Basic fibroblast growth factor and its receptor messenger ribonucleic acids are expressed in human ovarian epithelial neoplasms
Anna Maria Di Blasio, MD,” Laura Cremonesi, PhD, Paola Viganó, PhD,“ Maurizio Ferrari, MD,’ Denis Gospodarowicz, PhD,“ Mario Vignali, MD,” and Robert B. Jaffe, MDe
Milan, Italy, and San Francisco, California
OBJECTIVE: Our purpose was to determine whether basic fibroblast growth factor is present in, and synthesized by, human ovarian epithelial neoplasms and to evaluate the expression of gene for the basic fibroblast growth factor receptor.
STUDY DESIGN: The synthesis of basic fibroblast growth factor and its receptor was investigated in seven primary human ovarian epithelial neoplasms. Neoplastic tissues were homogenized and the cytoplasmic extracts purified by heparin-sepharose chromatography with a linear salt gradient of 0.6 to 3 mol/L sodium chloride in Tris-hydrochloric acid. The in situ synthesis of basic fibroblast growth factor and its receptor was demonstrated by polymerase chain reaction. Total ribonucleic acid was reverse transcribed and then amplified with two oligonucleotide primers specific for the bovine and human basic fibroblast growth factor gene and its human receptor gene.
RESULTS: As assessed by both bioassay and radioimmunoassay a peak of basic fibroblast growth factor-like activity was present in all tumors in the chromatographic fractions eluted with 3 mol/L sodium chloride. The mitogenic effect on bovine adrenocortical endothelial cell proliferation varied from 35% to 153% above control cultures. Levels of basic fibroblast growth factor-like immunoreactivity were between 4 and 33 ng/ml. Qualitatively similar results were obtained after purifying the cytoplasmic extract of dispersed human ovarian tumor cells. The mitogenic effect was completely abolished by a specific neutralizing anti-basic fibroblast growth factor antibody. Single major deoxyribonucleic acid bands of the expected size (354 and 661 bp) were detected in all tumors studied. The identity of this material with the human basic fibroblast growth factor sequence was confirmed by restriction enzyme analysis. CONCLUSION: These data demonstrate that both basic fibroblast growth factor and its receptor are present in and synthesized by human ovarian tumor cells. Thus basic fibroblast growth factor might stimulate their abnormal proliferation through an autocrine mechanism. (AM J OBSTET GYNECOL 1993;169:1517-23.)
Key words: Basic fibroblast growth factor, basic fibroblast growth factor receptor, ovarian cancer, growth factor
Ovarian epithelial cancer is the most common malig- nant ovarian neoplasm and is a leading cause of death among women.’ The factor(s) that regulates the rapid growth of ovarian epithelial carcinoma and other types of malignant tumors is still largely unknown. Recently the search for the molecular and cellular basis of ma- lignant transformation resulted in the integration of
From the Departments of Obstetrics and Gynecologyª and Laboratory Medicine, University of Milano, Istituto Scientific San Raffaele, and the Reproductive Endocrinology Center, University of California, San Francisco.“
Presented in part at the Fortieth Annual Meeting of the Society for Gynecologic Investigation, Toronto, Ontario, Canada, March 31- April 3, 1993.
Reprint requests: Anna Maria Di Blasio, MD, Department of Obstet- rics and Gynecology, H. San Raffaele, Via Olgettina 60, 20132 Milano, Italy.
Copyright @ 1993 by Mosby-Year Book, Inc. 0002-9378/93 $1.00 + .20 6/6/50958
studies of oncogenes and peptide growth factors. Evi- dence indicating that oncogenes can confer growth factor autonomy on cancer cells2. 3 has accumulated. The autocrine secretion hypothesis proposes that, as a result of oncogene activation, neoplastic cells can es- cape growth-restraining mechanisms by independently producing, and responding to, their own growth fac- tors. This hypothesis can be broadened to include the autonomous synthesis of growth factor receptors and postreceptor signaling.4. 5 As a component of studies designed to explore the factors responsible for the rapid growth and proliferation of human ovarian can- cer, we investigated the possible role of basic fibroblast growth factor (bFGF). It is a likely candidate because it (1) has angiogenic properties,6 (2) has been found in a variety of other neoplasms,7-11 and (3) has been found in normal ovarian tissue,12. 13 which undergoes
profound cyclic changes involving both proliferation and angiogenesis. Thus we designed experiments to determine whether bFGF is present in, and synthesized by, human ovarian epithelial neoplasms. Moreover, we also examined the expression of the gene for the bFGF receptor.
Material and methods
Reagents. Tissue culture media, calf serum, fetal calf serum, gentamicin, fungizone, and trypsin were ob- tained from Flow (Opera, Italy). Multiwell tissue culture dishes were from Corning (Milan); 35 and 10 cm culture dishes were obtained from Falcon (Milan). Heparin-Sepharose was purchased from Pharmacia (Milan). Bovine pituitary bFGF was purified as previ- ously described.14 The polyclonal antibody against bFGF was prepared as previously described.15 The GeneAmp ribonucleic acid (RNA) amplification kit was obtained from Perkin-Elmer (Milan).
Cell culture. Solid tissue from primary ovarian car- cinoma was minced into small pieces (1 to 2 mm3) and enzymatically dispersed in RPMI 1640 containing 0.1% collagenase and 0.01% deoxyribonuclease. The cells were incubated for 1 hour at 37° C, then resuspended in RPMI 1640 and carefully layered on a Ficoll gradient. After centrifugation at 1500g for 30 minutes, tumor cells were recovered from the interface and plated in RPMI 1640 supplemented with 10% fetal calf serum, 2 mmol/L glutamine, and 50 µg/ml gentamicin. Medium was changed every other day until cells became conflu- ent and were subjected to heparin-Sepharose chroma- tography.
Capillary endothelial cells derived from bovine adre- nal cortex were prepared as previously described.16 Stock cultures of these cells were cultured in Dulbecco’s modified Eagle medium supplemented with 2.5 µg/ml fungizone, 50 µg/ml gentamicin, 2 mmol/L glutamine, and 10% calf serum. Then bFGF (1 ng/ml) was added every other day until confluence was achieved; the cells were then passaged weekly at a split ratio of 1:64.
Purification of tissue and cell extracts. Samples of seven human ovarian epithelial neoplasms were ob- tained at laparotomy. Tissue was homogenized in 0.5% Triton X-100. An aliquot of 10 mmol/L Tris-hydrochlo- ric acid, pH 7.0, containing 3 mol/L sodium chloride was added to obtain a final concentration of 0.3 mol/L sodium chloride. The extract was then centrifuged (50,000g for 30 minutes at 4° C), and the supernatant was applied to a heparin-Sepharose column (bed vol- ume 0.5 ml) that had been preequilibrated at room temperature with 10 mmol/L Tris-hydrochloric acid containing 0.6 mol/L sodium chloride. The column was sequentially eluted with the same buffer containing 0.6, 1.0, or 3.0 mol/L sodium chloride. Aliquots of the fractions eluted were diluted with 0.2% gelatin in
calcium- and magnesium-free phosphate-buffered saline solution and tested for their ability to stimulate cell proliferation.15, 16 This same procedure was per- formed on the pellet derived from human ovarian tumor cells.
Mitogenic assay. The mitogenic activity of the col- umn fractions was assessed with a bioassay of capillary endothelial cells from bovine adrenal cortex as target. Cells were seeded at a density of 5 x 103 in 12-well plates containing 1 ml of Dulbecco’s modified Eagle medium supplemented with 10% calf serum and anti- biotics. Aliquots of 10 ul of the appropriate dilutions of each sample were then added to the wells on days 1 and 3. After 5 days in culture triplicate plates were trypsinized, and cell proliferation was determined with a Coulter counter (Coulter, Hialeah, Fla.).
Radioimmunoassay (RIA). The presence of bFGF immunoreactivity in the chromatographic fractions was assessed with a commercial RIA kit (Amersham, Milan).
Deoxyribonucleic acid (DNA) amplification. Total RNA was extracted from human epithelial neoplasms according to the method of Chomczynski and Sacchi.17 The presence of bFGF and bFGF receptor messenger (m) RNAs was demonstrated by amplifying respective target sequences with polymerase chain reaction ac- cording to the instructions provided with the GeneAmp amplification reagent kit. One microgram of total RNA was reverse transcribed to prepare complementary (c) DNA. Polymerase chain reaction was performed on the entire cDNA product with Taq (Thermus acquaticus) DNA polymerase with the manufacturer’s recom- mended buffers. Twenty and twenty-two nucleotide primers beginning with codon 215 at the 5’ end of the bovine bFGF cDNA18 and complementary to codons 547 to 568 at the 3’ end were prepared. The predicted amplification product was 354 bp. For the bFGF recep- tor 22 nucleotide primers corresponding to nucleotides 955 to 976 at the 5’ end and complementary to nucleo- tides 1594 to 1615 at the 3’ end of bFGF receptor mRNA9 were used.
Reaction conditions for reverse transcription were as follows: 1 mmol/L each deoxynitrophenyl triphos- phate, 1 unit of RNasin, 100 pmol of random hex- amer, and 200 units of reverse transcriptase. The reaction was run at 42º C for 1 hour. The reaction mixture was then heated at 99º C for 5 minutes and quick chilled on ice.
The amplification reaction mixture consisted of add- ing 10 pmol each of upstream and downstream primers and 1 unit of Taq DNA polymerase to the reverse transcription mixture. Samples were incubated for 30 polymerase chain reaction cycles. To amplify the bFGF cDNA fragment, during each cycle the samples were heated at 94° C to denature template complexes (120 seconds initially and 30 seconds during all subsequent
NaCl
0.6M
1.0M
3.0M
7
6
D Cells/well (X 10”*)
5
4
30
· bFGF (ng/ml )
3
20
2
1
10
0
0
1
5
10
15
20
25
30
35
Fraction
cycles), cooled at 58° C to permit templates and primers to anneal (30 seconds), and heated at 72° C to allow extension to occur (1 minute). For the receptor cDNA fragment amplification, conditions were as follows: 94° C for 1 minute (denaturation), 63º C for 1 minute (primer annealing), and 72º C for 1 minute (primer extension).
Each time the polymerase chain reaction was carried out, a blank was prepared that used all reagents and substituted 1 ul of water for the RNA. Amplification products were visualized on a 4% agarose gel. The gel was stained with ethidium bromide and viewed on an ultraviolet light box.
Results
To determine whether bFGF is present in human ovarian neoplasms, extracts of these tissues were sub- jected to heparin-Sepharose affinity chromatography. The eluted fractions were analyzed for their ability to stimulate the proliferation of capillary endothelial cells from bovine adrenal cortex. As shown in Fig. 1, elution of the column with 3 mol/L sodium chloride yielded a major peak of bioactivity. The same results were ob- tained when the tumor cells were dispersed, and their cytoplasmic extract was subjected to heparin-Sepharose chromatography (Fig. 2). This chromatographic profile is similar to that of bFGF, which is eluted with the highest salt concentration. The chromatographic frac- tions positive for the presence of mitogenic activity also
showed bFGF-like immunoreactivity when analyzed by a RIA that does not cross react with acidic fibroblast growth factor and other related peptides (Fig. 1). Fig. 3 shows that increasing concentrations of native bFGF or purified tumor extract could displace bound iodine 125 bFGF in a parallel manner. The amount of bFGF immunoreactivity present in the positive fractions of all of the tumors studied correlated well with the results obtained by bioassay (Table I). Finally, the identity of the mitogenic substance present in the ovarian tumors and bFGF was further demonstrated with a polyclonal antibody. When capillary endothelial cells from bovine adrenal cortex were cultured in the presence of an aliquot of purified tumor extract with an anti-bFGF antibody, the mitogenic effect previously observed was completely abolished (Fig. 4).
To assess whether the presence of the protein in the ovarian tumors was caused by in situ synthesis, we evaluated the expression of bFGF mRNA. This was accomplished by amplifying bFGF mRNA target se- quences with polymerase chain reaction. Total RNA from capillary endothelial cells from bovine adrenal cortex was used as a positive control in the validation of the reverse transcription-polymerase chain reaction, be- cause it previously was demonstrated that this cell type expresses the bFGF gene.15
Fig. 5 shows that reverse transcription of ovarian tumor mRNA generated a DNA product corresponding to the predicted length, 354 bp, of the bFGF amplifi-
0.6 M
NaCl
1.0 M
3.0 M
3
Cells/well (x10-5)
2
1
0
1
5
10
15
20
25
30
35
40
Fraction
· Tumor extract (ul)
100
3.125
6.25
12.5
25
80
B/Bo (%)
60
40
20
0
1.25
2.5
5
10
20
40
80
D bFGF (fmol/tube)
cation product. The primer sequences used have been successfully used to detect bFGF transcript in rat ova- ries.13 They were specific for regions separated by the first intron of the bFGF gene18; thus the oserved band could only have originated from mRNA and not ge- nomic DNA. The identity of the amplified product with the primer-defined bFGF DNA sequence was further
| Patient No. | Fraction | Bioassay (% stimulation)* | RIA (ng/ml) |
|---|---|---|---|
| 1 | 30 | 35 | 4 |
| 2 | 26 | 100 | 18 |
| 27 | 153 | 26 | |
| 3 | 28 | 69 | 8 |
| 29 | 87 | 10 | |
| 30 | 53 | 6 | |
| 4 | 28 | 56 | 9 |
| 5 | 27 | 74 | 12 |
| 28 | 90 | 33 | |
| 6 | 27 | 38 | 10 |
| 28 | 87 | 12 | |
| 7 | 28 | 30 | 10 |
| 29 | 42 | 15 |
*Increase in target cell proliferation expressed as percent- age of control cultures.
demonstrated by restriction enzyme analysis. The am- plified fragment of the human bFGF gene contains both an HinfI and a BamHI site.19 In contrast, the bovine gene does not have a BamHI site because of the presence of a proline at amino acid position 127 rather than the serine found in the human sequence.18 Con- sistent with this observation, digestion with BamHI and HinfI cut the bFGF polymerase chain reaction product derived from human ovarian tumors into two fragments of the expected size. In contrast, the bFGF polymerase chain reaction product derived from bovine endothelial
3
Cells/well (X10-5)
2
1
0
Control
bFGF Ab
Ovarian Tumor
Ovarian Tumor
+ bFGF Ab
cells was completely digested only by Hinfl and was unaffected by BamHI digestion (Fig. 6).
Finally, to evaluate whether bFGF might act through an autocrine mechanism, we also determined if bFGF receptor gene was expressed in these neoplastic tissues. As shown in Fig. 7, with specific oligonucleotide prim- ers, a 661 bp DNA fragment of the human bFGF receptor sequence was amplified in all the samples studied.
Comment
The results of this study demonstrate that human ovarian epithelial neoplasms contain a mitogenic sub- stance indistinguishable from bFGF, as judged by sev- eral criteria. Indeed, the protein present in the tumor extracts has high affinity for heparin and has an elution profile similar to that of bFGF. Furthermore, it demon- strates bFGF-like immunoreactivity and is able to stim- ulate the proliferation of bFGF target cells. This bio- logic action is completely abolished in the presence of a bFGF antibody. The amount of this bFGF-like sub- stance, as judged by RIA, varies in the different tumors studied but always correlates with the biologic activity. Moreover, reverse transcription polymerase chain reac- tion indicates that the presence of this protein is most likely the result of in situ synthesis, because the bFGF gene is expressed in all of these tumors. In addition, the bFGF receptor gene is also expressed in the same samples.
1114
-
320
1 2 3 4 5 6 7 8
It has long been known that cancer cells require fewer exogenous growth factors than do their normal coun- terparts.20 The potential role of growth factors in neo- plastic transformation has recently been emphasized. Experimental evidence indicates a close similarity be- tween the epidermal growth factor receptor and the v-erb-B oncogene protein sequence.4 Furthermore, platelet-derived growth factor is structurally related to the protein p28sis of simian sarcoma virus.3 Thus it would appear that oncogene activation can induce the synthesis of growth factors or their receptors by neo- plastic cells and that these processes could lead to autonomous cell growth.
bFGF has both mitogenic and angiogenic properties. Therefore it has been implicated as a potential regula- tor of tumor cell growth in several human tumors, including melanoma,7 bladder and kidney carcinoma,10 rhabdomyosarcoma,8 and human gliomas.9 It also has been reported that an oncogene isolated from Kaposi sarcoma DNA encodes for a growth factor that has significant homology with bFGF and is indeed a new member of the fibroblast growth factor family.21 Further support for the possible role of bFGF in neoplastic cell proliferation comes from a recent study by Morrison.9
This study demonstrated that both transformed and nontransformed human astrocytes express bFGF mRNA. However, modulation of bFGF synthesis by antisense oligonucleotides inhibits the growth of trans- formed human astrocytes, although it does not affect
1114
320
190
147
1 2 3 4
5 6 7 8
1114
692
1 2 3 4 5 6 7 8
the growth potential of normal astrocytes. This may indicate that the mere presence of the growth factor is not sufficient to promote abnormal cell growth. It is possible that bFGF overexpression, changes in its re-
ceptor, or the presence of bFGF isoforms might induce and sustain neoplastic transformation.
bFGF synthesis has been demonstrated in normal bovine12 and human (Di Blasio AM. Unpublished data) ovarian granulosa cells. Moreover, this growth factor can modulate both their proliferation22 and differenti- ated function.23 We now show that transformed human ovarian cells also are a site of bFGF and its receptor synthesis. These findings are consonant with other data supporting a role for the autocrine synthesis of growth factors in the development of human ovarian neo- plasms. Indeed, it has previously been reported that an epidermal growth factor-like protein is present in ex- tracts of ovarian carcinoma and that its levels are increased in those neoplasms that respond poorly to chemotherapy.24 Furthermore, ascitic fluid from ovar- ian cancer patients does contain mitogenic activity from at least one unique growth factor.25 The presence of this substance is necessary for the intraperitoneal growth of human ovarian cancer cells in immunodeficient nude mice.26
The results presented herein indicate that ovarian cancer cells are able to synthetize bFGF and its receptor but do not elucidate the mechanisms by which bFGF may play a role in the abnormal proliferation of these cells. Our data suggest that quantitative differences in growth factor synthesis might be present in the differ- ent tumors studied. Whether overexpression of bFGF might be a function of the histologic differences be-
tween tumors needs to be established in a larger num- ber of samples. Finally, studies are currently underway to evaluate whether unique bFGF isoforms with en- hanced mitogenic activity or differences in bFGF recep- tors are present in these neoplasms.
REFERENCES
1. Pfleiderer A. A diagnosis and staging of ovarian cancer. J Cancer Res Clin Oncol 1984;197:81-8.
2. Sporn MB, Roberts AB. Autocrine growth factors and cancer. Nature 1985;313:745-7.
3. Waterfield MD, Scrace GT, Whittle N, et al. Platelet- derived growth factor is structurally related to the putative transforming protein p28sis of simian sarcoma virus. Na- ture 1983;304:35-9.
4. Downward J, Yarden Y, Mayes E, et al. Close similarity of epidermal growth factor receptor and v-erb-B oncogene protein sequences. Nature 1984;307:521-6.
5. Landis CL, Master SB, Spada A, Pace AM, Bourne HR, Vallar L. GTPase inhibiting mutations activate the & chain of Gs and stimulate adenylyl cyclase in human pituitary tumors. Nature 1989;340:692-6.
6. Gospodarowicz D, Ferrara N, Schweigerer L, Neufeld G. Structural characterization and biological functions of fi- broblast growth factor. Endoc Rev 1987;8:95-114.
7. Baird A, Esch F, Mormede P. Molecular characterization of fibroblast growth factor: distribution and biological activ- ities in various tissues. Recent Prog Horm Res 1986;42: 143-205.
8. Schweigerer L, Neufeld G, Mergia A, Abraham JA, Fiddes JC, Gospodarowicz D. Basic fibroblast growth factor in human rhabdomyosarcoma cells: implication for the pro- liferation and neovascularization of myoblast-derived tu- mors. Proc Natl Acad Sci U S A 1987;84:842-6.
9. Morrison RS. Suppression of basic fibroblast growth factor expression by antisense oligodeoxynucleotide inhibits the growth of transformed human astrocytes. J Biol Chem 1991;266:728-34.
10. Chodak GW, Hospelhorn V, Judge SM. Increased levels of basic fibroblast growth factor-like activity in urine from patients with bladder or kidney cancer. Cancer Res 1988; 48:2083-8.
11. Presta M. Sex hormones modulate the synthesis of basic fibroblast growth factor in human endometrial adenocar- cinoma cells: implication for the neovascularization of normal and neoplastic endometrium. J Cell Physiol 1988; 137:593-7.
12. Neufeld G, Ferrara N, Schweigerer L, Mitchell R, Gospo- darowicz D. Bovine granulosa cells produce basic fibro- blast growth factor. Endocrinology 1987;121:597-603.
13. Koss RD, Olson CE. Expression of basic fibroblast growth
factor in the rat ovary: detection of mRNA using reverse transcription-polymerase chain reaction amplification. Mol Endocrinol 1989;3:2041-8.
14. Gospodarowicz D, Cheng J, Lui GM, Baird A, Bohlen P. Isolation by heparin-sepharose affinity chromatography of brain fibroblast growth factor: identity with pituitary fibro- blast growth factor. Proc Natl Acad Sci U S A 1984;81: 6963-7.
15. Schweigerer L, Neufeld G, Friedman J, Abraham JA, Fiddes JC, Gospodarowicz D. Capillary endothelial cells express basic fibroblast growth factor, a mitogen that promotes their own growth. Nature 1987;325:257-9.
16. Gospodarowicz D, Massoglia S, Cheng J, Fujii DK. Effect of fibroblast growth factor and lipoproteins on the prolif- eration of endothelial cells derived from bovine adrenal cortex, brain cortex and corpus luteum capillaries. J Cell Physiol 1986;127:121-36.
17. Chomczynski P, Sacchi N. Single-step method of RNA isolation by acid guanidinium thiocyanate-phenol-chloro- form extraction. Anal Biochem 1987;162:156-9.
18. Abraham JA, Mergia A, Whang JL, et al. Nucleotide sequence of a bovine clone encoding the angiogenic pro- tein, basic fibroblast growth factor. Science 1986;233: 545-8.
19. Abraham JA, Whang JL, Tumolo A, et al. Human basic fibroblast growth factor: nucleotide sequence and genomic organization. EMBO J 1986;5:2523-8.
20. Holley RW. Control of growth of mammalian cells in cell culture. Nature 1975;258:487-90.
21. Delli Bovi P, Curatola AM, Kern FG, Greco A, Ittman M, Basilico C. An oncogene isolated by transfection of Kapo- si’s sarcoma DNA encodes a growth factor that is a mem- ber of the FGF family. Cell 1987;50:729-37.
22. Tapanainen J, Leinonen P, Tapanainen P, Yamamoto M, Jaffe RB. Regulation of human granulosa-luteal cell pro- gesterone production and proliferation by gonadotropins and growth factors. Fertil Steril 1987;48:576-80.
23. Adashi EY, Resnick CE, Croft CS, May JV, Gospodarowicz D. Basic fibroblast growth factor as a regulator of ovarian granulosa cell differentiation: a novel non-mitogenic role. Mol Cell Endocrinol 1988;55:7-14.
24. Bauknecht T, Kiechle M, Bauer G, Siebers JW. Character- ization of growth factors in human ovarian carcinomas. Cancer Res 1986;26:2614-8.
25. Mills GB, May C, McGill M, Roifman CM, Mellors A. A putative new growth factor in ascitic fluid from ovarian cancer patients: identification, characterization and mech- anism of action. Cancer Res 1988;48:1066-71.
26. Mills GB, May C, Hill M, Campbell S, Shaw P, Marks A. Ascitic fluid from ovarian cancer patients contains growth factors necessary for intraperitoneal growth of human ovarian adenocarcinoma cells. J Clin Invest 1990;86: 851-5.