Female C57BL/6 mice (4–8 weeks old) were brought from Zydus Research Center, Ahmedabad, Gujarat, India. Mice were housed at the central facility of Nirma University and handled as per the institutional animal handling guidelines. All animal experiments were reviewed and approved by the Institutional Animal Ethical Committee of Nirma University (CPCSEA Reg. No: 883/PO/ReBi/S/05/CPCSEA).

Plasmodium berghei (Pb) (ANKA strain) sporozoites were isolated from salivary glands by dissecting the infected Anopheles stephensi mosquitoes on day 18 after feeding with infected blood. The purified sporozoites were attenuated by exposing to 12,000 rads of gamma-radiation (Bhabhatron) in the facility of BV Patel PERD Center, Ahmedabad. All experiments were reviewed and approved by the Institutional Biosafety Committee (IBSC) of Institute of Science, Nirma University (IBSC Reg. No: UA3K1PKH).

Mice were anesthetized by ketamine (80 mg/kg body weight) and xylazine (6 mg/kg body weight). Upon dissecting the mouse, liver was perfused by flowing 10–15 ml PBS at the rate of 10 ml/min and inferior vena cava was cut for blood out flow. Gall bladder was removed and liver was pass through a 70-µM cell strainer (Himedia, India) with the help of 10 ml syringe plunger. Cell suspension was washed, resuspended in the 33% Percoll (GE healthcare) in PBS, and centrifuged for 20 min at 800 × g. Cell suspensions were prepared by following the established protocol and RBCs were lysed with ACK RBC lysis solution, washed twice with PBS. The cell counting was done on Neubauer Chamber.

Three groups of five mice each were inoculated i.v. with 50K RAS, 10K Inf. Spz, or sham dissected salivary gland material from uninfected mosquitoes with similar dilution. After 72 h postinoculation, mice were euthanized and livers were perfused using PBS and were immediately submerged in RNA later solution to protect degradation of RNA.

Transcriptome sequencing was carried out on an Illumina NextSeq platform that generated approximately 75-bp paired-end (PE) raw reads. The raw data were filtered using Trimmomatic v0.30 software (29). After removing adaptor sequences, ambiguous “N” nucleotides (with the ratio of “N” greater than 5%) and low-quality sequences (with quality score less than 10), the remaining high-quality reads were aligned and pull together using TopHat and Bowtie software (30) as described for aligning the high-quality reads with a Mus musculus reference genome. Raw sequencing data of all the samples is available through the NCBI Sequence Read Archive accession: SRP129582.¹ For homology annotation, non-redundant sequences were subjected to public databases, including NCBI non-redundant protein and non-redundant nucleotide, Swiss-Prot,² gene ontology (GO) by panther classification system,³ and the pathway analysis by Kyoto Encyclopedia of Genes and Genomes (KEGG).⁴

To analyze initial host response against RAS and Inf. Spz, differentially expressed genes in M. musculus, the number of reads for each of the sequenced reads from the two samples was converted to fragments per kilo base of transcript per million mapped reads (FPKM) (31). The criteria of each gene having a read depth of minimum 5 reads was also applied for finding differentially expressed genes. Apart from the FPKM value, p-value and a fold change value were also calculated for sham and Inf. Spz inoculated groups. Subsequently, pathway enrichment analysis was done by KEGG pathway analyser server,⁵ which is widely used in functional annotation and enrichment analysis. The ontology covers three domains namely Cellular Component, Molecular Function and Biological Process. The GO annotations in panther are either obtained from public databases (panther classification system or KEGG pathway analyser server) or produced by computational predictions.

Genes identified in the transcriptome sequencing analysis were validated by quantitative real-time PCR (RT-qPCR). The primers (Table S1 in Supplementary Material) were taken from https://pga.mgh.harvard.edu/primerbank/. The prepared total RNA used in quantitative real-time PCR analysis was isolated from the same sample as that used for RNA sequencing. The quantitative real-time PCR was performed on the ThermoScientific Quant studio 3 real-time PCR system using 2 × SYBR green real-time PCR mix (Clontech, Takara). The PCR amplification was performed in triplicate, using the following cycling parameters: 94°C for 2 min, followed by 40 cycles of 10 s at 94°C, 10 s at 59°C, and 34 s at 72°C. All samples were analyzed in triplicate and the expression of target genes was calculated as relative fold values using the 2−ΔΔCT method.

The average values ± SEM are presented in each graph and table. Statistical analysis was performed using GraphPad Prism software version 7.0 for Windows (GraphPad Software). p values were calculated by using non-parametric Mann–Whitney U-test. p Value ≤0.05 was considered statistically significant.

Induction of liver-stage immunity to Plasmodium infection in animals and humans requires multiple immunizations of RAS (4, 6, 32). Whole sporozoites vaccination (WSV) has been shown to render sterile and protracted protection against multiple challenges (4). In B6 mouse model, we found that Pb RAS-induced sterile protection is maintained for 6 months (11). Interestingly, the intermittent Inf. Spz challenge following last-boost immunization was observed to prolong the sterile protection up to 18 months (11). While the loss of protection within 6 months was associated with the sharp decline of CD8⁺ T cells in liver, the intermittent challenge had rescued the CD8⁺ T cells from attrition. These observations were based on phenotypic and functional characterization of the memory CD8⁺ T cells without knowing their antigen specificities. Here, we wanted to understand the qualitative changes in LS-specific CD8⁺ T cell response upon challenge that might be responsible for ensuing longer-lived protection. We immunized B6 mice with four doses of Pb RAS (75K-20K-10K-10K) with an interval of 2 weeks apart as described previously (19). On day 60, some of the immunized mice challenged with 10K Inf. Spz (now onward termed as Inf. Spz group), while others were challenged with 10K RAS (now onward termed as RAS group). Since RAS immunized mice were protected for 6 months, we sought to determine the reactivation potential of CD8⁺ T cells after 6 months in order to find out the differences between RAS and Inf. Spz groups. On day 232, both the groups were challenged with 10K Inf. Spz as shown in Figure 1A and LS-specific memory CD8⁺ T cells were characterized on day 3 post-challenge.

Generally, the antigen-specific T cell responses are measured in the form of IFN-γ expression after in vitro stimulation with specific peptide or protein antigens. Since majority of the Plasmodium LS-specific antigens are yet to be discovered, measuring the LS-specific CD8⁺ T cell response through in vitro stimulation by limited available antigens we may bias our observation and the same might not truly reflect the nature of sterile protection. Moreover, inducing the LS-specific sterilizing immunity requires the hepatocytes to be invaded by the sporozoites for the expression of the LS antigens to effectively activate the CD8⁺ T cells. To reactivate the memory CD8⁺ T cells to any possible LS antigens, we followed the in vivo T cell reactivation approach (33). For this, CD8⁺ T cells generated following RAS immunization were reactivated by sporozoites challenge and IFN-γ response was measured by the in vivo ICS (33) along with cytotoxic response (CD107a⁺).

Upon challenging the mice with Inf. Spz for reactivating the T cells, we found significant elevation of effector/effector memory CD8⁺ T cells (T_E/EM; CD44^hiCD62L^lo) in their liver (post-challenge; day 235) compared to the same in mice without challenge (pre-challenge; day 232), suggesting that T cells respond to antigen challenge. Interestingly, the elevation of T_E/EM cells was found higher in the Inf. Spz group (46.06–82.37%) compared to RAS group (44.38–66.59%) (Figure 1B). Previously, we have shown that LS-specific CD8⁺ T_CM cells are critical for protracted protective immunity (7). The reduction in CD8⁺ T_CM cells associated with the elevation of CD8⁺ T_E/EM cells is used as a measure of the memory T cells being reactivated to antigenic challenge, thus converting T_CM to T_E/EM. In this study, we found the reduction of central memory CD8⁺ T cells (T_CM; CD44^hiCD62L^hi) upon reactivation on day 232 post last immunization. While the RAS group showed about 5% reduction (i.e., 19.64–13.99%), the Inf. Spz group demonstrated twofold higher reduction in T_CM population 19.81–9.04% (Figure 1B). The above results depict that either the reactivation potential of memory T cells is higher or the frequencies of memory T cells maintained in the Inf. Spz group compared to RAS group as reflected in the conversion of T_CM to T_E/EM. To further understand the qualitative changes in the effector function of the responding T cells, we measured the IC IFN-γ production and cytotoxic activity (CD107a⁺) of the CD8⁺ T_E/EM cells (Figure S1 in Supplementary Material). The results suggest the significant activation of cytotoxic (CD107a⁺) and IFN-γ-producing CD8⁺ T cells in both the groups after challenge (Figure 1C). The percentage of IFN-γ⁺ cells is lower in general that we observed consistently in our experiments. However, we found threefold higher accumulation of cytotoxic T cells (p < 0.05) in the liver of Inf. Spz group compared to RAS (Figure 1Di). Similarly, the frequencies of IFN-γ producing CD8⁺ T cells was 2.5-fold (p < 0.0001) higher in the Inf. Spz group (Figure 1Dii). It is noteworthy that we found not just quantitative, but also the qualitative difference in terms of IFN-γ production per cell. The intermittent challenge helped generate high IFN-γ producing T cells compared as presented in the form of mean fluorescent intensity (MFI) (Figure 1Diii). The reactivation of high frequencies of cytotoxic (CD107a⁺) and IFN-γ-producing CD8⁺ T cells in Inf. Spz group correlates with the higher percentage CD8⁺ T_E/EM cells. Collectively, these results suggest that intermittent challenge with Inf. Spz following RAS vaccination generates LS-specific CD8⁺ T cells, which are qualitatively better. It is possible that the generation of qualitatively different CD8⁺ T cells could be the result of differential innate immune response between RAS and Inf. Spz groups.

The pattern of LS-specific CD8⁺ T cell activation in response to intermittent sporozoites challenge prompted us to look for the differential microenvironment created in the liver during the challenge (34–37) that might modulate the antigen-presenting cells (APCs). Hence, we did a comparative study involving RNA-seq liver transcriptome analysis for which mice were inoculated with RAS, Inf. Spz, or sham control. Results revealed statistically significant differences in the expression of 590 transcripts (p < 0.05), out of which 430 genes were upregulated in case of Inf. Spz as compared to sham control mice. Of those 430 genes, 363 transcripts were at least onefold higher (p < 0.05). Subsequently, we found that 105 genes out of 363 genes involved in immune response (Figure S2A and Table S2 in Supplementary Material). Further, 31 genes out of 105 immune response-related genes found for their involvement in the TLR signaling pathway (Figure 2A; Table S2 in Supplementary Material) that induce interferon response associated with activation of APCs (38). By quantitative RT-PCR, we validated 9 upregulated genes (Figure S2B in Supplementary Material) that are found to be involved in TLR signaling as analyzed by KEGG pathway database server (Figure 2B).

Earlier it was reported in mouse models of various viral infections as well as IC parasites that the DCs accumulate at the site of infection (16, 39–41). Since the activation of malaria LS-specific CD8⁺ T cells is also shown to be critically dependent on the CD8α⁺ DCs (16, 25, 26, 42), we wanted to determine whether the inoculation of Inf. Spz would promote higher accumulation of DCs in the liver compared to RAS inoculation. Mice were given inoculums of 10K RAS or Inf. Spz and at 72 h postinoculation the DCs characterized as CD11c⁺CD11b⁻ population were further analyzed for CD8α⁺ subset of DCs in the liver, spleen, and hepatic LNs (Figure 3) as shown in earlier findings (24, 43, 44). We found that 7.8 ± 0.21% (p < 0.05) of DCs in the liver of Inf. Spz inoculated mice expressed CD8α, whereas only 3.2 ± 0.24% DCs in the liver of RAS mice expressed CD8α (Figure 3C). The status of DCs in spleen or LNs was not significantly different between Inf. Spz or RAS inoculated mice. Nevertheless, the presence of relatively high frequencies of CD8α⁺ DCs in Inf. Spz inoculated mice suggests that exposure to Inf. Spz enhances the accumulation of CD8α⁺ DCs in the liver compared to RAS. We suspect that accumulation of CD8α⁺ DCs in the liver could be the outcome of enhanced migration of these cells which is critically dependent on the availability of specific chemokines required for the migration of DCs to the site of infection (45). Therefore, we next determined whether infectious status of sporozoites modulates the inflammatory milieu of liver in terms of the expression of chemokine ligands, CCL20, and CCL21. We performed real-time RT-PCR from the liver of sham, RAS, or Inf. Spz inoculated mice. We observed about twofold increased expression of CCL20 (p < 0.05) and CCL21 (p < 0.05) in response to Inf. Spz inoculation as compared to that of RAS inoculation (Figure 3D). Collectively, these results suggest that the Inf. Spz inoculation promotes the greater accumulation of CD8α⁺ DCs to the site of infection probably by modulating the liver microenvironment.

The accumulation of high frequency CD8α⁺ DCs in the liver might not guarantee better CD8⁺ T cell responses as the effective adaptive immunity is critically dependent on the immune status of APCs (46). In our study, we observed the differences in the frequencies of CD8α⁺ DCs; next we determined the activation status of these DCs in the liver by measuring the expression of MHC-II molecules, costimulatory molecules CD80 (B7.1), CD86 (B7.2), and CD40 on the CD8α⁺ DCs by flow cytometry to understand whether Inf. Spz inoculation modulates the activation of CD8α⁺ DCs found in the liver. First, we segregated the CD8α⁺ DCs, for their MHC-II upregulation, into MHC-II^lo and MHC-II^hi populations (Figure 4Ai). We found that Inf. Spz significantly increased frequencies of MHC-II^hi (3.65 ± 0.13%, p < 0.05) as well as MHC-II^lo (2.7 ± 0.33%, p < 0.05) CD8α⁺ DCs in the liver compared to that of RAS (1.4 ± 0.29%, 1.3 ± 0.22%) and sham control (0.7 ± 0.16%, 1.2 ± 0.21%) mice (Figure 4Aii).

As DCs having higher MHC-II are considered matured APCs, we focused on MHC-II^hi populations for their expression of costimulatory molecules B7.1 (CD80) and B7.2 (CD86). We found that 35–50% of MHCII^hi CD8α⁺ DCs in the liver of Inf. Spz mice expressed CD80 and CD86 (Figure 4Bi,ii), whereas significantly low% of the same population expressed the costimulatory molecules in RAS (~10%, p < 0.05) and sham control (5%, p < 0.05) mice (Figure 4Bi,ii). These observations became more interesting when we individually determined the levels of CD80 or CD86 expression on MHCII^hi populations. While the levels of CD80 (MFI = 1,554 ± 46.4) or CD86 (MFI = 390 ± 43.9) expression on MHCII^hiCD8α⁺ DCs was higher in Inf. Spz inoculated mice compared to RAS CD80 (MFI = 805 ± 73) or CD86 (MFI = 194 ± 34.6), the same was not significantly different in RAS inoculated mice CD80 (MFI = 994 ± 103) or CD86 (MFI = 221 ± 15.9) (Figures 4B,C).

The CD40 expression on DCs is shown to be associated with enhanced CD8⁺ T cell mediated protective response against Plasmodium (26). Hence, we measured expression of CD40 on the matured liver CD8α⁺ DCs in response to RAS or Inf. Spz inoculation. We found higher % of MHCII^hiCD8α⁺ DCs express CD40 molecules in response to Inf. Spz inoculation (1.60 ± 0.09%; p < 0.05). The CD40 expression levels were estimated highest on the above population in Inf. Spz inoculated mice (p < 0.05) (Figure 4Biii) followed by RAS and sham control (MFI: 1,421 ± 91.76 in Inf. Spz, 397 ± 49.25 in RAS, 230 ± 24.6 in sham) (Figure 4Ciii).

The immune status of CD8α⁺ DCs is highly influenced by the cytokine microenvironment. Particularly, type I interferons play critical role to fully activate the DCs. We measured the expression of type I interferons (IFN-α, β) and the transcription factor IRF-7 (interferon regulatory factor-7) in the liver to determine the differential responses in RAS vs. Inf. Spz inoculated mice. The expression of cytokines at mRNA level was assessed by quantitative real-time PCR in liver after 72 h postinoculation. We observed significantly higher expression of IRF-7 (p < 0.05), IFN-α (p < 0.05) and IFN-β in the liver upon Inf. Spz than RAS inoculation (Figure 4D). These results suggest that the activation status of liver CD8α⁺ DCs in RAS or Inf. Spz inoculated mice could be the outcome of the differential expression of type I interferons.

Because Inf. Spz have ability to complete the liver-stage developmental cycle, the enhanced CD8⁺ T cell responses by Inf. Spz challenge may be associated with the development of parasite in liver. Unlike Inf. Spz, RAS undergoes aborted development in the liver which might limit the expression profiles of pathogen associated molecules in relation to inflammatory responses. Therefore, we wanted to dissect whether increased dose of RAS inoculation could surpass the accumulation and activation levels of CD8α⁺ DCs in the liver that was found with 10K RAS or inf. Spz inoculation. We inoculated mice with 10K or 50K RAS and evaluated the status of CD8α⁺ DCs in the liver 72 h postinoculation by comparing the same with mice inoculated with 10K Inf. Spz. We found that the accumulation of CD8α⁺ DCs in the liver of mice inoculated with 50K RAS doubled compared to that of 10K RAS, which is still lower than 10K Inf. Spz inoculation (Figure 5A). Upon determining the modulatory role of sporozoites inoculum on CD8α⁺DCs in liver, we found that 50K RAS significantly enhanced the upregulation of MHC-II (Figure 5B) as well as expression of CD80 (MFI = 1,340 ± 42.4, p < 0.05), CD86 (MFI = 1,301 ± 49.9, p < 0.05), and CD40 (MFI = 533 ± 15.4, p < 0.05) compared to 10K RAS (Figure 5D). Interestingly, we found that the % of MHCII^hiCD8α⁺ DC expressing CD80, CD86 or CD40 in the liver of 50K RAS inoculated mice are still appreciably lower than that of Inf. Spz; CD80 (0.338 ± 0.06%, 50K RAS vs. 1.13 ± 0.18%, Inf. Spz p < 0.05), CD86 (0.58 ± 0.06%, 50K RAS vs. 1.01 ± 0.16%, Inf. Spz p < 0.05), and CD40 (0.32 ± 0.046%, 50K RAS vs. 0.613 ± 0.10%, Inf. Spz p < 0.05) (Figure 5C).

Whole sporozoites vaccination approach requires a multiple immunizations to elicit the protective threshold frequencies of LS-specific CD8⁺ T cells, required to confer sterile protection against Plasmodia infection (19). As explained earlier, Inf. Spz challenge following multiple RAS immunizations plays a critical role in enhancing the longevity of the sterile protection by maintaining the frequencies of LS-specific CD8⁺ T cells (Figure 1). Therefore, we asked whether Inf. Spz challenge modulates the accumulation of CD8α⁺ DCs in RAS inoculated mice. The mice were primed with RAS (50K) i.v. and on day 15 post-priming, one of the groups of mice were challenged with 10K Inf. Spz, whereas another group received 10K RAS i.v. After 72 h, the mice were sacrificed and the accumulation and activation of CD8α⁺ DCs in the liver was determined (Figure 6A). We found that challenge with Inf. Spz promoted significantly higher accumulation of CD8α⁺ DCs population (p < 0.05) in the liver of primed mice compared to that of RAS challenge (Figure 6B).

Subsequently, the evaluation of activation status of CD8α⁺ DCs suggests that the pattern of difference between the two challenged groups with regard to the population having MHC-II^hi (Figure 6B) is very similar to the naive mice inoculated with Inf. Spz or RAS (Figure 5). However, % of MHC-II^hiCD8α⁺ DCs expressing CD80 (1.54 ± 0.17%, Inf. Spz challenge vs. 0.21 ± 0.07%, RAS challenge, p < 0.05), CD86 (1.42 ± 0.19%, Inf. Spz challenge vs. 0.35 ± 0.04%, RAS challenge, p < 0.05), or CD40 (1.43 ± 0.42%, Inf. Spz challenge vs. 0.15 ± 0.03%, RAS challenge, p < 0.05) in the liver of Inf. Spz challenged mice were found much higher (Figure 6) compared to that of direct inoculation (Figure 5). Further, we found that Inf. Spz challenge induced enhanced expression of CD80 (MFI = 59,675 ± 1,524, Inf. Spz challenge vs. MFI = 46,113 ± 991, RAS challenge, p < 0.05), CD86 (MFI = 22,793 ± 651, Inf. Spz challenge vs. MFI = 16,406 ± 260, RAS challenge, p < 0.05) and CD40 (MFI = 20,380 ± 1,277, Inf. Spz challenge vs. MFI = 13,288 ± 112, RAS challenge, p < 0.05) (Figure 6C).

The role of CD40 and CD40L (CD154) in APC licensing is well established (47). Moreover, CD4⁺ T cells via APC licensing is known to help CD8⁺ T cells via CD40 signaling and promotes CD8⁺ T cell activation and memory formation by providing the survival signal to T cells (48). We wanted to find out whether Inf. Spz challenge would induce the expression of CD40L in activated CD4⁺ T cells (Figure 7A). We looked for IFN-γ production and CD40L expression on the surface of activated CD4⁺ T cells in the liver in response to RAS or Inf. Spz challenge as part of the above experiment. We found significantly higher % of IFN-γ producing CD4⁺ T cells in the liver of Inf. Spz challenge mice to that of RAS challenge mice (2.23 ± 0.29% Inf. Spz challenge vs. 0.545 ± 0.075% RAS challenge, p < 0.05) (Figure 7Bi); also the % of CD40L expressed cells was significantly higher with at least two folds higher expression on the surface of CD4⁺ T cells of Inf. Spz challenge group mice (2.65 ± 0.43%, Inf. Spz challenge vs. 1.24 ± 0.12%, RAS challenge, p < 0.05) (Figure 7Bii). We also found that the effector functions of CD8⁺ T cells in response to infectious sporozoite challenge were significantly improved compared to RAS challenge (Figure 7C). It is worth noting that both type of challenge induced cytotoxic activity (CD107a) (5.72 ± 0.88%, Inf. Spz challenge vs. 3.36 ± 0.39%, RAS challenge, p < 0.05), while infectious sporozoite challenge specifically promoted IFN-γ response among CD8⁺ T cells (1.67 ± 0.22%, Inf. Spz challenge vs. 0.43 ± 0.08%, RAS challenge, p < 0.05) (Figure 7Cii,iii). Further, we have analyzed the CD40L expression on the surface of CD8⁺ T cells (T_E/EM and T_CM) in response to Inf. Spz challenge in liver. While 4.3 ± 0.62% CD8⁺ T_E/EM cells in the liver of RAS primed mice expressed CD40L upon Inf. Spz challenge, the same was found to be 1.36 ± 0.39% following RAS challenge. The expression of CD40L on CD8⁺ T_CM population in the livers of mice followed the similar trend (p < 0.05) (Figure 7Civ,v).