Chikungunya virus cell-to-cell transmission is mediated by intercellular extensions in vitro and in vivo

Abstract

Chikungunya virus (CHIKV) has recently emerged to cause millions of human infections worldwide. Infection can induce the formation of long intercellular extensions that project from infected cells and form stable non-continuous membrane bridges with neighboring cells. The mechanistic role of these intercellular extensions in CHIKV infection was unclear. Here we developed a co-culture system and flow cytometry methods to quantitatively evaluate transmission of CHIKV from infected to uninfected cells in the presence of neutralizing antibody. Endocytosis and endosomal acidification were critical for virus cell-to-cell transmission, while the CHIKV receptor MXRA8 was not. By using distinct antibodies to block formation of extensions and by evaluation of transmission in HeLa cells that did not form extensions, we showed that intercellular extensions mediate CHIKV cell-to-cell transmission. In vivo, pretreatment of mice with a neutralizing antibody blocked infection by direct virus inoculation, while adoptive transfer of infected cells produced antibody-resistant host infection. Together our data suggest a model in which the contact sites of intercellular extensions on target cells shield CHIKV from neutralizing antibodies and promote efficient intercellular virus transmission both in vitro and in vivo.

INTRODUCTION

Chikungunya virus (CHIKV) is an enveloped positive-strand RNA virus that belongs to the Alphavirus genus in the Togaviridae family¹. The genus includes important human pathogens such as CHIKV, Ross River virus, and the Eastern, Western and Venezuelan equine encephalitis viruses, as well as less pathogenic viruses such as Sindbis virus (SINV) and Semliki Forest virus (SFV)^1,2. Like most alphaviruses, CHIKV is transmitted to its vertebrate hosts by mosquito vectors. In humans, acute CHIKV infection causes high-titer viremia, rash, fever and severe muscle and joint pain^3,4. Chronic and debilitating arthritis and joint pain can persist for months to years after the initial infection⁵. CHIKV was originally discovered in Tanzania⁶ and caused sporadic outbreaks in parts of Africa, but starting in 2004 it caused a multi-year pandemic in countries around the Indian Ocean² and since 2013 it has spread across the Americas^7,8. To date there are no approved vaccines or antiviral therapies for CHIKV infection.

The alphavirus RNA genome encodes four non-structural proteins that mediate RNA replication, and six structural proteins: capsid, E3, E2, 6K, TF and E1^1,9. The alphavirus particle is highly organized and approximately 70nm in diameter, with a central nucleocapsid core that is comprised of the RNA genome surrounded by the capsid protein, enveloped by the virus lipid membrane containing 240 copies of heterodimers of the E2 and E1 transmembrane proteins. Alphavirus entry is initiated by virus binding to cell surface receptors, followed by clathrin-mediated endocytosis and delivery to endosome compartments^10,11. The acidic pH in endosomes triggers refolding of the E1 membrane fusion protein to drive the fusion of the virus and endosome membranes. The nucleocapsid is released into the cytoplasm where genome replication and viral protein synthesis occur. The alphavirus structural proteins are synthesized as a polyprotein. The capsid is released by autoproteolysis, and the E2/E1 envelope proteins are then transported through the secretory pathway to the plasma membrane, where assembly, budding and release of progeny virus take place^1,9. We will here refer to this pathway of infection by virus particles released into the extracellular fluid as “free” virus infection.

In addition to the better-studied process of infection by free virus particles, some animal viruses infect via pathways known collectively as cell-to-cell transmission^12,13. The mechanisms differ between viruses, but typically involve transfer of virions or virus components from infected cells to uninfected target cells, often with a dramatic remodeling of cellular architecture. For example, human immunodeficiency virus (HIV-1) infection can induce the formation of virological synapses that promote close cellular apposition and virus transmission¹⁴. HIV-1 particles also may be transferred through nanotubes¹⁵. During murine leukemia virus infection, filopodial bridges form between uninfected and infected cells, mediating virus transit toward the uninfected cell by retrograde actin flow and resulting in efficient intercellular spread¹⁶. These types of cell-cell contacts can shield viruses from neutralizing antibodies, circumventing immune barriers that exist for virus spread via extracellular fluid. Resistance to neutralizing antibodies has also been used as an assay for cell-to-cell transmission¹⁷.

Alphavirus infection produces high titers of free infectious virus particles either in cell cultures or in vivo¹. However, earlier studies also described antibody-resistant CHIKV spread in cell culture, detected as foci of CHIKV infection in the presence of antiviral serum¹⁸ or as antibody-resistant infection of labeled target cells¹⁹. We and others previously showed that alphaviruses can induce the formation of filopodia-like extensions in infected cells^20,21. We termed these structures intercellular long extensions (ILE), and defined them as actin and tubulin-positive, >10 μm in length, and emanating from the infected cell to form stable contacts with a neighboring cell²⁰. ILE do not mediate intercellular transfer of cytoplasmic dyes or membrane markers, and thus are close-ended structures. Infection by a fusion-inactive SFV mutant induces ILE but not infection of neighboring cells, suggesting that alphavirus cell-to-cell transmission involves infectious virus particles rather than transfer of viral RNA or replication complexes²⁰. Expression of the alphavirus structural proteins alone is sufficient to induce ILE, thus ruling out a required role for the non-structural proteins and virus replication.

Despite the description of the properties of ILE, data connecting them to a role in alphavirus cell-to-cell transmission in vitro and in vivo are lacking. Here we established a co-culture system and flow cytometry methods to quantitate CHIKV cell-to-cell transmission in the presence of chCHK-152, a potently neutralizing CHIKV antibody. We defined cellular requirements for ILE formation and cell-to-cell transmission, and demonstrated that ILE mediate CHIKV cell-to-cell transmission. Using adoptive transfer of CHIKV-infected cells into mice, we provide evidence for ILE-mediated cell-to-cell transmission in vivo.

RESULTS

Assay and role of CHIKV cell-to-cell transmission in cell culture.

To establish a quantitative assay for CHIKV cell-to-cell transmission, we used a CHIKV vaccine strain-based reporter virus that expresses cytoplasmic GFP during infection (CHIKV-GFP)²² (Extended Data Fig. 1a) and has comparable growth kinetics in mouse embryo fibroblast (MEF) cells as the unmodified 181/25 vaccine strain (Extended Data Fig. 1b). To inhibit infection by free virus in the culture medium we used a chimeric mouse-human neutralizing monoclonal antibody, mAb chCHK-152, which binds across E2 domain A, domain B and the ß-ribbon connector, preventing exposure of the E1 fusion loop and inhibiting virus fusion^23,24. chCHK-152 potently neutralized both CHIKV and CHIKV-GFP infection of MEF cells (IC₅₀ 0.69 ng/mL and 0.76 ng/mL, respectively), whereas a dengue virus-specific mAb DEN-4G2 showed no effect (Fig. 1a). Studies with retroviruses show that some neutralizing antibodies can prevent formation of virological synapses²⁵. To test the effect of chCHK-152 on formation of CHIKV-induced ILE, MEF cells were infected with CHIKV-GFP and then cultured in the presence of increasing concentrations of mAb. Even at the highest mAb concentration tested (20 μg/mL), chCHK-152 did not inhibit formation of ILE (Fig. 1b and Fig. 1c).

Fig. 1. Cell-to-cell transmission promotes CHIKV infection in cell culture.

a, mAb chCHK-152 neutralizes CHIKV and CHIKV-GFP. Viruses were incubated with the indicated concentrations of mAb chCHK-152 or the negative control mAb DEN-4G2 for 1 h at 37°C. Neutralization was determined by FFA on MEF cells. Data shown represent the mean ± S.D. of 3 independent experiments. b, Visualization of ILE in the presence of mAb chCHK-152. MEF cells were infected with CHIKV-GFP (MOI=0.5) for 2 h or mock-infected, then cultured in complete medium +/− 20 μg/mL of chCHK-152 for 9 h. Cells were permeabilized and immunostained to detect E2/E1 proteins (red, pseudo color) and tubulin (green, pseudo color). The GFP reporter channel is not shown. White arrowheads indicate ILE. Bar = 20 μm. c, ILE formation is not affected by chCHK-152. MEF cells were infected and treated as in panel 1b, and ILE in infected cells were quantitated based on positive staining for E2/E1 and tubulin, contact with a neighboring cell, and length ≥10 μm. Mean ± S.D. of 3 independent experiments (shown as points), each with n > 130 infected cells. d, Antibody-resistant cell-to-cell transmission. “Producer” MEF cells were infected with CHIKV-GFP at MOI=10, incubated at 37°C for 2 h, and washed to remove extracellular virus. “Target” MEF cells were prestained with CMRA dye, plated onto producer cells in the presence of chCHK-152 as indicated, and co-cultured for 12 h. The co-culture media were collected and the cells were fixed and analyzed by flow cytometry. A representative example of 3 independent experiments is shown. e, Quantitation of target cell infection. The fraction of target cells infected (Q2/(Q1+Q2)) was quantitated in samples prepared as in panel 1d. Mean ± S.D of 3 independent experiments (shown as points). f, Infectious virus in the co-culture media was quantitated by infectious center assay (ICA). Mean ± S.D. of 3 independent experiments (shown as points). Statistical significance in panels c, e, and f was calculated by unpaired two-tailed t-tests.

We then used mAb chCHK-152 to establish a quantitative assay system for CHIKV cell-to-cell transmission in cell culture. MEF “producer” cells were infected with CHIKV-GFP, and a separate population of MEF “target” cells was stained with Cell Tracker Orange CMRA dye (CMRA). The infected producer cells were washed, CMRA-labeled target cells were added at an approximately 2:1 ratio (target:producer), and the co-cultures were incubated for 12 h in media alone, or in media containing mAb chCHK-152 or the negative control mAb DEN-4G2 (schematic in Extended Data Fig. 1c). The culture media were then collected and titered, and the cell mixtures were analyzed by flow cytometry. As shown in Fig. 1d (mock-infected panel), flow cytometry clearly distinguished unlabeled cells (Q4) and CMRA dye-labeled cells (Q1) after 12 h co-culture, confirming the expected lack of dye transfer and absence of signal in the GFP channel. When the unlabeled cells were infected with CHIKV-GFP and the co-culture was carried out in the absence of chCHK-152, efficient infection of the target cells was observed as monitored by the GFP reporter (Fig. 1d and 1e, 0 μg/mL antibody samples). Target cell infection was decreased in a dose-dependent manner by the addition of 1–5 μg/mL chCHK-152. However, no further reduction of target cell infection was observed at antibody concentrations above 5 μg/mL. Thus ~17% of target cell infection was resistant to neutralization even by 20 μg/mL chCHK-152. Titration of the co-culture media confirmed high levels of infectious CHIKV (~4×10⁷ IC/ml) in cell cultures without antibody, which was efficiently neutralized when the media contained chCHK-152 concentrations ≥ 5 μg/mL (Fig. 1f). The control DEN-4G2 mAb had no effect on target cell infection or infectious virus in the co-culture media (Extended Data Fig. 2a and 2b, Fig. 1a), in keeping with the fact that under control mAb conditions target cells can be infected by free virus in the media and by cell-to-cell transmission. Similar chCHK-152-resistant cell-to-cell transmission was observed using human U-2 OS cells (Extended Data Fig. 2c and 2d). Thus, we established a co-culture system demonstrating CHIKV cell-to-cell transmission that is resistant to high concentrations of a strongly neutralizing antibody.

CHIKV ILE and cell-to-cell transmission are independent of MXRA8.

MXRA8 is an attachment and entry receptor for CHIKV that binds to E2/E1 heterodimers on the virus particle and promotes infection in cell culture and in mice^26,27. To test the role of MXRA8 in cell-to-cell transmission we made use of MEF-MXRA8-KO cells²⁶. We confirmed the loss of MXRA8 surface expression in the MEF-MXRA8-KO cells (Fig. 2a). Consistent with prior studies, infection by CHIKV-GFP and CHIKV was decreased by about 10-fold in MEF-MXRA8-KO cells vs. the parental MEF, whereas SINV infection was comparable in the two cell types (Fig. 2b). Infection by CHIKV-GFP induced comparable numbers of ILE in MEF vs, MEF-MXRA8-KO cells (Fig. 2c and d). To test the role of MXRA8 in cell-to-cell transmission, MEF producer cells were infected with CHIKV-GFP and co-cultured with MEF or MEF-MXRA8-KO target cells in the presence of 5 μg/mL chCHK-152 or control mAb. Flow cytometry analysis showed no significant difference in chCHK-152-resistant infection of MEF vs. MEF-MXRA8-KO target cells (Fig. 2e). Thus, the MXRA8 receptor is not required for either formation of ILE or CHIKV cell-to-cell transmission.

Fig. 2. Role of MXRA8 and the endocytic pathway in cell-to-cell transmission.

a, Cell surface expression of MXRA8 in MEF or MEF-MXRA8-KO cells measured by flow cytometry. Representative example of 3 independent experiments. b, Infectivity of CHIKV-GFP, CHIKV or SINV in MEF or MEF-MXRA8-KO cells by ICA, normalized to 10⁸ IC/mL on MEF cells. Mean ± S.D of 3 independent experiments (shown as points). c, Visualization of ILE in MEF and MEF-MXRA8-KO cells. Cells were infected with CHIKV-GFP (MOI=0.5, based on Fig. 2b) for 11 h and ILE visualized as in Fig. 1b. White arrowheads indicate ILE. Bar = 20 μm. d, ILE are independent of MXRA8. Cells were infected as in panel 2c and ILE in infected cells quantitated as in Fig. 1c. Data shown represent the mean ± S.D. of 3 independent experiments (shown as points), each with n > 130 infected cells. e, Cell-to-cell transmission is independent of MXRA8. MEF producer cells were infected as in Fig. 1d, and MEF or MEF-MXRA8-KO cells were used as target cells. After co-culture for 12 h in the presence of 5 μg/ml chCHK-152 or a control mAb DEN-4G2, the cells were analyzed by flow cytometry. Mean ± S.D. of 3 independent experiments (shown as points). f, Dyngo-4a inhibits CHIKV infection. U-2 OS cells were pretreated with Dyngo-4a or DMSO vehicle for 1 h, infected with CHIKV-GFP (MOI=0.5), cultured for 12 h in the presence (12h) or absence of inhibitor (pretreat), and analyzed by flow cytometry. Mean ± S.D of 3 independent experiments (shown as points). g, Dyngo-4a inhibits CHIKV cell-to-cell transmission. Target U-2 OS cells were pretreated with Dyngo-4a or DMSO for 1 h, then co-cultured for 12h with CHIKV-GFP-infected producer cells and analyzed, as in 2e. Mean ± S.D of 3 independent experiments (shown as points). h, Dominant-negative Rab5 inhibits CHIKV infection. U-2 OS cell lines inducibly-expressing Rab5-WT or Rab5-DN were incubated +/− doxycycline (dox) for 16 h, infected with CHIKV (MOI=1) for 1 h, cultured in media containing 20 mM NH₄Cl for 12 h, and primary infection quantitated by flow cytometry. Mean ± S.D of 3 independent experiments (shown as points). i, Dominant-negative Rab5 inhibits CHIKV cell-to-cell transmission. Target U-2 OS cell lines inducibly-expressing Rab5-WT or Rab5-DN were incubated +/− doxycycline for 16 h, then co-cultured with CHIKV-infected producer cells for 12 h and analyzed, as in 2e. Mean ± S.D of 4 independent experiments (shown as points). j, Bafilomycin A1 inhibits CHIKV infection. U-2 OS cells were pretreated with Bafilomycin A1 for 1 h, infected with CHIKV-GFP (MOI=1), cultured for 12 h in the presence (12h) or absence (pretreat) of inhibitor, and analyzed by flow cytometry. Mean ± S.D of 3 independent experiments (shown as points). k, Bafilomycin A1 inhibits CHIKV cell-to-cell transmission. Target U-2 OS cells were pretreated with Bafilomycin A1 or DMSO for 1 h, then co-cultured with CHIKV-GFP-infected producer cells for 12 h and analyzed, as in 2e. Mean ± S.D of 3 independent experiments (shown as points). Statistical significance in panels b and d-k was determined using unpaired two-tailed multiple t-tests.

CHIKV cell-to-cell transmission requires endocytic uptake.

We next tested the role of the endocytic pathway in cell-to-cell transmission using U-2 OS cells that endogenously express MXRA8. Dyngo-4a is a dynamin inhibitor that blocks the scission of clathrin-coated vesicles from the plasma membrane²⁸. To verify Dyngo-4a inhibition of virus infection, U-2 OS cells were pretreated with Dyngo-4a for 1 h, then inoculated with CHIKV-GFP in the absence or continued presence of Dyngo-4a. As shown in Fig. 2f, incubation of cells with Dyngo-4a significantly decreased CHIKV infection. Similar levels of CHIKV inhibition were observed when the cells were pretreated or continually cultured for 12 h in the presence of Dyngo-4a, demonstrating that inhibition was not rapidly reversed. No effect of Dyngo-4a treatment on ILE formation was observed (Extended Data Fig. 3). To test whether Dyngo-4a inhibits cell-to-cell transmission, target cells were pretreated with Dyngo-4a and then co-cultured with CHIKV-infected producer cells in the presence of chCHK-152 or control mAb. The results showed a dose-dependent inhibition by Dyngo-4a, demonstrating that endocytic vesicle formation is involved in CHIKV cell-to-cell transmission (Fig. 2g).

To evaluate the requirement for early endosomes in CHIKV cell-to-cell transmission, we generated U-2 OS cell lines that inducibly over-express Rab5-WT or a dominant-negative Rab5 mutant (Rab5-DN) that inhibits delivery to the early endosome compartment²⁹ and CHIKV infection³⁰. Flow cytometry confirmed the inducible expression of Rab5-WT, or of the Rab5-DN in two independent cell lines (Extended Data Fig. 4a). As expected, expression of Rab5-DN but not Rab5-WT significantly decreased infection by both CHIKV (Fig. 2h) and SFV (Extended Data Fig. 4b). No effect of Rab5-WT or Rab5-DN expression on ILE formation was observed (Extended Data Fig. 4c and 4d). Co-culture experiments showed that CHIKV cell-to-cell transmission was significantly reduced in two independent target cell lines expressing Rab5-DN (Fig. 2i). In contrast, a small but significant increase in CHIKV cell-to-cell transmission was observed for Rab5-WT-expressing target cells (Fig. 2i), suggesting that increased Rab5 levels may promote virus delivery to early endosomes.

To determine the importance of endosomal acidification in CHIKV cell-to-cell transmission, the non-reversible inhibitor bafilomycin A1 was used to inhibit the vacuolar ATPase and neutralize endosomal pH³¹. Pretreatment of cells with bafilomycin A1 inhibited >90% of CHIKV infection (Fig. 2j). Co-culture experiments showed that CHIKV cell-to-cell transmission was significantly reduced by pretreatment of target cells with bafilomycin A1 (Fig. 2k). ILE formation was not affected by bafilomycin A1 (Extended Data Fig. 3).

Taken together, our results demonstrated that, similar to infection of cells by exogenous virus addition, CHIKV cell-to-cell transmission exploits the target cell endocytic pathway including dynamin-dependent uptake, Rab5-mediated delivery to early endosomes, and exposure to endosomal low pH.

Tetherin retention inhibits CHIKV cell-to-cell transmission.

Tetherin is an interferon-induced cellular protein with membrane anchors at its N- and C-termini³². Tetherin inhibits the release of many enveloped viruses from infected cells by tethering budded particles to the plasma membrane. We previously reported that expression of either the long tetherin isoform (L-tetherin) or WT-tetherin producing both long and short isoforms reduced SFV release through a process involving tetherin-mediated virus internalization^33,34. To evaluate the effect of tetherin retention on CHIKV cell-to-cell transmission, we used U-2 OS cells that inducibly express WT-tetherin or L-tetherin. Doxycycline-induction and flow cytometry analysis confirmed robust tetherin surface expression (Fig. 3a). As previously observed for SFV, expression of either WT-tetherin or L-tetherin caused a significant decrease in infectious CHIKV release from infected cells (Fig. 3b). Neither WT-tetherin nor L-tetherin expression produced intercellular extensions in the absence of virus infection, or inhibited ILE formation upon CHIKV infection (Fig. 3c and Extended Data Fig. 5). We analyzed cell-to-cell transmission using CHIKV-infected U2-OS producer cells expressing WT-tetherin or L-tetherin, co-cultured with U-2 OS target cells. Tetherin expression in the producer cells significantly decreased antibody-resistant CHIKV transmission (Fig. 3d), suggesting that cell-to-cell transmission requires the release of virus particles at the contact site between the producer and target cells.

Fig. 3. Tetherin expression inhibits CHIKV cell-to-cell transmission.

a, Parental U-2 OS or U-2 OS cells inducibly-expressing WT- or L-tetherin were incubated with doxycycline (dox) for 16 h, surface-stained with anti-tetherin mAb, and analyzed by flow cytometry. Representative example of 3 independent experiments. b, The indicated U-2 OS cells were incubated with or without doxycycline for 16 h and infected with CHIKV-GFP (MOI=10) for 10 h. Infectious virus release was measured by ICA. Mean ± S.D. of 3 independent experiments (shown as points). c, ILE are independent of tetherin. The indicated U-2 OS cells were incubated with or without doxycycline for 16 h, infected with CHIKV-GFP (MOI=0.5) for 11 h, and ILE in infected cells quantitated as in Fig. 1c. Data shown represent the mean ± S.D. of 3 independent experiments (shown as points), each with n > 130 infected cells. d, Tetherin inhibits CHIKV cell-to-cell transmission. The indicated U-2 OS producer cells were incubated with or without doxycycline for 16 h, infected with CHIKV-GFP (MOI=10), and co-cultured with U-2 OS cells (target cells) for 12 h in the presence of 5 μg/mL DEN-4G2 or chCHK-152 mAb. The fraction of target cells infected was determined by flow cytometry. Mean ± S.D of 3 independent experiments (shown as points). Statistical significance of panels b-d was calculated using unpaired two-tailed multiple t-tests.

HeLa cells lack ILE and cell-to-cell transmission.

Our prior studies determined that ILE are induced by infection of Vero cells or primary human endothelial cells with SINV, SFV or CHIKV²⁰, and we here describe ILE formation in CHIKV-infected MEF and U-2 OS cells. However, we observed that CHIKV-infected HeLa cells did not form ILE (e.g., see Fig. 4c and 4d). While the reasons for this difference in ILE formation are not currently understood, we hypothesized that it could provide an approach to test the role of ILE in CHIKV cell-to-cell transmission. To minimize differences in infection and virus growth between cell lines, we compared CHIKV primary infection and virus production in U-2 OS, HeLa, and HeLa-MXRA8 cells stably expressing MXRA8²⁶. The infectivity of CHIKV-GFP or CHIKV in U-2 OS and HeLa-MXRA8 cells was similar, while infectivity by both viruses was reduced about 10-fold in the parental HeLa cells that lack MXRA8 (Fig. 4a). Single-step growth curves showed robust CHIKV production in HeLa-MXRA8 cells, which had slightly faster virus growth kinetics than U-2 OS cells up to the plateau of virus production at ~13 h post-infection (Fig. 4b). Despite their robust virus production, however, the formation of ILE by CHIKV-infected HeLa-MXRA8 cells was significantly reduced compared to the levels in CHIKV-infected U-2 OS cells, and was similar to the background levels in mock-infected cells (Fig. 4c, d). Cell-to-cell transmission was then evaluated using HeLa-MXRA8 cells as producer and target cells, and compared with transmission using U-2 OS cells as producer and target cells. After co-culturing for either 12 h or 16 h, the efficiency of chCHK-152-resistant cell-to-cell transmission in HeLa-MXRA8 cells was significantly lower than that in U-2 OS cells (Fig. 4e). We observed a somewhat lower infection of the HeLa-MXRA8 target cells vs. U2-OS in the presence of the control mAb (Fig. 4e). This may be because of the lack of ILE and thus lack of cell-to-cell transmission in HeLa-MXRA8 cells, or other effects of the co-culture system. Together these results suggest a key role for CHIKV-induced ILE in CHIKV cell-to-cell transmission.

Fig. 4. Properties of CHIKV infection in HeLa cells.

a, Infectivity of CHIKV-GFP and CHIKV on U-2 OS, HeLa-MXRA8 or HeLa cells was determined by ICA. Data were normalized to 10⁸ IC/mL on U-2 OS cells. Mean ± S.D. of 3 independent experiments (shown as points). b, Growth of CHIKV in U-2 OS or HeLa-MXRA8 cells. U-2 OS or HeLa-MXRA8 cells were infected with CHIKV (MOI=10) for 2 h, washed to remove virus, and virus production at the indicated times quantitated by ICA. The graphs represent the means and range of 2 independent experiments. c, Lack of ILE in HeLa-MXRA8 cells. U-2 OS and HeLa-MXRA8 cells were infected with CHIKV for 11 h, and ILE visualized as in Fig. 1b. White arrowheads indicate ILE. Bar = 20 μm. d, Quantitation of ILE in HeLa-MXRA8 cells. Samples infected as in panel 4c were quantitated as described in Fig. 1c. Data shown represent the mean ± S.D. of 3 independent experiments (shown as points), each with n > 130 infected cells. e, Lack of CHIKV cell-to-cell transmission in HeLa-MXRA8 cells. U-2 OS or HeLa-MXRA8 producer cells were infected with CHIKV (MOI=10) for 2 h, then co-cultured with their respective target U-2 OS or HeLa-MXRA8 cells, for 12 or 16 h in the presence of 5 μg/mL DEN-4G2 control or chCHK-152 mAb. Infection of target cells was determined by flow cytometry. Mean ± S.D of 3 independent experiments (shown as points). Statistical significance of panels a, d, e was calculated using unpaired two-tailed multiple t-tests.

Antibody inhibition of ILE blocks cell-to-cell transmission.

A direct test of the role of ILE in CHIKV cell-to-cell transmission would be to inhibit ILE formation in a normally permissive cell. We tested several mAbs against the CHIKV envelope proteins by adding them to MEF cells 2h after infection. Two mAbs against the CHIKV E1 protein, E10–18 and chCHK-166, caused a dose-dependent inhibition of subsequent ILE formation whereas chCHK-263, a mAb against the CHIKV E2 protein, did not inhibit (Fig. 5a and b). As detailed above, CHK-152 binds across E2 domains A and B^23,24. CHK-166 binds E1 domain II near the fusion loop and CHK-263 binds E2 domain B²³. We used an ELISA to quantitate the binding of these mAbs to the envelope proteins on the surface of infected cells. E10–18, chCHK-166, chCHK-152 and chCHK-263 all bound efficiently to infected cells as compared to the control antibody DEN-4G2 (Fig. 5c). Comparison of the neutralization activity of this set of mAbs showed that either a non-neutralizing mAb such as E10–18 or a neutralizing mAb such as chCHK-166 was able to block ILE formation, while the chCHK-152 and chCHK-263 mAbs efficiently neutralized but did not block (Fig. 5d). Treatment with E10–18 alone did not decrease target cell infection in the co-culture system (Extended Data Fig. 6), suggesting that in the absence of a neutralizing Ab the ongoing free virus infection was sufficient to cause robust CHIKV infection of target cells. To test the effect of blocking ILE on cell-to-cell transmission, we used 5 μg/mL of chCHK-152 to inhibit free virus infection in the co-culture system, and added increasing concentrations of chCHK-263, chCHK-166, E10–18, or the control mAb 4G2 (Fig. 5e). Both chCHK-166 and E10–18 inhibited CHIKV cell-to-cell transmission in a dose-dependent manner whereas the control mAb DEN-4G2 had no effect. mAb chCHK-263 did not reduce cell-to-cell transmission even though it very efficiently neutralized CHIKV infection. Thus, these data indicate that CHIKV cell-to-cell transmission requires the presence of ILE.

Fig. 5. Inhibition of ILE blocks cell-to-cell transmission.

a, Treatment with mAb chCHK-166 or E10–18 inhibits ILE. MEF cells were infected with CHIKV-GFP for 2 h, then cultured for 9 h in the presence of mAbs as indicated. ILE visualized as in Fig. 1b. White arrowheads indicate ILE. Bar = 20 μm. b, Dose dependence of ILE inhibition by mAb chCHK-166 or E10–18. Samples infected as in 5a were quantitated as in Fig. 1c. Mean ± S.D. of 3 independent experiments (shown as points), each with n > 130 infected cells. Statistical significance was calculated by unpaired two-tailed multiple t-tests. c, Binding of CHIKV mAbs to CHIKV-infected cells. MEF cells were infected with CHIKV-GFP (MOI=5) for 11 h, fixed, incubated with serial dilutions of mAbs, followed by ELISA. Mean of 3 independent experiments. d, Neutralization of CHIKV-GFP by the indicated mAbs (as in Fig. 1a). Mean ± S.D. of 3 independent experiments. e, chCHK-166 or mAb E10–18 inhibits cell-to-cell transmission. MEF producer cells were infected with CHIKV-GFP (MOI=10) for 2 h, washed, and co-cultured for 12 h with MEF target cells in the presence of 5 μg/mL chCHK-152 to inhibit free virus infection, plus additional mAbs as indicated. Infection of target cells was quantitated by flow cytometry. Mean ± S.D. of 3 independent experiments (shown as points). Statistical significance was calculated by unpaired two-tailed t-tests. f, Neutralization of CHIKV-GFP or CHIKV-GFP E1-K61T-G64S by the indicated mAbs was determined as in Fig. 1a. Mean ± S.D. of 3 independent experiments. g, mAb chCHK-166 does not bind to CHIKV E1-K61T-G64S-infected cells. MEF cells were infected with CHIKV-GFP E1-K61T-G64S (MOI=5) for 11 h, followed by ELISA as in Fig. 5c. Mean ± S.D. of 3 independent experiments. h, CHIKV E1-K61T-G64S escapes chCHK-166 inhibition of ILE formation. MEF cells were infected for 2 h then cultured in the presence of 20 ug/ml mAb chCHK-166 for 9 h. ILE were quantitated as in Fig. 1c. Mean ± S.D. of 3 independent experiments (shown as points), each with n > 130 infected cells. Statistical significance was calculated by unpaired two-tailed t-tests. i, CHIKV E1-K61T-G64S escapes chCHK-166 inhibition of cell-to-cell transmission. MEF producer cells were infected (MOI=10) for 2 h, washed, and co-cultured for 12 h with MEF target cells in the presence of 5 μg/mL chCHK-152 plus additional mAbs as indicated. Infection of target cells was quantitated by flow cytometry. Mean of 2 independent experiments (shown as points).

While the viral structural protein requirements for ILE are not yet well-defined, blocking of specific epitopes may be critical to ILE inhibition by mAbs. To address the mechanisms by which chCHK-166 inhibits ILE and cell-cell transmission, we took advantage of a prior neutralization escape selection that identified the resistance mutations E1-K61T and E1-G64S ^23,35. We generated CHIKV-GFP E1-K61T-G64S and showed that this virus was resistant to neutralization by chCHK-166 but not by chCHK-152 (Fig. 5f). MEF cells infected with the escape mutant lost binding to chCHK-166 but not chCHK-152 (Fig. 5g). The escape mutant efficiently induced ILE in MEF cells, but these ILE were no longer inhibited by chCHK-166 (Fig. 5h, Extended Data Fig. 7). The mutant showed comparable levels of cell-to-cell transmission as the WT CHIKV, but transmission was resistant to chCHK-166 (Fig. 5i). Thus, chCHK-166 binding, neutralization, and ILE inhibition activities were all dependent on a specific E1 epitope, but disruption of this epitope alone was not sufficient to inhibit ILE formation.

CHIKV cell-to-cell transmission in vivo.

To address the role of cell-to-cell transmission in vivo, we developed a model system based on adoptive transfer of infected MEF cells into recipient C57BL/6 mice pretreated with a neutralizing mAb. For virus infection, we used CHIKV AF15561-Venus (WT CHIKV-Venus), a reporter virus based on a pathogenic clinical isolate that is the parent virus to CHIKV 181/25³⁶. For a neutralizing mAb we used chCHK-152 carrying an N297Q mutation that abolishes a critical N-linked glycosylation site in the CH2 domain of the antibody Fc. The resultant aglycosylated antibody has an unaltered half-life in mouse serum, but does not appreciably interact with C1q or Fcγ receptors that could promote clearance of CHIKV-infected cells^23,37.

We first investigated whether pretreatment with chCHK-152 N297Q can block WT CHIKV-Venus infection. Mice were injected i.p. with either chCHK-152 N297Q or PBS, and 6 h later were inoculated with CHIKV-Venus in the left footpad. Virus infection in the ipsilateral ankle tissue was quantitated 24 h later by flow cytometry analysis of the CD45-negative cell population (Extended Data Fig. 8). The PBS-treated control showed a substantial population of Venus-positive cells, but infection was effectively blocked by chCHK-152 N297Q treatment.

We then tested if infected MEF cells could mediate antibody-resistant cell-to-cell transmission in mice. MEF “donor” cells were stained with CellTrace Violet dye (Violet) and inoculated with WT CHIKV-Venus for 2 h. Cells were washed extensively and incubated with mAb chCHK-152 for 1 h to neutralize any remaining input virus (Fig. 6a and methods). Mice were pretreated with mAb chCHK-152 N297Q or PBS, then either inoculated with WT CHIKV-Venus or adoptively transferred with CHIKV-Venus infected MEF cells. 24 h later ipsilateral ankle tissue was harvested and Venus-reporter protein expression was quantitated in the Violet-positive donor cells and the Violet-negative host cells by flow cytometry (Fig. 6b, Extended Data Fig. 9). The control, PBS-treated mice showed similar levels of infection after either virus inoculation or adoptive transfer of infected MEF, indicating comparable exposure to infectious virus (Fig. 6c and 6d). However, while infection of host cells by direct virus inoculation was blocked in the chCHK-152 N297Q-treated mice, antibody-resistant infection of host cells was observed after adoptive transfer of infected MEF cells. The % antibody-resistant infection in the direct virus inoculation vs. adoptive transfer samples was 0.02 vs. 0.67% (mean of 6–8 mice in two experiments). To confirm that the detected Venus-positive cells were not infected donor cells, infected host or donor cells were back-gated onto the FSC/SSC flow plot. The results showed that these two populations were physically distinct (Fig. 6e), demonstrating that the Venus-positive, Violet-negative cells are host-derived. To test whether CHIKV induces ILE in the ankle tissue, we isolated the joint cells and infected them ex vivo with the WT CHIKV pathogenic parental strain (AF15561) or the CHIKV vaccine strain (181/25). Both CHIKV strains induced ILE in the primary joint cells, and comparable levels of ILE were observed in the two samples (Fig. 6f, Extended Data Fig. 10). Thus, together our results provide evidence for ILE-mediated cell-to-cell transmission in vivo.

Fig. 6. CHIKV cell-to-cell transmission in mice.

a, Experimental flow diagram of MEF infection. MEF cells were loaded with CellTrace-Violet (Violet), cultured 24 h, and inoculated with WT CHIKV-Venus. Cells were then washed, incubated for 1 hour with chCHK-152 at 5 μg/mL, and prepped for adoptive transfer. b, Experimental flow diagram of in vivo infection experiments. WT C57BL/6 mice were pretreated by i.p. injection with 100 μg of chCHK-152 N297Q or PBS. At 6 h post-treatment, previously infected MEF cells (see panel a) were adoptively transferred by s.c. injection of 10⁶ cells per mouse into the left footpad. As a control group, mice were directly inoculated with 10³ PFU WT CHIKV-Venus by s.c. injection into the left footpad. At 24 h post-cell transfer or inoculation, ipsilateral ankle tissues were harvested, single cell suspensions generated and stained for viability and CD45. Panels a, b created with BioRender.com. c, Representative flow plots showing host cell infection in mice that were pretreated with either PBS or chCHK-152 N297Q and then received 10³ PFU of WT CHIKV-Venus (top two rows) or 10⁶ WT CHIKV-Venus-infected MEF cells (lower two rows). All plots were gated on viable singlet CD45⁻ cells. d, Frequency of Venus⁺ cells within the host CD45⁻ Violet⁻ cell population. The PBS-treated controls from each group of mice were averaged and set to 100% (maximum Venus signal). The chCHK-152 N297Q-treated samples were averaged and graphed as %Venus⁺ as compared to the PBS Venus⁺ signal. Mean ± S.D. of two independent experiments (n = 6 for CHK-152 N297Q +10³ PFU CHIKV-Venus group, n=8 for other 3 groups), with points showing the results from each mouse. e, Flow analysis of the Venus+ MEF vs. Venus+ host cells. Venus⁺ donor MEF cells (CD45⁻Violet⁺) or Venus⁺ host cells (CD45⁻Violet⁻) were back-gated onto the FSC/SSC plot to compare size and internal complexity. f, ILE formation in primary joint cells. Single cells from ankle tissue of C57BL/6 mice were infected with CHIKV strain 181/25 or WT CHIKV strain AF15561 for 16 h. ILE were quantitated as described in Fig. 1c. Data shown represent the mean ± S.D. of 2 independent experiments, each with 4 mice, with points showing the results from each mouse. Total infected cells evaluated per condition > 130. Statistical significance of panels d and f was calculated by unpaired two-tailed t-test.

DISCUSSION

For alphaviruses, infection by free virus particles and cell-to-cell virus transmission occur in parallel. To specifically study cell-to-cell transmission in vitro, we used flow cytometry to quantitate infection of labeled target cells by virus-infected producer cells in the presence of an excess of a potently neutralizing antibody. This co-culture system also permitted separate inhibitor treatment of producer and target cells, and differential cellular protein expression in the two cell populations.

While our previous results indicated that alphavirus cell-to-cell transmission involved infectious virus particles²⁰, the pathway used to mediate transmission remained undefined. We used our co-culture system to compare the steps in CHIKV cell-to-cell transmission vs. free CHIKV infection. While knock-out of the MXRA8 receptor in MEF cells reduced free CHIKV infection by ~1 log, the receptor was not required for either ILE formation or for CHIKV cell-to-cell transmission. This may reflect a high local concentration of virus particles in the ILE contact site, making target cell infection relatively MXRA8-independent. Notably, however, the endocytic pathway was clearly required for intercellular transmission. Thus, pharmacologic inhibition of dynamin, which mediates vesicle scission, Rab5DN inhibition of delivery to early endosomes, or neutralization of endosomal low pH by bafilomycin A1 all decreased CHIKV cell-to-cell transmission. The interferon-induced host protein tetherin effectively inhibits particle release into the extracellular milieu for many viruses, but cell-to-cell transmission is often cited as a potential escape route from tetherin inhibition³⁸. For example, tetherin promotes the accumulation of HIV-1 particles at the virological synapse and stabilizes the synapse, and thus can actually increase cell-to-cell spread in T cells³⁹. By contrast, we found that, although tetherin did not affect ILE formation, it inhibited CHIKV cell-to-cell transmission. Together with our previous studies³⁴, these results suggest that tetherin reduces CHIKV cell-to-cell transmission by decreasing the amount of virus at the ILE contact sites through decreased virus release and a tetherin-mediated virus internalization process.

We found that HeLa-MXRA8 cells support efficient CHIKV infection and growth, but do not form ILE and do not mediate efficient CHIKV cell-to-cell transmission. Two different mAbs against the CHIKV E1 protein, E10–18 and chCHK-166, blocked formation of ILE in CHIKV-infected MEF cells and reduced cell-to-cell transmission. Studies of the chCHK-166 escape virus showed that chCHK-166 binding, neutralization and ILE inhibition all required the epitope identified by the resistance mutations. However, the escape mutant still robustly induced ILE and cell-to-cell transmission, and thus further studies are needed to dissect the specific virus protein regions required. It will also be important to define the host factors that promote formation of ILE, and the difference in ILE induction between CHIKV-infected HeLa and U-2 OS cells may provide a system to do this.

We demonstrated that adoptive transfer of infected MEF cells produced antibody-resistant infection of host cells in mouse joint tissues, while free virus infection was blocked. Primary cells derived from mouse joint tissues showed ILE formation upon ex vivo CHIKV infection. Together our data thus provide evidence for ILE-mediated cell-to-cell transmission both in vitro and in vivo. Chronic CHIKV disease appears to involve viral persistence in joint-associated tissues, and long-term persistence of a CHIKV-specific antibody response has also been observed^3,40–42. As chronic CHIKV infection likely involves evasion of antiviral and immune barriers, the potential roles of cell-to-cell transmission and ILE in vivo should be explored as one possible mechanism for CHIKV persistence. Given that cell-to-cell transmission provides a route for evading antibody neutralization, our study emphasizes that preventive and therapeutic approaches should take the unique aspects of cell-to-cell transmission into consideration.

METHODS

Ethics Statement.

This study was performed in accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. All animals were handled according to approved institutional animal care and use committee (IACUC) protocols (#00026) of the University of Colorado School of Medicine (Assurance Number A3269‐01). Experimental animals were humanely euthanized at defined endpoints by exposure to isoflurane vapors followed by bilateral thoracotomy.

Viruses.

Virus stocks were produced from the following infectious clones: CHIKV (CHIKV 181/25 vaccine strain infectious clone pSinRep5–181/25ic⁴³, provided by Dr. Terence S. Dermody), CHIKV-GFP (CHIKV 181/25 reporter virus that expresses cytoplasmic GFP during infection²², provided by Dr. Elena I. Frolova), SFV (pSP6-SFV4⁴⁴), and SINV (dsTE12Q⁴⁵, provided by Dr. Beth Levine). Mouse experiments used CHIKV AF15561-Venus, a reporter virus based on the clinical isolate that is the parent strain to 181/25 and expressing cytoplasmic Venus, and here referred to as WT CHIKV-Venus. Viral RNAs were prepared by in vitro transcription and electroporated into BHK-21/WI-2 cells. The cells were cultured for 24 h at 37°C, the media harvested and clarified by centrifugation, and HEPES pH 8.0 was added to a final concentration of 10 mM. Virus stocks were stored at −80°C and titered by infectious center assays (ICA) in the indicated cell lines, as follows: Cells were inoculated for 2h with serial dilutions of virus stocks. Further secondary infection was blocked by addition of medium containing 20mM NH₄Cl and cells were cultured for ~16h at 30°C. Fluorescence microscopy was then used to quantitate cells positive for cytoplasmic GFP expression or for staining with antibodies against CHIKV E2/E1 or SINV E1 and E2. All infection experiments were based on the MOI specified in the figure legends and determined by ICA on the relevant cell type.

Rab5 plasmids.

The pcDNA4/TO-GFP-Rab5-WT or pcDNA4/TO-GFP-Rab5-DN expression plasmids were generated by PCR amplification of GFP-Rab5-WT or GFPRab5-DN (the dominant-negative S34N mutant) from Rab5-WT-GFP-C1 or Rab5-DN-GFP-C1 plasmid templates⁴⁶. The PCR primers were as follows: 5’-GTTTAAACTTAAGCTTGGTACCGAGCTCATGGTGAGCAAGGGCGAGGAGCTG -3′ (forward); 5′-GCAGAATTCCACCACACTGGACTAGTTTAGTTACTACAACACTGACTCCTG -3′ (reverse). The PCR fragments were then inserted into pcDNA/TO linearized by digestion with BamHI, using the Gibson assembly kit according to the manufacturer’s instructions (#E2611, NEB, Ipswich, MA).

Cell lines.

Mouse embryonic fibroblasts (MEF), MEF-MXRA8-KO, HeLa (ATCC #CCL-2) and HeLa-MXRA8 cells were described previously²⁶. Cells were cultured in high glucose DMEM supplemented with 10% FBS, 10 mM HEPES, 4 mM L-glutamine, 1 mM sodium pyruvate, 1X non-essential amino acids, 100 U penicillin/mL, and 100 μg streptomycin/mL. The MEF cells were derived from C57/BL/6 mice and immortalized by expression of the SV2 large T antigen⁴⁷. The MEF-MXRA8-KO cells were generated by CRISPR-Cas9 editing²⁶. HeLa cells were transduced with lentivirus expressing human MXRA2 isoform 2 in pCSII-EF1-IRES-Venus vector and then sorted for Venus-positive cells by flow cytometry²⁶.

U-2 OS cells were cultured in modified McCoy’s 5A medium supplemented with 10% FBS, 100 U penicillin/mL, and 100 μg streptomycin/mL (U-2 OS complete growth medium). T-REx U-2 OS parent cells that stably express the tetracycline repressor from pCEP4/tetR were cultured in U-2 OS complete growth medium supplemented with 60 μg/mL hygromycin B (#10687010, Invitrogen, Waltham, MA). TET-inducible U-2 OS cell lines expressing GFP-Rab5-WT or GFP-Rab5-DN were generated using methods previously described³³. In brief, pcDNA/TO Rab5 plasmids were transfected into T-REx U-2 OS parent cells using Lipofectamine 2000 transfection reagent (#11668019, Invitrogen, Waltham, MA). Starting at 24 h post transfection cells were selected using 60 μg/mL hygromycin B and 250 μg/mL Zeocin (Thermo Fisher Scientific, Waltham, MA) for 1 week and then maintained in 60 μg/mL hygromycin B and 100 μg/mL Zeocin. Cell lines were isolated using cloning cylinders and screened by flow cytometry for inducible expression. The TET-inducible U-2 OS cells expressing WT-tetherin and L-tetherin were previously generated³⁴. Expression was induced by addition of 1 μg/mL doxycycline (dox) as indicated in the figure legends.

Antibodies.

mAb DEN-4G2 to the flavivirus E protein was purified from hybridoma cell culture supernatant⁴⁸. Recombinant mAb chCHK-152 was provided by Dr. Zachary Bornholdt of Mapp Biopharmaceutial^23,49. The E10–18 mAb was generated as follows: a construct encoding a soluble form of the p62–E1 heterodimer from the clinical isolate 05–115 was expressed in Drosophila melanogaster S2 cells⁵⁰. Mice were immunized with the purified heterodimer and hybridomas were obtained. The E10–18 mAb was mapped to the E1 protein. Sequences for chCHK-166 and chCHK-263 were cloned into pMAZ-IgL and pMAZ-IgH for recombinant expression on a human constant domain backbone⁴⁹. Antibodies were expressed in ExpiCHO-S cells at 32°C following the manufacturer’s instructions³⁷. The chCHK-152 N297Q mAb was generated by using in vitro mutagenesis to produce the N297Q substitution²³. Antibodies were purified by standard protein A or protein G chromatography according to the manufacturer's instructions (Thermo Fisher Scientific, Waltham, MA).

For immunostaining, primary antibodies included: mouse anti-tubulin mAb (clone E7, IF-1:500, Developmental Studies Hybridoma Bank, E7), hamster anti-mouse Mxra8 mAb (9G2.D6, flow-2.5 μg/mL)²⁶, mouse mAb to SFV/CHIKV E2 (E2–1, Focus reduction neutralization test (FRNT) or ICA, 1:10 dilution of the clarified hybridoma supernatant)⁵¹, mouse mAbs to SINV E1 (R2, ICA-1:2 dilution of the clarified hybridoma supernatant) and SINV E2 (R6, ICA-1:2 dilution of the clarified hybridoma supernatant)⁵², rabbit pAb to SFV/CHIKV E2/E1(IF-1:500, flow-1:200)⁵¹, and phycoerythrin (PE)-conjugated anti-human tetherin (clone 26F8, flow-1:10, eBioscience, Waltham, MA, 12–3179-42). Secondary antibodies included: goat anti-mouse Alexa Fluor 488 (IF-1:500, ThermoFisher Scientific, A11001), goat anti-rabbit Alexa Fluor 488 (IF-1:500, ThermoFisher Scientific, A11008), goat anti-mouse Alexa Fluor 555 (IF-1:500, ThermoFisher Scientific, A21422), goat anti-rabbit Alexa Fluor 633 (IF-1:500, ThermoFisher Scientific, A21070), goat anti-rabbit Alexa Fluor 647 (flow-1:500, ThermoFisher Scientific, A21244), goat anti-Armenian hamster Alexa Fluor 647 (flow-1:200, Abcam, ab173004), horseradish peroxidase (HRP)-conjugated goat anti-mouse (FRNT or ELISA -1:1000, SeraCare, Milford, MA, 5450–0011) or goat anti-humanIgG (ELISA -1:1000, SeraCare, Milford, MA, 5450–0009).

Inhibitor studies.

Dyngo-4a (Hydroxy-Dynasore) and Bafilomycin A1 were obtained from MilliporeSigma (Burlington, MA). Inhibitors were dissolved in DMSO and stored according to the manufacturer’s instructions. For testing the effect of inhibitors in free virus infection, cells were seeded into 12-well plates and cultured for ~24h. Cell were then pretreated with Dyngo-4a (30 μM, 60 μM) or Bafilomycin A1 (100 nM) for 1 h, washed twice with PBS, and infected with CHIKV-GFP for 12 h in the absence or presence of inhibitor as indicated. For the cell-to-cell transmission assay, target cells were labeled with CMRA dye (1:000, Cell Tracker Orange CMRA, C34551, Thermo Fisher Scientific, Waltham, MA, emission maximum at 575nm) for 30 min, washed with PBS three times and then pretreated with Dyngo-4a (30 μM, 60 μM) or Bafilomycin A1(100 nM) for 1 h before plating onto producer cells.

Focus reduction neutralization test (FRNT).

100 to 150 focus forming units (FFU) of CHIKV or CHIKV-GFP were incubated with serial dilutions of the indicated mAbs in MEM plus 0.2% BSA and 10mM HEPES pH 7.0 for 1 h at 37°C. MEF cells in 96 well-plates were infected with antibody plus virus for 2 h then overlaid with 1% carboxylmethylcellulose in modified Eagle’s Medium supplemented with heat-inactivated 2% FBS and 10 mM Hepes pH 7.4. At 18 h post-infection, cells were fixed by adding 100 μl of a solution of 1% paraformaldehyde (PFA, Electron Microscopy Science) in PBS prewarmed to 37°C to the overlay and incubating for 1 h. After 5 washes with PBS the cells were permeabilized with 0.1% saponin in PBS containing 0.1% BSA, and incubated with mAb to E2 (E2–1, 1:10 dilution of the clarified hybridoma supernatant) followed by incubation with horseradish peroxidase-conjugated goat anti-mouse IgG (1:1000, SeraCare, 5450–0011). Foci were developed using TrueBlue Peroxidase substrate (Seracare, Milford, MA) and quantified on an ImmunoSpot S6 Macroanalyzer with Biospot 7.0.9.10 software (Cellular Technologies, Shaker Heights, OH). The number of foci in wells containing mAb was normalized to wells infected with CHIKV or CHIKV-GFP in the absence of antibody. Non-linear regression analysis was performed, and IC50 was calculated using Prism 8 software (GraphPad Software, La Jolla, CA).

Immunofluorescence microscopy and quantitation of extensions.

To quantitate intercellular extensions the indicated cells were seeded onto glass coverslips at 2 × 10⁵ cells/well in 6-well plates, cultured for ~24 h at 37 °C, inoculated with CHIKV or CHIKV-GFP at an MOI of 0.5 for 2 h at 37 °C, and then cultured for 11 h. Cells were fixed with 4% PFA in PBS for 10 min at room temperature and incubated in permeabilizing buffer (PBS containing 0.1% BSA and 0.2% saponin) for 5 min at room temperature. Cells were then stained for 1 h at room temperature with primary antibodies to E2/E1 and tubulin (rabbit pAb to E2/E1⁵¹,1:500; mouse anti-tubulin mAb, 1:500, Developmental Studies Hybridoma Bank, E7) diluted in permeabilizing buffer, washed three times with PBS, and incubated for 1 h with Hoechst dye, and the appropriate secondary antibodies (antibodies and dilutions as listed above). The coverslips were washed three times with PBS and mounted with Prolong Glass Antifade Mounting medium (Invitrogen, Waltham MA). Confocal images were acquired using 40X oil immersion lens and a Leica TCS SP5 microscope in the Einstein Analytical Imaging Facility.

To quantitate extensions, ≥130 infected cells (or randomly selected mock-infected cells) were counted for each sample of 3 independent experiments. Intercellular extensions were identified as tubulin- and E2/E1-positive structures >10 μM in length that emanated from infected cells and made physical contact with a target cell. The length of extensions was measured using ImageJ Version1.5c⁵³.

Cell-to-cell transmission assay.

Producer cells were seeded at 0.5 × 10⁵ cells/well in 12-well plates, cultured for ~24 h at 37 °C, and inoculated with CHIKV-GFP or CHIKV at a multiplicity of infection (MOI) of 10 for 2 h at 37 °C. During this time, ~1 × 10⁷ target cells in a 10 cm dish were stained with 2.5 μg/mL CMRA dye in Opti-MEM for 30 min at room temperature, washed with PBS for four times and harvested using trypsin-EDTA solution. At 2 h post-infection, producer cells were washed four times with PBS and 2 × 10⁵ CMRA-stained target cells were plated onto the producer cells in the presence of the indicated concentrations of antibodies. At 12 h after the start of co-culture the medium was collected and the cells were analyzed by flow cytometry.

Flow cytometry.

For flow cytometry 1 × 10⁴ cells were analyzed using a BD LSR-II analyzer (BD Biosciences, San Jose, CA, USA) in the Einstein Flow Cytometry core. The flow data were processed using FlowJo 10.2 software.

To quantitate infection of target cells by CHIKV-GFP, cells were harvested using trypsin-EDTA solution, washed two times with PBS, fixed with 2% PFA for 10 min at room temperature, washed two times with PBS, and analyzed by flow cytometry. Untreated cells, infected cells, and CMRA-stained cells were used to calculate the compensation and delineate the gates for flow analysis (Extended Data Fig. 1d). To quantitate infection of target cells by CHIKV, cells were harvested and fixed as above and incubated with rabbit pAb⁵¹ against E2/E1 diluted 1:200 in staining buffer (15 mM HEPES pH 7.0 and 2% FBS in PBS) for 1h at 4 °C. Cells were washed two times with staining buffer and incubated with goat anti-rabbit Alexa Fluor 647 (1:500, ThermoFisher Scientific, A21244) for 30 min at 4°C, washed twice with staining buffer and analyzed by flow cytometry.

For cell surface staining of MXRA8, equivalent numbers of MEF or MEF-MXRA8-KO cells were harvested, washed twice with staining buffer and incubated with 2.5 μg/mL anti-mouse MXRA8 mAb for 30 min at 4°C. After washing with staining buffer two times, cells were stained with goat anti-Armenian hamster-Alexa Fluor 647 (1:200, Abcam, ab173004) for 30 min at 4 °C, washed twice and fixed with 2 % PFA for 10 min at room temperature, then washed twice and analyzed by flow cytometry. For surface staining of tetherin, equivalent number of T-REx U-2 OS parent cells or TET-inducible U-2 OS cells were seeded one day before and then treated with 1 μg/ml tetracycline for 16 h. Cells were harvested, washed twice with staining buffer and incubated with PE-conjugated anti-human tetherin antibody (1:10, eBioscience, 12–3179-42) for 30 min at 4°C. Cells were washed twice and fixed with 2 % PFA for 10 min at room temperature, then washed twice and analyzed by flow cytometry with FACSDiva software V9.0. The same concentration of control antibody (rabbit IgG, ThermoFisher Scientific, 31188) or no antibody were used to delineate the gates for flow analysis.

Cell surface ELISA.

Binding of CHIKV mAbs to the surface of CHIKV-infected cells was quantitated by ELISA: MEF cells were seeded at 1.5 × 10⁴ cells/well in 96-well plates, cultured for ~24 h at 37 °C, infected with CHIKV-GFP at an MOI of 5 for 3 h at 37 °C, and cultured for 8 h. Cells were fixed with 4% fresh PFA in PBS for 10 min at room temperature without permeabilization. Cells were then washed and incubated with serial dilutions of mAbs in staining buffer (1% BSA in PBS) for 1 h at room temperature, and stained for 1 h with horseradish peroxidase (HRP)-conjugated goat anti-mouse (1:1000, SeraCare, 5450–0011) or goat anti-human IgG (1:1000, SeraCare, 5450–0009). Cells were washed three times with PBS, incubated with Ultra-TMB colorimetric substrate (Thermo Fisher #34028) for 20 min, and the reaction stopped by the addition of H₂SO₄ to a final concentration of 1N. Absorbance at 450 was acquired on a plate reader within 30 min.

Isolation and infection of primary mouse ankle cells.

4-week-old WT C57BL/6 mice (000664; Jackson Laboratories) were euthanized, ankles were dissected, and single-cell suspensions were generated by horizontal shaking of tissue at 37°C for 2 h with 5 mm glass beads in digestion media consisting of DMEM plus 10%FBS, 100 U/mL of penicillin-streptomycin, 2.5 mg/mL collagenase type I (Worthington Biochemical) and 1.7 mg/mL DNase I (Roche). The cell suspension was filtered through a 100 μm sterile filter, washed 1x with PBS, and resuspended in DMEM plus 10% FBS and 100 U/mL of penicillin-streptomycin (D:10 media). Approximately 1.5 × 10⁶ ankle tissue cells were added to a single well of a 6-well plate containing four 12-mm coverslips and 4 mL of D:10 media and cultured at 37°C for 24 h. Media were aspirated and cells were absorbed with 3 × 10⁶ PFU of WT CHIKV or 181/25 CHIKV in 200 μl PBS with 1% FBS, 1 mM Ca²⁺, and 1 mM Mg²⁺ for one hour with gentle rocking. Cells were washed 1x with PBS and then incubated for 16 h at 37°C in 4 mL D:10. Coverslips were fixed in 4% PFA for 15 min at room temperature, and ILE quantitated by staining and confocal microscopy as described above.

MEF preparation for adoptive transfer into mice.

Single-cell suspensions of MEFs were loaded with CellTrace violet (Violet) (Invitrogen; C34557) using the manufacturer’s protocols. Briefly, MEFs were resuspended at a concentration of 1 × 10⁶ cells/mL in PBS supplemented with 2 μM Violet and incubated at 37°C for 20 min. Violet loading was quenched by the addition of D:10 media, and cells were incubated for 5 min at room temperature and washed two additional times in D:10 media. Violet-loaded MEFs were seeded into T75 tissue culture flasks at 5 × 10⁶ cells per flask and cultured overnight. Cells were then inoculated for 2 h with WT CHIKV-Venus (MOI of 10 FFU/cell) diluted in DMEM supplemented with 2% FBS and 10 mM HEPES, or mock-infected. Cell monolayers were washed three times with D:10, then cultured in D:10 supplemented with chCHK-152 at 5 μg/mL for one hour. Cells were collected by trypsinization, washed twice with PBS, and resuspended in PBS prior to adoptive transfer into mice. Under these conditions, ~30% of the MEFs were infected and no residual virus was detected in the supernatant.

Adoptive transfer of MEFs, cell retrieval, and flow cytometry.

4-week-old male WT C57BL/6 mice (000664; Jackson Laboratories) were pretreated by i.p. injection with 100 μg of chCHK-152 N297Q or an equal volume of PBS. At 6 h post-treatment, mice were adoptively transferred with 1 × 10⁶ of WT CHIKV-Venus-infected or mock-infected MEFs via footpad injection. As controls, separate cohorts of mice were inoculated with 10³ PFU of WT CHIKV-Venus via footpad injection. At 24 h following adoptive transfer or infection, mice were euthanized and perfused with 5 mL of 1x PBS. Ipsilateral ankle tissue was harvested and incubated with 5 mm glass beads in digestion media containing RPMI media (Hyclone; SH30027.01) supplemented with 10% FBS, 100 U/mL of penicillin-streptomycin, 2.5 mg/mL collagenase type I (Worthington Biochemical) and 1.7 mg/mL DNase I (Roche) with vigorous shaking at 37°C for two hours. Cells were clarified by passing through 100 μm cell strainers and cleared of red-blood cells by lysis with ammonium chloride-potassium red blood cell lysis buffer. Single-cell suspensions were stained with the fixable viability stain Zombie NIR (Biolegend; 423105). To segregate hematopoietic and non-hematopoietic host cells, cells were stained with anti-mouse FcγRIII/FcγRII (Ly-17, 1:400, Biolegend, 101320) and anti-mouse CD45 (30-F11, 1:800, BD Biosciences, 564279). Following staining, cells were washed three times and fixed in 1x PBS containing 1% paraformaldehyde. Samples were processed on a BD Fortessa X-20 flow cytometer. Analysis and plot representation of flow cytometry data was performed using FlowJo 10.8.1 (Becton Dickinson).

Animals were assigned randomly to experimental groups. No mice or data points were excluded. Data collection and analysis were not performed blind due to the conditions of the experiments. No statistical methods were used to predetermine sample sizes, but our sample sizes are similar to those reported in previous publications^35,42.

Statistical analysis.

Statistical significance was determined using unpaired two-tailed t-tests or unpaired two-tailed multiple t-tests in GraphPad Prism, version 8. The number of replicates per experiment and the P-values are indicated in the figures and legends. NS, not significant. Data distribution was assumed to be normal but this was not formally tested.

Extended Data

Extended Data Fig. 1.

Growth properties of CHIKV and CHIKV-GFP in MEF cells.

Extended Data Fig. 2.

Infection of MEF or U-2 OS target cells in the presence of mAb DEN-4G2 or chCHK-152.

Extended Data Fig. 3.

Effect of Bafilomycin or Dyngo-4a on ILE formation.

Extended Data Fig. 4.

Characterization of inducible Rab5 U-2 OS cell lines.

Extended Data Fig. 5.

ILE formation in U-2 OS-tetherin cell lines.

Extended Data Fig. 6.

MAb E10–18 does not inhibit free virus infection of target cells.

Extended Data Fig. 7.

E1-K61T-G64S escapes chCHK-166 inhibition of ILE formation.

Extended Data Fig. 8.

CHK-152 N297Q pre-treatment can block CHIKV infection in ipsilateral ankle.

Extended Data Fig. 9.

FACS gating strategies for Fig. 6.

Extended Data Fig. 10.

ILE formation in primary joint cells infected with CHIKV ex vivo.

Data Availability.

All data supporting the findings of this study are available within the paper, its Extended data or Source data files. Source data are provided with this paper. Representative microscopy images are included in the main or extended data figures. Additional microscopy image files are available from the corresponding author upon request.