Novel Cytotoxic Oxopyridoindolizines: iso-Propyl-7,8,9-tri-chloro-6,7,8,9-tetrahydro-5-oxopyrido[2,3-a]-indolizine-10-carboxylates (OPIC)

A series of eight new alkyl-7,8,9-trichloro-6,7,8,9-tetrahydro-5-oxopyrido[2,3-a]-indolizine-10-carboxylates (OPIC), analogues of camptothecin (CPT), were prepared in a one-pot reaction of 2,2'-bipyridine-3,3'-dicarboxylic acid (BPA) with a mixture of thionyl chloride/chlorine, followed by addition of the appropriate alcohol. This led to a mixture of OPIC compounds 3a-d, 4a-d and 3,3'-dialkoxycarbonyl-2,2'-bipyridines (BPE, 2a-d). The isopropyl OPIC 3c and its corresponding diastereoisomer 4c showed marked activity against three cancer cell lines compared to other analogs. These same diastereoisomers also displayed high cytotoxic activity against five leukemia cell lines, thus the presence of an isopropyl substituent on the carboxylic ester, as opposed to other alkyl substituents, appears to play a key role in the cytotoxic potency of this new class of compounds.


Introduction
The natural alkaloid camptothecin (CPT, Scheme 1), extracted from Camptotheca acuminata by Wall and Wani [1], has demonstrated efficacy in the treatment of several cancers. However, interest in its application as an anti-tumour agent has declined due to its toxic secondary effects. The cellular target of the alkaloid has been shown to be topoisomerase I [2], an enzyme essential for religation of DNA during a number of critical cellular processes, including replication, transcription and repair [3]. Camptothecin, topoisomerase I and DNA form a so-called "cleavable complex" [4] that results in topoisomerase I-mediated DNA breaks by preventing DNA religation. These results have prompted the synthesis of a variety of derivatives and analogues of camptothecin. Topotecan (Hycamtin) and irinotecan (Camptosar) are two of the leading examples [5][6][7]. Much effort, including that leading to the two drugs above cited, has been devoted towards increasing the water solubility of the camptothecin analogues in order to obtain compounds with an improved pharmacological profile and enhanced efficacy against human tumors [8][9][10]. From structure-activity relationship studies [11] it appears that the E-ring lactone and the natural 20S-configuration are essential for anti-tumour activity. While the activity of compounds with substitutions in rings C and D is critically dependent on the size and type of the substituents [12], most structural modifications have concerned rings A and B where wide possibilities of variation exist, especially at positions 7, 9, 10 and 11. Recently the structure of topoisomerase I covalent and noncovalent complexes with a 22-base pair DNA duplex has been solved by X-ray analysis. Based on the crystal structure and structure-activity relationships, a mode of binding for camptothecin has been proposed [13]. In this and in the analogous compound [14], there is plenty of room for substitution at position 7 without steric interference.
A number of camptothecin derivatives have been developed and tested against various cancer cell lines in the past few years and have indicated the importance of lipophilic groups at position 7 of camptothecin for potent cytotoxic activity. The low solubility in water of these compounds does not amount to a grave disadvantage due to the possibility of successful administration per os of camptothecin derivatives [15,16].
We report the synthesis and in vitro anti-tumour activity of a new series of OPIC compounds (Scheme 1) that are readily separated into their pure diastereoisomers (3a-d and 4a-d). These new OPIC compounds, deriving from a pyridyl-oxo cyclisation, can be obtained through a simpler and more economical synthetic method than CPT. The starting material, 2,2'-bipyridine-3,3'-dicarboxylic acid (1), was obtained in good yield from commercial 1,10-phenanthroline by oxidation using the KMnO 4 procedure described by Ben-Hadda et al [17]. The reaction of 1 with thionyl chloride under an inert atmosphere provides an unstable chloro-addition intermediate, the 3,3'-dichlorocarbonyl-2,2'bipyridine, which is converted into the 3,3'-dialkoxycorbonyl-2,2'-bipyridine (BPE) upon addition of the desired alcohol. The same method for the synthesis of BPE could be slightly modified and usefully exploited to prepare the new family of OPIC compounds.

Results and Discussion
Chemistry Scheme 2 depicts the general reaction. In our first attempt, oxopyridoindolizine-carboxylates (OPIC 3, 4) were prepared from BPA (1) and freshly distilled thionyl chloride. After gentle reflux for 5-24 h, followed by subsequent addition of the alcohol, the title compounds were isolated in very poor yield (2-5%). The majority of products (62-93%) were the 3,3'-dialkoxycarbonyl-2,2'-bipyridines (BPE). 2a: The use of technical grade thionyl chloride (which contains small amounts of chlorine), under the same conditions as above led to higher, but still modest, yields of the OPIC compounds: 3-7% for 3a-d and 12-31% for 4a-d. The thionyl chloride contained a small amount of gaseous chlorine, most likely produced in situ by thermo and/or photodecomposition of the reagent. This observation led us to believe that the deliberate addition of chlorine would greatly improve the yield of OPIC. We found that in addition to improving the yield of OPIC, the addition of gaseous chlorine to the reaction decreased the reaction time from 5h to 2h. The optimisation and mechanism of the formation of OPIC has been reported recently in more detail [20].

Evaluation of in vitro Anti-tumour Activity
A series of substituted OPIC (3 and 4) and BPE (2) differing in the substituents on the carboxylate groups were selected by the National Cancer Institute for evaluation of their in vitro anticancer activity (Table 1). From the data in Table 1 it appears that only the oxopyridoindolizines 3c and 4c show any marked cytotoxic activity compared to the rest of series (3 and 4). In this case, the degree of lipophilicity of the carboxylate substituent does not correlate positively with the cytotoxic activity. A direct influence of the lipophilic group from the pyridyl nucleus appears to be more important for activity. Compounds 3c and 4c were further tested for their in vitro activity against 60 cell lines (Tables 2-4).

Conclusions
Chlorination of BPA with thionyl chloride in the presence of gaseous chlorine has been demonstrated to be a powerful and optimal method for pyridyl-oxo cyclisation in the preparation of a new heterocyclic family; the oxopyridoindolizine carboxylates (OPIC). This work provides for the first time a simple one-pot synthetic methodology for the preparation of a wide range of OPIC compounds which are analogues or hybrids of the natural alkaloid Camptothecin (CPT). The cytotoxic activity of this new family suggests a promising novel approach to the design of prospective compounds for the treatment of cancer.
As a guide for future work, the data reported here indicate that the OPIC compounds have a definite potential efficacy that merits development through modification to both the pyridyl and the chlorinated rings. We note, however, that although the compounds are nominally camptothecin analogues there is no necessary presumption that they do in fact resemble camptothecin and thereby target topoisomerase I. It might be possible to address this question by testing their effect(s) on two cell lines, one of which has normal topoisomerase I and the other has a mutant (camptothecin-resistant) enzyme. If there were a difference in GI 50 value this would indicate that topoisomerase I is a critical target for the drugs. Given the rather low GI 50 values of the present compounds, around 200 µM, it would be prudent first to seek new derivatives having GI 50 values around 1 micromolar or less so that the assay can be conducted more efficiently, though the interest in compounds 3c and 4c because of their particular stereochemistry remains undiminished. A further consideration relates to the poor solubility of the chloro derivatives in water; for the cell line assays to work it is important to prepare derivatives that are more soluble. Substituting the chloro atoms with some other groups such as alkyloxy or amino is possible and might produce the desired effect.

Acknowledgements
This work was supported by grants from the Ministry of Education of Morocco (PROTARS No. P1T2/27), Cancer Research U.K., France (AI N° 98/ 160/ SM), and Willaya of Oujda. We wish to thank Professor Bruce Baguley, University of Auckland School of Medicine, New Zealand, for helpful discussions. We would also like to thank Advanced Chemistry Development, Inc., Toronto, Canada for the free on-line nomenclature of organic compounds. We are indebted to TAACF and NCI of USA for biological tests.

General
Melting points are determined by using a Buchi 510 apparatus and are uncorrected. 1 H-and 13 C-NMR spectra were recorded on a Bruker WP-80 operating at 200.131 MHz and AC 250 at 250.14 MHz or an AM 300 at 300.134 MHz spectrometer. The numbering used in the assignment of the NMR spectra is shown in Scheme 4. Mass spectra were recorded on a Platform II Micro Mass spectrometer, and FTIR spectra on a Nicolet 205 instrument.