Search

We have examined the ability of 7-(2,3-epoxypropoxy)actinomycin D (EPA) to bind covalently to DNA and to 2'-deoxyribonucleoside 5'-monophosphates in a simple system in vitro. We have observed initially that EPA binds to DNA and deoxymono- and deoxydinucleotides with intercalative or stacking interactions that are characteristic of actinomycin D (AMD). When EPA is incubated (37 degrees C) for a prolonged period (pH 7.4, 6 h) in contact with either DNA or deoxyribonucleotides, it forms covalent adducts. Deoxyguanosine is always the preferred site of reaction by EPA. After enzymatic digestion of EPA-DNA adduct, three deoxyguanosine (EPA-dG) adducts, one major and two minor, were isolated. These adducts are separable from one another and from other deoxyribonucleoside adducts, e.g., EPA-dA and EPA-dC by reverse-phase HPLC. The authentic EPA-dG, EPA-dA, and EPA-dC adducts were synthesized by a chemical reaction of the epoxide in EPA with the deoxyribonucleotides followed by enzymatic dephosphorylation of the products. From the EPA-DNA adduct the EPA-dG adducts accounted for congruent to 2.2% of EPA employed; the remainder of EPA was completely hydrolyzed to an epoxide ring opened diol derivative, DHPA. DHPA binds to DNA by intercalation only and it does not form covalent adducts. Another model analogue of EPA (EPAMDEA) has the same epoxide-substituted chromophore but lacks the peptide lactone functions; it fails to associate with DNA and consequently it shows no covalent binding of its epoxide with DNA. Formation of a noncovalent intercalation complex between EPA and DNA appears to be a prerequisite for the covalent reaction. Presumably because of these dual interactions, EPA demonstrates superior antitumor activities both in human leukemic cells (CCRF-CEM) in vitro and P388 and L1210 cells in mice. The DNA base specific alkylating activity of EPA, which is derived from a combination of the actinomycin D (AMD) structure and the new epoxide function in the molecule of EPA, attributes to EPA a potentially novel pharmacological behavior that is not inherent of AMD.

The synthesis and biological activity of three 7-substituted actinomycin D derivatives are reported. Three such derivatives, 7-nitro-, 7-amino-, and 7-hydroxyactinomycin D, were synthesized via new methods which were first tested successfully with a chromophore model system. Of these, 7-nitro- and 7-aminoactinomycin D were assayed for growth inhibitory activity against mammalian cells (CCRF-CEM human lymphoblastic leukemia) in vitro and against the Ridgway osteogenic sarcoma and the L1210, P1534, and P388 murine leukemias in vivo. In these systems, the inhibitory activity of the 7-substituted analogs was comparable to actinomycin D. In two bacterial systems ( (L. casei and L. arabinosus) in vitro, on the other hand, these compounds showed inhibitory profiles which are distinctly different from actinomycin D. These studies demonstrate that substitution at the 7 position, which does not interfere with DNA binding, is capable of yielding experimental antitumor agents with significant activity against a variety of tumors.

7-(2,3-Epoxypropoxy)actinomycin D has been synthesized along with its major companion product, 7-(2,3-dihydroxypropoxy)actinomycin D. They were characterized by UV-visible and CD spectra and by NMR studies. According to UV-visible absorptiometry, circular dichroism, and thermal denaturation studies, they bind to DNA in a manner that is comparable to actinomycin D. The analogues are, like actinomycin D, extremely cytotoxic to human lymphoblastic leukemic cells (CCRF-CEM) in vitro but are significantly less toxic than actinomycin D to normal CDF1 mice is vivo. Unlike actinomycin, these analogues are metabolized in rats, and the metabolites are excreted in rat urine at a rapid rate. Compared to actinomycin D, the antitumor activity of the 7-(2,3-epoxypropoxy)actinomycin analogue against P-388 leukemia in mice is decidedly superior, and the therapeutic index is improved several fold.

The optical absorption and fluorescence characteristics of 7-animo-actinomycin D were determined to evaluate its potential as a fluorescent cytochemical probe. At pH 7.0, the absorption maximum and fluorescence excitation maximum are both at 503 nm; the fluorescence emission is at 675 nm. When this compound forms complexes with DNA in solution, the absorption and fluorescence excitation maxima shift to 543 nm and the fluorescence emission shifts to 655 nm. The fluorescence quantum yield is 0.016 for 7-amino-actinomycin D free in solution and 0.01-0.02 for complexes with native DNA. The 7-amino-actinomycin D also exhibits fluorescence shifts characteristic of binding when put into solution with poly(dG-dC) poly(dG-dC), but not with poly(dI-dC) poly(dI-dC). The spectral characteristics are the same at pH 7.0 whether the solvent is 0.01 M PO4 with 0.0001 M EDTA or Earle's salts with 0.025 M N-2-hydroxyethylpiperazine-N1-2-ethanesulfonic acid.

Two new classes of actinomycin D analogues, tetracyclic "reverse" analogues and a tricyclic "symmetrical" analogue of actinomycin D, are reported. These analogues bind to DNA and the binding does not occur by an intercalation mechanism. The analogues inhibit the synthesis of DNA and RNA in P388 tumor cells and the growth of CCRF-CEM cells in vitro at nanomolar concentrations. The tetracyclic "reverse" analogues, which are structurally related to the previously reported actinomycin D oxazolyl analogues, are metabolized in the presence of rat hepatic microsomes and tumor cell homogenates. The metabolism takes place with the loss of the oxazole ring; thus the "reverse" analogues produce a major metabolite which is the "symmetrical" analogue; the actinomycin oxazolyl analogues generate 7-hydroxyactinomycin D. Further, the microsomes activate the analogues to free-radical states which catalyze the production of superoxide as shown by stimulation of epinephrine oxidation and also indicated by electron paramagnetic resonance studies. The "symmetrical" and "reverse" analogues also demonstrate very high activities in these systems. In in vivo studies using P388/S, P388/ADR leukemia, and B16 melanoma in mice, the analogues showed increased activity and superior therapeutic index values, in comparison to actinomycin D.

The interaction of actinomycin D and three new 7-substituted analogs with calf thymus DNA has been studied by a number of physical techniques. The methods utilized in this investigation include visible absorption spectrometry and ultrafiltration methodology for the determination of equilibrium binding constants; viscometry; and circular dichroism. The studies show that the 7-substituted actinomycin D analogs retain the G . C base pair specific DNA binding demonstrated by actinomycin D. The mode of binding to native DNA, despite substitution at position 7, is practically unaltered. The retention of this binding specificity by these analogs seems to be unaffected by changes in the electon properties of the chromophore.

Enantiomeric forms of (+/-)-EPA [racemic 7-(2,3-epoxypropoxy)actinomycin D] have been synthesized; these are (R)-(+)- and (S)-(-)-EPA, which are active against a range of actinomycin resistant and marginally responsive tumors. The R-(+) enantiomer is uniformly superior to the other forms in all the tumor lines tested. These enantiomers act by binding to DNA, both by intercalation and alkylation at the guanine base of DNA. They are superior to actinomycin D in their in vitro activity against mouse leukemias (L1210 and P388/ADR) and mouse melanoma B16. This superior activity is also evident against all the preceding mouse leukemias and against solid tumors B16 and C26 in vivo. In biochemical action, the enantiomers behave similarly and act primarily by inhibiting DNA synthesis in tumor cells; the only difference found was in their preference for sites in DNA bases during alkylation. The R-(+) enantiomer generates an adduct that is believed to be bonded to the N7-site of guanosine; conversely, the S-(-) isomer forms two adducts with DNA that are different from the preceding one by HPLC and are tentatively assigned O6-guanosine-substituted structures on the basis of their UV, CD, and other chemical behaviors.

"Reverse" analogues (RAD's) of actinomycin D (AMD) and their antitumor activity against mouse and human colon tumor cells are reported. RAD's are tetracyclic, and they have an oxazole ring fused on the tricyclic phenoxazine chromophore of AMD. The oxazole ring in RAD is substituted at the C-2 carbon with either a CH3 (in RAD I), a C6H5 (in RAD II), or a CH2CONH(CH2)4NH2 (in RAD III) group. In tumor cells and rat hepatic microsomes, RAD's are metabolized to a tricyclic "symmetrical" analogue of AMD (SAD) with the loss of the oxazole ring and its substituents. RAD and SAD are very active in priming superoxides in the presence of microsomal enzymes as well as in inhibiting the synthesis of DNA and the growth of human colon tumor HT-29 cells in vitro. RAD III and SAD efficiently cleave closed circular plasmid pBR322 DNA like the antitumor agent bleomycin. In addition to their strong inhibitory activity against P388 and B16 tumors in vitro and in vivo, RAD III and SAD demonstrate high levels of activity against primary C26 and advanced C38 colon tumors in mice and against a xenograft of human colon adenocarcinoma CX-1 in athymic mice. In all these biological activities, the analogues demonstrate superiority to AMD in several experimental tumors. Also, the analogues, in contrast to AMD, show reduced toxicity in tumor-free mice, which is possibly due to the metabolic deactivation of SAD in host organs.

1,4-Oxazinone derivatives of the phenoxazinone chromophore in actinomycin D (AMD) have been synthesized by condensation of AMD with alpha-keto acids. By varying the starting alpha-keto acid, the substitutions on the oxazinone ring and, consequently, the lipophilicity of the molecule could be altered. These oxazinone derivatives revert to AMD in physiological media and it appears that these oxazinones are "depot" forms of AMD and possess physicochemical and DNA-binding properties which are significantly different from those of AMD. The oxazinones, which have bulky and lipophilic substituents at position 3, demonstrate more pronounced antitumor activity against P388 mouse leukemia and are less toxic than AMD.

N2-n-Alkyl- and omega-amino-n-alkylactinomycin D and 7-alkoxy-, 7-aralkoxy-, and 7-(acyloxy)actinomycin D were synthesized by modification of the parent actinomycin D molecule at the N2 and C-7 positions of the phenoxazinone moiety. The intermediate for N2 substitution was 2-deamino-2-chloroactinomycin D. For C-7 substitution, 7-hydroxyactinomycin D was used as the intermediate. Treatment of 2-deamino-2-chloroactinomycin D with an excess of the appropriate amine produced the N2-substituted derivatives. Condensation of the required alkyl or acyl halides with 7-hydroxyactinomycin D, aided by solid anhydrous potassium carbonate, yielded the C-7-substituted analogues. Calf thymus DNA-binding affinity was determined by equilibrium binding and also by thermal denaturation of DNA techniques, inhibitory activity of nucleic acid synthesis was examined using P388 cells in vitro, cytotoxicity measurements to tumor cells in vitro employed human lymphoblastic leukemic cells (CCRF-CEM), and antitumor activity was assayed against P388 mouse leukemia in CDF1 mice. Synthesis of a number of new analogues in each series and determination of the biophysical, biochemical, and biological properties established a more thorough structure-activity relationship in these analogues. These results establish that with the selection of omega-(n-alkylamino) groups at the N2 site or O-n-alkyl or O-acyl groups at the C-7 site a variety of modifications can be carried out on the actinomycin molecule while preserving biological activity. N2-3'-Amino-n-propyl- and N2-10'-amino-n-decylactinomycin D, 7-methoxy- and 7-ethoxyactinomycin D, and the 7-O-(1'-adamantoyl) ester of 7-hydroxyactinomycin D were found to be the most effective antitumor agents in vivo and in vitro. They also strongly inhibit cellular RNA and DNA synthesis and, with the exception of the ester, retain high DNA-binding affinity.

The growing importance of functionalized aziridines in numerous organic biomolecules led us to develop syntheses of novel actinomycin D (AMD) analogues substituted with an aziridine. Reaction of 7-hydroxyactinomycin D with 2-(iodomethyl)aziridine produced the desired 7-(2-aziridinylmethoxy)actinomycin analogue. In an attempt to develop an alternate route to this analogue, 7-(2-azido-3-iodopropoxy)actinomycin was subjected to reduction with dimethylamine-borane complex; the reaction did not produce the three-membered aziridine; instead the reaction product was found to be linear 7-(2-aminopropoxy)actinomycin D. Calf-thymus-DNA binding of these analogues was comparable to that of AMD as examined by UV-visible difference spectral measurements, thermal denaturation of DNA, and CD techniques. The analogues were found to be about 1/4 to 1/30 as cytotoxic to human lymphoblastic CCRF-CEM leukemia and B16 melanoma cells in vitro as AMD.

A series of N7-substituted 7-aminoactinomycin D analogues with alkyl, aralkyl, and heteroaralkyl substituents was synthesized and their biological properties were studied. All of these analogues proved to be 22- to 28-fold less toxic than actinomycin D when tested against human lymphoblastic leukemia cells (CCRF-CEM) in vitro. Against the P388 mouse leukemia in vivo, most of the analogues had activity comparable to actinomycin D and one was significantly more active. The results show that substitutions of this kind do not interfere with the antitumor activity of actinomycin D and may be useful for the design of modified actinomycin D analogues with greater selectivity.

The antitumour activity of actinomycin D (AMD) has been proposed to result, in part, from its intercalation into DNA dG-dC base-pairs leading to an inhibition of RNA synthesis. We have recently prepared 2-deamino-2-nitroactinomycin D and 2-deamino-actinomycin D and determined that, unlike AMD, these analogues do not intercalate into calf-thymus DNA. In the present study we show that these analogues and their corresponding peptide-free diethylamino derivatives are more effective than the parent AMD in forming ion radicals, in stimulating oxygen uptake and in forming superoxide anion when incubated in the presence of NADPH and NADPH cytochrome P-450 reductase. NaBH4-mediated reduction of these compounds yielded free radicals as shown by electron paramagnetic resonance (e.p.r.) spectroscopy. Free radicals could also be generated by incubation of these actinomycins with NADPH and either liver microsomes or purified NADPH cytochrome P-450 reductase. In the presence of molecular oxygen these free radicals spontaneously reoxidized by transfer of a single electron to molecular oxygen to form superoxide. Relative rates of superoxide formation were established for these substrates with the 2-deamino-2-nitroactinomycin D exhibiting the highest activity. It is proposed that the antitumour activity of these AMD analogues results, in part, from their ability to form reactive reduced oxygen species and, as such, these actinomycin derivatives may serve as useful probes for the tumouricidal mechanism of this family of agents.

Two different tetracyclic chromophoric analogues of actinomycin D have been synthesized by engaging two chromophoric DNA-binding functions in actinomycin D, i.e., 2-amino and 3-oxo, into either a 1,4-oxazin-2-one or an oxazole ring system. A third analogue has an extra quinone function at C-8 of the oxazole analogue. In all the analogues the chemical integrity of the peptide lactones of the parent antibiotic is kept intact, but their sterochemistry is altered. The analogues are designed as transport-modified prodrug forms of either the tricyclic active analogues of actinomycin D or actinomycin D itself. All analogues exhibit cytotoxicity that is several-fold less potent than AMD; they also have no binding affinity toward extracellular DNA. Nonetheless, the analogues of the first and the third series show improved antitumor activities (P388 leukemia, CDF1 mice). In fact, two of these analogues having a phenyl substituent at the C-3 site of the oxazinone ring or the C-2 position of the 8-oxo-8H-oxazole ring exhibit the highest antitumor effects. Most of the analogues are active over a broader dose range than actinomycin D and are 6- to 16-fold less cytotoxic to human lymphoblastic leukemia (CCFR-CEM) cells in vitro. The analogues with the most pronounced antitumor activity are those that retain most elements in the peptide stereochemistry of actinomycin D and have a quinone function or demonstrate susceptibility of their chromophores to biotransformation.

The growing importance of functionalized tricyclic rings, e.g., cyclopropyl and aziridine, in numerous organic biomolecules led us to develop syntheses of novel actinomycin D (AMD) analogues substituted with aziridine and cyclopropyl functions. Reaction of 7-hydroxyactinomycin D with 1-aziridineethyl iodide and bromomethylcycloporopane afforded the desired 7-[2-(1-aziridinyl)ethoxy] and cyclopropylmethoxy analogues, respectively. Calf thymus DNA binding of these analogues was comparable to that of AMD as examined by UV-vis difference spectral measurements, CD techniques, and relaxation of supercoiled closed circular SV40 DNA, indicating an intercalative mode of binding to the DNA duplex. Thermal denaturation of DNA experiments employing higher temperatures than room temperature exhibit a thermal lability of the DNA analogue complexes, suggestive of a probable covalent bond formation with DNA bases. The analogues were found to be 1/4-1/40 as cytotoxic to human lymphoblastic CCRF-CEM leukemia and B16 melanoma cells in vitro as AMD, with ID50 values in the nanomolar concentration range.

The DNA binding properties of several 7-substituted aralkylaminoactinomycin D analogues have been studied by spectrophotometry, DNA melting temperature studies, DNA-drug dissociation studies, and circular dichroism. Despite the presence of such bulky groups as 2-pyrrolylmethylamino or 3,4-dichlorobenzylamino at the 7 position, these analogues bind to DNA, inhibit RNA synthesis, and exhibit antitumor activity. A model is proposed for the interaction of the pyrrolyl analogue with phosphate groups of the DNA binding site, explaining the increased binding affinity for DNA of this actinomycin D analogue.