The cell biology of Chikungunya virus infection

Authors

  • Bor Luen Tang

    Corresponding author
    • Department of Biochemistry, Yong Loo Lin School of Medicine and NUS Graduate School of Integrative Sciences and Engineering, National University of Singapore, Singapore, Singapore
    Search for more papers by this author

For correspondence. E-mail bchtbl@nus.edu.sg; Tel. (+65) 6516 1040; Fax (+65) 6779 1453.

Summary

Chikungunya virus (CHIKV) infection causes a disease which appears to affect multiple cell types and tissues. The acute phase is manifested by a non-fatal febrile illness, polyarthralgia and maculopapular rashes in adults, but with recurrent arthralgia that may linger for months during convalescence. The issue of cellular and tissue tropism of CHIKV has elicited interest primarily because of this lingering incapacitating chronic joint pain, as well as clear encephalopathy in severe cases among neonates during the re-emergence of the virus in recent epidemics. The principle cell types productively infected by CHIKV are skin fibroblasts, epithelial cells and lymphoid tissues. There is controversy as to whether CHIKV productively infects haematopoietic cells and neurones/glia. CHIKV infection triggers rapid and robust innate immune responses which quickly clears the acute phase infection. However, significant acute as well as chronic infection of less obvious cell types, such as monocytes, neurones/glia or even CNS neural progenitors may conceivably occur. There is therefore a need to ascertain the full range potential of CHIKV tropism, fully understand the cellular responses triggered during the acute the convalescent phases, and explore possible cell types that might be the source of chronic problems associated with CHIKV infection.

Introduction

Chikungunya fever as a viral epidemic transmitted by mosquitoes was first recognized in Tanzania in the 1950s. The term ‘Chikungunya’ originated from an ethnic phrase that means ‘to walk bend over’, which aptly pictures the incapacitating joint pain suffered by infected individuals. The disease re-emerged (Powers and Logue, 2007) in Kenya in 2004, and spread to the islands of the Indian Ocean, Comoros and La Réunion (Bonn, 2006; Kariuki Njenga et al., 2008). In the latter, the epidemic was both widespread and severe, with an estimated 270 000 cases of Chikungunya virus (CHIKV) infections (nearly half the population), and 237 deaths reported (Renault et al., 2007; Staikowsky et al., 2009). The re-emerging epidemic subsequently spread to regions like India and Southeast Asia, with documented outbreak in Europe. In all, cases of Chikungunya fever have now been identified in more than 40 countries (Powers and Logue, 2007; Rezza et al., 2007; Staples et al., 2009).

Taxonomically, CHIKV is a positive single-strand (ss) RNA virus of the Alphavirus (group IV) genus of the Togaviridae family, belonging to the Semliki Forest Virus (SFV) complex. Like close members of the genus such as O'Nyong Nyong Virus and Ross River Virus (RRV), these Old World Alphaviruses all cause illnesses characterized by fever, joint pain and rashes. The 11 kb CHIKV RNA genome encodes two open reading frames (ORFs). The 5′ ORF encodes the virus non-structural proteins (NSP) 1–4, which together form the virus replicase, whereas the 3′ ORF encodes the capsid and the envelope glycoproteins (Solignat et al., 2009). Transmitted via mosquitoes of the Aedes genus, CHIKV replicates in the dermal site of inoculation and is then disseminated to other parts of the body. An incubation period of 2–4 days is followed by a sudden onset of clinical symptoms that include high fever (usually > 38.5°C), rigours, celphagia, myalgia, petechial or maculopapular rashes, and often incapacitating arthralgia (Mourya and Mishra, 2006; Yazdani and Kaushik, 2007). The acute phase viraemia could have a viral load of 109–1012 viral particle ml−1, and a strong type 1 interferon (IFN) response with the production of inflammatory cytokines are usually observed (Ng et al., 2009; Schilte et al., 2010). Acute symptoms are typically resolved within 2 weeks. Convalescence from acute symptoms may, however, be accompanied by a chronic arthralgia in 30–40% of patients that may last for weeks, months or longer (Hoarau et al., 2010).

Several fundamentally important aspects of CHIKV infection of human cells and tissues are not well resolved. Early findings indicate that blood cells do not contribute to the high viraemic load (Sourisseau et al., 2007) or interferon response (Schilte et al., 2010), but this view has been disputed by other findings (Her et al., 2010). The mechanism underlying chronic persistence of arthralgia/arthritis in CHIKV-infected individuals in spite of a robust innate and adaptive immune response (Hoarau et al., 2010) is also unclear, as chronic viraemia could not be clearly demonstrated in these patients. However, it is known that variants of the related RRV could persistently infect macrophages and showed significantly enhanced resistance to IFN-β-stimulated antiviral activity (Lidbury et al., 2011).

Another issue that has remained unclear pertains to CHIKV's ability to infect brain cells. Chikungunya fever is generally considered a non-fatal disease, and its Old World Alphavirus taxonomic affinity tends to limit perceptions of its pathology beyond joint diseases. In particular, although there are some association between CHIKV infection and encephalopathy in the some early epidemic episodes, the virus is not usually viewed as truly neurotrophic (Arpino et al., 2009). However, this notion is challenged in some re-emergence epidemic episodes, prominently that of La Réunion, which had a mortality rate of 1:1000 (Renault et al., 2007; Staikowsky et al., 2009; Das et al., 2010). Clinical facilities on this département d'outre-mer of France had allowed a good clinical documentation of disease and fatalities among newborns (mother-to-child infection), infants and elderly patients. Some cases with severe encephalitis and peripheral neuropathies were noted (Das et al., 2010), which suggest that CHIKV may be encephalitogenic. In the paragraphs that follow, controversies with regards to cell type susceptibility to productive CHIKV infection are discussed in the light of recent findings.

When CHIKV enters the cell – a brief review of known and inferred events

As per the transmission cycle of all arboviruses, CHIKV infects female mosquitoes via a viraemic blood meal of a zoonotic or human host, which passes the virus to another human host during subsequent feeding. The recent re-emergence of CHIKV is associated with an interesting evolutionary event of enhanced vector host range. Aedes aegypti mosquitoes were the classical vector of CHIKV. However, an extension in viral host range has likely resulted from CHIKV's adoption of Aedes albopictus as a vector (Schuffenecker et al., 2006; Tsetsarkin et al., 2007; Vazeille et al., 2007). A specific mutation in the envelope protein E1 (A226V) was apparently responsible for a significant increase in CHIKV infectivity for A. albopictus, and led to more efficient viral dissemination into mosquito secondary organs and transmission to neonate mice (Tsetsarkin et al., 2007). The E1-A226V mutated La Réunion isolate (LR-OPY1) replicated more efficiently than the African reference strain (37997) in A. albopictus cells (Gay et al., 2012). Geographically, A. albopictus is apparently more common than A. aegypti in some regions. The species survives well in both rural and urban environments, and may have thus facilitated CHIKV spread to immunologically naive human populations in urban settings.

Earlier work had shown that CHIKV could infect a variety of non-human and human cell lines, including some common laboratory cell lines such as the cervical carcinoma epithelial cell line HeLa, the kidney epithelial cell line HEK-293T, the hepatocarcinoma epithelial cell line HUH7, and the neuroblastoma cell line SH-SY5Y (Tsetsarkin et al., 2006; Solignat et al., 2009). Sourisseau and colleagues published the first extensive characterization of human cell types that support the replication of recent CHIKV isolates in vitro (Sourisseau et al., 2007). The authors found that the best cellular host for productive viral infection in humans are fibroblasts, epithelial/endothelial cells, and to a lesser extent, macrophages. However, CHIKV failed to replicate in either lymphoid and monocytoid cell lines, or primary lymphocytes and monocytes, neither did it infect monocyte-derived dendritic cells (Sourisseau et al., 2007). Another report showed that CHIKV could replicate in human muscle progenitors, or satellite cells, but not myotubes, which is demonstrated by immunohistochemical identification of viral antigens in satellite cells but not muscle fibres (Ozden et al., 2007).

Neither mosquito nor human cell surface receptor(s) specific for CHIKV are known (Solignat et al., 2009), but surface molecules such as the laminin receptor, C-type lectins [such as Dendritic Cell-Specific Intercellular adhesion molecule-3-Grabbing Non-integrin (DC-SIGN) and Liver/Lymph node-specific (L)-SIGN] as well as heparan sulfate proteoglycans have been implicated as receptors for other Alphaviruses (Leung et al., 2011). Another possibility is the natural resistance-associated macrophage protein (NRAMP), which has recently been shown to mediate Sindbis virus binding and entry in both Drosophila and mammalian cells (Rose et al., 2011). CHIKV could well use one or more of these for cell entry via receptor-mediated endocytosis. CHIKV entry into both mosquito (Gay et al., 2012) and human cells (Sourisseau et al., 2007) are inhibited by disruption of endosomal acidification and cholesterol depletion, and for the latter also by silencing of dynamin-2 (Sourisseau et al., 2007) and functional disruption of endosomal membrane trafficking factors such as Ep15 and Rab5 (Bernard et al., 2010), indicating that viral entry occurs through endosomal pH-dependent endocytosis. Although the general events known from the analysis of other Alphaviruses (Strauss and Strauss, 1994; Jose et al., 2009) could predict the following sequence of events to some degree, the details would need to be filled in. Upon endocytosis, the acidic environment of the endosome triggers conformational changes in the viral envelope (Li et al., 2010; Voss et al., 2010), resulting in the dissociation of E1 from E2, and exposes the previously buried fusion sequence of E1. E1-mediated viral-cell membrane fusion allows cytoplasmic delivery of the core and release of the viral ssRNA genome. Translation from the viral mRNA generates a 2474-amino-acid (aa) polyprotein precursor, whose cleavage generates the non-structural proteins nsP1–nsP4, which form the viral replication complex. A full-length negative-strand RNA intermediate is first synthesized, and this serves as the template for the synthesis of a subgenomic (26S) and the viral genomic (49S) RNA.

The subsequent events are largely inferred from the studies on SINV and SFV (Jose et al., 2009). The structural C-pE2-6K-E1 1244 aa polyprotein precursor is translated from the 26S RNA, and is processed by an autoproteolytic serine protease activity, releasing the capsid protein (C), which remains in the cytoplasm. The E3 portion contains a signal sequence that translocates the remaining polyprotein, E3-E2-6K-E1 (the spike proteins), into the endoplasmic reticulum (ER). Upon insertion, ER signalase cleaves the polyprotein into pE2 (E2-E3), 6K and E1 (Liljeström and Garoff, 1991). pE2 and E1 associate into heterodimers within the ER, and oligomerize to form non-fusogenic spikes that are then transported through the Golgi apparatus to the plasma membrane. During this step, pE2 is cleaved by a host cell furin or furin-like protease activity to form E2 and E3 (de Curtis and Simons, 1988; Ozden et al., 2008). Binding of the viral nucleocapsid to the viral RNA and the recruitment of the membrane-associated envelope glycoproteins completes viral particle assembly, and the assembled particle buds at the cell membrane.

Cellular responses and events associated with CHIKV infection

CHIKV infection of cells is typically cytopathic, with rapid onset of syncytium formation and induction of apoptosis in infected cells (Sourisseau et al., 2007; Dhanwani et al., 2012; Wikan et al., 2012). Apoptosis appears to occur by both the intrinsic and extrinsic pathways (Krejbich-Trotot et al., 2011a; Wikan et al., 2012), and could, at least partly, result from the innate immune response discussed below, augmented by the general inhibition of cellular translation by CHIKV. Interestingly, apoptotic host destruction could still be used by the virus to some degree, as CHIKV is able to infect neighbouring cells by hiding in apoptotic blebs and being subsequently phagocytosed (Krejbich-Trotot et al., 2011a). A recent investigation using HEK293 cells also showed that CHIKV infection induces autophagy (Krejbich-Trotot et al., 2011b). Although autophagy has been shown to be an important component of the innate immune response and is involved in host defence elimination of pathogens (Kuballa et al., 2012), some viruses have been shown to subvert this host process to aid their replication (Jackson et al., 2005). This may well be the case for CHIKV, as its replication was reduced by inhibition of autophagy (Krejbich-Trotot et al., 2011b). This property is unlike that of SFV, for example, whose glycoproteins induce autophagosome formation, but disruption of autophagy had no effect on viral replication rate or formation of viral replication complexes (Eng et al., 2012).

Very recent work has in fact indicated that simultaneous autophagy induction delays or attenuates apoptotic death by CHIKV infection (Joubert et al., 2012). During CHIKV infection, both oxidative stress [via reactive oxygen species-mediated inhibition of the mechanistic Target of Rapamycin (mTOR)] and endoplasmic reticulum stress pathways [through the Inositol-requiring protein 1 α (IRE1α) pathway] could potentially trigger cell death, but these pathways maybe be more relevant in triggering an early autophagic response rather than apoptosis in CHIKV-infected cells. Clear evidence on a role for autophagy in the CHIKV infection pathology is shown by experimental manipulation of key autophagy components. Atg5−/− MEFs showed a dramatic increase in CHIKV infection-induced cell death compared with wild-type MEFs, and Atg16LHM mice with reduced autophagy has increased lethality and sensitivity to CHIKV-induced apoptosis. Interestingly, based on kinetic studies of pro-caspase 3 cleavage, the authors inferred that viral-induced autophagy actually delays caspase-dependent cell death, but infected cells are ultimately overwhelmed by viral replication. Inducers of autophagy may thus limit the pathogenesis of acute CHIKV infection, and may be useful and important for irreplaceable cell types such as neurones (see below).

CHIKV infection rapidly results in the induction of type I interferons (IFNs) (IFN-α by typically leucocytes and IFN-β by fibroblasts) (Stetson and Medzhitov, 2006; Sourisseau et al., 2007; Her et al., 2010; Schilte et al., 2010; Werneke et al., 2011), and production of pro-inflammatory cytokines (Ng et al., 2009; Broz and Monack, 2011) as part of the cellular innate immune response (Takeuchi and Akira, 2007; Yan and Chen, 2012). Historically, CHIKV has in fact been used to stimulate IFN production from chick embryo fibroblast-like cells in the early days of investigation on type 1 IFNs (Friedman, 1964; Wagner, 1964). In general, IFN-α/β from infected cells stimulate the IFN-α/β receptor (IFNAR), which through its associated Janus (JAK) kinases activate in turn the signal transducers and activators of transcription (STATs) 1 and 2. The STATs associate with IFN regulatory factors (IRFs) and bind to IFN-stimulated response elements (ISREs) upstream of a bunch of IFN-stimulated genes (ISGs) encoding antiviral effector molecules (Takaoka and Yanai, 2006). IFN-α/β is indeed effectively suppressive of in vitro growth of CHIKV (Sourisseau et al., 2007) and other alphaviruses, and infection of mice lacking either IFNAR (Couderc et al., 2008; Schilte et al., 2010) or STAT1 (Schilte et al., 2010) by CHIKV causes heightened disease severity and invariable fatality.

There is some current controversy as to whether CHIKV infect circulating immune cells (particularly PBMCs) and induce IFN response in these (Her et al., 2010; Schilte et al., 2010), a point that shall be explore further below. In human fibroblasts, CHIKV infection was shown to activate IRF3 via an innate immune signalling pathway that includes the adaptor molecule interferon promoter stimulator 1 (IPS-1) (White et al., 2011). Interestingly, however, translation of the corresponding proteins and the IRF3-dependent antiviral genes is blocked by an evasion mechanism exhibited by CHIKV that causes widespread shutdown of cellular protein synthesis. One way whereby this could be achieved is through phosphorylation and inactivation of the eukaryotic initiation factor subunit 2α (eIF2α) by double-stranded RNA sensor protein kinase R (PKR) (Dauber and Wolff, 2009), as has been previously demonstrated for both SFV (McInerney et al., 2005) and Sindbis (Gorchakov et al., 2004). Interestingly, although CHIKV infection does trigger inactivation of eIF2α by PKR, this response is apparently not required for the block to protein synthesis of IRF3-dependent gene products (White et al., 2011).

Furthermore, there was also an apparent specific blocked of transcription of IRF3-dependent antiviral genes late in CHIKV infection (White et al., 2011). The non-structural protein 2 (nsP2) of alphaviruses are well known as important modulators of virus–host cell interactions (Frolova et al., 2002; Frolov et al., 2012). Most extensively studied in Sindbis virus, nsP2 is a cytopathic determinant and an inhibitor of cellular transcription, the latter is a major mechanism of viral silencing of the IFN response (Garmashova et al., 2006). SFV's nsP2 is nuclear localized, and elimination of its nuclear localization signal resulted in a significantly more robust IFN response in infected cells (Breakwell et al., 2007). CHIKV's nsP2 has likewise been recently shown to be a potent inhibitor of IFN-induced JAK-STAT signalling (Fros et al., 2010). Another recent report characterized the immunoregulatory role of the ubiquitin-like protein ISG15 in CHIKV infection using neonatal mice lacking ISG15, for which CHIKV infection is highly lethal (Werneke et al., 2011). As mice lacking UbE1L (the ISG15 E1 conjugating enzyme) and unable to form ISG15 conjugates displayed no increase in lethality following CHIKV infection, with ISG15 apparently acting in a non-classical manner. Although viral loads were similar between wild-type and ISG15−/− mice, the latter exhibited a dramatic increase in pro-inflammatory cytokines and chemokines which may contribute to their lethality. Although details are yet unclear, ISG15 apparently acts to reduce the pathological levels of effector molecules of the innate immune response elicited during viral infection.

The events and cellular responses that are triggered upon viral entry are summarized in Fig. 1. We turn next to two particularly controversial issues pertaining to CHIKV infection.

Figure 1.

A schematic diagram depicting cellular responses and events in a typical susceptible cell type (e.g. fibroblasts) upon CHIKV infection. Major cellular organelles involved are shown and major classes of physiological/pathological responses are highlighted in boxes. Note that viral stress-induced autophagy could be initiated from multiple cellular locations and only ER and TGN are illustrated here. The nature of the CHIKV receptor is not yet known (?). Also not shown in the diagram is the possibility of viral envelope protein activation of TLR4, as well as the activity of the double-stranded RNA sensor protein kinase R (PKR) on eukaryotic elongation factor 2 α (eIF2α) (see text). Cytokine receptors generally depicted here would include IL-1 receptors and TNF receptors. ER, endoplasmic reticulum; GA, Golgi apparatus; TGN, trans-Golgi network; Ω, omegasome; CHIKV-RC, CHIKV replicative complex; TLR, Toll-like receptor; MYD88, myeloid differentiation primary response gene 88; TRIF, TIR-domain-containing adapter-inducing interferon-β; cytc, cytochrome c; CARDIF, CARD adapter-inducing interferon β; RIG-1, retinoic acid-inducible gene 1; MDA5, melanoma differentiation-associated protein 5; IFNAR, interferon-α/β receptor; JAK2, Janus kinase 2; STAT, signal transducers and activators of transcription.

Controversy 1: How do immune cells feature in CHIKV infection?

Earlier work has indicated that most immune cells such as lymphocytes, monocytes and monocyte-derived dendritic cells either are not infected by CHIKV in vitro, or do not support a significant degree of productive CHIKV infection (Sourisseau et al., 2007). An exception to this is monocyte-derived macrophages, which did support viral infection to some degree. A more recent study by Schilte and colleagues also indicated that although high levels of IFN-α is detected in serum samples from La Réunion patients, no infection of haematopoietic cells were detected from localized injection of CHIKV into either wild-type or IFNR−/− mice. This is evidenced by three lines of findings, the first being that no CD45-positive cells were co-labelled with GFP-tagged CHIKV (Schilte et al., 2010).

IFN production in viral infected cells is triggered by the pattern-recognition receptors (PRRs) (Thompson et al., 2011), which recognize pathogen-associated molecular patterns (PAMPs) (Kumar et al., 2011), cellular sensors for viral glycoproteins, and nucleic acid replication products. Among the members of the Toll-like receptor TLR family, TLR3 senses extracellular dsRNA, while TLR7 and TLR8 recognizes ssRNA (Bowie and Unterholzner, 2008; Carty and Bowie, 2010). As CHIKV is an ssRNA virus and likely makes a dsRNA intermediate product during replication, engagement of TLR3 and TLR7 was expected. However, CHIKV was shown to not directly activate haematopoietic cells expressing these TLRs. On the other hand, the IFN response in infected mice appears to be dependent on the adaptor molecule CARD adapter-inducing interferon β (CARDIF/IPS-1/Visa/MAVS) (Hiscott et al., 2006), which acts downstream of intracellular sensors that include the RNA helicases melanoma differentiation-associated gene 5 (MDA5) and retinoic acid-inducible gene I (RIG-1) (Loo and Gale, 2011). CARDIF−/− mouse embryonic fibroblasts (MEFs) were more sensitive to CHIKV, and fail to mount a proper IFN response upon infection (Schilte et al., 2010). This observation appears to be in line with the notion that fibroblasts being the primary cell type infected and producing interferon during a CHIKV infection. However, RIG-I−/−, Mda5−/− and CARDIF−/− mice all had only subtle CHIKV infection phenotypes, with moderately higher viral titres at later time points, and none is close in terms of severity suffered by IFNAR−/− mice. In fact such a subtle phenotype is also observed in the TLR adaptor myeloid differential primary response protein 88 (MyD88) deficient mice. Although TLRs on haematopoietic cells appear not be involved in IFN response to CHIKV infection, there remains a possibility that endosomal TLRs in these cells are engaged as a result of them phagocytosing infected cells (Krejbich-Trotot et al., 2011b), or MyD88 might act through its role as adaptor for the IL-1β and IL-18 receptors (Schwartz and Albert, 2010).

A third line of evidence against active participation of haematopoietic cells during CHIKV infection came from experiments using bone marrow chimeric mice. When cells in the bone marrow of wild-type mice were lethally irradiated followed by adoptive transfer of IFNAR−/− bone marrow, the bone marrow chimeras were still capable of limiting CHIKV infection, indicating that the host's stromal cells are sufficient in this regard. Conversely, wild-type bone marrow was unable to confer a similar degree of CHIKV tolerance to IFNAR−/− mice. These data showed that IFNAR responsiveness in stromal cells, and not bone marrow-derived haematopoietic cells, is important for limiting CHIKV infection in mice. In this aspect, CHIKV is clearly different from other viruses such as reovirus, which is also more lethal towards IFNAR−/− mice compared with wild type, but which susceptibility phenotype could be rescued by bone marrow transplant from wild-type mice (Johansson et al., 2007).

As CHIKV infection resulted in very high viral load, that blood cells do not contribute significantly to viraemia and viral dissemination appears counterintuitive, and not in line with that observed for some other arboviruses. In another report which provided an almost diametrically opposite conclusion, Her and colleagues showed that monocytes from patients acutely infected with CHIKV could indeed harbour the virus (Her et al., 2010). In vitro infection experiments with whole blood or purified monocytes also indicate productive infection of monocytes and to lesser extents, B lymphocytes and myeloid dendritic cells. Further contrasting the results of Schilte et al. discussed above, a robust and rapid innate immune response high levels of IFN-α were produced rapidly after CHIKV incubation with monocytes. This swift IFN response may explain the rapid control of CHIKV replication in the blood. As a result, there is a high degree of lymphopenia, and extensive generation of virions by monocytes was also not observed.

The discrepancy between the reports highlighted above is difficult to reconcile. It is possible that the rapid and robust IFN response clear low viral load infections too quickly for easy observations of monocyte infection. A similar sensitivity to IFN precludes the spread of the neurotrophic SFV, another Alphavirus in the same antigenic complex as CHIKV, to extraneural tissues beyond the CNS (Fazakerley et al., 2006), and the extraneural virulence exhibited a mouse-adapted SFV strain is basically due to a reduction in susceptibility to IFN (Deuber and Pavlovic, 2007). It should be noted that Her et al.'s observations of monocytes infection in patient blood is made with early acute phase samples and the in vitro infections were done with high multiplicity of infection (moi), although the authors have checked that monocyte infection still occurs at the moi used by the other authors.

Although the glaring difference between the above results could not be easily explained, that monocytes could be productively infected during the early phase of CHIKV infection has important implications. The reason for recurrent joint pain in some CHIKV infected patients has been unclear as the virus can no longer be detected in the blood. However, in a recently reported primate CHIKV model in macaques, long-term CHIKV infection with extensive mononuclear cell invasion of infected tissues was observed in joints, muscles, lymphoid organs and liver. Macrophages are apparently the main cellular reservoirs during the late stages of CHIKV infection in vivo (Labadie et al., 2010). Infected blood monocytes may therefore disseminate the virus to sanctuaries sites supporting persistent viral replication in the chronic phase of the disease. The related Alphavirus RRV has been shown to persistently infect activated macrophages in culture (Lidbury et al., 2011), and this may well be the case for CHIKV. Detection of CHIKV (RNA and proteins) was in fact reported in perivascular synovial macrophages of one patient 18 months post-infection (Hoarau et al., 2010). IFN-α as well as IL-10 are expressed, but not pro-inflammatory cytokines such as TNF-α and IL1β, indicating a chronic immune distinct from rheumatoid arthritis (Jaffar-Bandjee et al., 2009; 2010; Hoarau et al., 2010).

Controversy 2: CHIKV infection of CNS cell types

Clinically documented encephalopathy among the neonate and elderly of re-emerging CHIKV infection epidemics brings forth the possibility of substantial CHIKV neurotropism. Traditionally, unlike members of the Flavivirus genus, West Nile virus and Japanese Encephalitis virus, which are obviously encephalitogenic, CHIKV has not been considered a neurotrophic virus. Even within the Alphavirus genus, CHIKV is classified under the Old World arthritogenic and not the encephalitogenic type, with the latter type primarily including the New World Alphaviruses [such as the Western equine encephalitis virus (WEEV)]. However, neurological symptoms had indeed been associated with reports of CHIKV epidemic since the 1960s and 1070s (Arpino et al., 2009). In the La Réunion epidemic, encephalitis and meningoencephalitis were two of the major causes of death among patients with severe ‘atypical’ CHIKV fever (Economopoulou et al., 2009). Encephalopathy is particularly prevalent among newborns infected via mother-to-infant transmission (Ramful et al., 2007; Gérardin et al., 2008; Robin et al., 2008), and there are also cases of Chikungunya fever patients presented with peripheral neuropathy and Guillain–Barré syndrome (Wielanek et al., 2007; Lebrun et al., 2009).

Early studies have shown that CHIKV could readily infect mouse brain cells and also productively replicate in primary brain cell mix cultures (Chatterjee and Sarkar, 1965; Precious et al., 1974), and intracerebral injection in newborn mice can apparently be used to amplify the virus (Schuffenecker et al., 2006). In the above, however, productive CHIKV infection of neurones and glia was not particularly well characterized. Different CHIKV mouse infection models produce varying and somewhat conflicting degree of severity in terms of CNS infection. In the work of Ziegler et al., where newborn and 14-day-old mice were inoculated subcutaneously with CHIKV, major histopathological changes are associated with muscle and skin, while brain tissue are not infected (Ziegler et al., 2008). In another model by Couderc et al., severe CHIKV infection inoculated intradermally into IFNAR−/− mouse disseminated to the choroid plexus [which form the blood-cerebrospinal fluid (CSF) barrier], ependymal walls and leptomeninges (Couderc et al., 2008). However, microvascular endothelial cells that constitute the blood–brain barrier (BBB) were not infected. The brain parenchyma was also not infected and there was no CHIKV immunolabelling of microglial and astrocytes. More severe brain infection by CHIKV was shown in another report where a mouse-adapted virulent CHIKV Ross strain was inoculated intra nasally (Wang et al., 2008). The virus infected the brain, resulting in multifocal inflammation and necrosis in the cerebral cortex. Neurones with CHIKV-positive cytoplasmic staining were observed in the cerebral cortex by day 3 after infection. By day 7 after infection, the number of CHIKV-positive neurones increased, and were primarily located in the necrotic areas of the cerebral cortex and hippocampus. In preliminary results mentioned in the authors’ review article, Das et al. infected mouse brain mixed cultures with CHIKV at low moi. Both βIII-tubulin-positive neurones and glial fibrilliary acidic protein (GFAP)-positive astrocytes appeared able to efficiently support viral replication (Das et al., 2010).

From the findings highlighted above, it appears that a typical CHIKV infection of otherwise healthy would normally not result in dissemination of the virus to the CNS. One reason for this, namely the efficiency of the IFN response in clearing the infection, has been discussed above. CHIKV infection of the CNS could be a problem in neonates if the inoculating dose is high and when the innate immune response could not effective clear viraemia in time. CNS infection would be a concern, as CNS cell types (neurones, astroglia and microglia) are all likely to induce an innate immune response (Ransohoff and Brown, 2012) as do fibroblast and other cell types. A robust expression of ISG15 by an astrocyte cell line in response to CHIKV infection was in fact noted (Das et al., 2010). On one hand, mechanisms that favour non-cytolytic viral clearance may preserve postmitotic neurones that could not be replenished or neuronal connections that could not be re-established may promote latent viral infection. On the other, there is a danger that an associated pro-inflammatory response would have undesirable consequences of neuroinflammation that has been seen for other viruses (Peterson and Du, 2009), even for those that are not obviously encephalitogenic (Jurgens et al., 2012). As the human brain has a particularly lengthy period of postnatal development (Johnson, 2001), any sort of neuroinflammatory damage to neonatal brain regions could affect cognitive and psychological functions later in life (Cohly and Panja, 2005; Hagberg and Mallard, 2005). It has been speculated that CHIKV could target progenitor and stem cells in the neurogenic regions of the brain such as the subventricular zone (SVZ) (Zhao et al., 2008), thus affecting postnatal and adult neurogenesis (Das et al., 2010). Although cases of this nature have not yet been conclusively reported, there is a need to be wary of their potential emergence.

Epilogue

The re-emergence of CHIKV and its spread to developed countries has heightened public attention and intensified interest in public health agencies in aspects of clinical transmission and epidemiology. The preceding paragraphs highlighted more basic aspects of CHIKV infection, namely its cellular host range and responses. Through investigations with both in vitro and in vivo models of infection, new understandings of the basic biology of the virus have emerged. Although typically eliciting a mild disease because of effective viral clearance by the innate immune response, there is a need to understand the complications associated with CHIKV infection. In the acute phase this takes the form of significant rate of encephalitis particularly among infants and young children, while in the convalescence phase this is reflected by recurrent joint pain in the absence of viraemia. The former may be consequential to CHIKV being encephalitogenic, although not obviously neurotrophic, while the latter may reflect CHIKV's previously underappreciated ability to persistently infect monocytes and its differentiated macrophagic forms. Further investigations and understandings of CHIKV's interaction with various human cell types would be helpful for its effective control and eradication.

Acknowledgements

Work is supported by incentive funding from the NUS Graduate School of Integrative Sciences and Engineering. B.L.T. is grateful to the constructive comments and criticisms of all three reviewers, which improved the manuscript.

Ancillary