Use of the Human Granulysin Transgenic Mice To Evaluate the Role of Granulysin Expression by CD8 T Cells in Immunity To Mycobacterium tuberculosis

Abstract

The cytotoxic granules of human NK and CD8 T cells contain the effector molecule granulysin. Although in vitro studies indicate that granulysin is bactericidal to Mycobacterium tuberculosis and human CD8 T cells restrict intracellular M. tuberculosis by granule exocytosis, the role of granulysin in cell-mediated immunity against infection is incompletely understood, in part because a granulysin gene ortholog is absent in mice. Transgenic mice that express human granulysin (GNLY-Tg) under the control of human regulatory DNA sequences permit the study of granulysin in vivo. We assessed whether granulysin expression by murine CD8 T cells enhances their control of M. tuberculosis infection. GNLY-Tg mice did not control pulmonary M. tuberculosis infection better than non-Tg control mice, and purified GNLY-Tg and non-Tg CD8 T cells had a similar ability to transfer protection to T cell deficient mice. Lung CD8 T cells from infected control and GNLY-transgenic mice similarly controlled intracellular M. tuberculosis growth in macrophages in vitro. Importantly, after M. tuberculosis infection of GNLY-Tg mice, granulysin was detected in NK cells but not in CD8 T cells. Only after prolonged in vitro stimulation could granulysin expression be detected in antigen-specific CD8 T cells. GNLY-Tg mice are an imperfect model to determine whether granulysin expression by CD8 T cells enhances immunity against M. tuberculosis. Better models expressing granulysin are needed to explore the role of this antimicrobial effector molecule in vivo. IMPORTANCE Human CD8 T cells express the antimicrobial peptide granulysin in their cytotoxic granules, and in vitro analysis suggest that it restricts growth of Mycobacterium tuberculosis and other intracellular pathogens. The murine model of tuberculosis cannot assess granulysin's role in vivo, as rodents lack the granulysin gene. A long-held hypothesis is that murine CD8 T cells inefficiently control M. tuberculosis infection because they lack granulysin. We used human granulysin transgenic (GNLY-Tg) mice to test this hypothesis. GNLY-Tg mice did not differ in their susceptibility to tuberculosis. However, granulysin expression by pulmonary CD8 T cells could not be detected after M. tuberculosis infection. As the pattern of granulysin expression in human CD8 T cells and GNLY-Tg mice seem to differ, GNLY-Tg mice are an imperfect model to study the role of granulysin. An improved model is needed to answer the importance of granulysin expression by CD8 T cells in different diseases.

Keywords: CD8 T cells; Mycobacterium tuberculosis; antimicrobial peptides; granulysin; tuberculosis.

FIG 1

GNLY-Tg mice do not control of M. tuberculosis infection better than non-Tg mice. Control (C57BL/6, non-Tg) and GNLY-Tg mice were infected with M. tuberculosis, and after 16 weeks, the lung bacterial burden was determined, and flow cytometry was performed. (A) Proportion of CD8 T cells and NK cells in the lungs of non-Tg (red) and GNLY-Tg (blue) infected mice. (B) Proportion of tetramer positive cells among lung CD8 T cells from non-Tg (red) and GNLY-Tg (blue) infected mice. (C) Total numbers of pulmonary CD8 T cells, antigen-specific CD8 T cells, and NK cells. (D) Frequencies of CD8 T cells expressing CD69, GzmB, or having a SLEC or MPEC phenotype among GNLY-Tg (blue) and non-Tg (red) TB10.4 tetramer⁺ CD8 T cells, or non-Tg (black) or GNLY-Tg (gray) tetramer^–ve CD8 T cells. (E) Lung CFU from non-Tg littermate control or GNLY-Tg mice. (F to J) Purified CD8 T cells from GNLY-Tg or non-Tg mice were transferred to TCRα KO recipients 24 h prior to aerosol infection and analyzed 5 weeks later. (F) Total lung and spleen CD8 T cells, 5 weeks after transfer of TCRα KO mice with non-Tg (red) or GNLY-Tg (blue) CD8 T cells. (G) Frequencies of MPEC, SLEC, and activated CD8 T cells in the lungs of TCRα infected mice with transferred non-Tg (red) or GNLY-Tg (blue) CD8 T cells. Bacterial burden in the (H) lungs and (I) spleens of TCRα KO mice after transfer of CD8 T cells from non-Tg (red) or GNLY-Tg mice (blue). (J) Survival of M. tuberculosis-infected TCRα KO mice after transfer of CD8 T cells from non-Tg (red) or GNLY-Tg (blue) mice. (K) In vitro control of M. tuberculosis growth in macrophages by CD8 T cells isolated from the lungs of GNLY-Tg (filled) and non-Tg (open) infected mice at various effector:target ratios compared to macrophage only (red). (A to E) is representative data from five independent infections analyzed 4-, 8-, and 16-wpi, with n = 5 to 10 mice/group. (F to J) is data from one experiment each with n = 5/group. (K) is representative of two independent experiments, each using pooled cells from n = 5 mice/group. Statistical analysis was performed using a t test (A to G), one-way ANOVA (H, I), log-rank test (J), and two-way ANOVA (K). Comparison to infected macrophages is represented by vertical asterisks. **, P < 0.01; ***, P < 0.001; ****, P < 0.0001; ns, not significant.

FIG 2

Granulysin is not expressed by CD8 T cells in the lungs of M. tuberculosis-infected GNLY-Tg mice. (A) Granulysin (GNLY) and granzyme B (GzmB) expression by human CD56⁺CD3^– NK and CD56^–CD3⁺ CD8 T cells in unstimulated state (unstim) and 8 days after in vitro activation with anti-CD3/CD28 + IL-2. RB1 detecting total granulysin (upper row) and DH2 detecting 9 kDa (lower row). (B) Representative flow cytometry plots of granulysin and granzyme expression by NK cells, tetramer positive and tetramer negative CD8 T cells in non-Tg (upper panel) and GNLY-Tg (lower panel). Data representative data from five independent infections analyzed 4-, 8-, and 16-wpi, with n = 5 to 10 mice/group. (C) Expression of granulysin and granzyme B by NK cells in the lungs of M. tuberculosis-infected mice.

FIG 3

Granulysin is expressed in GNLY-Tg NK cells, but not CD8 T cells. (A, B) Lung mononuclear cells from non-Tg littermate control and GNLY-Tg mice were isolated from M. tuberculosis-infected mice 4 weeks after infection. The cells were either stained immediately (“A,” ex vivo) or after stimulation with TB10.4_4-11 and 32A_309-318 peptides and IL-2 (“B,” day 7, in vitro). Granulysin and granzyme expression by NK cells (CD3^–NK1.1⁺) and CD8 T cells (CD3⁺CD8⁺NK1.1^–) was determined by intracellular staining. The CD8 T cells were further divided based on expression of CD44 and CD62L as CD44⁺CD62L^– effector T cells (T_E), CD44⁺CD62L⁺ central memory T cells (T_CM), CD62L⁺CD44^– naive T cells (T_N). Quadrants were based on FMO, isotype-matched antibody controls, and non-Tg littermates. Representative of three independent experiments each with four mice/group.

FIG 4

Granulysin expression by GNLY-Tg B8R_20-27-specific CD8 T cell lines. (A) Schematic of strategy for generating GNLY-Tg B8R_20-27-specific CD8 T cell lines. (B) Granulysin and granzyme B expression by NK or CD8 T cells was measured 5 days after in vitro stimulation of immune splenocytes with B8R peptide and cytokines (IL-2 or IL-15). (C) Splenic CD8 T cells were purified by negative selection and stimulated with B8R_20-27-pulsed irradiated splenocytes and IL-2. Three weeks after stimulation, granulysin and granzyme B expression was measured. Performed once. (D) Non-Tg and GNLY-Tg B8R_20-27-specific CD8 T cell lines were stimulated in vitro with B8R_20-27-pulsed irradiated splenocytes and indicated cytokines. At indicated days post stimulations, granulysin and granzyme B expression were determined by flow cytometry. Gates by isotype controls. Performed twice using cell lines.