Five-chlorodeoxycytidine, a tumor-selective enzyme-driven radiosensitizer,effectively controls five advanced human tumors in nude mice

Purpose: The study's goals were as follows: (1) to extend our past findings with rodent tumors to human tumors in nude mice, (2) to determine if the d rug protocol could be simplified so that only CldC and one modulator, tetra hydrouridine (H4U), would be sufficient to obtain efficacy, (3) to determin e the levels of deoxycytidine kinase and dCMP deaminase in human tumors, co mpared to adjacent normal tissue, and (4) to determine the effect of CldC o n normal tissue radiation damage to the cervical spinal cord of nude mice. Methods and Materials: The five human tumors used were as follows: prostate tumors, PC-3 and H-1579; glioblastoma, SF-295; breast tumor, GI-101; and l ung tumor, H-165. The duration of treatment was 3-5 weeks, with drugs admin istered on Days 1-4 and radiation on Days 3-5 of each week. The biomodulato rs of CldC were N-(Phosphonacetyl)-L-aspartate (PALA), an inhibitor of aspa rtyl transcarbamoylase, 5-fluorodeoxycytidine (FdC), resulting in tumor-dir ected inhibition of thymidylate synthetase, and HU, an inhibitor of cytidin e deaminase. The total dose of focused irradiation of the tumors was usuall y 45 Gy in 12 fractions. Results: Marked radiosensitization was obtained with CldC and the three mod ulators. The average days in tumor regrowth delay for X-ray compared to dru gs plus X-ray, respectively, were: PC-3 prostate, 42-97; H-1579 prostate, 2 9-115; glioblastoma, 5-51; breast, 50-80; lung, 32-123. Comparative studies with PC-3 and H-1579 using CldC. coadministered with H4U, showed that both PALA and FdC are dispensable, and the protocol can be simplified with equa l and possibly heightened efficacy. For example, PC-3 with X-ray and (1) no drugs, (2) CldC plus the three modulators, (3) a high dose of CldC, and (4 ) escalating doses of CldC resulted in 0/10, 3/9, 5/10, and 6/9 cures, resp ectively. The tumor regrowth delay data followed a similar pattern. After t reating mice only 1(1)/(2) weeks with CldC + H4U, 92% of the PC-3 tumor cel ls were found to possess CldU in their DNA. The great majority of head-and- neck tumors from patient material had markedly higher levels of dC kinase a nd dCMP deaminase than found in adjacent normal tissue. Physiologic and his tologic studies showed that CldC + H,U combined with X-ray, focused on the cervical spinal cord, did not result in damage to that tissue. Conclusions: 5-CldC. coadministered with only H4U is an effective radiosens itizer of human tumors. Ninety-two percent of PC-3 tumor cells have been sh own to take up ClUra derived from CldC in their DNA after only 11/2 weeks a nd 2 weeks of bolus i.p. injections. Enzymatic alterations that make tumors successful have been exploited for a therapeutic advantage. The great elec tronegativity, coupled with the relatively small Van der Waal radius of the Cl atom, may result in CldC's possessing the dual advantageous properties of FdC on one hand and BrdU and IdU on the other hand. These advantages inc lude autoenhancing the incorporation of CldUTP into DNA by not only overrun ning but also inhibiting the formation of competing TTP pools in tumors. A clinical trial is about to begin, with head-and-neck tumors as a first targ et of CldC radiosensitization. (C) 2001 Elsevier Science Inc.