0013-7227/90/1273-1033$02.00/0 Endocrinology Copyright © 1990 by The Endocrine Society

Vol. 127, No. 3 Printed in U.S.A.

Expression of Insulin-Like Growth Factor I Stimulates Normal Somatic Growth in Growth Hormone-Deficient Transgenic Mice RICHARD R. BEHRINGER,* TAL M. LEWIN, CAROL J. QUAIFE, RICHARD D. PALMITER, RALPH L. BRINSTER, AND A. JOSEPH D'ERCOLE Department of Pediatrics, University of North Carolina at Chapel Hill (T.M.L., A.J.D.), Chapel Hill, North Carolina 27599; Department of Biochemistry and the Howard Hughes Medical Institute, University of Washington (C.J.Q., R.D.P.), Seattle, Washington 98195; and the Laboratory of Reproductive Physiology, School of Veterinary Medicine, University of Pennsylvania (R.R.B., R.L.B.), Philadelphia, Pennyslvania 19104

ABSTRACT. A line of transgenic mice expressing insulin-like growth factor-I (IGF-I) under the control of the mouse metallothionien-1 promoter was crossed to a line of dwarf transgenic mice lacking GH expressing cells that were genetically ablated by diphtheria toxin expression. Mice generated from this cross that carry both transgenes express IGF-I in the absence of GH. These mice grew larger than their GH-deficient transgenic littermates and exhibited weight and linear growth indistinguishable from that of their nontransgenic siblings. These results

G

confirm the suspected role of IGF-I in mediating GH's stimulation of somatic growth, including that of long bones, and illustrates the essential role of GH and IGF-I in the modulation of postnatal growth. Analysis of differences in organ growth among these mice, however, suggests that GH and IGF-I also have growth promoting actions that are independent of one another; GH appears to be necessary for the attainment of normal liver size, while IGF-I can stimulate brain growth. (Endocrinology 127: 1033-1040,1990)

H PLAYS a central role in the regulation of postnatal mammalian growth (1, 2). Its precise mechanism of action, however, remains controversial. GH stimulates the expression of insulin-like growth factor-I (IGF-I), a peptide with mitogenic and differentialve activity that is synthesized in many tissues (3-5). It is not clear, however, whether GH has direct effects on somatic growth in addition to those exerted by IGF-I. Administration of IGF-I to hypophysectomized rats (69), or GH-deficient rats (10) or mice (11, 12), as well as intact rats (13, 14), for relatively short durations (6-18 days) has been shown to result in incremental somatic and linear growth. Because GH is about 20-fold more potent than IGF-I on a molar basis in effecting these increases in somatic and linear growth (6-8), IGF-I's capacity to replicate GH's growth-promoting activity has been questioned. Our studies of a transgenic (Tg) mouse line expressing a chimeric IGF-I gene driven by the mouse metallothionien-I promotor, MT-IGF-I, did not resolve this issue (15). While these mice exhibit over-

growth, manifested by a 30% increase in body weight and selective organomegaly (brain, pancreas, kidney, spleen, and carcass), they do not appear to grow longer, as estimated by radiographs of long bones or by tibial epiphyseal growth over a 10-day period. In addition, MTIGF-I Tg mice do not grow to the size of Tg mice that overexpress GH (16, 17); however, MT-IGF-I Tg mice express less IGF-I than do Tg mice overexpressing GH (18). To study the effects of IGF-I expression in the absence of GH, and thereby, to more clearly dissect the independent growth-promoting effects of IGF-I and GH, we crossed the MT-IGF-I Tg mice with GH-deficient mice generated by the expression of a transgene made by fusing rat GH 5' flanking genomic DNA to the gene for the A chain of diphtheria toxin (GH-DT) (19). We report that IGF-I expression in the absence of GH stimulates normal weight and linear growth, and that GH and IGFI also appear to exert unique growth promoting effects on liver and brain, respectively.

Received April 16, 1990. Address correspondence and reprint requests to: A. Joseph D'Ercole, M.D., Department of Pediatrics. CB#7220, University of North Carolina, Chapel Hill, North Carolina 27599-7220. * Present address: Department of Molecular Genetics, University of Texas, M. D. Anderson Cancer Center, Houston, Texas 77030.

Materials and Methods Transgenic mice MT-IGF-I Tg mice (15) were crossed with GH-DT Tg mice (19) by in vitro fertilization using sperm from GH-DT Tg males 1033

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 19 November 2015. at 06:45 For personal use only. No other uses without permission. . All rights reserved.

IGF-I EXPRESSION IN TRANSGENIC MICE

1034

and eggs from superovulated MT-IGF-I Tg females implanted into pseudopregnant nontransgenic females, employing established methods (20). The MT-IGF-I and the GH-DT Tg mice bear the full designations, Tg(Mt-l, IGF-l)Bri45 and Tg (GH, DT-A -I- Mt-1, GHRF)Bri78, respectively. The GH-DT mice are also transgenic for a fusion of the mMT-I promoter and a GH releasing factor minigene, that was co-injected in an attempt to stimulate the expression of the GH-DT fusion gene (19). Because both Tg mouse lines are hemizygous for their respective transgenes, progeny with four genotypes result, and are denoted herein by the transgene(s) that are integrated in their genomes: DT/-, DT/IGF-I, -/IGF-I, and - / - (normal), with the first annotation referring to the presence or absence of the GH-DT transgene and the second to the MT-IGF-I transgene. Genotype was determined by DNA dot hybridization, using nick translated labeled transgenes as probes (20). In vitro fertilization was performed on three separate occasions. In one group of mice, called experiment (exp.) 1, body weights and tail lengths were measured at weekly intervals. At 56 days of age, mice were anesthetized by injecting ketamine hydrochloride (Quad Pharm. Inc., Indianapolis, IN), 40 mg/100 g body weight, ip. Blood was collected by retro-orbital puncture, serum separated by centrifugation and stored at —20 C. Brain, duodenum, liver, lung, and pancreas were dissected, weighed, and immediately frozen in liquid nitrogen. Pituitaries also were collected and placed in Carnoy's fixative. Finally, to assess accumulated bone growth, right tibias were dissected and placed in absolute ethanol. These mice had been injected with tetracycline hydrochloride (Lederle Laboratories, Wayne, NJ), 1 mg/100 g body weight ip, on day 44 of age; therefore, the changes in tibial epiphyseal width, measured by a described histiologic method (21), were assessed during a 12-day period (days 44-56 of age). Mice from the other two in vitro fertilizations, exp. 2 and 3, represented our initial experiments and were less completely studied: body weight and tail length were serially measured (tail lengths were not measured in exp. 2 mice), and serum was collected at 56 days of age. These experiments were performed in accordance with IACUC approved protocols at the Universities of North Carolina-Chapel Hill and the University of Pennsylvania. Histology

To assess the expression of the GH-DT transgene, fixed pituitaries from mice in exp. 1 were embedded in paraffin and immunocytochemistry was performed to identify GH-containing cells using a described, indirect immunofluorescent method (19) that employed a monkey-derived antibody to rat GH (obtained from Dr. A. F. Parlow through the NIADDK distribution program). The number of GH-containing cells was estimated by counting all the immunostained cells in every fifth deparaffinized 5-/*m section and multiplying the result by 5. To investigate microscopic phenotypic differences, multiple tissues (brain, gonad, heart, liver, lung, muscle, and rib cage) from mice of each genotype were fixed in Carnoy's and processed for histologic exam by standard methods. Measurement of IGF-I IGF-I was quantified with a heterologous RIA using a specific antibody raised against human IGF-I (hIGF-I), as described

Endo • 1990 Vol 127 • No 3

previously (22-24). The antibody does not distinguish between human and mouse IGF-I (mIGF-I), which are 94% homologous. Because pure mIGF-I is not available to react with our antibody, we do not know whether our antibody recognizes each species of IGF-I equally; therefore, results are expressed as relative values per ml serum by assigning the mean IGF-I concentration per ml serum from normal mice a value of 1. Normal mice of 45-60 days of age have approximately 360 ng/ ml IGF-I when compared to purified hIGF-I (15). DNA, RNA, and protein assays Tissues (10-20 mg) were sonicated in 250 n\ Of 1 N NaOH and protein content was measured by the method of Bradford (25). Total nucleic acid (TNA) content was measured by spectrophotometry (Beckman DU-65, Fullerton, CA) using a program employing coefficients determined by Warburg and Christian (26). DNA was measured fluorometrically using a DNA fluorometer (model TKO 100, Hoefer Scientific Instruments, San Francisco, CA) (27). RNA was calculated by subtracting the DNA concentration from that of TNA. TNA preparation Frozen tissues (100-200 mg) were homogenized in 4 ml of a solution containing 0.1 mg/ml proteinase K (BoehringerMannheim, Indianapolis, IN), l x SET (1% sodium dodecyl sulfate, 10 mM Tris, pH 7.5, and 5 mM EDTA), and 0.2 M NaCl. These homogenates were then incubated for 1 h at 68 C, phenol/ chloroform/isoamyl alcohol (25:24:1) extracted, chloroform extracted X 3, and precipitated with a double volume of ethanol. The pellet was then resuspended in 0.2x SET (28). TNA and DNA were quantified, and RNA concentrations were calculated on each sample, as described above. Solution hybridization/S1 nuclease protection assays Liver IGF-I and metallothionein transcripts were quantified in TNA preparations employing a previously described solution hybridization/Si nuclease protection assay method (28, 29). T4 kinase 32P end-labeled synthetic oligonucleotides (oligomers) were used as probes and in vitro transcribed sense strand RNAs were used as standards (except in the case of metallothionein; see below). To assay transcripts derived from the transgene, we used a 24-base oligomer, 5' GCATGTCACT CTTCACTCCT CAGG 3', complimentary to bases 508-533 of the hIGF-IA cDNA encoding a portion of the 3' untranslated end of the mRNA as a probe. Standard was generated from a plasmid (Bluescript; Stratagene, LaJolla, CA) containing a 337-base pair (bp) hIGF-I cDNA fragment (Rsal to Pstl) with complimentary sequence (30). This oligomeric sequence was chosen because it contains 12 base mismatches with the homologous mouse sequence (31), and under the conditions used does not hybridize with mouse IGF-I mRNA. To assay endogenous mouse IGF-I transcripts, we used a 32 base oligomer, 5' CGGGCTGCTT TTGTAGGCTT CAGTGGGGCA CA 3', complimentary to sequence encoding the carboxy terminal end of the mature mIGF-I molecule (31) as a probe, and a 150-bp cDNA fragment with the complimentary sequence in pGEM (Promega, Madison, WI) to generate RNA standards (32). This

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 19 November 2015. at 06:45 For personal use only. No other uses without permission. . All rights reserved.

IGF-I EXPRESSION IN TRANSGENIC MICE oligomer had 9 mismatches with comparable sequences in the transgene and was capable of hybridizing with less than 3% of hIGF-I mRNA under the conditions of the assay. Quantification of IGF-I mRNA in TNA preparations was validated by demonstrating that nearly identical values are obtained when poly(A+) RNA and TNA prepared from the same tissues are assayed. In addition, when known quantities of in vitro transcribed IGF-I mRNA is added to TNA preparations, it is accurately quantified. Liver metallothionein-1 mRNA was estimated by a described method (33), using an end-labeled 25base oligomer directed at the 3' UT region of mMT-1 mRNA as a probe and single stranded M13 DNA containing the mMT1 cDNA as a standard. In each assay total nucleic acids with experimentally determined quantities of DNA and RNA were assayed in duplicate (28). Results are expressed as molecules of RNA/cell for all mice from each genotype, assuming 6.4 pg of DNA per cell. Statistics Data were analyzed using Systat programs (Systat Inc., Evanston, IL) to perform analysis of variance and Pearson correlations.

Results Pituitaries from mice in exp. 1 were evaluated for GHcontaining cells (Fig. 1). Mice not expressing the GHDT transgene (-/IGF-I and -/-) had greater than 200,000 GH immunostaining cells/pituitary, and, with one exception, mice expressing the GH-DT transgene (DT/- and

FIG. 1. Pituitary GH Immunocytochemistry. Indirect immunofluorescence photomicrographs of a normal (-/-) pituitary {panel A), the pituitary of the DT/- mouse with partial expression of the GH-DT transgene {panel B; see Results), and a typical pituitary section of a DT/- or DT/IGF-I mouse {panel C), showing several immuno-positive cells.

1035

DT/IGF-I) had less than 250 such cells/pituitary (range 0-250). The pituitary of one DT/- mouse had approximately 2,500 GH-containing cells (Fig. 1, panel B). This mouse exhibited a normal abundance of hepatic mIGF-I mRNA (1550 mol/cell; see Table 2), a normal serum IGF-I concentration (1.2; see Table 3), and grew normally. These data indicate that 2,500 somatotrophs are sufficient to maintain normal growth. Because this mouse was not GH-deficient, data from this animal was excluded of the analysis of DT/- mice. Mice of each genotype appeared similar in size at birth. Expression of the MT-IGF-I transgene increased the growth of GH-deficient mice to that of normal littermates (Fig. 2, panels C and D). The DT/IGF-I mice weighed more and had longer tail lengths than the DT/ - mice at 30 days of age and thereafter (body weight at 30 days = 15.2 ± 3.2 g, mean ± SD, in DT/IGF-I and 9.88 ± 1.9 g in DT/- mice, P < 0.05, tail length at 30 days = 62.2 ± 6.9 mm in DT/IGF-I and 49.8 ± 5.8 mm in DT/- mice, P < 0.05 at 40 days of age and P < 0.001 at 49 and 56 days of age for both body weight and tail length), and the increased growth of the DT/IGF-I mice was apparent by 15 days of age (Fig. 2, panels A and B). The growth of DT/IGF-I mice did not differ for either parameter from normal (-/-) mice at any time studied. At 56 days of age, the female -/IGF-I mice weighed more than their normal female littermates (-/-; P < 0.05). Although there were no other significant differences between -/IGF-I and - / - mice, the mean body weights and tail lengths of both male and female -/IGF-I mice were greater than those of normal littermates at 30 days of age and after, suggesting that significant differences would have been apparent had more mice been available for study. To be certain that tail length growth reflected bone growth, tibial epiphyseal accumulated growth over a 12day period (days 44-56) was assessed. Growth of the tibial epiphysis was much less in the DT/- mice (47.6 ± 54.6 Aim, mean ± SD; P< 0.005) than in any other group (DT/IGF-I, 270.4 ± 151; -/IGF-I, 298.3 ± 79; and -/-, 262.9 ± 114.5). To further evaluate growth among these mice, selected organs were assessed for size (Table 1). Brain size provided the most dramatic phenotypic difference among the genotypes in that weights differed in mice of each genotype (P < 0.0001), and both groups of mice expressing the MT-IGF-I transgene (DT/IGF-I and -/IGF-I) had greater brain weights than their siblings without the transgene (DT/- and - / - ) . Liver weights exhibited a different pattern among the genotypes. Weights were less in mice not expressing GH (DT/- and DT/IGF-I) compared to siblings expressing GH (DT/- < -/-, P < 0.001; DT/IGF-I < -/IGF-I, P < 0.0005). In addition, the livers of GH-deficient mice appeared to comprise a

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 19 November 2015. at 06:45 For personal use only. No other uses without permission. . All rights reserved.

IGF-I EXPRESSION IN TRANSGENIC MICE

1036

Endo • 1990 Vol 127 • No 3

TABLE 1. Size of organs at 56 days of age Genotype Organ Brain: Weight % body weight Liver: Weight % body weight Cell number Protein/DNA Duodenum Weight % body weight Pancreas Weight % body weight Lung: Weight % body weight Cell number Protein/DNA

DT/-

DT/IGF-I

-/IGF-I

333 ± 11** 2.88 ± 0.12

594 ± 45** 2.78 ± 0.75

739 ± 35** 2.72 ± 0.44

361 ± 69* 3.11 ± 0.47 73.9 ± 28* 158 ± 6

808 ± 3.4 ± 153.4 ± 139 ±

1125 ± 4.06 ± 195.2 ± 174 ±

165 0.73 15 20

170 0.38 39 20

-/469 ± 38** 2.06 ± 0.22 946 ± 4.1 ± 160.7 ± 162 ±

199 0.57 43 20

105 ± 12* 0.9 ± 0.1

196 ± 49 0.9 ± 0.2

250 ± 42 0.9 ± 0.2

196 ± 58 0.8 ± 0.2

55 ± 23* 0.5 ± 0.2

175 ± 60 0.8 ± 0.2

240 ± 67 0.9 ± 0.2

173 ± 32 0.8 ± 0.2

183 ± 0.83 ± 32.1 ± 182 ±

256 ± 85 0.92 ± 0.26 43.9 ± 18.2 166 ± 92

216 ± 0.94 ± 33 ± 165 ±

131 ± 1.14 ± 14.95 ± 167 ±

28 0.24 7.35 60

43 0.18 16.5 156

72 0.26 16 65

Organ weights are reported in mg, cell number in millions, and protein to DNA ratio (protein/DNA) in fig/ng. Means ± SD are shown for all mice in each genotype, because there were no differences between males and females. N = 5 for DT/- and DT/IGF-I mice, 11 for -/IGF-I mice, and 9 for - / - mice. Differences among a genotype and all other genotypes are denoted by the following symbols: * = P < 0.01, and ** = P < 0.0001. In addition, some measures in DT/- mice (liver % body weight, lung weight, and lung cell number) were different from those in -/IGF-I mice (P < 0.05) or those of - / - mice (liver % body weight, P < 0.01); and finally, liver weight and prot/DNA ratio were lower in DT/IGF-I mice than - / IGF-I mice (P