Improved clonal and nonclonal growth of human, rat and bovine adrenocortical cells in culture

The emerging role of senescent cells in tissue homeostasis and pathophysiology

Cellular senescence is a state of permanent growth arrest and is thought to play a pivotal role in tumor suppression. Cellular senescence may play an important role in tumor suppression, wound healing, and protection against tissue fibrosis in physiological conditions in vivo. However, accumulating evidence that senescent cells may have harmful effects in vivo and may contribute to tissue remodeling, organismal aging, and many age-related diseases also exists. Cellular senescence can be induced by various intrinsic and extrinsic factors. Both p53/p21 and p16/RB pathways are important for irreversible growth arrest in senescent cells. Senescent cells secret numerous biologically active factors. This specific secretion phenotype by senescent cells may largely contribute to physiological and pathological consequences in organisms. Here I review the molecular basis of cell cycle arrest and the specific secretion phenotype in cellular senescence. I also summarize the current knowledge of the role of cellular senescence in vivo in physiological and pathological settings.

Human adrenocortical carcinoma cell lines

The human adrenal cortex secretes mineralocorticoids, glucocorticoids and adrenal androgens. These steroids are produced from unique cell types located within the three distinct zones of the adrenal cortex. Disruption of adrenal steroid production results in a variety of diseases that can lead to hypertension, metabolic syndrome, infertility and androgen excess. The adrenal cortex is also a common site for the development of adenomas, and rarely the site for the development of carcinomas. The adenomas can lead to diseases associated with adrenal steroid excess, while the carcinomas are particularly aggressive and have a poor prognosis. In vitro cell culture models provide important tools to examine molecular and cellular mechanisms controlling both the normal and pathologic function of the adrenal cortex. Herein, we discuss currently available human adrenocortical carcinoma cell lines and their use as model systems for adrenal studies.

Regulation of the p21Sdi1/Cip1/Waf1DNA Synthesis Inhibitor in Senescent Human Diploid Fibroblasts

A large body of evidence has demonstrated that normal human fibroblasts have a limited division potential in culture and underwent senescence, a process whereby cells became arrested in the G1 phase of the cell cycle and overexpressed a DNA synthesis inhibitor(s). Cyclin-dependent kinase two (Cdk2) is required for the promotion of the Gi-to-S phase transition in human cells. Senescent fibroblasts contain intact cyclin-Cdk2 complexes but cannot induce Cdk2 protein kinase activity in response to mitogen stimulation. Recently, we cloned p21Sdi1, a potent inhibitor of DNA synthesis and Cdk2 kinase activity, from a senescent cell cDNA library and demonstrated that it was expressed at significantly higher levels in senescent cells than actively proliferating cells. In contrast to actively dividing cells, mitogen-stimulated senescent cells do not down-regulate the expression of p21Sdi1and do not express late G1 phase gene products that are required for entry into S phase. We suggest that the inability of mitogen-stimulated senescent cells to down-regulate p21Sdi1levels contributes to the resulting lack of late Gi gene expression and failure to traverse the G1/S phase boundary.

Human adrenocortical cell xenotransplantation: model of cotransplantation of human adrenocortical cells and 3T3 cells in scid mice to form vascularized functional tissue and prevent adrenal insufficiency

To establish an experimental model for replacement of endocrine organ function by xenotransplantation, human adrenocortical cells from postnatal donors were transplanted beneath the kidney capsule of adrenalectomized scidmice together with mitomycin C-treated 3T3 cells that secrete FGF. Adrenocortical cells from seven donors, male and female, ranging from 6-50 years of age, were used. 12 of 13 animals survived > 16 days following surgery. After 50 days they were sacrified to allow assessment of the histology and ultrastructure of tissue formed from the transplanted cells. Only 1 of 23 adrenalectomized sham-operated animals survived > 16 days. In all surviving animals, vascularized adrenocortical tissue formed at the site of transplantation. Cortisol, the normal human glucocorticoid, was present in the plasma of these animals, replacing corticosterone, the mouse glucocorticoid. Some animals, but not most, had measurable aldosterone. The tissue formed from the transplanted cells showed histological and ultrastructural features of normal adrenal cortex. Mitochondria had tubulo-vesicular cristae and there were prominent microvilli between cells. Tissues had a well-developed vasculature, sometimes with large sinusoidal vessels. Proliferation in the transplant tissues was very low. These results show that tissue formed from transplanted human adrenocortical cells is able to replace the essential functions of the adrenal gland in scid mice. This demonstrates that transplanted human endocrine cells can functionally replace a surgically removed endocrine organ in a host animal.

Telomere shortening and decline in replicative potential as a function of donor age in human adrenocortical cells

Telomere shortening is the cause of replicative senescence of mammalian cells in culture and may be a cause of cellular aging in vivo. Some tissues clearly show telomere shortening during aging in humans, but the relationship between replication history and telomere length is obscured by complex relationships between stem cells and more differentiated cell types. Previous experiments on the adrenal cortex and human adrenocortical cells in culture indicate that the proliferative biology of this tissue is relatively simple; cell division occurs continuously throughout life, without evidence for a distinct stem cell compartment. In this tissue we investigated the relationship between telomere biology and replicative senescence by measuring replicative capacity and telomere length as a function of donor age. Cells cultured from adrenal tissue from donors of different ages showed a strong age-related decline in total replicative capacity, falling from about 50 population doublings for fetal cells to an almost total lack of division in culture for cells from older donors. Telomere restriction fragment (TRF) length was analyzed in the same sets of cells and decreased from a value of about 12 kb in fetal cells to approximately 7 kb in cells from older donors. The latter value is consistent with that in fibroblasts which have reached replicative senescence. Furthermore, there was a good correlation in individual donor samples between TRF length and replicative capacity in culture. To confirm the relationship between telomere length, telomerase, and replicative capacity, we measured telomere length in cells before and after infection with a retrovirus encoding hTERT, the catalytic component of human telomerase. The adult adrenal cortex does not have telomerase activity; cells after transduction with the hTERT retrovirus had high telomerase activity. Whereas control cells underwent a replication-dependent shortening in telomeres during long-term growth in culture, hTERT-modified cells maintained telomere length and are probably immortalized. Symmetric cell division in human adrenocortical cells, occurring slowly over the life span, is associated with progressive telomere shortening and may result in proliferative defects in vivo in old age, which could partly account for the age-related changes in the structure and function of the human adrenal cortex.

Differentially expressed genes in zona reticularis cells of the human adrenal cortex

The zona reticularis (ZR) cell in the human adrenal cortex is responsible for the secretion of dehydroepiandrosterone, but its biology, origin, and putative decrease in number during aging are poorly understood. In the present experiments, we investigated to what extent ZR and zona fasciculata (ZF) cells differ in patterns of gene expression. Both cell types were purified by microdissection from adult adrenal cortex specimens. After a brief period in culture, RNA was harvested from the cells and used to prepare radioactively labeled probes following amplification by PCR. Probes were used in hybridizations of arrays of cDNAs on nylon membranes (PCR products or plasmids obtained from an adrenal cDNA library). Analysis of hybridization intensities showed that 17 of the 750 genes studied differed in expression by more than 2-fold. Several genes expressed at higher levels in ZR cells encode components of the major histocompatibility complex or enzymes involved in peroxide metabolism. Members of the tubulin gene family were expressed at higher levels in ZF cells. Differential expression of four of the genes was confirmed by Northern blotting. These differences show that although ZR and ZF cells are similar in gene expression, ZR cells have a gene expression pattern related to the unique biology of this cell type.

Subcutaneous transplantation of bovine and human adrenocortical cells in collagen gel in scid mice

Adrenocortical cells of bovine origin and of adult and fetal human origin were transplanted subcutaneously (s.c.) in scid mice after being embedded in collagen gel. In this site the cells survived, became vascularized by invasion of host endothelial cells, and secreted steroids into the circulation. The animals' own adrenal glands were removed at the time of cell transplantation. Steroids secreted by the transplants replaced the essential functions of the animals' own adrenal glands. Adrenalectomized animals without transplanted cells died after several days, but most animals with transplanted bovine or adult human adrenocortical cells survived; fewer animals survived with transplanted fetal human adrenocortical cells. The histology of the tissues formed from transplanted cells resembled that of the normal adrenal cortex. A few proliferating cells were observed in tissue from bovine or adult human cells; there was a greater percentage of dividing cells in tissue derived from fetal cells. Subcutaneous transplantation of bovine or human primary adrenocortical cells in collagen provides a model for the study of the physiology, cell biology, and molecular biology of adrenocortical cells in a three-dimensional vascularized tissue structure in a host animal.

Induction of steroidogenic enzyme genes by insulin and IGF-I in cultured adult human adrenocortical cells

Insulin and the insulin-like growth factors (IGFs) have multiple role in gene expression in steroidogenic cells. We investigated the regulation of steroidogenic enzyme gene expression by insulin and IGF-I in primary cultures of human adrenocortical cells from donors of ages 19-77 years. The effects of insulin and IGF-I observed here were independent of age and sex of the donor. After 5 days in serum-containing medium, cultures were exposed to insulin or IGF-I together with cyclic AMP analogs or ACTH in serum-free defined medium. Insulin and IGF-I at physiological concentrations increased mRNA levels for 17 alpha-hydroxylase and type II 3 beta-hydroxysteroid dehydrogenase (3 beta-HSD) in the absence of cyclic AMP or ACTH. They had lesser effects on 21-hydroxylase and cholesterol side-chain cleavage enzyme mRNA levels and were3 without effect on 11 beta-hydroxylase mRNA. All steroidogenic enzyme mRNAs were strongly increased by cyclic AMP or ACTH, and this increase was potentiated by insulin or IGF-I. These effects of insulin and IGF-I were accompanied by decreases in the ratio of dehydroepiandrosterone/cortisol synthesized from pregnenolone by the cultures. Induction of steroidogenic enzyme genes in adult human adrenocortical cells by insulin and IGF-I is unlikely to occur by means of a cyclic AMP-dependent mechanism. These data increase the evidence for an important regulation of steroidogenesis by these hormones.

Replicative senescence: implications for in vivo aging and tumor suppression

Normal cells have limited proliferative potential in culture, a fact that has been the basis of their use as a model for replicative senescence for many years. Recent molecular analyses have identified numerous changes in gene expression that occur as cells become senescent, and the results indicate that multiple levels of control contribute to the irreversible growth arrest. These include repression of growth stimulatory genes, overexpression of growth inhibitory genes, and interference with downstream pathways. Studies with cell types other than fibroblasts will better define the role of cell senescence in the aging process and in tumorigenesis.

Expression of p21(WAF1/CIP1/SDI1) and p53 in apoptotic cells in the adrenal cortex and induction by ischemia/reperfusion injury

p21(WAF1/CIP1/SDI1), an inhibitor of cyclin-dependent kinases, is expressed at varying levels in human adrenal glands removed during surgery or organ recovery. In glands with p21 mRNA, nuclear p21 immunoreactivity, which was occasionally extensive, colocalized with p53 immunoreactivity and DNA damage, as evidenced by in situ end-labeling. Many cells showed morphological features of apoptosis when observed by fluorescent DNA dye staining and electron microscopy. This pattern was also associated with high levels of cytoplasmic heat shock protein 70. To address the question of the origin of p21 expression in some human adrenal glands, rat adrenal glands were subjected to 30 min of ischemia followed by 8 h of reperfusion. Cells with nuclear p21 and p53 appeared in the adrenal cortex together with DNA damage detected by in situ end-labeling. Nuclear p21 immunoreactivity was also produced in adrenal tissue fragments incubated at 37 degrees C in vitro. However, in this case, p21 expression was confined to the cut edge of the tissue. In contrast, p21 in human adrenal glands, as in ischemic rat glands, was within the inner regions of the cortex, supporting an origin of the protein in vivo rather than postmortem. The p53/p21 pathway of reaction to cellular injury, potentially leading to apoptosis, may play a role in tissue damage such as that resulting from ischemia/reperfusion. In the human adrenal cortex this process may be a precursor of adrenal failure.

Studies demonstrating the complexity of regulation and action of the growth inhibitory gene SDI1

The identification of the DNA synthesis inhibitory gene SDI1 by investigators studying cell senescence, tumor suppression, cell cycle control and differentiation suggest a key regulatory role for this gene. To better understand the growth regulatory activity of this gene we proceeded with the experiments described here. The data demonstrate that SDI1 is an important downstream effector of p53, but here we report that it can also cause inhibition of DNA synthesis in several immortal human cell lines, independent of p53 or Rb status. Levels of SDI1 mRNA expression in immortal cells are consistently much lower than that of normal cells, indicating that immortalization is not compatible with high expression of SDI1. These results highlight the complex nature of regulation of this gene and its mode of action.

The NCI-H295 cell line: a pluripotent model for human adrenocortical studies

The human adrenal cortex is a complex endocrine organ that secretes mineralocorticoids, glucocorticoids and adrenal androgens. These steroids arise from morphologically and biochemically distinct zones of the adrenal gland. Studying secretion of these distinct steroid hormones has, in the past, required the isolation of cells from each of the adrenocortical zones. Indeed, the lack of a human adrenocortical cell line retaining the ability to produce any of the major adrenal steroid products has slowed studies on normal and abnormal adrenal function. This obstacle has now been largely overcome with the availability of H295 cells, which represents the first adrenocortical cell line to maintain the ability, under specified conditions, to produce all the adrenocortical steroids (i.e., mineralocorticoids, glucocorticoids, and adrenal androgens). Thus, H295 cells appear to act as pluripotent adrenocortical cells capable of being directed to produce each of the zone-specific steroids. The H295 cell line should prove to be of value in studying the molecular and biochemical mechanisms controlling adrenal steroidogenesis.

Interactions between TGF-beta and adrenocorticotropin in growth regulation of human adrenal fetal zone cells

The factors that regulate growth and function of the human adrenal gland during intrauterine development and thereafter are ill defined. Whereas others have reported that adrenocorticotropic hormone (ACTH) augments the inhibitory effect of transforming growth factor-beta (TGF-beta) on growth of fetal zone (FZ) cells of the human fetal adrenal, we recently found that ACTH interferes with TGF-beta's inhibition of growth of fetal adrenal neocortex cells. In this study we sought to assess independently the effects of TGF-beta in the absence and presence of ACTH on growth of FZ cells. TGF-beta, in a time- and dose-dependent manner, inhibited growth (i.e., [3H]thymidine incorporation) of FZ cells. ACTH (Cortrosyn), at 90 pM to 90 nM, was found to interfere with the TGF-beta inhibition of FZ growth. ACTH 1-24 and human ACTH 1-39, both from Sigma Chemical, also were found to blunt the response of FZ cells to TGF-beta. Growth inhibition due to TGF-beta action and the reversal by ACTH of TGF-beta effects on FZ cell growth were confirmed by the results of immunohistochemical analyses of 5'-bromo-2'-deoxyuridine incorporation into nuclei of FZ cells and by indirect evaluations of cell numbers. Both forskolin (10 microM) and dibutyryl adenosine 3',5'-cyclic monophosphate (1 mM), but not phorbol 12-myristate 13-acetate (1 or 100 mM), were able to mimic ACTH actions in blunting the inhibitory effects of TGF-beta on DNA synthesis. We conclude that ACTH, possibly via activation of adenylate cyclase, interferes with, rather than augments, the growth-inhibitory effect of TGF-beta on FZ cell growth.

Changes in gene expression during senescence of adrenocortical cells in culture

Bovine adrenocortical cells undergo a process in which expression of steroid hydroxylases is lost progressively as a function of population doubling level (PDL) in culture. Each cytochrome P450 shows a characteristic rate of loss of expression as a function of PDL (in order of rates of loss: CYP11B >CYP21 >CYP17 >CYP11A). CYP11B and CYP21 require insulin-like growth factor I as well as cyclic AMP; these are the only factors required for induction in the primary culture. Middle- and later passage cells do not express CYP11B and CYP21 under the same conditions, but will do so when cells are grown in extracellular matrix Matrigel. In late-passage cells neither CYP17, CYP21, nor CYP11B are expressed, even in the presence of Matrigel; only CYP11A is expressed in late-passage cultures. When the different environmental factors required for induction of CYP11B and CYP21 are taken into account, induction of these genes disappears with the same kinetics as previously shown for CYP17 as a function of PDL. The primary cause of the loss of expression of these genes is likely to be a phenotypic switching event similar to that previously demonstrated for CYP17 by in situ hybridization. The mechanism of phenotypic switching is unknown. However, one HpaII site at -2.3 kb of CYP17 was methylated in the bovine adrenal cortex in vivo but showed rapid and complete demethylation when adrenocortical cells were placed in culture. This indicates a unique, reproducible, environmentally determined change in methylation, with as yet undetermined consequences. However, data from reporter constructs suggest that phenotypic switching does not result from a simple loss of regulatory factors that act within 2.5 kb of the promoter. Previous data suggested that SV40 T antigen may affect phenotypic switching, and thus that SV40 may be useful for the derivation of functional adrenocortical cell lines. Adaptation of methods previously used for bovine cells to human adrenocortical cells to produce SV40 T antigen-transfected clones yielded data indicating preservation of essential aspects of the human adrenocortical cell differentiated phenotype.

Changes in gene expression and DNA methylation in adrenocortical cells senescing in culture

Recent experiments in cultured bovine adrenocortical cells show that the previously observed phenotypic switching of CYP17 (steroid 17 alpha-hydroxylase) expression is preceded at a much earlier time by changes in methylation in the CYP17 5' flanking region. Two CpG sites that are methylated in the adrenal cortex in vivo were observed to undergo rapid demethylation when adrenocortical cells were placed in culture. Two adjacent CpG sites that are also methylated in vivo did not demethylate; these two sites are completely nonmethylated in fibroblasts. All CpG sites downstream, in the promoter or coding region, are always methylated in all tissues and in bovine adrenocortical cells even after many population doublings in culture. In contrast to the specific and rapid demethylation of sites in CYP17, satellite I shows a slower and apparently random loss of methylation that extends over the entire replicative life span. These changes in methylation provide examples of genetic instability in cells that undergo senescence in culture. Future experiments will focus on the relationship of these events to the phenotypic switching process.

Regulation of steroid hydroxylases in normal and SV40 T antigen-transfected bovine adrenocortical cells in long-term culture

Over long periods of growth in culture, bovine adrenocortical cells lose the ability to express the steroid hydroxylase genes. For 17 alpha-hydroxylase, cells show a stochastic pattern of phenotypic switching from a state in which they express this gene in response to cyclic AMP to a state in which the gene is no longer inducible. Introducing SV40 T antigen into bovine adrenocortical cells greatly increases their replicative potential; steroid hydroxylase expression in these clones resembles that of the precursor cells before transfection. The other steroid hydroxylases (21-hydroxylase and 11 beta-hydroxylase) appear to undergo phenotypic switching like 17 alpha-hydroxylase. The loss of expression of these genes appears to be more rapid, but there are differences in the requirements of 21-hydroxylase and 11 beta-hydroxylase versus 17 alpha-hydroxylase for induction by cyclic AMP; additionally, growth of cells in extracellular matrix Matrigel was required for expression of 21-hydroxylase and 11 beta-hydroxylase in long-term cultures of either normal or SV40 T antigen-transfected cells. Understanding the molecular basis for the phenotypic switching of steroid hydroxylases that occurs in bovine adrenocortical cells may elucidate mechanisms for cellular senescence and for maintenance of tissue-specific functions during long-term growth in culture.

Density separation of rat adrenocortical cells: morphology, steroidogenesis, and P-450scc expression in primary culture

We have developed a method that separates rat adrenocortical cells by density into populations which retain zone specific properties in primary culture. Two different parenchymal populations were obtained and designated 2FASC (1.034 g/ml, 18.0 microns cell diameter) and 7GLOM (1.069 g/ml, 11.7 microns cell diameter). In freshly isolated cell suspensions the physical characteristics and differential steroidogenic responses to adrenocorticotropin and angiotensin II suggested that 2FASC cells originated predominantly from the zona fasciculata and 7GLOM cells from the zona glomerulosa. In primary culture (Dulbecco's Modified Eagle's Medium-F12 medium with 15% horse serum and 2.5% fetal bovine serum) the two populations exhibited different morphologies. 2FASC cells retained lipid and formed cohesive epithelial monolayers that remained stationary for 3 wk. 7GLOM cells were initially epithelial but rapidly lost lipid, spread, and assumed fibroblastic shapes. Both cell types were positive for the cholesterol side-chain cleavage cytochrome P-450 by immunofluorescence. Therefore, the morphologic changes seen in 7GLOM cultures were due to modulation, not fibroblastic overgrowth. This phenotypic plasticity may reflect the mesodermal origin of the adrenal cortex, and the subcapsular location of 7GLOM cells in vivo. In contrast, cells such as 2FASC which are located deeper in the cortex seem to have a more restricted, fully committed parenchymal phenotype.

Cytotoxic effects of expression of human superoxide dismutase in bovine adrenocortical cells

Oxygen radicals and the cellular antioxidant enzymes may play a role in cellular senescence. We studied the feasibility of altering oxygen radical metabolism in a normal differentiated cell that undergoes senescence in culture by transfection of an expression vector containing human CuZn-SOD cDNA. Plasmid pRSV2-cSOD was constructed to contain the cDNA for human CuZn-SOD under the regulation of the Rous sarcoma virus long terminal repeat. Early passage cultures of bovine adrenocortical cells were cotransfected with pRSV2-cSOD and a plasmid (pSV3neo) allowing initial selection and continued growth of transfectants. Three passages after isolation of the polyclonal population, as cells grew to confluence, cultures showed focal cell death that spread outward to affect neighboring cells, so that by 72 h most cells had detached from the culture dish. Long term growth of the polyclonal population of transfectants without extensive cell death was achieved by continuous maintenance of low cell density during growth. Southern blot analysis of DNA from the pooled polyclonal population of transfected cells showed the presence of the expected 625 bp band from human CuZn-SOD. However, the intensity of this band indicated that only a minority of cells in the population had integrated the SOD plasmid, and DNA isolated from cells after 25 passages at low cell density showed plasmid sequences only of an altered form, suggesting that cells containing intact human SOD cDNA had been selectively lost from the population. When early passage low density transfectants were allowed to grow back to high cell density, cell death foci were again observed. Additionally, cell fusion with the formation of giant cells with massive multinucleation was observed by fluorescence microscopy after staining cultures with a DNA binding dye. In later stages of this process, the large nuclear mass in such a giant cell became fragmented as the cell detached from the dish and formed the center of a focus of cell death in the surrounding cells. Because cell death prevented the growth of large numbers of transfected cells, it was not possible to demonstrate the involvement of CuZn-SOD in the cytotoxic effect by direct means, but the control plasmid without the CuZn-SOD cDNA insert had no cytotoxic effect. Thus, the introduction of a vector for human CuZn-SOD in a normal differentiated cell caused a cytotoxic effect involving cell death, cell fusion, and nuclear fragmentation.

Cyclic AMP-mediated cytoskeletal effects in adrenal cells are modified by serum, insulin, insulin-like growth factor-I, and an antibody against urokinase plasminogen activator

In adrenocortical cells in culture, increased intracellular cyclic AMP resulting from exposure to agents such as ACTH and cholera toxin causes a change in cell morphology termed 'retraction' or 'rounding'. The breakdown of actin-containing stress fibers in rounding suggested a role for microfilaments in steroidogenesis. Previously, we showed that cultured bovine adrenal cells under standard conditions (medium with 10% fetal bovine serum) do not round in response to intracellular cyclic AMP. Here, we show that these cells do round in defined, serum-free medium. Rounding was maximal within 1 h of addition of 1 nM cholera toxin and after 10 h most cells remained rounded. Cycloheximide at 100 micrograms/ml did not inhibit the response to cholera toxin. The rounding response was abolished when 10% fetal bovine serum, horse serum, or ether-extracted fetal bovine serum was included in the medium. The inhibitory effect of serum was not mimicked by growth factors with the exception that insulin and insulin-like growth factor-I (IGF-I), while not preventing rounding, accelerated the return of cells to a flattened morphology. A monoclonal antibody against urokinase plasminogen activator completely prevented rounding whereas a monoclonal antibody against tissue plasminogen activator had only a slight effect. Fluorescence visualization of F-actin with N-(7-nitrobenz-2-oxa-1,3-diazol-4-yl)-phallacidin showed that rounding in cultured bovine adrenocortical cells resembles that defined earlier for human and rat adrenocortical cells and includes depolymerization of actin microfilaments. These cytoskeletal changes in adrenal cells are unlikely to play a role in steroidogenesis; however, they may be involved in tissue remodeling occurring as part of the indirect mitogenic effects of ACTH.

Cellular senescence involves stochastic processes causing loss of expression of differentiated function genes: visualization by in situ hybridization for steroid 17 alpha-hydroxylase in bovine adrenocortical cells

When grown for long periods in culture, bovine adrenocortical cells lose the expression of a differentiated function gene, steroid 17 alpha-hydroxylase. Previously, we documented a decline in 17 alpha-hydroxylase mRNA with increasing culture passage level after induction with cyclic AMP (P. J. Hornsby et al., 1987, Proc. Natl. Acad. Sci. USA 84, 1580). We used in situ hybridization to investigate the loss of expression of this gene during cellular senescence at an individual cell level. In primary cultures, cells were uniformly positive for hybridization with cDNA for 17 alpha-hydroxylase after cyclic AMP induction. After two passages, cultures comprised a mixture of hybridizing and nonhybridizing cells. Cells appeared either to hybridize at a level comparable to that in primary cultures or to be nonhybridizing. When in situ hybridization was combined with immunofluorescence, cells positive for immunofluorescence were also positive for hybridization. Senescing mass cultures showed decreasing numbers of positive cells, and after 30 passages cultures comprised entirely nonhybridizing cells. Thus, the previously observed decline in overall 17 alpha-hydroxylase mRNA levels results from a decline in the fraction of expressing cells in the culture, and the rate of loss of expressing cells is in agreement with the rate of loss of total 17 alpha-hydroxylase mRNA. Primary clones, even when isolated at an early stage of clonal expansion, had mixtures of subclones of hybridizing and nonhybridizing cells. On recloning, hybridizing subclones usually produced uniformly nonhybridizing sub-subclones. Some subclones within primary clones had a morphology associated with replicative senescence (flattened cells with sparse intercellular contacts), yet had high numbers of hybridizing cells. We conclude that, in both mass and clonal populations, cells initially expressing 17 alpha-hydroxylase rapidly give rise to clones of nonexpressing cells. Such cells are continually derived by a stochastic process from cells originally expressing the gene.

Replicative senescence and differentiated gene expression in cultured adrenocortical cells

We have used a differentiated endocrine cell type, the adrenocortical cell, to investigate the interrelationship of senescence and differentiation, the effects of the environment on differentiated gene expression, and the interrelationship of differentiated gene expression and proliferation. In bovine adrenocortical cells, expression of some differentiated functions is maintained to very late points in the replicative life span, whereas expression of others is lost at various times prior to senescence. There is clonal variation in the rate and extent of loss of functions. For steroid 17 alpha-hydroxylase, in situ hybridization shows that the observed decline in induction of 17 alpha-hydroxylase mRNA during senescence results from a decline in the fraction of cells expression the gene. Descendants of expressing cells in the primary cell population randomly become nonexpressing. Among clones there is a correlation between the fraction of cells expressing the gene and remaining replicative potential, although several experiments show no direct mechanism linking replicative senescence and 17 alpha-hydroxylase expression. Transfection with SV40 early region genes also dissociates the decline in growth and the change in 17 alpha-hydroxylase expression. SV40 T antigen selectively affects growth; expression of 17 alpha-hydroxylase is stabilized either in the on state, when cells are transfected early in the culture life span, or in the off state, when senescent cells are transfected. Thus, although the switching off of 17 alpha-hydroxylase expression and the loss of replicative potential are independent events, the switching process requires DNA replication. Because the switch is irreversible, changes in replicative potential occurring after the switch-off event do not affect the state of expression of the switched-off gene. Changes in differentiated cell properties and changes in replicative potential may be two facets of a general phenomenon of stochastic changes in gene expression in normal cells during senescence.

Regulation of steroid 17 alpha-hydroxylase in adrenocortical cells in culture

The regulation of steroid 17 alpha-hydroxylase in human and bovine adrenocortical cells in culture is reviewed. It is shown that (i) the long-term growth and cloning of normal human fetal adrenocortical cells in culture is feasible; (ii) the phorbol ester 12-O-tetradecanoyl phorbol 13-acetate (TPA) is an effective mitogen in both clonal cultures and non-clonal early cultures of human adrenocortical cells, and shows a selective action in promoting adrenocortical cell growth and inhibiting fibroblast growth; (iii) the key steroidogenic enzymes, 17 alpha-hydroxylase and 3 beta-hydroxysteroid dehydrogenase (3 beta-HSD), are under dual regulation by the cyclic AMP and protein kinase C second messenger systems; (iv) for 17 alpha-hydroxylase, this dual regulation is mediated by changes in 17 alpha-hydroxylase mRNA levels; (v) bovine adrenocortical cells can be transfected with SV40 T antigen, producing lines with elevated differentiated functions, including stabilized high expression of 17 alpha-hydroxylase; (vi) human adrenocortical cells can also be transfected with SV40 large T antigen, giving rise to functional cell lines which may be useful in future studies of 17 alpha-hydroxylase regulation.

Cellular senescence involves stochastic processes causing loss of expression of differentiated function genes: transfection with SV40 as a means for dissociating effects of senescence on growth and on differentiated function gene expression

In the accompanying work we demonstrated that the decline in expression of steroid 17 alpha-hydroxylase in mass cultures and clones of adrenocortical cells is the result of a stochastic switching process which yields mixtures of expressing and nonexpressing cells. There is an apparent positive correlation between the replicative potential of adrenocortical cell cultures and the number of cells in the culture that can express 17 alpha-hydroxylase. We investigated this by extending the cells' replicative potential by transfecting them with cloned SV40 virus. Cells from a senescent subclone, with very limited remaining replicative potential, were transfected. The cell population showed a progressive increase in growth rate and gave rise to a line of cells that expressed T antigen and which was apparently immortalized. Induction of mRNA for 17 alpha-hydroxylase by cyclic AMP was absent in this line of cells, as it was in the senescent cells prior to transfection. The cells remained responsive to gene induction by cyclic AMP as evidenced by increases in mRNA and activity for cholesterol side-chain cleavage. The absence of 17 alpha-hydroxylase expression in this line was not the result of interference by SV40 T antigen. When early passage cells were transfected with pSV3neo, which contains the early region of SV40 and neo, and were selected with G418, SV40 T antigen-expressing lines were derived which showed high levels of expression of 17 alpha-hydroxylase after induction with cyclic AMP. These cells maintained high levels of expression of 17 alpha-hydroxylase through four successive recloning events, over a period of replication much longer than that achievable by nontransfected cells. Thus, transfection by SV40 can be used to dissociate effects of senescence on growth and differentiated gene expression. T antigen expression selectively affects growth, but preserves the state of expression of a differentiated function gene as it was prior to transfection.