Antiviral Research 111 (2014) 82–92

Contents lists available at ScienceDirect

Antiviral Research

journal homepage: www.elsevier .com/locate /ant iv i ra l

MEK/ERK activation plays a decisive role in yellow fever virusreplication: Implication as an antiviral therapeutic target

http://dx.doi.org/10.1016/j.antiviral.2014.09.0040166-3542/� 2014 Elsevier B.V. All rights reserved.

⇑ Corresponding author at: Signal Transduction Group/Flaviviruses – VirusesLaboratory, Department of Microbiology, Institute of Biological Sciences, Univer-sidade Federal de Minas Gerais, Av. Antônio Carlos, 6627 – Campus Pampulha,31270-901 Belo Horizonte, MG, Brazil. Tel.: +55 31 3409 2752; fax: +55 31 34092733.

E-mail address: claudio.bonjardim@pq.cnpq.br (C.A. Bonjardim).1 These authors contributed equally to this work.

Jonas D. Albarnaz a,b,1, Leonardo C. De Oliveira a,b,1, Alice A. Torres a,b,1, Rafael M. Palhares a,b,Marisa C. Casteluber a,b, Claudiney M. Rodrigues a,b, Pablo L. Cardozo a,b, Aryádina M.R. De Souza a,b,Carolina C. Pacca c, Paulo C.P. Ferreira b, Erna G. Kroon b, Maurício L. Nogueira c, Cláudio A. Bonjardim a,b,⇑a Grupo de Transdução de Sinal/Flavivirus, Departamento de Microbiologia, Instituto de Ciências Biológicas, Universidade Federal de Minas Gerais (UFMG), Minas Gerais, Brazilb Laboratório de Vírus, Departamento de Microbiologia, Instituto de Ciências Biológicas, Universidade Federal de Minas Gerais (UFMG), Minas Gerais, Brazilc Faculdade de Medicina de São José do Rio Preto, São Paulo, Brazil

a r t i c l e i n f o

Article history:Received 26 May 2014Revised 4 September 2014Accepted 9 September 2014Available online 19 September 2014

Keywords:Yellow fever virusFlavivirusU0126MEK1/2

a b s t r a c t

Exploiting the inhibition of host signaling pathways aiming for discovery of potential antiflaviviralcompounds is clearly a beneficial strategy for the control of life-threatening diseases caused by flavivi-ruses. Here we describe the antiviral activity of the MEK1/2 inhibitor U0126 against Yellow fever virus17D vaccine strain (YFV-17D). Infection of VERO cells with YFV-17D stimulates ERK1/2 phosphorylationearly during infection. Pharmacological inhibition of MEK1/2 through U0126 treatment of VERO cellsblockades not only the YFV-stimulated ERK1/2 phosphorylation, but also inhibits YFV replication by�99%. U0126 was also effective against dengue virus (DENV-2 and -3) and Saint-Louis encephalitis virus(SLEV). Levels of NS4AB, as detected by immunofluorescence, are diminished upon treatment with theinhibitor, as well as the characteristic endoplasmic reticulum membrane invagination stimulated duringthe infection. Though not protective, treatment of YFV-infected, adult BALB/c mice with U0126 resultedin significant reduction of virus titers in brains. Collectively, our data suggest the potential targeting ofthe MEK1/2 kinase as a therapeutic tool against diseases caused by flaviviruses such as yellow fever,adverse events associated with yellow fever vaccination and dengue.

� 2014 Elsevier B.V. All rights reserved.

1. Introduction the genome, while the remaining of the genome encodes the

Yellow fever virus (YFV), the causative agent of Yellow Fever(YF), is the prototypic member of the genus Flavivirus, family Flavi-viridae. Two genera in this family contain major human pathogens,including dengue (DENV), Saint-Louis encephalitis (SLEV), Japaneseencephalitis (JEV) and West Nile (WNV) viruses in the Flavivirusgenus and hepatitis C virus (HCV) in the Hepacivirus genus(Lindenbach et al., 2013). The family comprises a group of posi-tive-sense, single-stranded RNA viruses with a genome approxi-mately 11 kb in length. The YFV genome encodes a polyproteinfrom a single open-reading frame that is cleaved by host and virusproteases into 10 polypeptides (Gardner and Ryman, 2010). Thestructural proteins C, prM/M and E are encoded by the 50 end of

non-structural (NS) proteins NS1, NS2A, NS2B, NS3, NS4A, NS4B,and NS5 which are associated with virus-host interactions, viralassembly and replication (Fernandez-Garcia et al., 2009;Lindenbach et al., 2013; Morais et al., 2013; Rice et al., 1985).The urban YFV transmission is initiated through the bite of infectedAedes aegypti mosquito to human. It is estimated that 900 millionpeople are at risk of infection at endemic zones, which includetropical regions of Africa, Central and South America (Briandet al., 2009; Gardner and Ryman, 2010). YF is a viral hemorrhagicfever (VHF) that shares clinical features with other VHFs such asDengue hemorrhagic fever (DHF), Lassa fever, and Crimean-Congohemorrhagic fever, but is distinguished from the others VHFs bythe characteristic severe liver damage followed by jaundice. Itslethality ranges from 20% to 50%. In humans, after a 3- to 6-dayincubation period, it is followed by the onset of symptoms whichvary from non-specific, such as fever and headache in mild cases,to high fever and hemorrhagic manifestations with the character-istic ‘‘black vomit’’ (hematemesis, gastrointestinal hemorrhage)in the more severe form of the disease (Gardner and Ryman,2010; Lindenbach et al., 2013).

J.D. Albarnaz et al. / Antiviral Research 111 (2014) 82–92 83

Although a live-attenuated vaccine against YFV (e.g. 17D strain)is available, severe adverse events associated with vaccinationhave been recognized leading to complications such as neurotropicand viscerotropic diseases (Bonaldo et al., 2014; Martin et al.,2001; Vasconcelos et al., 2001). In addition, although potentialantiflavivirus compounds have been tested in vitro (Furuta et al.,2013; Mastrangelo et al., 2012; Stahla-Beek et al., 2012), there isno therapeutics currently available to treat clinically apparent YF(Bray, 2008; De Clercq, 2013). Targeting of host cell signaling hasbeen proven useful to prospect antiviral agents, based on the factthat, in order to ensure progeny generation and spread, virusesmanipulate diverse critical host signaling pathways (Andradeet al., 2004; Pastorino et al., 2010) associated with survival/prolif-eration (Soares et al., 2009) and cytoskeleton network and cell traf-ficking (Chu and Ng, 2004; Pereira et al., 2012).

Activation of the MEK1/2 [mitogen-activated protein kinase(MAPK) extracellular signal-regulated kinase (ERK) kinase]/ERK1/2 signaling pathway is involved in diverse cellular responses suchas motility, proliferation, differentiation and survival (Raman et al.,2007). A number of RNA and DNA viruses relies on MEK/ERK sig-nals in order to support a productive infection, including some fla-viviruses, such as including hepatitis C virus (Menzel et al., 2012;Zhao et al., 2006, 2013), West Nile fever virus (Scherbik andBrinton, 2010), Japanese encephalitis virus (Gupta et al., 2011;Yang et al., 2010) and Dengue virus (Smith et al., 2014), makingthis signaling pathway an attractive antiflaviviral therapeutic tar-get. In this study we provide evidence both in vitro and in vivo thatthe pharmacological inhibitor of MEK1/2 – U0126, presented anantiviral efficacy against other flavivirus, YFV.

2. Materials and methods

2.1. Cell culture, antibodies and chemicals

VERO E6 and BHK-21 cells were cultured in autoclaved Eagle’sminimum essential medium (MEM, Auto Pow, Gibco, USA) supple-mented with 5% (v/v) heat-inactivated fetal bovine serum (FBS) (Cul-tilab – Campinas, SP, Brazil), 200 mM Glutamin (Invitrogen – SãoPaulo, Brazil), and antibiotics (40 lg/ml Gentamicin, 200 U/ml Pen-icillin and 1.5 lg/ml Fungizone) and incubated in 5% CO2 at 37 �C. C6/36 cells were cultured in Leibowitz (L-15) medium (Gibco, USA) sup-plemented with 5% heat-inactivated FBS plus the above antibioticsand incubated at 28 �C. Rabbit polyclonal anti-phospho-ERK1/2(Thr202/Tyr204) and anti-PDI antibodies were purchased from CellSignaling Technology (Beverly, MA, USA). Mouse monoclonal anti-b-actin antibody and DAPI (4,6-diamino-2-phenylindole) were pur-chased from Sigma (São Paulo, Brazil). AlexaFluor 488-conjugatedanti-rabbit IgG and rhodamine-conjugated phalloidin were pur-chased from Molecular Probes. The non-competitive inhibitor ofMEK1/2 U0126 (1,4-diamino-2,3-dicyano-1,4-bis[2-aminophenyl-thio] butadiene), and PI3K inhibitor – LY294002 (2-morpholino-8-phenyl-4H-1-benzopyran-4-one) were purchased from Cell SignalingTechnology (Beverly, MA, USA). JNK inhibitor VIII – JNKi (N-(4-Amino-5-cyano-6-ethoxypyridin-2-yl)-2-(2,5-dimethoxyphenyl)acetamide), Akt inhibitor – Akt X (10-(40-(N-diethylamino)-butyl)-2-chlorophenoxazine), and p38 MAPK inhibitor – SB203580 (4-(4-fluorophenyl)-2-(4-methylsulfinylphenyl)-5-(4-pyridyl)-imidazole)were purchased from Calbiochem – Merck (Darmstadt, Germany).Rabbit anti-YFV NS4AB antibody was a generous gift from Dr CharlesRice, Rockfeller University, USA.

2.2. Viruses and viral infection

YFV strain 17D (vaccine strain) was propagated in VERO cells.DENV-2 genotype III (PI-59/2006 – Piauí-59) (Figueiredo et al.,

2008) and DENV-3 genotype I (MG-20/2004) (Ferreira et al.,2010) were propagated in C6/36 cells. SLEV (BeH 355964) (DeMorais Bronzoni et al., 2005) was propagated in BHK-21 cells.Briefly, VERO, C6/36 or BHK-21 monolayers at 80–90% confluencewere infected at a multiplicity of infection (MOI) of 0.01 in thepresence of 1% FBS. After 72 h post-infection (hpi) supernatantswere harvested and clarified by centrifugation (900 g, 4 �C,15 min). Aliquots were stored at �80 �C. The same procedure wascarried out with the supernatants of non-infected cells, which wereused as mock controls in the subsequent experiments. Procedurefor UV inactivation of virus stocks was described elsewhere (DeMagalhães et al., 2001). Viral infections were carried out withserum-starved cells. Twelve hours prior to infection, cells grownto 80–90% confluence were starved with 1% FBS MEM. Cells werethen infected at the indicated MOI for the times shown. Cells weretreated with the indicated pharmacological inhibitor for 30 minprior to viral infection and then incubated in its continued pres-ence for the indicated times.

2.3. Virus infectivity assays

VERO or BHK-21 cells were cultured and starved as above at adensity of 5 � 105 cells per well, in a 6-well culture dish and theninfected. Infection of VERO cells (YFV) was carried out at an MOI of1.0 for 6, 12, 24 and 48 h, and infection of BHK-21 cells (DENV andSLEV) was carried out at an MOI of 1.0 for 48 h, either in theabsence (DMSO 0.06%) or in the presence of U0126 (15 lM), Akt-X (15 lM), LY294002 (20 lM), JNK inhibitor VIII (4 lM) orSB203580 (20 lM). At the indicated times, cultures (cells plusmedium) were frozen and thawed once. Virus was collected fromthe supernatants after centrifugation and immediately assayedfor infectivity as described (Figueiredo et al., 2008). Each experi-ment was run in triplicate and the results are reported as the aver-age values. The data were confirmed by three independentexperiments with very similar results.

2.4. Cytotoxicity assays

Confluent 6-well dishes of VERO cells were treated withincreasing concentrations of U0126 (5, 10, 20 and 40 lM), of JNKiVIII (0.4, 4 and 40 lM), of SB203580 (5, 10, 20 and 40 lM), ofLY294002 (5, 10, 20 and 40 lM), and of Akt X (3.75, 7.5, 15 and20 lM). BHK-21 cells were treated with 15 lM of U0126. At 48 h,an equal volume of Trypan Blue stain was added to the trypsin-detached cells. VERO and BHK-21 cells were stained for 10 minat room temperature and after that cells were observed for any evi-dence of stain absorption (an indication of cellular membrane per-meability and death). We found that P90% of the cells pretreatedwith U0126 at 15 lM; JNKi at 4 lM; SB203580 at 20 lM andLY294002 at 20 lM, were not stained. These concentrations werethen used throughout the experiments.

2.5. Electron microscopy

VERO cells were grown in 150-cm2 culture flasks to 80–90%confluence and FBS-starved as described. Cells were treated withDMSO (0.06%) or U0126 (15 lM) as described above and theninfected with YFV-17D at MOI of 1.0. Uninfected cells were treatedin the same way and used as controls. After 1, 24 and 48 hpi, cellswere washed twice with FBS-free MEM, fixed with 2.5% glutaralde-hyde grade I (Electron Microscopy Sciences, Germany) for 1 h atroom temperature (RT) and processed as described previously(Pereira et al., 2012). Cells were examined using a Tecnai G2-SpiritFEI 2006 transmission electron microscope operating at 80 kV atthe Microscopy Center, UFMG, Brazil. A quantitative analysis ofYFV morphogenesis intermediates was carried out. Three struc-

84 J.D. Albarnaz et al. / Antiviral Research 111 (2014) 82–92

tures were counted: (i) empty vesicles, vesicles encompassed inER-derived vesicle packets induced in the early stages of flavivirusmorphogenesis; (ii) replication complex (RC)-containing vesicles,vesicles encompassed in vesicle packets with the RC assembledin their interior; and (iii) virions, contained in ER/golgi-derivedvesicles. Percents of each structure in the total amount of YFV mor-phogenesis intermediates were calculated for each treatment andtime of infection.

2.6. RNA extraction, cDNA synthesis and quantitative PCR

Total RNA extraction was performed by using the TRIzol reagent(Invitrogen) according to the manufacturer’s procedures. Comple-mentary DNAs from (�) and (+) strands were reverse transcribedwith mFU1 and CFD2 Flavivirus-specific primers (Chao et al.,2007), respectively, or without primers as a self-priming control.Five hundred nanograms of primers and 2 lg of RNA were incu-bated at 70 �C for 5 min and then placed on ice. ComplementaryDNA was synthesized with M-MLV Reverse Transcriptase (Pro-mega) according to the manufacturer’s instructions. QuantitativePCR (qPCR) was performed with the mFU1 and CFD2 primers. Eachreaction was performed by using the SYBR Green PCR Master Mix(Applied Biosystems), 10 lM of forward and reverse primers and1 lL of 1:3 diluted cDNA, in a Step ONE Real Time PCR System(Applied Biosystems), following manufacturer’s recommendations.The Relative Standard Curve method was used to analyze the data.

2.7. Immunofluorescence microscopy

8 � 104 VERO cells were grown on coverslips in 24-well dishes,FBS-starved for 12 h, and infected with YFV at an MOI of 1.0 in thepresence of DMSO (0.06%) or U0126 (15 lM) for 24 h. After, cellswere rinsed with cold PBS and fixed with 4% paraformaldehyde(PFA) in PBS for 10 min at room temperature (RT). Cells werestained with specific primary antibodies: rabbit anti-YFV NS4A/B(1:1000) or rabbit anti-PDI (1:50), as described previously(Pereira et al., 2012). AlexaFluor 488-conjugated anti-rabbit IgG(1:400) was used as secondary antibody. Fluorescently labeledcells were visualized using a Zeiss LSM 510 META confocal laserscanning microscope at the Center for the Acquisition and Process-ing of Images (CAPI/UFMG). Images were processed with ImageJ(NIH).

2.8. Western blotting

VERO cells were grown in 25-cm2 flasks, FBS-starved and leftuninfected or infected with YFV-17D at an MOI of 1.0 for the indi-cated times, in the absence or presence of U0126. Cells were thenwashed twice with cold PBS. Whole cell protein extracts andimmunoblotting were prepared as described (Pereira et al.,2012). Thirty-five micrograms of the cell lysate per sample wereseparated by electrophoresis on a 10% SDS–polyacrylamide geland then transferred onto nitrocellulose membranes. Immunoblot-ting was performed with anti-phospho-ERK1/2 (1:1000) or anti-b-actin (1:3000), and horseradish peroxidase-conjugated secondaryanti-rabbit IgG (1:3000) or anti-mouse IgG (1:3000). Immunoreac-tive bands were visualized by using ECL Plus detection system asdescribed in the manufacturer’s instructions (GE Healthcare, UK).

2.9. In vivo infection

This study was carried out in accordance with the recommen-dations in the Guide for the Care and Use of Laboratory Animals,NIH, USA. The protocol was approved by the Institutional AnimalCare and Use Ethics Committee at UFMG (Assurance Number:80/2011). All inoculation and experimental manipulation were

performed under anesthesia that was induced and maintainedwith ketamine hydrochloride and xylazine. Male BALB/c mice wereobtained from UFMG’s animal facility and kept in an ABSL2 at theDepartment of Microbiology, UFMG. At age of 8 week-old, micewere intravenously injected with 30 lg of U0126 or vehicle (2.5%Dimethyl-Sulfoxide, 10% Ethanol, 5% Cremophor EL and 10% PEG-400 in saline) through the tail vein. One day after starting theU0126 treatment, inoculum containing 103 PFU of YFV-17D(10 lL) was administered via intracranial injection to mice.Mock-infected animals were injected with the same volume ofthe supernatant of mock-infected VERO cells. Daily, animalsreceived intravenous single dose-injection of U0126 (30 lg/ani-mal/day) or vehicle and were monitored for body weight lossand disease appearance. Once the animals reached 25% of bodyweight loss or presented lower limbs paralysis they received alethal dose of anesthetics (300 mg/kg Ketamine, 30 mg/kg Xyla-zine), followed by cervical dislocation. Percent mortality andweight loss were calculated, and brains, spleens and livers werecollected for further analysis. Brain halves, spleen and liver of eachanimal were individually macerated in MEM containing 2% FBS,clarified by centrifugation (900 g, 5 min, 4 �C) and the supernatantswere inoculated onto VERO cell monolayers for virus titration, asdescribed above. For in vivo evaluation of ERK1/2 phosphorylation,mice brain halves were individually macerated in cell lysis buffer,containing proteases and phosphatases inhibitors, and the totalprotein extract was prepared and quantified as described. Poolsof whole brain protein extracts of 5 animals were prepared for eachgroup (5 lg/animal) for Western blot analysis.

2.10. Densitometric analysis

Phosphorylated-ERK1/2 was quantified using densitometricanalysis tool from ImageJ software (NIH) and the levels were nor-malized to the levels of b-actin in the same sample. The changes inprotein phosphorylation with respect to control values were esti-mated. The results were expressed as the phospho-ERK to b-actinratio measured in arbitrary units.

2.11. Statistics

Data gathered in animal studies were analyzed by Mantel–Coxsurvival analysis and Two-way Anova with Bonferroni post-test forweight loss curves. Virus yield means in brains of drug treated ver-sus non-treated animals were compared through Student’s t test.

3. Results

3.1. MEK/ERK pathway inhibition affects YFV replication

We initially tested specific cell signaling pathways inhibitors,such as MEK/ERK (U0126 – 15 lM), Akt (inhibitor-X – 15 lM),PI3K (LY294002 – 20 lM), JNK (inhibitor VIII – 4 lM) and p38MAPK (SB203580 – 20 lM). VERO cells were treated with the indi-cated concentration of inhibitor for 30 min prior to infection withYFV at an MOI of 1.0 for 48 h. After that, YFV-containing superna-tant was collected and assayed for infectivity. As shown in Fig. 1A,U0126 reduced virus replication more effectively. While U0126caused about 2-log10 units inhibition of YFV replication, the otherpharmacological inhibitors tested did not inhibit YFV replicationwith similar effectiveness. Further, we analyzed whether viralgrowth inhibition was dose-dependent. The infections were car-ried out in concentrations of U0126 ranging from 5 to 40 lM andthen the viral titers were assayed, showing that viral growth inhi-bition is dose-dependent (Fig. 1B). Thus, we conclude that the dis-turbance in MEK/ERK signaling is deleterious for YFV replication. In

Fig. 1. MEK/ERK plays a specific role in YFV replication. Cells were either left untreated, or treated for 30 min with the indicated pharmacological inhibitor prior to virusinfection and then infected with YFV, DENV or SLEV at an MOI of 1.0 either in the absence or in the continued presence of the inhibitor for 48 h. Viruses were then collectedand their infectivity determined. (A) Effect of diverse pharmacological inhibitors on YFV replication. VERO cells were incubated with the specific pharmacological inhibitors:MEK/ERK (U0126 – 15 lM), Akt (Akti X – 15 lM), PI3K (LY294002 – 20 lM), JNK1/2 (JNKi VIII – 4 lM) and p38 MAPK (SB203580 – 20 lM), for 48 h. (B) Effect of U0126 on YFVreplication is dose-dependent. VERO cells were infected with YFV-17D at an MOI of 1.0 for 48 h in the absence or in the continued presence of the MEK/ERK inhibitor (U0126)at 5, 10, 20 and 40 lM, for 48 h. Viral yield (PFU/mL) is indicated by black bars. Cell viability in the presence of U0126 for 48 h, as assessed by trypan blue exclusion, isexpressed as the percent (%) of viable cells and is indicated by the black dots. (C) Effect of U0126 on DENV-2/-3 and SLEV replication. BHK21 cells were infected with DENV-2,DENV-3 or SLEV at an MOI of 1.0 for 48 h either in the absence or in the continued presence of U0126 – 15 lM. (D) YFV stimulates ERK1/2 phosphorylation. VERO cells wereeither mock-infected or infected with YFV at an MOI of 1.0 for the indicated times. Infected-cell lysates were subjected to western blot and probed with anti-phospho-ERK1/2(Thr202/Tyr204-P) antibody (upper panel) or with anti-b actin antibody (lower panel), used as a loading control. Cells were infected either in the absence (lanes 1–3, 6–8, 10–12, 14–16, 18–20) or in the presence of U0126 (15 lM) (lanes 4, 5, 9, 13, 17 and 21). Cells were also exposed to UV-inactivated YFV (lanes 3, 5, 8, 12, 16 and 20). The levels ofphosphorylated ERK1/2 were quantified by densitometric analysis and the phospho-ERK/b-actin ratio (arbitrary units) is shown. The molecular masses (kDa) are indicated onthe left. (E) The MEK/ERK signaling pathway is required for viral replication. VERO cells were infected with YFV at an MOI of 1.0 for 6, 12, 24 and 48 h, either in the absence orin the presence of U0126 (15 lM). Virus was collected and their infectivity assayed. Similar results were obtained in three independent experiments.

J.D. Albarnaz et al. / Antiviral Research 111 (2014) 82–92 85

Fig. 2. The MEK/ERK signaling pathway is required for YFV replication. (A) Cells were infected with YFV at an MOI 1.0 for 48 h and treated with the U0126 (15 lM) accordingto the diagram depicted in the right panel. Heavy and dotted lines: in the presence or in the absence of U0126, respectively. Virus was collected and its infectivity assayed. Thedata are representative of three independent experiments with similar results. (B) Removal of U0126 releases YFV-stimulated MEK/ERK pathway. Western blot analysiscarried out with whole cell extracts of mock-infected (lanes 1–3) or infected with YFV at an MOI of 1.0 for the indicated times (lanes 4–6), either in the absence (lanes 1, 4) orin the presence of U0126 (lanes 2, 3, 5,6). Cells were washed after 2.5 h of U0126 treatment (lanes 3 and 6) and the infection was allowed to continue for additional 30 min.

86 J.D. Albarnaz et al. / Antiviral Research 111 (2014) 82–92

order to investigate whether the antiviral activity of U0126 wasrestricted to YFV, its effect on the replication of DENV and SLEVwas tested further. To that end BHK-21 cells were treated withU0126 at 15 lM for 30 min. prior to infection with DENV-2,DENV-3 and SLEV at an MOI of 1.0 for 48 h. After that, DENV- orSLEV-containing supernatants were collected and assayed forinfectivity. As shown in Fig. 1C, U0126 presented a broad antifla-viviral effect, affecting not only YFV, but other members of the Fla-vivirus genus such as DENV and SLEV.

Next, we examined MEK/ERK stimulation upon YFV infectionand whether it might depend on YFV replication. To approach that,VERO cells were exposed to the UV-inactivated YFV, or infectedwith YFV at an MOI of 1.0 either in the presence or absence ofU0126 for the indicated times. Fig. 1D shows that ERK1/2 phosphor-ylation upon YFV infection is an early event starting 15 min post-infection (lane 2), prolonged up to 6 hpi (lane 15) and declined afterthat time (lane 19). Our data also demonstrates that replicationcompetent virus is not an essential requirement for ERK1/2 activa-tion, since the levels of ERK1/2 phosphorylation (lanes 3, 8, 12, and16) with UV-inactivated YFV were about the same as those verifiedwith the non-inactivated virus (lanes 2, 7, 11 and 15).

To get some insight to what extent U0126 would impact the YFVlife cycle, VERO cells were virus-infected at an MOI of 1.0 for 6, 12,24 and 48 h either in continued presence or absence of U0126(15 lM). After that, virus was collected and assayed for infectivity.Our data are consistent with the idea that MEK/ERK plays, in fact, adecisive role in YFV replication, since the block imposed by theinhibitor affected the whole virus infective cycle (Fig. 1E).

3.2. The MEK/ERK signaling pathway is required for YFV replication

Next, we designed an experiment in the attempt to define towhat extent MEK/ERK could affect YFV replication, depicted at

Fig. 2A. Cells were left untreated (lane 1) or incubated withU0126 (15 lM) prior to (lanes 2, 3, 5 and 7) or after (lanes 4, 6and 8) YFV infection at an MOI of 1.0 for 48 h. Virus was then col-lected and assayed for infectivity. The experiments were carriedout at the time-frame where YFV-stimulated MEK/ERK activationwas verified (Fig. 1D). As demonstrated in Fig. 2A, when the blockthat interrupts the signal transmission initiated by MEK/ERK (lanes3, 5 and 7) is released, YFV replicates to the same extent as verifiedwith the control (lane 1). Moreover, whenever U0126 is addedafter the ongoing infection, YFV titers are similar to that verifiedwhen the infection is carried out in the continued presence ofthe inhibitor (lane 2). To confirm that the signal transmission byMEK/ERK is released after removing U0126, VERO cells were eithermock-infected or infected with YFV in the presence or absence ofU0126 as shown in Fig. 2B. Treatment with the inhibitor was inter-rupted by washing the cells with fresh media and the infection wasallowed to proceed for further 30 min, as indicated (lanes 3 and 6).Cell extracts were then collected, subjected to Western blot andprobed for phosphorylated ERK1/2. Our data shows that in contrastto the observed inhibition of phospho-ERK1/2 upon infection in thepresence of U0126 (lane 5), after removing the inhibitor YFV is ableto stimulate the kinases (lane 6). Although just one time is shown,the same result was also verified with the other investigated timesup to 12 hpi (data not shown).

3.3. Inhibition of MEK/ERK affects YFV genome replication

Considering that the UV-inactivated YFV is still able to stimu-late MEK/ERK, we firstly investigated whether the signal transmis-sion was altered at the virus adsorption/entry step. To address that,cells were infected with YFV at an MOI of 1.0 either in the absenceor presence of U0126 (15 lM) for 1 h. Cells were then fixed andprepared for electron microscopy (Fig. 3A). Results obtained here

Fig. 3. Inhibition of MEK/ERK leads to decreased YFV genome replication. (A) YFV entry is not affected by MEK/ERK interruption. VERO cells were left uninfected (a) orinfected with YFV (b, c) at an MOI of 1.0 for 1 h, either in the absence (a, b) or in the presence of U0126 (15 lM) (c). Cells were then fixed and prepared for transmissionelectron microscopy. Electron micrographs are shown with their scale indicated by the bars. Details show YFV virions (arrows) entering into the cell through endocytosis(arrowheads). Abbreviations: M, mitochondria, ER, endoplasmic reticulum. G, Golgi complex. Data is representative of two independent experiments with identical results. (B)Time-course of viral RNA (+) and (�) synthesis upon YFV infection. Cells were infected with YFV at an MOI of 1.0 for 3, 6, 9, 18, 30 and 48 h (B), or for 18, 24 and 48 h either inthe absence or in the presence of U0126 (15 lM) (C). RNA was then isolated and subjected to quantitative RT-PCR with specific primers.

J.D. Albarnaz et al. / Antiviral Research 111 (2014) 82–92 87

suggest that the treatment of VERO cells with U0126 did not affectYFV entry into cells (Fig. 3A – panel b compared to panel c).

We then raised the question whether the MEK/ERK blockadewas affecting the YFV genome replication. To approach this ques-tion, VERO cells were infected with YFV at an MOI of 1.0 for thetimes shown and total RNA was isolated and subjected to quanti-tative RT-PCR with specific primers for both the (+) and (�) RNAstrands.

Fig. 3B depicts the relative quantification of both plus andminus RNA strands during the time-course of YFV infection. RNA(+) and (�) strand levels remain steady during the early stages ofinfection (3–9 hpi). After 9 hpi, there is an exponential incrementin the accumulation of both strands. RNA (+) strand accumulationis increased approximately 100 times after 18 hpi and reaches thepeak at 30 hpi. RNA (�) strand synthesis follows similar kinetics,even though its quantification is commonly 2-log10 units lowerthan RNA (+) strand over time.

Although U0126 treatment does not block, it significantlyimpairs RNA (+) strand synthesis as demonstrated in Fig. 3C. Inthe presence of U0126, levels of RNA (+) strand are reducedapproximately 51.1%, 73.8% (P < 0.01) and 50.2% (P < 0.001), at18, 24 and 48 hpi, respectively. In the same conditions, RNA (�)strand synthesis levels are diminished 23.2%, 43.56% and 40.4%.Together, these observations suggest that the blockade of MEK/ERK pathway significantly affects YFV-17D genome replication.However, the inhibition levels observed here may not be sufficient

to account for a reduction of �2-log10 units in YFV-17D viral yieldin the presence of U0126. Additional downstream targets of MEK/ERK pathway inhibited by U0126 remain to be further addressed.

3.4. MEK/ERK inhibition affects the accumulation of YFV NS4AB andendoplasmic reticulum resident protein PDI

To gain insight into the mechanism(s) underlying the decreasedviral yield upon treatment with U0126, we investigated whetherYFV NS4AB protein accumulation was affected following the treat-ment with the inhibitor. VERO cells were mock-infected or infectedwith YFV either in the presence or absence of U0126 (15 lM), at anMOI of 1.0 for 24 h and NS4AB expression was analyzed by immu-nofluorescence. As shown in Fig. 4A, detection of NS4AB proteinwas impaired upon treatment with U0126 (compare panels band c).

Flavivirus genome replication occurs at the rough endoplasmicreticulum (rER) within vesicle packets, a site where genome repli-cation complex is assembled in order to maximize virus replication(Mackenzie and Westaway, 2001; Welsch et al., 2009). We theninvestigated whether the interruption of MEK/ERK signaling wasalso associated with altered organization of rER. To approach that,VERO cells were mock-infected or infected with YFV (MOI of 1.0)for 24 h either in the absence or presence of U0126 (Fig. 4B). Cellswere fixed and stained for Protein Disulfide Isomerase (PDI), an ER-resident protein (panels a, d, g, j) and rhodamine-conjugated phal-

Fig. 4. MEK/ERK is required for detection of viral NS4AB and host PDI proteins.VERO cells were grown onto coverslips and infected with YFV at an MOI of 1.0 for24 h, or left uninfected. Where indicated, VERO cells were infected with YFV in thepresence of U0126 (15 lM) for 24 h. Cells were then fixed, permeabilized andstained with an anti-NS4AB antibody (panels a, c, e) and DAPI (panels b, d, f) (A), orstained with anti-Protein Disulfide Isomerase (PDI) antibody (panels a, d, g, j) andrhodamine-conjugated phalloidin (actin cytoskeleton – panels b, e, h, k) (B). Thepanels c, f, i, and l in B depict the merge of PDI and actin stainings. The images werevisualized by confocal microscopy.

88 J.D. Albarnaz et al. / Antiviral Research 111 (2014) 82–92

loidin (actin cytoskeleton – panels b, e, h, k). Images were visual-ized by confocal microscopy. As observed, while an accentuatedincrease in PDI accumulation is evident at the perinuclear regionand cytoplasm of YFV-infected cells (panels g–i), in contrast, PDIlevels in U0126-treated cells (panels j–l) were similar to thoseobserved in mock-infected cells (panels a–c).

3.5. Blockade of MEK/ERK pathway affects YFV morphogenesis

We then asked whether the blockade of MEK/ERK would affectYFV morphogenesis. Therefore, VERO cells were infected with YFV

as described, in either the absence or presence of U0126. Cells werethen prepared and analyzed by electron microscopy (Fig. 5A). Asobserved, YFV infection stimulates the invagination of the ER-derived membranes leading to formation of diverse vesicle packets(VP) (Fig. 5A – panels b and e). The electron-dense materialobserved within the VPs is thought to contain the Replication Com-plex (RC) and serve as sites for the viral RNA replication (panel f –arrowhead) (Gillespie et al., 2010; Welsch et al., 2009). While theseVPs are consistently found in untreated and YFV-infected cells(inset from panel e highlighted in panel f), in remarkable contrast,they are barely detected at 24 h and just scarcely verified at 48 hwhen the infection is carried out in the presence of U0126(Fig. 5A – panel h). In this circumstance empty vesicles accumulate(panel h and enlarged inset in i). A morphometric analysis was car-ried out and the data is presented in Fig. 5B.

3.6. In vivo antiviral activity

Next, we sought to evaluate the in vivo antiviral potential of theMEK1/2 inhibitor U0126. Eight-week old BALB/c mice wereinfected with 103 PFU of the YFV-17D through intracranial injec-tion. Animals were treated with 30 lg of U0126, or vehicle only,through intravenous injection in the tail vein one day prior infec-tion and daily throughout the course of infection. Intravenousadministration of U0126 in mouse models has been described else-where (Mori et al., 2002; Namura et al., 2001; Wang et al., 2004).

Despite the in vitro efficiency of U0126 in halting YFV replica-tion in VERO cells, treatment of YFV-infected BALB/c mice withU0126 was not protective against infection, given all YFV-infectedanimals died within 9 days after challenge. Also, there were no sig-nificant differences in weight loss kinetics or survival betweenYFV-infected animals, treated or not with MEK1/2 inhibitorU0126 (not shown). Nonetheless� when animals reached 25% ofbody weight loss individual tissues such as liver, spleen and brainwere collected and both viral load and ERK1/2 phosphorylationwere determined. However, we were unable to detect YFV replica-tion through plaque assay in any tissues other than brains. Inter-estingly, YFV titers in brains of U0126-treated animals after 9 dpi(6.76 � 106 ± 1.51 � 106 PFU/g, n = 9) were significantly lower thanthose observed in vehicle-only treated mice (1.47 � 107 ± -3.36 � 106 PFU/g, n = 8) (reduction of 54.3%, p 6 0.05) (Fig. 6A).Accordingly, Fig. 6B demonstrates that treatment of YFV-infectedmice with U0126 reduces the levels of ERK1/2 phosphorylationin brains by approximately 48.8% (Fig. 6B, lane 3) in comparisonwith the levels in vehicle-only treated mice (Fig. 6B, lane 2). Theseresults are in accordance with the in vitro observed effect(s) ofU0126 over YFV replication and reinforces the potential therapeu-tic employment of MEK1/2 inhibitors as anti-YFV drugs.

4. Discussion

In an attempt to find chemicals with antiflaviviral activity, wecarried out the infection of VERO cells with YFV-17D in presenceof a series of pharmacological inhibitors targeting host signalingpathways that could be important to the virus replication cycle.Specific inhibitors of MEK1/2 (U0126), JNK1/2-SAPK (JNK inhibitorVIII), p38 MAPK (SB203580), PI3K (LY294002) and Akt/PKB (Aktinhibitor X) were then tested as potential antiviral and U0126proved to convey greatest efficiency in inhibition of YFV(Fig. 1A). A significant inhibition of approximately 2-log10 unitsin YFV yield was observed in the presence of U0126, in a dose-dependent manner (Fig. 1B).

The involvement of MEK/ERK pathway has already beenreported during the infection with several members of Flaviviridaefamily, including hepatitis C virus (Menzel et al., 2012; Zhao et al.,

Fig. 5. Blockade of MEK/ERK pathway affects YFV morphogenesis. (A) Representative micrographs of the effects of U0126 on YFV-17D morphogenesis. FBS-starved VERO cellswere treated with DMSO (a–c, e, f) or U0126 (d, g–i) and infected with YFV-17D at an MOI of 1.0 (b, c, e–i), or left uninfected as controls (a, d). After 24 (a–d, g) or 48 hpi (e, f, h,i), cells were processed for transmission electron microscopy. Micrographs c, f, and i are higher magnification insets of micrographs b, e, and h, respectively. Electronmicrographs are shown with their scale indicated by the bars. Abbreviations: M, mitochondria; Nu, nucleus; VP, vesicle packets; v, vesicles containing virions; arrow indicatesa typical YFV virion; arrowheads indicate replication complexes (RC) inside vesicles; asterisks (⁄) indicate typical ER organization. (B) Graphical representation of the percentsof YFV morphogenesis intermediates, according to the accompanying panel. Alongside the bars are the numbers of YFV morphogenesis intermediates counted in each sample(n = 10 cells). Data is representative of two independent experiments with similar results.

J.D. Albarnaz et al. / Antiviral Research 111 (2014) 82–92 89

2006, 2013), West Nile fever virus (Scherbik and Brinton, 2010),Japanese encephalitis virus (Gupta et al., 2011; Yang et al., 2010)and Dengue virus (Smith et al., 2014), thus emphasizing its impor-tant role played in flaviviruses’ replication. Here we showed thatYFV stimulates the activation of MEK/ERK pathway in VERO cellsfrom as early as 15 min. post-infection up to 6 hpi (Fig. 1D). Inter-estingly, infection carried out with UV-inactivated YFV is still ableto stimulate ERK1/2 phosphorylation (Fig. 1D), but the same wasnot observed for heat-inactivated YFV (data not shown). A previousreport showed that UV-inactivated WNV induces an early transientactivation ERK1/2, although sustained activation of these kinasesrequired replication competent viruses (Scherbik and Brinton,2010). More recently, a study described that treatment of DC-SIGNexpressing cells with soluble HCV E2 protein is responsible for arapid and sustained activation of MEK/ERK pathway (Zhao et al.,2013). Over the last decade several reports showed that HCV E2protein is actively responsible for MAPK stimulation, which might

support progression of disease and pathogenesis associated withHCV (Mazzocca et al., 2005; Zhao et al., 2005, 2006, 2013).Together, these observations demonstrate that structural compo-nents of flaviviruses accomplish important roles in initial steps ofhost-virus interaction.

Considering that a significant decrease in viral titers was veri-fied only if the drug is continuously present or added to an ongoinginfection (Fig. 2A) and that U0126 inhibition of MEK1 and MEK2catalytic activity is reversible (Favata et al., 1998), that couldexplain why the removal of the inhibitor would enable YFV to re-induce ERK1/2 phosphorylation (Fig. 2C). However, the mechanis-tic details underlying U0126 reversible effects on YFV replicationrequires further investigation.

Our data consistently shows that U0126 did not affect YFV vir-ion entry into host cell (Fig. 3A). We also observed that U0126treatment during adsorption at 4 �C does not alter either YFV entryor YFV replication (data not shown). Flavivirus entry by clathrin-

Fig. 6. Reduction of virus titers in brains of YFV-infected BALB/c treated with U0126. Adult BALB/c mice were injected in the tail vein with 30 lg of U0126, or vehicle, one dayprior and daily after intracranial inoculation with 103 PFU of YFV-17D. Control groups received the same volume of VERO cells supernatant. Animals were monitored fordisease signs and loss of weight and were euthanized when they reached the threshold of 25% body weight loss or presented lower limbs paralysis. (A) YFV virus productionin individual brains was determined through plaque assay in VERO cells. The graph represents the YFV virus yield in individual brains of untreated (YFV + vehicle) and U0126treated mice. The means of virus yield were statistically analyzed with Student’s t test (p = 0.0274). (B) Western blot analysis of ERK1/2 phosphorylation in brains ofuninfected mice (lane 1), YFV-infected and untreated (YFV + vehicle – lane 2) and YFV-infected and treated with U0126 (YFV + U0126 – lane 3). Protein extracts of mice brainhalves were prepared and quantified as described. Pools of whole brain protein extracts of 5 animals were prepared for each group (5 lg/animal) for western blot analysis.The levels of phosphorylated ERK1/2 were quantified by densitometric analysis and the phospho-ERK/b-actin ratio (arbitrary units) is shown. The molecular masses (kDa) areindicated in the left.

90 J.D. Albarnaz et al. / Antiviral Research 111 (2014) 82–92

mediated endocytosis is largely supported by experimental evi-dence (Pierson and Kielian, 2012) and it has been shown that theclassical clathrin-dependent endocytic pathway seems to be resis-tant to disruption of MEK/ERK signaling (Robertson et al., 2006).Our findings are in line with these observations and demonstratethat signaling through MEK/ERK is not required for YFV entry.

We showed here that treatment of YFV-infected cells withU0126 does not blockade, but significantly impairs, the accumula-tion levels of positive RNA strands at late stages of infection(Fig. 3C). Instead, accumulation of NS4A/B protein after 24 hpi isuntraceable in YFV-infected cells treated with U0126 (Fig. 4A).Since generation of NS proteins is a prerequisite for RNA replica-tion, we hypothesize that either canonical cap-dependent or non-canonical internal ribosome entry site (IRES)-independent transla-tion mechanisms are somehow functional in U0126-treated cells.The co-expression of both NS4A and NS4B proteins, concomitantwith other NS proteins, is required for the formation of character-istic membrane structures induced by flavivirus infection(Roosendaal et al., 2006; Miller et al., 2007). These virus-inducedmembranes termed paracrystalline array, convoluted membranesand vesicle packets all have specific functions during flavivirusgenome replication (Lindenbach et al., 2013; Gillespie et al.,2010; Welsch et al., 2009). Therefore, the reduction in accumula-tion of NS4A/B protein in the presence of U0126 strongly suggeststhat inhibition of MEK1/2 somehow disrupts YFV-induced mem-brane rearrangement. Actually, VERO cells infected with YFV andtreated with U0126 presented with decreased staining for the ERmarker PDI (Fig. 4B – panels j–l) when compared to untreated,YFV-infected VERO cells (Fig. 4B – panels g–i) (Op De Beecket al., 2004). Consistently, our electron microscopy analysis corrob-orates the defects in YFV-induced ultrastructural changes of mem-branes in infected cells treated with U0126 (Fig. 5A – panel hcompared to panels b/e). Together, these experimental observa-tions demonstrate that disruption of MEK/ERK signaling affects atleast three critical YFV-induced intracellular events: NS proteinsaccumulation, membrane rearrangements and genome replication.It has been shown that both HCV and DENV requires group IVAphospholipase A2 (PLA2G4A) for assembly of highly infectiousprogeny, whose lipolysis-dependent enzymatic activity is under

MEK/ERK regulation (Menzel et al., 2012). Whether PLA2G4A isalso a downstream target of MEK/ERK upon YFV infection is aninteresting issue that remains to be further investigated.

Finally, the observed antiviral activity conveyed by U0126 treat-ment suggested the potential therapeutic use of MEK1/2 inhibitorsin the treatment of diseases caused by flaviviruses. Even thoughthe protocol used in this study was not protective, since we didnot observe differences in morbidity and survival rates betweenanimals challenged with YFV and treated or not with U0126, theobservation that YFV titers are significantly reduced in the brainsof individuals treated with U0126 is remarkable (Fig. 6). Theseresults corroborate the in vitro analysis, in which MEK/ERK disrup-tion affects YFV replication.

U0126 has been largely employed to investigate the roles ofMEK/ERK in viral infections, cancer and other pathological pro-cesses, but this drug could not be further developed for clinicaluse due to pharmacological limitations, such as low oral bioavail-ability (Planz, 2013). However, intravenous injection (IV) ofU0126 in mouse models was found to be protective against braininjuries, such as ischemia and physical trauma (Namura et al.,2001; Mori et al., 2002; Wang et al., 2004), demonstrating theavailability of U0126 in the brain. MEK1/2 inhibition was also pro-tective in mice challenged with influenza virus upon treatmentwith U0126 by inhalation (IN) (Droebner et al., 2011) or by intra-peritoneal injection (IP) (Pinto et al., 2011). All treatment routes(IV, IN or IP) were able to reduce ERK1/2 phosphorylation in animaltissues, as observed in this study (Fig. 6B), and did not cause anygross physiological alteration in treated mice (Droebner et al.,2011; Namura et al., 2001; Mori et al., 2002; Pinto et al., 2011;Wang et al., 2004). Moreover, other small molecule inhibitors ofMEK/ERK are under development for clinical use, with better phar-macological performance, and can be tested for their efficacyagainst flavivirus infections (Planz, 2013).

5. Conclusions

In conclusion, our results suggest the use of MEK1/2 as a poten-tial therapeutic target for the development of antivirals against fla-viviruses, and U0126 as a potential lead for the design of such

J.D. Albarnaz et al. / Antiviral Research 111 (2014) 82–92 91

antiflaviviral therapeutics. Our study also highlights the need tocontinue pursuing better strategies for YFV infection control aswell as alternative YFV mouse models and routes for drug deliveryin order to foster our search for antiviral agents.

Acknowledgements

This work was supported by the Minas Gerais State’s Founda-tion for Research Support (FAPEMIG) – (CBB – APQ-01670-11 toC.A.B.); the National Council for Scientific and Technological Devel-opment (CNPq) – (477187/2010-2 to C.A.B.) and PRONEX-Dengue –(550106/2010-3 to E.G.K. and M.L.N.).The authors are grateful toDr. Charles Rice (Rockfeller University, USA) who kindly providedus with Antibody anti-YFV NS4AB. J.D.A., L.C.O. and P.L.C. wererecipients of pre-doctoral and scientific initiation fellowships,respectively, from CNPq. A.A.T. and A.M.R.S. are recipients of fel-lowships from Brazilian Federal Agency for Support and Evaluationof Graduate Education (CAPES) and FAPEMIG, respectively. R.M.P.and C.C.P. were recipients of master fellowships from CAPES.M.C.C. was recipient of a post-doctoral fellowship from FAPEMIG.C.M.R. is recipient of a post-doctoral fellowship from CNPq.C.A.B., E.G.K., M.L.N. and P.C.P.F. are recipients of research fellow-ships from CNPq.

The authors would also like to thank the Microscopy Center andthe Center for Image Acquisition and Processing, both from Univer-sidade Federal de Minas Gerais, for providing technical support forelectron and confocal microscopy analysis, respectively.

References

Andrade, A., Silva, P.N., Pereira, A.C., et al., 2004. The vaccinia virus-stimulatedmitogen-activated protein kinase (MAPK) pathway is required for virusmultiplication. Biochem. J. 381, 437–446.

Bonaldo, M.C., Sequeira, P.C., Galler, R., 2014. The yellow fever 17D virus as aplatform for new live attenuated vaccines. Hum. Vaccine Immunother. 10,1256–1265.

Bray, M., 2008. Highly pathogenic RNA viral infections: challenges for antiviralresearch. Antiviral Res. 78, 1–8.

Briand, S., Beresniak, A., Nguyen, T., et al., 2009. Assessment of yellow feverepidemic risk: an original multi-criteria modeling approach. PLoS Negl. Trop.Dis. 3, e483.

Chao, D.Y., Davis, B.S., Chang, G.J., 2007. Development of multiplex real-time reversetranscriptase PCR assays for detecting eight medically important flaviviruses inmosquitoes. J. Clin. Microbiol. 45, 584–589.

Chu, J.J., Ng, M.L., 2004. Infectious entry of West Nile virus occurs through a clathrinmediated endocytic pathway. J. Virol. 78, 10543–10555.

De Clercq, E., 2013. Antivirals: past, present and future. Biochem. Pharmacol. 85,727–744.

de Magalhães, J.C., Andrade, A.A., Silva, P.N., et al., 2001. A mitogenic signaltriggered at an early stage of vaccinia virus infection: implication of MEK/ERKand protein kinase A in virus multiplication. J. Biol. Chem. 276, 38353–38360.

De Morais Bronzoni, R.V., Baleotti, F.G., Ribeiro Nogueira, R.M., et al., 2005. Duplexreverse transcription-PCR followed by nested PCR assays for detection andidentification of Brazilian alphaviruses and flaviviruses. J. Clin. Microbiol. 43,696–702.

Droebner, K., Pleschka, S., Ludwig, S., et al., 2011. Antiviral activity of the MEK-inhibitor U0126 against pandemic H1N1v and highly pathogenic avianinfluenza virus in vitro and in vivo. Antiviral Res. 92, 195–203.

Favata, M.F., Horiuchi, K.Y., Manos, E.J., et al., 1998. Identification of a novelinhibitor of mitogen-activated protein kinase kinase. J. Biol. Chem. 273, 18623–18632.

Fernandez-Garcia, M.D., Mazzon, M., Jacobs, M., et al., 2009. Pathogenesis of flavivirusinfections: using and abusing the host cell. Cell Host Microbe 5, 318–328.

Ferreira, G.P., Figueiredo, L.B., Coelho, L.F., et al., 2010. Dengue virus 3 clinicalisolates show different patterns of virulence in experimental mice infection.Microbes Infect. 12, 546–554.

Figueiredo, L.B., Batista Cecílio, A., Portela Ferreira, G., et al., 2008. Dengue virus 3genotype 1 associated with dengue fever and dengue hemorrhagic fever, Brazil.Emerg. Infect. Dis. 14, 314–316.

Furuta, Y., Gowen, B.B., Takahashi, K., et al., 2013. Favipiravir (T-705), a novel viralRNA polymerase inhibitor. Antiviral Res. 100, 446–454.

Gardner, C.L., Ryman, K.D., 2010. Yellow fever: a reemerging threat. Clin. Lab. Med.30, 237–260.

Gillespie, L.K., Hoenen, A., Morgan, G., et al., 2010. The endoplasmic reticulumprovides the membrane platform for biogenesis of the flavivirus replicationcomplex. J. Virol. 84, 10438–10447.

Gupta, N., Bhaskar, A.S., Lakshmana Rao, P.V., 2011. Transcriptional regulation andactivation of the mitogen-activated protein kinase pathway after Japaneseencephalitis virus infection in neuroblastoma cells. FEMS Immunol. Med.Microbiol. 62, 110–121.

Lindenbach, B.D., Murray, C.L., Thiel, H.J., et al., 2013. Flaviviridae: the viruses andtheir replication. In: Knipe, D.M., Howley, P.M. (Eds.), Fields Virology. LippincottWilliams & Wilkins, Philadelphia, pp. 712–746.

Mackenzie, J.M., Westaway, E.G., 2001. Assembly and maturation of the flavivirusKunjin virus appear to occur in the rough endoplasmic reticulum and along thesecretory pathway, respectively. J. Virol. 75, 10787–10799.

Martin, M., Tsai, T.F., Cropp, B., et al., 2001. Fever and multisystem organ failureassociated with 17D-204 yellow fever vaccination: a report of four cases. Lancet358, 98–104.

Mastrangelo, E., Pezzullo, M., De Burghgraeve, T., et al., 2012. Ivermectin is a potentinhibitor of flavivirus replication specifically targeting NS3 helicase activity:new prospects for an old drug. J. Antimicrob. Chemother. 67,1884–1894.

Mazzocca, A., Sciammetta, S.C., Carloni, V., et al., 2005. Binding of hepatitis C virusenvelope protein E2 to CD81 up-regulates matrix metalloproteinase-2 inhuman hepatic stellate cells. J. Biol. Chem. 280, 11329–11339.

Menzel, N., Fischl, W., Hueging, K., et al., 2012. MAP-kinase regulated cytosolicphospholipase A2 activity is essential for production of infectious hepatitis Cvirus particles. PLoS Pathog. 8, e1002829.

Miller, S., Kastner, S., Krijnse-Locker, J., et al., 2007. The non-structural protein 4A ofdengue virus is an integral membrane protein inducing membrane alterationsin a 2K-regulated manner. J. Biol. Chem. 282, 8873–8882.

Morais, A.T., Terzian, A.C., Duarte, D.V., et al., 2013. The eukaryotic translationinitiation factor 3 subunit L protein interacts with Flavivirus NS5 and maymodulate yellow fever virus replication. Virol. J. 10, 205.

Mori, T., Wang, X., Aoki, T., et al., 2002. Downregulation of matrixmetalloproteinase-9 and attenuation of edema via inhibition of ERK mitogenactivated protein kinase in traumatic brain injury. J. Neurotrauma 19, 1411–1419.

Namura, S., Iihara, K., Takami, S., et al., 2001. Intravenous administration of MEKinhibitor U0126 affords brain protection against forebrain ischemia and focalcerebral ischemia. Proc. Natl. Acad. Sci. U.S.A. 98, 11569–11574.

Op De Beeck, A., Rouillé, Y., Caron, M., et al., 2004. The transmembrane domains ofthe prM and E proteins of yellow fever virus are endoplasmic reticulumlocalization signals. J. Virol. 78, 12591–12602.

Pastorino, B., Nougairède, A., Wurtz, N., et al., 2010. Role of host cell factors inflavivirus infection: implications for pathogenesis and development of antiviraldrugs. Antiviral Res. 87, 281–294.

Pereira, A.C., Leite, F.G., Brasil, B.S., et al., 2012. A vaccinia virus-driven interplaybetween the MKK4/7-JNK1/2 pathway and cytoskeleton reorganization. J. Virol.86, 172–184.

Pierson, T.C., Kielian, M., 2012. Flaviviruses: braking the entering. Curr. Opin. Virol.3, 3–12.

Pinto, R., Herold, S., Cakarova, L., et al., 2011. Inhibition of influenza virus-inducedNF-kappaB and Raf/MEK/ERK activation can reduce both virus titers andcytokine expression simultaneously in vitro and in vivo. Antiviral Res. 92, 45–56.

Planz, O., 2013. Development of cellular signaling pathway inhibitors as newantivirals against influenza. Antiviral Res. 98, 457–468.

Raman, M., Chen, W., Cobb, M.H., 2007. Differential regulation and properties ofMAPKs. Oncogene 26, 3100–3112.

Rice, C.M., Lenches, E.M., Eddy, S.R., et al., 1985. Nucleotide sequence of yellow fevervirus: implications for flavivirus gene expression and evolution. Science 229,726–733.

Robertson, S.E., Setty, S.R., Sitaram, A., et al., 2006. Extracellular signal-regulatedkinase regulates clathrin-independent endosomal trafficking. Mol. Biol. Cell 17,645–657.

Roosendaal, J., Westaway, E.G., Khromykh, A., et al., 2006. Regulated cleavages atthe West Nile virus NS4A-2K-NS4B junctions play a major role in rearrangingcytoplasmic membranes and Golgi trafficking of the NS4A protein. J. Virol. 80,4623–4632.

Scherbik, S.V., Brinton, M.A., 2010. Virus-induced Ca2+ influx extends survival ofwest nile virus-infected cells. J. Virol. 84, 8721–8731.

Smith, J.L., Stein, D.A., Shum, D., et al., 2014. Inhibition of dengue virus replicationby a class of small molecule compounds that antagonize dopamine receptor 4and downstream mitogen activated protein kinase signaling. J. Virol.(in press).

Soares, J.A., Leite, F.G., Andrade, L.G., et al., 2009. Activation of the PI3K/Akt pathwayearly during vaccinia and cowpox virus infection is required for both hostsurvival and viral replication. J. Virol. 83, 6883–6899.

Stahla-Beek, H.J., April, D.G., Saeedi, B.J., et al., 2012. Identification of a novelantiviral inhibitor of the flavivirus guanylyltransferase enzyme. J. Virol. 86,8730–8739.

Vasconcelos, P.F., Luna, E.J., Galler, R., et al., 2001. Serious adverse events associatedwith yellow fever 17DD vaccine in Brazil: a report of two cases. Lancet 358, 91–97.

Wang, Z.Q., Wu, D.C., Huang, F.P., et al., 2004. Inhibition of MEK/ERK 1/2 pathwayreduces pro-inflammatory cytokine interleukin-1 expression in focal cerebralischemia. Brain Res. 996, 55–66.

Welsch, S., Miller, S., Romero-Brey, I., et al., 2009. Composition and three-dimensional architecture of the dengue virus replication and assembly sites.Cell Host Microbe 5, 365–375.

92 J.D. Albarnaz et al. / Antiviral Research 111 (2014) 82–92

Yang, T.C., Lai, C.C., Shiu, S.L., et al., 2010. Japanese encephalitis virus down-regulates thioredoxin and induces ROS-mediated ASK1-ERK/p38 MAPKactivation in human promonocyte cells. Microbes Infect. 12, 643–651.

Zhao, L.J., Wang, L., Ren, H., et al., 2005. Hepatitis C virus E2 protein promoteshuman hepatoma cell proliferation through the MAPK/ERK signaling pathwayvia cellular receptors. Exp. Cell Res. 305, 23–32.

Zhao, L.J., Wang, W., Ren, H., et al., 2013. ERK signaling is triggered by hepatitis Cvirus E2 protein through DC-SIGN. Cell Stress Chaperones 18, 495–501.

Zhao, L.J., Zhang, X.L., Zhao, P., et al., 2006. Up-regulation of ERK and p38 MAPKsignaling pathways by hepatitis C virus E2 envelope protein in human Tlymphoma cell line. J. Leukoc. Biol. 80, 424–432.