Expression of insulin-like growth factor I in developing lens is compartmentalized.

We, and others, have recently reported that insulin-like growth factor I (IGF-I) mRNA is expressed in multiple tissues during embryogenesis and in whole embryos during early organogenesis. Therefore, it is likely that, in addition to any effect on embryo growth, IGF-I plays a paracrine/autocrine role in development. The embryonic chicken lens, an avascular organ composed by a single type of cell that undergoes differentiation in vivo and in vitro, is an ideal model to characterize the paracrine/autocrine action of IGF-I. The lens cells express IGF-I receptors, and respond to exogenous IGF-I with induction of fiber cells differentiation and stimulation of delta-crystallin gene transcription. Whether embryonic lens cells express IGF-I was uncertain. In the present study, we used a sensitive semiquantitative method (reverse transcription of RNA followed by amplification with the polymerase chain reaction) to analyze IGF-I gene expression. An amplified product of the expected length (209 base pairs) was found in days 8, 12, 15, and 19 lenses. At all embryo ages studied, the product was more readily detected in the lens than in the liver, while in eye tissues (excluding lens), IGF-I expression was relatively high. After microdissection of the epithelial cells from the fully differentiated fiber cells, IGF-I expression was detected exclusively in the epithelial cells. IGF-I immunoactivity was found using high performance liquid chromatography followed by radioimmunoassay in the days 8-19 lens extracts, and in primary cultures of isolated epithelial cells. Our previous and present findings show that the lens has all the elements for IGF-I autocrine/paracrine action in development.

We, and others, have recently reported that insulinlike growth factor I (IGF-I) mRNA is expressed in multiple tissues during embryogenesis and in whole embryos during early organogenesis. Therefore, it is likely that, in addition to any effect on embryo growth, IGF-I plays a paracrine/autocrine role in development. The embryonic chicken lens, an avascular organ composed by a single type of cell that undergoes differentiation in vivo and in vitro, is an ideal model to characterize the paracrine/autocrine action of IGF-I. The lens cells express IGF-I receptors, and respond to exogenous IGF-I with induction of fiber cells differentiation and stimulation of 6-crystallin gene transcription. Whether embryonic lens cells express IGF-I was uncertain. In the present study, we used a sensitive semiquantitative method (reverse transcription of RNA followed by amplification with the polymerase chain reaction) to analyze IGF-I gene expression. An amplified product of the expected length (209 base pairs) was found in days 8, 12, 15, and 19 lenses. At all embryo ages studied, the product was more readily detected in the lens than in the liver, while in eye tissues (excluding lens), IGF-I expression was relatively high. After microdissection of the epithelial cells from the fully differentiated fiber cells, IGF-I expression was detected exclusively in the epithelial cells. IGF-I immunoactivity was found using high performance liquid chromatography followed by radioimmunoassay in the days 8-19 lens extracts, and in primary cultures of isolated epithelial cells. Our previous and present findings show that the lens has all the elements for IGF-I autocrine/paracrine action in development.
The lens, an encapsulated organ that develops early in embryogenesis, lacks blood vessels and nerves. It is a tissue of epithelial origin, formed by a front layer of cubical epithelial cells that divide, migrate, elongate, and finally become fully differentiated fiber cells that constitute the main body of the mature lens (reviewed in Ref. 1). Lentropin, a protein that stimulates elongation, was partially purified from chick embryo vitlreous humor, and was found by Beebe et al. (2, 3) to * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
On be related to IGF-I.' A similar differentiation effect had previously been observed with fetal serum, vitreous humor (1, 4), and with high concentrations of insulin (5-7). Thus, growth factors of the insulin family and, perhaps, fibroblast growth factor (8,9) and platelet-derived growth factor (lo), may be essential for normal development of the lens. Fiber cell differentiation during organogenesis of the chick embryo lens is typically associated with an increase in 6-crystallin expression ( 7 , l l ) . We have previously shown that exogenous IGF-I stimulates transcription of the endogenous 6-crystallin gene, as well as of a transfected 61-crystallin/chloramphenicol actetyltransferase hybrid gene (11,12). We, and others, have demonstrated that both epithelial and fiber cells display IGF-I receptors that are highly regulated in development (13)(14)(15).
While IGF-I has been largely considered a postnatal growth factor in mammals, there is evidence suggesting that IGF-I also has autocrine/paracrine actions during development. IGF-I mRNA and IGF-I have been detected in multiple fetal tissues in mammals (16)(17)(18)(19)(20)(21). In the chick embryo, IGF-I mRNA and IGF-I are expressed in the whole embryo during gastrulation and neurulation, before circulation is established (22,23). Immunoactive IGF-I is detected in serum by day 6, peaks at days 14-16, and then decreases markedly by day 20 (hatching is at day 21) (23). In a model of chick embryo cultured ex ouo, with marked retarded growth, we have shown that the IGF-I midembryogenesis serum peak is abolished (23). While in this situation and others (18) the serum concentration of IGF-I appears to relate to the overall growth of an organism, to understand the local action (paracrine or autocrine) of IGF-I we need information on the levels of IGF-I mRNA and IGF-I in individual tissues.
The goal of our present study was to document that embryonic lens contained IGF-I mRNA and IGF-I, and to study the changes in ontogeny. We found IGF-I expression in the lens at all ages studied, from late organogenesis until near birth. Furthermore, the changes in IGF-I mRNA and IGF-I during development in the lens do not correlate with serum IGF-I. Within the lens, only the epithelial cells, and not the fully differentiated fiber cells express the IGF-I gene. This selectivity is a clear example of cell-compartmentalized expression of IGF-I within a developing tissue that has the ability to differentiate in response to IGF-I.

20786
humidity. On appropriate days of development, the embryos were morphologically staged (24) and the lens was microdissected from the rest of the eye.
The embryos were decapitated and a lateral incision at the corneal limbus of the eye was made. The vitreous body with the attached lens was gently expressed to the outside of the eye and was grasped with fine forceps (25). The lens was carefully lifted with curved forceps, rolled over filter paper to remove adherent iris, washed in unsupplemented DMEM, and then frozen at -70 "C until the protein or RNA was extracted.
Primary Lens Cells Culture-Patches of lens epithelial cells from day 12 chick embryos (El2) were prepared as described (26). Epithelia from nine lenses were used per 35-mm dish (total of 90 epithelia). After 24 h of incubation in 2 ml of DMEM, supplemented with 10% fetal calf serum, at 37 "C in 5% COB, the cells were washed three times with 3 ml of DMEM. The cells were then harvested by adding 0.5 ml of 0.01% trypsin/0.02% EDTA solution for 2 min, scraped from the dish and centrifuged for 1 min at 500 X g. The cell pellet was washed twice in DMEM and then homogenized in 0.25 M Tris, pH 7.5. The homogenate was desalted on a column of Sephadex G-25 and the eluate was lyophilized and resuspended in HPLC buffer (see later).
Isolation of RNA-Whole lenses (either 10 of Ea, E12, and El5 or 5 EI9) were homogenized in RNazol (2 m1/100 mg of tissue) for 1 min with a glass-teflon homogenizer, and total RNA was extracted according to the method of the RNazol manufacturer (Cinna/Biotec, Friendswood, TX, Ref. 27). For the study of epithelial and fiber cells, 15 lenses of EI9 were microdissected and washed in DMEM. The epithelial cells attached to the capsule were separated from the fiber cells. Each group of cells was homogenized and processed as described above for the whole lenses. The eye of a day 19 embryo was microdissected and the vitreous body and the lens were discarded; the residual tissue was processed as described for the lens. In all preparations the RNA concentration was determined by absorbance at 260 nm and the purity by the 260 nm:280 nm ratio. The integrity of ribosomal RNA was confirmed by gel electrophoresis and ethidium bromide staining (results not shown).
Design of Oligonucleotides-Oligonucleotides complementary to the sequence of chicken IGF-I determined previously (22,28)  After heating at 80 "C for 3-4 min and cooling to 42 'C, 1 p1 of AMV reverse transcriptase was added and the incubation was continued a t 42 "C for 1 h. For the PCR a 25-pl aliquot of the cDNA reaction mixture was used. In a final volume of 100 pl the following were added 200 ng of each primer U and D, 1.25 mM dNTPs and 1 unit of Taq polymerase, as recommended in the GeneAmp Kit (Perkin-Elmer-Cetus Corp., Emeryville, CA). 40 cycles were performed; each cycle consisted of denaturation (94 "C for 1 min), annealing (50 "C for 1.5 min), and extension (70 "C for 2 min); the extension time of the last cycle was 10 min.
DNA Blot Hybridization-An aliquot of the PCR-amplified product was fractionated on a 2% agarose gel containing ethidium bromide in Tris-borate-EDTA (TBE 1 X) buffer. Following electrophoresis the products were visualized by UV transillumination and the position of the expected bands was compared to a 6x174 ladder (Promega). The gels were denatured for 1 h in 0.5 N NaOH in 1.5 M NaCl, and washed for 1 h in 1 M Tris containing 1.5 M NaCl. The DNA was transferred to nylon filters (Nytran, Schleicher and Schuell; Keene, NH) by capillary action, using 5 X SSC. The filters were prehybridized and then hybridized (29) to a 32P-labeled oligonucleotide (Fig. lA), at 37 "C for 16-20 h. The filters were then washed in 2 X SSPE with primers used for the RT-PCR, and the predicted length (209 bp) of tha amplified product are shown. Note that this product spans the putative position of a large intron (>21 kb) and only mRNA and not any contaminant DNA should result in the amplification of a product of the predicted size. The position of the oligonucleotide used as a probe in Southern hybridization is indicated. B, IGF-I gene expression in whole lens during ontogeny. RT-PCR was performed using 10 pg of total RNA, purified from pools of lenses at days 8, 12, 15, and 19 of development (see "Experimental Procedures"). The same samples digested with RNase A (+ lanes) were used as negative controls. The amplified products were fractionated on a 2% agarose gel and transferred to a nylon filter (Nytran). The filter was hybridized with the 32P-labeled chicken IGF-I oligonucleotide probe shown in A. An autoradiogram was generated after exposure for 4 weeks at -70 "C. The arrow indicates the 209-bp predicted band corresponding to the specific amplified product. Results were comparable with other independent RNA preparations of the same ages, and in several PCRs. Also, an RNA preparation from day 9 lens showed a level of signal similar to day 8, higher than in day 12 lens. 0.05% sodium dodecyl sulfate at 37 "C. Autoradiograms were generated using XAR-2 Kodak film, with one intensifying screen at -70°C.
SepPak Chromatography-Prior to the chromatographic step to remove the IGF-I binding proteins (23), the lenses were homogenized in 1-2 ml of 0.25 N HCl using a hand-held glass/teflon homogenizer with 10 strokes. The homogenate was centrifuged in an Eppendorf microcentrifuge for 5 min at 14,000 rpm at 4 "C, and the supernatant was transferred to a tube and diluted 1:l with 0.5 N HCl. After incubation for 1 h at 20 "C the mixture was loaded onto a C-18 SepPak cartridge; the eluate was recycled once, the column was washed with 4% acetic acid, and the retained material was eluted with 10% methanol. The eluate was evaporated to dryness in a vacuum centrifuge. All the samples were dissolved in 1 ml of HPLC buffer (750 pl of solvent A and 250 p1 of solvent B, described below).
Reverse-phase HPLC-Chromatography was performed on a Gilson system, using a Vidac C-18 hydrophobic interaction TP104 column as described (30). The two-solvent linear gradient consisted of solvent A (0.1% trifluoroacetic acid in H20) and solvent B (80% acetonitrile/20% H20/0.1% trifluoroacetic acid). Aliquots of the extracts (as indicated in the corresponding figures) were injected in 25% solvent B/75% solvent A, and solvent B was maintained at 25% for 20 min. A linear gradient was then established from 25 to 65% over 40 min. Fractions were collected every min (1 ml) for the 60-min period. They were evaporated in a vacuum centrifuge and then resuspended in RIA buffer. 12'I-IGF-I (Amersham, Arlington Heights, IL) and human IGF-I (AmGen Biologicals, Thousand Oaks, CA) were used as markers. Prior to the injection of each extract, a blank sample was run and the appropriate fractions were processed and tested in the RIA; this blank served as a control. In each case a protein profile was automatically recorded by measuring the absorbance at 214 nm.
IGF-I RIA-The IGF-I RIA is a heterologous assay, the label is human 1251-IGF-I, and the antibody is antihuman IGF-I antibody UBK 487 (kind gift of the National Pituitary Program of the National Institute of Diabetes, Digestive and Kidney Diseases). The conditions of the RIA were according to Furlanetto (31). This assay system was validated previously for measuring chicken IGF-I (23, 32). Aliquots of the extracts from whole lens and epithelial cell cultures that had been fractionated by SepPak and HPLC were dissolved in RIA buffer and tested in the IGF-I RIA as described (23).

IGF-I mRNA and IGF-IAre Expressed in Developing Whole
Lens-We searched for IGF-I mRNA in the lens during ontogeny from day 8 through day 19 (hatching is at day 21). Suspecting that the level of IGF-I gene expression might be very low, based on our experiments in whole embryos (22), we used a very sensitive RT-PCR strategy to amplify IGF-I mRNA followed by gene-specific hybridization (Fig. L4). We detected IGF-I mRNA in the whole lens from embryos at all ages studied (Fig. 1B). The signal at day 19 was higher in several independent experiments when compared with lenses obtained from younger age embryos. The product amplified was of the expected size, and RNase pretreatment of the samples prevented its appearance, indicating that the band represented mRNA in the sample and not contaminating DNA.
To determine the expression of IGF-I, whole lenses from the same ages were studied. All the samples tested contained immunoactive IGF-I after HPLC (Fig. 2). The immunoactivity eluted in a single peak in a position similar to that of chicken serum IGF-I (which in this type of gradient also coincides with the position of human IGF-I) (23). The content of IGF-I per lens increased slightly during the second week of embryogenesis and there was a further increase at day 19 (Table I), consistent with the increase in IGF-I mRNA at that age.

Differential IGF-I Gene Expression in the Embryo
Lens Cells, Eye, and Liver-Since the lens is composed of a single layer of undifferentiated epithelial cells located next to the anterior part of the capsule and a large mass of fiber cells, the two cell groups can be studied separately. To analyze whether both epithelial and fiber lens cells expressed IGF-I mRNA, the RNA extracted from the cells after careful microdissection of the cell layers was amplified by RT-PCR. The epithelial lens cells, but not the fiber cells, were found to express the IGF-I gene (Fig. 3). In addition, the eye tissues minus the lens also expressed IGF-I mRNA in relatively high levels (Fig. 3). Although our RT-PCR protocol does not allow strict quantitation of mRNA, the amount of RNA in the reaction had a direct relationship with the signal obtained after amplification. Serial dilutions of the eye total RNA resulted in a progressive decrease in the specific signal (Fig. 3). The lens of the same age embryos (day 19) had a signal that fell between the 1:25 and 1:150 dilutions of the whole eye minus lens, i.e. the level of IGF-I mRNA in the lens is probably 2 orders of magnitude lower than in the structures of the eye minus the lens. In contrast with the IGF-I mRNA in the lens and the eye during embryogenesis, its presence in developing liver was barely detectable, using similar amounts of total RNA in the RT-PCR reaction (Fig. 3C). This suggests that the liver is not an important source of IGF-I prenatally.
To confirm that the epithelial cells of the lens contained IGF-I, we established primary cultures with patches of epithelial cells. After 24 h in culture an extract was chromatographed on C-18. The HPLC reactions contained immunoactive IGF-I (Fig. 4). The concentration of IGF-I per lensequivalent was slightly higher than in the extract of the whole   RNA from either day 19 whole lens, epithelial cells, fiber cells, or the eye minus lens, as described in the legend for Fig. 1. The initial RNA samples were either untreated (-lanes) or treated with RNase A (+ lanes). The exposure of the autoradiogram was for 3 days a t -70 "C; a short exposure (4 h) of the eye lanes is also shown. The presence of a n amplified product in RNA from epithelial cells and the absence in RNA from the fiber cells were confirmed in several independent RT-PCRs, and with another series of RNA preparations. Integrity of the RNA from fiber cells had been confirmed by gel electrophoresis and ethidium bromide staining of rRNA (results not shown). B, RNA concentration dependency of the hybridization signal. Prior to the RT-PCR, 10 pg of total RNA from day 19 eye (minus lens and vitreous humor) were diluted as indicated. As negative control we used distilled water ( W ) . 10 pg of total RNA of the day 19 lens were also analyzed. The amplification reaction and the hybridization were performed as described in the legend for Fig. 3. The amplified band corresponds to the 209-bp product and is indicated by an arrowhead. The autoradiogram was exposed for 18 h a t -70 "C. C, IGF-I gene expression in developing liver. RT-PCR was performed using 10 pg of total RNA from liver of embryos a t days 12, 14, 16, 19, and 50day-old chicken ( G O ) . The samples were either untreated (-lanes) or treated with RNase A (+ lanes) prior to RT-PCR. The amplified products were processed and the hybridization performed as described in the legend for Fig. 3. The exposure time of the autoradiogram was 3 days. The results presented in all panels (A-C) were obtained with products from the same PCR experiment. lens at the corresponding age (day 12). On a per cell basis, the concentration of IGF-I in epithelial cultures was slightly higher than that found in the embryo lens extracts of days 8, 12, and 15; this is a further suggestion that the cells in culture produced the peptide (Table I).

DISCUSSION
In the present study we show evidence that the embryonic lens, a tissue responsive to IGF-I, expresses the IGF-I gene. IGF-I mRNA is found in the lens in a subset of its cells, the epithelial cells, that are precursors of the differentiated fiber cells. The embryonic lens contains also IGF-I. The developmental profile of the IGF-I detected in lens extracts roughly parallels the levels of IGF-I mRNA in whole lens. The highest values of both peptide and mRNA are found at day 19,2 days before hatching (Table I and Fig. 1B). It is worth noting that this is not the developmental pattern exhibited by serum IGF-I, that increases from day 6 to a peak concentration at day 15 and then decreases markedly the days before hatching (23). This divergency between circulating IGF-I and lens IGF-I supports the local origin and autocrine/paracrine function of IGF-I in the lens.
Since IGF-I mRNA appeared to be restricted only to the epithelial, mitotic cells and absent from the fiber cells (Fig.  3), the peptide found in whole lens extracts most likely was produced in the epithelial layer. We could find IGF-I immunoactive protein in primary cultures of lens epithelial cells from day 12 embryos in comparable amounts to the IGF-I from fresh-frozen lens, supporting the concept of cell-specific expression of IGF-I (Fig. 4). Apparently, IGF-I is expressed by the least differentiated cells, suggesting that it may stimulate their differentiation process in uiuo. (We have not excluded that partially differentiated cells in vitro or equatorial cells in vivo also expresses IGF-I.) Parenthetically, the expression of IGF-I in one subset of cells, while the neighbor cells also respond to the peptide, occurs postnatally in the ovary; only the granulose cells contain IGF-I mRNA, but not theca cells, while both respond to IGF-I (33).
Lentropin, a factor related immunologically to IGF-I and clearly distinct from insulin, had been partially purified from day 15 chick embryo vitreous humor (2,3) and proposed as a physiological factor responsible for the differentiation of epithelial lens cells into fiber cells. The reported size of lentropin, M, -60 kDa (3) is much larger than the chicken IGF-I gene primary translation product (prepro-IGF-I, M , -17.5 kDa) (28). It remains to be clarified whether lentropin represents IGF-I bound to binding proteins, or an IGF-I binding protein alone, with IGF-I-like effects, or another molecule. We have confirmed that the vitreous humor from chicken embryo as early as day 6 contains both IGF-I and IGF-I binding proteins.2 Several other aspects related to IGF-I differ between fully differentiated lens fiber cells and epithelial cells. While both cells contain IGF-I receptors, epithelial cells did, but fiber cells did not, internalize gold-labeled IGF-I (34). The level of IGF-I binding to membrane preparations was slightly lower T. CaldBs, J. Alemany, H. L. Robcis, and F. de Pablo, unpublished observations. in fiber cells than in epithelial cells at all ages studied. The concentration of receptors decreased in both cells with embryo age from day 6 to day 19 (13). Therefore, the effect of IGF-I in uiuo is modulated by changes not only in the local peptide levels, but also changes in the level of cell receptors and, probably, the level of IGF-I binding proteins.
The whole eye (minus the lens) contains significant amounts of IGF-I mRNA (Fig. 3). We do not know the specific location of this mRNA or the peptide. (Preliminary attempts to localize IGF-I mRNA by in situ hybridization or IGF-I by immunocytochemistry in the eye of day 7 embryos have not been successful, perhaps due to their low concentration.) In the human fetus, IGF-I mRNA has been found in the sclera (19) and in the rat fetus in the iris (35). We were surprised that the IGF-I mRNA in the eye was markedly higher than in the embryonic liver (Fig. 3). However, in a previous study we had obtained similar results in liver using different oligonucleotides as primers for the RT-PCR (22); others have confirmed the very low expression of the IGF-I gene in developing liver during chicken embryogenesis by an independent technique, solution hybridization, and RNase protection assay (36). It is intriguing that in the only species in which a complete developmental profile for circulating IGF-I is available, the chicken, the contribution by hepatic IGF-I appears negligible and the regulation by growth hormone seems very unlikely during embryogenesis (36). Since multiple embryonic tissues from both fetal rat (19,21) and mouse (37, 38) as well as chicken embryos (22) express IGF-I mRNA, there are multiple potential sources for IGF-I in the circulation. Whether the eye is a contributor to IGF-I serum levels in the embryo remains to be studied.
The panoply of effects of IGF-I in developing lens includes stimulation of protein and RNA synthesis, morphological elongation and &crystallin gene transcription, Le. IGF-I is both a growth and a differentiation factor (2, 7, 11,12,39). It should be interesting to analyze in detail the dose dependency of these effects at different developmental points. Perhaps small differences in IGF-I concentration in the lens local milieu (aqueous humor uerszu lens, uerszu vitreous humor) can regulate both growth and differentiation of the lens cells in an orderly manner, as it has been proposed for fibroblast growth factor (8,9). We speculate that a concentration gradient may exist between the IGF-I locally produced by the epithelial cells and the vitreous humor IGF-I, that is in contact with fiber cells. Differences in IGF-I concentration may induce the lens cells either to divide or to differentiate. The embryonic lens, thus, provides a good example of cellcompartmentalized IGF-I gene expression and a confined model for IGF-I autocrine/paracrine action in early embryogenesis.