Novel vitamin K-dependent pathways regulating cell survival.

Historically, the vitamin K1-dependent proteins have been associated primarily with blood coagulation and secondarily with bone formation. Recent identification of K1-dependent proteins as specific ligands for the receptor tyrosine kinases (RTKs) that can stimulate cell replication and transformation and participate in cell survival highlighted a previously unrecognized and potentially important role for vitamin K1 in cell signaling. Growing evidence suggests that most normal and tumor cells possess an active K1-dependent gamma-carboxylation mechanism necessary for the production of gamma-carboxyglutamic acid (Gla)-containing proteins. Gla residues in proteins facilitate calcium-dependent protein/phospholipid interaction. Recent studies demonstrating the potentially positive effects of a vitamin K-dependent receptor:ligand system on cell growth and survival in general and the effects of the overexpression of these RTKs on malignant cell survival provide a new perspective on the role of vitamin K1, its dependent protein ligands, and their receptors. These cumulative observations also provide an explanation for the rigidly controlled K1 levels in the mammalian fetus and the minimal hepatic stores in the adult.

Apoptosis.

Mechanisms in Hematology is a book with an accompanying interactive CD-ROM designed to assemble basic concepts that underlie clinical understanding and progress. It is presented as a concise text with a series of diagrams that distill diffuse information into a compact form. The interactive CD, in particular, brings many of the processes "to life" as details of the more complex pathways are conveyed in clear visual images. The text begins with the basic molecular biology that underlies hematological and oncological physiology/pathology-cell signaling, adhesion molecules, and apoptosis. This is followed by sections, among others, on hematopoiesis, iron, B12, and folate metabolism, neutrophil function, immunoproteins, chemotherapy, and coagulation. With the permission of the authors and publisher, Stem Cells has reproduced the section on apoptosis, which we think our readers will enjoy.

Apoptosis.

Mechanisms in Hematology is a book with an accompanying interactive CD-ROM designed to assemble basic concepts that underlie clinical understanding and progress. It is presented as a concise text with a series of diagrams that distill diffuse information into a compact form. The interactive CD, in particular, brings many of the processes "to life" as details of the more complex pathways are conveyed in clear visual images. The text begins with the basic molecular biology that underlies hematological and oncological physiology/pathology--cell signaling, adhesion molecules, and apoptosis. This is followed by sections, among others, on hematopoiesis, iron, B12, and folate metabolism, neutrophil function, immunoproteins, chemotherapy, and coagulation. With the permission of the authors and publisher. The Oncologist has reproduced the section on apoptosis, which we think our readers will enjoy.

P53, MDM-2, BAX and BCL-2 and drug resistance in chronic lymphocytic leukemia.

Most antitumor agents exert their cytotoxic effect through the induction of apoptosis, and this process may be mediated through an elevation in p53 protein, with a subsequent increase in bax and decrease in bcl-2. p53 also increases mdm-2 expression and mdm-2 may then bind and inactivate p53. Cells from 31 patients with chronic lymphocytic leukemia (CLL) were treated in vitro with 2-chlorodeoxyadenosine (CdA), arabinosyl-2-fluoroadenine (F-ara-A), or chlorambucil (CLB) and drug sensitivity measured using the MTT assay. The protein levels of bax and bcl-2 were measured in CLL cells from 25 patients, and were found to be higher in leukemic cells than in normal B cells. The bcl-2 levels varied three-fold, the bax levels fifteen-fold, and the bax:bcl-2 ratios ranged from 0.44 to 2.91. The expression of mdm-2 mRNA was measured in CLL cells from 28 patients and was found to vary twenty-fold. However, no correlation was observed between drug sensitivity to CdA, F-ara-A, or CLB and the cellular levels of mdm-2 mRNA, or the protein levels of bax or bcl-2, or the bax:bcl-2 ratio. Treatment of CLL cells having wild type p53 with CdA, F-ara-A or CLB produced an increase in p53 protein and mdm-2 mRNA. This was not observed in cells having a p53 mutation, and these cells were highly resistant to both CLB and the nucleoside analogs. In contrast to the nucleoside analogs and CLB, dexamethasone and vincristine had no effect on mdm-2 mRNA levels. Treatment of CLL cells containing a wild type p53 gene with CdA, F-ara-A, or CLB, did not produce any consistent changes in bax or bcl-2. Thus, CdA, F-ara-A and CLB appear to act in CLL cells through a p53-dependent pathway, whereas this does not occur with dexamethasone or vincristine. The cellular levels of mdm-2, bcl-2, bax or the bax:bcl-2 ratios are not predictive indicators of clinical sensitivity in CLL, but an increase in mdm-2 levels after drug treatment is indicative of p53 function in these cells.

The riddle of vitamin K1 deficit in the newborn.

Vitamin K in the fetus and newborn is maintained at levels less than that necessary to achieve full gamma-carboxylation of the K-dependent proteins, including those required for hemostasis. As the infant matures and even into adulthood, there is no significant storage pool for this vitamin, and a K1-deficient state can be produced by placing an adult on a K-deficient diet for 7 to 10 days. Questions arise as to why the level of vitamin K is so rigidly controlled and why the placental gradient in humans and other mammals maintains the fetus in a K-"deficient" state. The evidence is reviewed that suggests that K-dependent proteins are ligands for receptor tyrosine kinases, which, in the rapidly proliferating cell milieu of the fetus, control growth regulation. Increased stimuli may result in growth dysregulation whereas conversely, the further depletion of vitamin K-dependent proteins, as in warfarin toxicity, depletes the required stimuli for normal embryogenesis. These findings argue for the need for tightly controlled levels of vitamin K consistent with normal embryogenesis.

A novel role for vitamin K1 in a tyrosine phosphorylation cascade during chick embryogenesis.

The development of the embryo is dependent upon a highly coordinated repertoire of cell division, differentiation, and migration. Protein-tyrosine phosphorylation plays a pivotal role in the regulation of these processes. Vitamin K-dependent gamma-carboxylated proteins have been identified as ligands for a unique family (Tyro 3 and 7) of receptor tyrosine kinases (RTKs) with transforming ability. The involvement of vitamin K metabolism and function in two well characterized birth defects, warfarin embryopathy and vitamin K epoxide reductase deficiency, suggests that developmental signals from K-dependent pathways may be required for normal embryogenesis. Using a chick embryogenesis model, we now demonstrate the existence of a vitamin K1-dependent protein-tyrosine phosphorylation cascade involving c-Eyk, a member of the Tyro 12 family, and key intracellular proteins, including focal adhesion kinase (pp125FAK), paxillin, and pp60src. This cascade is sensitive to alteration in levels or metabolism of vitamin K1. These findings provide a major clue as to why, in the mammalian (and human) fetus, the K-dependent proteins are maintained in an undercarboxylated state, even to the point of placing the newborn at hemorrhagic risk. The precise regulation of vitamin K1-dependent regulatory pathways would appear to be critical for orderly embryogenesis.

In vitro cytotoxicity of 2-chlorodeoxyadenosine and chlorambucil in chronic lymphocytic leukemia.

Chronic lymphocytic leukemia (CLL) is most commonly treated with the alkylating agent chlorambucil (CLB), although the nucleoside analogs, fludarabine (Flu) and 2-chlorodeoxyadenosine (CdA), are also effective in this disease. In this study, we investigated the in vitro cytotoxicity of CdA and CLB in CLL cells from 12 patients in vitro. Treatment with CLB for 6 h, followed by CdA for 18 h, resulted in 2.3- to 7.5-fold synergistic cytotoxicity in leukemic cells from 10 patients and an additive effect in cells from two patients. In general, synergy was greatest in patients who were sensitive to CLB or CdA, and could be enhanced by increasing the concentrations of CLB or CdA. Synergy was only observed if the cells were treated with CLB prior to CdA. Synergy could not be explained by an increase in the incorporation of CdA into DNA, or by the inhibition of repair of CLB-induced DNA crosslinks by CdA. In contrast to CLL cells, treatment of human marrow in vitro with CLB and CdA resulted in a low level of synergy for CFU-GM cells, and additive cell kill in erythroid progenitors. Thus, treatment with CdA and CLB can produce selective synergistic cell kill in CLL cells, and combination therapy may improve the therapeutic index of these agents in chemosensitive patients.

Chlorambucil in chronic lymphocytic leukemia: mechanism of action.

Chronic lymphocytic leukemia (CLL) is the most common leukemia in Western countries but the clinical presentation and rate of disease progression are highly variable. When treatment is required the most commonly used therapy is the nitrogen mustard alkylating agent, chlorambucil (CLB), with or without prednisone. Although CLB has been used in the treatment of CLL for forty years the exact mechanism of action of this agent in CLL is still unclear. Studies in proliferating model tumor systems have demonstrated that CLB can bind to a variety of cellular structures such as membranes, RNA, proteins and DNA; however, DNA crosslinking appears to be most important for antitumor activity in these systems. In addition, a number of different mechanisms can contribute to CLB resistance in these tumor models including increased drug metabolism, DNA repair and CLB detoxification resulting from elevated levels of glutathione (GSH) and glutathione S-transferase (GST) activity. However, unlike tumor models in vitro, CLL cells are generally not proliferating and studies in CLL cells have raised questions about the hypothesis that DNA crosslinking is the major mechanism of antitumor action for CLB in this disease. CLB induces apoptosis in CLL cells and this appears to correlate with the clinical effects of this agent. Thus, alkylation of cellular targets other than DNA, which can also induce apoptosis, may contribute to the activity of CLB. Alterations in genes such as p53, mdm-2, bcl-2 and bax which control entry into apoptosis may cause drug resistance. Loss of wild-type p53 by mutation or deletion occurs in 10 to 15% of CLL patients and appears to correlate strongly with poor clinical response to CLB. The induction of apoptosis by CLB is paralleled by an increase in P53 and Mdm-2 but this increase in not observed in patients with p53 mutations indicating that with high drug concentrations CLB can produce cell death through P53 independent pathways. The level of Mdm-2 mRNA in the CLL cells is not a useful predictor of drug sensitivity. In addition, although Bax and Bcl-2 are important regulators of apoptosis and the levels of these proteins are elevated in CLL cells compared with normal B cells, the levels of Bax and Bcl-2, or the Bax:Bcl-2 ratio, are not important determinants of drug sensitivity in this leukemia. Finally, whereas CLB and nucleoside analogs may produce cell death in CLL by a P53 dependent pathway other agents, such as dexamethasone or vincristine, may act through P53-independent pathways.

Comparison of antitumor activities of 2-chlorodeoxyadenosine and 9-beta-arabinosyl-2-fluoroadenine in chronic lymphocytic leukemia and marrow cells in vitro.

The in vitro antitumor activities of the nucleoside analogs, 2-chlorodeoxyadenosine (CdA) and 9-beta-arabinosyl-2-fluoroadenine monophosphate (Flu), and the alkylating agent, chlorambucil (CLB), were compared in leukemic cells from 28 patients with chronic lymphocytic leukemia (CLL). On a molar basis, the median sensitivities of the cells to these agents were CLB > CdA > Flu. CLL cells from 90% of the patients had similar relative orders of sensitivities to CdA and Flu, while cells from 10% of the patients showed differential sensitivities to these agents. There was no relationship between the sensitivities of the cells to the nucleoside analogs and sensitivity to CLB. CdA and CLB produced similar toxicities to human marrow progenitor cells in vitro, while Flu was less toxic to these cells. An 18 h exposure to CdA produced significantly greater cell kill of both CLL and marrow progenitor cells than an equivalent 2 h treatment; however, the difference in cytotoxicity was greater for the tumor cells resulting in a higher therapeutic index with the 18 h treatment. The intracellular accumulation of drug varied 5-fold for CdA, with the major metabolite being CdAMP, and 15-fold for Flu, with the major metabolite being F-ara-ATP. However, the accumulation of CdA, Flu or their metabolites did not predict for drug sensitivity. These studies suggest that CdA and Flu cross-resistance cannot be assumed in all CLL patients. The therapeutic effectiveness of CdA may be enhanced by use of a prolonged, low-dose drug regimen.

Deposition of transforming growth factor-beta in the marrow in myelofibrosis, and the intracellular localization and secretion of TGF-beta by leukemic cells.

The marrows of 10 patients with hematologic malignancies were examined by immunohistochemistry using anti TGF-beta antibody, CC(1-30), which detects secreted TGF-beta, and compared with four normal marrows. TGF-beta was not demonstrated in marrows with a normal level of reticulin fibrosis; however, TGF-beta was observed within collagen in marrows having collagen fibrosis or increased reticulin fibrosis. The extent of TGF-beta deposition paralleled the severity of fibrosis (P < .0001), and occurred even with normal or reduced numbers of megakaryocytes. Using another TGF-beta antibody, LC(1-30), which detects intracellular TGF-beta, TGF-beta was detected by immunofluorescence in discrete sites in the cytoplasm of immature and mature myeloid and large granular lymphocytic leukemia cells. These sites colocalized with areas detected by an anti-granule antibody (D545) suggesting that TGF-beta was stored in granules. However, neither the TGF-beta mRNA content nor the degree of TGF-beta secretion by these leukemic cells correlated with the extent of TGF-beta deposition in the marrow. Thus, TGF-beta deposition in marrow may contribute to myelofibrosis, but the source of this cytokine in the absence of megakaryocytes requires further study.

Observations on vitamin K deficiency in the fetus and newborn: has nature made a mistake?

The microsomal mixed function oxidase system metabolizes xenobiotics (Phase I) to products that, if not activated and conjugated for excretion (Phase II), are capable of forming conjugates with cellular macromolecules, including DNA, resulting in toxic, mutagenic, or carcinogenic events. Benzo(a)pyrene (BP), a polycyclic aromatic hydrocarbon, is a model carcinogen for this system. Vitamin K1 (phylloquinone) is a regulator of BP metabolism. These studies demonstrate that K1 is capable of increasing Phase I metabolism and decreasing glutathione transferase activity (Phase II) in chick embryo liver; that deprivation of K1 reduces BP/DNA adducts in mouse liver and reduces tumor formation in mice given intraperitoneal BP; and that K1 supplementation increases BP induced tumor formation in mice. However, epidemiologic studies indicate that children of mothers who smoke during pregnancy may not be at increased risk of cancer. It is known that the placentas from these pregnancies exhibit markedly increased levels of arylhydrocarbon hydroxylase induced by the polycyclic aromatic hydrocarbons in tobacco smoke, but there is no corresponding increase in this enzyme activity in the fetus in such pregnancies. We suggest that the low vitamin K level is a secondary protective mechanism for xenobiotics, such as BP, that may escape the primary placental screen. The recently described role of vitamin K-dependent Gla protein as ligands for receptor tyrosine kinases, also establishes K as a link in cell growth and transformation. It is proposed that the small total body pool of K1 in the adult, which is sufficient only to meet continuing needs, and the even smaller pool in the fetus are protective. This protective effect of low K1 levels is particularly important in the presence of the high mitotic rates and rapid cell turnover in the avian embryo and mammalian fetus.

Chlorambucil induced apoptosis in chronic lymphocytic leukemia (CLL) and its relationship to clinical efficacy.

Chlorambucil induced apoptosis was measured in CLL cells treated with clinically achievable drug doses in vitro. While spontaneous apoptosis occurred in CLL cells incubated in vitro in the absence of drug, the level of apoptosis, as measured by the extent of DNA fragmentation, was greater in cells treated with chlorambucil. In addition, macrophages were shown to engulf drug treated CLL cells in vitro. To determine if chlorambucil can also induce apoptosis in vivo, CLL cells were isolated from patients before treatment and at intervals after clinical therapy with chlorambucil (0.9 mg/kg given over 3 days). Apoptosis was measured in these cells immediately after isolation and following incubation in vitro for 72 hr. No apoptotic changes were detected in cells immediately after isolation either before or after clinical treatment. In contrast, apoptosis was observed in cells that were incubated ex vivo for 72 hr, and the level of apoptosis was greater in cells that were isolated after chlorambucil treatment compared with cells obtained prior to therapy. The increased apoptosis observed in CLL cells ex vivo after therapy was related to the fall in the patient's lymphocyte count. In general, a large increase in apoptosis ex vivo after treatment was followed by a significant decrease in the patient's lymphocyte count. Thus, chlorambucil may produce its antitumor effect in CLL by inducing apoptosis-associated membrane changes that result in rapid clearance of the apoptotic cells by the immune system.

Combination therapy with nucleoside analogs and alkylating agents.

The nucleoside analog, 2'-deoxycoformycin (dCF), and the alkylating agents, chlorambucil (CLB) and cyclophosphamide, are effective agents in the treatment of chronic B cell leukemias and lymphomas. The cyclophosphamide analog, 4-hydroperoxycyclophosphamide (4-HC), generates the same active metabolite as cyclophosphamide in cells and has been used extensively for bone marrow purging in vitro. We have observed that deoxyadenosine (dAdo) plus dCF (dAdo/dCF) inhibit the repair of x-irradiation-induced and bleomycin-induced DNA damage in vitro, and that this results in either synergistic or additive cytotoxicity, respectively. In the present study we examined whether dAdo/dCF, can enhance the antitumor activity of CLB and 4-HC in chronic lymphocytic leukemia (CLL) cells in vitro. CLL cells were treated with CLB for 6 hr and then with dAdo/dCF for 18 hr and cytotoxicity was measured by the MTT assay. Synergy was observed between CLB and dAdo/dCF in CLL cells from 2 patients, with synergy increasing as the CLB dose was raised. In contrast, similar treatment of human bone marrow cells resulted in little or no synergistic cell kill. Treatment of CLL cells from 2 patients with 4-HC for 30 min followed by dAdo/dCF for 18 hr resulted in little synergistic cytotoxicity, although this drug combination did produce an additive cell kill. Thus, combination therapy with nucleoside analogs and alkylating agents may be useful for improving treatment of CLL.

Role of transforming growth factor-beta in chronic lymphocytic leukemia.

TGF-beta is an important immunoregulator as it suppresses proliferation and function of B- and T-lymphocytes. In the present study we have examined the cellular localization and secretion of TGF-beta in B-cells from normal donors and patients with CLL and have assessed the influence of TGF-beta 1 on DNA synthesis in these cells. Using anti-LC(1-30)--a polyclonal anti-TGF-beta 1 antibody--TGF-beta was localized to discrete sites within the cytoplasm of both normal and malignant lymphocytes. These areas co-localized with areas detected by an antigranule antibody (D545), suggesting that TGF-beta may be stored within cytoplasmic secretory vesicles. Both normal B- and CLL cells contained low or undetectable levels of TGF-beta mRNA and secreted low and equivalent amounts of TGF-beta. Compared to untreated cells, DNA synthesis was reduced by TGF-beta 1 to a mean +/- S. E. of 0.84 +/- 0.07 in CLL cells and this was significantly less (p < 0.001) than that observed in normal B-cells (mean +/- S. E. of control, 0.12 +/- 0.02). In 3 of the 18 patients, TGF-beta 1 stimulated DNA synthesis. The reduced inhibition of leukemic cell DNA synthesis by TGF-beta 1 in CLL may provide these cells with a growth or survival advantage over normal lymphocytes and contribute to their selective accumulation.

Inhibition of repair of bleomycin-induced DNA strand breaks by 2'-deoxycoformycin and its effect on antitumor activity in L5178Y lymphoblasts.

We have observed previously that treatment of plateau-phase L5178Y murine lymphoblasts in vitro with 2'-deoxycoformycin plus deoxyadenosine (dCF/dAdo) can inhibit the repair of X-irradiation-induced DNA single-strand breaks (SSB) in these cells and that this effect is associated with synergistic cell kill. In this study we examined the effect of a combination treatment of plateau-phase L5178Y cells with bleomycin (BLM) plus dCF/dAdo. Incubation of BLM-treated cells with dCF/dAdo resulted in significant inhibition of the repair of BLM-induced DNA SSB. However, an additive, but not a synergistic, increase in cell kill was observed when cells were treated with a combination of BLM plus dCF/dAdo.

Induction of apoptosis in CD4+ prolymphocytic leukemia by deoxyadenosine and 2'-deoxycoformycin.

The leukemic cells of a patient with CD4+ prolymphocytic leukemia were treated in vitro with 5 microM deoxyadenosine and 60 microM 2'-deoxycoformycin (dCF), an inhibitor of adenosine deaminase (ADA). Following treatment, the leukemic cell dATP level increased to 378 pmol/10(6) cells on day 3, after which the level plateaued. Apoptosis was apparent following 4 h of incubation, and by day 8 34% of the chromatin was fragmented. Apoptosis also occurred in control cells, but to a lesser extent than in drug-treated cells. When the patient was treated with dCF, 4 mg/M2 i.v. the leukemic cell ADA activity was inhibited 24 h following treatment, and the lymphocyte dATP content increased to 303 pmol/10(6) cells by day 6. The lymphocyte count fell 60% in 1 week, but during this time there was no evidence of apoptosis in these cells. Thus, if dCF induces apoptosis in vivo, the effete cells may be rapidly cleared from the circulation and thus elude detection.