Analysis of myeloid-associated genes in human hematopoietic progenitor cells.

The distribution of myeloid lineage-associated cytokine receptors and lysosomal proteins was analyzed in human CD34+ cord blood cell (CB) subsets at different stages of myeloid commitment by reverse-transcriptase polymerase chain reaction (RT-PCR). The highly specific granulomonocyte-associated lysosomal proteins myeloperoxidase (MPO) and lysozyme (LZ), as well as the transcription factor PU.1, were already detectable in the most immature CD34+Thy-1+ subset. Messenger RNA (mRNA) levels for the granulocyte-colony stimulating factor (G-CSF) receptor, granulocyte-macrophage (GM)-CSF receptor alpha subunit and tumor necrosis factor (TNF) receptors I (p55) and II (p75) were also detected in this subset in addition to c-kit and flt-3, receptors known to be expressed on progenitor cells. By contrast, the monocyte-macrophage colony stimulating factor (M-CSF) receptor was largely absent at this stage and in the CD34+Thy-1-CD45RA- subsets. The M-CSF receptor was first detectable in the myeloid-committed CD34+Thy-l-CD45RA+ subset. All other molecules studied were found to be expressed at this stage of differentiation. Different cocktails of the identified ligands were added to sorted CD34+Thy-1+ single cells. Low proliferative capacity was observed after 1 week in culture in the presence of stem cell factor (SCF) + Flt-3 ligand (FL) + G-CSF. Addition of GM-CSF to this basic cocktail consistently increased the clonogenic capacity of single CD34+Thy-1+ cells, and this effect was further enhanced (up to 72.3 +/- 4.3% on day 7) by the inclusion of TNF-alpha. In conclusion, the presence of myeloid-associated growth factor receptor transcripts in CD34+ CB subsets does not discriminate the various stages of differentiation, with the exception of the M-CSF receptor. In addition, we show that TNF-alpha is a potent costimulatory factor of the very immature CD34+Thy-1+ CB subset.

Efficient retrovirus-mediated gene transfer of dendritic cells generated from CD34+ cord blood cells under serum-free conditions.

A retroviral-vector encoding the low affinity nerve growth factor receptor (LNGFR) was used to transduce dendritic cells (DCs) generated from CD34+ cord blood (CB) progenitor cells under serum-free conditions. Transduction efficiency was monitored by flow cytometry (FACS) using a specific monoclonal antibody. Prior to retroviral infections, CD34+ CB cells were stimulated for 60 h in a serum-free medium containing a DC differentiation inducing cytokine cocktail: stem cell factor (SCF), granulocyte/macrophage-colony stimulating factor (GM-CSF), tumor necrosis factor alpha (TNFalpha), and transforming growth factor beta 1 (TGF-beta1). Addition of flt3-ligand (FL) to the aforementioned growth factors significantly enhanced cell expansion (41.7+/-11.5 fold vs. 22.5+/-4.7 fold without FL) and generation of CD1a+ DCs (mean 45.7+/-9.8% vs. 28+/-6.5% without FL, n = 4,p = 0.01). Furthermore, FL significantly increased the proportion of CD1a+LNGFR+ cells (mean 10%+/-4.4% vs. 6%+/-2.4 without FL n = 4, p = 0.03). When serum-free viral supernatants were used to infect DCs progenitors under entirely serum-free conditions and with the most potent cytokine combination, approximately one-third of the CD1a+ DCs generated co-expressed the LNGFR gene. Moreover, the transduced gene was also identified in more mature CD1a+CD80+ and CD1a+CD86+ DCs after 12-14 days of culture. In addition, transduced CD1a+ DCs maintained their functional properties, stimulating allogeneic T cells with similar efficiency as nontransduced CD1a+ DCs. Thus, the serum-free system described allows efficient generation and transduction of CD1a+ DCs derived from CD34+ progenitor cells and may be very useful for future therapeutic applications of DCs.

Epidermal Langerhans cell development and differentiation.

Epidermal Langerhans cells (LC) play a critical role in host defense. Still we know rather little about the development and functional specialization of these bone marrow-derived dendritic cells (DC) located in the most peripheral ectodermal tissue of the mammalian organism. How LC develop from their primitive progenitors in bone marrow and to what extent LC are related in their development to other lineages of the hemopoietic system is still under debate. There are currently 3 major areas of debate: 1) which are the signals required for LC development and differentiation to occur, 2) what are the (molecular) characteristics of the intermediate stages of LC differentiation, and 3) how are LC related in their development and/or function to other cells of the hemopoietic system? A better understanding of LC development and answers to these questions can be expected from recently developed technologies which allow the in vitro generation of DC with the typical molecular, morphological and functional features of LC from purified CD34+ progenitor cells under defined serum-free culture conditions. TGF-beta 1 was found to be an absolute requirement for in vitro LC development under serum-free conditions upon stimulation with the classical DC growth and differentiation factors GM-CSF, TNF-alpha and SCF. The recently identified cytokine FLT3 ligand further dramatically enhanced in vitro LC development and even allowed efficient in vitro generation of LC colonies from serum-free single cell cultures of CD34+ hemopoietic progenitor cells.

Recombinant human GM-CSF enhances T cell-mediated cytotoxic function after ABMT for hematological malignancies.

The interactions of GM-CSF with cells of lymphoid lineage are not well understood and their clinical use has been focused on the acceleration of hematopoietic recovery. However, several reports have shown that human GM-CSF can affect certain T lymphocyte in vitro cytotoxic functions. To assess whether recombinant human GM-CSF (rhGM-CSF) has a more broadly based activity in the immune system, we studied its in vivo effects on endogenously-generated killer function in patients undergoing ABMT for hematologic malignancies. Eleven patients received rhGM-CSF after ABMT: eight received rhGM-CSF as a 2-h infusion daily from days +3 to +17 and three received rhGM-CSF until reaching > 500 x 10(6)/l granulocytes. Eight patients not enrolled in the rhGM-CSF therapy protocol served as controls. Natural killer (NK) cell activity and activated killer (AK) cell activity were studied before conditioning, during rhGM-CSF therapy and after withdrawal of GM-CSF. rhGM-CSF therapy does not affect NK activity. Circulating lymphocytes with the ability to kill AK-sensitive targets appear spontaneously in control ABMT patients. AK activity was 1.6 +/- 0.8% before ABMT increasing to 9 +/- 2.5% and 14 +/- 2.1% at 2 and 3 weeks after ABMT, respectively (p = 0.002). In rhGM-CSF-treated patients this phenomenon also occurs. AK activity increased from 2.4 +/- 1.5% before ABMT to 33.6 +/- 8.1% during rhGM-CSF administration (p = 0.001) and 17.5 +/- 3.4% after withdrawal (p = 0.001).(ABSTRACT TRUNCATED AT 250 WORDS)

Recombinant human granulocyte-macrophage colony stimulating factor (rhGM-CSF) administration after autologous bone marrow transplantation for acute myeloblastic leukemia enhances activated killer cell function and may diminish leukemic relapse.

Leukemic relapse is the major complication following autologous bone marrow transplantation (BMT) in acute myeloblastic leukemia (AML). Previously, we have shown that recombinant human granulocyte-macrophage colony-stimulating factor (rhGM-CSF) infusion after autologous BMT has the ability to augment endogenous activated killer (AK) cell function which may play a role in the eradication of minimal residual disease. However, the clinical application of rhGM-CSF in patients with AML has been limited by its potential stimulatory effect on the malignant clone. Here we report the effect of rhGM-CSF 5 micrograms/kg/day infusion on AK cell function in 20 patients with AML undergoing autologous BMT. AK cell function was investigated before autologous BMT, during rhGM-CSF therapy and after withdrawal. In addition, its influence on the actuarial risk of relapse is analyzed and compared with a historical control group of 20 patients transplanted immediately before initiation of this study. rhGM-CSF significantly enhanced AK cell function. During rhGM-CSF treatment, median AK cell function rose from 1.8% before autologous BMT (range 0-8%) to 35% (range 3-80%) and remained increased after cessation of rhGM-CSF (median 20%; range 0-36%; P < 0.001). After a median follow-up of 24 months, the actuarial risk of relapse is 37.4% in rhGM-CSF-treated patients compared with 49.5% in controls (P = 0.05). Interestingly, none of the 7 patients with an AK cell activity > or = 20% in the first 2-5 weeks after autologous BMT have relapsed compared with 6 of 9 patients with an AK cell activity < 20% (P < 0.02).(ABSTRACT TRUNCATED AT 250 WORDS)

Autologous bone marrow transplantation as consolidation therapy for non-Hodgkin's lymphoma patients with poor prognostic features.

Theoretical considerations and preliminary results of clinical trials support the earlier use of autologous bone marrow transplantation (ABMT) in poor prognosis non-Hodgkin's lymphoma (NHL). A prognostic analysis of 50 patients with intermediate or high grade NHL younger than 60 years, who achieved at least one complete remission and were not treated with BMT, was performed. Patients with bulky tumor at diagnosis and/or serum LDH greater than or equal to 600 U/l do poorly with conventional chemotherapy. Twelve patients with these high-risk initial characteristics in first complete remission (CR) and six patients in second or third CR were treated with cyclophosphamide (60 mg/kg x 2) and total body irradiation (1000-1200 cGy) followed by ABMT. Overall disease-free survival was 65% at a median follow-up of 35 months. No differences were found between the first and later CR patients. The rate of toxic death was 11%. Disease-free survival after first CR was better for 1st CR ABMT patients than for a historical chemotherapy control group with similar poor prognosis features (p = 0.008). These results support the use of ABMT in selected, high-risk NHL patients in first CR.

The ornithine decarboxylase gene is a transcriptional target of c-Myc.

Constitutive c-myc expression suppresses cell cycle arrest, promotes entry into S phase, and results in the growth factor-independent expression of ornithine decarboxylase (ODC; EC 4.1.1.17). The ODC gene contains a conserved repeat of the Myc binding site, CACGTG, in intron 1. In this report, we demonstrate that c-Myc is a potent transactivator of ODC promoter-reporter gene constructs in fibroblasts that requires the CACGTG repeat. These sites conferred Myc responsiveness on heterologous promoter constructs, suggesting that ODC is regulated by Myc at the level of transcription initiation. Analysis of deletion and point mutants of c-myc revealed that domains required for transactivation of the ODC promoter did not include the leucine zipper of the Myc protein. This suggests that Myc may interact with transcription factors other than Max to transactivate the ODC gene.

Position and orientation independent transactivation by c-Myc.

The c-myc oncogene c-Myc is commonly activated in cancer and transactivates gene expression by binding to CACGTG DNA sequences as a heterodimeric complex with Max. The ornithine decarboxylase (ODC), p53, prothymosin alpha and ECA39 promoters are transactivated by c-Myc, and are considered direct targets, as activation is mediated by CACGTG sequences. Interestingly, the c-Myc-responsive CACGTG sequences in the p53, prothymosin alpha, ECA39 and murine ODC genes are all downstream of the RNA CAP site, suggesting that downstream sequences are preferred c-Myc targets. Using a series of heterologous reporter constructs, we have tested the effects of position and orientation of c-Myc-responsive CACGTG sequences on c-Myc's ability to activate transcription. A single binding site conferred c-Myc-responsiveness independent of position and orientation, and over distances of 1.7 kbp. The extent of transactivation was not significantly influenced by position of the responsive elements. By contrast, the extent of transactivation was dependent upon the number of c-Myc binding sites. The results demonstrate that c-Myc activates transcription independent of position and orientation and that considerable flexibility exists in the interaction of c-Myc transactivation domains with the general transcription machinery.

flt3 ligand in cooperation with transforming growth factor-beta1 potentiates in vitro development of Langerhans-type dendritic cells and allows single-cell dendritic cell cluster formation under serum-free conditions.

Using a recently described serum-free culture system of purified human CD34+ progenitor cells, we show here a critical cooperation of flt3 ligand (FL) with transforming growth factor-beta1 (TGF-beta1) in the induction of in vitro dendritic cell/Langerhans cell (DC/LC) development. The addition of FL to serum-free cultures of CD34+ cells supplemented with TGF-beta1, granulocyte-macrophage colony-stimulating factor, tumor necrosis factor alpha, and stem cell factor strongly increases both percentages (mean, 36% +/- 5% v 64% +/- 4%; P = .001) and total numbers (4.4- +/- 0.8-fold) of CD1a+ dendritic cells. These in vitro-generated CD1a+ cells molecularly closely resemble a particular type of DC known as an epidermal Langerhans cell. Generation of DC under serum-free conditions was found to strictly require supplementation of culture medium with TGF-beta1. Upon omission of TGF-beta1, percentages of CD1a+ DC decreased (to mean, 10% +/- 8%; P = .001) and, in turn, percentages of granulomonocytic cells (CD1a- cells that are lysozyme [LZ+]; myeloperoxidase [MPO+]; CD14+) increased approximately threefold (P < .05). Furthermore, in the absence of TGF-beta1, FL consistently promotes generation of LZ+, MPO+, and CD14+ cells, but not of CD1a+ cells. Serum-free single-cell cultures set up under identical TGF-beta1- and FL-supplemented culture conditions showed that high percentages of CD34+ cells (mean, 18% +/- 2%; n = 4) give rise to day-10 DC colony formation. The majority of cells in these DC-containing colonies expressed the Langerhans cell/Birbeck granule specific marker molecule Lag. Without TGF-beta1 supplementation, Lag+ colony formation is minimal and formation of monocyte/macrophage-containing colonies predominates. Total cloning efficiency in the absence and presence of TGF-beta1 is virtually identical (mean, 41% +/- 6% v 41% +/- 4%). Thus, FL has the potential to strongly stimulate DC/LC generation, but has a strict requirement for TGF-beta1 to show this costimulatory effect.

Expression of Wiskott-Aldrich syndrome protein (WASP) gene during hematopoietic differentiation.

The Wiskott-Aldrich syndrome (WAS) is an X-linked recessive disorder described as a clinical triad of thrombocytopenia, eczema, and immunodeficiency. The gene responsible for WAS encodes a 502-amino acid proline-rich protein (WASp) that is likely to play a role in the cytoskeleton reorganization and/or in signal transduction of hematopoietic cells. However, the function and the regulation of the WAS gene (WASP) have not yet been clearly defined. We have studied WASP expression at the transcriptional level in freshly isolated mature peripheral blood cells and during hematopoietic development. For this purpose, we have isolated CD34+ hematopoietic precursor cells from cord blood. These cells were cultured in vitro with various growth factors to generate committed or mature cells belonging to different hematopoietic differentiation pathways, such as granulocytic (CD15+) cells, monocytic (CD14+) cells, dendritic (CD1a+) cells, erythroid lineage (glycophorin A+) cells, and megakaryocytic cells (CD41+). We have shown by reverse transcriptase polymerase chain reaction analysis that the WASP transcript is ubiquitously detectable throughout differentiation from early hematopoietic progenitors, including CD34+CD45RA- and CD34+CD45RA+ cells, to cells belonging to different hematopoietic lineages, including erythroid-committed and dendritic cells. In addition, Northern blot analysis showed that peripheral blood circulating lymphocytes (CD3+ and CD19+ cells) and monocytes express WASP mRNA. Several hematopoietic cell lines were tested and higher levels of expression were consistently detected in myelomonocytic cell types. By contrast, primary nonhematopoietic cells, including fibroblasts, endothelial cells, and keratinocytes, were consistently negative for WASP mRNA.

TGF-beta 1 promotes in vitro development of dendritic cells from CD34+ hemopoietic progenitors.

Several studies have demonstrated that dendritic cells (DC) can be generated in vitro from CD34+ hemopoietic progenitor cells. The growth requirements for these cells are poorly characterized, however. In particular, undefined serum/plasma components seem to significantly contribute to in vitro DC development. We report here that the cytokine combination granulocyte-macrophage CSF (GM-CSF) plus TNF-alpha and stem cell factor (SCF) commonly used for the in vitro generation of DC in serum/plasma-supplemented medium is, in the absence of serum supplementation, very inefficient in inducing DC development. We further demonstrate that supplementation with TGF-beta 1 is required for substantial DC development to occur in the absence of serum. Culture of CD34+ cells under serum-free conditions with TGF-beta 1 plus GM-CSF, TNF-alpha, and SCF strongly induces DC differentiation. This culture condition is even more efficient than culturing CD34+ cells with GM-CSF plus TNF-alpha and SCF in the presence of cord blood plasma. The proportions and total yields of cells with typical DC morphology and CD1a molecule expression are higher. The allostimulatory capacity of DC from TGF-beta 1-supplemented, cultures exceeds allostimulation by cells grown in plasma-containing medium. Substantial numbers (21 +/- 7%) of cells grown in TGF-beta 1-supplemented, but not plasma-supplemented, cultures express the Birbeck granule marker molecule Lag and display numerous Birbeck granules. Cells with distinct monocytic features are less frequently observed in TGF-beta 1-supplemented serum-free cultures. The addition of neutralizing anti-TGF-beta 1 Ab abrogates the observed TGF-beta 1 effects.

Molecular and functional characteristics of dendritic cells generated from highly purified CD14+ peripheral blood monocytes.

Dendritic cells (DC) are the most potent APCs within the immune system. We show here that highly purified CD14(bright) peripheral blood monocytes supplemented with granulocyte-monocyte (GM)-CSF plus IL-4 develop with high efficacy (>95% of input cells) into DC. They neo-expressed CD1a, CD1b, CD1c, CD80, and CD5; they massively up-regulated CD40 (109-fold) and HLA-DQ and DP (125- and 87-fold); and significantly (>5-fold) up-regulated HLA-DR, CD4, CD11b, CD11c, CD43, CD45, CD45R0, CD54, CD58, and CD59. CD14, CD15s, CD64, and CDw65 molecules were down-regulated to background levels, and no major changes were observed for HLA class I, CD11a, CD32, CD33, CD48, CD50, CD86, CDw92, CD93, or CD97. Monocytes cultured in parallel with GM-CSF plus TNF-alpha were more heterogeneous in expression densities but otherwise similar in their surface molecule repertoire. They clearly differed, however, in their accessory cell capacity. Only GM-CSF plus IL-4-cultured cells were found to be potent stimulators in allogeneic and autologous MLR and they presented tetanus toxoid 100- to 1000-fold more efficiently than other cell populations tested. Furthermore, only cytokine-treated monocytes formed clusters with resting T cells. At variance from all these similarities between in vitro-generated monocyte-derived DC and in vivo-developing DC, the DC populations generated by us contained significant amounts of myeloperoxidase and also expressed lysozyme. At least in this respect they, thus, differ from "classical" DC types.

Identification of CD68+lin- peripheral blood cells with dendritic precursor characteristics.

Expression of CD68 (macrosialin) in the absence of surface and lysosomal lineage marker molecules is a characteristic feature of T zone-associated plasmacytoid monocytes, which were recently shown to represent precursors of dendritic cells (DC). We demonstrate here a minor population of strongly CD68-positive (CD68bright) blood cells that lack all analyzed myeloid surface (CD14-, CD33-, CD13-, CD11b-, CD11c-) and lysosomal (myeloperoxidase, MPO- and lysozyme, LZ-) marker molecules (0.4 +/- 2% of the total mononuclear cells). These CD68bright, lineage marker-negative (lin-) cells can be induced to proliferate in the presence of IL-3. They do not acquire myeloid features even upon stimulation with granulocyte-macrophage CSF plus IL-1, IL-3, and IL-6. Instead, these cells develop typical DC characteristics upon culture. Furthermore, these CD68brightlin- DC precursors acquire mature DC characteristics (CD86+, CD83+, CD54bright) upon stimulation with CD40 ligand plus IL-3. A second subset of DC precursor-like blood cells was found to weakly express CD68 (0.3 +/- 0.2% of the total mononuclear cells) and to coexpress several myeloid lineage associated molecules (LZ+, CD11c+, CD33+, CD13+). Cells of this second subset resemble both previously described myeloid-related peripheral blood DC and germinal center DC. Analysis of peripheral blood leukocytes for CD68 thus revealed the existence of two cell subsets that phenotypically resemble lymphoid tissue-associated DC. The unique phenotype CD68brightlin- is highly reminiscent of T zone-associated plasmacytoid monocytes. CD68brightlin- blood leukocytes also functionally resemble plasmacytoid monocytes. The lack of all analyzed myeloid features by CD68brightlin- blood leukocytes suggests that these cells arise from a novel nonmyeloid human DC differentiation pathway.

IL-4 acts as a homeostatic regulator of IL-2-induced TNF and IFN-gamma.

Interleukin-4 (IL-4) is a cytokine secreted by interleukin-2 (IL-2)-activated lymphocytes. IL-2-stimulated lymphocytes also secrete two cytokines, tumour necrosis factor (TNF) and gamma-interferon (IFN-gamma), which contribute to effector function and which may themselves recruit fresh, cytokine-secreting effector cells. We have now investigated whether the IL-4 induced is able to homeostatically regulate secretion of the TNF and IFN-gamma. Peripheral blood mononuclear cells or lymphocytes from normal donors and from patients with neoplastic disease were cultured in the presence of IL-2 alone, IL-4 alone or with both cytokines. IL-2 induced high levels of TNF and IFN-gamma secretion in both groups. The addition of recombinant IL-4 to these IL-2-stimulated cultures lead to significant inhibition of IFN-gamma and TNF production. IFN-gamma secretion was reduced by 50-99% in normal donors and by between 11% and 99% in patients (P less than 0.001). TNF levels induced by IL-2 were similarly reduced by IL-4 both in normal donors (P less than 0.003) and in patients (P less than 0.01). These inhibitory effects were produced by IL-4 at doses of IL-2 attainable in vivo. Inhibition appears to represent a homeostatic regulatory mechanism which may limit recruitment of fresh activated killer (AK) cells. When endogenous IL-4 activity in IL-2-activated lymphocytes was blocked by anti-IL-4 antibody, significantly higher levels of IFN-gamma and TNF were secreted (P less than 0.05). Since both TNF and IFN-gamma may contribute to the anti-neoplastic action of IL-2, manipulating the level of IL-4 activity in vivo could augment the benefits of IL-2 immunotherapy.

Homeostatic action of interleukin-4 on endogenous and recombinant interleukin-2-induced activated killer cell function.

Cytokine-secreting, major histocompatibility complex-unrestricted activated killer (AK) cells are toxic to a wide range of virus-infected or malignant target cells and may be generated endogenously, eg, after bone marrow transplantation, or by infusion of cytokines such as recombinant interleukin-2 (rIL-2). Although AK cells secrete cytokines such as gamma-interferon and tumor necrosis factor, which are themselves able to recruit fresh cytokine-secreting AK cells, activation in both settings is short-lived, implying the existence of homeostatic regulatory mechanisms. We now demonstrate one mechanism by which rapid homeostasis is achieved. We show that IL-4 is produced in patients with both endogenously and exogenously generated AK cells. The cytokine was detected in serum after marrow transplantation, and IL-4 transcripts appeared in circulating lymphocytes during rIL-2 infusion. Although IL-4 inhibited the induction phase of AK cell function, it had no significant inhibitory effect on the ability of AK cells from these individuals to respond to restimulation. Nonetheless, neutralization of the IL-4 induced during cell activation doubled the half-life of AK function, once activating stimuli were removed, from 18 to 44 hours and produced a 2-log increase in AK cell secretion of tumor necrosis factor and gamma-interferon. These data suggest that IL-4 induced in vivo during lymphocyte activation abbreviates AK cell responses once the triggering stimuli have been removed. Neutralization of endogenous IL-4 in vivo by appropriate monoclonal antibodies might prolong the duration of AK function.

Effects of recombinant interleukin-2 administration on cytotoxic function following high-dose chemo-radiotherapy for hematological malignancy.

Activated killer cells, unrestricted by major histocompatibility (MHC) antigens circulate in the peripheral blood of patients who have undergone autologous and allogeneic bone marrow transplant (BMT) and may contribute to the reduced risk of leukemic relapse observed after these procedures. Interleukin-2 (IL-2) in vitro augments this cytotoxicity and used therapeutically might thereby promote the eradication of minimal residual disease. In order to assess whether these effects on cytotoxicity can be reproduced in vivo, we studied changes in number, phenotype, and MHC unrestricted cytotoxicity of peripheral blood mononuclear cells obtained from patients with hematologic malignancy receiving IL-2 infusions. Patients with acute myeloid leukemia and multiple myeloma were treated after cytotoxic chemotherapy or autologous BMT. IL-2 infusions produced an initial lymphopenia, followed by a progressive recovery in mononuclear cell numbers and a rebound lymphocytosis after the termination of treatment. This affected all lymphocyte subsets; in particular CD25 (IL-2 receptor) positive cell numbers rose sevenfold. Cells with the ability to kill a natural killer (NK)-resistant, lymphokine activated killer cell (LAK)-sensitive target appeared in the circulation during 16 of 19 infusions and mean LAK activity rose from 5.9% to 15.5% during infusion (E:T ratio, 50:1; P less than .001). During IL-2 infusion, cells present in the peripheral blood inhibited the growth of myeloid leukemia blasts in agar after overnight co-culture. Depletion experiments showed that LAK activity was mediated by cells of both CD3- CD16+ (NK derived) and CD3+ CD16- (T derived) subsets. LAK precursor activity in peripheral blood also significantly increased during IL-2 infusion. Increases in major histocompatibility complex (MHC) unrestricted cytotoxicity can be produced by IL-2 infusions in vivo and may result in improved relapse-free survival following chemotherapy or BMT.

Possible mechanism of selective killing of myeloid leukemic blast cells by lymphokine-activated killer cells.

Major histocompatibility complex-unrestricted lymphokine-activated killer (LAK) cells have been proposed as therapy for a variety of hematologic malignancies. Because these cells recognize and kill their targets independently of their antigen specific CD3 receptor, it is unclear how they might discriminate between normal and malignant cells. We now propose one such mechanism for the selective killing of myeloid leukemia blasts. While both CD2+ and CD2- activated killer cells may inhibit the clonogenic growth of myeloid leukemia cells, only the CD2+ subset effectively inhibits the growth of normal myeloid (granulocyte-macrophage and granulocyte) progenitors. This difference appears to reflect differential requirements for cell adhesion molecule recognition between normal and malignant progenitor cells. Inhibition of the growth of normal granulocyte-macrophage colonies by CD2+ LAK cells is blocked by antibodies to the CD2-lymphocyte function-associated antigen 3 (LFA-3) (CD58) cell adhesion system. In contrast, these antibodies have no effect on CD2+ LAK-mediated inhibition of malignant cell clonogenic growth. Instead, antibodies to the LFA-1 (CD11a/CD18)-intercellular adhesion molecule 1 (ICAM-1) (CD54) adhesion system reduce inhibition. These differences correspond to differential expression of the CD54 cell adhesion molecule by normal and malignant myeloid progenitor cells because less than 15% of normal CD34 positive cells are CD54+ while greater than 85% of CD34+ acute myeloid leukemia blasts express the CD54 antigen. LFA-3, the ligand for CD2, is strongly expressed by erythrocytes, and these cells competitively inhibit killing of normal but not malignant clonogenic cells in an analogous way to the effects of monoclonal antibody to the CD2-LFA-3 adhesion system. The operation of this effect in vivo may be a basis for selective cytotoxicity by CD2+ LAK against clonogenic myeloid blast cells, and could be exploited further with infusion of appropriate monoclonal antibodies.