Growth factors produced in the uterine endometrium are considered to be involved in the proliferation of the mouse uterine stromal cells induced by estradiol-17β (E2) and progesterone (P). The effect of epidermal growth factor (EGF) and transforming growth factor-α (TGF-α), one of EGF-related growth factors, on the proliferation of mouse uterine stromal cells was studied in a serum-free culture. The growth of the uterine stromal cells was measured by MTT assay. EGF was found to increase the number of uterine stromal cells in a dose-dependent manner. The DNA-replicating cells were investigated using the immunocytochemical detection of bromodeoxyuridine (BrdU)-labeled cells. EGF and TGF-α increased the percentage of BrdU-labeled cells in a dose-dependent manner. Administration of the combination of E2 (10−9 M) and P (10−7 M) for 2 days increased the percentage of BrdU-labeled cells 2.3-fold. The stimulatory effect of EGF, TGF-α and the combination of E2 and P on DNA replication in the uterine stromal cells was repressed by RG-13022 (10−5 M, the inhibitor of the EGF receptor tyrosine kinase). RT-PCR analysis of EGF-receptor-, TGF-α-, and EGF-mRNA was carried out in the cultured uterine stromal cells, and revealed the expression of those mRNAs. These data supported the hypothesis that uterine endometrial stromal growth induced by sex steroids required the EGF family of ligands such as EGF and TGF-α, both produced in the stromal cells, acting for DNA synthesis through EGF receptors.
The mammalian uterine endometrium consists of epithelial cells and stromal cells. The proliferation and differentiation of both the uterine endometrial epithelial and stromal cells are regulated by estrogen and progestin. Estrogen stimulates the proliferation of luminal and glandular epithelial cells in the uterus, whereas the proliferation of stromal cells is induced by a combined treatment of estradiol and progesterone (Huet-Hudson et al., 1989). Much evidence has suggested that the steroid hormone-induced growth of uterine endometrial cells is mediated by an autocrine or paracrine action of polypeptide growth factors synthesized within uterine tissues (Cooke et al., 1998). Several reports have indicated that estrogen-induced proliferation of uterine epithelial cells is mediated by epidermal growth factor (EGF) (Iguchi et al., 1985; Tomooka et al., 1986; Nelson et al., 1991; Uchima et al., 1991; Shiraga et al., 2000), while other reports have shown that transforming growth factor-α (TGF-α, one of the EGF-related growth factors) (Nelson et al., 1992), insulin-like growth factor-I (IGF-I) (Murphy et al., 1987; Shiraga et al., 1997) and heparin-binding epidermal growth factor-like growth factor (HB-EGF) (Zhang et al., 1994) are implicated in the proliferation of the uterine epithelial cells. As reported above, the regulatory mechanism of the proliferation of uterine epithelial cells has been extensively studied. However, the molecular mechanism that regulates the sex steroid hormone-induced proliferation of uterine stromal cells is not clear.
Immunoreactive EGF has been localized in the neonatal and adult mouse uterine stromal cells (Falck and Forsberg, 1996). TGF-α and its mRNA are also detected in both uterine epithelial cells and stromal cells (Tamada et al., 1991). Both EGF and TGF-α are known to bind to EGF receptors. EGF-receptor mRNA has been detected in the mouse uterine stromal cells (Das et al., 1994; Tong et al., 1996). These findings imply that EGF and EGF-related growth factors act on the uterine stromal cells in a paracrine manner. Therefore, the present study was aimed at clarifying the involvement of EGF and TGF-α in the estradiol and progesterone-induced proliferation of uterine stromal cells, since in our culture system, treatment with a combination of estradiol and progesterone stimulated the proliferation of the mouse uterine stromal cells.
MATERIALS AND METHODS
Twenty one- to twenty three-day-old female mice of the ICR strain (CLEA Japan Inc., Osaka, Japan) were used in the present study. They were housed in a temperature-controlled animal room, and were given a commercial diet of CA-1 (CLEA Japan Inc. Osaka, Japan) and tap water ad libitum. All animal experiments were performed in accordance with the Guidelines for Animal Experimentation, Faculty of Science, Okayama University, Japan.
Uterine stromal cell isolation
Uterine stromal cells were isolated according to the method of Ross et al. (1993). Briefly, uterine horns were removed from the mice and longitudinally cut to expose the endometrial luminal surface. The tissue fragments were treated with 0.1% trypsin (w/v, Type III, Sigma Chemical Co., St. Louis, MO, USA) at 4°C for 1 hr, and then at 37°C for 55 min. Trypsin was inactivated by 0.1% soybean trypsin inhibitor (STI, w/v, Sigma Chemical Co.) at 37°C for 10 min. The epithelial tissues were separated from the stromal tissues with forceps under a stereoscopic microscope. The endometrial stromal and myometrial tissues separated from the epithelial tissues were incubated with 0.1% trypsin and 1 mM EDTA at 37°C for 1 hr. Then they were treated with 0.1% STI at 37°C for 10 min. After gentle pipeting and DNase I (Sigma Chemical Co.) treatment, the stromal cells separated from the myometrium were collected by centrifugation. The cell viability was assessed by trypan blue dye exclusion test and was usually more than 95% in each study. All chemicals were dissolved in Ca2+- and Mg2+-free Hanks' solution containing 20 mM HEPES and 0.3% bovine serum albumin (BSA, w/v, Fraction V, Sigma Chemical Co.).
Uterine stromal cell culture
The isolated uterine stromal cells were cultured in a 1:1 mixture of Dulbecco's modified Eagle's medium and Ham's F12 medium without phenol red (DME/F12 medium, Sigma Chemical Co.) containing 2% dextran-coated charcoal-treated fetal bovine serum (v/v, Gibco BRL, Gaithersburg, MD, USA) (DC-FBS) (Lagace et al., 1980) for 1 day. After a 1-day pre-culture, the cells were cultured in DME/F12 medium supplemented with BSA (1 g/l), hydrocortisone (100 μg/l), triiodothyronine (400 ng/l), transferrin (10 mg/l), glucagon (10 ng/l), parathormone (200 ng/l), sodium selenite (5 μg/l) and insulin (100 μg/l). Insulin was stocked at a concentration of 10 μg/ml in sterile 1% acetic acid solution. E2 (Sigma Chemical Co.), P (Sigma Chemical Co.) and RG-13022 (BIOMOL Research Laboratories, Inc., PA, USA), a specific inhibitor of tyrosine kinase of EGF receptor (Yoneda et al., 1991), were initially dissolved in a small amount of sterile ethanol. They were all diluted with the culture medium, and the final ethanol concentrations were less than 0.01%. Mouse EGF was obtained from Sigma Chemical Co., and human TGF-α was obtained from Amersham Pharmacia (Uppsala, Sweden).
The isolated uterine stromal cells were seeded in 24-well tissue culture plates (Becton Dickinson, Lincoln Park, NJ, USA) at a density of 1–2×104 cells/ml per well for the determination of the cell number and the DNA-replicating cells, and in 9-cm tissue culture plates (Becton Dickinson) at a density of 2×106 cells per plate for RNA extraction. The culture was continued at 37°C in humidified atmosphere of 5% CO2 and 95% air.
Immunocytochemical analysis of uterine stromal cells
Uterine stromal cells were cultured on a poly-L-lysine-coated slide glass. The cells were fixed by pre-chilled 100% methanol (−20°C) for 20 min and dried at room temperature. For vimentin or cytokeratin detection, the cells were treated with 0.01% trypsin in 0.01 M PBS for 5 min at 37°C, then incubated with mouse anti-vimentin antibody (Sigma Chemical Co.) or pkk-1 mouse anti-cytokeratin antibody (Organon Teknika, NC, USA) overnight at 4°C. Vimentin-, and cytokeratin-immunoreactivity were visualized using ABC kit (Vector Laboratories, Burlingame, CA, USA).
Determination of the number of uterine stromal cells
The number of cultured stromal cells was determined by a modification of the tetrazolium assay using 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT, Chemical Co.) as a substrate (MTT assay; Alley et al., 1988). The amount of formazan accumulated after 4-hr incubation at 37°C was spectrophotometrically determined by measuring absorbance at 540 nm and was used to estimate the number of the cells in a well. The amount of formazan (the optical density observed at 540 nm) was well correlated with the number of cells in culture (Shiraga et al., 1997). The cell number was expressed as a percentage by regarding the minimal cell number as 100%.
Determination of the proliferating cells
DNA-replicating cells were immunocytochemically detected using a cell proliferation kit (Amersham Pharmacia). Uterine stromal cells were seeded on a poly-L-lysine-coated slide glass. Bromodeoxyuridine (BrdU, 3 mg/ml) was added to the culture medium, and the culture was continued for 5 hr. About one thousand stromal cells were observed from each slide. The percentage of BrdU-labeled cells to total cells was calculated.
Primers of EGF receptor, TGF-α and EGF were designed based upon the mouse sequence and were obtained from the Gibco BRL Custom Primers, Life Technologies Asia Pacific (Yokohama, Japan). The primer sequences and the location on the nucleotide sequences were as follows: EGF receptor 5′ (GTGTGAAGAAGTGCCCCCGAAAC, 931–953), EGF receptor 3′ (AACGACCGCCAAAGAAAACTGACC, 1359–1382), TGF-α 5′ (AGCCAGAAGAAGCAAGCCATCACT, 367–390), TGF-α 3′ (CTCATTCTCGGTGTGGGTTAGCAA, 845–868). EGF 5′(AGATGAGTGTGTGCTGGCTAGATC, 2987–3010), and EGF 3′ (TCCGAGTCCTGTAGTAGTAAGTCC, 3529–3552).
cDNA synthesis and PCR amplification
Total RNA (2 μg) in a final volume of 20 μl was subjected to reverse transcription (RT) using Superscript II Reverse Transcriptase (Gibco-BRL) with random primers according to the manufacturer's instructions. Each PCR was performed using primer pairs specific for EGF-R, EGF and TGF-α as described above. PCR was performed using Takara Taq (Takara Shuzo Co.) and a thermal cycler (Gene Amp PCR System 9700, Applied Biosystems, Branch-burg, NJ, USA). The conditions for PCR were as follows: after activation of DNA polymerase by a 20-sec incubation at 95°C, 35 cycles (for EGF-R cDNA) and 40 cycles (for EGF and TGF-α cDNA) of reactions, including denaturation for 30 sec at 95°C and extension for 1 min at 60°C, followed by an additional extension for 10 min at 60°C. A 10-μl aliquot of each reaction was electrophoresed on 2% agarose gel, stained with ethidium bromide, and photographed under ultraviolet rays.
The size of the resulting DNA fragment following PCR amplification was expected to be 452 base pairs (bp) for EGF-R, 502 bp for TGF-α and 566 bp for EGF. The cDNA fragments were sub-cloned into pGEM3zf (+), and sequenced using fluorescent primers and an automated DNA sequencer.
Immunocytochemical analysis of cultured uterine stromal cells
Uterine stromal cells were attached to a substratum within a few hours after seeding. Phase contrast microscopy revealed that the stromal cells in the culture were flattened, and were mostly fibroblast-like in appearance (Fig. 1A). Immunocytochemical study showed that the cultured cells expressed vimentin (a marker for stromal cells) (Fig. 1B), but not cytokeratin (a marker for epithelial cells) (Fig. 1C). These observations indicate that the cultured cells consisted of uterine endometrial stromal cells.
The growth of uterine stromal cells
After a 1-day pre-culture in the medium containing 2% DC-FBS, the uterine stromal cells were cultured in the serum-free medium. Under the serum-free conditions, the uterine stromal cells continued to grow until Day 5 in culture, and the number increased 4.5-fold (Fig. 2).
Effect of RG-13022 on the incorporation of BrdU to the nucleus of uterine stromal cells
The nucleus of DNA-replicating stromal cells was labeled with BrdU, which was detected immunocytochemically. To examine whether EGF receptor-signaling was involved in the growth of the uterine stromal cells, the stromal cells were treated with RG-13022, an inhibitor of the EGF receptor tyrosine kinase, for 2 days, starting on Day 1 in culture (Fig. 3). RG-13022 treatment decreased the percentage of BrdU-labeled stromal cells in a dose-dependent manner, and in the following experiments, 10−5 M RG-13022 was used.
Effect of EGF on the number and DNA replication of uterine stromal cells
Uterine stromal cells were cultured with various concen trations of EGF (0.1 to 10 ng/ml) for 6 days. EGF increased the number of uterine stromal cells in a dose-dependent manner (Fig. 4). When uterine stromal cells were treated with EGF (10 ng/ml) from Day 5 in culture, they had increased 1.3-fold in the number over the control level at Day 7 in culture. This increase continued, 1.5-fold at Day 9, and remaining 1.6-fold at Day 11 (Fig. 5). EGF increased the percentage of BrdU-labeled stromal cells in a dose dependent manner (Fig. 6).
Effect of TGF-α on the DNA replication of uterine stromal cells
Uterine stromal cells were cultured with various concentrations of TGF-α (0.1 to 10 ng/ml) for 6 days. TGF-α increased the percentage of BrdU-labeled stromal cells in a dose-dependent manner (Fig. 7).
Effect of RG-13022 on the DNA replication of uterine stromal cells induced by exogenous growth factors and sex steroids
E2 (10−9 M) or P (10−7 M) treatment did not increase the number of BrdU-labeled cells (data not shown), whereas the combination of E2 and P significantly increased the number of BrdU-labeled cells. Similarly, EGF and TGF-α caused the number of BrdU-labeled cells to increase. To clarify how EGF receptor signaling is implicated in mediating the action of E2 and P, the uterine stromal cells were treated with RG-13022 (10−7 M) in the presence of the combination of E2 and P. The RG-13022 treatment nullified the E2 and P-induced DNA replication of the uterine stromal cells as well as the EGF- or TGF-α-induced DNA-replication (Fig. 8).
Detection of EGF-receptor-, TGF-α-, and EGF-mRNAs in cultured uterine stromal cells
RT-PCR was performed to evaluate EGF receptor-, TGF-α-, and EGF-mRNA expression in the cultured uterine stromal cells using primer pairs for the amplification of EGF receptor cDNA (Fig. 9, lane 1), TGF-α cDNA (Fig. 9, lane 3) and EGF cDNA (Fig. 9, lane 5). These amplifications resulted in single bands in the ethidium bromide-stained gels. From the size of the RT-PCR products, it was concluded that the 452-bp product, 502-bp product and 566-bp product were amplified cDNAs of EGF receptors, TGF-α, and EGF, respectively. The sequencing analysis revealed that the amplified cDNA fragments were cDNAs coding parts of the mouse EGF receptor, TGF-α and EGF, respectively (data not shown). To exclude the possibility of genomic contamination, RNA samples were amplified by PCR without RT, and no PCR products were found (Fig. 9, lanes 2, 4 and 6).
The uterine stromal cells continued to grow under the serum-free condition. RG-13022 treatment inhibited the proliferation of the uterine stromal cells even in the absence of EGF and TGF-α under serum-free conditions, which suggests the involvement of EGF-receptor signaling in the growth of uterine stromal cells. In addition, we demonstrated the synthesis of EGF and TGF-α in the uterine stromal cells and their stimulatory effect on DNA replication of uterine stromal cells. From these findings, we concluded that endogenous EGF-related growth factors were involved in the regulatory mechanism of the growth of mouse uterine stromal cells.
In our culture system, the uterine stromal cells ceased to increase in number on Day 5 in culture. However, when the medium was treated with EGF beginning on Day 5, the uterine stromal cells started to increase in number, indicating that the cessation of the growth of uterine stromal cells was partly due to the depletion of endogenous growth factors, probably EGF-related growth factors.
In the adult ovariectomized mouse uterus, the proliferation of stromal cells is induced by a combination of E2 and P (Huet-Hudson et al., 1989). In the present study treatment with E2 or P failed to stimulate DNA replication in the uterine stromal cells, whereas the combination of E2 and P stimulated DNA replication in vitro. The proliferation of the uterine stromal cells induced by the combination of E2 and P in a serum-free culture is quite consistent with the proliferation of uterine stromal cells induced by E2 and P in vivo.
The DNA replication that had been enhanced by the combined treatment with E2 and P was nullified by an RG-13022 treatment. This finding suggests that the action of E2 and P on the DNA replication of uterine stromal cells requires the EGF-receptor signaling. Therefore, EGF-related growth factors are considered to mediate the action of the combination of E2 and P through the EGF receptor. This hypothesis is supported by a previous report that in response to estrogen treatment DNA synthesis in uterine endometrial stromal cells of EGF-receptor knockout mice is depressed compared with that of wild type mice, suggesting that EGF-receptor signaling is required in estrogen-induced proliferation of uterine endometrial stromal cells (Hom et al., 1998).
EGF and EGF mRNA levels are increased by estrogen in the mouse uterus (DiAugustine et al., 1988; Falck and Forsberg, 1996). The TGF-α mRNA level is also increased by estrogen in the adult ovariectomized rat uterus (Borgundvaag et al., 1992) and in the immature mouse uterus (Nelson et al., 1992). Immunoreactive TGF-α protein is secreted at high levels into the mouse uterine luminal fluid after estrogen treatment (Nelson et al., 1992). As mentioned above in vivo analysis of EGF and TGF-α expression in the uterus has been well reported. However, physiological growth factors which are involved in steroid hormone-induced proliferation of the stromal cells remains to be determined. Interestingly, in a study of pituitary growth, we found that TGF-α is a mediator of estrogen-induced proliferation of mammotrophs, since treatment of mouse pituitary cells with antisense TGF-α oligodeoxynucleotide or anti-TGF-α anti-bodies inhibited estrogen-induced proliferation of mammotrophs (Oomizu et al., 2000; Takahashi et al., 2002; Sharma et al., 2003). These results altogether suggest that estrogen-induced growth of various tissues is controlled by a common mechanism of an autocrine or paracrine action of growth factors.
The stromal cells do not proliferate when progesterone receptor antagonists are used to block the receptor function (Cullingford and Pollard, 1988; Rider and Psychoyos, 1994) or when anti-progesterone antibodies are administered during early pregnancy (Rider et al., 1986). Progesterone is essential for the growth of uterine stromal cells. The increase in the progesterone receptor level is evoked by estrogen (Aronica and Katzenellenbogen, 1991). The mechanism of cooperative action between E2 and P needs to be clarified.
In conclusion, we found that estradiol and progesterone co-operatively stimulated the proliferation of uterine stromal cells, and that EGF-receptor ligands were involved in the steroid-induced growth of uterine endometrial stromal cells.
This study was supported in part by Grant-in-Aid for Scientific Research from Japan Society for the Promotion of Science to S.T.