INFECTIOUS AGENTS AS CAUSES OF CANCER
In 1913, Johannes Fibiger (cf. Clemmesen, 1978) reported the occurrence of papillomas of the stomach in rats infected with Spiroptera, a small parasitic worm. The intermediate host of the parasite was the cockroach. By feeding either cockroaches infested with the worm or the worm itself to rats, Fibiger was able to produce lesions that were interpreted as papillomatous growths of the stomach. In 1927, Fibiger was awarded the Nobel Prize for this work, the first such award for cancer research. Unfortunately, later studies did not bear out Fibiger’s thesis that the stomach lesions were neoplasms but rather indicated that the tumors resulted from the combination of a deficiency of vitamin A in his experimental animals and the infestation by the parasite.
Despite this setback, over the years many investigators have reported that certain biologi- cal factors are important in the causation of cancer. The parasitic worm Spirocerca lupi is associ- ated with esophageal sarcomas in the dog (Thrasher et al., 1963). Studies have demonstrated that the worms encyst in the wall of the esophagus, and a sarcoma may arise around this cyst. Shortly after Fibiger’s discovery, Bullock and associates demonstrated genetic factors involved in the induction of sarcomas in rats by infection with Cysticercus fasciolaris, the common tapeworm of the cat (Curtis et al., 1933). Later studies demonstrated that sarcomas could be induced by the intraperitoneal injection of washed, ground larvae of the parasite in rats (Dunning and Curtis, 1953). The frequent association of Schistosoma haematobium infection with bladder cancer in various parts of the world, such as Egypt and other regions of Africa, has clearly indicated the association of the parasitic with the neoplastic disease (cf. Shimkin, 1977). Bladder cell hyper- plasia and occasionally bladder carcinoma can be induced with some regularity in monkeys by infection with S. haematobium, and infection of mice with Schistosoma mansoni enhances the hepatocarcinogenicity of various chemicals (cf. Cheever, 1978). Chronic infection with the latter organism has also been associated with hepatic and colon cancer in the human (Chapter 12).
Infectious Agents in Plant Neoplasia
The first reported example of a biological agent inducing a neoplasm in plants was that of Smith and Townsend (1907), who reported that a bacterial infection in the region of wounds in a plant resulted in growths called crown galls. Since those classic experiments, investigations have dem- onstrated repeatedly that under suitable experimental conditions the crown gall plant tumor can be induced in practically all wounds in a large number of dicotyledonous plants inoculated with Agrobacter tumefaciens (cf. Braun, 1975; Kupila-Ahvenniemi and Therman, 1968). These au- thors pointed out that crown gall tumors cannot be induced in monocotyledonous plants, which
show little or no reaction to wounding, whereas wounds in dicotyledonous plants stimulate a marked reaction in the plants, accompanied by DNA synthesis, mitosis, and increased ploidy of adjacent cells. Once infection by the bacterium has occurred and the crown gall tumor has formed, the bacterium is no longer necessary to maintain the existence and growth of the tumor.
Within the last three decades, rapid advances have been made in our understanding of the mechanism of tumor induction by A. tumefaciens. Only those bacteria containing a large extra- chromosomal plasmid are capable of inducing tumors (Watson et al., 1975). This plasmid, termed the Ti-plasmid, consists of a single circular molecule of DNA of molecular weight 90 to
150 million. That the plasmid is important in the transformation of normal to neoplastic cells was indicated by the fact that genes of the plasmid coded for the formation of enzymes synthe- sizing “opines,” amino acids not found in normal plants but synthesized in crown gall tumors that resulted from the bacterial infection (cf. Schell, 1982). Chilton et al. (1977) and others showed the incorporation of Ti-plasmid DNA from the bacteria into the nuclear DNA of plant cells; these cells ultimately develop into the crown gall tumors. This incorporated DNA thus becomes part of the host cell genome and codes for enzymes involved in opine synthesis. This transfer of genetic information from the bacteria to the plant cell is a natural example of “genetic engineering.” The soil bacterium forces plants, through the transfer of specific genes, to produce substrates, or opines, which are used solely by the soil bacterium containing the Ti-plasmids; this equips them to metabolize the opines for survival (Schell, 1982).
Thomashow et al. (1980) demonstrated that each line of crown gall tumor contained a “core” segment of Ti-plasmid DNA, which was postulated to be responsible for maintaining the transformed state. The infection of the plant by A. tumefaciens induces both chemical signaling from the plant cells by the production of phenolic compounds and the direct binding of the bac- terium to the plant cell itself (Stachel and Zambryski, 1986; Winans, 1992). Roughly 25 Ti plas- mid–encoded genes, termed vir genes, may be required for the development of tumors within the plants (cf. Winans, 1992). Many of these vir genes are involved in the processing of the T-DNA of the Ti plasmid. The gene products of the vir region facilitate the T-DNA transfer from the bacterium to the plant cell by a process very similar to bacterial conjugation (Beijersbergen et al., 1992). Furthermore, Hadley and Szalay (1982) and White et al. (1983) have presented evi- dence that the genome of the untransformed host plant cell contains sequences in its DNA that are homologous to sequences in or near the core region of the Ti plasmid. That DNA from the Ti plasmid is required for the maintenance of the neoplastic state in the plant cell is demonstrated by the fact that the neoplastic phenotype was lost concomitant with the loss of foreign DNA in the plant cell (Yang et al., 1980). Such loss of tumorigenicity always involved the loss of the vir region incorporated into all neoplasms studied (Yang and Simpson, 1981; Winans, 1992). This loss likely occurs during meiosis of the plant cell (Turgeon et al., 1976).
These findings are of considerable interest, especially in view of the findings of the infec- tious nature of some neoplasms in higher vertebrate animals. Some of the characteristics of crown gall tumor induction by A. tumefaciens are listed in Table 4.1. The reader should remem- ber these characteristics and refer back to them after considering the later discussions in this chapter. This comparison will underline the similarity of the mechanisms involved in the induc- tion of crown gall tumors in plants and the induction of malignant neoplasms in animals by on- cogenic viruses.
Bacteria as Causal Agents of Neoplasia
As noted in Chapter 1, during the latter part of the nineteenth century, when bacteria were shown to be the causal agents of many infectious diseases, it was natural for scientists to consider neo- plasia as the result of a similar cause. The studies by Doven and the claim that Micrococcus
Table 4.1 Characteristics of Crown Gall Tumor Induction by Infection with Agrobacter tumefaciens
1. Tissue regeneration, mitosis, DNA synthesis of host cells of infected wound
2. Infection of plant with plasmid-containing bacterium
3. Incorporation of plasmid DNA into host plant cell genome
4. Presence of DNA sequences common to all plasmid-transformed host cells; these sequences are apparently necessary for maintenance of the neoplastic state
5. Presence of sequences in untransformed plant cells that are homol- ogous to DNA sequences in plasmids of A. tumefaciens
neoformans was causative for many if not all human neoplasms are certainly examples of this. However, bacteria have actually only recently been clearly demonstrated to be causative agents in specific human neoplasms. In 1983, Warren reported the presence of an unidentified curved bacillus closely associated with the gastric epithelium in humans exhibiting active chronic gas- tritis. Subsequently, the organism was classified as Campylobacter pylori and shown by many others to be associated with gastritis and ulceration of the upper gastrointestinal tract (cf. Dooley and Cohen, 1988). Today this organism is known as Helicobacter pylori (DeCross and Marshall,1993). Chronic infection with this bacterium in the human has now been associated, probably causally, not only with the development of chronic gastritis and ulceration of the gastric and upper intestinal mucosa, but also with the development of intestinal metaplasia of the gastric mucosa (Rugge et al., 1996), gastric carcinoma (Parsonnet et al., 1991; Hansson et al., 1993), and gastric lymphoma (Parsonnet et al., 1994) (Chapter 12). There is no evidence that the bacte- rium induces neoplasia by mechanisms seen with Agrobacter induction of crown gall tumors in plants. Mechanisms whereby this bacterium may induce neoplasia in the human are further con- sidered in Chapter 12. Following investigations in the human, it is now clear that species of Helicobacter occur in other mammals as well. Helicobacter mustelae has been associated with gastric carcinoma in the ferret (Fox et al., 1997) as well as with a gastric lymphoma (Erdman et al., 1997). Similar gastric lymphomas have been induced in mice with another member of this bacterial family, Helico- bacter felis (Enno et al., 1995). A somewhat more practial problem is the demonstration that Helicobacter hepaticus is capable of inducing chronic hepatitis in specific strains of mice with the subsequent development of hepatocellular neoplasms (Ward et al., 1994). Infection causes a chronic stimulus to hepatocyte proliferation which increases the risk for the development of hepatocellular neoplasia (Nyska et al., 1997). Thus, with the discovery of this bacterial class, it is clear that such organisms may induce neoplasia both in humans and in experimental animals. The exact mechanisms for such induction will be considered later. The remainder of this chapter considers the primary infectious causative agents of neoplasia—viruses.
Intracellular parasites have long been implicated as the cause of certain types of neo- plasms in vertebrate animals. In 1909, Ellerman and Bang demonstrated a viral causation for avian leukosis. In 1911, Rous reported that cell-free extracts of a sarcoma in chickens would in some instances produce sarcomas when injected into other chickens. In 1932, Shope described the Shope papillomavirus of rabbits, and in the 1930s the Bittner mouse mammary tumor virus was discovered (see below; Bittner, 1936). However, the general significance and potential im- portance of viruses in the causation of cancer was not generally appreciated until after the dis- covery of lysogeny in bacteria and the importance of latency in viral infections. In 1951, Gross reported that injection of cell-free filtrates from AKR mice with leukemia into newborn AKR mice resulted in a high percentage of leukemia in the inoculated animals at a later period. Since Gross’s experiments, numerous individuals have extended these studies, and considerable bio- logical knowledge is now available on the viral causation of neoplasms.