**Activity of Selected Nucleoside Analogue ProTides against Zika Virus in Human Neural Stem Cells**

**Jean A. Bernatchez 1,2,**†**, Michael Coste 3,**†**, Sungjun Beck 1,**†**, Grace A. Wells 3, Lucas A. Luna 3, Alex E. Clark 1,4, Zhe Zhu 5,6, David Hecht 3,7, Jeremy N. Rich 5,6, Christal D. Sohl 3,\*, Byron W. Purse 3,8,\* and Jair L. Siqueira-Neto 1,2,\***

	- Tel.: +1-619-594-2053 (C.D.S.); +1-619-594-6215 (B.W.P.); +1-858-822-5595 (J.L.S.-N.)

Received: 27 March 2019; Accepted: 18 April 2019; Published: 20 April 2019

**Abstract:** Zika virus (ZIKV), an emerging flavivirus that causes neurodevelopmental impairment to fetuses and has been linked to Guillain-Barré syndrome continues to threaten global health due to the absence of targeted prophylaxis or treatment. Nucleoside analogues are good examples of efficient anti-viral inhibitors, and prodrug strategies using phosphate masking groups (ProTides) have been employed to improve the bioavailability of ribonucleoside analogues. Here, we synthesized and tested a small library of 13 ProTides against ZIKV in human neural stem cells. Strong activity was observed for 2- -*C*-methyluridine and 2- -*C*-ethynyluridine ProTides with an aryloxyl phosphoramidate masking group. Substitution of a 2-(methylthio) ethyl phosphoramidate for the aryloxyl phosphoramidate ProTide group of 2- -*C*-methyluridine completely abolished antiviral activity of the compound. The aryloxyl phosphoramidate ProTide of 2- -*C*-methyluridine outperformed the hepatitis C virus (HCV) drug sofosbuvir in suppression of viral titers and protection from cytopathic effect, while the former compound's triphosphate active metabolite was better incorporated by purified ZIKV NS5 polymerase over time. These findings suggest both a nucleobase and ProTide group bias for the anti-ZIKV activity of nucleoside analogue ProTides in a disease-relevant cell model.

**Keywords:** Zika virus; nucleoside analogues; antiviral agents; NS5; prodrugs; ProTides; neural stem cells; RNA-dependent RNA polymerase

#### **1. Introduction**

The explosive spread of Zika virus (ZIKV) during the 2015–2016 epidemics in Latin America attracted worldwide attention to this previously neglected disease. The lack of effective vaccines or small molecules to prevent or treat this infection remains a cause for concern and emphasizes the urgent need for new therapeutic options [1]. ZIKV, an emerging flavivirus infection, causes several serious neurodevelopmental anomalies in the fetus, including microcephaly, congenital ZIKV syndrome (CZS), and even fetal demise [2]. While most cases of ZIKV infection are asymptomatic, reports of rash, conjunctivitis, pain in the muscles or joints, and fever have been recorded in clinical manifestations of the disease, and in rare cases, ZIKV infection has been linked to the neuroinflammatory Guillain-Barré syndrome [2].

Viral polymerases remain attractive drug targets for the development of selective antiviral therapies [3–5]. Generally, clinically-approved inhibitors that target these proteins fall into two broad classes [6,7]. The first class consists of nucleoside analogues that mimic the natural substrate of the enzyme. Upon analogue incorporation by the virally-encoded polymerase, DNA or RNA synthesis is abrogated by preventing further nucleotide incorporation (chain termination), thereby arresting viral replication. The second class is known as non-nucleoside inhibitors, which bind allosterically and arrest viral nucleic acid synthesis by distorting the polymerase active site geometry to interfere with nucleotide binding or nucleotide incorporation.

A common prodrug strategy used for antiviral ribonucleoside analogues involves the chemical synthesis of nucleoside analogue monophosphates with metabolically-removable masking groups [8–10]. These masking groups neutralize the negative charge of the phosphate on the nucleoside and allow for better membrane penetrance of these compounds. Chemical addition of the first phosphate group is crucial for improving intracellular levels of the active triphosphate metabolite form of ribonucleoside analogues. Metabolic addition of the first phosphate group is relatively slow for this class of compounds and represents a rate-limiting step for the triphosphorylation of ribonucleoside analogues to active drug molecules.

This "ProTide" strategy has been successfully deployed in the synthesis of bioactive ribonucleoside analogues [11,12], including the Food and Drug Administration-(FDA) approved hepatitis C virus (HCV) drug, sofosbuvir [13]. Sofosbuvir has recently been shown by multiple groups to also be active against ZIKV in vitro and in vivo [14–17], demonstrating that ribonucleoside analogue ProTides are an attractive avenue for the development of novel, selective antivirals against ZIKV.

Numerous nucleoside analogues have recently been explored as antiviral agents against ZIKV [18–23]. However, many of these compounds were tested in either cell-free systems or using cells that are not clinically relevant for the disease (such as Vero cells). Further, sofosbuvir is the only FDA approved inhibitor tested against ZIKV thus far in ProTide form. In addition, recent work by our group and others has suggested that sofosbuvir and ProTides in general have differential activity depending on the cell line used, which may be linked to the cell-specific metabolism of ProTides [17,24]. In this work, we chemically synthesized a library of 13 ProTides and tested them for activity against ZIKV in human neural stem cells, a disease-specific cell model of infection. Interestingly, the activity of 2- -*C*-modified aryloxyl phosphoramidate ProTides was strongly biased towards uridylate derivatives, with only modest activity observed for the 2- -*C*-modified adenylate and cytidylate aryloxyl phosphoramidate ProTides we tested. Changing the aryloxyl phosphoramidate masking group to a 2-(methylthio)ethyl phosphoramidate abolished the activity of the 2- -*C*-methyluridine ProTide, a representative hit from our library. In a head-to-head comparison between the 2- -*C*-methyluridine aryloxyl phosphoramidate ProTide and sofosbuvir, better suppression of viral titers was observed for the former compound. In addition, better protection from ZIKV-induced cytopathic effect was observed for the 2- -*C*-methyluridine aryloxyl phosphoramidate ProTide over sofosbuvir in a highly ZIKV-susceptible stem cell model of infection with strong phenotypic changes during ZIKV infection. These results represent the first broad study of nucleoside analogue ProTides against ZIKV.
