Studies of the biochemistry of neoplasia during the first half of this century centered on qualita- tive and quantitative differences in enzyme levels reflecting gene expression. Approaches to an understanding of the most critical characteristic of neoplasia, that of relative autonomy, had to await methods for the investigation in the regulation of genetic expression in eukaryotic cells. Early investigations were concerned with changes in enzyme activities and, later, actual rates of enzyme synthesis and mRNA expression in neoplasms as compared with their appropriate tissue of origin in response to specific environmental agents including hormones, drugs, dietary fac- tors, and other environmental changes.
The Regulatory Cybernetics of Preneoplasia and Neoplasia
Although the term cybernetics refers to the theoretical aspects of control processes in various systems, it is used here in a more general reference to the control of the variety of processes necessary for the viability of a cell. It is the regulation of these control processes, the regulation of genetic expression, whose alteration is likely the key to our ultimate understanding of the molecular nature of the neoplastic transformation. Here, a number of examples of the alteration of the regulation of genetic expression in neoplastic cells are considered; where possible, this discussion is extended to the cell and molecular mechanisms of such alteration, the basis of the relative autonomy of neoplasia (Chapter 2).
Compared with studies on neoplastic tissues, relatively few investigations have attempted to in- vestigate the regulation of gene expression in preneoplastic tissues. In epidermal papillomas of mouse skin, overexpression of some genes occurs, including 12-lipoxygenase (Krieg et al.,
1995), metallothionein (Hashiba et al., 1989), and mRNAs for which a function has not been defined (Krieg et al., 1993). In other instances, regulation of gene expression is abnormal or absent, such as with K14, a keratin protein (Roop et al., 1988), and enzymes of xenobiotic me- tabolism (Reiners et al., 1998).
Earlier studies by Kitagawa (1971) noted that partial hepatectomy did induce mitotic activ- ity in preneoplastic nodules and hyperplasia of liver but less so in hepatomas in the rat. Studies of gene expression in altered hepatic foci have shown numerous differences in gene expression (Pitot, 1990) as well as altered responses to environmental agents compared with normal hepato- cytes (Olsson et al., 1995). Removal of the inducing agent or promoting agent may cause not only a loss of cells through apoptosis, but also a reversal of the altered gene expression prior to cell death (Neveu et al., 1990; Li and Rozman, 1995). This latter point is indicative of the tran- sient nature of cells in the stage of promotion as well as their altered gene expression.
Studies on the cybernetics or regulation of gene expression in hepatocellular carcinoma cells in the stage of progression were initiated some four decades ago in studies on the hormonal and substrate regulation of several different enzymes of amino acid metabolism (Pitot, 1959; Pitot et al., 1961). Some of these earlier studies had been reviewed previously (cf. Goldfarb and Pitot,1976; Sabine, 1975), but no generalization could be made. Table 15.3 gives examples of some of these earlier studies, which indicate both a normal response as well as an absence of response of gene expression to the administration of specific agents and under certain physiological condi- tions. While no specific regulatory mechanism is defective in all neoplastic cells when compared with these effects in their normal counterparts, even in most instances where an effect is noted in the neoplasm, this can be distinguished by differential sensitivity or other factors from that seen in the normal circumstance. An abnormality in control mechanisms that was thought to be char- acteristic of all hepatomas was that of the loss of the repressive control of cholesterol synthesis (Brown et al., 1973). However, in cell culture and in “preneoplastic” livers, the feedback both by dietary cholesterol in vivo and addition of the agent to cell culture resulted in some degree of repression of cholesterol biosynthesis (Sabine, 1976; Depass and Morris, 1982). The effect, which centers around the regulation of the synthesis of the key enzyme, hydroxymethylglutaryl-coenzyme A (HMG-CoA) reductase, a microsomal enzyme, has been described in both experi- mental and human neoplasms of the liver as well as leukemic cells. In view of the fact that hepatoma cells growing in vitro do exhibit cholesterol feedback of the expression of HMG-CoA reductase (Beirne and Watson, 1976), the original generalization developed in vivo was ques- tioned as to its significance on the basis of the availability of the regulatory molecule(s) to the neoplastic cells because of differences in cholesterol accumulation and transport to the neoplasm itself (Harry et al., 1971). As with a number of other circumstances where enzyme regulation appears to be abnormal, the lack of effect of cholesterol on HMG-CoA reductase expression does not appear to be a result of structural alterations in the enzyme itself (e.g., Beirne et al., 1977).
In contrast to the attempted generalization of loss of cholesterol feedback in hepatomas, xenobiotic metabolism in such lesions is for the most part qualitatively not defective when neo- plasms are compared with controlled tissues. Again, studies on the structure of the enzymes in- volved did not reveal significant differences (Ohmachi et al., 1985). The expression of several enzymes of amino acid metabolism in response to substrate and hormonal factors is abnormal (Tryfiates et al., 1974), while in other circumstances neoplasms appear to respond to the more general stimulus of dietary protein influx like normal tissues in liver (Brebnor et al., 1981). However, as noted in the table, the lack of response of hepatomas to the hydrocortisone induc- tion of tyrosine aminotransferase seen in vivo in pyridoxine (B6) deficiency is likely a result of the inability of the neoplasm to synthesize sufficient active forms of the vitamin in vivo (Tha-nassi et al., 1981). This is of note since the regulation of this specific gene has been extensively studied both in normal and neoplastic tissue in vivo and in vitro for a number of years as a model of the regulation of enzyme synthesis (Sorimachi et al., 1981; Thompson et al., 1966). Thyroid hormone is also capable of inducing mitochondrial α-glycerophosphate dehydrogenase in well- differentiated but not poorly differentiated hepatomas (Hunt et al., 1970). This finding, perhaps unexpected, indicates that as the stage of progression continues, regulation of specific genes will likely be further altered or lost altogether.
Thus, from these earlier studies it is not possible to draw any qualitative conclusions with respect to specific gene up- or downregulation. However, if one investigates quantitative differ- ences in gene expression in different neoplasms, an interesting phenomenon occurs, at least in rat liver. This finding is seen in Figure 15.1, in which the activities of four different enzymes and the glycogen content of rat liver and of nine highly and well-differentiated hepatocellular carci- nomas exhibiting chromosome numbers as shown is demonstrated. In each box there are three distinct points related to activities and content on a 12%, 30%, and 60% protein diet. The dark areas indicate the enzyme activity. While there are relatively few qualitative differences in nor- mal and neoplastic tissue, the striking characteristic is that the set of data for each neoplasm is unique to that neoplasm and also distinguishes it from normal liver. Other studies have also indi- cated the heterogeneity of phenotypes of each neoplasm studied (Mehle et al., 1993; cf. Arvan,1992). While this finding may be interpreted in many ways, one scenario is that the alterations reflect the beginning of karyotypic instability, which is the basic alteration in the stage of pro- gression (Chapter 9). Interestingly, some degree of this has also been seen in preneoplastic le- sions in the liver (see below).
This phenotypic heterogeneity of the regulation of gene expression in neoplasms, mostly hepatic in origin, is the complete opposite of the Warburg and Greenstein concepts of the bio- chemistry of neoplasia as described earlier in this chapter. The establishment of the Morris series of transplanted hepatocellular carcinomas as a major model system for the study of the biochem- istry of neoplasia served to accelerate the search for critical biochemical differences between normal and neoplastic cells. Well-differentiated and, to a lesser extent, even poorly differentiated hepatic neoplasms (cf. Potter and Watanabe, 1968) exhibit phenotypic heterogeneity of altered regulation of genetic expression. Such diverse populations of neoplasms as mouse hepatomas
Figure 15.1 Activities of four enzymes and the glycogen content of rat liver and nine Morris hepatomas having the chromosome numbers shown in the bottom line. Each box indicates, by means of the darkened areas, the level of each of the four enzymes and the glycogen content of animals on a 12%, 30%, and 60% protein diet fed for at least 1 week. The abbreviations are as follows: TAKG, tyrosine aminotransferase; SDH, serine dehydratase; G6PDH, glucose 6-phosphate dehydrogenase; CCE, citrate cleavage enzyme. (From Potter et al., 1969, with permission of the authors and publisher.)
(Reynolds et al., 1971), myelomas (M. Potter, 1972), rat mammary adenocarcinomas (Hilf et al.,1970), human endometrial cancers (Siracky, 1979), human renal cell carcinomas (Mehle et al.,1993), and many others (Dexter and Calabresi, 1982) exhibit metabolic, genomic, and regulatory heterogeneity. The phenotypic heterogeneity shown in Figure 15.1 indicates the variation in quantitative enzyme levels under three different environmental conditions (dietary protein con- tents) in normal liver and in nine different, well-differentiated hepatocellular carcinomas. As the reader can appreciate, no vertical set of five boxes is identical with any other. On the other hand, there is no single enzyme abnormality or regulatory abnormality common to all neoplasms stud- ied. Further investigations of the Morris tumors as well as other well-differentiated hepatic neo- plasms have not uncovered any duplicate phenotypes. Unfortunately, with the possible exception of the regulation of cell replication or DNA synthesis, which may be considered to be abnormal in all neoplasms as the operational part of our definition (Chapter 2), no single regulatory mecha- nism has thus far been found to be abnormal in all neoplasms investigated. Thus, the search for a specific biochemical abnormality present in all neoplasms and directly related to the neoplastic transformation has not borne fruit, but all neoplasms exhibit one or more abnormalities in the reg- ulation of genetic expression as might be expected from the definition of neoplasia (Chapter 2).
A somewhat analogous situation to the findings in Figure 15.1 may be implied by data on the phenotypic heterogeneity of preneoplastic enzyme-altered foci developing during the stage of promotion in rat hepatocarcinogenesis (Buchmann et al., 1992; Hendrich et al., 1987). While these data of phenotypic heterogeneity of preneoplastic lesions are not quantitative, the data are certainly compatible with the concept that preneoplastic cells and neoplastic cells during the early stages of progression show a degree of heterogeneity of gene expression that suggests that no single qualitative chemical defect common to all neoplasms will ever be discovered. How- ever, such data do indicate that an understanding of the mechanism of this variability may be one key that helps to unlock the door to our understanding of the entire process of the neoplastic transformation.