Growth kinetics of CHIKV during co-infection

We infected PBMCs from 3 different donors with one or both viruses at various multiplicities of infection (MOIs) as described in Materials and Methods. DENV and CHIKV production was measured in the supernatants to determine the growth kinetics of the viruses during mono- and co-infection (Fig 1 panels A and B). Interestingly, despite the fact that CHIKV is usually considered to have a shorter replication cycle than DENV (approximately 12 h and 24 h, respectively), both viruses were detected in the cell supernatant as early as 12 hpi. Furthermore, in general, viral titers increased over time however CHIKV production reached a plateau between 24 and 48 hpi while DENV continued to replicate until 72 hpi. At 48 hpi, titers of both viruses were comparable between the donors, although donor B was clearly more susceptible to CHIKV than to DENV as DENV titers were 1 log lower than that of CHIKV at all harvesting times (Fig 1A vs 1B).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Time-course analysis of DENV and CHIKV (co)- infection in hPBMCs cell-free supernatants from PBMCs of three different donors infected with DENV or CHIKV or co-infected with both viruses were collected at 12, 24, 48 and 72hpi and analysed by qPCR to determine the number of specific viral particles expressed as genome equivalent copies per mL (GEc/mL).

(A) Time-course analysis of DENV production during mono-infection at MOI of 1 and during the co-infection with CHIKV at different MOIs. (B) Time-course analysis of CHIKV production during mono-infection at MOI of 1 and during the co-infection with DENV at different MOIs. (C) Effect of co-infection on CHIKV production in hPBMCs, of DENV MOI 1 during co-infection with CHIKV or UV-CHIKV relative to mono-infection. (D) Effect of the co-infection with replicative and non-replicative DENV on CHIKV production in hPBMCs, For C and D, bars represent an average % of infection normalized to relative mono-infections of three donor’s ± SEM.

https://doi.org/10.1371/journal.pntd.0005712.g001

In case of co-infections, higher levels of DENV in the supernatants from co-infected cells were observed when compared to mono-infected cells [Fig 1A and 1C for (co)-infections at DENV MOI of 1 and S1A and S2A Figs for DENV MOI of 5]. Depending on the MOIs, the increase was statistically significant at different time points; for co-infections with DENV at MOI of 1 at 12 and 48 hpi (p<0.05) and for co-infections with DENV at MOI of 5 at 24 hpi (p<0.05) and 72 hpi (p<0.01). Interestingly, although at 12 hpi an increase in DENV replication (p<0.05) was seen during co-infection, a significant decrease (p<0.001) in replication was observed during co-infection with replication-incompetent UV-inactivated CHIKV (UV-CHIKV). At later time point, co-exposure with UV-CHIKV had no effect on DENV infection, suggesting that enhancement of DENV replication during co-infections relied on CHIKV replication. Intriguingly, when CHIKV production was assessed, we noticed a profound decrease of CHIKV titres during co-infections in all donors tested (Fig 1B and 1D for infections at CHIKV MOI of 1 and S1B and S2B Figs for CHIKV MOI of 5). The relative fold-change analysis of data from all the donors (Fig 1D) demonstrated that CHIKV production was significantly reduced from 24 hpi onwards during co-infections with DENV (p<0.05, at 24 hpi and p<0.01 at 48 and 72 hpi). To gain insight into the observed inhibition on CHIKV replication, we next compared the growth of CHIKV during co-infection with UV-DENV. Remarkably, the antagonistic effect of co-infection on CHIKV was independent of DENV replication (Fig 1D). In fact, inhibition was stronger and more consistent between the donors and different MOI conditions in presence of UV-DENV (S2D Fig).

Mixed signature of innate immune responses during DENV/CHIKV co-infection

Next, we sought to assess whether co-infection modulated the innate immune responses of the PBMCs. To this end, we selected a pool of cytokines and chemokines that are considered to play an important role in the confinement and/or pathogenesis of DENV and CHIKV infections [32–36]. We determined the concentrations of IFN-α, IFN-β, IFN-ω, IL-6, IL-8, IP-10, MCP-1 and TNF-α in the cell supernatants at 6, 24 and 48 hpi by multiplex immunoassay. As expected, the concentrations of different cytokines varied considerably between the donors, different multiplicities of infections and time points. Therefore, in Fig 2 we displayed fold-changes in the concentrations of cytokines released during co-infections relative to the corresponding mono-infection for each donor at 24 hpi, the peak time point for the vast majority of the measured cytokines (S4 Fig). When we compared fold-changes between co-infections and mono-infections within a donor, the directional trend was found to be consistent across the donors. To analyze this in more detail, we modelled log-cytokine expression values using a linear mixed model, allowing a random donor effect and a random interaction between donor and time. This model was used for all cytokines but IFN-β. For this cytokine, an additional interaction between experimental effect and time point was included due to significant, directionally opposite effects occurring between 6–24 hpi and 24–48 hpi (S4 Fig). The results of the tests are summarized in Table 1.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Innate immune signature of the co-infection vs mono-infection.

Chemokine and cytokine levels produced following different mono- and co-infection regimens in hPBMCs were measured by ProcartaPlex. Each graph shows the fold-change of concentration of the immune factor during co-infection relative to the indicated mono-infection. All concentrations shown correspond to 24 hpi, except for that of IFN-ω, which peaked in all donors at 48 hpi.

https://doi.org/10.1371/journal.pntd.0005712.g002

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Significant changes in cytokine levels following CHIKV and DENV mono- and co-infection in hPBMCs.

https://doi.org/10.1371/journal.pntd.0005712.t001

As expected, based on the published literature [25,37,38], CHIKV mono-infection resulted in significant increase of secreted IFN-β (3.08 fold, p = 0.004), IP-10 (2.09 fold, p<0.001) and MCP-1 (2.53 fold, p<0.001) as compared to mock-infection. Interestingly, exposure of the PBMCs to UV-CHIKV also caused a significant increase in production of these immune mediators with additional stimulation of IFN-α (1.71 fold, p<0.001). The only cytokine that was found to be significantly decreased during CHIKV mono-infections was IFN-ω (0.51 fold, p = 0.004). In case of DENV mono-infections, significantly increased production of IFN-β was observed for replicative (4.66 fold, p<0.001) as well as non-replicative virus (9.53 fold, p<0.001). Exposure of the cells to UV-DENV also induced IFN-α (1.33 fold, p<0.05). Furthermore, a modest but significant inhibition of TNF-α (0.63 fold, p = 0.004) during DENV infection. None of the virus mono-infections significantly altered the levels of IL-6 and IL-8.

Immune modulators that increased during co-infections compared to CHIKV mono-infection were IFN-ω (2.8 fold, p<0.001), TNF-α (1.38 fold, p = 0.011) and IFN-β (5.26 fold, p = 0.003 at 24 hpi). IFN-ω levels were higher (2.9 fold, p<0.001) during co-infection with UV-DENV than during CHIKV mono-infection, suggesting that this IFN could have contributed to the suppression of CHIKV during co-infections. The immune modulators that were found in lower concentrations during co-infection compared to CHIKV mono-infection were IL-8 (0.64 fold, p = 0.006), IP-10 (0.51 fold, p<0.001) and MCP-1 (0.43 fold, p<0.001). These two last chemokines, IP-10 (0.57 fold, p<0.001) and MCP-1 (0.40 fold, p<0.001), also had lower concentrations during co-infections with UV-DENV than during CHIKV mono-infections. Finally, cytokines that increased during co-infection when compared to DENV mono-infection were IFN-ω (2.28 fold, p<0.001), TNF-α (1.94 fold, p<0.001) and IL-6 (1.64 fold, p = 0.003).

In order to rule out that the observed changes in the cytokine response was solely due to an increase of the total number of particles added to cells, we compared the fold changes triggered by co-infections vs mono-infections with the same total number of particles (Table 2). Thus, the mono-infections at MOI of 2, 6, 10 were used as controls of the effect of viral particle number on innate responses during co-infections (1:1), (1:5)/(5:1), (5:5), respectively. Importantly, the majority of the cytokines that were significantly altered during co-infection as presented in Fig 2 and summarized in Table 1 did not withstand the particle control analysis. Only IFN-ω, TNF-α, IL-6 and IP-10 were in fact differentially modulated during co-infections. The levels of IFN-ω during co-infection were significantly higher than those of the mono-infections with DENV (3.34 fold, p<0.001) and CHIKV (2.83 fold, p<0.001) with comparable MOIs. The same pattern was found for TNF-α, levels of which were significantly increased in co-infection with 1.86 fold (p = 0.007) and 2.16 fold (p = 0.001) as compared to CHIKV and DENV mono-infections, respectively. Interestingly, the significance in differential modulation of IP-10 and IL-6 depended on which virus mono-infection they were compared to. The level of IP-10 was lower during co-infection (0.66 fold, p = 0.0244) than during particle-matched mono-infection of CHIKV. Yet, when compared to the same amount of DENV particles, there was no significant difference in the IP-10 levels between co- and mono-infections. This implies that co-infection altered the relative IP-10 response to CHIKV but not to DENV. In contrast, IL-6 was upregulated during co-infection when compared to DENV particle-matched mono-infection (1.72 fold, p = 0.047) but not to that of CHIKV. In summary, our data strongly indicate that co-infection modulates the innate immune response of the respective mono-infections.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Changes in cytokine signature: Co-infection vs particle control.

https://doi.org/10.1371/journal.pntd.0005712.t002

Type I IFN response is different in co-infection vs mono-infection and potentially contributes to decreased CHIKV replication during co-infection

Lastly, we sought to test whether suppression of CHIKV infections observed in our experiments could be attributed to differential induction of type I interferons during co-infection (in Table 1). Type I interferons are part of the first line of defense against many viral infections and their rapid induction can limit infection and viral spread [39]. Since inhibition of CHIKV replication during co-infection occurred independently of DENV replication (Fig 1D), we reasoned that co-infections, in particular those with non-replicating UV-DENV, led to higher levels of one or more antiviral interferons that ultimately inhibited CHIKV infection. Therefore, we re-analysed the data to depict changes in the concentrations of individual IFNs (IFN-α, IFN-β and IFN-ω) as well as the collective IFN type I response in co- vs mono-infection (Fig 3A). Indeed, the cumulative level of type I IFN was clearly higher in co-infections than in mono-infections with CHIKV. Since IFN-β and IFN-ω were the most prevalent cytokines in supernatants from mono- and co-infections, we next focussed on their role as potential antiviral effectors. To the best of our knowledge, the antiviral effect of IFN-ω on CHIKV infection has never been evaluated. To assess the antiviral activity of interferons on the individual viruses we used an assay based on IFN-deficient Vero-WHO cells [40]. Briefly, cells were pre-treated with mock- or increasing concentrations of IFN-β or IFN-ω for 24 h and subsequently infected with CHIKV or DENV at various MOIs (Fig 3B and 3C respectively). Interestingly, IFN-ω had no effect on CHIKV infection whereas it significantly inhibited (p = 0.011) DENV infection already at 20 IU/mL (Fig 3B). Hence, IFN-ω might have contributed the transient but significant reduction/delay of DENV production observed at 12 hpi (Fig 1C). As expected [40,41], IFN-β pre-treatment inhibited infection of both viruses. Notably, however, IFN-β pre-treatment had a much stronger effect on CHIKV than on DENV (50-fold vs 8-fold infection inhibition, respectively). Consequently, the differential susceptibility to IFN-β might explain why increased production of IFN-β during co-infection selectively inhibited CHIKV but not DENV infection.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Differential role of type I interferons in CHIKV and DENV infection.

(A) Concentration of IFN-α, IFN-β and IFN-ω in the supernatants of mono- and co-infected cells at 24 hpi. (B) Concentration of IFN-α, IFN-β and IFN-ω in the supernatants of mono- and CHIKV/UV-DENV co-infected cells at 24 hpi. (C/D) Effect of IFN-β and IFN-ω on CHIKV and DENV replication. Vero-WHO cells were pre-treated with IFN-β or IFN-ω for 24 h and infected with CHIKV (C) or DENV (D) at MOI of 1. The number of genome equivalent copies per mL (GEc/mL) in cell-free supernatant harvested at 24 hpi was determined by RT-qPCR. Results are represented as mean ± SEM of three independent experiments.

https://doi.org/10.1371/journal.pntd.0005712.g003

Cells

Vero E6 (a gift from dr. G. Pijlman, Wageningen University) and Vero WHO (ATCC) were cultured in DMEM (Life Technologies) containing 10% Fetal Bovine Serum (FBS), penicillin (100 U/ml), streptomycin (100 μg/ml), 10 mM HEPES, and 200 mM glutamine. PBMCs were maintained in RPMI 1640 medium supplemented with 10% FBS. Human PBMCs were isolated from Buffy coats using standard density gradient centrifugation procedures with Ficoll-Paque PLUS (GE Healthcare), as described previously. The buffy coats were obtained from healthy volunteers with informed consent from Sanquin blood bank, in line with the declaration of Helsinki. The PBMCs were cryopreserved at -196°C.

Viruses and virus titrations

CHIKV (La Reunion OPY1) was a gift from A. Merits (University of Tartu, Estonia), and was produced from infectious cDNA clones and passaged twice in Vero E6 cells [49]. DENV-2 strain 16681 was propagated in Aedes albopictus cell line C6/36, as described before [50]. Both virus preparations were analysed with respect to the infectious titre and the number of genome equivalents copies as described previously. Briefly, the infectivity of DENV was determined by measuring the number of plaque-forming units (PFU) by plaque assay on BHK-15 cells and the number of genome-equivalent copies (GEc) by quantitative RT-PCR (qRT-PCR), as described previously [51,52]. For CHIKV, the infectious virus titer was determined by standard plaque assay on Vero-WHO cells at 37°C and reverse transcriptase quantitative PCR (RT-qPCR) was used to determine the number of genome equivalents copies (GEc) [53].

UV inactivation

Virus inactivation was obtained by 1.5 h incubation of virus aliquots under UVS-28 8 watt Lamp. Inactivation to below level of detection 35 PFU/mL was assessed using standard plaque assay in Vero-WHO cells or BHK-15 as described previously, for CHIKV or DENV, respectively [51,53].

Infection of PBMCs

PBMCs from three different donors were infected with DENV and/or CHIKV multiplicity of infection (MOI) 1 or 5 at 37°C as described previously for mono-infections [38]. Cells were also exposed to UV-inactivated CHIKV and DENV in single exposures or in combination with the other virus replicative form. The mono-infections at MOI of 2, 6, 10 were used as controls of the effect of viral particle number. After 2 h incubation at 37°C, the inoculum was removed and fresh medium was added to the cells. At each indicated time point, cell-free supernatant was collected, divided into 2 aliquots and stored for subsequent analyses of cytokine and virus production.

Cytokines and chemokines determinations

Concentrations of IFN-α, IFN-β, IFN-ω, IL-6, IL-8, MCP-1, IP-10, and TNF-α were measured in cell-free supernatants using ProcartaPlex and human IFN-ω platinum ELISA (both from eBioscience) according to respective manufacturer’s instructions.

Effect of type I IFNs on CHIKV and DENV replication

Vero WHO cells (2x10⁵ cells/well) were (mock)treated with various concentrations of IFN-β (50, 70, 100, 250 and 500 IU/mL) or IFN-ω (20, 40, 100 and 200 IU/mL) (PROSPEC, cat #: CYT-040) at 37°C. After 24 h cells were washed twice with DMEM and (mock)infected with CHIKV-LR or DENV-2 at the MOI of 1at 37°C. At 2 hpi the inoculum was removed, cells were washed once with DMEM and finally resuspended in DMEM supplemented with 5% FBS. Cell-free supernatants were collected after 24 h of incubation at 37°C and stored for subsequent analysis of virus production by qPCR.

Statistics

All data are expressed as mean with bars representing standard error of the mean (SEM), unless specified otherwise. Linear mixed models were estimated using R package ‘lme4’ and contrasts tested with ‘multcomp’. Changes in viral titers were analysed using Student’s t-test in GraphPad Prism 5 application. In all tests, values of *p<0.05, **p<0.01 and ***p<0.001 were considered significant.