The search for novel anticancer agents: a differentiation-based assay and analysis of a folklore product.

One alternative approach to the current use of cytotoxic anticancer drugs involves the use of differentiation-inducing agents. However, a wider application of this strategy would require the development of assays to search for new differentiation-inducing agents. In this report we describe an in vitro assay using the murine erythroleukemia (clone 3-1) cells. Tests for the efficacy of this assay for the analysis of antineoplastic activity in natural products led to studies on pau d'arco, a South American folklore product used in the treatment of cancer. Purification of the activity in aqueous extracts by solvent partition and thin layer chromatography (TLC) indicated the presence of two activities, one of which was identified as lapachol. The activity in the pau d'arco extracts and of lapachol was inhibited by vitamin K1. As a vitamin K antagonist, lapachol might target such vitamin K-dependent reactions as the activation of a ligand for the Axl receptor tyrosine kinase.

An endogenous signal triggering erythroid differentiation: identification as thyroid hormone.

The identification of thyroid hormone as an endogenous signal for erythroid differentiation began with our studies on the spontaneously differentiating murine erythroleukemia clone 3-1. We observed that the spontaneous differentiation frequency was dependent on a heat stable factor present in fetal calf serum or calf bone marrow. We also noted that the bone marrow extract stimulated erythroid colony-forming units in mouse bone marrow cells, suggesting the relevance of this factor in normal erythroid differentiation. The bone marrow extract did not supplant the requirement of erythropoietin but was synergistic. Purification of the bone marrow extract indicated that the differentiation-inducing activity for clone 3-1 cells cochromatographed with a low-molecular-weight, UV (280 nm)-absorbing component(s). These observations and previous reports identifying the avian erythroblastosis virus oncogene v-erbA as a mutated thyroid hormone receptor which blocked erythroid differentiation led us to test thyroid hormone in our assay. Both triiodothyronine and thyroxine were highly active, and the active constituents in the chromatographically purified fraction were identified as triiodothyronine and thyroxine. Although thyroid hormone action has been associated with both in vivo and in vitro erythroid differentiation, its role has been often relegated to a secondary status. We suggest that thyroid hormone is required for the commitment of erythroid cells to terminal differentiation.

An anticancer drug-sensitive murine erythroleukemia clone: implications for the mechanism of action of antineoplastic drugs.

The mechanism(s) by which anticancer drugs kill tumor cells remains obscure. The studies reported here were undertaken with the view that the mechanism may be understood in part through an analysis of anticancer drug-sensitive clones. We have isolated a murine (Friend) erythroleukemia clone in which drug sensitivity was correlated with increased differentiation, suggesting that anticancer drug-induced cell death may be based on differentiation or a differentiation-dependent mechanism. In addition, this clone showed a high propensity for constitutive differentiation and frequent appearance of large multinucleate cells. Morphologically similar large aberrant cells were observed after the treatment of parental (745A) cells with Adriamycin (or bleomycin). We attribute these morphological defects occurring in clone 3-1 or in the parental cell line after anticancer drug treatment to a defective or inhibited cell cycle function. We suggest that the putative cell cycle defect in clone 3-1 is coupled to the increased drug-induced differentiation and resulting cell death. From a broader perspective, the studies reported here suggest that the search for and design of new anticancer drugs be directed at agents that modulate differentiation functions.

The linking of anticancer drugs, cell cycle blocks, and differentiation: implications in the search for antineoplastic drugs.

The quest for anticancer drugs has been primarily directed at agents that interfere with cell replication, yet the basis for drug-induced cytotoxicity remains unsolved. In our previous studies we noted a relationship between a mitotic block and commitment to terminal differentiation in the murine (Friend) erythroleukemia (FEL) cell. Since anticancer drugs are known to often block cell cycle transit typically in G2/mitosis, we tested a number of anticancer drugs with various modes of action and found that they all committed FEL cells to differentiate. Furthermore, other G2/mitosis-blocking drugs were also effective in inducing commitment. These results suggest (1) a causal relationship involving anticancer drugs, cell cycle block and differentiation, (2) that the search for new anticancer drugs utilize a differentiation assay and include G2/mitosis-blocking agents.

Temporal mapping of the differentiation pathway of the murine erythroleukemia cell.

These studies are concerned with "mapping" the temporal order of the precommitment events in the differentiation pathway of the Friend erythroleukemia cell. We have used a single-block procedure in which a differentiation-specific inhibitor of a temperature-sensitive (ts) differentiation-defective mutation was used to block the differentiation program. Later, the block was removed, differentiation was allowed to proceed, and the time required to reach a reference marker was monitored. These studies have indicated that the mutations tsC2GP1 and tsB5A, the poly(ADP-ribose) polymerase inhibitor 3-aminobenzamide, and the glucocorticoid hormone dexamethasone blocked functions which are required just prior to commitment. We have also used a double-block procedure involving two consecutive restrictive conditions, which suggests that the 3-aminobenzamide- and tsC2GP1-blocked functions constitute a part of a sequentially ordered pathway leading to terminal differentiation. The convergence of the 3-aminobenzamide, dexamethasone, and ts mutational blocks just prior to commitment suggests that the blocked functions may be part of a major control mechanism for commitment. In these studies, we have introduced a cytochalasin B-based assay to monitor commitment. The use of cytochalasin B permits a direct assay for commitment and obviates the need for colony-forming assays using semisolid medium, which have inherent problems such as efficiency of plating.

In vivo benzo[a]pyrene diol epoxide-induced alkali-labile sites are not apurinic sites.

We have used endonuclease IV from Escherichia coli as a probe for apurinic sites in the DNA of HeLa cells following treatment with an activated diol epoxide derivative of benzo[a]pyrene. DNA strand breaks and alkali-labile sites were observed that were repaired following exposure to the carcinogenic alkylating agent. The alkali-labile sites were not substrates for the apurinic site-specific endonuclease IV. We conclude that the alkali-labile sites formed in vivo by benzo[a]pyrene derivatives are not apurinic sites and probably arise as a consequence of rearrangement of the abundant N2-guanine adducts. This finding questions the involvement of apurinic sites in the mutagenic activity of benzo[a]pyrene.

Mitosis may be an obligatory route to terminal differentiation in the Friend erythroleukemia cell.

In previous studies, it was shown that treatment of Friend erythroleukemia (FEL) cells with dimethylsulfoxide (DMSO) and the poly(ADP-ribose) polymerase inhibitor 3-aminobenzamide (3AB) blocked the differentiation pathway just prior to commitment. These studies show that the exposure of DMSO(+3AB)-induced cells to the mitotic inhibitors colcemid or nocodazole resulted in commitment to terminal differentiation. Expression of differentiated phenotype required further incubation without the mitotic inhibitors. Microscopic examination indicated that the number of cells blocked in mitosis and those that differentiated were approximately equivalent. These observations suggest that commitment had occurred during mitosis and that expression of the differentiated state occurred after completion of mitosis. Since commitment was not inhibited by blocking DNA replication by aphidicolin or cytokinesis by cytochalasin B, mitosis may be the only phase of the cell cycle required for commitment.

O6-methylguanine (O6-MeG) and cytotoxicity: reversion analysis involving an N-methyl-N'-nitro-N-nitrosoguanidine-sensitive, O6-MeG-DNA methyltransferase-deficient HeLa cell mutant.

In a previous communication, we proposed that N-methyl-N'-nitro-N-nitrosoguanidine (MNNG)-induced cytotoxicity in an O6-methylguanine (O6-MeG)-DNA methyltransferase-deficient (MT-) HeLa cell mutant was mainly the consequence of DNA strand breaks resulting from the failure to remove O6-MeG lesions. If MNNG-induced cytotoxicity, DNA strand breaks and O6-MeG lesions are related, there should be a corresponding relationship of these properties in MNNG-resistant clones derived from the MT- strain. A study of such revertants indicated that they were a heterogeneous group with increased repair of DNA strand breaks and O6-MeG lesions and increased resistance to the cytotoxic effects of MNNG. These observations support the hypothesis relating O6-MeG, DNA strand breaks and cytotoxicity. The relationship of these 'revertants' to the MT- and wild-type strains is discussed.

p66 an in vivo target for poly(ADP-ribosyl)ation co-purifies with poly(ADP-ribose) polymerase.

A direct immunoassay has been applied for the quantitation of poly(ADP-ribose) polymerase and its automodification in BALB/3T3 (A31) cells. As the cell population reached a high density, a 66 kDa protein (designated p66) which co-purified with the enzyme became highly poly(ADP-ribosyl)ated. Pulse-chase experiments as well as a Western blot analysis indicated that the p66 was not a degradation product of poly(ADP-ribose) polymerase.

O6-methylguanine-DNA methyltransferase-defective human cell mutant: O6-methylguanine, DNA strand breaks and cytotoxicity.

We have isolated an isogenic O6-methylguanine (O6-MeG)-DNA methyltransferase-defective mutant from a HeLa cell line. This mutant exhibits excess DNA strand breaks and considerable cytotoxicity after N-methyl-N'-nitro-N-nitrosoguanidine (MNNG) treatment. The increased frequency of strand breaks after MNNG treatment was not abolished by DNA synthesis inhibitors. We propose that the presence of unrepaired O6-MeG lesions leads to excess strand breaks and these, in turn, are mainly responsible for the cytotoxicity.

Base excision repair of DNA in gamma-irradiated human cells.

Escherichia coli endonuclease IV was used to incise cellular DNA specifically at apurinic/apyrimidinic (AP) sites prior to alkaline elution to measure the resulting DNA strand breaks. gamma-Irradiated HeLa cells initially contained DNA strand breaks and no AP sites. Upon incubation at 37 degrees C the strand breaks were rapidly repaired and AP sites were generated and subsequently repaired. The transient nature of the AP sites indicates the in vivo operation of a base excision repair pathway whereby damaged bases are removed from DNA by DNA glycosylases to produce AP intermediates that are then substrates for AP endonucleases.

PolyADP ribosylation and Friend erythroleukemic-cell differentiation: action of poly(ADP-ribose) polymerase inhibitors.

Previous studies have shown that benzamide and nicotinamide, two inhibitors of poly(ADP-ribose) polymerase, induce erythroid differentiation in Friend erythroleukemic cells. In contrast, we observed that two other commonly used inhibitors, i.e., 3-aminobenzamide and 3-methoxybenzamide, not merely failed to induce differentiation but actually inhibited it. Furthermore, we observed that benzamide at high concentrations induced differentiation, whereas at low concentrations, it inhibited differentiation. We propose that the induction occurring at high concentrations is due to the polar-planar properties of the molecule, while the inhibition at low concentrations might be due to the inhibition of poly(ADP-ribose) polymerase. Cells that were first exposed to an inducer (hexamethylene bisacetamide) and an inhibitor (3-aminobenzamide) and were subsequently incubated in medium lacking these substances did not differentiate. However, re-exposure to the inducer resulted in a very rapid increase in commitment to terminal differentiation. Therefore, 3-aminobenzamide appears to block differentiation prior to the commitment stage. We infer from these findings that polyADP-ribosylation is required for the terminal differentiation of Friend erythroleukemia cells.

DNA repair pathway in alkylated human cells: apurinic/apyrimidinic intermediate resolved by Escherichia coli endonuclease IV-coupled alkaline elution.

Apurinic/apyrimidinic (AP) sites were measured in HeLa cells by digestion of cellular DNA with Escherichia coli endonuclease IV, an AP-specific endonuclease, prior to alkaline elution. The absence of non-specific endonuclease activity allowed endonuclease IV-sensitive AP sites to be detected with the sensitivity of conventional alkaline elution. Cells that were alkylated with dimethyl sulfate contained AP sites that were repaired along with DNA single-strand breaks during a post-alkylation recovery period. These results show that DNA alkylation products are repaired, at least in part, via an AP intermediate suggesting a repair pathway initiated by DNA glycosylases followed by DNA incision by AP endonuclease.

Inhibition of poly(ADP-ribose)polymerase causes increased DNA strand breaks without decreasing strand rejoining in alkylated HeLa cells.

Treatment of alkylated HeLa cells with 3-aminobenzamide, an inhibitor of poly(ADP-ribose) polymerase, increased the number of DNA strand breaks but did not affect the rate of strand rejoining. This suggests that an increase in DNA incision, not a decrease in ligation, results from the inhibition of poly(ADP-ribose) polymerase in cells recovering from DNA damaged by alkylating agents.

Superhelical DNA-dependent ATPase from calf thymus.

Two physically and catalytically distinct DNA-dependent ATPases were isolated from a purified preparation of calf thymus poly(ADP-ribose) polymerase. A unique feature of these two ATPases was the high stimulation by supercoiled DNA. Other nucleic acids (including denatured DNA and ribosomal RNA) and certain polynucleotides differentially stimulated the two enzymes. We have not detected any other DNA-related activity associated with these ATPases.

Genetic analysis of dar, uvsW, and uvsY in bacteriophage T4: dar and uvsW are alleles.

A genetic study of the T4 dar (DNA arrested synthesis restoration) mutations was performed by two- and three-factor crosses. The dar mutations restore the viability of gene 59 mutants. Mapping studies of the dar mutations indicated that the dar gene extended over 16 map units. This high recombination frequency is attributed to an increased level of recombination in the dar region. Two other mutations, uvsY and uvsW, known to be located in the vicinity of dar, were studied. These studies indicated that the uvsY and dar mutations were located in separate genes but that dar and uvsW were allelic. The genes are ordered as follows: gene 24, dar(uvsW), uvsY, and gene 25 in clockwise order.

An analysis of DNA repair and recombination functions of bacteriophage T4 by means of suppressors: the role of das.

Previous studies have indicated that the bacteriophage T4 das mutations partially suppressed the DNA replication defects in gene 46 and 47 mutations. Here it is shown that the das mutation also suppresses the DNA repair defects but not the DNA replication defects of the uvsX and uvsY mutations. In contrast, the das mutation suppressed both the DNA replication and repair defects of gene 46 and 47 mutations. These characteristics of das as well as those of the other suppressors, including uvsW(dar) and two new suppressors sur and uvsU have been used for the analysis of the DNA repair pathway. Based on the functions of these suppressors, a sequence in which the gene products in this pathway might act is suggested.

A new suppressor of mutations in the DNA repair-recombination genes of bacteriophage T4: sur.

A new mutation designated sur was isolated as a suppressor of a mutation in the uvsX gene of T4 phage. Unlike the other suppressors of mutations in genes involved in DNA repair and recombination, sur has a wide range, suppressing both DNA repair and replication defects in mutations in genes uvsX, uvsY, 46, 47, and 59. However, its suppressor functions may be confined to the uvsX-uvsY DNA repair pathway since sur did not suppress a mutation in the denV gene. The sur mutation results in an increased degradation of host DNA to an acid-soluble form, but this increase was blocked by a mutation in gene 46 (nuclease) indicating that the sur function is involved in an earlier step in the degradation of host DNA. This increased degradation of host DNA might be a reflection of a compensatory increase in an alternate DNA repair activity in the [sur] mutant.

The coupling of DNA repair-recombination functions with DNA replication in bacteriophage T4: a new DNA repair mutant.

The requirement of DNA repair-recombination functions for T4 phage DNA replication has been known for some time but the underlying basis for this relationship has been unclear. This report is concerned with a new uv-sensitive gene [uvsU], whose function appears to bridge these two major activities of DNA. The [uvsU] mutant fails to complement [uvsX] mutants but uvsU maps in a region distinct from uvsX. Furthermore, the uvsU mutation specifically suppressed the DNA replication defect but not the uv sensitivity of the uvsX mutation. The previously discovered uvsW gene, whose mutations suppress the DNA replication defects of gene 59, 46, and 47 mutations, seems to have an analogous role. As a possible explanation for these observations, it is suggested that the uvsW and uvsU gene products (gps) couple the DNA repair-recombination and replication functions by controlling the entry of DNA intermediates from the replication pool into the DNA repair-recombination pathway. Furthermore the suppression data are interpreted to suggest that the gps uvsW, 59, 46, and 47 function together. Similarly the gps uvsU and uvsX may form a functional unit.

Poly(ADP-ribose) polymerase is a zinc metalloenzyme.

Purified poly(ADP-ribose) polymerase was inhibited by 1,10-phenanthroline at pH less than 8. This inhibition and the inhibition by other chelating agents suggested that this enzyme was a metalloprotein. Atomic absorption spectroscopy showed the presence of one atom of zinc per protein molecule. Dialysis of the enzyme against buffers containing 1,10-phenanthroline resulted in the loss of activity and the removal of zinc from the enzyme. Initial rate kinetics showed that 1,10-phenanthroline was non-competitive with NAD+ and competitive with DNA. The binding of DNA to the enzyme was unaffected by the inhibitor. These results suggest that a metal-containing site is involved as part of the interaction of DNA and poly(ADP-ribose) polymerase.