Vol. 189, No. 2, 1992 December
BIOCHEMICAL
AND BIOPHYSICAL
RESEARCH COMMUNICATIONS Pages
15, 1992
1031-l
037
POST-TRANSCRIPTIONAL REGULATION OF INSULIN-LIKE GROWTH FACTOR BINDING PROTEIN-2 mRNA IN DIABETIC RAT LIVER Guck T. Ooi*, Lucy Y-H. Tseng, and Matthew
M. Rechler
Growth and Development Section, Molecular and Cellular Endocrinology Branch, National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health, Bethesda, MD 20892 Received
October
16,
1992
SUMMARY: IGFBP-1 and IGFBP-2 ‘mRNAs are increased in the livers of streptozotocin-diabetic rats. A corresponding increase is observed in transcription of the IGFBP-1 but not the IGFBP-2 gene, indicating that the increase in steadystate levels of IGFBP-2 mRNA is a post-transcriptional effect. IGFBP-1 and IGFBP-2 mRNAs also differ in the rapidity of their response to insulin treatment: hepatic IGFBP-1 mRNA is normalized within 1 h, IGFBP-2 mRNA decreases more slowly. These differences suggest that IGFBP-2 may provide more chronic adaptation to metabolic change than IGFBP-1. Q 1992Academic P=-~, WC.
The insulin-like growth factor binding proteins (IGFBPs) are a family of related proteins that specifically bind IGF-I and IGF-II, and modulate their bioavailability and biological activity (1,2). The levels of IGFBPs in plasma and tissues are highly regulated, and are determined by the balance between synthesis and degradation (2,3). Our laboratory has studied the expression of two of the IGFBP genes, IGFBP-1 and IGFBP-2, in rat liver under various physiological conditions (4). Both IGFBP-1 and IGFBP-2 mRNAs are increased in diabetes (5,6), fasting (5,7,8), and growth hormone deficiency (5,7,9,10). However, treatment of severely diabetic and ketotic rats with insulin for 3 days normalized IGFBP-1 but not IGFBP-2 mRNA (5). This suggested that the two IGFBP mRNAs had different mechanisms of regulation by insulin, different sensitivities to insulin, or both. The increase in IGFBP- 1 mRNA in diabetic rat liver and the rapid decrease following insulin treatment reflect changes in IGFBP-1 gene transcription (11). Transcription of the IGFBP-2 gene is increased in fasted adult rat liver and in the liver of neonatal rats (11,12). The present study examines whether the increase in IGFBP-2 mRNA in diabetic rat liver is transcriptional or posttranscriptional, and compares the kinetics of normalization of IGFBP- 1 and IGFBP-2 mRNAs following insulin treatment. METHODS Animals. Diabetes was induced in -300 g adult male Sprague-Dawley rats (Taconic Farms, Germantown, NY) by a single inuaperitoneal injection of streptozotocin (100 mg/kg body weight; *To whom correspondence and reprint requests should be addressed. National Institutes of Health, Building lo--Room 8D14, Bethesda, MD 20892. Fax: 301-402-3146; 301-496-1649. 0006-291X/92
1031
$4.00
Copyright 0 1992 by Academic Press, Inc. All rights of reproduction in any form reserved.
Vol.
189, No. 2, 1992
BIOCHEMICAL
AND BIOPHYSICAL
RESEARCH COMMUNICATIONS
Sigma Chemical Co., St. Louis, MO). By 7 days after streptozotocin administration, rats that were glucosuric and did not gain weight were considered diabetic and randomly assigned either to be sacrificed (together with control rats not treated with streptozotocin) or to be treated with insulii. Metabolic data are presented in reference 11 (Table 1, experiments 1,2, and 4). Different insulin treatment regimens were used in different experiments: a) subcutaneous injection of 2.4 U / 100 g human recombinant regular insulin (Humulin R, Eli Lilly and Co., Indianapolis, IN), sacrificed after 1 h (Fig. 1; reference 11, experiment 4); b) injection of 8U / 100 g recombinant NPH insulin (Humulin N, intermediate-acting) at 0 and 12 h, sacrificed 6 or 24 h after the first injection (Fig. 3; reference 11, experiment 1); or c) injection with a mixture of 2.4 U of Humulin R and 3.2 U Humulin N / lOOg, sacrificed 2 or 4 h later (Fig. 4, reference 11, experiment 2). Serum glucose was -3.5 times higher in diabetic rats than in nondiabetic controls; euglycemia was restored by insulin treatment. Northern Blot Hybridization. Total RNA was prepared using the guanidine thiocyanate-LiCl method (13) or the acid guanidine thiocyanate-phenol-chloroform method (14). The RNA (12-15 pg) was fractionated by electrophoresis on 1.5 % agarose/formaldehyde gels in 3morpholinopropanesulfonic acid (MOPS) buffer and blotted to nylon membranes (GeneScreen, Dupont-NEN, Boston, MA) as previously described (13). Ethidium bromide staining confii that the ribosomal RNAs were intact and that equal amounts of RNA had been loaded in each lane. Hybridization probes were generated by polymerase chain reaction amplification of nucleoddes - 116 to +557 (with respect to the ATG translation initiation site, +l) of a rat IGFBP-1 cDNA clone (G.T. Goi, unpublished results; GenBank accession No. M89791), or nucleotides 216-836 of a rat IGFBP-2 cDNA clone (15). Both cDNA probes were labelled with [c@P]dCI’P (3000 Ci/mmol; Amersham Corp., Arlington Heights, IL) by random priming using a kit obtained from Stratagene (La Jolla, CA). The blots were hybridized at 50 C as previously described [ 1.2 x 106 cpm/ml; 50% formamide, 10% dextran sulfate, 5x SSPE (0.9 M NaCl, 5 mM EDTA, 50 mM Na phosphate, pH 6.8), 1% sodium dodecyl sulfate (SDS), 10x Denhardt’s solution, and 100 &ml salmon sperm DNA] (13), washed in 2x SSPE-0.2% SDS at room temperature, followed by 0.1 x SSPE at 60 C, and exposed to Kodak XAR-2 film for l-2 days. Nuclear Transcription Assays. Nuclei were prepared from l-2 g frozen liver as previously described (11,12). Nuclear run-on transcripts were prepared using [c+P]Cl’P and [c+P]GTP (Amersham Corp.) (12). Denatured target DNAs (1 l&lot) were applied onto nitrocellulose membranes (Schleicher and Schuell, Keene, NH) using a slot-blot apparatus. The IGFBP- 1 DNA target was a 1374 base pair fragment (nucleotides -116 to 1258 of rat IGFBP-1 cDNA) synthesized by polymerase chain reaction amplification. Since intron 1 of the rat IGFBP-2 gene is >25 kb (3), two different rat IGFBP-2 genomic DNAs were used as targets: a 2.1 kb W-W fragment containing parts of exon 1 and intron 1, and a 4.2 kb Hindm-HindIII fragment beginning in intron 1 and extending into exon 4 (12). Filters without DNA (“Blank”) or containing linearized pGem7Zf(-) plasmid DNA were used as control hybridization targets. Aliquots of RNA transcripts having equal amounts of radioactivity (1 million cpm) were added to 1.8~ml capped polypropylene tubes containing the multiple DNA target filters in 250 pl of hybridization buffer [0.5 M NaCl, 50 mM piperazine-N,N’-bis[2-ethanesulfonic acid] (PIPES), pH 7.5,33% formamide, 2 mM EDTA, 100 l.tg/ml tRNA, 10 pg denatured linearized pGem 7 plasmid DNA without insert, 0.4 % SDS], overlaid with mineral oil, and hybridized at 45 C for 3 days. Filters were washed extensively, treated with 5 ml of ribonuclease A (10 pg/ml) and ribonuclease Tl(2 l@nl), mounted on a filter paper template, and autoradiographed (12). Quantitation of hybridization results. The hybridization signal was analysed quantitatively by &canning using a computer-driven radioactivity scanner (AMBIS Scanning System II, Automated Microbiology Systems, Inc, San Diego, CA). Results are expressed as counts hybridized, following subtraction of the appropriate blank: radioactivity hybridized to an equivalent area of the Northern blot, or nuclear transcript radioactivity hybrid&d to blank filters (containing either no DNA or linearized plasmid DNA) after incubation with the same radiolabeled transcripts. Radioactivity (uncorrected) hybridized to control samples was compared with radioactivity hybridized to blank filters by a l-tail paired t-test (16) using a statistical program (StatView SE+ Graphics, Abacus Concepts, Berkeley, CA). Regression analysis of the relationship betweeen IGFBP-1 and IGFBP-2 mRNA abundance (Northern blot) and transcript abundance (based on runon transcription assays) in samples from the same animals was performed using the StatView program. 1032
Vol.
189,
No.
2,
BIOCHEMICAL
1992
AND
BIOPHYSICAL
RESEARCH
COMMUNICATIONS
RESULTS The increase in IGFBP-2 mRNA in diabetic rat liver does not result from increased transcription. Transcription of the IGFBP-1 and IGFBP-2 genes, and the abundance of the respective mRNAs, were compared in the livers of 6 non-ketotic diabetic rats and 5 control rats (Fig. 1). Steady-state levels of both mRNAs were increased in the livers of all of the diabetic rats. The magnitude of the increase varied, with the increases being greater in animals 19 and 20 than in the other four animals. Transcription of the IGFBP-1 and IGFBP-2 genes was examined in transcription-elongation assays using nuclei from the same animals (Fig. 1). Transcript radioactivity hybridizing to an immobilized IGFBP-1 cDNA target was greater in diabetic than control animals. Because of the variation in mRNA abundance in individual animals, we have compared mRNA abundance and transcription in paired samples from the same animal (Fig. 2). For IGFBP-1, transcription increased in proportion to mRNA levels (lefppane0; r=O.88, p
BIOCHEMICAL
AND BIOPHYSICAL
RESEARCH COMMUNICATIONS Pages
15, 1992
1031-l
037
POST-TRANSCRIPTIONAL REGULATION OF INSULIN-LIKE GROWTH FACTOR BINDING PROTEIN-2 mRNA IN DIABETIC RAT LIVER Guck T. Ooi*, Lucy Y-H. Tseng, and Matthew
M. Rechler
Growth and Development Section, Molecular and Cellular Endocrinology Branch, National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health, Bethesda, MD 20892 Received
October
16,
1992
SUMMARY: IGFBP-1 and IGFBP-2 ‘mRNAs are increased in the livers of streptozotocin-diabetic rats. A corresponding increase is observed in transcription of the IGFBP-1 but not the IGFBP-2 gene, indicating that the increase in steadystate levels of IGFBP-2 mRNA is a post-transcriptional effect. IGFBP-1 and IGFBP-2 mRNAs also differ in the rapidity of their response to insulin treatment: hepatic IGFBP-1 mRNA is normalized within 1 h, IGFBP-2 mRNA decreases more slowly. These differences suggest that IGFBP-2 may provide more chronic adaptation to metabolic change than IGFBP-1. Q 1992Academic P=-~, WC.
The insulin-like growth factor binding proteins (IGFBPs) are a family of related proteins that specifically bind IGF-I and IGF-II, and modulate their bioavailability and biological activity (1,2). The levels of IGFBPs in plasma and tissues are highly regulated, and are determined by the balance between synthesis and degradation (2,3). Our laboratory has studied the expression of two of the IGFBP genes, IGFBP-1 and IGFBP-2, in rat liver under various physiological conditions (4). Both IGFBP-1 and IGFBP-2 mRNAs are increased in diabetes (5,6), fasting (5,7,8), and growth hormone deficiency (5,7,9,10). However, treatment of severely diabetic and ketotic rats with insulin for 3 days normalized IGFBP-1 but not IGFBP-2 mRNA (5). This suggested that the two IGFBP mRNAs had different mechanisms of regulation by insulin, different sensitivities to insulin, or both. The increase in IGFBP- 1 mRNA in diabetic rat liver and the rapid decrease following insulin treatment reflect changes in IGFBP-1 gene transcription (11). Transcription of the IGFBP-2 gene is increased in fasted adult rat liver and in the liver of neonatal rats (11,12). The present study examines whether the increase in IGFBP-2 mRNA in diabetic rat liver is transcriptional or posttranscriptional, and compares the kinetics of normalization of IGFBP- 1 and IGFBP-2 mRNAs following insulin treatment. METHODS Animals. Diabetes was induced in -300 g adult male Sprague-Dawley rats (Taconic Farms, Germantown, NY) by a single inuaperitoneal injection of streptozotocin (100 mg/kg body weight; *To whom correspondence and reprint requests should be addressed. National Institutes of Health, Building lo--Room 8D14, Bethesda, MD 20892. Fax: 301-402-3146; 301-496-1649. 0006-291X/92
1031
$4.00
Copyright 0 1992 by Academic Press, Inc. All rights of reproduction in any form reserved.
Vol.
189, No. 2, 1992
BIOCHEMICAL
AND BIOPHYSICAL
RESEARCH COMMUNICATIONS
Sigma Chemical Co., St. Louis, MO). By 7 days after streptozotocin administration, rats that were glucosuric and did not gain weight were considered diabetic and randomly assigned either to be sacrificed (together with control rats not treated with streptozotocin) or to be treated with insulii. Metabolic data are presented in reference 11 (Table 1, experiments 1,2, and 4). Different insulin treatment regimens were used in different experiments: a) subcutaneous injection of 2.4 U / 100 g human recombinant regular insulin (Humulin R, Eli Lilly and Co., Indianapolis, IN), sacrificed after 1 h (Fig. 1; reference 11, experiment 4); b) injection of 8U / 100 g recombinant NPH insulin (Humulin N, intermediate-acting) at 0 and 12 h, sacrificed 6 or 24 h after the first injection (Fig. 3; reference 11, experiment 1); or c) injection with a mixture of 2.4 U of Humulin R and 3.2 U Humulin N / lOOg, sacrificed 2 or 4 h later (Fig. 4, reference 11, experiment 2). Serum glucose was -3.5 times higher in diabetic rats than in nondiabetic controls; euglycemia was restored by insulin treatment. Northern Blot Hybridization. Total RNA was prepared using the guanidine thiocyanate-LiCl method (13) or the acid guanidine thiocyanate-phenol-chloroform method (14). The RNA (12-15 pg) was fractionated by electrophoresis on 1.5 % agarose/formaldehyde gels in 3morpholinopropanesulfonic acid (MOPS) buffer and blotted to nylon membranes (GeneScreen, Dupont-NEN, Boston, MA) as previously described (13). Ethidium bromide staining confii that the ribosomal RNAs were intact and that equal amounts of RNA had been loaded in each lane. Hybridization probes were generated by polymerase chain reaction amplification of nucleoddes - 116 to +557 (with respect to the ATG translation initiation site, +l) of a rat IGFBP-1 cDNA clone (G.T. Goi, unpublished results; GenBank accession No. M89791), or nucleotides 216-836 of a rat IGFBP-2 cDNA clone (15). Both cDNA probes were labelled with [c@P]dCI’P (3000 Ci/mmol; Amersham Corp., Arlington Heights, IL) by random priming using a kit obtained from Stratagene (La Jolla, CA). The blots were hybridized at 50 C as previously described [ 1.2 x 106 cpm/ml; 50% formamide, 10% dextran sulfate, 5x SSPE (0.9 M NaCl, 5 mM EDTA, 50 mM Na phosphate, pH 6.8), 1% sodium dodecyl sulfate (SDS), 10x Denhardt’s solution, and 100 &ml salmon sperm DNA] (13), washed in 2x SSPE-0.2% SDS at room temperature, followed by 0.1 x SSPE at 60 C, and exposed to Kodak XAR-2 film for l-2 days. Nuclear Transcription Assays. Nuclei were prepared from l-2 g frozen liver as previously described (11,12). Nuclear run-on transcripts were prepared using [c+P]Cl’P and [c+P]GTP (Amersham Corp.) (12). Denatured target DNAs (1 l&lot) were applied onto nitrocellulose membranes (Schleicher and Schuell, Keene, NH) using a slot-blot apparatus. The IGFBP- 1 DNA target was a 1374 base pair fragment (nucleotides -116 to 1258 of rat IGFBP-1 cDNA) synthesized by polymerase chain reaction amplification. Since intron 1 of the rat IGFBP-2 gene is >25 kb (3), two different rat IGFBP-2 genomic DNAs were used as targets: a 2.1 kb W-W fragment containing parts of exon 1 and intron 1, and a 4.2 kb Hindm-HindIII fragment beginning in intron 1 and extending into exon 4 (12). Filters without DNA (“Blank”) or containing linearized pGem7Zf(-) plasmid DNA were used as control hybridization targets. Aliquots of RNA transcripts having equal amounts of radioactivity (1 million cpm) were added to 1.8~ml capped polypropylene tubes containing the multiple DNA target filters in 250 pl of hybridization buffer [0.5 M NaCl, 50 mM piperazine-N,N’-bis[2-ethanesulfonic acid] (PIPES), pH 7.5,33% formamide, 2 mM EDTA, 100 l.tg/ml tRNA, 10 pg denatured linearized pGem 7 plasmid DNA without insert, 0.4 % SDS], overlaid with mineral oil, and hybridized at 45 C for 3 days. Filters were washed extensively, treated with 5 ml of ribonuclease A (10 pg/ml) and ribonuclease Tl(2 l@nl), mounted on a filter paper template, and autoradiographed (12). Quantitation of hybridization results. The hybridization signal was analysed quantitatively by &canning using a computer-driven radioactivity scanner (AMBIS Scanning System II, Automated Microbiology Systems, Inc, San Diego, CA). Results are expressed as counts hybridized, following subtraction of the appropriate blank: radioactivity hybridized to an equivalent area of the Northern blot, or nuclear transcript radioactivity hybrid&d to blank filters (containing either no DNA or linearized plasmid DNA) after incubation with the same radiolabeled transcripts. Radioactivity (uncorrected) hybridized to control samples was compared with radioactivity hybridized to blank filters by a l-tail paired t-test (16) using a statistical program (StatView SE+ Graphics, Abacus Concepts, Berkeley, CA). Regression analysis of the relationship betweeen IGFBP-1 and IGFBP-2 mRNA abundance (Northern blot) and transcript abundance (based on runon transcription assays) in samples from the same animals was performed using the StatView program. 1032
Vol.
189,
No.
2,
BIOCHEMICAL
1992
AND
BIOPHYSICAL
RESEARCH
COMMUNICATIONS
RESULTS The increase in IGFBP-2 mRNA in diabetic rat liver does not result from increased transcription. Transcription of the IGFBP-1 and IGFBP-2 genes, and the abundance of the respective mRNAs, were compared in the livers of 6 non-ketotic diabetic rats and 5 control rats (Fig. 1). Steady-state levels of both mRNAs were increased in the livers of all of the diabetic rats. The magnitude of the increase varied, with the increases being greater in animals 19 and 20 than in the other four animals. Transcription of the IGFBP-1 and IGFBP-2 genes was examined in transcription-elongation assays using nuclei from the same animals (Fig. 1). Transcript radioactivity hybridizing to an immobilized IGFBP-1 cDNA target was greater in diabetic than control animals. Because of the variation in mRNA abundance in individual animals, we have compared mRNA abundance and transcription in paired samples from the same animal (Fig. 2). For IGFBP-1, transcription increased in proportion to mRNA levels (lefppane0; r=O.88, p
