INFECTION AND IMMUNITY, Nov. 1989, p. 3653-3659 Vol. 57, No. 11 0019-9567/89/113653-07$02.00/0 Copyright X3 1989, American Society for Microbiology Inhibition of Human Lymphoproliferative Responses by Mycobacterial Phenolic Glycolipids JEAN-JACQUES FOURNIE,t* ELIZABETH ADAMS, RAYMOND J. MULLINS, AND ANTONY BASTEN Clinical Immunology Research Centre, University of Sydney, 2006 Sydney, Australia Received 17 April 1989/Accepted 8 August 1989 The effect of mycobacterial phenolic glycolipids from Mycobacterium leprae, M. bovis BCG, and M. kansasii on in vitro proliferative responses by human blood mononuclear cells from healthy BCG vaccinees was investigated. All three phenolic glycolipids inhibited proliferation in a concentration-dependent manner. Inhibition was independent of the stimulus used and involved neither antigen-presenting cells nor antigen- specific CD8+ suppressor T cells. It was concluded that the phenomenon may be a general property of mycobacterial phenolic glycolipids, perhaps analogous to the growth-modulating properties of gangliosides. Despite the lack of specificity of inhibition in vitro, de facto specificity may occur in vivo by virtue of the localization of glycolipid in the leprosy lesions. The apparent lack of protective immunity in patients with surface of the antigen-presenting cell (APC) for several days leprosy is associated with the persistence of viable myco- (18), thereby interfering with the subsequent proliferative bacteria within host tissues and a lack of degradation by responses of murine lymphocytes to a variety of stimuli (8). macrophages. Because the microbicidal function of macro- Thus, the studies with both these antigens emphasize the phages is largely dependent on activation by antigen-specific importance of the potential interaction between mycobacte- T lymphocytes, the occurrence of infection could be second- rial glycolipids and macrophages in interfering with T-cell ary to interference in the activation process by mycobacte- activation. rial antigens. Among these antigens, the abundant mycobac- The most widely studied glycolipid from M. leprae is Phe terial cell wall glycolipids constitute the active interface GL-I, largely on account of its capacity to elicit a strong between the pathogen and its environment and therefore antibody response in leprosy patients, particularly those at have the potential to modulate the host response. The the lepromatous end of the spectrum (2). However, this Phe glycolipids are of different types: the phosphatidylinositol GL-I also interacts with the cellular arm of the immune mannosides, the phosphorylated lipoarabinomannam (LAM) system at the level of both macrophage function and lym- found in most mycobacteria, the glycopeptidolipids from the phocyte proliferation. For example, pretreatment of mono- Mycobacterium avium-M. intracellulare-M. scrofulaceum cytes with Phe GL-I from M. leprae results in inhibition of complex, the trehalose-containing lipooligosaccharides from the oxidative burst in response to bacterial antigens (30), M. kansasii and M. xenopi, and the phenolic glycolipids (Phe whereas the ability of macrophages to kill Phe GL-I-coated GL) found in M. bovis BCG, M. leprae, M. kansasii, M. Staphylococcus aureus cells has been shown to be grossly marinum, and M. gastri (reviewed in reference 5). impaired (26). In addition, there is some evidence that the Data about the effects of some but not all of these Phe GLs may act as virulence factors by scavenging oxygen glycolipids on the in vitro response by human lymphoid cells radicals released by activated macrophages (10). are available. LAM is a potent inhibitor of in vitro prolifer- Phe GL-I was initially reported to cause inhibition of ation of peripheral blood mononuclear cells (PBM) from concanavalin A (ConA)-induced proliferation of PBM from healthy subjects or patients with tuberculosis when stimu- patients with lepromatous disease but not from those with lated by mycobacterial or other antigens, the effect being tuberculoid disease or from healthy controls (23). The mech- dependent on the presence of monocytes in culture (14). anism of this effect was thought to be due to triggering of M. Further studies have revealed that this molecule can also leprae-specific T suppressor cells by the specific oligosac- inhibit purified protein derivative (PPD)- or M. Ieprae- charide epitope on Phe GL-I (24). On the other hand, induced responses by PBM from leprosy and patients attempts to induce M. leprae-specific T lymphocytes in mice healthy individuals (20) as well as influenza virus-specific proliferation of CD4+ T-cell clones (25). Since proliferation with Phe GL-I have yielded little evidence of a specific cellular response to the glycolipid (6, 21). induced by macrophage-independent stimuli such as inter- More recently, the effect of Phe GL-I on proliferative leukin-2 (IL-2) and anti-CD3 antibody was not inhibited by responses to ConA in leprosy patients across the clinical LAM, the findings indicate that LAM exerted its effect at the level of antigen processing by macrophages. spectrum as well as in healthy contacts has been studied. Similar results have been found with glycopeptidolipids, The results indicated that Phe GL-I usually exerts a nonspe- (the serovar-specific antigens of M. avium, M. intracellulare cific inhibitory effect on proliferation, which occurs regard- less of the type or even the presence of leprosy, although in and M. scrofulaceum). This type of glycolipid is resistant to a minority of patients and healthy contacts the mitogen degradation by macrophages, so that it remains intact on the response was actually enhanced (27). These findings sug- gested that Phe GL-I was unlikely to determine the type of * Corresponding author. leprosy or to induce a specific T-suppressor cell response. t Present address: Centre de Recherches de Biochimie et de The conflicting results in the literature prompted the G6ndtique Cellulaires du Centre National de la Recherche Scienti- current study, which was designed to test whether higher fique, 118 route de Narbonne, 31062 Toulouse, France. purified mycobacterial glycolipids could in fact influence the 3653 3654 FOURNIE ET AL. INFECT. IMMUN. in vitro proliferative response by T cells and, if so, by what 30 ng/ml. Leu-O-Me was freshly diluted in PBS and used at mechanism. A variety of antigens and mitogens were used to a final concentration of 5 nM as previously described (29). control for specificity, and cells from healthy BCG vaccinees The IL-2 source was the culture supernatant from the gibbon and continuous cell lines served as targets to avoid the cell line MLA 144 used at a final concentration of 20%. potentially confounding effects of chronic infection and Monoclonal antibodies. The anti-CD3, CD4, CD8, and treatment associated with the use of PBM from leprosy CD25 (Tac) monoclonal antibodies were purchased from patients. Initial experiments were performed with the Phe Becton Dickinson and Co., Mountain View, Calif., and the GLs from M. leprae (Phe GL-I), M. kansasii (Phe GLK-I), secondary anti-mouse antibodies fluorescein isothiocyanate and M. bovis BCG (Phe GLB-I) (which differ only in their and phycoerythrin conjugate were obtained from Silenius oligosaccharide portion), and Phe GLK-I was used in the Laboratories, Hawthorn, Australia, and Amersham Interna- remainder because of its fully elucidated structure and ready tional Plc, Amersham, United Kingdom, respectively. availability in pure form (15, 16). Preparation of cells. Samples of blood from healthy BCG Inhibition of proliferation was found to be independent of vaccinees were collected in preservative-free heparin at 10 the Phe GL studied, of the stimulus to proliferation, and of IU/mi (Weddel Pharmaceuticals, Thomleigh, New South the presence of APC or CD8' T cells. The mechanism Wales, Australia). PBM were separated on Ficoll-Hypaque appeared to be a nontoxic functional effect on proliferating gradients. T lymphocytes and APC were further separated CD4+ lymphocytes. The degree of inhibition was correlated from PBM by AET-treated sheep erythrocyte rosetting, as in most cases with the concentration of glycolipid in culture, described previously (28), into E+ and E- cell-enriched although there was a large degree of individual variation. populations, respectively. Fluorescence-activated cell sorter Despite the apparent lack of specificity of the effect, it was (FACS) profiles of the E+ population revealed consistently concluded that inhibition may be a general property of less than 3% of remaining monocytic cells. CD8+-depleted glycolipids, with the potential for de facto specificity owing and CD8+-enriched T-cell populations were obtained from to their high concentration within the lesions of leprosy. E+ cells by using an anti-CD4 monoclonal antibody panning procedure (32). The nonadherent (CD8+-enriched) popula- MATERIALS AND METHODS tion was 52% CD8+, and the adherent (CD8+-depleted) population was less than 4% CD8+, as assessed by FACS Preparation of purified Phe GLs. Phe GLK-I, the major analysis. Monocyte-macrophage-selective depletion of the tetrasaccharidic Phe GL from M. kansasii, was extracted lymphocyte population was performed by treatment of E+ from cultures with chloroform-methanol and purified by cells with 5 mM Leu-O-Me in serum-free medium (5 x 106 liquid chromatography as already described (15). Purified M. cells per ml) for 40 min at room temperature. The remaining bovis BCG Phe GL B-I and M. leprae Phe GL-I were kind population was found by FACS analysis to contain less than gifts from, respectively, A. Vercelonne and C. Asselineau, 0.5% monocytes or macrophages, as described previously Centre de Recherches de Biochimie et de Genetique Cellu- (29), and consequently could not be stimulated to proliferate laires du Centre National de la Recherche Scientifique, by lectins (data not shown). Toulouse, France. The glycolipids were stored in sterile The antigen-specific helper T-cell line EABCG reactive to filtered ethanol stock solutions at 4°C and a concentration of M. bovis BCG sonic extract is part of a larger collection of 100 ,ug/ml. BCG-specific T-cell lines and was generated from PBM of a Reagents.