ABSTRACT. Insulin-like growth factor I (IGF-I) mRNA expression was studied after 90% partial pancreatectomy in the rat to determine whether IGF-I was ...
Proc. Natl. Acad. Sci. USA Vol. 88, pp. 6152-6156, July 1991 Cell Biology
Enhanced insulin-like growth factor I gene expression in regenerating rat pancreas (somatomedin C/pancreatic f8 cell/pancreatic ductules/in situ hybridization)
FANNIE E. SMITH*t, KENNETH M. ROSENt, LYDIA VILLA-KOMAROFFt, GORDON C. WEIR*, AND SUSAN BONNER-WEIR* *E. P. Joslin Research Laboratory, Joslin Diabetes Center, Harvard Medical School, Boston, MA 02215; and tDepartment of Neurology, Children's Hospital, Harvard Medical School, Boston, MA 02115
Communicated by Judah Folkman, April 11, 1991 (received for review August 14, 1990)
ABSTRACT Insulin-like growth factor I (IGF-I) mRNA expression was studied after 90% partial pancreatectomy in the rat to determine whether IGF-I was associated with pancreatic regeneration. The level of IGF-I mRNA was maximally increased (4-fold above control value) 3 days after pancreatectomy, but thereafter gradually decreased, returning to control levels by 14 days after surgery. By in situ hybridization, IGF-I mRNA in both pancreatectomized and sham-operated rats was localized to capillary endothelial cells, indicating that this is the site of IGF-I expression in the normal rat pancreas. However, enhanced IGF-I mRNA expression was localized to focal areas of regeneration unique to pancreatectomized rats. In these areas, epithelial cells of proliferating ductules and individual connective tissue cells expressed IGF-I, suggesting that IGF-I may play an important role in the growth or differentiation of pancreatic tissue.
MATERIALS AND METHODS Experimental Animals. Under sodium amytal anesthesia, 100-g male Sprague-Dawley rats underwent a 90% pancreatectomy by a modification of the method of Folgia as described by Bonner-Weir et al. (21). Briefly, the pancreatic tissue was gently abraded from the blood vessels except for an anatomically well-defined portion bounded by the common bile duct and the duodenum; for sham animals, the pancreas was rubbed between the fingers but no tissue was
removed. At 1, 3, 7, and 14 days after surgery, blood for plasma glucose and insulin values was obtained by tail snipping from the nonfasted animals, and then the animals were killed. The remnant pancreas, a corresponding area of the pancreas in the sham animals, and the right lobe of the liver were removed for RNA isolation. Plasma glucose values were determined using a glucose analyzer (model II, Beckman) and plasma insulin values were determined by radioimmunoassay (22). RNA Preparation. Total pancreatic RNA was isolated by the guanidinium thiocyanate method (23). RNA concentrations were determined spectrophotometrically at 260 nm. The integrity of the RNA was monitored by ethidium bromide staining after electrophoresis of 2 pug of total RNA in 2.2 M formaldehyde/1.2% agarose surface-tension gels (24). IGF-I Probe. The rat IGF-I cDNA was a generous gift of D. LeRoith and C. Roberts (National Institutes of Health). A 240-base-pair fragment of IGF-I cDNA derived from the coding region of the IGF-I mRNA was subcloned into bacteriophage M13mpl9. DNA was uniformly labeled with [a-32P]dCTP (800 Ci/mmol; Amersham; 1 Ci = 37 GBq) to a specific activity of 108 cpm/,ug by using recombinant phage DNA as a template to generate single-stranded antisense DNA for use in S1 nuclease protection assays and in situ hybridization. To improve resolution of in situ hybridizations, the DNA was also labeled with [a-[35S]thio]dCTP and [a-[35S]thio]dATP (1100 Ci/mmol; Amersham). As a negative control for in situ hybridizations, a labeled probe was synthesized using a phage containing the antisense IGF-I DNA fragment. For dot blot hybridization, the 240-base-pair IGF-I cDNA was labeled to a specific activity of 5 x 108 cpm/,ug by using a random-hexanucleotide priming kit (Boehringer Mannheim) with [a-32P]dCTP (3000 Ci/mmol; Amersham). S1 Nuclease Protection Assay. S1 nuclease protection assays were used to increase the sensitivity of detection of IGF-I mRNA (25). After 50,000 cpm of labeled probe was hybridized to 10 ,g of each RNA sample at 47°C for 18 hr in 75% formamide/400 mM NaCl, the reaction mixture was digested with S1 nuclease (Pharmacia; 150-225 units/ml) for 60 min at 37°C. Protected hybrids were precipitated with ethanol, dried, resuspended in 80% formamide loading buffer, and run in 8 M urea/5% polyacrylamide sequencing gels to size the protected fragment. Dried gels were exposed to x-ray film with intensifying screens for 1-3 days. Dot Blot Assay. For simultaneous quantification of multiple samples at different time points, 25 ,g of total RNA was applied to charged nylon membranes (Nytran) with a Minifold dot blotter (Schleicher & Schuell). Hybridization buffer contained 50% deionized formamide, 10% dextran sulfate, 0.1% SDS, 5x SSC (final concentration, 750 mM NaCl/75
The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact.
Abbreviation: IGF-I, insulin-like growth factor I. tTo whom reprint requests should be addressed at: Research Division, Joslin Diabetes Center, One Joslin Place, Boston, MA 02215.
Insulin-like growth factor I (IGF-I) belongs to a family of peptide growth factors that stimulate cell proliferation and differentiation (1, 2) and are expressed in developing tissues (3-6) and in response to injury (7-11). Several lines of evidence suggest that IGF-I may be involved in both endocrine and exocrine pancreatic growth. Specific receptors for IGF-I have been described on rat pancreatic a and p cells and murine acinar tissue (12, 13). Other studies have shown that cultured fetal rat islets and human fetal pancreatic explants release immunoreactive IGF-I into the medium (14-16) and IGF-I has been localized to islet l cells by immunohistochemistry (16). However, the response elicited by exogenous IGF-I on cultured islets (17, 18) and acinar tissue (19, 20) varied with the age and the culture conditions of the animals from which the islets were isolated. The role of IGF-I in pancreatic growth is not clear, since the results obtained under culture conditions may not reflect the in vivo state. Therefore, we used a rat model of pancreatic regeneration, the 90% partial pancreatectomy, to investigate changes in the expression of IGF-I and to localize the cells synthesizing this factor.
6152
Cell
Biology: Smith et A
Proc. Natl. Acad. Sci. USA 88 (1991)
mM sodium citrate), 2x Denhardt's solution (final concentration, 0.04% bovine serum albumin/0.04% Ficoll/0.04% polyvinylpyrrolidone), 50 mM sodium phosphate, 2 mM vanadyl ribonucleoside, and 100 Ag of denatured salmon sperm DNA per ml. Following hybridization for 18 hr at 420C, blots were washed twice in 2x SSC/0.1% SDS at room temperature for 30 min, twice in lx SSC/O.1% SDS at 420C for 30 min, and once in 0.1 x SSC/O.1% SDS at 550C for 1 hr. Blots were then autoradiographed for 1-5 days and autoradiograms were quantified on a scanning video densitometer (Zeineh; Biomedical Instruments, Fullerton, CA). Blots were reprobed as described above with a randomly primed 32Plabeled chicken a-tubulin cDNA to confirm equivalent loading of each sample on the filter. In Situ Hybridization. Sections of paraffin-embedded pancreas (5 ,tm) from pancreatectomized or sham rats that had been perfused with 4% paraformaldehyde in vivo were mounted on gelatin-coated slides and deparaffinized and rehydrated through graded concentrations of ethanol. Pretreatment consisted of the following steps: 0.2 M HCI for 10 min, 0.01% Triton X-100 for 1.5 min, proteinase K at 1 Ag/ml in phosphate-buffered saline for 30 min at 370C, and 0.1 M triethanolamine/0.25% acetic anhydride for 10 min. The sections were prehybridized for 3-4 hr at 40°C in a solution containing 50%o formamide, 2 x SSC, 25 ,ug of yeast tRNA per ml, Sx Denhardt's solution, 25 mM EDTA, 0.2% SDS, and 250 ,ug of denatured salmon sperm DNA per ml. To reduce nonspecific binding with 35S-labeled probes, 10 mM dithiothreitol was added to all solutions and washes. The sections were hybridized for 18 hr at 45°C in the prehybridization solution containing 5% dextran sulfate and 50,000-100,000 cpm of 32P-labeled probe or 200,000 cpm of 35S-labeled probe per slide. Adjacent sections were hybridized with the labeled sense strand as the negative control. The sections were washed four times in 2 x SSC at room temperature for 30 min, once in 1 x SSC at room temperature for 30 min, and finally in 0.1 x SSC at 450C for 1 hr. Sections were then dehydrated through graded concentrations of ethanol containing 0.3 M ammonium acetate and air-dried. For autoradiography, the slides were dipped in Kodak NTB-2 emulsion diluted 1:1 with 0.6 M ammonium acetate. Sections hybridized with 32p_ labeled probe were exposed for 48 hr and those with 35Slabeled probe for 2 weeks before development. Sections were counterstained with hematoxylin and examined by both brightfield and darkfield microscopy. A Standards
C D
B
** *
E
F
G
0
6153
344-'
220-
t
154-1 if
MP 1
3
7 14 1
3
PX
7 14 LS LP C
S
FIG. 1. Analysis of IGF-I mRNA by S1 nuclease protection. Pancreatic RNA from pancreatectomized (PX) and sham (S) rats at 1, 3, 7, and 14 days after surgery was hybridized to a 240-base 32P-labeled IGF-I probe. Sham and 14-day PX samples represent RNA from one animal, but two to three animals were used for 1-, 3-, and 7-day PX samples because ofthe small amount of remnant tissue. Molecular size markers (lane M; length in bases at left), undigested probe (lane P), and a negative control (lane C) of tRNA hybridized to the probe are shown. Lanes containing 5 ,g of liver from 3-day sham (LS) and 3-day PX (LP) rats were run for comparison. The light band in the 14-day PX sample is a small amount of undigested probe.
RESULTS Seven days after 90%o pancreatectomy, plasma glucose values were elevated by 1.1-1.7 mM over sham levels (10.0 + 0.6, n = 6 pancreatectomized rats, vs. 8.4 + 0.4, n = 3, sham rats). In contrast, plasma insulin values were not significantly different at any time point, as previously reported (21). To examine the expression of IGF-I mRNA in regenerating pancreas, we first hybridized an IGF-I probe to a Northern blot containing 10 ,g of total RNA from pancreas and 5 ,ug
of RNA from liver. Liver RNA contained transcripts of 7.5, 4.7, 1.7, and 1.2 kilobases, as previously described (26). Identical transcripts were visible in total pancreatic RNA before and after pancreatectomy but were near the limit of detection (unpublished data). Therefore an S1 nuclease protection assay was used to increase the sensitivity of detection of IGF-I mRNA (Fig. 1). Densitometry of three different autoradiograms revealed an 8-fold increase in the IGF-I protected fragment 3 days after pancreatectomy. At 7 days, the levels of IGF-I were 4-fold higher than in sham animals but were equal to those of sham animals by 14 days. This change in IGF-I mRNA levels was pancreas-specific, since
H
*
w
60-
Controls 0
1
* 0
3
0S@-
Days
0O
.:...
I*
40-
0
10 201
74 1 14
4;E
ID----------
Sham
r v r
0 1
7 14 3 Time after pancreatectomy, days FIG. 2. (Left) Dot blot analysis of pancreatic IGF-I mRNA from pancreatectomized (PX) and sham (S) rats at 1, 3, 7, and 14 days. Each dot represents 20 ,ug of pooled RNA from one to three pancreatectomized rats or one sham rat hybridized to a 32P-labeled cDNA IGF-I probe. Decreasing concentrations of rat liver mRNA (2.0-0.1 ,ug of rat liver mRNA) were applied in the top row, wells A-H, to assess the linearity of exposure in the densitometric analysis. A series of unlabeled IGF-I cDNA samples (50, 10, and 5 pg) were applied in the second row, wells A-C, as positive controls. A negative control (20 ,ug of yeast tRNA) was applied in the second row, wells F and G. After exposure to x-ray film, the hybridized label was removed and the blot was reprobed with a 32P-labeled chicken a-tubulin probe to verify that equivalent amounts of RNA were present in each sample. (Right) Densitometric analysis of the dot blot, with arbitrary densitometry units alongsthe y axis.
PX
S
6154
Cell Biology: Smith et A
Proc. Natl. Acad. Sci. USA 88 (1991)
lated with the S1 nuclease assays. The level of IGF-I mRNA was 3-fold increased 7 days after surgery and returned to
normal by 14 days (Fig. 2 Right). The level of IGF-I mRNA in sham animals was slightly elevated 1 day after pancreatectomy compared with 14 days and probably reflected the small sample size, although an effect of surgical manipulation or anesthesia has not been ruled out. To identify the cell type(s) with increased IGF-I mRNA, in situ hybridization was performed on sections of rat pancreas 3 days after pancreatectomy, when IGF-I expression was highest. In pancreatectomized and sham animals, autoradiographic grain labeling corresponded only to the location of the capillary endothelial cells and not to the endocrine or exocrine cells. An example of this localization is seen in Fig. 3. Thus it appeared that capillary endothelial cells were the major site of IGF-I expression in the normal adult rat pancreas, although very low levels of expression by f3 cells cannot be ruled out. However, highest labeling was found in focal areas of regeneration, unique to the pancreatectomized animals (Fig. 4). These foci of regeneration accounted for 10-15% of the pancreatic remnant and consisted of proliferating ductules and blood vessels surrounded by loose connective tissue. In the connective tissue, numerous cells were heavily labeled, while other cells in the same area were totally devoid of grains (Fig. SA). We have identified both activated macrophages and fibroblasts in these areas but have not determined whether these were the cell types expressing IGF-I (unpublished data). Within the focal areas of regeneration, proliferating ductules were abundant and specific labeling was found over the ductule epithelial cells (Fig. 5 B and C). These focal areas of regeneration were found 3 days after pancreatectomy, were rarely seen at 7 days, and were not present in sham animals. Therefore the relative increase in IGF-I expression in the pancreatectomized animals could be accounted for by the induction of IGF-I mRNA specifically in these focal areas of regeneration.
FIG. 3. Localization of IGF-I mRNA by in situ hybridization in rat pancreas 3 days after 90% pancreatectomy. The section contains a pancreatic islet with adjacent exocrine tissue. The arrows point out several islet capillaries lined with endothelial cells with localized autoradiographic grains (black). (Brightfield micrograph of a hematoxylin-stained section hybridized with a 32P-labeled IGF-I probe; bar = 25 jam.)
the IGF-I mRNA levels were not significantly different in liver from pancreatectomized or sham animals by dot blot assays (unpublished data). Dot blot hybridizations were performed to test multiple samples at each time point in the same assay. Although there was variability in the IGF-I mRNA levels 1 day after pancreatectomy, all pancreatectomized animals had more IGF-I mRNA than corresponding sham animals at 3 days. Later the expression of IGF-I was less variable in the samples and mRNA levels showed the same pattern found by S1 nuclease assays (Fig. 2 Left). Reprobing the dot blots with tubulin DNA verified equivalent application of RNA samples. Densitometric analysis of the dot blot autoradiogram found the peak of IGF-I expression at 3 days after pancreatectomy (4-fold increase), which corre-
DISCUSSION Little is known about the regulation of pancreatic growth or regeneration. Previous investigations have used culture tech-
I I FIG. 4. Darkfield micrographs of IGF-I mRNA localization by in situ hybridization with an 35S-labeled IGF-I probe in a rat pancreas 3 days after 90% pancreatectomy. Exposed silver grains appear as white dots. (A) The edge (arrows) of a large focal area of regeneration adjacent to exocrine tissue (EX) hybridized with the antisense IGF-I probe. A blood vessel (BV) is located in the center. (B) The same area from an adjacent section hybridized with the sense IGF-I probe, showing absence of labeling. The exocrine tissue (EX) appears whiter due to a longer exposure time but has no increased labeling over background. (Bar = 500 jim.)
Cell
t\:i:~ ~.
Biology: Smith et aL
Proc. Natd. Acad. Sci. USA 88 (1991)
/-fiir.
b,..X 4
..i.
6155
PD
D . .-
,.-
. S
_
A
.4..).
B
*
t
.gPD
ze #r 'j.
FIG. 5. Brightfield micrographs of hematoxylin-stained sections of a 3-day pancreatectomy pancreas after in situ hybridization with an 35S-labeled IGF-I probe. (A) Multiple individual cells in the loose connective tissue at the edge of a focal area of regeneration. Some cells are heavily labeled with black autoradiographic grains (large arrows), whereas other cells are devoid of grains (small arrows). (B) Proliferating
ductules (PD) within a focal area of regeneration with numerous labeled epithelial cells (arrows). (C) Proliferating ductules (PD) in adjacent section hybridized with the sense IGF-I probe as negative control. (Bar = 25 ,um.)
niques to study the effects of hormones and growth factors on pancreatic islets (27) and exocrine tissue (19, 20). We have used the 90% pancreatectomized rat model, in which there is marked pancreatic regeneration, to study the expression of IGF-I. In this model, growth occurs by neogenesis from proliferating ductules and by increased replication of existing differentiated cells (28). Enhanced IGF-I mRNA expression in regenerating pancreatic tissue peaked 3 days after pancreatectomy but then declined to control levels by 14 days. This peak coincided with a rise in the mitotic index of both /3 and exocrine cells following pancreatectomy (28), suggesting that IGF-I may act in an autocrine or paracrine manner to stimulate pancreatic growth. This conclusion is supported by studies that show increased IGF-I biosynthesis in regeneration after injury in a number of tissues (7-11). In addition, the temporal expression of IGF-I after pancreatectomy is similar to the time course of maximal IGF-I biosynthesis in regenerating muscle (7, 8) and in rat kidney during compensatory renal hypertrophy (11). Using in situ hybridization, we have demonstrated the specific localization of IGF-I mRNA in the pancreas. In both pancreatectomized and sham animals IGF-I mRNA was localized to capillary endothelial cells, suggesting that this is the predominant site of IGF-I expression in the normal rat pancreas. IGF-I synthesis also occurs in endothelial cells in numerous other tissues (29, 30). We detected no differences in the level of IGF-I expression in pancreatic endothelial cells from pancreatectomized and sham animals, while arterial endothelial cells appear to increase IGF-I synthesis after injury (9). The majority of the enhanced IGF-I expression was localized to unique focal areas of regeneration in the pancreatectomized animals. In these areas both individual connective tissue cells and ductule epithelial cells expressed IGF-I. The connective tissue cells, which have yet to be fully identified, are likely to include cells of mesenchymal origin, fibroblasts, and macrophages, which are all known to synthesize IGF-I (31-33). Proliferating ductules were abundant in the focal areas of regeneration 3 days after pancreatectomy. In this model, pancreatic acinar and islet tissue developed by budding from these ductules in a process similar to neogenesis from protodifferentiated ductules in embryonic development (34). Since levels of IGF-I mRNA decreased concordantly with a decrease in proliferating ductules by 7 days, IGF-I may play a critical role in cellular differentiation rather than proliferation in ductule epithelial cells. This view is supported by reports that IGF-I is involved in cellular differentiation in a number of other cell types (35-37).
IGF-I mRNA levels returned to control levels by 14 days but there was still sustained growth of pancreatic cells, particularly the ,3 cells (28). This growth may be the result of a cascade effect of IGF-I stimulation or a secondary effect of other mitogens such as glucose. Mild hyperglycemia was present 7 days after pancreatectomy, and studies in vivo (38) and in culture (39, 40) demonstrated that glucose stimulated ,8-cell growth and DNA synthesis. We found no evidence of IGF-I expression in normal or regenerating adult endocrine or exocrine tissue. Immunoreactive IGF-I has been found in culture medium from pancreatic islets and explants (14-16) and IGF-I was localized to human fetal /8 cells by immunohistochemistry (16). IGF-I mRNA was not found in pancreatic acinar, ductule, or islet cells in the human fetus; it was present in "retroperitoneal tissue," but the pancreatic cell types were unspecified (31). These differences are most likely explained by the age of the animals studied, since fetal tissues often express IGFs not present in the same adult tissue (2). Furthermore, immunohistochemical localization may represent receptor-bound IGF-I and not cellular biosynthesis. The results of our study demonstrate that following 90% pancreatectomy, IGF-I mRNA levels increase rapidly and are associated with pancreatic growth. Although capillary endothelial cells express IGF-I mRNA, the majority of IGF-I expression in the regenerating pancreas is localized to proliferating ductule epithelial cells and the surrounding connective tissue cells. The cellular signals for the induction of IGF-I in this model of regeneration are unknown. IGF-I may play an important role in pancreatic regeneration by autocrine or paracrine mechanisms to stimulate DNA synthesis or may act as a differentiation factor in the proliferating ductules. We thank Drs. John Leahy, George King, Christopher Rhodes, and Robin Mozell for helpful discussions; Genie Brainerd and Leslie Baxter for expert technical assistance; Chris Cahill for expert photographic work; and Ellen Moore for preparation of the manuscript. This study was supported by National Institutes of Health Training Grant DK07260 to the Joslin Diabetes Center, Mental Retardation Center Grant HD18655 to Children's Hospital, a grant from the Diabetes Research and Education Foundation, National Institutes of Health Grants DK35449 and NS27832, and the Animal Core Facility of Diabetes Endocrinology Research Center (Grant DK36836). 1. Daughaday, W. H. & Rotwein, P. (1989) Endocr. Rev. 10, 68-91. 2. Sara, V. R. & Hall, K. (1990) Physiol. Rev. 70, 591-614.
6156
Cell Biology: Smith et A
3. Han, V. K. M., Lund, P. K., Lee, D. C. & D'Ercole, A. J. (1988) J. Clin. Endocrinol. Metab. 66, 422-429. 4. Lund, P. K., Moats-Staats, B. M., Hynes, M. A., Simmons, J. G., Jansen, M., D'Ercole, A. J. & Van Wyk, J. J. (1986) J. Biol. Chem. 261, 14539-14544. 5. Rotwein, P., Pollack, K. M., Watson, M. & Milbrandt, J. D. (1987) Endocrinology 121, 2141-2144. 6. Adamo, M., Lowe, W. L., Jr., LeRoith, D. & Roberts, C. T., Jr. (1989) Endocrinology 124, 2737-2744. 7. Edwall, D., Schalling, M., Jennische, E. & Norstedt, G. (1989) Endocrinology 124, 820-825. 8. Jennische, E., Skottner, A. & Hansson, H.-A. (1987) Acta Physiol. Scand. 129, 9-15. 9. Hansson, H.-A., Jennische, E. & Skottner, A. (1987) Cell Tissue Res. 250, 499-505. 10. Hansson, H.-A., Dahlin, L. B., Danielsen, N., Fryklund, L., Nachemson, A. K., Polleryd, P., Rozell, B., Skottner, A., Stemme, S. & Lundborg, G. (1986) Acta Physiol. Scand. 126, 609-614. 11. Fagin, J. A. & Melmed, S. (1987) Endocrinology 120, 718-724. 12. Van Schravendijk, C. F. H., Foriers, A., Van den Brande, J. L. & Pipeleers, D. G. (1987) Endocrinology 121, 1784-1788. 13. Williams, J. A., Bailey, A., Humbel, R. & Goldfine, I. D. (1984) Am. J. Physiol. 246, G%-G99. 14. Romanus, J. A., Rabinovitch, A. & Rechler, M. M. (1985) Diabetes 34, 696-702. 15. Scharfmann, R., Corvol, M. & Czernichow, P. (1989) Diabetes 38, 686-690. 16. Hill, D. J., Frazer, A., Swenne, I., Wirdnam, P. K. & Milner, R. D. G. (1987) Diabetes 36, 465-471. 17. Swenne, I., Hill, D. J., Strain, A. J. & Milner, R. D. G. (1987) Diabetes 36, 288-294. 18. Rabinovitch, A., Quigley, C., Russell, T., Patel, Y. & Mintz, D. H. (1982) Diabetes 31, 160-164. 19. Logsdon, C. D. (1986) Am. J. Physiol. 251, G487-G494.
Proc. Nati. Acad. Sci. USA 88 (1991) 20. Brannon, P. M., Hirschi, K. & Korc, M. (1988) Pancreas 3, 41-48. 21. Bonner-Weir, S., Trent, D. F. & Weir, G. C. (1983) J. Clin. Invest. 71, 1544-1553. 22. Albano, J. D. M., Ekins, R. P., Maritz, G. & Turner, R. C. (1972) Acta Endocrinol. 70, 487-509. 23. Chirgwin, J. M., Przybyla, A. E., MacDonald, R. J. & Rutter, W. J. (1979) Biochemistry 18, 5294-5299. 24. Rosen, K. M., Lamperti, E. D. & Villa-Komaroff, L. (1990) Biotechniques 8, 398-403. 25. Berk, A. J. & Sharp, P. A. (1977) Cell 12, 721-732. 26. Hynes, M. A., Van Wyk, J. J., Brooks, P. J., D'Ercole, A. J., Jansen, M. & Lund, P. K. (1987) Mol. Endocrinol. 1, 233-242. 27. Nielsen, J. H. (1985) Acta Endocrinol. Suppl. 266, 7-39. 28. Brockenbrough, J. S., Weir, G. C. & Bonner-Weir, S. (1988) Diabetes 37, 232-236. 29. Kern, P. A., Svoboda, M. E., Eckel, R. H. & Van Wyk, J. J. (1989) Diabetes 38, 710-717. 30. Bar, R. S., Boes, M., Dake, B. L., Booth, B. A., Henley, S. A. & Sandra, A. (1988) Am. J. Med. 85, 59-70. 31. Han, V. K. M. & D'Ercole, A. J. (1987) Science 236, 193-197. 32. Clemmons, D. R. & Van Wyk, J. J. (1985) J. Clin. Invest. 75, 1914-1918. 33. Rappolee, D. A., Mark, D., Banda, M. J. & Werb, Z. (1988) Science 241, 708-712. 34. Pictet, R. L., Clark, W. R., Williams, R. H. & Rutter, W. J. (1972) Dev. Biol. 29, 436-467. 35. Florini, J. R., Ewton, D. Z., Falen, S. L. & Van Wyk, J. J. (1986) Am. J. Physiol. 250, C771-C778. 36. Schmid, C., Steiner, T. & Froesch, E. R. (1984) FEBS Lett. 173, 48-52. 37. Smith, P. J., Wise, L. S., Berkowitz, R., Wan, C. & Rubin, C. S. (1988) J. Biol. Chem. 263, 9402-9408. 38. Bonner-Weir, S., Deery, D., Leahy, J. L. & Weir, G. C. (1989) Diabetes 38, 49-53. 39. Swenne, I. (1982) Diabetes 31, 754-760. 40. Chick, W. L. (1973) Diabetes 22, 687-693.
