Heterogeneity of the human immune response to malaria infection and vaccination driven by latent cytomegalovirus infection

Abstract

Background: Human immune responses to infection and vaccination are heterogenous, driven by multiple factors including genetics, environmental exposures and personal infection histories. For malaria caused by Plasmodium falciparum parasites, host factors that impact on humoral immunity are poorly understood.

Methods: We investigated the role of latent cytomegalovirus (CMV) on the host immune response to malaria using samples obtained from individuals in previously conducted Phase 1 trials of blood stage P. falciparum Controlled Human Malaria Infection (CHMI) and in a MSP1 vaccine clinical trial. Induced antibody and functions of antibodies, as well as CD4 T cell responses were quantified.

Findings: CMV seropositivity was associated with reduced induction of parasite specific antibodies following malaria infection and vaccination. During infection, reduced antibody induction was associated with modifications to the T -follicular helper (Tfh) cell compartment. CMV seropositivity was associated with a skew towards Tfh1 cell subsets before and after malaria infection, and reduced activation of Tfh2 cells. Protective Tfh2 cell activation was only associated with antibody development in individuals who were CMV seronegative, and a higher proportion of Tfh1 cells was associated with lower antibody development in individuals who were CMV seropositive. During MSP1 vaccination, reduced antibody induction in individuals who were CMV seropositive was associated with CD4 T cell expression of terminal differentiation marker CD57.

Interpretation: These findings suggest that CMV seropositivity may be negatively associated with malaria antibody development. Further studies in larger cohorts, particularly in malaria endemic regions are required to investigate whether CMV infection may modify immunity to malaria gained during infection or vaccination in children.

Funding: Work was funded by National Health and Medical Research Council of Australia, CSL Australia and Snow Medical Foundation. Funders had no role in data generation, writing of manuscript of decision to submit for publication.

Keywords: Antibodies; CD4 T cells; Cytomegalovirus; Malaria.

Fig. 1

Antibodies induced against merozoite by controlled human malaria infection are reduced in adults with latent CMV infection a) Individuals were inoculated with blood stage malaria. IgM, IgG and functional antibodies to merozoite and major merozoite antigen MSP2 were quantified at end of study and categorised as negative, low or high responses with a score 0, 1, 2 respectively. Scores were combined to calculate an antibody score used to capture the total magnitude, breadth and functionality of the responses. Antibody score was stratified CMV serostatus (CMV seronegative n = 19 [NEG], CMV seropositive n = 21 [POS]), (b) In the same individuals, during blood stage malaria, cumulative parasitemia was measured by Area Under Curve (AUC) from day 4 post inoculation to end of study. Data is stratified by CMV serostatus. Data is Tukey boxplots with the median, 25th and 75th percentiles. The upper and lower hinges extend to the largest and smallest values, respectively but no further then 1.5XIQR from the hinge. P are Mann–Whitney U test. See also Supplementary Fig. S1.

Fig. 2

IgG1 and functional antibodies to the merozoite induced by controlled human malaria infection are reduced in CMV infected adults (a) IgG and IgM antibodies, (b) IgG subclass antibodies and levels of functional antibodies that can (c) fix C1q complement component, (d) bind FcγRII and FcγRIII, and (e) drive opsonic phagocytosis by THP1 cells, targeting 3D7 strain whole merozoite parasites, were quantified and stratified by CMV serostatus. For IgG and IgM antibodies were measured at day 0, 8, 14/15 and end of study (EOS) timepoints after controlled human malaria infection. For IgG subclasses and functional antibodies, responses were measured at day 0 and end of study timepoints. CMV seronegative white bars (n = 19), CMV seropositive grey bars (n = 21), data is Tukey boxplots with the median, 25th and 75th percentiles. The upper and lower hinges extend to the largest and smallest values, respectively but no further then 1.5 XIQR from the hinge. P are Mann–Whitney U test. See also Supplementary Fig. S2.

Fig. 3

Tfh cell compartment is skewed to Tfh1 subsets in CMV seropositive adults (a) Activation of Tfh cells as measured by quantifying PD1+, ICOS+, and CD38+ cells (% of Tfh CD4 T cells) on whole blood in individuals during CHMI at day 0, 8, 14/15 and end of study (EOS) timepoints. Data is stratified by CMV serostatus. (b) Tfh cell subsets (% of Tfh CXCR5+ CD4 T cells), stratified by CMV status. (c) Tfh cell subsets as a proportion of activated Tfh cells, stratified by CMV serostatus at day 14/15. CMV seronegative white bars (n = 19 for day 0, 8, 15, n = 17 for day EOS), CMV seropositive grey bars (n = 21 for day 0, 8, 15 and n = 20 for EOS), data is Tukey boxplots with the median, 25th and 75th percentiles. The upper and lower hinges extend to the largest and smallest values, respectively but no further then 1.5XIQR from the hinge. P are Mann–Whitney U test. See also Supplementary Fig. S3.

Fig. 4

Correlation between Tfh cell subsets and antibody development based on CMV serostatus (a) Correlation between activated Tfh2 cells (ICOS + Tfh2 (% of Tfh CXCR5+ CD4 T cells) and (b) the proportion of Tfh1 (% of Tfh CXCR5+ CD4 T cells) at day 0, 8, 14/15, end of study (EOS) and Antibody score measured at EOS. CMV seronegative green (n = 19 for day 0, 8, 15, n = 17 for day EOS), CMV seropositive orange (n = 21 for day 0, 8, 15 and n = 20 for EOS). Spearman's Rho and p are indicated.

Fig. 5

Impact of CMV serostatus on antibody responses to MSP1 vaccination. Antibodies induced by MSP1 vaccination at day 85 were measured to quantify MSP1 targeted IgM, IgG subclasses (IgG1, IgG2, IgG3, IgG4) and functional antibodies that could mediated C1q complement fixation, opsonic phagocytosis (OPA) from THP1 cells and neutrophils, neutrophil antibody dependent respiratory burst (ADRB), and NK production of IFNγ and degranulation (CD107a). Antibody responses were normalised and centred and represented as z-scores. (a) Unbiased clustering of all z-score response in individuals within vaccination cohort Individuals aggregated into low and high responders. CMV status is shown (n = 24). (b) Impact of CMV serostatus on antibody responses for IgG, IgM, C1q, THP1 OPA, neutrophil OPA and neutrophil ADRB (n = 24). Manova p is shown, ttest for C1q is shown. (c) IgG subclasses, and NK responses between individuals who were CMV negative and positive (n = 11–13). (d) Proportion of CD28+ and CD57+ non-Tfh CD4 T cells (CXCR5-) at day 0 and day 85 stratified by CMV infection status. (e) Correlation between frequency of CD57+ non-Tfh CD4 T cells at day 0 and day 85, and C1q antibodies at day 85 following vaccination. p values is Mann Whitney U test for D and for E Spearman's Rho and p are indicated. See also Supplementary Fig. S4.