Immune responses to malaria pre-erythrocytic stages: Implications for vaccine development

Radiation-attenuated sporozoites induce sterilizing immunity and remain the ‘gold standard’ for malaria vaccine development. Despite practical challenges in translating these whole sporozoite vaccines to large-scale intervention programmes, they have provided an excellent platform to dissect the immune responses to malaria pre-erythrocytic (PE) stages, comprising both sporozoites and exoerythrocytic forms. Investigations in rodent models have provided insights that led to the clinical translation of various vaccine candidates—including RTS,S/AS01, the most advanced candidate currently in a trial implementation programme in three African countries. With advances in immunology, transcriptomics and proteomics, and application of lessons from past failures, an effective, long-lasting and wide-scale malaria PE vaccine remains feasible. This review underscores the progress in PE vaccine development, focusing on our understanding of host-parasite immunological crosstalk in the tissue environments of the skin and the liver. We highlight possible gaps in the current knowledge of PE immunity that can impact future malaria vaccine development efforts.


| BACKG ROU N D
Malaria remains an intractable global public health problem with an estimated 228 million cases and 405 000 deaths in 2018 alone. 1 A vast majority of these deaths occur in sub-Saharan Africa, where malaria is associated with a 24% prevalence and 94% of the malaria-associated deaths globally. 1,2 Recent advances in malaria control including improved diagnostic approaches, artemisinin-combination treatments (ACTs), intermittent preventive treatment (IPT) in pregnancy and vector control saw a 48% decrease in mortality rates between 2000 and 2015. 1 Whilst these strategies have unquestionably contributed to reduction in incidence and mortality rates, an effective vaccine would provide the ultimate solution to malaria elimination and should be an urgent public health priority.
Malaria biology is complex. Our understanding of the pre-erythrocytic (PE) stage infections is based on model systems with Plasmodium berghei (Pb) and P. yoelii (Py), with limited information on P. falciparum (Pf). The PE stage begins when an infected female Anopheles mosquito inoculates a few (typically <100) infective sporozoites into the host skin. 3,4 Quantitative studies with Pb and Py indicate that a large proportion (~60%) lose motility and remain localized at the site of inoculation where they can develop into skin exoerythrocytic forms (EEF) and initiate an immune response. [5][6][7][8] Some sporozoites 'trickle out' of the skin into the blood (~25%) and the lymphatic drainage (~15%). 6,9 Most of the sporozoites that enter the bloodstream reach and invade the liver, where they traverse through several hepatocytes in a transient vacuole. The sporozoites then invade a final hepatocyte and form parasitophorous vacuoles (PV), where the liver EEFs develop. 6,10 The circumsporozoite protein (CSP), which is the major antigen on the sporozoite surface, and thrombospondin-related anonymous protein (TRAP), a micronemal protein, are thought to facilitate invasion into the hepatocytes. 11,12 In the liver, the parasites undergo asexual development for a number of days depending on the Plasmodium species (ie 7-10 days for human malaria vs 42-44 hours for Pb infection in mice), pre-existing immunity and concomitant malaria prophylaxis. 13 They differentiate into multinucleated schizonts that form thousands of merozoites via nuclear division. In the late stages of development, the PV membrane is lysed, and the merozoites become packaged together inside merosomes. 14,15 These merosomes egress out of the liver, circulate F I G U R E 1 The malaria life cycle. An infected mosquito deposits motile infective sporozoites into the dermis of a susceptible host. Some sporozoites migrate to the liver, where they invade hepatocytes, multiply asexually to produce thousands of merozoites which egress in merosomes and rupture inside microvasculature of lungs. The merozoites invade the red blood cells (RBC), and undergo multiple cycles of ring, trophozoite and schizont stages, to initiate the clinical phase of the disease. Some parasites differentiate into male and female gametocytes, which are taken up mosquitoes during their next blood meal. Different immune cells interact with the malaria sporozoites during its journey from the skin to the liver and may be exploited in the development of an effective and long-lasting vaccine. NK denotes natural killer cells Administration of sporozoites followed by antimalarial chemoprophylaxis with chloroquine or mefloquine (CPS vaccines), which acts on blood stage but not liver-stage parasites, yields comparable efficacy levels to RAS and confers protection against PE stages in both rodent models and humans. [32][33][34][35][36] CPS vaccines may provide more robust immunity as the sporozoites undergo complete liver-stage development. Alternative CPS approaches involve using antibiotics, such as clindamycin and azithromycin, which allow full parasite development in the liver, but lead to delayed death in the resulting merozoites. 37 In rodent models, CPS vaccines have been shown to induce robust, long-lived immunity that not only protects against PE stages, but also against blood stages. 38 A clinical trial using PfGAP lacking two genes (p36 − p52 − ) reported favourable anti-sporozoite immune responses. 44 The triple gene PfGAP was reported to fully attenuate sporozoite development in the early liver stages in in vitro and humanized mice studies. 45,46 PfGAP3KO was reported to be safe and immunogenic in human volunteers after 150-200 mosquito bites, but is yet to complete clinical trials. 47 Other GAP vaccine development efforts include targeting the late EEF stages, such as deletion of fabb/f, PlasMei2, and liver-specific protein 2 (LISP2) genes. 31

| Sub-unit vaccines
Sporozoites are covered with a dense coat, and CSP-a 40-66 kDa protein, with ~40 NANP repeats in the central region of PfCSP -is the major surface protein. 58 Inadvertently, many approaches have been explored to target and improve immune responses to CSP.
RTS,S/AS01 (Mosquirix TM ), the most advanced malaria vaccine to date, contains a section of the CSP central repeat region (18 NANP repeats with B-cell epitopes) and C-terminal (with T-cell epitopes).
In a large phase III study involving 15 459 infants (6-12 weeks old) and young children (5-17 months old) at 11 sites, RTS,S showed moderate vaccine-induced protection at 18 months (26% and 45%, respectively) which waned on follow-up. 59 In subjects receiving a booster at 20 months, the vaccine efficacy was ~36% in children (vs 28% in controls without the booster) and ~25% in infants (vs 18% in controls) at the end of a 48-month study period. 60 Fractional dosing of the third dose may increase the vaccine efficacy up to ~86%, 61 but this remains to be seen in endemic areas where efficacy in adult declined with an increase in malaria transmission. 62 After a positive review by the European Medicines Agency, RTS,S was recently rolled out for implementation in three African countries (Malawi, Kenya and Ghana). 63 Assuring earlier concerns that CSP diversity may impact vaccine efficiency, it is noteworthy that in the above large phase III trials, <10% of the parasites corresponded the CSP alleles used in the RTS,S. 64 Prime-boost viral vectored delivery platforms using chimpanzee adenoviruses (eg ChAd63) prime and a modified vaccinia strain Ankara (MVA) have been explored as alternative approaches to improve the efficacy of CSP-based vaccines. ChAd63-MVA CSP vaccine candidate induced high levels of antigen-specific antibodies and T-cell responses. 65 Nevertheless, its efficacy in a CHMI trial was poor, protecting only 1/15 subjects. 66 In vitro and rodent studies have suggested that CSP is dispensable in achieving sterile immunity and low levels of anti-CSP antibodies may aid in sporozoite invasion. 58,67,68 Other studies reported that the CSP repeat region, but not the C-terminal domain, induced antibody-dependent phagocytic activity that was protective against infection. 69 Thus, the modest protection induced by CSP-based vaccines, as compared to the sterile immunity observed in RAS, calls for exploration of alternative adjuvants, antigens and/or CSP epitopes as vaccine targets and increased focus on antibody functionality rather than quantity.
The genome of Pf reference strain 3D7 contains ~5400 genes. 70 Some of these genes encode for proteins that are essential for cell traversal (sporozoite microneme protein essential for cell traversal

| Innate host responses in the skin and the liver
The skin is the first defence layer against the malaria parasites. Apart from being a physical barrier, the skin harbours a diverse range of phenotypically and functionally distinct dendritic cells (DCs) and macrophages that interact with sporozoites, as described in mouse malarias ( Figure 1). 5,6 The contribution of these cells is challenging to study in humans considering the 'silent' clinical nature of malaria PE stages. Neutrophils and monocytes infiltrate the site of sporozoite inoculation, and mast cells have been reported to induce DCs and T-cell recruitment. 78,79 Remarkably, a rodent study reported that neutrophils and monocytes may not be critical in the development of sterile immunity. 78 Further work is needed to dissect the roles of neutrophils and monocytes in PE stage immunity.
Whilst the liver is known to be an autonomous and competent priming site for naïve CD8 + T cells, 80

Nutritional immunity may play a role in protection against
Plasmodium infections. In endemic settings, children with iron deficiency are protected against malaria. 96,97 The hepatic hormone hepcidin has been reported to increase across the malaria season in these settings. 98,99 Hepcidin restricts iron availability in the liver hence denying Plasmodium parasites a vital nutrient, and may protect against secondary liver-stage infections. 100 Supplementing children with iron in a malaria-endemic region was associated with increased malaria incidences and mortality. 101 Accordingly, targeting the nutritional requirements of the parasite is an alternative innate response to malaria infections. Early malaria vaccine studies reported increased production of anti-CSP antibodies in response to RAS, and these antibodies are associated with protection against reinfection. 18,22,24,103 In field and CHMI studies, antibody responses to other PE antigens such as LSA-1, TRAP and STARP have also been reported [104][105][106] and protected individuals may have higher antibody titres. [105][106][107] Passive transfer of monoclonal anti-sporozoite antibodies delayed patency of Pb infection in mice. 108 The effector activity of these antibodies may include blocking sporozoite motility, dermal exit and subsequent invasion of hepatocytes. 78,109 Antibodies may remove the surface coat protein of sporozoites in the skin and expose the parasites to their own pore-forming proteins. 110 Beyond inhibiting sporozoite mobility, antibodies also aid in sporozoite destruction through activation of the complement system, phagocytosis and Fc-mediated innate cell functions. 94,[111][112][113] Various field studies have reported that high antibody levels against sporozoites are required for effective and long-term protection. 105 [105][106][107] antibody titres have generally performed poorly as correlates of protection in malaria vaccine studies. 94,120 The modest efficacy of RTS,S in endemic regions suggests that the functionality and avidity of the antibodies, rather than the antibody titres, is a better correlate of immune protection to malaria. 94,113 In recent serological profiling studies, the functionality of antibodies was reported to be a better predictor of protection. 94 These antibodies were reported to induce NK cell effector functions, including activation and phagocytosis.

| Antibody responses, including targeting the parasites whilst in the skin
The hurdle with malaria infections is the inability to generate long-lasting protective immunity. This is compounded by the lack of appropriate surrogates of protection in field and CHMI studies.
Malaria-specific MBCs are elicited at levels comparable to conventional licensed vaccines 121 and can persist in naturally infected and travellers to endemic regions. 122 Like antibodies, malaria MBCs appear to increase with age and exposure. 123   The functional roles of CD4 + T cells are not limited to direct activity. As discussed before, CD4 + T cells may be involved in the licensing of the antigen-presenting DCs that prime effector CD8 + T cells.

| CD4 + T-cell effector mechanisms
The cytokines generally produced by CD4 + T cells may influence other immune cells involved in response to malaria and development of immunity. IL-4-producing CD4 + T cells sustain and expand the effector and memory Py-specific CD8 + T-cell pool. 87,137,138 In the absence of CD4 + T cells, the sporozoite-specific memory CD8 + T cells fail to protect against challenge infections in mice. 137 Some of the cytokines produced by CD4 + T cells, such as IFN-γ, IL-4, IL-5 and IL-10, enable B cells to undergo immunoglobulin class-switching. 102 A subset of CD4 + T cells, FOXP3 + regulatory T cells (T REG s), has been associated with poor development of CPS vaccine-induced immunity. 139 A recent study implicated a subset of T FH CD4 + T cells in the poor response of participants receiving RTS,S and ME-TRAP combinations. 77 Nonetheless, further studies are required to elucidate induction, regulation, maintenance and tissue requirements of CD4 + T cells in malaria PE stage immunity.

| CD8 + T-cell effector mechanisms, including liver-resident memory CD8 + T cells
CD8 + T cells are the primary effector cells against PE stages as seen in rodent, non-human primate and human studies. [140][141][142][143][144] As observed in Py, the responses by CD8 + T cells begin after they are primed by mature CD11c + DCs in the skin-draining lymph nodes. 8 Naïve CD8 + T cells do not exert antiparasitic activity, unless previously primed by antigen-presenting cells. 145 The CD8 + T cells with cognate receptors to the antigens presented by the DCs will differentiate to short-lived effector cells (SLEC) or memory precursor effector cells (MPEC) depending on the local cytokine environment and transcriptional factor profile. [146][147][148] Activated CD8 + T then undergo clonal expansion, which requires the presence of IL-2/IL-4 produced by CD4 + T cells. 87 The numbers of CD8 + T cells have been shown to increase rapidly after sporozoite inoculation. 86,145,149,150 The activation and proliferation of naïve CD8 + T cells are dose-dependent, and a successful response requires viable sporozoites. 5,53,151 The SLEC migrate to the liver to exert their effector properties whilst MPEC further differentiate to memory cells. 152,153 CD8 + T cells confer sterile immunity against Pb-independent of B cells or CD4 + T cells. 18 In rodent and nonhuman primate models, depletion of CD8 + T cells abrogates sterile immunity after RAS immunization, whilst their restoration reinstates the protection. 140,143 However, the effector mechanisms of these malaria PE-specific CD8 + T cells are not well characterized. In vivo imaging studies report that CD8 + T cells recognize cognate epitopes on the infected hepatocyte MHC-I and cluster around these cells. 154 Murine and vaccine studies have reported elevated CD8 + T-cell effector mediators including cytokines (IFN-γ and TNF) and/or proteins involved in contact-mediated cytotoxicity (perforin, TRAIL, FAS ligand and granzyme). 18,35,134,151,155 Surprisingly, CD8 + T cells lacking perforin, FAS ligand and perforin can still eliminate Py-and Pb-infected hepatocytes. 156,157 Malaria memory T cells are involved in patrolling, surveillance and rapid recruitment to the site of infection. 34,155,158 This enables a fast, effective, specific and durable protection against subsequent malaria infections. Pre-clinical and CHMI trials have shown induction and persistence of Pf-specific CD4 +159, 160 and CD8 + T cells. 144 In Pb and Py, CD8 + T memory cells have been described as CXCR3 hi CXCR6 hi CD62L -CD69 + liver-resident (T RM ), CXCR3 lo CXCR6 lo CD44 + CD62L -CD122circulating effector (T EM ), and CD44 + CD62L + CD122 + central memory (T CM ) cells, 157,161,162 and their effector immune responses are species-specific. 157 Nonetheless, the epitope signatures and correlates of CD8 + T memory cell protection are yet to be characterized.
Majority of the circulating CD8 + T memory cells in mouse studies are T EM but a small proportion of T CM has also been observed. 150,162 A large population of T EM cells is required for effective and longterm protection. 150

| Perspectives on immune responses to PE stages
Naturally acquired immunity in endemic areas is short-lived and non- immunity. 105,114,115,150,163 There is paucity of data on the quantity of CD4 + T cells required to induce sterile immunity. More work is also needed to understand how trained immunity of innate cells, which has recently been described, 165,166 may contribute to immune protection in PE stages. Various adjuvants including alum, ASO1 and viral vectors have been employed as immunostimulants and/or delivery systems for the existing vaccine candidates. 167 Adjuvants have the potential to induce and maintain large numbers of effector and memory immune cells, and the appropriate choice or combination of adjuvants may be the key to unlocking a malaria vaccine that confers sterile and long-lasting protection.
Very little is known regarding the regulation of immune responses to PE stages-the possible roles for regulatory T cells, cytokines and T H 1/T FH have been thoroughly explored in malaria blood stages. 168 Additionally, malaria blood-stage infections have been reported to downregulate PE stage immunity. 169,170 Checkpoint blockade has been explored in cancer and malaria blood-stage research, 171 and it is possible that some answers to the regulation of frequencies of anti-PE stage immune responses lie here. The contribution of inhibitory and other regulatory proteins, and their tissue-specific regulation, has not been widely studied in the context of malaria PE stages, but it is plausible that they are involved in a complex web of factors influencing protection against malaria.

| CON CLUS ION
Delivery of an efficient and long-lasting vaccine protection remains an ambitious goal that requires sustained efforts of all stakeholders.
Gaps in the existing parasite-host immunological crosstalk in both the skin and the liver during malaria PE stages need to be addressed first. Quantification and characterization of immune mechanisms have only started to emerge recently despite decades of research into an efficient malaria vaccine. Nonetheless, the identification of correlates of protection and protective malaria PE stage epitopes remain a work in progress. In the current review, we highlighted how protection to malaria sporozoites may rely on a fine, yet to be adequately described, balance between innate and adaptive immune responses. Utilizing advances in other fields such as systems biology and bioinformatics can inform the study of more immunological processes, which have proven challenging to study in the setting of a natural infection. Alternative efforts should include targeting novel sporozoite proteins, a multi-stage and multi-antigen vaccine, or a 'nutritional' vaccine that targets metabolic requirements of sporozoites.

CO N FLI C T O F I NTE R E S T
The authors declare no conflict of interest.

AUTH O R CO NTR I B UTI O N S
KMA and JCRH conceptualized the study and wrote the original draft of the manuscript. KMA, WJW and JCRH revised subsequent drafts and approved the final draft for publication.

PE E R R E V I E W
The peer review history for this article is available at https://publo ns.com/publo n/10.1111/pim.12795.

DATA AVA I L A B I L I T Y S TAT E M E N T
Data sharing is not applicable to this article as no data sets were generated or analysed during the current study.