Marilyn J. Lemmon

SR-4233: A NEW BIOREDUCTIVE AGENT WITH HIGH SELECTIVE TOXICITY FOR HYPOXIC MAMMALIAN CELLS

We have examined the effects of the benzotriazine di-N-oxide SR-4233 (3-amino-1,2,4-benzotriazine-1,4 dioxide) on a variety of aerobic and hypoxic cells in culture, and on tumors in mice. The cell lines used were Chinese hamster ovary (HA-1), mouse 10T1/2, RIF-1, and SCC VII cells, and the human cell lines HCT-8, AG1522, and A549. The effect of SR 4233 in combination with irradiation was also examined in the SCC VII tumor growing in the flank of C3H mice using clonogenic assay (tumors excised 24 hr after irradiation). We found SR-4233 to be a potent and selective killer of hypoxic cells. Cell killing as a function of time for the various cell lines was exponential, with no shoulder. Drug concentrations producing equivalent levels of cell killing were 75-200 fold lower in hypoxic than in aerobic cells for the mouse and hamster lines, and 15-50 fold lower for the human cells. In vivo experiments showed that the non-toxic dose of 0.3 mmole/kg of SR-4233 enhanced radiation-induced tumor cell kill when the drug was given between 1 hr before and 2 hr after the radiation dose. We have also shown that the drug metabolizes more rapidly under hypoxic than aerobic conditions, both in vitro and in vivo. The toxic product(s) is unknown, but could be the 1-electron reduction product, the radical anion, because the mono N-oxide (the 2-electron reduction product) did not display cytotoxicity or selective killing under hypoxic conditions. This compound could therefore be a useful tool in tumor biology, as well as being a new lead in the development of bioreductive cytotoxic agents for cancer therapy.

Anaerobic bacteria as a gene delivery system that is controlled by the tumor microenvironment

A fundamental obstacle in gene therapy for cancer treatment is the specific delivery of an anticancer gene product to a solid tumor. Although several strategies exist to control gene expression once a vector is directly introduced into a tumor, as yet no systemic delivery system exists that specifically targets solid tumors. Nonpathogenic, obligate anaerobic bacteria of the genus Clostridium have been used experimentally as anticancer agents because of their selective growth in the hypoxic regions of solid tumors after systemic application. In this report we further describe a novel approach to cancer gene therapy in which genetically engineered clostridia are used as tumor-specific vectors for the delivery of antitumor genes. We have introduced into a strain of C. beijerinckii the gene for an E. coli nitroreductase known to activate the nontoxic prodrug CB 1954 to a toxic anticancer drug. Nitroreductase produced by these clostridia enhanced the killing of tumor cells in vitro by CB 1954, by a factor of 22. To demonstrate the specificity of this approach for tumor targeting, we intravenously injected the inactive spore form of C. beijerinckii, which upon transition to a reproductive state will express the E. coli nitroreductase gene. Nitroreductase activity was detectable in 10 of 10 tumors during the first 5 days after intravenous injection of inactive clostridial spores, indicating a rapid transition from spore to reproductive state. Tumors harboring clostridial spores which did not possess the E. coli nitroreductase gene were devoid of nitroreductase activity. Most importantly, E. coli nitroreductase protein was not found in a large survey of normal mouse tissues following intravenous injection of nitroreductase containing clostridia, strongly suggesting that obligate anaerobic bacteria such as clostridia can be utilized as highly specific gene delivery vectors for cancer therapy.

Enhancement of radiation-induced tumor cell killing by the hypoxic cell toxin SR 4233

SR 4233 (3-amino-1,2,4-benzotriazine 1,4-dioxide) is the lead compound in a series of benzotriazine di-N-oxides which exhibit high selective killing of hypoxic mammalian cells in vitro. Drug concentrations to produce equivalent levels of cell killing of SCC VII murine carcinoma cells under hypoxia were nearly 200-fold lower than under aerobic conditions. Following a one hour hypoxic incubation with drug, 20 microM SR 4233 killed 99.9% of SCC VII cells. The hypoxia-specific cytotoxicity of SR 4233 is due to bioreductive metabolism. For in vivo studies, pharmacokinetic measurements showed that drug concentrations well in excess of 20 microM were achievable in SCC VII tumors in mice for approximately one hour after a single injection of SR 4233. Under these conditions, cell killing was considerably enhanced in SCC VII tumors when SR 4233 was combined with a single X-ray dose of 20 Gy. The enhancement was seen whether SR 4233 was given for up to 2 h before or for up to an hour after radiation, and was comparable to the enhanced cell killing achievable with a single large dose of the radiosensitizer misonidazole. While this finding is consistent with the selective killing of at least some subset of hypoxic tumor cells by SR 4233, other interactions between the drug and radiation damage may contribute to the overall effect observed.

TUMOR HYPOXIA CAN BE EXPLOITED TO PREFERENTIALLY SENSITIZE TUMORS TO FRACTIONATED IRRADIATION

The present study describes a new way in which tumors may be made more sensitive to fractionated irradiation without affecting the sensitivity of surrounding normal tissues. It involves exploiting the cycling or intermittent hypoxia that occurs in at least some solid tumors, but not in normal tissues, using a new drug SR 4233, a benzotriazine di-N-oxide, which is rapidly metabolized in hypoxic cells to a product that kills these cells. Using this approach with a rodent tumor in a fractionated x-ray treatment regimen similar to that used in human radiotherapy, the addition of SR 4233 produced a large enhancement of the radiation response of the tumor with no change in the sensitivity of normal mouse skin. Under identical circumstances, there was no effect of the hypoxic cell radiosensitizer SR 2508, showing that SR 4233 with intermittent hypoxia was superior to a protocol which sensitized the hypoxic cells to doses of 2.5 Gy per fraction.

Mechanism of action of the selective tumor radiosensitizer nicotinamide.

Nicotinamide has been shown to selectively enhance the radiation damage of tumors in preference to normal tissues. Our present study was an investigation into the mechanism responsible for this effect in the SCCVII/St tumor model grown on the backs of C3H/km mice. A large single injection of nicotinamide (1000 mg/kg), given intraperitoneally 60 minutes before whole body irradiation, significantly enhanced the radiation response of SCCVII tumors as measured by an in vivo/in vitro excision assay performed 24 hr following irradiation. It also gave rise to an almost 4-fold reduction in the binding of 14C-misonidazole, injected 1 hr after the nicotinamide and measured by scintillation counting of excised tumor material 24 hr later. This suggested that nicotinamide was decreasing the degree of tumor hypoxia. Attempts were made to correlate these results with nicotinamide-induced changes in tumor blood flow using the techniques of 133Xe clearance, 86RbCl extraction and Hoechst 33342 fluorescent labelling. Nicotinamide produced between a 30-40% increase in mean tumor cell fluorescence of Hoechst 33342, which was consistent with an increase in tumor blood flow. A similar response was obtained using the uptake of 86RbCl as the end point. However, no statistically significant difference was seen between the tumor blood flow of control and nicotinamide treated mice using the 133Xe clearance procedure. These results are discussed with respect to their clinical implications.

STRUCTURE-ACTIVITY RELATIONSHIPS FOR BENZOTRIAZINE DI-N-OXIDES

SR 4233 (3-amino-1,2,4-benzotriazine 1,4-dioxide) is a bioreductive agent that selectively kills and radiosensitizes hypoxic mammalian cells in vitro and murine tumors in vivo. In an attempt to better understand the mechanism of action of the drug, and to determine whether a superior analog may exist, 15 benzotriazine-di-N-oxide analogs of SR 4233 have been evaluated to date for the following properties: hypoxic and aerobic toxicity toward CHO cells in vitro, drug-induced stimulation of oxygen consumption by incubation with respiration-inhibited cells, and acute LD50 evaluated in BALB/c mice. We noted several correlations between these biological properties of the drugs and some of their physicochemical characteristics. Both the hypoxic cytotoxicity and stimulation of oxygen consumption by respiration-inhibited cells were positively correlated with E1/2, the polarographic half-wave reduction potential, and a measure of electron affinity. The air-to-nitrogen differential cytotoxicity reached a maximum (corresponding to SR 4233) and then declined with increasing E1/2. The acute LD50 of each analog in mice decreased with increasing E1/2. One new compound, SR 4482, was found to be more toxic to hypoxic cells in vitro, but less toxic to mice, than SR 4233. It is similar in structure to SR 4233, but lacks any substituent in the 3-position of the triazine ring. This promising drug may represent a member of a new subseries of 1,2,4-benzotriazines with different structure-activity relationships.

Preferential tumor radiosensitization by analogs of nicotinamide and benzamide

The effects of a range of different analogs of nicotinamide and benzamide on the X ray response of the EMT-6 tumor in vivo was investigated. Using an in vivo/in vitro survival assay, sensitization was seen at a dose of 2 mmole/kg for all but one of the analogs tested. The enhancement ratios (ERs) ranged from 1.0 to 1.5. Of particular interest were nicotinamide and SR-4350 which gave large ERs (1.5 and 1.4 respectively) at doses which were only about 12% of the LD50 values. In one normal tissue studied (skin reaction) a large single dose of nicotinamide (8 mmole/kg) only gave an ER of 1.1. These results will be discussed with reference to the mechanisms involved and the clinical implications.

SR 4233: A tumor specific radiosensitizer active in fractionated radiation regimes

The benzotriazine SR 4233, in addition to preferential killing of hypoxic cells both in vitro and in vivo, also radiosensitizes aerobic cells in vitro if the cells are exposed to the drug under hypoxic conditions, either before or after irradiation. We have attempted to exploit this aerobic radiosensitization in vivo, by giving SR 4233 with the hypoxia inducing agent, hydralazine, after each radiation dose in a 8 x 2.5 Gy fractionated regime. The results show greater than additive cytotoxicity using both cell survival and regrowth delay as the endpoints of radiation response, but no radiosensitization in parallel groups treated with the hypoxic cell radiosensitizer SR 2508. The data are, therefore, consistent with radiosensitization of the tumor aerobic cells by the SR 4233 treatments. Significantly, the effect occurred with equal magnitude with or without hydralazine. Further, there was no radiosensitization to radiation induced leg contraction in the thighs of mice, a late responding normal tissue endpoint. The results, therefore, demonstrate a selective radiosensitization of tumors to a multifraction regime and suggest that SR 4233, or a close analog, may be useful in radiation therapy.

Pharmacological modification of tumor blood flow: Lack of correlation between alteration of mean arterial blood pressure and changes in tumor perfusion

The correlation between mean arterial blood pressure (MABP) and vascular perfusion in SCC-VII/St tumors in mice was compared following administration of three vasoactive drugs: flavone acetic acid (200 mg/kg), hydralazine (5 mg/kg), or nicotinamide (500, 750, and 1000 mg/kg). MABP was measured by the direct method in unanesthetized, unrestrained mice bearing a carotid catheter. Vascular perfusion of the tumor was measured using the 86RbCl extraction method. Body temperature was maintained at 36 degrees to 37 degrees C after drug administration when necessary. All three drugs reduced MABP from a control value of 125 +/- 2 (s.e.) mm Hg in mice without tumors. Flavone acetic acid at this dose had the least effect on blood pressure, with a minimum of 86% of control values at 10 to 20 min, and a return to control values by 1 hr. However, it produced a profound reduction in tumor perfusion that lasted more than 48 hr. Hydralazine and nicotinamide reduced blood pressure to minima between 55% and 69% of control values within 30 min, followed by a gradual return toward control values by about 8 hr. The reduction in tumor perfusion by hydralazine paralleled its effect on blood pressure. However, nicotinamide produced a transitory, although not statistically significant, increase in tumor perfusion at the highest dose given. These data demonstrate that tumor blood flow modification by drugs is not necessarily the result of changes in MABP, and blood pressure changes alone do not inevitably lead to changes in tumor perfusion.

Second-generation 1,2,4-benzotriazine 1,4-di-N-oxide bioreductive anti-tumor agents: Pharmacology and activity in vitro and in vivo

SR 4233 (1,2,4-benzotriazine-3-amine 1,4-dioxide) will soon be entering Phase I clinical trials as a new bioreductive cytotoxic agent for the treatment of solid tumors in combination with fractionated radiotherapy. We have selected 3 from over 50 analogues of SR 4233 which showed particular promise as second generation bioreductive antitumor agents. These compounds, when compared to SR 4233, have higher hypoxic toxicity and comparable or higher oxic to hypoxic cytotoxicity ratios in vitro and similar animal toxicity. We have compared the effectiveness of these three compounds with SR 4233 in two tumor systems and have examined some pharmacokinetic properties. The results show that replacement of the amino group at the 3-position of SR 4233 with either a hydrogen or an N,N-dialkylaminoalkylamino group shortens the half-life of these compounds in the blood because of the combined effects of partition coefficients, basicity, and higher reactivity. SR 4754 and SR 4755, the N,N-dialkylaminoalkylamino derivatives, exhibited shorter plasma half-lives than SR 4233 but exhibited lower anti-tumor activity than SR 4233 based on equal mouse toxicity in a fractionated regimen. SR 4482, with the hydrogen substitution and very high electron affinity, possessed a very short blood half life yet retained similar anti-tumor activity as SR 4233.