Research ArticleMalaria

Complete attenuation of genetically engineered Plasmodium falciparum sporozoites in human subjects

See allHide authors and affiliations

Science Translational Medicine  04 Jan 2017:
Vol. 9, Issue 371, eaad9099
DOI: 10.1126/scitranslmed.aad9099

A triple punch knocks out the malaria parasite

Vaccination with weakened infectious forms of the malaria parasite is the most promising approach to protect against malaria infection. However, creating genetically defined and weakened parasite strains that are safe for vaccination remains challenging. In a new study, Kublin et al. show that genetic engineering of the malaria parasite by the precise removal of three genes creates a parasite strain that infects humans and is well tolerated but cannot cause malaria. These genetically attenuated parasites thus appear safe for vaccination and stimulate the human immune system to generate responses that have the potential to block infection.


Immunization of humans with whole sporozoites confers complete, sterilizing immunity against malaria infection. However, achieving consistent safety while maintaining immunogenicity of whole parasite vaccines remains a formidable challenge. We generated a genetically attenuated Plasmodium falciparum (Pf) malaria parasite by deleting three genes expressed in the pre-erythrocytic stage (Pf p52/p36/sap1). We then tested the safety and immunogenicity of the genetically engineered (Pf GAP3KO) sporozoites in human volunteers. Pf GAP3KO sporozoites were delivered to 10 volunteers using infected mosquito bites with a single exposure consisting of 150 to 200 bites per subject. All subjects remained blood stage–negative and developed inhibitory antibodies to sporozoites. GAP3KO rodent malaria parasites engendered complete, protracted immunity against infectious sporozoite challenge in mice. The results warrant further clinical testing of Pf GAP3KO and its potential development into a vaccine strain.


Fifty percent of the global population lives at risk of malaria infection. In 2015, the World Health Organization (WHO) reported 214 million cases and 438,000 deaths due to malaria, despite major strides in malaria control through increased use of insecticide-treated bed nets, improved diagnostics, and scale-up of antimalarial drug treatments (1). Control efforts remain under constant threat due to insecticide-resistant mosquitoes, drug-resistant parasites, and fatigue in sustaining all control efforts over long periods of time. Because of these challenges, the development of an efficacious malaria vaccine remains a high global health priority, particularly for the malaria parasite Plasmodium falciparum (Pf), which is a major cause of morbidity and mortality among children and pregnant women in sub-Saharan Africa.

Malaria parasite transmission occurs through the bite of an infected Anopheles mosquito and the resulting deposition of a relatively small number of Plasmodium sporozoites into the skin of the host. Sporozoites traverse numerous cells in the skin, enter the bloodstream, exit the bloodstream by traversing the hepatic sinusoidal endothelium, and infect hepatocytes. The sporozoites then transform into liver stages, which go on to replicate asymptomatically within hepatocytes, each generating tens of thousands of merozoites that exit the liver and infect red blood cells (RBCs). RBC infection is cyclical and rapidly expands the parasite population, causing malaria-associated symptoms, pathology, and mortality. It is clear that effectively targeting the pre-erythrocytic sporozoite and liver stages by vaccination would prevent disease and death, and would prevent onward transmission. The most clinically advanced malaria vaccine candidate to date is a subunit vaccine formulation, RTS,S/AS01, which targets the host immune response to a major Pf sporozoite surface protein, the circumsporozoite protein (CSP). A recent phase 3 clinical trial with RTS,S/AS01 showed some efficacy (2, 3), but this was well below the goal of 75% efficacy against malaria, set forth by the WHO Malaria Vaccine Technology Roadmap. Nevertheless, the RTS,S vaccine data show that even partial pre-erythrocytic immunity can reduce clinical malaria episodes. Generating a high degree of sterilizing pre-erythrocytic immunity by vaccination would not only prevent disease and death but also contribute to elimination of malaria.

Numerous animal studies and experimental clinical studies have demonstrated that complete sterilizing immunity to malaria infection can be achieved by immunization with live parasites. However, immunization with whole, live parasites faces a formidable safety problem because they have the potential to cause malaria infection. The best-studied live parasite immunization strategy uses whole radiation-attenuated sporozoites (RAS), in which parasites are subjected to random irradiation-induced DNA damage, thereby preventing parasite replication in the liver (4). To achieve complete protection, multiple immunizations with large numbers of RAS are required (5). Another strategy is immunization with infectious sporozoites and the concurrent administration of antimalarial drugs [chloroquine prophylaxis with sporozoites (CPS)], which allows for completion of liver infection and eliminates asexual parasites once they initiate replication within RBCs. CPS requires about 20-fold fewer sporozoites than RAS to induce complete protection (6) but is limited by the need to provide continuous antimalarial drug cover during immunization.

Advances in genetic engineering of Plasmodium have enabled a third potential path to the development of a whole parasite vaccine, the creation of live-attenuated parasite strains using gene deletions. The concept of genetically attenuated parasites (GAP) has been extensively explored in rodent malaria models, with the discovery of numerous gene knockouts that arrest parasite development at various points during liver infection [reviewed in (7)]. Immunization with GAPs confers long-lasting, complete sterilizing immunity in mice (8). When considered for human immunization, a major advantage of GAPs compared to RAS or CPS is that GAPs constitute a homogeneous population of parasites with a defined attenuation phenotype that can be further modified to ensure immunization safety and induction of optimal protective immunity.

We have created a Pf GAP that carries deletions of three genes that are critical for pre-erythrocytic infection. This Pf GAP3KO (Pf p52/p36/sap1) was unable to complete pre-erythrocytic stage development and transition to blood-stage infection in a humanized mouse model transplanted with human hepatocytes and human RBCs (9). Here, we demonstrate complete attenuation of Pf GAP3KO in human volunteers, show that it elicits functional immune responses in a single-dose immunization, and provide evidence using a rodent malaria model that GAP3KO completely protects against infectious sporozoite challenge.


Complete attenuation of Py GAP3KO in a rodent malaria model

We generated a Plasmodium yoelii (Py) triple gene deletion parasite (Py p52/p36/sap1) to model Pf GAP3KO attenuation and protective efficacy (fig. S1 and table S1). When 30 BALB/cByJ mice (a BALB/c substrain that is extremely susceptible to Py pre-erythrocytic infection) (10) were intravenously challenged with 7.5 × 104 sporozoites of a previously generated Py double gene deletion parasite (p52/p36), we had observed breakthrough blood-stage parasitemia in 4 of 30 mice (9). Therefore, to show that Py GAP3KO was fully attenuated in this stringent test, we injected 30 BALB/cByJ mice intravenously with 1 × 105 Py GAP3KO sporozoites each and monitored for the development of blood-stage parasitemia for 14 days. In contrast to mice injected with wild-type (WT) Py sporozoites that all became blood-stage patent after sporozoite injection, none of the 30 mice injected with Py GAP3KO sporozoites became patently infected (Table 1). Furthermore, five BALB/cByJ mice were injected intravenously with 1 × 106 Py GAP3KO sporozoites and showed no occurrence of blood-stage infection (Table 1). Thus, Py GAP3KO was completely attenuated during pre-erythrocytic infection, and this provided further rationale for evaluating the Pf GAP3KO in a phase 1 clinical trial in human volunteers.

Table 1. Attenuation and immunogenicity of Py GAP3KO parasites.

For attenuation experiments, BALB/cByJ mice were intravenously injected with indicated doses of Py GAP 3KO or Py WT salivary gland sporozoites and monitored for development of blood-stage parasitemia for 15 days. For immunization experiments, BALB/cByJ mice were immunized twice (2×) with indicated doses of Py GAP 3KO or uninfected mosquito salivary gland debris (mock) 2 weeks apart, challenged by intravenous injection of 104 Py WT salivary gland sporozoites at the indicated days. Patency was monitored for 15 days. na, not available.

View this table:

Protection by immunization with Py GAP3KO sporozoites in mice

We next determined whether immunization with Py GAP3KO sporozoites conferred protection against WT Py sporozoite challenge. BALB/cJ mice were immunized intravenously twice with 1 × 104 Py GAP3KO sporozoites 2 weeks apart and then challenged intravenously with 1 × 104 WT Py sporozoites either 7, 30, or 180 days after the final immunization. All mice immunized with Py GAP3KO were completely protected against challenge, whereas all mock-immunized mice became blood-stage patent (Table 1). Thus, a two-dose immunization with Py GAP3KO sporozoites conferred complete protection in the rodent malaria model.

Production of Pf GAP3KO and experimental exposureof volunteers

For experimental administration of Pf GAP3KO to human volunteers by mosquito bite, parasites were thawed from a Pf GAP3KO working cell bank and expanded in normal human erythrocytes using standard culture conditions, from which gametocyte cultures were established. Laboratory-reared Anopheles stephensi mosquitoes were infected with Pf GAP3KO gametocyte cultures using membrane feeds. Evaluation of mosquito infection before the experimental bite administration revealed mosquito salivary gland infection prevalence to be 91%. Pf GAP3KO salivary gland sporozoite loads by a smash test (a rapid test that uses microscopic examination of infected mosquito salivary glands ruptured on glass slides to score sporozoite infection) were rated +2 to +4, indicating robust infection (table S2). Pf GAP3KO clinical run smash test results were compared with multiple previous preclinical runs of Pf GAP3KO, where sporozoite loads were precisely enumerated. This indicated total salivary gland sporozoite loads of >40,000 Pf GAP3KO sporozoites per mosquito in mosquitoes used for bite administration.

All study volunteers were eligible to participate in the trial if they were healthy, malaria naïve, and male or nonpregnant females aged 18 to 50 years in good general health as demonstrated by medical history, physical examination, electrocardiography (ECG), and laboratory assessment within 90 days of enrollment (table S3). The Pf GAP3KO administration and follow-up schedule (Fig. 1) and baseline physical and demographic characteristics of the volunteers (table S4) are shown. All enrolled volunteers who received Pf GAP3KO were free of major protocol violations and completed the final study visit. On the day of administration, 150 to 200 Pf GAP3KO–infected mosquitoes were allowed to feed on the forearm of each subject for a total of 10 min to allow transmission of Pf GAP3KO sporozoites from infected mosquitoes to volunteers (Table 2).

Fig. 1. Schematic showing study design.

Pf GAP3KO sporozoites were delivered to 10 eligible (screened) volunteers via infected mosquito bites with a single administration of 150 to 200 bites per volunteer. After day 7 (day 8 after infection), the volunteers entered the hotel phase, where they were monitored daily until day 18. After the hotel phase, the volunteers entered a 3- to 6-month follow-up phase. D denotes days of clinic visits; observational hotel phase was days 8 to 18.

Table 2. Administration of Pf GAP3KO to volunteers through mosquito bites.

Total number of mosquitoes in cup refers to the number of mosquitoes in the challenge cup that was used for volunteers’ exposure; total number of bites refers to the estimated number of infectious bites volunteers received; proportion of fed mosquitoes refers to the proportion of mosquitoes in the cup that fed on volunteers. Microscopy was used to determine patency (parasite presence); qRT-PCR was used to determine the presence of P. falciparum 18S rRNA/rDNA in the sample tested. A minus sign denotes negative for blood-stage parasites.

View this table:

Full attenuation of Pf GAP3KO confirmed by microscopy and qRT-PCR

All 10 subjects who received Pf GAP3KO completed the 28-day study period without exhibiting any malaria symptoms and remained negative for patent blood-stage parasitemia as demonstrated by microscopic evaluation of peripheral thick blood smears throughout the entire 28-day study period (Table 2). To additionally test for potential subpatent blood-stage infection, we evaluated blood samples from days 0, 7, 10, and 28 after administration for the presence of the P. falciparum 18S ribosomal RNA (rRNA) marker by quantitative reverse transcription polymerase chain reaction (qRT-PCR). At a limit of detection of 20 parasites per milliliter of whole blood, the P. falciparum 18S rRNA was undetectable in all blood samples tested (Table 2). This demonstrated complete attenuation of Pf GAP3KO with developmental arrest during pre-erythrocytic infection.

Local and systemic adverse reactions to mosquito bites and Pf GAP3KO

Most of the adverse reactions reported in this study were related to the large number of mosquito bites sustained during product administration. All local and systemic solicited adverse events reported as primary end points were classified as grade 1 (mild) or grade 2 (moderate) (Figs. 2 and 3). No grade 3 (severe) or grade 4 (potentially life-threatening) adverse events occurred. Seven subjects (70%) experienced grade 2 (moderate) adverse events; most of these were administration site erythema. Other grade 2 events included administration site swelling, administration site pruritus, and fatigue.

Fig. 2. Proportion and severity of solicited adverse events.

The x axis shows the percentage of subjects who did not demonstrate the indicated solicited adverse events (grade 0, green) or who demonstrated a grade 1 (mild severity, yellow) or grade 2 (moderate severity, purple) solicited adverse events.

Fig. 3. Select adverse events for each of the 10 participants.

The x axis indicates the days after GAP3KO administration (day 0 is the day of administration). Study participants are individually listed by the last three digits of the subject ID, and their individual adverse events are color-coded to indicate no adverse events (grade 0, green), grade 1 (mild severity, yellow), or grade 2 (moderate severity, purple). The letter A indicates that the adverse events were reported to be ongoing at the time of the clinic visit.

Three classes of solicited adverse events were evaluated: local reactogenicity, systemic reactogenicity, and malaria signs and symptoms (tables S5 and S6). All subjects experienced at least one solicited local reactogenicity event. The most frequently reported solicited local reactogenicity symptoms were administration site erythema (10 of 10 subjects), administration site swelling (10 of 10 subjects), and pruritus (9 of 10 subjects) (Figs. 2 and 3 and fig. S2). Five subjects experienced at least one solicited systemic reactogenicity event. The most frequently reported solicited systemic adverse events were fatigue (5 of 10 subjects) and headache (3 of 10 subjects). Three subjects experienced grade 1 (mild) solicited potential malaria signs and symptoms, the most frequent of which was fatigue (3 of 10 subjects). However, because no subject was parasitemic at any time throughout the study, these events were later determined to be systemic reactogenicity events that persisted past the time frame during which the adverse events were solicited. Additional details regarding grading and study day resolution of all adverse events are noted in table S7.

All unsolicited adverse events related to Pf GAP3KO administration were considered administration site reactions. Most of these events were unsolicited because they were reported outside the 5-day period during which adverse events were solicited. The most frequently reported events were administration site swelling (10 of 10 subjects), administration site erythema (10 of 10 subjects), and administration site pruritus (7 of 10 subjects). All cases of administration site pruritus had resolved by the end of the 28-day study period; however, administration site swelling (3 of 10 subjects) and administration site erythema (7 of 10 subjects) that did not resolve by the end of the 28-day study period resolved by the 3-month follow-up visit. There were three reports of grade 1 (mild) administration site discoloration observed on day 28 that were reported resolved by the 3-month follow-up visit. An additional subject reported grade 1 administration site discoloration at the 3-month follow-up call that resolved by the 6-month follow-up phone call. There were no clinically significant laboratory toxicities reported. The most frequent laboratory abnormalities were low white blood cell count (3 of 10 subjects), low platelet count (2 of 10 subjects), high white blood cell count (2 of 10 subjects), and high red cell distribution width (2 of 10 subjects) (table S8).

Induction of high levels of IgG antibodies by single administration of Pf GAP3KO

Sera from the volunteers who received the single administration of 150 to 200 Pf GAP3KO–infected mosquitoes were collected on day 0 (preimmune sera) and days 7, 13, and 28 after immunization (immune sera) and analyzed for anti-CSP antibody responses. Anti-CSP immunoglobulin G (IgG) was measured by enzyme-linked immunosorbent assay (ELISA) using full-length recombinant PfCSP protein. At day 0, volunteer sera showed an average of 1436 ± 257.3 arbitrary units (AU) of anti-CSP titers by ELISA, which was below the previously published cutoff of 2000 AU for positivity (11); three volunteers were slightly above this cutoff, with titers of 2443 and 2256 AU (Fig. 4A). At day 7 after administration, the frequency of positivity and average titer both increased such that 8 of 10 volunteers were ELISA-positive, with an average titer of 3821 ± 808.1 AU. Anti-CSP titers peaked at day 13 after administration when all volunteers were positive for anti-CSP IgG, and the group average increased to 11547 ± 2084 AU, ranging from 3874 to 24,445 AU. At day 28, anti-CSP titers declined to an average of 5774 ± 840 AU, with all volunteers maintaining titers above the cutoff for positivity (Fig. 4A). (Additional CSP titer data are provided in table S9.)

Fig. 4. Humoral immune responses in volunteers after Pf GAP3KO administration.

(A) Antibodies recognizing full-length PfCSP were quantified by ELISA at each time point. Antibody titers are expressed in AU with a positivity cutoff at 2000 (indicated by dashed line). Each data point represents the mean of duplicate ELISA titers for one volunteer, with the middle bar of the box as the median and the borders as the SDs. Error bars represent the minimum and maximum for each day. The symbol “#” represents a statistical difference in the mean compared to day 0; bars with asterisks indicate comparison made via one-way analysis of variance (ANOVA) and Tukey’s post-test; *P < 0.05, **0.05 > P ≥ 0.01, ****P < 0.001. Sera from immunized volunteers were also evaluated for functional inhibition of sporozoite invasion (B) and traversal (C) in an in vitro ISTI assay. Sera from days 7, 13, and 28 after immunization were normalized to volunteer-matched sera at day 0 to determine the percent inhibition of invasion and traversal. Each data point represents one volunteer and is the mean of three independent experiments. The middle bar of the box is the median, the borders are the SDs, and error bars represent the minimum and maximum for each day. Asterisks above each data set indicate a significant difference from 0% inhibition using one-sample t test, whereas bars with asterisks indicate comparisons via one-way ANOVA and Tukey’s post-test; *P < 0.05, ****P < 0.001. (D) Correlation between CSP titer and inhibition of invasion (orange) or traversal (blue). For each curve, R2 values and P value testing for a slope significantly different than zero are shown.

Functional activity of Pf GAP3KO–elicited antibodies against Pf sporozoites

Given that the immune serum of Pf GAP3KO–immunized volunteers showed considerable anti-CSP IgG titers, we next tested whether these serum antibodies were functional in blocking sporozoite infection using an in vitro inhibition of sporozoite traversal and invasion (ISTI) assay. The ISTI activity of immune serum (from days 7, 13, and 28) was measured by determining the percent inhibition using day 0 preimmune sera as a baseline (Figs. 4B and 4C). We determined that immune serum inhibited sporozoite invasion by 51.8 ± 7.3% for day 7 and 59.1 ± 2.97% for day 13 (Fig. 4B). Only one volunteer day 7 serum showed no inhibition, and day 13 sera from all volunteers had invasion-inhibitory activity ranging from 43.2 to 74.6%. Day 28 serum inhibited invasion as vigorously as day 13 serum (mean of 62.8 ± 9.3%) despite a drop in CSP titer (Fig. 4A). Strikingly, inhibition of invasion in the ISTI assay did not correlate with anti-CSP titers at any time point (Fig. 4D), and the serum of one volunteer with a negative day 7 CSP titer still exhibited a 44.6% inhibition of sporozoite invasion (Fig. 4, B and D). (Additional subject-specific data on invasion inhibition are provided in table S10.)

Fig. 5. FRG huHep passive transfer of immune sera shows inhibition of infection.

(A) IgG was purified from day 0 and day 13 serum samples from five volunteers that exhibited high and consistent inhibition in the ISTI assay. This IgG was passively transferred into FRG huHep mice (five mice per volunteer per time point) at 7 mg per mouse. The following day, mice were infected by bites of 50 PfGFP-luciferase–infected mosquitoes, and parasite liver burden was assessed at day 6 after infection using an in vivo imaging system (IVIS). Liver burden of each mouse that received day 13 immune IgG was normalized to the average liver burden of five mice receiving volunteer-matched day 0 preimmune serum. Each data point represents one mouse, with bars indicating means ± SD. Each group was tested via one-sample t test to determine whether the percent inhibition was significantly different than a 0 value, and significant results are indicated by asterisks, where *** is 0.001 ≤ P < 0.05. (B) Pf NF54 sporozoites or day 5 liver-stage parasites from FRG huHep mouse infections were stained using an immunofluorescence assay, with sera pooled from all 10 volunteers taken either before immunization (preimmune sera) or 28 days after immunization (immune sera). Parasites were counterstained in sporozoite indirect immunofluorescence assays with an anti-CSP monoclonal antibody or with a rabbit polyclonal antibody directed against the Plasmodium endoplasmic reticulum resident protein Bip in day 5 liver stages. For both sporozoite and liver-stage indirect immunofluorescence assays, human sera reactivity was visualized using a secondary rabbit anti-human antibody; DNA was visualized using 4′,6-diamidino-2-phenylindole (DAPI) stain.

Similarly, immune serum from volunteers inhibited in vitro cell traversal of sporozoites by an average of 41.9 ± 7.9% and 44.4 ± 8.3% on days 7 and 13, respectively (Fig. 4C). Inhibition ranged from 30.3 to 80.4%, with only two volunteers at each time point negative for traversal inhibition. Despite a drop in anti-CSP titer, inhibition of traversal stayed constant until day 28 with a mean of 40.9 ± 11.8%, with all volunteers exhibiting inhibition, which ranged from 17.9 to 63%. Again, inhibition of cell traversal did not correlate with anti-CSP titers at any time point (Fig. 4D). In every case where inhibition of traversal was absent, CSP IgG titers were above the threshold. (Subject-specific data on traversal inhibition are provided in table S9.)

To determine whether the antibodies engendered by Pf GAP3KO administration could also inhibit Pf sporozoite invasion in vivo, we used FRG huHep liver-chimeric humanized mice (12). These mice allow Pf liver infection and complete development of liver-stage parasites, mimicking natural human infection (13). We have recently shown that this model can robustly detect inhibition of sporozoite liver infection by monoclonal antibodies (14) using bioluminescence imaging of challenges that were carried out with sporozoites of a green fluorescent protein (GFP)–luciferase–expressing P. falciparum strain, PfGFP-luc (15). We selected five volunteers whose immune serum showed consistent, high levels of invasion inhibition in vitro and varying CSP IgG titers. We purified IgG from these volunteers from day 0 preimmune and day 13 immune samples using protein G enrichment and passively transferred this IgG into FRG huHep mice at a dose of 7 mg per mouse. Mice were then challenged with the bites of 50 PfGFP-luc–infected mosquitoes and monitored for liver-stage burden by bioluminescence imaging at day 6 after infection. Percent inhibition was calculated by comparing the liver-stage burden of mice receiving day 13 IgG samples to the mean value of liver-stage burden in mice receiving day 0 preimmune IgG samples. By this measure, three of five of the day 13 IgG inhibited liver infection by an average of 69.4 ± 8.3% and 76.6 ± 8.9% and 88 ± 7% (Fig. 5A and table S11). As in the in vitro ISTI, the three volunteers (003, 006, and 007) whose serum exhibited the highest inhibition had the lowest Pf CSP titers of the five samples. Given that a single immunization with GAP3KO-infected mosquitoes can elicit anti-CSP antibody titers and that immune sera can mediate inhibition of infection both in vitro and in vivo, we wanted to further characterize antibody specificity. To do this, we performed immunofluorescence assays on both Pf sporozoites and liver stages with pooled preimmune sera from volunteers and 28-day immune sera. In immunofluorescence assays with immune sera on fixed permeabilized Pf sporozoites, we saw predominant reactivity in a circumferential pattern, which overlapped with anti-CSP staining (Fig. 5B). Preimmune serum had no detectable reactivity. To investigate whether immune sera could recognize liver-stage antigens, we performed immunofluorescence assays on liver tissue sections from FRG huHep mice infected with PfNF54 that contained day 5 liver stages. Staining of these liver stages with preimmune serum revealed no reactivity, whereas immune serum from day 28 post-Pf GAP3KO immuniza tion revealed staining on both the parasite periphery and the interior (Fig. 5B). Together, these data indicate that a single Pf GAP3KO immunization delivered by mosquito bite elicited antibodies with specificity for both sporozoites and liver-stage parasites.


Pursuit of a highly efficacious pre-erythrocytic malaria vaccine that protects against parasite infection has followed two distinct approaches: immunization with subunit parasite antigens (either adjuvanted recombinant protein or virally vectored) and immunization with whole sporozoites (16, 17). The latter approach is the only method known to experimentally achieve complete, sterile protection in challenge trials using controlled human malaria infection (CHMI) with Pf sporozoites that are administered by mosquito bite. Although some subunit immunization approaches are showing improved efficacy (18), they have all failed to achieve complete, sterile immunity in CHMI trials (19). In contrast, RAS, administered by large numbers of mosquito bites (>1000 bites), and immunization with nonattenuated sporozoites by low-dose mosquito bite administration (45 bites) under chloroquine prophylaxis both confer complete protection against Pf CHMI (5, 6). Furthermore, recent parenteral administration of RAS in human trials demonstrated that live sporozoites could be manufactured, formulated, and delivered by modalities that are safe, immunogenic, protective, and practical for human vaccination (20, 21). GAPs may offer advantages compared to vaccination with RAS or CPS. Complete and consistent intrinsic attenuation eliminates the risk of over-attenuation of the Pf parasite and batch-to-batch variation, which hampers safety and immunogenicity of RAS. GAP also eliminates the risk of noncompliance with the antimalarial prophylaxis, which is essential for the safety of CPS. Thus, the rational design of an intrinsically pre-erythrocytic stage–attenuated Pf strain that can be used for immunization is of great importance.

The GAP strategy was enabled by the development of gene deletion methodologies in Plasmodium (22, 23), delineation of the full Plasmodium genome sequence, and transcriptional profiling that identified pre-erythrocytic stage–specific gene expression (24, 25). Extensive research with rodent malaria parasites identified a number of gene products that are critical for parasite development within hepatocytes, and deletions of these genes arrested liver-stage development at distinct points in the growth and maturation process. However, to date few of the orthologous gene deletions in Pf have yielded parasite strains that are uniquely attenuated only during pre-erythrocytic infection. We have previously shown that deletion of two pre-erythrocytic stage–expressed genes in Pf P52 and P36 (Pf GAP2KO), causes early developmental arrest of the parasite soon after hepatocyte infection (26). In a phase 1 trial, a five-bite administration of the Pf GAP2KO–infected mosquitoes to subjects showed no breakthrough blood-stage infection, but one of six subjects administered 200 Pf GAP2KO–infected mosquito bites developed patent blood-stage infection (11), indicating severe but incomplete attenuation. To achieve complete attenuation, we introduced a third gene deletion of the sap1 gene to create a triple gene deletion p52/p36/sap1 Pf (Pf GAP3KO) (9) that retained robust sporozoite production and a unique pre-erythrocytic stage–attenuated phenotype (9). Here, we have shown the successful translation of the Pf GAP3KO strain to a first-in-human phase 1 safety and immunogenicity trial. Pf GAP3KO was administered by a single administration of 150 to 200 infected mosquito bites under controlled conditions. The results demonstrated that Pf GAP3KO is safe and well tolerated in healthy adult volunteers. Pf GAP3KO was fully attenuated during pre-erythrocytic infection, and no blood-stage parasites or parasite 18S rRNA markers were detected in the peripheral blood of volunteers by microscopy or qRT-PCR, respectively. Local reactions to the large number of Pf GAP3KO–infected mosquito bites were the most significant adverse events. Such administration site reactions were managed with topical steroids or (in one subject) oral antihistamines. Although all subjects experienced local and systemic adverse events, they were all mild to moderate, and there were no clinically significant laboratory abnormalities. Local solicited adverse events remained ongoing at the end of the 28-day study period; however, none were greater than grade 1 (mild) at the time of the last in-clinic study visit. Unsolicited adverse events were infrequent and mild in nature, with most resolving by 28 days and all by 3 months.

On the basis of previous experimental determination of the number of sporozoites transmitted by mosquito bite, we estimate that each volunteer received between 15,000 and 200,000 Pf GAP3KO sporozoites (27, 28). The absence of liver toxicity suggests that, although Pf GAP3KO sporozoites target the liver and infect hepatocytes, infection does not lead to significant damage of the parenchyma. The clinical data from this study indicate that Pf GAP3KO is fully attenuated during pre-erythrocytic infection in humans when administered by 150 to 200 mosquito bites, a finding that is in accordance with preclinical data, where no breakthrough blood-stage infections occurred when humanized FRG huHep mice that carry human hepatocytes and human RBCs were inoculated intravenously with large numbers of Pf GAP3KO sporozoites (9).

To model Pf GAP3KO efficacy, we created a triple gene deletion parasite (p52/p36/sap1) in the Py rodent malaria model. The Py GAP3KO was completely attenuated in highly susceptible mice. Only two immunizations with Py GAP3KO conferred complete, protracted protection against infectious sporozoite challenge for up to 6 months. Because our Pf GAP3KO phase 1 study was only intended for safety testing, limited immunogenicity studies were performed. Nonetheless, our initial analysis of the samples produced encouraging results. The single 150- to 200-bite Pf GAP3KO immunization induced antibody responses to pre-erythrocytic stages in all subjects. Functional antibody responses revealed by in vitro sporozoite inhibition of invasion and traversal assays as well as FRG huHep liver humanized mouse challenges were also demonstrated. All volunteers developed antibody titers to CSP, but these did not necessarily correlate with functional activity. Given that immunization with Pf GAP3KO whole parasites likely primed immune responses to antigens other than CSP (2931), it is reasonable to speculate that GAP3KO-elicited antibodies might recognize other protein targets of the parasite and thus additionally block infection through targeting non-CSP sporozoite antigens. Although immune serum from subjects showed in a predominant circumferential staining of sporozoites that colocalized with anti-CSP staining, the specificities of this reactivity might include non-CSP surface proteins of sporozoites, which have recently been identified by proteomics (32). In addition, immune serum recognized day 5 liver stages and bound to the interior of the parasite, suggesting that the serum recognized non-CSP antigens (13).

A wealth of malaria literature indicates that interferon-γ (IFN-γ)–producing T cells are also important for protection against pre-erythrocytic infection (3335). Human Pf GAP2KO vaccination by mosquito bite showed that 5 or 200 bites induced antigen-specific responses of interleukin-2– and IFN-γ–producing CD4 and CD8 T cells (11), which is consistent with immunophenotyping performed on subjects protected from CHMI in the Pf RAS vaccination model (21). Here, subjects received a single high dose of Pf GAP3KO sporozoites. Hence, it is not possible to evaluate immune response recall. Such questions will be a focus of subsequent multidose immunization challenge studies with Pf GAP3KO.

Together, our data support the continued development of Pf GAP3KO as a promising candidate for whole-organism vaccination. Conclusions from this study are restricted by the limited number of Pf GAP3KO–immunized subjects. In addition, only humoral immune responses of subjects were analyzed, and we do not yet know the degree of protection that can be achieved with Pf GAP3KO immunization.

The GAP platform is flexible. Further genetic engineering that introduces transgenes expressing asexual blood stage and gametocyte antigens during pre-erythrocytic infection might broaden the immune responses to include additional protection against asexual-stage parasitemia and onward transmission of sexual stages when pre-erythrocytic immunity is waning and incomplete. However, a live-attenuated whole sporozoite vaccine, such as Pf GAP3KO, continues to face challenges. Mosquito bite administration of attenuated sporozoites is not a practical or scalable route of immunization. Thus, although we are currently pursuing a phase 1b immunization CHMI study with Pf GAP3KO using mosquito bite administration to assess preliminary efficacy, advanced clinical testing will require a formulation of Pf GAP3KO sporozoites that enables parenteral immunization. However, even for such a formulation, Pf GAP3KO sporozoites can currently only be produced in mosquitoes and need to be isolated from their salivary glands, limiting scalable manufacturing of this type of vaccine. Nonetheless, the successful creation of a completely attenuated P. falciparum strain by means of genetic engineering is an important step on the path to development of a live-attenuated sporozoite vaccine that protects against malaria.


Study design

Protocol MC-004 was a single-center, single-arm, open-label, phase 1 experimental medicine study designed to evaluate the safety and tolerability of a genetically attenuated (p52/p36/sap1) P. falciparum (Pf GAP3KO) that arrests early in the parasite liver stage. The study was also designed to confirm attenuation of the Pf GAP3KO parasites using peripheral blood smears and qRT-PCR and to evaluate immune responses to Pf GAP3KO. Before initiation, the study was approved by the Western Institutional Review Board and the U.S. Army Medical Research and Material Command, Office of Research Protections, Human Research Protection Office.

Malaria-naïve study subjects received a single Pf GAP3KO administration delivered via the bites of 150 to 200 A. stephensi Pf GAP3KO–infected mosquitoes under controlled conditions. Mosquitoes were raised under phase-appropriate Current Good Manufacturing Practice conditions in the Center for Mosquito Production and Malaria Infection Research facility at the Center for Infectious Disease Research. For administration, Pf GAP3KO–infected mosquitoes were allowed to feed on the subject’s arm for a total of 10 min. The dose was measured according to procedures described in the Supplementary Materials.

Delineation of the site and sponsor for the conduct of the trial was defined by separate policies and standard operating procedures. The site-specific procedures governing the conduct of the trial also followed, where applicable, Center for Infectious Disease Research procedures and quality systems. An independent medical monitor was available for oversight of the safety and welfare of subjects participating in this trial and to advise the investigator on trial-related medical questions or problems.

Study participants and eligibility criteria

Subjects were eligible to participate in the trial if they were healthy, malaria-naïve men or nonpregnant women aged 18 to 50 years in good general health, as demonstrated by medical history, physical examination, ECG, and laboratory assessment within 90 days of enrollment. All inclusion and exclusion criteria are listed in table S2.

Objectives and end point measures

Blood samples were collected from volunteers at predefined time points in accordance with the study design depicted in Fig. 1. The primary study objectives were to monitor safety and tolerability of Pf GAP3KO in healthy malaria-naïve adults and confirm attenuation. Secondary objectives included assessment of humoral immune responses and inhibition of sporozoite infection after Pf GAP3KO administration. Primary end point measures were frequency of solicited and unsolicited adverse events, frequency of serious adverse events, and patent parasitemia via thick blood smear. Secondary end point measures included CSP antibody titer by ELISA and percent inhibition of in vitro sporozoite cell invasion and traversal by immune sera after Pf GAP3KO administration. For detailed description of frequency of unsolicited adverse events, solicited adverse events, and serious adverse events and end point measurements, see the Supplementary Materials.

Production of the Pf GAP3KO parasite

The Pf GAP3KO parasite was engineered to remove three key infectivity genes—p52, p36, and sap1—from the haploid parasite genome. Although essential for the liver stage, these genes are not essential for the asexual blood stage, and therefore, Pf GAP3KO can be maintained in asexual blood-stage culture. The gene deletion procedure and plasmids used for gene deletion were described previously (22, 23). The parent strain was Pf NF54. For more detailed description of Pf GAP3KO master cell bank production, see the Supplementary Materials.

Administration of Pf GAP3KO parasite

Pf GAP3KO was administered in a secure insectary facility under Arthropod Containment Level 2 conditions. About 150 to 220 potentially infected mosquitoes were placed in an administration container and allowed to feed for a single 10-min period on a research subject’s arm. The total number of mosquitoes placed in the container was adjusted to achieve the target number of infectious bites (150 to 200) and allows for adjustment based on lot-specific infection rates. An estimate of the lot-specific infection rate was made 1 day before administration by dissecting the salivary glands from a sample of mosquitoes taken from the source cages. Infection rates were scored (table S2), with gland grades of 2+ or higher considered infected. For measurement of mosquito infection rate, see the Supplementary Materials.


Peripheral thick blood smears were prepared in duplicate from EDTA-anticoagulated whole blood and examined using CHMI-specific procedures by trained study microscopists (36). Methods used align with consensus thick blood smear microscopy methods published for monitoring patency end points in CHMI studies (37). Sporozoite and liver-stage immunofluorescence assays were conducted as previously described (13). For sporozoite immunofluorescence assay, sporozoites were stained with a 1:200 dilution of pooled preimmune or day 28 serum for 2 hours at 4°C. A 1:800 dilution of Alexa Fluor 594–conjugated goat anti-human IgG (Life Technologies, catalog #A11014) was used as a secondary antibody. Directly conjugated mouse monoclonal anti-PfCSP antibody (clone 2A10) was used as a parasite marker at 2 μg/ml. To generate the day 5 Pf-infected liver, FRG huHep mice were intravenously infected with 106 PfNF54 sporozoites and sacrificed them 5 days later. Liver sections were stained with a 1:200 dilution of serum and a 1:1000 dilution of rabbit anti-Bip antibody to visualize parasites. Goat anti-rabbit Alexa Fluor 488 antibody (Life Technologies, catalog #A11008) at 1:500 dilution and a 1:600 dilution of Alexa Fluor 594–conjugated anti-human IgG (Life Technologies, catalog #A11014) were used as secondary antibodies. All images were acquired using Olympus IX-70 DeltaVision deconvolution microscope at ×100 magnification, with identical exposure settings and post-acquisition contrast/brightness adjustments made for comparing day 0 with day 28 serum staining.

Diagnostic qRT-PCR on human samples

Concurrent with tris-buffered saline preparation, 50-μl aliquots of EDTA-anticoagulated whole blood were preserved in 2 ml of NucliSENS lysis buffer (bioMérieux) and stored at −80°C until qRT-PCR testing. Nucleic acid extraction and amplification were performed on the Abbott m2000 RealTime System as previously described (38). This assay sensitively detects and quantifies blood-stage P. falciparum parasites based on detection of the P. falciparum 18S rRNA/ribosomal DNA (rDNA) (limit of detection, 20 parasites per milliliter).

Detection of CSP by ELISA

Full-length Pf CSP was used in ELISA to detect antibodies from serum from Pf GAP3KO–exposed subjects. Pf CSP was produced by cloning Pf csp (encoding amino acids 21 to 289) into the cytomegalovirus promoter–based pTT3 vector (39), transfecting into human embryonic kidney (HEK) 293F cells (40), and purifying by HisTrap FF nickel affinity column chromatography (GE Healthcare) and a HiLoad 16/60 Superdex 200 size exclusion column (GE Healthcare). All ELISA values were repeated and are reported as the mean of two independent runs. For detailed description of CSP ELISA performed, please refer to the Supplementary Materials.

ISTI assay

The ability of serum from immunized volunteers to inhibit sporozoite traversal and invasion was performed as described previously (41, 42), and it is described in detail in the Supplementary Materials. Data were reported as the average of three independent experiments.

FRG humanized mouse studies

The FRG huHep mouse model was used to determine whether antibodies generated by immunization could inhibit Pf sporozoite infection of the liver in vivo. This model has been reported to support full liver-stage development of Pf and can be used to test antibody-mediated reduction of liver infection after Pf infection (13). Sera and plasma from five volunteers with varying anti-CSP ELISA titers and positive inhibition of invasion in vitro were used for IgG purification using protein G columns (GE Healthcare, catalog #28-9852-55). Purified IgG from preimmune and day 13 post–Pf GAP3KO samples were intravenously injected into five mice each per day (preimmune or day 13) for each of the four subjects. Twenty-four hours later, all mice were infected by the bites of 50 mosquitoes infected with luciferase-expressing Pf parasites (15). Six days later, liver-stage burden was determined by measuring bioluminescence with an IVIS. Percent inhibition of liver-stage burden was determined by dividing the total flux (pixels per second) of mice receiving post-immunization IgG by the average of mice injected with preimmune IgG. Statistical significance of inhibition was determined by performing a one-sample t test with a hypothetical value of 0% inhibition.

Creation of a P. yoelii GAP3KO parasite

A P. yoelii 17XNL parasite was created with deletions in p36, p52, and sap1 as outlined in fig. S1. This approach was based on the published gene insertion/marker out strategy (1). Primers used for creation and analysis of Py GAP3KO are in table S1.

Py mouse studies

For attenuation and immunization studies with Py GAP3KO and WT parasites, salivary gland sporozoites were generated and isolated as previously described (9). Sporozoites were diluted in phosphate-buffered saline and injected intravenously into BALB/cByJ or BALB/cJ mice (Jackson Laboratory) at indicated doses for attenuation, immunization, and challenge studies. Mouse experiments adhered to the National Institutes of Health Guide for the Care and Use of Laboratory Animals.


Because the study was designed to characterize safety, attenuation, and immunogenicity of Pf GAP3KO, most data were descriptive. Where performed, statistical analyses included all evaluable participants at each time point. For immunology assays, qualitative assay data analysis was performed by tabulating the frequency of positive responses for each assay at each time point that an assessment was performed, with crude response rates presented with corresponding exact 95% confidence interval estimates. No formal sample size calculation was conducted, because this was an exploratory study and contained only one treatment group. The ability of the study to detect parasitemia was expressed by the true event rate above which at least one participant would likely be confirmed, via peripheral blood smear, to have experienced parasitemia and the true event rate below which no parasitemic events would likely be observed. There was a 90% chance of observing at least one event if the true rate of such an event was 20.5% or more; conversely, there was a 90% chance of observing no events if the true event rate was 1% or less.


Materials and Methods

Fig. S1. The creation of P. yoelii lacking p52, p36, and sap1 (Py GAP3KO).

Fig. S2. Photograph of application site illustrating spectrum of local reactogenicity.

Table S1. Primers used to produce P. yoelii lacking p52, p36, and sap1 (Py GAP3KO).

Table S2. Rating scale for sporozoite salivary gland load.

Table S3. Eligibility criteria for study volunteers.

Table S4. Participant demographics.

Table S5. Solicited adverse events—post-administration reactogenicity.

Table S6. Solicited adverse events—malaria signs and symptoms.

Table S7. All adverse events by system organ class/preferred term and severity grade.

Table S8. Abnormal laboratory values.

Table S9. CSP titer values.

Table S10. Functional assay values.

Table S11. In vivo inhibition values.


  1. Acknowledgments: We thank J. Williams for assistance with microscopy during the hotel phase of this study. Funding: This research and development program was made possible by a cooperative agreement that was awarded and administered by the U.S. Army Medical Research and Materiel Command and the Telemedicine and Advanced Technology Research Center at Fort Detrick, MD, under contract number W81XWH-11-2-0184. The views, opinions, and/or findings contained in this publication are those of the author(s) and do not necessarily reflect the views of the U.S. Department of Defense (DoD) and should not be construed as an official DoD/Army position, policy, or decision unless so designated by official documentation. No official endorsement should be made. Humanized mouse experiments in this study were funded by the Bill and Melinda Gates Foundation (investment ID, 24922). Author contributions: S.A.M., B.K.S., M.E.F., A.S., D.A.D., R.W.J.S., A.M.V., D.N.S., and S.C.M. performed preclinical experiments. A.M.V. and D.A.D. designed and created Py transgenic parasites. B.K.S., D.A.D., and A.M.V. performed Py attenuation and immunization/challenge studies. B.K.S. and M.E.F. performed ELISAs. B.K.S. and R.W.J.S. performed ISTI. B.K.S., S.A.M., and M.E.F. performed FRG huHep passive transfer. B.K.S. performed immunofluorescence assays. J.G.K., S.A.M., S.C.M., and S.H.I.K. designed the clinical study. J.G.K., S.A.M., L.S., T.V.G., M.F., M.E.F. S.M., E.F., W.B., H.S.K., S.C.M., and S.H.I.K. contributed to the execution of the clinical trial. J.G.K., S.A.M., B.K.S., S.C.M., and S.H.I.K. led the writing of the manuscript that was reviewed by all authors. Competing interests: S.H.I.K. is an inventor listed on U.S. Patent 7,22,179, U.S. Patent 7,261,884, and international patent application PCT/US2004/043023 entitled “Live genetically attenuated malaria vaccine.” Data and materials availability: All requests should be directed to S.H.I.K.
View Abstract

Stay Connected to Science Translational Medicine

Navigate This Article