Insulin gene expression in chicken ontogeny: Pancreatic, extrapancreatic, and prepancreatic

DEVELOPMENTAL

BIOLOGY

132,410-418

(1989)

Insulin Gene Expression in Chicken Ontogeny: Extrapancreatic, and Prepancreatic JOSE SERRANO,

CHARLES

L. BEVINS, SCOTT W. YOUNG,*

AND FLORA

Pancreatic, DE PABLO

Receptors and Hormone Action Section, Diabetes Branch, National Institute of Diabetes, Digestive and Kidney Diseases, and *Laboratory of Cell Biolcgg, National Institute of Mental Health, National Institutes of Health, Bethesda, Maryland 20892 Accepted Nmxmber 28, 1988 Insulin has metabolic, growth, and differentiation effects in chicken embryos in vivo and it is required for normal development. Whether the pancreas is the sole source of insulin in embryogenesis is controversial. In the present study we investigated (1) the developmental pattern of expression of the chicken insulin gene in the pancreas; (2) the expression of the insulin gene in three nonpancreatic tissues, liver, brain, and lower limb, during chicken development; and (3) the expression of the insulin gene at prepancreatic stages and during chicken embryo organogenesis. Hybridization of synthetic species-specific insulin oligonueleotides to pancreatic frozen section in situ and to Northern blots revealed a major increase in insulin messenger RNA (mRNA) levels during the third week of embryonic development. The hybridization histochemistry showed both an increase in the levels of insulin mRNA per pancreatic islet and, in addition, an increase in the number of insulin mRNA containing islets with development. By Northern analysis there was a major polyadenylated transcript of 0.6 kb, which increased in abundance approximately 30-fold during this interval. Under the same stringency conditions used for pancreatic RNA an insulin transcript was detected in liver RNA blots. The abundance of this hepatic insulin mRNA was about loo-fold less than the pancreatic insulin mRNA and, in contrast to the latter, did not increase in late development. Primer extension experiments demonstrated that the insulin transcripts of pancreas and liver had similar 5’ ends. No insulin mRNA was detected by Northern analysis or primer extension either in whole brain or lower limb total RNA from several developmental stages. A very low abundance insulin mRNA was detected in whole embryo at Day 8 and body regions at Day 4 and Day 5 when organogenesis of the pancreas takes place. Interestingly, a polyadenylated insulin transcript was detected, as well, in whole Day 2 and Day 3 embryos (stages 10 to 20, with 20 to 40 somites) before differentiation of p cells occurs. Thus, there is differential developmental regulation of the insulin gene in several chicken embryo tissues and the expression of insulin preceeds pancreatic maturation. These findings support the proposed role of insulin in differentiation and development in vivo and suggest a paracrine type of action of the hormone in early embryos before blood circulation begins. 0 1989 Academic Press, Inc. INTRODUCTION

Depending upon cell type, developmental age, and state of differentiation of the cells, insulin can stimulate mitogenesis or differentiation in embryonic tissues in culture (Ridpath et ab, 1984; Milstone and Piatigorski, 1977; Ewton and Florini, 1981; Puro and Agardh, 1984). These in vitro studies suggest that insulin may be a multifunctional growth and differentiation factor in development. While studying the effects of exogenous insulin on embryo tissues gives information on their potential ability to respond to the hormone, it is critical to understand the expression of the endogenous gene to define the role of insulin in normal development. The chicken embryo allows studies of early organogenesis. We have previously shown in this model that insulin may be essential for normal early development (Days 2 to 5) (De Pablo et al., 1985). We have also detected insulin immunoreactivity and bioactivity in embryos at Day 2 of development (De Pablo et al, 1982), at the onset of organogenesis, prior to pancreatic /3 cells’ 0012-1606/89 $3.00 Copyright All rights

0 1989 by Academic Press, Inc. of reproduction in any form reserved.

410

differentiation. In addition, the Day 2 whole embryo and all early embryo tissues that we have analyzed (brain, liver, heart, lens, limb buds, muscle) display insulin receptors which are regulated in chicken ontogeny in a tissue-specific manner (Bassas et ah, 1987a,b). To examine in more detail the possible role of insulin in growth and differentiation of these tissues we have started to explore whether the pancreas is the sole site of insulin synthesis or whether other embryonic tissues may express the insulin gene. In addition, we studied the possible insulin mRNA expression in the whole embryo at prepancreatic stages (Days 2 and 3). In the present study we demonstrate the presence and determine the size of chicken insulin mRNA in two tissues: the pancreas and the liver. In situ hybridization, Northern blot analysis and primer extension were carried out using specific chicken insulin probes. Pancreas and liver showed differential developmental regulation of the steady-state levels of insulin mRNA. Although in extremely low levels compared to postnatal pancreas (about three to four orders of magnitude less),

SERRANO ET AL.

Insulin

mRNA

there is also insulin mRNA in embryos prior to pancreatic differentiation and the beginning of blood circulation.

--)

MATERIAL3

in Chick y”~l?,S 37jlj

411

Embryos

ATG 166

AND METHODS

Chicken Embryos and Organs

Fertilized eggs of white Leghorn chickens were purchased from Truslow Farms, Inc. (Chestertown, MD) and incubated at 37.5”C, 60-90% relative humidity in an egg incubator. On selected days of the 21-day period of chicken development until hatching, embryos were staged according to Hamburger and Hamilton (1951). Whole embryos, bodies minus head, or specific organs were dissected on wet ice, washed in cold saline, and immediately frozen on dry ice. Tissues from 3 to 50 embryos, depending on the age, were pooled before isolation of RNA or were processed for hybridization histochemistry. Isolation of RNA

Frozen tissues were homogenized for 30 set at high speed with a Polytron homogenizer (Brickman Instruments, Westbury, NY) in 6 M guanidinium thiocyanate (Fluka, Ronkonkoma, NY) and total RNA was extracted according to the method of Cathala et al. (1983). The RNA concentration and purity were determined by absorbance at 260 nm and the 260 nm:280 nm ratio, respectively. Polyadenylated RNA was prepared by affinity chromatography using poly(dT)-cellulose (BRL, Gaithersburg, MD) following the protocol of the manufacturer. Gel Electrophoresis and Northern Blot Hybridization

RNA was resolved by electrophoresis in 1% agarose gels in the presence of formaldehyde (Maniatis et al, 1982) and blotted to nylon membranes (GeneScreen, DuPont, Boston, MA) by capillary blotting (Thomas, 1980) or by electroblotting. For this study we synthesized, based on the genomic sequence (Perler et al., 1980), two oligodeoxynucleotides complementary to the chicken insulin mRNA (Fig. 1). These 48-base oligomers, 1 and 2, correspond to protein coding and noncoding regions: 1,5’GGCTTTGGGGGAGTAGAAGAAGCCACGCTCTCCACACACCAGGTAGAG3’, complementary to the mRNA coding for the last 16 amino acids of insulin B chain; and 2, S’AGAGTAAGTGTATGTCTGTGCCCGCTTCTGGCTTCTTGGCTAGTTGCA3’, complementary to the mRNA coding for the last 2 amino acids of the insulin A chain and 42 nucleotides of the 3’ untranslated region (UTR) (both oligomers were kindly synthesized by Dr. M. J. Brownstein, NIMH). These oligomers were used as specific probes to

cRNA

FIG. 1. Schematic representation of the chicken insulin gene (redrawn from Perler et al., 1980). The 48-base oligonueleotides and the complementary RNA (cRNA) used as probes are indicated below the gene.

study insulin gene expression in pancreas and extrapancreatic tissues by Northern blot analysis, in situ hybridization histochemistry, and primer extension. Blots were hybridized with the oligonucleotides 32P-5’-end-labeled to a specific activity of 0.5 X 10” cpm/pmole using [T-~~P]ATP (Amersham, 3000 mCi/mmole) and T4 polynucleotide kinase (BRL). The hybridization was carried out in 50% formamide, 4~ SSC (1X SSC: 0.15 M NaCl, 0.015 M sodium citrate), 1X Denhardt (0.02% polyvinylpyrrolidone, 0.02% Ficoll, 0.02% bovine serum albumin), 1% SDS, and 10 pg/ml ssDNA for 16 hr at 37°C. Filters were washed twice at 37°C with 50% formamide, 4~ SSC, 1% SDS and once at 3’7°C with 2X SSC, 1% SDS, for 20 min each, and subsequently exposed to XAR-2 film (Eastman Kodak, Rochester, NY) with intensifying screens (Cronex lightning Plus, DuPont) at -70°C for 2 days to 2 weeks. The chicken type I collagen probe was a cDNA (pCg45) kindly provided by Dr. Helga Boedtker (Department of Biochemistry and Molecular Biology, Harvard University). It was labeled by random priming to a specific activity of 10’ cpm/pg and hybridized to Northern blots as described above. The stringency wash performed was in 0.1X SSPE, 0.1% SDS at 40°C. In situ Hybridization

Histochemistry

For in situ hybridization histochemistry, 12- to 14-pm pancreas and liver frozen sections were cut with a Lipshaw microtome (Lipshaw, Detroit, MI). The sections were mounted on gelatin double-coated glass slides. The oligomers were labeled with [a-35S]dATP (New England Nuclear) and terminal deoxynucleotidyltransferase (BRL) to a specific activity of 0.5 X lo6 cpm/pmole. The hybridization and wash conditions were according to Young et al. (1986). The tissue sections were covered with NTB-3 photographic emulsion (Eastman Kodak, Rochester, NY) and exposed at 4°C for the indicated length of time. They were processed, using Dektol 19 developer and Rapid fixer (Eastman Kodak) following manufacturer’s instructions. The sections were counterstained with 0.2% toluidine blue (Sigma, St. Louis, MO) and photographed under bright-field with a Zeiss

412

DEVELOPMENTALBIOLOGY V0~~~~132,1989

optical microscope. For a low power magnification view of tissues we obtained direct autoradiograms of the sections on XAR-2 film prior to the exposure to photographic emulsion. Primer

gram prior to RNAase treatment. The filter was subsequently incubated with 2X SSPE, 10 pg/ml RNAase A at 50°C for 1 hr and washed twice with 1X SSPE, 1% SDS, 50 mM MgClz for 20 min at 50°C. The filter was again exposed to XAR-2 film at -70°C for 7 days.

Extension

Total RNA (50 pg) was mixed with 0.3 pmole of a 48mer primer (oligodeoxynucleotide 1 labeled with [y32P]ATP and T4 polynucleotide kinase as described above) in hybridization buffer (50% formamide, 0.4 M NaCl, 40 mM Pipes, pH 6.4) and was incubated overnight at 37°C. After precipitation with 2.5 vol of ethanol added, the RNA-DNA hybrid was resuspended in primer extension reaction buffer (50 mM Tris base, pH 8.3, 6 mM MgCIB, 50 mM KCl, 10 mM dithiothreitol, 0.5 mM each of dATP, dGTP, dCTP, and dTTP (Pharmacia). Actinomycin D (100 fig/ml) and AMV reverse transcriptase (25 units; Boehringer-Mannhein, Indianapolis, IN) were then added and the mixture was incubated for 1 hr at 42°C. The reaction mixture containing the extended cDNA was precipitated with ethanol and subsequently resuspended in 90% formamidei0.2 M NaOH. The sample was heated to 85°C for 3 min and loaded onto a 5% polyacrylamideL3 M urea gel. The gel was run at 60 W for 2 hr, dried, and exposed to film at -70°C. Filter-RNAase

Protection

Assay

An insulin cRNA probe was generated by standard subcloning procedures as follows: a 166-bp BamHlZ’aql genomic DNA fragment was isolated from chicken insulin clone cl15 (phage kindly provided by Dr. A. E. Efstratiadis, Columbia University, New York); the DNA was purified by gel electrophoresis. This fragment was then ligated to a pGem2 vector (Promega Biotec, Madison, WI) previously digested with BamHl and HincII. For the generation of antisense riboprobes which would hybridize to mRNA this construct was linearized with EcoRl and gel purified. Synthesis of uniformly 32P-labeled antisense RNA was performed with T7 RNA polymerase (Promega Biotec) following the protocol of the manufacturer. A Northern blot containing 20 pg of pancreatic total RNA and 50 pg of each Day 2 embryos poly(A)+, Day 2 embryos poly(A)-, Day 3 embryos poly(A)+, Day 3 embryos poly(A)-, and Day 12 limb total RNA was prehybridized and hybridized (in 50% formamide, 4X SSPE, 1X Denhardt, 0.1% SDS, 200 fig/ml sperm salmon DNA, at 50°C) with a uniformly labeled, high specific aciivity insulin cRNA probe. The filter was then washed at high stringency, i.e., 0.1~ SSPE, 0.1% SDS, 15 min at 5O”C, the last wash. The filter was apposed to a XAR-2 Kodak film at -70°C for a few hours to obtain an autoradio-

Densitometry

The specific bands in Northern blots and the signal obtained in occasional slot blots were quantitated by densitometric analysis with a GS-300 densitometer (Hoefer Scientific Institute, San Francisco, CA). RESULTS

Insulin

mRNA

in Developing

Pancreas

In situ hybridization histochemistry. Insulin mRNA expression in the pancreas was compared between growing chickens (3 weeks posthatching) and middevelopment embryos (Day 11). Despite similar spatial patterns, the levels of expression were dramatically different (Fig. 2). The embryonic pancreas required four times longer exposure (1 month) than the posthatched chicken pancreas (1 week) and still gave a much less intense signal. This quantitative difference, reflecting at least a 20-fold difference in insulin mRNA, was further studied by filter hybridization (see below). In addition to lower insulin mRNA per islet, the embryonic pancreas appeared to have less islets, roughly half the number, containing insulin mRNA-producing cells. The islet’s specific signal was effectively competed when hybridization was carried out in the presence of an excess of unlabeled oligonucleotide and was not observed using as a labeled probe an oligonucleotide “sense” to the insulin mRNA sequence (data not shown). Northern blot hybridization. When pancreatic total RNA from developing embryos (Days 10 to 18) and posthatched chickens was analyzed by Northern blotting, a major transcript of 0.6 kb was detected in all samples (Fig. 3). With longer exposure, another transcript of 2.4 kb was also detected (Fig. 4, pancreas lane). The insulin mRNA from embryonic and posthatched chicken pancreas was polyadenylated, since affinity chromatography on poly(dT)-cellulose resulted in its enrichment (result not shown). The specific insulin mRNA signal increased approximately 30-fold from Day 10 embryos to 3 weeks posthatching. Similar to the results obtained by the hybridization histochemistry, the level of expression in pancreas from Day 10 to Day 12 embryos is quite low. A major increase in the level of insulin mRNA is evident by Day 16 of development.

SERRANO ET AL.

Ins&n

mRNA

in Chick

Embryos

FIG. 2. Proinsulin mRNA detected by in situ hybridization histochemistry in chicken pancreas. (A) Section from 3-week posthatched chick. (B) Section from Day 11 (stage 37) chicken embryo. The microphotographs shown are bright-field images at a magnification of 100X. Note the exquisite localization of the signal in, presumably, the islet cells. Sections were hybridized with oligomer 1 as described under Materials and Methods and exposed either 1 week (A) or 1 month (B) at -70°C. Comparable results were obtained with Day 18 embryo pancreas.

Insulin Gene Expression in Embryonic Liver

In searching for tissues which might express the insulin gene during embryogenesis, we analyzed total and poly(A)+ RNA from liver and brain by Northern blotting. Hybridization of oligomer 1 to embryonic liver total RNA obtained from Day 12 to Day 21 embryo is shown in Fig. 4. A specific signal similar in size to the main pancreatic transcript was present in liver at Days 12 and 18 and was barely detectable at Day 21. Since hybridization and washes were carried out at high stringency conditions for a 4%base oligonucleotide (50% formamide, 4~ SSC, 1% SDS, 4O”C), the degree of similarity of this hepatic transcript to the probe must be high. No insulin transcript was detected in an equal amount of total RNA extracted from embryonic limb or brain. To confirm the specificity of the hepatic transcript, the blot was rehybridized with the oligomer 2 (complementary to the 3’ UTR region of the insulin gene, see Fig. l), under stringent hybridization conditions (same buffer, 50°C). An identical pattern, al-

though with even weaker signal, was obtained (data not shown), indicating that the liver transcript contains similarities to the pancreatic transcript also in the 3’ untranslated region. The abundance of the putative insulin mRNA in liver relative to pancreas is approximately loo-fold less. To assess the integrity of the RNA from limb tissues and the amounts of total RNA in the liver samples, the blot was rehybridized to a chicken type I collagen probe. A transcript of 5.7-5.1 kb was detected; it was very abundant in the limb, low but present at comparable levels in all the liver samples, and absent in the pancreas. Primer extension analysis. It was of interest to compare the embryonic liver and pancreas transcripts with regard to their 5’ end. Because the chicken proinsulin mRNA has not been previously characterized by cDNA cloning, it was also interesting to verify the length of the putative 5’ end which had been assigned based on analogy with the rat insulin cDNA (Perler et ah, 1980). cDNA primer extension was performed with oligonucleotide 1 and AMV reverse transcriptase. The oligonucle-

414

DEVELOPMENTAL BIOLOGY

-n AGE

10

Embryos 12

Post-hatched

(day) 16

18

0

3 weeks

- 2% -18s

p;

c

FIG. 3. Northern blot analysis of pancreatic RNA during chick development. Total RNA, 20 pg/lane, was electrophoresed in a denaturing agarose gel, blotted to a nylon filter, and hybridized with the oligomer 1 as described under Materials and Methods. The arrow indicates the major transcript detected, of 0.6 kb. Similar results were obtained with the other oligonucleotide. The filter was rehybridized sequentially to a P-actin and a 28 S rRNA probe. There is some regulation of /?-actin mRNA in development and the signal obtained at Days 10 and 12 was lower than that at later stages. The hybridization to the 28 S rRNA probe confirmed that similar RNA amounts had been loaded in all lanes (results not shown). The mobility of the 18 and 28 S ribosomal RNAs is indicated. The autoradiogram was exposed for 2 days at -70°C with one intensifying screen.

otide, complementary to the 48 bases encoding the 16 carboxyterminal amino acids of the insulin B chain, was end-labeled with 32P. The expected extended product length, assuming initiation of transcription at the putative start site proposed by Perler et ah (1980), was 219 nucleotides. RNAs from developing pancreas (Day 18), liver (Day 18), and lower limb (Day 12) were used as templates. An extended cDNA of similar length was obtained in the two insulin-expressing tissues, pancreas and liver, indicating that the 5’ ends of insulin mRNA transcripts are comparable. The cDNA, however, appeared slightly longer than expected which may in part be due to differences in migration between the standard phage DNA digest and our cDNA in a gel of limited resolution. The sensitivity of this technique was lower than that with Northern analysis and not as quantitative. Nevertheless, the liver from Day 18 embryos showed an extended product, many-fold lower in abundance compared to the one detected in pancreas (Fig. 5). No extended product was obtained with total RNA extracted from limb (and neither in brain, not shown). In situ hybridization. In an attempt to exclude gross tissue contamination of the liver preparation with pancreas and confirm by an independent technique the expression of insulin mRNA, we performed in situ hybridization experiments. We analyzed multiple large sections of liver tissue at the macroscopic level. The frozen sections were hybridized to oligonucleotide 1 as a probe. Two types of controls were used to assess the specificity of the signal: (a) competition of the radioactive labeling with excess unlabeled oligonucleotide, and (b) hybridization with an oligonucleotide complementary to oligo-

VOLUME 132,1989

nucleotide 1, therefore nonhybridizable to insulin mRNA. As shown in Fig. 6, the radioactive grains display a diffuse pattern of insulin mRNA expression in liver of embryos at Day 16 and Day 18, very different from pancreas. In addition, there is partial competition of the signal in the presence of unlabeled oligonucleotide (Fig. 6B) and reasonably low nonspecific hybridization (Fig. SD). Insulin

Gene Expression

in Early Organogenesis

The low levels of expression of insulin mRNA in Day lo-11 pancreas and in Day 12 liver and the difficulty of accurately dissecting individual organs during earlier stages of organogenesis led us to prepare large amounts of poly(A)+ RNA from whole Day 8 embryos and from the body region of Day 4 and Day 5 embryos. Northern blot hybridization to the 32P-labeled oligomer 1 under the same conditions used for the pancreatic Northern blot, revealed a very low abundance transcript of similar size to the pancreatic 0.6-kb insulin mRNA (Fig. 7). Quantitatively, this transcript changed very little from Day 4 to Day 8. Two larger bands were apparent in the

Liver

* Embryo

age

12

18

21

Leg

Pancreas

12

Bweeks

posth.

- 28s c -1%

FIG. 4. Northern blot of insulin in developing liver. (A) Total RNA, 50 pg from embryonic livers and lower limbs (days are indicated) and 30 pg from pancreas (3 weeks posthatched), was hybridized with oligomer 1. The conditions were the same as for the blot in Fig. 3. The chicken embryonic limb RNA and the pancreatic RNA were used as expected negative and positive controls, respectively. The arrows indicate the major and minor pancreatic transcripts detected in this overexposed autoradiogram (exposed for 15 days at -70°C with one intensifying screen). The broad band in Day 12 liver may be due to slight degradation of the RNA sample as indicated by the ethidium bromide stain of the gel. The signal in Day 21 liver was weak but present in the original autoradiogram. The mobility of the 18 and 28 S rRNAs is indicated. (B) The same blot was rehybridized to a type I collagen cDNA and a 5.7- to 5.1-kb transcript was readily detected in the limb lane. A much weaker signal is visible in all the liver samples, while no detectable transcript is present in the pancreas. Exposure of this autoradiogram was for 2 days at -70°C.

SERRANO ET AL.

Pancreas Limb El8 El2

Liver El8

Insulin

mRNA

in Chick

415

E-OS

DISCUSSION

M - 281 - 271

- 234

-194

FIG. 5. Primer extension of insulin mRNA. Fifty micrograms of total RNA from liver (Day 18 embryo), lower limb (Day 12 embryo), and pancreas (Day 18 embryo) was hybridized with oligonucleotide 1 (see Figure 1). 32P-5’-end-labeled $X174 Hoe111 DNA digest was used as size marker on a urea-polyacrylamide gel. The extended cDNA expected based on genomic sequences was 219 nucleotides in length. Note that in both pancreas and liver there are some partially extended products, while the full length products appear between the 234- and 194-bp markers (arrow). In a repetition of this experiment the full length was the predominant band. The intensity of the signal is weaker in liver, as expected based on the abundance of insulin mRNA demonstrated in RNA blots. There were no extended products with RNA from limb, which was estimated intact by hybridization to a collagen probe (see Fig. 4) even if the background in this photograph is lower in the limb lane than in the positive tissues. The gel was dried and exposed at -70°C for 10 days.

The chicken insulin gene (Perler et al, 1980) is a single copy gene, structurally similar to the rat insulin II gene. No studies on insulin gene expression have been reported in chicken species, either in development or postnatally, and no insulin cDNA has been cloned. We have used a variety of techniques and several types of probes (synthetic insulin oligodeoxynucleotides and a cRNA) to analyze the localization, size, and developmental regulation of the insulin mRNA in chicken embryo tissues. In the developing pancreas, the insulin gene is expressed as a 0.6-kb polyadenylated mRNA. This size (based on the relative mobility of the 28 and 18 S rRNAs and rat insulin mRNA) is similar to the rat insulin I and II mRNAs (Cordell et ak, 1982) and the human mRNA (Bell et al, 1980), but smaller than the anglerfish 0.8-kb insulin mRNA (Hobart et al, 1980).

intentionally overexposed blot. The larger coincided with the migration of the 28 S ribosomal RNA, which hybridized strongly to a human 28 S ribosomal probe (data not shown), suggesting nonspecific hybridization of the insulin oligomer to rRNA. An intermediate band was detected in between the 28 and 18 S rRNAs and may represent specific hybridization to another insulin mRNA species-perhaps similar to the less abundant transcript found in the pancreas (Fig. 4). Insulin

Gene Expression

in Prepancreatic

Embryos

Northern blot hybridization of large amounts of RNA with oligonucleotide probes (Fig. 7) was not fully satisfactory because we could not apply very stringent washing conditions and nonspecific hybridization remained high. Thus, we subcloned a 166-bp fragment of the chicken insulin gene in a riboprobe vector and combined the techniques of filter hybridization and RNAase protection to achieve the desired sensitivity and specificity. Using the high specific activity cRNA as a probe (Fig. 1) and very large amounts of poly(A)+ RNA from whole Day 2 and Day 3 embryos, a transcript resistant to RNAase digestion was demonstrated. This transcript, of approximately the same size as the pancreatic insulin mRNA, was absent in the poly(A)- RNA from Day 2 and Day 3 embryos as well as in total RNA from embryonic limbs (Fig. 8).

FIG. 6. In situ hybridization of chicken embryo liver. (A, B) Twelvemicrometer frozen serial sections from Day 18 (stage 44) embryo liver. (C, D) Sections from Day 16 (stage 42) embryo liver. A and C were hybridized to oligonucleotide 1 to detect proinsulin mRNA. B was hybridized to @S-labeled oligonucleotide 1 in the presence of an excess of unlabeled oligonucleotide. D was hybridized to a %S-labeled oligonucleotide complementary to oligonucleotide 1 and, therefore, noncomplementary to insulin mRNA. Thus, the black grains in D represent background as opposed to grains in A and C which represent diffuse signal. The photographs shown are direct prints from the original autoradiogram obtained with X-ray film and, thus, are only intended to show overall intensity of hybridization and macroscopic distribution. This low magnification allows a view of a large piece of liver tissue but does not provide cellular resolution. The sections presented are representative of eight sections per group analyzed.

416

DEVELOPMENTAL BIOLOGY

VOLUME 132,1989

portional to the amount of protein synthetized (MeKnight et al, 1980; Paevy et aa, 1978), the increase in Whole e. Pancreas insulin mRNA observed in late development in pancreas does not correlate with changes in plasma insulin levels, which have been reported low and fairly constant during chicken ontogeny until hatching (Benz0 and Green, 1974; Liebson et al., 1976). This discrepancy re28s flects an important post-transcriptional control of insulin expression and, possibly, that most of the insulin 18s synthesized is stored in the /3 cells for postnatal secre+ tion (Milner et al, 1973). The production of insulin by tissues other than the pancreas has been suspected in chickens since total pancreatectomy did not eliminate circulating insulin (Colca RNA: L30 pg polyp21 pg total and Hazelwood, 1982). In addition it has been a subject FIG. 7. Northern blot of insulin mRNA in early chick embryos. of debate in rat and human tissues (Rozenzweig et al., Thirty micrograms of poly(A)+ RNA from the body region (separated 1980; Muglia and Locker, 1984; Giddings et ah, 1985; Liu from the head by a cut below the otic vesicles) of Day 4 and Day 5 et ah, 1985; Young, 1986). Quite interestingly, the levels embryos and from Day 8 whole embryos were hybridized with oliof insulin mRNA in chicken embryo liver were many gomer 1. Twenty-one micrograms of total pancreatic RNA was used fold lower than those in the pancreas and did not paralas a positive control. The arrow indicates the typical insulin transcript detected in all the samples. lel the pancreatic mRNA changes in late development. It is clear that the liver 0.6-kb signal corresponds to an insulin transcript because it hybridizes under high stringency conditions to two probes (oligomers 1 and 2) Since the putative 5’ and 3’ untranslated regions plus and the similar 5’ cDNA extension of RNA from panthe protein coding region of the chicken insulin mRNA creas and liver. The possibility of gross tissue contamiadd up to 458 bases, the mRNA has an approximately nation with pancreas is very unlikely since four inde150-base long poly(A) tail. contained the With very long exposures of Northern blots there is pendently obtained liver preparations another transcript of 2.4 kb in chicken pancreas RNA. This may be similar to the 2.4-kb transcript also detected in rat pancreas and developing rat yolk sac (Muglia and Locker, 1984). This longer transcript could E2 E3 arise by a number of mechanisms, including transcripq* -2 ~LYIA)+ n n Kb (A,FOLYIAl+ (Altion from a second promoter 5’ to the normal insulin v gene, termination of transcription at further downstream sequence, alternative splicing including sequences within the long second intron of the chicken insulin gene, or represent not fully processed nuclear RNA. A transcript longer than the typical mature mRNA has also been reported in human pancreas (Giddings et al., 1983). The abundance of this second transcript in late chicken development in pancreas is very FIG. 8. Northern blot of insulin mRNA in prepancreatic embryos. low relative to that of the 0.6-kb insulin mRNA. Fifty micrograms of poly(A)+ RNA or poly(A)) RNA from whole chicken embryos after 2 days (E2, stage 10 to 14) and 3 days (E3, stage The increase in abundance of the mature 0.6-kb insu18 to 20) of development were fractionated on a formaldehyde gel. lin mRNA in the developing pancreas (at least 30-fold Fifty and thirty micrograms, respectively, of total RNA from embryfrom Day 10 chick embryo to 3 weeks posthatching) can onic limbs and pancreas were loaded as well. The RNA was transbe accounted for by an increase in the level of expresferred to a nylon membrane and hybridized to the cRNA probe indision per islet as well as by an increase in the number of cated in Fig. 1. After stringency washes the filter was treated with functional islets. A combination of both mechanisms is RNAase A as described under Materials and Methods. The autoradiogram generated by the RNAase A-resistant hybrids is shown. Note suggested by hybridization histochemistry of pancrethe presence of an insulin transcript (arrow) in the pancreas, the E3 atic sections from Day 11 embryos and S-week-old poly(A)+ lane, and the E2 poly(A)+ lane. The slight difference in of this chickens (Fig. 2). The functional implications mobility in the E3 sample is probably a reflection of a minor salt increase are unclear. Although it is considered that the difference in the sample which affected the electrophoretic run, based levels of mRNA coding for a protein are roughly pro- on the ethidium bromide stain of the gel. Embryos

AGE ’

4 Body

5 Body

Iday)

Post-hatched

a ‘13

6d

SERRANO ET AL.

Insulin

specific transcript. Also, very different developmental patterns of insulin mRNA expression were found between the two tissues. In addition, preliminary in situ hybridization studies show a diffuse pattern of labeling (Fig. 6), distinct from the pattern obtained in pancreas. Although one cannot completely rule out that the mRNA localized in situ in liver is not insulin but a related species, for instance, an insulin-like growth factor mRNA, the additional confirmation of the specificity of the probe (pancreatic pattern, primer extension, etc.) is reassuring of the identity of the signal. On the other hand, our failure to detect insulin mRNA in whole brain RNA in the present studies should not be interpreted as a final negative. If the insulin gene is expressed at low level in only a few cells in certain areas of the developing nervous system, the techniques used might have been too insensitive to detect it. Ongoing studies by in situ hybridization may prove more adequate to analyze insulin gene expression in the nervous system. In the early embryo at Days 4,5, and 8, the low abundance insulin mRNA could be contributed by a number of tissues: the rudimentary pancreas, liver, and many others. (Attempts to obtain adequate hybridization histochemistry with sagittal frozen sections of whole embryos have repeatedly failed due to the extreme fragility of the early embryo tissues and their high water content.) There are no changes in the insulin mRNA signals between Days 4 and 8 per equal amount of poly(A)+ RNA, which may suggest absence of insulin gene transcription positive regulators or prevalence of negative regulators in the immature p cells or constitutive expression by either p cell precursor or other cells. Finally, the insulin transcript that we detect in Day 2 and Day 3 embryos antedates the differentiation of the pancreatic p cells as assessed by an immunocytochemistry report, exclusively focused in the pancreas (Dieterlen-Lievre and Beaupain, 1976). Although we have not proven that this prepancreatic insulin mRNA is translatable in z&-o, we previously found a material with the immunological and biological activities of authentic chicken insulin in whole Day 2 embryos, as well as in the head and body regions of Day 3 chick embryos (De Pablo et al., 1982). It appears reasonable to interpret that the prepancreatic insulin is the product of the insulin mRNA now detected, thus, embryonic insulin and not maternal, egg-derived hormone. Indeed, we have detected insulin mRNA in chicken ontogeny much earlier than it had been found in mammalian embryos (Kakita et al., 1983). The location and precise function of the prepancreatic insulin remains to be established. It is worth emphasizing that insulin receptors and insulin-like growth factor receptors are abundant in nervous system-related structures during chicken embryo

mRNA

in

Chick

417

Embrgos

neurulation (Girbau M. and De Pablo F., manuscript submitted for publication). Recently, neurons have been shown to transcribe the insulin gene in mammalian neonatal brain (Schechter et al, 1988) as well as in an invertebrate (Smit et al+, 1988). The debate on the origin of pancreatic endocrine cells (Le Douarin, 1988; Alpert et al., 1988) and the debate on the significance of nonpancreatic insulin expression continue. This paper is dedicated to Jesse Roth who, at the beginning of this decade, predicted the findings of insulin early in evolution and ontogeny. We thank Michael J. Brownstein for providing us with synthetic oligonucleotides, Helga Boedtker for the collagen cDNA, Charles T. Roberts for helpful technical advice, and Joan Blanchette-Mackie for help with the microphotography. We are grateful to Ms. Esther Bergman for preparing the manuscript. C.L.B. and S.W.Y. were supported by Pharmacology Research Associate Training Fellowships of the National Institutes of General Medical Sciences. REFERENCES ALPERT, S., HANAHAN, D., and TEITELMAN, G. (1988). Hybrid insulin genes reveal a developmental lineage for pancreatic endocrine cells and imply a relationship with neurons. CeU 53,169-171. BASSAS, L., DE PABLO, F., LESNIAK, M. A., and ROTH, J. (1987a). The insulin receptors of chick embryo show tissue-specific structural differences which parallel those of the insulin-like growth factor I. Endocrinology 121,1468-14’76. BASSAS, L., LESNIAK, M. A., GIRBAU, M., and DE PABLO, F. (1987b). Insulin-related receptors in the early chick embryo: From tissue patterns to possible function. J. Exp. ZooL 1, (Suppl.), 299-307. BELL, G. I., PICTET, R. L., RUTTER, W. J., CORDELL, B., TISCHER, E., and GOODMAN, H. M. (1980). Sequence of the human insulin gene. Nature

(London)

284.26-32.

BENZO, C. A., and GREEN, T. D. (1974). Functional differentiation of the chick endocrine pancreas: Insulin storage and secretion. Anat. Rec. 180,491-496. CATHALA, G., SAVOURET, J. F., MENDEZ, B., WETS, B. L., KARIN, M., MARTIAL, J. A., and BAXTER, J. D. (1983). A method for isolation of intact, translationally active ribonucleic acid. DNA 2,329-335. COLCA, J. R., and HAZELWOOD, R. L. (1982). Persistence of insulin, glucagon and pancreatic polypeptide in the plasma of depancreatized chickens. J. EndoctinoL 92,317-326. CORDELL, B., DRAMOND, D., SMITH, S., PUNTER, J., SCHONE,H. H., and GOODMAN, H. M. (1982). Disproportionate expression of the two nonallelic rat insulin genes in a pancreatic tumor is due to translational control. Cell 31,531-542. DE PABLO, F., GIRBAU, M., GOMEZ, J. A., HERNANDEZ, E., and ROTH, J. (1985). Insulin antibodies retard and insulin accelerates growth and differentiation in early embryos. Diabetes 34,1063-1067. DE PABLO, F., ROTH, J., HERNANDEZ, E., and PRUSS, R. M. (1982). Insulin is present in chicken eggs and early chick embryos. Endowinology 111,1909-1916. DIETERLEN-LIEVRE, F., and BEAUPAIN, D. (1976). Immunocytological study of endocrine pancreas ontogeny in the chick embryo: Normal development and pancreatic potentialities in the early splachnopleura. In “The Evolution of Pancreatic Islets” (T. Adensaya, I. Grillo, L. Leibson, and A. Epple, Eds.), pp. 37-50. Pergamon, Oxford. EWTON, D., and FLORINI, J. R. (1981). Effects of the somatomedins and insulin on myoblast differentiation in vitro. Deu. BioL 86,31-39. GIDDINGS, S. J., CHIRGWIN, J., and PERMUTT, M. A. (1985). Evaluation

418

DEVELOPMENTALBIOLOGY

of rat insulin messenger RNA in pancreatic and extrapancreatic tissues. L3iobetologio 28,343-347. GIDDINGS, S. J., ROTWEIN, P., CHIRGWIN, J. M., SCHARP, D., and PERMUTT, M. A. (1983). Analysis of insulin gene expression in human pancreas. Drizbetes 32,777-780. HAMBURGER, V., and HAMILTON, H. L. (1951). A series of normal stages in the development of the chick embryo. J. Morphol. 88, 49-92. HOBART, P. M., SHEN, L. P., CRAWFORD,R., PICTET, R. L., and RUTTER, W. J. (1980). Comparison of the nucleic acid sequence of anglerfish and mammalian insulin mRNA’s cloned cDNA’s. Science 210, 1360-1363. KAKITA, K., GIDDINGS, S. J., ROTWEIN, P. S., and PERMUTT, M. A. (1983). Insulin gene expression in the developing rat pancreas. Diabeta 32,691-696. LE DOUARIN, N. M. (1988). On the origin of pancreatic endocrine cells. Cell 53.169-171. LIEBSON, L., BONDAREVA, V., and SOLTITSKAYA, L. (1976). The secretion and role of insulin in chick embryos and chickens. In “The Evolution of Pancreatic Islets” (T. Adensaya, I. Grillo, L. Leibson, and A. Epple, Ed.), pp. 69-79. Pergamon, Oxford. LIU, K. S., WANG, C. Y., MILLS, N., GYVES, M., and ILAN, J. (1985). Insulin-related genes expressed in human placenta from normal and diabetic pregnancies. Proc. Nat1 Acad Sci USA 82,3868-3870. MANIATIS, T., FRITSCH, E. F., and SAMBROOK, J. (Eds.). (1982). Extraction, purification and analysis of mRNA from eukaryotic cells. In “Molecular Cloning. A Laboratory Manual,” pp. 187-210. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY. MCKNIGHT, G. S., LEE, D. C., HAMMARPLARDH, D., FINCH, C. A. S., and PALMITER, R. D. (1980). Transferrin gene expression: Effects of nutritional iron deficiency. J. BioL Chem. 255,144-147. MILNER, R. D. G., LEACH, F. W., and JACK, P. M. B. (1973). Reactivity of the fetal islet. In “Carbohydrate Metabolism in Pregnancy and the Newborn” (H. W. Sutherland and J. M. Stower, Ed.), pp. 83-104. Churchill-Livingstone, Edinburgh. MILSTONE, L. M., and PIATIGORSKY, J. (1977). &Crystallin gene ex-

VOLUME132,1989 pression in embryonic chick lens epithelia cultured in the presence of insulin. Exp. Cell Res. 105,9-14. MUGLIA, L., and LOCKER, J. (1984). Extrapancreatic insulin gene expression in the fetal rat. Proc. Nat1 Acad Sci USA 81,3635-3639. PAEVY, D. E., TAYLOR, J. M., and JEFFERSON, L. S. (1978). Correlation of albumin production rates and albumin mRNA levels in livers of normal, diabetic, and insulin-treated diabetic rats. Proc Nat1 Acad Ski. USA 75,5879-5883. PERLER, F., EFSTRATIADIS, A., LOMEDICO, P., GILBERT, W., KOLODNER, R., and DODGSON, J. (1980). The evolution of genes: The chicken preproinsulin gene. Cell 20,555-566. PURO, D. G., and AGARDH, E. (1984). Insulin mediated regulation of neuronal maturation. Science 225,1170-11’72. RIDPATH, J. F., HUIATT, T. W., TRENKLE, A. H., ROBSON, R. M., and BECHTEL, P. J. (1984). Growth and differentiation of chicken embryo muscle cell cultures derived from fast- and slow-growing lines. Intrinsic differences in growth characteristics and insulin response. D@erentiatim 26,121-126. ROSENZWEIG, J. L., HAVRANKOVA, J., LESNIAK, M. A., BROWNSTEIN, M. J., and ROTH, J. (1980). Insulin is ubiquitous in extrapancreatic tissues in rats and humans. Proc NatL Acad Sci. USA 77.572-576. SCHECHTER, R., HOLTZCLAW, L., SADIQ, F., KAHN, A., and DEVASKAR, S. (1988). Insulin synthesis by isolated rabbit neurons. Endocrinology 123,505-513. SMIT, A. B., VREUGDENHIL, E., EBBERINK, R. H. M., GERAERTS, W. P. M., KLOOTWIJK, J., and JOOSE, J. (1988). Growth controlling molluscan neurons produce the precursor of an insulin-related peptide. Nature(Lmdon)331,535-538. THOMAS, P. S. (1980). Hybridization of denatured RNA and small DNA fragments transferred to nitrocellulose. Proc Natl. Ad. Sci. USA 77.5201-5205. YOUNG, S. W. (1986). Periventricular hypothalamic cells in the rat brain contain insulin mRNA. Neuropeptides 8,93-97. YOUNG, S. W., BONNER, T. I., and BRAUN, M. R. (1986). Mesencephalic dopamine neurons regulate the expression of neuropeptide mRNAs in the rat forebrain. Proc NatL Acad. Sci USA 83,9827-9831.