Low immunogenicity of malaria pre‐erythrocytic stages can be overcome by vaccination

Abstract

Immunogenicity is considered one important criterion for progression of candidate vaccines to further clinical evaluation. We tested this assumption in an infection and vaccination model for malaria pre‐erythrocytic stages. We engineered Plasmodium berghei parasites that harbour a well‐characterised epitope for stimulation of CD8⁺ T cells, either as an antigen in the sporozoite surface‐expressed circumsporozoite protein or the parasitophorous vacuole membrane associated protein upregulated in sporozoites 4 (UIS4) expressed in exo‐erythrocytic forms (EEFs). We show that the antigen origin results in profound differences in immunogenicity with a sporozoite antigen eliciting robust, superior antigen‐specific CD8⁺ T‐cell responses, whilst an EEF antigen evokes poor responses. Despite their contrasting immunogenic properties, both sporozoite and EEF antigens gain access to antigen presentation pathways in hepatocytes, as recognition and targeting by vaccine‐induced effector CD8⁺ T cells results in high levels of protection when targeting either antigen. Our study is the first demonstration that poorly immunogenic EEF antigens do not preclude their susceptibility to antigen‐specific CD8⁺ T‐cell killing, which has wide‐ranging implications on antigen prioritisation for next‐generation pre‐erythrocytic malaria vaccines.

Key benchmarks for malaria vaccine design were investigated. Antigen immunogenicity and accessibility were studied with results indicating the proof‐of‐concept that a poorly immunogenic exo‐erythrocytic form (EEF) antigen is comparably vulnerable as a strongly immunogenic sporozoite antigen to targeting by vaccine‐induced effector CD8⁺ T cells.

The paper explained.

Problem

Antigen selection is critical for malaria vaccine discovery. To date, sporozoite antigens are prioritised according to their immunogenicity, accessibility and capacity to elicit inhibitory antibodies, and, to a limited extent, cellular responses. Very little is known about the vaccine potential of Plasmodium antigens expressed during liver infection.

Results

We have demonstrated in a mouse malaria model that despite having poor immunogenicity, antigens expressed during liver infection are excellent targets of vaccine‐induced protection through cellular immunity.

Impact

Our data provide a rationale for systematic evaluation of previously unrecognised parasite‐derived antigens for malaria vaccine discovery and pre‐clinical evaluation in formulations aimed at eliciting strong cell‐mediated immunity.

Introduction

Malaria caused by the apicomplexan parasite Plasmodium is responsible for more than 229 million clinical cases and over 409,000 deaths annually worldwide, with more than 97% of cases attributable to Plasmodium falciparum (WHO, 2020). Whilst current malaria control strategies have led to marked reduction in incidence rate, cases and mortality for the past 16 years, a highly efficacious vaccine is likely essential to approach the ambitious World Health Organisation’s (WHO) vision of “a world free of malaria”. Targeting the malaria pre‐erythrocytic stages, an obligatory and clinically silent phase of the parasite’s life cycle, is considered an ideal and attractive strategy for vaccination; inhibiting parasite infection of, and development in hepatocytes results in preclusion of both disease‐causing blood stages and transmissible sexual stages. Yet, despite intensive research for over 35 years, a highly efficacious pre‐erythrocytic stage vaccine remains elusive (Draper et al, 2018). An in‐depth characterisation of how the complex biology of pre‐erythrocytic stages influences the generation and protective efficacy of immune responses is warranted to inform the design of future malaria vaccines.

CD8⁺ T cells are crucial mediators of protective immunity to malaria pre‐erythrocytic stages (Doolan & Hoffman, 2000). Whilst often considered as a single phase of the parasite’s life cycle, the malaria pre‐erythrocytic stage is comprised of two different parasite forms: (i) sporozoites, which are motile extracellular parasites that are delivered by infected mosquitoes to the mammalian host, and (ii) exo‐erythrocytic forms (EEF; also known as liver stages), which are intracellular parasites resulting from the differentiation and growth of sporozoites inside a parasitophorous vacuole (PV) within hepatocytes (Hafalla et al, 2011). How these two spatially different parasite forms and the ensuing temporal expression of parasite‐derived antigens impact the magnitudes, kinetics and phenotypes of CD8⁺ T‐cell responses elicited following infection is poorly understood. Furthermore, the complexity within the pre‐erythrocytic stages has fuelled a long‐standing debate focussed on the contributions of distinct sporozoite and EEF antigens in parasite‐induced responses, and whether sporozoite or EEF proteins are better targets of vaccines.

Our current understanding of CD8⁺ T‐cell responses to malaria pre‐erythrocytic stages has been largely based on measuring responses to the H‐2‐K^d‐restricted epitopes of rodent P. yoelii (Py) (Weiss et al, 1990) and P. berghei (Pb) (Romero et al, 1989) circumsporozoite proteins (CSP), the major surface antigen of sporozoites. Many of these fundamental studies have focussed on using infections with irradiated sporozoites, the benchmark vaccine model for malaria. Infection with Py sporozoites elicits an expected T‐cell response typified by early activation and induction of effector CSP‐specific CD8⁺ T cells followed by contraction and establishment of quantifiable memory populations (Sano et al, 2001). CSP‐specific CD8⁺ T cells are primed by dendritic cells that cross‐present sporozoite antigens via the endosome‐to‐cytosol pathway (Cockburn et al, 2011). Yet, CSP is a unique antigen because it is expressed in both sporozoites and EEFs (Hollingdale et al, 1983). Whilst the expression of CSP mRNA ceases after sporozoite invasion, the protein on the parasite surface is stable and endures in EEFs during development in hepatocytes (Silvie et al, 2014). In vitro data indicate that primary hepatocytes process and present PbCSP‐derived peptides to CD8⁺ T cells in a proteasome‐dependent manner, involving export of antigen to the cytosol (Cockburn et al, 2011). Taken together, these data imply that sporozoite antigens induce quantifiable CD8⁺ T‐cell responses after infection. Antigens that have similar expression to the CSP, persisting to EEFs and with epitope determinants presented on hepatocytes, are excellent targets of CD8⁺ T cell‐based vaccines.

The paucity of EEF only‐specific epitopes has hindered not only our ability to understand the immune responses that are evoked whilst the parasite is in the liver, but also their utility as targets of vaccination. Accordingly, the contribution of EEF‐infected hepatocytes in the in vivo induction of CD8⁺ T‐cell responses is poorly understood. The liver is an organ where the primary activation of CD8⁺ T cells is generally biased towards the induction of tolerance (Thomson & Knolle, 2010; Bertolino & Bowen, 2015). Yet, studies in other model systems have demonstrated antigen‐specific primary activation within the liver (Bertolino et al, 2001). Another confounding issue with EEFs is their development in PVs with constrained access to the hepatocyte’s cytosol (Hafalla et al, 2011). Nonetheless, if CD8⁺ T cells specific for EEF antigens are primed, do they expand and contract with distinct kinetics? Moreover, are EEF‐specific epitopes efficiently generated for recognition and targeting by vaccine‐induced CD8⁺ T cells? Answers to these questions will be key for antigen selection and design of future malaria vaccines.

In this study, we compared the initiation and development of CD8⁺ T‐cell responses—elicited following parasite infection—to CSP, a sporozoite antigen, and to upregulated in infective sporozoites gene 4 (UIS4), an EEF‐specific vacuolar protein (Mueller et al, 2005). UIS4, a member of the early transcribed membrane protein (ETRAMP) family, is abundantly expressed in EEFs and associates with the PVM (Mueller et al, 2005). Whilst UIS4 mRNA expression is present in sporozoites, translation is repressed until when EEFs develop (Silvie et al, 2014). To control for epitope specificity, we generated Pb transgenic parasites that incorporate the MHC class I H‐2‐K^b epitope SIINFEKL, from ovalbumin, in either the CSP or UIS4 protein. The resulting transgenic parasites develop normally as wild‐type (WT) Pb in the mosquito vector and mammalian host. However, SIINFEKL would be expressed at the same time and space as its respective Plasmodium protein, enabling the CD8⁺ T‐cell response against these proteins to be tracked in an epitope‐specific physiological manner. In line with previous studies (Cockburn et al, 2011; Montagna et al, 2014), to augment low numbers of CD8⁺ T cells in the naïve response, cells from OT‐I mice, which express SIINFEKL‐specific T‐cell receptors (TCRs) on their CD8⁺ T cells, were initially adoptively transferred to mice prior to them receiving sporozoite immunisations. Furthermore, we evaluated the capacity of vaccine‐induced CD8⁺ T cells to target these parasites in a mouse challenge model. Our data show disparate immunogenic properties between a sporozoite and an EEF vacuolar membrane antigen but equivalent susceptibility to vaccine‐induced CD8⁺ T cells.

Results

Transgenic CSP^SIINFKEL and UIS4^SIINFEKL parasites display normal sporozoite motility and liver invasion

We generated, by double homologous recombination, transgenic Pb parasites expressing the immunodominant H‐2‐K^b‐restricted CD8⁺ T‐cell epitope of ovalbumin (SIINFEKL) in the context of the sporozoite surface antigen CSP or the EEF vacuolar membrane antigen UIS4 (Figs 1A and EV1A and B). Constructs included the TgDHFR/TS‐positive selection cassette and incorporated SIINFEKL in the context of the gene open reading frame. For CSP^SIINFEKL, SIINFEKL replaced SYIPSAEKI, the immunodominant H‐2‐K^d‐restricted CD8⁺ T‐cell epitope of CSP, which allowed for recognition in H‐2‐K^b‐carrying C57BL/6 mice. For UIS4^SIINFEKL, the SIINFEKL epitope was added to the immediate C terminus of the UIS4 protein. Appending the C terminus was chosen because it had been shown in Toxoplasma gondii that the potency of the immunodominant epitope of GRA6 was associated with its C‐terminal location, which may have enhanced the presentation by parasite‐infected cells (Feliu et al, 2013). Whilst undefined for UIS4 itself, it has been shown for several other ETRAMPs that the C terminus faces the host cell cytoplasm (Spielmann et al, 2003), which is likely to facilitate exposure to the MHC I complex.

Figure 1. Generation and characterisation of recombinant CSP^SIINFEKL and UIS4^SIINFEKL P. berghei parasites.

Plasmodium berghei parasites expressing the CD8⁺ T‐cell epitope of ovalbumin, SIINFEKL, in the context of CSP or UIS4 were generated using double homologous recombination.

1. Schematic overview of transgenic parasite lines. To generate CSP^SIINFEKL, SIINFEKL replaced amino acids SYIPSAEK in CSP. To generate UIS4^SIINFEKL, SIINFEKL was adjoined to the carboxyl‐terminus of the UIS4 protein.
2. Sporozoite immunofluorescent antibody staining of WT, CSP^SIINFEKL or UIS4^SIINFEKL sporozoites after gliding on BSA‐coated glass slides. Shown are microscopic images of the respective sporozoites that were stained with anti‐CSP (green), anti‐UIS4 (red) and nuclear stain Hoechst (blue). Scale bars, 10 μm. The numbers show mean percentage (± SD) of sporozoites with trails assessed from ≥ 220 sporozoites.
3. Fluorescent‐microscopic images of EEF‐infected Huh7 hepatoma cells. 24 and 48 h after infection with WT, CSP^SIINFEKL or UIS4^SIINFEKL sporozoites, the cells were fixed and stained with anti‐UIS4 (red), anti‐HSP70 (green) and the nuclear stain Hoechst (blue). Scale bars, 10 µm. The numbers show mean numbers (± SD) of intracellular parasites counted per well of 8‐well Labtek slides.

Data information: The data shown are representative from one of two independent experiments.

Figure EV1. Generation of transgenic CSP^SIINFEKL and UIS4^SIINFEKL P. berghei lines.

Plasmodium berghei parasites expressing the CD8⁺ T‐cell epitope of ovalbumin, SIINFEKL, in the context of CSP or UIS4 were generated using double homologous recombination, combining drug‐resistance selection (through incorporation of the dhfr/ts gene from Toxoplasma gondii) and cloning by limiting dilution to select for correctly recombined parasites.

• A, B
  Diagrams illustrate the reverse genetics strategy. (A) In CSP^SIINFEKL, SIINFEKL replaces the immunodominant CD8⁺ T‐cell epitope SYIPSAEK(I) of CSP. (B) In UIS4^SIINFEKL SIINFEKL is adjoined to the carboxyl‐terminus of the UIS4 protein. Purified schizonts of WT P. berghei ANKA were transfected with linearised plasmid by electroporation as described (Janse et al, 2006), and immediately injected intravenously in the tail vein of a mouse. The day after transfection, pyrimethamine (70 mg/l) was orally administered in the drinking water for selection of transgenic parasites. Transgenic clones were generated in mice by in vivo cloning by limiting dilution. Correct integration of the constructs and purity of the transgenic lines were verified by diagnostic PCR using primer combinations specific for the unmodified CSP or UIS4 locus, and for the 5’ and 3’ recombination events as indicated by lines, arrows and expected fragment sizes.
• C
  Oocyst midgut infectivity of mosquitoes infected with WT, CSP^SIINFEKL or UIS4^SIINFEKL. The mean percentage (± SD) of infected midguts was enumerated 10–14 days after infection (14 biological replicates).
• D
  Salivary glands were isolated from WT‐, CSP^SIINFEKL‐ or UIS4^SIINFEKL‐infected mosquitoes and mean sporozoite numbers (± SD) were enumerated between 18–30 days after infection (CSP^SIINFEKL, 21 biological replicates; UIS4^SIINFEKL, 18 biological replicates; WT, 18 biological replicates).
• E
  Numbers of intracellular parasites of hepatoma cells 24, 48 and 72 h after infection with WT, CSP^SIINFEKL or UIS4^SIINFEKL sporozoites. The cells were fixed, stained and then were enumerated from two independent experiments. Values are percent of intracellular parasites compared with the mean of WT intracellular parasites ± SD (*P < 0.05; one‐way ANOVA with Tukey’s multiple comparison test). See Appendix Table S4 for exact P‐values.
• F
  Kaplan–Meier curve showing the prepatency of WT, CSP^SIINFEKL or UIS4^SIINFEKL sporozoite infections after intravenous injection of 800 sporozoites (n = 4 per group).

The resulting parasites showed a phenotype comparable to WT parasites, with comparable mosquito infectivity and number of salivary gland sporozoites (Fig EV1C and D), functional sporozoite motility (Fig 1B) and normal invasive capacity and development inside hepatocytes (Figs 1C and EV1E). Thus, the introduced mutations to generate CSP^SIINFEKL and UIS4^SIINFEKL parasites did not interfere with the completion of the life cycle, in either mosquito vector or mouse. All C57BL/6 mice that received 800 sporozoites of either CSP^SIINFKEL or UIS4^SIINFEKL intravenously developed a detectable (patent) blood stage infection by day 4, comparable to infection with WT sporozoites (Fig EV1F).

Peripheral blood CD8⁺ T‐cell responses are superior if elicited by a sporozoite surface protein in contrast to a vacuolar membrane protein in the infected liver

We first wanted to determine whether the generated transgenic parasites allow antigen‐specific responses to be tracked using SIINFEKL as a surrogate CD8⁺ T‐cell epitope for sporozoite surface and EEF vacuolar membrane antigens. To this end, we assessed the kinetics of the CD8⁺ T‐cell response following intravenous immunisation with CSP^SIINFEKL or UIS4^SIINFEKL sporozoites. To augment the CD8⁺ T‐cell response, mice were adoptively transferred with 2 × 10⁶ OT‐I cells expressing a SIINFEKL‐specific TCR (Cockburn et al, 2011), prior to receiving 10,000 γ‐radiation attenuated WT, CSP^SIINFEKL or UIS4^SIINFEKL sporozoites. Prior work showed that γ‐radiation attenuation of P. berghei sporozoites does not impact host cell invasion and UIS4 expression (Mac‐Daniel et al, 2014).

Peripheral blood was taken at days 4, 7, 14, 21, 42 and 88 after immunisation and CD8⁺ T‐cell responses were analysed after staining with H‐2‐K^b‐SIINFEKL pentamers and for CD11a, a marker for antigen‐experienced T cells (Rai et al, 2009; Schmidt et al, 2010) (Fig 2A and B). Additional gating strategies for the characterisation of K^b‐SIINFEKL⁺ CD8⁺ T cells are shown in Appendix Fig S1. A substantial proportion of K^b‐SIINFEKL⁺ CD11a^hi CD8⁺ T cells were observed in mice immunised with CSP^SIINFEKL; the response was highest on day 4, reaching 5% of all antigen‐experienced CD8⁺ T cells, and declined steadily until day 21, when the response stabilised and remained unchanged for several weeks (Fig 2C). In marked contrast, UIS4^SIINFEKL immunisation induced a poor CD8⁺ T‐cell response; the proportion of K^b‐SIINFEKL⁺ CD11a^hi CD8⁺ T cells was only higher than the control groups at day 4 after immunisation, and the response remained within background levels for the duration of the experiment. Control groups included mice receiving OT‐I cells only or in addition to γ‐radiation attenuated WT sporozoites, which lack SIINFEKL sequences.

Figure 2. Sporozoite antigen exposure elicits superior peripheral blood CD8⁺ T‐cell responses.

• A
  Schematic overview of experimental design. C57BL/6 mice received 2 × 10⁶ OT‐I cells alone or were additionally immunised with 10,000 γ‐radiation attenuated parasites intravenously.
• B
  Flow cytometry plots show the gating strategy for identifying K^b‐SIINFEKL⁺ CD11a^hi CD8⁺ T cells.
• C
  Temporal analysis of antigen‐specific CD8⁺ T‐cell responses. Peripheral blood was obtained on days 4, 7, 14, 21, 42 and 88 post‐immunisation with WT (triangles; n = 3–9), CSP^SIINFEKL (orange squares; n = 4–8) or UIS4^SIINFEKL (blue circles; n = 4–10) sporozoites, or no parasites (diamonds; n = 2–5) and stained for K^b‐SIINFEKL⁺ CD11a^hi CD8⁺ T cells. Line graph shows mean values (± SEM) from representative experiments. (*P < 0.05; **P < 0.01; ***P < 0.001; Welch’s t‐test comparing CSP^SIINFEKL and UIS4^SIINFEKL). See Appendix Table S1 for the number of mice used per timepoint and per group. See Appendix Table S4 for exact P‐values.
• D–H
  C57BL/6 mice (n = 4 per group), which received 2 × 10⁶ CFSE‐labelled OT‐I splenocytes, were immunised with 10,000 γ‐radiation attenuated WT, CSP^SIINFEKL or UIS4^SIINFEKL sporozoites intravenously. 5 days later, mice were sacrificed, spleens harvested and splenocytes assessed for (D) CFSE dilution of CD8⁺ T cells, (E) CFSE dilution of antigen‐experienced K^b‐SIINFEKL⁺ CD11a^hi CD8⁺ T cells and stained ex vivo (F‐H) for effector CD8⁺ T‐cell surface markers. Shown are flow cytometry plots of K^b‐SIINFEKL co‐staining with markers of effector phenotypes: (F) CD11a^hi, (G) CD62L^lo and (H) CD49d^hi.

The poor CD8⁺ T‐cell response induced by UIS4^SIINFEKL sporozoites, as compared to CSP^SIINFEKL, led us to characterise the early events in the proliferation and differentiation of these cells. Mice were adoptively transferred with CFSE‐labelled OT‐I cells and immunised with γ‐radiation attenuated WT, CSP^SIINFEKL or UIS4^SIINFEKL sporozoites. As shown by gating on CD8⁺ T cells (Fig 2D and E), after 5 days, immunisation with CSP^SIINFEKL sporozoites led to greater expansion of K^b‐SIINFEKL⁺ CD8⁺ T cells, as compared to that observed with UIS4^SIINFEKL sporozoites (Fig EV2A), in good agreement with the peripheral blood data described above (Fig 2C). Consistent with the activation of these cells, the proliferation of antigen‐specific CD8⁺ T cells by both parasites was associated with the development of effector and effector memory phenotypes as evidenced by upregulation of CD11a and CD49d, and downregulation of CD62L, respectively (Figs 2F–H and EV2B–D).

Figure EV2. Sporozoite surface antigen leads to greater expansion of antigen‐specific CD8⁺ T cells than EEF vacuolar membrane antigen.

• A–D
  C57BL/6 mice (n = 4 per group), which received 2 × 10⁶ CFSE‐labelled OT‐I splenocytes, were immunised with 10,000 γ‐radiation attenuated WT, CSP^SIINFEKL or UIS4^SIINFEKL sporozoites intravenously. 5 days later, mice were sacrificed, spleens harvested and splenocytes assessed for (A) CFSE dilution of antigen‐experienced K^b‐SIINFEKL⁺ CD8⁺ T cells and stained ex vivo for effector CD8⁺ T‐cell surface markers (B) CD11a^hi, (C) CD62L^lo and (D) CD49d^hi. Bar charts show mean values (± SEM) from representative experiments (***P < 0.001; one‐way ANOVA with Tukey’s multiple comparison test). See Appendix Table S4 for exact P‐values.

Taken together, these findings establish that immunisations with γ‐radiation attenuated CSP^SIINFEKL and UIS4^SIINFEKL sporozoites permit antigen‐specific responses to be tracked longitudinally in the peripheral blood. Importantly, we demonstrate that a sporozoite surface protein evokes a CD8⁺ T‐cell response of superior magnitude than an EEF vacuolar membrane protein following immunisation with malaria sporozoites.

Higher magnitude of splenic and intrahepatic CD8 ⁺ T‐cell responses to a sporozoite antigen

Previous research has shown that CD8⁺ T cells are primed primarily in the spleen following intravenous immunisation with malaria sporozoites (Lau et al, 2014) and that liver lymphocytes form a front‐line defence against developing EEFs in hepatocytes (Guebre‐Xabier et al, 1999; Fernandez‐Ruiz et al, 2016). Thus, we further analysed the development of CD8⁺ T‐cell responses in the spleens and livers of mice adoptively transferred with OT‐I cells and intravenously immunised with γ‐radiation attenuated WT, CSP^SIINFEKL or UIS4^SIINFEKL sporozoites. Consistent with our aforementioned results, surface staining of splenic and liver lymphocytes showed higher proportion and absolute numbers of K^b‐SIINFEKL⁺ CD11a^hi CD8⁺ T cells at day 14 and day 42 following immunisation with CSP^SIINFEKL compared with UIS4^SIINFEKL sporozoites (Fig 3A–D). In addition, the proportions of CD11a^hi, CD62L^lo, CD49d^hi and CD44^hi expressing splenic and liver CD8⁺ T cells, elicited by both CSP^SIINFEKL and UIS4^SIINFEKL sporozoites, were comparable, indicative of effector and effector memory cell phenotypes (Fig EV3). Although low, the numbers of antigen‐specific CD8⁺ T cells induced by UIS4^SIINFEKL sporozoites were within the detection limits of the assay.

Figure 3. Sporozoite surface antigen induces a higher CD8⁺ T‐cell response than EEF vacuolar membrane antigen in the spleen and liver.

• A
  Schematic overview of experimental design. C57BL/6 mice received 2 × 10⁶ OT‐I cells alone (n = 2–4) or were additionally immunised with 10,000 γ‐radiation attenuated WT (n = 3–6), CSP^SIINFEKL (n = 4) or UIS4^SIINFEKL (n = 4) sporozoites intravenously. See Appendix Table S2 and S3 for the number of mice used for pentamer staining and peptide restimulation. Spleens and livers were harvested either at day 14 or day 42.
• B–G
  Proportions and numbers of (B‐D) K^b‐SIINFEKL⁺ CD8⁺ T cells were enumerated, or (E‐G) IFN‐γ‐secreting CD8⁺ T cells following restimulation ex vivo with SIINFEKL peptide were quantified. Flow cytometry plots show representative percentages of CD8⁺ T cells co‐stained with CD11a and (B) K^b‐SIINFEKL or (E) IFN‐γ. The upper panel of bar charts (C, F) show the percentage of co‐stained CD8⁺ T cells and the lower panel (D, G) the absolute cell counts. Bar charts show mean values (± SEM) from representative experiments (*P < 0.05; **P < 0.01; ***P < 0.001; one‐way ANOVA with Tukey’s multiple comparison test). See Appendix Table S4 for exact P‐values.

Figure EV3. Sporozoite surface antigen induces a greater effector CD8⁺ T‐cell phenotype than EEF vacuolar membrane antigen.

C57BL/6 mice received 2 × 10⁶ OT‐I cells alone (n = 2) or were additionally immunised with 10,000 γ‐radiation attenuated WT (n = 3), CSP^SIINFEKL (n = 4) or UIS4^SIINFEKL (n = 4) sporozoites intravenously. Spleens and livers were harvested either 14 or 42 days later, and proportions of CD8⁺ T cells expressing effector surface markers were quantified. Flow cytometry plots show representative percentages of CD8⁺ T cells co‐staining K^b‐SIINFEKL and markers of effector phenotype (CD11a^hi, CD49d^hi, CD62L^lo, CD44^hi).

To assess for effector functions, splenic and liver lymphocytes were stimulated ex vivo with the SIINFEKL peptide. Generally, higher numbers (proportion and absolute numbers) of IFN‐γ‐secreting CD8⁺ T cells were observed at day 14 and day 42 following immunisation with CSP^SIINFEKL compared with UIS4^SIINFEKL sporozoites (Fig 3E–G). In addition, these CD8⁺ T cells also expressed TNF and IL‐2 in a similar differential pattern, suggesting potential polyfunctionality (Fig EV4). Additional gating strategies for the characterisation of IFN‐γ‐secreting CD8⁺ T cells are shown in Appendix Fig S2.

Figure EV4. Antigen‐experienced SIINFEKL‐specific CD8⁺ T cells also produce TNF and IL‐2.

C57BL/6 mice received 2 × 10⁶ OT‐I cells alone (n = 4) or were additionally immunised with 10,000 γ‐radiation attenuated WT (n = 6), CSP^SIINFEKL (n = 4) or UIS4^SIINFEKL (n = 4) sporozoites intravenously. Spleens and livers were harvested either 14 or 42 days after immunisation and lymphocytes restimulated ex vivo with SIINFEKL peptide at 10 μg/ml per well for 5–6 h. The upper panel of bar charts show the percentage of CD11a^hi TNF secreting CD8⁺ T cells, the bottom panel CD11a^hi IL‐2 secreting CD8⁺ T cells. This is a representation of one experiment from two experiments performed. Bar charts show mean values (±SEM) from representative experiments (*P < 0.05, **P < 0.01, ***P < 0.001; one‐way ANOVA with Tukey’s multiple comparison test). See Appendix Table S4 for exact P‐values.

Altogether, even though effector and effector memory CD8⁺ T‐cell responses can be detected against both sporozoite surface protein and EEF vacuolar membrane protein antigens following immunisation with γ‐radiation attenuated sporozoites, the two antigens show a striking difference in the magnitude of CD8⁺ T‐cell responses they induce.

Higher magnitude of endogenously produced splenic and intrahepatic antigen‐specific CD8 ⁺ T‐cell responses to a sporozoite antigen

Previous work tracking responses to SIINFEKL‐tagged proteins has used adoptively transferred cells from OT‐I mice, with all T cells from these mice expressing T‐cell receptors specific to SIINFEKL (Cockburn et al, 2011; Montagna et al, 2014). We employed this robust approach by adoptively transferring a fixed amount of OT‐I splenocytes in order to augment the response and allow visualisation (Figs 2 and 3). Next, we wanted to explore whether we can capture the endogenous K^b‐SIINFEKL⁺ CD11a^hi CD8⁺ T‐cell population, which is elicited by intravenous immunisation with γ‐radiation attenuated sporozoites without OT‐I cell transfer. We performed ex vivo restimulation of lymphocytes with SIINFEKL peptide followed by flow cytometry and were able to clearly identify the endogenous population with a trend complementary to our earlier results (Fig 4A–D). Immunisation with CSP^SIINFEKL sporozoites elicited a superior splenic and liver CD8⁺ T‐cell response than with UIS4^SIINFEKL sporozoites. As expected, the proportion and absolute cell numbers were considerably lower than with adoptive transfer of OT‐I cells, but this did not preclude the ability to detect IFN‐γ‐secreting CD8⁺ T cells and capture the differences between the two groups.

Figure 4. Higher magnitude of endogenous SIINFEKL‐specific CD8⁺ T‐cell responses induced by a sporozoite antigen.

• A
  Schematic overview of experimental design.
• B–G
  C57BL/6 mice received 10,000 γ‐radiation attenuated WT, CSP^SIINFEKL or UIS4^SIINFEKL sporozoites, either (B‐D) intravenously (n = 6–8 mice per group) or (E‐G) intradermally (n = 4 mice per group). Additional control mice did not receive sporozoites (n = 6–8 intravenously, n = 4 intradermally). Spleens and livers were harvested either at day 14 (n = 5 livers and n = 6 spleens for intravenous immunisation) or day 42 (n = 8 for intravenous immunisation), and IFN‐γ‐secreting lymphocytes following restimulation ex vivo with SIINFEKL peptide were quantified. Flow cytometry plots show representative percentages of CD8⁺ T cells co‐stained with IFN‐γ and CD11a (B, E). The upper panel of bar charts (C, F) show the percentage of CD11a^hi IFN‐γ⁺ CD8⁺ T cells and the lower panel (D, G) the absolute cell counts. Bar charts show mean values (± SEM) from representative experiments (*P < 0.05; **P < 0.01; ***P < 0.001; one‐way ANOVA with Tukey’s multiple comparison test). See Appendix Table S4 for exact P‐values.

Under normal conditions of transmission, sporozoites are delivered into the host skin by mosquito bite. All preceding immunisation experiments were performed with parasites injected intravenously. As a proxy for the natural route of infection, whilst ensuring consistent quantities of parasites were inoculated, γ‐radiation‐attenuated CSP^SIINFEKL and UIS4^SIINFEKL sporozoites were injected via the intradermal route into the ear pinnae. Under these conditions, CSP still induced a greater number of IFN‐γ‐secreting SIINFEKL‐specific CD8⁺ T cells following restimulation with SIINFEKL compared with UIS4, with a comparable magnitude as after intravenous injection (Fig 4E–G; Appendix Fig S3A). Thus, these biologically and immunologically more appropriate data entirely recapitulate the strong immunogenicity of a sporozoite surface antigen compared with an EEF vacuolar membrane protein.

Increasing the amount of EEF vacuolar membrane antigen does not impact its immunogenicity

CSP and UIS4 are critical proteins expressed by the sporozoite and EEF, respectively, and both proteins are important for survival and succession into the subsequent life stage and parasite form (Mueller et al, 2005; Coppi et al, 2011; Silvie et al, 2014). Previous studies have shown that the magnitude of the CD8⁺ T‐cell response to a sporozoite surface antigen depended on the amount of parasites used for immunisation (Hafalla et al, 2002). Hence, poor immunogenicity of an EEF vacuolar membrane protein could be a result of the lower level of protein expression during parasite infection. It is possible to enhance CD8⁺ T‐cell responses by increasing the number of parasites used for immunisation (Hafalla et al, 2002). Therefore, we immunised groups of mice with 8,000 γ‐radiation‐attenuated CSP^SIINFEKL, 8,000 UIS4^SIINFEKL or 64,000 UIS4^SIINFEKL sporozoites and compared the magnitude of the elicited antigen‐specific responses. Strikingly, the CD8⁺ T‐cell response following 8× sporozoite immunisation dose with UIS4^SIINFEKL did not increase proportionally and was not significantly higher than immunisation with a 1× dose (Fig 5A–C; Appendix Fig S3B). This result suggests that, in the context of attenuated sporozoite immunisation, EEF vacuolar membrane antigens are poorly immunogenic and increasing antigen fails to substantially improve the magnitude of CD8⁺ T‐cell responses.

Figure 5. Increasing antigen dose does not improve antigen‐specific CD8⁺ T‐cell responses to an EEF vacuolar membrane protein.

1. Schematic overview of experimental design. C57BL/6 mice (n = 4 per group) received an intravenous dose of 8,000 γ‐radiation attenuated CSP^SIINFEKL or UIS4^SIINFEKL sporozoites or 64,000 γ‐radiation attenuated UIS4^SIINFEKL sporozoites. Spleens and livers were harvested at day 12 and IFN‐γ‐secreting lymphocytes following restimulation ex vivo with SIINFEKL peptide were quantified.
2. Flow cytometry plots show representative CD8⁺ T cells co‐stained with IFN‐γ and CD11a.
3. The upper panel of bar charts show the percentage of CD11a^hi IFN‐γ⁺ CD8⁺ T cells and the lower panel the absolute cell counts. Bar charts show mean values (± SEM) from representative experiments (***P < 0.001; one‐way ANOVA with Tukey’s multiple comparison test). See Appendix Table S4 for exact P‐values.

Immunogenicity of parasite antigens does not predict effector responses following vaccination

Our findings thus far showed that sporozoite surface proteins appear more immunogenic than EEF vacuolar membrane proteins and raised an intriguing and important question; does immunogenicity predict susceptibility to vaccine‐induced effector responses? To address this, we vaccinated mice, which had received OT‐I cells, with a recombinant adenovirus expressing full‐length ovalbumin (de Cassan et al, 2011) (Fig 6A). This vaccination protocol resulted in frequencies of ~7.5% SIINFEKL‐specific CD8⁺ T cells in peripheral blood (Fig 6B and C; Appendix Fig S3C). Vaccinated mice were then challenged with live CSP^SIINFEKL or UIS4^SIINFEKL sporozoites, and protection was assessed 19 days after vaccination by two complementary assays; (i) determination of the reduction of parasite load in the liver 42 h after sporozoite challenge (Fig 6D) and (ii) induction of sterile protection (Fig 6E). Vaccinated mice challenged with CSP^SIINFEKL or UIS4^SIINFEKL sporozoites showed a dramatic reduction in parasite load in the liver (Fig 6D) as compared to vaccinated mice challenged with WT parasites. Strikingly, contrary to the differential CD8⁺ T‐cell responses induced by CSP and UIS4, there was no statistical difference in the protection observed when vaccinated mice were challenged with either CSP^SIINFEKL or UIS4^SIINFEKL sporozoites. Consistent with these findings, both groups of vaccinated mice challenged with either CSP^SIINFEKL or UIS4^SIINFEKL sporozoites exhibited sterile protection of comparable levels (Fig 6E; Table 1) determined by microscopic analysis of blood smears post‐challenge. These findings indicate that spatial and temporal aspects of antigen expression may affect protein immunogenicity in the context of parasitic infection but not necessarily the same target’s susceptibility for antigen‐specific CD8⁺ T‐cell killing.

Figure 6. Both sporozoite surface and EEF vacuolar membrane antigens are susceptible to vaccine‐induced CD8⁺ T cells for killing, resulting in sterile protection.

• A
  Schematic overview of experimental design. Mice received 2 × 10⁶ OT‐I splenocytes, and 1 day later, they were either vaccinated with 1 × 10⁸ ifu recombinant AdHu5 expressing whole ovalbumin (AdOVA) (n = 5 per challenge group) or left untreated (n = 4 per challenge group).
• B, C
  Flow cytometry (B) and scatter plots (C) represent CD8⁺ T cells derived from peripheral blood co‐stained with IFN‐γ and CD11a, following ex vivo restimulation with SIINFEKL.
• D
  Protective efficacy as measured by quantitative real‐time PCR. Groups of mice were vaccinated as described and challenged 19 days later with 10,000 WT (n = 7 non‐vaccinated, n = 9 vaccinated), CSP^SIINFEKL (n = 6 non‐vaccinated, n = 4 vaccinated) or UIS4^SIINFEKL sporozoites (n = 3 non‐vaccinated, n = 11 vaccinated). 42 h later livers were removed, and parasite load was assessed by qPCR. Plots show the relative parasite load of mice in each condition (**P < 0.01; Mann–Whitney U‐test). See Appendix Table S4 for exact P‐values.
• E
  Proportion of sterile protection after immunisation. Mice (vaccinated group, n = 8; control group, n = 4) were vaccinated as described and were challenged with 800 WT, CSP^SIINFEKL or UIS4^SIINFEKL sporozoites. Daily blood smears were taken from day 3 to 14 post‐challenge to check for parasitaemia.

Data information: Data for B‐E are representative of two experiments performed with scatter plots showing mean values (± SEM).

Table 1.

Prepatency of AdOVA immunised and challenged mice. ^a

Challenge	WT	WT
AdOVA Immunisation	−	+	−	+	−	+
Infected / Challenged	4 / 4	8 / 8	4 / 4	2 / 8	4 / 4	1 / 8
Prepatency (days)	3.75	4	3.35	(5) b	3.5	(5) b

^aAll mice received 2×10⁶ OT‐1 cells 1 day prior to immunisation.

^bBrackets are indicative of the small proportion of mice that became blood stage‐positive at the day indicated, while all others remained negative until day 14.

Discussion

The malaria pre‐erythrocytic stages have been a prime target for the development of a Pf vaccine for more than 35 years. Indeed, RTS,S/AS01, the most advanced malaria sub‐unit vaccine candidate to date, is based on PfCSP, the major surface protein of sporozoites (Cohen et al, 2010). Yet, final results of the Phase III trial showed that RTS,S/AS01 offers only modest efficacy, which rapidly wanes over time (Rts SCTP, 2015). Thus, there is an imperative need not only to widen the pursuit for new sub‐unit vaccine candidates, but also to radically improve the antigen selection process. Antigens are generally prioritised based on a range of criteria, including their immunogenicity in the context of parasitic infection. We examined this notion in an infection and vaccination model for Pb pre‐erythrocytic stages. We evaluated whether antigen immunogenicity and accessibility are decisive features of vaccine design.

The malaria pre‐erythrocytic stages consist of two spatially different parasite forms: extracellular sporozoites and intracellular EEFs. The transformation of sporozoites into EEFs involves regulation at both transcriptional (Iwanaga et al, 2012) and translational (Muller et al, 2011; Zhang et al, 2016) levels, resulting in divergent spatial and temporal expression of many antigens that are distinct for each parasite form (Tarun et al, 2008). Whilst our current understanding of immune responses to malaria pre‐erythrocytic stages has focussed on CSP, the lack of well‐defined epitopes that are expressed only by EEFs has restrained fundamental studies investigating the contributions of EEF antigens in parasite‐induced CD8⁺ T‐cell responses and their value as target of vaccines.

In this study, we contrasted the development of CD8⁺ T‐cell responses induced by CSP and UIS4, two major proteins expressed by sporozoites and EEFs, respectively. We generated transgenic Pb parasites where SIINFEKL is expressed as part of either CSP or UIS4, allowing the presentation of the epitope at the same space and time as the respective protein. This approach is in contrast to a more common strategy of expressing the whole, or a fragment of, ovalbumin inserted as a transgene into the Pb genome, and then tracking the immune response elicited by an extraneous molecule (Lin et al, 2014; Montagna et al, 2014). Since CSP is expressed in both sporozoites and EEFs, the processing and presentation of the SIINFEKL in CSP^SIINFEKL occurs as soon as sporozoites are inoculated and are able to interact with CD11c⁺ dendritic cells, which cross‐present antigens via an endosome‐to‐cytosol pathway (Cockburn et al, 2011); CSP also has direct access to the hepatocyte’s cytosol for processing and presentation of the CSP‐derived epitope (Cockburn et al, 2011). Since UIS4 is expressed only in the PVM of EEFs, processing and presentation of the epitope in UIS4^SIINFEKL are restricted to just hepatocytes, although a recent report has suggested a role for monocyte‐derived CD11c⁺ cells in acquiring parasites during EEF stages (Kurup et al, 2019).

Our results establish that following sporozoite immunisation, a sporozoite surface protein induces superior CD8⁺ T‐cell responses—as measured both by pentamer staining and by IFN‐γ secretion following peptide stimulation—than an EEF vacuolar membrane protein. Detailed kinetic and phenotypic analysis of the development of antigen‐specific CD8⁺ T cells to both CSP and UIS4 revealed that the responses differ in magnitude, demonstrating the ability of both antigens to elicit effector and effector memory responses. There was no difference in our results whether sporozoites are delivered using the commonly used intravenous immunisation or the more physiological intradermal route. We also showed that increasing the number of UIS4^SIINFEKL parasites used for immunisation did not augment CD8⁺ T‐cell responses, signifying that the poor immunogenicity of an EEF vacuolar membrane protein is not due to the level of UIS4 expression during parasite infection. Our findings support the idea that EEF antigens have minimal contributions to the magnitude of immune responses following whole sporozoite immunisation, which corroborates with prior data showing that hepatocytes are poor at priming T‐cell responses (Thomson & Knolle, 2010; Bertolino & Bowen, 2015).

In our initial experiments, mice received cells from transgenic OT‐I mice, which express SIINFEKL‐specific TCRs on their CD8⁺ T cells, prior to sporozoite immunisation. The results from these experiments were compared with experimental set‐ups, where mice did not receive OT‐I cells; the results of the latter solely quantified the more physiological endogenous CD8⁺ T‐cell response. As expected, mice that received transgenic CD8⁺ T cells elicited higher CD8⁺ T‐cell responses—of one order of magnitude—as compared to endogenous responses. Nonetheless, the trend of frequencies is consistent in establishing that a sporozoite surface antigen induces a higher CD8⁺ T‐cell response than EEF vacuolar membrane antigen. It is of fundamental interest to understand further the fine specificities (i.e. proliferative capacity, phenotypic and functional characteristics) and contributions of the transferred versus endogenous CD8⁺ T cells in the resulting immune response. Further investigations using congenic mice are warranted to address these limitations.

CD8⁺ T cells recognise peptides directly processed and presented by parasitised hepatocytes to exert their functions (Bongfen et al, 2007; Chakravarty et al, 2007; Cockburn et al, 2011; Balam et al, 2012; Huang et al, 2015). Thus, we also evaluated whether sporozoite or EEF‐derived epitopes are accessible, and therefore susceptible to vaccine‐induced effector CD8⁺ T cells. Regardless of their differing immunogenicities in the context of parasitic infection, we demonstrated that both sporozoite and EEF antigens are effectively targeted by antigen‐specific effector CD8⁺ T cells, which were generated by vaccinating mice that received OT‐I cells with a recombinant adenovirus expressing the cognate epitope. The transfer of OT‐I ensured the induction of high levels of vaccine‐induced effector CD8⁺ T cells in a single vaccination, with the ensuing frequencies comparable to those obtained by prime‐boost vaccinations with recombinant adeno‐ followed by vaccinia viruses (Bruna‐Romero et al, 2001; Gilbert et al, 2002; Reyes‐Sandoval et al, 2010; Reyes‐Sandoval et al, 2011), or with peptide‐loaded dendritic cells followed by recombinant Listeria (Doll et al, 2016). These prime‐boost strategies have consistently shown to induce high numbers of antigen‐specific CD8⁺ T cells (Bruna‐Romero et al, 2001; Gilbert et al, 2002; Reyes‐Sandoval et al, 2010; Reyes‐Sandoval et al, 2011; Doll et al, 2016) necessary for protection (Schmidt et al, 2010). Previous work in mouse models and CSP‐based adenovirus vaccines showed the generation of endogenous responses that yielded only up to 40% sterile immunity, despite two orders of magnitude in the reduction in parasite load in the liver after challenge with viable sporozoites (Rodrigues et al, 1997; Rodrigues et al, 1998). Importantly, using the OT‐I transfer model, our study shows that mice harbouring similarly high levels of vaccine‐induced, antigen‐specific CD8⁺ T cells were comparably protected when challenged with either CSP^SIINFEKL or UIS4^SIINFEKL. These findings imply that both sporozoite and EEF antigens comparably access the antigen presentation pathways in hepatocytes leading to recognition of defined epitopes. Our study reiterates that high levels of vaccine‐induced, antigen‐specific CD8⁺ T cells are needed to achieve protection in mice, and likely in humans. Our work provides proof‐of‐concept for vaccines targeting the complete malaria pre‐erythrocytic stages. The translational challenge will now be to design vaccine formulations, which evoke and maintain high levels of antigen‐specific human CD8⁺ T cells either given as single dose or as part of a prime‐boost approach.

Our study is the first demonstration that poor natural immunogenicity, in this case of an EEF antigen, does not preclude antigen‐specific CD8⁺ T‐cell killing. Our findings that antigen immunogenicity in this context is an inadequate predictor of vaccine efficacy have wide‐ranging translational implications on antigen prioritisation for the design and testing of next‐generation pre‐erythrocytic Pf vaccines. Thus, the strategy to screen for T‐cell responses in naturally infected or sporozoite‐immune human volunteers to pre‐clinically prioritise vaccine candidates requires some form of reassessment. Conventional immunological assays, aimed at identifying highly immunogenic antigens, may fail to discover those candidates with the potential to induce superior levels of protective immunity. For this proof‐of‐concept study, we selected CSP and UIS4 as the best characterised representatives of sporozoite and EEF vacuolar antigens. It will be interesting to systematically compare global allelic diversity of sporozoite and EEF antigens. The observation of allele‐specific RTS,S/AS01 vaccine efficacy and low proportions of vaccine‐matched CSP alleles in sub‐Saharan Africa (Neafsey et al, 2015) underscores the potential critical advantage of EEF antigens, which likely display only moderate diversity across P. falciparum strains vis‐à‐vis extended diversity of sporozoite surface antigens. It is noteworthy that for other stages of malaria infection, conserved antigens that give limited or no responses, e.g. PfRH5 (Douglas et al, 2011; Osier et al, 2014) and sexual stage antigens (Kapulu et al, 2015), are promising antibody targets for malaria vaccines under clinical development. Together, our study establishes that EEF‐expressed antigens are prospective vaccine targets and thus effectively increases the pool of existing pre‐erythrocytic stage vaccine targets for the malaria community to explore and translate to clinical trials.

A key direction for future research will be finding new assays to easily distinguish and validate promising vaccine targets, namely those antigens that can protect, via susceptibility to vaccine‐induced CD8⁺ T cells, rather than those that naturally induce strongly immunogenic responses. In addition, our findings open new questions as to whether EEF‐expressed antigens are optimal antigens for the generation of endogenous liver‐resident memory (T_RM) CD8⁺ T cells. T_RM cells were recently identified to persist in and continuously perambulate the liver to evoke immune responses following antigen challenge (Fernandez‐Ruiz et al, 2016; Valencia‐Hernandez et al, 2020). Further work, utilising congenic mice, is warranted to characterise in detail the intrahepatic cells induced by CSP^SIINFEKL and UIS4^SIINFEKL parasites. There is current translational interest in exploiting T_RM cells for vaccine strategies against malaria pre‐erythrocytic stages. Current pre‐clinical approaches are aimed at initially activating T cells in the spleen by a priming vaccine, and another delivery system is used to trap these cells in the liver to form T_RM cells. It is conceivable that EEF‐expressed antigens could already prime T_RM cells, which can be boosted by novel vaccine delivery systems.

Ultimately, the molecular mechanisms of presentation of EEF antigens, those expressed in the PVM and within the parasite itself, onto the surface of infected hepatocytes remains to be fully understood. Determination of the processes involved in parasite antigen presentation in the pre‐erythrocytic stages of malaria may elucidate links to protection and the identification of hitherto neglected antigens that could drive the development of a highly efficacious malaria vaccine.

Materials and Methods

Reagents and Tools table

Reagent	Source	Identifier
Antibodies
Anti‐mouse CD3e‐V500 (500A2)	BD Biosciences	Cat# 560771, RRID: AB_1937314
Anti‐mouse CD8a‐PerCP‐Cy5.5 (53‐6.7)	Thermo Fisher Scientific	Cat# 45‐0081‐82, RRID: AB_1107004
Anti‐mouse CD11a‐e450 (M17/4)	Thermo Fisher Scientific	Cat#48‐0111‐82, RRID: AB_11064445
Anti‐mouse CD11a‐FITC (M17/4)	Thermo Fisher Scientific	Cat#11‐0111‐82, RRID: AB_464931
Anti‐mouse CD44‐e450 (IM7)	Thermo Fisher Scientific	Cat# 48‐0441‐82, RRID: AB_1272246
Anti‐mouse CD49d‐PE (R1‐2)	Thermo Fisher Scientific	Cat# 12‐0492‐82, RRID: AB_465697
Anti‐mouse CD62L‐PE‐Cy7 (MEL‐14)	Thermo Fisher Scientific	Cat# 25‐0621‐82, RRID: AB_469633
Anti‐mouse IFN‐γ‐APC (XMG1.2)	Thermo Fisher Scientific	Cat# 17‐7311‐82, RRID: AB_469504
Anti‐mouse TNF‐α‐FITC (MP6‐XT22)	Thermo Fisher Scientific	Cat# 11‐7321‐82, RRID: AB_465418
Anti‐mouse IL‐2‐PE‐Cy7 (JES6‐5H4)	Thermo Fisher Scientific	Cat# 25‐7021‐82, RRID: AB_1235004
APC‐labelled Pro5 MHC pentamer H‐2‐Kb‐SIINFEKL	ProImmune	Customised
Anti‐mouse CD45.2‐Alexa647 (104)	Biolegend	Cat# 109818, RRID:AB_492870
Mouse anti‐Plasmodium berghei CSP	Potocnjak et al, 1980; PMID: 6991628	RRID :A B_2650479
Rabbit polyclonal anti‐Plasmodium berghei UIS4	Muller et al, 2011: PMID: 21673790	N/A
Mouse anti‐Plasmodium berghei HSP70	Tsuji et al, 1994: PMID: 8153120	RRID :A B_2650482
Alexa Fluor 488 goat anti‐mouse	Thermo Fisher Scientific	Cat # A‐11029 RRID:AB_2534088
Alexa Fluor 546 goat anti‐rabbit	Thermo Fisher Scientific	Cat # A‐11010 RRID:AB_2534077
Experimental Models: Cell Lines
Huh‐7	Human hepatoma cells	Nakabayashi, H., Taketa, K., Miyano, K., Yamane, T. & Sato, J. Growth of human hepatoma cells lines with differentiated functions in chemically defined medium. Cancer Res. 42, 3858–3863 (1982).
Experimental Models: Organisms/Strains
C57BL/6J mice	Charles River or LSHTM	RRID:IMSR_JAX:000664
CD‐1 mice	LSHTM	RRID:IMSR_CRL:22
NMRI mice	Charles River	RRID:IMSR_CRL:605
OT‐I mice	Charles River	RRID:IMSR_CRL:642
Anopheles stephensi SK	(Feldmann and Ponnudurai, 1989)

Methods and Protocols

Ethics and animal experimentation

Animal procedures were performed in accordance with the German “Tierschutzgesetz in der Fassung vom 18. Mai 2006 (BGBl. I S. 1207)” which implements the directive 2010/63/EU from the European Union. Animal experiments at London School of Hygiene and Tropical Medicine were conducted under licence from the United Kingdom Home Office under the Animals (Scientific Procedures) Act 1986. All animal experiments were carried out under institutional care protocols. NMRI, CD‐1, C57BL/6 and OT‐I laboratory mouse strains were bred in house at LSHTM or purchased from Charles River Laboratories (Margate, UK or Sulzfeld, Germany). Female mice were used for experiments at the age of 6–8 weeks. Mice were placed into experimental cages of the required sample size by animal staff and not investigators; thus, an element of random group allocation occurred. Where more animals were required for a particular cage, selected animals were chosen at random for relocation. No blinding was performed. Sample size was estimated based on our previous experience and mouse availability. This project is in compliance with the ARRIVE guidelines.

Generation of transgenic parasites

Transgenic P. berghei ANKA mutants CSP^SIINFEKL and UIS4^SIINFEKL were developed using double homologous recombination. In the CSP^SIINFEKL mutant, the CSP gene is altered so the epitope SYIPSAEKI (residues 252‐260) is replaced with the H‐2^b restricted Gallus gallus ovalbumin epitope SIINFEKL. In the UIS4^SIINFEKL mutant, the SIINFEKL epitope is appended to the C‐terminal end of the UIS4 protein. As such, B3D‐CSP^SIINFEKL plasmid was assembled by successive cloning of three fragments, CSP‐C, CSP‐B and CSP‐A, obtained by PCR amplification from P. berghei ANKA genomic DNA followed by restriction enzyme digestion. These fragments correspond, respectively, to a 3′ homology region downstream of CSP (CSP‐C, 0.7 kb), a fragment comprising the CSP ORF downstream of the SYIPSAEKI epitope followed by the CSP 3′ UTR (CSP‐B, 0.8 kb) and a fragment comprising a 5′ promoter region followed by the CSP‐modified ORF where the SYIPSAEK coding sequence has been replaced by a SIINFEKL coding sequence (CSP‐A, 1.8 kb). The resulting B3D‐CSP^SIINFEKL plasmid, containing the Toxoplasma gondii dihydrofolate reductase/thymidylate synthase (TgDHFR/TS) pyrimethamine resistance cassette flanked by CSP‐A and CSP‐B on one side, and CSP‐C on the other, was linearised with NotI and SacII before transfection. Integration of the construct after double crossover homologous recombination results in replacement of the WT CSP gene by a modified copy containing the SIINFEKL coding sequence instead of the SYIPSAEKI coding sequence. The B3D‐UIS4^SIINFEKL plasmid was assembled by successive cloning of three fragments, UIS4‐A, UIS4‐B and UIS4‐C, obtained by PCR amplification from P. berghei ANKA genomic DNA followed by restriction enzyme digestion. These fragments correspond, respectively, to a fragment comprising a 5′ upstream sequence followed by the UIS4 entire ORF fused in frame to the SIINFEKL coding sequence (UIS4‐A, 1.2 kb), to the UIS4 3′ UTR sequence (UIS4‐B, 0.6 kb) and to a 3′ homology region downstream of UIS4 (UIS4‐C, 0.9 kb). The resulting B3D‐UIS4^SIINFEKL plasmid, containing the TgDHFR/TS pyrimethamine resistance cassette flanked by UIS4‐A and UIS4‐B on one side, and UIS4‐C on the other, was linearised with SacII and KpnI before transfection. Integration of the construct after double crossover homologous recombination results in replacement of the WT UIS4 gene by a modified copy containing the SIINFEKL coding sequence just upstream of a STOP codon. P. berghei CSP^SIINFEKL and UIS4^SIINFEKL parasites were generated by transfection of P. berghei ANKA with linearised B3D‐CSP^SIINFEKL and B3D‐UIS4^SIINFEKL plasmids, respectively. Purified schizonts of WT P. berghei ANKA (clone c15cy1) were transfected with 5–10 μg of linearised plasmid by electroporation using the AMAXA Nucleofector™ device (program U33), as described (Janse et al, 2006), and immediately injected intravenously in the tail vein of a mouse. The day after transfection, pyrimethamine (70 mg/l) was administrated in the mouse drinking water, for selection of transgenic parasites. Transgenic clones were isolated after limiting dilution and injection into mice. Correct integration of the constructs and purity of the transgenic lines was verified by analytical PCR using primer combinations specific for the unmodified CSP or UIS4 locus, and for the 5′ and 3′ recombination events. All primers used in this study are indicated in Table 2.

Table 2.

Primers used to generate plasmids and genotype parasites.

	Oligonucleotide	Sequence 5′ 3′
Production of B3D‐CSPSIINFEKL construct	CSP‐A forward	ATAAGAATGCGGCCGCATGGTTATATTTTGTGCAATGCTAAAATGG
	CSP‐A reverse	CGGAATTCTAGTATCAGTTTTTCAAAGTTGATTATACTATCGTCATTATTATTATTTTTGTTATTG
	CSP‐B forward	GGACTAGTGAATTCGTTAAACAGATCAGGGATAGTATCACAGAGG
	CSP‐B reverse	CCGCAATTGTACAAAAAATATTTTCGACAAAGGATAACG
	CSP‐C forward	CCCAAGCTTTGGGAATCTATTTTACAATATTATTTAAGGG
	CSP‐C reverse	CGGGGTACCCCGCGGTTATTGAAAAAGACACAAAATAGCTAG
Production of B3D‐UIS4SIINFEKL construct	UIS4‐A forward	TCCCCGCGGATAGCTATATTTTATGGTTGATCCTTTCC
	UIS4‐A reverse	GGACTAGTTTACAGTTTTTCAAAGTTGATTATACTTATGTATGGGCCGAATGATTTATTTTCC
	UIS4‐B forward	GGACTAGTTTCATTATGAGTAGTGTAATTCAGAAAGAG
	UIS4‐B reverse	CCGGAATTCTATGTAAAAAAGTTTGCATATACGGCTG
	UIS4‐C forward	CCCAAGCTTAGTGAAATATAAATATGAATGGAAGCAGCC
	UIS4‐C reverse	CGGGGTACCAGCAGCTAATGTCAATATATTTTATGCAC
Genotying of transgenic parasites	TgDHFR forward	CGCATTATATGAGTTCATTTTACACAATCC
	OVA reverse	CTAGTTTACAGTTTTTCAAAGTTGATTATAC
	CSP WT forward	TGTGAACTTTTCCTTATTTATTACGATTATG
	CSP test forward	AATATGAGCACGCTTTTACTTTGTCCAGG
	CSP test reverse	ACGAATCGAAATAAGTTACTATTCGTGCC
	UIS4 test forward	TGGTTCTTAATATTATTTTGGATACATGC
	UIS4 test reverse	CTCGTGTCCTTTGTAGTAAAAATAAACC

Restriction sites in the primer sequences are underlined.

Plasmodium berghei ANKA immunisation

Plasmodium berghei WT (strain ANKA clone c15cy1 or clone 507) parasites and CSP^SIINFEKL and UIS4^SIINFEKL (clone c15cy1) parasites were maintained by continuous cycling between murine hosts (NMRI or CD‐1) and Anopheles stephensi mosquitoes. Infected mosquitoes were kept in incubators (Panasonic and Mytron) at 80% humidity and 20°C. Sporozoites were isolated from salivary glands and γ‐irradiated at 1.2 × 10⁴ cGy. Mice were immunised intravenously in the lateral tail vein or intradermally in the ear pinnae with 10,000 sporozoites, unless otherwise stated, and challenged with either 1,000 or 10,000 sporozoites injected intravenously.

Indirect fluorescent antibody staining (IFA) of sporozoites

Epoxy‐covered 8‐well glass slides were coated with 3% BSA‐RPMI. 10,000 sporozoites were added per well in 3% BSA‐RPMI and incubated for 45 min during which the shed surface proteins are deposited in the gliding motility process. Sporozoites and their trails were stained with a mouse anti‐CSP (Potocnjak et al, 1980) (1:300) primary antibody and a rabbit polyclonal anti‐PbUIS4 (Muller et al, 2011) (1:500) primary antibody and the respective fluorescently labelled secondary antibodies (1:2,000). Nuclei were stained with Hoechst 33342 (1:2,000) and slides mounted with “Fluoromount‐G” (Southern Biotech). Sporozoites and trails were analysed by fluorescent microscopy (Zeiss Axio Observer).

In vitro infection of hepatoma cells and fluorescent staining

In vitro EEF development was analysed in infected Huh7 hepatoma cells for 24 and 48 h. Triplicate Labtek (Permanox plastic—Nunc) wells were infected with 10,000 transgenic CSP^SIINFEKL or UIS4^SIINFEKL parasites and duplicate wells were infected with 10,000 WT parasites. Infected cells were analysed by fluorescence microscopy using a mouse anti‐PbHSP70 (Tsuji et al, 1994) (1:300) and a rabbit polyclonal anti‐PbUIS4 (Muller et al, 2011) (1:500) primary antibody, the respective fluorescently labelled secondary antibodies (1:2,000) and nuclear staining with Hoechst 33342 (1:2,000). Stainings were analysed by fluorescent microscopy (Zeiss Axio Observer).

Quantification of SIINFEKL‐specific CD8 ⁺ T‐cell responses

Spleens and livers were harvested from immunised or naïve mice and perfused with PBS. Lymphocytes were derived from spleens by passing through 40 or 70 μm cell strainers (Corning) and from 1× PBS‐perfused livers by passing through 70 μm cell strainers (Corning). Hepatocytes were removed from homogenised livers using a modified 35% Percoll gradient (Goossens et al, 1990) (prepared using ratio of 1.6 10× PBS (Gibco): 12.4 Percoll (GE Healthcare) : 24.1 incomplete RPMI (Gibco)). Red blood cells were lysed with PharmLyse (BD), and lymphocytes were resuspended in complete RPMI (cRPMI‐ RPMI + 10% FCS + 2% Penicillin‐Streptomycin + 1% l‐glutamine (Gibco)). For cell counting, lymphocytes were diluted 40× with Trypan Blue (Thermo Fisher Scientific) and enumerated using a Neubauer “Improved” haemocytometer (Biochrom). Alternatively, lymphocytes were counted using a MACSQuant flow cytometer (Miltenyi Biotec), using propidium iodide (PI) (1:1,250) (Sigma‐Aldrich) or, in the case of intrahepatic lymphocytes, using CD45.2‐Alexa647 (1:400) (Biolegend) to distinguish between hepatocytes and lymphocytes, prior to PI administration and counting. Peripheral blood was acquired by tail vein puncture collected in Na⁺ heparin capillary tubes (Brand) and assayed in 96‐well flat bottom plates (Corning). For CD8⁺ T‐cell stimulations, 2–3 × 10⁶ splenocytes or 1–2 × 10⁵ liver cells were incubated with SIINFEKL peptide (Peptides and Elephants, Henningsdorf) at a final concentration of 10 μg/ml in the presence of Brefeldin A (eBioScience). Cells were incubated at 37°C, 5% CO₂ for 5–6 h, before incubation at 4°C overnight. For staining of cell surface markers and intracellular cytokines, cells were incubated for 1 h at 4°C. Cells derived from the spleen or liver were fixed with 4% paraformaldehyde, and cells from peripheral blood were fixed with 1% paraformaldehyde between the extra‐ and intracellular staining steps. Data were acquired by flow cytometry using an LSRII or LSRFortessa (BD). Antibodies used for staining were as follows; BD: CD3 (500A2) (1:300); eBioScience: CD8 (53‐6.7) (1:225), CD11a (M17/4) (1:300), CD49d (R1‐2) (1:225), CD62L (MEL‐14) (1:250), CD44 (IM7) (1:250), IFN‐γ (XMG1.2) (1:225), TNF (MP6‐XT22) (1:250) and IL‐2 (JES6‐5H4) (1:250); ProImmune: H‐2‐K^b‐SIINFEKL pentamers (3 µl/test).

CFSE labelling of OT‐I cells

Spleens from OT‐I mice were lysed and cells washed twice in PBS without serum. Splenocytes resuspended at a density of 5 × 10⁶ cells/ml in PBS had 1:5,000 CFSE (Thermo Fisher Scientific) added and were incubated in the dark at room temperature, with gentle inversion for 4 min. The labelling reaction was quenched with cRPMI and cells washed twice in cRPMI. Cells were recounted and 2 × 10⁶ cells were injected per mouse.

Vaccination with OVA expressing recombinant adenovirus

To assess parasite liver load after vaccination with virus‐expressed OVA, groups of C57BL/6 mice were immunised with recombinant human adenovirus serotype 5 (AdHu5) expressing full‐length chicken ovalbumin (de Cassan et al, 2011). Each mouse received 1 × 10⁸ infective units (ifu) in a volume of 100 µl administered intramuscularly (50 µl into each thigh). One day prior, mice received OT‐I splenocytes intravenously (2 × 10⁶ cells/mouse). 19 days after vaccination, vaccinated and control mice were challenged with 10,000 WT, CSP^SIINFEKL or UIS4^SIINFEKL sporozoites administered intravenously. 42 h after the challenge the livers were harvested and homogenised in TRIzol (Thermo Fisher Scientific) for total RNA isolation. Afterwards, cDNA was generated using the RETROScript Kit (Ambion). Quantitative real‐time PCR was performed using the StepOnePlus Real‐Time PCR System and Power SYBR Green PCR Master Mix (Applied Biosystems). Relative liver parasite levels were quantified using the ΔΔCt method comparing levels of P. berghei 18S rRNA normalised to mouse GAPDH mRNA (Muller et al, 2011). To assess sterile protection, AdHu5 OVA‐vaccinated and control mice received 2 × 10⁶ OT‐I splenocytes 1 day prior to vaccination. 14 days later, all mice were challenged with 800 WT, CSP^SIINFEKL or UIS4^SIINFEKL sporozoites. Blood smears were taken from day 3 to 14 after challenge to determine the presence of blood stage parasites.

Statistics

Data were analysed using FlowJo version 9.5.3 (Tree Star Inc., Oregon, USA), Microsoft Excel and GraphPad Prism v9 (GraphPad Software Inc., CA, USA). We used Mann–Whitney U‐test for analysing data that were not normally distributed and Welch’s t‐test or one‐way ANOVA with Tukey’s multiple comparison test for normally distributed data. Normal distribution was assessed using Shapiro–Wilk test.

Author contributions

Experiment design: KMa, OS and JCRH; Transgenic parasites CSP^SIINFEKL and UIS4^SIINFEKL : OS; Experiments and data analysis: KMü, MPG, MR, OS and JCRH; Adenovirus AdOVA: AR‐S, AVSH and SJD; Manuscript writing: MPG and JCRH; Manuscript comments and revision: All authors.

Conflict of interest

A.R.‐S., A.V.S.H. and S.J.D. are named inventors on patent applications relating to malaria vaccines, adenovirus vaccines and immunisation regimens.

Supporting information

Appendix

Expanded View Figures PDF

Review Process File

EMBO Mol Med (2021) 13: e13390.

Data availability

This study includes no data deposited in external repositories.