Aug 7, 1990 - Kidney Diseases, National Institutes of Health, Bethesda, Maryland. Address correspondence and reprint requests to Derek LeRoith, MD, PhD,.
Experimental Diabetes Increases Insulinlike Growth Factor I and II Receptor Concentration and Gene Expression in Kidney HAIM WERNER, ZILA SHEN-ORR, BETHEL STANNARD, BARTOLOME BURGUERA, CHARLES T. ROBERTS, JR., AND DEREK LEROITH
Insulinlike growth factor I (IGF-I) is a mitogenic hormone with important regulatory roles in growth and development. One of the target organs for IGF-I action is the kidney, which synthesizes abundant IGF-I receptors and IGF-I itself. To study the involvement of IGF-I and the IGF-I receptor in the development of nephropathy, one of the major complications of diabetes mellitus, we measured the expression of these genes in the kidney and in other tissues of the streptozocin-induced diabetic rat. The binding of 12SIlabeled IGF-I to crude membranes was measured in the same tissues. We observed a 2.5-fold increase in the steady-state level of IGF-l-receptor mRNA in the diabetic kidney, which was accompanied by a 2.3-fold increase in IGF-I binding. In addition to this increase in IGF-I binding to the IGF-I receptor, there was also binding to a lower-molecular-weight material that may represent an IGF-binding protein. No change was detected in the level of IGF-l-peptide mRNA. Similarly, IGF-ll-receptor mRNA levels and IGF-II binding were significantly increased in the diabetic kidney. IGF-I— and IGF-ll-receptor mRNA levels and IGF-I and IGF-II binding returned to control values after insulin treatment. Because the IGF-I receptor is able to transduce mitogenic signals on activation of its tyrosine kinase domain, we hypothesize that, among other factors, high levels of receptor in the diabetic kidney may also be involved in the development of diabetic nephropathy. Increased IGF-ll-receptor expression in the diabetic kidney may be important for the intracellular transport and packaging of lysosomal enzymes, although a role for this receptor in signal transduction cannot be excluded. Finally, the possible role of IGF-binding proteins requires further study. Diabetes 39:1490-97, 1990
From the Diabetes Branch, National Institutes of Diabetes and Digestive and Kidney Diseases, National Institutes of Health, Bethesda, Maryland. Address correspondence and reprint requests to Derek LeRoith, MD, PhD, Chief, Section of Molecular and Cellular Physiology, National Institutes of Health, Building 10, Room 8S 243, 9000 Rockville Pike, Bethesda, MD 20892. Received for publication 21 March 1990 and accepted in revised form 7 August 1990.
1490
N
ephropathy, the loss of glomerular function and subsequent kidney failure due to structural lesions in the glomeruli, is one of the major complications of both insulin-dependent and non-insulin-dependent diabetes mellitus and clearly constitutes one of the main causes of death among diabetic patients (1). Early nephropathy leads to an expansion of the mesangium, which results in a reduction in the capillary lumenal space and, as a consequence, a diminution in filtration surface and rate (2). Different paracrine and autocrine factors are synthesized by mesangial cells in culture, and it has been shown that conditioned medium from these cultures strongly stimulates mesangial cell proliferation (3). One of these mitogenic factors is insulinlike growth factor I (IGF-I; 4), an anabolic hormone that is structurally and functionally related to insulin and has important regulatory roles in both growth and development (5). Although the liver produces >90% of circulating IGF-I, other tissues are capable of synthesizing IGF-I (6,7), and indeed, IGF-I purified from mesangial cells was shown to coelute with authentic IGF-I on a high-performance liquid-chromatography column (4). Furthermore, immunostainable IGF-I colocalizes with IGF-I mRNA in the collecting duct, consistent with a focal expression of the IGF-I gene at this site (8). In a recent study, both the administration of growth hormone to hypophysectomized rats and compensatory hypertrophy subsequent to unilateral nephrectomy led to an increase in IGF-I in collecting duct cells, suggesting a causative role for this growth factor in kidney hypertrophy (9). Administration of IGF-I to hypophysectomized rats resulted in a significant increase in kidney weight (10), whereas IGF-I infusion in nondiabetic human subjects induced a significant increase in glomerular filtration rate and kidney plasma flow (11,12). The biological effects of IGF-I, both short and long term, are initiated by its binding to a specific heterotetrameric cell surface receptor that is composed of two extracellular ligand-binding a-subunits linked by disulfide bonds to two transmembrane p-subunits that contain IGF-I—inducible tyDIABETES, VOL. 39, DECEMBER 1990
H. WERNER AND ASSOCIATES
rosine kinase domains in their cytoplasmic portions (13). Receptors for IGF-I have been described in kidney membranes, glomerular mesangial cells, and tubules (14-20). To study the possible involvement of IGF-I and its receptor in diabetic nephropathy, we measured the expression of these genes in the kidney and other tissues of the streptozocin-induced diabetic (STZ-D) rat with sensitive solution hybridization-RNase protection assays. In addition, 125l-labeled insulinlike growth factor I (125(—IGF-I) binding to plasma membranes from the same tissues was measured. Because IGF-I is a member of a family of closely related peptides that includes IGF-I I, which may also interact with the IGF-I receptor, we determined the binding of IGF-II and the levels of IGF-II and IGF-Il-receptor mRNA in the diabetic kidney. Our results indicate that levels of IGF-I- and IGF-I l-receptor mRNAs and binding are significantly increased in the STZD kidney.
RESEARCH DESIGN AND METHODS
After an overnight fast, male Sprague-Dawley rats (weighing 178.0 ± 1.8 g; Zivic-Miller, Allison Park, PA) were given a single injection of STZ (Upjohn, Kalamazoo, Ml; 100 mg/kg body wt i.p. in 0.01 M citrate buffer, pH 4) or vehicle. Administration of STZ resulted in the induction of hyperglycemia in 85% of the animals 24 h after treatment. Rats were given free access to laboratory chow and water, except for a group of control animals that were pair fed according to the food intake of the diabetic rats. The metabolic status of the animals was monitored by daily semiquantitative estimation of the levels of urinary glucose (Dia,stix, Ames, Elkhart, IN) and ketone (Ketostix, Ames). Only rats with urine glucose levels > 112 mM were used in this study. No ketones were detected in urine at any stage of the study. In addition, body weight was measured daily. Two groups of STZ-D rats were treated with human recombinant insulin (Humulin, Lilly, Indianapolis, IN; 8 U • 100 g~1 body wt • day 1 ) beginning either 3 or 7 days after the STZ injection. Fourteen days after initiation of the experiment, rats were killed by decapitation, and the following tissues were immediately frozen in liquid N2: brain, liver, kidney, testes, heart, and skeletal muscle. Trunk blood was collected and serums were separated to measure glucose, insulin, and IGF-I. Because no significant differences were found in the circulating levels of glucose or these hormones, in the levels of IGF-I or IGF-l-receptor mRNA between animals treated with insulin for 7 or 11 days, or between the two control groups (see below), the results obtained for both insulintreated groups were combined, as were the results obtained for both control groups. To quantitate the concentration of IGF-I in serum, circulating IGF-I was first dissociated from its carrier protein by treatment with 0.5 N HCI, followed by chromatography through a C-18 Sep-Pak column (Waters, Milford, MA). IGFI radioimmunoassays were performed as described previously (21) with 125I-IGF-I from Amersham (Arlington Heights, IL), unlabeled Thr-59-IGF-l from Amgen (Thousand Oaks, CA), and anti-IGF-l serum from the National Hormone and Pituitary Agency (Univ. of Maryland, Baltimore). To determine levels of IGF-I in kidney, frozen tissue was homogenized in 0.1 N acetic acid, neutralized with NaOH, loaded on a C-18 Sep-Pak column, and assayed as above. DIABETES, VOL. 39, DECEMBER 1990
The levels of insulin in serum were measured with a kit obtained from Cambridge Medical Diagnostics (Billerica, MA), and serum glucose was determined with a Beckman glucose analyzer. Plasma membranes from control, diabetic, and insulintreated tissues were prepared as described previously (22). Briefly, frozen tissues were homogenized in 20 vol ice-cold 1 mM NaHCO3 containing 20 fiM leupeptin, 1 KlU/ml aprotinin, and 2 mM phenylmethylsulfonyl fluoride with a motordriven glass-Teflon homogenizer for brain, liver, kidney, and testes or a polytron for heart and skeletal muscle. After centrifugation of homogenates at 600 x g for 10 min at 4°C, the resulting supematants were centrifuged at 20,000 x g for 30 min at 4°C. Pellets were washed in homogenization buffer and recentrifuged. The final membrane pellets were resuspended in Ca2+-free Krebs-Ringer phosphate buffer (pH 7.8) and stored frozen at -70°C. Protein concentration was determined with the method of Lowry et al. (23). Binding of 125I-IGF-I and 125I-IGF-II to crude membranes was performed as described previously (22,24). Binding affinities and receptor concentrations were determined from Scatchard analysis. Cross-linking of 125I-IGF-I to receptor a-subunits was performed as described previously (25). Crude microsomal membranes from control, diabetic, and insulin-treated kidneys were cross-linked to 125I-IGF-I with disuccinimidyl suberate (0.1 mM) in the absence or presence of 130 nM of unlabeled IGF-I or 170 nM insulin. Total RNA was prepared from tissues of individual rats with a modification (26) of the lithium chloride extraction method of Cathala et al. (27) and quantified by absorbance at 260 nm. The integrity of the RNA and the accuracy of the spectrophotometric determinations were assessed by visual inspection of the ethidium bromide-stained 28S and 18S ribosomal RNA bands after agarose formaldehyde gel electrophoresis of 10-|xg aliquots as described previously (26). The antisense RNA probe used to detect IGF-l-receptor mRNA has been previously described (28). This transcript contains 40 bases of vector sequence and 265 bases complementary to 15 bases of 5'-untranslated sequence and to the region encoding the signal peptide and the first 53 amino acids of the IGF-l-receptor a-subunit. On hybridization of this RNA probe with IGF-l-receptor mRNA and subsequent RNase digestion, a protected band of 265 bases was obtained. The levels of IGF-I l-receptor message were measured with a riboprobe derived from rat IGF-I l-receptor cDNA (clone K3) provided by W.J. Rutter (Univ. of California, San Francisco). A 500-base pair (bp) Eco Rl-6am HI fragment of this cDNA was subcloned into pGEM-4Z, linearized with Rsa I, and transcribed via T7 RNA polymerase. The size of the protected band obtained by hybridizing this antisense RNA probe with IGF-II—receptor mRNA was - 3 0 0 bases. The riboprobe employed to measure the levels of IGF-I mRNA was described previously (29). This probe allows the detection of both IGF-I mRNA species encoding the IGF-la and IGF-lb prohormones. Only the levels of IGF-la mRNA, which constitute >90% of the total IGF-I message and correlate with the levels of IGF-lb mRNA, were measured in this study. IGF-ll-mRNA levels were determined with a riboprobe constructed from a Pst \-Bam HI fragment of rat IGF-II cDNA 1491
IGF-I AND IGF-II RECEPTORS IN DIABETIC KIDNEY
TABLE 1 Body weight, serum glucose, insulin, and insulinlike growth factor I (IGF-I) in control, diabetic, and insulin-treated rats
Control Diabetic Insulin treatedf
Weight (g)
Glucose (mM)
Insulin (pM)
IGF-I (nM)
312.6 ± 8.7 208.4 ± 11.9 292.0 ± 15.7
7.3 ± 0.2 23.7 ± 1.1 3.2 ± 0.2
234.0 :t 4.5 25.1 :t 3.6 5905.0 :t 892.0
182.5 ± 8.3 44.9 ± 8.5* 96.5 ±8.81:
Values are means ± SE; n = 5. Levels of glucose, insulin, and IGF-I in serums were determined 14 days after administration of streptozocin (diabetic and insulin-treated rats) or saline (control rats). For details, see RESEARCH DESIGN AND METHODS. Weights are at time of death. *P < 0.001 vs. control. fRats treated with insulin for 7 or 11 days. tP < 0.005 vs. diabetic.
provided by M.M. Rechler (Natl. Inst. of Health, Bethesda, MD; 30) subcloned into pGEM-3. The size of the protected band obtained on hybridization of this antisense RNA probe with IGF-II mRNA and subsequent RNase digestion was - 5 7 0 bases. Solution hybridization-RNase protection assays were performed as described (31). Briefly, 20 p.g of total RNA were hybridized with 2 x 105 dpm 32P-labeled antisense RNA probes. The hybridization was carried out at 45°C for 16 h in a buffer containing 75% formamide. After hybridization, RNA samples were digested with RNase A and T,, and the protected hybrids were extracted with phenol-chloroform, ethanol precipitated, and electrophoresed on an 8% polyacrylamide-8 M urea denaturing gel. Multiple autoradiograms from each gel were scanned with a GS300 Hoefer Scientific densitometer (Palo Alto, CA) connected to an Apple Macintosh computer. The significance of the differences observed between experimental groups was determined via one-way analysis of variance. The Newman-Keuls test was used to compare individual means. P < 0.05 was considered statistically significant.
RESULTS
The levels of glucose, insulin, and IGF-I determined in trunk blood 14 days after injection of STZ are presented in Table 1. As previously reported (32-34), the concentration of IGFI in serums of diabetic rats was significantly lowered with respect to control animals (mean ± SE 44.9 ± 8.5 vs. 182.5 ± 8.3nM,n = 5, P < 0.001), whereas insulin treatment for 7 or 11 days partially reversed this effect (96.5 ± 8.8 nM, P < 0.005 vs. untreated diabetic rats). No significant differences in IGF-I values were seen between rats treated with insulin for 7 or 11 days (data not shown). Body weights were decreased in diabetic animals but after insulin therapy were no different from controls (Table 1). To determine the effect of STZ-D on IGF-I binding to rat tissues, plasma membranes were prepared, and the specific binding of 125I-IGF-I was measured. An ~2.3-fold increase in 125I—IGF-I binding was displayed by membranes prepared from diabetic kidneys compared with control rats (10.77 ± 2.16 vs. 4.72 ± 0.18% specific binding/60 |xg protein, P < 0.005). This increased binding was reduced to control levels (5.21 ± 0.39%) after 7 or 11 days of insulin treatment (Fig. 1; Table 2). No increase in 125I-IGF-I binding was observed in the diabetic brain, testes, or heart. As previously shown (22), very low levels of 125I-IGF-I specific binding were exhibited by liver and muscle (
