- Email: [email protected]
Is There a Ubiquitous Growth Factor in the Eye?
Differentiation
Differentiation 18, 29-42 (1981)
0 Springer-Verlag 1981
Is There a Ubiquitous Growth Factor in the Eye? Proliferation Induced in Different Cell Types by Eye-Derived Growth Factors D. BARRITAULT, C. ARRUTI, and Y.COURTOIS Unit6 de Recherches GCrontologiques, INSERM U. 118, ERA CNRS 842, 29, rue Wilhelm, 75016 Pans, France
In a previous work [l]we showed that a neutral extract of bovine adult retina RE can stimulate the growth and modify the morphology of bovine epithelial lens (BEL) cells in vitro. We were also able to demonstrate that the differences in cell shape are closely related to the cell growth properties induced by RE and are mediated by cytoskeletal protein organization as well as external proteins. In this study, we report the results of further investigations on this retinal extract. We show that it possesses all the characteristics of other growth factors such as promoting proliferation in low serum concentration or of enhancing the colony-forming efficiency of BEL cells considerably. By comparing the morphological response of BEL cells treated with RE with the response of other cells to other growth factors, we propose that the phenotypic modifications are cell specific, but not growth factor specific. We report-also that RE has a broad spectrum of activity since it is able to stimulate cells from different origins and species (vascular and corneal endothelial cells, myoblasts, chondrocytes, neuroblastoma cells, and keratinocytes), but not all of them, since it can be toxic for fibroblasts. In this respect, it has an activity similar in many aspects to FGF and EGF, while it differs from them for some target cells. Its action has also been compared with the effects of retinoic acid derivatives and shown to be strikingly different. RE-like activity can be found in other ocular tissues from bovine and other species. The .highest growth-promoting capacities were found in extracts of iris, pigmented epithelium with choroid, and vitreous body. The nature of all these extracts has not yet been determined. Since they are prepared in a similar way and since they have similar growth-promoting activity, we postulate that there is an ubiquitous growth factor in the eye called eye-derived growth factor (EDGF) which may play an important role in physiology and pathology of the eye.
Introduction The use of an in vitro system to study growth interactions between tissues has several advantages insofar as most parameters can be controlled. For instance, it is possible to control all essential nutrients, except there is still a requirement for a certain amount of serum until a synthetic medium is developed. It is therefore possible to look at the interaction between two tissues by techniques such as co-culture or the use of tissue extracts. Because of the progress in cell-culture technology one can hope to understand better the relationships
between two tissues whose maintenance is otherwise vital, and thus mostly inaccessible to in vivo experiments. Thus, the discovery of growth factors is closely linked to the in vitro technology. Conversely, it is possible that growth in these artificial conditions modifies the behaviour of cells. These general considerations were in mind when we decided to investigate the presence of putative factors which might control the growth of the lens. The stimulus to look for such factors in the eye came from many experimental data concerning both normal or pathological development and the control of tissue reconstitution after injury. On the one hand, 0301-4681/81/0018/0029/$ 02.80
30 many investigators have pointed out that the retina must produce some diffusible factors which control proliferation and differentiation of lens epithelial cells during development [2-41. On the other hand, the lens increases continuously in size during the entire lifespan of the individual [5] and this growth corresponds to the differentiation of a pool of epithelial cells very precisely located at the periphery of the lens. Paracentesis of the aqueous humour [6] has demonstrated the presence of a factor which stimulates cell division in a lens maintained in vitro in the appropriate medium. Bovine lens epithelial cells (BEL) are easy to culture, and our laboratory has investigated the growth and synthetic properties of these cells in standard conditions, i.e., in 6% foetal calf serum. These cells transformed spontaneously, but nevertheless synthesized lens capsule material [7].We thus decided to investigate the effect of a retinal extract (RE) on these cells and found that this extract was able to stimulate the growth of these cells and to modulate their morphology [l]. We have studied separately how these phenotypic modifications correlated with the organization of the cytoskeletal proteins and with the extracellular deposition of fibronectin [8]. In this work, we will report further investigations on the retinal extract. We will show that RE fulfils the criteria of other potent growth factors. Several different cell types are stimulated by this extract, and their response is of a similar nature to that observed with other growth factors such as epidermal growth factor (EGF) and fibroblast growth factor (FGF). Preliminary results show that RE tends to differ from both of them and from retinoic acid. We will report that similar RE-like activities are found in ocular tissues other than the retina; we suggest that these activities could be involved in the interactions between all these tissues.
Methods Tissue Extracts Adult bovine eyes from freshly slaughtered animals were used. Retinal extract was prepared as previously described [l].(See also accompanying article.) Extracts from tissues such as iris, choroid, pigmented epithelium, cristallin, vitreous body, and aqueous humour were prepared as for the retina, except that the tissues were first homogenized using a Turax blendor. For vitreous body extract, a concentration was made by ammonium sulfate (AS) precipitation and the fraction used for centrifugation was the
D. Bamtault et al.: Eye Derived Growth Factors supernatant of 20% AS precipitated by 60% AS. Through all these extractions, a phosphate saline buffer (PBS) was used and the protein concentration was estimated according to Bradfort [9] and adjusted to 5 mg/ml after dialysis. The extracts were sterilized by filtration through 0.22 pm millipore filters, aliquoted, and kept at -20" C. A crude extract from bovine brain was made at pH 4-5, followed by ammonium sulfate precipitation like in the preparation of FGF [lo].
Other Growth Factors and Chemicak Purified pituitary and brain FGF were generously given by D. Gospodarowicz (San Francisco). Purified EGF was generously given by G. Orth (Paris). H q d r thymidine (25 C. mM-') came from CEA, France, and retinoic acid from Sigma.
Cell Culture Calf and adult bovine epithelial lens (BEL) cells were cultured as in [ll]; the stimulation experiments were performed on serially subcultured cells (3-30 passages). Human lens cells were cultured as described by Tassin et al. [12]. Foetal bovine and chick embryo myoblasts were cultured by M. Fizman (Paris) [13, 141. Primary cultures of human adult epidermal cells were done as by Guedon [15]. Aortic and foetal heart bovine endothelial cells were cultured by D. Gospodarowia [16]. Corneal endothelial cells were grown from adult bovine cornea. The same culture media was used as with BEL cells, but after addition of 10% foetal calf serum. Since RE is a potent growth factor for these cells (171, its presence allowed serial subculture without a sign of in vitro senescence. RE was constantly present in our culture and removed 3 days before stimulation experiments. Primary cultures of rabbit ear and knee chondrocytes were grown by M. T. Corvol as in [IS]. Human Foreskin fibroblasts, Swiss mouse 3T3, bovine from skin, lung, and kidney fibroblasts, as well as a rat smooth muscle cell line (gift from J. P. Mauger) were all cultured in MEM medium supplemented with 10% foetal calf serum and subcultured as in [ll]. Mouse NIE115 adrenergic neuroblastoma cells were grown by L. Legault-Demare as in [19]; experiments for RE stimulation were done after decreasing the foetal calf serum concentration to 0.2%.
Measure of DNA Synthesis and Cell Multiplication Unless otherwise specified, H?dr thymidine incorporation was measured in triplicate experiments on BEL cells cultured in Linbro trays (24 wells, 2cm2 per well) treated or non-treated with the different tissue extracts at different doses for 44 h and labelled for 4 h using 2pCilml of culture. Cells were washed three times with PBS buffer lysed in 0.5 ml of 0.1 N NaOH, 0.1% SDS for 5 min and cold TCA was added to a fiial concentration of 20%. The TCA precipitate was filtered in a Whatman GFlC glass filter, washed with 5 vol. through 5% TCA, air dried, and the radioactivity was counted in aliquid scintillation counter. . Parallel experiments were done without addition of 3H Tdr thymidine, and the cells were dissociated by a trypsin-EDTA solution (0.2g EDTA, 0.5 g trypsin, 1g glucose, 0.58 sodium bicarbonate, 0.4 g KCl, and 8 g Na€l per liter) and counted in a hemacytometer or a Coultertronic cell counter.
31
D. Barritault et al.: Eye Derived Growth Factors For all the other cells studied, petri dishes of 3.5 cm or 5 cm diameter were used. After different time of cultures in the presence or absence of tissue extracts, cells were trypsinized and counted in triplicate. The increase in cell number of myoblasts was estimated after 44 h or 5 days'of culture after lysing the cells by an overnight incubation in 0.1 N citric acid and counting the nuclei [20].This allowed a measure of the total number of cells, including the cells forming the myotubes. (More details about the formation of myotubes, the dosage of specific receptors and the percentage of fusion after RE treatment will be published elsewhere.) Radio-immunoassay Recently brain FGF was shown to be part of the basic protein of m y e h (211.To detect if there was any FGF activity in the different tissues extracts, a radio-immunoassay designed by J a q u e [22]was used. In this experiment, a serum against human basic protein was prepared by J a q u e and was shown to cross-react with bovine FGF. Iso-electrofocusing of Retinal Extract Retinal extract prepared as above was further purified by precipitation of a 20% ammonium sulfate supernatant with 60% ammonium sulfate. The pellet was redissolved in PBS and dialyzed against 1% glycine, and 250 mg of the protein solution in 15 ml was separated on a LKB iso-electrofocusingpreparation slab apparatus using Serva Ampholines pH 2-11, 2% w/v, as described in application note 198 supplied with the LKB equipment. After separation, the gel slab was cut into 30 bands, pH was measured, and the protein content was eluted from each of the bands by transferring the granulated gel into a small column (stopped at one end with a nylon net) and washing this column with PBS buffer. Ampholines were removed by dialysis in the same buffer and protein content was measured as in [ll]. Each fraction was sterilized by filtration and tested for stimulation of thymidine incorporation on BEL cells.
dine incorporation) in 6% foetal calf serum. We show here that it can also stimulate cell proliferation in complete absence of serum. This is illustrated in Table 1. Cells in monolayers were washed several times with a medium without serum. They were then cultured in the different media, described in Table 1, for 24 h. Three hours before processing they were given 1pCi ml-' of 3H thymidine. The results indicated that an increase in both cell number and thymidine incorporation was brought about by the RE and the serum. They also showed that there may be a synergism between the serum and the RE since the combination of both resulted in greater stimulation than the addition of either one separately.
-
Table 1. Effect of RE on 3Hthymidhe incorporation in the DNA of BEL cells in the absence of serum Medium MEM MEM RE (50 pg-ml-') SM SM + RE (50 pg * I&')
+
Cell number
cpm. 10-4 cells
344
5.6 10' 7.6 10' 7.0 10'
2934 1514
9.2 10"
6460
SM: Standard Medium, i s ; MEM containing 6% vlv calf S e m MEM: Eagle's Minimum essential Medium without sem Each result is the mean of two experiments Increase in cell number was measured after 24-h exposure to the different media; (p) thymidine incorporation was measured in the last 3 h
Molecular Weight Estimation After partial purification with ammonium sulfate, RE was dialyzed against PBS or 8 M urea and centrifuged at 2,000g on Amicon membrane cones Centriflow CF 25 and CF 50 (molecular weight cut off at about 25,ooO and 50,000 daltons, respectively). Protein content of the fractions retained and nonretained by the cones was estimated, and each fraction was dialyzed back against PBS and tested for stimulation of thymidine incorporation in BEL cells. A 75 Amicon Macrosolute Concentrator Protein was used to eliminate macromolecules of molecular weight below 75,000 in the RE. The retained proteins were then tested for biological activity.
Results
I. The Retinal Extract (RE) has all the Characterhtics of a Growth Factor In a previous work [l],we showed that RE could stimulate cell proliferation (recorded by 3H thymi-
Fig. 1. Effect of increasing RE concentration (abscissa pg ml-l) on BEL cell proliferation in 0.5% serum. Ordinate: cell number per dish (Scm diameter, mean of 4 petri dishes after 10 days of treatment). The value of the control was 6 X 6 10" cells
32
D. Bamtault et al.: Eye Derived Growth Factors
One way to explain these results is that RE makes the cells more susceptible to the growth factors present in the serum. As suggested later, if one mechanism of RE action is to control the cell basement membrane deposition, this would be in accordance with the most recent data from D. Gospodarowicz [23] on the control of cell growth rate by the cells’ interaction with the basement membrane P41 The addition of RE by itself is not sufficient, however, to maintain a high proliferative rate in the complete absence of serum for long periods. In these conditions, BEL cells become highly vacuolated. In a f
low serum concentration - 0.5% - the cells can be maintained in a continuous, although low state of proliferation. This proliferation is also a function of the RE concentration. If RE is added at a concentration greater than 5-pg total protein per ml culture medium, the cell number increases significantly after 10 days of culture (Fig. 1). Labelling experiments (results not shown here) have indicated that a plateau of H3 thymidine incorporation was reached for 100 pg * ml-1. BEL cells can be maintained at a very low proliferation level by culturing them without transfer in the standard medium for long periods of up to
a
Fig. 2a and b. Effect of RE (50pl/ml-1)
b
on BEL cell proliferation. Top: The cells have been maintained for 85 days at confluency. Note the enlarged irregular shaped nuclei and the extracellular material. The medium was changed twice a week. Bottom: The same cells which received RE daily during the last 3 weeks before staining; note the dense homogeneous multilayers. No extracellular material is visible. Stained with Giemsa 1 0 0 ~
D. Barritault et al.: Eye Derived Growth Factors
3 months. In this case, they formed some multilayers but also formed a very characteristic pattern which was described previously and is illustrated (Fig. 2a). The mitotic activity in this type of culture is negligible [20]. However, if the same cells were treated for the final 3 weeks with a daily RE addition, they
33 proliferated again with the formation of regular multilayers (Fig. 2b), the organization of which we will describe elsewhere. These results clearly demonstrate that the growth-promoting activity of RE is still efficient on these long-term quiescent cultures. While this has not been fully documented, we think that the
Fig.3.Effect of RE on cloning of BEL cells. After 13 days, they were stained with Giemsa. Cells were seeded at 6 (A-B), 12 (C-D), or 26 (E-F) cells per cm2.The medium was changed every third day, and RE (50 ml-l) was added at the same time to the dishes at right. After 13 days, dishes were stained with Giemsa
34
D. Barntault et al.: Eye Derived Growth Factors
stimulation does not affect all the cells in this population, but is mostly confined to the small cells in the center of round multilayers. In another set of experiments, it was shown that RE was very efficient in promoting the cloning efficiency of the BEL cell lines. In standard medium, the efficiency of cloning the BEL cells was very low. The use of X-irradiated fibroblasts as a feeder layer
did not enhance cloning significantly. In these experiments, RE was added one day after plating and renewed for 13 days before staining. As shown (Fig. 3), the colonies developed much faster in the presence of RE. In the treated colonies, cells remained smaller than in the control, and their distribution in the colony was much different, with sharper edges. These results clearly demonstrate that
Table 2. Effects of RE on the proliferation of various cell types. Comparison with EGF and FGF Cell type
Species
REa
FGP
EGP
Epithelial lens Epithelial lens Epithelial lens Epithelial lens Epithelial lens 6. Epithelial lens 7. Epidermal 8. Epidermal 9. Aortic endothelial 10. Heart endothelial 11. Corneal endothelial 12. Myoblasts 13. Myoblasts 14. BH3LL(smooth muscle) 15. Chondrocytes (cartilage) 16. Chondrocytes (articular) 17. Skin fibroblasts 18. Kidney fibroblasts 19. Skin fibroblasts 20. Lung fibroblasts 21. 3T3 Swiss 22. Neuroblastoma
Bovine Calf Chick normal Chick strain Hyl Chick strain Hy2 Human Human adult Human newborn Bovine Bovine (foetus) Bovine Bovine (foetus) Chick (embryo) Rat Rabbit Rabbit Human Bovine Bovine Bovine Mouse Mouse
+++ +++
+++b
-c
1. 2. 3. 4. 5.
++ ++ +++ +++ +++ +++ +++ +++ +++ +++ ++ + -
++
+++b
+++b
+++ + +b
+b
++ ++ ++ +
+++
++
++
+++ ++
RE is used at 50pg of total proteins per ml of culture medium Experiments also performed in our laboratory with brain FGF usually in parallel with RE. Results previously reported by others (26) c, The same as b, but with EGF All the data with FGF and EGF are from [24]. For FGF and EGF the stimulation is obtained with pg or fg ml-' (+ +), or ng ml-* (+ +), no effect (-). Blanks correspond to non-tested. For RE the stimulation is measured by an increase in cell number or compared with the response obtained in parallel experiments with FGF or EGF and non-treated cells. Cells usually counted after 8 days of stimulation +++ mean of increase by a factor of 4 or more in cell number ++ an increase by 2-4 a weak increase - absence of stimulation
+
+
1. Lens epithelial cells as described here and in 7 2. Cells donated by H. Bloemendal (Nijmegen, Holland) 3., 4., 5. Unpublished experiments with R. Clayton (Edinburgh) 6. J. Tassin: this laboratory. Despite the lack of stimulation in long-term experiments, RE induced a morphological change 7., 8. In collaboration with I. Guedon and M. Prunieras (Pans) 9 . , 10. Experiments performed in D. Gospodarowicz's laboratory (San Francisco) 11. C. Arruti (Montevideo and this laboratory) 12., 13. In collaboration with M. Fizman, Paris (see also Fig. 5) 14. Cell line gift from J. P. Mauger, Pans. The real nature of this cell line derived from brain is unknown 15., 16. Articular and cartilage rabbit chondrocytes in collaboration with M. T. Corvol, Pans 17., 21. The behaviour of RE towards fibroblasts is different from other cells. In several cases after 6 days or more, the addition of the growth factor was toxic to the culture (see also Fig. 6) 22. In collaboration with Dr. Legault, Paris (see also Fig. 4)
D. Barritault et al.: Eye Derived Growth Factors
RE is a cloning factor and can be used to clone single cells. This finding provides a very useful assay for further experiments. LI. Are There Specific Target Cells for RE Action?
To study the mechanism of RE action as well as its physiological significance, we asked the following question: Does RE have the ability to stimulate the
Fig. 4. Effect of RE on mice neuroblastorna cells (NIE-115).The cells were cultured in 0.2% serum for 9 days. Top: Note the appearance of long processes in control culture. In these conditions, the cells stop dividing and differentiate. Bottom: In the presence of RE, the cells have piled up to form large clumps. Few neuronal outgrowths can be seen. Giemsa staining 80x
35 proliferation of other ocular or non-ocular cells from other species and origins? Different types of cells from various species in primary cultures or in established cell lines were cultured in the presence or in the absence of RE. The effect of RE was measured by counting exponentially growing cells or by measuring H3 thymidine incorporation after 4 h of labelling following 44 h of treatment. In Table 2, we present the results summarizing some of the data obtained in our laboratory and in collaboration with other laboratories. Since we found that the RE had a
36
D. Barritault et al.: Eye Derived Growth Factors
Fig. 5. Effect of RE on calf myoblast cultured in 0.5% serum for 44 h. Len:Note the myotubes formed in control culture. Right: Cell density in =-treated cultures (50 kg d - l ) is higher than in controls. Myotubes formation is delayed. Giemsa staining
broad spectrum of activity, we present the results obtained by others with two comparable growth factors: EGF and FGF [25]. The resuls concerning the corneal endothelial stimulation have been already partially reported [17] and will be published in detail elsewhere. It is clear from these data that RE has an activity similar to both EGF or FGF, while the differences observed here may allow us to think that the active fraction from RE is different from both. For example, both RE and EFG stimulate adult human epidermal cells [15],while FGF has no effect: conversely, FGF stimulatesfibroblasts, while RE does not (see below). The difference between these factors is further elucidated when the chemical characterization of RE is completed. We present here an illustration of the RE action on two stimulated cell types (Fig. 4). In low serum condition, mouse neuroblastoma cells diff erentiate to form characteristic dendrites [26]. Upon RE addition,these cells, maintain a high proliferative rate.
However, no difference is observed at high serum concentrations. Similarly, in Fig. 5 we illustrate the stimulationof proliferation of bovine foetus myoblasts in low serum. The stimulation delays, but does not prevent myotube formation or the appearance of markers of differentiation (M. Fizman, unpublished results). Thus, even if RE has a broad spectrum of action, like that of a mitogenic agent crossing over the barrier of species, it does have some specificity towards fibroblast cells. It seems to stimulate slightly their proliferation for one or two divisions. Then, depending on the RE concentration, there is a loss of cells. In several instances, we have observed a general toxicity of RE to fibroblasts at a concentration highly active in BEL cells. This phenomenon may also be correlated with a loss of adhesiveness since the cells elongate considerably upon 3 days of RE treatment (Fig. 6). Further investigations are in progress to elucidate this point.
37
D . Barritault et al.: Eye Derived Growth Factors
Fig. 6. Effect of RE on fibroblast morphology. Top: Lung bovine fibroblast cultures in 10% foetal calf serum. Bottom: The same cells after 3day treatment with RE (50pg . ml-l). Giemsa staining 2 0 0 ~
IZI. R E is Different from FGF and Retinoic Acid Derivatives The similarity of action of FGF and RE on some cells is very striking. An experiment was performed by Jacque [22] to detect if there was any relationship between the basic protein and RE. The results obtained by a very sensitive radioimmunassay designed to titrate the basic protein were negative for RE and for extracts from other ocular components such as iris, pigmented epithelium with choroid, and vitreous body. In these experiments the tissue
extracts were at a final protein concentration of 5 mg ml-l, and the threshold of detection of the basic proteinwasat600pg m1-l. Controlexperimentsusing braincrudeextractsorpurifiedbrainFGFshowa highly significant activity. For example, in a solution of 2 pg ml-' of purified brain FGF, 425 ng of material cross-reacted with antihuman myelin basic protein (21%). Interestingly, no cross-reactionwas found with pituitary FGF in similar condition. Iso-electrophoretic separation of the RE extract was performed and each fraction was tested for its ability to stimulate 3H Tdr thymidine incorporation in
-
-
-
38
D. Barritault et al.: Eye Derived Growth Factors
BEL cells (see Methods). For each point different amounts of eluted proteins were used to estimate the minimum dose of protein to add to obtain the maximum stimulation. Figure 7 shows the profile of the pH gradient and the TCA precipitable radioactivity over the background incorporated in BEL cells per pg of proteins for each fraction of the gel separation. Almost all of the mitogenic activity is found in the region correspondingto pH 4.5-5. Since a large amount of material was separated and since most of the proteins in this extract had an iso-electrical point between pH 4 and 6, a perturbation in the
1
0000
Fig. 7. Iso-electrofocusing separation pattern of RE tested for mitogenic activity on BEL cells (I-P Tdr thymidine incorporation). Abscissa: fractions numbers obtained by cutting out the gel slab into 30 pieces. Ordinate: (right) pH; (left) radioactivity above background in cpm per Fg protein added per ml culture medium. x x x Value obtained with the starting sample before the iso-electrofocusing. pH values. A ---A 3H Tdr thymidine incorporation
linearity of the pH gradient was generated, creating a loss of resolution. However, no significant activity could be detected in the region where FGF from brain or pituitary migrate (pH 9.3) [lo]. Similar results (not shown) were obtained with iso-electrofocusing of vitreous body extract. An estimation of the molecular weight of the RE growth factor was performed by centrifugation on Amicon centriflow cone in different conditions (Table 3). In these experiments, the same total amount of protein was added to the culture. No apparent loss of activity was found in the fractions of molecular weights greater than 25,000 or 50,000 daltons, although a small activity was found in fractions of small molecular weight. The recovery of activity is total in the high molecular weight fractions; this is probably because the dose of 50-pg RE used is slightly above the minimal dose for the maximum stimulation, and a drop of activity by 10%-20% would not be detected. Therefore more precise dose-response studies are in progress. We note that 8-M urea treatment has no effect on the activity of the RE. Thatmostoftheactivityrecoveredisofamolecular weight above 50,000 daltons suggests it is unlikely that the mitogenic activity found in the RE is present in a smallmolecularweight protein associatedwith a higher molecular weight protein by ionic bounds, since the samples were centrifugedin 8-Murea and the fractions dialyzed back against PBS for biological activity tests. The eye contains a large amount of vitamin A and its derivatives known for decades to participate in the mechanism of vision. Vitamin A and retinoic acid are also known to be involved in the control and differentiation of many tissues of ectodermal origin
Table 3. All the RE fractions were tested for biological activity at a dose of 50 pg ml-' of culture medium as described in Methods. Measures were done in triplicate, and each experiment was standardized to give 100% for the non-fractionated RE. The increase in counts induced at the maximum of stimulation after the addition of RE was 4- 10 x above the control (PBS)according to the experiments. Since the exact minimal dose of RE for the 100% stimulation could not be determined precisely, the values given for the 100% recoveries indicate that no significant loss of activity was found at the dose used. The molecular weight cut-off was obtained with Amicon membrane cones, Centriflow CF25 and CF50 (-+ 25,000 and -+ 50,000 Daltons) and with the A 75 Amicon macrosolute concentrator protein Untreated RE
Treated RE Percentage of stimulation
PBS
0%
RE RE RE RE RE RE
100% 100% 5% 100% 10% 0%
> 25,000 < 25,000 > 50,000 < 50,000 > 75,000
Treatment and molecular weight separations in presence of 8 M urea
PBS RE (PBS) 8M urea RE RE > 50,000 RE < 50,000 RE > 25,000 RE < 25,000
0YO 100% 100% 100% 20% 100% 5%
39
D. Bamtault et al.: Eye Derived Growth Factors
[26,27]. It was thus of interest to assess the effects of retinoic acid on BEL cells in culture and to compare its action with RE. For instance, several reports have pointed out that retinoic acid can change the morphology of cells in vitro [28]. In the presence of retinoic acid (2 x 10-6-2 X 10-8M) the BEL cell morphology is not affected after short (2 days) or long treatments (25 days). However, there seems to be a slight increase in cell population with 2.10-8 M retinoic acid, but it is much smaller than if the cells had received 50 pg - ml-' RE. At higher concentraM, retinoic acid slightly inhibits cell tion than proliferation. Retinoic acid, M, added at the same time as RE, does not prevent cell elongation and stimulation. At a higher concentration such as 2.10-6 M, it can also decrease RE growth stimulation (Fig. 8).
Thus, retinoic acid and RE clearly have distinct actions on BEL cells, and some of them are antagonistic. A similar effect has been described following long treatment with cortisol (Arruti, unpublished work). Cortisol does not increase the mitogenic effect of RE on BEL cells, in contrast to its role on 3T3 fibroblast stimulated by FGF (results not shown). UV irradiation at a very high dose (10 jm-2) of RE for inactivating retinoic acid or its derivatives had no effect upon RE mitogenic activity of RE on BEL cells.
N.RE-like Activity can be Found in Other Ocular Tissuesfrom Bovine or Other Species Using the same technique for preparing neutral buffered extracts from ocular tissues as used for RE,
Fig. 8. Effect of RE and retinoic acid RA on BEL cell proliferation. This multilayer dish illustrates the general aspect of multilayers of BEL cells obtained after 26 days in vitro in standard condition (Dl-D2-D3), or RE (50 pg ml-l)treatment (Cl-C2-C3) as well as in retinoic acid administered alone at several concentrations (2.10-6 M = D4-D5-D6; 2.lO-'M = Al-A2-A3; 2.10-' M = Bl-B2-B3) or in the presence of RE. RE 50 pg ml-1 + RA 2.10-6 M = A4-A5-A6; RE 50 pg ml-' FL4 2.lO-' M = B4-B5-B6; RE 50 pg ml-' RA 2.10-' M = C4-C5-C6
+
+
40
D . Barritault et al.: Eye Derived Growth Factors
Table 4. Effects of different ocular tissue extracts on H3 thymidme incorporation and cell number in BEL cultures
Tissue of origin PBS alone Retina Iris Choroid Vitreous body Aqueous humour Crystalline cortex
cpm H3 thymidine
pg of total protein per ml culture medium
Number of cefls/cm2
(4 h)
0 100 100 100 100 50 100
56.10' 112.10' l00.ld 110.103 90.103 S0.ld 55.103
10,000 170,000 160,000 130,000 140,000 12,000 10,000
Retina, iris, and choroid were crude extracts in PBS prepared after 100,oOog centrifugation and dialysis. Vitreous body was first concentrated by ammonium sulfate (AS) precipitation; the fraction used was the supernatant of 20% AS (precipitated by 60% AS). Aqueous humour was concentratedusing a B 15 Amicon macrosolute concentration (mol. wt. cut-off above 15,OOO Daltons). Undialyzed samples had a similar stimulatory effect These results were obtained from experimentsin which BEL cells were maintained at confluency stimulated for 2 days and labelled for 4 h. In each experiment, one control was done by adding PBS alone. Since the extracts are not purified, a dose-response curve has to be determined for every sample. Only the maximum stimulation is shown here. Similar results have been obtained with extracts from human, calf. and horse retina
we examined their action on BEL cell proliferation and cell elongation. The incorporation of H3 thymidine in these cells after 2 days of treatment indicated that extracts from iris, pigmented epithelium, choroid, and the vitreous body had the same activity, while extracts from lens cortex or aqueous humour had no stimulatory effect (Table 4). These extracts were tested on other cells and also gave the same results, except for the vitreous extract which elicited a weaker elongation response. While until they are purified we have no evidence that these factors are similar, that they are prepared by the same technique, that they elicit the same response on target cells, and that mitogenic activities for RE and VE are found in the same PI range, all argue strongly in favour of one factor (or one family) common to these different tissues. A similar growth factor can be obtained also from human, horse, or embryonic calf retina (results not shown). We decided to call them EDGF (eye-derived growth factors) by analogy with PDGF (platelet-derived growth factors).
Discussion Is there an ubiquitous growth factor in the eye? The data presented in this work clearly demonstrate that such a growth factor (or several growth factors) is present in the adult eye, and is easily accessible for extraction. Since extracts from the different sources have not yet been purified to homogeneity, we cannot exclude that the growth stimulation is due to more
than one molecule. In addition, there may be complicated interactions between them and the serum proteins. However, the few biochemical properties already known (molecular weight, iso-electric point, and stability to denaturation) seem to indicate that they are similar or belong to the same family. EDGF has in vitro a very potent activity on various cell types stimulating their proliferation and maintaining some of their differentiated properties. That it is an endogenous factor leads us to explore its role in tissue regeneration and development as in the case of other well characterized growth factors like EGF and FGF. In contrast to these factors whose physiological target cells have not been fully characterized, RE or EDGF may act to maintain the integrity and differentiation of topographically and even ontogenically related tissues such as the corneal endothelium (which it does so precisely in vitro) or the lens epithelium. It could be stored in an inactivated form in normal tissue or circulate at a low level to maintain normal growth. If released in greater amounts under pathological conditions, it could be responsible for several disorders such as vascular endothelium proliferation. Our results showed, for instance (Table 2), that RE can induce proliferation of the vascular endothelial cells in vitro. Recent experiments using the growth of capillaries induced by pieces of agar dipped in RE and deposited on egg chorioallantoic membrane, a technique described recently for other growth factors [29], have shown that it may be also angiogenic (unpublished results). If these experiments are confirmed by other techniques, they will point to RE as an important
41
D. Bamtault et al.: Eye Derived Growth Factors
potential factor in the etiology of the main diseases of the retina where hypervascular proliferation occurs. Thepresenceofmitogenicandneurotrophicfactors in the retina have already been proposed to explainthe control of lens regeneration in the newt [30]. Recently the presence in the vitreous of a factor which promotes lens fiber cell differentiation was described [31]. Many of its chemical properties are shared with EDGF from adult bovine eye. It would be spectacular if these two factorswere related, sincetheywoulddemonstratethat the same material regulates cell proliferation and cell differentiation. The preceding paper has demonstrated that BEL cell morphology could be modulated by RE. From the results on cytoskeleton organization and extracellular deposition of fibronectin, it was proposed that these modifications were brought about by the specific response of these cells to a growth stimulus. We proposed that cells with similar architecture would respond in a similar way to any factor which has the growth-promoting capacity. This hypothesis that we have confirmed in studying the effect of RE on bovine corneal endothelial cells [32] was also recently illustrated by Gospodarowicz on the same cells and other cells using FGF and EGF [23]. Cells which do not respond to RE by an elongation such as myoblast as illustrated in Fig. 5 may have a different cytoskeleton organization. Conversely, the striking elongation observed in fibroblasts which are already spindle-shaped could be a deterrent to their growth. Thus the possibility of response of a given cell to a growth factor would be dictated not by its embryonic lineage but by its intracellular organization. the other hypothesis which does not exclude the first one could be that RE directly controls the deposition of the extracellular material which in turn controls the susceptibility of these cells to serum growth factors. Such a hypothesis was recently proposed by Gospodarowicz and illustrated in several cell types [24]. These results outlined the general interest of studying these growth factors in the understanding of the control of cell division and cell differentiation.
Acknowledgements. We want to express our gratitude to J. Tassin and M. Olivit5 for their excellent technical assistance during this work, to D. Gospodarowia, R. Clayton, M. T. Corvol, I. Guedon, L. Legaut, M. Fizman for their expertise in growing cells, and to C. Jacque for the immunological titration of FGF. This work was supported by a grant from the INSERM (ATP 79-114) and was presented at the NEI Ocular Tissue Culture Symposium in Bethesda, Md., October 22-23, 1979.
References 1. Arruti C, Courtois Y (1978) Morphological changes and growth stimulation of bovine epithelial lens cells by a retinal extract in vitro. Exp Cell Res 117: 283 2. Coulombre JL, Coulombre AJ (1963) Lens development. Fiber elongation and lens orientation. Science 142: 1489 3. Jacobson AG (1966) Inductive processes in embryonic development. Science 152: 25 4. Karkinen-Jaaskelainen M (1978) Transfilter lens induction in avian embryo. Differentiation 12: 31 5. Nordman J (1977) Au sujet du vieillissement du cristallin humain et de la cataracte dnile. Adv Ophthalmol 34: 1 6. Reddan JR, Weinsieder A, Wilson D (1979) Aqueous humor from traumatized eyes triggers cell division in the epithelial of cultured lenses. Exp Eye Res 28: 267 7. Laurent MV, Lonchampt MO, Regnault F, Tassin J, Courtois Y (1978) Biochemical, ultrastructural and immunological study of in vitro production of collagen by bovine lens epithelia cells in culture. Exp Cell Res 115: 127 8. HughesRC,MillsG,CourtoisY(1979)Roleoffibronectininthe adhesivenessof bovine epithelial cells. Biol Cell 36: 321 9. Bradford MM (1976) A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal Biochem 72: 248 10. GospodarowiaD,BialeckiH,GreenburgG (1978)Purification of the fibroblast growth factor activity from bovine brain. J Biol Chem 252 :3736 11. HughesRC,LaurentMV,LongchamptMO,CourtoisY (1975) Biosynthesis in cultured epithelial cells of bovine lens. Eur J Biochem 52: 143 12. TassinJ, Malaise E, CourtoisY (1979) Human lens cells have an in vitroproliferative capacityinverselyproportional to the donor age. Exp Cell Res 123:388 13. Buckingham ME, Caput D, Cohen A, Whalen RG, Gros F (1974) The synthesis and stability of cytoplasmic messenger RNAduMgmyoblast differentiationin culture. ProcNatl Acad Sci USA 71 : 1466 14. Koenigsberg IR (1963) Clonal analysis of myogenesis. Science 140: 1273 15. Guedon I, Bamtault D, Courtois Y,Prunieras M (Differentiation in press) Karyotypes and sister chromatid exchanges in normal human adult keratinocytes as a routine test 16. Gospodarowicz D, Moran JS, Braun DL (1977) Control of proliferation of bovine vascular endothelial cells. J Cell Physiol 91 : 377 17. Arruti C, Courtois Y (1979) Retinotrophic stimulation of proliferation of bovine corneal endothelial cells in vitro. Proceedingsfor 20th Meeting of Association for Eye Research. Paris, 25-28 September 18. Corvol MT, Dumontier MF, Rappaport R (1975) Culture of chondrocytes from the proliferative zone of epiphyseal growth plate cartilage from prepubertal rabbits. Biomedicine 23: 103 19. Legault Demare L, Zeitour Y, Lando D, Lamande N, Grasso A, Gros F (1980) Expression of a specific neuronal protein 143-2 during in vitro differentiation of neuroblastoma cells. Exp Cell Res 125:233 20. HervC B, Jacquemin E,Courtois Y (1979) Histones biosynthesis and turnover in epithelial lens cultured in vitro. Cell Biol Int Rep 3: 271 21. Westall F, Lennon V,GospodarowiczD (1978) Brain-derived fibroblast growth factor: identity with a fragment of the basic protein of myelin. Proc Natl Acad Sci USA 75: 4675
42 22. Delassale A, Jacque C, Drouet J, Raoul M, Legrand JC (1980) Radioimmunoassay of the myelin basic protein in biological fluids: conditions improving sensitivity and specificity. Biochimie 62: 159 23. GospodarowiczD, Greenburg G, Birdwell CR (1978) Determination of cellular shape by the extracellular matrix and its correlation with the control of cellular growth. Cancer Res 38: 4155 24. Vlodavsky I, Lui GM, GospodarowiczD (1980) Morphological appearance, growth behaviour and migratory activity of human tumor cells maintained on extracellular matrix versus plastic. Cell 19: 607 25. Gospodarowicz D, Greenburg G, Bialecki H, Zetter BR (1978) Factors involved in the modulation of cell proliferation in vivo and in vitro: The role of fibroblast and epidermal growth factors in the proliferative response of mammalian cells. In Vitro 14: 85 26. Seeds NW, Gilman AG, Amano T, Niremberg MW (1970) Regulation of Axon formation by clonal lines of a neural tumor. Proc Natl Acad Sci USA 66: 160
Note Added in Proofs. Recently the angiogenic nature of EDGF was confirmed by our own laboratory using a perfusion system in the rabbit cornea (Stimulation of neovascularisation of the rabbit cornea by EDGF. P. Thompson, D. Maurice, D. Bamtault, J. Plouet, Y.Courtois, in press) and by two other laboratories (Chen CH and Chen SC 1980 Invest. Ophtalm. 19 5% and Glaser B, d'Amore P, Michels RG, Patz A. and Feselan A (1980) J Cell Biol 84: 298) using different technics.
D. Barritault et al.: Eye Derived Growth Factors 27. Saari JC, Futterman S, Bredberg L (1978) Retinal binding protein in the retina. J Biol Chem 253: 6432 28. Patt LM, Itaya K,Hakomori S (1978) Retinal induces density dependent growth inhibition and changes in glycolipids and lets. Nature 273: 379 29. Folkman J, Merler E, Abernathy C, Willums G (1971) Isolation of a tumor factor responsible for angiogenesis. J Exp Med 133: 275 30. Yamada T (1977) Control mechanisms in cell type conversion in newt lens regeneration. In: Wolsky A (ed) Monographs in developmental biology, vol. 13. S. Karger, Base1 31. Beebe D, Feagans DE, Jebens HA (1980) Lentropin a factor in vitreous humour which promotes lens fiber cell differentiation. Proc Natl Acad Sci USA 77: 490 32. Arruti C, Courtois Y (1980) The in vitro control of proliferation and cell interaction of bovine corneal endothelial cells by a retinal growth factor. Eur J Cell Biol 22: 386 Received March 1980lAccepted July 1980
Differentiation 18, 29-42 (1981)
0 Springer-Verlag 1981
Is There a Ubiquitous Growth Factor in the Eye? Proliferation Induced in Different Cell Types by Eye-Derived Growth Factors D. BARRITAULT, C. ARRUTI, and Y.COURTOIS Unit6 de Recherches GCrontologiques, INSERM U. 118, ERA CNRS 842, 29, rue Wilhelm, 75016 Pans, France
In a previous work [l]we showed that a neutral extract of bovine adult retina RE can stimulate the growth and modify the morphology of bovine epithelial lens (BEL) cells in vitro. We were also able to demonstrate that the differences in cell shape are closely related to the cell growth properties induced by RE and are mediated by cytoskeletal protein organization as well as external proteins. In this study, we report the results of further investigations on this retinal extract. We show that it possesses all the characteristics of other growth factors such as promoting proliferation in low serum concentration or of enhancing the colony-forming efficiency of BEL cells considerably. By comparing the morphological response of BEL cells treated with RE with the response of other cells to other growth factors, we propose that the phenotypic modifications are cell specific, but not growth factor specific. We report-also that RE has a broad spectrum of activity since it is able to stimulate cells from different origins and species (vascular and corneal endothelial cells, myoblasts, chondrocytes, neuroblastoma cells, and keratinocytes), but not all of them, since it can be toxic for fibroblasts. In this respect, it has an activity similar in many aspects to FGF and EGF, while it differs from them for some target cells. Its action has also been compared with the effects of retinoic acid derivatives and shown to be strikingly different. RE-like activity can be found in other ocular tissues from bovine and other species. The .highest growth-promoting capacities were found in extracts of iris, pigmented epithelium with choroid, and vitreous body. The nature of all these extracts has not yet been determined. Since they are prepared in a similar way and since they have similar growth-promoting activity, we postulate that there is an ubiquitous growth factor in the eye called eye-derived growth factor (EDGF) which may play an important role in physiology and pathology of the eye.
Introduction The use of an in vitro system to study growth interactions between tissues has several advantages insofar as most parameters can be controlled. For instance, it is possible to control all essential nutrients, except there is still a requirement for a certain amount of serum until a synthetic medium is developed. It is therefore possible to look at the interaction between two tissues by techniques such as co-culture or the use of tissue extracts. Because of the progress in cell-culture technology one can hope to understand better the relationships
between two tissues whose maintenance is otherwise vital, and thus mostly inaccessible to in vivo experiments. Thus, the discovery of growth factors is closely linked to the in vitro technology. Conversely, it is possible that growth in these artificial conditions modifies the behaviour of cells. These general considerations were in mind when we decided to investigate the presence of putative factors which might control the growth of the lens. The stimulus to look for such factors in the eye came from many experimental data concerning both normal or pathological development and the control of tissue reconstitution after injury. On the one hand, 0301-4681/81/0018/0029/$ 02.80
30 many investigators have pointed out that the retina must produce some diffusible factors which control proliferation and differentiation of lens epithelial cells during development [2-41. On the other hand, the lens increases continuously in size during the entire lifespan of the individual [5] and this growth corresponds to the differentiation of a pool of epithelial cells very precisely located at the periphery of the lens. Paracentesis of the aqueous humour [6] has demonstrated the presence of a factor which stimulates cell division in a lens maintained in vitro in the appropriate medium. Bovine lens epithelial cells (BEL) are easy to culture, and our laboratory has investigated the growth and synthetic properties of these cells in standard conditions, i.e., in 6% foetal calf serum. These cells transformed spontaneously, but nevertheless synthesized lens capsule material [7].We thus decided to investigate the effect of a retinal extract (RE) on these cells and found that this extract was able to stimulate the growth of these cells and to modulate their morphology [l]. We have studied separately how these phenotypic modifications correlated with the organization of the cytoskeletal proteins and with the extracellular deposition of fibronectin [8]. In this work, we will report further investigations on the retinal extract. We will show that RE fulfils the criteria of other potent growth factors. Several different cell types are stimulated by this extract, and their response is of a similar nature to that observed with other growth factors such as epidermal growth factor (EGF) and fibroblast growth factor (FGF). Preliminary results show that RE tends to differ from both of them and from retinoic acid. We will report that similar RE-like activities are found in ocular tissues other than the retina; we suggest that these activities could be involved in the interactions between all these tissues.
Methods Tissue Extracts Adult bovine eyes from freshly slaughtered animals were used. Retinal extract was prepared as previously described [l].(See also accompanying article.) Extracts from tissues such as iris, choroid, pigmented epithelium, cristallin, vitreous body, and aqueous humour were prepared as for the retina, except that the tissues were first homogenized using a Turax blendor. For vitreous body extract, a concentration was made by ammonium sulfate (AS) precipitation and the fraction used for centrifugation was the
D. Bamtault et al.: Eye Derived Growth Factors supernatant of 20% AS precipitated by 60% AS. Through all these extractions, a phosphate saline buffer (PBS) was used and the protein concentration was estimated according to Bradfort [9] and adjusted to 5 mg/ml after dialysis. The extracts were sterilized by filtration through 0.22 pm millipore filters, aliquoted, and kept at -20" C. A crude extract from bovine brain was made at pH 4-5, followed by ammonium sulfate precipitation like in the preparation of FGF [lo].
Other Growth Factors and Chemicak Purified pituitary and brain FGF were generously given by D. Gospodarowicz (San Francisco). Purified EGF was generously given by G. Orth (Paris). H q d r thymidine (25 C. mM-') came from CEA, France, and retinoic acid from Sigma.
Cell Culture Calf and adult bovine epithelial lens (BEL) cells were cultured as in [ll]; the stimulation experiments were performed on serially subcultured cells (3-30 passages). Human lens cells were cultured as described by Tassin et al. [12]. Foetal bovine and chick embryo myoblasts were cultured by M. Fizman (Paris) [13, 141. Primary cultures of human adult epidermal cells were done as by Guedon [15]. Aortic and foetal heart bovine endothelial cells were cultured by D. Gospodarowia [16]. Corneal endothelial cells were grown from adult bovine cornea. The same culture media was used as with BEL cells, but after addition of 10% foetal calf serum. Since RE is a potent growth factor for these cells (171, its presence allowed serial subculture without a sign of in vitro senescence. RE was constantly present in our culture and removed 3 days before stimulation experiments. Primary cultures of rabbit ear and knee chondrocytes were grown by M. T. Corvol as in [IS]. Human Foreskin fibroblasts, Swiss mouse 3T3, bovine from skin, lung, and kidney fibroblasts, as well as a rat smooth muscle cell line (gift from J. P. Mauger) were all cultured in MEM medium supplemented with 10% foetal calf serum and subcultured as in [ll]. Mouse NIE115 adrenergic neuroblastoma cells were grown by L. Legault-Demare as in [19]; experiments for RE stimulation were done after decreasing the foetal calf serum concentration to 0.2%.
Measure of DNA Synthesis and Cell Multiplication Unless otherwise specified, H?dr thymidine incorporation was measured in triplicate experiments on BEL cells cultured in Linbro trays (24 wells, 2cm2 per well) treated or non-treated with the different tissue extracts at different doses for 44 h and labelled for 4 h using 2pCilml of culture. Cells were washed three times with PBS buffer lysed in 0.5 ml of 0.1 N NaOH, 0.1% SDS for 5 min and cold TCA was added to a fiial concentration of 20%. The TCA precipitate was filtered in a Whatman GFlC glass filter, washed with 5 vol. through 5% TCA, air dried, and the radioactivity was counted in aliquid scintillation counter. . Parallel experiments were done without addition of 3H Tdr thymidine, and the cells were dissociated by a trypsin-EDTA solution (0.2g EDTA, 0.5 g trypsin, 1g glucose, 0.58 sodium bicarbonate, 0.4 g KCl, and 8 g Na€l per liter) and counted in a hemacytometer or a Coultertronic cell counter.
31
D. Barritault et al.: Eye Derived Growth Factors For all the other cells studied, petri dishes of 3.5 cm or 5 cm diameter were used. After different time of cultures in the presence or absence of tissue extracts, cells were trypsinized and counted in triplicate. The increase in cell number of myoblasts was estimated after 44 h or 5 days'of culture after lysing the cells by an overnight incubation in 0.1 N citric acid and counting the nuclei [20].This allowed a measure of the total number of cells, including the cells forming the myotubes. (More details about the formation of myotubes, the dosage of specific receptors and the percentage of fusion after RE treatment will be published elsewhere.) Radio-immunoassay Recently brain FGF was shown to be part of the basic protein of m y e h (211.To detect if there was any FGF activity in the different tissues extracts, a radio-immunoassay designed by J a q u e [22]was used. In this experiment, a serum against human basic protein was prepared by J a q u e and was shown to cross-react with bovine FGF. Iso-electrofocusing of Retinal Extract Retinal extract prepared as above was further purified by precipitation of a 20% ammonium sulfate supernatant with 60% ammonium sulfate. The pellet was redissolved in PBS and dialyzed against 1% glycine, and 250 mg of the protein solution in 15 ml was separated on a LKB iso-electrofocusingpreparation slab apparatus using Serva Ampholines pH 2-11, 2% w/v, as described in application note 198 supplied with the LKB equipment. After separation, the gel slab was cut into 30 bands, pH was measured, and the protein content was eluted from each of the bands by transferring the granulated gel into a small column (stopped at one end with a nylon net) and washing this column with PBS buffer. Ampholines were removed by dialysis in the same buffer and protein content was measured as in [ll]. Each fraction was sterilized by filtration and tested for stimulation of thymidine incorporation on BEL cells.
dine incorporation) in 6% foetal calf serum. We show here that it can also stimulate cell proliferation in complete absence of serum. This is illustrated in Table 1. Cells in monolayers were washed several times with a medium without serum. They were then cultured in the different media, described in Table 1, for 24 h. Three hours before processing they were given 1pCi ml-' of 3H thymidine. The results indicated that an increase in both cell number and thymidine incorporation was brought about by the RE and the serum. They also showed that there may be a synergism between the serum and the RE since the combination of both resulted in greater stimulation than the addition of either one separately.
-
Table 1. Effect of RE on 3Hthymidhe incorporation in the DNA of BEL cells in the absence of serum Medium MEM MEM RE (50 pg-ml-') SM SM + RE (50 pg * I&')
+
Cell number
cpm. 10-4 cells
344
5.6 10' 7.6 10' 7.0 10'
2934 1514
9.2 10"
6460
SM: Standard Medium, i s ; MEM containing 6% vlv calf S e m MEM: Eagle's Minimum essential Medium without sem Each result is the mean of two experiments Increase in cell number was measured after 24-h exposure to the different media; (p) thymidine incorporation was measured in the last 3 h
Molecular Weight Estimation After partial purification with ammonium sulfate, RE was dialyzed against PBS or 8 M urea and centrifuged at 2,000g on Amicon membrane cones Centriflow CF 25 and CF 50 (molecular weight cut off at about 25,ooO and 50,000 daltons, respectively). Protein content of the fractions retained and nonretained by the cones was estimated, and each fraction was dialyzed back against PBS and tested for stimulation of thymidine incorporation in BEL cells. A 75 Amicon Macrosolute Concentrator Protein was used to eliminate macromolecules of molecular weight below 75,000 in the RE. The retained proteins were then tested for biological activity.
Results
I. The Retinal Extract (RE) has all the Characterhtics of a Growth Factor In a previous work [l],we showed that RE could stimulate cell proliferation (recorded by 3H thymi-
Fig. 1. Effect of increasing RE concentration (abscissa pg ml-l) on BEL cell proliferation in 0.5% serum. Ordinate: cell number per dish (Scm diameter, mean of 4 petri dishes after 10 days of treatment). The value of the control was 6 X 6 10" cells
32
D. Bamtault et al.: Eye Derived Growth Factors
One way to explain these results is that RE makes the cells more susceptible to the growth factors present in the serum. As suggested later, if one mechanism of RE action is to control the cell basement membrane deposition, this would be in accordance with the most recent data from D. Gospodarowicz [23] on the control of cell growth rate by the cells’ interaction with the basement membrane P41 The addition of RE by itself is not sufficient, however, to maintain a high proliferative rate in the complete absence of serum for long periods. In these conditions, BEL cells become highly vacuolated. In a f
low serum concentration - 0.5% - the cells can be maintained in a continuous, although low state of proliferation. This proliferation is also a function of the RE concentration. If RE is added at a concentration greater than 5-pg total protein per ml culture medium, the cell number increases significantly after 10 days of culture (Fig. 1). Labelling experiments (results not shown here) have indicated that a plateau of H3 thymidine incorporation was reached for 100 pg * ml-1. BEL cells can be maintained at a very low proliferation level by culturing them without transfer in the standard medium for long periods of up to
a
Fig. 2a and b. Effect of RE (50pl/ml-1)
b
on BEL cell proliferation. Top: The cells have been maintained for 85 days at confluency. Note the enlarged irregular shaped nuclei and the extracellular material. The medium was changed twice a week. Bottom: The same cells which received RE daily during the last 3 weeks before staining; note the dense homogeneous multilayers. No extracellular material is visible. Stained with Giemsa 1 0 0 ~
D. Barritault et al.: Eye Derived Growth Factors
3 months. In this case, they formed some multilayers but also formed a very characteristic pattern which was described previously and is illustrated (Fig. 2a). The mitotic activity in this type of culture is negligible [20]. However, if the same cells were treated for the final 3 weeks with a daily RE addition, they
33 proliferated again with the formation of regular multilayers (Fig. 2b), the organization of which we will describe elsewhere. These results clearly demonstrate that the growth-promoting activity of RE is still efficient on these long-term quiescent cultures. While this has not been fully documented, we think that the
Fig.3.Effect of RE on cloning of BEL cells. After 13 days, they were stained with Giemsa. Cells were seeded at 6 (A-B), 12 (C-D), or 26 (E-F) cells per cm2.The medium was changed every third day, and RE (50 ml-l) was added at the same time to the dishes at right. After 13 days, dishes were stained with Giemsa
34
D. Barntault et al.: Eye Derived Growth Factors
stimulation does not affect all the cells in this population, but is mostly confined to the small cells in the center of round multilayers. In another set of experiments, it was shown that RE was very efficient in promoting the cloning efficiency of the BEL cell lines. In standard medium, the efficiency of cloning the BEL cells was very low. The use of X-irradiated fibroblasts as a feeder layer
did not enhance cloning significantly. In these experiments, RE was added one day after plating and renewed for 13 days before staining. As shown (Fig. 3), the colonies developed much faster in the presence of RE. In the treated colonies, cells remained smaller than in the control, and their distribution in the colony was much different, with sharper edges. These results clearly demonstrate that
Table 2. Effects of RE on the proliferation of various cell types. Comparison with EGF and FGF Cell type
Species
REa
FGP
EGP
Epithelial lens Epithelial lens Epithelial lens Epithelial lens Epithelial lens 6. Epithelial lens 7. Epidermal 8. Epidermal 9. Aortic endothelial 10. Heart endothelial 11. Corneal endothelial 12. Myoblasts 13. Myoblasts 14. BH3LL(smooth muscle) 15. Chondrocytes (cartilage) 16. Chondrocytes (articular) 17. Skin fibroblasts 18. Kidney fibroblasts 19. Skin fibroblasts 20. Lung fibroblasts 21. 3T3 Swiss 22. Neuroblastoma
Bovine Calf Chick normal Chick strain Hyl Chick strain Hy2 Human Human adult Human newborn Bovine Bovine (foetus) Bovine Bovine (foetus) Chick (embryo) Rat Rabbit Rabbit Human Bovine Bovine Bovine Mouse Mouse
+++ +++
+++b
-c
1. 2. 3. 4. 5.
++ ++ +++ +++ +++ +++ +++ +++ +++ +++ ++ + -
++
+++b
+++b
+++ + +b
+b
++ ++ ++ +
+++
++
++
+++ ++
RE is used at 50pg of total proteins per ml of culture medium Experiments also performed in our laboratory with brain FGF usually in parallel with RE. Results previously reported by others (26) c, The same as b, but with EGF All the data with FGF and EGF are from [24]. For FGF and EGF the stimulation is obtained with pg or fg ml-' (+ +), or ng ml-* (+ +), no effect (-). Blanks correspond to non-tested. For RE the stimulation is measured by an increase in cell number or compared with the response obtained in parallel experiments with FGF or EGF and non-treated cells. Cells usually counted after 8 days of stimulation +++ mean of increase by a factor of 4 or more in cell number ++ an increase by 2-4 a weak increase - absence of stimulation
+
+
1. Lens epithelial cells as described here and in 7 2. Cells donated by H. Bloemendal (Nijmegen, Holland) 3., 4., 5. Unpublished experiments with R. Clayton (Edinburgh) 6. J. Tassin: this laboratory. Despite the lack of stimulation in long-term experiments, RE induced a morphological change 7., 8. In collaboration with I. Guedon and M. Prunieras (Pans) 9 . , 10. Experiments performed in D. Gospodarowicz's laboratory (San Francisco) 11. C. Arruti (Montevideo and this laboratory) 12., 13. In collaboration with M. Fizman, Paris (see also Fig. 5) 14. Cell line gift from J. P. Mauger, Pans. The real nature of this cell line derived from brain is unknown 15., 16. Articular and cartilage rabbit chondrocytes in collaboration with M. T. Corvol, Pans 17., 21. The behaviour of RE towards fibroblasts is different from other cells. In several cases after 6 days or more, the addition of the growth factor was toxic to the culture (see also Fig. 6) 22. In collaboration with Dr. Legault, Paris (see also Fig. 4)
D. Barritault et al.: Eye Derived Growth Factors
RE is a cloning factor and can be used to clone single cells. This finding provides a very useful assay for further experiments. LI. Are There Specific Target Cells for RE Action?
To study the mechanism of RE action as well as its physiological significance, we asked the following question: Does RE have the ability to stimulate the
Fig. 4. Effect of RE on mice neuroblastorna cells (NIE-115).The cells were cultured in 0.2% serum for 9 days. Top: Note the appearance of long processes in control culture. In these conditions, the cells stop dividing and differentiate. Bottom: In the presence of RE, the cells have piled up to form large clumps. Few neuronal outgrowths can be seen. Giemsa staining 80x
35 proliferation of other ocular or non-ocular cells from other species and origins? Different types of cells from various species in primary cultures or in established cell lines were cultured in the presence or in the absence of RE. The effect of RE was measured by counting exponentially growing cells or by measuring H3 thymidine incorporation after 4 h of labelling following 44 h of treatment. In Table 2, we present the results summarizing some of the data obtained in our laboratory and in collaboration with other laboratories. Since we found that the RE had a
36
D. Barritault et al.: Eye Derived Growth Factors
Fig. 5. Effect of RE on calf myoblast cultured in 0.5% serum for 44 h. Len:Note the myotubes formed in control culture. Right: Cell density in =-treated cultures (50 kg d - l ) is higher than in controls. Myotubes formation is delayed. Giemsa staining
broad spectrum of activity, we present the results obtained by others with two comparable growth factors: EGF and FGF [25]. The resuls concerning the corneal endothelial stimulation have been already partially reported [17] and will be published in detail elsewhere. It is clear from these data that RE has an activity similar to both EGF or FGF, while the differences observed here may allow us to think that the active fraction from RE is different from both. For example, both RE and EFG stimulate adult human epidermal cells [15],while FGF has no effect: conversely, FGF stimulatesfibroblasts, while RE does not (see below). The difference between these factors is further elucidated when the chemical characterization of RE is completed. We present here an illustration of the RE action on two stimulated cell types (Fig. 4). In low serum condition, mouse neuroblastoma cells diff erentiate to form characteristic dendrites [26]. Upon RE addition,these cells, maintain a high proliferative rate.
However, no difference is observed at high serum concentrations. Similarly, in Fig. 5 we illustrate the stimulationof proliferation of bovine foetus myoblasts in low serum. The stimulation delays, but does not prevent myotube formation or the appearance of markers of differentiation (M. Fizman, unpublished results). Thus, even if RE has a broad spectrum of action, like that of a mitogenic agent crossing over the barrier of species, it does have some specificity towards fibroblast cells. It seems to stimulate slightly their proliferation for one or two divisions. Then, depending on the RE concentration, there is a loss of cells. In several instances, we have observed a general toxicity of RE to fibroblasts at a concentration highly active in BEL cells. This phenomenon may also be correlated with a loss of adhesiveness since the cells elongate considerably upon 3 days of RE treatment (Fig. 6). Further investigations are in progress to elucidate this point.
37
D . Barritault et al.: Eye Derived Growth Factors
Fig. 6. Effect of RE on fibroblast morphology. Top: Lung bovine fibroblast cultures in 10% foetal calf serum. Bottom: The same cells after 3day treatment with RE (50pg . ml-l). Giemsa staining 2 0 0 ~
IZI. R E is Different from FGF and Retinoic Acid Derivatives The similarity of action of FGF and RE on some cells is very striking. An experiment was performed by Jacque [22] to detect if there was any relationship between the basic protein and RE. The results obtained by a very sensitive radioimmunassay designed to titrate the basic protein were negative for RE and for extracts from other ocular components such as iris, pigmented epithelium with choroid, and vitreous body. In these experiments the tissue
extracts were at a final protein concentration of 5 mg ml-l, and the threshold of detection of the basic proteinwasat600pg m1-l. Controlexperimentsusing braincrudeextractsorpurifiedbrainFGFshowa highly significant activity. For example, in a solution of 2 pg ml-' of purified brain FGF, 425 ng of material cross-reacted with antihuman myelin basic protein (21%). Interestingly, no cross-reactionwas found with pituitary FGF in similar condition. Iso-electrophoretic separation of the RE extract was performed and each fraction was tested for its ability to stimulate 3H Tdr thymidine incorporation in
-
-
-
38
D. Barritault et al.: Eye Derived Growth Factors
BEL cells (see Methods). For each point different amounts of eluted proteins were used to estimate the minimum dose of protein to add to obtain the maximum stimulation. Figure 7 shows the profile of the pH gradient and the TCA precipitable radioactivity over the background incorporated in BEL cells per pg of proteins for each fraction of the gel separation. Almost all of the mitogenic activity is found in the region correspondingto pH 4.5-5. Since a large amount of material was separated and since most of the proteins in this extract had an iso-electrical point between pH 4 and 6, a perturbation in the
1
0000
Fig. 7. Iso-electrofocusing separation pattern of RE tested for mitogenic activity on BEL cells (I-P Tdr thymidine incorporation). Abscissa: fractions numbers obtained by cutting out the gel slab into 30 pieces. Ordinate: (right) pH; (left) radioactivity above background in cpm per Fg protein added per ml culture medium. x x x Value obtained with the starting sample before the iso-electrofocusing. pH values. A ---A 3H Tdr thymidine incorporation
linearity of the pH gradient was generated, creating a loss of resolution. However, no significant activity could be detected in the region where FGF from brain or pituitary migrate (pH 9.3) [lo]. Similar results (not shown) were obtained with iso-electrofocusing of vitreous body extract. An estimation of the molecular weight of the RE growth factor was performed by centrifugation on Amicon centriflow cone in different conditions (Table 3). In these experiments, the same total amount of protein was added to the culture. No apparent loss of activity was found in the fractions of molecular weights greater than 25,000 or 50,000 daltons, although a small activity was found in fractions of small molecular weight. The recovery of activity is total in the high molecular weight fractions; this is probably because the dose of 50-pg RE used is slightly above the minimal dose for the maximum stimulation, and a drop of activity by 10%-20% would not be detected. Therefore more precise dose-response studies are in progress. We note that 8-M urea treatment has no effect on the activity of the RE. Thatmostoftheactivityrecoveredisofamolecular weight above 50,000 daltons suggests it is unlikely that the mitogenic activity found in the RE is present in a smallmolecularweight protein associatedwith a higher molecular weight protein by ionic bounds, since the samples were centrifugedin 8-Murea and the fractions dialyzed back against PBS for biological activity tests. The eye contains a large amount of vitamin A and its derivatives known for decades to participate in the mechanism of vision. Vitamin A and retinoic acid are also known to be involved in the control and differentiation of many tissues of ectodermal origin
Table 3. All the RE fractions were tested for biological activity at a dose of 50 pg ml-' of culture medium as described in Methods. Measures were done in triplicate, and each experiment was standardized to give 100% for the non-fractionated RE. The increase in counts induced at the maximum of stimulation after the addition of RE was 4- 10 x above the control (PBS)according to the experiments. Since the exact minimal dose of RE for the 100% stimulation could not be determined precisely, the values given for the 100% recoveries indicate that no significant loss of activity was found at the dose used. The molecular weight cut-off was obtained with Amicon membrane cones, Centriflow CF25 and CF50 (-+ 25,000 and -+ 50,000 Daltons) and with the A 75 Amicon macrosolute concentrator protein Untreated RE
Treated RE Percentage of stimulation
PBS
0%
RE RE RE RE RE RE
100% 100% 5% 100% 10% 0%
> 25,000 < 25,000 > 50,000 < 50,000 > 75,000
Treatment and molecular weight separations in presence of 8 M urea
PBS RE (PBS) 8M urea RE RE > 50,000 RE < 50,000 RE > 25,000 RE < 25,000
0YO 100% 100% 100% 20% 100% 5%
39
D. Bamtault et al.: Eye Derived Growth Factors
[26,27]. It was thus of interest to assess the effects of retinoic acid on BEL cells in culture and to compare its action with RE. For instance, several reports have pointed out that retinoic acid can change the morphology of cells in vitro [28]. In the presence of retinoic acid (2 x 10-6-2 X 10-8M) the BEL cell morphology is not affected after short (2 days) or long treatments (25 days). However, there seems to be a slight increase in cell population with 2.10-8 M retinoic acid, but it is much smaller than if the cells had received 50 pg - ml-' RE. At higher concentraM, retinoic acid slightly inhibits cell tion than proliferation. Retinoic acid, M, added at the same time as RE, does not prevent cell elongation and stimulation. At a higher concentration such as 2.10-6 M, it can also decrease RE growth stimulation (Fig. 8).
Thus, retinoic acid and RE clearly have distinct actions on BEL cells, and some of them are antagonistic. A similar effect has been described following long treatment with cortisol (Arruti, unpublished work). Cortisol does not increase the mitogenic effect of RE on BEL cells, in contrast to its role on 3T3 fibroblast stimulated by FGF (results not shown). UV irradiation at a very high dose (10 jm-2) of RE for inactivating retinoic acid or its derivatives had no effect upon RE mitogenic activity of RE on BEL cells.
N.RE-like Activity can be Found in Other Ocular Tissuesfrom Bovine or Other Species Using the same technique for preparing neutral buffered extracts from ocular tissues as used for RE,
Fig. 8. Effect of RE and retinoic acid RA on BEL cell proliferation. This multilayer dish illustrates the general aspect of multilayers of BEL cells obtained after 26 days in vitro in standard condition (Dl-D2-D3), or RE (50 pg ml-l)treatment (Cl-C2-C3) as well as in retinoic acid administered alone at several concentrations (2.10-6 M = D4-D5-D6; 2.lO-'M = Al-A2-A3; 2.10-' M = Bl-B2-B3) or in the presence of RE. RE 50 pg ml-1 + RA 2.10-6 M = A4-A5-A6; RE 50 pg ml-' FL4 2.lO-' M = B4-B5-B6; RE 50 pg ml-' RA 2.10-' M = C4-C5-C6
+
+
40
D . Barritault et al.: Eye Derived Growth Factors
Table 4. Effects of different ocular tissue extracts on H3 thymidme incorporation and cell number in BEL cultures
Tissue of origin PBS alone Retina Iris Choroid Vitreous body Aqueous humour Crystalline cortex
cpm H3 thymidine
pg of total protein per ml culture medium
Number of cefls/cm2
(4 h)
0 100 100 100 100 50 100
56.10' 112.10' l00.ld 110.103 90.103 S0.ld 55.103
10,000 170,000 160,000 130,000 140,000 12,000 10,000
Retina, iris, and choroid were crude extracts in PBS prepared after 100,oOog centrifugation and dialysis. Vitreous body was first concentrated by ammonium sulfate (AS) precipitation; the fraction used was the supernatant of 20% AS (precipitated by 60% AS). Aqueous humour was concentratedusing a B 15 Amicon macrosolute concentration (mol. wt. cut-off above 15,OOO Daltons). Undialyzed samples had a similar stimulatory effect These results were obtained from experimentsin which BEL cells were maintained at confluency stimulated for 2 days and labelled for 4 h. In each experiment, one control was done by adding PBS alone. Since the extracts are not purified, a dose-response curve has to be determined for every sample. Only the maximum stimulation is shown here. Similar results have been obtained with extracts from human, calf. and horse retina
we examined their action on BEL cell proliferation and cell elongation. The incorporation of H3 thymidine in these cells after 2 days of treatment indicated that extracts from iris, pigmented epithelium, choroid, and the vitreous body had the same activity, while extracts from lens cortex or aqueous humour had no stimulatory effect (Table 4). These extracts were tested on other cells and also gave the same results, except for the vitreous extract which elicited a weaker elongation response. While until they are purified we have no evidence that these factors are similar, that they are prepared by the same technique, that they elicit the same response on target cells, and that mitogenic activities for RE and VE are found in the same PI range, all argue strongly in favour of one factor (or one family) common to these different tissues. A similar growth factor can be obtained also from human, horse, or embryonic calf retina (results not shown). We decided to call them EDGF (eye-derived growth factors) by analogy with PDGF (platelet-derived growth factors).
Discussion Is there an ubiquitous growth factor in the eye? The data presented in this work clearly demonstrate that such a growth factor (or several growth factors) is present in the adult eye, and is easily accessible for extraction. Since extracts from the different sources have not yet been purified to homogeneity, we cannot exclude that the growth stimulation is due to more
than one molecule. In addition, there may be complicated interactions between them and the serum proteins. However, the few biochemical properties already known (molecular weight, iso-electric point, and stability to denaturation) seem to indicate that they are similar or belong to the same family. EDGF has in vitro a very potent activity on various cell types stimulating their proliferation and maintaining some of their differentiated properties. That it is an endogenous factor leads us to explore its role in tissue regeneration and development as in the case of other well characterized growth factors like EGF and FGF. In contrast to these factors whose physiological target cells have not been fully characterized, RE or EDGF may act to maintain the integrity and differentiation of topographically and even ontogenically related tissues such as the corneal endothelium (which it does so precisely in vitro) or the lens epithelium. It could be stored in an inactivated form in normal tissue or circulate at a low level to maintain normal growth. If released in greater amounts under pathological conditions, it could be responsible for several disorders such as vascular endothelium proliferation. Our results showed, for instance (Table 2), that RE can induce proliferation of the vascular endothelial cells in vitro. Recent experiments using the growth of capillaries induced by pieces of agar dipped in RE and deposited on egg chorioallantoic membrane, a technique described recently for other growth factors [29], have shown that it may be also angiogenic (unpublished results). If these experiments are confirmed by other techniques, they will point to RE as an important
41
D. Bamtault et al.: Eye Derived Growth Factors
potential factor in the etiology of the main diseases of the retina where hypervascular proliferation occurs. Thepresenceofmitogenicandneurotrophicfactors in the retina have already been proposed to explainthe control of lens regeneration in the newt [30]. Recently the presence in the vitreous of a factor which promotes lens fiber cell differentiation was described [31]. Many of its chemical properties are shared with EDGF from adult bovine eye. It would be spectacular if these two factorswere related, sincetheywoulddemonstratethat the same material regulates cell proliferation and cell differentiation. The preceding paper has demonstrated that BEL cell morphology could be modulated by RE. From the results on cytoskeleton organization and extracellular deposition of fibronectin, it was proposed that these modifications were brought about by the specific response of these cells to a growth stimulus. We proposed that cells with similar architecture would respond in a similar way to any factor which has the growth-promoting capacity. This hypothesis that we have confirmed in studying the effect of RE on bovine corneal endothelial cells [32] was also recently illustrated by Gospodarowicz on the same cells and other cells using FGF and EGF [23]. Cells which do not respond to RE by an elongation such as myoblast as illustrated in Fig. 5 may have a different cytoskeleton organization. Conversely, the striking elongation observed in fibroblasts which are already spindle-shaped could be a deterrent to their growth. Thus the possibility of response of a given cell to a growth factor would be dictated not by its embryonic lineage but by its intracellular organization. the other hypothesis which does not exclude the first one could be that RE directly controls the deposition of the extracellular material which in turn controls the susceptibility of these cells to serum growth factors. Such a hypothesis was recently proposed by Gospodarowicz and illustrated in several cell types [24]. These results outlined the general interest of studying these growth factors in the understanding of the control of cell division and cell differentiation.
Acknowledgements. We want to express our gratitude to J. Tassin and M. Olivit5 for their excellent technical assistance during this work, to D. Gospodarowia, R. Clayton, M. T. Corvol, I. Guedon, L. Legaut, M. Fizman for their expertise in growing cells, and to C. Jacque for the immunological titration of FGF. This work was supported by a grant from the INSERM (ATP 79-114) and was presented at the NEI Ocular Tissue Culture Symposium in Bethesda, Md., October 22-23, 1979.
References 1. Arruti C, Courtois Y (1978) Morphological changes and growth stimulation of bovine epithelial lens cells by a retinal extract in vitro. Exp Cell Res 117: 283 2. Coulombre JL, Coulombre AJ (1963) Lens development. Fiber elongation and lens orientation. Science 142: 1489 3. Jacobson AG (1966) Inductive processes in embryonic development. Science 152: 25 4. Karkinen-Jaaskelainen M (1978) Transfilter lens induction in avian embryo. Differentiation 12: 31 5. Nordman J (1977) Au sujet du vieillissement du cristallin humain et de la cataracte dnile. Adv Ophthalmol 34: 1 6. Reddan JR, Weinsieder A, Wilson D (1979) Aqueous humor from traumatized eyes triggers cell division in the epithelial of cultured lenses. Exp Eye Res 28: 267 7. Laurent MV, Lonchampt MO, Regnault F, Tassin J, Courtois Y (1978) Biochemical, ultrastructural and immunological study of in vitro production of collagen by bovine lens epithelia cells in culture. Exp Cell Res 115: 127 8. HughesRC,MillsG,CourtoisY(1979)Roleoffibronectininthe adhesivenessof bovine epithelial cells. Biol Cell 36: 321 9. Bradford MM (1976) A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal Biochem 72: 248 10. GospodarowiaD,BialeckiH,GreenburgG (1978)Purification of the fibroblast growth factor activity from bovine brain. J Biol Chem 252 :3736 11. HughesRC,LaurentMV,LongchamptMO,CourtoisY (1975) Biosynthesis in cultured epithelial cells of bovine lens. Eur J Biochem 52: 143 12. TassinJ, Malaise E, CourtoisY (1979) Human lens cells have an in vitroproliferative capacityinverselyproportional to the donor age. Exp Cell Res 123:388 13. Buckingham ME, Caput D, Cohen A, Whalen RG, Gros F (1974) The synthesis and stability of cytoplasmic messenger RNAduMgmyoblast differentiationin culture. ProcNatl Acad Sci USA 71 : 1466 14. Koenigsberg IR (1963) Clonal analysis of myogenesis. Science 140: 1273 15. Guedon I, Bamtault D, Courtois Y,Prunieras M (Differentiation in press) Karyotypes and sister chromatid exchanges in normal human adult keratinocytes as a routine test 16. Gospodarowicz D, Moran JS, Braun DL (1977) Control of proliferation of bovine vascular endothelial cells. J Cell Physiol 91 : 377 17. Arruti C, Courtois Y (1979) Retinotrophic stimulation of proliferation of bovine corneal endothelial cells in vitro. Proceedingsfor 20th Meeting of Association for Eye Research. Paris, 25-28 September 18. Corvol MT, Dumontier MF, Rappaport R (1975) Culture of chondrocytes from the proliferative zone of epiphyseal growth plate cartilage from prepubertal rabbits. Biomedicine 23: 103 19. Legault Demare L, Zeitour Y, Lando D, Lamande N, Grasso A, Gros F (1980) Expression of a specific neuronal protein 143-2 during in vitro differentiation of neuroblastoma cells. Exp Cell Res 125:233 20. HervC B, Jacquemin E,Courtois Y (1979) Histones biosynthesis and turnover in epithelial lens cultured in vitro. Cell Biol Int Rep 3: 271 21. Westall F, Lennon V,GospodarowiczD (1978) Brain-derived fibroblast growth factor: identity with a fragment of the basic protein of myelin. Proc Natl Acad Sci USA 75: 4675
42 22. Delassale A, Jacque C, Drouet J, Raoul M, Legrand JC (1980) Radioimmunoassay of the myelin basic protein in biological fluids: conditions improving sensitivity and specificity. Biochimie 62: 159 23. GospodarowiczD, Greenburg G, Birdwell CR (1978) Determination of cellular shape by the extracellular matrix and its correlation with the control of cellular growth. Cancer Res 38: 4155 24. Vlodavsky I, Lui GM, GospodarowiczD (1980) Morphological appearance, growth behaviour and migratory activity of human tumor cells maintained on extracellular matrix versus plastic. Cell 19: 607 25. Gospodarowicz D, Greenburg G, Bialecki H, Zetter BR (1978) Factors involved in the modulation of cell proliferation in vivo and in vitro: The role of fibroblast and epidermal growth factors in the proliferative response of mammalian cells. In Vitro 14: 85 26. Seeds NW, Gilman AG, Amano T, Niremberg MW (1970) Regulation of Axon formation by clonal lines of a neural tumor. Proc Natl Acad Sci USA 66: 160
Note Added in Proofs. Recently the angiogenic nature of EDGF was confirmed by our own laboratory using a perfusion system in the rabbit cornea (Stimulation of neovascularisation of the rabbit cornea by EDGF. P. Thompson, D. Maurice, D. Bamtault, J. Plouet, Y.Courtois, in press) and by two other laboratories (Chen CH and Chen SC 1980 Invest. Ophtalm. 19 5% and Glaser B, d'Amore P, Michels RG, Patz A. and Feselan A (1980) J Cell Biol 84: 298) using different technics.
D. Barritault et al.: Eye Derived Growth Factors 27. Saari JC, Futterman S, Bredberg L (1978) Retinal binding protein in the retina. J Biol Chem 253: 6432 28. Patt LM, Itaya K,Hakomori S (1978) Retinal induces density dependent growth inhibition and changes in glycolipids and lets. Nature 273: 379 29. Folkman J, Merler E, Abernathy C, Willums G (1971) Isolation of a tumor factor responsible for angiogenesis. J Exp Med 133: 275 30. Yamada T (1977) Control mechanisms in cell type conversion in newt lens regeneration. In: Wolsky A (ed) Monographs in developmental biology, vol. 13. S. Karger, Base1 31. Beebe D, Feagans DE, Jebens HA (1980) Lentropin a factor in vitreous humour which promotes lens fiber cell differentiation. Proc Natl Acad Sci USA 77: 490 32. Arruti C, Courtois Y (1980) The in vitro control of proliferation and cell interaction of bovine corneal endothelial cells by a retinal growth factor. Eur J Cell Biol 22: 386 Received March 1980lAccepted July 1980