Conceived and designed the experiments: CYL, JF, BM, LR, LFPN, LW. Performed the experiments: CYL, YWK, JF, BM, EGLK, WH, LW. Analyzed the data: CYL, YWK, JF, BM, EGLK, CP, WWLL, LR, CIW, LFPN, LW. Contributed reagents/materials/analysis tools: CL, RTPL. Wrote the paper: LW.
I have read the journal's policy and have the following conflicts; Chia Yin LEE and Lucile WARTER are inventors in a pending patent.
Chikungunya virus (CHIKV) is an alphavirus responsible for numerous epidemics throughout Africa and Asia, causing infectious arthritis and reportedly linked with fatal infections in newborns and elderly. Previous studies in animal models indicate that humoral immunity can protect against CHIKV infection, but despite the potential efficacy of B-cell-driven intervention strategies, there are no virus-specific vaccines or therapies currently available. In addition, CHIKV has been reported to elicit long-lasting virus-specific IgM in humans, and to establish long-term persistence in non-human primates, suggesting that the virus might evade immune defenses to establish chronic infections in man. However, the mechanisms of immune evasion potentially employed by CHIKV remain uncharacterized. We previously described two human monoclonal antibodies that potently neutralize CHIKV infection. In the current report, we have characterized CHIKV mutants that escape antibody-dependent neutralization to identify the CHIKV E2 domain B and fusion loop “groove” as the primary determinants of CHIKV interaction with these antibodies. Furthermore, for the first time, we have also demonstrated direct CHIKV cell-to-cell transmission, as a mechanism that involves the E2 domain A and that is associated with viral resistance to antibody-dependent neutralization. Identification of CHIKV sub-domains that are associated with human protective immunity, will pave the way for the development of CHIKV-specific sub-domain vaccination strategies. Moreover, the clear demonstration of CHIKV cell-to-cell transmission and its possible role in the establishment of CHIKV persistence, will also inform the development of future anti-viral interventions. These data shed new light on CHIKV-host interactions that will help to combat human CHIKV infection and inform future studies of CHIKV pathogenesis.
Chikungunya virus (CHIKV) is transmitted by mosquito bites and causes a febrile disease that is often characterized by persistent joint pain. Until recently, CHIKV outbreaks were limited to tropical areas of Africa and Asia. However, since 2007, following a large CHIKV epidemic in the Indian Ocean and South-East Asia, CHIKV has also been reported in temperate European regions. As mosquito habitats expand, virus dissemination may become more prevalent, but there are currently no vaccines or CHIKV-specific treatments available. We previously described two human antibodies that potently block cellular CHIKV infection. In the current report, we have characterized CHIKV mutants that escape neutralization to identify sub-domains of the virus envelope which are involved in CHIKV interaction with these antibodies, thereby opening the door for the development of CHIKV-specific sub-domain vaccination strategies. For the first time, we have also demonstrated that CHIKV can be directly transmitted between cells, bypassing transport through the extra-cellular space. This mode of dissemination, which is associated with viral resistance to antibody neutralization, may play a critical role in the establishment of persistent CHIKV infection. Together, these findings will aid the design of new strategies to combat CHIKV infection and will inform future studies of CHIKV pathogenesis.
Chikungunya virus (CHIKV) belongs to the
While CHIKV infection in humans is often associated with only mild clinical symptoms that resolve over 1–2 weeks
CHIKV exhibits a positive strand RNA genome that encodes 4 non-structural proteins (NSP1–4) and 5 structural proteins: the capsid (C), the E1, E2, and E3 envelope glycoproteins (E2 and E3 are initially synthesized as a single precursor molecule, p62, which is subsequently cleaved), and a small polypeptide molecule, 6K
The structure of the E1 protein in alphaviruses has previously been determined using the representative Semliki Forest virus family member
Human antibodies isolated from the plasma of a CHIKV convalescent patient have previously been shown to both prevent and cure CHIKV infection in mice
Viruses can escape neutralizing antibody responses by undergoing genetic mutations that abolish antibody binding, or by indirect evasion strategies such as cell-to-cell transmission. It is currently unclear whether CHIKV is capable of exploiting these strategies to persist in human hosts.
In this study, we aimed to characterize CHIKV antigens targeted by neutralizing human antibodies to inform the subsequent design of CHIKV-specific sub-domain vaccination strategies. We also sought to identify potential mechanisms of immune evasion that CHIKV might exploit to establish persistent infections in man. We recently identified two human monoclonal antibodies (mAb), designated 5F10 and 8B10, which broadly and potently neutralize CHIKV
We recently identified two human mAb designated 5F10 and 8B10, which broadly and potently neutralize CHIKV
CHIK/Irr, CHIK/5F, CHIK/8B and CHIK/5F+8B indicate CHIKV rescued from 8 serial cell passages under the continuous mAb pressure of Irr.IgG1, 5F10, 8B10 and 5F10+8B10, respectively. (
To investigate whether CHIKV resistance to mAb 5F10 and/or 8B10 was associated with specific mutation(s), viral RNA was isolated from each CHIKV mutant and reverse-transcribed into cDNA for sequencing. When compared with the CHIK/Irr control genome, we identified two nucleotide (nt) substitutions within the CHIK/5F genome that resulted in amino acid (aa) changes at positions 82 and 216 in the E2 protein (E2.R82G and E2.V216E). In the CHIK/8B genome, we identified one nt substitution that resulted in an aa change at position 101 in the E1 protein (E1.T101M), and also observed 1 mix of wild-type (wt)/mutated nt, with the mutated nt leading to one
CHIKVs | E1 |
E2 |
||
101 | 12 | 82 | 216 | |
CHIK/Irr | T | T | R | V |
CHIK/5F | T | T |
|
|
CHIK/8B |
|
|
R | V |
CHIK/5F+8B | T | T |
|
V |
Numbers refer to the aa positions within the E1 and E2 CHIKV proteins, respectively. The aa variations associated with the CHIKV-specific mAb are highlighted in bold.
Intriguingly, the mutated nts associated with the residue substitutions E1.T101M, E2.T12I, E2.R82G and E2.V216E were also detectable in the polyclonal CHIK/Irr control cDNA preparation (
aa substitution |
nt substitution |
number of reads |
nt (%) |
|||
A | T | C | G | |||
E2.T12I | C1010T | 7907 | 0.10 |
|
99.71 | 0.11 |
E2.R82G | A1219G | 7826 | 99.58 | 0.04 | 0.18 |
|
E2.V216E | T1622A | 7840 |
|
99.55 | 0.31 | 0.09 |
E1.T101M | C2729T | 7999 | 0.14 |
|
99.76 | 0.01 |
Numbers refer to the aa position within the indicated protein, E1 or E2.
Nt substitutions associated with aa substitutions indicated in a.
Numbers refer to the nt position within the C-E1 encoding sequence.
Indicates the number of sequenced viral cDNA copies, as performed by Next Generation Sequencing.
Indicates the percentage of each nt identified at each position.
To investigate the roles of the different CHIKV mutations in mediating resistance to mAb 5F10 and 8B10, we isolated CHIKV clones that exhibited either single or dual mutations and then further probed their sensitivity to mAb-dependent neutralization.
CHIKV populations were cultured under agarose-medium before 6 individual CHIK/5F, CHIK/8B, or CHIK/5F+8B colonies (or 2 CHIK/Irr control colonies) were isolated for amplification and sequencing. The sequencing data from the plaque-purified CHIKV clones are shown in
Clonal CHIKV | E1 |
E2 |
||||
101 | 396 | 427 | 12 | 82 | 216 | |
CHIK/5F #1,3, 4, 5 and 6 | T | T | I | T |
|
|
CHIK/5F #2 | T |
|
|
T |
|
|
CHIK/8B #1,3, 4, 5 and 6 |
|
T | I |
|
R | V |
CHIK/8B #2 |
|
T | I | T | R | V |
CHIK/5F+8B #1-6 | T | T | I | T |
|
V |
CHIK/Irr #1-2 | T | T | I | T | R | V |
Numbers refer to the
In PRNT assays, mAb 5F10 neutralized 5F/E2.R82G+V216E and 5F/E1.TMm+E2.R82G+V216E far less efficiently than CHIKwt (
CHIKV variants have been labelled according to their residue substitutions and corresponding mAb added to the culture medium during the serial passages. (
In parallel, mAb 8B10 neutralized 8B/E1.T101M less efficiently than CHIKwt (
Interestingly, the 8B10-derived CHIKV mutants remained efficiently neutralized by mAb 5F10, and
As the two mutations in the E1 trans-membrane domain were not associated with resistance to mAb-dependent neutralization, 5F/E1.TMm+E2.R82G+V216E was excluded from further analyses. Although the E2.R82G mutation was not found to be associated with significant resistance to mAb-dependent neutralization in PRNT assay, as this was the only mutation to be selected under dual treatment with mAb 5F10 and 8B10 (
To clarify the mechanism(s) associated with the neutralization escape mutations, we next analyzed the capacity of mAb 5F10 and 8B10 to bind to clonal CHIKVs (
(
Binding tests performed on CHIKV-infected cells (
Taken together, these data demonstrate that mutations E2.V216E and E1.T101M abolish the binding of mAb 5F10 and 8B10 to CHIKV, respectively.
To further clarify the CHIKV domains involved in virus interaction with mAb 5F10 and 8B10, we used Chimera software to locate residues E1.101, E2.12, E2.82 and E2.216 within the CHIKV E1/E2 heterodimer for which the crystal structure was recently resolved under neutral pH conditions
Based on structural data retrieved from protein database records, the 3D organization of E1 (pale yellow), E2 (white), and the E1 fusion loop (pink) are shown for 3N44 under neutral pH conditions. (
The E2.82 residue is exposed at the surface of the E2 domain A (
The E1.101 residue is located within the E1 domain II next to the fusion loop (
The E2.12 residue, which is involved in resistance to 8B10-dependent neutralization together with E1.101 (
Noteworthy, under acidic pH conditions, the alphavirus domain B becomes disordered and releases the fusion loop
Indicated residues were located based on structural data retrieved from protein data base records (3MUU, under acidic pH conditions).
Taken together, these structural data strongly suggest that the 5F10 epitope is located at the tip of the CHIKV E2 domain B which contains the E2.216 residue. While the 8B10 target antigen remains somewhat ambiguous, our data suggest that this antibody may recognize a transitional epitope closely associated with the CHIKV fusion peptide.
The alignment of available CHIKV E1 and E2 protein sequences obtained from GeneBank (data not shown) indicated that all published CHIKV variants contain the residues E1.T101, E2.T12, E2.V216 and E2.G82, except for one strain: TSI-GSD-218, which contains E2.I12 and E2.R82. Thus, among the mutations selected here, E1.T101M, E2.T12I and E2.V216E clearly modify highly conserved CHIKV residues, while E2.R82G instead restores a highly conserved viral residue. These data indicate that the CHIKV11 isolate used in the current study is atypical with regards to the E2.82 residue, and suggest an important role for the E1.T101, E2.T12, E2.G82 and E2.V216 residues in the CHIKV life cycle. Therefore, the substitution of these conserved residues was expected to modify CHIKV fitness, and thus, we next investigated fitness characteristics of 5F10/8B10-resistant CHIKV escape mutants.
We first assessed the
(
We next investigated the
Mice were next infected with either CHIKwt or 5F+8B/E2.R82G and viral load was determined in serum and liver at 48 h post-infection (
Since measuring the relative size of virus-induced plaques is commonly performed to monitor viral cell-to-cell transfer
Direct cell-to-cell transmission was previously proposed to occur during CHIKV infection
CHIKV-infected HEK293T cells (producer cells) were co-cultured with CFSE-labeled naïve HEK293T cells (target cells) in the absence or presence of mAb 8B10. (
To further confirm direct cell-to-cell transmission of CHIKV, some co-cultures were visualized by confocal microscopy. For both CHIKwt and 5F+8B/E2.R82G, in the absence of mAb 8B10, CHIKV staining was detected uniformly on the cell surface, whereas, under 8B10 neutralizing pressure, CHIKV staining was strongly polarized and virus was often detected in areas of cell-cell contact (
CHIKV-infected HEK293T cells were co-cultured with CFSE-labeled naïve HEK293T cells (green) in the absence or presence of mAb 8B10. After 16 h of co-culture the cells were fixed, permeabilized, and stained for alphavirus expression (Alexa 647, red). Newly infected target cells appear orange. Magnification: ×40 or ×80.
We previously demonstrated potent
Escape mutants have previously been described for three members of the alphavirus family; Sindbis virus
Our results strongly suggest that the 5F10 mAb epitope is located at the tip of the CHIKV E2 domain B, while 8B10 might recognize a transitional epitope, close to the fusion loop, which is likely to be exposed under acidic pH conditions. Interestingly, in addition to the disordering of the alphavirus envelope at acidic pH, virus binding to a host cellular receptor is believed to induce pH-independent shifting of the E2 domain B, leading to exposure of transitional epitopes prior to virion internalization
Based on their proposed epitope specificity, we speculate that 5F10 and 8B10 inhibit viral entry and fusion to the cell membrane, respectively. However, it has also been suggested that antibodies which target the E2 domain B might also affect the viral-cell fusion step, possibly by inhibiting domain disordering and fusion loop exposure
Although RNA viruses are usually prone to high nucleotide sequence evolution due to their lack of a proofreading polymerase
After 8 cycles of neutralization/amplification, only partial CHIKV resistance to mAb 8B10 or 5F10+8B10 was observed. We then performed 5 additional rounds of neutralization/amplification while using increasing concentrations of mAb. However, even after 13 neutralization/amplification rounds under mAb selective pressure, we did not manage to select a CHIKV population fully resistant to 8B10 or 5F10+8B10 (data not shown). It is likely therefore that the mutation(s) required to fully escape 8B10 and 5F10+8B10 mAb give(s) rise to viruses that are unable to replicate robustly.
Cell-to-cell virus transmission is faster than extra-cellular spreading and enables viruses to evade the immune response
We demonstrate in this report, for the first time, CHIKV direct cell-to-cell transmission, and further show that this mode of dissemination is enhanced by the E2.R82G mutation. Interestingly, as the alignment of available CHIKV E2 protein sequences obtained from GeneBank (data not shown) revealed that the majority of CHIKV variants contain the residue E2.G82 (and not E2.R82), we speculate that direct cell-to-cell transmission is commonly used by CHIKV to disseminate in the presence of extra-cellular neutralizing antibodies. However, this remains to be shown with CHIKV isolates containing the residue E2.G82.
Viruses have evolved various mechanisms to disseminate from cell to cell. Pre-existing cell-cell contacts may be exploited, or virus-induced new contacts can be established between infected and uninfected target cells.
The fact that CHIKV cell-to-cell transmission is enhanced by E2.R82G, suggests the involvement of the CHIKV E2 domain A in this mode of dissemination. Interestingly, the E2 domain A of both Venezuelan Equine Encephalitis virus and Sindbis virus has previously been shown to contain residues that are important for virus binding to cells, notably to heparan sulfate located at the cell surface
For the first time, we have identified CHIKV envelope domains that are recognized by human neutralizing immune responses, and we have been able to demonstrate direct cell-to-cell transmission of CHIKV. This mode of dissemination, which protects CHIKV from neutralizing host antibodies, might play an important role in establishment of CHIKV persistence. These findings advance our understanding of CHIKV-human host interactions and will aid the rational design of future domain-based vaccines against CHIKV, as well as inform further studies of CHIKV pathogenesis.
This study was carried out in strict accordance with the guidelines of the Agri-Food and Veterinary Authority (AVA) and the National Advisory Committee for Laboratory Animal Research (NACLAR) of Singapore. All animal procedures were approved by the Institutional Animal Care and Use Committee (IACUC) of Biological Research Center, Biomedical Sciences Institutes, A*STAR, Singapore (IACUC number: #100515).
Vero cells (ATTC CCL-81) and HEK293T cells (ATCC CRL-N268) were cultured in DMEM-10% FCS (Gibco-Invitrogen). The CHIKV-neutralizing human mAb 5F10 and 8B10, the irrelevant human mAb HA4 (referred as Irr.IgG1), and the CHIKV isolate CHK/Singapore/11/2008 (referred as CHIKV11) have been described previously
CHIKV11 (200 PFU) was incubated for 1 h at 37°C with 100 ng/ml 5F10 and/or 8B10 in DMEM-10% FCS. HEK293T cells were then incubated at 37°C for 1.5 h with mAb/CHIKV mixtures, prior to being further cultured in DMEM-10% FCS supplemented with additional mAb as before for 2 days. Cell supernatants were then collected and their infectious viral titer was determined by Plaque Assay; 200 PFU of rescued virus was subjected to a second round of neutralization/amplification in the presence of the same antibodies as employed in previous steps. Eight neutralization/amplification rounds were performed in total.
Vero cells were infected for 1.5 h with CHIKV (10 PFU/well) before being washed with PBS and cultured in DMEM-0.25% agarose for 2 days. Individual CHIKV colonies were then selected through the agarose layer and amplified separately in Vero cells. Plaque-purified clonal CHIKV genomes were then sequenced.
Viral RNA was extracted from 140 µl supernatant from CHIKV infected-cells using the QIAamp Viral RNA Mini kit (Qiagen). For each viral RNA, two independent full-length cDNAs were synthesized using random hexamers and SuperScript III First-Strand kit (both from Invitrogen). Purified cDNAs were PCR-amplified using Taq PCR Master Mix Kit (Qiagen) and several primer pairs designed to cover the C-E1 CHIKV polyprotein encoding sequence and to generate ∼1000 bp-long overlapping PCR fragments. PCR fragments were sequenced (Aitbiotech) and results were analyzed using Lasergene 7 software.
A 200 bp cDNA library was synthesized from 56 ng of extracted CHIK/Irr RNA using the mRNA-seq Sample Prep Kit (Illumina) according to the manufacturer's instructions. The cDNA library was then sequenced using the Illumina GAIIX genome analyzer (Next Generation Sequencing Core facility, Genomic Institute of Singapore) at the coverage of 67436×. Unique reads were subsequently aligned with the consensus sequence encoding CHIKV11 structural proteins (C-E1) using the Burrows-Wheeler aligner, and site-specific nucleotide frequencies were determined using SAMtools Pileup.
The Plaque Reduction Neutralization test and determination of mAb potency were performed as previously described
Immunofluorescence Assay: HEK293T cells were infected with CHIKV and then fixed as previously described
In some experiments, 96-well plates were coated with UV-inactivated plaque-purified CHIKVs (104 PFU/well) for analysis by ELISA. A range of mAb concentrations (0.1 ng-10 µg/mL) were added to the wells for 1 h at RT. Bound mAb were detected using HRP-conjugated goat anti-human IgG (Jackson ImmunoResearch), followed by incubation with 3,3′,5,5′-tetramethybenzidine substrate (Sigma). The reaction was stopped by addition of HCl (1 M) and absorbance was measured at 450 nm using the TECAN Infinite M200 Monochromator Microplate Reader (TECAN).
Vero cells were infected with plaque-purified CHIKV mutants (MOI = 0.1) for 1.5 h prior to being cultured in DMEM-10% FCS. The number of PFU in each supernatant was determined by Plaque Assay at various times post-infection.
AGR129 mice (IFN-α/ß/γR−/−/− and RAG-2 deficient,
HEK293T cells were infected with CHIKV for 1.5 h (MOI = 0.1), then washed and cultured in DMEM-10% FCS. At 10 h post-infection, naïve “target” HEK293T cells were labeled with 10 µM CFSE (Sigma) and 3×105 labeled “target” cells were co-seeded into 12-well plates with 3×105 extensively washed infected or non-infected “producer” cells. The cells were then co-cultured in DMEM-10% FCS, supplemented or not with 200 µg/mL 8B10. At 0 h and 16 h post-co-culture the cells were harvested, washed and fixed/permeabilized (BD Cytoperm/Cytofix, BD Biosciences), prior to intracellular staining with 5 µg/mL mouse IgG2a anti-alphavirus (3581) (Santa Cruz Biotechnology) followed addition of 4 µg/mL Alexa 647-conjugated goat anti-mouse IgG (Invitrogen). The proportion of infected cells was then determined by flow-cytometry (FACSCalibur, BD Biosciences). Cell supernatants were also collected in parallel after 16 h co-culture for analysis by Plaque Assay to quantify infectious extra-cellular CHIKV. Alternatively, at 14 h post-infection, 105 HEK293T cells were seeded into μ-Slide 8 well plates (Ibidi) with 105 CFSE-labeled uninfected cells and then cultured as described above. After 16 h of co-culture the cells were washed with PBS, fixed with PBS-4% paraformaldehyde and then permeabilized in PBS-0.5% Triton X-100. The permeabilized cells were then stained with 3 µg/mL mouse IgG2a anti-alphavirus (3581) followed by addition of 3 µg/mL Alexa 647-conjugated goat anti-mouse IgG (Invitrogen). ProLong Gold (Invitrogen) was added to the wells and fluorescence was analyzed using an Olympus FV1000 confocal microscope at ×40 magnification (or at ×80 in highlighted panels).
Data were analyzed using GraphPad Prism 5. The specific statistical tests used are indicated in the respective figure legends.
We thank Mark Suter (University of Zurich) for providing AGR129 mice. We are very grateful to Alessandra Nardin (Singapore Immunology Network, BMSI, A*STAR, Singapore) and Annette Martin (Institut Pasteur, Paris, France) for their critical reading of the manuscript, and to Jean-Pierre Abastado and Sebastien Bertin-Maghit (both from Singapore Immunology Network, BMSI, A*STAR, Singapore) for helpful discussions. We are grateful to Lucy Robinson and Neil McCarthy for manuscript editing.