Nrf2-Regulated Antioxidant System in Cells of In Vivo Drug-Resistant P388 Murine Leukemia Strains.

We studied the expression of Nrf2 transcription factor and antioxidant system proteins in drug-resistant murine leukemia strains P388 in vivo, as well as the redox status of cells under conditions of induced oxidative stress. Immunoblotting and real-time PCR showed that the cyclophosphamide-resistant strain P388 (P388/CP) exhibits Nrf2-mediated drug resistance. Cells of the P388/CP strain are characterized by high expression of Nrf2, which leads to a significant increase in the expression of ARE genes and antioxidant system proteins, as well as to the effective maintenance of redox homeostasis under conditions of induced oxidative stress. Taking into account the important role of Nrf2 overexpression in reducing the effectiveness of chemotherapy in patients with different leukemias, the P388/CP strain can be of great interest as a model in the development of new drugs for the treatment of malignant neoplasms.

Crystal structure of SsfS6, the putative C-glycosyltransferase involved in SF2575 biosynthesis.

The molecule known as SF2575 from Streptomyces sp. is a tetracycline polyketide natural product that displays antitumor activity against murine leukemia P388 in vivo. In the SF2575 biosynthetic pathway, SsfS6 has been implicated as the crucial C-glycosyltransferase (C-GT) that forms the C-C glycosidic bond between the sugar and the SF2575 tetracycline-like scaffold. Here, we report the crystal structure of SsfS6 in the free form and in complex with TDP, both at 2.4 Å resolution. The structures reveal SsfS6 to adopt a GT-B fold wherein the TDP and docked putative aglycon are consistent with the overall C-glycosylation reaction. As one of only a few existing structures for C-glycosyltransferases, the structures described herein may serve as a guide to better understand and engineer C-glycosylation.

Overexpression of apoptosis-associated speck-like protein in P388D1 murine lymphoma cells affects metastatic properties.

Apoptosis-associated speck-like protein (ASC) is a bipartite adaptor molecule that participates in inflammation and apoptosis. ASC silencing has been observed in a significant proportion of human cancers. Here, we examined the role of ASC overexpression in the metastasis of the P388D1 murine lymphoma cell line to the liver, lung, spleen and kidney. First, we determined that the P388D1 cells express ASC. Then, ASC overexpression in P388D1 was achieved by transfecting pEGFP-ASC-C2 into the P388D1 cells. Furthermore, after the ASC-overexpressing P388D1 cells were injected into DBA/2 mice through the vena caudalis, their metastasis to the lung and the liver was significantly reduced in the pEGFP-ASC-C2-transfected group. These data indicate that ASC overexpression affects the in vivo metastatic properties of P388D1 cells.

Extracellular ATP induces CD44 shedding from macrophage-like P388D1 cells via the P2X7 receptor.

The P2X7 receptor (P2X7R) is a nucleotide receptor expressed predominantly on hemopoietic, bone, and epithelial cells. The P2X7R can be activated by extracellular ATP and induces the influx of calcium, releases cytokines, and participates in cell proliferation and apoptosis. CD44 is an adhesion molecule. The effects of CD44 include cell-cell and cell-matrix adhesion interactions, lymphocyte activation, and cell migration. Many studies have shown that P2X7R and CD44 play important roles in hematological malignancies, but no study exists regarding the relationship between P2X7R and CD44. In the present study, we characterized P388D1 cells for the surface expression of CD44 and analyzed ATP-induced shedding. The data showed that P388D1 cells express CD44. Incubation of P388D1 cells with ATP induced a rapid loss of CD44 from the P388D1 cell surface. In addition, using a receptor inhibitor and P2X7R short hairpin RNA, we showed that the loss of CD44 is mediated via the P2X7R. Finally, we demonstrated that activation of P2X7R by ATP induces CD44 shedding.

Histoplasma capsulatum modulates the acidification of phagolysosomes.

The phagolysosome is perhaps the most effective antimicrobial site within macrophages due both to its acidity and to its variety of hydrolytic enzymes. Few species of pathogens survive and multiply in these vesicles. However, one strategy for microbial survival would be to induce a higher pH within these organelles, thus interfering with the activity of many lysosomal enzymes. Altering the intravesicular milieu might also profoundly influence antigen processing, antimicrobial drug delivery, and drug activity. Here we report the first example of an organism proliferating within phagolysosomes that maintain a relatively neutral pH for a sustained period of time. We inoculated P388D1 macrophages with fluorescein isothiocyanate (FITC)-labeled Histoplasma capsulatum or zymosan. Using the ratio of fluorescence excitations at 495 and 450 nm, we determined that vesicles containing either virulent or avirulent FITC-labeled H. capsulatum yeasts had a pH one to two units higher than vesicles containing either zymosan or methanol-killed H. capsulatum. The difference in pH remained stable for at least 5.5 h postinoculation. Longer-term studies using cells preincubated with acridine orange indicated that phagolysosomes containing live Histoplasma continued to maintain a relatively neutral pH for at least 30 h. Many agents raise the pH of multiple vesicles within the same cell. In contrast, H. capsulatum affects only the phagolysosome in which it is located; during coinoculation of cells with unlabeled Histoplasma and labeled zymosan, organelles containing zymosan still acidified normally. Similarly, unlabeled zymosan had no influence on the elevated pH of vesicles housing labeled Histoplasma. Thus, zymosan and Histoplasma were segregated into separate phagolysosomes that responded independently to their phagocytized contents. This localized effect might reflect an intrinsic difference between phagosomes housing the two particle types, active buffering by the microbe, or altered ion transport across the phagolysosomal membrane such that acidification is inhibited.

Studies of the biosynthesis of the mannose 6-phosphate receptor in receptor-positive and -deficient cell lines.

Phosphomannosyl residues present on lysosomal enzymes are specifically recognized by the mannose 6-phosphate receptor protein. This interaction results in the selective targeting of lysosomal enzymes to lysosomes. While this pathway is operative in many cell types, we have found four cultured cell lines that are deficient in the ability to bind lysosomal enzymes containing phosphomannosyl residues to their intracellular or surface membranes (Gabel, C., D. Goldberg, and S. Kornfeld, 1983, Proc. Natl. Acad. Sci. USA, 80:775-779). These cells appear to segregate lysosomal enzymes by an alternate intracellular pathway. To determine the basis for the lack of mannose 6-phosphate receptor activity in these cell lines, we studied the biosynthesis of the receptor in receptor-positive (BW5147) and receptor-deficient (P388D1 and MOPC 315) cells. The cells were labeled with [2-3H]mannose or [35S]methionine and the receptor was immunoprecipitated with an antireceptor antiserum. BW5147 cells synthesize a receptor protein whose size increases after translation/glycosylation. MOPC 315 cells produce an apparently normal receptor and degrade it rapidly. P388D1 cells fail to synthesize any detectable receptor. The receptor from BW5147 and MOPC 315 cells is a glycoprotein with both high mannose and complex asparagine-linked oligosaccharides. The complex-type units become fully sialylated and remain so during long periods of chase.

Heparin inhibits P388D1 cells adherence and metastasis to peripheral lymph nodes in vitro and in vivo.

Heparin is known to attenuate tumor metastasis mainly by inhibiting the interaction between L-selectin and its ligand. However, the mechanism of heparin on lymphoma is unclear. This report demonstrates that both L-selectin/h IgG chimeric protein and heparin treatment can significantly inhibit the adhesion of P388D1 cells onto lymphatic sinusoids and marginal sinusoids in vitro, that heparin can attenuate P388D1 cell homing to lymph nodes in vivo at 12 hours, and that heparin significantly reduced P388D1 cells metastasis to lymph nodes 18 days after injection. These results indicate that heparin may act as a ligand for L-selectin on the P388D1 macrophage-like lymphoma cell line to attenuate tumor growth and metastasis.

Mouse CD40-transfected cell lines cannot exhibit the binding and RANTES-stimulating activity of exogenous heat shock protein 70.

Here we demonstrate the inducible mouse Hsp72 binds markedly to lymphoid neoplastic macrophage-like P388D1 cells. To examine whether mouse CD40 can play a role in signaling exogenously administered HSP70 in a fashion similar to that of human CD40, we established mouse CD40-transfectants of both human 293 cells and murine-pro-B cell line Ba/F3. A small portion of mouse CD40 expressed on 293-derived transfectants was the mature form with a signal-transducible C-terminal domain, whereas a majority of expressed antigen showed the molecular size smaller than we expect. Flow cytometry showed that mouse Hsp72, but neither its deletion variants nor the related Escherichia coli DnaK, bound to the 293-derived transfectants regardless of CD40 expression. CD40 molecules expressed on the transfectants showed the binding of soluble form of CD40L but this binding was not inhibited by excess amount of HSP70. CD40L, but not any HSP70 recombinant proteins, stimulated the production of chemokine RANTES in the transfectants. Furthermore, no RANTES production was induced by HSP70-RCMLA complex in the transfectants, although it binds to 293-derived cells in a CD40-independent manner. No interaction between mouse CD40 and HSP70 recombinant proteins was detected by using the Ba/F3-derived transfectants that express the mature form of mouse CD40. The present results imply that mouse CD40 expressed on the transfectants differs from its human homolog in the binding of exogenously administered HSP70.

Distinct murine macrophage receptor pathway for human triglyceride-rich lipoproteins.

Murine P388D1 macrophages have a receptor pathway that binds human hypertriglyceridemic very low density lipoproteins (HTG-VLDL) that is fundamentally distinct from the LDL receptor pathway. Trypsin-treated HTG-VLDL (tryp-VLDL), devoid of apolipoprotein (apo)-E, fail to bind to the LDL receptor, yet tryp-VLDL and HTG-VLDL cross-compete for binding to P388D1 macrophage receptors, indicating that these lipoproteins bind to the same sites. The specific, high affinity binding of tryp-VLDL and HTG-VLDL to macrophages at 4 degrees C is equivalent and at 37 degrees C both produce rapid, massive, curvilinear (receptor-mediated) triglyceride accumulation in macrophages. Ligand blots show that P388D1 macrophages express a membrane protein of approximately 190 kD (MBP190) that binds both tryp-VLDL and HTG-VLDL; this binding is competed by HTG-VLDL, trypsinized HTG-VLDL, and trypsinized normal VLDL but not by normal VLDL or LDL. The macrophage LDL receptor (approximately 130 kD) and cellular uptake of beta-VLDL, but not MBP 190 nor uptake of tryp-VLDL, are induced when cells are exposed to lipoprotein-deficient medium and decreased when cells are cholesterol loaded. Unlike the macrophage LDL receptor, MBP 190 partitions into the aqueous phase after phase separation of Triton X-114 extracts. An anti-LDL receptor polyclonal antibody blocks binding of HTG-VLDL to the LDL receptor and blocks receptor-mediated uptake of beta-VLDL by P388D1 cells but fails to inhibit specific cellular uptake of tryp-VLDL or to block binding of tryp-VLDL to MBP 190. Human monocytes, but not human fibroblasts, also express a binding protein for HTG-VLDL and tryp-VLDL similar to MBP 190. We conclude that macrophages possess receptors for abnormal human triglyceride-rich lipoproteins that are distinct from LDL receptors in ligand specificity, regulation, immunological characteristics, and cellular distribution. MBP 190 shares these properties and is a likely receptor candidate for the high affinity uptake of TG-rich lipoproteins by macrophages.

Glutathione cycle activity and pyridine nucleotide levels in oxidant-induced injury of cells.

Exposure of target cells to a bolus of H2O2 induced cell lysis after a latent period of several hours, which was prevented only when the H2O2 was removed within the first 30 min of injury by addition of catalase. This indicated that early metabolic events take place that are important in the fate of the cell exposed to oxidants. In this study, we described two early and independent events of H2O2-induced injury in P388D1 macrophagelike tumor cells: activation of the glutathione cycle and depletion of cellular NAD. Glutathione cycle and hexose monophosphate shunt (HMPS) were activated within seconds after the addition of H2O2. High HMPS activity maintained glutathione that was largely reduced. However, when HMPS activity was inhibited--by glucose depletion or by incubation at 4 degrees C--glutathione remained in the oxidized state. Total pyridine nucleotide levels were diminished when cells were exposed to H2O2, and the breakdown product, nicotinamide, was recovered in the extracellular medium. Intracellular NAD levels fell by 80% within 20 min of exposure of cells to H2O2. The loss of NADP(H) and stimulation of the HMPS could be prevented when the glutathione cycle was inhibited by either blocking glutathione synthesis with buthionine sulfoximine (BSO) or by inhibiting glutathione reductase with (1,3-bis) 2 chlorethyl-1-nitrosourea. The loss of NAD developed independently of glutathione cycle and HMPS activity, as it also occurred in BSO-treated cells.

Membrane transport influences the rate of accumulation of cytosine arabinoside in human leukemia cells.

The role of membrane transport in the cellular accumulation of 1-beta-D-arabinofuranosylcytosine (ara-C) was studied in freshly isolated human acute leukemia cells. Patient cells had low rates for ara-C transport as compared with human and murine experimental cells and correspondingly low binding capacities for the nucleoside transport inhibitor, nitrobenzylmercaptopurine riboside (NBMPR). At 1 microM ara-C, the rate of net cellular accumulation was close to the membrane transport rate, and NBMPR inhibited transport and accumulation to the same extent. The rate of ara-C accumulation was half maximal at only 3-5 microM, a level much lower than that required for murine cells (67-85 microM). At concentrations below 1 microM the rate of ara-C accumulation was determined primarily by the transport rate, but at higher concentrations above 10 microM, phosphorylation capacity was the principal determinant of the net uptake rate. This difference in the role of transport at high and low ara-C concentrations may explain, in part, the efficacy of high-dose ara-C in patients refractory to standard dose protocols.

Oxidant injury of cells. DNA strand-breaks activate polyadenosine diphosphate-ribose polymerase and lead to depletion of nicotinamide adenine dinucleotide.

To determine the biochemical basis of the oxidant-induced injury of cells, we have studied early changes after exposure of P388D1 murine macrophages to hydrogen peroxide. Total intracellular NAD+ levels in P388D1 cells decreased with H2O2 concentrations of 40 microM or higher. Doses of H2O2 between 0.1 and 2.5 mM led to an 80% depletion of NAD within 20 min. With doses of H2O2 of 250 microM or lower, the fall in NAD and, as shown previously, ATP, was reversible. Higher doses of H2O2 that cause ultimate lysis of the cells, induced an irreversible depletion of NAD and ATP. Poly-ADP-ribose polymerase, a nuclear enzyme associated with DNA damage and repair, which catalyzes conversion of NAD to nicotinamide and protein-bound poly-ADP-ribose, was activated by exposure of the cells to concentrations of 40 microM H2O2 or higher. Activation of poly-ADP-ribose polymerase was also observed in peripheral lymphocytes incubated in the presence of phorbol myristate acetate-stimulated polymorphonuclear neutrophils. Examination of the possibility that DNA alteration was involved was performed by measurement of thymidine incorporation and determination of DNA single-strand breaks (SSB) in cells exposed to H2O2. H2O2 at 40 microM or higher inhibited DNA synthesis, and induced SSB within less than 30 s. These results suggest that DNA damage induced within seconds after addition of oxidant may lead to stimulation of poly-ADP-ribose polymerase, and a consequent fall in NAD. Excessive stimulation of poly-ADP-ribose polymerase leads to a fall in NAD sufficient to interfere with ATP synthesis.

Functional evaluation of multidrug resistance transporter activity in surgical samples of solid tumors.

Determination of multidrug resistance (MDR) activity of tumor cells could provide important information for the personalized therapy of cancer patients. The functional calcein assay (MultiDrug Quant Assay, Solvo Biotechnology, Budaörs, Hungary) has been proven to be clinically valuable in hematological malignancies by determining the transporter activity of MDR protein 1 (MDR1, ATP-binding cassette protein [ABC] B1, P-glycoprotein-170) and MDR-related protein 1 (MRP1, ABCC1). In this study, we evaluated if the same functional test was adaptable for the analysis of MDR activity in solid tumors. For this purpose, tissue specimens of human colorectal cancer samples were subjected to limited enzymatic digestion by collagenase to provide a single-cell suspension; dead cells were excluded by 7-aminoactinomycin D staining, and epithelial cancer cells were detected by Cy5-conjugated anti-BerEP4 monoclonal antibody. The transporter functions of MDR1 and MRP1 in viable epithelial cells were assessed by flow cytometry detecting the intracellular accumulation of calcein dye after exposing cells to various MDR inhibitors. Collagenase disintegration preserved the MDR activity and the antigenicity of tumor cells. Thus using the extended calcein assay provided sufficient viable and functionally active tumor cells from surgical biopsies to determine the functional MDR activity. In conclusion, the newly described modified calcein assay may be applicable for evaluating the MDR phenotype in solid tissue specimens from colorectal forceps biopsy to surgical samples.

Cycloimide bacteriochlorin p derivatives: photodynamic properties and cellular and tissue distribution.

Reactive oxygen species generated by photosensitizers are efficacious remedy for tumor eradication. Eleven cycloimide derivatives of bacteriochlorin p (CIBCs) with different N-substituents at the fused imide ring and various substituents replacing the 3-acetyl group were evaluated as photosensitizers with special emphasis on structure-activity relationships. The studied CIBCs absorb light within a tissue transparency window (780-830 nm) and possess high photostability at prolonged light irradiation. The most active derivatives are 300-fold more phototoxic toward HeLa and A549 cells than the clinically used photosensitizer Photogem due to the substituents that improve intracellular accumulation (distribution ratio of 8-13) and provide efficient photoinduced singlet oxygen generation (quantum yields of 0.54-0.57). The substituents predefine selective CIBC targeting to lipid droplets, Golgi apparatus, and lysosomes or provide mixed lipid droplets and Golgi apparatus localization in cancer cells. Lipid droplets and Golgi apparatus are critically sensitive to photoinduced damage. The average lethal dose of CIBC-generated singlet oxygen per volume unit of cell was estimated to be 0.22 mM. Confocal fluorescence analysis of tissue sections of tumor-bearing mice revealed the features of tissue distribution of selected CIBCs and, in particular, their ability to accumulate in tumor nodules and surrounding connective tissues. Considering the short-range action of singlet oxygen, these properties of CIBCs are prerequisite to efficient antitumor photodynamic therapy.

Binding with serum components favorably affects cellular uptake of 111In-oligonucleotide in a leukemia cell line.

PURPOSE: The influence of serum components on the intracellular uptake of an 111In-oligonucleotide (ODN) against mdr1 mRNA was investigated in the murine leukemia cell line, P388/S, and its mdr1-overexpressed P388/R. METHODS: 111In-ODNs naked and vectorized with lipids were analyzed for binding with serum components using high-performance liquid chromatography (HPLC). 111In-ODN was incubated in albumin and transferrin solutions. 111In-DTPA and 111In-mononucleotide were incubated in serum. Degradation of naked 111In-ODN was detected in phosphate buffered saline (PBS) and serum containing endonuclease S1. Cellular uptakes of naked and vectorized 111In-ODN in the above cells were examined with and without fetal calf serum (FCS). RESULTS: Time-dependent binding of naked and vectorized 111In- ODN with serum components was observed throughout 24 hours. Transchelation of 111In to transferrin was not detected. HPLC profiles of 111In-DTPA and 111In-mononucleotide did not change in serum. Degradation of 111In-ODN by S1 was less remarkable in serum than in PBS. Specific accumulation of vectorized 111In-ODN in P388/R cells was achieved in culture with and without 10% FCS. CONCLUSIONS: This study verified the intense binding of ODN with serum components, leading to no inhibition on ODN intracellular specific uptake. Binding with serum components protects 111In-ODN from degradation by endonuclease and thus may facilitate ODN transmembrane delivery.

Isoquino[4,5-bc]acridines: design, synthesis and evaluation of DNA binding, anti-tumor and DNA photo-damaging ability.

Several novel isoquino[4,5-bc]acridine derivatives have been designed and synthesized. Their DNA-binding, anti-tumor and DNA-photo-damaging properties were investigated. A4 exhibited the highest anti-tumor activities against both A 549 (human lung cancer cell) and P388 (murine leukemia cells). All these compounds were found to be more cytotoxic against P388 than against A549. Under 365-nm light irradiation, A3 damaged plasmid DNA pBR322 at <2 microM and cleaved DNA from form I to 100% form II by 50 microM. The mechanism studies revealed that A3 damaged DNA by electron transfer mechanism and singlet oxygen species.

Intracellular adriamycin levels and cytotoxicity in adriamycin-sensitive and adriamycin-resistant P388 mouse leukemia cells.

Adriamycin (ADR) (NSC-123127) uptake and retention in ADR-sensitive P388 leukemia (P388/S) and ADR-resistant P388 leukemia (p388/R) cells were compared by fluorometry and laser flow cytometry (FCM) and were correlated with cytotoxic effects. Drug levels in P388/R cells treated in vitro with ADR (1-10 micrograms/ml) were twofold to fourfold lower than were levels in similarly treated P388/S cells FCM analysis of P388/S and P388/R cells exposed in vitro to ADR showed qualitative and quantitative differences in ADR fluorescence profiles of drug-treated cells (1-5 micrograms/ml) but not of the isolated nuclei (0.5- 10 micrograms/ml). Drug-induced perturbations in cell cycle traverse and chromosome aberrations were seen in P388/S but not in P388/R cells treated with 0.5-5 micrograms ADR/ml in vitro or 4-8 mg ADR/kg in vivo. The role of FCM in rapidly comparing and quantitating cellular ADR fluorescence profiles of ADR-sensitive and ADR-resistant tumors was demonstrated.

Double labeling of S-phase murine cells with bromodeoxyuridine and a second DNA-specific probe.

A rapid method has been developed which combines immunofluorescence and autoradiography and permits the double labeling of DNA. P388 murine leukemic cells were incubated with bromodeoxyuridine and tritiated thymidine simultaneously. After fixation, the sample was first processed with a monoclonal antibody to bromodeoxyuridine (RPMB I) so that any cell in S-phase was brightly fluorescent (RPMB technique). Next, tritiated thymidine grains were developed by autoradiography, and the result demonstrated fluorescence as well as black grains in each S-phase cell. P388 cells sensitive (P388S) and resistant (P388R) to 1-beta-D-arabinofuranosylcytosine (ara-C) were incubated with bromodeoxyuridine and [3H]ara-C simultaneously. Processing by autoradiography and RPMB techniques revealed that all S-phase cells in the P388S sample demonstrated vivid "double labeling," whereas P388R cells only revealed bright green fluorescence in S-phase cells, but no grains, confirming a lack of ara-C incorporation into the DNA by this line. Finally, a computerized digital analysis system attached to a microphotometer was used to quantitate fluorescence and grains per cell, and the data demonstrated that the number of [3H]ara-C grains in each P388S cell was inversely proportional to the degree of fluorescence in that cell, indicating that DNA synthesis was inhibited by ara-C. In conclusion, a simple, easy-to-use double-labeling method has been introduced which will be useful to a wide variety of researchers, because this technique together with the digital analysis system offers the possibility of measuring drug sensitivities in individual cells.

Characterization of multidrug resistance by fluorescent dyes.

Fluorimetric techniques were used to examine accumulation of fluorescent probes by the P388 murine leukemia and an anthracycline-resistant subline, P388/Adriamycin(ADR), which expresses the multidrug-resistant phenotype. P388 could be differentiated from P388/ADR on the basis of fluorescence intensity measurements using 3 classes of cationic dyes that are sensitive to membrane potential differences: rhodamine esters, cyanines, and styrylpyridinium dyes. But fluorescence intensity differences were also observed with potential-insensitive dyes: zwitterionic rhodamines and an acridine orange derivative. In all cases, fluorescence intensity differences were caused by impaired dye accumulation, and could be eliminated by treatment of P388/ADR cells with verapamil. Moreover, fluorescence signals from 2 anionic potential-sensitive dyes, merocyanine 540 and a bis-oxonol, were identical in P388 and P388/ADR. None of these dyes could be used to delineate CCRF-CEM, a lymphoblastic leukemia of human origin from the CEM/VM-1 subline that exhibits a markedly atypical drug resistance pattern not based on an enhanced outward transport. But accumulation of both neutral and cationic dyes was impaired in CEM/VLB100, a subline of CCRF-CEM expressing mdr. These studies show that many cationic and neutral fluorescent probes are substrates for the enhanced outward drug transport system associated with P388/ADR cells, and cannot be used to probe membrane-potential differences in cells expressing the mdr phenotype. With several dyes, differences in fluorescence intensity were sufficient so that flow cytometry could be used to delineate P388 from P388/ADR and CCRF-CEM from CEM-VLB100. The latter technique may be useful for identifying malignant cell populations expressing multidrug resistance in patients with neoplastic disease.

Metabolism of hexamethylene bisacetamide and its metabolites in leukemic cells.

We investigated whether leukemic cell lines could convert hexamethylene bisacetamide (HMBA) to any of the metabolites previously identified and quantified in the urine and plasma of patients treated with HMBA. After 5-7 days of incubation with 1-2 mM HMBA, HL60 human promyelocytic leukemic cells, L1210 and P388 murine lymphoblastic leukemic cells, and Friend murine erythroleukemia cells contained 4 of the previously identified metabolites of HMBA. Gas chromatography/mass spectrometry confirmed the presence of N-acetyl-1,6-diaminohexane (NADAH), 1,6-diaminohexane (DAH), 6-acetamidohexanoic acid (AcHA), and 6-aminohexanoic acid (AmHA). Gas chromatography with nitrogen-phosphorus selective detection was used to quantify cellular concentrations of each metabolite. Cellular concentrations of AmHA and DAH were greater than those of NADAH and AcHA but no concentration of a metabolite exceeded that of HMBA. Metabolites were not detected in media from cells incubated with HMBA. Friend murine erythroleukemia cells that were resistant to HMBA contained only HMBA and NADAH. Moreover, the concentrations of NADAH in Friend murine erythroleukemia cells that were resistant to HMBA were less than those in the other cell lines studied. HL60 cells accumulated HMBA rapidly. NADAH, DAH, AcHA, and AmHA appeared sequentially in HL60 cells that were incubated with HMBA. NADAH appeared very rapidly, but concentrations of DAH were greater than or equal to those of NADAH by 8 h. AcHA and AmHA were not detected in cells before 24-48 h of incubation with HMBA. HL60 cells incubated with individual HMBA metabolites were able to accumulate each compound and to interconvert some: cells incubated with NADAH also contained DAH, AcHA, and AmHA; cells incubated with AcHA also contained low concentrations of AmHA; cells incubated with DAH also contained AmHA; and cells incubated with AmHA contained no other HMBA metabolites. HMBA was not present in cells incubated with any of its known metabolites. These results document the ability of various leukemic cells to metabolize HMBA, indicate the unidirectional catabolism of that compound, and may have implications as to its mechanism of action.