Inflammation is ultimately a reasonable and beneficial biological reaction. If a sequence of inflammation, hemorrhage, and necrosis occurs at a carcinoma site, it is reasonable to assume that the reaction leads to the necrosis of carcinoma cells, even if it is a nonspecific reaction that does not specifically attack the metabolic system or the antigens of carcinoma cells. The Shwartzman reaction is characterized by a strong necrotic reaction and a subsequent effective repair process and is thus suitable for this purpose.
Many investigators, including Shwartzman 1), Boivin 2), Westphal 3), 4), Stauch 5), Kawanishi 6), Coley 7), Shear 8), Okamoto 9), Nagashima 10), and Kayama 11), have reported that lipopolysaccharide fragments containing a certain amount of protein have antitumor activity and serve as inducers of a Shwartzman reaction. Substances that can induce a Shwartzman reaction are widely distributed in nature 12). While the standard procedure for inducing a Shwartzman reaction consists of a preparatory injection followed by an eliciting injection 13), the reaction can also be induced by performing the eliciting injection before the preparatory injection 14, 15) or by repeated preparatory injections 16). In addition, reactions occurring against transplanted carcinoma 17) are considered a form of a Shwartzman reaction. This report describes the results of our extensive study on a Shwartzman-type biological reaction, demonstrating the effectiveness of the reaction, rather than its pharmacological effect, against Ehrlich carcinoma, and our observation of an anti-transplant immunological phenomenon against carcinoma cells, which was considered cellular immunity, in studied animals.

Experimental procedures and results

In the previous study of experimentally induced gastric ulcer in rabbits, we learned that a preparatory injection of killed E. coli bacteria induced a stronger Shwartzman-type local reaction than that of purified endotoxin with or without adjuvant. Therefore, in the present study we used E. coli, S. typhi, and S. marcescens bacteria killed by heating at 60 °C for 30 min and ovalbumin as experimental materials and we evaluated their effect on Ehrlich carcinoma by monitoring the biphasic response of a pre-transplant injection (into the abdominal cavity of mice) and a intra-mass injection (into the carcinoma cell mass) after transplantation. A total of 1 mg of saline suspension of killed bacteria or 5 mg of ovalbumin was injected into the abdominal cavity of mice in several doses at an interval of a few days. One week after the last injection, 0.25?0.5 ml of saline suspension of killed bacteria (1 mg/ml) or 0.25?0.5 ml of 0.2% ovalbumin in saline was injected into the base of and around a mass of Ehrlich carcinoma cells that had grown to the size of a red bean from subcutaneous transplantation of 5 x 106 carcinoma cells in the lumbar region over 5?10 days prior to injection. The experiment was repeated with different combinations of bacterial species, and the results are shown in Table 1. Before transplantation of carcinoma cells, a total of 1 mg of killed bacteria was intraperitoneally injected in several doses at an interval of a few days, which stimulated the reticuloendothelial system but did not interfere with carcinoma cell transplantation, resulting in death of all animals due to carcinoma. When carcinoma cells were subcutaneously transplanted before intraperitoneal injection of killed bacteria, no significant regression in mass size (as observed with the biphasic method) was observed, and all animals died. These results indicate that the pharmacologic effect of the crude endotoxin contained in bacteria is not immediately manifested as an antitumor effect. Intra-mass injection alone did not lead to a cure in any mice. With the biphasic method, a reduced volume of intraperitoneal injection also resulted in the reduced effect of intra-mass injection. These observations demonstrate that the antitumor effect of bacterial components varies substantially according to the mode of injection.
The observed effectiveness of intra-mass injection of killed E. coli or S. typhi bacteria after repeated intraperitoneal injections of the same killed bacteria seemed consistent with the Arthus phenomenon. However, the fact that a better outcome was obtained with the combination of E. coli and S. marcescens and the ineffectiveness of ovalbumin suggest that the actual mechanism of action of this reaction is more similar to the mechanism of a Shwartzman reaction mediated by allergic reaction to lipopolysaccharide-protein complexes, rather than to the Arthus phenomenon mediated by specific antigen-antibody reaction. Further investigations are needed to test all substances that induce a Shwartzman reaction for the optimal sequence of administration, comparative effectiveness of different combinations, and relationship between effectiveness and substance purity.
Histological findings observed in the mass after intraperitoneal injection of killed E. coli bacteria combined with intra-mass injection of killed S. marcescenes bacteria included hemorrhage, necrosis, fibrosis, sparse collagenization, nuclear degeneration, giant cell transformation, and exudation of leukocytes and plasma cell reaction in all animals, as well as the appearance of histiocytes in half of the animals. Other histological findings observed in each organ were as follows: edema, capillary congestion, and peritracheal cellular response in the lung; opacification and vacuolization of hepatocytes, vacuolated and irregular-sized nuclei, swelling of Kupffer's cells, and perivascular leukocyte response in the liver; and congestion, enlargement of follicles and germinal centers, and other findings suggestive of hyperimmunization in the spleen. Similar findings were observed in lymph nodes, including sinus catarrh and plasmocytosis, and in the kidney, including increased nuclei in glomeruli, and congestion and opacification of tubular epithelia, in all animals.
Mice that recovered through this method were rechallenged with 5 x 106 carcinoma cells. In some animals, carcinoma grew to a rice grain-sized mass, but was then gradually absorbed and reduced in size, resulting in non-engraftment of cells in all animals. This suggests that a Shwartzman-type reaction not only induces necrotic reaction spreading to carcinoma cells, but also provides the animal with anti-transplant immunity, probably mediated by very robust cellular immunity. There has been no evidence of specific antibodies against Ehrlich carcinoma in animal studies. The process of acquisition of this anti-transplant immunity is similar to that of immunity against skin allografts. When offspring born to mice immunized against Ehrlich carcinoma were raised and transplanted with Ehrlich carcinoma cells, cells were engrafted in all animals. It is therefore very unlikely that repeating this training results in the production of a mouse strain resistant to carcinoma cell transplantation.
The experiment was successful with Ehrlich solid carcinoma, but not with Ehrlich ascites carcinoma. After multiple subcutaneous injections of killed E. coli bacteria into mice, Ehrlich ascites carcinoma cells were intraperitoneally transplanted. After 3 days, a saline suspension of killed E. coli bacteria was intraperitoneally injected, resulting in the death of all animals due to carcinoma-associated ascites accumulation.

This abstract was presented at the 11th Bacterial Toxin Symposium and at the 14th Annual Meeting of Japanese Society of Allergology.
The author would like to thank Dr. Yamanaka for reviewing this manuscript and Drs. Hosokawa and Yamashita of Yamaguchi Medical College (currently Yamaguchi University School of Medicine) for their assistance preparing tissue specimens.

Discussion

The mechanism by which normal cells undergo malignant transformation is not related to biological interaction between carcinomas and carcinoma-bearing animals; the two issues should thus be considered separately. The former issue is still being extensively studied in many areas. For the latter issue, however, in particular from the immunological point of view, carcinomas can be regarded as highly compatible homografts or isografts. With regard to homografts, although attention is focused on only graft survival in skin and organ transplantation, attention should also be paid to the opposite phenomenon. It would therefore be helpful to study the underlying mechanisms of biological reactions that preclude engraftment of homo-, iso, and auto-grafts. In this sense, idiopathic carcinomas may be regarded as autografts that undergo metastasis, a form of transplantation, without external manipulation.
While studies are needed regarding issues related to reticuloendothelial systems, pharmacological effects of endotoxin, dynamics of leukocytes and plasma cells, cell response induced by nonspecific inflammation and other issues, it is also meaningful to further characterize Shwartzman-type biological reactions for application to cancer treatment and for better understanding of associated cell/antibody-type immunological phenomena.

This abstract was presented at the 11th Bacterial Toxin Symposium, the 23rd Annual Meeting of Japanese Cancer Association and the 14th Annual Meeting of Japanese Society of Allergology.
The author would like to thank Dr. Yamanaka for his reviewing this manuscript and Drs. Hosokawa and Yamashita of Yamaguchi Medical College (currently Yamaguchi University School of Medicine) for their assistance in preparation of tissue specimens.