Phenotypic and functional analysis of CD4+ NKRP1A+ human T lymphocytes. Direct evidence that the NKRP1A molecule is involved in transendothelial migration

In this report, we show that among human CD4+ T lymphocytes 5-20% express the C-type lectin molecule NKRP1A. This lymphocyte subset displays a slightly more limited T cell receptor V beta repertoire than the CD4+ NKRP1A- counterpart. CD4+ NKRP1A+ T lymphocytes are characterized by a high expression of beta 1 and beta 2 integrins, thus representing a T lymphocyte subset that can possibly adhere and migrate through vascular endothelium. Indeed, resting CD4+ NKRP1A+ lymphocytes, differently from the CD4+ NKRP1A- subset, migrated across endothelial cell monolayers in a Transwell chamber system. Pretreatment of CD4+ NKRP1A+ T lymphocytes with an anti-NKRP1A monoclonal antibody (mAb) strongly reduced transendothelial migration, suggesting the involvement of the NKRP1A molecule in the transmigration process. Furthermore, cells of the NKRP1A- Jurkat CD4+ T cell line stably transfected with NKRP1A cDNA migrated more rapidly and efficiently than either untransfected or mock-transfected Jurkat cells. Finally, mAb-mediated cross-linking of NKRP1A molecules in CD4+ T lymphocytes induced the up-regulation of the lymphocyte function-associated antigen 1 Mg(2+)-binding site as well as beta 1 and beta 2 integrin chains. Altogether, these findings suggest that the NKRP1A molecule is involved in transendothelial migration of resting CD4+ T lymphocytes.

The selective engulfment of apoptotic bodies by dendritic cells is mediated by the alpha(v)beta3 integrin and requires intracellular and extracellular calcium

Dendritic cells derived in vitro from monocytes are known to be poor phagocytes. Here we show that, unlike macrophages, monocyte-derived dendritic cells indeed fail to take up opsonized particles or necrotic cells; however, apoptotic bodies are efficiently engulfed by dendritic cells. The temperature dependence and the sensitivity to cytochalasin D indicate that the apoptotic body engulfment is representative of early stages of phagocytosis. Inhibition studies with ligands for surface molecules involved in recognition of apoptotic bodies, such as vitronectin receptor, CD36 and phosphatidylserine receptor, revealed that apoptotic body engulfment by dendritic cells is mediated preferentially by the vitronectin receptor alpha(v)beta3, while all the receptors, with different efficiency, are engaged in phagocytosis of apoptotic bodies by macrophages. The interaction between apoptotic bodies and dendritic cells elicits a rise in intracellular free calcium concentration ([Ca2+]i) which is essential for the process of engulfment. Either intra- or extracellular Ca2+ buffering inhibits apoptotic body engulfment by dendritic cells and [Ca2+]i increases, indicating the involvement of both intra- and extracellular Ca2+. In contrast, Ca2+ mobilization is dispensable for macrophage phagocytosis of apoptotic bodies. The different requirements of Ca2+ in macrophages and dendritic cells is possibly due to the differential usage of phagocytic receptors (CD36 vs. alpha(v)beta3) and might reflect different fates of apoptotic bodies in the two cell types.

The RGD-containing domain of exogenous HIV-1 Tat inhibits the engulfment of apoptotic bodies by dendritic cells

OBJECTIVE

HIV-1 Tat can be released by infected cells and exert various extracellular functions on bystander cells, possibly contributing to immunodeficiency. In order to investigate whether exogenous Tat can affect antigen presentation, the effects of synthetic Tat on the function of dendritic cells displaying antigen presenting cell phenotype were studied.

DESIGN

Cultured dendritic cells were challenged with apoptotic bodies and monitored for cell engulfment and free intracellular calcium ([Ca2+]i) increase. The effect of synthetic HIV-1 Tat and its RGD-containing domain (peptide 65-80) or basic domain (peptide 46-60) on both functions was investigated.

METHODS

Dendritic cells were obtained by culture of monocytes with granulocyte-macrophage colony-stimulating factor. Apoptosis was induced in Jurkat cells by sub-lethal irradiation. Engulfment of radiolabelled apoptotic bodies by dendritic cells was obtained by a 45 min co-incubation at 37 degrees C. Non-ingested apoptotic bodies were removed and cell-associated radioactivity evaluated in a gamma-counter after cell lysis. Single cell analysis of calcium fluxes was performed by video-microscopy and ratio-imaging, after cell staining with the fluorescent calcium chelator FURA-2.

RESULTS

Apoptotic bodies were engulfed by dendritic cells: this process was accompanied by [Ca2+]i rise. Synthetic HIV-1 Tat inhibited both apoptotic body engulfment and [Ca2+]i increase. The same inhibition was obtained with the RGD-containing domain (peptide 65-80), but not with the basic domain (peptide 46-60) of Tat, suggesting the involvement of an integrin. This integrin is likely to be alpha v beta 3, since RGD-containing peptides from vitronectin, but not from fibronectin, inhibited apoptotic body engulfment. Furthermore, both HIV-1 Tat and its 65-80 peptide blocked [Ca2+]i increase due to beta 3-integrin cross-linking.

CONCLUSIONS

Our results support a role for HIV-1 Tat in decreasing the function of dendritic cells, possibly impairing antigen presentation.

NKRP1A molecule is involved in transendothelial migration of CD4+ human T lymphocytes

Among human CD4+ T lymphocytes, 5-20% express the C-type lectin molecule NKRP1A. Interestingly, CD4+ NKRP1A+ T lymphocytes express high levels of beta 1 and beta 2 integrins, thus representing a T lymphocyte subset that can possibly adhere and migrate through vascular endothelium. Indeed, resting CD4+ NKRP1A+ lymphocytes, differently from the CD4+ NKRP1A- subset, migrated across endothelial cell monolayers in a Transwell chamber system. This transendothelial migration was strongly reduced after pre-treatment with an anti-NKRP1A monoclonal antibody (mAb). In addition, the NKRP1A negative Jurkatt CD4+ T-cell line that had been stably transfected with NKRP1A cDNA, migrated more rapidly and efficiently than untransfected Jurkatt cells. Finally, mAb-mediated cross-linking of NKRP1A molecule in CD4+ T lymphocytes induced the upregulation of the LFA1 Mg2+ binding site as well as beta 1 and beta 2 integrin chains. Altogether, these findings indicate that NKRP1A molecule is involved in transendothelial migration of resting CD4+ T lymphocytes.

Physical and functional association of CD45 and CD3-TCR complex on CD1+ human thymocytes. Evidence that the engagement of CD45 molecules can prevent CD1+ thymocytes from apoptosis

In this study the effects of CD45 engagement on CD3-TCR-driven stimulation of CD1+ human immature thymocytes have been analyzed. Simultaneous cross-linking of CD45 and CD3 antigens on highly purified CD1+ thymocytes reduced the number of cells undergoing apoptosis after 16 h of in vitro culture. This cell population might represent immature thymocytes committed in vivo to die by programmed cell death (PCD). CD45 engagement could also increase the number of cycling CD1+ thymocytes; of note, the large majority (> 95%) of dividing cells expressed the CD1 molecule at the cell surface, indicating that proliferating cells were actually represented by immature thymocytes. These data suggest that the CD45 molecule might play a role in the rescue of immature thymocytes from PCD during differentiation. Along this line, we found that activation of CD1+ thymocytes via the CD3-TCR complex could be enhanced by CD45, both in terms of transcription and surface expression of IL-2R. These effects might be explained by the finding that the CD45 molecule (but not its isoforms CD45RO and RA) was physically associated with the CD3-TCR complex at the cell surface of CD1+ human thymocytes, as shown by co-precipitation and co-capping experiments. Finally, cross-linking of CD45 and CD3 antigens led to the expansion of CD3+ thymocytes co-expressing CD4 and CD8, indicating that simultaneous engagement of CD45 and CD3 molecules can block CD1+ cells at the double-positive (CD3+CD4+CD8+) differentiation stage. On the other hand, stimulation through CD3 resulted in the expansion of thymocytes showing a mature phenotype (CD3+CD4+ or CD3+CD8+). Altogether, these findings suggest that the CD45 molecule is involved both in early activation and in the regulation of CD1+ thymocyte differentiation.

Tumor necrosis factor alpha-induced vascular leakage involves PECAM1 phosphorylation

Herein we show that exposure of human umbilical vein endothelial cells to tumor necrosis factor alpha (TNFalpha) led to platelet endothelial cell adhesion molecule-1 (PECAM1) surface redistribution, disruption of cytoskeleton connections, and increased PECAM1 phosphorylation, accompanied by increased permeability to macromolecules. The in vitro use of inhibitors of tyrosine or serine-threonine kinases could prevent both PECAM1 surface redistribution and the increase in permeability induced by the cytokine. In vivo administration of lavendustin A, a natural tyrosine kinase inhibitor, protected endothelial cells from TNFalpha-dependent vascular leakage in mouse liver. We propose that the involvement of PECAM1 in TNFalpha-mediated effects on vascular permeability may depend on a dynamically regulated cytoskeletal association, related to the degree of PECAM1 phosphorylation.

A functional monoclonal antibody recognizing the human alpha 1-integrin I-domain

Alpha 1 beta 1 heterodimer is a member of the integrin receptor superfamily that has been described to be involved in cell-matrix binding through its interaction with collagens, fibronectin and laminin. The alpha 1 integrin belongs to a subset of I-domain containing integrins that includes alpha M, alpha L, alpha X and alpha 2. In this study we describe an anti-alpha 1 mAb (FB12) that recognizes an epitope located in the human alpha 1 I-domain, since the mAb can bind to human, but not to rat, recombinant I-domain GST fusion protein. FB12 mAb efficiently and specifically inhibits the binding of activated human lymphocytes to laminin, collagen and fibronectin. These data support the notion that the alpha 1 I-domain itself has an important role in receptor-ligand binding. In particular, we show that alpha 1 integrin-dependent lymphocyte adhesion to fibronectin is I-domain mediated, at variance with the RGD-dependent adhesion which seems to be mediated by the beta 1 rather than the alpha 1 integrin chain. Lastly, the overexpression of the alpha 1-integrin by stromal cells and blood vessels of solid tumors may suggest a role for this integrin in tumor biology.

Drug-induced in vitro inhibition of neutrophil-endothelial cell adhesion

1. Leukocyte-endothelial cell interactions play an important role during ischaemia-reperfusion events. Adhesion molecules are specifically implicated in this interaction process. 2. Since defibrotide has been shown to be an efficient drug in reducing damage due to ischaemia-reperfusion in many experimental models, we analysed the effect of defibrotide in vitro on leukocyte adhesion to endothelial cells in basal conditions and after their stimulation. 3. In basal conditions, defibrotide (1000 micrograms ml-1) partially inhibited leukocyte adhesion to endothelial cells by 17.3% +/- 3.6 (P < 0.05), and after endothelial cell stimulation (TNF-alpha, 500 u ml-1) or after leukocyte stimulation (fMLP, 10(-7) M), it inhibited leukocyte adhesion by 26.5% +/- 3.4 and 32.4% +/- 1.8, respectively (P < 0.05). 4. In adhesion blockage experiments, the use of the monoclonal antibody anti-CD31 (5 micrograms ml-1) did not demonstrate a significant inhibitory effect whereas use of the monoclonal antibody anti-LFA-1 (5 micrograms ml-1) significantly interfered with the effect of defibrotide. 5. This result was confirmed in NIH/3T3-ICAM-1 transfected cells. 6. We conclude that defibrotide is able to interfere with leukocyte adhesion to endothelial cells mainly in activated conditions and that the ICAM-1/LFA-1 adhesion system is involved in the defibrotide mechanism of action.

Dissection of lymphocyte function-associated antigen 1-dependent adhesion and signal transduction in human natural killer cells shown by the use of cholera or pertussis toxin

The effect of the guanosine triphosphate-binding protein (G-protein) inhibitors cholera toxin (Ctx) and pertussis toxin (Ptx) has been analyzed on lymphocyte function-associated antigen 1 (LFA-1)-dependent adhesion and signal transduction in human natural killer (NK) cells. Ctx, but not Ptx, inhibited the LFA-1-dependent adhesion of NK cells to tumor target cells which constitutively express the intercellular cell adhesion molecule-1 (ICAM-1) and to NIH/3T3 mouse fibroblasts stably transfected with human ICAM-1. This effect was detectable only by the use of the entire Ctx but not of the Ctx B subunit. In addition, Ctx could inhibit both NK cell binding and spreading to purified ICAM-1 protein. NK cell treatment with Ctx modified neither the surface expression of LFA-1 nor its Mg2+ binding site. These findings, together with the absence of any detectable effect of Ctx on the constitutive phosphorylation of LFA-1 alpha, suggests that this toxin modifies the avidity of LFA-1 for ICAM-1 by acting on LFA-1-cytoskeletal protein association. Unlike Ctx, Ptx did not affect NK cell adhesion. The effects of Ctx and Ptx are unlikely to depend on intracellular levels of cyclic adenosine 3',5'-monophosphate (cAMP), since a strong increase of cAMP was induced by both toxins. Moreover, this was confirmed by the observation that the LFA-1-dependent adhesion was not inhibited by the adenylate cyclase activator forskolin (FSK), the phosphodiesterase inhibitor isobutyl-1-methylxanthine (IBMX), or both, which increase intracellular cAMP levels. Unlike the differential effect on cell adhesion, both the intracellular calcium [Ca2+]i increase and phosphoinositide breakdown mediated via LFA-1 were consistently inhibited in a dose-dependent manner by both Ctx and Ptx. Also in this case, the inhibitory effect did not depend on an increase of intracellular cAMP as indicated by NK cell treatment with FSK, IBMX, or both. Further evidence of the involvement of G-proteins in LFA-1-mediated signal transduction was the inhibitory effect of the GDP analog guanosine-5'-O-2-thiodiphosphate (GDP beta S) on LFA-1-mediated calcium mobilization. Taken together, our data provide evidence that the LFA-1-mediated NK cell adhesion and signal transduction are partially independent phenomena which may be regulated by different G-proteins.

CD31/PECAM-1-driven chemokine-independent transmigration of human T lymphocytes

We assessed the relative contribution of CD31/PECAM-1 (platelet-endothelial cell adhesion molecule-1) to T lymphocyte transmigration by the use of transfected murine fibroblasts stably expressing either the human CD31/PECAM-1 or the intercellular adhesion molecule-1 (CD54/ICAM-1). Unlike CD54/ICAM-1, CD31/PECAM-1 supported migration of activated T cells in the absence of chemokines: most of the migrating lymphocytes were CD31+ and displayed a phenotype corresponding to the naive subpopulation (LFA-1 dull and CD45RA+). Migration of activated T lymphocytes through CD54/ICAM-1+ transfected monolayers could be induced by creating a chemotactic gradient with the chemokine monocyte chemotactic protein-1, and the migrating cells mainly displayed a memory phenotype (LFA-1 bright CD45RO+) under these conditions. Furthermore, we found that in transfected cells CD54/ICAM-1 is uniformly distributed along the apical surface of the cells, while CD31/PECAM-1 is concentrated at the intercellular junctions, suggesting the existence of a haptotactic gradient (i.e. a gradient of substrate- or cell-bound molecules) responsible for T cell migration. This was also confirmed by the finding that monolayers of murine fibroblasts transfected with a CD31/PECAM-1 mutant lacking the cytoplasmic domain (CD31/PECAM-1-delta cyto), which has a reduced tendency to localize at cell-cell contact areas, supported efficient adhesion but were unable to induce migration of activated T cells unless a chemotactic gradient was created. We propose that in lymphocytes, homophilic CD31/PECAM-1 adhesion may be primarily involved in transmigration of naive T cells and that its role is complementary to that of CD54/ICAM-1.

CD31-triggered rearrangement of the actin cytoskeleton in human natural killer cells

In this report, we analyze whether CD31, also known as platelet-endothelial cell adhesion molecule-1 (PECAM-1), can transduce an outside-in signal in human natural killer (NK) lymphocytes in vitro. We show that CD31, but not HLA class-I cross-linking triggers an outside-in transmembrane signal in NK lymphocytes, mediating cell spreading and cytoskeletal rearrangement. These phenomena are Mg2+, but not Ca2+ dependent, suggesting that signal transduction elicited by CD31 cross-linking may involve an associated integrin. Two possible candidates would be alpha v and alpha L, whose function is known to depend on Mg2+. However, the CD31-induced cytoskeletal rearrangement was not reduced by the use of alpha v- or alpha L-specific F(ab')2, suggesting that CD31 could transduce a signal by itself or by association with a still-undefined integrin. Moreover, talin, but not vinculin or tubulin, appears to co-localize with actin microfilaments in the membrane ruffles of NK cells that undergo cytoskeleton rearrangement following CD31 cross-linking. Both spreading and cytoskeletal rearrangement appear to be regulated by intracellular cyclic-3',5'-adenosine monophosphate (cAMP). Indeed, the activator of the adenylyl cyclase, forskolin, inhibited cell spreading and cytoskeletal rearrangement induced by CD31 cross-linking. This phenomenon was also observed using the membrane-permeants cAMP analog Sp adenosine-3', 5' -cyclic monophosphothioate (Sp-cAMPS), but not its inactive isomer Rp-cAMPS. Likewise, adhesion of NK lymphocytes to NIH/3T3 murine fibroblasts transfected with the cDNA encoding human CD31 was blocked by increasing intracellular cAMPS levels. We suggest that intracellular cAMP may be involved in CD31-mediated signal transduction, and may regulate NK-endothelial cell adhesion and possibly, the tissue localization of NK cells.

The platelet endothelial cell adhesion molecule-1 (PECAM1) contributes to endothelial barrier function

In this study we have analyzed the role of the platelet-endothelial cell adhesion molecule-1 (PECAM1) in vascular barrier function. PECAM1 is an immunoglobulin gene superfamily member expressed by endothelial cells at the cell boundaries. Macromolecule permeability assays performed on cell monolayers that express native or transfected PECAM1, indicated that the molecule participates in the establishment and maintenance of vascular barrier function in vitro. This hypothesis was confirmed by the finding that in vivo injection of the specific monoclonal antibody directed against the murine vascular PECAM1 led to a detectable leakage of hepatic and renal blood vessels.

Inducible nitric oxide synthase modulates fibronectin production in the EA.hy926 cell line and cultured human umbilical vein endothelial cells

We wished to test the hypothesis of a connection existing between inducible nitric oxide (NO) synthesis and production of extracellular matrix proteins in endothelial cells (EC). We recently reported that the inducible-NO pathway contributes to cytokine-induced enhancement of tumor cell (TC) adhesion to cultured vascular endothelium, independent of changes in E-selectin expression on endothelial cells (EC). We now show that inducible NO-synthase is involved in enhancing fibronectin production by EC. Indeed, fibronectin synthesis and secretion increased both in the EA.hy926 EC line and in human umbilical vein EC (HUVEC) after prolonged exposure to tumor necrosis factor-alpha (TNF-alpha) or interferon-gamma (IFN-gamma). This effect was reversed by the reported inhibitor of NO synthase N omega-nitro-L-arginine methyl ester (L-NAME 10(-5) M). The two cytokines exerted no additive effect, suggesting that they trigger a common metabolic pathway. NO production by cytokine-stimulated EC was dependent on the inducible NO-pathway, as demonstrated by studies of EC-dependent inhibition of platelet aggregation. This inhibition was also evident in calcium-free medium and was reversed by L-NAME and by two inhibitors of protein synthesis that are reported to block the inducible-NO synthase, such as dexamethasone (Dex 10(-7) M) and cycloheximide (Chx 10(-6) M). We conclude that modulation of the inducible NO-synthase may regulate matrix protein production by vascular endothelium during inflammation.

Expression of N-CAM by human renal cell carcinomas correlates with growth rate and adhesive properties

In the present study we provide evidence for the involvement of N-CAM in the spreading of human renal cell carcinomas (RCC) through the interaction with the subendothelial matrix. We found that in tumor cell lines derived from human RCC the increase of growth rate and the loss of adhesiveness to inert substrate were accompanied by N-CAM expression and by the appearance of specific binding to endothelial heparan sulfate. Indeed, the adhesion of tumor cells to human endothelial cells and heparan sulfate in vitro was inhibited by monoclonal antibodies able to bind and inactivate N-CAM and was abrogated by endothelial cell treatment with heparitinase. Furthermore, when the renal epithelial cell line COS7 was transfected with a cDNA coding for N-CAM a significant increase in the ability to bind both endothelium and heparan sulfate in vitro was observed. Of note, HS complexed with epithelial growth factor could enhance the proliferation of RCC-derived tumor cells; this effect was also achieved by cross-linking of N-CAM at the surface of tumor cells, suggesting that N-CAM could transduce an activation signal across the cell membrane. This was also supported by the finding that N-CAM cross-linking induced a strong calcium mobilization from internal stores and opening of surface calcium channels in such tumor cells. N-CAM was detectable in vivo at the tumor site in the areas of active proliferation, as judged by the coexpression of Ki67 nuclear antigen, and heparan sulfate was present in the wall of blood vessels in the proximity of the tumor. These findings would suggest that growing kidney tumors might use N-CAM to bind the subendothelial matrix and complexed growth factors during tissue invasion and spreading.

Distinct pattern of HSP72 and monomeric laminin receptor expression in human lung cancers infiltrated by gamma/delta T lymphocytes

We have previously demonstrated that gamma/delta T lymphocytes may participate in the host immune response against lung adenocarcinomas. Here we show that, in about one-fourth of human lung cancers, gamma/delta T cells represented a significant proportion of freshly isolated tumor-infiltrating lymphocytes. Moreover, these cells selectively expand in vitro upon culture in the presence of IL-2, thus suggesting a prior activation in vivo. Finally, when we evaluated the expression of heat shock proteins and of a panel of tumor-associated antigens in lung cancers infiltrated by gamma/delta vs. alpha/beta T cells, we found that the former displayed a distinct antigenic pattern, characterized by over-expression of HSP72 and of the 67-kDa high-affinity laminin receptor, which might account for gamma/delta T-cell recognition.

T-cell receptor V delta gene usage by tumour reactive gamma delta T lymphocytes infiltrating human lung cancer

In seven human adenocarcinomas and a non-neoplastic granulomatous disease of the lung, gamma delta+ infiltrating lymphocytes (TIL) could be isolated and selectively expanded in vitro upon culture in interleukin-2 (IL-2), without any additional stimuli, indicating a prior activation in vivo. In most cases gamma delta TIL were predominantly V delta 1+, despite a normal V delta 2:V delta 1 ratio in paired peripheral blood lymphocytes, suggesting a possible expansion of this subset in response to localized antigens/superantigens. Moreover, in five patients it was possible to identify a V delta 1- V delta 2- TIL population which by polymerase chain reaction (PCR) analysis was shown to be heterogeneous as V delta gene usage, inclusive of V delta 3,4,5,6,7 and 8. Of note, these V delta regions have not been found in peripheral blood so far. Finally, in all cases, gamma delta TIL displayed killing activity of the autologous tumour, which appeared to be more restricted in the case of V delta 1+ cells. Altogether, these findings suggest a preferential expansion, at the tumour site, of V delta 1+ cells and of cells expressing V delta genes other than V delta 2.

Involvement of CD56/N-CAM molecule in the adhesion of human solid tumor cell lines to endothelial cells

Tumor-endothelial cell adhesion has been investigated, as this event might represent one of the first steps in the generation of tumor metastasis. We focused our attention on the N-CAM homologous CD56 (besides the CD44, CD15, and CDw49d antigens), as we have recently reported that it can mediate lymphocyte-endothelial cell adhesion [11]. Herein we show that CD56 is involved in tumor cell-endothelial cell binding. This phenomenon is independent of Ca2+ and unlikely to be influenced by the degree of CD56 sialylation.

Unusual expression and localization of heat-shock proteins in human tumor cells

It has been suggested that members of HSP families represent the surface target of immune responses leading to tumor rejection in mice. Here we report that tumor cells, compared with normal cells, constitutively expressed 2- to 10-fold higher levels of intracellular HSP90. Moreover, in the absence of environmental stress, 2 lines (out of 6) expressed the "inducible" HSP72, which was also detectable in fresh tumor cells. HSP72 expression was not regulated during the cell cycle, in contrast with what has been observed with normal cells. Both HSP90 and HSP72 proteins exhibited a heterogeneous pattern of intracellular distribution in most cells, HSP72 being confined mainly to the nuclear compartment. Finally, we could detect both HSP90 and, to a lesser extent, HSP72 (that are generally believed to be located intracellularly) at the surface of some tumor cell lines. We conclude that tumor cells differ from normal cells in their pattern of HSP expression; this might imply a role of HSPs in eliciting an immune response against cancer.

Antigen-independent pathways of T-cell activation are functional in human immature thymocytes

The signal requirements for proliferation of CD1+CD3- immature thymocytes have been studied in order to define whether this immature cell population could function despite the lack of the CD3/T-cell receptor complex. We found that CD1+CD3- cells proliferate upon stimulation with anti-CD28 monoclonal antibody as well as with a pair of anti-CD2 monoclonal antibodies in the presence of low doses (0.5 ng/ml) of phorbol-13-myristate-12-acetate and/or recombinant interleukin-2. A minor fraction of CD3+ cells (15%-20%) was also present in the proliferating cell population originating from CD1+CD3- thymocytes stimulated with phorbol-13-myristate-12-acetate and recombinant interleukin-2, either in the presence or in the absence of specific monoclonal antibodies. We further observed that the anti-CD3 monoclonal antibody did not induce the proliferation of CD1+CD3- cells, as expected, and efficiently triggered unfractionated or CD1+CD3+ thymocytes only if exogenous recombinant interleukin-2 was provided. Unexpectedly, we noted that highly purified (greater than 99%), CD1+CD3- immature thymocytes could mobilize calcium via CD3, besides CD2 and CD28 surface molecules, suggesting that a minor undetectable fraction (less than 1%) of CD3+ cells was still present in the purified CD3- population. Nevertheless, the preferential expansion of CD3-CD8+ cells (about one-third of proliferating cells) after triggering via CD28, and to a lesser extent via CD2, support the notion that the alternative pathways of T-cell activation are actually functional in CD1+CD3- immature thymocytes.

Cultured human thymocytes lacking CD2 and CD11a/CD18 antigens are functional and adhere to endothelial cells via CD56 or CDw49d molecules

The preferential growth of CD3-CD2-CD11a/CD18- thymocytes was obtained by stimulation of CD2-CD3- thymic cells with low doses of PMA (0.5 ng/ml) and subsequent culture in the presence of recombinant interleukin-2 (100 U/ml). After 2-3 weeks, CD3-CD2-CD11a/CD18- thymocytes represented 40-60% of the total proliferating cells. Highly purified CD3-CD2-CD11a/CD18- cell populations were obtained by depletion of the CD11a/CD18+ thymocytes by immunomagnetic beads. Moreover, these populations proliferated for 2-5 weeks and did not change their surface phenotype. It is of note that these cells, despite the lack of CD2 and CD11a/CD18 adhesion molecules, could bind to umbilical vein endothelial cells as efficiently as did CD3+CD2+CD11a/CD18+ thymocytes. Furthermore we demonstrate that (a) CD56 molecule is involved in the adhesion of CD3-CD2-CD11a/CD18- thymic cells, but not of peripheral CD3-CD56+ lymphocytes, to untreated or IFN-gamma- and/or TNF-alpha-treated endothelium, (b) anti-CDw49d mAb could inhibit the adhesion of this thymus-derived population to either IFN-gamma- or TNF-alpha-treated endothelial cells but not to untreated endothelium, and (c) CD56 antigen expressed by these cultured thymocytes has a sialic acid content different from that of peripheral lymphocytes. Indeed, isoelectrofocusing analysis showed that CD56 molecule expressed on CD3-CD2-CD11a/CD18- thymocytes displayed an isoelectric point (pI 5.0) different from that of CD56 antigen expressed by peripheral NK cells (pI 4.7 and 5.4). Further, we noted that CD56 antigen showed the same pI 5.8 after desialylation obtained using neuraminidase treatment. Finally, CD3-CD2-CD11a/CD18- thymocytes mobilized Ca2+ and released TNF-alpha and IFN-gamma after treatment with lectins.

Signalling in human tumour infiltrating lymphocytes: The CD28 molecule is functional and is physically associated with the CD45R0 molecule

The CD28 T cell activation pathway was functional in human tumour infiltrating lymphocytes (TIL) and can induce strong proliferation, lymphokine release and calcium mobilisation. Conversely, TIL responded poorly to stimulation via CD2, and CD28 did not synergise with CD2, which is at variance with that observed using peripheral lymphocytes from the same patients. On stimulation with anti-CD28 the monoclonal antibody, most TILs, which were CD3+, CD28+ and CD45R0+ at the beginning of culture, co-expressed both high (CD45RA) and low (CD45R0) molecular weight isoforms of CD45. CD28 was associated with the CD45R0 isoform at the cell surface of activated TIL, as demonstrated by immunoprecipitation and immunoenzymatic assay. Thus CD28 can substitute for CD3 in TIL leading to the expansion of functional lymphocytes and to the amplification of antitumour immune response.

Involvement of nitric oxide in tumor cell adhesion to cytokine-activated endothelial cells

The aim of the present experiments was to test the possible involvement of nitric oxide (NO) in cytokine-induced enhancement of tumor cell (TC) adhesion to endothelial cells (ECs). Exposure of EA hyb 926 cells to TNF (500 U/ml) plus IFN (100 U/ml) for 24 h significantly enhanced their adhesivity for the 51Cr-labeled GLC1 (small cell lung carcinoma) TCs. Conversely, exposure of TCs to cytokines did not result in an increased adhesion of these cells to ECs. TC-stimulated adhesion to EA hyb 926 was abrogated by the glucocorticoid dexamethasone (Dex, 10(-7) M), the NO synthase inhibitors N omega-nitro-L-arginine methyl ester (L-NAME, 10(-5) M) and NG-monomethyl-L-arginine (L-NMMA, 10(-5) M) and the protein synthesis inhibitor cycloheximide (Cex, 10(-6) M). Furthermore, GLC1-stimulated adhesion to EA hyb 926 was reversed following removal of L-arginine from the medium or pretreatment with the guanylate cyclase inhibitor methylene blue. TC-stimulated adhesion was also prevented when TCs were pretreated with the monoclonal antibody CD15 directed against the endothelial-leukocyte adhesion molecule (ELAM-1) ligand or following exposure of ECs to anti-ELAM-1 monoclonal antibody. Although suppressing TC-stimulated adhesion, L-NMMA failed to modify significantly cytokine-induced ELAM-1 expression in EA hyb 926. These results (a) provide evidence for the NO-inducible pathway contributing to cytokine-induced enhancement of tumor cell adhesion to the vascular endothelium and (b) demonstrate the involvement of the ELAM-1/CD15 adhesion system in tumor cell-stimulated adhesion to ECs.