss mice were infected with ZIKV (2 x10 7 PFU) and treated with sofosbuvir (SF) or not (nil) starting at day 1 before the infection. Animals that survived were kept and tested for behavior sequela in learning and memory after 60 days . Time to find the platform, according to MWM test, was performed (A). Trial in the absence of the platform was conducted (B) (*p<0.05, Student T test). In the panel B, latency represents the time spent exploring the quadrant where the platform was located before removal, and data are expressed as the mean ± SEM (n= 5-10).
Aversive memory was evaluated by freezing behavior 24 h post-training session, whereas mice were allowed to explore the aversive environment during 180 followed by two foot-shock (0.6 mA, 3 s) (C). In panel C, latency represents the time spent without movement (freezing) during 180 s, and data are expressed as the mean ± SEM (n= 2-5).