As published in Molecular Human Reproduction vol. 3 no. 8 pp. 685-691, 1997

Expression of the fibroblast growth factor receptor in women with leiomyomas and abnormal uterine bleeding.

Departments of Obstetrics, Gynecology and Reproductive Biology and Pathology, Brigham and Women's Hospital, Harvard Medical School, Boston, Massachusetts, USA

¹ To whom correspondence should be addressed at: Laboratory for Human Reproduction and Reproductive Biology, Room 449, Boston Lying-In, 221 Longwood Avenue, Boston, Massachusetts 02115, USA

Basic fibroblast growth factor (bFGF) is a regulator of angiogenesis which is overexpressed in leiomyomas compared to matched myometrium. To understand the physiological significance of this finding we characterized the expression of the type 1 receptor for this ligand (FGFR1). Utilizing RT-PCR we identified the complete and alternatively spliced transmembrane forms and two secreted forms of the FGFR1 in endometrium, myometrium and leiomyomas from all patients. This is the first report of secreted forms in uterine tissue. Proteins consistent with each of these isoforms were identified by western blot analysis in all three tissues. Menstrual cycle specific regulation of FGFR1 protein was seen in the endometrial stroma of normal women but not in women with leiomyomas and abnormal uterine bleeding utilizing immunohistochemistry. Stromal FGFR1 expression is suppressed in the early luteal phase in normal women, but not in women with leiomyoma-related bleeding. These findings support the role of the bFGF ligand-receptor system in the pathogenesis of leiomyoma-related bleeding and may have implications for fertility and contraception since the differential FGFR1 expression occurs in the peri-implantation period of the early luteal phase.
Key words: angiogenesis/fibroblast growth factor/leiomyoma/ uterine bleeding

Introduction

Despite the central role that the endometrial vasculature must play in any gynecologic bleeding disorder, few studies have evaluated endometrial microvasculature and angiogenesis in women with abnormal uterine bleeding. Angiogenesis is a complex process that occurs in many pathological conditions. As a physiological process, angiogenesis occurs primarily in the female reproductive tract, in particular, vascularization of the endometrium during the menstrual cycle and development of the corpus luteum.

Ovarian steroids are the classic regulators of endometrial development, but recent evidence indicates that growth factors, either alone or in conjunction with steroids play an important role in endometrial growth (Irwin et al., 1991). Altered expression of these factors may affect endometrial development by causing local abnormalities in vascular structure and/or function and these abnormalities may be important in the pathogenesis of abnormal bleeding.

Basic fibroblast growth factor (bFGF) is an angiogenic growth factor that is highly mitogenic for capillary endothelial cells in vitro and can induce angiogenesis in vivo (Folkman and Klagbrun, 1987). It is present in the endometrium, myometrium and corpus luteum of women throughout the normal menstrual cycle (Ferriani et al., 1993; Rusnati et al., 1990; Salat-Baroux et al., 1994). In addition, bFGF has been shown to be stored in extracellular matrix (ECM) and can initiate remodeling of ECM, an important step in angiogenesis (Gospodarowicz, 1983). Uterine leiomyomas are benign smooth-muscle tumours that often cause abnormal bleeding, although the pathogenesis of this process is poorly understood. Leiomyomas are characterized by large amounts of ECM, and contain abundant amounts of bFGF in this ECM (Mangrulkar et al., 1995). Leiomyomas may therefore serve as a reservoir for bFGF and impact endometrial vasculature through a paracrine or local endocrine effect.

Although the expression of bFGF in the uterus has been described, to fully understand its function requires knowledge of its receptors in endometrium, myometrium and leiomyomas throughout the menstrual cycle. The gene for bFGF receptor type 1 (FGFR1) has been cloned and is known to generate multiple isoforms by alternative splicing of mRNA (Johnson et al., 1991). The complete form of the receptor is composed of three extracellular immunoglobulin-like domains, a transmembrane region, and an intracellular portion exhibiting intrinsic tyrosine kinase activity (Lee et al., 1989). An alternatively spliced transmembrane variant has been characterized that has only two immunoglobulin-like domains (Lee et al., 1989). A third isoform generated by alternative splicing encodes for a secreted form of the receptor that lacks the transmembrane and intracytoplasmic domains (Johnson et al., 1991). The functional significance of the secreted form is unknown, but it may act as a binding protein for bFGF and thus may prevent the binding between bFGF and the membrane-bound receptor. Using women with leiomyomas as a model system, we thus examined differences in receptor expression in women with and without abnormal bleeding to gain insight into the pathogenesis of abnormal uterine bleeding.

Materials and Methods

Preparation of human tissue

Samples of myometrium, endometrium and leiomyomas from the same uterus were obtained under an approved human subjects protocol at the Brigham and Women's Hospital. Tissue was obtained from fourteen premenopausal women undergoing hysterectomy for leiomyoma-related bleeding. All preoperatively had complaints of menorrhagia and documented anemia. Pathologic examination of the uterus confirmed leiomyomas with mean uterine weight of 496 ± 431 gm and no other pathologic reasons for bleeding. Normal control endometrial biopsy specimens were obtained from 14 premenopausal women who were undergoing bilateral tubal ligation who had no known uterine pathology and no evidence of abnormal uterine bleeding. Histologic dating of the endometrium was performed for all samples utilizing classic criteria (Noyes et al., 1950). Patients with leiomyoma-related bleeding were comparable to control patients in terms of age (45.8 ± 3.2 vs. 37.3 ± 6.5), gravidity (2.9 ± 2.1 vs. 3.3 ± 0.7) and parity (2.6 ± 1.7 vs. 2.7 ± 0.8). No patient in either group received any hormonal medication for at least two months before hysterectomy or biopsy.

Tissues were obtained within 30 minutes of surgical resection and divided into three sections: one was minced and immediately homogenized in 4M guanidine isothiocyanate for RNA processing, one was placed in 10% formalin for histological analysis, and the third was homogenized and processed for membrane protein extraction.

Immunohistochemistry

Paraffin sections were first deparaffinized and rehydrated in phosphate buffered saline. Sections were then incubated with a mouse monoclonal anti-FGFR1 primary antibody [Santa Cruz Biotechnology, Inc., Santa Cruz, California] 2 mg/ml for 30 minutes at room temperature. Nonspecific mouse IgG (Sigma, St. Louis, MO) was used as a negative control at 5ug/ml. Binding was visualized with the Vecta-Stain kit (Vector Laboratories, Burlingame, CA) using diaminobenzidine tetrahydrochloride (DAB) for detection with hematoxylin counterstaining. Negative controls were examined for every slide. Staining intensity in tissue sections was evaluated and graded (1 weak; 2 moderate; 3 strong) in a blinded fashion by two examiners and assigned a semiquantitaive HSCORE, assessing both intensity of staining and percentage of stained cells (Lessey et al., 1992). Endometrial glands and stroma were assessed separately. Intra-assay and interassay variations for this technique have been previously validated (Budwit-Novotny et al., 1986).

Statistical analysis

HSCORE data was analyzed using the Wilcoxon rank sum test for independent nonparametric variables utilizing STATA statistical software (Computing Resource Center, College Station, TX). Data are reported as the mean ± standard deviation.

RNA isolation and cDNA synthesis

Total cellular RNA was extracted from fresh tissue using guanidine isothiocyanate as previously described (Chirgwin et al., 1979). cDNA was synthesized using 1 ug of total RNA with oligo (dT) priming (Pharmacia, Piscataway, NJ) using a standard protocol (Nowak et al., 1993; Yeh et al., 1993).

Polymerase chain reaction (PCR) amplification

Three different primer pairs were used to detect the three different forms of the FGFR1. To detect the complete transmembrane form of FGFR1, primers used were 5'-ATGTGGAGCTGGAAGTGCCTC-3' and 5'-GCTGCTGGTTACGCAAGCATA-3' (Isacchi et al., 1990; Yeh and Osanthanondh, 1993). These primers amplify a 315-bp region corresponding to amino acids 1 to 319 (Isacchi et al., 1990). The FGFR1 alternatively spliced transmembrane variant was amplified with the same upstream primer and a different downstream primer (5'-GGTGTTATCTGTTTCTTTCTC-3') corresponding to amino acids 422 to 431 (Isacchi et al., 1990; Yeh and Osanthanondh, 1993). These primers produce expected bands of 432 bp (complete form) and 165 bp (alternatively spliced form). The FGFR1 secreted form was amplified utilizing a different primer pair: 5'-TATGCCACCTGGAGCATCATAATGGACTCTGTGGTGCCCTCTGAC-3' and 5'- GACTGGCCCACGAAGACTGGTGCCAT-3' with an expected amplified band of 341 bp (Johnson et al., 1991). These primers encompasses bases 877 to 1217. Amplification was carried out as per our standard protocol for forty cycles with the following thermocycling conditions: 94¡C for 30 seconds, 55¡C for 2 minutes and 72¡C for 1 minute. Umbilical vein RNA was used as a positive control and negative controls were used in each amplification and included all reagents, except cDNA.

Analysis and sequencing of PCR products

The amplified products were analyzed by electrophoresis on 2% agarose gels with a 100-bp DNA ladder (Pharmacia, Piscataway, NJ) as a molecular weight standard. All products were restriction digested: restriction digestion of DNA from the complete form of FGFR1 by AvaII produces two fragments, 183 and 132 bp; digestion of the alternative form of FGFR1 by Hinf1 results in fragments of 130 and 35 bp; and digestion of the secreted form by AvaII gives fragments of 122 and 219 bp. DNA sequence analysis was also performed on all products to confirm their identities. Gel-purified PCR products were sequenced using an ABI 373A DNA sequencer (Perkin-Elmer, Foster City CA) using dye-primer cycle sequencing.

Isolation of cellular membranes

Membranes were isolated from surgical specimens as described by Mukku and Stancel (Mukko and Stancel, 1985). Tissue samples were homogenized in 1,4-piperazine diethane sulfonic acid (PIPES) buffer with protease inhibitors leupeptin, antipan, benzamidine, chymostatin and pepstatin to prevent protein degredation (Mukko and Stancel, 1985). The homogenate was centrifuged twice at 800 x g for 15 minutes at 4¡C. The resulting supernatants were combined and centrifuged at 105,000 x g for 60 minutes at 22¡C. Membrane pellets were resuspended and samples were centrifuged at 105,000 x g for 60 minutes at 22¡C. The final membrane pellet was suspended in PIPES buffer containing 0.15M sodium chloride. Samples were then stored at -70¡C. The protein concentration was determined with the Coomassie Protein Assay Reagent Kit (Pierce, Rockford, IL). The final concentration of membrane proteins ranged from 425 mg/ml to 1.5 mg/ml.

Immunoblotting

Immunoblots were performed with slight modification from our standard protocol (Mangrulkar et al., 1995). Membrane proteins (30 mg) were treated with denaturing buffer at 68¡C for three minutes and separated on an 8.0% SDS-polyacrylamide gels under nonreducing conditions. Proteins were transferred electrophoretically to nitrocellulose membranes [0.1um pore size (Protran, Keene, NH)] in 0.025 M Tris, 0.192 M glycine and 20% methanol transfer buffer overnight at 40 volts in 4¡C. To prevent non-specific protein binding, blots were pretreated in Tris-buffered saline (TBS) containing 10% nonfat dry milk at 37¡C for 60 minutes. The blots were reacted with a polyclonal rabbit anti-FGFR1 antibody [Upstate Biotechnology, Lake Placid, NY] diluted 1:250 in TBS containing 0.05% Tween (TBS-T) for 30 minutes at 22¡C. The FGFR1 antibody is reported by the manufacturer to recognize the extracellular region of chicken, rat, and human FGFR1. The polyclonal antibody was used for immunoblotting after we found that the monoclonal FGFR1 antibody was not as sensitive when used on the membrane extracts. Blots were washed in TBS-T and reacted with anti-rabbit IgG alkaline phosphatase conjugate [1:1000 dilution (Sigma, St. Louis, MO)] for 30 minutes at 22¡C. After a final wash in TBS-T, immunoreactive proteins were visualized with NBT/BCIP (Pierce, Rockford, IL), a modified alkaline phosphatase substrate. Rat liver was used as a negative control and rat brain as a positive control, for FGFR1; both were received as a gift from Dr. Virginia Rider.


Results

Immunohistochemical analysis of FGFR1 protein localization demonstrated a significant difference in stromal cell staining between normal women and women with leiomyoma related bleeding (Figures 1 and 2). In normal women, stromal cells demonstrated significant cell-associated staining in the early proliferative and late luteal phases of the cycle with suppression of staining that begins in the late proliferative phase and extends through the mid-luteal phase. In the women with abnormal uterine bleeding and leiomyomas, this suppression does not occur. A significant difference in stromal HSCORE was seen (Table I). Although there appears to be a suppression of receptor expression in women with abnormal uterine bleeding in the late luteal phase, no significant difference was seen in stromal HSCORE in either the proliferative phase or late luteal phase. Staining for endometrial glandular cells was similar for both groups with intense homogeneous staining of the cytoplasm in both proliferative and secretory phases of the menstrual cycle

.

 

Figure 1. HSCORE for immunohistochemical staining of endometrial stroma in women with normal menstrual cycles (left-hand panel) and leiomyoma-related bleeding (right-hand panel). Luteal phase samples were dated and proliferative phase samples were classified as early, mid or late (E, M, or L). Day 28 samples were considered menstrual samples for the purpose of analysis. A suppression of FGFR1 expression is seen in the normal women that may begin as early as the late proliferative phase and extends through the mid-luteal phase. This suppression is absent in most women with abnormal uterine bleeding.



In both groups, the myometrium demonstrated homogeneous cytoplasmic staining that did not vary through the menstrual cycle (Figure 2G, H). Overall, the intensity of the cellular staining for the receptor was greater in the myometrium than in leiomyomas (Figure 2G, H, I, J). Leiomyomas, however, contained large amounts of ECM, which did show some immunoreactivity for the receptor (Figure 2I, J). Leiomyomas also exhibited homogeneous cytoplasmic staining that did not vary through the menstrual cycle. Matched negative controls were consistently immunonegative.

Figure 2. Immunohistochemical localization of fibroblast grwoth factor receptor type 1 (FGFR1) in human uterine tissues throughout the menstrual cycle. Panels A, C, E, G and I are negative controls and were treated with nonspecific (Ig)G at a concentration of 5 ug/ml. Panels B, D, F, H, and J are treated with a mouse monoclonal antibody against FGFR1 at a concentration of 2 mg/ml. All endometrial samples shown are from patients with no abnormal bleeding. Panels A and B show early proliferative endometrium with moderate glandular staining and stromal cells showing significant cytoplasmic staining. Panels C and D show a sample of day 17 endometrium. The glandular staining is similar with no staining of the endometrial stroma. Panels E and F show day 24 endometrium. Significant cytoplasmic staining is seen in the stroma as well as the glandular staining. Panels G and H show day 22 endometrium with part of the underlying myometrium. Discrete staining is seen in the glands with minimal endometrial stromal staining. The myometrial smooth muscle cells shows strong staining. Panels I and J show a leiomyoma with light staining throughout the extracellular matrix and some cell associated staining but overall decreased staining when compared to that of normal myometrium.


















Utilizing reverse transcription (RT)-PCR, mRNA for all three forms of the FGFR1: 432 bp (complete transmembrane), 165 bp (alternatively spliced transmembrane) and 341 bp (secreted), were identified in the endometrium, myometrium and leiomyomas (Figure 3 A, B). Additionally, the primers designed to amplify the transmembrane forms also amplified a 540 bp fragment of a second secreted form of FGFR1 (Figure 3A, B) (Eisemann et al., 1991). For ease of discussion we will hereafter identify the secreted form amplified as the 540bp fragment as secreted form A and the form detected as the 341 bp fragment as secreted form B (Eisemann et al., 1991; Johnson et al., 1991).

There was no menstrual cycle variability seen in any tissue for any of the identified bands. Restriction digestion supported the identification of these RT-PCR products and sequencing confirmed these findings (Figure 3C, D).

Figure 3. Reverse transcription-polymerase chain reaction (RT-PCR) detection of fibroblast grwoth factor receptor type 1 (FGFR1) isotypes. (A) A schematic diagram of the primers used and fragments amplified. The three extracellular immunoglobulin-like domains (Ig) are diagrammed leading to the transmembrane portion of the molecule. Primers are indicated with capital letters for ease of reference and fragments are diagrammed in portion and with relative sizes represented. The sequence of downstream primer C is repeated in both of the first two immunoglobulin domains (Ig I and Ig II) and thus led to amplification of secreted form A in addition to the anticipated fragments. (B) Amplification of the complete and alternatively spliced transmembrane forms of the FGFR1 from human endometrium. Molecular weight standards are seen in lane 1. Lane 2: a 315 bp product representing the complete form, Lane 3: 540, 432, and 165 bp products are seen. The 432 and 165 bp products are consistent with both the complete and alternatively spliced transmembrane forms, respectively. The 540 bp product was sequenced and identified as a secreted form of the receptor (form A). Lane 4 represents the 341 bp product for secreted form B. Lane 5 shows a sample that did not receive reverse transcriptase and thus serves as a negative control. Identical results were seen from 8 other endometrial specimens and similar numbers of leiomyoma and myometrial samples. (C) Restriction digests for both the complete and alternatively spliced forms of the receptor. Lane 1 is the molecular weight standard. Lanes 2, 3 and 4 represent Ava II digestions of amplified samples from amplification of endometrium, myometrium and leiomyoma, respectively from a single patient. Both the 183 and the 132 bp fragment are seen in all lanes. Lanes 5, 6 and 7 show HinfI digestion of the same samples to identify the alternatively spliced form. The 130 bp fragment is seen in all lanes and the 35 bp fragment is unable to be visualized due to its small size. Identical results were seen for 7 other triplicate samples and leiomyoma/myometrium pairs for normal women. (D) Restriction digestion of the amplified secreted FGFR1 form B utilizing AvaII shows both the 219 and the 122 bp product for leiomyomas (lane 2), myometrium (lane 3) and endometrium (lane 4) from a single patient. Molecular weight markers are seen in lane 1. Identical results were seen in tissue samples from 8 other patients and myometrial and endometrial samples obtained from normal women.



To determine whether mRNAs for FGFR1 identified by RT-PCR were translated into proteins, samples of human myometrium, endometrium and leiomyomas were analyzed using western immunoblots to identify the different protein isoforms (Figure 4). Utilizing a polyclonal FGFR1 antibody that recognizes all forms of FGFR1 four major bands were identified in myometrium, endometrium and in leiomyomas. The two largest proteins (167 kD and 150 kD) were consistent with the complete and alternatively spliced transmembrane forms of the FGFR1. Two smaller proteins (70 kD and 61 kD) are consistent with secreted forms of the receptor (Duan et al., 1992). As expected, the molecular weights did not change under reducing conditions.

Figure 4. Immunoblot for fibroblast grwoth factor receptor type 1 (FGFR1) isotypes. Protein molecular weight standards (st) are shown. The remaining lanes show proteins obtained from endometrium (E), leiomyomas (L) and myometrium (M) from a single patient. All four proteins were detected equally in each tissue types although due to the efficiency of protein detection each sample did not always contain all four bands. A representative gel of the five gels run is shown with all four bands can be seen for leiomyoma and three of the four bands for myometrium and endometrium.

Discussion

This study demonstrates that expression of FGFR1 in human uterine tissues shows menstrual cycle-specific regulation and differs in women with abnormal uterine bleeding and normal women. In women with abnormal uterine bleeding and leiomyomas, expression of FGFR1 in the endometrial stromal cells is dysregulated in the early luteal phase. This time coincides with the time of embryo apposition and implantation. Since the FGF system appears to play an important role in pattern formation in the early embryo, implantation and trophoblast outgrowth, this finding may have implications for women with leiomyomas who also experience infertility or recurrent miscarriage (Carlone and Rider, 1993; Cornell et al., 1995; Haimovici and Anderson, 1993).

This study also provides anatomic evidence that the bFGF receptor/ligand system may be important in the pathogenesis of leiomyoma-related bleeding. A number of studies have suggested that vascularity is abnormal in the myomatous uterus (Stewart and Nowak, 1996). Abnormalities of bFGF and FGFR1 can lead to vascular abnormalities in a number of ways that could result in abnormal bleeding. First, bFGF is an angiogenic growth factor that can induce endothelial cell proliferation thus leading to an increased number of vessels (Presta, 1988). Alternatively, bFGF regulates the production of enzymes that cause remodeling of the ECM, including collagenase and plasminogen activator which could lead to vascular dilatation and resulting in increased bleeding in this way (Presta, 1988). Finally, the bFGF receptor/ligand system may cause bleeding by disrupting the expression of integrins, cell adhesion molecules integral to the process of angiogenesis (Klein et al., 1993). bFGF had been shown to upregulate the synthesis of §3 integrin and the expression of the av§3 integrin shows the same luteal phase suppression seen in our study for FGFR1. (Lessey et al., 1992; Lessey et al., 1994).

The secreted form of the receptor identified in this study may play a unique role in the physiology or pathophysiology of the adult human uterus, particularly the leiomyomatous uterus. Although secreted forms of the FGFR1 have been identified in other adult tissues, neither of the secreted forms we report were previously identified in studies of the human fetal uterus or mature rat uterus (Hanneken et al., 1994; Hanneken and Baird, 1995; Isacchi et al., 1990; Rider et al., 1995). Previous studies from our lab demonstrated that leiomyomas contain more bFGF than normal myometrium (Mangrulkar et al., 1995). The immunostaining for FGFR1 in this study in the extracellular matrix of leiomyomas probably represents secreted receptor since there are no cell membrane to anchor the transmembrane forms. The extracellular matrix of leiomyomas may act as a reservoir of bFGF bound to secreted receptor. Thus, in addition to disordered expression in the endometrium of FGFR1, women with leiomyoma-related bleeding may also have a reservoir of bound bFGF ligand that can act in a local endocrine manner on the endometrium.

This dysregulation of the FGF receptor/ligand system may be relevant to other states characterized by abnormal uterine bleeding. For example, women using progestin-only contraceptives have abnormal uterine bleeding and their endometrium is characterized by morphologically abnormal blood vessels (Hourihan et al., 1991; Rogers et al., 1993). Further work is needed to examine the FGF system in other states associated with abnormal bleeding.

If the bFGF receptor/ligand system is important in the pathogenesis of abnormal uterine bleeding, therapeutic strategies may be developed that target these molecules. Interferons and the coupling of bFGF to cytotoxic agents including saporin are two such agents that are currently available for this purpose (David, 1995; Ezekowitz et al., 1992; Lappi and Baird, 1991; Oleszask, 1988; Real et al., 1986; White, et al., 1989). Other agents that interfere with this receptor/ligand system may also be therapeutic options for abnormal uterine bleeding.

The current study also highlights an inherent problem in many studies that utilize endometrial tissue from unspecified hysterectomy specimens as normal control samples. Better study design will lead to a better understanding of the pathophysiology of this process. By better understanding the biology of processes resulting in abnormal uterine bleeding, we may be able to design innovative therapies with greater specificity and decreased side-effects.

Acknowledgements

The authors would like to thank Dr. Virginia Rider for her expert technical advice on the cell membrane preparations and for the rat membrane extracts. This study was supported by HD-30496, National Institutes of Health, Bethesda, MD (to R.A.N.) and F32 CA-60447-03 (to B.J.Q.)

References

Budwit-Novotny, D.A., McCarty, K.S., Cox, E.B. et al. (1986) Immunohistochemical analysis of estrogen receptor in endometrial adenocarcinoma using a monoclonal antibody. Canc Res. 46:5419-5425.

Carlone D.L. and Rider V. (1993) Embryonic modulation of basic fibroblast growth factor in the rat uterus. Biol. Reprod. 49:653-65.

Chirgwin, J.M., Przbyla,A.E., MacDonald, R.J., et al. (1979) Isolation of biologicaly active ribonucleic acid from sources enriched in ribonuclease. Biochemistry 18:5294-5300.

Cornell, R.A., Musci, T.J., and Kimelman, D. (1995) FGF is a prospective competence factor for early activin-type signals in Xenopus mesoderm induction. Development 121:2429-37.

David, M. (1995) Transcription factors in interferon signaling. Pharmacol Ther. 65:149-61.

Duan, D.R., Werner, S., and Williams L. (1992) A naturally ocurring secreted form of fibroblast growth factor receptor 1 binds basic FGF in preference over acidic FGF. J Biol Chem.267:16076-80.

Eisemann, A., Ahn, J.A., Graziani, G., et al. (1991) Alternative splicing generates at least five diffferent isoforms of the human basic-FGF receptor. Oncogene 6:1195-1202. Ezekowitz, R.A., Mulliken, J.B., and Folkman, J. (1992) Interferon alpha-2a therapy for life-threatening hemangiomas of infancy. N Engl J Med. 326:1456-63.

Ferriani, R.A., Charnock-Jones D.S., Prentice, A., et al. (1993) Immunohistochemical localization of acidic and basic fibroblast growth factors in normal human endometrium and endometriosis and the detection of their mRNA by polymerase chain reaction. Hum Reprod. 8:11-16.

Folkman J. and Klagsbrun M. (1987) Angiogenic factors. Science, 235:442-4.

Gospodarowicz, D. (1983) Growth factors and their action in vivo in in vitro. J. Pathol. 141:201-33.

Haimovici F. and Anderson D.J. (1993) Effects of growth factors and growth factor-extracellular matrix interactions on mouse trophoblast outgrowth in vitro. Biol. Reprod. 49:124-30.

Hanneken, A., Ying, W., Ling, N., et al. (1994) Identification of soluble forms of the fibroblast growth factor receptor in blood. Proc. Natl Acad Sci USA. 91:9170-4.

Hanneken, A. and Baird, A. (1995) Soluble forms of the high-affinity fibroblast growth factor receptor in human vitreous fluid. Invest. Opthalmol. Vis. Sci. 36:1192-6.

Hourihan, H.M., Sheppard, E.M., Belsey, I.A., et al. (1991) Endometrial vascular features prior to and following exposure to levonorgestrel. Contraception 43:375-85.

Irwin J.C., Utian, W.H., and Eckert R.L. (1991) Sex steroids and growth factors differentially regulate the growth and differentiation of cultured human endometrial stromal cells. Endocrinology 129:2385-92.

Isacchi, A. Bergonzoni, L., and Sarmientos, P. (1990) Complete sequence of a human receptor for acidic and basic fibroblast growth factors. Nucleic Acids Res 18:1906.

Johnson, D.E., Lu, J. Chen, H., et al. (1991) The human fibroblast growth factor receptor genes: a common structural arrangement underlies the mechanism for generating receptor forms that differ in their third immunoglobulin domain. Mol. Cell.Biol. 9:4627-34.

Klein, S., Giancott, M., Presta, M., et al. (1993) Basic fibroblast growth factor modulates integrin expression in microvascular endothelial cells. Mol. Biol. Cell. 4:973-82.

Lappi, D.A. and Baird, A. (1991) Mitotoxins: growth factors-targeted cyctotoxic molecules. Prog Growth Factor Res. 2:223-36.

Lee, P.L, Johnson, D.E., Cousens, L.S., et al. (1989) Purification and complementary DNA cloning of a receptor for basic fibroblast growth factor. Science 245:57-60.

Lessey, B.A., Castelbaum, A.J., Buck, C.A., et al. (1994) Further characterization of endometrial integrins during the menstrual cycle and in pregnancy. Fertil Steril. 62:497-506.

Lessey, B.A., Damjanovich, L., Coutifaris, C., et al. (1992) Integrin adhesion molecules in the human endometrium. Correlation with the normal and abnormal menstrual cycle. J Clin Invest. 90:188-95.

Mangrulkar, R.S., Ono, M., Ishikawa, M., et al. (1995) Isolation and characterization of heparin-binding growth factors in human leiomyomas and normal myometrium. Biol Reprod 53:636-46.

Mukko, V.R., and Stancel, G.M. (1985) Regulation of epidermal growth factor receptor in human fetal ovary and uterus. J. Biol Chem. 260:9820-4.

Nowak, R.A., Rein, M.S., Heffner, L.G., et al. (1993) Production of prolactin by smooth muscle cells cultured from human uterine fibroid tumours. J Clin Endorinol Metab. 5:1308-13

Noyes, R.W., Hertig,A.T., and Rock, J. (1950) Dating the endometrial biopsy. Fertil Steril. 1:3-25.

Oleszask, E. (1988) Inhibition of mitogenic activity of PDGF, EGF, and FGGF by interferon-gamma. Exp Cell Res. 179:575-80.

Presta, M. (1988) Sex hormones modulate the synthesis of basic fibroblast growth factor in human endometrial adenocarcinoma cells: implication for the neovascularization of normal and neoplastic endometium. J Cell Physiol. 137:593-7.

Real, F.X., Oettgen, H.F., and Krown, S.E. (1986) Kaposi's sarcoma and the acquired immunodeficiency syndrome:treatment with high and low doses of recombinant leukocyte A interferon. J Clin Oncol. 4:544-51.

Rider, V., Piva, M., Cohen, M., et al. (1995) Alternative splicing and differential targeting of fibroblast growth factor receptor 1 in the pregnant rat uterus. Endocrinology 136:3137-45.

Rogers, P.A., Au, C.L., and Affandi, B. (1993) Endometrial microvascular density during the normal menstrual cycle and following exposure to long-term levonorgestrel. Hum Reprod. 8:1396-1404.

Rusnati, M., Casarotti, G., Pecorelli, S., et al. (1990) Basic fibroblast growth factor in ovulatory cycle and postmenopausal human endometrium. Growth Factors 3:299-307.

Salat-Baroux J., Romain S., Alvarez S., et al. (1994) Biochemical and immunohistochemical multiparametric analysis of steroid receptors and growth factor receptors in human normal enndometrium in spontaneous cycles and after the induction of ovulation. Hum Reprod. 9:200-8.

Stewart, E.A. and Nowak, R.A. (1996) Leiomyoma-related bleeding: a classic hypothesis updated for the molecular era. Hum Reprod Update 2:295-306.

White, C.W., Sondheimer, H.M., Crouch, E.C., et al. (1989) Treatment of pulmonary hemangiomatosis with recombinant interferon alpha-2a. N Engl J Med 320:1197-200.

Yeh, J., Osathanondh, R., and Villa-Komaroff L. (1993) Expression of messenger ribonucleic acid for epidermal growth factor and transforming growth factor alpha, in human first-and second trimester fetal ovary and uterus. Am J Obstet Gynecol. 168:1569-73.

Yeh, J. and Osanthanondh R. (1993) Expression of messenger ribonucleic acids encoding for basic fibroblast growth factor (FGF) receptor in human fetal ovary and uterus. J Clin Endorinol Metab. 77:1367-71.