Antitumor effects of hydrogen peroxide in vivo

Abstract

Glucose oxidase, covalently coupled to polystyrene microspheres (GOL), produced H(2)0(2) at an average rate of 3.6 nmol/min per 10(9) beads under standard assay conditions. Injection of 1.3 × 10(10) to 1.1 × 10(11) GOL i.p. prolonged the survival of mice by 27 percent after injection of 10(6) P388 lymphoma cells in the same site, consistent with destruction of 97.6 percent of the tumor cells. Placing mice for several hours in 100 percent O(2), the probable rate-limiting substrate for GOL, afforded a 42 percent prolongation of survival from P388 lymphoma, consistent with destruction of 99.6 percent of the tumor cells. When the P388 inoculum was 10(5), 10(4), or 10(3) cells, GOL led to long-term survival (presumed cure) of 23 percent, 77 percent, and 92 percent of the mice, respectively, consistent with reduction of the injected tumor dose to less than 10 cells. Subcutaneous growth of 10(5) P388 cells (approximately 300 lethal dose to 50 percent of mice) was suppressed in 83 percent of mice by admixture of GOL with the tumor cell inoculum. GOL alone had no effect against a more peroxide-resistant tumor, P815 mastocytoma. However, P815 cell glutathione reductase could be inhibited in vivo by well-tolerated doses of the antitumor agent, 1,3-bis(2-chloroethyl)- 1-nitrosourea (BCNU). BCNU alone cured few mice with P815. Together, BCNU and GOL apparently cured 86 percent of mice injected with 10(6) P815 cells i.p. The protective effect of GOL was abolished by boiling it to inactivate the enzyme, by co-injection of catalase coupled to latex beads, or by delaying the injection of tumor cells for 3 h, by which time the beads had formed aggregates. Soluble glucose oxidase, in doses threefold higher than that bound to GOL, had no detectable antitumor effect. A single injection of preformed H(2)0(2) readily killed P388 cells in the peritoneal cavity, but only at doses nearly lethal to the mice. In contrast, GOL had very little toxicity, as judged by the normal appearance of the mice for over 400 d, gross and microscopic findings at autopsy, and various blood tests. GOL injected i.p. remained in the peritoneal cavity, where it was gradually organized into granulomata by macrophages, without generalized inflammation. Thus, an H(2)0(2)-generating system confined to the tumor bed exerted clear- cut antitumor effects with little toxicity to the host.