A Statistical Interaction between Circumsporozoite Protein-Specific T Cell and Antibody Responses and Risk of Clinical Malaria Episodes following Vaccination with RTS,S/AS01E

The candidate malaria vaccine RTS,S/AS01E provides significant but partial protection from clinical malaria. On in vitro circumsporozoite protein (CSP) peptide stimulation and intra-cellular cytokine staining of whole blood taken from 407 5–17 month-old children in a phase IIb trial of RTS,S/AS01E, we identified significantly increased frequencies of two CSP-specific CD4+ T cells phenotypes among RTS,S/AS01E vaccinees (IFNγ-IL2+TNF− and IFNγ-IL2+TNF+ CD4+ T cells), and increased frequency of IFNγ-IL2-TNF+ CD4+ T cells after natural exposure. All these T cells phenotypes were individually associated with reductions in the risk of clinical malaria, but IFNγ-IL2-TNF+ CD4+ T cells independently predicted reduced risk of clinical malaria on multi-variable analysis (HR = 0.29, 95% confidence intervals 0.15–0.54, p<0.0005). Furthermore, there was a strongly significant synergistic interaction between CSP-specific IFNγ-IL2-TNF+ CD4+ T cells and anti-CSP antibodies in determining protection against clinical malaria (p = 0.002). Vaccination strategies that combine potent cellular and antibody responses may enhance protection against malaria.


Introduction
The current lead candidate malaria vaccine is RTS,S/AS01 E [1]. The RTS,S antigen consists of the C-terminal region of the P falciparum CSP including 19 copies of the central tandem repeats, fused to the hepatitis B surface antigen (HBsAg), co-expressed with unfused HBsAg in Saccharomyces cerevisiae cells. The RTS,S antigen has been formulated with different adjuvants to enhance immunogenicity [2,3]. AS01 contains the immunostimulants monophosphorly lipid A and QS21 in liposomes. RTS,S, formulated with AS01 and at a paediatric dose, is referred to as RTS,S/AS01 E .
The vaccine induces high concentrations and frequencies of antibodies and CD4+ T cells, respectively, specific for CSP [4,5]. Anti-CSP antibodies correlate with protection against infection in malaria-naïve-adult challenge studies [4] and field studies in young children [6], against clinical malaria in trials with young children in Kenya/Tanzania [7] and in Gabon/Ghana/Tanzania [8], but anti-CSP antibodies did not correlate with protection against clinical malaria in a trial with older children in Mozambique [9]. Anti-CSP antibodies could protect by a variety of mechanisms including complement activation, antibody dependent cellular cytotoxicity, sporozoite neutralization, and/or FccR mediated phagocytosis [10]. CD4+ T cells might mediate protection indirectly by providing help to B cells for the production of highly effective anti-CSP Abs, or directly by secreting effector/cytotoxic cytokines (e.g. TNF or IFNc) [11,12].
The correlations between CD4+ T cell responses and clinical outcomes are not consistent in the literature, and this may reflect the different clinical settings (ranging from challenge studies in malaria-naïve adults [4] to Phase II field studies in African children [7]) and/or the different methods used to measure vaccine induced T cell cytokine responses (including from ex vivo or cultured ELISpots [13] and intracellular cytokine staining (ICS) performed on isolated PBMC [4] or ICS on whole blood assays [7,14,15]).
Correlations between polyfunctional T cell phenotypes and protection against malaria infection have been reported in challenge studies [4,16], and recently, these results have been extended to identify central memory and effector/effector memory subpopulations, both of which secreted high levels of IL2, and whose frequencies were elevated in the protected relative to the unprotected groups [16].
We have previously reported our findings using a whole blood ICS assay to assess cellular responses after vaccination with RTS,S/AS01 E in a field trial of 447 5-17 month-old children in Kenya [7]. In that previous report, we were unable to assess polyfunctionality of T cell phenotypes, but nevertheless identified an association between the frequency of CD4+ T cells producing at least TNF on stimulation with CSP peptides and protection against clinical malaria. We have now conducted a further analysis of the flow cytometry (FACS) data using alternative software to identify polyfunctional CD4+ T cell responses, and tested for the associations of T cell phenotype with protection from clinical malaria in Kenyan children vaccinated with RTS,S/AS01 E .

Re-analysis and quality control
We-reanalysed the FACS acquired data, following ICS, done previously in a randomized controlled trial of the candidate malaria vaccine RTS,S/AS01 E in 447 5-17 month-old children in Kilifi, Kenya [7]. Samples were stained with fluorescently labelled monoclonal antibodies to IL2, TNF and IFNc in addition to T cell markers (i.e., CD3, CD4+ and CD8+). We did not include CD40L as two previous studies using the same whole blood assay had found CD40L to be undetectable in T cells in samples taken from African children [14,15].
From the full dataset, 6 (0.5% of 1200) samples failed quality control because of high background (.5% IFNc+ CD4+ or CD8+ T cells on media-only control conditions), and 38 samples failed because the positive control (i.e. staphylococcal enterotoxin B (SEB) stimulation) did not result in a four-fold increase in IFNc+ CD4+ T cells over media-only control. There were too few CD4+ T cell events acquired (,10,000) from 2 samples and too few CD8+ T cell events (,5,000) from 10 samples.
After these exclusions, data were available from 1,104 samples for CD4+ cells and 1,100 samples for CD8+ T cells. Cytokine responses were expressed as frequencies of positive cells out of 1 million CD4 or CD8 T cells. The mean background (i.e. IFNc+ cells on media-only control) was 40 per million CD4+ T cells and 90 per million CD8+ T cells. The mean positive control results (i.e. IFNc+ cells on SEB stimulation) were 4,000 per million CD4+ T cells and 4,000 per million CD8+ T cells.
The previous presentation of these data used FACSdiva analysis had been used to identify at least TNF+, IL2+ and IFNc+ CD4+ T cells, without reference to polyfunctionality [7]. The present analysis with Kaluza software identified polyfunctional T cell phenotypes, but numbers of at least TNF+ and at least IL2+ CD4+ T cells could be calculated by summing all the relevant T cell phenotypes.  In order to quality control the analysis of polyfunctionality using Kaluza software, data from 8 volunteers were analysed twice,  blind to subject labelling: the correlation coefficient between these repeated analyses was r = 0.99.

CD4+ and CD8+ T cell cytokine responses
The geometric mean frequencies of responsive CD4 and CD8 T cells (per 1 million CD4/CD8 T cells) results for all time points (i.e. pre-vaccination, 1 months after the last vaccination and 12 months after the last vaccination) and vaccination group are given in table 1, and the distributions of CD4+ T cell responses are shown in figure 1.
The mean frequencies of CSP-specific cytokine positive CD8+ T cells were low, with means of 10 to 50 per million CD8+ T cells, and there were no significant differences between vaccination groups or by time-point. Therefore, data for CD8+ T cells are not described further.

Time course of CD4+ T cell cytokine responses
In both RTS,S/AS01 E and control vaccinees, there were significant increases in the frequencies of CSP-specific IFNc-IL2+TNF2, IFNc-IL2+TNF+, and IFNc-IL2-TNF+ CD4+ T cells during the 4 months between pre-vaccination levels and 1 month post vaccination, and a subsequent decrease in frequencies by 12 months post vaccination ( Figure 1). This temporary increase in CSP specific CD4+ T cell cytokine responses among control vaccinees parallels the increase and decrease in antibodies to blood stage malaria antigens seen in the same children [17] and the timing of the malaria transmission season [17], suggesting that CSP specific CD4+ T cell responses are also acquired naturally following exposure to malaria parasites, as suggested by previous studies in adults [18], and as observed for TRAP specific CD4+ T cell responses following exposure to malaria parasites [19].

Correlates of protection against clinical malaria
We examined the associations between the frequencies of IFNc-IL2+TNF2, IFNc-IL2+TNF+, and IFNc-IL2-TNF+ T cells and T cells expressing at least TNF or at least IL2 on stimulation with CSP and subsequent risk of clinical malaria in the 6 months that followed the measurement. Data from samples taken prevaccination were not included in this analysis.
We examined these associations among the RTS,S/AS01 E vaccinees alone, among the control vaccinees, and among both groups combined (with adjustment for vaccination group). We also examined for statistical interactions between anti-CSP antibody titres and cellular responses in determining risk of clinical malaria (Table 2). However, we could not examine these interactions among the control vaccinees, since antibody responses were detectable in less than 1% of this group, confirming earlier reports that natural exposure to malaria does not induce durable levels anti-CSP antibodies [20,21]. The interaction between TNF+ CD4+ T cells and anti-CSP antibody titers is shown graphically in figure 2, which is based on medians, interquartile ranges, and 5 th to 95 th centiles.
The main effects (i.e. without considering an interaction) of IFNc-IL2+TNF2, IFNc-IL2+TNF+, and IFNc-IL2-TNF+ CD4+ T cells were reductions in the risk of clinical malaria of varying statistical significance. These associations were significant after a Bonferroni correction for IFNc-IL2-TNF+ CD4+ T cells in two of the three cohorts examined (i.e. among RTS,S/AS01 E vaccinees, and among RTS,S/AS01 E and control vaccinees combined), but not among rabies control vaccinees (Table 2, significant results in bold).
The interaction between the effect of IFNc-IL2-TNF+ CD4+ T cell frequency and anti-CSP antibodies was significant after Bonferroni correction among RTS,S/AS01 E vaccinees and controls combined, and significant at p = 0.033 among RTS,S/ AS01 E vaccinees alone. The associations between cells positive for other combinations of cytokines (i.e. IFNc-IL2+TNF2, IFNc-IL2+TNF+, at least IL2+ and at least TNF+ T cells) were smaller in magnitude and less significant than those between IFNc-IL2-TNF+ CD4+ T cells and outcome, and IFNc-IL2-TNF+ CD4+ T cells were the only significant independent cellular responses in multivariate analysis (HR = 0.57, 95%CI 0.39-0.82, p = 0.002).

Discussion
We find that vaccination of malaria-exposed children with RTS,S/AS01 E induces IFNc-IL2+TNF2 and IFNc-IL2+TNF+ CD4+ T cell responses upon in vitro stimulation of whole blood with CSP peptides. In agreement with the results from malarianaïve adult challenge studies [4], we find no evidence of RTS,S/ AS01 E vaccine induced CSP-reactive CD8+ T cell responses. These findings are consistent with studies conducted in Ghana [15], where most responses were IL2+ only CD4+ T cells or IL2+TNF+ CD4+ T cells, and in Gabon [14] where responses were primarily IL2+ only CD4+ T cells with a lower IL2+TNF+ CD4+ T cell response. Although CD40L induction was seen in separated PBMC from malaria-naive adult vaccinees [4], it was not identified in either of the studies in Gabon and Ghana using the whole blood assay, and so was not measured in our study in Kenya using the whole blood assay.
We also show that IFNc-IL2+TNF2, IFNc-IL2+TNF+, and IFNc-IL2-TNF+ CD4+ T cells were induced by natural exposure to malaria in the control vaccinees. The time course of acquisition and loss of these T cell phenotypes among control vaccinees closely parallels the acquisition and loss of antibody responses to blood stage parasite antigens seen our cohort [17].
Over and above the naturally acquired responses, vaccination with RTS,S/AS01 E induced larger and more durable CD4+ T cell cytokine responses for IFNc-IL2+TNF2 and IFNc-IL2+TNF+ CD4+ T cells. However, vaccination with RTS,S/AS01 E did not significantly enhance IFNc-IL2-TNF+ CD4+ T cell responses above those induced by natural exposure to malaria alone. This lack of IFNc production may not be attributable to the concentration of peptides used to stimulate blood for ICS, as Olutu et al [7] used higher concentrations of the same peptides, and similarly could not detect IFNc from ELISpot assays.
There was an association between the frequency of RTS,S/ AS01 E induced CSP-specific CD4+ T cells and protection from clinical malaria, most strongly seen for IFNc-IL2-TNF+ CD4+ T cells. Furthermore, there were significant interactions between CSP-specific TNF+ CD4+ T cell responses and anti-CSP antibodies induced by RTS,S/AS01 E vaccination. This interaction was synergistic, suggesting that the protection afforded by the combination of CD4+ T cells and anti-CSP antibodies is greater than would be predicted by their sum.
These data raise the possibility that naturally acquired cellular immunity interacts synergistically with vaccine-induced antibodymediated immunity to enhance protection. The existence of naturally acquired cellular immunity to CSP is further supported by genetic evidence of variant specific selection pressure among T cell epitopes in CSP [22,23], and a previously described association between CSP-peptide reactive T cells detected by cultured ELISPOT and protection against malaria in both RTS,S/AS01 E vaccinees and control vaccinees [13]. Furthermore, T cells responding to sporozoites may be induced by a single exposure to malaria infection [24]. On the other hand, there was no evidence for selection pressure based on CSP T epitopes in studies of RTS,S vaccinees [25,26], and T cells responding to CSP after vaccination or natural exposure appear to be much lower in frequency than those required to demonstrate protection after vaccination with viral-vectored vaccinations [12,27].
Our study examines associations with protection against clinical malaria and hence we must be cautious in making inferences regarding causality. Nevertheless, it is clear from studies in the field [28,29] and in experimental challenge [4,30] that RTS,S confers protection against asymptomatic infection and in malaria-naive volunteers, and when protection is partial this appears to result from a reduced liver to blood parasite inoculums rather than reduced blood-stage multiplication.
T cells may mediate pre-erythrocytic protection either by their direct effects on parasitized cells [31,32], by stimulating other effector cells including natural killer cells or phagocytes [33], or may be only associated with better quality antibody responses in the absence of a causal role [34]. However, although IL2 production is associated with longer-lived anti-CSP antibodies in our study (as has been found previously, [16]), the frequency of IFNc-IL2-TNF+ CD4+ T cells was not associated with anti-CSP antibody levels, suggesting that this is an unlikely source of confounding.
Synergistic interactions between antibody titers and magnitudes of T cell responses have been identified in mouse models of vaccination [35]. Synergy may occur because antibodies and T cells act as sequential filters, with T cells more able to protect against the reduced numbers of hepatocyte-infecting parasites that are not neutralised by antibodies. Alternatively, TNF may activate phagocytes and other innate cells like NK cells, which then act in conjunction with antibodies in mechanisms that kill and clear the opsonized pathogen [36].
Taking together the observations that TNF+ CD4+ T cells a) are induced by natural exposure to malaria (and therefore likely to vary according to transmission intensity) b) do not correlate with anti-CSP antibodies and c) interact with anti-CSP antibodies to predict outcome, it is not surprising that analyses for correlates of protection in field studies have given varying results in different cohorts [10].
In summary, we conclude that RTS,S/AS01 E induces CSPspecific IFNc-IL2+TNF2 and IFNc-IL2+TNF+ CD4+ T cells. However, IFNc-IL2-TNF+ CD4+ T cells were independently associated with protection against clinical malaria, and were induced by natural exposure. Furthermore, these T cells interacted synergistically with anti-CSP antibody to afford greater protection than either immunological response alone. These findings suggest that vaccination strategies that induce stronger cellular and antibody responses will lead to enhanced protection in the field.

Methods
The study protocol and its amendments received ethical and scientific approval from Kenya Medical Research Institute National Ethics Committee, National Institute for Medical Research of Tanzania, the Oxford Tropical Research Ethics Committee, the London School of Hygiene and Tropical Medicine Ethics committee and the Western Institutional Review Board in Seattle. The study was conducted in accordance with the Helsinki Declaration of 1964 (revised 1996) and Good Clinical Practice guidelines and was overseen by an Independent data monitoring committee and local safety monitors. Written informed consent was obtained from parents/guardians using approved Swahili or Giriama consent forms. Illiterate parents thumb printed the consent form, which was then countersigned by an independent, literate witness.
Details on randomization, immunization and surveillance have been published previously [29]. With parental consent, children aged 5-17 months old were randomized to receive either RTS,S/ AS01 E or rabies vaccine in a 1:1 ratio according to 0, 1, 2 month schedule. The primary end point was clinical malaria, defined as the presence of fever (axillary temperature $37.5uC) and P. falciparum parasitaemia $2500/mL. Active and passive surveillance for malaria was conducted by field workers and study personnel at local dispensaries. Children

CSP antibody measurement
Antibodies to the P. falciparum CSP tandem repeat epitope were assessed by ELISA at the Center for Vaccinology, Ghent University Hospital, Belgium. Results were reported in EU/mL. Plates were adsorbed with the recombinant antigen R32LR that contained the sequence [NVDP(NANP)15]2LR [37].

Whole blood ICS assay
As previously published [7], whole blood was stimulated in Kilifi within 2 hours of being drawn. 350 ml of whole blood plus 100 ml of phosphate buffered saline (PBS) was incubated in three different 15 ml Falcon tube, with 1 mg/ml of anti-CD28 anti-CD4+9d monoclonal antibodies (supplied by BD). After 2 hours, Brefeldin A was added to a final concentration of 1 mg/ml and incubation was continued overnight at 37uc61-CO 2 5 to 7%. EDTA was then added to a final concentration at 5 mM, and after 15 minutes 1 ml FACS lysing solution (BD). The positive control was stimulated using Staphylococcal Enterotoxin B (SEB) and negative control was PBS without peptides (media control). CSP antigen peptides were added to the third tube to a final concentration of 1 mg/ml. A set of 32 15-mer, peptides were used, overlapping by 11 amino acids to cover the full length of the CS antigen used in the vaccine (3D7 strain).

Re-analysis of raw FACS data
An automated batch analysis of these data has already been published using the FACSDiva software (BD Biosciences) [7]. We have re-analysed the raw FACS acquired data using Kaluza software (Beckman Coulter) in order to determine polyfunctionality. Data from each individual vaccinee was analysed manually. Briefly, the pre-vaccination data was used to set up the analysis protocol for each sample. Analyses for cytokine production were done for CD3+CD4++, and for CD3+CD8++ Lymphocytes (Figure 3). A gating tree was used to hierarchically identify unique functional subsets of CD4+ T cells based on their of CD4+, IL2, TNF and IFNc.

Statistical analysis
Geometric mean responses are calculated and a Student's T test was performed on log-transformed values to compare between vaccination groups. Cox regression for the primary endpoint (clinical malaria with P. falciparum density $2500/mL) was adjusted for age at first vaccination, village, distance from the health facility, bed net use and anti-CSP antibody levels. Anti-CSP levels were included by dichotomizing concurrent anti-CS titers at 40 EU/ mL, since this had been found to most closely correlate with protection in previous analyses [38]. Cellular responses were analyzed as time-varying covariates, applying the result from the time of the most recent clinic visit. Responses were log transformed to produce normal distributions before inclusion in the Cox regression models. The significance of interactions between cellular and antibody responses was assessed using the log-likelihood ratio test. STATA version 10 was used.