Sequences essential for activity of the thyroid hormone responsive transcription stimulatory element of the rat growth hormone gene.

Thyroid hormone dependent transcription stimulatory and inhibitory elements exist at the 5'-end of the rat GH (rGH) gene (TSE and TIE, respectively). In this study, the location of the sequences essential for TSE activity was examined using stably transfected GC cells. Because the TIE may influence TSE activity, we investigated TSE activity both on the rGH promoter, in the presence of the TIE, and on the viral thymidine kinase promoter, with the TIE deleted. The results of these studies indicate that the minimum sequences essential for TSE activity exist between positions -194 and -169 of the rGH gene.

Positive and negative thyroid hormone response elements are composed of strong and weak half-sites 10 nucleotides in length.

The steroid-thyroid hormone receptors bind to imperfect repeats of two or more half-sites. It is generally accepted that a T3 response element (TRE) half-site consists of a six-nucleotide core motif (5'-AGGT(C/A)A-3'). It is less widely appreciated that the nucleotides flanking this core motif also have a major influence on the affinity of T3 receptor (TR) for its response element. We analyzed TR-DNA interactions under conditions in which the affinity of receptor monomers for individual TRE half-sites of the rat GH (rGH) gene was measured. These studies avoided the effects of half-site spacing and orientation on receptor binding. Variations in the nucleotides flanking the core sequence can modulate receptor binding by more than 15-fold. Systematic mutational analysis of TRE half-site structure demonstrated that at least two nucleotides flanking either side of the half-site core motif strongly influence TR binding affinity and activity, indicating that half-sites are approximately 10 nucleotides long. Thus, the half-sites of most TREs overlap, and mutations in one half-site may affect the activity of its partner. The TRE half-site sequence 5'-CTGAGGTAACG-3' was bound with highest affinity by TRs. The negatively T3-responsive promoter of the rGH gene was used to investigate the functional significance of the nucleotides flanking the core motif in vivo. A promoter consisting of only 22 rGH nucleotides, containing two functional TRE half-sites which overlap the rGH TATA box, directed T3-inhibited transcription. Mutation of nucleotides flanking the core sequence of the weaker half-site dramatically reduced the activity of the element, demonstrating that the flanking sequences of the half-sites can profoundly affect TRE activity.

Cyclic regulation of growth hormone gene transcription in vivo and in vitro.

The in vivo thyroid and glucocorticoid hormone regulation of GH gene transcription was compared with that found in cultured GH rat pituitary tumor cells. The GH cell lines have been widely used to study GH gene expression, but their relevance to the in vivo regulation of the gene has not been well established. The in vivo studies described here utilized rats that were both thyroparathyroidectomized and adrenalectomized to remove the organ sources of these hormones. The in vitro studies described utilized GC cells hormonally deinduced in medium lacking the hormones. Continuous administration of glucocorticoid or thyroid hormones to either system induced multiple cycles of GH transcriptional activation and deactivation. These cycles were accompanied by cycles of increasing and decreasing GH messenger RNA. In both systems, a brief transcription cycle occurred within hours of thyroid or glucocorticoid hormone addition, and a second broad occurred between 3 and 11 days later. These cycles were independent of changes in receptor levels. The similarities in the responses found in vivo and in cell culture suggest that the molecular mechanisms regulating expression of the GH gene appropriately function in GC cells, despite their transformed phenotype and prolonged maintenance in culture. Thus, these cell lines appear to be appropriate model systems for studies of thyroid and glucocorticoid hormone action.

Age-related decrease of growth hormone and prolactin gene expression in the mouse pituitary.

The effects of aging on pituitary GH, PRL, and alpha-tubulin messenger RNA (mRNA) levels were measured in 3-, 12-, and 27-month-old male C57BL/6J mice by dot-blot hybridization. The amount of GH and PRL mRNA in the pituitary deceased dramatically with age. However, total poly(A+) RNA (mRNA), as measured by hybridization with radioactively labeled oligo-(dT), was not altered during aging. In addition, there were no age-related changes in the level of alpha-tubulin mRNA. Thus, the effects of aging on GH and PRL mRNA levels are specific; the levels of the majority of cellular mRNAs are not altered with age. GH and mRNA levels decreased 35% between 3 and 12 months (P less than 0.05) and a total of 75% after 27 months (P less than 0.01). PRL mRNA levels decreased 65% between 12 and 27 months (P less than 0.01), although there was no significant decrease before 12 months. Whereas T3 is the most potent regulator of GH gene expression, we did not detect any significant age-related change in serum T3 levels. These results suggest that factors other than T3 play a role in the age-related decline in GH and PRL gene expression.

Structure and regulation of the mouse GRP78 (BiP) promoter by glucose and calcium ionophore.

Dietary calorie restriction, also termed energy restriction, increases mean and maximum life span, reduces the incidence of tumors and increases the mean age of onset of diseases and tumors in every animal tested. Because life-span is genetically determined, we are studying the mechanisms by which energy restriction regulates the expression of genes. We found that energy restriction reduces hepatic glucose-regulated protein-78 (GRP78) and protein-94 mRNA levels by 2-3-fold in mice [Spindler et al., J. Nutr. 20 (1990) 1412-1417]. To investigate this down-regulation, we have cloned the mouse GRP78 promoter (pGRP78) and studied its regulation by glucose. The mouse pGRP78 and the previously cloned rat promoter mediate responsiveness to glucose deprivation, as well as to the calcium ionophore A23187. These studies are the first demonstration that cis-elements in the pGRP78 mediate responsiveness to glucose deprivation.

Glucose regulation of GRP78 gene expression.

The endoplasmic reticulum chaperone glucose-regulated protein 78 (GRP78) is essential for the proper glycosylation, folding and assembly of many membrane bound and secreted proteins. GRP78 mRNA is well known to be induced in cultured cells by lowering medium glucose concentrations from 4.5 to 0 mg/ml. Here we report a study designed to determine the effects of intermediate concentrations of glucose on GRP78 mRNA abundance. Progressive reduction in culture medium glucose from 4.5 to 1.0 mg/ml progressively reduced GRP78 mRNA to approximately 30% of the initial level. Induction of GRP78 mRNA by glucose starvation was observed in medium containing less than 1 mg/ml glucose. Determination of the amount of glucose consumed in these cultures showed that reduction of glucose concentrations led first to repression of GRP78 mRNA abundance, followed by induction of the mRNA only when glucose is nearly exhausted. Caloric restriction in mice both reduces fasting and mean 24 h glucose blood concentrations and GRP78 mRNA abundance in the liver. Thus, it is possible that negative regulation of GRP78 mRNA in the liver is due directly to reduced blood glucose concentrations.

Age-related down regulation of hepatic cytochrome P1-450, P3-450, catalase and CuZn-superoxide dismutase RNA.

The influence of age on liver gene expression was investigated in two strains of H-2 congenic mice. In B10.RIII mice (H-2r), basal P1- and P3-450 RNA levels progressively decreased 65 and 95%, respectively, between 4 and 28 months of age (P less than or equal to 0.05). Polyaromatic hydrocarbon (PAH) induced P1- and P3-450 RNA levels decreased about 50% during this time (P less than or equal to 0.05). In contrast, in C57BL/10 mice (H-2b) little or no change was detected in basal or induced P1- or P3-450 RNA levels. CuZn-superoxide dismutase RNA decreased 80 to 90% between 4 and 9 months of age in B10.RIII mice, while a quantitatively smaller decrease of 50 to 65% was found in C57BL/10 mice (P less than or equal to 0.05). Catalase RNA decreased approximately 80% between 4 and 9 months of age in B10.RIII mice, and a similar decrease was found in C57BL/10 mice. Down regulation of these genes may explain the reduced activities of the cognate hepatic enzymes, and reduced xenobiotic metabolism found in older animals.

Caloric restriction alters the feeding response of key metabolic enzyme genes.

Differential 'fuel usage' has been proposed as a mechanism for life-span extension by caloric restriction (CR). Here, we report the effects of CR, initiated after weaning, on metabolic enzyme gene expression 0, 1.5, 5, and 12 h after feeding of 24-month-old mice. Plasma glucose and insulin were reduced by approximately 20 and 80%. Therefore, apparent insulin sensitivity, as judged by the glucose to insulin ratio, increased 3.3-fold in CR mice. Phosphoenolpyruvate carboxykinase mRNA and activity were transiently reduced 1.5 h after feeding, but were 20-100% higher in CR mice at other times. Glucose-6-phosphatase mRNA was induced in CR mice and repressed in control mice before, and for 5 h following feeding. Feeding transiently induced glucokinase mRNA fourfold in control mice, but only slightly in CR mice. Pyruvate kinase and pyruvate dehydrogenase activities were reduced approximately 50% in CR mice at most times. Feeding induced glutaminase mRNA, and carbamyl phosphate synthetase I and glutamine synthase activity (and mRNA). They were each approximately twofold or higher in CR mice. These results indicate that in mice, CR maintains higher rates of gluconeogenesis and protein catabolism, even in the hours after feeding. The data are consistent with the idea that CR continuously promotes the turnover and replacement of extrahepatic proteins.

Promoter independent down-regulation of the firefly luciferase gene by T3 and T3 receptor in CV1 cells.

We report that the activity of the firefly luciferase (LUC) reporter gene is down-regulated by T3 and T3 receptor (TR) in the CV1 mammalian cell line, which is widely used for studies of TR action. Repression was highly reproducible, T3 and TR dependent, promoter independent, and observed regardless of whether an internal control for transfection efficiency was used. Cotransfections with normal and mutant TRs indicate that the negative T3 response is mediated by sequences within the LUC gene coding region, and is not due to the interaction of TR with a limiting transcription factor. Negative regulation of the LUC reporter was overcome by a strong, cis-linked T3 response element (TRE), but continued in the presence of a TRE of moderate strength. The results described here demonstrate that conclusions drawn from studies of TRE structure and activity performed using the LUC reporter in CV1 cells should be interpreted with caution.

Calorie restriction enhances the expression of key metabolic enzymes associated with protein renewal during aging.

Our studies show that dietary caloric restriction (CR) alters the expression of key metabolic enzymes in a manner consistent with an increased rate of extrahepatic protein turnover and renewal during aging. Of the key hepatic gluconeogenic enzyme genes affected by CR, glucose 6-phosphatase mRNA increased 1.7- and 2.3-fold in young and old CR mice. Phosphoenolpyruvate carboxykinase mRNA increased 2-fold in young mice, and its mRNA and activity increased 2.5- and 1.7-fold in old mice. These changes indicate that CR enhances the enzymatic capacity for gluconeogenesis. The carbon required for gluconeogenesis appears to be generated from peripheral protein turnover. Muscle glutamine synthetase mRNA increased 1.3- and 2.1-fold in young and old CR mice, suggesting increased disposal of nitrogen and carbon derived from protein catabolism for energy. mRNA for the key liver nitrogen disposal enzymes glutaminase, carbamyl phosphate synthase I, and tyrosine aminotransferase were increased by 2.4-, 1.8-, and 1.8-fold in CR mice. Consistent with increased hepatic nitrogen disposal, hepatic glutamine synthetase mRNA and activity were each decreased about 40% in CR mice. Together, these and our other published data suggest that CR enhances and maintains protein turnover, and thus protein renewal, into old age. These effects are likely to resist the well-documented decline in whole body protein renewal with age. Enhanced renewal may reduce the level of damaged and toxic proteins that accumulate during aging, contributing to the extension of life span by CR.

The response to calorie restriction in mammals shows features also common to hibernation: a cross-adaptation hypothesis.

The marked elevation in hepatic carbamyl phosphate synthetase I (CPSI) in calorie-restricted mice, and the changes in erythrocyte 2,3-diphosphoglycerate (2,3-DPG) and in hemoglobin oxygen affinity in calorie-restricted and hypoxic humans living in Biosphere 2 suggest similarities between physiologic events in calorie restriction and hibernation. Other data from the literature strengthen this comparison. Accordingly, we hypothesize that the response to the calorie restriction regime as studied by gerontologists, rather than being a laboratory artifact, is part of a spectrum of responses to food deprivation which have adaptive value in the wild, and whose triggering mechanism may primarily involve the neuroendocrine system.

Caloric intake alters the efficiency of catalase mRNA translation in the liver of old female mice.

The free radical theory of aging predicts that calorie restriction, which extends life span, should reduce oxidant damage. In mammals, the oxidative processes centered in the liver are a major source of free radicals. Liver catalase has the dominant role in the intracellular detoxification of hydrogen peroxide. In male rodents, published studies indicate that aging decreases catalase gene transcription and that calorie restriction obviates this effect. In females, published studies are inconsistent, and no molecular mechanisms have been identified. Here we report that, in female mice, aging can lead to an increase in the translational efficiency of hepatic catalase mRNA, and that calorie restriction obviates this effect. Consideration of these results and published studies leads us to propose that the variability in catalase results in females may arise from the small number of studies or from unique aspects of female physiology, perhaps the estrous cycle and its cessation with age.

Caloric restriction mimetics: metabolic interventions.

Caloric restriction (CR) retards diseases and aging in laboratory rodents and is now being tested in nonhuman primates. One way to apply these findings to human health is to identify and test agents that may mimic critical actions of CR. Panel 2 focused on two outcomes of CR, reduction of oxidative stress and improved glucoregulation, for which candidate metabolic mimics exist. It was recommended that studies on oxidative stress should emphasize mitochondrial function and to test the efficacy of nitrone and other antioxidants in mimicking CR's effects. Studies should also focus on the long-term effects of compounds known to lower circulating glucose and insulin concentrations or to increase insulin sensitivity. Also, four other developing areas were identified: intermediary metabolism, response to infection, stress responses, and source of dietary fat. These areas are important because either they hold promise for the discovery of new mimetics or they need to be explored prior to initiation of CR trials in humans. Other recommendations were that transgenic approaches and adult-onset CR should be emphasized in future studies.

P1-450 and P3-450 gene expression and maximum life span in mice.

The effects of beta-naphthoflavone on the inducibility of hepatic P1-450 and P3-450 mRNA were investigated in male B10.RIII/Sn, C57BL/10Sn, C3H/HeSnJ, and A/WYSn mice. Previous work has shown that the maximum level of aryl hydrocarbon hydroxylase induction in these strains correlates with maximum life span. In this study we found that the maximum inducible levels of P1- and P3-450 RNA were significantly different among the strains, and these levels also correlate with life span. The differences were not due to strain-specific differences in the kinetics of P1- or P3-450 RNA induction. The differences were specific to expression of the P-450 genes, since the levels of hepatic alpha-actin and albumin RNA were not significantly different among the strains, and specific RNA levels were normalized to the level of total polyadenylated RNA. beta-Naphthoflavone was found to induce alpha-actin mRNA approximately 2-fold and to transiently repress albumin RNA about 50% in all mouse strains. Maximum P1- and P3-450 gene expression correlated directly with the 10th deciles of survival of the mouse strains. Longer-lived strains expressed higher combined levels of P1- and P3-450 RNAs. Maximum P1- and P3-450 gene expression also correlated generally with the reported aryl hydrocarbon hydroxylase receptor levels of each strain. It is unlikely that the hepatic P1- and P3-450 genes are ever maximally induced under the sheltered laboratory conditions used to determine maximum life span, as we consistently find very low levels of P-450 expression in uninduced animals. These uninduced levels were not statistically different between the strains. Therefore, the reason for the relationship between maximum life span and maximum P1- and P3-450 inducibility is unclear at present.

Aging alters hepatic expression of insulin receptor and c-jun mRNA in the mouse.

The clear association between species and life span suggests that aging, like development, is genetically orchestrated. To explore this hypothesis, the expression of mRNA for a number of transcription regulatory and signal transduction proteins was investigated during aging of B10.RIII, C57BL/10 and B10.BR mice. mRNA for glucocorticoid receptor, CCAAT and enhancer binding protein, transcription factor Sp1 and RNA polymerase II elongation factor S-II were unchanged between 4 and 24 months of age in these mice. These factors are required for the normal transcription of many genes, perhaps explaining their steady rates of expression throughout life. Insulin-like growth factor I mRNA also remained unchanged. By contrast, mRNA for the insulin receptor and transcription factor c-jun changed significantly during aging. c-Jun mRNA decreased approximately 55% between 4 and 12 months of age and then increased by 24-25 months of age to levels approximately equal to those found in young mice. Insulin receptor mRNA increased approximately 30% by 24-25 months of age in all strains of mice. These results suggest that factors determining the steady state level of these mRNAs are altered in level or activity during aging. Assessing the causes and significance of these changes will require further study. However, our results demonstrate that alterations in the expression of specific regulatory genes occur during aging.

Reversing the negative genomic effects of aging with short-term calorie restriction.

According to government figures, total health care spending in the U.S. in 1999 was 1.316 trillion dollars. The government projects an increase in health care costs to 2.176 trillion dollars by 2008. If we project this growth rate to 2020, health care costs will reach 4.009 trillion dollars. Today, people often spend more health care dollars during the last year of their lives than in all previous years combined. Medical treatment in the last few years of life is usually very expensive and often futile. With the baby-boom generation now moving through middle age, the prescription for the U.S. health care system will be disastrous unless we learn how to keep people healthier longer. This dramatic increase in health care costs leaves us with only one acceptable alternative to rationed health care or financial ruin--to discover interventions that make people functionally younger, healthier, and less susceptible to debilitating, age-related diseases.

Thyroid hormone regulation of the transfected rat growth hormone promoter.

A region of the rat growth hormone gene and 5' flanking DNA has been identified which promotes accurate, thyroid hormone-regulated transcriptional initiation. GC rat pituitary tumor cells were transfected with chimaeric plasmids containing various lengths of rat growth hormone gene and 5' flanking DNA fused to the coding region of the dominant selectable marker gene neo. Thyroid hormone induction of rGH-neo RNA was observed by Northern and dot blot analysis of cells transfected with rGH-neo chimaeric genes sharing the rat growth hormone gene and upstream regions from -235 to +11. Initiation of rGH-neo transcription was mapped by S1 nuclease protection to the in vivo initiation site of the natural growth hormone gene. Transcription of the most deleted thyroid hormone responsive construct involved an induction-attenuation cycle qualitatively similar to the response of the natural gene. However, the 3,5,3'-triiodo-L-thyronine responsiveness of this deleted construct was approximately 2- to 3-fold less than that of less deleted rGH-neo genes tested. These results suggest that, at a minimum, the sequences required for the cyclic 3,5,3'-triiodo-L-thyronine transcriptional response are located within the region of the gene from -235 to +11. Other sequences essential for full responsiveness appear to be located elsewhere in the 5'-flanking DNA. Rat growth hormone promoter utilization appears to be strongly cell-type dependent. We obtained stable transfectants with rGH-neo constructs only in GC cells.

3,5,3'-L-triiodothyronine (thyroid hormone)-induced protein-DNA interactions in the thyroid hormone response elements and cell type-specific elements of the rat growth hormone gene revealed by in vivo dimethyl sulfate footprinting.

The cell type-specific element and 3,5,3'-L-triiodothyronine (thyroid hormone) (T3) response element of the rat growth hormone gene act synergistically to produce cell type-specific, T3-regulated expression. Pit-1 is a pituitary cell type-specific transcription factor that binds the cell type-specific element and is essential to its activity. T3 receptors bind as homodimers and heterodimers to the T3 response element and are essential for its activity. Here, we report the use of ligation-mediated polymerase chain reaction in vivo dimethyl sulfate footprinting to study the effects of T3 on protein-DNA interactions in the rat growth hormone gene promoter in vivo. T3-responsive guanine methylation was detected only in and near the T3 response element and the Pit-1 binding sites. The results indicate that within 2 h, T3 induces occupancy of the T3 response element and Pit-1 sites by their respective trans-acting factors in vivo.

Alterations in local chromatin structure accompany thyroid hormone induction of growth hormone gene transcription.

The chromatin structure of the growth hormone gene and flanking DNA was analyzed in GC rat pituitary tumor cells, which appropriately express and regulate the gene. Thyroid hormone induced the DNase I hypersensitivity of a chromatin domain spanning the transcriptional initiation site from approximately -200 to +150. Three inducible hypersensitive sites were discerned in this region. One of these sites was in the first intron of the gene. The second site was within a region of 5' flanking DNA which promotes accurate, thyroid hormone-regulated transcriptional initiation. Centered between these two sites was a third hormone-inducible hypersensitive site mapping at the position of the TATA sequence. The hormone responsiveness of these hypersensitive sites suggests that occupied 3,5,3'-triiodo-L-thyronine receptor interacts with sequences flanking the growth hormone promoter, perhaps facilitating the binding or activation of a transcription factor at the TATA homology. Three hormone-independent hypersensitive sites were identified near or within intron or flanking DNA sequences similar to Chinese hamster ovary type 2 repetitive DNA. Two additional hormone-independent hypersensitive sites were located more than 1000 nucleotides up-stream of the growth hormone promoter. These sites were present in the cloned genomic DNA and were the only distinct sites found in rat spleen and cultured rat liver cells. Thyroid hormone treatment of GC cells appeared to increase the moderate DNase I sensitivity of the growth hormone gene region beyond that found in deinduced or glucocorticoid-treated cells. Dexamethasone had no discernible effect on the chromatin structure of the gene or flanking DNA in these studies.

Thyroid hormone transcriptional regulatory region of the growth hormone gene.

By deletion-transfection analysis, a region of the rat growth hormone gene has been identified which directs accurate, thyroid hormone responsive transcriptional initiation in vivo. In addition, a thyroid hormone-responsive DNase I hypersensitive domain containing three discrete hypersensitive sites has been identified near the GH promoter. One site is coincident with the TATA homology, and the others lie approximately 150 nucleotides 5' and 3' of this sequence. The TATA and 5' flanking DNA hypersensitive sites are located in the region of the gene which promotes hormone-responsive gene transcription. Based on these results, it is possible that the molecular basis for thyroid hormone induction of GH gene transcription includes binding of the occupied receptor to chromatin sites flanking the TATA homology, promoting binding of the TATA activating protein to this sequence. Together, these events may enhance the rate of RNA polymerase II initiation at the promoter.