Adoptive Immunotherapy against Experimental Visceral Leishmaniasis with CD8⁺ T Cells Requires the Presence of Cognate Antigen

Abstract

CD8⁺ T cells have a protective role in experimental visceral leishmaniasis. However, the observation that inflammatory cytokines induce bystander activation of CD8⁺ T cells questions the need for antigen-dependent effector function. Here, we demonstrate that successful adoptive immunotherapy with CD8⁺ T cells is strictly dependent upon the presence of cognate antigen.

CD8⁺ T cells provide a major component of host defense against a variety of viral and nonviral pathogens and have been implicated as a major component of vaccine-induced immunity against some of the most important human pathogens (1-3). CD8⁺ T cells have been demonstrated to play a role in various experimental models of leishmaniasis, including infection with Leishmania major (2, 13), L. mexicana (10), L. infantum (21), and L. donovani (8, 14, 19, 20). However, the study of CD8⁺ T-cell function in leishmaniasis has been hampered compared to the study of their role in viral and bacterial infections by the lack of suitable tools for directly measuring the function of Leishmania-specific CD8⁺ T cells and for discriminating between antigen-dependent and potentially antigen-independent effector functions (10). In two models of cutaneous leishmaniasis, antigen-specific CD8⁺ T cells generated to defined Leishmania-expressed antigens have failed either to exert protection in vivo (22) or to kill infected macrophages in vitro (10). Recent studies demonstrating that a variety of inflammatory cytokines can mediate antigen-independent “bystander activation” of CD8⁺ memory and effector T cells, leading to host protection in different disease models (3, 6, 9), now question whether the effector function of CD8⁺ T cells in experimental leishmaniasis requires the presence of cognate antigen. Here, we have used adoptive CD8⁺ T-cell immunotherapy during experimental visceral leishmaniasis to directly examine this question.

We recently described a vaccination model in which host protection against L. donovani is mediated at least in part by CD8⁺ T cells with specificity for an amastigote-expressed antigen, hydrophilic acylated surface protein B1 (HASPB1) (18). HASPB1 is unusual in being targeted for expression at the plasma membrane by a unique N terminus which specifies its dual acylation (4), and this property may facilitate phagosomal presentation for class I processing (5). We therefore made use of this property of HASPB1 and generated transgenic L. donovani expressing a HASP_1-18-OVA fusion protein. Full details of the generation of these and other Leishmania lines expressing HASP_1-18-OVA will be published elsewhere. Briefly, the sequence encoding the N-terminal 18 amino acids of the HASPs (4) was cloned in frame with the OVA open reading frame and inserted into the pSSU vector for gene targeting by homologous recombination into a ribosomal DNA locus (11). Log-phase promastigotes of L. donovani (LV9) were electroporated with 40 μg of PacI/PmeI-digested DNA, and the resulting transfectants were maintained in 40 μg/ml hygromycin (Invitrogen) prior to single-cell cloning. Clones were screened for integration of the linear targeting fragment by Southern blot analysis of genomic DNA and for OVA expression by immunoblotting analysis of parasite lysate (Fig. 1). These transgenic parasites are formally designated Ld18SrRNA::HASP_1-18-OVA (herein referred to as PINK). In order to recover infectivity, positive PINK clones were transformed to amastigotes (16) and maintained by serial passage in hamsters. Expression of OVA was maintained over several months of passage in the absence of drug selection. Semiquantitative analysis of the levels of OVA expression indicated that these transgenic parasites expressed approximately 2 ng OVA/10⁶ promastigotes or approximately 0.05% of the total amount of protein, similar to the expression level of HASPB in Leishmania (4) and somewhat higher than the level of OVA expressed in Leishmania from the episomal vector pX (7). Clones of transgenic OVA-expressing L. donovani were fully infective to mice, and host immunity was similarly organ specific, as has been described for wild-type infections. Clone 6, with maximal expression of OVA as determined by immunoblotting, was chosen for in vivo studies.

FIG. 1.

Transgenic L. donovani organisms expressing ovalbumin. A, Southern blot analysis of wild-type L. donovani and various clones of transgenic L. donovani (PINK clones 2, 3, 6, and 11) to indicate genomic integration of the HASP:OVA transgene. Genomic DNA was digested with XmaI and probed with a 380-bp ³²P-labeled fragment of the ovalbumin gene. A single ∼11-kb chromosomal fragment was detected in all transgenic lines. B, Immunoblot analysis of wild-type L. donovani promastigotes and various clones of transgenic L. donovani (PINK clones 1, 2, 3, 6, and 11) probed with anti-OVA antiserum. Native OVA is shown as a control. Blots were reprobed for the constitutive protein N-myristoyl transferase, as a control for equivalent loadings.

To determine whether adoptive immunotherapy with CD8⁺ T cells altered the course of experimental visceral leishmaniasis and to define whether such an effect required in vivo antigen recognition, B6.CD45.1 mice were infected intravenously with 2 × 10⁷ wild-type L. donovani (LV9) amastigotes or transgenic PINK amastigotes. At day 21 postinfection, adoptive transfer of 1 × 10⁶ T-cell receptor transgenic OVA-specific OT-I CD8⁺ T cells was performed (see below). Hepatic and splenic parasite burdens were determined by examining methanol-fixed, Giemsa-stained tissue impression smears 7 days after transfer (24). All animal procedures were approved by the LSHTM Animal Procedures Ethics Committee and the United Kingdom Home Office.

In the first series of experiments, we wished to evaluate whether in vitro-generated cytotoxic CD8⁺ T cells were host protective. OT-I cytotoxic T lymphocytes (CTL) were generated by incubation of naïve OT-I cells with 1 nM OVA_257-264, 350 ng/ml recombinant interleukin 4 (rIL-4), and 10 ng/ml rIL-2 for 3 days, followed by a further 2 days with cytokine and no peptide (1). CTL activity was confirmed by lysis of EG7-OVA (12) but not parental EL4 cells (data not shown). As shown in Fig. 2, the transfer of OT-I CTL into mice infected with wild-type L. donovani had no impact on parasite growth in either the liver or the spleen. In contrast, OT-I cells exerted significant control over the course of infection with PINK amastigotes in both target tissues. In the liver, this was evident as leishmanicidal activity, whereas in the spleen, the effect was restricted to being leishmanistatic. Together, these data both indicate a requirement for cognate antigen for successful immunotherapy by CTL and, by implication, demonstrate that the inflammatory environment created by wild-type L. donovani infection is insufficient to induce bystander effector responses from these CTL.

FIG. 2.

Adoptive immunotherapy for experimental visceral leishmaniasis using CTL. B6 mice infected with wild-type L. donovani (A) or OVA-expressing L. donovani PINK (B) received either 10⁶ in vitro-generated OT-I CTL (open circles) or no transfer (filled circles) at day 21 postinfection. Parasite burdens in liver (left panels) and spleen (right panels) were determined at day 28 by direct counting of amastigotes in tissue impression smears. Data are presented as Leishman Donovan units (LDU) ± standard errors (five mice per time point). **, P was <0.01 versus values for control mice.

In the second series of experiments, we confirmed and extended these findings to other populations of CD8⁺ T cells. CD8⁺ memory T cells have been divided into two subsets. Effector memory T cells continually recirculate through peripheral tissues, whereas central memory T cells, by virtue of their high levels of expression of CD62L and CCR7, are endowed with lymph node homing capacity (17). CD8⁺ T cells resembling effector and central memory cells were derived in vitro as described previously (23). Briefly, splenocytes from naïve OT-1 transgenic mice were incubated with 10 μg/ml OVA_257-264 for 1 h at 37°C, washed, and cultured for a further 48 h. Cells were then washed and incubated for a further 9 days with either 20 ng/ml rIL-2 (effector memory-like OT-I^IL-2) or 20 ng/ml rIL-15 (central memory-like OT-I^IL-15). Cytokines were changed every second day. OT-I^IL-2 and OT-I^IL-15 cells used for transfer were >88% CD62L^lo and >75% CD62L^hi, respectively. In some experiments, cells were sorted on the basis of CD62L to >99% purity, with similar results.

As with OT-I CTL, neither OT-I^IL-2 nor OT-I^IL-15 cells had any effect on the course of wild-type L. donovani infection (Fig. 3A). In contrast, the transfer of OT-I^IL-2 cells had a rapid impact on the course of PINK infection (Fig. 3B). Hepatic parasite burden 48 h after transfer was reduced by 61% ± 6% compared to levels in control mice (P < 0.001), an effect mirrored in the spleen (59% ± 5%; P < 0.001). This therapeutic effect was sustained and/or improved when determined 7 days after transfer (54% ± 6% and 84% ± 6% reduction in liver and spleen, respectively). In comparison, OT-I^IL-15 and OT-I cells adoptively transferred directly after isolation from naïve OT-I transgenic mice did not exert a statistically significant effect on host resistance until day 7 after transfer, though at this time they were as effective as OT-I^IL-2 cells (Fig. 3B). Collectively, these data indicate that cells with memory cell characteristics, as well as naïve OVA-specific CD8⁺ T cells, are capable of transferring significant protection against ongoing infection, albeit with differing kinetics of action. As with transferred CTL, we found no evidence for bystander activation of host-protective function. Rather, therapeutic benefit was strictly dependent upon infection with parasites expressing the cognate antigen recognized by these CD8⁺ T cells.

FIG. 3.

Adoptive transfer of naïve and memory phenotype OT-I cells. OT-I^IL-2 cells (open circles), OT-I^IL-15 cells (filled triangles), or OT-I cells directly from naïve mice (open triangles) (10⁶) were transferred into B6 mice infected 21 days previously with either wild-type L. donovani (A) or L. donovani PINK (B). Control mice received no cell transfer (filled circles). Parasite burden was determined in liver (left panels) and spleen (right panels) prior to transfer (day 21) and at day 23 and day 28, and data are shown as mean Leishman Donovan units (LDU) ± standard errors (five mice per time point). *, P was <0.05; **, P was <0.01 versus values for control mice.

In summary, we have addressed a basic yet fundamental issue related to CD8⁺ T-cell function in the later stages of experimental visceral leishmaniasis, namely, whether cognate antigen is required for CD8⁺ T cells to exert host-protective function. Our data indicate that this is indeed the case and suggest that future interventions based on increasing antigen recognition by endogenous CD8⁺ T cells during infection may be beneficial. Additional studies are required to confirm why bystander activation of effector function in CD8⁺ T cells does not occur to any significant extent in this model, in spite of extensive inflammation and evidence for bystander proliferation of both CD4⁺ (15) and CD8⁺ (15a) T cells. Furthermore, the model system that we describe here should also be useful for analyzing other facets of CD8⁺ T-cell biology, including the differential requirements for antigen-dependent activation of naïve versus effector/memory CD8⁺ T cells and the mechanisms underlying their protective effects.