Inhibition of human lymphocyte-mediated mitogen-induced cytotoxicity of murine L-929 cells by heterologous anti-human lymphotoxin antisera in vitro.

Heterologous anti-human lymphotoxin (LT) antisera have been employed to investigate the role of LT in mitogen-(Con-A, PHA) induced destruction of murine L-929 cells by human lymphocytes in vitro. These various antisera will effectively neutralize human LT molecules associated with the stable (70 to 90,000 dalton) alpha-LT class of cytotoxin (anti-alpha-LT), the more unstable (35 to 50,000 dalton) beta-LT class of cytotoxins (anti-beta-LT), and antisera which will neutralize all classes of these cytotoxins in vitro, anti-whole supernatant (anti-W.S.). These anti-LT sera will greatly inhibit lysis of L-929 cells by using mitogen-activated human effector lymphocytes in vitro. This blocking was shown to be mediated by whole serum, purified IgG, or IgG-Fab fragments, which had been extensively absorbed with bovine serum, human serum, mitogens, and normal human lymphocytes. Inhibition of lysis was not apparently due to interference with either lymphocyte-target cell contact or lymphocyte activation step(s). The blocking effects of these sera were also shown to occur during the lymphocyte-independent phase of the lytic reaction. These data support the concept that the lymphocyte deposits an LT-like effector molecule on the target-L cell surface during the lymphocyte-dependent phase, which mediates cell lysis at a later time during the lymphocyte-independent phase.     

The human LT system. V. A comparison of the relative lytic effectiveness of various MW human LT classes on 51Cr-labeled allogeneic target cells in vitro: enhanced lysis by LT complexes associated with Ig-like receptor(s).

The present studies examine the in vitro cell-lytic capacity of various molecular weight (MW) human lymphotoxin (LT) classes obtained from lectin-activated normal or immune lymphocytes on allogeneic target cells. The findings reveal that the high-MW complex class of LT is up to 100 times more effective than the smaller MW LT forms (α, β, and γ) in causing lysis of various allogeneic cell types including lymphoid cells in vitro. Moreover, the data suggest that lectin-stimulated alloimmune cells (MLC sensitized) release complex LT forms in association with a specific antigen-binding receptor(s), and that these complexes are from 3 to 10 times more effective on the sensitizing target cell than complexes obtained from lectin-stimulated nonimmune cells. Positive evidence that complex-induced lysis involved LT was indicated by the finding that lysis was completely neutralized by incubation with heterologous antisera directed against a refined human α₂-LT subclass (anti-α₂) and partially neutralized with anti-human Fab₂′ serum. These findings support the concept that LT molecules may represent a system of related cell-lytic molecules. While the smaller MW forms are only weakly lytic by themselves, they can be assembled into highly lytic complexes which may be focused or directed by an antigen-binding receptor(s).     

Relationship of CIF, LT, and PIF released in vitro by activated human lymphocytes. II. A further functional comparison of LT and PIF activities on HeLa and L-929 target cells.

Lymphotoxin (LT), and proliferation inhibition factor (PIF) activities found in 5-day supernatants of mitogen-activated human lymphocytes (SAL) were further compared. In agreement with previous results, the activities could not be distinguished functionally. Quantitative differences in the amount of activity detected in the SAL could be accounted for on the basis of target cell differences, concentration of the lymphocyte effector molecules in the supernatant, and the parameter employed to assess cell function. Growth inhibitory activity detected at high supernatant dilutions was completely reversible, whereas the cytotoxic activity detected at low supernatant dilutions was irreversible. When the active medium was fractionated on DEAE, two peaks of inhibitory activity were detected. Depending upon the amount of activity and target cell, both peaks of activity were growth inhibitory or cytotoxic. Since both peaks of material affected HeLa and L-929 cells, the materials were not species specific. Thus, it appears that cloning inhibitor factor, LT, and PIF activities may actually be measures of the same stable materials found in 5-day activated lymphocyte supernatants.     

The human LT system. II. Immunological relationships of LT molecules released by mitogen activated human lymphocytes in vitro.

Materials with LT activity present in supernatants from PHA stimulated human lymphocytes in vitro are very heterogeneous and can be separated into multiple molecular weight classes, termed complex, α, β, and γ. Several of these classes can be further resolved into subclasses by other physical and chemical methods. The immunologic relationships of these materials one to another were examined employing various rabbit anti-humn LT sera which will neutralize LT activity on L-929 cells in vitro. These studies reveal: (a) LT activities are due to a distinct group of substances which are immunologically related one to another and can exist in several molecular weight forms; (b) a high MW class of molecules, termed complex, appears to contain all currently known LT classes and subclasses; (c) LT classes and subclasses both have common (public) and discrete (private) antigenic specificities; (d) human LT classes and subclasses do not appear to share Ag determinants with materials with LT activity released by lectin stimulated lymphoid cells from rabbit, rat, hamster, guinea pig, or mouse; and (e) human LT molecules are not immunologically related to cell toxins released by glass adherent human peripheral blood monocytes or PMN cells. These data indicate human LT molecules form a “discrete system” of lymphocyte derived cell toxins, which can associate together into various related but different MW forms in the supernatant.     

A physicochemical and immunologic comparison of the cell growth inhibitory activity of human lymphotoxins and interferons in vitro.

The physicochemical, immunologic, and biologic relationships between humam lymphotoxins (LT) and interferons (IF) present in supernatant fluids from lectin-stimulated peripheral blood lymphocytes (PBL) and a continuous B-lymphoblastoid cell line (PGLC-33h) were analyzed. LT activity obtained from lectin-activated PBL could not be resolved from IF activity by gel filtration chromatography. LT activity eluted in multiple peaks of activity at 70 to 90,000, and 40 to 50,000 m.w., characteristic of alpha and beta LT, respectively. IF activity in these supernatant fluids eluted as a broad band between 35 and 80,000 m.w., also suggestive of molecular heterogeneity. In contrast, this m.w. heterogeneity was not observed in LT and IF activities obtained from the PGLC-33h cell line. LT and IF eluted as separate peaks of activity at 90,000 and 25,000 m.w., respectively. In addition, acid and heat lability of PGLC-33h IF suggested similarity to type II IF. Immunologic studies, with a rabbit anti-alpha class serum that neutralized LT activity from both PBL and PGLC-33h, did not affect IF activity from either of these sources. Supernatant fluids from PGLC-33h cultures were also capable of inhibiting the proliferation of HeLa cells in vitro. The growth inhibitory activity was attributed to LT- and IF-like molecules. This evidence suggests that although cytotoxic and anti-viral activities were due to separate molecules, LT and IF have overlapping biologic activities in their ability to inhibit the proliferation of cells in vitro.     

Inhibition of human NK-induced cell lysis and soluble cell-lytic molecules with anti-human LT antisera and various saccharides

The present study examines and compares the cytolysis of K-562 and MOLT-4 cells mediated by human natural killer (NK) cells from fresh peripheral blood and lymphotoxins (LT) derived from human lymphoid cell populations after lectin stimulation in vitro. Lymphotoxins were obtained from 5-hr concanavalin A (Con A)-restimulated human peripheral blood lymphocytes (PBL) which were precultured for 5 days in medium and fetal calf serum or with allogeneic human B-lymphoid cell lines. Two classes of probes were employed in both direct (cell) and indirect (supernatant) induced target-cell lysis: (a) various saccharides and (b) antibodies reactive with human LT forms. Two sugars, N-acetylglucosamine and α-methylmannoside, were able to inhibit direct cell lysis of both MOLT-4 and K-562 target cells. However, saccharide inhibition was distinct for each type of target even when effector cells were obtained from the same donor. These same saccharides were also able to inhibit 20–30% of the total LT activity in a supernatant for L-929 cells and 50–90% of the lytic activity on MOLT-4 cells. Anti-human F(ab′)₂ (IgG) and rabbit anti-α₂ LT sera blocked direct cell lysis of MOLT-4 and K-562 targets in 50% of the experiments. The anti-α₂ LT serum only recognizes a portion of the LT forms in these supernatants. These results reveal that, while both direct and indirect cell lysis are complex phenomena, they may both occur in some cases by a common mechanism(s).     

Identification of human lymphocyte-derived lymphotoxins with binding and cell-lytic activity on NK-sensitive cell lines in vitro

Supernatants obtained from lectin-restimulated, preactivated, human peripheral blood lymphocytes rapidly released (5–24 hr) high levels of lymphotoxin (LT) activity in vitro. Peripheral blood lymphocytes were preactivated by coculturing with either fetal calf serum or with allogeneic continuous B-cell lines (LCCL) which were treated with mitomycin C. These Supernatants contained a population of L-929 cell-lytic LT forms which also selectively bind to the NK-sensitive K-562 cell. However, lytic LT forms for L-929 cells from cPBL and LCCL cultures did not bind to the NK-sensitive MOLT-4 or NK-resistant Raji cells. Additional studies reveal these supernatants contain a second set of LT forms which have cell-binding and cell-lytic activity detectable on MOLT-4 and K-562 cells in a 12 to 18 hr ⁵¹Cr-release assay. Cell-lytic form(s) for the MOLT-4 and K-562 cells were not stable for more than a week at ?20°C. These findings indicate that materials with LT activity are heterogeneous with respect to their capacity to recognize common and discrete cell-surface components on different types of target cells in vitro.     

Mechanisms of lymphocyte-mediated cytotoxicity. I. The effects of anti-human lymphotoxin antisera on the cytolysis of allogeneic B cell lines by MLC-sensitized human lymphocytes in vitro

Goat and rabbit anti-human lymphotoxin sera, IgG and F(ab')2 reagents were investigated for their capacity to effect a specific alloimmune lymphocyte-mediated cytotoxic reaction. The cytotoxic reaction employed human peripheral blood or adenoid lymphocytes sensitized in MLC to allogeneic B lymphocyte cell lines and lysis was measured in a short-term 51Cr-release assay. A polyspecific anti-LT sera (anti-WS), made against unfractionated whole supernatants from lectin-activated lymphocytes and its IgG and F(ab')2 fragments, was found to be a potent inhibitor of this reaction when the anti-WS reagent was present throughout the assay period. Absorption studies indicated the anti-WS was inhibiting cytolysis at the level of effector cell or its products. Two broadly defined antibody specificities were involved in the cytolytic-inhibitory activity of the polyspecific anti-LT; i) antigens present on the normal lymphocyte cell surface; and ii) lymphocyte surface antigens associated with activated cells. These results correlate with the previously defined antigenic structure of the LT Cx and alpha H classes. Anti-LT sera reactive with the smaller m.w. alpha and beta classes and subclasses were not inhibitory, although the anti-beta sera showed a moderate enhancing activity. The results indicated that several anti-LT antibody specificities may be required to inhibit alloimmune cytolysis. The results suggest LT molecules may mediate lymphocyte-induced alloimmune cytolysis as a multi-component toxin system, rather than as an individual toxin.     

In vitro depression of cellular immunity by Friend virus leukemic spleen cells.

Four distinct sublines of mouse L 929 cells (termed alpha, beta, gamma, and delta) were derived and shown to differ markedly in their in vitro sensitivity to human lymphotoxin (LT). The alpha L cell is most sensitive and is rapidly destroyed by very low dilutions of LT. This cell is 100 times more sensitive to LT than the most resistant (delta) L cell. The highly lymphotoxin-sensitive alpha cell makes it possible to reproducibly detect LT activity in as little as 0.0005 ml of supernatant medium. Additional studies revealed a direct correlation between the sensitivities of the four L cell sublines to LT and to direct cytolysis mediated by mitogen-stimulated human lymphocytes. The alpha, beta, gamma, and delta L cells were shown to be equally sensitive to antibody-mediated complement-dependent lysis, indicating that the sequence of sensitivities of these L cell sublines to the direct lymphocyte and to LT does not merely reflect a general susceptibility to cell destruction. These results lend further support to the view that lymphotoxin is an important mediator of in vitro target cell destruction by human effector lymphocytes.     

Identification of membrane-associated lymphotoxin (LT) on mitogen-activated human lymphocytes using heterologous anti-LT antisera in vitro.

Surface-associated lymphotoxin (LT) molecules have been identified on mitogen-activated human lymphocytes employing heterologous anti-α-LT serum in vitro. These membrane-associated LT molecules are present on PHA- or Con A-activated lymphocytes but do not appear to be expressed on unstimulated cells. Furthermore, these molecules were detected primarily on activated T lymphocytes, with little detectable on activated B- or null-cell populations. The removal of surface LT-bearing lymphocytes, using anti-α-LT serum + C′, does not dramatically affect the capacity of the remaining cells to release LT after mitogen restimulation. In addition, the presence of toxic molecules on the surface of activated lymphocytes suggests that these materials may be expressed in an inactive, noncytotoxic form.     

Effect of Puromycin on Guinea Pig Lymphotoxin Release and Lectin-Induced Cellular Cytotoxicity

In order to confirm the role of guinea pig lymphotoxin (GLT) in lectin-induced cellular cytotoxicity (LICC), the effect of puromycin, a potent enhancer of GLT activity, on the LICC to target L · P3 cells induced by phytohemagglutinin (PHA) was investigated under serum-free conditions. LICC was completely inhibited by puromycin, when it was added at the initiation of LICC culture, because of the inhibition of the release of GLT from the effector lymph node cells. However, LICC was markedly enhanced when puromycin was added several hours after the initiation of LICC culture. The interpretation of these facts is that GLT release can be inhibited by puromycin, but that the GLT already released exerts an enhanced cytotoxic effect on the target cells in the presence of puromycin. Enhancement of the cytotoxicity by the addition of puromycin several hours after the initiation of LICC culture was observed even after the removal of the GLT present in the supernatant, suggesting that the morphologically intact target cells were already affected by GLT in the early stages of LICC culture.     

Inhibition of the lytic phase of murine T-cell-mediated alloimmune cytotoxicity by a rat antiactivated T-cell antiserum

Antisera produced in rats by immunization with alloimmune murine C57Bl/6 anti-P815 splenic lymphocytes or purified T cells activated in vitro by coculture with phytohemagglutinincoated L-929 cells were found to inhibit the in vitro cytolytic action of in vivo and in vitro alloimmune C57Bl/6 anti-P815 cytotoxic T cells in a 4-hr chromium-51 release assay. The rat anti-murine-activated lymphocyte (anti-MAL) or antiactivated T-cell (anti-ATC) serum inhibited lysis in the absence of exogenously added complement activity and were not directly cytotoxic to CTL. Absorption of anti-MAL with target cells P815, L-929, EL-4, and normal C57Bl/6 lymphocytes removed a limited amount of the CTL-inhibitory activity. In contrast, lectin-activated alloimmune lymphocytes fully absorbed the inhibitory activity indicating these antisera preferentially recognize unique antigenic determinants associated with the activated CTL cell surface. The anti-ATC was found to block alloimmune lysis by CTL from several inbred mouse strains suggesting these antisera recognized antigenic determinants of a common lytic mechanism. A kinetic analysis of the inhibitory activity of the anti-MAL on the CTL reaction scheme revealed this antiserum inhibited lysis at a post-Ca²⁺-dependent step, presumably during the target cell lytic phase. This result suggests the rat antiserum can neutralize the CTL lytic mechanism.     

Potentiation of cytotoxic effect of lymphotoxin by anti-cancer drugs and elevated temperatures

The cytotoxic lymphokine, lymphotoxin (LT), has been shown to possess antitumor effect in vitro and in vivo. We examined the effect of the combination of partially purified LT with anti-cancer drugs and elevated temperatures on mouse transformed fibroblast cell line, L-929, and two human carcinoma of the cervix cell lines, HeLa and ME180. The cells were treated for 7 hr with Adriamycin, cisplatin, or bleomycin. These cells were then incubated for 24 hr in the presence of LT. At the end of the incubation period, cytotoxicity was measured by the neutral red dye uptake assay. There was 10- to 47-fold potentiation of cytotoxicity of LT on L-929 cells. The potentiation of cytotoxicity on human carcinoma of cervix cell lines ranged from 3- to 23-fold. L-929 cells and ME180 cells were incubated for 7 hr at 40 or 42 degrees C followed by 24 hr of incubation in the presence of LT. The elevated temperature treatment also enhanced (5- to 9-fold) the cytotoxic effect of LT. DNA, RNA, and protein syntheses of the ME180 cells was measured following incubation at 42 degrees C. It was observed that all three parameters were suppressed by incubation at this temperature. It was, therefore, possible that the repair of LT damaged cells was hampered by the elevated temperature treatment. It is suggested that LT may have a potential as an anti-tumor agent in combination with selected therapeutic drugs and hyperthermia.     

In vitro lymphocyte cytotoxicity. I. Evidence of multiple cytotoxic molecules secreted by mitogen activated human lymphoid cells in vitro.

Multiple families of cytotoxic molecules [Lymphotoxin (LT)] have been identified in phytohemagglutinin (PHA-P) activated human lymphocyte supernatants and lymphocyte homogenates, using gel filtration chromatography on Sephadex G-150. These macromolecules have molecular weights of 80–90,000, 50,000, and 10–15,000 daltons and have been termed LT₂, LT₂ and LT₃, respectively. They are secreted by cells from a variety of lympboid tissues, i.e., tonsil, adenoid, and peripheral blood. The kinetics of appearance of the cytotoxins indicate that all three are present within 16 hr after lymphocyte activation. However, while LT₁ and LT₂ persist in these cultures through day 5, LT₃ is not detectable after day 3. These molecules can also be detected when either PHA or concanavalin A are employed as the stimulating agent. Moreover, the relative amounts of LT₁, LT₂ and LT₃ activity in a given supernatant vary dramatically from culture to culture. Extracellular levels of LT accumulate and peak by 4 to 5 days in culture, however, intracellular levels of LT reach a maximum on day 3 and decrease to very low levels on day 5. Mitogen-stimulated lymphocytes at 3 days contain intracellular levels of LT which are several logs higher than that detectable in unstimulated cells. This observation suggests that both the biosynthesis and secretion of lymphotoxin is governed by a regulatory control process(es).     

Activated lymphocytes increase expression of 5-lipoxygenase and its activating protein in THP-1 cells

The aim of this study was to investigate the regulation of the5-lipoxygenase pathway of arachidonic acid metabolism by lymphocytes using the monocyte-like cell line, THP-1. When THP-1 cells were incubated over 4-7 days in 10% supernatant from lectin-activated human lymphocytes, their capacity to synthesize 5-lipoxygenase productswas significantly increased. In contrast, the supernatant fromnonactivated lymphocytes had no effect. The increase in capacity tosynthesize 5-lipoxygenase products was mimicked by the addition ofeither granulocyte macrophage colony-stimulating factor (GM-CSF) orinterleukin-3. These increases in synthetic capacity reflected increased enzymatic activity. Increased immunoreactive protein and mRNAfor the enzymes 5-lipoxygenase and 5-lipoxygenase-activating protein were also found in cells conditioned with activated lymphocyte supernatants. Furthermore, the increase in mRNA for both enzymes wasnot blocked by cycloheximide, suggesting that the effect on steady-state mRNA levels does not require the synthesis of new protein.The increase in mRNA could be reproduced by GM-CSF. We conclude thatlymphocytes can regulate the expression of 5-lipoxygenase in THP-1cells over a period of days via the release of soluble factors.

     

The role of lymphotoxin in target cell destruction induced by mitogen-activated human lymphocytes in vitro. II. The correlation of temperature and trypsin-sensitive phases of lymphotoxin-induced and lymphocyte-mediated cytotoxicity.

The in vitro destruction of phytohemagglutinin (PHA) coated Beta L cells by non-immune human lymphocytes was resolved into two distinct phases--lymphocyte dependent and lymphocyte independent. The initial or lymphocyte-dependent phase occurred within the first 2 hr and proceeded equally well at 34 and 37 degrees C. The amount of lymphotoxin (LT) secreted by PHA-activated human lymphocytes in vitro to PHA stimulation was the same at 34 and 37 degrees C. Antiserum and complement inactivation of the aggressor lymphocytes at various intervals revealed that target cell lysis was lymphocyte independent. However, the latter phase was temperature dependent, i.e., proceeding at the permissive temperature of 37 degrees C, but inhibited at the restrictive temperature of 34 degrees C. Further experiments revealed that LT-induced destruction had the same temperature sensitivity as target cell cytolysis occurring during the lymphocyte-independent step. Trypsin treatment of target cells during an early period of the lymphocyte-independent phase protected the target cell from subsequent death, indicating the aggressor lymphocyte has deposited a cytotoxic effector material on its surface. These results suggest the lymphocyte-dependent stage involves the processes required for the induction of LT synthesis and secretion. The actual cytolysis occurring during the lymphocyte-independent stage may be caused by LT or LT-like material(s) deposited on the target cell surface by the mitogen-activated human lymphocyte.     

The human LT system. I. Physical-chemical heterogeneity of LT molecules released by mitogen activated human lymphocytes in vitro.

Molecules with LT activity which can be identified in supernatants from PHA stimulated human lymphocytes by lysis of murine L-929 cell in vitro are heterogeneous. They can be separated by MW on sephadex or ultrogel columns into four separate classes: (a) complex (>200,000 daltons); (b) α (70–90,000 d); (c) β (25–50,000 d); and (d) γ (10–20,000 d). The amount of activity in a supernatant due to each class varies but is approximately: Cx—5 to 20%, α—40 to 60%, β—20 to 40%, and γ—0 to 10%. These classes differ one from another in their stability and kinetics of appearance in culture. Furthermore, they may aggregate together with the complex class under conditions of low ionic strength. Each class, except γ, can be further separated into subclasses by ion exchange chromatography and polyacrylamide gel electrophoresis. Alpha class can be separated on DEAE into the three subclasses, termed, α₁ (rf 0.25), α₂ (rf 0.37), and α₃ (rf 0.50). Alpha₂ subclasses can be further separated on phosphocellulose at pH 6.6 into α_2a (rf 0.37) and α_2b (rf 0.30). However, α₂ contains additional subclasses, which were resolved on PC columns at pH 5.5. Beta class activity can be resolved by DEAE and PAGE into two subclasses, termed β₁ (rf 0.28) and β₂ (rf 0.49). Gamma class activity was not studied, because of its instability. The complex class of LT activity is a macromolecular aggregate greater than 200,000 daltons which appears to contain smaller MW LT class(es). This study demonstrates that materials with LT activity in supernatants from PHA activated human lymphocytes in vitro are very heterogeneous.     

Spontaneous cytotoxicity (SCMC) of normal human lymphocytes against a human melanoma cell line: a phenomenon due to a lymphotoxin-like mediator.

Spontaneous cell-mediated cytotoxicity (SCMC) of normal human lymphocytes against various allogenic tumor cell lines has recently been identified as a non-T lymphocyte function. In this study evidence is presented that SCMC against a human melanoma cell line (IGR3) involves a nonspecific lymphotoxin-like mediator(s) (LT), which is rapidly produced by lymphocytes that are retained on IgG-anti-IgG columns. LT was shown to inhibit in 48-hr cultures the DNA synthesis of growing IGR3 and HeLa cell monolayers. Furthermore, in short term 51Cr-release assays using IGR3 target cells, LT increased strongly the SCMC of normal allogeneic lymphocytes, although it exhibited little cytotoxicity by itself in this assay. LT was detectable in cell-free supernatants harvested after 6 hr from co-cultures of IGR3 melanoma cells and normal effector lymphocytes, but not in supernatants from melanoma cells alone or lymphocytes alone. Lymphocyte preparations that had been passed through IgG-anti-IgG columns had lost the capacity to generate LT and were poor effectors in SCMC. However, in the presence of LT, a small proportion of null cells, which pass through IgG-anti-IgG columns, was capable of inducing strong SCMC. Absorption of the supernatants, containing LT activity, with an insolubilized rabbit-anti-human IgG antiserum did not remove the mediator, suggesting that it is not an immunoglobulin.     

Cytotoxic activities of normal cultured human T cells.

Normal human T cells grown in continued cultures in medium containing conditioned medium (CM) from PHA-stimulated lymphocytes were studied for their ability to manifest three known forms of cell-mediated cytotoxicity: lectin-induced cellular cytotoxicity (LICC), natural killer cell (NK) activity, and antibody-dependent cellular cytotoxicity (ADCC). The cultured T cells (CTC) were very effective mediators of LICC, being cytotoxic even at very low attacker-target cell ratios in the presence of different lectins, and against different types of targets. When tested without the addition of lectin, the CTC demonstrated a low degree of spontaneous cytotoxicity. This spontaneous cytotoxicity might not be due to conventional NK cells however, since the CTC failed to show significant numbers of cells with Fc receptors (FcR) for IgG, and had no detectable ADCC activity. CTC could represent a population enriched in polyclonal activated T cells with low spontaneous cytotoxicity against a variety of allogeneic target cells, which is greatly enhanced by the addition of lectins dur ing the 51Cr release assay.     

Cellular origin of cytotoxic effectors and secondary educated lymphocytes in human mixed leukocyte reaction

Human lymphocytes sensitized in vitro during a mixed leucocyte reaction (MLR) against an allogeneic-stimulating cell respond by blast transformation and generation of specific cytotoxic effector cells. Both proliferation and cytotoxicity are maximum on Days 6 and 7 of culture. On Day 14, no more dividing cells or cytotoxic cells are detected in such primary cultures. Restimulation by the specific priming cell triggers a secondary proliferative response and rapid reappearance of specific cytotoxic effector cells. The velocity sedimentation cell separation method which separates cells according to their size was applied to human lymphocytes sensitized in vitro during an MLR on Day 7 of culture. Blast cells were separated from nondividing small lymphocytes. It was shown that: (1) cytotoxic effectors generated at the peak of a primary response are exclusively present in the isolated blast population; (2) highly cytotoxic secondary effector cells are induced to reappear mainly from the blast-derived population upon restimulation; and (3) secondary educated proliferative cells mainly derive from the blast population. Conversely, the blast-depleted small lymphocyte population is operationally depleted of cells able to respond by proliferation to the priming cell while responding normally against third party control cells. HLA-D region specificity of the secondary proliferative response is suggested.