Deficiency of Antigen-Specific B Cells Results in Decreased Trypanosoma cruzi Systemic but Not Mucosal Immunity Due to CD8 T Cell Exhaustion

Trypanosoma cruzi is a protozoan parasite and the etiological agent of Chagas’ disease. Prevention and vector control practices throughout Latin America have reduced the current number of infected individuals to ∼8–11 million people (1). However, movement of infected individuals to nonendemic areas poses an emerging public health problem. Up to 40% of infected individuals develop serious cardiac and/or gastrointestinal problems 1–30 y after infection, leading to significant morbidity and mortality. T. cruzi is transmitted to both humans and animals by reduviid insects of the subfamily Triatominae. Infectious parasites are present in the excreta of infected Triatominae insects and can transmit via breaks in the skin, mucosal tissues associated with the eye and gastrointestinal tract, congenital transmission from mother to child, as well as blood and tissue donation from infected individuals.

T cells and B cells have been shown to play critical roles in protection against T. cruzi. CD8⁺ T cells are important in both primary (2) and secondary (3, 4) protective T. cruzi immunity. There are several highly immunodominant CD8⁺ epitopes encoded in the T. cruzi trans-sialidase (TS) family of genes. In the BALB/c background, up to 30–40% of CD8⁺ T cells during acute infection are responsive to the immunodominant H-2K^d–restricted TS epitope IYNVGQVSI (TSKd1) (5, 6). CD4 T cells have been shown to be necessary (7–9) but not sufficient (10) for generation of protective immunity against T. cruzi infection. B cells have also been shown to play an important role in systemic T. cruzi protection. Early work demonstrated that T. cruzi–specific Abs can provide systemic protection via complement activation and lysis (11–13), opsonization (14), and Ab-dependent cellular cytotoxicity (15). These early results demonstrated that B cells play an important role in systemic T. cruzi protection through the production of T. cruzi–specific Abs (16). Additionally, systemic infection of μMT B cell knockout (KO) mice with T. cruzi resulted in initial control of parasite replication but the mice eventually died due to increased parasitemia (16). Previous work by our laboratory demonstrated that T. cruzi mucosal infection induces protective immunity against subsequent challenge (17, 18). This mucosal protection was associated with increased levels of T. cruzi–specific IgG and IgA Ab-secreting cells in the gastric mucosa (17), the initial site of invasion after oral parasite exposures. However, a causal role for B cells in mucosal T. cruzi protection has not been mechanistically defined.

In the present study, we have further examined the importance of B cells for both mucosal and systemic T. cruzi immunity. First, we demonstrate that in contrast to what we initially hypothesized, B cells are not required for mucosal T. cruzi protection. We predicted that B cells producing secretory IgA would be very important in mucosal protection against an extracellular parasite life stage that invades through nasal and gastrointestinal epithelia, but this was found not to be the case. In contrast, we demonstrate that CD8⁺ T cells are critical for mucosal T. cruzi protection. We confirm that B cells are important for systemic T. cruzi protection in both KO and transient depletion models. After virulent systemic challenge, B cell–deficient/depleted mice are unable to control parasitemia and develop increased morbidity and mortality. We further demonstrate that T. cruzi–specific B cells are required for systemic protection, producing parasite-specific Abs that reduce parasite load, resulting in decreased CD8⁺ T cell exhaustion and mortality in B cell KO mice. Finally, we demonstrate that B cells play a critical role in development and maintenance of multifunctional CD8⁺ T cells, as well as in the prevention of CD8⁺ T cell exhaustion. Overall, our new data suggest that memory CD8⁺ T cells primed in mice lacking T. cruzi–specific B cells are sufficient to control low levels of tissue parasitism. However, these CD8⁺ T cells eventually become exhausted after high-level systemic challenge due to prolonged exposure to high levels of Ag.

BALB/c, SCID BALB/c, JhD B cell KO BALB/c, and hen egg lysozyme (HEL)–specific BCR transgenic and C57BL/6 mice were used throughout this study. In some instances, female BALB/c mice were B cell depleted or not using anti-CD20 (18B12) or IgG2a (2B8) mAb (provided by Biogen Idec). Mice were treated with 250 μg Ab in PBS i.p. every 2–3 wk. Mice were housed under pathogen-free conditions. All studies were conducted with the approval of the Institutional Animal Care and Use Committee/Animal Care Committee in an American Association for the Accreditation of Laboratory Animal Care–accredited facility at Saint Louis University. Tulahuén strain culture-derived metacyclic trypomastigotes (CMT) were used for mucosal infections/challenges and blood-form trypomastigotes (BFT) were used for systemic challenges as described (18, 19) and detailed in Fig. 1. Briefly, CMT can infect both mucosally (via oral, intranasal, conjunctival routes) and systemically whereas BFT can only infect systemically. These differences in infectivity are due to differential glycosylation resulting in increased inflammation induced by the BFT life stage, which is hypothesized to prevent infection at mucosal surfaces (20, 21). T. cruzi infection–induced immune mice (referred to as Tc immune mice throughout this study) mice were generated by repeated low-dose infection (1–3 × 10⁶ CMT intragastrically [i.g.]). For i.g. infection of mice, mice were first given 0.5 ml 1.5% sodium bicarbonate in HBSS i.g. using a ball-ended 1.5-inch, 22-gauge animal feeding needle and rested for 15 min to neutralize stomach pH. Parasites were then diluted in PBS plus 1% glucose, and 0.1 ml was delivered i.g. These mice are referred to as Tc immune mice throughout this study.

To generate mucosal immunity, naive BALB/c mice (anti-CD20/IgG2a mAb treated) were vaccinated with 50 μg recombinant TS plus 10 μg CpG-containing oligodeoxynucleotide 1826 (InvivoGen, San Diego, CA) intranasally twice, 2 wk apart (TS/CpG). To generate systemic immunity, naive wild-type (WT), B cell KO, or anti-CD20/IgG2a mAb–treated BALB/c mice were vaccinated with our heterologous TS vaccination protocol (100 μg TS-DNA on days 0 and 14 i.m., and boosted on days 44 and 58 with 1 × 10⁸ PFU adenovirus-expressing TS s.c. (base of tail) and intranasally (4). These mice are referred to as TS vaccine–induced immune (TS immune) mice throughout this study.

To assess protective mucosal immunity, mice were challenged with 1–2 × 10⁷ CMT i.g. Twelve days later, which is the peak time for T. cruzi replication in the gastric mucosa (17), mice were sacrificed and gastric DNA was used for quantitative PCR as described (18). Briefly, 100–200 ng gastric DNA purified using Qiagen DNeasy blood and tissue kits was added to each real-time PCR reaction containing 900 nM each primer (5′-AACCACCACGACAACCACAA-3′ and 5′-TGCAGGACATCTGCACAAAGTA-3′), 250 nM TaqMan probe (FAM/TAM 5′-TGCCCCAGGACCGTCCCCA-3′), and 1× TaqMan PCR master mix. Thermocycling conditions using an Applied Biosystems 7500 fast real-time PCR instrument were 95°C for 10 min, followed by 40 cycles of 95°C for 15 s and 60°C for 1 min. A standard curve was generated using DNA purified from a known number of T. cruzi epimastigotes.

To assess protective systemic immunity, mice were challenged with 5000 T. cruzi (BFT) s.c. Hindlimb paralysis was assessed via paralysis scores similar to those used in experimental autoimmune encephalomyelitis studies: 0, no obvious changes in motor functions; 1, hindlimb weakness; 2, complete paralysis of hindlimbs; 3, complete hindlimb paralysis and partial front leg paralysis; 4, death. Morbidity and mortality were assessed weekly for 60–200 d after systemic BFT challenge.

To determine whether total lymphocytes from immune mice could transfer mucosal T. cruzi protection, one mouse equivalent (4-8 × 10⁷) of spleen cells plus draining lymph node cells from Tc immune and naive BALB/c mice were transferred i.v. into SCID mice. Both spleen and draining lymph nodes were used for transfer experiments, as Butcher and colleagues (22) have shown that there are more Ag-specific Ab-secreting cells (ASC) in the spleen after rotavirus infection. Furthermore, it was also shown that splenocytes from immune mice contain lymphocytes with mucosal homing capacity (α₄β₇) that can transfer mucosal protection against rotavirus intestinal infection (22, 23). To determine whether B cells were necessary for the transfer of mucosal T. cruzi protection, B cells were depleted using anti-CD19 microbeads (Miltenyi Biotec, Auburn, CA) resulting in >93% CD19⁺ depletion (Supplemental Fig. 2). CD19⁺-depleted cells (2.4 × 10⁷) from Tc immune or naive BALB/c mice were transferred i.v. into naive SCID mice. To determine whether CD8⁺ T cells alone were sufficient to transfer mucosal T. cruzi protection, CD8⁺ T cells were positively selected using anti-CD8 microbeads (Miltenyi Biotec) resulting in >98% CD8⁺ purity. CD8⁺ T cells (5.2–6.95 × 10⁶) from Tc immune and naive BALB/c mice were transferred i.v. into naive SCID mice. One day after transfer, recipient mice were challenged with 1.2 × 10⁷ T. cruzi (CMT) i.g. Twelve days later, recipient mice were sacrificed and T. cruzi DNA was assessed in the gastric mucosa via real-time PCR. In some instances, CD8⁺Thy1.2⁺ T cells (>95% CD8⁺) were purified from B cell KO, WT BALB/c Tc immune, and naive BALB/c mice and transferred into naive BALB/c Thy1.1⁺ mice. To normalize the number of CD8⁺TSKd1⁺ T cells transferred, purified CD8⁺ T cells were stained with a TSKd1 tetramer and analyzed with an LSR II flow cytometer (BD Biosciences) and FlowJo flow cytometry software (Tree Star, Ashland, OR). Equivalent numbers of CD8⁺TSKd1⁺tetramer⁺ cells from B cell KO and WT Tc immune mice were transferred i.v. into naive BALB/c Thy1.1 mice. One day after transfer, recipient mice, along with naive control mice, were challenged systemically with 5000 T. cruzi (BFT) s.c. and followed for parasitemia and survival.

Serum was collected from naive and WT Tc immune BALB/c mice (generated by repeated low-dose oral infection using CMT i.g.). B cell KO BALB/c Tc immune mice were given either 300 μl WT BALB/c Tc immune serum, WT BALB/c naive serum, or PBS i.p. every week starting either 2 wk prior to or at the time of T. cruzi (BFT) s.c. systemic challenge. T. cruzi–specific serum IgG endpoint titers were measured every 1–2 wk.

Serum was collected at different time points and assessed for TS-specific serum IgG responses via ELISA as described (24). In some instances, 8 M urea was added to evaluate for the presence of high-affinity Abs. Following serum incubation, the ELISA plate was washed with PBS plus 0.05% Tween 20. Urea (8 M) was added for 5 min at room temperature, washed four times with PBS–Tween 20 with the last wash incubated for 5 min at room temperature before removal. Endpoint titers were calculated using Unitwin ELISA software (PhPlate, Stockholm, Sweden).

Millititer HA 96-well microtiter plates with nitrocellulose bases (Millipore, Bedford, MA) were coated with either purified rat anti-mouse IFN-γ (clone R46A2; BD Pharmingen, San Diego, CA) or recombinant TS as described (25). For IFN-γ ELISPOT analyses, spleen cells (2.5 × 10⁵ cells/well) plus APCs (1 × 10⁵ A20J cells alone [control A20; American Type Culture Collection], A20J cells stably transfected with the TS gene [A20-TS], or A20J cells pulsed with 2.5 μg/ml immunogenic CD8⁺ H-2K^d–restricted TS peptide IYNVGQVSI [A20-TSKd1]) were cultured overnight at 37°C. TS-specific Ab ELISPOT analyses were done as previously described (24). Briefly, 1–5 × 10⁵ spleen cells were added to TS-coated ELISPOT plates. Following overnight incubation at 37°C, anti-mouse IgG or IgA biotin was added to detect T. cruzi–specific IgG and IgA Ab-secreting cells, respectively (SouthernBiotech). Results are represented as the number of spot-forming cells or ASC per million spleen cells or absolute number per spleen. Background responses to A20J cells were subtracted from experimental Ag responses in results shown from IFN-γ ELISPOT assays.

All Abs were purchased from BD Biosciences (San Diego, CA) with the exception of anti–PD-1, anti–Lag-3, and anti-CD62L (eBioscience, San Diego, CA). The TS^IYNVGQVSI/H-2K^d–APC (TSKd1) tetramer was provided through the National Institutes of Health Tetramer Core Facility (Atlanta, GA). To evaluate in vitro TSKd1 expansion, spleen cells were labeled with CFSE (Vybrant CFDA SE cell tracer; Invitrogen) and cultured with 1 × 10⁵ irradiated (12,500 rad) APCs (A20-negative control or A20-TS) for 6 d at 37°C. Cells were then counted using Guava ViaCount reagent and the EasyCyte flow cytometer (Millipore, Billerica, MA). Half of the cells were used for direct cell surface staining and the other half were stimulated with PMA (10 ng/ml), ionomycin (500 ng/ml), and cultured with monensin (GolgiPlug; 1 μl/ml) and brefeldin A (GolgiStop; 0.67 μl/ml) for 3 h at 37°C. Cells were stained as described (4) and analyzed with an LSR II flow cytometer (BD Biosciences) and FlowJo v7 software (Tree Star).

CD8⁺TSKd1⁺ stimulation indices were calculated by taking the absolute number of live CD19⁻CD3⁺CD4⁻CD8⁺tetramer⁺CFSE^lo cells present after A20-TS stimulation divided by the number present after A20-negative control incubation. The percentages and absolute numbers of CD8⁺TSKd1⁺cytokine⁺ T cells were calculated by taking the percentages or numbers of live CD19⁻CD3⁺CD4⁻CD8⁺tetramer⁺CFSE^locytokine⁺ cells present after stimulation with A20-TS subtracted by the number present after A20-negative control incubation.

Cardiac and skeletal muscle inflammation was assessed at days 30 or 87 after systemic T. cruzi (BFT) challenge. The heart and skeletal muscles from individual mice were fixed in 10% formalin, paraffin embedded, sectioned, and stained with H&E. Pathology slides were graded by a blinded pathologist as follows: 1, minimal; 2, mild; 3, moderate; 4, marked.

Statistical analyses were performed using Prism v4 software (GraphPad Software, La Jolla, CA). Mann–Whitney U tests or unpaired Student t tests were used to compare responses between groups. Log-rank tests were used to compare survival between groups.

Previous work has shown that repeated T. cruzi mucosal infections result in chronically infected mice that develop concomitant immunity protective against subsequent mucosal challenges (Tc immune) (17). These Tc immune mice developed high numbers of T. cruzi–specific ASC in the gastric mucosa, suggesting that B cells played an important role in mucosal T. cruzi protection (17). We first sought to directly address the potential causative role of B cells in protective mucosal T. cruzi immunity. Tc immune B cell KO (JhD) and WT BALB/c mice were generated as shown in Fig. 1A. One month after the last infection, Tc immune and naive B cell KO and WT BALB/c mice were challenged orally with a high dose of T. cruzi. Mucosal protection was assessed 12 d later via real-time PCR, as we have previously shown that the highest level of T. cruzi DNA can be measured in the gastric mucosa, the point of initial mucosal invasion after oral parasite inoculation, 10–14 d postchallenge (17) (Fig. 2A). B cell KO mice had significantly higher levels of T. cruzi DNA in the gastric mucosa compared with WT mice after primary mucosal parasite infection (p = 0.0006; Mann–Whitney U test), suggesting that B cells are important in the primary control of parasite replication. However, more relevant for vaccine development is the role of B cells in memory immune responses. Tc immune B cell KO mice did develop mucosal immunity, as they were significantly protected compared with the primary challenged B cell KO mice (Fig. 2A, p < 0.0005; Mann–Whitney U test). Additionally, Tc immune B cell KO mice developed similar levels of mucosal protection compared with WT Tc immune mice (p = 0.1719; Mann–Whitney U test). We also assessed the frequency of T. cruzi–specific T cells in B cell KO and WT Tc immune mice. There were similar frequencies of total T cells specific for the T. cruzi protein TS (A20-TS response) and T cells reactive with an immunodominant CD8 T cell epitope (A20-TSKd1 response) in WT and B cell KO mice (Fig. 2B). We measured immune responses in the spleen, as it has been shown previously that immunity to mucosal pathogens can be efficiently measured in the spleen (22, 23). Ag-specific cells generated in mucosal lymph nodes traffic through the spleen before migrating back to mucosal tissues. Additionally, higher total numbers of rotavirus-specific sIgA ASC have been recovered from spleens than from mucosal tissues (22), and splenic CD8⁺ T cells from rotavirus-immune mice can efficiently transfer mucosal protection (22, 23). Overall, our results demonstrate that B cells are not required for T. cruzi mucosal immunity.

Next, we wanted to determine whether B cells play an important role in TS vaccine–induced mucosal immunity (Fig. 1B). TS is an enzyme on the surface of T. cruzi that is important for parasite infectivity, as it has both neuraminidase and sialic acid transfer activity. We have previously shown that mice vaccinated with TS plus the TLR9 agonist CpG 1826 are protected against both mucosal and systemic T. cruzi challenges (25). B cells were first depleted in naive WT BALB/c mice using anti-CD20 mAb (Biogen Idec). TS immune mice (Fig. 1B) were intranasally vaccinated twice with recombinant TS plus CpG 1826 two week apart, rested for 1 mo, challenged with a high dose of T. cruzi orally (10⁷), and assessed for mucosal protection via real-time PCR. Similar to the results seen in Tc immune mice, B cells were not required for TS vaccine–induced protective mucosal immunity (Fig. 2C). TS- and TSKd1-specific T cell responses were comparable between B cell–depleted and isotype control–treated TS-vaccinated mice (Fig. 2D), demonstrating that B cells are not required for the development, maintenance, and/or recall of TS- or TSKd1-specific T cells after mucosal T. cruzi challenge. Thus, B cells are not required for TS-specific vaccine-induced mucosal immunity.

To confirm that B cells are not critical for mucosal T. cruzi protection, we performed adoptive transfer studies. We first determined whether cells from Tc immune mice or TS immune mice generated as shown in Fig. 1C with a more potent vaccination regimen (heterologous TS DNA primed, adenovirus-TS boosted) could transfer mucosal protection, as this has not been shown previously. Owing to the low numbers of cells recovered from the gastric lymph nodes, mixed populations of splenocytes plus gastric lymph node cells were used to transfer maximal numbers of parasite-specific cells. Previous work by others has shown that splenocytes from rotavirus-immune mice contain lymphocytes with mucosal homing capacity that can transfer mucosal protection (22, 23). Both Tc immune (Fig. 3A) and TS immune (Fig. 3B) donor cells were able to transfer mucosal protection to naive SCID mice.

Next, we studied whether B cells are necessary and/or sufficient for the transfer of mucosal protection. We transferred highly purified CD19⁺ B cells from either naive or Tc immune BALB/c mice into SCID mice. However, we could not confirm whether B cells were sufficient, as SCID mice receiving purified B cells (Supplemental Fig. 1A) alone did not develop T. cruzi–specific Ab responses (serum IgG by ELISA or ASC detected by ELISPOT; Supplemental Fig. 1C and 1D, respectively), even though B cells were clearly present (Supplemental Fig. 1B). Thus, we decided to perform B cell depletion studies to determine whether B cells were necessary for the transfer of mucosal protection. B cells were depleted from Tc immune and naive BALB/c splenocytes and gastric lymph node cells (Supplemental Fig. 2). These B cell–depleted fractions were transferred into naive SCID mice and recipient mice were orally challenged the following day with a high dose of T. cruzi i.g. Twelve days after challenge, recipient mice were assessed for the presence of B cells and T. cruzi–specific serum IgG and IgG ASC. Supplemental Fig. 2 shows that in SCID mice given CD19-depleted cells, B cells were undetectable. Fig. 3C demonstrates that B cells are not necessary for the transfer of mucosal protection, as SCID mice that received either total or B cell–depleted splenocytes plus gastric lymph node cells from Tc immune mice were similarly protected (p < 0.05, Mann–Whitney U test compared with transfer of cells from naive mice). There was no significant difference in protection transferred by naive total or naive B cell–depleted cell populations (p = 0.22, Mann–Whitney U test). These results confirm our earlier results detected in the B cell KO model that B cells are not required for mucosal T. cruzi protection.

Since B cells were not critical for mucosal T. cruzi protection, we examined whether CD8⁺ T cells alone from Tc immune mice could transfer mucosal protection. Previous work has shown that CD8⁺ T cells from Tc immune mice could transfer protection against normally lethal systemic parasite challenges (4, 26). However, in order for CD8⁺ T cells to mediate mucosal protection, these cells must be able to migrate into the gastric tissue and provide effector functions. Highly purified CD8⁺ T cells (Supplemental Fig. 3A) from Tc immune and naive BALB/c mice were transferred into naive SCID mice 1 d prior to oral challenge. SCID mice that received Tc immune CD8⁺ T cells were significantly protected against oral T. cruzi challenge as compared with mice receiving naive control CD8⁺ T cells (Fig. 3D), despite the absence of detectable T. cruzi–specific serum IgG (Supplemental Fig. 3B). Thus, CD8⁺ T cells alone are able to confer mucosal T. cruzi protection.

We next investigated the ability of B cells to contribute to systemic T. cruzi protection in our model systems. Previous work by Kumar and Tarleton (16) showed that B cell KO mice could initially control parasitemia after systemic infection with T. cruzi but eventually succumbed to the infection. First, we verified whether B cell KO mice could develop systemic protection. B cell KO and WT BALB/c mice were orally infected with a low dose of T. cruzi twice, 2 mo apart, to generate Tc immune mice (Fig. 1A). These mice did not succumb to infection or require antiparasitic drugs (data not shown). In contrast, we have previously shown that mice deficient in IL-12, IFN-γ, or β₂-microglobulin are unable to survive low-dose parasite infections even with antiparasite drug intervention (19). We then challenged B cell KO and WT Tc immune mice systemically with BFT, which are more virulent and induce higher levels of parasitemia in naive mice (25, 27). B cell KO Tc immune mice survived significantly longer than did their naive cohorts, but they eventually succumbed to infection (Fig. 4A), developing increased parasitemia (Fig. 4B), weight loss (Fig. 4C), and decreased TS- and TSKd1-specific CD8⁺ T cell responses (Fig. 4D, 4E). Importantly, the frequencies and absolute numbers of IFN-γ–producing TSKd1-specific CD8⁺ T cells were significantly reduced. These results indicate that B cells are important for the maintenance of T. cruzi–specific, CD8⁺ T cell functional responses when systemic Ag levels are high. Previous work by our laboratory has shown that WT BALB/c mice require ∼5000 TSKd1-specific IFN-γ–producing CD8⁺ T cells to be protected from death (C.S. Eickhoff, N.L. Sullivan, and D.F. Hoft, manuscript in preparation). As shown in Fig. 4E, B cell KO Tc immune mice developed significantly lower numbers of TSKd1-specific IFN-γ–producing CD8⁺ T cells per spleen (738 ± 198) compared with WT Tc immune mice (11,374 ± 1,764). Thus, B cell KO Tc immune mice probably were not protected because they either did not develop or sustain sufficient numbers of IFN-γ–producing, TSKd1-specific CD8⁺ T cells. Consistent with the higher levels of parasitemia, Tc immune mice lacking B cells developed significantly higher levels of inflammation in cardiac and skeletal tissues, hallmarks of disease pathology induced by chronic T. cruzi infection (Fig. 4F, 4G). Overall, these results demonstrate that B cells play a critical role in protective T. cruzi systemic immunity.

B cells provide many important functions such as Ab production, Ag presentation, costimulation, and cytokine production. Thus, B cell deficiency not only results in the absence of Abs, but it also could affect the development of T cell responses. To determine whether T. cruzi–specific B cells are critical for systemic protection, we studied HEL-specific BCR transgenic mice. Advantages of using these mice are that in contrast to complete B cell KO mice, lymphoid tissue development and cytokine production are relatively normal. However, these mice cannot generate B cells that produce high-affinity T. cruzi–specific Abs, which could be important for both direct protective effects as well as enhanced Ag presentation. HEL and WT Tc immune mice were generated by repeated low-dose T. cruzi oral infections (Fig. 1A). Prior to and following systemic T. cruzi (BFT) challenge (at least 1 mo after the second oral infection), mice were assessed for T. cruzi–specific IgG Abs, parasitemia, weight loss, and death (Fig. 5).

HEL Tc immune sera did have some background reactivity in our recombinant TS-specific ELISA, but very little of this reactivity persisted in the presence of 8 M urea, demonstrating that minimal if any T. cruzi–specific, high-affinity Abs were present (Fig. 5A). In contrast, WT Tc immune sera had markedly higher reactivity with recombinant TS and much of this persisted despite 8 M urea treatment.

Even before the high-level T. cruzi BFT systemic challenge, parasitemia levels were increased in the HEL Tc immune mice (Fig. 5B). These results could be interpreted to suggest that the absence of high-affinity, T. cruzi–specific B cell responses resulted in a deficiency in protective mucosal memory immunity, despite our earlier results shown in Figs. 2 and 3 demonstrating that B cells were not important for either infection- or vaccine-induced mucosal T. cruzi–specific protective immunity. However, parasitemia is a systemic endpoint largely controlled by systemic immunity. After oral T. cruzi challenges, the parasite drains to the gastric lymph node and disseminates throughout the body, causing systemic infection. All Tc immune mice develop chronic systemic infection requiring systemic immunity to control parasitemia. The fact that HEL Tc immune mice had higher parasitemias prior to the high-level BFT systemic challenge indicates that these mice have profound defects in systemic immunity, resulting in the inability to control systemic parasite infection even after low-dose T. cruzi mucosal challenge.

Consistent with the results presented in Fig. 4, parasitemia levels increased further in HEL Tc immune mice after systemic challenge (Fig. 5B), resulting in increased weight loss (Fig. 5C) and death (Fig. 5D). These results further demonstrate that T. cruzi–specific B cells are critical for systemic T. cruzi protection.

In our Tc immune model, mice are chronically infected, making it difficult to study the development of memory CD8⁺ T cell responses in the absence of persistently high levels of T. cruzi Ag. Thus, we examined the development of CD8⁺TSKd1⁺ T cells in our TS immune memory model (heterologous TS DNA primed, adenovirus-TS boosted; Fig. 1C). As described earlier, B cells were depleted in vivo with anti-CD20 Ab (or isotype control) throughout the vaccination/treatment regimen. TS immune B cell–depleted and control mice developed similar levels of immunodominant CD8⁺TSKd1⁺ T cells detectable in blood (Fig. 6A–C). Thus, Ag-specific CD8⁺ T cells can be primed normally in B cell–depleted mice. These mice were then challenged systemically with T. cruzi and assessed for immune responses and protection. The percentages of the total CD8 populations that were CD8⁺TSKd1⁺ T cells expanded from ∼5.3 ± 1.0% at 4 mo after boost (Fig. 6C) to 38.5 ± 4.4% 15 d after the systemic T. cruzi challenge in isotype control–treated TS immune mice (Fig. 6D). Anti-CD20–treated TS immune mice had frequencies of CD8⁺TSKd1⁺ T cells similar to control mice prior to systemic challenge (4.62 ± 1.5%; Fig. 6C). However, these cells did not expand optimally after systemic T. cruzi challenge (Fig. 6D), resulting in significantly fewer CD8⁺TSKd1⁺ T cells after challenge than in control TS immune mice. These reduced numbers of CD8⁺TSKd1⁺ T cells in anti-CD20–treated TS immune mice were associated with increased parasitemia (Fig. 6E) and death (Fig. 6F). Although similar levels of immunodominant CD8⁺ T cells developed in response to TS vaccination in the absence of B cells, these CD8⁺ T cells were functionally deficient and failed to protect against T. cruzi systemic challenge. Thus, memory CD8⁺ T cells that develop in the absence of B cells were impaired in their ability to respond to and protect against virulent systemic T. cruzi challenges.

As we have shown in Figs. 4–6, in the absence of B cells, mice are unable to develop immunity that prevents death after virulent T. cruzi systemic challenge. Interestingly, these mice are able to develop CD8⁺ T cell responses and survive longer than do their naive cohorts (Fig. 4A). However, parasitemia increases over time in these B cell–deficient immune mice after virulent systemic challenge, ultimately resulting in death. These data suggested that although similar numbers of T. cruzi–specific CD8⁺ T cells could be induced in B cell–deficient mice, these CD8⁺ T cells were functionally deficient, especially when confronted with high Ag loads. To test this hypothesis, we first examined the ability of TSKd1-specific CD8⁺ T cells to expand and produce cytokines after ex vivo restimulation. B cell KO and WT Tc immune mice were challenged systemically with high doses of T. cruzi and immune responses were evaluated 30 d later. Splenocytes were labeled with CFSE and cocultured with either negative control APCs or APCs that were stably transduced with TS (27). Significant reductions in Ag-specific CD8⁺TSKd1⁺ T cell proliferative capacity (Fig. 7A) and cytokine production (IFN-γ and TNF-α; Fig. 7B, 7C) were detected in mice lacking B cells compared with WT mice. To evaluate CD8⁺TSKd1⁺ in vivo expansion, equivalent numbers of CD8⁺Thy1.2⁺TSKd1⁺-specific T cells from B cell KO and WT Tc immune mice (based on tetramer staining) were transferred into naive Thy1.1⁺ BALB/c mice with normal B cell numbers and function (Fig. 7D–H). Recipient mice were then challenged systemically 1 d later. CD8⁺Thy1.2⁺ T cells from B cell KO Tc immune mice failed to expand normally in vivo compared with WT infection-immune CD8⁺ T cells (Fig. 7E), resulting in a reduced ability to control parasitemia (Fig. 7F). These results identify intrinsic functional defects in T. cruzi–specific CD8⁺ T cells generated in the absence of B cells.

To further examine the mechanism responsible for CD8⁺ T cell dysfunction in B cell–deficient mice, we evaluated expression of the inhibitory receptors PD-1 and Lag-3 on tetramer⁺CD8⁺ T cells. There were significant increases in the percentages (Fig. 7G) and absolute numbers (Fig. 7H) of CD8⁺TSKd1⁺PD-1⁺Lag-3⁺ T cells in the spleens of B cell KO Tc immune mice compared with WT Tc immune mice. We also found significant increases in CD8⁺TSKd1⁺Lag-3⁺ T cells in the blood of TS-vaccinated and B cell–depleted mice 15 d after systemic T. cruzi challenge (Supplemental Fig. 4). Overall, these results demonstrate that B cells play a critical role for both the induction of optimally functional CD8⁺ T cell responses and the prevention of CD8⁺ T cell exhaustion.

Previous work by others has shown that passive serum transfers from Tc immune into naive mice can transfer systemic protection (11–15). We hypothesized that the lack of T. cruzi–specific Abs in B cell KO Tc immune mice resulted in increased tissue parasitism and death due to the inability of functionally exhausted CD8⁺ T cells to properly control T. cruzi infection. To test this hypothesis, we performed passive serum transfers. Prior to and after systemic T. cruzi challenge, B cell KO Tc immune mice received WT Tc immune serum, naive serum, or PBS. As shown in Fig. 8A, T. cruzi–specific Ab responses were present in B cell KO mice after transfer of WT Tc immune serum. The immune serum protected B cell KO mice from increased morbidity and mortality (Fig. 8B, 8C). This protection was associated with increased TSKd1-specific CD8⁺ T cell numbers (Fig. 8D) and total TS-specific and TSKd1-specific T cell function (Fig. 8E). Finally, Tc immune serum transfer resulted in reduced expression of the T cell exhaustion markers PD-1 and Lag-3 on TSKd1-specific CD8⁺ T cells (Fig. 8F). Overall, these results demonstrate that T. cruzi–specific Abs could prevent death and CD8⁺ T cell exhaustion.

In this study, we evaluated the importance of B cells in mucosal and systemic T. cruzi protection. We first examined whether B cells were important for mucosal T. cruzi protection. Previous results suggested that B cells would play a critical role in mucosal T. cruzi protection (17, 25). In fact, B cells have been shown to play important roles in mucosal protection against several mucosal pathogens such as helminthes (28, 29) and rotavirus (30, 31). Although we failed to identify a critical role for B cells in mucosal T. cruzi protection, this does not rule out the possibility that B cell responses could provide some level of mucosal protection. However, using both B cell–deficient mice and adoptive transfer studies, we showed that CD8⁺ T cells, but not B cells, are critical for mucosal T. cruzi immunity.

Next, we examined the importance of B cells in systemic protection induced by multiple low-dose parasite infections or highly potent vaccine regimens (TS DNA prime/adenovirus-TS boost). B cell–deficient mice infected orally with T. cruzi developed higher levels of parasitemia compared with WT mice, but this did not result in increased morbidity and mortality. These results indicate that B cells are not required when systemic Ag load is low after an oral infection. In contrast, we did show that B cell responses are absolutely required for protection after systemic challenge with the more virulent BFT that leads to higher systemic Ag loads. Tc immune B cell–deficient mice survived longer than did their naive cohorts after virulent systemic parasite challenge. However, Tc immune B cell–deficient mice developed increased parasitemia over time compared with WT Tc immune mice, resulting in significant morbidity and eventual death. We also determined that T. cruzi–specific B cells, not polyclonal nonspecific B cells, are critically important for T. cruzi systemic protection using HEL BCR transgenic mice. Decreased survival in B cell KO μMT^−/− mice after T. cruzi systemic challenge has been shown previously (16, 25). Additionally, our results are similar to recently published data demonstrating that after disseminating virus infection (lymphocytic choriomeningitis virus [LCMV]-t1b), μMT KO and anti-CD20–treated mice could not control viral replication, resulting in weight loss and decreased LCMV-specific CD8⁺ T cell functionality (32). Overall, these previous results combined with our new data suggest that memory CD8⁺ T cells primed in mice lacking T. cruzi–specific B cells are sufficient to control low levels of tissue parasitism, but they eventually become exhausted after systemic challenge due to prolonged exposure to high levels of Ag. These findings are consistent with previous data showing that high Ag load results in CD8⁺ T cell exhaustion (33–35). The role of B cells in preventing CD8⁺ T cell exhaustion is nicely shown when examining the CD8⁺ T cell response after TS vaccination. B cell–deficient mice developed similar frequencies of TSKd1-specific CD8⁺ T cells compared with WT mice after our heterologous TS vaccine regimen. These results confirm that B cells are not required for the initial priming of CD8⁺ T cell memory. However, after systemic challenge (resulting in high Ag load), CD8⁺TSKd1⁺ T cells from B cell–deficient TS immune mice did not expand as well as did CD8⁺ T cells primed in WT mice. This suboptimal CD8⁺TSKd1⁺ T cell secondary expansion capacity in B cell–deficient, TS-vaccinated mice was associated with increased parasitemia and death. Thus, although T. cruzi–specific CD8⁺ T cells can be primed in the absence of B cells, they become functionally deficient, severely impairing protective immunity when Ag level is high (Fig. 9).

To determine whether the defect in memory CD8⁺ T cell secondary expansion present in Tc immune B cell KO mice was due to poor priming or deficient support for re-expansion, we studied TSKd1-specific memory CD8⁺ T cells generated in B cell KO and WT Tc immune mice after subsequent transfer into immunocompetent BALB/c mice (Fig. 7D–H). Memory CD8⁺ T cells from B cell KO mice failed to expand in response to systemic parasite challenge after transfer into normal hosts containing polyclonal B cells, demonstrating an intrinsic defect in CD8⁺ T cell responses generated in mice lacking B cells. This failure of expansion resulted in significantly reduced protection. We then examined expression of the inhibitory receptors PD-1 and Lag-3 by CD8⁺TSKd1⁺ T cells after TS vaccination and T. cruzi infection. After systemic challenge of Tc immune and TS immune B cell–deficient mice, significant increases in CD8⁺TSKd1⁺ T cells expressing PD-1⁺ and Lag-3⁺ were observed. Overall, the lack of both direct ex vivo and in vivo expansion capacity, decreased cytokine production, increased inhibitory receptor expression, and reduced protection associated with memory T cells primed in the absence of B cells all indicate an important role for B cells in the prevention of CD8⁺ T cell exhaustion. To our knowledge, these results are the first to mechanistically demonstrate the importance of B cells for prevention of T. cruzi CD8⁺ T cell exhaustion.

Few studies have evaluated the role of B cells in the development of CD8⁺ T cell memory. During primary responses to an acute LCMV Armstrong infection, CD8⁺ T cell responses in mice lacking B cells were normal at early time points (36), but they were reduced at 154 d after infection (36, 37). In a vaccination model using Listeria monocytogenes, B cells were not required for the initial activation and expansion of Ag-specific CD8⁺ T cell responses (38). However, in the absence of B cells, a more profound contraction phase occurred leading to lower numbers of persisting Ag-specific memory CD8⁺ T cells (38). Another group showed that lack of B cells affected the long-term maintenance of CD8⁺ T cell memory (39). Previous T. cruzi studies have also indicated that B cells might be important in T. cruzi–specific memory T cell development, as lack of mature B cells (μMT^−/− mice) resulted in increased susceptibility to primary parasite infection and failure to develop robust vaccine-induced protection (16, 25). However, none of these previous studies investigated the specific role of CD8 T cell exhaustion in the absence of B cells. Our present results confirm these previous data and further demonstrate that B cells are critically important for both the development of fully functional CD8⁺ T cell responses as well as the prevention of CD8⁺ T cell exhaustion.

B cells may provide help for optimal CD8⁺ T cell memory responses by several mechanisms. First, B cells can present Ag and produce cytokines. Prior to high-level systemic challenge, the frequencies of CD8⁺TSKd1⁺ T cells were similar in Tc immune B cell KO and WT mice, suggesting that general T cell priming is not affected by the absence of B cells. However, CD8⁺TSKd1⁺ T cells induced in the absence of B cells and restimulated in the presence of high systemic Ag load in vivo were functionally deficient, producing reduced multifunctional cytokine responses (IFN-γ⁺TNF-α⁺). It has been reported previously that B cells amplify IFN-γ production by T cells in a contact-dependent manner (40). Our data also indicate that the production of Ag-specific Abs is another function of B cells critical for protective T. cruzi systemic immunity. Thus, in the absence of T. cruzi–specific IgG, increases in parasitemia and subsequent increases in overall Ag load occurred. Furthermore, passive transfer of T. cruzi immune sera known to reduce systemic parasite burden also reduced CD8⁺ T cell exhaustion and mortality in B cell KO Tc immune mice after virulent systemic challenge.

Although we did not study T. cruzi–specific CD4⁺ T cells, we do not think that CD4⁺ T cells are irrelevant for inducing protective CD8⁺ T cells. We and others have shown that CD4⁺ T cell help is required for effective CD8⁺ T cell memory (10, 41) and can rescue exhausted CD8⁺ T cells (42). Work recently published showed that CD4⁺ T cells are dysfunctional in the absence of B cells in the LCMV model (32). These data support the importance of our work, which, to our knowledge for the first time, demonstrates the principle that B cell responses are important for preventing the exhaustion of T cells directed against a major human pathogen.

Our work clearly demonstrates that previously activated memory CD8⁺ T cells alone can protect against mucosal T. cruzi challenge. However, after highly virulent T. cruzi systemic challenges, CD8⁺ T cells failed to provide optimal protection in the absence of B cells because CD8⁺ T cells became functionally exhausted. We further showed that HEL-specific B cells cannot prevent CD8⁺ T cell exhaustion, indicating that T. cruzi–specific B cells/Ab are needed for protection. Using both infection- and vaccine-induced memory immune models, we demonstrate that the absence of B cells leads to CD8⁺ T cell exhaustion in mice with high Ag load. Overall, these data show that B cells are important for two interrelated aspects of Ag-specific CD8⁺ T cell responses. First, Ag-specific B cells are critically important for the optimal development of multifunctional CD8⁺ T cell memory. Second, T. cruzi–specific Abs can directly reduce overall parasite load, preventing CD8⁺ T cell exhaustion. In conclusion, our results demonstrate that although B cells are not directly important for mucosal immunity, they do play critical roles for the induction and functional responses of CD8⁺ memory T cells, which are absolutely essential for both mucosal and systemic protective T. cruzi immunity. These studies identify an important role for B cells in the prevention of CD8⁺ T cell exhaustion, providing critical new information for vaccine design.

We thank Robert Dunn and Biogen Idec for providing the anti-CD20/isotype control Abs, Paul Garside at the University of Glasgow for providing the HEL BCR transgenic mice, Joy Eslick, Sherri Koehm, Erin Touchette, Kelly O’Shea, and Stella Hoft for technical assistance, Andreas Wieland at Emory University for critical reading of this manuscript, and the National Institutes of Health Tetramer Core Facility at Emory University for providing the TS^IYNVGQVSI/H-2K^d–APC tetramer.

The authors have no financial conflicts of interest.