Search

Two patients with agranulocytosis associated with diphenylhydantoin (DPH) therapy and clinical data suggesting suppression of granulopoiesis were investigated using in vitro culture techniques for committed granulocyte/macrophage precursors. Addition of DPH to cultures containing the patients' sera resulted in significant suppression of colony growth. Extensive studies on the acute serum from one patient revealed the drug-dependent inhibitory activity to be nondialyzable, resistant to chloroform extraction, heat stable, active in the presence of heat-inactivated fetal bovine serum, active against autologous as well as allogeneic cells, and absent from convalescent sera. Drug-dependent bone marrow colony-suppressing activity was removed by absorption on an antiimmunoglobulin-Sepharose column but not by IgG-Sepharose. The serum show non-drug dependent suppression of oxygen consumption by normal polymorphonuclear leukocytes engaged in phagocytosis and also showed evidence of ability to opsonize these cells. When the serum was incubated with mitogen-stimulated lymphocytes, suppression of 3H-thymidine uptake by autologous but not allogeneic cells was noted. Similarly, the serum suppressed short-term 3H-thymidine uptake by autologous but not allogeneic bone marrow. Absorption of the patients' sera with allogeneic polymorphonuclear leukocytes, autologous polymorphonuclear leukocytes, or autologous lymphocytes removed the drug-dependent inhibitory activity, but absorption with allogeneic lymphocytes did not. These data are most consistent with the presence of a noncomplement dependent antibody capable of suppressing granulopoiesis, mediating peripheral destruction of polymorphonuclear leukocytes, and cross-reacting with a lymphocyte antigen of limited population distribution.

Human myeloid leukemia cells (HL60) and malignant lymphocytes (Namalwa) were grown in protein-free, Fe-supplemented media and used to study growth responses to insulin and insulin-like growth factor 1 (IGF-I). HL60 cells previously grown in serum-free medium containing microgram quantities of insulin showed an 18-fold reduction in cumulative cell production when grown without insulin. However, the same cells showed reduced or absent growth stimulation with 1 to 100 ng/mL insulin or IGF-I for at least four days following insulin deprivation, indicating that culture conditions modified insulin and IGF-I responses. When the same cells were grown in Fe-supplemented, protein-free medium (RPMI-Fe), insulin and IGF-I caused dose-dependent stimulation of HL60 cell growth with half-maximal stimulation at nanogram concentrations. Namalwa cells grown in protein-free medium showed no response to either hormone. Radioligand binding showed the presence of insulin and IGF-I receptors on both HL60 and Namalwa cells grown in RPMI-Fe. HL60 cells grown in fetal bovine serum had higher, and cells grown with microgram quantities of insulin dramatically reduced, insulin binding. Competitive binding studies and cultures with anti-IGF-I receptor antibody showed insulin and IGF-I stimulated growth through their respective specific receptors. Both insulin and IGF-I stimulate growth of some cultured human leukemia cells, but the presence of insulin or IGF-I receptors alone does not predict growth responses. Culture conditions affect both cellular responses and ligand binding by these hormones and must be closely controlled to study growth responses.

The response of human myeloid leukemia cells to various sources of colony-stimulating activity (CSA) and media conditioned by phytohemagglutinin-stimulated mononuclear cells (PHA-LCM) was investigated in liquid and colony culture. PHA-LCM, placenta-conditioned medium, GCT cell line-conditioned medium, leukocyte-conditioned medium, and partially purified CSA for human and murine cells were tested for ability to support growth of granulocyte-macrophage colonies from adherent cell-depleted human bone marrow. This activity was correlated with ability to support leukemia colony growth in methylcellulose, and [3H]thymidine incorporation in liquid culture by normal bone marrow cells, leukemia cells, and the KG-1 myeloid leukemia cell line. For normal cells, growth and liquid culture responses were highly correlated for various sources of CSA (r = 0.92), and addition of data using PHA-LCM changed results only slightly (r = 0.89). [3H]thymidine incorporation by leukemia cells from patients without a prior history of a myeloproliferative disorder was also highly correlated with normal CSA (r = 0.97) for sources other than PHA-LCM. Responses of leukemia blasts and KG-1 cells in liquid culture to PHA-LCM appeared in excess of its CSA for normal cells. Colony growth by leukemia cells was not clearly correlated with either liquid culture activity for leukemia cells or CSA for normal cells. PHA-LCM was also not statistically superior to placenta-conditioned medium as stimulus for leukemia colony growth, but was superior to placenta-conditioned medium for some patients. Differentiation in culture did not appear to depend on CSA source. We conclude that normal myeloid cells respond to CSA in a highly correlated fashion in both colony and liquid cultures. The majority of myeloid leukemia cells respond to either PHA-LCM or CSA, but the ability of PHA-LCM to support leukemia cell growth is greater than its CSA content. The possibility exists that overlapping populations responsive to CSA and to PHA-LCM are present simultaneously in patients with myeloid leukemia.

A liquid culture technique was used to study 23 synovial fluids (SF) (21 from inflammatory joint diseases and 2 noninflammatory SF) and supernatants of two cultured rheumatoid arthritis (RA) synovial tissues for colony-stimulating factor (CSF). The proliferative responses of human peripheral blood macrophage-depleted non-T cells treated with synovial fluids, supernatants of synovial tissue explants, and recombinant granulocyte-macrophage (rGM)-CSF were compared. Aggregates of cells that formed in long-term cultures (15 d) were similar for each applied agent and consisted of macrophages, eosinophils, and large blasts. Tritiated thymidine incorporation was proportional to the concentration of rGM-CSF and was accompanied by an increase in number and size of cellular aggregates formed in the cultures. CSF activity was observed in inflammatory SF, with tritiated thymidine uptake of 3,501 +/- 1,140 cpm in the presence of RA samples (n = 15) compared to 1,985 +/- 628 for non-RA inflammatory SF (n = 7) (P less than 0.05) and 583 +/- 525 for medium (n = 6) (P less than 0.01). The proliferative response to RA SF was often more apparent when the samples were diluted, because at higher concentrations the RA SF was inhibitory. Two RA SF were fractionated by Sephadex G100 column chromatography; low levels of CSF activity were detected in fractions corresponding to Mr of 70-100 kD, but the major CSF activity was found in the 20-24-kD fractions. A polyclonal rabbit anti-GM-CSF antibody eliminated the stimulating activity from both rGM-CSF and RA SF. Finally, a specific RIA identified significant levels of GM-CSF (40-140 U/ml) in the culture supernatants of 3 additional RA synovial tissues. These data document the local production of GM-CSF in rheumatoid synovitis and are the first description of this cytokine at a site of disease activity.

Untreated, late passage HL60 promyelocytic and KG1 myeloblastic leukaemia cells did not increase proliferation with placenta or Mo T cell conditioned medium containing colony-stimulating factor (CSF) nor with partially purified, recombinant granulocyte/macrophage (GM)-CSF. However, after induction with DMSO or 1,25-dihydroxy-vitamin D3, HL60 cells showed dose-dependent increases in proliferation with crude and purified CSFs. CSF responses and macrophage differentiation were induced in KG1 cells by treatment with tetradecanoylphorbol-acetate (TPA). When cells were exposed to inducing agents for varying periods, washed and exposed to CSF, proliferative responses were related to time of exposure. Cells exposed for 1-4 d showed post-induction CSF-induced proliferation, but cells induced for 5-6 d were inhibited by CSF. Induction of CSF response appeared linked to differentiation, since KG1 cells differentiated with TPA and developed CSF-induced proliferative responses, but showed no differentiation or CSF induced proliferation after treatment with vitamin D3. When HL60 cells were continuously exposed to DMSO or vitamin D3, overall cell production was increased by placenta conditioned medium, but cultures still became senescent and died after several weeks. Cells continuously cultured with DMSO were predominantly macrophages, indicating lineages of DMSO-induced differentiation were modified by continuous culture or the presence of CSF. After treatment with chemical inducers, proliferation of myeloid leukaemia lines is stimulated by CSF, providing a model for post-deterministic regulation of normal and malignant myeloid cell production.

Prostaglandins of the E series (PGE) have varying effects in vitro on normal granulocyte/macrophage (CFU-GM) progenitors. When added directly to cultures, PGE inhibits growth of normal 7 and 14 day bone marrow CFU-GM in a dose-dependent manner, while CML CFU-GM are refractory to PGE inhibition. The plant diterpene forskalin, a potent activator of intracellular cyclic AMP, inhibited normal and CML CFU-GM, indicating that the response of CML cells to increased intracellular cyclic AMP is normal. Low dose forskalin, which potentiates activators of adenyl cyclase, showed additive effects with PGE on normal CFU-GM cells, however, showed no additive effects of PGE and forskalin, suggesting that PGE failed to activate adenyl cyclase. Normal CFU-GM incubated with PGE for two hours showed a significant increase in CFU-GM growth, and removal of adherent cells prior to exposure to PGE abrogated this effect. CML cells did not respond to a 2-h exposure to PGE, and removal of adherent cells from these cultures had no effect. Normal and CML CFU-GM showed dose-dependent increases in proliferation in the presence of PGF2. These studies confirm that CML CFC-GM are refractory to inhibition by PGE, and show that these cells respond normally to increased levels of intracellular cyclic AMP. Response of CML cells to PGF2 alpha is normal, and conversion of PGE to PGF could result in increased granulocyte progenitor proliferation.

Fifty-six patients with small cell carcinoma of the lung were treated with a two-cyclic induction course of hexamethylmelamine, vincristine, doxorubicin, and cyclophosphamide. Patients with limited disease (LD) who responded and patients with extensive disease (ED) who had a complete response received prophylactic whole-brain radiotherapy, as well as radiotherapy to thoracic and abdominal sites of disease. Concurrently with radiotherapy, consolidation chemotherapy was given with doxorubicin, cyclophosphamide, methotrexate, and etoposide. The complete response rate was 35% for ED patients and 68% for LD patients. The median survival time for complete responders was 54 weeks for ED patients and 65 weeks for LD patients. The toxicity of the program was moderate, and the effectiveness was comparable to that of other reported combined-modality treatment programs.

Concurrent administration of sodium thiosulfate reduced the toxicity of cis-dichlorodiammineplatinum(II) (DDP) in a dose-related manner in mice. Sodium thiosulfate protected mice against an otherwise lethal dose of DDP (20 mg/kg), and reduced DDP-induced weight loss. Sodium thiosulfate (800 mg/kg) injected within 1 hour before or 1/2 hour after DDP blocked nephrotoxicity as measured by a rise in BUN, an increase in kidney weight, and medullary hemorrhage. In culture, sodium thiosulfate markedly reduced the toxicity of DDP to mouse colony-forming units. Concurrent injection of sodium thiosulfate partially reduced the antitumor activity of DDP. The therapeutic dose range of DDP was expanded, but the maximum increase in lifespan of mice bearing L1210 leukemia was reduced by 40%. Sodium thiosulfate offers the possibility of systemic protection against the cytotoxicity of regionally administered DDP in man.

The peripheral blood of 2 patients with acute monocytic leukemia of the undifferentiated type was studied for the presence of leukemic progenitor cells in colony forming assays. The peripheral blood of both patients contained only one type of progenitor cell as determined with these assay systems. Daughter cells of these progenitor cells morphologically and histochemically resembled monoblasts and immature macrophages. Similar progenitor cells were not encountered in the study of 5 patients with acute myelogenous leukemia whose peripheral blood cells formed colonies in the same system. The finding of this unusual progenitor cell supports the existence of acute monocytic leukemia as a separate clinicopathologic entity, and suggests that it represents a malignant transformation of a progenitor cell of the monocyte/macrophage series. An alternative explanation of these observations would be provided by single phenotypic expressions of a multi-potent stem cell which has undergone malignant transformation.

Thymidine (dThd) concentrations have been measured in the sera of normal subjects and solid tumor cancer patients by means of a sensitive high-pressure liquid chromatographic assay to determine whether natural and methotrexate (MTX)-induced fluctuations were large enough to alter the toxicity of MTX to marrow. The mean concentration in normal subjects with measurable levels was 1.3 X 10(-7) M (range less than 4 X 10(-8) to 6 X 10(-7) M). In cancer patients it was 2.0 X 10(-7) M (range less than 4 X 10(-8) to 8.7 X 10(-7)), and in malignant effusions 1.2 X 10(-7) M (range less than 4 X 10(-8) to 2.2 X 10(-7) M). The wide range of variation in random samples was also found when multiple samples were obtained from the same patient during a 24-h period where dThd concentration varied from a minimum of two- to greater than six-fold. Treatment with MTX 3 mg/m2 caused an average 59% reduction in serum dThd during the first 24 h after injection during nine courses of therapy. dThd was tested for its ability to modulate the toxicity of MTX to human granulocate colony-forming units in culture across the concentration range found in vivo: changes in dThd concentration equivalent to normal fluctuations in vivo altered colony survival by 31% to greater than 72%. A reduction in culture dThd equivalent to that produced in vivo by high-dose TMX increased colony kill by 25%. The results indicate that in vivo variations in serum dThd are in an appropriate range and of a sufficient magnitude to alter the toxicity of MTX to marrow, and they demonstrate that MTX can modulate its own toxicity by reducing serum dThd.

Twenty-nine patients with extensive disease, small-cell carcinoma of the lung, were treated with two cycles of intensive combination chemotherapy: HexaVAC (hexamethylmelamine, vincristine, Adriamycin, cyclophosphamide). Responders received prophylactic cranial radiation (2000 rad/10 fractions) and non cross resistant chemotherapy via a schedule of alternating cycles of CMV (cyclophosphamide, methotrexate, VP-16-213) and AMV (Adriamycin, methotrexate, VP-16-213). Whenever a complete response was achieved, consolidation radiotherapy was given to the lung primary (4000 rad/20 fractions, split dose) and abdominal metastases (2000 rad/10 fractions) synchronous with CMV therapy. The complete response rate was 14% with HexaVAC, but increased to 38% during CMV/AMV. Total response rate (complete and partial) was 59% and median survival was 42 weeks. Prophylactic brain radiation prevented clinical relapse in the brain in all 14 patients who received it. However, consolidation radiotherapy failed to prevent clinical relapse in the lung and/or liver, and therapeutic brain radiation (3000 rad) failed to prevent relapse in that site. The simultaneous administration of radiotherapy and chemotherapy was well-tolerated although two patients with poor performance status died of infectious complications while leukopenic. In spite of the high response rate, durable remissions with prolonged disease free survival were rare. Further evaluation of induction, consolidation, and maintenance modes of therapy are indicated.

A case of fatal pulmonary fibrosis and atypical epithelial proliferation (AEP) in a patient with multiple myeloma treated with melphalan is presented. Review of 10 other autopsied patients with myeloma treated with melphalan but no thoracic radiation, other cytotoxic agents, or highdose oxygen therapy revealed one other patient who died with extensive pulmonary fibrosis and AEP. Four other patients with AEP not associated with pneumonitis or fibrosis were also found, while no such changes were found in 11 autopsy controls or 11 patients with myeloma who did not receive cytotoxic agents. Melphalan should be added to the growing list of agents capable of causing severe fibrotic pulmonary reactions.

Oxipurinol, the major metabolite of allopurinol, decreased the toxicity of 5-fluorouracil (5-FU) to human granulocyte colony-forming units in vitro by a factor of four. The ability of allopurinol to reduce 5-FU toxicity in vivo was studied in 23 advanced cancer patients during 42 courses of treatment. 5-FU was administered by continuous intravenous infusion for five days; allopurinol, 300 mg, po, every 8 hours was started 2 hours before and continued during and for 24 hours after 5-FU infusion. 5-FU was escalated from 1.5 to 2.25 g/m2/day on separate courses; the dose-limiting toxicity was mucositis which occurred at a level of 2.0 g/m2/day. At a 5-FU dose rate of greater than 2.0 g/m2/day 5-FU pharmacokinetics were nonlinear, reflecting saturation of catabolic pathways, and the steady-state 5-FU serum concentration was approximately 4 times that which was tolerable without allopurinol. At these concentrations of 5-FU oxipurinol significantly influenced the clearance of 5-FU. Thus concurrent allopurinol therapy permitted a doubling of the maximum tolerated dose of 5-FU and a four-fold increase in the tolerated concentration x time exposure to 5-FU.

Ricin A-chain was purified from native ricin using lactosyl-Sepharose. It was non-toxic to whole cells at a concentration of 1 microM yet nearly as effective as an equimolar concentration of ricin in blocking in vitro protein synthesis. Hybridomas secreting monoclonal antibodies to ricin A-chain were produced using the murine myeloma cell line NS-1. These anti-ricin A-chain antibodies cross-reacted with whole ricin but exhibited little cross-reactivity with purified ricin B-chain. Antibodies 2F2 and 2F5 both immunoprecipitated ricin A-chain. Both antibodies also precipitated ricin B-chain, as did the irrelevant control antibodies MOPC-21 and MPC-11. Pre-incubation of B-chain with 0.1 M galactose eliminated greater than 90% of precipitation by 2F2, MOPC-21 and MPC-11 but effected minimally precipitation by 2F5. Enzyme-linked immunosorbent assays using antibodies 2F2 and 2F5 to detect ricin A-chain in murine or human serum were linear between 40 and 800 ng ricin A-chain per ml. Anti-ricin A-chain antibodies 2F2 and 2F5 produced some inhibition of in vitro A-chain catalytic activity. Specific monoclonal antibodies to A-chain hemitoxin will be useful for characterization of functional hemitoxin domains, in in vitro assays for the stability of A-chain immunotoxins, and in characterizing the cellular internalization and processing of conjugates containing ricin or ricin A-chain.

In previous studies, antitransferrin receptor antibody 42/6 inhibited growth of normal granulocyte/macrophage progenitors and some malignant myeloid cells. In these studies, leukemia cell lines cultured without serum and fresh leukemia cells were used to investigate the roles of Fe, transferrin receptors, and transferrin in leukemia cell growth, and mechanisms of 42/6 inhibition and resistance. HL60 and KG-1 leukemia cells grown in serum-free medium were inhibited by 42/6. In contrast to results in fetal calf serum (FCS), soluble Fe (ferric nitriloacetate) reversed 42/6 growth inhibition of serum-free HL60 cells. When HL60 cells were adapted for growth in serum-free, transferrin-free medium, they became refractory to 42/6 growth inhibition. By using radiolabeled transferrin and 42/6, HL60 cells cultured in FCS and transferrin displayed similar quantities of transferrin receptors (29,000-30,000/cell) and similar Kd's (3.8-4.9 X 10(-9) M). Cells grown in transferrin-free medium showed a similar Kd (3.1 X 10(-9) M), but fewer transferrin binding sites (5,000/cell). Transferrin-independent cells contained a log higher concentration of intracellular ferritin. For both FCS and serum-free HL60 cells, calculated affinities for 42/6 were lower (5.7-10.0 X 10(-9) M), but the number of binding sites was three- to fourfold higher. To investigate further the relationship between receptor display and antibody inhibition in proliferating normal and malignant myeloid cells, simultaneous immunofluorescence was used to determine the cell cycle status of transferrin receptor-positive cells. Malignant cells in S + G2/M displayed approximately 50% of the amount of transferrin receptors detected in normal dividing colony-stimulating factor-stimulated marrow cells. Receptor display by dividing cells from two patients with acute nonlymphocytic leukemia was variable. When HL60 cells were exposed to dimethyl sulfoxide, transferrin receptor display decreased, and 42/6 growth inhibition was abrogated or greatly diminished. The presence of 42/6 did not prevent dimethyl sulfoxide-induced HL60 differentiation in serum-containing or serum-free cultures. We conclude that human leukemia cells require Fe for growth and that 42/6 inhibits transferrin-dependent cells by Fe deprivation. Some dividing normal and differentiating malignant cells display reduced transferrin receptors, and can also escape antibody inhibition. The increased ferritin levels and decreased transferrin receptors in transferrin-independent HL60 cells confirm the inverse relationship between cell ferritin content and transferrin receptor display. These studies indicate a critical role for Fe in leukemia cell growth and possible roles in cellular differentiation.

In a previous study, colony-stimulating factor (CSF) activity assayed in colony culture correlated closely with 3HTdR uptake by human marrow cells depleted of adherent cells. To use this assay for screening media for CSF and immunotoxins for marrow toxicity, cells growing in liquid culture were compared to conventional granulocyte/macrophage (CFU-gm) colony assays. CSF dose-response relationships for liquid and colony-forming assays were nearly identical. 3HTdR uptake by nonadherent marrow cells was CSF dose-related, and there was a linear relationship between number of cells cultured and 3HTdR uptake. Ricin cytotoxicity curves for liquid cultures and CFU-gm were identical on day 7 but showed some disparity with day 14 cultures. Results with all cultures showed 3HTdR uptake to be most closely correlated with CFU-gm colony, rather than cluster, growth. Myeloid cell differentiation in liquid culture was similar to colony cultures, producing mixtures of granulocytes, macrophages and eosinophils. By combining cell and differential counts, production of various myeloid cells could be quantitated. Cytotoxicity of anti-Ia for CFU-gm and liquid culture cells was compared and the majority of both cell populations expressed Ia-like antigens. Simultaneous staining for surface antigens and DNA content was used to characterize proliferating marrow cells, and the vast majority of cells expressed myeloid markers. Transferrin receptors were displayed by cells in S/G2/M and appeared after CSF stimulation on G0/G1 cells. We conclude liquid cultures can be used to screen conditioned media for human CSF and to screen for cytotoxicity to normal myeloid precursor cells. Behavior of CSF-responsive cells in liquid culture appears most closely related to that of CFU-gm colony-forming cells, and characterization of CSF-stimulated cells allows quantitative as well as qualitative estimates of myeloid cell production.

The effects of antiepidermal growth factor (EGF) receptor monoclonal antibodies (MAbs) 528 and 225 and a 528-ricin A conjugate on the growth of normal and malignant human cells were tested in vitro. Malignant human cell lines with EGF receptor numbers ranging from 0 to 4 X 10(5) receptors/cell, human fetal fibroblasts, and normal marrow granulocyte/macrophage progenitors (CFU-gm) showed no effect when grown with 10(-12) M to 10(-7) M MAb 225 or 528. MAbs 225 and 528 and EGF also had no effect on the ability of marrow stromal cells to maintain CFU-gm viability in long-term marrow cultures. Reversible growth inhibition of A431 epidermoid and MDA-468 breast carcinoma cells with 2 and 3 X 10(6) EGF receptors/cell, respectively, was observed with both antibodies and with 10(-8) M EGF. In contrast, an immunoconjugate prepared with MAb 528 and recombinant ricin A chain (528-rRA) showed dose-dependent killing over a concentration range of 10(-12) M to 10(-8) M against cells with greater than or equal to 1.2 X 10(5) EGF receptors/cell [concentration that causes 50% inhibition of growth (IC50) values, approximately 10(-12) M to 10(-10) M]. Human fetal fibroblasts (5.6 X 10(4) EGF receptors/cell), melanoma cells without detectable EGF receptors, and human CFU-gm showed IC50 values of greater than 10(-8) M. Killing of KB epidermoid carcinoma cells and 547 ovarian carcinoma cells with 4 and 1.2 X 10(5) EGF receptors/cell by 10(-10) or 10(-11) M 528-rRA was time dependent, but cytotoxicity to 547 cells was not complete even with 48 hours of immunotoxin treatment. Cytotoxicity of 528-rRA was not enhanced by chloroquine or verapamil. In vitro, anti-EGF receptor MAbs cause reversible antiproliferative effects only against malignant cell lines with amplified EGF receptor expression. In contrast, 528-rRA shows potent, specific toxicity to cells with greater than 50,000 EGF receptors/cell. However, kinetics of cell killing with 528-rRA are protracted, suggesting that prolonged exposure may be required for in vivo antitumor effects.

The potential use of nude mouse xenografts as a source of human tumor tissue for preclinical assessment of drug activity was examined by a comparison of in vitro sensitivity to four anthracycline derivatives of eight xenografts maintained in nude mice and tumor colony-forming units (CFU) from the xenografts grown in agar. Results obtained in the two systems with doxorubicin (Dx) and a new, closely related derivative (epi-doxorubicin, epi-Dx) correlated well. In vivo tumor growth delay and maximum in vitro tumor CFU kill for these two drugs showed a significant correlation (r = 0.64, P less than .01). Separation of tumors into "sensitive" and "insensitive" tumor populations on the basis of maximum in vitro CFU kill also predicted the in vivo response to these two drugs in 88% of cases. Similar analyses performed on in vitro and in vivo results with two other new anthracyclines (4'-deoxy-doxorubicin, deoxy-Dx and 4'-O-methyl-doxorubicin, O-Me-Dx) showed a significant negative correlation between in vivo and in vitro results; the in vitro system failed to predict in vivo activity of these two drugs. No significant differences in in vitro activity against normal, granulocyte/macrophage progenitors (CFU-GM) or against various tumor CFUs were detected. Thus, the selectivity (activity against normal tissue compared with that against tumor) of the four drugs appeared equal. These data suggest that in vitro screening of drugs using tumor CFUs from nude mouse xenografts may predict the in vivo activity of drugs for which pharmacologic data are available, but illustrate the difficulties in attempting to predict the in vivo activity of new drugs for which no such data are available.

We performed a series of studies to further clarify the nature of lymphocyte colony-forming cells (CFC) from normal peripheral blood. Mononuclear cells were separated into E-rosette-enriched (E+) and E-rosette-depleted (E-) populations and cultured in methylcellulose with conditioned media and irradiated mononuclear cells. Linear plating relationships were obtained with plating efficiencies of 0.26% +/- .02% (mean +/- SE) for E+ CFC and 0.18% +/- .02% for E- CFC. Cells in E+ colonies were T lymphocytes and in E- colonies were B lymphocytes as determined by cell surface marker analysis. Using the thymidine suicide technique, approximately one-half of CFC were found to be in cycle at any moment, and plating efficiencies and cell cycle status of E+ CFC were not changed by preincubation with PHA in liquid culture for 48 hr. Using antibody complement-mediated cytotoxicity, E+ CFC were found to be T101+, OKT3+, and Ia-, while E- CFC were OKT3- and Ia+. Using monocyte-depleted populations obtained by sedimentation at unit gravity, lymphocyte colony growth was absent in monocyte-depleted fractions, and optimal growth occurred with 40% monocytes in culture. In contrast to some previous studies, we find that lymphocyte CFC originate from a small, cycling population of cells bearing mature T or B lymphocyte markers. Entry into cell division, however, does not confer colony-forming capacity on lymphocytes. Monocytes are critical to growth of E+ CFC, and cultures severely depleted of monocytes would not be expected to form colonies.