Protection Against Malaria by Intravenous Immunization with a Nonreplicating Sporozoite Vaccine

Malaria Sporozoite Vaccine Each year, hundreds of millions of people are infected with Plasmodium falciparum, the mosquito-borne parasite that causes malaria. A preventative vaccine is greatly needed. Seder et al. (p. 1359, published online 8 August; see the Perspective by Good) now report the results from a phase I clinical trial where subjects were immunized intravenously with a whole, attenuated sporozoite vaccine. Three of 9 subjects who received four doses and zero of 6 subjects who received five doses of the vaccine went on to develop malaria after controlled malaria infection. Both antibody titers and cellular immune responses correlated positively with the dose of vaccine received, suggesting that both arms of the adaptive immune response may have participated in the observed protection. Intravenous immunization with an attenuated whole malaria sporozoite vaccine protected volunteers in a phase I clinical trial. [Also see Perspective by Good] Consistent, high-level, vaccine-induced protection against human malaria has only been achieved by inoculation of Plasmodium falciparum (Pf) sporozoites (SPZ) by mosquito bites. We report that the PfSPZ Vaccine—composed of attenuated, aseptic, purified, cryopreserved PfSPZ—was safe and wel-tolerated when administered four to six times intravenously (IV) to 40 adults. Zero of six subjects receiving five doses and three of nine subjects receiving four doses of 1.35 × 105 PfSPZ Vaccine and five of six nonvaccinated controls developed malaria after controlled human malaria infection (P = 0.015 in the five-dose group and P = 0.028 for overall, both versus controls). PfSPZ-specific antibody and T cell responses were dose-dependent. These data indicate that there is a dose-dependent immunological threshold for establishing high-level protection against malaria that can be achieved with IV administration of a vaccine that is safe and meets regulatory standards.

M alaria control interventions, including insecticide-impregnated bednets, insecticide spraying, and antimalarial drugs, have substantially reduced malaria morbidity and mortality (1). However, despite these measures there were in 2010 an estimated 220 million clinical cases and 0.66 to 1.24 million deaths caused by malaria (1,2). A highly effective vaccine will be ideal for preventing malaria in individuals and eliminating malaria in defined geographic areas. It would optimally target the parasite at asymptomatic, pre-erythrocytic stages (3,4). The World Health Organization malaria vaccine technology roadmap set a vaccine efficacy goal of 80% by 2025 (5). Heretofore, no injectable malaria vaccine candidate has consistently approached that level of efficacy.
RTS,S/AS01 is the most advanced subunit malaria vaccine, protecting~50% of subjects against controlled human malaria infection (CHMI) 2 to 3 weeks after the last dose of vaccine and 22% at 5 months after last immunization (6). In a largescale phase 3 efficacy trial in African infants 6 to 12 weeks of age, RTS,S/AS01 reduced the rates of clinical and severe malaria acquired over a 12-month period by 31.3 and 36.6%, respectively (7).
It has been known for 40 years that it is possible to achieve high-level, sustained, protective immunity against the pre-erythrocytic stages of the parasite through immunization by the bites of >1000 irradiated mosquitoes carrying Plasmodium falciparum (Pf ) sporozoites (SPZ) (8)(9)(10)(11). Advancing this technique beyond administration by mosquitoes required the capacity to manufacture aseptic, radiation-attenuated, metabolically active, purified, cryopreserved PfSPZ for an injectable vaccine that met regulatory standards (12). This was achieved (13,14), and the first clinical trial of the PfSPZ Vaccine, composed of Pf NF54 strain SPZ (15), was conducted in 80 adults (14), who received up to 6 doses of 1.35 × 10 5 PfSPZ Vaccine subcutaneously (SC) or intradermally (ID). The PfSPZ Vaccine was safe and welltolerated but elicited low-level immune responses and minimal protection. We hypothesized that the limited efficacy was due to the inefficiency of ID and SC administration for sufficient presentation of Pf SPZ antigens at critical inductive sites required to achieve a protective immunological threshold (16). Therefore, we immunized nonhuman primates (NHPs) with the PfSPZ Vaccine and showed that intravenous (IV), but not SC, administration of the PfSPZ Vaccine elicited potent and durable PfSPZ-specific T cell responses in peripheral blood and most notably in the liver (14), the likely site of immune protection (17). On the basis of these data, we hypothesized that IV administration of the PfSPZ Vaccine would be protective in humans (14) and conducted a phase 1 clinical trial in order to determine safety, immunogenicity, and protective efficacy against CHMI of IV immunization with the PfSPZ Vaccine (18).

Study Population
Fifty-seven subjects, composed of 40 vaccine recipients, 12 CHMI controls, and five back-up controls, enrolled from October 2011 to October 2012 (baseline characteristics are shown in table S1). Thirty-six of 40 (90%) subjects completed all scheduled vaccinations ( Fig. 1 and fig. S1). Of these 36 subjects, two were immunized with 2 × 10 3 PfSPZ Vaccine per dose so as to assess safety only on schedules without CHMI, whereas at higher doses, CHMI was administered to 17 in July 2012 and to 15 in October 2012. Two subjects in the 1.35 × 10 5 PfSPZ Vaccine-per-dose group did not undergo CHMI because of an unrelated serious adverse event (SAE) and travel, respectively ( Fig. 1).
Transient, asymptomatic increases in aspartate aminotransferase (AST) and/or alanine aminotransferase (ALT), assessed as possibly related to vaccination, were observed in 16 (40%) of vaccinees and were not dose-dependent (table S4 and figs. S2 and S3). Two SAEs occurred: One subject was diagnosed with colon cancer and did not participate in CHMI. After CHMI, one vaccine recipient developed anxiety and headache 14 days after completing chloroquine treatment and was hospitalized overnight for diagnostic purposes. The symptoms, which were attributed to chloroquine, resolved within 1 month.

Vaccine Efficacy CHMI #1
In July 2012, subjects who received 7.5 × 10 3 and 3 × 10 4 PfSPZ Vaccine per dose underwent CHMI~3 weeks after last immunization, and 16 of 17 developed parasitemia (Fig. 2). Among the nine subjects who had received four doses of 3 × 10 4 PfSPZ Vaccine, one did not develop parasitemia, whereas the other eight had a 1.4day prolongation of time to parasitemia by thick blood smear (prepatent period) as compared with the six nonvaccinated controls (P = 0.007, Log-Rank) ( Fig. 2 and table S5). This observation suggests a modest reduction in the numbers of liver stage parasites. The prepatent periods in the 7.5 × 10 3 PfSPZ Vaccine-per-dose group were not significantly different than those of controls ( Fig. 2 and table S5).

Second Vaccine Efficacy CHMI
In October 2012, subjects who received 1.35 × 10 5 PfSPZ Vaccine per dose underwent CHMĨ 3 weeks after last immunization along with six new controls. Five of six controls developed parasitemia. Three of nine subjects in the four-dose group and none of six in the five-dose group developed parasitemia (P = 0.015 for the fivedose group versus controls, Fisher's exact test) ( Fig. 2 and table S5). All subjects who did not develop parasitemia detected with thick blood smear were negative as determined by means of quantitative polymerase chain reaction (PCR) at 28 days after CHMI. In the three vaccinated subjects that became infected, there was a modest delay in the time to positive PCR (table S5). Overall, 12 of 15 subjects immunized with 1.35 × 10 5 PfSPZ Vaccine per dose were protected (P = 0.028) ( Fig. 2 and table S5).

Antibody Responses
Monoclonal antibodies against the major surface protein on SPZ, the circumsporozoite protein (CSP), prevent malaria by blocking SPZ invasion and development in hepatocytes (19). Furthermore, the RTS,S/AS01 malaria vaccine is thought to mediate protection primarily through the induction of high levels of antibodies against PfCSP (20). Antibodies mediate some protection in rodents immunized with irradiated SPZ (21). In human subjects immunized via irradiated PfSPZinfected mosquito bites, PfCSP or PfSPZ antibody titers have been detected by means of enzyme-

RESEARCH ARTICLE
linked immunosorbent assay (ELISA) or immunofluorescence assay (IFA), respectively (22). On the basis of these data, we assessed antibodies against PfCSP and the entire PfSPZ (IFA) as well as performed a functional assay using subject serum so as to assess inhibition of sporozoite invasion (ISI) in a hepatocyte line in vitro. Two weeks after the final vaccination, there was a correlation between the total dosage of PfSPZ Vaccine administered and results of PfCSP ELISA, PfSPZ IFA, and PfSPZ ISI (Spearman R = 0.77, 0.83 and 0.84, respectively; P < 0.001) (Fig. 3, A to C). Among the nine subjects who received four doses of 1.35 × 10 5 PfSPZ Vaccine, antibodies for all three assays were numerically   higher in the six protected than the three nonprotected subjects and significantly higher for the PfSPZ ISI (P = 0.05) (Fig. 3, A to C). Antibodies to asexual or sexual erythrocytic stage parasites and antigens were undetectable, indicating that the irradiated PfSPZ were attenuated and did not develop beyond the early liver stage.

Cellular Immune Responses
Studies in mice and NHPs show that protective immunity induced with irradiated SPZ requires cellular immunity, and most preclinical data indicate that CD8 + T cells and interferon-g (IFN-g) are critical mediators of protection (21,(23)(24)(25). CD4 + T cells, CD3 + gd T cells, and natural killer (NK) cells can also play a role in protection (24,26,27). Therefore, we used multiparameter flow cytometry to assess the frequency of PfSPZspecific IFN-g, interleukin-2 (IL-2), tumor necrosis factor (TNF), or perforin-producing CD3 + CD4 + , CD3 + CD8 + , and CD3 + gd T cells and NK cells from cryopreserved peripheral blood mononuclear cells (PBMCs) prior to immunization and 2 weeks after the final immunization ( fig. S4). The memory phenotype of PfSPZ-specific cells was assessed on the basis of differential expression of the cell-surface markers, CD45RA and CCR7 ( fig. S4, C and G).
We assessed PfSPZ-specific, IFN-g-producing T cells in subjects who received 1.35 × 10 5 PfSPZ Vaccine per injection because of the critical role of IFN-g in mediating attenuated SPZ-induced protective immunity. For PfSPZ-specific CD3 + CD4 + IFN-g-producing T cells, there was no significant difference between protected and unprotected subjects (Fig. 4A). For PfSPZ-specific CD3 + CD8 + IFN-g-producing T cells, protected subjects who received four doses showed a trend toward higher and more consistent responses (Fig. 4A). Although all subjects who received five doses were protected, there was considerable variability in their responses. In terms of memory phenotype, PfSPZ-specific cytokine-producing CD3 + CD4 + T cells were detected in central (T CM ) and effector (T EM ) memory subsets (Fig. 4B and  figs. S4G and S5). In contrast, Pf SPZ-specific CD3 + CD8 + T cells were within effector (T EM ) and terminal effector (T TE ) memory subsets (Fig. 4B  and fig. S4G).
We next characterized the quality of PfSPZspecific T cell cytokine responses. Among PfSPZspecific CD3 + CD4 + T cells, there were three primary populations, comprising IFN-g, IL-2, and TNF; cytometry for PfSPZ-specific cytokine-producing (IFN-g, IL-2, or TNF) memory CD3 + CD4 + and CD3 + CD8 + T cells from all subjects at week 2 after the final immunization with PfSPZ Vaccine after in vitro restimulation (fig. S4). The frequency of PfSPZ-specific memory CD3 + CD4 + and CD3 + CD8 + T cells was calculated as the fraction of cells making any cytokine in response to PfSPZ minus the response to 1% human serum albumin (diluent). (D) A representative dot plot showing PfSPZ-specific IFN-g-, IL-2-, or TNF-producing memory CD3 + CD4 + (top) and CD3 + CD8 + (bottom) T cells from a volunteer who received five doses of 1.35 × 10 5 PfSPZ Vaccine and was protected. (E) The frequency of total PfSPZ-specific cytokine-producing (IFN-g, IL-2, or TNF) memory CD3 + CD4 + (top) and CD3 + CD8 + (bottom) T cells from all subjects within each group. Data in all four graphs show a point for each subject. Horizontal bars denote the medians, and whiskers denote the interquartile ranges.
CD3 + gd T cells from naïve humans respond to malaria antigens after in vitro stimulation (28). Accordingly, we also detected IFN-g in CD3 + gd T cells from the majority of subjects before immunization in response to PfSPZ stimulation ( fig. S4D). Although the proportion of CD3 + gd T cells producing IFN-g in response to PfSPZ stimulation did not change after vaccination ( fig.  S4E), the overall frequency of CD3 + gd T cells increased in subjects who received 1.35 × 10 5 PfSPZ Vaccine per dose (Fig. 5).
Last, to assess cellular immunity with a method commonly reported for other malaria vaccine trials, we used enzyme-linked immunosorbent spot assays to enumerate the frequency of IFN-g responses to PfSPZ and six Pf proteins. PfSPZ-specific responses ranged from 100 to 300 spotforming cells (SFCs) per 10 6 PBMCs and were highest in the 1.35 × 10 5 PfSPZ Vaccine fivedose group (fig. S6A). Responses to stimulation with peptide pools spanning five individual Pf pre-erythrocytic stage antigens were low to undetectable (mostly <100 SFCs per 10 6 PBMCs) but generally higher than to one Pf blood-stage antigen included as a control (MSP1) (fig. S6B).

Discussion
The protection against infection we observed in all of the subjects who received the highest dosage of PfSPZ Vaccine has only been achieved previously through immunization with whole PfSPZ administered through mosquito bites (11, 29). The dose threshold at which we observed the highest rate of protection is consistent with prior studies in which >90% protection was associated with exposure to >1000 irradiated PfSPZinfected mosquitoes (11).
The five-dose group in which all vaccinees were protected differed in two potentially important ways from other groups. These subjects received the highest total number of PfSPZ Vaccine (6.75 × 10 5 ), and there was a 7-week interval between the fourth and fifth doses. We do not know whether the protection in all subjects who received five doses as compared with subjects who received four doses of 1.35 × 10 5 Pf SPZ Vaccine was due to the total dosage of Pf SPZ, the numbers of doses, or the increased interval between the fourth and fifth dose. Studies in NHPs show that extending the time between the fourth and fifth dose to~8 weeks boosted the PfSPZspecific CD8 + T cell responses as compared with that of 4 weeks. Demonstrating that modifications in dose and schedule can achieve a protective immunological threshold in humans will guide future clinical trials designed to optimize the immunization regimen.
Two clinical trials of the PfSPZ Vaccine have established that IV administration of Pf SPZ Vaccine induced superior immunogenicity and protective efficacy as compared with those of SC and ID administration (14). The findings are consistent with prior studies in NHP that show IV administration induced a higher frequency of liver resident CD8 + T cells specific for PfSPZ and antibodies against PfSPZ (14). Although the IV route is routinely used for therapeutic inter-ventions in humans, it has not previously been used for administration of preventive vaccines against infectious diseases. Accordingly, a series of clinical trials of the PfSPZ Vaccine administered with IV injection are now planned in Africa, Europe, and the United States to expand critical data on the vaccine for clinical development as a method to prevent malaria in travelers, military, and other high-risk groups and for mass immunization of populations to eliminate Pf malaria from geographically defined areas. The findings of the current study suggest that the protective immune threshold could potentially be achieved in humans with higher numbers of PfSPZ administered in fewer doses by a different schedule or by other routes, as has been done in mice (14). This could be facilitated with microneedle arrays or other novel devices.
The dose-dependent increase in the frequencies of PfSPZ-specific CD3 + CD4 + , CD3 + CD8 + , and CD3 + gd T cells in the peripheral blood after immunization is consistent with studies of Pf infection and treatment (28). Because CD3 + gd T cells and other IFN-g-producing cells such as NK cells represent a higher proportion of lymphocytes in liver as compared with blood in humans (30), they could contribute to protection.
Although we detected PfSPZ-specific CD3 + CD8 + T cells in 7 of 12 protected subjects at 1.35 × 10 5 PfSPZ Vaccine per dose, five protected subjects had low to undetectable responses. This finding is consistent with data obtained from NHP studies. We reported that some NHPs with a high frequency (~3%) of PfSPZ-specific CD3 + CD8 + T cells in the liver had no detectable responses in blood after IV immunization with PfSPZ Vaccine (14). We speculate that as in mice (21,23,24) and NHPs (25), PfSPZ-specific CD3 + CD8 + T cells are required for protection in most individuals and are primarily confined to the liver because of persistence of parasite antigens after vaccination and/or retained as tissue-resident memory cells (31).
There was a dose-dependent increase in PfSPZspecific antibodies using three assays. Although antibodies may have contributed to protection, such responses may best be used to predict vaccine take and serve as a biomarker or nonmechanistic correlate of protection.
Vaccine efficacy against CHMI in this trial was considerably higher than that induced by subunit vaccines (6,32) and consistent with data generated by immunization with mosquito bites (11). However, the antibody and cellular immune responses induced by the PfSPZ Vaccine against the specific malaria antigens tested were substantially lower than those induced by experimental Pf subunit vaccines (6,(32)(33)(34)(35).
It remains an open question whether protection results from the summation of multiple lowlevel antigen-specific responses or from robust responses to a small number of as-yet unidentified antigens from among the~1000 proteins expressed by sporozoites (36). We show that there are multiple populations of lymphocytes responding to PfSPZ, providing a breadth of effector responses. Furthermore, we speculate that the PfSPZ Vaccine induces immunity to a broad spectrum of antigens among the~1000 expressed by attenuated PfSPZ (36)(37)(38). The capacity to induce a cascade of multiple adaptive and innate effector cells and the breadth of antigenic specificity may explain the advantage of whole PfSPZ vaccines compared with existing subunit vaccines.
Demonstration of high-level protection through IV administration is a critical first step in the development of the PfSPZ Vaccine. Future studies will determine the duration of protection and degree of protection against heterologous strains of Pf, establish immune correlates of protection, and optimize approaches to deliver IV vaccine administration to achieve the coverage in mass administration campaigns needed to eliminate Pf malaria from defined areas.