Evolution of the innate and adaptive immune response in women with acute Zika virus infection

Abstract

Zika virus (ZIKV) is a flavivirus that is closely related to other human pathogens, such as dengue virus (DENV)¹. Primary transmission usually involves Aedes aegypti, which has expanded its distribution range considerably², although rarer infection routes, including mother-to-fetus transmission, sexual contact and blood transfusion, have also been observed^3-7. Primary ZIKV infection is usually asymptomatic or mild in adults, with quickly resolved blood viraemia, but ZIKV might persist for months in saliva, urine, semen, breast milk and the central nervous system^8-12. During a recent ZIKV outbreak in South America, substantial numbers of neurological complications, such as Guillain-Barré syndrome, were reported^13,14 together with cases of microcephaly and associated developmental problems in infants born to women infected with ZIKV during pregnancy^15-20, highlighting the clinical importance of this infection. Analyses of the human immune response to ZIKV are lacking^21-28, but the recent outbreak has provided an opportunity to assess ZIKV immunity using current immunological methods. Here, we comprehensively assess the acute innate and adaptive immune response to ZIKV infection in ten women who were recruited during early infection and followed through reconvalescence. We define a cascade of events that lead to immunological control of ZIKV, with previous exposure to DENV impacting some, but not all, mediators of antiviral immunity.

Extended Data Fig. 1. Patient’s symptoms and ZIKV-RNA detection data

(a) Patient’s symptoms information and associated plasma ZIKV-RNA detection, over time. (b) Quantification of ZIKV RNA in the plasma. Each patient and previous exposure to DENV status, as defined by the positive detection of anti-DENV IgG at symptom onset, is displayed through unique symbols and connecting lines. X-axis, Time (days) represents the time from onset of symptoms. Gray dashed line notes the assay’s detection limit. Data are representative of n=2 independent experiments.

Extended Data Fig. 2. Gating strategies for blood immunophenotyping

Flow cytometry gating strategies for the identification of the different cell subsets of monocytes (NC= non-classical; I= intermediate; C= classical), dendritic cells (pDCs= plasmacytoid dendritic cells; mDCs= myeloid dendritic cells) and MDSCs (myeloid-derived suppressor cells) (a), as well as plasmablasts and activated CD8+ T cells (b).

Extended Data Fig. 3. Changes in immune cell frequencies following acute ZIKV infection

Representative flow-cytometry plots showing changes in the frequency of monocytes (a), dendritic cells (b), plasmablasts (c) and activated CD8+ T cells (d), at different time points following ZIKV infection. Numerical values of the measured frequencies in a total n=10 patients, at different time points (acute n=10; recovery n=5; follow-up n=4) are displayed in Fig. 1d. Times (in days) from onset of symptoms are indicated.

Extended Data Fig. 4. Humoral immune response to acute ZIKV infection

Linear regression analysis to model the relationship between plasma anti-ZIKV IgM and IgA detection signals (n=41) using DIACHECK ELISA assays. (b) Plasma detection of anti-ZIKV IgM antibodies using EUROIMMUNE ELISA assay (left chart) and linear regression analysis to model the relationship between plasma anti-ZIKV IgM detection signals obtained with EUROIMMUNE and DIACHECK ELISA assays (n=41) (right chart). (c) Plasma detection of anti-ZIKV IgG antibodies using EUROIMMUNE ELISA assay (left chart) and linear regression analysis to model the relationship between plasma anti-ZIKV IgG detection signals obtained with EUROIMMUNE and DIACHECK ELISA assays (n=41) (right chart). (a-c) Each patient and previous exposure to DENV status, as defined by the positive detection of anti-DENV IgG at symptom onset, is displayed through unique symbols and connecting lines. Gray dashed lines note assay’s detection limits. X-axis, (Time (days)) represents the time from onset of symptoms. Data are expressed as mean values of OD/CO ratios from two independent experiments. Pearson correlation coefficient R and significance p (two-sided) values are reported from the linear regression analysis performed with GraphPad Prism software.

Extended Data Fig. 5. Titration of anti-ZIKV neutralizing antibodies

Representative titration assays for the detection of anti-ZIKV neutralizing antibodies, overtime. Titers were measured by endpoint titration. A color mapping of the OD/CO ratio values for the detection anti-ZIKV IgM and IgG using EUROIMUNE ELISA assays are indicated. The data are representative of n=2 (patients CR8587, CR8602, CR8663, CR4434, CR8597 and CR8622), n=3 (patients CR4565 and 8603) or n=4 (CR8623) independent experiments.

Extended Data Fig. 6. ZIKV-specific T cell memory differentiation following acute ZIKV infection

T cell memory differentiation based on CCR7 and CD45RA co-expression (naïve: CCR7+CD45RA+; CM: CCR7+CD45RA-; EM: CCR7−CD45RA−; TEMRA: CCR7−CD45RA+). Frequencies of ZIKV-specific CD4+ (a) and CD8+ (b) T cells across the different memory subsets over time, from 5 different patients, are indicated. The analysis was performed on a total of n=6 patients, at different time points (Data from patient CR4965 are available in Fig. 4a,b).

Extended Data Fig. 7. ZIKV-specific T cell functional profiles overtime

Detailed representation of the overlapping pie charts presented in Fig. 4d,e. The data represent the different sub-groups of cytokine secreting and cytotoxic CD154+CD4+ (a) and CD69+CD8+ (b) T cells after stimulation with 15-mer overlapping peptide pools covering all ZIKV-proteins, by ex vivo intracellular cytokine stainings (ICSs). Frequencies of IL-2, TNFa, IFNγ and CD107a co-expressing cells are indicated. Baseline signals of IL-2, TNFα, IFNγ and CD107a-expressing cells from unstimulated controls have been subtracted to the stimulated conditions to allow the visualization of ZIKV-specific CD4+ and CD8+ T cell signals. Only time points with detectable CD154+IFNγ+CD4+ (acute n=24, recovery n=26, follow-up n=29) and CD69+IFNγ+CD8+ T cells (acute n=17, recovery n=19, follow-up n=24) from patients CR4965, CR8623, CR8603 and CR8622, as depicted in Fig. 3a,b, have been used for this analysis. Black bars correspond to the median of expression in each condition.

Extended Data Fig. 8. ZIKV-specific T cell functional profiles across the different viral proteins targeted

Overlapping pie charts describing the polyfunctionality of ZIKV-specific CD4+ (a) and CD8+ (b) T cells according to the ZIKV-overlapping peptide pools used for T cell stimulation and determined by ex vivo intracellular cytokine staining (ICS) as defined in Fig. 3. Baseline signals of TNFα, IL-2, CD107a and IFNγ-producing cells in unstimulated controls have been subtracted from ZIKV-stimulated assays to allow the visualization of ZIKV-specific CD4+ and CD8+ T cell signals. Only time points with detectable CD154+IFNγ+CD4+ (acute n=24, recovery n=26, follow-up n=29) and CD69+IFNγ+CD8+ T cells (acute n=17, recovery n=19, follow-up n=24) from patients CR4965, CR8623, CR8603 and CR8622, as depicted in Fig. 3a,b, were used for this analysis. Distribution of the numbers (n) of T cell responses across the different ZIKV-peptide pools are reported.

Extended Data Fig. 9

General overview of the dynamics of immune responses following acute ZIKV infection in human.

Fig. 1:. Clinical and immunological features of ten ZIKV acutely-infected women.

(a) Complete clinical follow-up of a 31-year-old woman assessed at the Viral Hepatitis Ambulatory Clinic, Oswaldo Cruz Foundation, Rio de Janeiro, Brazil. Representative pictures of maculopapular cutaneous rash on arms and dry peeling skin on hands are depicted, as well as plasma, urine and vaginal fluid detection of ZIKV RNA, and plasma detection of anti-ZIKV IgM, IgA and IgG. (b) ZIKV-infected subjects and peripheral-blood specimen collection over time since symptom onset. Patients with previously documented Dengue (DENV) or Chikungunya (CHIKV) viruses or Yellow-fever (YFV) vaccine exposures are annotated. (c) Color mapping of the number of patients with designated symptoms or clinical signs, and who are positive for ZIKV-RNA detection in the plasma. *patient CR8622 had recent documented CHIKV infection and ZIKV-associated symptoms onset correspond to a rebound in symptoms previously experienced following CHIKV infection, as both infections have similar symptom patterns. (d) Flow cytometry analysis of the frequency of the different monocyte and dendritic cell subsets, as well as of myeloid-derived suppressor cells (MDSCs), plasmablasts and activated (CD38high) CD8+ T cells, in the peripheral blood of ZIKV-infected patients at different time points starting from symptom onset (acute n=10; recovery n=5; follow-up n=4). Flow cytometry gating strategies are depicted in Extended Data Fig. 2. Data are expressed as percentages of the indicated populations. Each patient and previous exposure to DENV status, as defined by the positive detection of anti-DENV IgG at symptom onset, are displayed through unique symbols and connecting lines. Cross symbols represent data from 9 healthy adults, as a control. X-axis, Time (days) represents the time from symptom onset. Abbreviations: pDCs= plasmacytoid dendritic cells; mDCs= myeloid dendritic cells.

Fig. 2:. Dynamics of the humoral response to acute ZIKV infection.

Plasma detection of anti-ZIKV IgM (a), IgA (b) and IgG (c) over time, as determined by ELISA (DIACHECK^®) in 10 ZIKV acutely-infected patients. Data are expressed as mean values of OD/CO ratios from n=2 independent experiments. (d) Median of anti-ZIKV antibody detection signals within DENV-naïve (n=4) and DENV pre-exposed (n=6) patient groups. (e) Titration of anti-ZIKV neutralizing antibodies, overtime. Titers were measured by endpoint titration. The data are expressed as mean values of n=2 (patients CR8587, CR8602, CR8663, CR4434, CR8597 and CR8622), n=3 (patients CR4565 and 8603) or n=4 (CR8623) independent experiments. (f) A linear regression model was applied to fit the relationship between anti-ZIKV IgG detection signal intensities and ZIKV-neutralizing antibodies according to patient’s previous exposure to DENV status (Total: n=36; DENV-naïve: n=15; DENV pre-exposed: n=21). Pearson correlation coefficient R and significance p (two-sided) values are reported from the linear regression analysis performed with GraphPad Prism software. (a-e) Each patient and previous exposure to DENV status, as defined by the positive detection of anti-DENV IgG at symptom onset, is displayed through unique symbols and connecting lines. Gray dashed lines note assay’s detection limits. X-axis, (Time (days)) represents the time from onset of symptoms.

Fig. 3:. Dynamics of the cellular immune response to ZIKV infection.

(a) Representative flow-plots of intra-cellular cytokine staining (ICS) for the detection of ZIKV-specific CD4+ (upper plots) and CD8+ T cells (lower plots) after stimulation with 15-mer overlapping peptide pools covering the structural as well as non-structural ZIKV proteins. Frequencies of activated-INFγ-producing cells are indicated. A positive ICS was determined by the detection of at least 10 ZIKV-specific T cells and a frequency of ZIKV-specific T cells being at least twice the unstimulated signal. (b) Detection of ZIKV-specific CD4+ (blue) and CD8+ (red) T cells over time and across six patients by ICS. Frequencies of activated-INFγ-producing cells are depicted through a blue (for CD4+ T cells) or red (for CD8+ T cells) color mapping and are expressed as percentages of total CD4+ or CD8+ T cells. (c) Frequencies of samples with positive detection of activated-INFγ-producing cells among total samples tested by ICS as grouped according to the ZIKV regions targeted; Structural (S) (n=47) or Non-structural (NS) (n=79) ZIKV proteins. (d) Cell frequencies of activated-INFγ-producing CD4+ (upper charts) or CD8+ (lower charts) T cells targeting Structural (S) or Non-structural (NS) ZIKV proteins. Each dot plot and line represent a single patient (acute n=6; recovery n=6; follow-up n=4).

Fig. 4:. Memory differentiation and polyfunctionality of ZIKV-specific CD4+ and CD8+ T cells.

T cell memory differentiation based on CCR7 and CD45RA co-expression (naïve: CCR7+CD45RA+; CM: CCR7+CD45RA−; EM: CCR7−CD45RA−; TEMRA: CCR7−CD45RA+). Representative flow-cytometry plots from the analysis of n=6 patients are indicated on the upper charts. Frequencies of ZIKV-specific CD4+ (n=8) (a) and CD8+ (n=6) (b) T cells across the different memory subsets over time are indicated on the lower charts for patient CR4965. Data from 5 additional patients are displayed in Extended Data Fig. 6. (c) Representative flow-cytometry plots, from the analysis of n=4 patients, of cytokine production and cytotoxicity within CD154+CD4+ (left panels) and CD69+CD8+ (right panels) T cells following stimulation with 15-mer overlapping peptide pools. Frequencies of TNFα, IL-2, CD107a and IFNγ-producing cells are indicated. (d-e) Overlapping pie charts describing the polyfunctionality of activated CD154+CD4+(d) and CD69+CD8+ (e) T cells following stimulation with ZIKV-overlapping peptide pools over time. Baseline signals of TNFα, IL-2, CD107a and IFNγ-producing cells in unstimulated controls have been subtracted from ZIKV-stimulated assays to allow the visualization of ZIKV-specific CD4+ and CD8+ T cell signals. Only time points with detectable CD154+IFNγ+CD4+ (acute n=24, recovery n=26, follow-up n=29) and CD69+IFNγ+CD8+ T cells (acute n=17, recovery n=19, follow-up n=24) from patients CR4965, CR8623, CR8603 and CR8622, as depicted in Fig. 3b, have been used for this analysis. Data represents the median of expression across the different ZIKV-specific T cell responses detected. More detailed representations of the data used in the overlapping pie charts are available in Extended Data Fig. 7.