γδ T Cells Are Required for the Induction of Sterile Immunity during Irradiated Sporozoite Vaccinations Free

Malaria caused over 430,000 deaths in 2014 (1), and effective vaccines are urgently required for malaria eradication. PfSPZ Vaccine (Sanaria) confers sterilizing immunity against homologous Plasmodium falciparum infection in malaria-naive individuals (2, 3). The vaccine is composed of radiation-attenuated aseptic purified cryopreserved P. falciparum sporozoites (PfSPZs) (4). Humoral and cellular immune responses are induced after vaccination, but there is still no consensus on the mechanisms of protection and no reliable immune correlates. Ab responses have correlated with protection in United States studies of the PfSPZ Vaccine, but protection is maintained in some individuals even as Ab wanes (5). Intriguingly, γδ T cells expanded in a dose-dependent manner in malaria-naive subjects immunized with PfSPZ Vaccine (3, 5), and the Vδ2 subset of γδ T cells was recently associated with protection in these vaccinees (5), suggesting a role in protective immunity.

γδ T cells constitute 1–5% of total T cells circulating in healthy adults and share features that are common to the innate and adaptive immune systems (6–8). Two major types of γδ T cells in humans are differentiated by the expression of Vδ2 or Vδ1 chains that recognize different classes of Ags. A subset of Vδ1 cells recognizes lipid Ags presented on the CD1d molecule (9), whereas Vδ2 T cells respond to phosphoantigens presented on butyrophilin receptors (10). Vδ2 cells recognize the blood stages of malaria and respond by producing cytokines and lytic molecules that are required to control parasite replication (11–15). In mice, in which the Vδ2 homolog has not been identified, γδ T cells induced by whole-sporozoite (SPZ) vaccination of αβ T cell–deficient animals inhibit intrahepatocytic parasitic development (16). In addition to their role as effectors, Vδ2 T cells can directly prime CD4 and CD8 T cell responses in vitro (17, 18). Hence, γδ T cells may have diverse functions that could contribute to PfSPZ Vaccine–induced responses.

The PfSPZ Vaccine trial that we conducted (ClinicalTrials.gov, number NCT01988636) in Mali, West Africa was the first to assess the efficacy of this vaccine against naturally occurring infection (19). In this article, we show that Vδ2 T cells were significantly higher in Malian adult vaccinees who remained uninfected throughout follow-up. In a mouse model of SPZ vaccination, we find that γδ T cells are required during vaccination, but not at the time of challenge for protection, indicating they do not function as effectors. The absence of γδ T cells during vaccination was associated with reduced accumulation of CD8α⁺ dendritic cells (DCs) in the liver and the ensuing development of αβ T cell responses, including CD8 T cells required for sterile protection; circumsporozoite protein (CSP)-specific Ab responses did not require γδ T cells. The data support a model wherein induction of protective immunity during SPZ vaccination requires γδ T cells and CD8α⁺ DCs.

Eighty-eight subjects gave informed consent and were randomized to receive five doses of Sanaria PfSPZ Vaccine (2.7 × 10⁵ PfSPZ), or normal saline as the placebo control, by direct venous inoculation. Vaccinations were given at 4-wk intervals, with the exception of the fifth vaccination, which was given 8 wk after the fourth vaccination. One milliliter of whole blood was collected in a sodium heparin tube from each volunteer for ex vivo assays 2 wk after the final vaccination. One hundred and fifty microliters of whole blood was stained using the Abs CD3-BV650, CD4-PerCP, CD8–allophycocyanin–H7, γδTCR-PE, CD11a allophycocyanin, CD38 BV421, Vδ2-FITC, CD45RO PECF594, and CD56 PE-Cy7. After RBC lysis using BD FACS Lysing Solution (Becton Dickinson), the cells were washed and acquired on an LSR II flow cytometer equipped with blue (488 nm), red (633 nm), and violet (405 nm) lasers. γδ T cells were enumerated as a percentage of total CD3 T cells 2 wk after the final vaccination. All Abs used in this study are listed in Supplemental Table I.

RNA was purified from whole blood collected in PAXgene tubes from 22 study volunteers (17 vaccinees and 5 placebo controls), 3 d after the final vaccination (day 143), using a PAXgene Blood RNA Kit (QIAGEN), per the manufacturer’s instructions. RNA quantity and quality were measured with a NanoDrop ND-1000 spectrophotometer (NanoDrop Technologies) and a 2100 Bioanalyzer (Agilent), respectively, to confirm a yield > 200 ng RNA with a quality score ≥ 7. RNA sequencing was performed at the National Institutes of Health (NIH) Intramural Sequencing Center on an Illumina HiSEquation 2500 Platform. Raw FASTQ read data were processed using in-house R package DuffyNGS, as originally described (20). Briefly, raw reads pass through a three-stage alignment pipeline: 1) a prealignment stage to filter out unwanted transcripts (e.g., rRNA, mitochondrial RNA, albumin, globin), 2) a main genomic-alignment stage against the genome(s) of interest, and 3) a splice junction–alignment stage against an index of standard and alternative exon splice junctions. All alignments were performed with Bowtie 2 (21), using command line option “–very-sensitive.” BAM files from stages 2 and 3 were combined into read-depth wiggle tracks that record uniquely mapped and multiply mapped reads to each of the forward and reverse strands of the genome(s) at single-nucleotide resolution. Multiply mapped reads were prorated over all highest quality–aligned locations. Gene transcript abundance was then measured by summing total reads landing inside annotated gene boundaries or exon boundaries (user selectable, based on quality of exon annotation of each genome) and expressed as RPKM (22) and raw read counts. Two stringencies of gene abundance were provided using all aligned reads and by using only uniquely aligned reads.

To minimize biases from the choice of algorithm for calling differentially expressed genes, a panel of five differential expression (DE) tools was used: Round Robin (in-house), rank product (23), significance analysis of microarrays (24), EdgeR (25), and DESeq (26). Each DE tool was called with appropriate default parameters and operates on the same set of transcription results, using RPKM abundance units for Round Robin, rank product, and significance analysis of microarrays and raw read count abundance units for DESeq and EdgeR. All five DE results were synthesized, by combining gene DE rank positions across all five DE tools. Specifically, a gene’s rank position in all five results was averaged, using a generalized mean to the 1/2 power, to yield the gene’s final net rank position. Each DE tool’s explicit measurements of DE (fold change) and significance (p value) were similarly combined via appropriate averaging (arithmetic and geometric mean, respectively). The final DE result was sorted by gene net rank position, so the top genes were those found in common by all DE tools. The sequencing data presented in this article have been submitted to the Gene Expression Omnibus (www.ncbi.nlm.nih.gov/geo) under accession number GSE86308.

Ten C57BL/6, ten γδTCR (B6.129P2-Tcrdtm1Mom/J, γδKO), and ten BATF3 (129S-Batf3tm1Kmm/J, BATF3KO) mice were vaccinated three times with 10⁴ irradiated Plasmodium berghei ANKA SPZs (SPZ vaccination) at 2-wk intervals via direct i.v. inoculation. The same vaccinations were given to groups of 10 C57BL/6 mice that received the GL3 mAb (depletes γδ T cells) or the UC310A6 mAb (depletes Vγ4 γδ T cells), either prior to each vaccination or immediately before challenge. For CD8 T cell depletion, mice were administered CD8β mAb (clone 53-5.8; Bio X Cell) 1 d prior to challenge. A separate group of C57BL/6 mice received an isotype-control mAb (B81-3) prior to each vaccination. Whole blood was collected for ex vivo assays prior to vaccination, 3 d after each vaccination, and 1 d before challenge. Five weeks after the last vaccination, five mice in each group were euthanized, and livers and spleens were used for in vitro Ag-stimulation assays and ex vivo staining. The remaining mice were challenged by i.v. injection of 10³ freshly isolated P. berghei SPZs (PbSPZs) and monitored for blood-stage parasites from day 3 onward by microscopy. All experiments were repeated twice.

Livers and spleens were prepared as previously described (27). One million cells, in a final volume of 100 μl, were plated out and stimulated with PbTRAP₁₃₀ (10 μg/ml, Peptide 2.0) or with RPMI 1640 medium containing 0.001% DMSO as the unstimulated control. Brefeldin A (Sigma) was added to the cells 2 h after the initial stimulation and cultured for an additional 16 h at 37°C, 5% CO₂ in an incubator. After the incubation period, the cells were washed and stained with aqua viability dye (Invitrogen) for 15 min. Cells were washed and surface stained with anti-CD3 BV650, CD4 PE-Cy5, CD8 allophycocyanin-Cy7, γδTCR-PE, KLRG1–PE–Cy7, and CD11a-FITC and incubated for 20 min. After washing, the cells were prepared for intracellular staining by the addition of 200 μl of BD Cytofix/Cytoperm Buffer (Becton Dickinson), as per the manufacturer’s instructions. IFN-γ Ab was added to all tubes and incubated for 20 min. Cells were washed and acquired on an LSR II flow cytometer.

One million splenocytes or hepatocytes were washed and stained with the aqua viability dye (Invitrogen) for 15 min. A mixture of Abs containing CD3 PE-Cy5, γδTCR PECF594, CD11c FITC, B220 PE-Cy7, Clec9A PE, CD8 Alexa Fluor 700, H2Db (or HLA Class II [A–E]) allophycocyanin, and XCR1 BV650 was prepared and used to stain the cells at room temperature. Cells were washed and acquired immediately on an LSR II flow cytometer.

The repeat region of the P. berghei CSP was a kind gift from Dr. U. Krych (Walter Reed Army Institute of Research). The CSP peptide was coated at a concentration of 1 μg/ml and incubated overnight at 4°C. Serum collected from immunized mice prior to challenge (study day 65) was used at a 1:100 dilution for the assay. After a 2-h incubation, the plates were developed using alkaline phosphatase–labeled goat anti-mouse IgG (1.0 mg catalog number 075-1806; Kirkegaard & Perry Laboratories) and Phosphatase Substrate tablets (catalog number S0942; Sigma). Plates were read on a SpectraMax 340PC instrument.

Four- to ten-day-old female Anopheles stephensi mosquitoes were a kind gift from the Laboratory of Malaria Vector Research, National Institute of Allergy and Infectious Diseases, NIH. The mosquitoes were infected by allowing them to feed on donor mice (C57BL/6) infected with blood-stage P. berghei ANKA parasites. The infected mosquitoes were reared in an environmental chamber set at 19–21°C, 80% relative humidity. The infected mosquitoes were exposed to 150 Gy using a cesium irradiator to attenuate the PbSPZs and were dissected to harvest PbSPZs from the salivary glands within 2 h after irradiation.

For pairwise analysis, data were analyzed using the Wilcoxon rank-sum test. Survival analysis was done using the log-rank test. DE between groups for the RNA-sequencing analysis was calculated using five independent algorithms (Supplemental Fig. 1). Each algorithm ranks the genes by DE using its own criteria (p value and/or fold change), and rank positions from the five algorithms were averaged (generalized mean to the 1/2 power) to give final DE rankings for each gene. The measurements of DE (fold change) and significance (p value) from the five algorithms were combined via arithmetic and geometric averaging, respectively.

Written informed consent was received from participants prior to inclusion in the study. All animal studies were approved and performed as per the guidelines of the Animal Care and Use Committee of National Institute of Allergy and Infectious Diseases/NIH.

Throughout the PfSPZ Vaccine trial, whole-blood samples were used for ex vivo staining to enumerate the proportions of CD4, CD8, and γδ T cells, as well as of NK cells, and to relate these to protection; however, no significant associations were observed (19). Interestingly, two populations of γδ T cells were visible that had varying expression levels of γδTCR (Fig. 1A, upper panel). The γδTCR^lo population was enriched for CD45RO expression, suggesting that these might represent the Vδ2 subset (28, 29). Therefore, we enumerated the Vδ2 subset after the last vaccine dose and assessed the relationship to protection. Costaining confirmed that the γδTCR^lo subset was enriched for Vδ2 γδ T cells (Fig. 1A).

The percentage of Vδ2 T cells 2 wk after the last vaccine dose was significantly higher (p = 0.005, Wilcoxon rank-sum test) in vaccinees who remained uninfected throughout the transmission season (median 3.255%, interquartile range [IQR] 1.688–6.725%) compared with vaccinees who became infected (median 1.82%, IQR 1.365–2.40%) or with controls (median 1.85%, IQR 0.98–3.43%, p = 0.008) (Fig. 1B). Similarly, the absolute numbers of Vδ2 T cells were higher in protected vaccinees compared with vaccinees who became infected or with controls (Supplemental Fig. 1A). Levels of non-Vδ2 γδ T cells did not differ among groups (p = 0.908, Fig. 1C). Activation of Vδ2 cells (measured by CD38 coexpression, Supplemental Fig. 1B) was significantly higher in the vaccinated group compared with the control group (p = 0.04, Supplemental Fig. 1C) 2 wk after the fifth vaccination, but it did not differ between the uninfected and infected vaccinees. The data indicate that Vδ2 cell numbers predict protection against naturally occurring infection, similar to the relationship seen between Vδ2 percentages and protection from homologous controlled human malaria infection in United States vaccinees (5).

To characterize global differences in host response profiles, RNA was purified and sequenced from whole blood of vaccinees who remained uninfected (n = 12), vaccinees who became infected (n = 6), and unvaccinated subjects (n = 7), 3 d after the last vaccination. Strikingly, the Vγ9 and Vδ2 genes were ranked as the most upregulated genes, being highest in the uninfected vaccinees, and their expression profile across subjects was most similar by hierarchical clustering (Fig. 1D). These data supported the findings from the ex vivo assays that Vδ2 T cells were associated with protection. Other genes in the top 15 included those associated with regulation of proliferative responses (TGM2, SIPR5), a marker of plasma cells and activated γδ T cells (SLAMF7) (28), and activating receptor for NK cells (NCR1).

To elucidate the role of γδ T cells, we used a mouse model of sterile protection against Plasmodium challenge that requires CD8 T cells for protection (30, 31). Vaccination with irradiated PbSPZs was carried out in wild-type C57BL/6 mice, γδKO mice, and C57BL/6 mice depleted of all γδ T cells or the Vγ4 subset alone, using the GL3 or UC310A6 mAb, respectively (Table I). The total number of circulating WBCs measured 3 d after the last vaccination was similar among mice in the different groups (Supplemental Fig. 2A).

Five mice in each group were challenged with 10³ freshly isolated PbSPZs 5 wk after the third vaccination. All unvaccinated naive mice became infected, and blood-stage parasites appeared at day 3 postinfection, whereas all of the wild-type and control mAb–treated C57BL/6 mice were protected and showed no evidence of blood-stage parasitemia during 14 d of follow-up after challenge. All of the γδKO mice became parasitemic by day 5 postchallenge, establishing that γδ T cells are required for sterile immunity. As previously reported (32), BATF3KO mice, which lack CD8α⁺ DCs, also failed to develop protective immunity (Fig. 2A). Administration of GL3 mAb or UC310A6 mAb resulted in complete depletion of total γδ T cells or the Vγ4 subset, respectively (Fig. 2B). γδ T cells were required for the induction of effective immunity, because C57BL/6 mice administered the GL3 mAb during vaccination, but not at the time of challenge, showed no protection (Fig. 2C). All mice that received the UC310A6 mAb, either during vaccination or prior to challenge, were protected after challenge, indicating that the Vγ4 subset of γδ T cells is not required for induction of protective immunity (Fig. 2D). The same results were obtained in a repeat experiment. To confirm that CD8 T cells were required for protection, previously protected animals had their CD8 T cells depleted (Fig. 2E) and then were rechallenged. As expected, depletion of CD8 T cells resulted in the loss of protection (Fig. 2F) (30).

We examined the requirement for γδ T cells during vaccination to induce αβ T cell and Ab responses. Activated CD4 and CD8 T cells were defined by the coexpression of CD11a and KLRG1, as previously described (27, 31). The levels of activated T cells were measured in the blood of all mice 3 d after each vaccination and immediately before challenge (day 66).

There was no expansion of either T cell subset in BATF3KO mice. Levels of activated CD8 and CD4 T cells were significantly higher in γδKO mice or in C57BL/6 mice treated with GL3 than in BATF3KO mice (p < 0.001, Mann–Whitney U test) but were significantly lower than in wild-type C57BL/6 and UC310A6-treated C57BL/6 mice (Fig. 3B, 3D, p < 0.001 for all comparisons). Ab responses were measured against the immunodominant repeat region of P. berghei CSP in γδKO mice and wild-type mice, using sera collected on study day 66 (prechallenge). Ab titers (expressed as OD values) in wild-type or γδKO mice did not differ significantly (Fig. 3E). Taken together, the evidence indicates that γδ T cells are required for the induction of protective CD8 T cell responses but not of CSP-specific Ab responses to SPZ vaccination.

We next determined the levels of Ag-specific T cell responses in livers and spleens of vaccinated γδKO and wild-type C57BL/6 mice. Cells from livers and spleens were stimulated overnight with the PbTRAP₁₃₀ peptide (33), and CD8⁺IFN-γ⁺ cells were enumerated. Interestingly, PbTRAP₁₃₀-specific CD8 T cells were in the CD11a^hi subset (Fig. 4A) and were significantly more frequent in the livers of C57BL/6 mice (median 2.3%, IQR 1.64–4.89%) compared with γδKO mice (median −0.19%, IQR −1.19– 0.33%) (p = 0.0001, Fig. 4B). Similarly, the percentage of PbTRAP₁₃₀-specific CD8⁺IFN-γ⁺ T cells was higher in the spleens of C57BL/6 mice compared with γδKO mice (median 1.1 versus 0.30%, p = 0.016, Fig. 4B). These data demonstrate that accumulation of Plasmodium-specific T cells is impaired in the absence of γδ T cells during SPZ vaccination.

The above data showed that γδ T cells and CD8α⁺ DCs are required to induce protective immune effectors. We sought to clarify their independent or interacting roles for induction of protective immune responses. CD8α⁺ DCs were enumerated in the livers of wild-type C57BL/6 mice, BATF3KO mice, GL3-treated C57BL/6 mice, and γδKO mice. CD8α⁺⁺ DCs were defined as CD11c⁺CD8⁺B220⁻CD3⁻NK1.1⁻ (Fig. 5A), as previously described (32). Following vaccination, CD8α⁺ DCs were increased in the livers of wild-type C57BL/6 mice compared with unvaccinated mice. Similarly, vaccinated γδKO and GL3-treated C57BL/6 mice had substantially reduced percentages (p = 0.016, Fig. 5B) and numbers (p = 0.04, Supplemental Fig. 2B) of CD8α⁺ DCs compared with wild-type C57BL/6 mice. The results suggest that γδ T cells are required for accumulation of CD8α⁺ DCs in the liver during the response to SPZ vaccination.

Vaccine confers sterilizing immunity against homologous infection in malaria-naive subjects (3, 5). In the recently concluded PfSPZ trial in Mali, we demonstrated vaccine-induced protection against naturally occurring infections (19), which allowed us to examine correlates of immunity. In malaria-naive vaccines, CSP Ab responses and fold increases in γδ T cell levels associated with protection against homologous challenge (3, 5). In Malian vaccinees who remained uninfected throughout our follow-up, neither response was convincing as a correlate of protection. The anti-CSP Ab titers were modest compared with United States vaccinees, and total γδ T cell levels were not significantly elevated (19). In this study, we used human data and animal models to provide definitive evidence that a subset of γδ T cells is required for the induction of protective CD8 T cell responses but do not themselves function as effectors mediating protection.

During the Mali trial, levels of γδ T cells were enumerated ex vivo along with other immune subsets using peripheral blood samples. The Vδ2 subset had lower expression of the γδTCR and a subset expressed CD45RO, which is consistent with a memory phenotype (28, 29). This prompted us to measure the Vδ2 subset in all study participants after the final vaccination. Vaccinees who remained uninfected during follow-up had a significantly higher percentage of Vδ2 T cells than vaccinees that became infected or the placebo group. Strikingly, the Vδ2 and Vγ9 transcript levels were also the most highly upregulated genes in the vaccinees who remained uninfected, independently supporting the association of Vδ2 T cells with protection. Our results concur with the recent finding that the expansion of Vδ2 T cells during vaccination of malaria-naive vaccinees is greatest among those who become protected from infection with homologous parasites (5). Taken together, these results highlight that Vδ2 T cells could be a valuable biomarker during PfSPZ vaccination in malaria-endemic populations.

γδ T cells may have diverse functions during vaccinations. For example, bacillus Calmette-Guérin–expanded human Vδ2 T cells can function as effectors (34), as accessory cells for APCs, or as APCs (35). Vδ2 T cells can be activated by P. falciparum (12, 36) and lyse infected erythrocytes in their role as effectors (13). In addition, it has been shown that, in vitro, Vδ2 T cells can induce CD8 T cell responses to P. falciparum Ags (37). To elucidate the precise role of γδ T cells during PfSPZ vaccination, we used a mouse model of SPZ vaccination that confers sterile protection and is dependent on CD8 T cells. The γδ T cell subpopulation in mice that corresponds to the human Vδ2 T cell subset has not been defined; therefore, all γδ T cells were depleted initially. These experiments clearly demonstrated that γδ T cells were not required as effector cells, inasmuch as depletion of these cells immediately before infectious challenge had no impact on protection.

Instead, γδ T cells were essential during vaccination to induce effector CD8 T cells that mediated sterile protection. In the absence of γδ T cells during vaccination, there was diminished activation of CD8 T cells in the periphery and reduced numbers of PbTRAP₁₃₀-specific CD8 T cells in the liver and the spleen that are required for sterile immunity (38). Because we did not specifically stain for markers of tissue-resident T cells, we cannot exclude the possibility that some of these responses are due to circulating effector memory T cells.

The absence of γδ T cells did not affect Ab titers to the immunodominant P. berghei CSP repeat region. Although γδ T cells can assist in induction of humoral responses (39–41), previous studies showed that Ab titers are not impaired in γδ T cell–deficient mice after oral immunization (42), similar to what we have observed in our model. These results clearly show that induction of protective CD8 T cells, but not CSP-specific Abs, is dependent on the presence of γδ T cells during SPZ vaccination. A caveat to our results is the possibility that the GL3 mAb we used may not deplete γδ T cells, but instead leads to downregulation of the TCR, as reported by one study (43). We found consistent results in the γδΚΟ mice and C57BL/6 mice that received the GL3 mAb. In both of these mice, SPZ vaccination did not lead to induction of effector CD8 T cells; consequently, they were not protected after challenge. Also, studies in multiple models have demonstrated that, at the very least, treatment with GL3 leads to functional impairment of γδT cells (44, 45). The fact that sterile immunity was not perturbed in mice that were treated with the GL3 mAb immediately prior to challenge strongly suggests that γδ T cells are not required as effectors for control of parasite replication in the liver.

We next attempted to identify the subset of γδ T cells in mice that were responsible for induction of protective immunity. We focused on the Vγ4⁺ subset because they were required for induction of CD8 T cells in a bacillus Calmette-Guérin vaccination model (46). Depletion of Vγ4 γδ T cells in our model had no effect on induction of sterile immunity. Further studies are needed to identify the γδ T cell subset that is required during the induction of effector CD8 T cell responses.

CD8α⁺ DCs are required for protection induced by SPZ vaccines (32, 47), and we replicated those findings in this study. Therefore, we examined the CD8α⁺ DCs in the liver after SPZ vaccine in the absence of γδ T cells. Absence of γδ T cells during vaccinations was associated with a dramatic reduction in CD8α⁺ DC numbers in the livers of vaccinated mice. These findings are consistent with those of Montagna et al. (32), who similarly showed accumulation of CD8α⁺ DCs in the liver in protected mice. This suggests that γδ T cells are required for the expansion of CD8α⁺ DCs in the liver, either by facilitating their migration or supporting their proliferation. Our results are suggestive, and we cannot conclude that there is direct interaction between γδ T cells and CD8α⁺ DCs or that this solely occurs in the liver. Further experiments are required to tease out the precise location and interaction of CD8α⁺ DCs and γδ T cells during the induction of protective CD8 T cell responses that impart sterile immunity. Taken together, the data from this study suggest that cross-talk between γδ T cells and CD8α⁺ DCs is required for induction of downstream effector T cell responses during SPZ vaccination. Although we did not examine the role of CD4 T cells in this study, there is ample evidence to show that they are also required for induction of protective CD8 T cell responses (48).

In summary, we showed that Vδ2 T cells are elevated in vaccinees with prior exposure to malaria who showed evidence of protection against natural infection. In murine vaccination studies, we definitively demonstrated that a subset of γδ T cells, along with CD8α⁺ DCs, are required for the induction of protective immunity by SPZ vaccination, but not as direct effectors. The understanding that γδ T cells are required for induction of protective CD8 T cell responses can guide strategies to improve the efficacy of SPZ vaccines in malaria-exposed populations.

We thank Eric James, Peter Billingsley, Anita Manoj, and Yonas Abebe for PfSPZ Vaccine preparations, quality assurance, and logistics in Mali, and we acknowledge the support from the Sanaria Manufacturing, Quality, Regulatory, and Clinical Teams. We acknowledge the support from Richard Sakai and Sharon Wong-Madden during the vaccine trial in Mali. We thank Jean Langhorne and J. Patrick Gorres for reviewing the manuscript.

The authors have no financial conflicts of interest.