recombinant vaccinia virus expressing Plasmodium yoelii CSP
Protective immune responses against malaria are induced by immunization with radiation-attenuated Plasmodium sporozoites. In contrast, non-viable, heat-killed sporozoites do not induce protection, emphasizing the requirement for live parasites to achieve effective immune responses. Using an experimental system with CD8+ T cells from T cell receptor-transgenic mice, we analyzed the primary CD8+ T cell responses elicited by heat-killed inactivated sporozoites. We found that the numbers of specific CD8+ T cells induced were much lower compared to when immunizing with attenuated sporozoites; however, the kinetics of activation and the phenotype of these T cells were similar in both groups. Despite their low frequency after priming, high numbers of specific CD8+ T cells were observed after boosting with a recombinant vaccinia virus. Upon induction of the recall response, the same level of protection was observed when either heat-killed or attenuated sporozoites were used for priming. We propose that live parasites are not critical for the induction of memory T cell populations against the malaria liver stages.
Immunization of rodents, non-human primates and humans with attenuated (γ-irradiated) sporozoites (IrrSpz) confers protective immunity against challenge with normal sporozoites 1. Indeed, the utilization of IrrSpz has been the gold standard in the development of experimental vaccines that target the sporozoite and liver stages of malaria. Protective immunity elicited following immunization with IrrSpz is multi-factorial, involving T cells and antibodies. However, studies with rodent parasites, particularly Plasmodium yoelii and Plasmodium berghei, have consistently demonstrated a significant role for parasite-specific CD8+ T cells that inhibit the intracellular development of malaria liver stages 2.
The activation of a protective T cell response against malaria liver stages appears to depend on parasite viability. IrrSpz sporozoites retain their ability to infect liver cells and to start early development, but further differentiation into late liver stages and blood stages is aborted 3, 4. Normal sporozoites complete development in the liver and also induce T cell-mediated protection against normal sporozoite challenge when the blood stages of infection are inhibited with drug treatments to avoid interference with liver stage-specific T cell responses 5–7. In contrast, early studies have shown that immunizations with non-viable sporozoites, e.g. heat-killed (HKSpz) or freeze-thawed sporozoites, provide insufficient protection against normal sporozoite challenge 8, 9. Whether non-viable sporozoites induce T cell responses is unknown.
As described in other infectious systems, it is likely that viable sporozoites induce ample inflammatory stimuli that support the generation of parasite-specific effector and memory T cells. The opposite may be true for non-viable sporozoites, which confer inadequate protection, similar to vaccination with sub-unit antigens in the absence of adjuvants. Thus, the factors that govern the efficiency of anti-Plasmodium immune responses remain poorly understood.
To elucidate the basis of the poor immunogenicity of HKSpz, we dissected the CD8+ T cell response induced after HKSpz immunization. For this purpose, we utilized an experimental system with CD8+ T cells from T cell receptor-transgenic mice. Recent studies with this transgenic system have facilitated the in vivo analysis of the events that lead to the development of parasite-specific effector and memory CD8+ T cells after immunization with IrrSpz. Due to the high sensitivity of this experimental system, we were able to determine the frequencies of the antigen-specific CD8+ T cell response induced following HKSpz immunization. Here, we show that this response involves very low numbers of specific T cells. Moreover, we demonstrate that this response can be recalled by recombinant vaccinia virus, and the resulting secondary response can confer sterile protection.
Induction of protective immunity following immunization with IrrSpz but not HKSpz
We compared the primary CD8+ T cell responses induced after priming with P. yoelii IrrSpz and HKSpz in normal mice. For this purpose, Balb/c mice were immunized i.v. with equal numbers of either IrrSpz or HKSpz. At 16 and 35 days after immunization, we measured the epitope-specific IFN-γ secretion by T cells in the spleen using an ELISPOT. We used a peptide derived from the P. yoelii circumsporozoite protein (CSP) corresponding to the well-characterized H2Kd-restricted CD8+ T cell epitope SYVPSAEQI, to coat target cells. Parasite-induced T cell responses were significantly larger in mice immunized with IrrSpz. Very low numbers of epitope-specific CD8+ T cells were detected in the spleens of mice immunized with HKSpz (Fig. 1A).
We also measured the protective immunity induced by priming with IrrSpz and HKSpz. Mice immunized with either IrrSpz or HKSpz were challenged 16 days after immunization with normal sporozoites. Forty-two hours later, the livers of these mice were excised for total RNA isolation. Following reverse transcription, we used a real-time PCR assay to quantitatively measure the copies of 18s rRNA of liver stage parasites against known standards 10. Immunization with IrrSpz conferred ∼95% reduction in liver stage burden as compared to a non-significant reduction when immunizing with HKSpz (Fig. 1B). These results confirm previous findings that HKSpz do not confer protective immunity 9.
Kinetics of CSP-specific CD8+ T cell responses following HKSpz immunization: Transgenic system
To characterize in detail the primary CD8+ T cell response induced by HKSpz, we used CD8+ T cells from transgenic mice that express a T cell receptor specific for the CSP epitope SYVPSAEQI. As previously reported, these naive transgenic CD8+ T cells can be adoptively transferred to syngeneic mice to augment the very low precursor frequency of anti-SYVPSAEQI T cells in normal mice. The activation of the CD8+ T cells can then be followed after immunization 11, 12.
To compare the primary CD8+ T cell responses induced by IrrSpz and HKSpz, we transferred equal numbers of transgenic CD8+ T cells to syngeneic Balb/c mice. The recipient mice were then immunized with equal numbers of either IrrSpz or HKSpz. The evolution of the CD8+ T cell response in the spleen and in the liver for both immunization groups was then measured by ELISPOT and flow cytometry. As expected, immunization of the recipient mice with IrrSpz induced the clonal expansion of the transferred SYVPSAEQI-specific CD8+ Thy1.1+ T cells, which correlated with the production of the pro-inflammatory cytokine IFN-γ, which peaked on day 4 to 5 (Fig. 2). This was followed by a contraction phase between days 6 and 8, and then by subsequent stabilization and establishment of memory populations that were detectable in the spleen and the liver for up to 2 months (Fig. 2).
Strikingly, in mice immunized with HKSpz, CD8+ T cells were induced with comparable kinetics (Fig. 2). That is, the CD8+ T cell response peaked and contracted on similar days, and this response remained detectable in the spleen for up to 2 months. Notably, priming with HKSpz appeared to result in very limited T cell expansion. Consistent with results from the non-transgenic system (Fig. 1), the number of CD8+ T cells was significantly lower (Δ1–2 logs) than the number of CD8+ T cells induced by IrrSpz.
Phenotypic analysis of CD8+ T cells induced by HKSpz: Transgenic system
The activation and differentiation of T cells are coupled to changes in the expression of cell surface markers, which can be analyzed by flow cytometry. To determine the phenotypic activation status of the CD8+ T cells induced by HKSpz in the transgenic adoptive transfer system, we characterized these cells by flow cytometry following staining with two widely used markers: CD62L and CD44 13. CD62L, a lymphoid homing receptor, is highly expressed by naive and central memory T cells, in contrast to effector memory T cells, which present heterogeneous expression of this marker. High expression of CD44 is characteristic of activated/memory T cells. Therefore, naive cells are generally CD62Lhi CD44lo, central memory T cells are CD62Lhi CD44hi, and effector memory T cells are CD62Llo CD44hi. As shown in Fig. 3A, SYVPSAEQI-specific CD8+ T cells, which were activated following IrrSpz immunization, showed a rapid decline in CD62L expression. After day 8, CD8+ T cells that heterogeneously were either CD62hi or CD62lo were found in the spleen. Constant up-regulation of CD44 expression was also detected after immunization, thus indicating effective activation of specific T cells and generation of memory populations.
Despite the limited expansion of CD8+ T cells in HKSpz-immunized mice, we were able to characterize the phenotype of these cells from recipient mice by flow cytometry (Fig. 3). Consistent with the expansion of transferred CD8+ T cells and their production of IFN-γ after HKSpz immunization, these cells also demonstrated phenotypic activation (Fig. 3). We found that a significant number of T cells down-regulated CD62L, indicating the generation of effector memory T cells, which can be detected up to 64 days after immunization. However, CD44 expression in these spleen-derived cells was up-regulated consistently from day 4 to 64, signifying that memory T cells were maintained in these animals. Notably, during the peak of activation (day 4/5), not all of the CD8+ T cells down-regulated CD62L and up-regulated CD44, suggesting that only a fraction of the transgenic T cells are recruited to the immune response.
We also analyzed CD8+ T cells isolated from the livers of mice immunized with IrrSpz or HKSpz (Fig. 3B). While the absolute numbers of CD8+ T cells differed (Fig. 2), all of the cells had the same activated phenotype, CD62Llo CD44hi. Consistent with previous studies in malaria and in other infectious systems, these results indicate that CD8+ T cells traffic to peripheral organs (i.e. liver) following antigen-driven activation 14, 15.
CD8+ T cells from HKSpz-primed mice can be boosted by recombinant vaccinia virus
The main goal of vaccination against malaria liver stages in this model is to induce high levels of T cells capable of inducing protective immunity. To this end, we determined whether we could augment the primary CD8+ T cell response induced by HKSpz immunization. A heterologous prime-boost approach with recombinant vaccinia virus expressing P. yoelii CSP (VacPyCS) as booster could yield a 20-fold increase in CD8+ T cell responses from pre-existing levels 16–18. We utilized the normal mouse system to test whether this experimental vaccine could boost the low levels of primary CD8+ T cells obtained after HKSpz immunization. Normal Balb/c mice were primed with either IrrSpz or HKSpz, and then boosted with VacPyCS 35 days later. Epitope-specific IFN-γ secretion by T cells in the spleen and the liver was measured by ELISPOT at intervals after boosting.
Remarkably, mice primed with HKSpz and boosted with VacPyCS induced comparable numbers of CSP-specific IFN-γ-secreting CD8+ T cells relative to mice primed with IrrSpz and boosted with the same recombinant virus (Fig. 4). Moreover, the high CD8+ T cell numbers were maintained at least 2 months after boosting. Mice immunized only with VacPyCS elicited a response that is ∼8–10× lower than that observed in mice previously immunized with either IrrSpz or HKSpz. As shown in previous work 11, 12, non-immunized mice do not have IFN-γ-producing antigen-specific CD8+ T cells. These results demonstrate that the very low CD8+ T cell response induced by HKSpz can be efficiently recalled by booster immunization.
CD8+ T cells mediate protective immunity following HKSpz prime-recombinant vaccinia boost immunization
Next, we asked whether the quality of boosting by VacPyCS in both IrrSpz- and HKSpz-immunized mice were similar in conferring protective immunity after challenge with viable normal sporozoites. For this purpose, IrrSpz- and HKSpz-primed mice were boosted with VacPyCS 35 days later. Sixteen and 64 days after boosting, groups of mice were challenged with viable normal sporozoites. Between 40 and 42 h later, the parasite burden in the livers was quantified by real-time PCR. As shown in Fig. 5A, both IrrSpz-primed VacPyCS-boosted and HKSpz-primed and VacPyCS-boosted groups conferred 99% inhibition of liver stages when challenged on day 16 after boosting. The high level of protection was also maintained for at least 2 months (Fig. 5B). Control mice that received VacPyCS only were essentially non-protected.
We then determined the mechanism by which protection was achieved in the HKSpz-primed VacPyCS-boosted group. The level of parasite burden in the CD4+ T cell-depleted group was similar to that in the undepleted controls (Fig. 5A). In sharp contrast, depletion of CD8+ T cells prior to challenge led to a significant decrease in inhibition of liver stages (p=0.002), thus indicating that CD8+ T cells play a major role in this vaccination model (Fig. 5A). The parasite burden of CD8+ T cell-depleted mice was lower than that observed in mice primed only with VacPyCS although this difference did not reach statistical significance (p=0.07). Nevertheless, since the VacPyCS contains the entire CSP, it is possible that antibodies to the central repeats, which are targets of protective antibodies, were also boosted in HKSpz-primed mice and thus, they may also play a protective role. As shown in Fig. 1, mice primed only with HKSpz do not induce protection.
Finally, we ascertained whether the efficient inhibition of liver stage burden (Fig. 5) translated into the induction of sterile protective immunity. For this purpose, HKSpz- or IrrSpz-primed mice were boosted with VacPyCS and then challenged with 125 sporozoites 16 days later. The occurrence of bood stage parasites was followed daily after challenge. In addition, HKSpz- or IrrSpz-primed mice were boosted with VacPyCS and challenged with 75 sporozoites 43 days later. Remarkably, more than 50% of the mice which had been primed with HKSpz or IrrSpz and boosted with VacPyCS were completely protected following normal sporozoite challenge (Table 1). For both groups, significant delays in prepatency were detected in mice that succumbed to blood stage infection, in contrast to mice that were immunized with recombinant vaccinia alone and to mice that were infection controls. Altogether, these results indicate that the HKSpz prime-VacPyCS boost strategy effectively induces high numbers of antigen-specific CD8+ T cells and provides high levels of protection against challenge with normal sporozoites.
|Priminga) (day 0)||Boostingb) (day 35)||Challenge||Sterile Protectionc)||Prepatent Periodd)|
|IrrSpz||VacPyCS||125 Spz||66% (8/12)||4.5±0.6 days|
|HKSpz||VacPyCS||16 days after boosting||50% (6/12)||4.0±0.8 days|
|PBS||VacPyCS||0% (0/6)||3.3±0.5 days|
|PBS||PBS||0% (0/8)||3.2±0.5 days|
|IrrSpz||VacPyCS||75 Spz||80% (4/5)||5.0±0.0 days|
|HKSpz||VacPyCS||43 days after boosting||83% (5/6)||5.0±0.0 days|
|PBS||VacPyCS||0% (0/3)||3.7±0.6 days|
|PBS||PBS||0% (0/3)||3.3±0.6 days|
Live vaccines, which generally induce high levels of protection, present a number of safety and logistic difficulties. On the other hand, vaccines based on killed or inactivated microbes are potentially safer to administer, but they tend to have poor immunogenicity and are usually effective only when delivered with an adjuvant. To date, the gold standard vaccine against the liver stages of Plasmodium has been IrrSpz. HKSpz have also been characterized but they fail to elicit high levels of protection against normal sporozoite challenge 9. Therefore, we sought to address the issue of the poor immunogenicity of HKSpz.
Our findings suggest that priming with inactivated vaccines results in minimal expansion of antigen-specific T cells and poor protection, but low numbers of effector and memory T cells are nevertheless generated. Recent studies have proposed that as IrrSpz enter the liver and invade hepatocytes, parasite-derived antigens may be taken up by antigen-presenting cells in this organ, leading to the induction of protective immunity 19, 20. As HKSpz are not able to reach the liver and infect hepatocytes, the induction of T cell responses probably occurs in the spleen and lymph nodes. In this context. the quantitative differences observed in the primary response induced by IrrSpz versus HKSpz could be due, at least in part, to the tissue compartment where antigen presentation occurs. Despite these differences, when recalled by an optimal boosting agent, HKSpz-induced T cells can generate a secondary response that provides a high level of protection.
While the response of endogenous CD8+ T cell precursors to HKSpz immunization was barely detectable in normal mice, adoptive transfer of transgenic T cell precursors to normal mice enabled the kinetic analysis of the development of antigen-specific CD8+ T cell responses following HKSpz immunization. Our results demonstrate that HKSpz follow similar kinetics of induction of CD8+ T cell responses as compared to IrrSpz, as measured by the clonal expansion of the transferred cells. However, the numbers of antigen-specific CD8+ T cells induced by HKSpz are consistently lower than those induced by IrrSpz. Nonetheless, despite their low numbers, these CD8+ T cells produce the pro-inflammatory cytokine IFN-γ, a hallmark of effector-like function. Moreover, these effector CD8+ T cells are also found in the liver, and they present the typical activated phenotypic markers, consistent with previous studies in malaria indicating that CD8+ T cells migrate to peripheral organs following antigen-driven activation 14.
Heterologous prime-boosting immunization strategies, which use a combination of different delivery vectors encoding the same epitopes or antigen, and which are delivered at certain time intervals, provide the best levels of protection 16, 21. HKSpz-primed mice mounted robust and significant CD8+ T cell response after heterologous boosting with VacPyCS. Remarkably, this recall response translated into high levels of protection against normal sporozoite challenge. This protection was measured either by reduction in the development of liver stages or by the induction of sterile immunity. Depletion studies indicate that T cells, particularly CD8+ T cells, mediate the protection induced by HKSpz prime-VacPyCS boost immunization.
As major scientific efforts are currently directed at the generation of anti-malaria vaccines based on IrrSpz 22 or genetically attenuated sporozoites 23, the safety and logistic problems associated with these approaches must be fully resolved. Our findings suggest that using HKSpz is an effective priming strategy for immunization against liver stages, with less safety and logistic concerns.
Materials and methods
Mice, parasites and viruses
Female Balb/c mice (Thy1.2+; 6–8 wk old) were obtained from the National Cancer Institute (Frederick, MD) or from Taconic Farms (Germantown, NY). The generation of the SYVPSAEQI-specific T cell receptor-transgenic mice was described previously 11, and these mice were maintained in the Balb/c Thy1.1+ background (a gift from Dr. Hao Shen, University of Pennsylvania, PA). For adoptive transfer experiments, spleen cells from the transgenic mice, containing ∼2×106 naive antigen-specific CD8+ T cells, were injected i.v. into syngeneic recipient Balb/c mice 24.
P. yoelii (17X NL) sporozoites were obtained by the dissection of infected Anopheles stephensi mosquito salivary glands. Attenuation by irradiation (IrrSpz) was accomplished by exposing dissected sporozoites to a γ-source (15 krad) HKSpz were prepared by incubating the parasites at 72°C for 15 min followed by 95°C for 15 min. For immunization with either IrrSpz or HKSpz, 7.5×104 sporozoites/mouse were injected i.v.. For challenge, 1.5×104 viable sporozoites were used.
The generation of VacPyCS was previously described 25. VacPyCS was administered i.v. at 2×107 PFU/mouse.
Enzyme-linked immunospot assay
The ex vivo IFN-γ ELISPOT for quantifying epitope-specific CD8+ T cells was performed as described 26, 27. A20.2J cells were used as antigen-presenting cells and were coated with SYVPSAEQI peptide (Biosynthesis, Lewisville, TX), the CD8+ T cell epitope of PyCSP 28. Anti-mouse IFN-γ (R4) and biotinylated anti-mouse IFN-γ (XMG1.2) were obtained from BD Biosciences (San Diego, CA). Bulk analysis of spleen and liver-infiltrating cells was performed. Isolation of liver-infiltrating cells was performed as described 29.
Flow cytometric analysis
Antibodies for flow cytometry were obtained from BD Biosciences. These include monoclonal antibodies to CD8 (53-6.7), Thy1.1 (OX-7), CD62L (MEL14) and CD44 (IM7). Spleen or liver-infiltrating cells were stained using standard protocols. Cells were analyzed using FACSCalibur and Cell Quest software (BD Biosciences).
Depletion of CD4+ and CD8+ T cells
Monoclonal antibodies to CD4 (GK1.5) and CD8 (YTS169) were obtained from Harlan (Indianapolis, IN) and were used to deplete T cell subsets from immunized mice. Depletion was performed by intraperitoneal administration of 0.2 mg of respective antibodies for three consecutive days prior to viable sporozoite challenge.
Assays for protection
Forty-two hours following challenge with viable sporozoites, the livers of immunized mice were excised and processed for total RNA isolation and reverse transcription as described 10. The resulting cDNA were used as templates for real-time PCR of P. yoelii 18s rRNA sequences. The primers used were as described 10, and the amplification was performed in an iCycler iQ Real-Time PCR Detection System (Bio-Rad, Hercules, CA).
The authors would like to thank Dr. Maria Thelma Ocampo-Hafalla (New York University) for reviewing the manuscript. Funding sources: NIH grants AI 044375 (to F.Z.) and AI 053698 (A.R.), and an American Liver Foundation Gene Varian Research Fellowship (to J.C.R.H.).