Long lived liver-resident memory T cells of biased specificities for abundant sporozoite antigens drive malaria protection by radiation-attenuated sporozoite vaccination

Abstract

Vaccination with radiation-attenuated sporozoites (RAS) can provide highly effective protection against malaria in both humans and mice. To extend understanding of malaria immunity and inform the development of future vaccines, we studied the protective response elicited by this vaccine in C57BL/6 mice. We reveal that successive doses of Plasmodium berghei RAS favour the generation of liver CD8+ tissue-resident memory T cells (TRM cells) over circulating memory cells and markedly enhance their longevity. Importantly, RAS immunisation strongly skews the composition of the liver CD8+ TRM compartment towards cells specific for abundant sporozoite antigens, such as thrombospondin-related anonymous protein (TRAP) and circumsporozoite protein (CSP), which become major mediators of protection. The increased prevalence of sporozoite specificities is associated with limited intrahepatic attenuated parasite development and inhibition of naïve T cell responses to all parasite antigens, whether formerly encountered or not, in previously vaccinated mice. This leads to the exclusive expansion of effector T cells formed upon initial immunisation, ultimately reducing the diversity of the liver TRM pool later established. However, stronger responses to less abundant epitopes can be achieved with higher initial doses of RAS. These findings provide novel insights into the mechanisms governing malaria immunity induced by attenuated sporozoite vaccination and highlight the susceptibility of this vaccine to limitations imposed by strain-specific immunity associated with the abundant, yet highly variable sporozoite antigens CSP and TRAP.

Fig 1. Repeated RAS vaccination of C57BL/6 mice provides efficient protection against P. berghei sporozoite infection that is dependent on CD8

⁺ T_RM cells. A-B. Mice were vaccinated with 1, 2 or 3 doses of 10,000 RAS (1xRAS, 2xRAS and 3xRAS respectively), one week apart and were challenged with 200 P. berghei sporozoites 30 days after the last dose of RAS. A. Rates of sterile protection. Numbers above columns denote numbers of protected mice/ total numbers of mice per group. B. Parasitemia at day 7 post-challenge. C-D. Mice were vaccinated using three doses of 5,142-10,000 RAS, one week apart, and treated with αCXCR3 mAb 3 and 1 days prior to challenge with 200 live sporozoites, which was performed on day 30-34 after the last vaccination. Parasitemia was monitored to evaluate protection. C. Rates of sterilising protection. D. Parasitemia on day 7 after sporozoite infection. Data were pooled from 2 independent experiments. Comparisons of sterile protection rates were done using Fisher’s exact tests. Parasitemia data were log-transformed and compared using one-way ANOVA and Tukey’s multiple comparisons test.

Fig 2. Abundance of liver CD8

⁺ T_RM cells specific for known Plasmodium antigens in mice vaccinated multiple times with WT RAS. A-D. CD8⁺ T_RM cell responses specific for TRAP, RPL6, RPA1, or other specificities (i.e., tetramer-negative T_RM cells) in the livers of mice vaccinated with 1, 2 or 3 rounds of 10,000 RAS, one week apart. Cell numbers were assessed 30 days after the last round of immunisation. Data were log-transformed and compared using two-way ANOVA and Tukey’s multiple comparisons test. A. Liver T_RM cell numbers. Data were log transformed and statistically compared using two-way ANOVA and Tukey’s multiple comparisons test. B. Frequencies of liver T_RM cells of different specificities amongst all T_RM cells in the liver. C. Fold change in the number of T_RM cells of the indicated specificities compared to those in 1xRAS vaccinated mice. Data were log transformed and statistically compared using one-way ANOVA and Tukey’s multiple comparisons test. D. Ratios of T_RM cell numbers in the liver vs numbers of circulating memory T cells (T_CIRCM, calculated by adding T_CM and T_EM cells) in the spleen. Data were compared using one-way ANOVA and Tukey’s multiple comparisons test. Data in Fig 2A-D were pooled from two independent experiments. E-H. CD8⁺ T_RM cells specific for mutated CSP (SIINFEKL), TRAP, RPL6, RPA1, or other specificities in the livers of mice vaccinated with 1 or 3 rounds of 5,050-7,700 CS5M RAS, 4-8 days apart. Cell numbers were assessed 30-63 days after the last round of immunisation. E. Liver T_RM cell numbers. Data were log transformed and statistically compared using two-way ANOVA and Tukey’s multiple comparisons test. F. Percentages of liver T_RM cells of different specificities amongst all T_RM cells in the liver. G. Fold change in the number of T_RM cells of the indicated specificities compared to those in 1xCS5M RAS vaccinated mice. Data were log transformed and statistically compared using one-way ANOVA and Tukey’s multiple comparisons test. H. Ratios of T_RM cell numbers in the liver vs numbers of circulating memory T cells (T_CIRCM, calculated by adding T_CM and T_EM cells) in the spleen. Data were log-transformed and compared using unpaired Student’s t-tests. Data in Fig 2E-H were pooled from two independent experiments. I-K. Liver T_RM cells generated by immunisation with a single dose of 10,000 or 50,000 RAS. I. CD8⁺ liver T_RM cell numbers specific for TRAP, RPL6, RPA1, or other specificities (i.e., tetramer-negative T_RM cells) in the livers of mice vaccinated with one dose of 10,000 on day 30 after vaccination (in Fig 2A) were compared with those in mice vaccinated with 50,000 RAS 25 days earlier. Data were log-transformed and compared using two-way ANOVA and Tukey’s multiple comparisons test. J. Frequencies of liver T_RM cells of different specificities amongst all T_RM cells in the liver. Data in 2B, 2F and 2J are represented as median with interquartile range. K. Fold change in the number of T_RM cells of the indicated specificities compared to those in naive mice. Data were log transformed and statistically compared using unpaired Student’s t-tests. Data in 2I-K were pooled from 2 (10k RAS) or 3 (50k RAS) independent experiments.

Fig 3. Repeated RAS vaccination generates long lived memory T cells.

Numbers of CD8⁺ T_RM cells in the liver (A-C) or T_EM cells in the spleen (D-F) specific for TRAP (A, D), RPL6 (B, E) and RPA1 (C, F) in mice vaccinated with one dose of 50,000 RAS (dotted line) or three doses of 10,000 RAS (solid line), one week apart. Cell numbers were assessed up to day 101 after the last round of immunisation. Data were pooled from 5 independent experiments for 1xRAS and 8 independent experiments for 3xRAS. Data were log-2 transformed and linear regression analyses of the log-transformed data were performed. Slopes were compared using F-tests.

Fig 4. TRAP-specific liver T_RM cells substantially contribute to protection induced by 3xWT RAS in C57BL/6 mice.

A. Experimental design. Red arrows denote intravenous injection of PbTRAP_130-138 or OVA_257-264 peptide dissolved in PBS. Numbers denote days after the first RAS vaccination. B. Sterile protection to live sporozoite challenge in 3xRAS vaccinated mice in which TRAP-specific cells (3xRAS-TRAPtol), or T cells specific for an irrelevant antigen (3xRAS-OVAtol) were removed (tol = tolerated). Data were compared using Fisher’s exact test. C. Parasitemia on day 7. Data were log-transformed and compared using one-way ANOVA and Tukey’s multiple comparisons test. D. Number of TRAP-specific liver CD8⁺ T_RM cells in protected vs non-protected mice. Data were log-transformed and compared using two-way ANOVA and uncorrected Fisher’s LSD test. Data in this figure were pooled from two independent experiments.

Fig 5. Mechanisms contributing to T cell specificity bias towards sporozoite antigens.

A-C. Protective capacity of effector responses elicited by RAS against incoming sporozoites. A. Representative experimental design. B-C. Mice received one or two doses of 5-10x10³ RAS 4-8 days apart, and 8-9 days later were injected with 1x10⁴ live sporozoites. Emergence of parasitemia was measured to evaluate protection. Data were pooled from two independent experiments. B. Parasitemia on day 5 after liver sporozoite challenge. Data were log-transformed and compared using one-way ANOVA and Tukey’s multiple comparisons test. C. Sterile protection. Data were compared using Fisher’s exact test. Numbers above columns denote numbers of protected mice/ total numbers of mice per group. D-K. Inhibition of naïve CD8⁺ T cell responses by prior RAS vaccination. D. Representative experimental design. Red arrows denote the time point in which mice were intravenously injected with 5x10⁴ naïve transgenic T cells (PbT-I cells in E-G; OT-I cells in H-K). E-G. Inhibition of naïve RPL6-specific (PbT-I) T cell responses by prior RAS vaccination. Distribution of memory PbT-I (E) and endogenous, RPL6- (F) and TRAP-specific (G) T cells in the liver 30-35 days after the last RAS injection. Mice received 1-3 doses of 5-10x10³ RAS. Data were pooled from two independent experiments. Statistical analyses, denoted by dark green and purple asterisks, denote comparisons of T_RM numbers. H-I. Inhibition of naïve Hsp70-OVA-specific (OT-I) T cell responses by prior HsOVA RAS vaccination. Distribution of memory OT-I (H) and endogenous, TRAP-specific (I) T cells in the liver 30 days after the last HsOVA RAS injection. Mice received 1-3 doses of 5.9-10x10³ HsOVA RAS. Data were pooled from two independent experiments. Statistical analyses, denoted by brown and purple asterisks, denote comparisons of T_RM numbers. J-K. Inhibition of naïve CS5M-CSP-specific (OT-I) T cell responses by prior CS5M RAS vaccination. Distribution of memory OT-I (J) and endogenous, TRAP-specific (K) T cells in the liver 30 days after the last CS5M RAS injection. Mice received 1 or 3 doses of 5.9-10x10³ CS5M RAS. Data were pooled from two independent experiments. Statistical analyses, denoted by brown and purple asterisks, denote comparisons of T_RM numbers. L-M. Naïve CD8⁺ T cell responses to RPL6 are inhibited in the absence of TRAP-specific T cells. Mice were vaccinated with 1-3 doses of 5-10x10³ WT RAS and received intravenous injections of either PbTRAP_130-138 (TRAPtol) or OVA_257-264 peptide (OVAtol) dissolved in PBS as explained in Fig 4A. PbT-Is were transferred 1 day before the last RAS vaccination and, on day 36 after the last RAS vaccination, mice were euthanised and numbers of memory PbT-I or TRAP-specific memory CD8⁺ T cells were examined. L. Number of memory PbT-I cells in the liver. M. Numbers of TRAP-specific memory T cells in the liver. Data were pooled from two independent experiments, log-transformed and analysed using one-way ANOVA and Tukey’s multiple comparisons test. Statistical analyses denote comparisons of T_RM numbers.

Fig 6. Memory T cell generation of prime-and-trap targeting OVA, and protection against challenge with parasites expressing SIINFEKL under the Hsp70 promoter or within CSP.

A-H. Mice received 50,000 naïve OT-I/uGFP cells (A-C, G, H) or not (D-F, G, H) and were vaccinated with either 2 μg Clec9A-OVA plus 5 nmol CpG B-P, 10⁹ rAAV-OVA, or all components combined. OT-I (A-C) or SIINFEKL-specific memory cells (D-F) were enumerated in the liver and the spleen 35 days later. Data were pooled from two independent experiments, log transformed and compared using one-way ANOVA and Tukey’s multiple comparisons test. Mice vaccinated with full P&T-OVA were challenged with 200 HsOVA (G) or CS5M (H) sporozoites on day 35 after vaccination, and rates of sterile protection were determined. Data were pooled from two independent experiments and compared using Fisher’s exact test. Asterisks over columns denote comparisons with the unvaccinated control group. Numbers over columns denote numbers of protected mice vs total numbers of mice. I-J. Generation of OVA-specific memory T cells in suboptimally vaccinated mice. Mice received 50,000 naïve OT-I/uGFP cells and were vaccinated with a low dose of 0.5 μg Clec9A-OVA plus 5 nmol CpG B-P. Numbers of OT-I T_RM cells were measured in the liver 30 days later (I). Separate cohorts of mice were challenged with 200 HsOVA or CS5M sporozoites (J), and rates of sterile protection were determined. Data were pooled from two independent experiments and compared using Fisher’s exact test.