Most parasite-specific CD8⁺ cells in Trypanosoma cruzi-infected chronic mice are down-regulated for T-cell receptor-αβ and CD8 molecules

Abstract

The present study shows that CD8⁺ T lymphocytes expressing low levels of T-cell receptor (TCR)αβ, CD8 and CD3 accumulate in the spleen, blood, peritoneum and liver, but not in the lymph nodes of mice chronically infected with Trypanosoma cruzi. Analysis of spleen lymphocytes reveals that most CD8^LOW TCR^LOW T cells have an experienced phenotype (CD44^HIGH CD62L^LOW and CD45RA,B,C^LOW). These cells have small size, lack activation markers such as CD69, CD25 and CD11b (Mac-1), and do not spontaneously secrete cytokines, suggesting they are at the resting state. When stimulated in vitro with T. cruzi-infected macrophages, TCR^LOW CD8^LOW T cells behave as parasite-specific memory cells, readily responding with interferon-γ (IFN-γ) production. Indeed, among parasite-activated CD8⁺ lymphocytes, IFN-γ production was mostly due to TCR^LOW CD8^LOW cells. Upon in vitro stimulation with anti-CD3/CD28 monoclonal antibodies, down-regulated cells produce IFN-γ and tumour necrosis factor-α, but not interleukin IL-10 or IL-4. Our results indicate that despite parasite persistence, most T. cruzi-specific experienced CD8⁺ cells are resting. Nevertheless, when encountering infected macrophages these cells differentiate to Tc1 effectors.

Introduction

In Chagas' disease, individuals who survive the acute phase of Trypanosoma cruzi infection develop a parasite-specific humoral and cellular immune response that efficiently reduces parasite levels in the tissues and blood. In spite of the protective role of the immune system, however, a small number of parasites persists during the host life and occasionally gain access to the blood. While the mechanisms underlying protection during the acute phase have been extensively analysed, it is poorly understood how the immune system deals with a life-long stimulation of effector cells.

T. cruzi parasites reproduce inside a wide variety of mammalian cells that are candidate targets for major histocompatibility complex (MHC) class I-restricted cell-mediated immune responses. The protective role of CD8⁺ T lymphocytes at the early phase of infection becomes evident from studies showing increased susceptibility of mice depleted of these cells or genetically deficient in CD8, TAP-1 or β₂ microglobulin,^1–6 as well as from adoptive transfer experiments using parasite-specific cytotoxic T lymphocyte (CTL) lines.⁷ At the chronic phase of the disease, CD8⁺ T cells constitute the predominant lymphocyte type in spleen,⁸ heart^9–11 and nervous tissue infiltrates.¹² Their in vivo contribution to intracellular parasite killing was demonstrated by the fact that long-term depletion of this population increases the number of heart parasite nests.¹¹ Nevertheless, the control of circulating trypomastigotes at this phase of the disease appears to be mediated by other effector mechanisms because parasitaemias do not recrudesce in chronic mice depleted of CD8⁺ T cells.⁵

Recently, we have shown that in chronically infected mice the number of CD8⁺ T lymphocytes in the spleen correlate with disease severity.⁸ This association corroborates the idea that CD8⁺ T lymphocytes play an important role at the chronic phase and raises the possibility that their increase could be a marker for disease progression. Aiming to evaluate the functional state of CD8⁺ T lymphocytes from chronic mice, here we characterized this population according to its tissue distribution, surface markers and cytokine production.

Materials and methods

Mice and parasites

Female A/J mice from our breeding colony at the Biotério de Animais Isogênicos, ICB, USP were used in most of the experiments. Female BALB/c, C3H/HePas or C57BL/6 mice were eventually used. Mice were infected i.p. with 1000 blood trypomastigotes of the reticulotropic Y strain of Trypanosoma cruzi, maintained by in vivo weekly passages in A/J mice. Because under these conditions the animals died around day 16 post-infection, for chronic phase studies mice were treated on day 8 with a single oral dose of benznidazole (Rochagan, Roche, Rio de Janeiro, Brazil), 1 g/kg body weight. This treatment allowed control of parasitaemia and animal survival, but not the cure from infection. In some experiments, mice were infected with 1000 blood trypomastigotes of the CL strain (maintained by biweekly passages in A/J mice) or with 10⁶ tissue culture trypomastigotes of the Silvio X10/4 strain. Parasitaemias were determined by microscopic examination of 5 µl blood samples obtained from the tail vein.

Tissue-culture trypomastigotes

Tissue culture trypomastigotes of the Y and Silvio X10/4 strains were obtained from LLCMK2-infected cultures as described.¹³

Cell suspensions

Spleen, lymph node (LN), blood, peritoneal exudate (PEC) and liver cells were obtained from acute or chronic infected animals. Peripheral blood lymphocytes (PBL) were obtained from the interface of a 60% Percoll (Sigma-Aldrich Co., St Louis, MO) gradient layered with heparinized blood. Liver lymphocytes were purified from the 35–60% interface of Percoll gradients layered with cell suspensions prepared from the livers, dissected after perfusion from the left heart ventricle with 30 ml of phosphate-buffered saline (PBS).

Phenotypic analysis of lymphocytes

Phenotypic analysis of lymphocyte subpopulations were assessed by three-colour fluorescence-activated cell sorting (FACS), using an Facscalibur Cytometer (Becton Dickinson, Mountain View, CA), after in vitro incubation with fluorescein isothiocyanate (FITC)-, phycoerythrin (PE)-or Cy-chrome-labelled monoclonal antibodies (mAb). Labelled mAb anti-CD4 (clone H129.19), anti-CD8 (clone 53-6.7), anti-B220 (clone RA3-6B2), anti-CD3 (clone 145-2C11), anti-T-cell receptor (TCR)αβ (clone H57-597), anti-CD45RA (clone 14.8), anti-CD45RB (clone 16A), anti-CD45RC (clone DNL-1.9), anti-Thy-1.2 (clone 30-H12), anti-class I (H-2K^k; clone AF3-12.1), anti-CD5 (clone 53-7.3), anti-CD69 (clone H1.2F3), anti-CD25 (clone 7D4) and anti-CD11b (Mac-1; clone M1/70) were purchased from PharMingen (Ambriex, São Paulo). Unlabelled anti-CD62L (clone MEL-14) and anti-CD44 (clone KM-201) were a kind gift from Dr L. Aroeira and Dr M. Rodrigues, respectively. When using these antibodies, biotin-labelled anti-rat immunoglobulin G (IgG; Southern Biotechnology Associates Inc., Birmingham, AL) was used as a second step reagent, followed by streptavidin–FITC or streptavidin–PE (PharMingen). Cell viability of the different subsets was estimated by incorporation of propidium iodine (Sigma-Aldrich Co.) and 7-AAD (7-amino-actinomycin D; Sigma-Aldrich Co., St Louis, MO).

Macrophage colony-stimulating factor (M-CSF) grown macrophages

Single-cell suspensions of A/J bone marrow cells were cultured in six-well culture dishes at 2 × 10⁶ cells/ml in RPMI medium (Sigma-Aldrich Co.), supplemented with 2 mm glutamine, 63 mg/l penicillin, 100 mg/l streptomycin sulphate, 50 µm 2-mercaptoethanol (2-ME) and 10% fetal calf serum (FCS) containing 30% (v/v) L929 cell-conditioned medium as a source of M-CSF.¹⁴ The cultures were rinsed and medium was replenished every 3 days in order to remove non-adherent cells. After day 7 of culture, the cultures consisted of adherent macrophages.

In vitro macrophage infection

Cultures of M-CSF-grown macrophages (M-MØ) were dispensed (2 × 10⁵ cells) into 96-well culture dishes (Costar, New York, NY), and infected with LLCMK2-derived T. cruzi culture trypomastigotes (Y strain) at 4 : 1 parasite : cell ratio for 6 hr at 37° in 5% CO₂. After that, cultures were washed to remove extracellular parasites and they were further incubated in complete RPMI medium for 18 hr until the incubation with splenocytes.

Intracellular detection of cytokines

Spleen cells were cultured for 5 hr with monensin-containing Golgistop (PharMingen) in the presence or absence of plate-bound anti-CD3 (10 µg/ml; clone 145–2C11; PharMingen) and soluble anti-CD28 (2 µg/ml; clone 37·51; PharMingen) antibodies, or in the presence of T. cruzi-infected or control noninfected M-MØ. Intracellular cytokine levels were measured by FACS using the Cytofix/Cytoperm Plus Kit (PharMingen), according to the manufacturer's instructions. PE-labelled monoclonal antibodies to anti-interferon-γ (IFN-γ) (clone XMG1·2), anti-interleukin IL-10 (clone JES5-16E3), anti-IL-4 (clone 11B11) and anti-tumour necrosis factor-α (TNF-α; clone MP6-XT22) were purchased from PharMingen.

Statistics

Comparisons between two groups were made by the Student's t-test. Multiple comparison analysis of IFN-γ-producing cells after stimulation with T. cruzi-infected or non-infected macrophages were made by the Tukey HSD test.

Results and discussion

CD8⁺ cells accumulate in the blood, peritoneal cavity and liver, but not in lymph nodes of chronic mice

Susceptible A/J mice infected i.p. with 1000 blood trypomastigotes of the reticulotropic Y strain of T. cruzi displayed a first peak of parasitaemia at day 9 post-infection, and a second peak, from day 14 onwards. Unless treated, these animals died around day 16 post-infection. Treatment with benznidazole on day 8 of infection allowed control of parasitaemia and animal survival, but not the cure from infection. Along the acute phase, spleen weight and cellularity progressively augmented (six and four times over the controls) with corresponding increase in the numbers of CD4⁺ and CD8⁺ T cells and B (B220⁺) cells (Fig. 1a). Later on, at the early chronic phase, while spleen cell numbers decreased towards control values, as did the numbers of B and CD4⁺ T cells, the CD8⁺ population kept expanding. In consequence, at day 64 of infection, CD8⁺ spleen T cell frequencies largely exceeded those of controls. Accumulation of CD8⁺ T cells was also observed in the blood, peritoneal cavity (port of entry of the parasite in this experimental model) and liver of chronic mice, but not in the lymph nodes (Fig. 1d). The wide distribution of this enlarged CD8⁺ population suggests that these lymphocytes play an important role at the chronic phase of the disease.

Figure 1.

CD8⁺ cells accumulate in T. cruzi-infected chronic mice. Total numbers of CD4⁺ (a), CD8⁺ (b) and B220⁺ (c) lymphocytes in the spleen of mice at different time points from the onset of infection. Infected mice (squares), non-infected benznidazole-treated control mice (triangles) and normal control mice (circles). (d) Increase in the number of CD8⁺ cells in different sites of mice infected for 65 days (ratio of total number of CD8⁺ cells in chronic mice/total number of CD8⁺ cells in age-matched noninfected benznidazole-treated control mice). Each point or bar represents the mean±standard deviation of individual values from five mice.

Most CD8⁺ cells in the spleen of chronic mice are down-regulated for receptors and coreceptors

Analysis of CD8⁺ T lymphocytes from the spleen of chronic mice revealed that a considerable part of these cells expressed low levels of TCRαβ, CD8 and CD3 on the surface (Fig. 2a). Down-regulation of the three molecules occurred in the same subpopulation. Cells with the CD8^LOW TCR^LOW CD3^LOW phenotype coexisted in the spleen of chronic mice with CD8⁺ cells expressing normal levels of receptors and coreceptors (CD8^HIGH TCR^HIGH CD3^HIGH). CD8^LOW TCR^LOW T cells accumulated in the spleen from day 11 of infection, progressively increasing up to day 60–100 when they represented more than 50% of total CD8⁺ T lymphocytes (Fig. 2b). Down-regulated cells persisted in the spleen at least until day 200 of infection, despite that, at this time, the numbers of total CD8⁺ T cells approached those of age-matched controls. Different from this population, splenic CD4⁺ T cells of chronic mice expressed normal levels of CD4 and CD3 molecules on the membrane (data not shown).

Figure 2.

Phenotypic characterization of CD8⁺ spleen cells from T. cruzi-infected chronic mice. Spleen cells from T. cruzi-infected chronic mice and from age-matched, non-infected, benznidazole-treated controls were screened for expression of different surface molecules. (a) Histograms of TCRαβ, CD8 and CD3 expression by gated CD8⁺ cells of chronic mice at day 75 of infection (solid line) and control mice (dashed line). (b) Frequency of CD8^LOW TCR^LOW T cells of chronic mice at different time points from the onset of infection (solid line) and control mice (dashed line). Each point represents the mean±standard deviation of individual values from five mice. (c) Expression of CD45RA, CD45RB and CD45RC isoforms by CD8⁺ T cells from control and chronic mice (day 70 of infection).

The most likely explanation for TCR down-regulation is that stimulation through MHC/peptide complexes led to receptor internalization. In this regard, it has been shown that a large number of TCR are serially engaged, triggered and down-regulated after T-cell priming.^15,16 Because of this, we can envisage that at the chronic phase of the disease a large fraction of CD8⁺ T cells must have engaged TCR molecules. Nonetheless, we can not disregard that other signals besides those mediated by the TCR might be required to sustain receptor down-regulation. In this context, it is worth noticing that TCR expression levels can also be regulated by the balance between protein kinase C and serine–threonine protein phosphatase activities, an alternative pathway for receptor down-regulation that could be influenced by cytokines.¹⁷ On the other hand, it is intriguing that receptor–coreceptor down-regulation in T. cruzi-infected chronic mice affected the CD8⁺, but not the CD4⁺ T-cell compartment. This observation suggests that the two subsets are receiving different signals, or alternatively, they differ in their sensitivity to down-regulation or capacity for receptor–coreceptor re-expression.

Down-regulation of receptor–coreceptors was not restricted to a particular T. cruzi–mouse strain combination. Thus, high numbers of CD8⁺ T lymphocytes expressing low levels of CD8, TCRαβ and CD3 were observed in A/J, C3H/HePas, C57BL/6 or BALB/c mice chronically infected with Y parasites and in A/J mice chronically infected with CL (myotropic) or Sylvio X10/4 parasites (data not shown). Our results extend and confirm previous observations showing reduction of CD8 expression in the course of T. cruzi infections with the metacyclic DM28c clone.¹⁸ Down-regulation of CD28, another surface receptor involved in T-cell activation, has been also reported in circulating CD8⁺ lymphocytes from human chagasic patients.¹⁹

Down-regulated CD8⁺ T cells are resting and display a memory–effector phenotype

Analysis of CD45, a tyrosine phosphatase essential for T cell activation, showed that CD8^LOW TCR^LOW T cells in the spleen of chronic mice displayed lower expression of CD45RA, CD45RB, and CD45RC isoforms, compared to CD8^HIGH TCR^HIGH T cells from the same mice, or to CD8⁺ T cells from control mice (Fig. 2c). Indeed, the great majority of spleen CD8⁺ CD45R^LOW cells expressed on their membranes low levels of CD8, CD3 and TCR. Reduction of CD45RA, B or C expression is known to result from switch to CD45 low-molecular-weight isoforms that easily associate with the TCR complex.^20–22 This occurs in memory and effector CD8⁺ T lymphocytes and may facilitate TCR signalling, allowing activation by low levels of antigen and costimulation.²³

To further characterize the CD8^LOW TCR^LOW spleen cell subset, we analysed the expression of several membrane molecules that could indicate a role for this subpopulation at the chronic phase of the disease. As shown in Fig. 3(a), the great majority of CD8^LOW CD45R^LOW T lymphocytes lacked activation markers such as CD69, CD25 (IL-2Rα chain) or CD11b (Mac-1, β₂ integrin), a homing molecule to inflammatory sites expressed in virus-specific CTL and memory cells.²⁴ These results suggest that CD8^LOW TCR^LOW cells were at the resting state. This conclusion is supported by the very low frequency of large cells in this subset as determined by forward scatter analysis (Fig. 3b). In contrast, the frequency of large cells was increased in the CD8^HIGH TCR^HIGH subset of mice at the early chronic phase of infection (day 26).

Figure 3.

CD8^LOW TCR^LOW T cells in the spleen of T. cruzi-infected chronic mice are experienced and resting. CD8⁺ spleen cells from T. cruzi-infected chronic mice and from age-matched, non-infected, benznidazole-treated controls were analysed for expression of different surface molecules and frequency of large cells. (a) Expression of CD69, CD11b (Mac-1), Thy-1, H-2K, CD62L, CD44, CD25 and CD5 by gated CD8⁺ cells from control and chronic mice (day 65 of infection), according to expression of CD45RC or CD45RB. (b) Frequency of large cells (estimated by forward scatter analysis) among gated CD8⁺ cells from untreated control mice (diagonal hatched bars), benznidazole-treated control mice (horizontal hatched bars), and gated CD8^HIGH TCR^HIGH cells (white bars) or CD8^LOW TCR^LOW cells (black bars) from chronic mice. Each bar represents the mean±standard deviation of individual values from five mice.

CD8^LOW CD45R^LOW T lymphocytes from infected mice expressed slightly lower surface levels of Thy-1, class I MHC (H-2K^k) and CD5 molecules (Fig. 3a). Analysis of CD62L (L-selectin) and CD44 (Pgp-1), adhesion molecules that discriminate naive, memory and effector T lymphocytes, revealed that the great majority of CD8^LOW TCR^LOW T cells displayed an experienced phenotype (CD44^HIGH CD62L^LOW). In contrast, CD8^HIGH TCR^HIGH T cells showed a heterogeneous expression of CD44 on their membranes, with 65% of them exhibiting a naive cell phenotype (CD44^LOW CD62L^HIGH). Simultaneous down-regulation of CD8 and TCR molecules after TCR engagement has been observed in vitro in memory or activated cells, but not in naive cells,²⁵ supporting the notion that CD8^LOW TCR^LOW T lymphocytes observed in chronic mice are experienced cells. FACS analysis of propidium iodide or 7-amino-actinomycin D stained cells revealed that less than 3% of CD8^LOW TCR^LOW lymphocytes were dead (data not shown).

Down-regulation of receptors and coreceptors also occurs in CD8⁺ T cells from blood, peritoneum and liver, but not from LN of chronic mice

Analysis of sites other than the spleen showed that CD8^LOW CD45R^LOW cells were present at very high numbers in the blood, peritoneum and liver of chronic mice, but not in the lymph nodes (Fig. 4). Among PBL, the vast majority of CD45R^LOW CD8⁺ cells (which made up 95·5% of CD8⁺ T cells) expressed low levels of CD8, demonstrating that down-regulated cells circulate through the blood. In the peritoneal cavity and liver, however, a third CD8⁺ subpopulation expressing low levels of CD45R, but relatively normal levels of CD8 (CD8^HIGH CD45R^LOW) was also observed. At these sites, the great majority of CD8⁺ T cells (90·8% and 75·7% for the peritoneal cavity and liver, respectively) were CD45R^LOW, but only half of them was down-regulated for CD8 expression. In these studies, lymphocytes expressing low levels of CD8 were also down-regulated for CD3 and TCR molecules (data not shown). Interestingly, the inguinal lymph nodes of chronic mice contained negligible numbers of CD8 down-regulated lymphocytes, an observation that correlates with the absence of an increase in total CD8⁺ cell numbers in this lymphoid organ (Fig. 1d). The fact that CD8^LOW cells circulate in the bloodstream but can not be found in the lymph nodes could be explained by their low CD62L expression that may hinder adhesion to high endothelial venules. Forward scatter analysis of CD8^LOW lymphocytes from the blood, peritoneum and liver confirmed that they were small cells.

Figure 4.

CD8⁺ subsets in different sites of T. cruzi-infected chronic mice. Distribution of CD8⁺ subsets was estimated by expression of CD8 and CD45RC molecules in different sites of chronic mice (day 75 of infection) and age-matched non-infected benznidazole-treated control mice. Numbers in each region indicate the percentage of each population in relation to the total number of CD8⁺ cells. Representative dot plots of each group are shown (three to four mice per group)

In non-infected controls, the frequency of down-regulated CD8⁺ cells was very low, except in the liver and among PBL, where they accounted for 14·3% and 16·0% of all CD8⁺ cells, respectively (Fig. 4).

Down-regulated CD8⁺ T cells are not secreting IFN-γ, IL-10, IL-4 and TNF-α in vivo, but they can be activated for cytokine production after in vitro stimulation

To clarify whether CD8^LOW TCR^LOW cells from chronic mice displayed effector functions or could differentiate into effector cells, we evaluated intracellular production of IFN-γ, IL-10, IL-4 and TNF-α. Spleen cells were cultured for 5 hr with Golgistop in the presence or absence of plate-bound anti-CD3 and soluble anti-CD28 antibodies, a culture condition that did not change the CD8⁺ cell subset distribution of chronic mice (data not shown). Spleen cells from acutely T. cruzi-infected mice were included as positive controls. At the chronic phase of the infection, the spontaneous production of IFN-γ, IL-10, IL-4 or TNF-α by CD8^LOW CD45R^LOW and CD8^HIGH CD45R^HIGH spleen cells was very low, similar to that of non-infected controls (Fig. 5). Nevertheless, in the acute phase of the infection, spontaneous production of IFN-γ, IL-10, IL-4 and TNF-α could be clearly detected. After stimulation with anti-CD3 and anti-CD28, both CD8⁺ subsets from chronic mice displayed a marked increase in the number of IFN-γ and TNF-α producers, but not of IL-4 and IL-10 producers. This was also observed for CD8⁺ cells from control mice and acutely T. cruzi-infected mice. The low spontaneous production of cytokines by CD8^LOW CD45R^LOW T lymphocytes is consistent with the view that these cells are resting in the chronic animal. In the other hand, their potential for IFN-γ and TNF-α production after in vitro stimulation suggests that these cells display a memory Tc1 phenotype, rather than being tolerant or anergic.

Figure 5.

Cytokine production by CD8^LOW TCR^LOW and CD8^HIGH TCR^HIGH spleen cells from T. cruzi-infected chronic mice. Spontaneous and anti-CD3 plus anti-CD28 induced production of IFN-γ, IL-10, IL-4 and TNF-α by gated CD8⁺ spleen cells (shown according to their CD45RB expression), from T. cruzi-infected acute mice (9 days post-infection), chronic mice (120 days post-infection) and age-matched non-infected benznidazole-treated control mice cultured for 5 hr in the presence of Golgistop. Representative dot plots of each group are shown (three to four mice per group)

Down-regulated CD8⁺ cells respond to in vitro stimulation with T. cruzi-infected macrophages

To clarify if down-regulated CD8⁺ T cells were parasite-specific, we measured the intracellular IFN-γ production by splenocytes from chronic and control mice after 5 hr culture with T. cruzi-infected or uninfected macrophages, in the presence of Golgistop. As shown in Fig. 6, CD8⁺ T cells from chronic mice responded to infected macrophages with production of IFN-γ, the frequency of producing cells being significantly higher than those obtained in the presence of noninfected macrophages. Most of the increase in IFN-γ production was restricted to CD8^LOW TCR^LOW T cells, but a small increase was also detected in the CD8^HIGH TCR^HIGH subset. Thus, for the CD8^LOW TCR^LOW subset, the frequency of IFN-γ producers raised from 1·7 to 7·9%, while for CD8^HIGH TCR^HIGH T cells, the increase was from 1·0 to 2·7%. Cells from control mice did not respond to infected macrophages. Similar results were obtained when CD8⁺ cells were gated according to TCR expression (data not shown).

Figure 6.

IFN-γ production by CD8^LOW TCR^LOW and CD8^HIGH TCR^HIGH spleen cells from T. cruzi-infected chronic mice after in vitro stimulation with T. cruzi-infected macrophages. Chronic mice (80 days postinfection) or age-matched benznidazole-treated controls were stimulated for 5 hr with Golgistop in the presence of T. cruzi-infected or non-infected M-CSF-grown macrophages (M-MØ). Intracellular IFN-γ production of gated CD8⁺ cells was analysed by FACS. Numbers in each quadrant indicate the percentage of each population in relation to the total number of CD8⁺ cells, while numbers in parenthesis indicate the percentage of IFN-γ⁺ cells in relation to the corresponding CD8^LOW or CD8^HIGH subsets. Representative dot plots of each group are shown (three to four mice per group). P < 0·01 for IFN-γ⁺ cells in the CD8^LOW subset from chronic mice in relation to all the other groups.

Taken together, our results show that most T. cruzi-specific experienced CD8⁺ T cells from chronic mice are down-regulated for receptors and coreceptors and, despite parasite persistence, display a resting phenotype without loosing their potential for IFN-γ production. Assuming the possibility that TCR down-regulation reflects a sustained receptor engagement,¹⁶ the resting phenotype displayed by CD8^LOW TCR^LOW T cells should not be attributed to lack of stimulation. In this case, it is intriguing how these cells would persist in a resting state despite TCR engagement. The response to this riddle could be in down-regulation itself, a possibility supported by in vitro data showing that the capacity of T cells to attain an activation threshold is severely compromised by TCR down-regulation.²⁶ Thus, in a situation of increased activation threshold, the persistent low antigen levels of chronic mice would be insufficient to induce proliferation and cytokine production by CD8^LOW TCR^LOW T cells, but could be sufficient to maintain these cells in the down-regulated state. Activation of CD8^LOW TCR^LOW T cells would follow, nevertheless, in the infrequent circumstances of facing a cell, such as an infected macrophage, that present in their membranes high levels of class I-parasite peptide complexes.

Accumulation of TCR^LOW CD8⁺ T cells has been reported in diverse tolerant states induced by continuous antigen stimulation, supporting the notion that sustained down-regulation of receptors and coreceptors is an adaptive mechanism that limits antigen-specific responses. Thus, down-regulation of TCR by CD8⁺ T lymphocytes was observed (i) in anti-H-Y TCR transgenic male mice, where the few peripheral CD8⁺ cells escaping negative selection expressed low levels of TCR and CD8,²⁷ (ii) in animals surviving kidney allografts for more than 3 weeks,²⁸ and (iii) in a TCR and antigen double transgenic model, in which the membrane expression of the antigen was exogenously manipulated.²⁹ In our model, by contrast, we clearly show that down-regulated cells are not anergic, suggesting that lowered expression of receptors and coreceptors does not inevitably leads to a tolerant state. CD8⁺ T cells with low TCRαβ expression exist in the normal mouse liver and, in more reduced levels, in other lymphoid organs.^30,31 These cells accumulate with age and produce high levels of IFN-γ, but not IL-4, upon in vitro stimulation with anti-CD3. While these cells may stand for the small number of CD8^LOWTCR^LOW lymphocytes we have observed in the liver and blood of non-infected control mice, their relationship with CD8^LOW TCR^LOW from T. cruzi-infected chronic mice remains as an interesting open possibility.

Our results constitute the first description of a simultaneous TCR and coreceptor down-regulation in the course of a parasitic infection. Further characterization of these lymphocytes may help us to understand their origin, as well as their participation in protection and pathology of Chagas'disease.