Differential surface expression and possible function of 9-O- and 7-O-acetylated GD3 (CD60 b and c) during activation and apoptosis of human tonsillar B and T lymphocytes

The disialoganglioside GD3 (CD60 a) and its O-acetylated variants have previously been described as surface molecules of human T lymphocytes of the peripheral blood system. Here we report the expression of the 9-O-, and 7-O-acetylated disialoglycans of GD3 (CD60 b and CD60 c respectively) on human tonsillar lymphocytes. CD60 b and c are surface-expressed on activated germinal centre B cells and colocalize in raft-like structures on the cell surface together with the cytoplasmic tyrosine kinase Lyn and Syk. Addition of CD60 b and c mAb together with anti-IgM/IL-4 to in vitro cultivated tonsillar B cells resulted in a costimulatory effect. During spontaneous and staurosporine-induced apoptosis a distinct population of activated annexin V+/CD60 b+/CD60 c- B cells was observed. CD60 b and c are also found on cells of the extrafollicular T cell area. On tonsillar T cells, CD60 b mAb had a costimulatory effect together with PHA while CD60 c mAb alone was sufficient to induce proliferation. In further contrast to B cells, during apoptosis a distinct CD60 b+ T cell subpopulation was not observed. Together, surface-expressed CD60 b and c are differently expressed on tonsillar B and T cells and may be involved in the regulation of activation and apoptosis of lymphocytes in secondary lymphatic tissue.     

Genes modulated by expression of GD3 synthase in Chinese hamster ovary cells. Evidence that the Tis21 gene is involved in the induction of GD3 9-O-acetylation

9-O-Acetylation is a common sialic acid modification, expressed in a developmentally regulated and tissue/cell type-specific manner. The relevant 9-O-acetyltransferase(s) have not been isolated or cloned; nor have mechanisms for their regulation been elucidated. We previously showed that transfection of the GD3 synthase (ST8Sia-I) gene into Chinese hamster ovary (CHO)-K1 cells gave expression of not only the disialoganglioside GD3 but also 9-O-acetyl-GD3. We now use differential display PCR between wild type CHO-K1 cells and clones stably expressing GD3 synthase (CHO-GD3 cells) to detect any increased expression of other genes and explore the possible induction of a 9-O-acetyltransferase. The four CHO mRNAs showing major up-regulation were homologous to VCAM-1, Tis21, the KC-protein-like protein, and a functionally unknown type II transmembrane protein. A moderate increase in expression of the FxC1 and SPR-1 genes was also seen. Interestingly, these are different from genes observed by others to be up-regulated after transfection of GD3 synthase into a neuroblastoma cell line. We also isolated a CHO-GD3 mutant lacking 9-O-acetyl-GD3 following chemical mutagenesis (CHO-GD3-OAc(-)). Analysis of the above differential display PCR-derived genes in these cells showed that expression of Tis21 was selectively reduced. Transfection of a mouse Tis21 cDNA into the CHO-GD3-OAc(-) mutant cells restored 9-O-acetyl-GD3 expression. Since the only major gangliosides expressed by CHO-GD3 cells are GD3 and 9-O-acetyl-GD3 (in addition to GM3, the predominant ganglioside type in wild-type CHO-K1 cells), we conclude that GD3 enhances its own 9-O-acetylation via induction of Tis21. This is the first known nuclear inducible factor for 9-O-acetylation and also the first proof that 9-O-acetylation can be directly regulated by GD3 synthase. Finally, transfection of CHO-GD3-OAc(-) mutant cells with ST6Gal-I induced 9-O-acetylation specifically on sialylated N-glycans, in a manner similar to wild-type cells. This indicates separate machineries for 9-O-acetylation on alpha2-8-linked sialic acids of gangliosides and on alpha2-6-linked sialic acids on N-glycans.     

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.     

M241 (CD1) expression on B lymphocytes

The human thymus leukemia-like antigens (CD1a-c) consist of three similar glycoproteins found on subpopulations of normal thymocytes, T cell acute leukemias, and cutaneous dendritic cells. The CD1c antigen recognized by the M241 monoclonal antibody was detected on the circulating mononuclear cells of three children with severe combined immunodeficiency disease (SCID). Two-color immunofluorescence analysis demonstrated that M241 expression (43 to 95%) was limited to cells expressing the B cell-restricted antigens B4 (CD19), B1 (CD20), and surface immunoglobulin. To confirm M241 expression on normal cells of the B lineage rather than aberrant expression limited to SCID B cells, its expression was demonstrated serologically and biochemically on purified B cells from spleen, tonsil, and peripheral blood. Parallel analyses with monoclonal antibodies NA1/34 and 4A76 demonstrated that the CD1a and CD1b molecules were negative on all B cells that were studied. It has been hypothesized that the CD1 molecules represent the human counterpart of the murine thymus leukemia antigens due to their similar size, limited tissue distribution, and association with beta 2-microglobulin. This study suggests that a subset of CD1 antigens detected by M241 (CD1c) may represent a human analog of a murine Qa antigen due to its extended distribution on normal peripheral B cells.     

Expression of the rat CD26 antigen (dipeptidyl peptidase IV) on subpopulations of rat lymphocytes.

The T cell activation antigen CD26 has been recently identified as the cell surface ectopeptidase dipeptidyl peptidase IV (DPP-IV). DPP-IV is found on many cell types, including lymphocytes, epithelial cells, and certain endothelial cells. The MRC OX61 monoclonal antibody (MAb) which specifically recognises rat DPP-IV was used to examine the expression of CD26/DPP-IV on rat lymphocytes. The molecular nature of the antigen was examined by immunoprecipitation from thymocytes, splenocytes, and hepatocytes. Analysis by one- and two-dimensional gel electrophoresis indicated that the native form of CD26 includes a 220-kDa homodimer. On tissue sections MRC OX61 MAb stained nearly all thymocytes and in the spleen and lymph nodes predominantly stained the T cell areas. However, in immunofluorescence experiments OX61 stained 80 to 87% of lymph node cells and 78 to 85% of spleen cells. Furthermore, two-colour immunofluorescence analysis of the CD4+, CD8+, and Ig+ lymphocyte subsets indicated that only 2 to 5% of each of these subsets lacked OX61 staining. Spleen cells and thymocytes of both CD4+ and CD8+ subsets stained much more intensely with OX61 after these cells were stimulated with phytohemagglutinin. These findings indicate that rat CD26 antigen expression is not confined to the T cell population as has been suggested, but also occurs on B cells, and is increased on T cells following their activation.     

A novel CD8 T cell-restricted CD45RB epitope shared by CD43 is differentially affected by glycosylation.

The mAb 1B11 has been characterized as recognizing the activation-associated glycoform of murine CD43, a heavily O-glycosylated protein implicated in leukocyte homing. When hemopoietic cells from CD43-/- mice were stained with 1B11, CD43-independent binding of 1B11 was observed on peripheral CD8 T cells and at low levels on thymocytes, while no binding was detected on CD4 T cells, B cells, or bone marrow cells. Levels of 1B11 staining were comparable in lymph node CD8+ T cells from both CD43-/- mice and CD43+/+ mice. We sought to identify the CD43-independent target of 1B11 expressed on CD8 T cells. Previous work had demonstrated that neuraminidase treatment of lymph node cells (LNC) enhanced 1B11 binding on CD43+/+ LNC; this enhancement was also observed in CD43-/- LNC. We show that neuraminidase-enhanced 1B11 binding in CD43-/- LNC and EL4 thymoma cells is CD43 independent and that 1B11 detects a novel target of apparent mass of approximately 200 kDa identified as a hyposialylated form of CD45RB preferentially expressed on peripheral CD8, but not CD4, T cells. Our data also show that the recognition of CD43 and CD45RB by 1B11 is differentially affected by O-linked glycosylation and sialic acid. Whereas 1B11 recognition of CD43 on activated T cells required both core 2 O-glycan branching and sialic acid, 1B11 recognition of CD45 only occurred in the absence of both core 2 glycosylation and sialic acid.     

Structural and functional characteristics of the CD3 (T3) molecular complex on human thymocytes

The CD3 (T3) molecular complex is noncovalently associated with the antigen receptor molecule on T cells. The mitogenic properties of anti-CD3 antibodies have suggested that this complex may be the transducer of the antigenic signal to the intracellular environment. In the present investigation, we studied some of the structural and functional characteristics of the CD3 complex on human thymocytes. In 11 specimens tested, we found that anti-CD3 antibodies react with 50 to 76% of the thymocytes. Two-color immunofluorescence analysis revealed that the majority (greater than 50%) of thymocytes express both CD3 and CD1 on their surfaces. The latter is a marker of immature thymocytes. However, a distinct subpopulation comprising 13 to 19% of the total cells displays only CD3, while an approximately equal percentage of cells expresses only CD1. The mitogenic potential of anti-CD3 antibodies on peripheral T cells is dependent on the presence of monocytes. Anti-CD3 antibodies by themselves cannot activate thymocytes, indicating that functionally active monocytes are absent from the thymocyte population. Even the addition of peripheral monocytes does not allow a response of thymocytes to anti-CD3 antibodies. However, when the anti-CD3 antibody 64.1 is added in the presence of exogenous rIL 2, a strong antibody and lymphokine dose-dependent response ensues. Only CD1- CD3+ thymocytes are stimulated by the addition of antibody and IL 2. The mere expression of CD3 on the CD1+ CD3+ subpopulation of thymocytes apparently is not sufficient to render the cells responsive to the signals of anti-CD3 and IL 2.     

The cytoplasmic expression of CD3 antigens in normal and malignant cells of the T lymphoid lineage

Anti-CD3 (T3) Ab reacting with different proportions of thymocytes (anti-CD3a: UCHT1, anti-CD3b: T10B9, and anti-CD3c: OKT3) were tested for cytoplasmic (cCD3) and membrane (mCD3) expression in the bone marrow, thymus, and blood in man and selected primates. The expression of cCD3a and cCD3c in the perinuclear and Golgi area of large, BrdU-incorporating, strongly TdT+ thymic blasts probably represents one of the earliest signs of T cell commitment, because these blast cells are CD1-, CD4-, CD8-, and mCD3-. The cCD3+, TdT+ cells are normally restricted to the thymus and are absent among the TdT+ cells of bone marrow. The anti-CD3b Ab used, T10B9, co-caps and co-modulates with the other anti-CD3 Ab and is a T cell-specific reagent at a membrane level but does not bind to perinuclear cCD3. Instead, this reagent cross-reacts with a filamentous cytoplasmic network in non-T cells in man and in primates S. oedipus and M. rhesus despite their T cell negativity for mCD3. The characteristics of all T-ALL cases studied: cCD3+, CD7+ along with nuclear TdT+ suggest lineage fidelity to early thymic blasts. As a marked contrast, cCD3 is absent in common ALL and in AML, including cases that concomitantly express CD7 and myeloid antigens. Thus, the cCD3, TdT combination provides a very sensitive assay for residual T-ALL blasts outside the normal thymus.     

Biosynthesis and turnover of O-acetyl and N-acetyl groups in the gangliosides of human melanoma cells

We and others previously described the melanoma-associated oncofetal glycosphingolipid antigen 9-O-acetyl-GD3, a disialoganglioside O-acetylated at the 9-position of the outer sialic acid residue. We have now developed methods to examine the biosynthesis and turnover of disialogangliosides in cultured melanoma cells and in Golgi-enriched vesicles from these cells. O-Acetylation was selectively expressed on di- and trisialogangliosides, but not on monosialogangliosides, nor on glycoprotein-bound sialic acids. Double-labeling of cells with [3H]acetate and [14C]glucosamine introduced easily detectable labels into each of the components of the ganglioside molecules. Pulse-chase studies of such doubly labeled molecules indicated that the O-acetyl groups turn over faster than the parent molecule. When Golgi-enriched vesicles from these cells were incubated with [acetyl-3H]acetyl-coenzyme A, the major labeled products were disialogangliosides. [Acetyl-3H]O-acetyl groups were found at both the 7- and the 9-positions, indicating that both 7-O-acetyl GD3 and 9-O-acetyl GD3 were synthesized by the action of O-acetyltransferase(s) on endogenous GD3. Analysis of the metabolically labeled molecules confirmed the existence of both 7- and 9-O-acetylated GD3 in the intact cells. Surprisingly, the major 3H-labeled product of the in vitro labeling reaction was not O-acetyl-GD3, but GD3, with the label exclusively in the sialic acid residues. Fragmentation of the labeled sialic acids by enzymatic and chemical methods showed that the 3H-label was exclusively in [3H]N-acetyl groups. Analyses of the double-labeled sialic acids from intact cells also showed that the 3H-label from [3H]acetate was exclusively in the form of [3H]N-acetyl groups, whereas the 14C-label was at the 4-position. Pulse-chase analysis of the 3H/14C ratio showed that the N-acetyl groups of both GD3 and of the monosialoganglioside GM3 were turning over faster than the parent molecules. Selective periodate oxidation showed that both the inner and outer sialic acid residues of GD3 incorporated 3H-label in the in vitro reaction, and showed similar turnover of N-acetylation in the pulse-chase study. Taken together, these results indicate that both the O- and N-acetyl groups of the sialic acid residues of gangliosides turn over faster than the parent molecules. They also demonstrate a novel re-N-acetylation reaction that predicts the existence of de-N-acetyl gangliosides in melanoma cells.     

Expression of inhibitory receptors Ly49E and CD94/NKG2 on fetal thymic and adult epidermal TCR V gamma 3 lymphocytes

Ly49 and CD94/NKG2 inhibitory receptors are predominantly expressed on murine NK cells, but they are also expressed on a subpopulation of peripheral CD8 memory TCR alphabeta lymphocytes. In this study we demonstrate that Ly49E and CD94/NKG2 receptors are expressed on mature TCR Vgamma3(+) cells in the fetal thymus. Expression correlated with a memory phenotype, such as expression of CD44, 2B4, and IL-2Rbeta (CD122), and absence of IL-2Ralpha (CD25) expression. No expression of Ly49A, C, D, G2, or I receptors was observed. This phenotype is similar to that of fetal thymic NK cells. Skin-located Vgamma3 T cells, the progeny of fetal thymic Vgamma3 cells, also expressed CD94/NKG2 and Ly49E but not the other members of the Ly49 family. The development and survival of Ly49E(+) or CD94/NKG2(+) Vgamma3 T lymphocytes was not dependent upon expression of MHC class I molecules. The cytotoxicity of TCR Vgamma3 cells was inhibited when Qdm, the ligand for CD94/NKG2, was presented by Qa1(b)-transfected target cells. Also, upon cross-linking of CD94/NKG2 with mAb 3S9, TCR Vgamma3 thymocytes were prevented from killing FcgammaR(+) P815 target cells. These effects were most pronounced in the CD94/NKG2(high) subpopulation as compared with the CD94/NKG2(low) subpopulation of Vgamma3 cells. Our data demonstrate that Vgamma3 T cells expressing inhibitory Ly49E and CD94/NKG2 receptors are mature and display a memory phenotype, and that CD94/NKG2 functions as an inhibitory receptor on these T lymphocytes.     

Expression of variable exon A-, B-, and C-specific CD45 determinants on peripheral and thymic T cell populations.

A mAb (I/24) has been generated that is specific for a determinant on mouse CD45 molecules. Reactivity of this mAb with a panel of CD45 transfected cell lines demonstrated that the determinant recognized is dependent upon expression of one or more CD45 variable exons and that exon C is sufficient for its expression. The exon C-specific epitope detected by I/24 is expressed at high density on essentially all B lymphocytes and at an intermediate density on the vast majority of CD8+ splenic T cells. Two distinct subpopulations of CD4+ splenic T cells were detected, a minor subpopulation that expresses this exon determinant at high density and a major subpopulation that expresses it at a much lower density. This first identification of a CD45RC-specific reagent allowed a comparison of the expression of exon A-, exon B-, and exon C-specific determinants on peripheral and thymic lymphoid populations. When splenic lymphocytes were analyzed for expression of CD45RA (reactive with mAb 14.8), CD45RB (reactive with mAb 23G2 or mAb 16.A), and CD45RC (reactive with mAb I/24) determinants, it was found that each of these CD45 determinants had a distinct pattern of expression on CD4+ and CD8+ T cells and B cells. CD45RB and RC epitopes were also detected at high density on a small proportion (0.7 to 4.1%) of thymocytes. Both CD45RB and RC epitopes were found predominantly on CD4-CD8- and CD4-CD8+ thymocytes but were also found on small numbers of CD4+CD8+ and CD4+CD8- cells. The population of thymocytes that expressed CD45RB and CD45RC determinants displayed a novel TCR CD3 phenotype characterized by a level of expression that was intermediate between that seen in the larger CD3 bright and CD3 dull populations of thymocytes.     

Deficiency of CD11b or CD11d results in reduced staphylococcal enterotoxin-induced T cell response and T cell phenotypic changes

The beta(2) integrin CD11a is involved in T cell-APC interactions, but the roles of CD11b, CD11c, and CD11d in such interactions have not been examined. To evaluate the roles of each CD11/CD18 integrin in T cell-APC interactions, we tested the ability of splenocytes of CD11-knockout (KO) mice to respond to staphylococcal enterotoxins (SEs), a commonly used superantigen. The defect in T cell proliferation with SEA was more severe in splenocytes from mice deficient in CD18, CD11b, or CD11d than in CD11a-deficient splenocytes, with a normal response in CD11c-deficient splenocytes. Mixing experiments showed that the defect of both CD11b-KO and CD11d-KO splenocytes was, unexpectedly, in T cells rather than in APC. Cytometric analysis failed to detect CD11b or CD11d on resting or activated T cells or on thymocytes of wild-type adult mice, nor did Abs directed to these integrins block responses in culture, suggesting that T cells educated in CD11b-KO or CD11d-KO mice were phenotypically altered. Consistent with this hypothesis, T cells from CD11b-KO and CD11d-KO splenocytes exhibited reduced intensity of CD3 and CD28 expression and decreased ratios of CD4/CD8 cells, and CD4(+) T cells were reduced among CD11b-KO and CD11d-KO thymocytes. CD11b and CD11d were coexpressed on a subset of early wild-type fetal thymocytes. We postulate that transient thymocyte expression of both CD11b and CD11d is nonredundantly required for normal thymocyte and T cell development, leading to phenotypic changes in T cells that result in the reduced response to SE stimulation.     

CD4+CD8- thymocytes from neonatal mice induce IgM production in SCID mice.

Defective recombination of both the TCR and Ig genes results in the absence of mature lymphocytes in mice with the scid mutation. We have shown previously that the transfer of neonatal, but not adult, thymocytes results in high levels of Ig production in 100% of C.B-17-scid (SCID) mice, in contrast to the 10 to 25% of SCID mice spontaneously producing low levels of oligoclonal Ig. In this report we demonstrate that neonatal CD4+8- thymocytes were able to induce this response; the CD4+8+ and CD4-8+ subpopulations were totally inactive and CD4-8- T cells had only limited activity several weeks after transfer. The stimulation of IgM production in SCID mice was detectable by 1 wk posttransfer of CD4+8- thymocytes or splenic T cells, and could be achieved with as few as 300 cells. The ability of neonatal CD4+8- thymocytes to induce Ig diminished gradually to insignificant levels at 3 wk postbirth; this loss of function was not associated with differential survival of neonatal T cells. Neonatal CD4+8- thymocytes from C.B-17 and other H-2d strains rescued Ig production, whereas cells from H-2b, H-2a, and H-2k strains were much less effective. These results suggest that a CD4+8- subpopulation found in both neonatal thymus and peripheral lymphoid tissues is able to induce the expansion or differentiation of the small numbers of functional B lymphocytes in SCID mice, and that the inducing T cell disappears shortly after birth, perhaps during the acquisition of self-tolerance.     

Prolonged expression of high molecular mass CD45RA isoform during the differentiation of human progenitor thymocytes to CD3+ cells in vitro.

CD45, the leukocyte common Ag, has been shown to characterize T cell development both within the thymus and among peripheral T cells. The work reported here demonstrates that human multinegative (MN) thymocytes, depleted of cells bearing CD3, CD4, CD8, and CD19, express predominantly the high molecular mass CD45RA isoform, and lack low molecular mass CD45RB isoforms and CD45R0 as detected by immunofluorescence. By immunoprecipitation of surface-labeled CD45 molecules from MN thymocytes, a proportion of the CD45 is in fact of low molecular mass but does not include epitopes recognized by CD45R0, nor by CD45RB mAb specific for the p190. This suggests either glycosylation variants of CD45RB/CD45R0 undetectable by our mAb, or underglycosylated CD45RA. MN thymocytes lack TCR-alpha beta mRNA confirming their early developmental stage. Upon culture with IL-2 or with mitogenic combinations of anti-CD2/CD28 mAb, MN thymocytes differentiate to acquire CD3, TCR-alpha beta, and in some cases CD4 and/or CD8. We have predicted that maintenance of CD45RA and lack of CD45R0 expression is fundamental to generative thymic development. If correct, this demands that unlike peripheral T cells, differentiation of MN thymocytes should be accompanied by prolonged expression of high molecular mass CD45 isoforms. Analysis of CD45 isoform expression during MN thymocyte development confirms this prediction and indicates that expression of CD45RA is maintained, at increasing density, for at least 8 to 12 days of culture. Unlike peripheral blood T cells, this is accompanied by the gradual acquisition of firstly the p190 isoforms of CD45RB and later by CD45R0, resulting in a population of CD3+TCR-alpha beta cells coexpressing CD45RA/RBp190/R0. Dot blot analysis of mRNA from differentiating MN thymocytes indicates prolonged expression of mRNA encoding CD45 exons a, b, and c, again in contrast to peripheral T cells which lose all mRNA for alternatively spliced CD45 exons within the first 24 h poststimulation. This is discussed in the context of negative selection during thymic development and interconversion of T cell subsets.     

Human T-cell leukemia virus type I-induced proliferation of human immature CD2+CD3- thymocytes.

The mitogenic activity of human T-cell leukemia virus type I (HTLV-I) is triggering the proliferation of human resting T lymphocytes through the induction of the interleukin-2 (IL-2)/IL-2 receptor autocrine loop. This HTLV-I-induced proliferation was found to be mainly mediated by the CD2 T-cell antigen, which is first expressed on double-negative lymphoid precursors after colonization of the thymus. Thus, immature thymocytes express the CD2 antigen before that of the CD3-TCR complex. We therefore investigated the responsiveness of these CD2+CD3- immature thymocytes and compared it with that of unseparated thymocytes, containing a majority of the CD2+CD3+ mature thymocytes, and that of the CD2-CD3- prothymocytes. Both immature and unseparated thymocytes were incorporating [3H]thymidine in response to the virus, provided that they were cultivated in the presence of submitogenic doses of phytohemagglutinin. In contrast, the prothymocytes did not proliferate. Downmodulation of the CD2 molecule by incubating unseparated and immature thymocytes with a single anti-CD2 monoclonal antibody inhibited the proliferative response to HTLV-I. These results clearly underline that the expression of the CD2 molecule is exclusively required in mediating the proliferative response to the synergistic effect of phytohemagglutinin and HTLV-I. Immature thymocytes treated with a pair of anti-CD2 monoclonal antibodies were shown to proliferate in response to HTLV-I, even in the absence of exogenous IL-2. We further verified that the proliferation of human thymocytes is consecutive to the expression of IL-2 receptors and the synthesis of IL-2. These observations provide evidence that the mitogenic stimulus delivered by HTLV-I is more efficient than that provided by other conventional mitogenic stimuli, which are unable to trigger the synthesis of endogenous IL-2. Collectively, these results show that the mitogenic activity of HTLV-I is able to trigger the proliferation of cells which are at an early stage of T-cell development. They might therefore represent target cells in which HTLV-I infection could favor the initiation of the multistep lymphoproliferative process leading to adult T-cell leukemia.