A sphingosine kinase inhibitor combined with temozolomide induces glioblastoma cell death through accumulation of dihydrosphingosine and dihydroceramide, endoplasmic reticulum stress and autophagy.

Glioblastomas (GBMs) are very aggressive tumors with low chemosensitivity. The DNA-alkylating agent temozolomide (TMZ) is currently the most efficient chemotoxic drug for GBM therapy; however, many patients develop resistance to TMZ. Combining TMZ with another agent could present an improved treatment option if it could overcome TMZ resistance and avoid side effects. Sphingosine kinase inhibitors (SKIs) have emerged as anticancer agents. Sphingosine kinases are often overexpressed in tumors where their activity of phosphorylating sphingosine (Sph) contributes to tumor growth and migration. They control the levels of the pro-apoptotic ceramide (Cer) and Sph and of the pro-survival sphingosine-1 phosphate. In the present work, TMZ was combined with a specific SKI, and the cytotoxic effect of each drug alone or in combination was tested on GBM cell lines. The combination of sublethal doses of both agents resulted in the cell death potentiation of GBM cell lines without affecting astrocyte viability. It triggered a caspase-3-dependent cell death that was preceded by accumulation of dihydrosphingosine (dhSph) and dihydroceramide (dhCer), oxidative stress, endoplasmic reticulum stress, and autophagy. Autophagy was identified as the crucial switch that facilitated induction of this cell death potentiation. The sublethal dose of the inhibitor induced these stress events, whereas that of TMZ induced the destructive autophagy switch. Remarkably, neither Cer nor Sph, but rather the Cer intermediates, dhSph and dhCer, was involved in the cytotoxicity from the combination. Cell lines sensitive to the combination expressed low levels of the antioxidant enzyme glutathione peroxidase-1, indicating this enzyme as a potential marker of sensitivity to such treatment. This work shows for the first time a strong interaction between a SKI and TMZ, leading to a tumor cell-specific death induction. It further demonstrates the biological relevance of dihydrosphingolipids in cell death mechanisms and emphasizes the potential of drugs that affect sphingolipid metabolism for cancer therapy.

Immunocompetent astrocytes and microglia display major differences in the processing of the invariant chain and in the expression of active cathepsin L and cathepsin S.

The role of astrocytes and microglia as antigen-presenting cells in the brain is still controversial. In this study we have analyzed and compared aspects of the molecular machinery that underlies MHC class II trafficking in immunocompetent astrocytes and microglia. We show that IFN-gamma-stimulated microglia possess active cathepsin L and cathepsin S, and efficiently degrade the invariant chain, unlike IFN-gamma-stimulated astrocytes that express cathepsin L but not cathepsin S. The lack of cathepsin S proves to be dramatic for the antigen-presentation capacity of astrocytes, which is nearly abolished when these cells are stimulated by a combination of IFN-gamma and TNF-alpha. TNF-alpha indeed decreases cathepsin L activity as we show here, leading to alterations in invariant chain processing, and hence in MHC class II trafficking in astrocytes. Cystatin C inhibits cathepsin L activity in astrocytes, but does not regulate cathepsin L and cathepsin S activity in microglia. We therefore identify cathepsin L and cathepsin S as key components in the regulation of the immune potential of astrocytes and microglia, and provide evidence for a cell-specific regulation exerted by IFN-gamma and TNF-alpha on the expression and activity of cathepsins.

Regional and cellular expression of the mannose receptor in the post-natal developing mouse brain.

The mannose receptor, a glycoprotein expressed in a soluble and membrane form by macrophages, plays an important role in homeostasis and immunity. Using biochemical and immunohistochemical analyses, we demonstrate that this receptor, both in its soluble and membrane forms, is expressed in vivo in the post-natal murine brain and that its expression is developmentally regulated. Its expression is at its highest in the first week of life and dramatically decreases thereafter, being maintained at a low level throughout adulthood. The receptor is present in most brain regions at an early post-natal age, the site of the most intense expression being the meninges followed by the cerebral cortex, brain stem and the cerebellum. With age, expression of the mannose receptor is maintained in regions such as the cerebral cortex and the brain stem, whereas it disappears from others such as the hippocampus or the striatum. In healthy brain, no expression can be detected in oligodendrocytes, ependymal cells, endothelial cells or parenchymal microglia. The mannose receptor is expressed by perivascular macrophages/microglia and meningeal macrophages, where it might be important for the brain immune defence, and by two populations of endogenous brain cells, astrocytes and neurons. The developmentally dependent, regionally regulated expression of the mannose receptor in glial and neuronal cells strongly suggests that this receptor plays an important role in homeostasis during brain development and/or neuronal function.

CD1 expression is differentially regulated by microglia, macrophages and T cells in the central nervous system upon inflammation and demyelination.

Expression of CD1 by microglia, macrophages and T cells was investigated ex vivo. In the healthy central nervous system (CNS), resident microglia, macrophages and T cells express levels of CD1 significantly lower than that expressed by splenic macrophages and T cells. During experimental autoimmune encephalomyelitis (EAE), CD1 expression by microglia and the number of CD1+ microglia increase. Macrophages and T cells strongly upregulate CD1 expression in the CNS, but not in the spleen. Whereas the function of CD1 expressed by T cells remains unclear, the expression by microglia and macrophages provides the CNS with a (glyco)lipidic-presenting molecule in an inflammatory and demyelinating environment.

Tumor necrosis factor alpha and interleukin-1 alpha inhibit through different pathways interferon-gamma-induced antigen presentation, processing and MHC class II surface expression on astrocytes, but not on microglia.

Astrocytes and microglia, two glial cell populations of the CNS, have been described to be involved in many immune processes. We used defined combinations of cytokines, interferon gamma (IFN-gamma)/interleukin-1 alpha (IL-1 alpha) and IFN-gamma/tumor necrosis factor alpha (TNF alpha), to simulate different in vitro immune environments observed in disease or inflammation. In these conditions, we analyzed and compared the regulating effects of these cytokines on cell surface and total expression of MHC II and on the capacity of murine astrocytes and microglia to present peptide and native antigens to specific primed T cells. Neither IL-1 alpha nor TNF alpha affected the IFN-gamma-induced antigen presentation capacity of microglia. Astrocytes, however, were severely impaired in their capacity to present native antigens and, to a minor extent, a peptide antigen. Total expression of MHC II was not affected by these cytokines in microglia, whereas in astrocytes it was reduced by IL-1 alpha and increased by TNF alpha. Both cytokines downregulated MHC II expression at the surface of astrocytes, but not of microglia. This shows that TNF alpha affects the of IFN-gamma-immunocompetent astrocytes to process and present antigen, probably either by altering membrane traffic of MHC II and of antigen and/or enzymatic activities associated with these mechanisms, while IL-1 alpha does so by downregulating MHC II expression. Altogether, our results illustrate how differently astrocytes and microglia react toward a defined, similar immune environment. One type of cell, the astrocytes, downregulate their T-cell stimulation and MHC II trafficking, and probably also their antigen processing, functions while the other, the microglia, maintain their antigen presentation potential.

Identification and functional characterization of the mannose receptor in astrocytes.

The immune competence of astrocytes is still ill defined, especially their endocytic capacity, a prerequisite for efficient antigen presentation. We show that mannose receptor, a very important conduit for internalization of infectious agents and self antigens, is functionally expressed in the murine CNS. By in vitro assays, astrocytes and microglia were shown to be the prime cells expressing this receptor. Studies on astrocytes demonstrate that its expression and function are inversely regulated by anti-and pro-inflammatory compounds. Downregulation of the mannose receptor by IFN-gamma is concomitant with the induction of the invariant chain, which is also induced by GM-CSF + IL-4. Mannose receptor-expressing astrocytes may thus act as scavenger not only in CNS development but also in defense, against soluble and particulate mannosylated pathogens, presenting fragments thereof at strategic locations in the CNS. These findings unravel a new and putatively very important role of astrocytes in innate immunity and possibly development.

A dominant-negative mutant of the Rab5 GTPase enhances T cell signaling by interfering with TCR down-modulation in transgenic mice.

TCR triggering results in the down-modulation of engaged receptors by endocytosis. As a result of this process, Ag-binding sites are depleted from the surface and signaling responses should be attenuated. To test the importance of TCR down-regulation on T cell signaling, we generated mice expressing a dominant-negative form of Rab5 (Rab5N133I) in T cells. Rab5, a monomeric GTPase of the Ras superfamily, has been implicated in the regulation of early steps in the endocytic pathway. In Rab5N133I mice, mature thymocytes developed, but the absolute number of CD4+CD8+ double positive thymocytes was reduced. Fluid phase endocytosis was severely impaired in the transgenic thymocytes. In peripheral T cells, the kinetics and rate of ligand-induced TCR down-modulation were delayed and reduced. These effects were correlated with enhanced early and late signaling responses. Analysis of thymocyte development in doubly transgenic mice for Rab5N133I and a lymphocytic choriomeningitis virus (LCMV) peptide-specific TCR demonstrated that TCR signaling was enhanced by dominant inhibition of Rab5 function, resulting in altered thymic selection. These findings suggest that TCR endocytosis is an important regulatory component of TCR signaling and that defects in this regulation can result in prolonged signaling and alter thymic development.

Selective storage of acetylcholine, but not catecholamines, in neuroendocrine synaptic-like microvesicles of early endosomal origin.

We have defined, in the neuroendocrine cell line PC12, the catecholamine- and acetylcholine-storing organelles in the context of the biogenesis of secretory granules and synaptic-like microvesicles (SLMVs). SLMVs were found to originate directly from early endosomes. Both early endosomes and SLMVs exhibited uptake and storage of biosynthetic acetylcholine. Surprisingly, however, despite the presence of a reserpine-sensitive vesicular amine transporter in early endosomes, SLMVs lacked detectable uptake and storage of catecholamines. This was confined to two populations of mature secretory granules, referred to as small and large mature secretory granules, which both derived from immature secretory granules. Our result show that PC12 cells lack small dense core vesicles, i.e., the catecholamine-storing, but secretory protein-lacking, vesicles found in sympathetic neurons and imply that the biogenesis of these vesicles requires the expression of a distinct type of vesicular amine transporter and/or a change in endosomal protein sorting.

Newly synthesized synaptophysin is transported to synaptic-like microvesicles via constitutive secretory vesicles and the plasma membrane.

The biogenesis of synaptic-like microvesicles (SLMVs) in neuroendocrine cells was investigated by studying the traffic of newly synthesized synaptophysin to SLMVs in PC12 cells. Synaptophysin was found to be sulfated, which facilitated the determination of its exit route from the trans-Golgi network (TGN). Virtually all [35S]sulfate-labeled synaptophysin was found to leave the TGN in vesicles which were indistinguishable from constitutive secretory vesicles but distinct from immature secretory granules and SLMVs. [35S]sulfate-labeled synaptophysin was rapidly transported from the TGN to the cell surface, with a t1/2 of approximately 10 min in resting cells. After arrival at the cell surface, [35S]sulfate-labeled synaptophysin cycled for at least 1 h between the plasma membrane and an intracellular compartment likely to be the early endosome. Up to approximately 40% of the [35S]sulfate-labeled synaptophysin eventually (after 3 h and later) reached SLMVs, which could be distinguished from the other post-TGN compartments by their lower buoyant density in a sucrose gradient and their selective inclusion upon permeation chromatography using a controlled-pore glass column. Our results suggest that newly synthesized membrane proteins of SLMVs in neuroendocrine cells, and possibly of small synaptic vesicles in neurons, reach these organelles via the TGN----plasma membrane----early endosome.

Epitope mapping of apamin by means of monoclonal antibodies raised against free or carrier-coupled peptide.

A panel of 20 monoclonal antibodies raised against the bee-venom peptide apamin (18 residues, 2 disulfide bridges) was prepared. Nine monoclonal antibodies (mAb) were obtained from a mouse immunized with free apamin and 11 from a mouse immunized with a mixture of free and carrier-coupled peptide. Using a panel of 11 synthetic apamin analogs, we examined the fine antigenic specificity of each antibody. The mAb generated against free apamin preferentially bound to the central part of the peptide and less frequently recognized the N- and C-terminal regions. However, monoclonal antibodies obtained by immunization with carrier-bound apamin showed a broader range of specificities, consistent with the possibility of the entire surface of this small antigen becoming immunogenic upon coupling to the carrier.

Identification of antigenic residues on apamin recognized by polyclonal antibodies.

The structural requirements for antigenic recognition of apamin--an 18-amino acid, disulfide-bridged peptide--by rabbit antibodies were defined using a set of 18 apamin analogs in a competition liquid-phase radioimmunoassay. Some residues contribute considerably to antigenic recognition, e.g. Ala10, Arg13, and others to a lesser extent, e.g. Arg14, Glu7 and Thr8. The N- and C-terminal moieties of apamin are less antigenically important. These findings suggest that a good part of antibody specificities are directed to the central tightly folded part of the molecule. They are consistent with the observation that in saturating conditions, labeled apamin can, on average, bind one specific Fab fragment.

Structural analysis of the interaction of apamin with Ia and its recognition by Ad- or Ab-restricted mouse T cells.

Apamin is a single-chain, disulfide-bonded, 18-amino acid peptide that elicits mouse T cell responses when presented by cells expressing syngeneic Ad or Ab class II MHC molecules. We previously showed that both the unfolding of this peptide by APC and the integrity of its N terminus segment were required for efficient apamin T cell recognition. To seek further information on the sites through which this peptide interacts with Ia and/or TCR, we used a panel of Ad- or Ab-restricted, apamin-specific THC to probe the antigenicity of a series of synthetic apamin analogs. These included peptides either truncated at the N terminus, or substituted by Ala at position 2, 4, 6, 7, 8, or 10. Analysis of THC responses to apamin analogs and use of the latter in competition assays for peptide presentation revealed the following: 1) optimal apamin T cell recognition critically involved Lys4, Ala5, Pro6, Glu7, and Leu10. The role of these residues in either "Ia or TCR binding regions" was found to depend upon the restricting Ia molecules at play. Thus, Lys4, Glu7, and Leu10 were TCR-binding residues in both Ad- and Ab-apamin complexes, whereas Lys4 participated in apamin/Ab but not, or to a marginal extent, in apamin/Ad interaction. Furthermore, Pro6 was associated either with an Ia contact region or a TCR interaction site when apamin was presented by Ab or Ad molecules, respectively. Unfolded apamin and the unrelated chicken OVA323-339 peptide were found to bind to the same, or closely related site(s) of Ad, as shown by their ability to compete reciprocally for recognition by appropriate Ad-restricted THC. Four distinct TCR V beta genes (V beta 2, V beta 4, V beta 6, and V beta 8) were found to be used in our panel of 16 apamin-specific THC. These data indicate that apamin interacts with Ad or TCR through a motif resembling other beta-sheeted, Ad-binding sequences; however, based on the spacing of the critical residues (i.e., 4, 7, and 10), the possibility exists that apamin processing permits the folding of this sequence into an alpha-helix.

Processing by accessory cells for presentation to murine T cells of apamin, a disulfide-bonded 18 amino acid peptide.

Apamin, an 18 amino acid peptide with two disulfide bonds, elicits specific T cell proliferative responses in H-2d and H-2b mouse strains. We evaluated the processing requirement of this compact peptide by accessory cells for presentation to apamin-reactive T hybridoma cells (THC) by analyzing the IL-2 responses of 16 THC from apamin-primed BALB/c or C57BL/6 mice, to various forms of either native or chemically synthesized apamin analogs. These included: unfolded peptides (whose four sulfhydryl groups were blocked by acetamidomethyl residues), N-and/or C-truncated peptides, and an analog with a single amino acid substitution at position 10. Assessment of the Ag-specific THC responses in the presence of either live or formaldehyde-prefixed APC indicated the following: 1) all THC stringently required Ag processing; 2) in 8 of 16 cases, the simple unfolding of apamin was sufficient to eliminate the need for Ag processing, or even induced increased THC IL-2 responses (other cells required further antigenic alterations in addition to unfolding, or rare processing steps dependent on the integrity of the two disulfide bonds); and 3) for most THC, the Leu10 and the N terminus arm of apamin were shown to be critical for expression of the epitopes involved in T cell recognition. These data indicate that apamin, a natural peptide having an appropriate size for T cell triggering, acquires its antigenic conformation after a processing by APC which primarily involves an alteration of a disulfide bond-dependent peptide folding.

A Thy-1.1-specific monoclonal alloantibody activates both mouse and rat T cells.

In an attempt to further evaluate the role of Thy-1 in the antigen-independent triggering of mouse T cells, we have examined the activating properties of two Thy-1.1-specific mouse monoclonal antibodies (mAb). These reagents were established from an (A.TH X A.TL)F1 hybrid mouse (Thy-1b) immunized with IL-2 producing (BALB/c (Thy-1b) X BW5147 (Thy-1a)) T hybridoma cells. Although both mAb recognized the same Thy-1.1 determinant, one mAb of the gamma 3,kappa class (H171-146) was found to induce several T hybridoma cells to produce IL-2, and AKR thymocytes or cloned helper T cells to proliferate, whereas another mAb of the gamma 1,kappa class (H171-112) failed to do so even in the presence of phorbol myristic acetate (PMA). Increased IL-2 responses of T hybridoma cells were observed when the cell bound Thy-1.1-specific mAb were crosslinked by goat anti-mouse Ig (GaMIg) antibodies. Both a T-cell activating rat anti-Thy-1.2 mAb and the anti-Thy-1.1 mAb H171-146, although directed at distinct cell surface molecules, synergistically stimulated IL-2 production by T hybridoma cells. In addition, the mouse mAb H171-146 was found to stimulate LOU/M rat thymocytes to proliferate in the presence of exogenous IL-2. These data demonstrate that T cells can use Thy-1 as a signal-transducing molecule in both mouse and rat species, and support the notion that the activating properties of Thy-1.1-specific mAb are influenced by their heavy chain isotypes.

Accessory molecules and T cell activation. II. Antibody binding to L3T4a inhibits Ia-independent mouse T cell proliferation.

Monoclonal antibody (mAb) blocking assays using L3T4- and lymphocyte function-associated antigen-1 (LFA-1)-specific reagents were used as an approach to investigate the involvement of these accessory molecules in various T cell activation pathways. As previously reported, rat mAb to L3T4a and LFA-1A functional epitopes efficiently blocked antigen-driven T helper cell proliferation. In contrast, antigen- and Ia-independent T cell triggering induced by appropriate mAb to the Thy-1 or the T cell receptor molecules were found to be inhibitable by L3T4a- but not LFA-1A-specific mAb, although the extent of inhibition varied, depending on the cell type and the activating signal examined. These results provide further evidence that the inhibiting effects of L3T4-specific mAb on T cell responses may be due, in addition to an impairment of L3T4-class II major histocompatibility complex molecular interaction, to a down regulatory signal possibly transmitted by the L3T4 molecule itself.

Accessory molecules and T cell activation. I. Antigen receptor avidity differentially influences T cell sensitivity to inhibition by monoclonal antibodies to LFA-1 and L3T4.

A series of BALB/c-derived T hybridoma cells, capable of producing interleukin 2 (IL 2) in response to poly(Glu60, Ala30, Tyr10) (GAT) presented by syngeneic B lymphoma cells in the context of Ad restriction determinants, was used as a model system to evaluate the roles of LFA-1 and L3T4 accessory molecules in antigen-specific T cell activation. Examination of the antigen requirement for optimal IL 2 responses revealed marked differences in the apparent avidity of these cells for GAT/Ad complexes. A relationship was observed between this parameter and the susceptibility of T hybridoma cells to inhibition by monoclonal antibodies directed at 5 distinct epitopes of LFA-1, and at A beta d allodeterminants. In contrast, L3T4a-specific monoclonal antibodies were found to block in a similar fashion the antigen-specific IL 2 responses of T hybridoma cells, regardless of the apparent avidity of their antigen receptors. It was also shown that both L3T4+ and L3T4- T hybridoma cells were capable of recognizing GAT plus Ad with high avidity. Thus, the quality of T cell antigen recognition appears to critically influence the involvement of LFA-1, and only to a marginal extent that of L3T4, in antigen-specific T cell activation. The implications of these findings are discussed in the context of recent data indicating that L3T4 may not only be an Ia-binding protein.

Identification of 5 topographic domains of the mouse LFA-1 molecule: subunit assignment and functional involvement in lymphoid cell interactions.

We have evaluated the serologic and T cell function inhibiting properties of 10 rat mAb reactive with the mouse LFA-1 molecule. Binding inhibition studies revealed that these mAb identified five topographic domains on LFA-1, including an immunodominant epitope region (A) defined by 6 mAb (H35-89, H68-96, H85-326, H129-37, H154-595, and H155-141) and four other spatially separate epitopes each defined by a single mAb (i.e., B, H154-266; C, H129-296; D, H154-163; and E, H155-78). Immunoprecipitation studies carried out with T cell hybridoma detergent lysate containing native or dissociated alpha and beta LFA-1 subunits permitted assignment of the epitopes A, C, and D to the alpha-chain, while expression of the epitopes B and E required homologous pairing of the alpha and beta LFA-1 subunits. These anti-LFA-1 mAb did not bind to the Mac-1 positive P388D1 cells. All the six mAb directed at epitope A inhibited, in the range of 50 to 95%, the proliferative responses of alloantigen- or soluble-antigen GAT-specific T cell clones and the cytolytic activity of I-Ak-specific CTL clones. MAb reactive with the epitopes C and D also blocked these T cell responses, although to a lesser extent. No inhibition was observed with mAb specific to epitope B, whereas the epitope E-specific mAb H155-78 potentiated control T cell responses by 20 to 40%. Suboptimal amounts of anti-L3T4 mAb H129-19 were found to synergistically enhance the T cell function inhibiting properties of mAb to LFA-1 epitopes A, C, and D. These studies reveal an unexpected diversification of LFA-1 between mouse and rat species and further the functional dissection of this molecule.

Analysis of the Thy-1 pathway of T cell hybridoma activation using 17 rat monoclonal antibodies reactive with distinct Thy-1 epitopes.

Seventeen monoclonal anti-Thy-1 antibodies (mAb) derived from LOU/M rats immunized with mouse T cell clones were used to study the role of Thy-1 in antigen-independent T cell activation. These mAb identified Thy-1.2 or monomorphic determinants and immunoprecipitated a molecule of 25-28 kDa from detergent-solubilized, 125I-labeled T cell surface proteins. Competitive cross-inhibition binding assays demonstrated that these reagents defined 3 epitope groups including either Thy-1.2 (group A) or Thy-1 monomorphic (groups B and C) determinants. Experiments using high titered culture supernatants revealed that all 6 IgG mAb defining the epitope group C, and one IgG2c mAb directed at a determinant in group A were capable of stimulating the terpolymer-L-glutamic acid60-L-alanine33-Ltyrosine10 (GAT) plus I-Ad-reactive BALB/c T cell hybridoma T14-117.9 to produce interleukin 2 (IL2) in the absence of accessory cells. Cross-linking of cell-bound rat mAb by a BALB/c anti-rat kappa chain mAb, or the presence of B cell lymphomas in the culture resulted in an increase of the Thy-1-mediated IL2 responses of this hybridoma. Some mAb from group B required antibody doses exceeding 80 micrograms/ml in order to activate T cells, while others remained nonstimulatory at any dose tested. Striking synergy in mAb-mediated T cell activation was observed when nonmitogenic doses of mAb group groups A and C were mixed in the same culture. Analysis of a panel of GAT plus I-Ad-specific T cell hybridomas revealed that these cells markedly differed in the magnitude of their IL2 responses induced by a given amount of stimulating anti-Thy-1 mAb. Such reagents also stimulated normal thymocytes to express IL2 receptor on their surface. These studies show that the epitopic specificity and the amount of anti-Thy-1 mAb, and the susceptibility of the T cell examined represent important parameters for the triggering of the Thy-1 pathway of T cell activation.