The ongoing epidemic of Chikungunya virus occurring in the Indian Ocean region has highlighted how little we understand about the pathogenesis of this virus. Epidemiological studies have provided the first documentation of vertical transmission, as well as providing detailed information about the severity of disease and long term sequelae 
. One important finding from these studies concerns the increased susceptibility of neonates and infants to severe forms of Chikungunya disease 
. In this study, we evaluated the response of infants to CHIKV infection using data from both human samples collected during the La Réunion outbreak, as well as taking advantage of a newly described mouse model of infection. Our results show that human infants and murine neonates mount a robust innate immune response to CHIKV infection, which includes the induction of type I IFNs, several cytokines and chemokines, and the induction of at least a subset of IFN induced genes, including ISG15. We establish a role for ISG15 in the pathogenesis of CHIKV infection with an absolutely essential role in the neonatal response to infection. Moreover, the reported data suggest that ISG15 acts independent of UbE1L mediated conjugation, and rather than exerting a direct anti-viral role, it appears to be implicated in limiting an exaggerated inflammatory response.
In general, neonates are more susceptible to microbial and viral infection. This vulnerability has been explained by two principal mechanisms: broader tropism of the infectious agent or a defective host response. Regarding the latter, many argue that neonatal susceptibility to infection is due to a delayed or weaker immune response 
. Factors contributing to this include delays in immune system maturation, decreased expression of activation receptors, or distinct regulation of signaling pathways in neonatal vs. adult immune cells. Since the type I IFN response is critical for controlling CHIKV infection 
, we hypothesized that neonates may have a defect in their ability to either produce and/or respond to IFN. Instead, we observed that the production of type I IFN was intact in both human infants and mouse neonates. Furthermore, the relative level of IFN produced in infants was higher than the responses observed in adult human patients, even when viral load was normalized between the two patient groups (–
). These data indicate that neonates do not have a defect in their ability to produce type I IFN during CHIKV infection, and may actually be hyper-responsive. Very little has been reported on the signaling through RLRs or other viral sensors in neonates. Most previous work has supported the notion that signaling through TLRs, including TLR4, are diminished in neonates 
. However, one study did find that neonatal mice exhibited increased lethality following LPS treatment, due in part to an exaggerated pro-inflammatory cytokine response as compared to adult mice 
. Future work is needed to characterize differences that may exist between neonatal and adult RLR signaling.
We also demonstrate that neonates can respond to the IFN that is produced. In both human infants and in neonatal mice, several known IFN-induced chemokines and cytokines were upregulated during the course of infection (–). These results suggest that the ability to respond to IFNs was at least partially intact. These observations were confirmed by demonstrating the greater susceptibility of neonatal mice lacking expression of the type I IFN receptor (); notably, the kinetics of viral replication and death were massively accelerated as compared to age-matched WT controls. Additionally, we demonstrate that the prophylactic exposure to an IFN-inducing adjuvant protected animals from challenge with lethal doses of CHIKV ().
To provide insight into the mechanism by which IFN participates in the host response to CHIKV infection, we evaluated the role of ISG15, an anti-viral host protein that has previously been shown to be important in the control of several viruses, including SINV 
. We found that ISG15−/−
mice were more susceptible to CHIKV infection, demonstrating a dramatic increase in lethality as compared to WT mice (). Moreover, ISG15 played a critical role in pIC induced protection of mice in a prophylactic setting (). While OAS has previously been shown to inhibit viral replication when over-expressed in vitro 
, to our knowledge our data is the first in vivo
demonstration of an IFN effector molecule having activity against CHIKV.
Most importantly, the results from our current study indicate that ISG15 is regulating CHIKV pathogenesis by a unique mechanism of action. First, we demonstrate that ISG15 regulates CHIKV infection independent of UbE1L mediated conjugation. The protection mediated by ISG15 during pIC prophylaxis also appeared to be UbE1L independent, as UbE1L−/−
mice displayed survival curves similar to WT mice (). These results suggest that the non-classical function of ISG15 is at work both during acute viral infection and in the pIC induced protection seen in our mice. This is in contrast to both influenza virus and SINV infection, where the anti-viral activity of ISG15 is dependent upon ISG15 conjugation and abrogated in UbE1L−/−
. To date, UbE1L is the only known E1 for the ISG15 pathway. A second E1 has recently been identified for the ubiquitin pathway 
, leaving open the possibility that another E1 may exist for ISG15. However, our current analysis () and previous studies 
have revealed no conjugation activity in UbE1L−/−
cells and mice. Therefore the actions of ISG15 during CHIKV infection are likely to be independent of conjugation, and mediated by free ISG15. Second, it appears that during CHIKV infection, ISG15 is not functioning as a direct antiviral molecule. In both the influenza and SINV models, the increase in lethality was accompanied by a dramatic increase in viral loads 
. In contrast, during CHIKV infection, ISG15−/−
mice did not show an increased CHIKV burden (). Instead, we noted a significant elevation of several proinflammatory cytokines and chemokines in the ISG15 deficient mice (). Therefore, as opposed to having direct antiviral activity, it appears that ISG15 modulates the immune response during CHIKV infection. Finally, in contrast to what was previously reported for control of SINV infection 
, a recombinant CHIKV expressing ISG15 was unable to rescue neonatal ISG15−/−
mice from viral induced lethality (Figure S4
). The inability to rescue the ISG15−/−
phenotype may be due to insufficient levels or inappropriate timing of ISG15 expression; or alternatively, may indicate that ISG15 expression is required in an uninfected cell. Since we detect no differences in viral load, and instead observed increased cytokine levels in the ISG15−/−
mice, we favor this latter possibility. Further analysis into the precise mechanism by which ISG15 regulates the host response to CHIKV should provide additional insight into this issue. Together, our data suggest a novel mechanism for ISG15, which is likely to be independent of conjugation and extrinsic to virally infected cells.
The precise mechanism by which ISG15, independent of UbE1L mediated conjugation, contributes to the control of viral infection is currently unclear. The most intriguing difference we have noted to date is the increased cytokine responses in the mice lacking ISG15 (). As noted above, while we cannot formally exclude the possibility of another E1 functioning in this system, it seems most likely that free, unconjugated ISG15 mediates the activity during CHIKV infection. Free ISG15 is found within the cell, but interestingly it may also be secreted by a still undefined mechanism 
. Previous work has shown that unanchored ISG15 can associate non-covalently with proteins (i.e. independent of conjugation). For example, the NS1 protein of influenza B virus can non-covalently bind ISG15, thereby inhibiting its interaction with UbE1L and blocking conjugation of target proteins 
. The over-expression of ISG15 has also been shown to disrupt Nedd4 ligase activity and inhibit Ebola virus VLP release 
. Recent research within the ubiquitin field has described a role for unanchored polyubiquitin chains, shown to regulate TRAF6 function, as well as promote RIG-I dimerization and signaling 
. It is therefore possible that intracellular, unanchored ISG15 interacts non-covalently with members of an innate immune signaling pathway to regulate cytokine and chemokine production or other host response pathways. Alternatively, released ISG15 could be contributing to the phenotype seen during CHIKV infection. Indeed, the 17 kDa form of ISG15 is released into the serum in both WT and UbE1L−/−
mice during CHIKV infection (). Released ISG15 has been reported to function as an immunomodulatory molecule, increasing NK cell proliferation and lytic activity, acting as a neutrophil chemoattractant, and upregulating E-cadherin expression on dendritic cells 
. Released ISG15 could function as in immunomodulatory cytokine by signaling through a receptor to regulate the cytokine response or through its ability to function as a chemoattractant. In order to characterize these effects in greater detail, a receptor for ISG15 must be identified. Future studies evaluating these possibilities will be required in order to further define the mechanism by which ISG15 is contributing to the host response to CHIKV.
In conclusion, we have demonstrated that neonates are capable of producing type I IFN in response to CHIKV, which serves to limit viral infection though remains insufficient to clear the virus. We have demonstrated a critical, age-dependant role for ISG15 during neonatal infection. We have also characterized the mechanism of ISG15 activity, revealing a novel mechanism for ISG15, independent of UbE1L mediated conjugation, and functioning as a putative immunomodulator of proinflammatory cytokines. The ability of pIC to protect neonatal mice against CHIKV infection suggests that manipulation of the IFN signaling pathway, and perhaps the induction of ISG15, may be an appropriate therapeutic target for combating CHIKV infection.