Neoplastic Transformation Part 1 (Molecular Biology)

The term neoplastic transformation first came into common use in the 1950s to refer to changes in cell morphology and growth behavior in cell culture related to those described for tumors in animals. The changes in culture came about either spontaneously during long-term culture of mouse embryo cells or as a result of infection of chicken embryo cells with Rous sarcoma virus (RSV). The term was then applied retrospectively to the changes occurring in the development of cancer itself in animals and to a lesser degree in plants. During the 1970s, individual transforming genes were identified with different RNA-containing tumor-inducing viruses, including RSV, known collectively as retroviruses. The RNA of these viruses was transcribed during infection into DNA and integrated into the genome of the host cell, where it brought about its transforming action. The DNA of the transforming genes, or oncogenes, was used to "transfect" normal chicken embryo cells and mouse embryo cells and to transform them into neoplastic cells in the same way as the virus itself did. The DNA of normal vertebrate cells contains nucleotide sequences similar to those of the retrovirus oncogenes, and a mutated form of one of these was isolated from the DNA of a cell culture derived from a human bladder cancer. It was inferred that oncogenes were mutated forms of cellular genes related to viral oncogenes, and they played an important part in the development of human cancer. This inference was related to much earlier results in chemical carcinogenesis in experimental animals, particularly mice, which suggested that the primary event in the development of a tumor was a mutation that then required a series of other events to bring the tumor to full expression. A later development was the concept of tumor suppressor genes. As methods are refined for detecting genetic changes in tumors, large numbers of such changes can be found in a single tumor, particularly in solid cancers of humans, and the role of any particular genetic change in development of the cancer is uncertain. The sum of these observations indicates that the development of cancer is an extremely complex process involving many genetic, as well as epigenetic, changes in the host.


1. Experimental Carcinogenesis in Animals

The first strong experimental indication that mutation plays a primary role in neoplastic development came from the application of polycyclic aromatic hydrocarbons (PAHs) to the skin of mice. The earliest work in rabbits and mice indicated that repeated application of a PAH was necessary to induce a tumor. In the 1940s, it was discovered that the combined application of a PAH and a non-carcinogenic substance (ie, croton oil) was more effective than the PAH alone. In fact, a single application of a PAH to the skin of a mouse produced no tumors, unless it was followed by repeated applications of croton oil over many weeks. If the repeated application of the croton oil was made before the PAH, no tumor resulted. A key finding was that the interval between the single PAH application and the croton oil could be as long as 10 months and a positive result would still be obtained. It was therefore concluded that the single dose of the PAH induced a permanent genetic change, called initiation, that required the promoting action of croton oil or further PAH application to produce a tumor. Because the croton oil was itself noncarcinogenic, promotion was considered an epigenetic or physiological effect. The most active ingredient of croton oil was 12-O-tetradecanoyl-phorbol-13-acetate (TPA), more recently termed phorbol-13-myristate-12-acetate (PMA) (see Phorbol esters).

The actions of initiators and promoters are subsumed under the rubric of the two-stage theory of carcinogenesis. While the two-stage theory has gained wide acceptance, it is not without its critics who interpret results such as the seeming unidirectional action of initiators and promoters in a quite different way, that is, "carcinogenesis is most probably composed of a multifarious combination of both genetic and epigenetic phenomena" (1).

The two-stage theory of carcinogenesis is based largely on studies of carcinogenesis of mouse skin. Another carefully studied model of carcinogenesis is in the rat liver. There a carcinogenic procedure consisting of systemically inoculating a carcinogen, such as dimethylhydrazine or 2-acetylaminofluorine, that does not initiate in the same way as the PAHs do in the skin, but inhibits proliferation of the vast majority of hepatocytes and selects a few resistant hepatocytes for proliferation. When coupled with mitotic stimulation, such as removal of part of the liver (partial hepatectomy), many nodules of altered, proliferating cells are formed. If the hepatectomy is delayed for 48 h or longer, the number of nodules is greatly reduced. In any case, 90% to 98% of the nodules redifferentiate into normal-appearing liver tissue. Changes typical of cancer appear only in a small fraction of the nodules. The rapid decrease in the number of nodules with increasing delay in applying the mitotic stimulation, along with the subsequent redifferentiation of most of the nodules, has suggested that the initiating treatment does not directly result in mutational change but is a physiological adaptation resulting in hepatocytes that resist growth-inhibitory toxic material. The cells in these nodules have an increased probability of undergoing the genetic changes that eventually result in liver cancer (2). It is apparent therefore that the simple relation between initiation/mutation and promotion that arose from PAH carcinogenesis in mouse skin is inadequate as a generalization for all forms of cancer.

2. Mutagenic Action of Carcinogens

The two-stage model of carcinogenesis led to attempts to demonstrate the mutagenic action of known carcinogens, particularly the PAH compounds, which were the only ones that regularly caused neoplastic transformation when painted on the skin. Attempts to demonstrate their mutagenicity in the fruit fly Drosophila melanogaster, a favorite organism for genetic studies, were largely negative or indecisive. Beginning in the 1950s, however, genetic studies turned increasingly to the use of microorganisms, which would provide many millions of dividing organisms that would exhibit mutational change within a day or two. A test was developed for mutations in the bacterium Salmonella typhimurium that seemed capable of detecting about 90% of the animal carcinogens (see Ames Test). The small number of presumed nongenotoxic, epigenetic carcinogens were considered a negligible problem. Broader studies later reduced the sensitivity of the assay to 53%, and they also reported that about 30% of noncarcinogens were mutagenic in the test. Further study confirmed these results and recommended dividing carcinogens into genotoxic and nongenotoxic categories (3). The picture is further complicated by the fact that a diet deficient in choline or methionine is carcinogenic. Some interpretations of these results could be used to question the significance of genetic change in the origin of cancer. However, such doubts are countered by the fact that the Salmonella test detects only base pair substitutions and frameshift mutations and therefore would not detect large-scale chromosome rearrangements and deletions. Such changes are, in fact, considered to be the most important ones in the origin of human cancer (4) and will be discussed further in the text below.

There is an efficient assay that detects larger-scale changes in the genome and that had identified many carcinogens that were negative in the Salmonella test. It uses a line of mouse lymphoma cells that are heterozygous for a gene that phosphorylates thymidine (5). When the active allele becomes inactive as a result of genetic alteration, the cell can survive under certain conditions that kill the unaltered cells containing the active allele. The inactivation can occur either by local (intragenic) changes in base sequence or by large-scale deletions or rearrangements of chromosomes. The latter types of change can be distinguished from the former because they also cause an inheritable reduction in growth rate of the altered cells. Probably because of the dual response, this assay has detected several important carcinogens (such as the hormone diethylstilbesterol) that were negative in the Salmonella test, and it exhibits a good quantitative correlation between mutagenicity and carcinogenicity in experimental animals (5). Many of the compounds that are positive in this assay do not interact directly with DNA, but their ultimate effect is to alter the genome regardless of their initial target. These would be termed epigenetic or dysgenetic agents (6).

3. Viral Transformation of Cells

The first clear-cut demonstration that some viruses could induce neoplastic transformation was the induction of sarcomas in chickens by inoculation of filtrates from a naturally occurring connective tissue tumor of chickens that had first been transplanted serially by intact cells in closely related chickens. The virus was later named Rous sarcoma virus (RSV) after its discoverer and was the source of most of the quantitative work done on viral carcinogenesis in animals for many years. Single infectious particles of RSV could initiate epithelial tumors on the chorioallantoic membrane of the developing chicken embryo, and progeny virus particles from the epithelial tumors initiated sarcomas or connective tissue tumors in the tissue underlying the epithelial tumors, or when inoculated into adult chickens. Every infected cell gave rise to a tumor, and unlike the cytocidal or cell-killing viruses, all the infected cells could proliferate. It was therefore obvious that a small amount of genetic material could initiate neoplastic transformation, the first clear indication that genetic change could cause cancer. There was as yet, however, no evidence that some change in the host cell genome itself could induce transformation. The viral genome was made of RNA instead of DNA. However, it was discovered that the virus contained an enzyme, reverse transcriptase, that converted the viral RNA into DNA, which would then integrate into the genome of the cell, where the protein readout of the viral oncogene would transform the cell. This confirmed experiments on radiation of cells before infection with RSV which showed that loss of ability of cells to reproduce prevented the establishment of virus infection and suggested that the viral genome had to integrate into cellular DNA before the virus could multiply. Because there was a nucleotide sequence in the normal cellular DNA that was very similar to that of the viral RNA oncogene, the idea arose that a simple mutation in the cellular gene would lead to neoplastic transformation.

The opportunity for precise quantification of viral transformation arose when a method for infecting and transforming chicken embryo cells in culture was developed. Infection resulted in morphological transformation of spindle-shaped fibroblastic or connective tissue cells into a more rounded sarcoma cell. More significant than the morphological change, however, was the altered growth behavior of the cell. Normal fibroblasts multiply rapidly in culture when they are sparsely distributed on the floor of the dish, but when they become crowded the cell-to-cell contact results in a marked decrease in growth rate (see Contact Inhibition). Net growth ceases when the cells multiply enough to form a confluent sheet. However, the cells transformed by RSV continue to multiply after they come in contact with other cells. If there is one transformed cell surrounded by normal cells, the latter form a single sheet of cells, but the former continues to multiply into several layers and form what is called a transformed focus. If the cells from the focus are dispersed from one another and transferred to a new culture dish along with a large majority of non-transformed cells, many neoplastically transformed foci result, each one like a small tumor.

Some strains of RSV are defective in the sense that they require coinfection of the same cell by a helper virus in order to produce new infectious virus. The reason for the defect is that the defective strains lack the information to produce the outer coat of the virus, which is required to adsorb to and penetrate into another cell. However, the defective virus can transform the cell to which it has gained entry even when no new infectious virus is produced. Clearly the coat protein does not participate in the transformation itself.

Just as there are a number of transforming retroviruses, there are a number of transforming viral oncogenes, and their protein translation products are associated with different functions in the cell. The transforming src gene of RSV codes for a protein kinase (an enzyme that phosphorylates protein), and considerable effort has been expended to find the cellular target of phosphorylation that causes transformation. However, the enzyme is promiscuous in the proteins it phosphorylates, and no single target protein has been found that transforms upon phosphorylation.

There is another oncogene called the ras gene, which stands for rat sarcoma virus from which it was isolated. A particular substrain, the Harvey- or H-ras gene, was the first gene from a human tumor shown to have transforming activity on transfection into a mouse cell culture line. Actually, the gene was only isolated from the tumor cells after they had been several years in culture, and the mutation could have occurred during that period. Later work showed that the H-ras gene, even when mutated, did not transform the target mouse cells directly but had first to recombine with a strong promoter, which led to a 100-fold increase in the protein product of the gene (7). Because there was no indication of a significant overproduction of the ras protein in the original tumor, it is unlikely the mutated gene played a major role in tumor development. It is also noteworthy that recombining the normal cellular ras gene with a strong promoter also leads to transformation, though with a lower efficiency than the recombined mutant ras gene does. The transfection assay in murine cells therefore leads to the questionable conclusion that the mutated ras gene derived from human tumor cells was the cause of the tumor, when it was in fact the artifact of recombination that produced the transformation in the test system. The results, however, do point up the potential significance of chromosome rearrangements in producing new gene combinations that are carcinogenic.

4. Induction of Neoplastic Transformation by Chemical and Physical Carcinogens

Although several unsuccessful attempts were made to transform mouse embryo cells in culture with carcinogenic PAHs (see Spontaneous transformation below), the first success was achieved with Syrian hamster embryo cells in 1965. The untreated hamster embryo cells became large, flat, and vacuolated about 3 weeks after they were explanted to culture from the embryo, slowed down in growth, and finally stopped multiplying after 6 to 7 weeks. The main body of PAH-treated cells exhibited the same early degenerative changes as the controls, but there were also foci of crisscrossing slender, spindle-shaped proliferative cells. Upon isolation, these early-appearing proliferative foci could not be subcultured. Their number and size increased with time of incubation in the original culture. The later transformed cells had a random pattern of organization, in contrast to the flattened, degenerating nontransformed cells of the background, and they continued to multiply on subculture. After about 10 weeks in culture, the cells produced slow-growing sarcomas when inoculated into hamsters. Longer periods in culture resulted in faster growing tumors in hamsters. When single cells were treated with PAH, even if only for a few hours, a significant fraction (2% to 26%) developed into transformed foci after they grew into full colonies. This would constitute a remarkably high frequency of transformation if it were the result of mutation in one or a few genes. There was considerable variation in cell morphology, both within and between the transformed foci. A fairly close relationship existed between the concentration of carcinogen that produced transformation and that which caused an initial inhibition of growth. Noncarcinogenic and weakly carcinogenic PAH produced neither growth inhibition nor transformation. The results suggest a relationship between a degree of cell damage and subsequent transformation, as had been proposed many years earlier. They suggest that PAH treatment introduces genetic instability into cells that results in varied forms of transformation among a fraction of the progeny. Attempts to transform primary or secondary cultures of mouse embryo cells were difficult to interpret, because even the untreated mouse cells underwent transformation. (See Spontaneous transformation below.)

Following the success in chemically transforming hamster embryo cells in culture, efforts were renewed with mouse embryo cells, only this time cell lines were used that had been adapted to grow in culture continuously. Single cells from a permanent line of fibroblasts derived from mouse prostate glands were exposed to a carcinogenic PAH. Under optimal conditions of treatment, 100% of the treated cells that could form a clone exhibited morphologic evidence of transformation and produced tumors upon inoculation into hamsters, while only 5% of the untreated clones did so.

Despite this remarkable result, further work with the prostate fibroblasts was discontinued, possibly because they developed higher frequencies of spontaneous transformation. Instead, a permanent line of fibroblasts was developed from embryonic body wall fibroblasts of the C3H mouse strain. These cells formed a stable, flat monolayer of cells when they reached confluence and did not appear to undergo spontaneous transformation. Brief exposure to a carcinogenic PAH led to the formation of transformed foci after a 6-week period of incubation. A peculiar feature of clonally isolated transformed cells was that they grew at a lower rate than their parental normal cells when seeded at low population density. This again indicates that transformation is associated with cell damage and will be considered further in the text below. The efficiency of chemically induced transformation in mass populations of the C3H cell line appeared much lower than had been obtained with clones of the prostate fibroblasts. However, later PAH treatments of low densities of the C3H cells resulted in transformation at high but variable frequencies. The same cell line was used to study the transforming effect of exposure to X-rays. These experiments gave the surprising result that the number of transformed foci produced by the X-rayed cells when they reached confluence was independent of the number of cells that was used to start the culture after the irradiation. The authors proposed that X-rays induce an ill-defined change in many or all of the surviving cells that is transmitted to their progeny. This change increases the probability of a second step, overt transformation, when the cells are maintained under the growth inhibitory conditions of confluence. The fraction of confluent cells that undergo transformation is small, but almost all the survivors of X-irradiation give rise to progeny that produce some transformed foci. As in the case of the hamster cells treated with PAH, it appears that the carcinogenic treatment induces a genetic instability in most of the cells that can result in transformation in a small fraction of their progeny.

The very high frequency of the first change, while surprising when contrasted with the rarity of specific genetic mutations, is correlated with several other findings. A similarly high proportion of X-rayed cells suffers a heterogeneous, heritable reduction in growth rate signaled by the formation of small colonies. The X-rayed cells and their progeny also display an increased frequency of mutations and an increased sensitivity to mutagenic treatment. The lesions that produce the instability appear to be related to double-strand breaks in DNA and to the elimination of chromosome fragments. Another corollary is the finding that chromosome aberrations are produced in a high proportion of cells by doses of X-rays only one-tenth the mean lethal dose. The aberrations consist mainly of chromatid and chromosome gaps and are found in several copies per cell as the X-ray dose approaches the mean lethal dose. Thus the genetic lesions at the chromosomal level are likely to be associated with the instability induced by X-rays in whole populations that increases the likelihood of neoplastic transformation.

Another example of a high-frequency transformation occurs when either thyroid or mammary cells are removed from rats, X-irradiated in culture, and reinoculated into rats made susceptible to tumor development. As many as one irradiated thyroid cell out of seven will produce a thyroid cancer, and about 1 in 100 irradiated mammary epithelia produce mammary cancer. Similar results are obtained in this procedure when the cells are treated with the chemical carcinogen and mutagen methylnitrosourea. It is likely that the same sequence of changes is produced in these cells as in those that produce transformed foci in culture, namely a pervasive chromosomal destabilization in a cell population that results in much rarer genetic lesions in the progeny that produce tumors.

Exposure of normal fibroblasts from baby mouse skin to white fluorescent light produced a marked increase in the number of transformed foci that appeared in the culture. A conspicuous feature of this treatment was the variation in morphology of cells between foci, contrasted with their relative similarity within a focus. Many combinations of sparse or dense growth, isometric or spindle shape of cells, and large or small, narrow or wide, flat or refractile cells were observed. Also observed were one or more characters commonly associated with spontaneous neoplastic transformation in culture, such as clumping, "cording" in linear arrays, poor spreading, criss-cross or disorderly orientation, multilayering, small cytoplasm, and variable shape. A similar effect was produced by exposing the cells to drugs, such as colchicine, that combine with tubulin and prevent the formation of spindles that are necessary for chromosome segregation during cell division. In the first few divisions after removal of the antitubulin drugs, many cells became tetraploid (double the normal diploid number of chromosomes). There was also widespread asymmetric nuclear division, which creates the potential for chromosome loss. The authors felt that the resultant abnormal chromosome composition of cells (aneuploidy or heteroploidy) along with DNA damage were responsible for transformation of the cells. A large number of the transformed foci were isolated and continued for long-term culture. Although the resulting cultures grew quickly at first, this was followed by slower growth, accompanied by many dying cells. Some of the cultures died out, some grew slowly, and about one out of five became an established cell line. All of them were entirely or mostly polyploid. There appeared to be a continuing but random genetic reshuffling process, with some cell death and selection of viable gene sets. The long-term survival of isolated foci of 20% was orders of magnitude greater than the bulk of the original culture, from which only about one in 105 cells survived. The surviving cells retained some of the features of neoplastic cells but did not produce tumors when inoculated into mice. The cell lines eventually became aneuploid (mainly subtetraploid), resembling in chromosome distribution the preneoplastic cells described in cultures of mouse salivary gland epithelium. Because focus formation is maximized by fluorescent light irradiation plus antitubulin treatment, the authors conclude that DNA damage, as well as chromosome reshuffling, is a significant factor in the neoplastic transformation (8).

5. Spontaneous Neoplastic Transformation in Newly Explanted Cells

When connective tissue is removed from the mouse embryo and the cells separated from one another in culture, they will multiply in a nutrient medium containing animal serum. They can be subcultured at 3-day intervals, but the growth rate decreases at each subculture until they reach a point of crisis. At that point, a new cell type appears that can grow indefinitely in culture. If the subcultures are kept at low population densities, the permanent cell line stops growing at confluence, behaving like normal cells. If the subcultures are made at high density, the ensuing cell line gains the capacity to multiply at high density, a characteristic of neoplastic and preneoplastic cells. Such cells produce tumors when inoculated into the same strain of mice from which the cells originated. This transforming effect of high population density also occurs in the normally behaving established lines of mouse cells. For example, if the cells are grown to confluence before subculture, transformed foci appear on a background of normal cells. These transformed cells can be maintained in culture indefinitely, and they will eventually gain the capacity to produce tumors in mice (9). This process was termed spontaneous malignant or spontaneous neoplastic transformation because it occurred in the absence of deliberate carcinogenic treatment. There was a trend toward a marked increase in chromosome abnormalities at about the time of neoplastic transformation. The abnormalities were chromosomal breaks and minute and biarmed chromosomes. The frequency of transformation depended on the type of serum used in the medium, and it could be increased by exposure to fluorescent light or high concentrations of oxygen.

Chinese hamster embryo fibroblasts also undergo spontaneous neoplastic transformation in culture. A sequence of changes in growth behavior preceding transformation occurs in these cells, beginning with a crisis followed by establishment of permanent growth. The early preneoplastic changes consist of growth to higher population densities in crowded cultures and increased capacity to multiply in low concentrations of protein growth factors. The capacity for isolated cells to multiply and form colonies increases, the cells gain the capacity to grow in suspension, and colonies with transformed morphology appear. After many more subcultures, the cells are able to produce tumors when inoculated into gelatin sponges that had been implanted into the subcutaneous space of the hamsters. Still later, the cells gain the capacity to produce tumors when inoculated in suspension directly into the subcutaneous space. Chromosome studies were conducted during sequential subcultures. After only six passages, about half the cells had an abnormal chromosome composition, with almost every abnormal cell having a unique karyotype (10). By the 19th passage, all the karyotypes were abnormal; about half of them were of the same type, and the others were quite variable. The presence of individual marker chromosomes increased with passage level. Despite the presence of diverse chromosome abnormalities in the early passages, the cells did not produce tumors in gelfoam sponges until about the 30th passage, nor in direct subcutaneous injection until the 54th passage. It is apparent, therefore, that only a small fraction of the total number of abnormal chromosomal combinations are associated with tumor-producing capacity, but many more combinations are associated with preneoplastic growth behavior of the cells in culture. No specific chromosome combination could be definitively linked to tumor formation, although trisomy of chromosome 5 seemed to play an important role. The sequence of changes in culture leading to tumor-forming capacity can be considered a form of progression similar to that which occurs during the development of tumors in animals.

Next post:

Previous post: