Background
Despite significant reductions in the prevalence of malaria during the last 15 years [
1], emerging drug and insecticide resistance and the significant ongoing burden of morbidity and mortality emphasize the need for an effective malaria vaccine. Such a vaccine is possible, as radiation-attenuated sporozoites (RAS) administered intravenously (IV) to mice [
2] or by mosquito bite [
3] to mice and non-human primates [
4] induce almost complete sterile protection. During the 1970s, 1980s and early 1990s a series of human studies using
Plasmodium falciparum RAS (PfRAS) delivered by bite of irradiated mosquitoes similarly induced nearly 100 % sterile protection as long as sufficient numbers of immunizing bites were administered [
5‐
9]; since parasitaemia was completely prevented in these volunteers, all clinical manifestations of malaria were avoided. Beginning in 1989, additional human subjects were immunized with PfRAS and the immunological outcomes were extensively published [
10‐
14]. Ten out of ten subjects (100 %) given greater than 1000 bites were fully protected against controlled human malaria infection (CHMI) conducted less than 10 weeks after immunization (one undergoing CHMI at 10 weeks was not protected), six of six (100 %) were protected on repeat CHMI within 10 weeks of primary CHMI, and five of six (83 %) were protected on repeat CHMI within 23–42 weeks of primary CHMI, indicating that protection was durable for at least 10 months [
15]. These studies also showed that protection extended to heterologous strain parasites (parasites genetically and antigenically different from the immunizing strain), as several subjects immunized with an African malaria strain (NF54) were protected against a parasite cloned from a Brazilian isolate (7G8) [
15].
Although these studies provided proof of concept that sporozoites could induce high-level immunity, as a vaccine for human use, PfRAS immunization was deemed impractical for many decades due to the complexity of administering a vaccine via mosquito bite, the requirement for a secure insectary and a laboratory for maintaining
P. falciparum in culture, and the perceived need for five or more immunization sessions to achieve a sufficient number of bites. Recently, it has been demonstrated that the Sanaria PfSPZ vaccine, composed of aseptic, purified, cryopreserved, PfRAS is safe, well-tolerated, easily administered by syringe using a variety of routes, and can induce 100 % protective efficacy against CHMI when administered intravenously [
16,
17]. PfRAS immunization by mosquito bites or by syringe therefore serves as a model for high-grade, cross-strain protective immunity in animals and humans, creating a strong rationale to develop a sub-unit vaccine approach that might provide equivalent protection, if the protective immune mechanisms and targeted antigens could be identified.
The sterile immunity induced by RAS appears to be mediated primarily by CD8+ and CD4+ T cell-dependent mechanisms targeting antigens expressed by sporozoites and liver-stage parasites [
11,
13,
14,
18]. Responses to a liver-stage antigen were also identified in one study [
14]. In rhesus monkeys, and in most murine studies, CD8+ cells were required for protection [
19‐
22]. Murine studies suggest that inflammatory cytokines such as interferon-gamma (IFN-γ) induce the nitric oxide pathway in hepatocytes to kill liver-stage parasites [
23], or that the infected hepatocytes are destroyed by direct cytotoxic activity [
24]. Sporozoite-neutralizing antibodies likely contribute significantly to protection [
10,
25], particularly when CHMI is conducted soon after immunization.
Antibodies also appear to contribute to protection. Studies in mice and humans show that immunization with RAS induces sporozoite-neutralizing antibodies [
10,
16] that recognize the circumsporozoite protein (CSP), an abundant protein forming the surface coat of the sporozoite [
4]. This finding led to the cloning of
P. falciparum CSP [
26,
27] and the formulation of several CSP-based sub-unit vaccines designed to induce protective antibodies [
28,
29]. Although efficacy was low, subsequent development of CSP using a particle-based approach has led to the currently most advanced malaria sub-unit vaccine, RTS,S/AS01, that elicits 30 % protection in young children [
30] primarily mediated by anti-CSP antibodies and CD4+ T cells [
31,
32]. When tested in the field, RTS,S does not induce sterile protection, but rather reduces the frequency of clinical malaria. The lack of CD8+ T cell responses may be one reason. In addition to protein-based sub-unit vaccines, gene-based vaccines have shown promise, particularly in heterologous prime-boost regimens [
33,
34], although efficacy is still well below that of RAS. The partial efficacy of these first generation sub-unit vaccines suggests that a better understanding of RAS-induced protective mechanisms may provide a rationale to develop alternative or improved sub-unit strategies using newly discovered antigens or more potently inducing cell-mediated immunity.
To address this objective, a new human trial using
P. falciparum (strain NF54) RAS was conducted from 1999 to 2002 with the primary goal to obtain samples for investigation of protective immune mechanisms and antigen discovery. Fifteen healthy adult subjects were immunized five to six times by mosquito bite at intervals of 5–7 weeks (with the exact timing based on the availability of sufficient numbers of PfRAS-infected mosquitoes) to achieve a total of at least 1000 bites of irradiated, infected
Anopheles stephensi mosquitoes, and seven received a similar number of non-infectious bites (mock-immunized controls) to control for the effects of mosquito salivary antigens injected during the mosquito bites. Ten of the true-immunized subjects and five non-immunized infectivity controls underwent CHMI with homologous strain (NF54) infectious sporozoites administered by five mosquito bites, to assess protection and to investigate possible correlation between the CHMI outcome and immune responses. Sera, plasma and peripheral blood mononuclear cells (PBMC), using leukapheresis to obtain large numbers of lymphocytes, were collected from all subjects. These samples have already been used to investigate the antigenic targets induced by PfRAS [
35‐
40], underscoring the usefulness of these studies for vaccine development. Here, the safety, tolerability, protection, and humoral response data collected during this clinical trial are reported. The trial increases from 14 to 24 the total number of subjects receiving more than 1000 infectious bites and undergoing CHMI within the published literature.
Methods
Objectives
The objectives of this study were to determine whether a minimum of 1000 bites of irradiated
P. falciparum-infected mosquitoes was safe and well tolerated and would elicit protection in up to 100 % immunized subjects as previously proposed [
15], and to provide immune samples to investigate correlates of protection, including both the immunological responses and the targeted
P. falciparum antigens. Research subjects were enrolled into three groups: (1) ‘true-immunized’ subjects who were immunized by the bites of
An. stephensi mosquitoes infected with irradiated sporozoites, followed by CHMI; (2) ‘mock-immunized’ subjects who were immunized by the bites of uninfected mosquitoes, but not exposed to CHMI; and, (3) infectivity controls who were not immunized but were exposed to CHMI at the same time as the true-immunized subjects in order to prove the viability of the CHMI.
Ethics
The study was conducted at the Naval Medical Research Center (NMRC) Clinical Trials Center between 1999 and 2002. At the time of these studies, the Food and Drug Administration (FDA) did not require the administration of infectious mosquito bites to be conducted under an Investigational New Drug (IND) allowance. This trial was retrospectively registered at ISRCTN ID 17372582. The study protocol was reviewed and approved by the NMRC Institutional Review Board in compliance with all federal regulations governing the protection of human subjects. Walter Reed Army Institute of Research (WRAIR) holds a Federal-wide Assurance from the Office of Human Research Protections (OHRP) under the Department of Health and Human Services as does NMRC. NMRC also holds a Department of Defense/Department of the Navy Federal-wide Assurance for human subject protections. All key personnel were certified as having completed mandatory human research ethics education curricula and training under the direction of the WRAIR Institutional Research Board or the NMRC Office of Research Administration (ORA) and Human Subjects Protections Program (HSPP). All potential study subjects provided written, informed consent before screening and enrolment and had to pass an assessment of understanding.
Study population
Healthy, malaria-naïve, non-pregnant adults between the ages of 18 and 50 were included in this study. Malaria-naïve status was confirmed by travel history, medical history and P. falciparum CSP ELISA screening.
True and mock immunization procedures
The infected mosquito batches used for true-immunization were infected 14–21 days prior to human biting with the chloroquine-sensitive NF54 strain of P. falciparum by membrane feeding on in vitro blood cultures at the Biological Research Institute, Rockville, MD, USA. Monitoring for salivary gland infections was conducted by hand dissection of a representative sample from the batch, grading infection rates as gland scores: 1–10 sporozoites = gland score 1; 11–100 sporozoites = gland score 2; 101–1000 sporozoites = gland score 3; and >1000 = gland score 4. A gland score of two or higher was used as the cut-off to count a mosquito as ‘infected’, although those with ten or fewer sporozoites on dissection could still be infectious and inject sporozoites during feeding. The morning of an immunization procedure, mosquito batches with 70 % or more of mosquitoes showing gland score 2 or higher were transported to NMRC/WRAIR, Silver Spring, MD, USA and subjected to 15,000 cGy using a Model 109-68 Cobalt60 irradiator.
Both true- and mock-immunizations were conducted in the secure WRAIR/NMRC insectary by placing two cylindrical cardboard containers with mosquito netting at one end, each holding approximately 200 mosquitoes, in contact with the volar surface of one forearm for 5 min, followed 2 min later by a second 5-min feed with the same mosquitoes at the same sites. Consistent with previous experience, approximately 70 % of the 400 mosquitoes in the two containers (200 × 2) took a blood meal. After the two 5-min feeding sessions, a sample of the engorged mosquitoes was hand-dissected to calculate infectivity rates. The total number of engorged mosquitoes was multiplied by the per cent of mosquitoes with mean gland grade at least 2 to estimate the dose of infectious bites. The goal for the full immunization series was for the true-immunized group to receive a minimum of 1000 irradiated infectious mosquito bites before CHMI. In practice, this required five to six immunization sessions. Similarly, the goal for the mock-immunized group was to receive a minimum of 1000 non-infectious mosquito bites, with the number of bites from each immunization session calculated in this latter case as the number of engorged mosquitoes. Mosquitoes used for mock immunization were raised, handled and irradiated in the same fashion as those for true immunization except they were not fed on P. falciparum blood cultures. Both true- and mock-immunized subjects were observed on site for at least 30 min after each immunization.
Controlled human malaria infection (CHMI)
Five non-irradiated mosquitoes, infected with the same NF54 strain of P. falciparum used for immunization were allowed to feed once for 5 min on the subjects. All fed mosquitoes were dissected to determine the infectivity rate. Replacement mosquitoes for those of the initial five not feeding or feeding but found on dissection to have gland grades of 1 or less (ten sporozoites or fewer) were then allowed to feed and this process was repeated until five infectious bites had been achieved. Beginning 7 days after CHMI, subjects were assembled each night in a regional hotel for clinical monitoring by study staff. Each morning, thick blood smears were made for microscopic examination, and sufficient passes over the slide were made using the high-power objective, such that approximately 40 µL of blood were examined. The presence of two parasites was required for a positive diagnosis, leading to immediate anti-malarial treatment with chloroquine phosphate. The treatment regimen was directly observed and included 1000 mg chloroquine phosphate salt (600 mg chloroquine phosphate base) immediately, 500 mg salt (300 mg base) at 6 h and again at 24 and 48 h. Subjects who were positive were monitored daily by symptom checks and blood smears until three consecutive negative smears were documented and subjects remaining negative were similarly monitored until day 21 post CHMI, then approximately every other day until day 28. Those remaining negative on day 28 were considered fully protected.
Adverse events (AEs)
Subjects were examined by physical examination and verbal questioning for local adverse events at 24, 48 and 72 h and at 1 and 2 weeks after each immunization. Although specific systemic symptoms were not actively solicited, subjects were asked in open-ended fashion to describe any systemic symptoms to the evaluating clinician, and these were recorded.
AE grading
Local AEs were subjectively graded as follows:
Mild: Minimally apparent symptoms noticed by the study subject (pain, tenderness, pruritus) or signs noticed by the examiner (erythaema, induration, swelling, lymphadenitis) but not requiring treatment.
Moderate: Symptoms or signs quite evident to the study subject (pain, tenderness, pruritus) or the examiner (erythaema, induration, swelling, lymphadenitis), potentially interfering with the activities of daily living (ADLs); treatment offered (i.e., study subject provided with topical corticosteroid cream to apply as needed).
Severe: Clinically significant findings interfering with daily activities; study subject requested or examiner recommended immediate local and/or systemic treatment with topical corticosteroids and/or oral antihistamines/corticosteroids/non-steroidal anti-inflammatory drugs.
Systemic AEs were subjectively graded as follows:
Mild: No treatment required; ADLs not compromised (subject able to work, or attend school).
Moderate: Outpatient treatment required, ADLs only minimally compromised (subject able to work, or attend school). Severe: Outpatient treatment required, ADLs compromised (subject not able to work or attend school). Serious: AEs resulting in death; AEs that were life-threatening, meaning that failure to intervene could result in hospitalization or death (example: bronchospasm requiring parenteral medication in the emergency room, or grand mal seizure evaluated in the emergency room but not resulting in hospitalization); AEs leading to or prolonging inpatient hospitalization; AEs resulting in persistent or significant disability or incapacity, including addiction; congenital anomaly or birth defect in an infant conceived by a subject.
Laboratory tests
Screening clinical laboratory tests were initially collected to determine enrolment eligibility. These tests included a complete blood count (CBC) and screens for hepatitis B virus, hepatitis C virus and human immunodeficiency virus. Once a study subject was enrolled but prior to immunization, additional sampling was performed by withdrawal of whole blood and by leukapheresis, to provide pre-immunization serum, PBMCs and plasma for banking. Additional blood collections were performed at various time points throughout the trial for banking serum, PBMCs and plasma. Leukapheresis was repeated halfway through the immunization series in some subjects and after the final immunization/prior to CHMI in all subjects that underwent CHMI. There was no systematic collection of safety laboratory data beyond screening for enrolment and a CBC prior to each leukapheresis. Additional samples for safety laboratory tests were collected only in study subjects as clinically indicated.
Immunofluorescence antibody assay (IFA) using sporozoites
Serum antibody levels were assessed by IFA against air-dried
P. falciparum 3D7 strain sporozoites; 3D7 is a clone of NF54 obtained by limiting dilution [
41]. To prepare the IFA slides, infected mosquitoes were suspended in 3 % bovine serum albumin (BSA) at a concentration of 10
6 sporozoites per mL. An aliquot of 10 µL containing 10
4 sporozoites was delivered into each well of the antigen slide. The antigen slides were allowed to air dry at room temperature and were kept at −70 °C until used. 20 µL of a twofold serial dilution of test or control serum in PBS containing 2 % BSA was added to each well of the antigen slides. The slides were incubated for 1 h at 37 °C, washed three times in PBS, 5 min each wash. Each well was incubated for 30 min at 37 °C with 20 µL of a 1:50 dilution of FITC-labelled goat anti-human IgG (H
+L) (Kirkegaard and Perry). The slides were washed again, mounted in a Vectashield mounting medium (Vector Laboratories, Inc.) and examined under an Olympus UV microscope and end-point titers were determined as the last dilution above the background that fluorescent parasites were observed.
Enzyme-linked immunosorbent assay (ELISA)
The
P. falciparum recombinant proteins used in the ELISA assays, CSP, SSP2/TRAP, EXP1, and LSA1 have been previously described [
42‐
44]. Stock solutions of
P. falciparum recombinant proteins were diluted in phosphate buffered saline, pH 7.2, to the optimal concentration of each (0.5 µg/mL for CSP, 1.0 µg/mL for SSP2/TRAP, 2.0 µg/mL for EXP1, 4.0 µg/mL for LSA1) as previously described [
45,
46]. The ELISA titre was defined as the calculated serum dilution yielding an optical density of 0.5 in the assay. Samples were considered positive if the titre of the sample post-immunization was greater than the titre plus two standard deviations of the sample pre-immunization and greater than twofold higher than the corresponding pre-immunization sample.
Sample size and statistical assessment
The primary objective of the study was to collect PBMCs, sera and plasma before and after PfRAS immunization, and then to characterize protective immune responses and identify protective antigens for malaria vaccine development by comparing protected and non-protected research subjects. The number of study subjects was constrained by the capacity to generate infected mosquitoes.
The log rank test was used to compare time to parasitaemia between infectivity control and non-protected immunized subjects. The Mann–Whitney U test was used to compare the interval between protected and unprotected subjects and the interval between leukapheresis and CHMI for protected and non-protected subjects. The repeated measure analysis of variance was used to compare the means of the IFA titres between protected and non-protected subjects. The IFA titres were log10 transformed prior to the analysis. The repeated measure analysis of variance was also used to compare the means of the ELISA titres between protected and non-protected subjects. Statistical significance was defined as a two-tailed P ≤ 0.05.
Discussion
Immunization with RAS by mosquito bite established the original gold standard for protection against falciparum malaria in the early 1970s, and earlier studies of PfRAS showed up to 93 % efficacy (13/14 subjects) [
15]. These studies established that a malaria vaccine was feasible, and that sporozoite-specific antibodies and CD4+ and CD8+ T cells were induced by this form of immunization. Murine and simian studies established that CD8+ cells were required for protection [
19‐
21]. PfRAS administered by mosquito bite is not being developed as a human vaccine, although considerable progress is being made using radiation-attenuated, aseptic, purified, cryopreserved sporozoites (PfSPZ Vaccine) administered by direct venous inoculation using a syringe [
16,
17]. The study reported here was undertaken to further elucidate the mechanisms and antigenic targets of protective immunity induced by immunization with PfRAS by mosquito bite through the collection of sera and PBMCs from protected and non-protected subjects. The results of several studies using the sera and PBMCs from this clinical study have been published [
35‐
37,
50,
51]. Here, the method of immunization, safety and tolerability, and antibody responses to selected
P. falciparum antigens are reported.
In earlier PfRAS studies, including a few subjects that received
Plasmodium vivax RAS, immunization was safe and well tolerated, highlighting as common events mild discomfort during mosquito feeding, erythaema, erythaematous papules, focal and sometimes more generalized local swelling, mild headaches and malaise that spontaneously resolved within 24 h [
15]. In the current study, PfRAS immunization was also generally well tolerated, although two subjects experienced significant large local reactions of the forearm (swelling from elbow to wrist) and two other subjects experienced sudden onset systemic symptoms 16 h after immunization. Due to concern regarding the potential for systemic allergic reactions if immunizations were to continue, both cases of large local reaction led to exclusion from further participation. Other than the two large local reactions, local AEs were consistent with prior reports and with reactions to mosquito bites in nature: erythaema, papules, swelling, and induration. Large local reactions have only rarely been reported following mosquito bites in nature [
52] and may relate to the very large antigenic load associated with hundreds of mosquito bites occurring simultaneously during immunization sessions. Because one of the two subjects experiencing a large local reaction to the mosquito bites was mock-immunized and that reaction was very similar to that of the true-immunized subject, it is probable that sporozoites themselves did not contribute to the large local reaction in the true-immunized subject.
Systemic AEs occurred at a lower rate than local AEs and were generally mild, consisting of headache, myalgia, nausea, and low-grade fever. However, two subjects experienced the abrupt onset of symptoms 16 h after immunization and these reactions were consistent with serum sickness-like reactions that can occur when pre-formed antibodies react with administered antigens. One of the subjects was withdrawn, and the second received an additional immunization without recurrence of systemic AEs. Both were true-immunized subjects, precluding differentiation of the effects of mosquito saliva from the effects of sporozoites.
When true- and mock-immunized research subjects were compared overall, there were no qualitative differences in the numbers of local AEs for each subject. More systemic AEs were recorded in the true-immunized group than in the mock-immunized group, but the number of systemic AEs was small in both groups, precluding meaningful comparison. Thus, these data support the hypothesis that the reactogenicity and systemic AEs seen following immunization with PfRAS via mosquito bite result primarily from the complex mixture of mosquito antigens in the saliva [
52,
53]. This is consistent with the minimal reactogenicity reported when aseptic, purified, cryopreserved sporozoites have been administered intradermally, subcutaneously or intravenously in the absence of mosquito saliva [
16,
17,
54‐
61].
PfRAS efficacy appears to be characterized by a threshold effect greater than 1000 infectious bites induced protection in 13/14 (93 %) subjects in the early studies, while protection was only 33 % (five out of 15) in subjects immunized with fewer than 1000 infectious bites, although there were subjects protected after receiving as few as 400 irradiated infectious bites [
15,
62]. A similar threshold has been identified in studies of PfSPZ vaccine [
16]. In this study, only five/ten (50 %) subjects, all of whom received at least 1000 infectious bites, were protected against CHMI. This was an unexpected decrease in sterile protection from the results obtained previously [
15]. Several hypotheses can be advanced to explain this difference. Firstly, the observed variability in sterile protection may be due to host genetics, biting behaviour of mosquitoes, salivary gland sporozoite counts, or other host-vector-parasite interactions. The efficacy induced by 1000 infectious PfRAS bites may range from a likely high-end estimate of >90 % to a low-end estimate of 50 % depending on these factors, particularly when the total sporozoite dose is near the protective threshold. Differences in the immunization and CHMI procedures may also have played a role, such as the longer interval between the final immunization and CHMI in these subjects compared to previous studies (Fig.
7). The efficacy induced by PfRAS may have declined beyond 3 weeks post immunization, due to a waning immune response. As a third hypothesis to explain the reduced protection seen in this study, the leukapheresis procedure used to collect PBMCs prior to CHMI in this study may have depleted the immune response.
The protection induced by PfRAS also appears to be durable. In the prior study, six protected subjects underwent additional CHMI 23–24 weeks after the last exposure to sporozoites, and five out of six were protected, indicating that once established, protection is durable. The same conclusion was reached in a study of PfSPZ Vaccine, where once a subject was protected using an adequate dose of sporozoites, protection against a second CHMI lasted at least 59 weeks [
65]. In the current study, one non-protected subject underwent a second CHMI, but received additional immunizations between the first and second CHMI, and was protected after the second CHMI.
It is noteworthy that the non-protected subjects in this trial all showed a delay to parasitaemia. This delay was interpreted to indicate partial immunity—decreased liver stage parasite burden leading to the release of fewer parasites into the blood and therefore later onset of parasitemia. This inference rests on the assumption that growth rates in the blood were similar for the non-protected true-immunized research subjects and the non-immunized infectivity controls.
As previously reported [
10,
63,
64], PfRAS-immunized subjects developed antibody responses to whole
P. falciparum sporozoites and
P. falciparum CSP. Although antibody responses to other pre-erythrocytic antigens were low or absent, cellular assays have demonstrated the presence of T cell responses to antigens expressed in sporozoites and liver stages (PfCSP, PfTRAP, PfEXP1), and in one study to a liver-(PfLSA1) and blood-stage antigens [
14]. Sera from these PfRAS-immunized subjects have been used to screen protein microarrays to identify novel antigens recognized by these subjects as potentially contributing to protective efficacy [
36,
50], as well as PBMC in cell-free transcription translation strategies [
51]. Recently, sera and PBMC from these subjects were used to identify and characterize a panel of 27 novel
P. falciparum antigens that provides evidence to further evaluate these antigens as candidate vaccines [
35].
While PfRAS administered by infectious mosquito bites is not being further developed as a vaccine, recently administration of radiation-attenuated (metabolically active, non-replicating), aseptic, purified, cryopreserved
P. falciparum sporozoites by intravenous inoculation has achieved 100 % efficacy in human trials [
16]. The dose required (1.35 × 10
5 sporozoites in five doses) was consistent with >1000 PfRAS mosquito bites that elicited up to 93 % protection, and protection last for at least 59 weeks after CHMI as tested in a small number of immunized subjects [
65]. Thus, partial efficacy in mice with radiation-attenuated
Plasmodium berghei first reported in 1967 [
2], has now led 50 years later, to a vaccine shown to be highly effective in clinical trials against
P. falciparum.
Authors’ contributions
DC, DD, SH, and TR conceived and designed the experiments; Clinical Trial PI: DF, LG, TL, and TR; Clinical trial performance: MS, MB, LB, PV, and DC; MS, YC, HG, EA, DD, GB, JA, AK, GJ, and ML performed the experiments; Chart/record review: BH, JL and SR. BH, JL, SR, MH, and TR analysed the data: Intellectual contributions: TN and SL. BH, JL, MH, SH, JE, and TR wrote the paper. All authors read and approved the final manuscript.