Prolactin induction of insulin gene transcription: roles of glucose and signal transducer and activator of transcription 5.

GH and PRL stimulate insulin production in pancreatic beta-cells through induction of insulin gene transcription. The transcriptional effects of GH are mediated through the binding of signal transducer and activator of transcription-5 (STAT5) to a consensus recognition sequence (TTCnnnGAA) in the rat insulin-1 promoter. In this study we demonstrate that PRL also induces the binding of STAT5 proteins to the rat insulin-1 STAT5 motif. However, the magnitude of binding of STAT5 nuclear proteins, as assessed by electrophoretic mobility shift assays, was only 1/30th that of the binding of the same STAT5 proteins to the beta-casein STAT5 site. The differences in the affinities of the rat insulin-1 and beta-casein STAT5 motifs are explained in part by differences in promoter sequences flanking the STAT5 sites. To assess the importance of the STAT motif in PRL induction of insulin gene transcription, we deleted the STAT5 consensus sequence in the rat insulin 1 promoter, cloned the truncated promoter upstream of the luciferase reporter gene, and transfected the construct into rat insulinoma (INS-1) cells. The transcriptional activity of this construct was compared with that of the wild-type promoter. Although deletion of the STAT5 site in the promoter reduced the basal luciferase activity, the response to PRL was unaffected. PRL also induced transcription of constructs containing the wild-type human insulin promoter or the rat insulin-2 promoter, which contain no classic STAT5 sequences. The transcriptional effect of PRL was manifest even when cells were incubated in glucose-free medium, indicating that the action of the hormone is not mediated solely through changes in glucose uptake or glucose metabolism. To identify PRL-responsive regions of the rat and human insulin promoters, we constructed a series of promoter truncations and assessed their responsiveness to PRL. A PRL-responsive region of the rat insulin-1 promoter was localized between nucleotides -165 and -109. A PRL-responsive region of the human insulin promoter was localized between nucleotides -346 and -250. Additional regions of the human and rat insulin-1 promoters were required for PRL induction of a heterologous, minimal thymidine kinase promoter, suggesting that there are multiple PRL-responsive elements in the insulin genes. These observations suggest a glucose- and STAT5-independent pathway by which PRL may induce insulin gene transcription.

Body weight and fat deposition in prolactin receptor-deficient mice.

To explore the roles of the lactogens in adipose tissue development and function, we measured body weight, abdominal fat content, and plasma leptin concentrations in a unique model of lactogen resistance: the PRL receptor (PRLR)-deficient mouse. The absence of PRLRs in knockout mice was accompanied by a small (5-12%), but progressive, reduction in body weight after 16 weeks of age. Females were affected to a greater degree than males. The reduction in weight in female PRLR-deficient mice (age 8-9 months) was associated with a 49% reduction in total abdominal fat mass and a 29% reduction in fat mass expressed as a percentage of body weight. Lesser reductions were noted in male mice. Plasma leptin concentrations were reduced in females but not in males. That the reductions in abdominal fat may reflect in part the absence of lactogen action in the adipocyte is suggested by the demonstration of PRLR messenger RNA in normal mouse white adipose tissue. Nevertheless, steady state levels of PRLR messenger RNA in mature adipocytes are very low, suggesting that the effects of lactogens might be mediated by other hormones or cellular growth factors. Our observations suggest roles for the lactogens in adipose tissue growth and metabolism in pregnancy and postnatal life.

Prolactin induction of insulin gene expression: the roles of glucose and glucose transporter-2.

Previous studies have shown that lactogenic hormones stimulate beta-cell proliferation and insulin production in pancreatic islets. However, all such studies have been conducted in cells incubated in medium containing glucose. Since glucose independently stimulates beta-cell replication and insulin production, it is unclear whether the effects of prolactin (PRL) on insulin gene expression are exerted directly or through the uptake and/or metabolism of glucose. We examined the interactions between glucose and PRL in the regulation of insulin gene transcription and the expression of glucose transporter-2 (glut-2) and glucokinase mRNAs in rat insulinoma (INS-1) cells. In the presence of 5.5 mM glucose, the levels of preproinsulin and glut-2 mRNAs in PRL-treated cells exceeded the levels in control cells (1.7-fold, P<0.05 and 2-fold, P<0.05 respectively). The maximal effects of PRL were noted at 24-48 h of incubation. PRL had no effect on the levels of glucokinase mRNA. The higher levels of glut-2 mRNA were accompanied by an increase in the number of cellular glucose transporters, as demonstrated by a 1. 4- to 2.4-fold increase in the uptake of 2-deoxy-d-[(3)H]glucose in PRL-treated INS-1 cells (P<0.001). These findings suggested that the insulinotropic effect of PRL is mediated, in part, by induction of glucose transport and/or glucose metabolism. Nevertheless, even in the absence of glucose, PRL stimulated increases in the levels of preproinsulin mRNA (3.4-fold higher than controls, P<0.0001) and glut-2 mRNA (2-fold higher than controls, P<0.01). These observations suggested that PRL exerts glucose-independent as well as glucose-dependent effects on insulin gene expression. Support for this hypothesis was provided by studies of insulin gene transcription using INS-1 cells transfected with a plasmid containing the rat insulin 1 promoter linked to a luciferase reporter gene. Glucose and PRL, alone and in combination, stimulated increases in cellular luciferase activity. The relative potencies of glucose (5.5 mM) alone, PRL alone, and glucose plus PRL in combination were 2.2 (P<0.001), 3.4 (P<0.01), and 7.9 (P<0.0001) respectively. Our findings suggest that glucose and PRL act synergistically to induce insulin gene transcription.

Constitutive expression of placental lactogen in pancreatic beta cells: effects on cell morphology, growth, and gene expression.

To explore the roles of lactogens in islet function, we generated a stable line of rat insulinoma (INS-1) cells that express rat placental lactogen II (rPLII) constitutively in culture. We used this cell line (Ins-rPLII) to examine the effects of endogenous rPLII on beta-cell growth, islet formation, and the expression of glucose transporter 2 (glut-2) and insulin mRNA. Growth and maturation of Ins-rPLII cells were compared with that of cells transfected stably with an empty expression plasmid (control) and of INS-1 cells treated with exogenous prolactin. The Ins-rPLII cells proliferated more rapidly than control cells in serumfree medium and showed distinct morphologic characteristics in culture. Whereas the control cells flattened readily on plastic and formed a branching monolayer, the Ins-rPLII cells remained more rounded, sent out fewer projections, and formed more numerous (p<0.01) and larger (p<0.01) beta-cell clusters. Larger clusters assumed a spherical form with well-delineated smooth borders and detached more readily from the culture plates. Maturational progression of the Ins-rPLII cells was associated with a 40% increase in preproinsulin mRNA (p<0.05) and a 2-3-fold increase in glut-2 mRNA (p<0.01). Induction of glut-2 mRNA was accompanied by a 1.4-2.4-fold increase (p< 0.01) in the uptake of radiolabeled 2-deoxyglucose. Similar effects were observed in INS-1 cells exposed for 48 h to exogenous prolactin. These findings suggest novel roles for the lactogenic hormones in the maturation and growth of pancreatic islets. Lactogen induction of beta-cell aggregation coupled with localized beta-cell growth may contribute to the expansion of islet mass that occurs in pregnancy and during the perinatal period. The induction of insulin and glut-2 mRNA provides a mechanism by which the lactogens may increase fetal and maternal insulin production and enhance the sensitivity of the pancreas to glucose.

Genomic structure and chromosomal localization of the novel ETS factor, PE-2 (ERF).

The members of the ETS family of transcription factors are grouped because they share a highly conserved DNA binding domain. These factors are involved in growth factor pathways and regulate both proliferation and differentiation. To identify ETS factors that may be involved in early hematopoietic progenitor regulation, we isolated a novel member of the ETS family by reverse transcriptase-PCR of the conserved DNA binding domain using degenerate oligonucleotides. This gene directs the synthesis of a 2704-nucleotide transcript whose largest open reading frame encodes a 548-amino-acid protein. Northern blot analysis reveals ubiquitous expression in all human tissues and cell lines tested, with highest levels in the testis, ovary, pancreas, and heart. Comparison of this gene with the available databases reveals very significant homology to the ETS factor PE-1 and probable near-identity with the recently cloned factor ERF. The PE-2 gene is composed of four exons spanning over 9 kb of genomic DNA. Sequence analysis of the promoter region reveals a GC-rich sequence without a TATA motif and with putative binding motifs for CREB, c-myb, and AP-1 factors. Using mouse-human somatic hybrids and FISH analysis, the PE-2 gene is localized to human chromosome 19q13.2, a region involved in translocations and deletions in leukemias and several solid tumors, suggesting that this novel ETS factor may play a role in carcinogenesis.

Comparison of human and Xenopus GATA-2 promoters.

GATA proteins comprise a family of transcription factors that are required for appropriate development of hematopoietic lineages. In order to understand the transcriptional regulation of GATA genes, we have cloned the human GATA-2 gene and identified and characterized its promoter. Comparison with the Xenopus GATA-2 promoter demonstrates highly conserved CCAAT box elements, which are essential for appropriate Xenopus expression. Unlike the Xenopus gene, the human GATA-2 gene lacks GATA binding motifs within the first 800 bp of 5' flanking sequence. In addition, the human GATA-2 promoter has two highly conserved ets sites that resemble the binding site for a recently described ets repressor factor, ERF. These conserved DNA sequence motifs provide strong candidate regions for the regulation of GATA-2.

The mouse CD7 gene: identification of a new element common to the human CD7 and mouse Thy-1 promoters.

Human CD7 (hCD7) is a 40 000 Mr member of the immunoglobulin gene superfamily that is expressed early in natural killer (NK) and T-lymphocyte development. CD7 is involved in lymphocyte activation, as ligation of CD7 activates NK and TCRgammadelta T lymphocytes, and ligation of CD7 on TCRalphabeta T lymphocytes induces a non-mitogenic calcium flux. We have previously cloned and characterized the gene for human CD7 (hCD7) and have described its expression in transgenic mice. Recently a mouse cDNA homologous to hCD7 was reported, which we mapped to the corresponding mouse chromosomal location as hCD7. We now report the identification and characterization of a mouse CD7 (mCD7) genomic clone. We demonstrated that the mCD7 gene was similar both in size and structural organization to hCD7. Comparison of the 5' flanking sequences of the mCD7 and hCD7 genes revealed two regions of sequence similarity. Electrophoretic mobility shift assay confirmed both of these regions to be sites of tissue-restricted protein binding in vitro. The more 3' similarity region also shared sequence with a region in the mouse Thy-1 gene 5' flanking region, suggesting that this sequence may be a cis-acting regulatory element common to all three genes. Thus, the promoter regions and exonic organization were similar in the human CD7, mouse CD7, and mouse Thy-1 genes.

Characterization of human CD7 transgenic mice.

CD7 is a 40-kDa transmembrane glycoprotein member of the lg gene superfamily expressed on most peripheral blood T lymphocytes and NK cells. CD7 is also expressed on myeloid, NK, B, and T cell precursors during adult hematopoiesis. Because Thy-1 is absent in human thymocytes and peripheral blood T cells and shows structural similarities to the human CD7 gene, we have suggested that human CD7 may be a functional homologue in humans of mouse Thy-1. To study the tissue-specific expression of the CD7 gene utilizing its own promoter, we constructed transgenic mice that contained both the coding and flanking regions of the human CD7 gene. We found that human CD7 was expressed in transgenic mice in T, B, NK, and myeloid lineages and was induced with T cell activation. Unlike the expression of CD7 in humans, the CD7 transgene was present on mature B lymphocytes and macrophages. Like mouse Thy-1, transgenic human CD7 was expressed in immature and mature T cells and in Sca-1+ bone marrow mononuclear cells. Unlike mouse Thy-1, the human CD7 transgene was not expressed in mouse brain or fibroblasts. The human CD7 transgene was expressed during fetal development before mouse Thy-1 in fetal liver mononuclear cells. Expression of the human CD7 transgene did not alter mouse thymopiesis or Thy-1 expression. Taken together, these data demonstrated that the CD7 transgene contained sufficient regulatory regions to direct hematopoietic expression and mitogenic induction. The pattern of CD7 transgene expression more closely resembled that of CD7 in humans than that of mouse Thy-1.

The structure of the urokinase-type plasminogen activator receptor gene.

The cellular receptor for urokinase-type plasminogen activator (uPAR) is a glycosylphosphatidylinositol (GPI)-anchored membrane protein that plays a central role in pericellular plasminogen activation. It contains 313 amino acid residues, including 28 cysteine residues in a pattern of three homologous repeats. The cysteine residue pattern suggests that uPAR belongs to a superfamily of proteins including CD59, murine Ly-6, and a variety of elapid snake venom toxins. A novel 1.7-kb uPAR cDNA was isolated that is missing exon 5 and that contains 380 bp not previously reported at the 5' end. This cDNA was used to probe a human genomic library from which three clones were isolated and analyzed. The uPAR gene consists of 7 exons spread over 23 kb of genomic DNA. Exons 2, 4, and 6 code for homologous domains within the mature protein, as do exons 3, 5, and 7; CD59-like homologous pairs are encoded by exons 2-3, 4-5, and 6-7, respectively. The structure of the gene for uPAR further confirms the relationship of this molecule to the superfamily containing CD59, Ly-6, and the elapid snake venom toxins.

A novel beta-globin mutation, beta Durham-NC [beta 114 Leu-->Pro], produces a dominant thalassemia-like phenotype.

Mutations within exon 3 of the beta-globin gene are relatively uncommon, and many of these mutations produce a dominant thalassemia-like phenotype. We describe a novel thalassemic hemoglobinopathy caused by a single nucleotide substitution (CTG-->CCG) at codon 114 resulting in a leucine to proline substitution and designate it beta Durham-NC [beta 114 Leu-->Pro]. The mutation producing this thalassemic hemoglobinopathy is located near to the beta Showa-Yakushiji mutation (beta 110 Leu-->Pro). Both of these hemoglobinopathies share similar phenotypic features with moderately severe microcytic anemia. Using computer imaging of the hemoglobin molecule, we examined several reported point mutations within exon 3 of the beta-globin gene. These point mutations cause a single amino acid substitution in the G helix, and result in a thalassemic and/or hemolytic phenotype. Computer imaging of nine separate examples suggests that amino acid substitutions affecting side chains that project into the heme pocket may destabilize the heme moiety within the beta-globin chain, resulting in a thalassemic phenotype. Hemolytic phenotypes may be the result of decreased alpha 1 beta 1 interactions. The beta Durham-NC mutation further characterizes a novel group of thalassemias/hemoglobinopathies that are clinically difficult to identify and require accessory laboratory testing.

Characterization of the DNase I hypersensitive site 3' of the human beta globin gene domain.

The members of the human beta globin gene family are flanked by strong DNase I hypersensitive sites. The collection of sites 5' to the epsilon globin gene is able to confer high levels of expression of linked globin genes, but a function has not been assigned to the site 3' to the beta globin gene (3'HS1). Our analysis of this DNase I super hypersensitive site shows that the region is composed of multiple DNase I sites. By examination of the DNA sequence, we have determined that the region is very A/T-rich and contains topoisomerase II recognition sequences, as well as several consensus binding motifs for GATA-1 and AP-1/NF-E2. Gel mobility shift assays indicate that the region can interact in vitro with GATA-1 and AP-1/NF-E2, and functional studies show that the region serves as a scaffold attachment region in both erythroid and nonerythroid cell lines. Whereas many of the physical features of 3'HS1 are shared by 5'HS2 (a component of the 5' locus control region), transient expression studies show that 3' HS1 does not share the erythroid-specific enhancer activity exhibited by 5'HS2.

Structure of the CD59-encoding gene: further evidence of a relationship to murine lymphocyte antigen Ly-6 protein.

The gene for CD59 [membrane inhibitor of reactive lysis (MIRL), protectin], a phosphatidylinositol-linked surface glycoprotein that regulates the formation of the polymeric C9 complex of complement and that is deficient on the abnormal hematopoietic cells of patients with paroxysmal nocturnal hemoglobinuria, consists of four exons spanning 20 kilobases. The untranslated first exon is preceded by a G+C-rich promoter region that lacks a consensus TATA or CAAT motif. The second exon encodes the hydrophobic leader sequence of the protein, and the third exon encodes the amino-terminal portion of the mature protein. The fourth exon encodes the remainder of the mature protein, including the hydrophobic sequence necessary for glycosyl-phosphatidylinositol anchor attachment. The structure of the CD59 gene is very similar to that encoding Ly-6, a murine glycoprotein with which CD59 has some structural similarity. The striking similarity in gene structure is further evidence that the two proteins belong to a superfamily of proteins that may also include the urokinase plasminogen-activator receptor and a squid glycoprotein of unknown function.

Avian cardiac tropomyosin gene produces tissue-specific isoforms through alternative RNA splicing.

We have isolated a quail cardiac tropomyosin gene which encodes three distinct isoforms through the use of alternative exon splicing. Characterization of cDNA clones produced by this gene indicate that the gene encodes a unique 284 amino acid cardiac tropomyosin isoform, along with a 248 amino acid cytoskeletal and 284 amino acid smooth muscle isoforms. Northern analyses indicate that the gene is primarily expressed in cardiac muscle, with only minor expression of the cytoskeletal and smooth muscle transcripts.

Isolation and characterization of the genomic human CD7 gene: structural similarity with the murine Thy-1 gene.

The human CD7 molecule is a 40-kDa member of the immunoglobulin gene superfamily that is expressed on T-lymphoid and myeloid precursors in fetal liver and bone marrow. CD7 is also expressed on T lymphocytes in multiple stages of T-cell development, including a major subset of mature peripheral T cells. In this paper we report the isolation and characterization of the human CD7 gene and 5' flanking region. Sequence analysis revealed that the CD7 gene comprises four exons that span 3.5 kilobases. The 5' flanking region (506 base pairs) has a high G + C content and no "TATA" or "CCAAT" elements. DNase I sensitivity analysis of chromatin from the CD7+ progenitor cell leukemia line, DU528, and the CD7-, CD4+, CD8+, TCR alpha beta + T-cell line, DU980 (where TCR is the T-cell receptor), revealed two distinct hypersensitive sites 5' of the CD7 gene. Hypersensitive site 1, present in the DU980 T-cell line, was located 4.5 kilobases upstream of the presumed CD7 transcription initiation site. Only DNase I hypersensitive site 2, which mapped to the promoter region, was found in the DU528 line. Comparison of the organization of the CD7 gene with that of other members of the immunoglobulin gene superfamily revealed that the human CD7 gene most closely resembles the murine Thy-1 gene. Both CD7 and Thy-1 are encoded by small genes with four exons, contain TATA-less promoters, and have a similar functional organization. These structural similarities suggest that human CD7 and murine Thy-1 may be functional homologues.

Avian tropomyosin gene expression.

Sequence analysis of overlapping fragments from a quail genomic library has revealed a tropomyosin gene consisting of 13 exons spaced over about 18 kilobase pairs of DNA. Skeletal muscle and smooth muscle transcripts share the same 5' untranslated sequence and may initiate from the same promoter. However, the regions encoding amino acids 39-80 and 258-284 are specific to each muscle type. The two sets of exons encoding these regions undergo mutually exclusive alternative splicing in a tissue-specific manner as determined by Northern blots and S1-nuclease protection. Similarly, the 3' ends of the transcripts are different in skeletal muscle and smooth muscle, and each contains two polyadenylation signals which appear to be utilized in vivo. The avian alpha-tropomyosin gene is not expressed in cardiac muscle. The sequence of the gene shows great homology with other muscle-specific tropomyosins and includes a region homologous to the amino terminus of nonmuscle tropomyosins.

Impaired up-regulation of type II corticosteroid receptors in hippocampus of aged rats.

Several recent investigations have reported a decline of rat hippocampal corticosteroid-binding receptors (CSRs) with aging. This decline has been proposed to be an initial cause (through disinhibition) of the elevated adrenal steroid secretion that apparently occurs with aging; however, it could instead be an effect of corticoid elevation (through down-regulation). In order to assess the effects of age on CSR biosynthetic capacity in the absence of down-regulatory influences of endogenous corticoids, as well as to study aging changes in CSR plasticity, we examined the up-regulation of hippocampal CSR that follows adrenalectomy (ADX). The rat hippocampus contains at least two types of CSR binding and differential analysis of types I and II CSR was accomplished by selective displacement of [3H]corticosterone with RU-28362, a specific type II agonist. In young (3 months old) Fischer-344 rat hippocampus, up-regulation of type II binding above 2-day ADX baseline was present by 3-7 days and increased still further by 8-10 days post-ADX; type I CSR density did not change significantly between 1 and 10 days post-ADX. However, in aged (24-26 months old) rats, type II CSR up-regulation did not occur over the 10 day post-ADX period. Thus, the age-related impairment of type II up-regulation may reflect an intrinsic deficit in CSR biosynthesis or lability that is independent of the acute endogenous adrenal steroid environment.

Analysis of xanthine dehydrogenase mRNA levels in mutants affecting the expression of the rosy locus.

Xanthine dehydrogenase (XDH) mRNA levels were measured in a number of mutants and natural variants affecting XDH gene expression. Two variants, ry+4 and ry+10, contain cis-acting elements which map to a region flanking the 5' end of the XDH gene. Ry+4, which has 2-3 times more XDH protein than a wild type strain, has 3.2 times more XDH mRNA. Ry+10 has 50% of the wild type XDH level and 54% of the wild type XDH mRNA level. Three rosy mutants which map within the structural gene were also examined. Two of these had little if any XDH mRNA, but the third mutant had 1.3 times more XDH mRNA than wild type flies. Another mutant, ry2 , which contains no XDH protein and has a 9KB transposable element inserted into the XDH gene, has normal levels of XDH mRNA transcripts which are also the same size as those found in the wild type strain. Changes in XDH mRNA levels were measured during Drosophila development and found to parallel changes in the amount of XDH protein. In addition, there were no large changes in the size of XDH mRNA during development.