insulin-like growth factor-related genes, receptors, and binding

Insulin-Like Growth Factor-Related Genes, Receptors,and Binding Proteins in Cultured HumanRetinal Pigment Epithelial Cells

Hitoshi Takagi, Nagahisa Yoshimura, Hidenobu Tanihara, and Yoshihito Honda

Purpose. It was determined in cultured human retinal pigment epithelial (HRPE) cells whetherthere is gene expression for insulin, insulin-like growth factors (IGFs), insulin and IGF recep-tors, and IGF-binding proteins (IGFBPs); if these peptides are secreted by the HRPE cells,and whether they have mitogenic effects on these cells; and if IGF-related peptides bind toHRPE cells.

Methods. Reverse transcriptase-polymerase chain reaction (RT-PCR), nucleotide sequencing,and Southern blot analyses were done to identify gene expression. The presence of IGFs in thetissue culture medium was detected by radioimmunoassay. For determination of mitogeniceffects and receptor binding characterization, [3H]-thymidine incorporation and radiorecep-tor binding measurements were performed.

Results. The genes for IGF-I and IGF-II, IGFBP-2, insulin receptor, and type I and II IGFreceptors were detected. In the tissue culture supernatant, there was immunoreactivity forIGF-I and IGF-II. Insulin, IGF-I, and IGF-II stimulated DNA synthesis with EC50s of 3, 10,and 30 to 100 nM, respectively. Scatchard analyses of [125I]-IGF-I binding and [125I]-insulinbinding to the cells showed that the cells have relatively abundant insulin and IGF-I bind-ing sites.

Conclusions. HRPE cells express genes for IGF-I and IGF-II, and conditioned medium fromthese cells is immunoreactive to their protein products. The cells express genes for insulinreceptor, type I and II IGF receptors, and IGFBP-2. IGF-I, IGF-II, and insulin are all mito-genic, possibly as a result of their interactions with either the insulin or type I IGF receptor.The cells bind insulin and IGF-I with high affinity. These results suggest that IGF-I and IGF-IIproduction by HRPE cells may be essential for autocrine/paracrine-mediated regulation ofproliferation. The presence of insulin receptor suggests that insulin has a role in the regulationof HRPE function. Invest Ophthalmol Vis Sci. 1994;35:916-923.

Insulin-like growth factors I and II (IGF-I and IGF-II) share a high degree of structural similarity withproinsulin. These three peptides and relaxin mole-cules comprise the insulin gene family.1 IGF-I andIGF-II are known to interact with specific receptors,designated type I and type II IGF receptors, and theyhave some mitogenic, metabolic, and growth-stimulat-

From the Department of Ophthalmology, Kyoto University Faculty of Medicine,Kyoto, Japan.Supported in part by a Grant-in-Aid for Scientific Research from the Ministry ofEducation, Science and Culture of the Japanese Government.Submitted for publication June 30, 1993; revised November 3, 1993; acceptedNovember 8, 1993.Proprietary interest category: N.Reprint requests: Nagahisa Yoshimura, M.D., Department of Ophthalmology, KyotoUniversity Faculty of Medicine, Kyoto 606, Japan.

ing activities in many cells and tissues.2 They arethought to be significant autocrine/paracrine factors.3

In addition to the type I and type II receptors, a familyof IGF-binding proteins (IGFBPs) that have no se-quence homology with the IGF receptors has beenidentified. IGFBPs may modulate activity of IGFs bybinding IGFs. Six distinct classes of IGFBPs have beencloned and characterized (IGFBP-1 to IGFBP-6).4

In the eye, IGF-related peptides are known to beinvolved in a variety of functions in the sensory retina.For example, insulin and IGF-I are thought to play arole in the visual transduction cascade,56 and to beinvolved in neural retinal differentiation.7"11 More re-cently, retinal pigment epithelium (RPE) cells were re-ported to express genes for IGF-I and type I and IIIGF receptors.12 This finding may be of importance

916Investigative Ophthalmology & Visual Science, March 1994, Vol. 35, No. 3Copyright © Association for Research in Vision and Ophthalmology

Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933180/ on 04/11/2018

IGF-Related Genes Expressed in RPE Cells 917

because the mannose receptor (type II IGF receptor)may be involved in phagocytosis of rod outer segmentsby RPE cells.1314 Furthermore, IGF-I is thought toplay an important role in the photoreceptor cell-RPEcell interaction because IGF-I and IGFBP are knownto exist not only in the interphotoreceptor matrix butalso in the culture supernatant of RPE cells.15 Thus,IGF-related peptides produced by RPE cells may playan important role in the maintenance of normal physi-ology of the retina as well as in disease processes.

In this report, through an extensive and system-atic study of IGF-related peptides, we show that cul-tured human RPE (HRPE) cells secrete products thatare immunoreactive to antibodies against IGF-I andIGF-II. We also demonstrate gene expression of theinsulin receptor as well as type I and II IGF receptors,and demonstrate that the cells express only theIGFBP-2 gene among the six types of IGFBPs. Fur-thermore, we show that the insulin gene family of pep-tides has mitogenic effects on HRPE cells; the recep-tor-binding properties of insulin and IGF-I also werecharacterized.

MATERIALS AND METHODS

Cell Culture

The research was approved by the Institutional Hu-man Experimental Committee, and in the authors'opinion, methods for securing human tissue were hu-mane, included proper consent and approval, andcomplied with the tenets of the Declaration of Hel-sinki. Cultured HRPE cells were established from hu-man fetal eyes as described previously.16 The cells weregrown in Dulbecco's modified Eagle's medium(DMEM) supplemented with 10% fetal bovine serumand antibiotics under a humidified atmosphere of 5%CO2 and 95% air at 37°C. Cells from the third to fifthpassages were used in the experiments.

Polymerase Chain ReactionTotal RNA was isolated from RPE cells cultured on10-cm dishes using the acid guanidinium thiocyanate-phenol-chloroform extraction method.17 First-strandcomplementary DNA (cDNA) was prepared from thetotal RNA using a first-strand cDNA synthesis kit(Pharmacia-LKB, Uppsala, Sweden). Total RNA of 5/ng was used to synthesize cDNA. Specific primers forinsulin, IGF-I, IGF-II, insulin receptor, type I and IIIGF receptors, and IGFBP-1 to IGFBP-6 were synthe-sized by an oligonucleotide synthesizer (Model 392;Applied Biophysics, Foster City, CA), using previouslypublished sequences18"28 so that each sequence wouldbe detected specifically and intron sequences thatwere excised during RNA processing would be in-cluded in genomic targets (Table 1). Polymerase chainreaction (PCR) was carried out by the method of Saikiet al29 with a slight modification. The following condi-tions were used: denaturation, at 94°C for 0.5 min-utes; annealing, 55°C or 65°C for 0.5 minutes; andpolymerization, 72°C for 1 minute. The reaction wasinitiated by adding two units of Taq polymerase, afterwhich 30 cycles were carried out. Taq DNA polymer-ase and reagents for PCR experiments were obtainedfrom Perkin-Elmer Cetus (Norwalk, CT). After PCR,reaction products were separated by 1.5% agarose gelelectrophoresis, and bands of the expected lengthwere extracted. The resultant DNAs were subclonedinto pBluescript II vector (Stratagene, La Jolla, CA).Before subcloning, the vector was treated with Eco RV,and the T vectors were made by Taq polymerase anddeoxythymidine triphosphate.30

Sequencing of PCR ProductsTo confirm that the PCR products were derived fromnucleotide sequences corresponding to IGF-relatedpeptides, their receptors, and binding proteins, se-quencing of subcloned DNA was carried out accord-ing to the dideoxynucleotide chain termination

TABLE i. Primers Used in the Experiment

5' Primer 3' PrimerLength(bp)

InsulinIGF-IIGF-IIInsulin receptorType I IGF receptorType II IGF receptorIGFB1MIGFBP-2IGFBP-3IGFBP-4IGFBP-5IGFBP-6

GGAACGAGGCTTCTTCTACATCGCATCTCTTCTACCTGGCCCCAGTGAGACCCTGTGTGAGAACAACGTCGTCCACTCACGAGGCTGAGAAGCTTGGAACCTCTGACAAGACCAGCAACCTCTGCACGCCCTCAAGACAATGGCGATGACCACTTGCTGCCTGACGTGCGCACTTGCTGCTGCTAGCTCTGCTGAACAGCTTCAGCCCCTGTAGCTGCACGGCCGCGGGGTTTG

AGTTGCAGTAGTTCTCCAGCGTAGTTCTTGTTTCCTGCACAGCACAGTACGTCTCCACGTAAAGCGGTCCCAGCGGCAGGCATACAGCACTCCATAATATGGCCGTTCTCCTGCCATTCCAAGGGTAGACGCACGGGTTCACACACCAGCACTCAGAAGTTCTGGGTATCTGTGTGCAGCACTGAGTCCAGATGCAGACTCAGACTCCACTCTGTGGCTGGAGTCGGGGCTGGG

200366174392502554434423397484491598


918 Investigative Ophthalmology & Visual Science, March 1994, Vol. 35, No. 3

method,31 using a Sequenase ver 2.0 DNA sequencingkit (United States Biochemicals, Cleveland, OH). Dou-ble-strand template DNAs were denatured by alkalinetreatment and the sequencing reaction was initiated byadding T3 or T7 primer.

Southern Blot Analysis

For Southern blot analysis, 10 1̂ of PCR product wasseparated by 1.5% agarose gel electrophoresis andthen transferred to a nylon filter (Hybond-N+; Amer-sham, Little Chalfont, United Kingdom) by capillarytransfer with 20X SSC. The DNA was fixed to the filterby baking at 80°C for 2 hours in a vacuum oven, DNAfragments corresponding to the target sequences werelabeled by the enhanced chemiluminescence gene de-tection kit (Amersham).

Detection of Immunoreactivities

After three washes with phosphate-buffered saline,the RPE cells cultured in a 10-cm dishes were incu-bated with 5 ml of serum-free DMEM containing0.25% bovine serum albumin (BSA). Twenty-fourhours later, the culture supernatant was collected.IGF-I and IGF-II in the supernatant were extractedwith acid-ethanol by the method of Daughaday et al.32

IGF-I was assayed by radioimmunoassay (RIA) using aRadioimmunoassay Kit for Somatomedin C (Med-genix Diagnostics, Fleurus, Belgium), and IGF-II wasassayed by RIA with anti-IGF-II monoclonal antibody(Amano Pharmaceutical Co., Tokyo, Japan) and [125I]-IGF-II and cold IGF-II (Amersham Japan, Tokyo, Ja-pan). For measuring insulin concentration, we appliedthe culture supernatant directly to RIA using the Pha-deseph Insulin RIA kit (Pharmacia Diagnostics).

[3H]-Thymidine Incorporation

Cells were grown to near confluence at 37°C in 24-well cluster plates. The growth medium was replacedwith DMEM containing 0.25% BSA for 48 hours. Cellswere incubated with insulin, IGF-I, and IGF-II (Amer-sham Japan) for 18 hours, after which they werepulsed with 2 MCi/ml [methyl-3H]-thymidine (Amer-sham Japan) for 2 hours at 37°C. The cells were solubi-lized with 0.1 N NaOH containing 0.1% sodium do-decyl sulfate, and a 500-^1 aliquot of each well wasadded to 2.5 ml of scintillation fluid and then counted.Two 100-/il aliquots of each well was used for pro-tein assay by a BCA Protein Assay Kit (Pierce, Rock-ford, IL).

Binding Study[125I]-insulin and [125I]-IGF-I binding studies wereperformed on cells attached to 12-well cluster plates(Corning, Corning, NY) according to the method ofKing et al.53 Incubation was at 15°C for 4 hours inHepes binding buffer (pH 7.8, 0.1 M Hepes, 0.12 M

NaCl, 5 mM KC1, 1.2 mM MgSO4, 8 mM glucose, and1% BSA). Concentrations of unlabeled insulin andIGF-I were added as noted in the figures. After bind-ing, the cells were washed with phosphate-bufferedsaline at 4°C and solubilized with 1 N NaOH. Afterneutralizing with 1 N HC1, an aliquot of 500 td wasremoved for determination of radioactivity by agamma counter, and two 100-/xl aliquots were used formeasurement of protein by a BCA Protein Assay Kit.[mI] -insulin and [125I]-IGF-I were obtainedfrom Amersham Japan. Cells were counted with aCoulter counter (Coulter Electronics, Luton, UnitedKingdom).

RESULTS

PCR and Nucleotide Sequencing

To ascertain whether cultured HRPE cells expressgenes for insulin, IGFs, insulin receptor, IGF recep-tors, and IGFBPs, PCR amplification using specificprimers was performed. Figure 1 indicates that PCRproducts of the expected length were detected for in-sulin, IGF-I, and IGF-II, although bands for insulinwere multiple. After subcloning of the PCR productsinto the pBluescript II vector, nucleotide sequenceanalysis was performed. Only IGF-I and IGF-II se-quences were found to correspond to the reported

M 1 2 3

FIGURE l. PCR studies for the detection of insulin, IGF-I,and IGF-II gene transcripts. An ethidium bromide-stainedagarose gel of the PCR products is shown. PCR was per-formed with the specific primers (shown in Table 1) for insu-lin, IGF-I, and IGF-II, using cDNA derived from culturedHRPE cells. PCR products of 10 ml were electrophoresed ina 1.5% agarose gel. Lane 1, PCR product for insulin; lane 2,PCR product for IGF-I; lane 3, PCR product for IGF-II;lane M, molecular weight marker of #X174 Haelll digest.PCR products of the expected length were detected for in-sulin, IGF-I, and IGF-II, but subsequent sequencing analy-ses revealed that only the PCR products for IGF-I and IGF-II corresponded to those reported sequences of the targetgenes.


IGF-Related Genes Expressed in RPE Cells

sequences of the target genes. As for IGFBP, PCRproducts of the expected length were detected onlyfor IGFBP-2 and IGFBP-6 among the six types ofIGFBPs (Fig. 2). The nucleotide sequences of IGFBP-2 and IGFBP-6 revealed that only IGFBP-2 corre-sponded to the reported sequence of the target gene.Insulin receptor and both types of IGF receptors weredetected by agarose gel electrophoresis and nucleo-tide sequencing (Fig. 3).

Southern Blot Analysis

Southern blot analysis was performed to confirm thatPCR products were derived from these genes. Weused the same gel indicated in Figures 1 to 3 for thisanalysis. As summarized in Table 2, DNA probes forIGF-I and IGF-II hybridized to the correspondingPCR products on the same nylon membrane. A DNAprobe of IGFBP-2 was found to hybridize to PCRproducts for IGFBP-2 and IGFBP-5, although theband for IGFBP-2 was dominant. A homology be-tween IGFBP-2 and IGFBP-5 can explain this cross-hybridization. DNA probes of insulin receptor andtype I and II IGF receptors hybridized to the corre-sponding PCR products, respectively. These data fur-ther confirm that the PCR product bands were de-rived from both types of IGFs, insulin receptor, bothtypes of IGF receptors, and IGFBP-2.

M 1 2 3 4 5 6

919

M 1 2 3

FIGURE 2. PCR studies for the detection of IGFBP gene tran-scripts. An ethidium bromide-stained agarose gel of the PCRproducts is shown. PCR was performed with the specificprimers (shown in Table 1) for IGFBP-1 to IGFBP-6, usingcDNA derived from cultured HRPE cells. Lanes 1 to 6 showPCR products for IGFBP-1 to IGFBP-6, respectively; laneM, molecular weight marker of 4>X174 Haelll digest. PCRproducts of the expected length were detected for IGFBP-2and IGFBP-6, but sequencing analyses revealed that only thePCR product for IGFBP-2 corresponded to those reportedsequences of the target genes.

FIGURE 3. PCR studies for the detection of insulin receptorand type I and II IGF receptors. An ethidium bromide-stained agarose gel of the PCR products is shown. PCR wasperformed with the specific primers (shown in Table 1) forthe receptors, using cDNA derived from cultured HRPEcells. Lane 1, PCR product for insulin receptor; lane 2, PCRproduct for type I IGF receptor; lane 3, PCR product fortype II IGF receptor; lane M, molecular weight marker of$X174 Haelll digest. PCR products of the expected lengthwere detected for all three receptors, and subsequent se-quencing analyses confirmed that these PCR products arederived from target genes.

Detection of ImmunoreactivitiesTo determine whether RPE cells produce and secreteIGF-I, IGF-II, and insulin, we measured the amountof immunoreactivities of these peptides released intothe culture medium. Duplicate measurement was per-formed for each sample, and the average value wasused for calculation. The concentrations of IGF-I- andIGF-II-like immunoreactivities in the culture superna-tant were 40 ± 3 ng/ml (n = 3) and 0.4 ± 0.0 ng/rnl (n= 3), respectively, whereas the fresh medium was notimmunoreactive to anti-IGF-I and anti-IGF-II antibod-ies. We could not detect insulin-like immunoreactivityin the culture supernatant (n = 6). These results indi-cate that cultured HRPE cells produce and secreteIGF-I and IGF-II immunoreactivities into the me-dium.

[3H]-Thymidine IncorporationTo determine whether insulin, IGF-I, and IGF-II aremitogenically active in fetal human eye cells, we mea-sured their effects on HRPE cells. In Figure 4, theresults indicate that insulin (1 f*M), IGF-I (30 nM), andIGF-II (30 nM) significantly stimulated DNA synthesisin the cells (P < 0.01). These three peptides increased[3H]-thymidine incorporation in a dose-dependentmanner, with a rank order of potency of IGF-I > IGF-II >> insulin. The values for 50% effective concentra-tion (EC50) for these stimulants were about 3? 10, and



TABLE 2. Southern Blot Analysis of the PCR ProductsProbe PCR Products Hybridization

IGF-I

IGF-II

IGFBP-2

Insulin receptor

Type I IGF receptor

Type II IGF receptor

InsulinIGF-IIGF-II

InsulinIGF-IIGF-II

IGFBP-1IGFBP-2IGFBP-3IGFBP-4IGFBP-5IGFBP-6

Insulin receptorType I IGF receptorType II IGF receptor



NoYesNo

NoNoYes

NoYesNoNoYesNo

YesNoNo

NoYesNo

NoNoYes

The probe for IGFBP-2 cross-hybridized to the IGFBP-5 PCR product because of the homology of thetwo IGFBPs.

30 to 100 nM, respectively. Insulin at low doses, how-ever, showed a relatively high mitogenic potency.

Binding Study

To ascertain to what extent the genes for insulin re-ceptor and type-I IGF receptor are functionally ex-

8

%o

fio

n (

irat

&8—<ulipiL

200

180

160

140

120

100

• insulin

• IGF-I

A IGF-II

Ligand Concentration (M, log10)

FIGURE 4. Effects of insulin, IGF-I, and IGF-II on [3H]-thymidine incorporation in HRPE cells. Quiescent cellswere incubated with these peptides at the indicated concen-tration for 24 hours, and [3H]-thymidine uptake by the cellsfor 1 hour was determined. Data are mean ± SEM for two orthree separate experiments, each done in triplicate. Thesethree peptides increased [3H]-thymidine incorporation in adose-dependent manner, with a rank order of potency ofIGF-I > IGF-II » insulin.

pressed in HRPE cells, studies were carried out to al-low direct counting of bound radioligand and Scat-chard analysis. Specific binding of [125I]-IGF-I wasmore than tenfold greater (11.9%/100 /xg cell protein)than the specific binding of [125I]-insulin (1.6%/100yug cell protein). Fifty percent inhibition of [125I]-IGF-Ibinding was obtained with 3 X 10~10 M unlabeled IGF-I, and that of [125I]-insulin required 2 X 10~9 M unla-beled insulin (Fig. 5). Scatchard analysis of [125I]-IGF-Ibinding using a two-site model revealed two relativelyhigh-affinity binding sites, with K-, values of 3.0 X 109

M"1 and 1.9 X 109 M~\ respectively. The calculatedreceptor site densities were 7.7 X 104/cell and 1.2 X105/cell, respectively (Fig. 5A). Scatchard analysis of[125I]-insulin binding using a two-site model revealedboth high- and low-affinity binding sites. The IC, valuesfor the high- and low-affinity binding sites were 3.2 X109 M"1 with 7.4 X 103 receptor sites per cell, and 4.2X 107M-1 with 6.6 X 104 receptor sites per cell, respec-tively (Fig. 5B).

DISCUSSION

In this study, cultured HRPE cells of fetal origin wereshown to express genes for IGF-I and IGF-II. In addi-tion, immunoreactivities to their protein productswere detected in conditioned culture medium. Fur-thermore, we demonstrated expression of genes forinsulin receptor, type I and II IGF receptors, butamong the six subtypes of IGFBP, we detected only



5 10 15Bound (fmol/100 jxg)

-12-11 -10 -9 -8 -7 -6 -5 -4

Insulin ( M, log,0)

FIGURE 5. Binding of IGF-I and insulin to HRPE cells. (A)Competitive inhibition of [125I]-IGF-I binding by unlabeledIGF-I. Inset shows Scatchard analysis of [I25I]-IGF-I bind-ing. (B) Competitive inhibition of [125I]-insulin binding byunlabeled insulin. Inset shows Scatchard analysis of [125I]-insulin binding. Each point represents the mean of measure-ments in quadruplicate. Typical results from three indepen-dent experiments are shown. The HRPE cells express a rela-tively abundant number of receptors for insulin and IGF-I.

IGFBP-2. Additional proof for the presence of thegenes for insulin and IGF receptors is provided by themitogenic activity of their ligands, and the abundanceof ligand-binding sites for insulin and IGF-I.

Expression of IGF-I in RPE cells has been re-ported, and this peptide is thought to act as an auto-crine/paracrine mediator15; IGF-I and type I and IIIGF receptor gene expression also has been reportedin RPE cells.12 Our study confirmed previous findingson the presence of IGF-I and IGF receptors, andshowed that there also is IGF-II gene expression andimmunoreactivity in conditioned culture medium. Aprevious report showing negative IGF-II gene expres-sion12 is in contradiction to the positive results in the

current study. Differences in the ages of the eyes (fetalcells in our study and adult origin in the previous re-port) from which the RPE cells were derived may ex-plain the differences in results. A comparative studyon RPE cells of fetal and adult origin is an importantfuture project.

We found stimulation of DNA synthesis by IGF-IIas well as IGF-I and insulin, with IGF-I being the mostpotent stimulator (Fig. 4). Our findings suggest thattype I IGF receptor-mediated stimulation of DNA syn-thesis can be elicited by both IGF-I and IGF-II. Re-ceptor antibody-blocking studies would be of value inconfirming specificity of ligand-receptor interaction.The dose-response curve for insulin was not parallelto that of IGF-I. Insulin induced greater stimulation atlow concentration, but was less potent at higher con-centrations (Fig. 4). A possible explanation may bethat activation of insulin receptors induces mitogene-sis more effectively than activation of IGF receptors.At lower concentrations, insulin induces greater stimu-lation because the insulin receptor is more potent thanthe IGF receptors. With high concentrations of IGF,much larger numbers of IGF receptors are activated,thus inducing more potent mitogenesis. This supportsthe idea that cultured HRPE cells express functionalinsulin receptors.

In cultured bovine RPE cells, the type II IGF re-ceptor is abundantly expressed, but only modest levelsof type I IGF receptor and virtually no insulin recep-tor are found.34 In contrast, in our study, we detecteda relatively high level of expression of the genes for theinsulin receptor as well as those for both IGF-I andIGF-II (Fig. 5B), levels comparable to those reportedin aortic endothelial and smooth muscle cells.33 Asshown in Figure 5A, the cells have a larger number of[125I]-IGF-I-specific binding sites than do monkeyRPE cells.35 Scatchard analysis indicates that theHRPE cells have more IGF-I binding sites comparedto those in vascular cells and glial cells.3637 Furtherstudies are required to determine if there is a uniqueexpression pattern for insulin and IGF receptors infetal HRPE cells.

Mannose-6-phosphate receptors (ie, type II IGFreceptor) are known to be located on the apical mem-brane of RPE cells, and may be involved in the phago-cytosis of photoreceptor outer segments. IGF-II is re-ported to modulate the surface expression and func-tion of type II IGF receptors.3839 The secretion ofIGF-II by RPE cells and the existence of the type IIIGF receptor gene in the cells may imply not only par-acrine- but also autocrine-mediated function of IGF-II, possibly including the regulation of phagocytosis.

IGFBPs have been implicated in the local as well asin the systemic regulation of IGF activities. Culturedbovine RPE cells secrete IGFBPs that have a molecularweight of about 31 kD, and they are immunologically



related to IGFBP-2.34 Cultured monkey RPE cells se-crete 36- to 49-kD IGFBPs, some of which contain anoligosaccharide side chain with iV-glycosidic linkage,whereas others have no side chains.35 The gene forIGFBP-2 was not detected in mouse RPE cells by insitu hybridization.40 We tried to identify which type ofIGFBP, among IGFBP-1 to IGFBP-6, is expressed inRPE cells, and found that only the gene for IGFBP-2was expressed (Fig. 2). Because we did not have posi-tive controls in our study, however, care must be takenin interpreting the absence of IGFBP gene expressionother than that of IGFBP-2. Based on Southern blotanalysis, probes for IGFBP-2 hybridized to the band ofIGFBP-5 (Table 2), but PCR analyses were negative(Fig. 2). Further confirmation of IGFBP-5 gene ex-pression is required. IGFBP-2 has noAMinked glycosyl-ation site, and is reported to have a molecular weightof 31.5 to 36 kD. These properties correspond tosome extent to those reported in RPE tissue culturesupernatant. IGFBP-2 has a marked tendency to bindpreferentially to IGF-II rather than IGF-I, which maysuggest that regulation of IGF-II function is impor-tant for HRPE cells.

In conclusion, we confirm the previously reportedexpression of IGF-I in cultured fetal HRPE cells, andthe secretion by these cells of IGF-II. Gene expressionof IGF receptors, insulin receptor, and IGFBP-2 re-ceptor also was shown. HRPE cell proliferation wasstimulated by IGF-II as well as IGF-I and insulin, possi-bly through their interactions with insulin and/or typeI IGF receptors. These findings imply possible auto-crine/paracrine roles for IGF-II and IGF-I.

Key Words

retinal pigment epithelial cells, insulin-like growth factor(IGF)-related genes, IGF-related receptors, IGF-relatedbinding proteins, gene expression

Acknowledgments

The authors thank Dr. George L. King, Joslin DiabetesCenter (Boston, MA) for his kind help in the binding studyof IGF-I and insulin, and Dr. Peter S. Reinach, Medical Col-lege of Georgia (Augusta, GA) for his critical reading of themanuscript.

References

1. Blundell TL, Humbell RE. Hormone families: pancre-atic hormones and homologous growth factors. Na-ture. 1980;287:781-787.

2. Zapf J, Schmid C, Froesch ER. Biological and immuno-logical properties of insulin-like growth factors (IGF)I and II. Clin Endocrinol Metab. 1984; 13:3-30.

3. Humbel RE. Insulin-like growth factors I and II. EurJBiochem. 1990; 190:445-462.

4. Shimasaki S, Ling N. Identification and molecularcharacterization of insulin-like growth factor binding

proteins (IGFBP-1, -2, -3, -4, -5 and -6). Progress inGrowth Factor Research. 1991; 3:243-266.

5. Waldbillig RJ, Fletcher RT, Somers RL, Chader GJ.IGF-I receptors in the bovine retina: structure, kinaseactivity and comparison with retinal insulin receptors.ExpEyeRes. 1988;47:587-607.

6. Zick Y, Spiegel A, Sagi-Eisenberg R. Insulin-likegrowth factor I receptors in retinal rod outer seg-ments. JBiol Chem. 1987; 262:10259-10264.

7. Hansson HA, Holmgren A, Norstedt G, Rozell B.Changes in the distribution of insulin-like growth fac-tor I, thioredoxin, thioredoxin reductase and ribonu-cleotide reductase during the development of the ret-ina. Exp Eye Res. 1989;48:411-420.

8. Tesoriere G, Vento R, Calvaruso G, Taibi G, GiulianoM. Identification of insulin in chick embryo retinaduring development and its inhibitory effect on DNAsynthesis. J Neurochem. 1992; 58:1353-1359.

9. Waldbillig RJ, Arnold DR, Fletcher RT, Chader GJ.Insulin and IGF-I binding in developing chick neuralretina and pigment epithelium: a characterization ofbinding and structural differences. Exp Eye Res.1991;53:13-22.

10. Roark EF, Paradies NE, Lagunowich LA, GrunwaldGB. Evidence for endogenous proteases, mRNA leveland insulin as multiple mechanisms of N-cadherindown-regulation during retinal development. Develop-ment. 1992; 114:973-984.

11. Cuthbertson RA, Beck F, Senior PV, et al. Insulin-likegrowth factor II may play a local role in the regulationof ocular size. Development. 1989; 107:123-130.

12. Martin DM, Yee D, Feldman EL. Gene expression ofthe insulin-like growth factors and their receptors incultured human retinal pigment epithelial cells. MolBrain Res. 1992; 12:181-186.

13. Shepherd VL, Tarnowski BI, McLaughlin BJ. Isola-tion and characterization of a mannose receptor fromhuman pigment epithelium. Invest Ophthalmol Vis Sci.1991;32:1779-1784.

14. Tarnowski BI, Shepherd VL, McLaughlin BJ. Expres-sion of mannose receptors for pinocytosis and phago-cytosis on rat retinal pigment epithelium. Invest Oph-thalmol Vis Sci. 1988; 29:742-8.

15. Waldbillig RJ, Pfeffer BA, Schoen TJ, et al. Evidencefor an insulin-like growth factor autocrine-paracrinesystem in the retinal photoreceptor-pigment epithelialcell complex. / Neurochem. 1991;57:1522-1533.

16. Kuriyama S, Ohuchi T, Yoshimura N, Honda Y.Growth factor-induced cytosolic calcium ion tran-sients in cultured human retinal pigment epithelialcells. Invest Ophthalmol Vis Sci. 1991;32:2882-2890.

17. Chomczynski P, Sacchi N. Single-step method of RNAisolation by acid guanidinium thiocyanate-PhOH-chlo-roform extraction. Anal Biochem. 1987; 162:156-159.

18. Bell G, Selby M, Rutter W. Sequence of the humaninsulin gene. Nature. 1980;284:26-32.

19. Bell G, Merrywether J, Sanchez-Pescador R, et al. Se-quence of a cDNA clone encoding human preproinsu-lin-like growth factor II. Nature. 1984;310:775-777.

20. Binkert C, Landwehr J, MaryJ, SchwanderJ, HeinrichG. Cloning, sequence analysis and expression of a



cDNA encoding a novel insulin-like growth factorbinding protein (IGFBP-2). EMBO J. 1989;8:2497-2502.

21. Brinkman A, Groffen C, Kortleve DJ, Geurts vKA,Drop SL. Isolation and characterization of a cDNAencoding the low molecular weight insulin-like growthfactor binding protein (IBP-1). EMBO J. 1988; 7:2417-2423.

22. Cubbage ML, Suwanichkul A, Powell DR. Insulin-likegrowth factor binding protein-3: organization of thehuman chromosomal gene and demonstration of pro-moter activity./Biol Chem. 1990;265:12642-12649.

23. Ebina Y, Ellis L, Jarnagin K, et al. The human insulinreceptor cDNA: the structural basis for hormone-acti-vated membrane signaling. Cell. 1985;40:747-758.

24. Kiefer MC, Masiarz FR, Bauer DM, Zapf J. Identifica-tion and molecular cloning of two new 30-kDa insulin-like growth factor binding proteins isolated fromadult human serum. / Biol Chem. 1991; 266:9043-9049.

25. Kiefer MC, Ioh RS, Bauer DM, Zapf J. Molecularcloning of a new human insulin-like growth factorbinding protein. Biochem Biophys Res Commun.1991;l76:219-225.

26. Morgan D, Edman J, Standring D, et al. Insulin-likegrowth factor II receptor as a multifunctional bindingprotein. Nature. 1987;329:301-307.

27. Rotwein P, Pollok K, Didier D, Krivi G. Organizationand sequence of the human insulin-like growth factorI gene./ Biol Chem. 1986;261:4828-4832.

28. Ulrich A, Gray A, Tarn A, et al. Insulin-like growthfactor 1 receptor primary structure: comparison withinsulin receptor suggests structural determinants thatdefine functional specificity. EMBOJ. 1986; 5:2503-2512.

29. Saiki R, Scharf S, Faloona F, et al. Enzymatic amplifi-cation of j8-globin genomic sequences and restrictionsite analysis for diagnosis of sickle cell anemia. Science.1985;230:1350-1354.

30. Marchuk D, Drumm M, Saulino A, Collins F. Con-struction of T-vectors, a rapid and general system fordirect cloning of unmodified PCR products. NuclAcids Res. 1991; 19:1154-1154.

31. Sanger F, Nicklen S, Coulson A. DNA sequencingwith chain- terminating inhibitors. Proc Natl Acad SciUSA. 1977;74:5463-5467.

32. Daughaday WH, Mariz IK, Blethen SL. Inhibition ofaccess of bound somatomedin to membrane receptorand immunobinding sites: a comparison of radiorecep-tor and radioimmunoassay of somatomedin in nativeand acid-ethanol-extracted serum. J Clin EndocrinolMetab. 1980;51:781-788.

33. King G, Buzney S, Kahn C, et al. Differential respon-siveness to insulin of endothelial and support cellsfrom micro- and macrovessels. / Clin Invest.1983;71:974-979.

34. Ocrant I, Fay CT, Parmelee JT. Expression of insulinand insulin-like growth factor receptors and bindingproteins by retinal pigment epithelium. Exp Eye Res.1991;52:581-589.

35. Waldbillig RJ, Schoen TJ, Chader GJ, Pfeffer BA.Monkey retinal pigment epithelial cells in vitro synthe-size, secrete, and degrade insulin-like growth factorbinding proteins. / Cell Physiol. 1992; 150:76-83.

36. King GL, Goodman AD, Buzney S, Moses A, KahnCR. Receptors and growth-promoting effects of insu-lin and insulinlike growth factors on cells from bovineretinal capillaries and aorta. / Clin Invest. 1985;75:1028-1036.

37. Han VKM, Lauder JM, D'Ercole AJ. Characterizationof somatomedin/insulin-like growth factor receptorsand correlation with biologic action in cultured neona-tal rat astroglial cells./Neurosci. 1987;7:501-511.

38. Braulke T, Tippmer S, Chao H-J, von Figura K. Insu-lin-like growth factor I and II stimulate endocytosisbut do not affect sorting of lysosomal enzymes in hu-man fibroblasts.y^o/ Chem. 1990;265:6650-6655.

39. Braulke T, Tippmer S, Neher E, von FK. Regulationof the mannose 6-phosphate/IGF II receptor expres-sion at the cell surface by mannose 6-phosphate, insu-lin like growth factors and epidermal growth factor.EMBOJ. 1989; 8:681-686.

40. Agarwal N, Hsieh C-L, Sills D, et al. Sequence analy-sis, expression and chromosomal localization of agene, isolated from a subtracted human retina cDNAlibrary, that encodes an insulin-like growth factor bind-ing protein {\G¥WP2). Exp Eye Res. 1991; 52:549-561.


insulin-like growth factor-related genes, receptors, and binding

Documents