BIOLOGY OF REPRODUCTION 73, 927–934 (2005) Published online before print 29 June 2005. DOI 10.1095/biolreprod.105.039875
Vascular Endothelial Growth Factor (VEGF) Production by the Monkey Corpus Luteum During the Menstrual Cycle: Isoform-Selective Messenger RNA Expression In Vivo and Hypoxia-Regulated Protein Secretion In Vitro1 Marta Tesone,3,5 Richard L. Stouffer,2,5,7 Sherri M. Borman,4,5 Jon D. Hennebold,5,6 and Theodore A. Molskness5 Division of Reproductive Sciences,5 Oregon National Primate Research Center, Oregon Health & Science University, Beaverton, Oregon 97006 Department of Obstetrics & Gynecology6 and Department of Physiology and Pharmacology,7 Oregon Health & Science University, Portland, Oregon 97239 ABSTRACT
INTRODUCTION
Experiments were designed to investigate the expression and regulation of vascular endothelial growth factor (VEGF) in the primate corpus luteum (CL) throughout the luteal life span in the natural menstrual cycle. Corpora lutea were collected during the early (ECL; Days 3–5 post-LH surge), mid (MCL; Day 6– 8 post-LH surge), mid-late (MLCL; Days 10–12 post-LH surge), late (LCL; Days 14–16 post-LH surge), and very late (Days 17– 18 post-LH surge) luteal phase. Specific primers were designed to amplify mRNAs encoding VEGF isoforms 206, 189, 183, 165, 145, and 121. Only two cDNA products were obtained by reverse transcription-polymerase chain reaction (RT-PCR) and rapid amplification of cDNA ends; cloning and sequencing confirmed their 98% homology to the corresponding human VEGF 165 and 121 sequences. Semiquantitative RT-PCR assays indicated that VEGF 165 mRNA levels increased (P , 0.05) from ECL to MLCL but then declined (P , 0.05) by LCL. Although VEGF 121 mRNA levels were limited in ECL, they increased significantly in MCL (P , 0.05). Levels of VEGF protein, as measured by Western blot analysis, were two- to fourfold higher for VEGF 165 versus VEGF 121. Also, VEGF 165 levels were higher (P , 0.05) in ECL and MCL compared to those at later stages. During 2-day culture, preparations of dispersed luteal cells secreted VEGF into the media; the highest levels were observed in ECL and declined (P , 0.05) by LCL. Regardless of luteal stage, hypoxic conditions increased (P , 0.05) VEGF levels, whereas LH exposure increased (P , 0.05) progesterone, but not VEGF, in the media. These results are consistent with a dynamic, local regulation of VEGF production during the life span of the primate CL that is not directly controlled by LH.
The ovarian vasculature, particularly that associated with the dominant structures (preovulatory follicle and corpus luteum [CL]) during the ovarian cycle, is one of the few sites where nonpathological development and regression of blood vessels occur in the adult. Vascular elements form in the theca layer of the antral follicle and, after ovulation, invade the avascular granulosa layer to establish an extensive capillary network that nourishes the developing CL [1, 2]. Conversely, the integrity and function of the luteal vasculature declines during regression of the CL near the end of the cycle [2]. Recently, local factors that act specifically on vascular endothelial or support (pericyte or smooth muscle) cells to control angiogenesis or angiolysis were identified in the growing follicle and CL of several species, including nonhuman primates and women [3, 4]. Moreover, experiments employing antagonist molecules have provided evidence supporting a critical role for these factors, notably vascular endothelial growth factor (VEGF; also referred to as VEGF-A [5]), in folliculogenesis, ovulation, and CL development in rodents [6] and monkeys [7, 8]. Since the discovery of VEGF in 1989 as a 40 000- to 46 000-Mr, disulfide-linked, dimeric glycoprotein, investigators have determined that at least seven isoforms of VEGF displaying different biochemical properties can be generated by alternative splicing from a single gene [5, 9]. The human isoforms include proteins of 206, 189, 183, 165, 148, 145, and 121 amino acids per monomer that differ in their ability to bind to heparin and to specific VEGF receptors/coreceptors (neuropilins). In addition, increasing evidence suggests dynamic expression of different isoforms within tissues [10] and that the different isoforms may promote specific functions in certain vascular beds [11]. Limited evidence indicates hormonal regulation of selected VEGF isoforms. For example, Ancelin et al. [12] reported that stromal cells in the human endometrium express mRNAs for the VEGF 189, 165 and 121 isoforms and that progesterone (P4) exposure selectively increases expression of VEGF 189 mRNA. To date, little detailed information exists regarding the expression or regulation of VEGF isoforms in the ovary, particularly in the CL. Our group [13, 14] and others [15, 16] have provided evidence that the gonadotropic hormone, LH, is a major stimulator of VEGF production by granulosa cells in the ovulatory, luteinizing follicle, but the control of VEGF production in the CL by hormonal or local factors remains unknown. Therefore, experiments were designed,
corpus luteum, corpus luteum function, luteinizing hormone, ovary, progesterone Supported by the National Institute of Child Health and Human Development (NICHD) through cooperative agreement U54-HD18185 as part of the Specialized Cooperative Centers Program in Reproduction Research, NICHD/Fogarty Fellowship D43 TW00668 (M.T.) and RR00163. 2 Correspondence: Richard L. Stouffer, Division of Reproductive Sciences, Oregon National Primate Research Center, OHSU West Campus, 505 NW 185th Ave., Beaverton, OR 97006. e-mail:
[email protected] 3 Current Address: Instituto de Biologia y Medicina Experimental, Obligado 2490, 1428 Buenos Aires, Argentina 4 Current Address: West Valley Fertility Center, Glendale, AZ 85308 1
Received: 11 January 2005. First decision: 10 February 2005. Accepted: 27 June 2005. Q 2005 by the Society for the Study of Reproduction, Inc. ISSN: 0006-3363. http://www.biolreprod.org
927
928
TESONE ET AL.
Cloning and Sequencing of Macaque VEGF mRNAs Primers corresponding to regions in VEGF exon 4 and exon 8 were used in a polymerase chain reaction (PCR) system to amplify VEGF cDNAs for mRNAs encoding isoforms 206, 189, 183, 165, 145, and 121 (Fig. 1). The primer sequences were as follows: sense, 59-TTATGCGGA TCAAACCTCACCAAGGC-39; antisense, 59-TGTGGGTGGGTGTGTC TA-39. The products obtained were cloned into a plasmid-based vector (pGEM; Invitrogen) and transformed into the appropriate bacterial host vector (Escherichia coli strain XL-19 Blue; Stratagene, San Diego, CA). Plasmids were purified from individual clones as described previously [22]. The plasmids were sequenced in the Molecular and Cellular Biology Core facility of the Oregon National Primate Research Center using an ABI 377 automated sequencer, with T7 and SP6 primers (Applied Biosystems, Foster City, CA). The sequence data were analyzed for homology to previously characterized genes deposited in the National Center for Biotechnology Informatics (NCBI) database using the Basic Local Alignment Tool (BLAST) program. FIG. 1. Reported [9] model for alternative splicing of VEGF mRNAs that generates seven VEGF isoforms from 206 to 121 amino acids (a.a.) in length. Gene exons are represented by numbers. Arrows over the 206a.a. isoform indicate the location of sense and antisense primers used to amplify, clone, and sequence six of the seven possible VEGF mRNAs. Arrows over the 165- and 121-a.a. isoforms designate the sequence location of specific primers employed for semiquantitative RT-PCR analysis of these individual mRNAs. See the Materials and Methods for primer sequences.
using the rhesus monkey as a primate model, to determine the following: 1) which VEGF isoforms are expressed by the primate CL, 2) if the VEGF isoforms are selectively expressed during the CL life span in the menstrual cycle, and 3) whether LH and/or hypoxia (the primary stimulant of VEGF production in various vascular beds [17]) increases VEGF production by macaque luteal cells at various stages of the CL life span. MATERIALS AND METHODS
Animal Protocols and Tissues Animal protocols were approved by the Oregon National Primate Research Center Animal Care and Use Committee and were conducted in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals. Daily blood samples were collected by saphenous venipuncture from adult, female rhesus monkeys (Macaca mulatta) exhibiting regular menstrual cycles; samples were collected beginning 6 days after menses and continuing to the time of lutectomy. Serum estradiol and P4 levels were determined using ELISA (Elecsys 2010; Roche Diagnostics, Indianapolis, IN) as described previously [18]. The first day of low (,100 pg/ml) serum estradiol levels following the midcycle estradiol peak corresponds with the day after the LH surge and, therefore, was considered to be Day 1 of the luteal phase [19]. The CL (n 5 4–6 per stage of the luteal phase) was removed from anesthetized rhesus monkeys [20] during the early (3–5 days post-LH surge), mid (6–8 days post-LH surge), mid-late (10–12 days post-LH surge), and late (14–16 days post-LH surge) luteal phase of the menstrual cycle and very late in menstruation (17–18 days post-LH surge). This provided tissues at intervals representing developing, functioning, on the verge of regressing, regressing, and regressed CL, respectively [20]. The individual CL were divided into three portions, two which were immediately frozen in liquid nitrogen and stored at 2808C for isolation of either total RNA or protein; the third portion was processed for other histological studies [18]. Other freshly isolated CL (n 5 4–6 per stage of the luteal phase) also were transferred to the laboratory for preparation of dispersed luteal cells.
RNA Isolation and Reverse Transcription Total RNA from individual CL collected at the different stages of the luteal phase was extracted using Trizol reagent (Invitrogen, Carlsbad, CA) according to the manufacturer’s instructions. The RNA was treated with 1 mg of DNase (Invitrogen). Reverse transcription (RT) was performed for 2 h at 378C in a 20-ml reaction volume using Moloney murine leukemia virus reverse transcriptase (Invitrogen) as described previously [21].
Rapid Amplification of cDNA Ends for VEGF mRNAs Rapid amplification of cDNA ends (RACE) was performed using the Invitrogen GeneRacer cDNA Amplification Kit to identify the 39-end of the VEGF cDNA. Touchdown PCR parameters were according to the manufacturer’s instructions. A cDNA template was created from polyA1 RNA isolated from a pool of CL that included tissues from all stages of the luteal phase. For 39-RACE, the Universal Primer Mix (Invitrogen) was used in combination with the VEGF sense primer corresponding to a region in exon 2 (59-CCATGGCAGAAGGAGGAGGGCAGAATC-39). Products were analyzed by electrophoresis using a 1% agarose gel and visualized by ethidium bromide staining. Products were sequenced and compared to the macaque and human VEGF sequence data as described in the previous section.
PCR Analysis of VEGF 165 and 121 mRNAs Primer sets that specifically amplify cDNA for the detectable VEGF transcripts were used in semiquantitative RT-PCR experiments using peptidylprolyl isomerase A (PPIA; also known as cyclophilin A) mRNA as an internal control. For the VEGF 165 transcript, primers (Fig. 1) were designed at the junctions of exons 5 and 7 (sense, 59-CAAGAAAATCCC TGTGGG-39) and exons 8 and 7 (antisense, 59-TTGTCACATCTGCAA GTA-39). For the VEGF 121 variant (Fig. 2), primers were designed using an area of exon 4 (sense, 59-TTATGCGGATCAAACCTCACCAAGG39) and the junctions of exons 8 and 5 (antisense, 59-CTTGTCACATTT TTCTTGTCT-39). The primer sequences were based on the rhesus macaque VEGF cDNA data provided by the cloning experiments. As positive controls and to prove the specificity of the VEGF primers, the 165 and 121 plasmids obtained in the cloning experiments also were included in the experiments. Validation of the RT-PCR assay was performed as described previously [23]. The PCR was carried out in a 25-ml volume containing an empirically determined amount of the RT reaction, 1 ml of the 10 mM specific primer set, 2.5 ml of 103 strength Taq buffer (BD Biosciences Clontech, Palo Alto, CA), 1 ml of 10 mM deoxy-NTPs, and 0.75 ml of Advantage 2 Taq (BD Biosciences Clontech). The reaction was initiated at 948C for 1.5 min, followed by 948C for 30 sec, 568C for 30 sec, and 728C for 2 min for 35 cycles, and then a final extension at 728C for 5 min. Aliquots of the PCR products were electrophoresed through a 1.4% agarose gel stained with 0.1 mg/ml of ethidium bromide. The PPIA mRNA was used as an internal control, because no apparent changes in macaque CL expression of this gene were observed at any stage of the menstrual cycle [24]. Gels were visualized on an ultraviolet transilluminator and photographed. The gel bands were quantified by densitometry using Quantity One software (Bio-Rad, Hercules, CA); both VEGF and PPIA mRNAs showed a linear response over defined PCR cycles.
Western Blot Analysis of VEGF-A Proteins Proteins were extracted from individual CL; the protein extracts were subjected to electrophoresis on 15% SDS-polyacrylamide gels and transferred onto nitrocellulose membranes as reported previously [25]. Goat polyclonal anti-human VEGF (catalog no. AF-293-NA; R&D Systems, Minneapolis, MN) that recognizes both the 165 and 121 isoforms was used as primary antibody (1:1000, 1 h). Protein bands were visualized by incubating the blots for 1 h with peroxidase-conjugated rabbit anti-goat immunoglobulin G (1:4000 dilution; Zymed Laboratories, San Francisco,
VEGF IN THE PRIMATE CORPUS LUTEUM
929
CA). Negative controls were performed in the absence of the primary antibody. In each experiment, 30 mg of protein were loaded for each CL tested, and the proteins obtained from different CL stages were loaded on the same gel. For quantification, an initial screening was performed on blots with Hyperfilm (Amersham, Piscataway, NJ) using different exposure times to optimize the signal. The levels of protein were analyzed by densitometry [26]; optical density data are expressed as arbitrary units 6 SEM (n 5 4–5 CL per stage of the luteal phase). The density in each VEGF band was normalized to the density of the b-actin band that was used as an internal control. In addition, VEGF proteins were analyzed in a pool of conditioned media (CM) from cultured luteal cells isolated from CL at the early, mid, and late luteal phases (see next section). Proteins were concentrated using a 10 000-Mr cutoff membrane (Microcon Centrifugal Filter Devices, Millipore, Burlington, MA). Western blot analysis was performed as described above.
Isolation and Serum-Free Culture of Luteal Cells Collagenase-dispersed luteal cells were prepared from CL obtained at the early, mid, and late luteal phases as described previously [27] and placed in fibronectin (Sigma Chemical Co., St. Louis, MO)-coated, 96well plates (40 000 cells/well; Costar; Corning, Inc., Corning, NY) containing Dulbecco modified Eagle medium/F-12 medium supplemented with insulin (5 mg/ml), transferrin (5 mg/ml), selenium (5 ng/ml), aprotinin (10 mg/ml), and low-density lipoprotein (25 mg/ml; all from Sigma) at 20% O2, 5% CO2 in air at 378C [13]. After 24 h, cell cultures were transferred to atmospheres of either 20%, 5%, or 0% O2 in control medium with or without 100 ng/ml of human LH (AFP-4261A; Pituitary Network Assoc., Thousand Oaks, CA) or 100 mM CoCl2 (Sigma) as described previously in studies of macaque granulosa cells [14]. After 48-h exposure to various O2 atmospheres and treatments, media were collected and assayed for VEGF (Quantikine Human VEGF ELISA; R&D Systems; validated for macaque VEGF [13]) and for P4 (Elecsys 2010 [18]) as described above. In all experiments, cell numbers (DNA content) in wells were determined 48 h after treatment using the crystal violet assay [14].
Statistical Analysis Statistical tests were performed using the SigmaStat software package (SPSS, Chicago, IL). Differences in VEGF mRNA and protein levels between stages of the luteal phase were analyzed using one-factor ANOVA followed by the Student-Neuman-Keuls method or Dunn test. Differences in VEGF and P4 levels between treatment groups within a stage of the luteal phase were determined by one-way ANOVA using a complete randomized block design followed by multiple-range tests to determine differences. Differences were considered to be significant at P , 0.05.
RESULTS
PCR, Cloning, and Sequencing of VEGF mRNA Isoforms in Macaque CL
When cDNAs from macaque CL at all stages (early through very late luteal phase) were amplified using VEGF primers that hybridize with the transcripts that encode six of the seven isomers (VEGF 206, 189, 183, 165, 145, and 121) (Fig. 1), all the PCR products corresponded to the mRNAs for VEGF 165 or 121. The products were inserted into a plasmid-based vector and transfected into the appropriate bacterial host. The resulting plasmids from individual bacteria clones were sequenced, and from nine plasmids, either the partial cDNA for VEGF 165 (n 5 5) or VEGF 121 (n 5 4) was identified. Both isoforms possessed 98% identity with the corresponding cDNAs for the human VEGF isoforms. Subsequently, the RACE technique was employed to determine whether the mRNA encoding VEGF 148 isoform (which is missing exon 8) (Fig. 1) is also expressed in the macaque CL and to confirm the RT-PCR results. Again, the cDNA products evident following agarose gel electrophoresis were limited to two bands, corresponding to VEGF 165 and 121 (data not shown).
FIG. 2. Levels of VEGF 165 and 121 mRNA, as determined by semiquantitative RT-PCR, in macaque CL collected during the early, mid, midlate, late, and very late luteal phase of the natural menstrual cycle. Values are presented as the mean 6 SEM (n 5 3–5 per luteal stage) relative to cyclophilin mRNA. Different superscripts indicate significant (P , 0.05) differences within an isoform between stages of the luteal phase. Representative ethidium bromide-stained gels (below the graph) illustrate VEGF and PPIA (cyclophilin A [cyclo]) mRNA levels in the CL at the different stages of the luteal phase. As positive controls, the plasmids containing cloned macaque VEGF 165 and 121 partial cDNAs are also represented.
Semiquantitative RT-PCR Analysis of VEGF 165 and 121 mRNAs
Figure 2 summarizes the levels of VEGF mRNAs in macaque CL at different stages of the luteal phase during the natural menstrual cycle. Levels of VEGF 165 mRNA were appreciable at the early stage, increased 2.3-fold (P , 0.05) by the mid-late stage, and then declined (P , 0.05) in the later stages. Levels of VEGF 121 mRNA were limited in the early stage, but they increased significantly (6.6-fold) by the midluteal phase, peaked at the mid-late luteal phase, and then declined (P , 0.05) by the very late luteal phase. The levels of VEGF 165 and 121 mRNAs correlated positively (r 5 0.51, P , 0.02 and r 5 0.50, P , 0.03, respectively) with serum P4 levels on the day of CL collection (data not shown). Although the levels of VEGF mRNAs were expressed relative to PPIA mRNA, the ratio of VEGF 165 to VEGF 121 mRNAs can be depicted (Fig. 3) throughout the luteal phase. A 35-fold greater expression of VEGF 165 mRNA was observed in the early luteal phase, but the ratio was variable. Nevertheless, the ratio declined (P , 0.05) fivefold by the midluteal phase, and it remained between 5 and 10 in favor of VEGF 165 mRNA through the later stages. Western Blot Analysis of VEGF 165 and 121 Proteins
Western blot analysis (Fig. 4) resulted in two distinct bands of the expected sizes for VEGF 165 and 121. When normalized to b-actin, a protein for which content was invariant during the CL life span, VEGF 165 levels were twoto fourfold higher than those of VEGF 121 in CL throughout the luteal phase. Levels of VEGF 165 were significantly higher in the early luteal phase, declined at the midluteal stage, and were lowest from mid-late luteal phase onward to menstruation. A similar pattern was observed for VEGF
930
TESONE ET AL.
FIG. 3. The ratio of VEGF 165 and VEGF 121 mRNA in macaque CL at the various stages of the luteal phase. Values are presented as the mean 6 SEM (n 5 3–5 per luteal stage) as derived from the data in Figure 2. An asterisk denotes a significant (P , 0.05) difference between the early luteal phase and all other stages.
121, except that levels did not decline between the early and mid stages. A single, broad band, spanning the size range of VEGF 165 to 121, was observed in CM (Fig. 4) from cultures of macaque luteal cells. VEGF Production by Macaque Luteal Cells In Vitro: Effects of LH and Hypoxia
After an initial plating interval, dispersed cell preparations from CL produced readily detectable levels of VEGF during 48 h of culture in control (serum-free, 20% O2) conditions (Table 1). However, levels appeared to decline as the luteal phase advanced, with luteal cells from the late stage producing lower (P , 0.05) levels of VEGF than cells from the early luteal phase. Notably, the presence of a receptor-saturating concentration of LH did not increase VEGF levels in cultures of luteal cells from any stage. In contrast, exposure to hypoxic (0% O2) conditions significantly increased VEGF levels in cultures of luteal cells from CL at the early, mid, or late luteal phase. In comparison, P4 levels (Table 1) also declined as the luteal phase advanced, with luteal cells from the midluteal stage producing less (P , 0.05) than those from the early stage. Exposure to LH increased (P , 0.05) P4 levels by four- to fivefold in cultures of cells from the mid and late luteal phase, but not in those from the early stage. In con-
FIG. 4. Levels of VEGF 165 and 121 proteins, as determined by Western blot analysis, in macaque CL collected at various stages of the luteal phase during the menstrual cycle. Values are presented as the mean 6 SEM (n 5 4–5 per luteal stage) normalized to b-actin. Different superscripts within each isoform indicate differences (P , 0.05) between luteal stages. Representative autoradiograms (below the graph) illustrate VEGF and b-actin levels at the different stages. In addition, results from CM pooled from cultured cells obtained at the early, mid, and late luteal phase are represented.
trast, hypoxic conditions suppressed (P , 0.05) P4 levels in control cultures of cells from the early luteal phase, but not in those from other stages. Figure 5 summarizes VEGF (Fig. 5A) and P4 (Fig. 5B) levels in cell cultures prepared from the CL at the late luteal phase and incubated in various O2 concentrations or with CoCl2. Pharmacologic hypoxia because of cobalt exposure increased (P , 0.05) VEGF levels in the presence of 20% or 5% O2 concentrations. Also, reducing the oxygen tension increased VEGF levels; at 0% O2, levels were fivefold higher (P , 0.05) than those at 20% O2. Levels of VEGF in the presence of Co12 or 0% O2 were comparable. Similar results were obtained with luteal cells from the early and midluteal phases (not shown), but the 5% O2 concentration significantly increased (P , 0.05) VEGF levels. Addition of LH, either alone or to any of the above treatment groups, did not alter VEGF levels in cell cultures from the late (Fig. 5A), mid (not shown), or early (not shown) luteal phase. Cobalt exposure significantly decreased P4 levels in cell cultures prepared from CL at the late luteal phase in the presence of LH (Fig. 5B) at 20% or 5% O2 concentrations. However, cobalt did not alter P4 levels in the absence of LH (control vs. CoCl2). Similar results were observed on reducing the oxygen concentration. Although 5% O2 had little effect, 0% O2 reduced (P , 0.05) LH-stimulated P4 levels compared to those in the 20% O2 environment. Similar results were obtained with cells from the early and mid-
TABLE 1. Effects of LH and hypoxia (0% O2) on in vitro production of VEGF and progesterone by cells dispersed from the macaque CL at various stages of the luteal phase. Stage of luteal phase Early Mid Late
VEGF (pg/ml)a
Progesterone (ng/ml)a
Control
LH
0% O2
Control
LH
0% O2
596 6 273 366 6 117 274 6 65*
683 6 310 365 6 99 273 6 69
1523 6 640** 866 6 242** 1283 6 305**
432 6 207 65 6 18* 22 6 10*
534 6 205 275 6 65** 121 6 85**
165 6 54** 53 6 21 12 6 5
a Per 4 3 104 cells per well for 2 days; mean 6 SEM; n 5 4–6 per group. * P , 0.05 (ANOVA), compared to early luteal phase. ** P , 0.05 (ANOVA, complete block design), compared to control at each stage.
VEGF IN THE PRIMATE CORPUS LUTEUM
931
luteal phase (not shown), except that Co12 or 0% O2 elicited a more pronounced (P , 0.05) suppression of P4 levels in cultures not incubated with LH (Table 1). The DNA analyses of cell cultures from all stages of the luteal phase, after 48-h exposure to the various conditions (20% to 0% oxygen, CoCl2) and with or without LH, revealed no significant differences among treatment groups (data not shown). DISCUSSION The present study is, to our knowledge, the first to demonstrate the selective expression of alternative-spliced mRNA and protein isoforms of VEGF in the CL throughout the specific stages of the luteal life span, notably during the menstrual cycle in a nonhuman primate model. Alternative splicing is an important mechanism for creating related proteins with potentially different bioactivity, distribution, or stability. Transcripts that give rise to the individual VEGF isoforms differ in the presence or absence of exons 6 and 7, which influence their acidity and affinity for binding sites [5, 9]. Using pools of macaque CL from all stages of the luteal phase of the natural menstrual cycle, RT-PCR and RACE techniques were only able to detect expression of mRNAs encoding the VEGF mRNA corresponding to isoforms 121 and 165. The smallest known isoform, VEGF 121 lacks both exon 6 and exon 7 and, consequently, is a weakly acidic protein that does not bind to heparin or heparin-sulfated glycoproteins (HSPGs). In contrast, the VEGF 165 mRNA includes exon 7, resulting in a more basic molecule with moderate affinity for heparin. It is believed that VEGF 121 generally is released and diffuses away from its cells of origin, whereas as much as 75% of VEGF 165 remains associated with the cell surface or nearby, depending on the HSPG composition of the extracellular matrix [5, 9]. The current findings are consistent with reports that VEGF 121 and 165 mRNAs were the predominant transcripts detected in whole ovaries and luteinizing granulosa cells of rats 1–4 h after administering an ovulatory hCG bolus [15] and in granulosa cells [28] and CL [29, 30] from sheep and cows. Examples can be cited in which VEGF 121 is the primary isoform expressed in cells/tissues [31]. However, evidence that VEGF 165 mRNA and protein levels typically were greater than those for VEGF 121 in the macaque CL is similar to that from the study by Redmer et al. [30], in which the latter represented about one-third of VEGF expression in the ovine CL. Notably, VEGF 145 mRNA was not detected in the CL, despite evidence that this isoform is produced by human luteinizing granulosa cells [16], reproductive tract carcinomas [32], the placenta [9], and the blastocyst [33]. Also, despite reports of weak expression of VEGF 189 mRNA in the CL of domestic animals [29, 30], we were unable to detect a signal by RTPCR from primate luteal tissue. Because evidence exists that P4 selectively increases the expression of VEGF 189 mRNA in the human uterus [12] and several progestin-regulated processes were recently discovered in the primate CL [34], we hypothesized that VEGF 189 would be dynamically expressed in macaque luteal tissue—but this does not appear to be the case in the macaque (present study) or human [16]. Rather, it appears that the larger, basic VEGF 189 and 209 isoforms that are not readily secreted but are completely sequestered at the cell surface and ECM [5, 9] are not major products of the CL at any stage of the menstrual cycle. Collectively, the data suggest that VEGF 121 and 165
FIG. 5. Levels of VEGF (A) and P4 (B) in cultures of dispersed cells obtained from CL in the late luteal phase following 2 days of incubation at various O2 concentrations (20%, 5%, and 0%) and in the presence or absence of 100 mM CoCl2 or 100 ng/ml of human LH. Values are presented as the mean 6 SEM (n 5 4 experiments). Lowercase letters (a–c) denote differences (P , 0.05; ANOVA using complete block design) between treatment groups within an O2 concentration. Asterisks (* 5% vs 20%; ** 0% vs 5% or 20%) denote differences within a treatment across O2 concentrations.
constitute the bioactive forms within the primate CL. Both VEGF isoforms interact with the two VEGF receptors, FLT1 (also known as VEGFR-1) and KDR (also known as VEGFR-2) [5, 9]. However, recent evidence suggests that additional molecules on the endothelial cell surface, notably the neuropilin (NRP) 1 and NRP2, can serve as coreceptors for VEGF 165, thereby selectively enhancing the binding and actions of this isoform [35, 36]. Both NRP1 and NRP2 were recently detected in the rat uterus [37], and we have evidence for their presence in the macaque CL (unpublished data). Our findings that VEGF 165 expression and protein levels were greater than those for VEGF 121 suggests that this VEGF isoform is the major regulator of angiogenesis and vascular function during the life span of the CL. Nevertheless, further studies are needed to evaluate the actions of VEGF 165 versus those of VEGF 121 in luteal
932
TESONE ET AL.
tissue; recent experiments on mice expressing single VEGF isoforms (VEGF120/120, VEGF164/164, and VEGF188/188 mice [11, 38]) demonstrate overlapping as well as isoform-specific functions in such diverse processes as retinal vascular development and bone development [5]. Whether the higher ratio of VEGF 165 to VEGF 121 in macaque CL during development in the early luteal phase is biologically significant remains unknown. The current data extend our previous observation [23] of dynamic VEGF expression in the macaque CL during luteal life span by establishing that the rise in mRNA levels from the early to mid-late luteal phase results from transcription of the VEGF 165 and 121 mRNAs. Previous investigators have not agreed on whether VEGF mRNA levels change or remain constant in the CL during the ovarian cycle in primate species [23, 39]. Nevertheless, our experiments suggest that such patterns of RNA expression do not necessarily portray the pattern of VEGF protein produced within the CL (Fig. 4) or secreted by luteal cells (Table 1) during the luteal life span. Under control conditions, VEGF levels produced by dispersed cells from the macaque CL decreased as a function of age in the menstrual cycle (i.e., from the early through late luteal phase). Although these in vitro studies employed short-term culture conditions, it is possible that cell activity does not reflect the in vivo situation. However, the pattern of declining VEGF production during the luteal phase is consistent with earlier evidence that immunocytochemical staining for VEGF in the macaque CL was intense during the early luteal phase and diminished by the later stages before menses [23]. Thus, these data, as well as our earlier findings of divergence in VEGF mRNA and protein levels in the preand periovulatory follicle [40], support the concept that the level of VEGF produced by luteinizing granulosa cells and luteal cells in the primate ovary is regulated at both the transcriptional and posttranscriptional levels. To our knowledge, the present study is the first examination of VEGF production by luteal cells in vitro, and the evidence suggests that following luteinization, the cells in the CL do not respond acutely (24–48 h) to gonadotropin with enhanced VEGF secretion. The observation that dispersed cells from the macaque CL did not respond to the luteotropic hormone LH with enhanced VEGF production was unexpected. It generally is believed that the midcycle LH surge promotes VEGF production by luteinizing granulosa cells in many [15, 16], but not all [41, 42], species, and we have reported previously that exposure to LH or chorionic gonadotropin increased VEGF production by nonluteinized granulosa cells from macaque preovulatory follicles by one to two orders of magnitude [13]. Evidence from primate species, wherein LH/chorionic gonadotropin are the primary luteotropic hormones, suggests that LH withdrawal (via administration of a GnRH antagonist [43]) or exposure to chorionic gonadotropin (mimicking CL rescue in early pregnancy [39]) decreased or increased, respectively, VEGF mRNA expression by the CL in vivo. The lack of an effect of LH on VEGF production was not a result of the luteal cell’s general insensitivity to LH (e.g., an artifact of enzymatic tissue dispersion), because P4 production increased by four- to fivefold in the presence of LH after the early (post-LH surge) interval. However, we cannot rule out specific lesions in LH-sensitive VEGF production because of cell dispersion or culture. Alternatively, the divergence with the results of in vivo studies may relate to the sampling interval; notably, changes in VEGF expression were noted after 3 days, but not after 1 or 2 days, of GnRH
antagonist treatment [43] and after 6–8 days of exposure to chorionic gonadotropin [39]. Thus, the gonadotropin effects in vivo may be indirect, resulting from other actions in the days following LH depletion or chorionic gonadotropin repletion. Our data are consistent with a model of hypoxic stress [17] stimulating VEGF production within the CL, either to promote angiogenesis or to maintain the microvasculature in developing or functional luteal tissue, respectively. Reduced O2 tension or pharmacologic (e.g., cobalt) hypoxia increases VEGF mRNA or protein expression in a number of cell types or tissues, including reproductive organs, such as the uterus [44]. The discovery that the genes for VEGF and its receptor KDR (VEGF-R1) contain a hypoxia-response element that binds the hypoxia-inducible transcription factor HIF-1 [45] established a fundamental link between oxygen availability and VEGF expression/action in physiologic (e.g., primate menstruation and endometrial repair) and pathologic (e.g., tumor growth and retinopathy) angiogenesis [46]. However, we recently reported [14] that macaque granulosa cells isolated from preovulatory follicles either before or 27 h after exposure to an ovulatory gonadotropin (hCG) stimulus did not respond in vitro to reduced O2 tension or cobalt with increased VEGF production. We speculated that in the developing primate antral follicle, VEGF production by mural granulosa cells was low because of a lack of response to hypoxia and was regulated primarily by the midcycle gonadotropin surge. Collectively, the data suggest that following the action by LH at midcycle, the local O2 milieu is a primary regulator of VEGF production by luteinizing [47] or luteinized (present study) cells in the primate CL. It remains unclear why VEGF protein and mRNA levels decline as the primate CL ages in the menstrual cycle, but it could involve both endocrine factors (e.g., luteotropic LH) and local factors (O2 tension and insulin-like growth factors [14]). Unlike VEGF production, hypoxic conditions reduced P4 production by luteal cells either in the absence of LH when basal secretion was high (i.e., early luteal phase) (Table 1) or in the presence of LH (all luteal stages; e.g., late luteal phase, as shown in Fig. 5). Because P4 synthesis requires a member of the cytochrome P-450 enzyme family (CYP11A or P-450 side-chain cleavage) that acts to insert one atom of oxygen into their products [48], it is not surprising that a pharmacologic competitor, such as cobalt or carbon monoxide, or low oxygen tension reduces P4 levels in short-term cultures of granulosa [14] or luteal (present study) cells. Considering the pronounced effects on LHstimulated steroidogenesis, we cannot rule out other effects on LH-stimulated signal transduction (e.g., cAMP), cholesterol transport (e.g., StAR), or other mitochondrial functions. In summary, these findings establish the dynamic and selective expression of VEGF isoforms in the macaque CL during the luteal life span in the natural menstrual cycle. Further studies are needed to elucidate the complementary and/or overlapping roles of VEGF 165 versus VEGF 121 during the development, function, and regression of the primate CL. Also, evidence suggests that the primary regulation of VEGF production changes as follicular granulosa cells differentiate into luteal cells, with LH losing its acute stimulatory activity and hypoxia gaining this capacity. How endocrine and local factors control the pattern of VEGF production at the cell or tissue level during the luteal life span awaits further investigation.
VEGF IN THE PRIMATE CORPUS LUTEUM
ACKNOWLEDGMENTS The dedicated assistance of the animal care staff and surgical unit of the Division of Animal Resources and the technical expertise of members of the Molecular and Cellular Biology (MCB) and Endocrine Services (ES) Core Laboratories is greatly appreciated. We give special thanks to Carol Gibbins for her assistance in the preparation and submission of this manuscript.
REFERENCES 1. Koos RD. Ovarian angiogenesis. In: Adashi EY, Leung PCK (eds.), The Ovary. New York: Raven Press; 1993:433–453. 2. Hazzard TM, Stouffer RL. Angiogenesis in ovarian follicular and luteal development. In: Arulkumaran S (ed.), Clinical Obstetrics and Gynaecology. Angiogenesis in the Female Reproductive Tract. London: Bailliere Tindall; 2000:883–900. 3. Reynolds LP, Redmer DA. Expression of the angiogenic factors, basic fibroblast growth factor, and vascular endothelial growth factor in the ovary. J Anim Sci 1998; 76:1671–1681. 4. Fraser HM, Lunn SF. Regulation and manipulation of angiogenesis in the primate corpus luteum. Reproduction 2001; 121:355–362. 5. Ferrara N. Vascular endothelial growth factor: basic science and clinical progress. Endocr Rev 2004; 25:581–611. 6. Zimmermann RC, Hartman T, Bohlen P, Sauer MV, Kitajewski J. Preovulatory treatment of mice with anti-VEGF receptor 2 antibody inhibits angiogenesis in corpora lutea. Microvasc Res 2001; 62:15–25. 7. Wulff C, Wilson H, Rudge JS, Wiegand SJ, Lunn SF, Fraser HM. Luteal angiogenesis: prevention and intervention by treatment with vascular endothelial growth factor trap (A40). J Clin Endocrinol Metab 2001; 86:3377–3386. 8. Hazzard TM, Xu F, Stouffer RL. Injection of soluble vascular endothelial growth factor receptor 1 into the preovulatory follicle disrupts ovulation and subsequent luteal function in rhesus monkeys. Biol Reprod 2002; 67:1305–1312. 9. Robinson CJ, Stringer SE. The splice variants of vascular endothelial growth factor (VEGF) and their receptors. J Cell Sci 2001; 114:853– 865. 10. Hovey RC, Goldhar AS, Baffi J, Vonderhaar BK. Transcriptional regulation of vascular endothelial growth factor expression in epithelial and stromal cells during mouse mammary gland development. Mol Endocrinol 2001; 15:819–831. 11. Maes C, Carmeliet P, Moermans K, Stockmans I, Smets N, Collen D, Bouillon R, Carmeliet G. Impaired angiogenesis and endochondral bone formation in mice lacking the vascular endothelial growth factor isoforms VEGF 164 and VEGF 188. Mech Dev 2002; 111:61–73. 12. Ancelin M, Buteau-Lozano H, Meduri G, Osborne-Pellegrin M, Sordello S, Plouet J, Perrot-Applanat M. A dynamic shift of VEGF isoforms with a transient and selective progesterone-induced expression of VEGF 189 regulates angiogenesis and vascular permeability in human uterus. Proc Natl Acad Sci U S A 2002; 99:6023–6028. 13. Christenson LK, Stouffer RL. Follicle-stimulating hormone and luteinizing hormone/chorionic gonadotropin stimulation of vascular endothelial growth factor production by macaque granulosa cells from pre- and periovulatory follicles. J Clin Endocrinol Metab 1997; 82: 2135–2142. 14. Martinez-Chequer JC, Stouffer RL, Hazzard TM, Patton PE, Molskness TA. Insulin-like growth factor (IGF)-1 and -2, but not hypoxia, synergize with gonadotropin hormone to promote vascular endothelial growth factor (VEGF)-A secretion by monkey granulosa cells from preovulatory follicles. Biol Reprod 2003; 68:1112–1118. 15. Koos RD. Increased expression of vascular endothelial growth/permeability factor in the rat ovary following an ovulatory gonadotropin stimulus: potential roles in follicle rupture. Biol Reprod 1995; 52: 1426–1435. 16. Laitinen M, Ristimaki A, Honkasalo M, Narko K, Paavonen K, Ritvos O. Differential hormonal regulation of vascular endothelial growth factors VEGF, VEGF-B, and VEGF-C messenger ribonucleic acid levels in cultured human granulosa-luteal cells. Endocrinology 1997; 138:4748–4756. 17. Neeman M, Abramovitch R, Schiffenbauer YS, Tempel C. Regulation of angiogenesis by hypoxic stress: from solid tumors to the ovarian follicle. Int J Exp Pathol 1997; 78:57–70. 18. Young KA, Stouffer RL. Gonadotropin and steroid regulation of matrix metalloproteinases and their endogenous tissue inhibitors in the developed corpus luteum of the rhesus monkey during the menstrual cycle. Biol Reprod 2004; 70:244–252.
933
19. Duffy DM, Abdelgadir SE, Stott KR, Resko JA, Stouffer RL, Zelinski-Wooten MB. Androgen-receptor messenger RNA expression in the rhesus monkey ovary. Endocrine 1999; 11:23–30. 20. Young KA, Hennebold JD, Stouffer RL. Dynamic expression of mRNAs and proteins for matrix metalloproteinases and their tissue inhibitors in the primate corpus luteum during the menstrual cycle. Mol Hum Reprod 2002; 8:833–840. 21. Chaffin CL, Stouffer RL. Expression of matrix metalloproteinases and their tissue inhibitor messenger ribonucleic acids in macaque periovulatory granulosa cells: time course and steroid regulation. Biol Reprod 1999; 61:14–21. 22. Hennebold JD, Tanaka M, Saito J, Hanson BR, Adashi EY. Ovaryselective genes I: The generation and characterization of an ovaryselective cDNA library. Endocrinology 2000; 141:2725–2734. 23. Hazzard TM, Christenson LK, Stouffer RL. Changes in expression of vascular endothelial growth factor and angiopoietin-1 and -2 in the macaque corpus luteum during the menstrual cycle. Mol Hum Reprod 2000; 6:993–998. 24. Duffy DM, Stouffer RL. Progesterone receptor messenger ribonucleic acid in the primate corpus luteum during the menstrual cycle: possible regulation by progesterone. Endocrinology 1995; 136:1869–1876. 25. Irusta G, Parborell F, Peluffo M, Manna PR, Gonzalez-Calvar SI, Calandra R, Stocco DM, Tesone M. Steroidogenic acute regulatory protein in ovarian follicles of gonadotropin-stimulated rats is regulated by a gonadotropin-releasing hormone agonist. Biol Reprod 2003; 68: 1577–1583. 26. Duffy DM, Wells TR, Haluska GJ, Stouffer RL. The ratio of progesterone receptor isoforms changes in the monkey corpus luteum during the luteal phase of the menstrual cycle. Biol Reprod 1997; 57:693– 699. 27. Brannian JD, Stouffer RL. Progesterone production by monkey luteal cell subpopulations at different stages of the menstrual cycle: changes in agonist responsiveness. Biol Reprod 1991; 44:141–149. 28. Garrido C, Saule S, Gospodarowicz D. Transcriptional regulation of vascular endothelial growth factor gene expression in ovarian bovine granulosa cells. Growth Factors 1993; 8:109–117. 29. Berisha B, Schams D, Kosmann M, Amselgruber W, Einspanier R. Expression and tissue concentration of vascular endothelial growth factor, its receptors, and localization in the bovine corpus luteum during the estrous cycle and pregnancy. Biol Reprod 2000; 63:1106– 1114. 30. Redmer DA, Dai Y, Li J, Charnock-Jones DS, Smith SK, Reynolds LP, Moor RM. Characterization and expression of vascular endothelial growth factor (VEGF) in the ovine corpus luteum (CL). J Reprod Fertil 1996; 108:157–165. 31. Renner W, Pilger E. Simultaneous in vivo quantitation of vascular endothelial growth factor mRNA splice variants. J Vasc Res 1999; 36: 133–138. 32. Poltorak Z, Cohen T, Sivan R, Kandelis Y, Spira G, Vlodavsky I, Keshet E, Neufeld G. VEGF145, a secreted vascular endothelial growth factor isoform that binds to extracellular matrix. J Biol Chem 1997; 272:7151–7158. 33. Krussel JS, Behr B, Milki AA, Hirchenhain J, Wen Y, Bielfeld P, Polan ML. Vascular endothelial growth factor (VEGF) mRNA splice variants are differentially expressed in human blastocysts. Mol Hum Reprod 2001; 7:57–63. 34. Stouffer RL. Progesterone as a mediator of gonadotropin action in the primate corpus luteum: beyond steroidogenesis. Hum Reprod Update 2003; 9:99–117. 35. Gagnon ML, Bielenberg DR, Gechtman Z, Miao H-Q, Takashima S, Soker S, Klagsbrun M. Identification of a natural soluble neuropilin1 that binds vascular endothelial growth factor: in vivo expression and antitumor activity. Proc Natl Acad Sci U S A 2000; 97:2573–2578. 36. Gluzman-Poltorak Z, Cohen T, Herzog Y, Neufeld G. Neuropilin-2 is a receptor for the vascular endothelial growth factor (VEGF) forms VEGF-145 and VEGF-165. J Biol Chem 2000; 275:18040–18045. 37. Pavelock K, Braas KM, Ouafik L, Osol G, May V. Differential expression and regulation of the vascular endothelial growth factor receptors neuropilin-1 and neuropilin-2 in rat uterus. Endocrinology 2001; 142:613–622. 38. Stalmans I, Ng Y-S, Rohan R, Fruttiger M, Bouche A, Yuce A, Fujisawa H, Hermans B, Shani M, Jansen S, Hicklin D, Anderson DJ, Gardiner T, Hammes H-P, Moons L, Dewerchin M, Collen D, Carmeliet P, D’Amore PA. Arteriolar and venular patterning in retinas of mice selectively expressing VEGF isoforms. J Clin Invest 2001; 109: 327–336. 39. Wulff C, Wilson H, Largue P, Duncan WC, Armstrong DG, Fraser
934
40.
41.
42.
43.
TESONE ET AL.
HM. Angiogenesis in the human corpus luteum: localization and changes in angiopoietins, Tie-2, and vascular endothelial growth factor messenger ribonucleic acid. J Clin Endocrinol Metab 2000; 85: 4302–4309. Hazzard TM, Molskness TA, Chaffin CL, Stouffer RL. Vascular endothelial growth factor (VEGF) and angiopoietin regulation by gonadotrophin and steroids in macaque granulosa cells during the periovulatory interval. Mol Hum Reprod 1999; 5:1115–1121. Barboni B, Turriani M, Galeati G, Spinaci M, Bacci ML, Forni M, Mattioli M. Vascular endothelial growth factor production in growing pig antral follicles. Biol Reprod 2003; 63:858–864. Redmer DA, Doraiswamy V, Bortnem BJ, Fisher K, Jablonka-Shariff A, Grazul-Bilska AT, Reynolds LP. Evidence for a role of capillary pericytes in vascular growth of the developing ovine corpus luteum. Biol Reprod 2001; 65:879–889. Ravindranath N, Little-Ihrig L, Phillips HS, Ferrara N, Zeleznik AJ.
44.
45. 46. 47.
48.
Vascular endothelial growth factor messenger ribonucleic acid expression in the primate ovary. Endocrinology 1992; 131:254–260. Sharkey AM, Day K, McPherson A, Malik S, Licence D, Smith SK, Charnock-Jones DS. Vascular endothelial growth factor expression in human endometrium is regulated by hypoxia. J Clin Endocrinol Metab 2000; 85:402–409. Giordano FJ, Johnson RS. Angiogenesis: the role of the microenvironment in flipping the switch. Curr Opin Genet Dev 2001; 11:35– 40. Pugh CW, Ratcliffe PJ. Regulation of angiogenesis by hypoxia: role of the HIF system. Nat Med 2003; 9:677–684. Friedman CI, Danforth DR, Herbosa-Encarnacion C, Arbogast L, Alak BM, Seifer DB. Follicular fluid vascular endothelial growth factor concentrations are elevated in women of advanced reproductive age undergoing ovulation induction. Fertil Steril 1997; 68:607–612. Hall PF. Cytochromes P-450 and the regulation of steroid synthesis. Steroids 1986; 48:131–196.
