Introduction Chikungunya virus is a member of the genus Alphavirus, which are enveloped positive-strand RNA viruses transmitted by mosquitoes. It was first isolated in Tanzania in 1952, and reported to cause severe fever as well as myalgia, joint pain and rash within 2–5 days of infection [1], [2]. Recently, CHIKV reemerged in Eastern Africa and has developed into a major epidemic in the Indian Ocean region. In 2006, an outbreak on the island of La Réunion resulted in the infection of approximately one-third of the inhabitants [3], [4]. It has since spread to India and Southeast Asia with estimates of between 1–6 million people having been infected [5], [6]. Concerns for the globalization of this virus has evolved given the continuation of this epidemic, the high serum viremia seen in infected patients, and mutations in the currently circulating strain of CHIKV that have allowed it to adapt to a more widely distributed mosquito vector [7]. These concerns have been raised by both an increase in the number of foreign travelers contracting CHIKV and returning to both Europe and the United States, and by the possibility of spread by infected individuals, the latter being exemplified by an outbreak in Italy in 2007 and in Southern France in 2010 [5], [8], [9]. The typical clinical presentation of adults infected with CHIKV includes fever, rash, arthralgias and severe myalgias. Infected neonates, however, display more severe disease, with symptoms including encephalopathy and cerebral hemorrhage, with a subset of these infants developing permanent disabilities [10]. During the recent epidemic, it was reported for the first time that CHIKV-infected mothers can transmit the virus to their newborns during delivery, with a vertical transmission rate of approximately 50%, and in some instances infection resulted in mortality [10]. This age dependence of disease severity has also been reported for other alphaviruses and can be reproduced in mouse models through the infection of suckling mice. Indeed, it has been shown that neonatal mice succumb to Ross River Virus (RRV) [11], Semliki Forest Virus (SFV) [12], Sindbis Virus (SINV) as well as CHIKV [13]–[15]. There are likely several factors that contribute to this increased sensitivity in neonates, including alterations in the neonatal immune response. Regarding immune function in neonates, developmental delays have been described for the adaptive immune responses, however less is known about neonatal innate responses. While still controversial, many reports indicate that neonatal responses are diminished as compared to adults. For example, cord blood cells stimulated with toll-like receptor (TLR) ligands produced low levels of TNFα, IL-1β, and IL-12 [16]. Neonatal plasmacytoid dendritic cells (DCs) have also been shown to have impaired production of type I IFN in response to CpG stimulation [17]. It has been shown that LPS does not effectively activate IRF-3 dependent responses, including the production of IFNβ [18]. The response to cytokines may also be impaired as there is evidence that STAT-1 recruitment following IFNγ stimulation is less efficient in neonatal than in adult leukocytes [19]. Based on the critical role of type I IFN in the control of CHIKV infection [20], we considered the possibility that neonates may have a developmental delay in their type I IFN response, possibly contributing to their increased susceptibility to CHIKV infection and viral dissemination. In response to CHIKV infection, the production of type I IFNs is triggered by the engagement of the Rig-I like receptor (RLR) pathway [20], [21]. IFNs then stimulate the induction of hundreds of interferon stimulated genes (ISGs) and it is through these ISGs that IFNs mediate their effector function [22]. One of the earliest ISGs induced following IFN stimulation is ISG15. ISG15 is a 17 kDa protein that contains two ubiquitin-like domains connected by a proline peptide linker [23]. Similar to ubiquitin, ISG15 can form conjugates with an array of intracellular host and viral proteins through the use of an E1 (UBE1L), E2 (UbcH8), and E3 (e.g., Herc5, EFP, HHARI) enzymatic cascade [24]–[28]. Conjugation of ISG15 to target proteins has been suggested to cause either a gain or loss of function of the targeted protein, although the consequences of ISG15 conjugate formation are not well understood [29]–[32]. An unconjugated form of ISG15 can also be found in the sera of humans treated with IFN-βser as well as in virally infected mice [33], [34]. The released form of ISG15 has been suggested to have cytokine like activity [35]–[37], however its role during an antiviral immune response has not been examined. The rapid induction of ISG15 following IFN stimulation has led to the identification of an antiviral role for ISG15 during infection. ISG15−/− mice display increased susceptibility to SINV, herpes simplex virus-1 (HSV-1), gamma herpes virus (γHV68), influenza A and B viruses, and vaccinia virus [34], [38]. ISG15−/− mice infected with influenza B virus display a 2–3 log increase in lung viral titers as well as elevated cytokine and chemokine levels [34]. The ability of ISG15 to conjugate to target proteins appears to be essential for ISG15's antiviral activity during certain viral infections, as UbE1L−/− mice, which lack the ability to form ISG15 conjugates, phenocopy ISG15−/− mice during both SINV and influenza B virus infection [39]–[41]. Further support for the importance of ISG15 conjugation during viral infection comes from the evolution of viral proteins that directly target ISG15 conjugate formation. Both the NS1 protein of influenza B virus and the E3L protein of vaccinia virus inhibit ISG15 conjugate formation, while OTU-domain containing viral proteins, such as the L protein of Crimean-Congo hemorrhagic fever virus (CCHFV) or the Nsp2 protein of Equine Arteritis virus (EAV), and the SARS coronavirus papain-like protease (SARS-CoV PLpro) exert both deubiquitinating and deISGylating activity [24], [38], [42], [43]. Therefore, the antiviral activity of ISG15 has been thought to be conjugation dependent. Herein we demonstrate that despite an increased susceptibility of neonates to CHIKV infection, they produce robust levels of type I IFNs. While insufficient to completely control infection, IFNs participate in limiting the infection. We also show that ISG15 is induced during CHIKV infection and plays a critical role in protecting neonatal mice from viral induced lethality. Surprisingly, the mechanism of action by which ISG15 limits infection is independent of UbE1L mediated conjugation, as UbE1L−/− mice displayed no phenotypic differences as compared to WT animals. Furthermore, ISG15 does not directly inhibit viral replication, as suggested by the similar viral loads in WT and ISG15−/− mice. Instead, ISG15 appears to function as an immunomodulatory molecule in this model. These data demonstrate a novel role for ISG15 during viral infection and suggests that prophylactic measures targeting the induction of IFN and ISG15 may help protect neonates during future CHIKV outbreaks. Results CHIKV infected infants produce high levels of IFN and IFN-induced chemokines/cytokines Based on the increased severity of neonatal disease that has been observed during the recent epidemic of CHIKV, we assessed the inflammatory response of infants during the acute phase of CHIKV infection. Patients were recruited at the time of presentation in the emergency room and sera samples were collected and stored. CHIKV infection was confirmed by RT-PCR; and all patients were negative for anti-CHIKV IgG and IgM, indicating acute infection. We performed multi-analyte testing using Luminex technology, with a focus on inflammatory cytokines and chemokines. The inflammatory signature was compared to uninfected patients presenting to the emergency room for reasons unrelated to acute infection (e.g., broken bone). The intensity of the immune response in the infant vs. adult cohorts was compared. Non-parametric tests were used for the statistical analysis, and a false discovery rate (FDR) correction was applied to all p-values in order to adjust for multiple testing. We detected elevated levels of both IFNα and IFNγ in both groups of patients when compared to their respective control group (p global median: high viral load); in both groups the levels of IFNα were more elevated in infected infants than in infected adults ( Figure 2D ). Furthermore, we performed univariate linear regression analysis to model the effect of infant status on plasma IFNα concentrations. Infants had higher IFNα concentrations compared to adults (RGM [95% CI] = 5.49 [3.16–9.53]; p median: high viral load; Median = 4.8×107) and statistic are represented (Mann Whitney test) (E) For analytes identified in Figure 1 that showed a correlation with viral load, IFN or age are plotted in a network array, illustrating the correlations identified in neonatal vs. adult individuals. Connecting lines indicate Spearman correlation (rs) values; positive correlations in red and negative correlations are depicted in blue. IFNα data from adult patients were previously reported [20], but are shown here for comparison to data from infected neonates. CHIKV infection in neonatal mice results in a robust type I interferon and proinflammatory cytokine response that is critical in controlling infection Previously, we reported an age-dependent susceptibility to CHIKV infection in mice. Infection of 6 day old animals resulted in 100% mortality; 9 day old animals developed paralysis, with approximately 50% of the animals succumbing to infection; while by 12 days of age the mice became refractory to symptoms of severe disease and lethality [15]. To compare our experimental mouse model of CHIKV infection to the response seen in human infants we assessed the IFN and proinflammatory responses in 8–9 day old mice. We first assessed the induction of the IFN response at the local site of infection by monitoring mRNA levels of IFNβ and selected interferon stimulated genes (ISGs) in the skin. Mice were inoculated with 2×105 PFU of CHIKV and the injection site was removed between 3–120 hrs post-infection. Increased expression of IFNβ mRNA could be detected as early as 3 hrs post-infection with peak levels being achieved at 16 hrs post-infection ( Figure 3A ). Similar to IFNβ mRNA induction, IRF7, Mx1 and ISG15 mRNA levels could also be detected as early as 3 hrs post-infection, with peak levels observed at 16 hrs post-infection ( Figure 3A ). Thus, at the site of infection, neonatal mice are able to induce IFNβ expression as well as a subset of known ISGs. Of note, IRF7 and Mx1 mRNA expression is indicative of signaling via the type I IFN receptor, suggesting that the production as well as reception of IFNαβ is intact in neonatal mice. 10.1371/journal.ppat.1002322.g003 Figure 3 Neonatal mice mount a robust IFN and pro-inflammatory cytokine/chemokine response following CHIKV infection. Nine day old C57BL/6 mice were injected s.c. in the right flank with CHIKV. (A) The skin at the site of infection was harvested at the indicated times post infection and IFNβ, IRF7, Mx1 and ISG15 gene expression was evaluated by qRT-PCR. The ΔCt was calculated using GAPD as a reference gene. At various time points, sera was harvested and tested for (B) IFNα by ELISA and IFNγusing Luminex. (C) Serum cytokines and chemokines were also analyzed using Luminex. Each graph represents 2–3 independent experiments with 3 mice per experiment. We next assessed the systemic inflammatory response in this model. Similar to our findings in human infants infected with CHIKV, we observed a strong induction of IFNα and IFNγ in infected pups ( Figure 3B ). Plasma concentration of CCL2, CCL4, CXCL9 and CXCL10 were also elevated, however IL-12p70 was only modestly induced ( Figure 3C ). Similar to the human data, a mixed Th1, Th2 and Th17 cytokine profile was observed with the induction of IL-12, IL-5, IL-13, IL-15, and IL-17 ( Figure 3C ). For all analytes, peak levels were seen 16–24 hrs post-infection ( Figure 3 ). Notably, there were some differences seen between the murine neonate and human infant inflammatory profiles. Most interestingly, the mice displayed increased levels of the pyrogenic cytokines, including IL-1β, IL-6 and TNFα, which were not seen in our studies of human infants ( Figure 3C ). While this may represent differences in pathogenesis, we believe it is more a reflection of the fact that the mice can be assessed within hours of viral inoculation, while the exact timing of the human infection is unknown. Overall, we find that the similarities between the mouse and human responses support the use of neonatal mice to study the response to CHIKV infection, and indicate that induction of IFNs, as well as the triggering of an ISG response, are both rapid and robust. Next, to confirm that endogenous IFN contributes to the control of CHIKV in neonatal mice, mice lacking subunit 1 of the type I IFN receptor (IFNAR−/−) mice were infected with CHIKV at 9 days of age. Consistent with our previous observations in adult mice [15], neonatal pups lacking IFNAR1 were highly susceptible to CHIKV infection with 100% of the pups dying by day 2 post-infection ( Figure 4A ). These pups developed a rapid, disseminated infection. Within 1 day of infection, the IFNAR−/− mice displayed viral loads at the injection site that were 100-fold higher than that detected in WT controls. We also observed a striking increase in viral titers in the serum and multiple organs, including the brain, liver, and lung ( Figure 4B ). These data indicate that endogenous IFN, while insufficient to protect mice, plays an important role in limiting CHIKV infection during disease pathogenesis in neonatal animals. 10.1371/journal.ppat.1002322.g004 Figure 4 Endogenous Type I IFN is required for the control of neonatal CHIKV infection. WT and IFNAR−/− mice were infected at 9 days of age with 2×105 PFU CHIKV s.c. (A) Mice were monitored for lethality for 21 days with data displayed as Kaplan-Meier curves. (B) Tissue and sera were collected on day 1 post-infection and viral titers were determined using a standard plaque assay. Mann-Whitney statistical comparison of WT and IFNAR−/− viral titers are shown where * indicates p 5s) to return and land on its feet when flipped on its back. For experiments with adult mice, 6–8 week old mice were infected as outlined above and followed for daily weight loss and lethality. For experiments utilizing the recombinant CHIKV clones, 6 day old pups were infected with 3×105 PFU of the indicated recombinant virus diluted in 30 µL of PBS by s.c. injection into the right flank. Viral titers were determined in organs harvested at the indicated days post-infection. Organs were harvested into 1 ml of DMEM without fetal bovine serum and homogenized with 1.0 mm diameter zirconia-silica beans at 3,200 rpm for 1 minute with a MagnaLyzer prior to plaque assay on BHK cells, protocol modified from [63]. Serial dilutions of organ homogenates in DMEM with 1% FBS was added to BHK cells (6×105 cells for 6 well plates) and incubated for 1 hr at 37°C. An agar overlay was then added to the cells and incubated for 28 hrs at 37°C. Plates were fixed with 1% formaldehyde (>30 min at room temperature), agar plugs were removed and plaques were visualized using a 1% crystal violet solution. Poly I∶C prophylactic treatment studies Eight day old mice were injected intraperitoneally (i.p.) with 10–25 µg of high molecular weight pIC (Invivogen). Twenty-four hours later, mice were challenged with 2×105 PFU CHIKV s.c. Mice were monitored daily for the development of symptoms and followed for lethality as described above. Cytokine analysis Sera were harvested and conserved at −80°C for analysis. Human cytokines were measured by Luminex (25 plex kits, Biosource, Invitrogen) following manufacturer's instructions. Human CXCL10 was re-titrated by ELISA (human quantikine ELISA kit, R&D). Mouse sera were obtained after coagulation of blood in T-MG tubes (Terumo). Mouse IFNα levels were quantified by ELISA (PBL biomedical) and other cytokines were measured using Luminex technology with either the 32 plex from Millipore (MPXMCYTO-70X) or by customized 10 plex from Biorad. qRT-PCR for quantification of viral load and IFN mRNA For determination of patient viral load, total nucleic acid extraction was performed on sera in a MagNa Pure automate using the Total Nucleic Acid Kit (Roche Diagnostics). CHIKV RNA was detected with specific taqman probes using a one step RT-PCR (Master RNA hybridization probes, Roche) performed on a Chromo 4 machine (Biorad). The 20 µL reaction mix contained 2 µL of extracted RNA, 7.5 µL of LightCycler RNA Master Hyb-Probe, 3.25 mmol/L Mn2, 450 nmol/L CHIKV-forward primer, 150 nmol/L CHIKV-reverse primer, 150 nmol/L CHIKV Probe (5 6-carboxyfluorescein-3 TAMRA) (TibMol-Biol). The thermal cycling consisted of a reverse transcription at 61°C for 20 min followed by 45 cycles at 95°C for 5 s and 60°C for 15 s. The fluorescence was measured at 530 nm. CHIKV load is measured using a synthetic RNA calibrator [64]. CHIKV-rev CCAAATTGTCCGGGTCCTCCT; CHIKV-forw AAGCTCCGCGTCCTTTACCAAG; Probe: Fam-CCATGTCTTCAGCCTGGACACCTTT-TAMRA. For mouse studies, skin tissue was harvested at the site of infection on days 1 and 2 post-infection. Tissue was snap frozen in liquid nitrogen and then homogenized in RLT+ with 0.04 M DTT, and RNA was extracted with the Qiasymphony robot (Qiagen) with a protease step and a DNase step. The quality and quantity of RNA was evaluated with the Agilent technology, with the RNA integrity number between 8 and 9.5. Reverse transcription was performed with random primers (Roche) using Superscript enzyme (Invivogen). cDNA for murine IFN-β and ISG15 were detected using Applied Biosystem Taqman probes (Mm00439546-s1, Mm01705338-s1). To analyze the relative fold induction of mRNA, GAPD expression levels were determined in parallel for normalization using the CT method. Western blot analysis Nine day old mice were infected with 2×105 PFU CHIKV. Tissue homogenates as well as serum samples were subjected to protein electrophoresis on a 12% Tris gel. The gel was transferred to a polyvinylidene fluoride membrane and probed for ISG15 expression using a rabbit anti-ISG15 polyclonal serum (1∶5000) as previously described [62]. The membrane was then developed with a horseradish peroxidase (HRP)-conjugated goat anti-rabbit antiserum (Jackson Immunoresearch, West Grove, Pennsylvania) diluted 1∶200,000. For loading controls, the same blot was re-probed with anti-β-actin mAb (clone AC-74; Sigma) and then developed with a HRP conjugated goat anti-mouse antibody (Jackson Immuno-Research). All blots were developed with chemiluminescent reagent (Millipore). Flow cytometry Spleens were harvested from naïve nine day old WT, UbE1L−/− and ISG15−/− mice (12 mice per strain). Lymphocyte subsets were stained using the following cell surface markers: CD3, CD4, CD8, NK1.1, CD19, F4/80 and Gr1. Data is represented as percent of the total cell population and Kruskal-Wallis test was used to compare the three genotypes. Statistical analysis Human data was analyzed using the OMNIVIZ statistical platform (BioWisdom, Cambridge, UK) to perform comparisons among data sets using nonparametric tests (Mann-Whitney U-test) and false discovery rate (FDR) procedures, a permutation-based method to correct for the increased probability of obtaining a false positives among all significant tests [65]. Additional data was analyzed using the Prism software (Graphpad software). Differences were considered significant if p 5 s). 3 independent experiments are compiled. (PPT) Click here for additional data file. Figure S2 Induction of ISG15 during neonatal CHIKV infection is largely dependent upon IFNAR1 signaling. Serum from nine day old WT, UbE1L−/−, ISG15−/− and IFNAR−/− mice infected s.c. with 2×105 PFU CHIKV was collected on day 1 post-infection and ISG15 expression was assessed by western blot. (PPT) Click here for additional data file. Figure S3 The role of ISG15 during CHIKV infection is age dependent. WT, UbE1L−/− and ISG15−/− mice were infected with 2×105µ PFU CHIKV s.c. at either (A) eleven days of age, (B) twelve days of age, or (C) 6–8 weeks of age and were monitored for survival for 21 days post-infection. Kaplan-Meier survival curves are shown. (PPT) Click here for additional data file. Figure S4 Recombinant CHIKV viruses expressing WT ISG15 do not rescue ISG15 −/− mice. Recombinant CHIK viruses were generated to express the following proteins: wild type ISG15(−LRLRGG), non-conjugatable ISG15(−LRLRAA), and GFP(−GFP). (A) Schematic representation of recombinant CHIK clones adapted from [61]. (B and C) BHK cells were infected with the indicated rCHIK viruses at an MOI = 1. (B) Viral titers were measured at 0,6,12,24,48 hrs post-infection by plaque assay. (C) Cell lysates were collected at 0,6,12, and 24 hrs post-infection and were analyzed for ISG15 expression by western blot. (D) Six day old ISG15−/− mice were infected with either CHIK-GFP or CHIK-LRLRGG at 3×105 PFU s.c. Mice were monitored for lethality for 21 days with data displayed as Kaplan-Meier curves. (PPT) Click here for additional data file. Table S1 Clinical signs in adult and infant cohort. (PPT) Click here for additional data file. Table S2 Splenic lymphocyte subsets in naïve WT, UbE1L −/− and ISG15 −/− neonatal mice (% total cell population, mean ± sem). (PPT) Click here for additional data file.
