Isolation of heparin-binding growth factors from dogfish (Mustela canis) brain and retina

We have purified two growth factors from dogfish brain and retina by using their binding affinity for heparin-Sepharose. These growth factors were eluted at 1 and 2 M NaCl similarly to those purified from bovine brain or retina. Their mitogenic activity was assayed in vitro on the same mammalian cells: bovine lens epithelial cells or human fibroblasts. All these data seem to indicate that these growth factors belong to the families of other well defined mammalian growth factors: EDGF I, brain basic FGF, AGF II, on the one hand and EDGF II, brain acidic FGF, AGF I, ECGF, on the other. Thus, these growth factors have been widely distributed during evolution and retain at least a conservative sequence to stimulate cell proliferation, in mammalian cells.     

Lens regeneration from cultured newt irises stimulated by retina-derived growth factors (EDGFs)

It has been shown that lens regeneration from the iris of the newt Notophthalmus viridescens is dependent on the presence of neural retinal tissue in organ culture and in vivo. The recent discovery of various eye-derived growth factors (EDGFs) in the bovine retina [14] prompted us to investigate whether one of these factors may be involved in the stimulation of lens regeneration. Dorsal irises were cultured for 20 days in serum-supplemented diluted Eagle's medium. Growth factors from bovine retina of various degrees of purification were added. Lens regeneration was assessed on the basis of morphological lens-regeneration stages and by immunofluorescent detection of a lens-specific marker protein, alpha-crystallin. Crude isotonic retinal extract at 80-800 micrograms/ml significantly augmented lens regeneration. Very similar results were obtained when EDGF III, the nonretained retinal factor after heparin-affinity chromatography, was present at 2-20 micrograms/ml. Lens regeneration was also significantly increased when EDGF II, the retinal form of acidic fibroblast growth factor (aFGF) at 50-500 ng/ml was added to the cultures. On the other hand, EDGF I at 4-40 ng/ml and brain basic FGF at 5-50 ng/ml did not seem to significantly stimulate lens regeneration under the conditions used. Our results suggest that at least two retina-derived growth factors (EDGF II and III) can stimulate lens regeneration. These growth factors may be the putative signal that is naturally produced by the retina during lens regeneration in the newt.     

Regulation of eye derived growth factor binding to membranes by light, ATP or GTP in photoreceptor outer segments.

Eye derived growth factor II (EDGF II), the retinal form of acidic fibroblast growth factor (aFGF) is present in rod outer segments (ROS) purified in the dark, which display higher EDGF specific activity than all other parts of the retina. EDGF binds to ROS disc membranes upon illumination. This binding is not reversible in the dark. ATP, but not GTP, readily releases EDGF from either dark-adapted or previously bleached ROS. The release of EDGF activity from ROS membranes would require a phosphorylation mechanism since AMP-PNP, an ATP analogue, is not efficient. ROS membranes compete with cellular EDGF receptors of retinal pigment epithelial cells in vitro for the binding of labelled EDGF II, suggesting that they also possess specific binding sites. These data suggest that EDGF II is involved in photoreceptor cell biology.     

Human and bovine vascular endothelial cells. Comparative effect on cell growth and longevity of an eye-derived growth factor (EDGF) and of extracellular matrix

Human and bovine vascular endothelial cells from the umbilical vein and the aorta, respectively, were cultured in the presence of EDGF (a growth factor prepared from bovine retina) on plastic or on extracellular matrix (ECM). Both EDGF and ECM are required to allow the maximal proliferation of human cells and their organization in a typical monolayer. Conversely, bovine aortic endothelial cells grow perfectly in the absence of both factors in 6% fetal calf serum. However, a requirement for EDGF can also be demonstrated in low serum conditions, or in cells at high passage number. ECM had no growth promoting activity by itself. Thrombin acts similarly to EDGF on bovine serum-starved cells. EDGF prolongs the in vitro lifespan of both types of cells. Cells at all stages still synthesize factor VIII antigen as revealed by immunofluorescence. Thus EDGF, like other growth factors from brain, FGF or ECGF, may have an important role in angiogenesis, a critical problem in pathological retinas.     

Influence of chromatographic fractions of extracts derived from bovine neural retina on newt (Notophthalmus viridescens) lens regeneration in vitro

Removal of the ocular lens in adult newts (Notophthalmus viridescens) is followed by a series of cellular events leading to regeneration of a new lens by cell type conversion of pigmented iris epithelial cells at the dorsal pupillary margin (Yamada, Curr. Top. Dev. Biol. 2:247-283, 1967). Following depigmentation and five to seven cell divisions, iris epithelial cells redifferentiate into lens fiber cells and synthesize crystallin proteins (Yamada, Curr. Top. Dev. Biol. 2:247-283, 1967). This process is dependent upon neural retina in vivo (Stone, Anat. Rec. 131:151-172, 1958; Reyer, Dev. Biol. 14:214-225, 1966) and in vitro (Yamada et al., Differentiation 1:65-82, 1973). Acting on the hypothesis that the role of the neural retina is to promote passage of iris epithelial cells through the requisite number of cell cycles which will then allow them to redifferentiate as lens fiber cells (Yamada, in: Cell Biology of the Eye. Academic Press, New York, 1982), we undertook testing of the effects of eye-derived mitogenic substances, as well as other mitogens, on regeneration of lens from iris in organ culture. We have previously defined a critical period for the retinal influence in vivo and in vitro, and have shown that crude extracts of retina can enhance regeneration of lenses in culture (Connelly et al., J. Exp. Zool., 240:343-351, 1986). In this paper, we report on the lens regeneration enhancing activity (LRA) of more highly purified fractions of the retinal extracts. Heparin-sepharose chromatography of the crude retinal extract yields three fractions (Courty et al., Biochemie 67:265-269, 1985) called EDGF I, II, and III. EDGF I and II have affinity for heparin, while EDGF III does not. In our bioassay, LRA appears only in the EDGF III fraction. Dialysis of EDGF III against 0.1 N acetic acid yields a fraction which has affinity for cibacron blue sepharose (eluting at 2.15 M salt) and also has significant LRA. Because insulin at high doses has a marginal effect on lens regeneration in culture (Williams and McGlinn, Am. Zool. 19:923, 1979; Connelly, Differentiation 16:85-91, 1980), we tested IGF-I. Because of the putative neurotrophic effects of transferrin (Tf) (Mescher and Munaim, J. Exp. Zool., 230:485-490, 1986), we tested Tf for its ability to enhance regeneration of the lens in culture. IGF-I seems to have an enhancing effect on lens regeneration; Tf does not.     

Separation of growth-stimulating activity of BSA fraction V from the bulk of albumin using Heparin Sepharose Chromatography

Bovine serum albumin (BSA) is a potential source of biological contamination in cell culture medium. The aim of this work was to attempt to replace BSA in low serum and serum-free medium (SFM). BSA fraction V was subjected to a variety of processes in order to determine if the growth promoting activity observed for NRK cells could be extracted from the BSA molecule. These included solvent extractions, diafiltration, reverse phase HPLC and affinity chromatography using heparin sepharose. Solvent extraction and diafiltration failed to remove the activity from the BSA. Affinity chromatography using heparin sepharose indicated that all of the activity observed with BSA was retained in the 0.5 M NaCl fraction and was associated with less than 3% of the original protein. The major protein band in the 0.5 M NaCl fraction had the same apparent molecular weight as albumin (as seen by SDS-PAGE and analytical reverse phase HPLC). Unlike the untreated BSA, the 0.5 M NaCl fraction was partially susceptible to proteolytic digestion and to variations in pH.Abbreviations HS heparin sepharose - DHS donor horse serum - SFM serum free-medium     

Regulation of fibroblast growth factor-like protein(s) in the androgen-responsive human prostate carcinoma cell line LNCaP

Growth of the normal and malignant prostate is known to be regulated by androgens. Part of their effect has been suggested to be mediated through coordinated regulation of secreted growth factors with autocrine function. We now examine the biological role of preferentially paracrine acting factors in growth control of prostate cancer, i.e. fibroblast growth factor(s) (FGF). Coculture experiments using the androgen-responsive human prostate carcinoma cell line LNCaP as feeder cells and the FGF-dependent human adrenal carcinoma SW-13 cell line as target cells show that (i) LNCaP cells induce growth of SW-13 cells, (ii) even higher stimulation of SW-13 cells is seen in the presence of androgen treated LNCaP cells and (iii) a specific anti-bFGF antibody inhibits growth of SW-13 cells induced by androgen treated LNCaP cells; no proliferation of SW-13 cells occurs in the absence of LNCaP cells. Partial purification of the secretory products of LNCaP cells was performed by affinity chromatography using a heparin sepharose column. Fractions were tested for biological activity in a soft agar assay with SW-13 cells. Several activities could be detected, the main activity was eluted with about 1.5 M NaCl. These data suggest that androgen treatment of LNCaP cells leads to enhanced secretion of proteins which belong to the FGF-family.     

Biochemical properties of the endothelium-derived growth factor: Comparision to other growth factors

Cultured bovine aortic endothelial cells (BAEC) can be maintained at saturation density for several weeks in the absence of serum. These cells retain viability and normal culture morphology, and continuously produce a growth factor for mesenchymally derived cells–the endothelium-derived growth factor (EDGF). The amount and specific activity of EDGF that is produced by BAEC under serum-free conditions remains constant for weeks. The levels of EDGF produced under these serum-free conditions is equivalent to levels produced in medium containing 5% plasma-derived serum. EDGF has been found to be trypsin sensitive, acetone and ammonium sulfate precipitable, and resistant to heat and sodium dodecyl sulfate treatment. Gel filtration on Sephacryl S-200 in the presence of formic acid (1%) yields two major peaks of activity corresponding to proteins of apparent molecular weights of approximately 24,000 and 14,000 daltons. This chromatographic step affords a ten-to 12-fold purification with a combined recovery of greater than 85%. Unlike brain or pituitary fibroblast growth factor, EDGF activity is destroyed by dithiothreitol or periodic acid. EDGF is not a somatomedin since it exhibits no detectable sulfation activity in a porcine cartilage assay. EDGF is not inhibited by antiserum to epidermal growth factor and is capable of stimulating DNA synthesis in a 3T3 variant cell line that is nonresponsive to and lacks receptors for epidermal growth factor. The majority of EDGF activity does not behave like the platelet-derived growth factor during ion exchange chromatography. Antisera prepared in rabbits and in mice to human platelet-derived growth factor has little effect on bivine or human EDGF activity. These biochemical and immunological properties of EDGF indicate that it is distinct from several other well-characterized polypeptide growth factors.     

Biochemical comparative studies between eye- and brain-derived growth factors

Two bovine brain-derived growth factors, BDGF I and BDGF II, were isolated using the same extraction procedure as previously described for eye-derived growth factors (EDGF). The hypothesis that these growth factors were identical to EDGF I and EDGF II, respectively, was supported by their similar molecular weights (16,000 and 15,000, respectively) and isoelectric points (9.0 and 5.0, respectively), their identical retention behavior on reverse-phase chromatography and their similar amino acid compositions. From studies on their binding properties to cell surfaces, competition between EDGF I and BDGF I as well as competition between EDGF II and BDGF II to the same receptor was observed. The amino terminal sequence of EDGF II (1-16) was shown to be identical to the amino acid residues (7-22) of the acidic FGF, strongly confirming our observations on the identity of the factors isolated from bovine brain and retina.     

Platelet derived growth factor induces ornithine decarboxylase activity in NIH 3T3 cells

Incubation with highly purified human Platelet Derived Growth Factor induced ornithine decarboxylase activity in quiescent NIH 3T3 cells concomitantly with mitogenic stimulation. Pretreatment of cells with a specific ornithine decarboxylase inhibitor, DL-alpha-difluoromethyl-ornithine significantly inhibited the effect of the mitogen on DNA synthesis. These experiments suggest that the mitogenic activity of Platelet Derived Growth Factor, similarly to that of other serum growth factors or tumor promoters, is mediated through rise in polyamine levels.     

Immunological study of acidic fibroblast growth factor (aFGF) distribution in the eye

During the last ten years, several groups, including the present authors, have detected growth factor activities in various ocular tissues, and the presence of a ubiquitous Eye-Derived Growth Factor (EDGF) has been described. More recently, isolation and characterization of this growth factor activity from the retina led to the identification of two molecules. These molecules were shown to be identical to other growth factors isolated from neuronal and non-neuronal tissues and are now designated as acidic and basic fibroblast growth factor (aFGF, bFGF). The biological function and the reason for the ubiquitous distribution of these factors remain unclear. Understanding may be improved by quantification of this distribution in various tissues during development. In the present study, specific polyclonal antibodies were raised against acidic FGF, aFGF was determined in various ocular tissues by enzyme immunoassay, and the localization of immunoreactive aFGF by immunohistological staining with fluorescent antibodies or with enzyme- or gold-labeled antibodies was studied. In almost all tissues tested aFGF was found; but the retina, cornea, and vitreous body contained the highest levels of aFGF per gram of tissue. In the retina, aFGF was associated primarily with the nerve fiber layer and the inner and outer segments of the photoreceptors, whereas corneal aFGF was detected in the cytoplasma of the basal layer of epithelial cells.     

A unique family of endothelial cell polypeptide mitogens: the antigenic and receptor cross-reactivity of bovine endothelial cell growth factor, brain-derived acidic fibroblast growth factor, and eye-derived growth factor-II

Bovine brain, hypothalamus, pituitary, and retina contain potent anionic polypeptide mitogens for endothelial cells. Immunological assays using murine monoclonal antibodies against bovine endothelial cell growth factor (ECGF) and radioreceptor assays using [125I]ECGF were performed to determine the cross-reactivity of ECGF with bovine acidic pI brain-derived fibroblast growth factor (acidic FGF) and bovine eye-derived growth factor-II [EDGF-II). We observed that acidic FGF and EDGF-II are recognized by anti-ECGF monoclonal antibodies and compete with [125I] ECGF for receptor occupancy. Furthermore, the biological activity of ECGF, acidic FGF, and EDGF-II is potentiated by the glycosaminoglycan, heparin. These results argue that ECGF, acidic FGF, and EDGF-II belong to a common family of polypeptide growth factors.     

Further purification of human pituitary-derived chondrocyte growth factor: heparin-binding and cross-reactivity with antiserum to basic FGF

The previously described human pituitary-derived chondrocyte growth factor (CGF), mitogenic for rabbit fetal chondrocytes, was found to bind to heparin-Sepharose and was eluted with 1.5M NaCl. Further characterization of CGF demonstrated a molecular weight of 18-20 kD and cross-reactivity with antiserum to synthetic bovine basic fibroblast growth factor (FGF1-24). When human pituitaries were homogenized in 0.15 ammonium sulfate (pH 5.5) and the extract chromatographed on heparin-Sepharose, 98% of the mitogenic activity was adsorbed to heparin and eluted with 3M NaCl. These findings indicate that CGF is closely related or identical to basic FGF and that the bulk of mitogenic activity in the human pituitary extracts binds to heparin.     

An eye-derived growth factor regulates epithelial cell proliferation in the cultured lens

Lenses in organ culture permit an analysis of factors acting on epithelial cell growth, while keeping the normal steric constraints of the cell population. By employing this technique with radioautography of epithelial whole mounts, we showed that the DNA synthesis found in the epithelia of cultured bovine lenses follows an organized spatial and temporal pattern during culture. Within the first 48 h, active cells were located at the preequatorial region ("germinative zone"), a distribution consistent with the in vivo spatial organization of multiplying cells. Starting at about 48 h, cells from the central region of the epithelium--a nonproliferating population--were triggered to synthesize DNA in the presence of eye-derived growth factor (EDGF). When cultured in serum-free medium, only a small fraction of the cells was labeled, but when a low serum concentration was present, this fraction reached 50% of the cell population. The stimulatory effect of EDGF required a lag period, but its effect reached a maximum exceeding that found for serum. However, the cells from the germinative region, having a cell density three- to four-fold higher than the central region, were not stimulated to proliferate. This occurred irrespective of the presence of EDGF or serum. If this growth-stimulatory activity derived from the retina were the actual factor controlling cell proliferation in the lens in vivo, then the results presented here would point to the presence of a regulatory mechanism similar to that known for some other hormones.     

Isolation of heparin-binding growth factors from bovine, porcine and canine hearts

Fresh bovine, porcine and canine hearts were homogenized and mitogens for mesoderm-derived cells were purified in three different steps. Extraction by two different ammonium sulfate precipitations was followed by cation-exchange chromatography and by heparin-Sepharose affinity chromatography. A heparin-Sepharose fraction from heart (eluted at 1.1 M NaCl) increased mitotic activity in serum-deprived cultures of porcine aortic endothelial and smooth muscle cells, and in human fibroblasts. This mitogenic activity is potentiated by heparin and inhibited by gamma-interferon. The heart mitogenic fraction showed one double peak on HPLC at A215 and one polypeptide band on SDS/PAGE. These peaks and bands were identical to those obtained from bovine brain. The heart acidic fibroblast growth factor (aFGF) showed a positive signal in Western blots using antibodies raised against brain aFGF. Gas-phase amino acid sequencing established that the mitogens were identical to aFGF and the N-terminally truncated aFGF. Extraction in the presence of a protease inhibitor (pepstatin A) produced a higher-molecular mass form of aFGF with a blocked amino terminus. Another mitogen, eluted at 1.6 M NaCl from heparin-Sepharose, reacted with polyclonal antiserum against human recombinant basic fibroblast growth factor (bFGF) and showed a 66% (12 from 18 amino acids determined by gas-phase sequencing) similarity with bFGF. This polypeptide increased the mitotic activity of the same cell lines but was more potent than aFGF.     

Identification of the hepatocyte mitogen in bovine spleen as heparin-binding growth factors.

Growth promoting activity for rat hepatocytes in bovine spleen was identified as three heparin-binding growth factors. All the features tested, such as heparin affinity, molecular mass, cross reactivity with antibody, and partial amino acid sequence, indicated that one of the three factors was identical to FGF-1 (fibroblast growth factor-1, acidic FGF), another one was related to FGF-2 (fibroblast growth factor-2, basic FGF), whereas it was more potent for hepatocytes than the FGF-2 purified from bovine brain. The third one was eluted from heparin-Sepharose column at 0.75M NaCl, of which activity was not abolished by anti-FGF-1 or FGF-2 antibodies. In addition, the mitogenic effect of this factor was synergistic with that of HGF (hepatocyte growth factor), a known potent hepatocyte mitogen, suggesting that it is a novel growth factor for hepatocytes.     

Endothelial cell growth factors in embryonic and adult chick brain are related to human acidic fibroblast growth factor.

We have investigated the nature of endothelial cell growth factors in 14-day embryonic and adult chick brain extracts. Mitogenic activity was isolated by a combination of cation-exchange, heparin-Sepharose affinity, and reverse-phase HPLC. Two major mitogenic fractions eluted from heparin-Sepharose at 0.8-1.3 M and 1.5-2 M. Biologically active proteins eluting at 0.8-1.3 M NaCl, after purification to homogeneity from embryonic and adult brain, were found to possess the same amino-terminal sequence as human acidic fibroblast growth factor (aFGF). The notion that the isolated mitogens represent chick aFGF is further supported by the findings that their affinity for heparin and their retention behavior in highly resolutive HPLC are indistinguishable from those of genuine aFGF. Mitogenic activities eluting at 1.5-2 M NaCl were also present in embryonic and adult brain, but in quantities insufficient for preliminary characterization. The high specific mitogenic activity for endothelial cells, high affinity for heparin and cross-reactivity with antibodies against bovine basic FGF (bFGF) suggest a relationship of those materials with basic FGF. Our data also suggest that the sequence of aFGF is highly conserved among vertebrates. While angiogenesis occurs predominantly in the embryonic brain, the absence of notable differences in the contents of the potent angiogenic factors aFGF and bFGF in embryonic versus adult chick brain is interesting.     

Identification of a low molecular weight form of epithelial transforming growth factor

We have purified a novel form of epithelial transforming growth factor (TGFe) from bovine kidney. Acid ethanol extracts of kidney were fractionated by size exclusion, reverse phase and cation exchange chromatography and activity was monitored by measuring growth of SW13 adrenocortical carcinoma cells in soft agar. The purified material was highly cationic, bound weakly to heparin and gave a band at 13-15000 Mr by SDS-PAGE following Bolton-Hunter iodination. This band correspond to the migration of biological activity extractable from gel slices. The results suggest that we have isolated a truncated form of TGFe which nonetheless retains biological activity.     

Heparin-like glycosaminoglycans influence growth and phenotype of human arterial smooth muscle cells in vitro. I. Evidence for reversible binding and inactivation of the platelet-derived growth factor by heparin

Summary We have investigated the effects of interactions between growth factors and heparin-like glycosaminoglycans on untransformed human arterial smooth muscle cells (hASMC) in vitro. The results indicate that heparin in the presence of serum mitogens prevents the cells from entering the S phase of the cell cycle by binding and inactivating reversibly some serum mitogen(s). Our results suggest that platelet-derived growth factor (PDGF) is one of them and that it is the most potent stimulator of hASMC growth in vitro. Thymidine incorporation as well as increase in DNA content was inhibited not only by the presence of heparin in serum-containing medium but also when serum was chromatographed on Heparin-Sepharose at physiologic salt concentrations before exposure to the cells. The mitogenic activity of the unretained serum fraction was restored by the addition of PDGF AA, AB, or BB dimers or of a fraction (RF I) that dissociated from Heparin-Sepharose at 0.2 to 0.6M NaCl. Radiolabeled recombinant PDGF (c-sis) dissociated from Heparin-Sepharose within a concentration range of NaCl similar to that of RF I. Neither the unretained material nor the RF I or PDGF dimers were effective alone. The effect of RF I was significantly decreased by the addition of an anti-PDGF IgG that is known to neutralize the PDGF mitogenic activity partially. Addition of heparin abolished DNA-synthesis when the PDGF dimers or RF I were combined with the unretained fraction. A second fraction (RF II) bound strongly to Heparin-Sepharose and eluted between 1.1 and 1.6M NaCl. The RF II also induced DNA synthesis but was neither as efficient as RF I nor depending on other serum fractions for growth promotion and it was not inhibited by anti-PDGF IgG. A similar strong affinity for Heparin-Sepharose was found for labeled basic fibroblast growth factor and we cannot exclude the possibility that RF II represent fibroblast growth factor. Under these culture conditions, inhibition of hASMC proliferation was directly correlated with the expression of smooth muscle specific alpha actin isoforms in stress fibers and the suppression of a proliferating cell-specific nuclear antigen. Conversely, stimulation of hASMC proliferation was associated with the opposite phenomenon. We conclude that heparin-like glycosaminoglycans influence growth and phenotype of hASMCs in vitro by binding and inactivating PDGF. Inasmuch as heparin-like substances constitute a significant proportion of the proteoglycan-associated glycosaminoglycans of the arterial wall, such mechanisms might be important for the development of atherosclerotic lesions.     

Glycosaminoglycan structures required for strong binding to midkine,a heparin-binding growth factor

Midkine (MK), a heparin-binding growth factor, binds strongly to oversulfated structures in chondroitin sulfates (CSs) and heparan sulfate. To elucidate the carbohydrate structure actually involved in the strong binding, dissected brains from 13-day mouse embryos were incubated with [14C]-glucosamine. The labeled glycosaminoglycans were fractionated by MK-agarose affinity chromatography to a weakly binding fraction, which was eluted by 0.5 M NaCl, and a strongly binding fraction, which was eluted by higher NaCl concentrations. Among the unsaturated disaccharides released from the strongly binding fraction by chondroitinase ABC, DeltaDi-diSE with 4,6-disulfated N-acetylgalactosamine accounted for 32.3%, whereas its content was lower in the weakly binding fraction. Artificial CS-E structure was formed using N-acetylgalactosamine 4-sulfate 6-O-sulfotransferase purified from squid or recombinant human enzyme. Analysis of the products and their interaction with MK revealed that E units without 3-O-sulfation of glucuronic acid are sufficient for strong binding, provided that they are present as a dense cluster. Among the sulfated disaccharides released by heparitinase digestion, the trisulfated one, DeltaDiHS-triS, was the most abundant in the strongly binding fraction and was lower in the weakly binding fraction. Together with results of previous studies, we concluded that the multivalent trisulfated heparin-like unit is another structure involved in strong binding to MK.