|Home | About | Journals | Submit | Contact Us | Français|
Granulysin is a cytolytic and proinflammatory molecule first identified by a screen for genes expressed ‘late’ (3–5 days) after activation of human peripheral blood mononuclear cells. Granulysin is present in cytolytic granules of cytotoxic T lymphocytes and natural killer cells. Granulysin is made in a 15-kDa form that is cleaved into a 9-kDa form at both the amino and the carboxy termini. The 15-kDa form is constitutively secreted, and its function remains poorly understood. The 9-kDa form is released by receptor-mediated granule exocytosis. Nine kiloDalton granulysin is broadly cytolytic against tumors and microbes, including gram-positive and gram-negative bacteria, fungi/yeast and parasites. It kills the causative agents of both tuberculosis and malaria. Granulysin is also a chemoattractant for T lymphocytes, monocytes and other inflammatory cells and activates the expression of a number of cytokines, including regulated upon activation T cell expressed and secreted (RANTES), monocyte chemoattractant protein (MCP)-1, MCP-3, macrophage inflammatory protein (MIP)-1α, interleukin (IL)-10, IL-1, IL-6 and interferon (IFN)-α. Granulysin is implicated in a myriad of diseases including infection, cancer, transplantation, autoimmunity, skin and reproductive maladies. Small synthetic forms of granulysin are being developed as novel antibiotics. Studies of the full-length forms may give rise to new diagnostics and therapeutics for use in a wide variety of diseases.
Granulysin is a cytolytic and proinflammatory molecule expressed by activated human cytotoxic T lymphocytes (CTLs) and natural killer (NK) cells (1). It was first identified by subtractive hybridization in a search for genes expressed by human T lymphocytes ‘late’ (3–5 days) after activation (2). Based on sequence homologies with sphingolipid hydrolase activators of the central nervous system, saposins A–D, it is a member of the saposin-like protein (SAPLIP) family of proteins (3, 4) (Table 1). The SAPLIP family members amoebapores A–C (Entamoeba histolytica pore-forming proteins) and NK-lysin (a porcine lytic granule protein) exhibit lytic activity, suggesting that granulysin might also be cytolytic.
Granulysin is found in cytolytic granules in CTL and NK cells along with the pore-forming protein, perforin, and granzymes (4, 5). It is synthesized as a 15-kDa molecule, and portions are then cleaved at the amino and carboxy termini to produce a 9-kDa form (4). Equivalent amounts of these two forms of granulysin are found in CTL and NK cells. However, the 9-kDa form is sequestered in cytolytic granules, while the 15-kDa form is constitutively secreted (6). Recombinant 9 kDa granulysin is dependent on perforin for killing intracellular pathogens (5). Recombinant 9 kDa granulysin is tumoricidal and broadly antimicrobial, killing gram-positive and gram-negative bacteria, yeast, fungi and parasites (5). Granulysin kills Mycobacterium tuberculosis, the causative agent in tuberculosis (5), and Plasmodium falciparum, a cause of malaria (7).
Recombinant 9 kDa granulysin is broadly tumoricidal. Extensive studies of recombinant granulysin lysis of Jurkat, a human model tumor, are summarized in Figure 1. Granulysin binds to the tumor cell surface based on charge (8). There is no evidence to date for a specific receptor for cytotoxicity. Granulysin appears to scissor through the cell membrane, causing ion fluxes (8, 9). Cytosolic calcium levels increase as a result of an influx of extracellular calcium and a release of calcium from intracellular stores, while cytosolic potassium levels decrease as a result of activation of a calcium-dependent potassium pump (10). Addition of agents that block either the increase in intracellular calcium or the decrease in intracellular potassium prevents activation of subsequent cell death pathways, and no tumor lysis occurs. The increase in intracellular calcium, decrease in potassium and the presence of granulysin (if ATP is present) all contribute to mitochondrial damage. Mitochondrial calcium overload disrupts the Krebs cycle and oxidative phosphorylation. Megachannels are opened, giving rise to release of both cytochrome c and apoptosis-inducing factor (AIF). Electron transport is blocked and reactive oxygen species increase. Cytochrome c release activates a caspase cascade, which together with AIF induces endonuclease activation and standard apoptosis. This pathway occurs in minutes (10).
In contrast, a slower cytotoxicity pathway is also induced by granulysin as it activates a sphingomyelinase associated with the cell membrane to generate ceramide (11). The importance of this ceramide pathway remains unclear.
Recombinant 9 kDa granulysin is broadly antimicrobial (5, 12). Its mechanism of action appears to mimic tumor lysis, interrupting oxidative metabolism and energy generation by the organism. The cell wall is damaged and lipid metabolism disrupted in M. tuberculosis.
Like defensins, which have both cytotoxic and proinflammatory functions, granulysin is a chemoattractant for a variety of cell types (12). Recombinant 9 kDa granulysin attracts T cells, monocytic and NK-cell tumor lines but not Epstein-Barr virus-transformed B-cell lines. Granulysin shows maximal chemotactic activity for CD4+ and CD8+ T cells and monocytes at 10 nM. Chemotaxis is inhibited by pertussis toxin, suggesting involvement of a G-protein coupled receptor, but the identity of the receptor(s) remains unknown.
In addition to chemotaxis, recombinant 9 kDa granulysin induces expression of a number of proinflammatory cytokines in monocytes and related tumor cell lines (U937), including RANTES, MCP-1, MCP-3, MIP-1α, IL-10, IL-1, IL-6 and IFN-α but not CD14, X-ray repair cross complementing protein 1 (XRCC-1), chemokine (C-X-C motif) receptor 4 (CXCR-4) or TNF-α (12).
Granulysin has broad clinical relevance to a myriad of diseases, including infections, cancer, transplantation, skin afflictions and reproductive complications.
Granulysin has been implicated in mycobacterial diseases and malaria. Relatively, recent experiments with gene-deleted mice (β2m, IFN-γ, perforin, Fas/FasL, reviewed in 13) and the characterization of CTL lines from patients indicate an important role for CTL in immunity to tuberculosis (13). With Modlin, Stenger and others, we identified a granulysin/perforin-mediated pathway responsible for protection against intracellular pathogens, including M. tuberculosis (5). Granulysin is present in granulomas in tuberculosis, and its expression is highly correlated with a curative host response and outcome (14). Children with tuberculosis show decreased serum granulysin levels that reverse with therapy (15). CD8+ T cells coordinately express CCL5 (RANTES), granulysin and perforin in tuberculosis (16). CCL5 attracts infected macrophages, and perforin and granulysin coordinate to kill the intracellular organism. Granulysin is highly active against both drug-resistant and drug-sensitive M. tuberculosis clinical isolates (17). There is impaired expression of perforin and granulysin in CD8+ T lymphocytes at the site of infection in chronic pulmonary tuberculosis patients (18). Recently, Klucar et al. showed that CD4+ T cells can lyse M. tuberculosis-infected target cells and that both perforin and granulysin contribute to lysis and inhibition of mycobacterial growth (19). Human natural killer T (NKT) cells express granulysin and exhibit antimycobacterial activity (20). Mycobacterium bovis Bacillus Calmette-Guerin (BCG) vaccination induces memory CD4+ T cells in cows with enhanced expression of granulysin homologue and perforin (21). Eukaryotic expression of granulysin in mice is protective against M. tuberculosis (22).
In a detailed examination of granulysin expression in leprosy, granulysin-expressing cells were detected in cutaneous leprosy lesions at a sixfold greater frequency in patients with the localized form of the disease (tuberculoid) compared with the disseminated (lepromatous) form of the disease (23). By contrast, perforin expression was similar across all spectra of the disease. Unexpectedly, within the leprosy granulomas, granulysin was found in CD4+ T cells, while the perforin was in the CD8+ T cells. These findings suggest a likely hypothesis: when the same effector cell expresses both perforin and granulysin, the Mycobacterium leprae is destroyed and there is no disease. When different cells express granulysin and perforin, the microbe is walled off in a granuloma. If there is little or no granulysin found, the disease disseminates.
Farouk et al. showed that γδ T cells inhibit growth of the asexual blood stages of P. falciparum, a causative agent for malaria, by a granule exocytosis pathway dependent on granulysin expression (7). The inhibitory capacity of effector T cells strongly correlated with granulysin expression in their cytotoxic granules. Zheng et al. showed granulysin-mediated killing of the fungus Cryptococcus neoformans by granulysin in cytotoxicCD4+ T cells (24). Furthermore, this pathway is defective in patients with HIV.
Granulysin expression has been widely correlated with good outcomes in a variety of cancers. Low intracellular expression of granulysin, but not perforin, in CD3-CD16+ cells correlates with progression of cancer (25). Flow cytometric analysis showed high levels of both perforin and granulysin in the NK cells in normal healthy controls. Tumor-free patients expressed granulysin at levels similar to controls, while cancer patients exhibited significantly decreased (P < 0.005) granulysin levels. In contrast, perforin levels were similar in all groups. This study suggests that impaired expression of granulysin in circulating NK cells correlates with tumor progression (25).
Additional studies indicate that granulysin expression in tumor infiltrates is associated with good outcomes. Pages et al. found low granulysin expression in effector memory T cells in tumor infiltrates to correlate early metastasis and poor survival rates, while high levels of granulysin correlated with good outcomes in colorectal carcinoma (26). Others have suggested that granulysin expression is a useful biomarker for outcomes in gastric carcinoma (27) and neuroblastoma (28). A Japanese group showed that expression of granulysin in a mouse model of small cell lung cancer was protective (29).
Granulysin is also a useful biomarker in transplantation (1). We showed that granulysin expression in situ is a marker for acute rejection and steroid resistance (30). Kotsch et al. reported that enhanced granulysin mRNA expression in urinary sediment is the best biomarker for both early and delayed renal allograft rejection (31). Hidalgo et al. examining the ‘transcriptome’ of human cytotoxic T cells found that granulysin was the most highly expressed transcript in both CD8+ and CD4+ CTLs (32). In allogeneic, but not autologous, stem cell transplantation, serum granulysin is transiently increased as measured by Enzyme-Linked ImmunoSorbent Assay (ELISA) 3 weeks after transplantation (33). Granulysin levels are markedly increased and correlate with severity of graft vs host disease, suggesting that it will be a useful biomarker in this context as well.
Granulysin is an important mediator of damage in a variety of skin diseases, including folliculitis (34), psoriasis (35), acne (36), lichen planus (37) and viral vesicles (38). Most recently, Chung et al. showed that granulysin is a key mediator for disseminated keratinocyte death in both Stevens–Johnson syndrome (SJS) and toxic epidermal necrolysis (TEN) (39). Granulysin was present in CTL and NK cells in the lesions, but the 15-kDa secretory form was present in blister fluid as well. Intradermal injection of recombinant 15 kDa granulysin caused skin necrosis and blisters mimicking the clinical features of both SJS and TEN.
A standard approach to examine the in vivo function of cytolytic molecules like granzymes and perforin has been to disrupt expression of these genes in mice. Because mice do not express granulysin or an apparent homologue, we generated mice transgenic for human granulysin to establish in vivo activity (45). In the first study, mice transgenic for human granulysin were shown to reject C6VL, a T-cell lymphoma of C57BL/6 origin, better than wild-type littermates. This tumor rejection is dependent on CD8+ T cells. In contrast, protection against RMA-S, a mutant lymphoma cell of C57BL/6 origin that is defective in peptide loading of major histocompatibility complex (MHC) class I and that is rejected by NK cells, was identical in granulysin transgenic and control animals. This parallels studies in human effector cell lysis of C. neoformans in which CD8+ T-cell-mediated killing is granulysin dependent, but NK-cell-mediated killing is not (46, 47). Rather, NK cell killing of C. neoformans depends on perforin (46).
We are in the process of backcrossing the granulysin transgenic animals onto perforin and granzyme knockouts and onto Balb/c gene backgrounds. These animals will be useful for a variety of in vivo studies.
Granulysin has potential as a new therapeutic for cancer and/or infectious diseases. To date, we have focused on the generation of synthetic peptides that are broadly antimicrobial (48). Granulysin derivatives offer the opportunity to be broad spectrum, fast acting, non-toxic to human cells and inexpensive and easy to manufacture. As antibiotic resistance increases in clinically important organisms like M. tuberculosis and Staphylococcus aureus, we focused on granulysin and its derivatives as new broad-spectrum antibiotics.
A series of several hundred synthetic peptides were examined (49, 50). This analysis indicated that the cysteine residues in granulysin are important for lysis of mammalian cells (tumors) but not bacteria. Substitution of some arginine residues also abrogates lysis of mammalian cells but not bacteria. Inclusion of some d-amino acids disrupts α-helices, abrogating mammalian cell lysis but not bacteria. The approach has been to synthesize test variant peptides for in vitro and in vivo evaluation. These studies have led to broad-spectrum lead compounds of improved activity and stability.
These peptides have been evaluated in model systems of Vibrio cholera in neonatal mice and S. aureus in wound infections with encouraging results (49).
Granulysin is constitutively expressed by NK cells and after activation in both CD8+ and CD4+ cytolytic T lymphocytes. Recombinant 9 kDa granulysin kills a variety of bacteria, fungi, parasites and tumors. In concert with perforin, it kills intracellular pathogens. Granulysin is also a chemoattractant for subsets of leukocytes and can activate both monocytes and dendritic cells. Mice transgenic for granulysin have been generated to evaluate in vivo activity. The first studies with these mice show increased tumor rejection by CD8+ T lymphocytes but no effect on NK-cell-mediated tumor protection. Lastly, synthetic peptides based on granulysin exhibit greatly enhanced activity and specificity in vitro and in vivo. Together, these studies suggest that granulysin may prove a useful diagnostic (biomarker) and/or therapeutic for a wide variety of diseases.
This work was supported by the Intramural Program of the National Institutes of Health, National Cancer Institute and the Center for Cancer Research.