J.F. CATTERALL, K.K. KONTULA, C.S. WATSON, P.J. SEPPÄNEN, B. FUNKENSTEIN, E. MELANITOU, N.J. HICKOK, C.W. BARDIN and O.A. JÄNNE,     The Population Council and The Rockefeller University, New York, New York

Publisher Summary

This chapter focuses on the regulation of gene expression by androgens in murine kidney. The murine kidney provides a unique experimental system to study the regulation of gene expression by androgens. Several genes in this tissue respond in a noncoordinate fashion to stimulation by testosterone. The three genes used in studies exhibit diverse characteristics in both their structure and the regulation of their expression by androgens. The ornithine decarboxylase genes form a family of several members, while those for KAP and β-glucuronidase appear to be unique in the mouse genome. Modulation of the affinity of the hormone-receptor complex for a specific binding site on chromatin could occur via a number of mechanisms, such as structural alterations in the receptor protein or the DNA sequence, changes in packaging of chromatin, and the presence of tissue-specific factors regulating this interaction. While androgens have been shown to accelerate the rate of gene transcription, increases in this rate were not sufficient to account for the androgen induction of the steady-state levels of mRNAs for proteins C1, C2, and C3 in the rat ventral prostate or of KAP, MK 908, and ornithine decarboxylase mRNAs in mouse kidney. Nucleotide incorporation into specific gene products is, therefore, proportional to the rate of RNA chain initiation that prevailed in vivo prior to the start of the assay in vitro. Methods that allow direct measurement of the absolute transcription rates or of mRNA turnover rates must be applied to the models for androgen regulation of gene expression, to determine with certainty the ultimate mechanisms by which male sex hormones control accumulation of specific gene products.

I Introduction

The phenotypic expression of steroid hormone action in different target organs varies widely, yet the initial steps in the mechanism of action of all steroids are similar. These hormones probably enter the cell by passive diffusion through the plasma membrane after which they interact with soluble receptor molecules that are specific for each steroid class (O’Malley and Means, 1974; Katzenellenbogen, 1980; Bardin and Catterall, 1981). Whether this interaction occurs in the cytoplasm followed by translocation of the receptor–steroid complex to the nucleus (see Schrader, 1984) or primarily in the nucleus (King and Greene, 1984; Welshons et al., 1984) does not change the basic concept that the receptor–steroid complex has a greater affinity for relevant sites on chromatin than the steroid–free receptor. These chromatin binding regions or “acceptor” sites may include DNA sequences which have been shown to specifically bind steroid receptors in vitro (Payvar et al., 1981, 1983; Mulvihill et al., 1982; Compton et al., 1983; Renkawitz et al., 1984). Although DNA binding alone is sufficient to account for interaction of the receptor–steroid complex with a specific DNA sequence in vitro, one or more nuclear proteins must be involved in determination of tissue– and cell-specific responses of genes to hormones in vivo (Spelsberg et al., 1984). Recent studies have identified nuclear proteins with properties consistent with tissue-specific regulation of the prolactin gene (White et al., 1985). It remains to be proven that these are indeed tissue-specific regulators and the mode of action of such regulatory proteins in steroid hormone action also awaits further elucidation. Whatever the precise nature of the interaction between steroid–receptor complexes and chromatin acceptor sites, the major result of this interaction is regulation of specific gene transcription (Ringold et al., 1977; McKnight and Palmiter, 1979; Swaneck et al., 1979; Brock and Shapiro, 1983a). It is clear, however, that in other cases, steroid hormone treatment affects mRNA stability (Brock and Shapiro, 1983b). Both mechanisms have been documented in the regulation of gene expression by androgenic steroids (Derman, 1981; Page and Parker, 1982; Berger et al., 1986).

The mechanism of androgen action is less well understood than that for other classes of steroids such as estrogens, progestins, and glucocorticoids. There are at least two major reasons for this: the main approaches to study steroid hormone action have recently been the analysis of the structure and function of steroid-regulated genes and the purification and characterization of steroid hormone receptors. Androgen-regulated gene products in general are not of the high abundance class. Therefore, recombinant DNA technology was first applied to those steroid-regulated systems that produced more readily obtainable gene products (O’Malley et al., 1979; Yamamoto et al., 1983; Groner et al., 1984). Similarly, androgen receptors have proven to be unusually difficult to purify owing to their lability and low concentration in target tissues. Recent advances in the stabilization and purification of androgen receptors (Isomaa et al., 1982; Chang et al., 1983) should facilitate their more complete characterization.

One line of investigation that has been particularly useful for the study of androgen action is the identification of genetic mutations that cause androgen insensitivity. Testicular feminization (TF) was the first inherited disorder for which the phenotype, male pseudohermaphroditism, could be attributed to deficient receptors. However, studies of the clinical features of the syndrome alone could not distinguish among several possible mechanisms including defects in steroid metabolism. It was subsequently shown that the defect was in an early step in the pathway of androgen action (French et al., 1966), but studies on the rat model of the syndrome (Stanley and Gumbreck, 1964; Stanley et al., 1973) were required to prove the correlation between the phenotype and abnormality of the androgen receptor (Bullock and Bardin, 1970, 1972). Similar correlations were subsequently made in the mouse (Bullock et al., 1971) and in man (Keenan et al., 1974).

The animal models of the TF syndrome differed in that the rat exhibited a reduced androgen response while the response in the Tfm/Y mouse was completely abolished. These alternatives were assumed to represent syndromes analogous to incomplete and complete testicular feminization as described in man (Bardin et al., 1973). This led to the conclusion that development of the male phenotype was directly proportional to the level of functional androgen receptors (i.e., the TF rat contained a reduced level of the receptor and the Tfm/Y mouse had no functional receptors (Bardin and Catterall, 1981). This view proved to be too simplistic because, as clinical studies continued, it became clear that androgen insensitivity in man was associated with a wide variety of receptor abnormalities (Table I). For instance, complete TF can be associated with a lack of functional receptor, a receptor that exhibits abnormal binding, or an apparently normal receptor complement. On the other end of the spectrum, infertile, but otherwise normal men can exhibit abnormal, apparently normal, or a reduced level of apparently normal receptors. Thus, the direct correlation of functional receptor level and androgen resistance phenotype was untenable.

It seems likely that, insofar as the male phenotype is represented by the sum of the expression of androgen-induced (or -attenuated) genes, whatever mechanism causes differential regulation of gene expression by the hormone would also be responsible for phenotypic variation. In an attempt to study this variation at the molecular level, we have used the expression of several genes regulated by androgens in mouse kidney as representative of the male phenotype. In this article, we present data on the characterization of the genes for ornithine decarboxylase, β-glucuronidase, and kidney androgen–regulated protein. In addition, we will show that each of these genes exhibits a unique pattern of response to androgen treatment. The data support the hypothesis that gene regulation based on differential sensitivity to androgen–receptor complexes can explain tissue–specific phenotypic variation in response to the...