Safety, infectivity and immunogenicity of a genetically attenuated blood-stage malaria vaccine

There is currently a strong commitment to eradicate malaria in the 21st century. This ambitious goal will be facilitated by the development of an effective vaccine. The past 20 years have seen over 100 clinical trials of malaria vaccine candidates, with the vast majority testing sub-unit vaccines [1]. To date, only one vaccine (RTS,S/AS01) has demonstrated protective efficacy against clinical malaria in a phase 3 trial, although protection is only partial [2]. RTS,S/AS01 has been approved by the European Medicines Agency, and a pilot implementation program coordinated by the World Health Organization is underway in Africa [3]. Given the overall slow progress of current vaccine strategies, there is a clear need for novel approaches.

Live, whole cell vaccines offer several potential advantages over subunit vaccines against a range of pathogens including malaria parasites. These include exposure to a greater range of antigens that are delivered to the correct anatomical compartments for immunological priming, as well as providing the correct pattern recognition signals to the immune system, which are specific for the pathogen class [4]. Various whole sporozoite vaccines that target the pre-erythrocytic stage of malaria parasite development have demonstrated promising results in preclinical studies, human challenge studies, and in clinical trials in malaria-endemic areas [1]. Strategies for sporozoite attenuation have included the use of radiation [5] or targeted gene deletion [6‐8] to produce genetically attenuated parasites (GAP). Naturally acquired immunity to malaria is complex, with evidence suggesting that repeated exposure to blood-stage infection is an important component and is associated with an antibody response against merozoites and infected erythrocytes [9]. However, in the modern era, only a single live whole parasite blood-stage vaccine candidate has been tested in humans, in a pilot study undertaken using chemically attenuated parasites [10]. We aimed to develop a blood-stage GAP malaria vaccine and characterize its safety, infectivity, and immunogenicity in a phase 1 clinical trial.

P. falciparum erythrocyte membrane protein 1 (PfEMP1) is a polymorphic adhesion ligand displayed on knob-like structures on the surface of P. falciparum-infected erythrocytes, with specific types binding a variety of host receptors present on the vascular endothelium [11]. This arrangement facilitates attachment and sequestration of the infected erythrocyte in the microvasculature, thereby protecting parasites from clearance by the spleen [12]. Further, parasite sequestration is a critical determinant of malaria pathogenesis as it leads to the accumulation of infected erythrocytes in vital organs [13]. PfEMP1 is also a target of protective immunity [14] and has been shown to have an immunomodulatory effect [15]. PfEMP1 is a variable surface antigen, encoded by a repertoire of around 60 var genes per genome [16]. Thus, creation of a genetically modified parasite lacking PfEMP1 would require a targeted genetic modification that would prevent transcription of any var gene allele, an approach that may be technically challenging. Further, if such a result were achieved it would result in no expression of PfEMP1, a protein which has been proposed to be the target of strain-specific immunity [14].

The in vivo virulence and immunogenicity of a knobless P. falciparum clone selected under in vitro culture was investigated in Aotus monkeys by Langreth and Peterson [17]. After multiple inoculations with the knobless parasites, non-splenectomized animals either did not develop patent parasitemia, or exhibited very low parasitemia, with infections self-resolving. In contrast, splenectomized animals developed significant infections when inoculated with knobless parasites. This study indicated the importance of the knobs on the surface of infected erythrocytes in enabling parasites to sequester to avoid splenic clearance. When monkeys with intact spleens were exposed to knobless parasites they developed a humoral immune response and exhibited partial protection when subsequently challenged with knob-positive parasites. Together, these data suggest a knob-minus live malaria vaccine represents an attractive approach.

Our approach to develop a blood-stage GAP vaccine was to target the gene encoding the knob-associated histidine-rich protein (KAHRP), which is responsible for the assembly of knob structures at the infected erythrocyte surface. In the absence of KAHRP, P. falciparum-infected erythrocytes lack knob-structures and cannot correctly display PfEMP1 [18]. Building on the early studies in Aotus monkeys, we hypothesized that the loss of KAHRP would result in a significant loss of fitness in vivo due to enhanced parasite clearance accompanied by priming of the immune system from lack of parasite sequestration and increased passage to the spleen. The production of the blood-stage GAP vaccine under Good Manufacturing Practice (GMP) by targeted deletion of the gene encoding KAHRP has been described previously [19]. Here we report the results of a clinical trial to characterize the safety, infectivity, and immunogenicity of the P. falciparum kahrp– blood-stage GAP vaccine in healthy volunteers.

This study represents the first clinical investigation of a genetically attenuated blood-stage human malaria vaccine. We demonstrate that P. falciparum 3D7 parasites lacking KAHRP are highly attenuated in vivo compared to their wild-type counterparts, with no parasitemia detectable up to 28 days following inoculation of up to 1.8× 10⁵ viable parasites. Inoculation of healthy subjects with 1800–2800 wild-type P. falciparum 3D7 parasites in CHMI studies using the IBSM model consistently results in detectable parasitemia within 5 days, and parasitemia typically reaches > 5000 parasites/mL 8 days after inoculation [21, 22, 25, 46, 47]. However, we found that P. falciparum kahrp– parasites were capable of generating a significant infection when administered at a high dose (3 × 10⁶ viable parasites).

The fact that P. falciparum kahrp– parasites are capable of generating a significant infection in humans was somewhat unexpected based on the well-characterized role of KAHRP in mediating the adherence of infected erythrocytes to the microvascular endothelium, which is an important process in promoting parasite survival by evading splenic clearance [18]. The in vivo growth observed in the current study appears to be consistent with the previously reported breakthrough infections (peak parasitemia of 0.02%) observed in 2 of 7 Aotus monkeys challenged with multiple doses of knobless P. falciparum parasites (up to 1× 10⁷ parasites) [17]. The delay in and variability of timing of onset of patent parasitemia in the subjects administered the highest dose suggests some in vivo adaptation of parasites to enable escape from host clearance mechanisms. In contrast to the highly reproducible growth pattern of parasites seen in experimental infection with unattenuated P. falciparum 3D7 parasites [48], there appears to be significant inter-individual variability in this process. Such variability would likely complicate any use of this approach for in vivo immunization. Because of the unexpected breakthrough in blood stage infection at the highest dose, we investigated for selection of phenotypic or genetic changes that had taken place to explain this observation. Genome sequencing of parasites rescued into culture at maximum parasitemia prior to antimalarial treatment confirmed that parasites had not reverted back to wild-type. Similarly, ex vivo cultured parasites retained their knob-minus phenotype, and exhibited impaired cytoadherence, particularly under flow conditions.

There was evidence that parasites switched their expression of var genes during in vivo growth. Parasites administered to subjects on day 0 expressed a single dominant group C var gene. This is consistent with previous evidence that, in the absence of phenotype selection, parasites often switch to expression of a group C var gene which appear to have a slower rate of silencing than other var gene types [49]. Following growth in vivo, parasites had switched away from the expression of the single group C var gene, suggesting selection in the naïve, human host, possibly for an adhesion phenotype. The increased diversity of var expression could also reflect the absence of selection for cytoadhesive variants, because the infected erythrocytes could not cytoadhere. If this were the case, then the altered environment within the host stimulated the de-repression of much of the var repertoire. Interestingly a similar observation of broad var de-repression has been made in CHMI studies using sporozoites [40]. In that case, passage through mosquitoes was inferred to have reset the epigenetic regulation of var genes, consistent with the broad, epigenetic reprogramming that occurs through sexual reproduction in higher order eukaryotes. The data we present suggests that the human host environment itself may stimulate var gene de-repression during the parasite’s asexual lifecycle.

Despite the well-characterized role of knobs in the cytoadherence and sequestration of parasite-infected erythrocytes, there is some evidence that erythrocytes infected with knob-minus parasites may be capable of binding weakly to the microvascular endothelium. Knobless P. falciparum clones were found to adhere to CD36 at a level comparable to the parent 3D7 strain under static conditions, but dramatic differences were observed under flow conditions [18]. Similar results were observed in the current study, with P. falciparum kahrp– parasites exhibiting some adhesion to CD36 under static conditions (albeit at considerably lower levels than the control strain) and almost no binding under flow conditions. These results are supported by recent research using engineered human capillaries which demonstrated that erythrocytes infected with knob-minus parasites can accumulate in the post-capillary space where shear rates are lower [50]. In contrast, knob-positive parasite-infected erythrocytes that were trypsinized to remove the surface presentation of PfEMP1 did not accumulate in vessels to any extent. Thus, we hypothesize that in the current study a limited number of P. falciparum kahrp– parasites were able to bind to the microvascular endothelium in areas where shear rates were low and avoid splenic clearance. This may have been facilitated by the switch in var gene expression observed during in vivo growth to present a range of PfEMP1s on the cell surface. An alternative hypothesis is that parasites may have accumulated in the spleen and were cleared by the immune system slowly enough to maintain an infection.

Immunogenicity results indicated that the GAP vaccine was capable of eliciting a malaria-specific antibody and cell-mediated immune response when administered at the highest dose (3 × 10⁶ viable parasites). Immunogenicity appeared to be dependent on parasite replication, evidenced by the fact that the subject administered the highest dose who did not develop parasitemia failed to generate a detectable antibody response. This finding may be consistent with evidence that killed whole parasite malaria vaccines require an adjuvant to induce robust immunity [51]. The anti-parasite humoral immune responses to the immunodominant blood-stage antigen MSP2 in this study were similar to those seen during natural infection against this and other blood-stage antigens, as well as when they were measured in CHMI studies using the IBSM model where subjects were infected with wild-type 3D7 parasites [52, 53]. As the preclinical primate data suggest a critical role of the spleen in the outcome of infection with knobless parasites, the effect of splenectomy on immunogenicity and clearance phenotype would require careful investigation in future clinical studies.

Although no serious safety concerns prevented completion of this stage of clinical development of the candidate vaccine, two considerations resulted in the vaccine not progressing to testing for protective efficacy in a challenge study with unattenuated parasites. Firstly, the fact that an immune response only developed in subjects who developed patent parasitemia some weeks after inoculation, and this parasitemia required termination by timed antimalarial chemotherapy, would impose significant practical limitations on the clinical development of this vaccine in the absence of other modifications to render the parasite avirulent. Secondly, two subjects administered the highest dose of GAP vaccine developed erythrocyte alloimmunization (anti-C and anti-P1 antibodies in one subject and anti-c antibodies in the other subject). This was unexpected based on our previous experience with CHMI studies using the IBSM model where alloimmunization attributed to the malaria challenge agent has not been reported previously (approximately 400 subjects in 32 studies). The fact that alloimmunization was observed in subjects who were administered a dose of parasites approximately 1000-fold higher than is typically administered in CHMI studies using the IBSM model involving wild-type P. falciparum 3D7 (3× 10⁶ viable parasites vs 2800 viable parasites) suggests that the dose of parasitized erythrocytes may be a factor. The total dose of erythrocytes (including non-parasitized) administered was similar to that routinely administered (approximately 1 × 10⁸ erythrocytes). We speculate that altered antigen presentation by parasitized erythrocytes with increased passage to the spleen may contribute to a higher probability of alloimmunization with the GAP vaccine. Of note, one of the 8 subjects administered a chemically attenuated blood-stage P. falciparum parasite vaccine in a pilot study also developed anti-c alloantibodies [10]. Although anti-P1 antibodies are not considered clinically important for transfusion reactions, anti-C and anti-c antibodies may result in a delayed transfusion reaction characterized by slow drop in hemoglobin over two weeks post-transfusion. Consultation with transfusion medicine specialists indicated there is no immediate risk if these subjects were to require emergency administration of unmatched Group O Rh (D) negative blood, and in the setting of routine blood transfusion a full cross-match would obviate such a reaction. Since both antibodies were of low titer, there is a possibility that they may diminish over time. Together, these two adverse safety findings precluded us continuing the study to test for protective immunity against virulent parasite challenge.

Although the results of this study overall do not support the further clinical development of the P. falciparum kahrp– strain as a blood-stage GAP malaria vaccine, they do contribute valuable new information on the pathogenicity of P. falciparum. Additionally, this study acts as a proof-of-concept for the use of genetically modified blood-stage malaria parasites in humans. We have previously developed IBSM models for P. falciparum [23], artemisinin-resistant P. falciparum [54], P. vivax [55], and P. malariae [56]. The use of genetically modified parasites in CHMI studies using the IBSM model offers significant opportunities in the ongoing search for interventions to reduce the burden of malaria.