|Home | About | Journals | Submit | Contact Us | Français|
Secondary dengue virus (DENV) infection is a major factor contributing to the risk for severe disease, an effect that depends upon the sequence of infection with different DENV serotypes. We previously reported sequence-dependent effects of secondary DENV infection on CD8+ T-cell responses in mice. To further evaluate the effect of infection sequence, we analyzed DENV-specific CD4+ T-cell responses and their relationship to the CD8+ T-cell response. Serotype cross-reactivity of CD4+ T-cell responses also depended upon the sequence of serotypes in this model. Furthermore, adoptive transfer of memory CD4+ T cells altered the response of memory CD8+ T cells to secondary infection. These data demonstrate the interaction of different components of the T-cell response in determining the immunological outcome of secondary DENV infection.
Dengue virus (DENV) is a mosquito-borne flavivirus and the causative agent of dengue fever (DF) and dengue hemorrhagic fever (DHF). There are four immunologically and antigenically distinct serotypes of DENV: DENV-1, DENV-2, DENV-3, and DENV-4. There is a significantly higher risk of DHF during heterologous secondary DENV infections (5,12). Dengue disease severity also depends on the sequence of infection by the serotypes (1,4,12). Immunopathological mechanisms have been proposed to explain the enhanced risk for disease during secondary infection, including a role for cross-reactive memory T lymphocytes.
To investigate the immune responses to secondary DENV infections in a model amenable to experimental manipulation, we infected Balb/c mice sequentially with heterologous DENV serotypes and measured T-cell cytokine responses (3). In this model, secondary infections enhanced peptide-specific CD8+ T-cell responses due to activation of serotype cross-reactive memory T-cells. We showed differences in epitope-specific immune responses depending on infection sequence. We utilized this mouse model to further study the mechanisms that affect the response to different sequences of DENV serotypes.
Among the sequences tested, we observed that DENV-2-immune mice subsequently challenged with DENV-1 showed a marked increase in the immune response to peptide D1/3 NS3 (GYISTRVGM), the variant in DENV-1 and DENV-3 of the immunodominant H-2Kd-restricted epitope in the NS3 protein (11). In contrast, there was no increase in response to this peptide in DENV-2-immune mice rechallenged with DENV-3. This phenomenon was of interest since the amino acid sequence for this peptide was the same for both DENV-1 and DENV-3. Both DENV-1 and DENV-3 induced comparably low CD8+ T-cell responses to this epitope in primary infection. We therefore hypothesized that the differential response in secondary DENV-1 versus DENV-3 infections in DENV-2-immune mice reflected differences in the CD4+ T-cell response. CD4+ T-cell help has been shown to be essential for the generation of effective CD8+ T-cell memory (6,7,9,14).
To assess serotype-dependent differences in the CD4+ T-cell response, we examined the magnitude of this response to primary and secondary DENV infections. Balb/c mice 4–6 wk of age (Jackson Laboratories, Bar Harbor, ME) were immunized with 2×105pfu IP of DENV-1 (strain Hawaii), DENV-2 (strain New Guinea C), or DENV-3 (strain CH53489). For secondary infection, DENV-2-immune mice were challenged 28 d after the primary infection with 2×105pfu IP of DENV-1 or DENV-3. Nine days post primary or secondary infection, the mice were sacrificed and splenectomized and single cell splenocyte suspensions were made. Kinetics of the T-cell responses were previously optimized (3). In addition, homologous rechallenge was also previously found to result in no change in the T-cell response, presumably due to neutralizing antibodies (3).
We measured the frequency of antigen-specific IFN-γ+ CD4+ T cells using intracellular cytokine staining (8). Although Roehrig et al. reported several candidate helper T-cell epitopes on the DENV E protein (10), we did not find any significant and reproducible responses to overlapping peptides corresponding to the E protein. Therefore, we tested the CD4+ T-cell response to whole virus antigens using glutaraldehyde-fixed DENV-infected Vero cell lysates corresponding to each serotype, as previously described (8). Splenocytes (5×105) were incubated overnight with a 1:20 dilution of DENV or control antigen, PMA/ionomycin, or media alone with 5μL/mL brefeldin A (GolgiPlug; BD Bioscience) at 37°C. Cells were permeabilized and stained with anti-mouse CD3 anti-mouse CD4α-FITC, and anti-mouse IFN-γ-APC. Data were acquired on a FACSCalibur in the UMMS Flow Cytometry Core. A small lymphocyte gate was drawn on forward and side-scatter low populations and further gated on CD3+ CD4+ cells.
DENV-specific CD4+ T-cell IFN-γ responses were comparably low during primary infection regardless of the infecting serotype (Fig. 1A–B). Responses in mice that received DENV-3 after DENV-2 were slightly higher (0–6.9% of CD4+ T cells), whereas the highest responses (1.9–10.0%) were observed in mice that received DENV-1 after DENV-2 (Fig. 1A). In all groups, stimulation with DENV-4 and DENV-3 antigens induced the highest percentages of IFN-γ-producing CD4+ T cells. IFN-γ responses to DENV-1 or DENV-2 antigens were not detectable above background after primary DENV infection. A preferential IFN-γ response to heterologous serotypes was observed for human DENV epitope-specific CD8+ T cells (2); however, while the comparison of murine CD4+ T-cell responses to each antigen across immunization groups is valid, we interpret any comparison of responses to antigens of different serotypes with caution, as these experiments used crude cell lysates and we were unable to standardize the content of the relevant DENV antigens. With secondary DENV-1 infection, however, there was a marked increase in response to all four DENV serotypes. The difference in the frequency of IFN-γ+ T cells in response to DENV-2 antigen stimulation was statistically significant; this finding is notable since the primary infection was DENV-2. These results demonstrate a differential CD4+ T-cell response between secondary DENV-1 and secondary DENV-3, especially to the DENV-2 antigen.
A potential explanation for the enhanced DENV-2–reactive CD4+ T-cell response in secondary DENV-1 infection compared to secondary DENV-3 infection was that greater cross-reactive proliferation of DENV-2–reactive CD4+ T cells would occur in response to DENV-1 than to DENV-3. To test this hypothesis, bulk culture lines were generated from splenocytes of DENV-2–immune mice in the memory phase (day 28). To mimic secondary DENV infection, splenocytes (3×107) were stimulated with DENV-1 or DENV-3 antigen at 1:100 dilution in RPMI medium supplemented with 10% FBS and 50U/mL rmIL-2 (BD Bioscience). The cells were assayed for DENV-specific IFN-γ production on days 7 (Fig. 1C and D) and 14 (Fig. 1E and F) post-stimulation by intracellular cytokine staining. Cell lines showed the highest responses to DENV-2 antigen (Fig. 1 D and F). In addition, the frequency of DENV-specific IFN-γ CD4+ T cells was higher in DENV-1 antigen-stimulated cultures than in DENV-3 antigen-stimulated cultures; this difference was statistically significant on day 14 of culture. These results suggest that DENV-2-specific memory CD4+ T cells are more cross-reactive to DENV-1 than DENV-3. Surprisingly, by day 14, there was minimal IFN-γ response to DENV-1 and DENV-3 antigens, which were used for stimulation of the lines. This suggests that the responding cells had higher avidity for DENV-2 than other serotypes, even those used for in vitro expansion. Fig. 1 shows a preferential expansion of DENV-2-specific memory T cells by a heterologous DENV-1 stimulus in vivo and in vitro, suggesting original antigenic sin.
Lastly, we hypothesized that memory CD4+ T cells generated from primary DENV-2 infection that are cross-reactive with DENV-1 and/or DENV-3 provide help for CD8+ T cells upon secondary infection. Therefore, we next examined whether adoptive transfer of memory CD4+ T cells affected the CD8+ T-cell response to secondary infection. CD4+ and CD8+ T cells were isolated (~90% purity) from DENV-2-immune splenocytes by negative selection using magnetic beads (Miltenyi Biotec, Gladbach, Germany). We injected combinations of CD8+ (1×106) and CD4+ (3×106) T cells from DENV-2-immune mice IV in 100μL into naïve mice. The mice were infected the following day with 2×105 pfu IP of DENV-1 or DENV-3. Nine days later, the mice were sacrificed and CD8+ T-cell IFN-γ responses to the D1/3 NS3 and D2/4 NS3 (GYISTRVEM) peptides were measured by intracellular cytokine staining. Cells were stimulated for 5h at 37°C with 10μg/mL peptide. Mice that received both memory CD4+ and CD8+ T cells and were then challenged with either DENV-1 or DENV-3 infection showed a higher CD8+ T-cell response to the D2/4 NS3 peptide than mice that received only CD8+ T cells prior to challenge (Fig. 2). The transfer of both memory CD4+ and CD8+ T cells also yielded a modest, although not statistically significant, increase in the CD8+ T-cell response to the D1/3 NS3 peptide in mice challenged with DENV-1, but not in mice challenged with DENV-3. These data indicate that the transfer of memory CD4+ T cells augmented the DENV-specific memory CD8+ T-cell response to a subsequent heterologous DENV challenge.
Our data parallel the epidemiologic observation that the DENV serotype infection sequence (serotype and/or strain) influences disease outcome (1,4,13), and identify one potential mechanism for this effect. Upon secondary DENV infection, the cross-reactive DENV-specific CD4+ memory cells are stimulated by antigen from the secondary infection. These CD4+ T cells then augment the response of memory CD8+ T cells. In patients, this could result in an increase in the production of inflammatory cytokines and an increased risk for severe disease. The effect of CD4+ T cells on the CD8+ T-cell response in DENV infections has not been addressed in clinical studies; therefore the relationship between this mouse model and DENV infection of humans remains speculative. There may also be other mechanisms that influence the memory CD8+ T-cell response. However, it is known that CD4+ T cells have an important role in the priming of memory CD8+ T cells (6,7,9,14). This mechanism could also be applicable to other diseases such as HIV and influenza.
This work was supported by the National Institutes of Health grants U19 AI57319 and P30 DK032520. Its content are solely the responsibility of the authors and do not necessarily represent the official views of the NIH.
We thank Jurand Janus for the propagation of viruses and preparation of antigens, and Dr. Sharone Green and Dr. Anuja Mathew for their advice and guidance.
The authors state that no competing financial interests exist.