Bernd L. Hartmann

Apoptosis induced by the histone deacetylase inhibitor sodium butyrate in human leukemic lymphoblasts

The histone deacetylase inhibitor and potential anti‐cancer drug sodium butyrate is a general inducer of growth arrest, differentiation, and in certain cell types, apoptosis. In human CCRF‐CEM, acute T lymphoblastic leukemia cells, butyrate, and other histone deacetylase inhibitors caused G2/M cell cycle arrest as well as apoptotic cell death. Forced G0/G1 arrest by tetracycline‐regulated expression of transgenic p16/INK4A protected the cells from butyrate‐induced cell death without affecting the extent of histone hyperacetylation, suggesting that the latter may be necessary, but not sufficient, for cell death induction. Nuclear apoptosis, but not G2/M arrest, was delayed but not prevented by the tripeptide broad‐range caspase inhibitor benzyloxycarbonyl‐Val‐Ala‐Asp·fluoromethylketone (zVAD) and, to a lesser extent, by the tetrapeptide ‘effector caspase’ inhibitors benzyloxycarbonyl‐Asp‐Glu‐Val‐Asp·fluoromethylketone (DEVD) and benzyloxycarbonyl‐Val‐Glu‐Ile‐Asp‐fluoromethyl‐ketone (VEID); however, the viral protein inhibitor of ‘inducer caspases’, crmA, had no effect. Bcl‐2 overexpression partially protected stably transfected CCRF‐CEM sublines from butyrate‐induced apoptosis, but showed no effect on butyrate‐induced growth inhibition, further distinguishing these two butyrate effects. c‐myc, constitutively expressed in CCRF‐CEM cells, was down‐regulated by butyrate, but this was not causative for cell death. On the contrary, tetracycline‐induced transgenic c‐myc sensitized stably transfected CCRF‐CEM derivatives to butyrate‐induced cell death.—Bernhard, D., Ausserlechner, M. J., Tonko, M., Löffler, M., Hartmann, B. L., Csordas, A., Kofler, R. Apoptosis induced by the histone deacetylase inhibitor sodium butyrate in human leukemic lymphoblasts. FASEB J. 13, 1991–2001 (1999)

Ceramides induce a form of apoptosis in human acute lymphoblastic leukemia cells that is inhibited by Bcl-2, but not by CrmA

The generation of ceramides by the action of acidic and/or neutral sphingomyelinases has been implicated in many forms of apoptosis. We investigated whether exposure to ceramides is sufficient to induce apoptosis in human leukemia cells and, if so, what the characteristics of this form of apoptosis might be. Treatment of the acute lymphoblastic T‐cell line CEM‐C7H2 with short‐ and medium‐chain ceramide analogs (C2‐, C6‐, and C8‐ceramide) resulted in apoptosis, whereas the inactive C2‐dihydroceramide had no effect on cell survival. Induction of apoptosis was relatively slow (∼40% after 24 h) and required high concentrations of ceramide analogs (40–100 μM). To investigate a possible involvement of interleukin 1‐β‐converting enzyme (ICE) or ICE‐related proteases, we treated CEM‐C7H2 sublines constitutively expressing the vaccinia virus protease inhibitor crmA with ceramide analogs. Although such cells were completely resistant to apoptosis induced by antibodies to the Apo‐1/Fas surface receptor (a form of apoptosis known to be inhibitable by CrmA), they were not protected from ceramide‐induced cell death. In contrast, tetracycline‐regulated overexpression of Bcl‐2 protected CEM‐C7H2 sublines stably transfected with corresponding constructs from ceramide‐induced apoptosis. Thus, in these human leukemia cells, ceramides induce a relatively slow death response that can be prevented by Bcl‐2, but is independent of CrmA‐inhibitable proteases. These characteristics distinguish ceramide‐induced from other forms of apoptosis, such as Apo‐1/Fas‐induced cell death where ceramide production has been causally implicated.

Glucocorticoid-receptor-gene defects and resistance to glucocorticoid-induced apoptosis in human leukemic cell lines

The application of giucocorticoids (GC3) in human leukemia is based on apoptosis induction but is often hampered by GC resistance. To delineate resistance mechanisms, we examined 5 GC‐resistant leukemic cell lines, termed CEM‐C7.RI–R5, isolated from the GC‐sensitive human acute‐T‐cell‐leukemic line, CCRF‐CEM‐C7, by selection in GC‐containing medium. GC resistance was ascertained by analyzing cell‐cycle progression, proliferation, and apoptosis. Radioreceptor assays revealed absence of ligand‐binding activity in all clones, suggesting that defects in GC‐receptor (GR) expression cause GC resistance. Analyses of the GR gene revealed that all but one (CEM‐C7.R5) of the clones were heterozygous for the previously described L753F mutation. CEM‐C7.R5 was either hemi‐ or homozygous for the L753F mutation and, hence, lacked a functional GR. Sequencing of the allele not carrying the L753F mutation of the other GC‐resistant sub‐lines revealed additional mutations in the GR gene in 3 cases: CEM‐C7.R1 and R2 had a base‐pair deletion in exon 9 (ΔT740) that resulted in a reading‐frame shift and a pre‐terminal in‐frame stop. Translation of this mutant mRNA would produce a protein lacking 32 amino acids and expressing 4 altered residues at its new C terminus. CEM‐C7.R3 harbored a non‐sense mutation (Q7 10X) in exon 8, and its mRNA would be translated into a protein lacking 67 residues. Only CEM‐C7.R4 cells were devoid of mutations in the coding region of the L753F negative allele. These data suggest that, in the CCRF‐CEM acute‐lymphatic‐leukemia model, mutations in the GR‐gene coding region represent one, but not the only, cause of GC resistance.

Bcl-2 interferes with the execution phase, but not upstream events, in glucocorticoid-induced leukemia apoptosis.

Due to their growth arrest- and apoptosis-inducing ability, glucocorticoids (GC) are widely used in the therapy of various lymphoid malignancies. Cell death is associated with activation of members of the interleukin-1β-converting enzyme (ICE) protease/caspase family and, is presumably prevented by the anti-apoptotic protein Bcl-2. To further address the role of Bcl-2 in GC-mediated cytotoxicity, we generated subclones of the GC-sensitive human T-cell acute lymphoblastic leukemia line CCRF-CEM, in which transgenic Bcl-2 expression is regulated by tetracycline. Up to about 48 h, exogenous Bcl-2 almost completely protected these cells from apoptosis, digestion of poly-ADP ribose polymerase (PARP) and generation of Asp-Glu-Val-Asp cleaving (DEVDase) activity. However, when the cells were cultured for another 24 h in the continuous presence of GC, they underwent massive apoptosis that was associated with DEVDase activity and PARP cleavage. Bcl-2 did not markedly affect GC-mediated growth arrest, thereby separating the anti-proliferative from the apoptosis-inducing effect of GC. Moreover, Bcl-2 did not prevent the dramatic reduction in the levels of several mRNAs observed during GC treatment, including the transgenic Bcl-2 mRNA. Thus, Bcl-2 can be placed upstream of effector caspase activation, but downstream of other GC-regulated events, such as growth arrest and the potentially critical repression of steady state levels of multiple mRNA.

Increased lactate production follows loss of mitochondrial membrane potential during apoptosis of human leukaemia cells

Acute tumour‐lysis syndrome (ATLS) is a frequently fatal complication after cytoreductive leukaemia therapy. Lactic acidosis is associated with ATLS and its extent is correlated with the severity of ATLS. In the course of cytoreductive therapy, apoptosis is induced in tumour cells, which results in loss of mitochondrial function. We hypothesize that loss of mitochondrial function leads to compensatory glycolysis, which is the main cause of lactate accumulation and acidosis. We tested this hypothesis using the model of glucocorticoid‐induced apoptosis in the human acute lymphoblastic leukaemia cell line CCRF‐CEM. After induction of glucocorticoid‐induced apoptosis, a biphasic course of lactate production was observed. Prior to the onset of apoptosis, i.e. prior to the loss of membrane potential, lactate production was reduced. However, subsequent to loss of mitochondrial membrane potential a massive increase in lactate production was observed (15·5 ± 0·5 versus 10·17 ± 0·09 mmol/106 cells, P = 0·001). We also demonstrated that inhibition of respiratory chain activity by antimycin A resulted in excess lactate production. In the model cell line used, conditional bcl‐2 expression delayed glucocorticoid‐induced apoptosis by protecting against loss of mitochondrial membrane potential; bcl‐2 expression delayed the increase in lactate production and had no effect on the pre‐apoptotic drop in lactate production. Apoptosis‐induced lactate production was also observed in other cell lines (HL60, THP1 and OPM2) with various cytotoxic agents [doxorubicin, gemcitabine and vumon (VM26)]. Thus, the data suggest that lactate acidosis can be caused by apoptotic loss of mitochondrial function and massive apoptosis of a tumour mass via lactic acidosis may be the essential pathological event in ATLS.

p53-induced apoptosis in the human T-ALL cell line CCRF-CEM

The tumor suppressor p53 has been implicated in apoptosis induction and is mutated in human T-ALL CCRF-CEM cells. To investigate possible consequences of wild-type p53 loss, we reconstituted CEM-C7H2, a subclone of CCRF-CEM, with a temperature-sensitive p53 allele (p53ts). Stably transfected lines expressed high levels of p53ts and shift to the permissive temperature (32°C) caused rapid induction of p53-regulated genes, such as p21CIP1/WAF1, mdm-2 and bax. This was followed by extensive apoptosis within 24 h to 36 h, supporting the notion that mutational p53 inactivation contributed to the malignant phenotype. p53-dependent apoptosis was preceded by digestion of poly(ADP-ribose) polymerase, a typical target of interleukin-1β-converting enzyme (ICE)-like proteases/caspases, and was markedly resistant to the ICE/caspase-1 and FLICE/caspase-8 inhibitor acetyl - Tyr - Val - Ala - Asp.chloromethylketone (YVAD), but sensitive to the CPP32/caspase-3 inhibitor benzyloxycarbonyl -Asp-Glu-Val- Asp.fluoromethylketone (DEVD) and benzyloxycarbonyl-Val-Ala-Asp.fluoromethylketone (zVAD), a caspase inhibitor with broader specificity. This indicated an essential involvement of caspases, but argued against a significant role of ICE/caspase-1 or FLICE/caspase-8. Actinomycin D or cycloheximide prevented cell death, suggesting that, in this system, p53-induced apoptosis depends upon macromolecule biosynthesis. Introduction of functional p53 into CEM cells enhanced their sensitivity to the DNA-damaging agent doxorubicin, but not to the tubulin-active compound vincristine. Thus, mutational p53 inactivation in ALL might entail relative resistance to DNA-damaging, but not to tubulin-destabilizing, chemotherapy.

The interleukin 1β-converting enzyme inhibitor CrmA prevents Apo1/Fas- but not glucocorticoid-induced poly(ADP-ribose) polymerase cleavage and apoptosis in lymphoblastic leukemia cells

Glucocorticoids (GC) induce programmed cell death (apoptosis) in immature lymphocytes and are an essential component in the therapy of acute lymphatic leukemia. The mechanism underlying GC‐induced apoptosis particularly in leukemia cells is, however, not well understood. Most forms of apoptosis seem to employ a common final effector pathway characterized by specific proteolytic events mediated by interleukin 1β‐converting enzyme (ICE) and/or other ICE‐like cysteine proteases. These events may result in the morphologic changes characteristic of apoptosis. To determine whether a similar proteolytic pathway is activated during GC‐induced leukemia cell apoptosis, we investigated poly(ADP‐ribose) polymerase (PARP), a typical target of ICE‐like proteases, during GC‐induced apoptosis of the human acute T‐cell leukemic cell line CEM‐C7H2. Our studies showed proteolytic PARP cleavage suggestive of activation of ICE‐like proteases that preceeded morphologic signs of apoptosis. We further established stably transfected CEM‐C7H2 sublines expressing the cowpox virus protein CrmA that inhibits some, but not all, ICE‐like proteases. GC‐induced PARP cleavage and apoptosis were neither inhibited nor delayed in crmA‐expressing cell lines. In contrast, crmA expression rendered the same lines resistant to Apo1/Fas‐induced PARP cleavage and apoptosis. Thus, different proteases might be activated during the effector phases of GC‐ and Apo1/Fas‐induced apoptosis in human leukemia cells.

c-Myc does not prevent glucocorticoid-induced apoptosis of human leukemic lymphoblasts

Due to their growth arrest- and apoptosis-inducing ability, glucocorticoids (GC) are widely used in the therapy of various lymphoid malignancies. The signal transduction pathways leading to this clinically-relevant form of apoptosis have, however, not been sufficiently elucidated. GC bind to their specific receptor, a ligand-activated transcription factor of the Zn-finger type, that activates or represses transcription of GC-responsive genes. Previous studies in leukemia cells suggested that transcriptional repression of c-myc expression might be the crucial event in GC-induced apoptosis, although in other systems, c-Myc apparently increased the sensitivity to cell-death inducers. To address this controversy, we stably transfected the GC-sensitive human T-ALL cell line CEM-C7H2 with constructs allowing tetracycline-regulated expression of c-Myc. Subsequent analyses of these cell lines showed that overexpression of c-Myc per se had little, if any, effect on cell viability, although it rendered the cells more sensitive to apoptosis induced by low serum, confirming the functionality of the expressed transgene. More importantly, however, when the cells were treated with GC in the presence of exogenous c-Myc, they underwent apoptosis exceeding that in cells treated in the absence of transgenic c-Myc. The data indicate that c-myc downregulation is not critical for induction of cell-death by GC in this system, and support the notion that c-Myc sensitizes cells to apoptosis-inducing agents.

Irradiation induces G2/M cell cycle arrest and apoptosis in p53-deficient lymphoblastic leukemia cells without affecting Bcl-2 and Bax expression

The tumor suppressor p53 has been implicated in gamma irradiation-induced apoptosis. To investigate possible consequences of wild-type p53 loss in leukemia, we studied the effect of a single dose of gamma irradiation upon p53-deficient human T-ALL (acute lymphoblastic leukemia) CCRF–CEM cells. Exposure to 3–96 Gy caused p53-independent cell death in a dose and time-dependent fashion. By electron microscopic and other criteria, this cell death was classified as apoptosis. At low to intermediate levels of irradiation, apoptosis was preceded by accumulation of cells in the G2/M phase of the cell division cycle. Expression of Bcl-2 and Bax were not detectably altered after irradiation. Expression of the temperature sensitive mouse p53 V135 mutant induced apoptosis on its own but only slightly increased the sensitivity of CCRF–CEM cells to gamma irradiation. Thus, in these, and perhaps other leukemia cells, a p53- and Bcl-2/Bax-independent mechanism is operative that efficiently senses irradiation effects and translates this signal into arrest in the G2/M phase of the cell cycle and subsequent apoptosis.

Interaction between dexamethasone and butyrate in apoptosis induction: non-additive in thymocytes and synergistic in a T cell-derived leukemia cell line

In thymocytes butyrate and trichostatin A are unable to augment dexamethasone-induced apoptosis. In cultured rat thymocytes the extent of apoptosis induced by dexamethasone alone did not increase by addition of 0.1–10 mM butyrate. Even more pronounced was the non-additive interrelationship between dexamethasone and trichostatin A, as trichostatin A-induced apoptosis was not only blocked by the presence of dexamethasone but dexamethasone-induced apoptosis was also partially inhibited in the presence of 0.1–0.5 μM trichostatin A. The fact that the non-additive relationship with dexamethasone for apoptosis induction was observed with both histone deacetylase inhibitors suggests that in thymocytes this phenomenon is related to histone acetylation. In contrast to this, in the human T cell-derived leukemia cell line CEM-C7H2, dexamethasone did not block butyrate- or trichostatin A-induced apoptosis; moreover, butyrate, in the concentration range of 0.1–1 mM, had a marked synergistic effect on dexamethasone-induced apoptosis. This synergism, however, was not mimicked by trichostatin A, indicating that the effect is not related to histone acetylation but rather due to a pleiotropic effect of butyrate. Furthermore, in CEM-C7H2 cells, at higher concentrations of butyrate (5–10 mM) or trichostatin A (0.4–0.8 μM), there was a minor but reproducible antagonistic effect of dexamethasone on apoptosis induced by each of the two histone deacetylase inhibitors, suggesting that this antagonistic effect too, is related to histone hyperacetylation.