Generation, genotyping, and phenotyping of the PbΔmei2Δlisp2 parasite line

A previous study identified a role for meiosis inhibited 2 (MEI2), a member of a family of RNA-binding proteins containing an RNA recognition motif23, during late liver stage development of Pb13. However, occasional breakthrough blood stage infections upon inoculation with high number (200 K) of PbΔmei2 spz were observed13. In addition to mei2, a single gene deletion of the liver-specific protein 2 (lisp2), a gene expressed on the mid-to-late liver stage parasitophorous vacuole membrane, was reported to lead to late liver stage developmental arrest in Pb24. We employed the available PbΔmei2 (2834cl2m1cl1) parasite line13 to create the double gene-deletion mutant PbΔmei2Δlisp2, as was previously reported for Py24 (Supplementary Fig. 1a). Three independent clones of mCherry-Luccon-expressing PbΔmei2Δlisp2 were generated and clone 3 (2900cl3) was selected for further analysis (Supplementary Fig. 1b). Our results show that the WT and PbΔmei2Δlisp2 parasite lines present comparable numbers of spz (Supplementary Fig. 1c), indicating that the deletion of lisp2 did not significantly impact the mosquito infectivity of the resulting transgenic parasites. Furthermore, mice injected with up to 3 ×105 spz of PbΔmei2Δlisp2 did not develop blood stage infections indicating complete liver stage growth arrest of PbΔmei2Δlisp2 (Supplementary Fig. 1d). Live imaging of mCherry-expressing parasites further revealed no differences between the size of WT and PbΔmei2Δlisp2 exoerythrocytic forms (EEFs) in cultured hepatocytes at 24 h post infection (hpi). However, at 48 and 72 hpi PbΔmei2Δlisp2 liver stages were significantly larger than WT parasites, which is most evident at the 72 h time-point (1399 µm2 for PbΔmei2Δlisp2 compared to 603 µm2 for WT parasites; Supplementary Fig. 1e, f). Immunofluorescence microscopy analysis additionally revealed that PbΔmei2Δlisp2 liver stage schizonts expressed merozoite surface protein 1 (MSP1) and apical membrane antigen 1 (AMA1), two proteins that are expressed by early and late Pb blood stage schizonts, respectively25,26, at 72 hpi (Supplementary Fig. 1g). These observations confirm that PbΔmei2Δlisp2 develops to late-stage hepatic parasites, including the expression of antigens associated with a late phase of schizont development.

Different WSpz formulations present distinct liver stage development profiles

Evidence has been presented that the extent of parasite development in the liver is an important determinant of the protective efficacy of different WSpz formulations15,16,17. In this study we compared the protective efficacy afforded by immunization of mice with four different Pb-based WSpz formulations: EA-GAP (PbΔb9Δslarp19), RAS20, LA-GAP (PbΔmei2Δlisp2) and CPS21. In order to assess the development profiles of the Pb parasites employed as surrogates of WSpz vaccines we first determined the parasite load of livers of C57BL/6J mice injected intravenously (I.V.) with 30 K spz of each formulation. As a control, mice were infected with 30 K luciferase-expressing Pb spz (hereafter referred to as Pb-Luci) without chloroquine treatment. Mouse livers were collected at specific time-points following spz inoculation, spanning the entire period of Pb hepatic development (12–60 h post inoculation - hpi), as well as the initial stages of erythrocytic infection (68–96 hpi)4, and the parasite liver load was determined by quantitative reverse transcription PCR (RT-qPCR) (Fig. 1a).

Fig. 1: Parasite liver load of C57BL/6J mice inoculated with different Pb-based surrogates of WSpz formulations.

a Study protocol. C57BL/6 J mice were inoculated intravenously (I.V.) with 30 K spz of Pb-Luci, EA-GAP, RAS, LA-GAP, or Pb-Luci administered under a chemoprotective chloroquine regimen, and euthanized at the indicated time-points post inoculation for liver collection and RT-qPCR analysis. b Parasite load in mouse livers collected at the indicated time-points post spz inoculation (n = 6–9 mice per time-point from 2-3 independent experiments). For each time-point, data are expressed as mean ± SD and were compared using the Kruskal–Wallis test with Dunn’s multiple comparison post-test (*P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001).

As expected, the parasite biomass in mice infected with Pb-Luci linearly increased up to 48 hpi, corresponding to the period of growth and multiplication of the parasites into mature liver stages, followed by a decrease in liver load between 60 and 72 hpi, presumably as a result of the completion of the parasite’s hepatic development and the release of liver merozoites. An increase in parasite biomass is observed between 84 and 96 hpi, likely due to the presence of Pb blood stages in the liver sinusoids (Fig. 1b). The parasite biomass in mice inoculated with EA-GAP and RAS WSpz is markedly lower than that of mice inoculated with Pb-Luci at all time-points, which is consistent with the early growth-arrest of those parasites in the liver19,27. However, small differences exist between these two WSpz formulations; while the parasite biomass in EA-GAP-inoculated mice is highest at 12 hpi, RAS-inoculated mice reach an equivalent load at 24 hpi (Fig. 1b), indicating that liver growth of EA-GAP arrests earlier than RAS. The parasite biomass in mice inoculated with LA-GAPs or CPS WSpz formulations displays a pattern similar to that of Pb-Luci up to 48 hpi, indicating wild-type-like growth of these parasites into late liver stages. However, parasite biomass at later time-points decreases more rapidly in LA-GAP WSpz- than in CPS WSpz-inoculated mice, presumably due to the former parasite’s developmental arrest and eventual elimination from the liver. Conversely, in CPS WSpz-inoculated mice, hepatic parasites develop into fully mature schizonts that release liver merozoites, which are eliminated by chloroquine (Fig. 1b).

Protective efficacy of the different WSpz formulations is dependent on the immunization dosage

WSpz immunization has resulted in variable levels of protective efficacy, both in rodent malaria models and in humans, depending on factors such as the route of administration, and immunization dosage and schedule28,29,30. Since previous studies have underscored the need for optimization of the WSpz dosage required to achieve robust and sterile protection30, we sought to assess the impact of the immunization dosage on the protective efficacy of the four different WSpz formulations. To this end, C57BL/6J mice were immunized following a P2B regimen with weekly intervals, an immunization regimen previously demonstrated to induce highly protective immune responses in mouse malaria models31,32, employing five distinct WSpz dosages (1, 10, 30, 90, and 270 K). On day 28, two weeks after the last immunization, mice were challenged with 30 K fully infectious Pb-Luci spz, and relative parasitaemia was monitored for at least 15 days thereafter by bioluminescence33. Control mice (hereafter referred to as Challenge controls) infected with 30 K Pb-Luci spz at the time of challenge (Fig. 2a) developed a blood stage infection with a 3-day pre-patent period (Supplementary Fig. 2a).

Fig. 2: Sterile protection of C57BL/6J mice immunized with different dosages of the various WSpz formulations following a P2B vaccination regimen.

a Study protocol. C57BL/6J mice were immunized following a P2B immunization regimen administered via I.V. injection of 1, 10, 30, 90 and 270 K of the EA-GAP, RAS, LA-GAP and CPS WSpz formulations. Non-immunized and immunized mice were infected/challenged on day 28 with 30 K spz of Pb-Luci. Blood was collected daily for a period of 15 days and analyzed by a bioluminescence assay. b Percentage of protected (in colored and dotted pattern bars) and non-protected (in gray bars) mice for each immunization dosage and vaccination approach (n = 5–20 mice per immunization dosage from 1–4 independent experiments). Sterile protection is defined as the absence of blood stage parasites for up to 15 days following the infectious challenge. Numbers above the bars indicate number of protected/total number of challenged mice. For each immunization dosage, the percentage of protected mice are expressed as mean ± SD.

The protective efficacy induced by the different dosages of the four WSpz formulations was assessed by induction of sterile protection, as defined by the absence of detectable blood stage infection until 15 days after challenge. The percentage of protected mice followed a similar, dosage-dependent increase up to a dosage of 30 K of all WSpz formulations (Fig. 2b). The dosage-dependent increase in protection is further observed for RAS, LA-GAP and CPS at WSpz dosages above 30 K. Between 90 and 100% of the immunized mice were sterile protected at immunization dosages of 90 and 270 K spz. Remarkably, however, not only was sterile protection not observed for all mice immunized with any given dosage of EA-GAP WSpz, but also immunization with 90 and 270 K WSpz dosages of this formulation led to a 1,6- and 4,3-fold reduction in sterile protection levels, respectively, relative to those observed for immunization with 30 K WSpz (Fig. 2b).

In addition to sterile protection, protective immunity was assessed by determining the pre-patent period in the mice that developed a blood infection after challenge with 30 K Pb-Luci spz. A longer pre-patent period has been correlated with a reduction in the inoculum of parasites emerging from the liver into the blood34. Overall, the pre-patency profile of the non-sterile protected mice followed a trend similar to that observed for sterile protection, with the pre-patent period of mice immunized with RAS, LA-GAP and CPS increasing with immunization dosage. In contrast, the maximum delay in patency for EA-GAP WSpz-immunized mice was observed at a 30 K dosage and decreased at the higher dosages of 90 and 270 K WSpz (Supplementary Fig. 2a). Pb ANKA-infected C57BL/6J mice constitute a well-established model of experimental cerebral malaria (ECM), enabling the assessment of protection against ECM induced by different immunization approaches. Overall, protection against the development of ECM follows a pattern similar to that observed for sterile protection afforded by the different dosages and WSpz formulations (Supplementary Fig. 2b and Fig. 2b). Nevertheless, 10 K or higher dosages of CPS WSpz appear to offer the highest protection against this severe syndrome. Additionally, while all non-sterile protected mice immunized with 270 K EA-GAP developed ECM35, only 20 ± 28.3% of the non-sterile protected mice immunized with 1 K EA-GAP succumbed to ECM (Supplementary Fig. 2b). Thus, high dosage immunization with EA-GAP affords low levels of not only sterile protection and delayed patency, but also protection against severe disease.

Overall, our results show that, whereas sterile protection was not observed for all mice immunized with any given dosage of EA-GAP WSpz, 100% of the animals immunized with 270 K RAS, 90 and 270 K LA-GAP, and 270 K CPS were sterile protected against challenge. Moreover, while protection conferred by RAS, LA-GAP and CPS correlates with WSpz dosages from 1 to 270 K, EA-GAP WSpz dosages higher than 30 K result in a decline in protective efficacy. Somewhat surprisingly, we also found that protection induced by immunization with RAS WSpz was comparable to that afforded by LA-GAP and CPS WSpz, indicating the absence of a strong correlation between protective immunity and the extent of parasite development and biomass in the liver.

Low protection following high-dosage EA-GAP immunization is not associated with evidence for liver damage or reduced magnitude of the immune responses

The low protection levels observed following immunization with EA-GAP WSpz dosages higher than 30 K prompted us to investigate possible causes behind this lower protective efficacy compared to the other WSpz formulations. A previous in vitro study showed that hepatocytes infected with another early-arresting parasite, PbΔp36p, presented higher levels of apoptosis than RAS-infected cells36. Similarly to the EA-GAP employed in the present study, PbΔp36p fails to maintain a parasitophorous vacuole and arrests soon after hepatocyte invasion36. Thus, we hypothesized that increased levels of apoptosis of infected hepatocytes following three immunizations with high EA-GAP dosages might compromise the ability of the booster parasites, to establish protective immune responses. To investigate the hepatocellular damage resulting from three EA-GAP WSpz immunizations, histopathological analyses were performed on livers from mice immunized with 270 K EA-GAP WSpz, collected 48 h after the challenge with Pb-Luci parasites. To limit the number of animals used in this study, and since RAS immunization is considered the gold-standard of WSpz immunization7, mice immunized with 270 K RAS WSpz were employed as controls in these experiments. In addition, to further control for the potential influence of the load of immunizing parasites, two additional groups of mice, immunized with 30 K EA-GAP or RAS WSpz were included in these analyses (Fig. 3a). Livers of all immunized animals presented infiltrates indicative of a chronic (longer than 2 days) inflammation (Supplementary Fig. 3a), which is absent from non-immunized, challenged mice, and thus likely results from the immunization rather than from the challenge. Nonetheless, no significant histopathological differences or distinct patterns of microscopically detectable hepatocellular damage were found between mice immunized with either dosage of the EA-GAP and RAS WSpz (Supplementary Fig. 3a). Additionally, we measured known biochemical parameters associated with liver function37 in serum collected from these immunized mice. In agreement with the histopathological analysis (Supplementary Fig. 3a), no significant differences in the serum levels of total bilirubin and glucose were found between the four groups of immunized mice. In addition, the levels of aspartate aminotransferase (AST) and alanine aminotransferase (ALT) are comparable between mice immunized with the same dosage of either WSpz formulation (Supplementary Fig. 3b). Collectively, our data indicate that the low protection conferred by high dosage EA-GAP immunization is not associated with evidence for WSpz-induced liver injury or disfunction.

Fig. 3: Immune population analyses of the spleens and livers of C57BL/6J mice immunized with either 30 or 270 K of EA-GAP or RAS WSpz following a P2B immunization regimen.

a Study protocol. C57BL/6J mice were immunized by I.V. injection of 30 or 270 K EA-GAP or RAS WSpz. Non-immunized and immunized mice were subsequently infected/challenged on day 28 with 30 K spz of Pb-Luci. Forty-eight hours post challenge (day 30), mice were euthanized for spleen, liver, and serum collections. b Number of CD4+ T, CD8+ T, NK, NKT and γδ T cells in the spleen (top) and liver (bottom) of naïve and challenge controls, as well as of immunized mice (n = 3–5 per group). Symbols represent the individual values of each mouse and floating bars indicate the mean ± SD. Statistically significant differences relative to naïve control mice (#), relative to challenge controls (* above the bars) or between experimental groups (* above the lines) were assessed through the Kruskal–Wallis test with Dunn’s multiple comparison post-test (*,#P < 0.05; **,##P < 0.01; ***,###P < 0.001; ****,####P < 0.0001).

Immunization with WSpz formulations induces strong immune responses against the pre-erythrocytic parasite stages7. RT-qPCR analysis of immunized mouse livers collected 2 days after challenge (Fig. 3a) revealed that the Pb-Luci parasite biomass was higher in EA-GAP WSpz- than in RAS WSpz-immunized mice, particularly for animals immunized with 270 K WSpz (Supplementary Fig. 3c), in agreement with the sterile protection results shown in Fig. 2b. A standard flow cytometry hierarchical gating strategy (Supplementary Fig. 3d) was employed to assess potential immune factors involved in the low protective efficacy conferred by immunization with high dosage EA-GAP WSpz. To this end, lymphocyte populations isolated from the spleens, as a proxy for systemic immune responses, and livers of mice immunized with 30 and 270 K EA-GAP WSpz 48 h after challenge were compared with those from spleens and livers of mice immunized with 30 and 270 K RAS WSpz (Fig. 3a). This analysis was performed after challenge because that enables elucidating the overall immune response in a vaccination/infection setting. In addition, we ensured the rigor of this analysis by the use of appropriate control groups, including naïve and non-immunized mice subjected to the same challenge, providing information on the contribution of the mobilization of immune cells from the challenge alone.

Overall, no significant differences were found in the magnitude of the splenic immune responses between the four groups of immunized mice (Fig. 3b). Nonetheless, it is worth mentioning that CD4+, CD8+ and γδ T cell populations generally increased in immunized mice relative to the naïve or to the challenge controls (Fig. 3b). We further analyzed the same set of immune populations in the livers of EA-GAP WSpz- and RAS WSpz-immunized mice (Fig. 3b). Our results show that, whereas immunization with 30 K EA-GAP WSpz did not lead to significant differences in the numbers of CD4+ T, CD8+ T, NKT, γδ T and NK cells in the liver compared to naïve and challenge controls, immunization with the same dosage of RAS WSpz induced significant increases in the cell numbers of all lymphocyte populations analyzed, except NKT cells (Fig. 3b). This suggests that the EA-GAP and RAS WSpz immunizations induce local immune responses of significantly different magnitudes, which may be related to differences in the extent of liver stage development of either WSpz formulation. On the other hand, mice immunized with 270 K spz of either EA-GAP or RAS WSpz exhibit numbers of CD4+, γδ and NK T cells equivalent to those found in the livers of mice immunized with 30 K RAS WSpz. This indicates that a 9-fold increase in immunization dosage with either EA-GAP or RAS WSpz relative to the 30 K RAS WSpz inoculum does not elicit a substantial difference in the magnitude of immune cell responses in the liver (Fig. 3b).

Overall, our analyses did not identify any significant differences in the magnitude of either systemic or liver immune responses that could be associated with the low protection levels observed upon immunization with high dosage EA-GAP WSpz.

High dosage EA-GAP immunization results in lower frequency of effector CD8+ T cells with an intermediate phenotype between TRM and SLECs

CD8+ T cells have been identified as key mediators of protective immunity elicited by different WSpz formulations in rodent models of malaria38. Notably, a relationship between antigen dose and functional avidity, memory phenotype and survival of CD8+ T cells has been demonstrated in different models, such as Leishmania major, HIV and Mycobacterium tuberculosis, with lower dosages providing increased protection, indicating that lower antigen doses lead to more favorable T cell responses (reviewed in ref. 39). Hence, we further investigated the composition of the liver CD8+ T cell population applying an unbiased clustering algorithm to our flow cytometry data40. To this end, the X-shift algorithm was applied to concatenated datasets consisting of CD8+ T cell populations from the livers of naïve and challenge control mice, as well as from livers of mice immunized with 30 and 270 K EA-GAP or RAS WSpz, collected 48 h post challenge and the clustering analysis was visualized using TriMap dimensionality reduction (Fig. 4a, b). The phenotypic classification of the CD8+ T cell subsets was primarily based on the expression of CD62L (L-selectin), a lymph node homing molecule, and CD44, a molecule involved in cell adhesion, migration, and signaling41. Three main populations were identified: i) naïve T cells (TN; CD62L+ CD44−, cluster 10); ii) central memory T cells (TCM; CD62L+ CD44+, cluster 4); and iii) effector/effector memory T cells (TEM; CD62L− CD44+; clusters 2, 3, 5, 6, 8; Fig. 4a). The latter subset has been reported to consist of a quite heterogeneous pool of CD8+ T cells, comprising distinct subpopulations of TEM cells42. Our analysis identified these subpopulations as including resident-memory cells (TRM; CD69+ KLRG1− CXCR3+, cluster 5), memory-precursor effector cells (MPECs; KLRG1- CD127+, cluster 6) and short-lived effector cells (SLEC; KLRG1+ CD127−, cluster 3), each with distinctive memory potential and functional properties43 (Fig. 4a). Interestingly, in addition to previously known cell populations, our X-shift analysis identified two populations of TEM co-expressing CD69 and KLRG1 (CD69+ KLRG1int, cluster 2 and CD69+ KLRG1+, cluster 8), which appear to present an intermediate phenotype between TRM and SLECs (clusters 8 and 2 phenotypically closer to SLECs and TRM, respectively; Fig. 4a, b).

Fig. 4: Clustering analysis of the CD8+ T cell pool in the liver of C57BL/6J mice immunized with either 30 or 270 K EA-GAPs or RAS WSpz following a P2B immunization regimen.

a Heatmap of clusters identified by X-shift analysis of concatenated data from CD8+ T cells of the livers of naïve and challenge control mice, as well as of immunized mice, 48 h post challenge (day 30). TN: naïve T cells; TCM: central memory T cells; TEM: effector/effector memory T cells; TRM: resident memory T cells; SLECs: short-lived effector cells; MPECs: memory-precursor effector cells; KLRG1: killer cell lectin-like receptor subfamily G, member 1; CXCR3: C-X-C motif chemokine receptor 3. b X-shift clustering analysis of concatenated data from the liver of mice under each condition. Data is presented using TriMap dimensionality reduction. Parts-of-whole bar graphs next to each plot represent the clusters within the CD8+ T cell population of each condition. Frequencies (c) or numbers (d) of each of the clusters identified in the X-shift analysis in the liver, for each condition. Symbols in c represent the individual values of each mouse and floating bars indicate the mean ± SD. Statistically significant differences relative to naïve control mice (#), relative to challenge controls (*above the bars) or between experimental groups (* above the lines) were assessed through the Kruskal–Wallis test with Dunn’s multiple comparison post-test (*,#P < 0.05; **,##P < 0.01; ***,###P < 0.001; ****,####P < 0.0001). Data in d are presented as mean ± SD. e Combined frequency of clusters 2 and 8, representing total CD8+ TEM CD69+ KLRG1+ T cells within liver leukocytes. Symbols represent the individual values of each mouse and floating bars indicate the mean ± SD. Statistically significant differences relative to naïve control mice (#), relative to challenge controls (*above the bars) or between experimental groups (* above the lines) were assessed through the Kruskal–Wallis test with Dunn’s multiple comparison post-test (*,#P < 0.05; **,##P < 0.01; ***,###P < 0.001; ****,####P < 0.0001).

As expected, a statistically significant increase in the frequency of CD8+ SLECs and TRM was observed in all four groups of immunized mice relative to both naïve and challenge controls, which is also observed in terms of cell numbers except for the mice immunized with 30 K EA-GAP (Fig. 4c, d, Supplementary Fig. 4). Of note, the numbers of CD8+ TRM are more comparable between mice immunized with the different dosages of RAS WSpz than between mice immunized with the same dosage of the two WSpz formulations (Fig. 4d, Supplementary Fig. 4). This indicates that RAS immunization induces higher levels of CD8+ TRM than EA-GAP immunization, irrespective of the immunization dosage. Likewise, mice immunized with either EA-GAP WSpz dosage present similar levels of hepatic CD8+ T clusters corresponding to TN and TCM, which were higher than those of RAS-immunized mice (Fig. 4c). Of particular interest, we observed an increase in the frequency of clusters 2 and 8 in all immunized mice, except for those immunized with 270 K EA-GAP WSpz, which showed significantly lower frequencies of those clusters than either group of RAS WSpz-immunized mice. A similar trend for lower frequencies of clusters 2 and 8 was observed in mice immunized with 270 K EA-GAP WSpz in comparison with those immunized with 30 K of this formulation, although it did not reach statistical significance (Fig. 4c). The results of a joint analysis of clusters 2 and 8 were similar to those obtained for the analyses of either individual cluster, including the statistically significant difference between the mice in the 270 K EA-GAP WSpz-immunized group and both groups of RAS WSpz-immunized mice (Fig. 4e).

In conclusion, our results indicate that, contrarily to RAS immunization, immunization with a high dosage of EA-GAP WSpz did not induce the expansion of an intermediate TEM population expressing markers of both effector and liver-resident memory CD8+ T cells, which may be related to the relatively low levels of protective efficacy observed with the latter immunization regimen.