Background
Chikungunya virus (CHIKV) is a mosquito-transmitted arbovirus belonging to the alphavirus genus of the
Togaviridae family [
1]. The major vectors of CHIKV are
Aedes aegypti and
Aedes albopictus. Importantly, CHIKV is the etiologic agent of chikungunya fever (CHIKF), a rheumatic-like disease typically characterized by high fever, prolonged polyarthralgia, myalgia, rash and sometimes death [
2‐
4]. However, to date no effective vaccine or specific therapeutic is available to prevent or treat CHIKV infection [
5].
CHIKV is an enveloped, spherical, positive sense, and single stranded RNA virus. The genome size of CHIKV is approximately 11.8Kb containing a 5′-methylguanylate cap and a 3′-polyadenylate tail as well as two open reading frames (ORFs). The first ORF encodes for four nonstructural proteins (nsP1 to nsP4), while the second ORF encodes for three structural proteins (C, E1 and E2) and two small peptides (E3 and 6 K) [
6,
7]. Based on the E2 gene sequence, CHIKV is classified into four CHIKV lineages, including the West African (WA) lineage, the East/Central/South lineage (ECSA), the Asian, and the Indian Ocean lineage (IOL) [
8,
9]. The IOL lineage was first distinguished from the ECSA lineage during an outbreak on the island of La Reunion in 2005–2006 [
5]. The Asian lineage originated in Africa and experienced independent evolution for several centuries before its first outbreak in 1958 in Asia [
10]. Since 2013, the Asian lineage has caused several epidemics in the Pacific islands and Americas [
11,
12].
To date, the cell surface receptors for CHIKV in both mosquito cells and vertebrate cells remain incompletely understood [
7]. Thus, to better understand the pathology of CHIKV infection, it is very important to confirm the cell types that CHIKV can attach to and productively infect. Previous studies found that different CHIKV lineages showed different cell tropisms in vitro [
13] and pathogenesis in vivo [
3]. For example, the
Aedes albopictus cell line C6/36 was found to be significantly more permissive to the recently prevalent CHIKV isolates of the ECSA lineage than the original ROSS strain [
13]. In another study,
Aedes albopictus showed a higher disseminated infection and a more rapid transmission of the IOL lineage sooner after ingesting viral blood meal, while
Aedes aegypti displayed a more severe infection and more rapid transmission of the Asian lineage after viral blood meal infection [
12]. Suckling mice infected with a CHIKV strain of the Asian lineage showed a lower weight gain and higher mortality than mice infected with a strain of the ECSA lineage after intra-cerebral inoculation, despite displaying similar viral load in the brains [
14]. Further gene expression studies found that the higher mortality caused by the Asian lineage was due to a differential gene expression profile involved in host immune response [
14]. However, studies that compared the differences between the Asian lineage and the IOL lineage on cell susceptibility in mammalian and mosquito cell lines are limited.
In our previous study, two virus strains, SZ1050 and SZ1239, were successfully isolated from human serum samples using C6/36 cells. SZ1050 was isolated in 2010 and was from a patient returned from India [
15]. SZ1239 was isolated in 2012 from a female traveler who had visited Indonesia [
16]. Here, we cultured these two strains with BHK-21 cells and sequenced their whole viral genomes. Phylogenetic analysis indicated that SZ1050 belonged to the IOL lineage while SZ1239 was a strain of the Asian lineage. Next, we inoculated these two virus strains into a range of cells lines derived from different tissues of various hosts, including 293 (human embryonic kidney), HepG2 (human hepatocarcinoma), RD (human Rhabdomyosarcoma), HeLa (human cervical epithelial), THP-1 (peripheral blood monocytes from monocytic leukemia), K562 (human erythroleukemia), U937 (human histiocytic lymphoma), Ana-1 (the murinal celiac macrophage), BHK-21 (baby hamster kidney, fibroblast), MDCK (dog kidney epithelial), Vero (African Green Monkey Kidney), C6/36 (
Aedes albopictus), and Aag-2 (
Aedes aegypti). The viral RNA loads in the supernatant and cell lysate were evaluated.
Discussion
From the 1960s to1980s, CHIKV outbreaks were limited to Africa and Asia. In 2004, it re-emerged in Kenya and rapidly spread to several islands in the Indian Ocean as well as many other regions, including South Asia, Central and West Africa, Europe, the Caribbean and Central, South and North America served to refocus global attention to this virus [
25]. Characterizing the whole viral genome and cell tropism contributes to better understand the pathogenesis and vector competence of CHIKV.
Based on the sequence analysis, we found that SZ1050 was phylogenetically most related to GZ1029, another reported CHIKV strain of IOL lineage isolated from an imported foreign case travelled from India to China in 2010 [
19]. This suggested that SZ1050 was potentially a circulating CHIKV strain in India in 2010. On the other hand, SZ1239 was indicated to belong to the Asian lineage. The highest genetic identity was found between SZ1239 and another strain DH130003 of the Asian lineage, which was isolated in 2012 from a patient who traveled back to Bali from Indonesia [
20]. CHIKV from the Asian lineage was the major causative agent for the increased CHIKV-infected cases in Indonesia in 2008, 2009 and 2011 [
23]. Therefore, the SZ1239 strain also represented a recently circulating isolate of the Asian lineage. More recently, CHIKV of Asian lineage was also found in the pacific region and America [
22,
25], indicating the Asian lineage as the major prevalent genotype of the current CHIKV outbreak.
Of note, the first RSE in S27 was deleted in the 3’ UTR of SZ1239. Similar gene gap was also found in other three CHIKV isolates from Indonesia (JMB-154, DH130003 and JMB-230), suggesting the absence of the first RSE might be due to an evolution process of CHIKV. It was reported that RSE regulated viral RNA synthesis [
7]. A deletion of RSEs in model alphaviruses might affect the interaction of unknown cellular proteins involved in virus production and/or tissue specificity and leads to a reduced and delayed viral release in different cell types [
26]. The consequence of the deletion of the first RSE in SZ1239 is currently unknown.
Compared to the amino acid sequences of S27, most variations in both SZ1050 and SZ1239 were observed in the nsP3 and the E2 protein (Table
1) in line with other studies [
17]. The mutations in nsP3 were focused in 326–524 aa, which was a variable region [
27,
28]. A gap of seven amino acids (1710–1716 in S27, 377–383 aa in the nsP3) located in the nsP3 of SZ1239 compared to S27 (Table
1). This gap should not be an occasional deletion caused by viral culture in vitro because it was also observed in an isolate in 2006 (MY/06/37350,GenBank: FN295484, Malaysia) and many recent circulating isolates of the Asian lineage such as NC/2011–568 (GenBank: HE806461.1, 2011, New Caledonia) [
29], DH130003 (GenBank: KM673291.1, 2013, Indonesia) [
20] and JMB-154 (GenBank: KX097982.1, 2015, Indonesia) [
30]. More interestingly, compared to S27, there was a small gap of four amino acids (LPSA, 1712–1715 in S27) in the middle of this seven amino acid-gap in the nsP3 of some CHIKV isolates found in Micronesia: Yap State in 2013 (strain 3807,GenBank: KJ451622.1, 2013) [
31] and America (isolate 14.02217, GenBank: KY435477.1, 2014) [
22]. However, there was no gap for the CHIKV isolates identified in Malaysia in 2007 or before (strain MY002IMR/06/BP, GenBank: EU703759.1; strain MY003IMR/06/BP, GenBank: EU703760.1; strain M125, GenBank: KM923917.1). This suggested that the four amino acids (1712–1715) might play a key role in the evolution of CHIKV.
It was reported that the C-terminal hypervariable domain of nsP3 (398–406 aa) prevented stress granule formation through sequestration of GTPase-activating protein (SH3 domain)-binding proteins (G3BPs) during the mammalian stress response [
32]. Depletion of G3BPs caused severely reduced levels of negative-stranded (and consequently also positive-stranded) RNA [
32]. Fross et al. also identified the hypervariable C-terminal domain (475–501 aa) of nsP3 as a critical factor for granular localization and sequestration of mosquito Rin (G3BP homologue Rin in live mosquitoes) [
33]. The 18 amino acid deletion in nsP3 (386–403 aa) in Sindbis virus strain AR86 affected neurovirulence in mice [
34]. Whether the gap (377–383 aa) in the nsP3 of SZ1239 would affect the interaction between G3BPs and nsP3 or contribute to the neurovirulence requires further investigation.
In addition, the E1-A226V mutation, which was reported to relate to the adaption in
Aedes albopictus [
35], was not detected in SZ1050 and SZ1239. In the previous two small-scale outbreaks in China in 2010, E1-226 V was observed in all four isolated CHIKV strains [
18]. Therefore, the possibility of the SZ1050- or SZ1239-induced CHIKV outbreak in China would be low because
Aedes albopictus is the major vector responsible for arbovirus transmission in most Chinese regions. Interestingly, another mutation, I211T, was found in E2 of both SZ1050 and SZ1239, which was also found in the West African lineage and Asian lineage [
10]. Virus containing E1-A226V and E2-I211T showed enhanced infectivity of CHIKV in
Aedes albopictus [
36]. Whether the single E2-I211T mutation could contribute to the transmission advantages of the virus in
Aedes remains largely unkown.
To further investigate the cell tropism of SZ1050 and SZ1239, we inoculated two viral isolates in thirteen cell lines. The viral RNA increasing speed in the first 24 h in RD, HepG2 and 293 was significantly higher than that in HeLa cells (Fig.
3d), which suggested that CHIKV virus might prefer to infect and replicate in human liver, kidney and muscle rather than in the cervix. This could partially explain why many patients showed symptoms of myalgia and dysfunction of liver.
Beside adherent cells, suspension blood cells were also detected to compare their susceptibility to these two CHIKV strains. Our results showed that U937 cells were refractory to CHIKV infection in agreement with other reports [
5,
13,
24]. A previous study found that CHIKV was able to bind to a cell membrane protein–prohibitins in U937 cells but could not replicate in them [
37]. The detailed mechanism is unclear. Similarly, our result showed that Ana-1 was refractory to infection by SZ1050 from the IOL lineage. In addition, we also infected PMA-stimulated THP-1 with SZ1050 and SZ1239. Although SZ1239 showed higher viral RNA copies in the supernatant than that of SZ1050, it was hard to confirm that it could effectively replicate in THP-1 due to the minimal viral RNA load changes between time points. It was reported that CHIKV (La Reunion isolate of IOL lineage) could neither bind THP-1 at 4 °C nor produce infective viruses at 37 °C [
24]. More studies should be performed to investigate the cell susceptibility for CHIKV infection in unstimulated THP-1 and PMA-stimulated THP-1 cells.
As CHIKV is an arbovirus, it is very important to evaluate the virus tropism in the mosquito vectors. In this study, Aedes aegypti cell line Aag-2 and Aedes albopictus cell line C6/36 were selected to serve as two mosquito cell models. Our results suggested that both cell lines were susceptible to these two strains. Of note, in our study, Aag-2 cells were very susceptible to both CHIKV isolates but had no CPE after infection.
This was not consistent with the findings of another study [
13]. In that study, the authors found chikungunya virus of the ECSA lineage not able to effectively infect the
Aedes aegypti cell line CCL-125 cells [
13]. Possible explanation might come from the differences on the sources and contaminated pathogens between the two
Aedes aegypti cell lines (Aag-2 and CCL-125). On one hand, Aag-2 was derived from embryonic
A. aegypti while CCL-125 was larvae originated; on the other hand, Aag-2 was contaminated with Phasi charoen-like virus (PCLV) and Cell-fusing agent virus (CFAV), while CCL-125 was onlyinfected with PCLV [
38]. The single infection of CFAV was believed to promote the infection of DENV [
39]. Therefore, it is possible that CFAV in Aag-2 can modulate the infection of CHIKV.