Differential Effects of Cytolytic T Cell Subsets on Intracellular Infection

See allHide authors and affiliations

Science  13 Jun 1997:
Vol. 276, Issue 5319, pp. 1684-1687
DOI: 10.1126/science.276.5319.1684


In analyzing mechanisms of protection against intracellular infections, a series of human CD1-restricted T cell lines of two distinct phenotypes were derived. Both CD4CD8 (double-negative) T cells and CD8+ T cells efficiently lysed macrophages infected withMycobacterium tuberculosis. The cytotoxicity of CD4CD8 T cells was mediated by Fas-FasL interaction and had no effect on the viability of the mycobacteria. The CD8+ T cells lysed infected macrophages by a Fas-independent, granule-dependent mechanism that resulted in killing of bacteria. These data indicate that two phenotypically distinct subsets of human cytolytic T lymphocytes use different mechanisms to kill infected cells and contribute in different ways to host defense against intracellular infection.

Effective immunity to intracellular bacterial infection often requires the lysis of infected cells as well as killing of the invading pathogen. A possible role for cytolytic T lymphocytes (CTLs) in protection against M. tuberculosis has been suggested by experiments in mice bearing a disruption in the β2-microglobulin gene. These mice are unable to express major histocompatibility complex (MHC) class I or class I–like molecules or to generate CTLs and were shown to be highly susceptible to infection (1). Despite numerous studies of CD4+ T cell responses and cytokine production in tuberculosis, there remain only a few reports of CD8+ CTLs that recognize mycobacterial antigens (2). This paradox led us to investigate whether other antigen-presenting systems could be essential for generation of M. tuberculosis–specific CTLs. CD1 is an MHC-like surface molecule with a unique ability to process and present nonpeptide antigens to T cells, including mycobacterial lipids (3, 4). We examined whether CD1-restricted CTLs have the capacity to recognize and lyse M. tuberculosis–infected macrophages.

CD1-restricted T cells were derived from patients with active tuberculosis as well as healthy donors (5). All of these T cell lines recognized M. tuberculosis lipid and lipoglycan antigens in a CD1b-restricted manner as assessed by antigen-specific T cell proliferation and interferon-γ (IFN-γ) secretion. T cells were either CD4CD8 (double negative, DN) or CD8+ and expressed αβ T cell receptors, consistent with our previous findings (3, 4, 6). We investigated whether CD1-restricted T cells recognized antigen-presenting cells (APCs) harboring live mycobacteria. CD1+ macrophages were infected with virulent M. tuberculosis with 90% efficiency, such that there were approximately three bacteria per macrophage (7). All DN and CD8+ T cell lines examined efficiently lysed infected macrophages in a dose-dependent manner (Fig. 1A). The restriction and specificity were shown by the inhibition of CTL-mediated lysis of infected targets and release of IFN-γ by antibody to CD1b (Fig. 1B). CTLs did not lyse uninfected CD1+ macrophages.

Figure 1

Cytotoxicity of CD1-restricted CTLs against macrophages infected with virulent M. tuberculosis. The cytotoxic response of (A) CD8+ line (CD8.TX) and (B) DN line (DN.PT, E:T = 10:1) against infected macrophages was measured in a51Cr-release assay in the presence or absence of blocking antibodies to CD1 (anti-CD1) (24). The results shown are representative of one out of three independent experiments, each performed in triplicate. Error bars correspond to the SEM.

CTLs lyse targets by two pathways, the exocytosis of granules containing perforin and granzymes and the interaction of Fas ligand on the CTL with Fas on the target cell (8). The mechanisms operate independently: For example, mice with a disrupted perforin gene retain the ability to exert Fas-FasL–dependent T cell lysis, but the biological roles and contribution to immunity of each remains unresolved. Because M. tuberculosis–infected macrophages were killed by two phenotypic subsets of cytotoxic T cells, we sought to clarify the mechanisms of lysis, specifically the relative importance of killing by Fas-FasL interaction and by the degranulation mechanism. The cytotoxicity mediated by two DN, CD1-restricted CTL lines was markedly inhibited by blocking antibodies to Fas or to FasL (Fig. 2A). In contrast, the cytotoxicity of two CD8+ CD1-restricted CTL lines was not affected by blocking of Fas or FasL (Fig. 2B). We also determined the contribution of the granule-dependent pathway to the target cell lysis. Strontium ions (Sr2+), which release histamine from mast cells by inducing granular degranulation, also induce degranulation of cytotoxic lymphocytes (9), thereby transiently inhibiting lytic activity. This effect was used to determine the extent to which the granule-dependent pathway participated in killing M. tuberculosis–infected macrophages (10). Preincubation with Sr2+ selectively inhibited the cytotoxicity of the CD8+, but not DN, CD1-restricted CTLs (Fig. 2, A and B). Granzyme A, characteristic of cytotoxic granules, was detected in Sr2+-induced supernatants of CD8+, but not DN, T cells (Fig. 2C). The capacity of lymphocytes to proliferate and release IFN-γ upon antigen-specific activation was not affected by treatment with Sr2+ (11). The differential ability of antibodies to Fas-FasL or of Sr2+ to inhibit CTL activity was not dependent on the level of killing (Fig. 2, A and B).

Figure 2

Distinct mechanisms of cytotoxicity of DN and CD8+ CTLs. Cytotoxicity of (A) DN (DN.PB and DN.OR) or (B) CD8+ (CD8.TX and CD8.1) CTLs against antigen-pulsed macrophages was determined in the presence or absence of blocking antibodies to FasL (5 μg/ml) or Fas (1 μg/ml) or after initial treatment of the CTLs with Sr2+ (25). The E:T ratio was 10:1. The result shown is representative of three independent experiments. Error bars correspond to the SD. (C) Release of BLT-esterase by CD1-restricted CTLs induced by treatment with Sr2+ (26). The data shown are representative of three independent experiments, each performed in triplicate. Data are given as the absorbance at 405 nm ± SD. (D) Expression of FasL and perforin mRNA by DN and CD8+ CTLs. CTLs (2 × 105) were stimulated with antigen for 12 hours, and total RNA was isolated as described (27). cDNA was synthesized and standardized to yield similar amounts of CD3δ PCR product within the linear range of amplification. cDNA specific for FasL (28) and perforin (29) was amplified by PCR and visualized by autoradiography (28). Lane 1: DN.OR; lane 2: DN1; lane 3: DN.LDN4; lane 4: CD8.1; lane 5: CD8.2; and lane 6: CD8.FP1.

A critical component of lymphocyte cytotoxic granules is perforin, which polymerizes on the target cell membrane after antigen activation and induces a nonselective pore that may be responsible for target cell lysis (12). Using reverse transcriptase–polymerase chain reaction (RT-PCR), we detected induction of perforin mRNA in all three CD8+ CTL lines examined, but not in three DN CTL lines (Fig. 2D). In contrast, mRNA for FasL was detected in stimulated DN CTL lines but not in the CD8+ lines.

The existence of two populations of human CTLs, differentiated by phenotype and by mechanism of cytotoxicity, was confirmed in a larger group of CTLs. Five DN CTL lines independently derived from different donors, all CD1-restricted, killed targets by the Fas-FasL pathway, with little contribution from the granule-dependent mechanism (Fig.3). Conversely, the cytotoxicity of three CD8+ CD1-restricted CTL lines was granule-dependent. In addition, the killing by two classical CD8+ MHC class I–restricted CTL lines specific for influenza peptide was almost completely dependent on cytotoxic granules (13).

Figure 3

Subset-dependent mechanisms of cytotoxicity. The ability of blocking antibodies to FasL or Fas and of Sr2+ to inhibit antigen-specific cytotoxicity of CD1- or MHC class I–restricted CTLs at an E:T ratio of 10:1 was determined. Inhibition of cytotoxicity was calculated as [(specific cytotoxicity in the absence of inhibitor − specific cytotoxicity in the presence of inhibitor)]/(specific cytotoxicity in the absence of inhibitor).

The question remains as to why the immune response has two virtually independent modes of cytotoxic responses. Lysis of infected macrophages will release intracellular bacteria, thus reducing the reservoir of infected cells. The bacteria will be dispersed and taken up at low multiplicities of infection (MOIs) by activated infiltrating macrophages, which can kill them (14). In addition, the process of lysing the infected target cell may directly or indirectly result in the death of the bacteria. To determine whether CD1-restricted T cell activation results in killing of intracellular mycobacteria, we cocultured CTL lines with M. tuberculosis–infected CD1+ cells and measured bacterial viability after 18 hours. Whereas four DN, CD1-restricted T cell lines had no effect on the number of colony-forming units (CFUs) of virulent M. tuberculosis, both CD8+CD1-restricted T cell lines examined reduced the number of CFUs by 35 to 50% (Fig. 4). In addition, two human influenza peptide–specific CD8+ CTL lines that cause lysis solely by a granule-dependent mechanism reduced the number of viable mycobacteria by lysing infected macrophages that had been simultaneously pulsed with influenza peptide. Although the percentage reduction of CFUs was within an order of magnitude, M. tuberculosis infection in vivo is slow and protracted, and the time of in vitro assay was only 18 hours, so that a cumulative antimicrobial effect mediated by these T cells over time could have a profound effect on the number of bacilli during the course of infection.

Figure 4

Perforin, but not Fas-FasL–mediated lysis, inhibits the growth of M. tuberculosis. Macrophages infected with live M. tuberculosis were coincubated with DN or CD8+ CTLs at an E:T ratio of 10:1. After 18 hours, cells were lysed with saponin to release intracellular bacteria. For the determination of mycobacterial viability, fivefold dilutions of the lysate were plated in duplicate on 7H11 agar plates. CFUs were counted after a 3-week incubation. The results shown are representative of two independent experiments. Error bars correspond to the SEM.

These data and a recent study of a murine model (15) suggest that the two defined mechanisms of cytotoxicity are associated with distinct phenotypic T cell subsets, yet have differential effects on microbial immunity. Consistent with the findings that Fas-FasL interactions appear to be most relevant to lysis of cells of the immune system itself, this mechanism may function primarily in immune regulation in vivo, particularly in eliminating antigen-expressing APCs, thereby down-regulating immune-mediated tissue injury (16). In contrast, the ability of CD8+ CTLs both to lyse infected cells by the granule-dependent mechanism and to kill intracellular M. tuberculosis suggests that they may have a special role in resistance to infectious pathogens. The finding in gene-disrupted mouse models that perforin is itself not essential for resolution of mycobacterial infection in vivo (17) raises the possibility that the antimicrobial activity may be independent of the lytic process or that there may be additional mediators in the cytotoxic granules, such as granzymes, defensins, or granulysin (18). Delineation of the mechanism whereby CD1-restricted CTLs kill intracellular mycobacteria may provide useful insights into mechanisms whereby other types of CTLs contribute to protection against microbial pathogens.

  • * To whom correspondence should be addressed. E-mail: rmodlin{at}


Stay Connected to Science

Navigate This Article