Intraperitoneal administration of Corynebacterium parvum in patients with ascitic ovarian tumors resistant to chemotherapy: effects on cytotoxicity of tumor-associated macrophages and NK cells

Eight ovarian cancer patients with ascites resistant to various chemotherapy protocols were treated i.p. with C. parvum in an effort to obtain a palliative reduction of peritoneal effusions. C. parvum(7–14 mg) was given i.p. on days 0, 7, 28 and then at monthly intervals. Three patients showed no clinical evidence of therapeutic benefit from C. parvum administration. Three subjects had a complete disappearance of ascites, and two additional patients showed a marked reduction of their effusions. Solid tumor masses were not affected by i.p. C. parvum. C. parvum injection caused a rapid decrease of the percentage of tumor cells in the ascites as early as 7–15 days after the first treatment. Concomitantly, the percentage of polymorphs and, to a lesser extent, of macrophages in the ascitic tumor increased, whereas tumor-associated lymphoid cells (TAL) were not altered in a consistent pattern. The tumoricidal activity of peripheral blood monocytes and natural killer (NK) cells was not consistently modified after C. parvum treatment. Pretreatment NK activity of TAL was usually very low and was not appreciably modified by i.p. C. parvum. In one patient, whose TAL had NK activity in the range of normal PBL, C. parvum administration was followed by a profound depression of intratumor NK cytotoxicity. The tumoricidal activity of tumor-associated macrophages (TAM) was not enhanced following C. parvum administration; 14–21 days after the first C. parvum inoculation, a marked decrease of the cytolytic capacity of TAM was observed, cytotoxicity returning to pretreatment values thereafter. These observations do not elucidate the mechanism of the limited, but significant, anti-tumor activity of i.p. C. parvum in human ovarian carcinomatous ascites. The lack of enhancement of macrophage cytotoxicity after i.p. C. parvum, predicted on the basis of data in rodents, cautions against the empirical use in humans of agents assumed to affect human neoplasia through activation of host defense mechanisms.