| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Division of Animal Physiology,3 University of Nottingham, Loughborough, LEICS LE12 5RD, United Kingdom
Institute of Physiology,4 Technical University of Munich, D-85350 Freising, Germany
ABSTRACT
Luteal inadequacy is a major cause of infertility in a number of species. During the early luteal phase, progesterone production requires the rapid growth of the corpus luteum (CL), which is in turn dependent on angiogenesis. In the present study, we examined the temporal changes in vascular endothelial growth factor A (VEGFA), fibroblast growth factor 2 (FGF2) and secreted protein, acidic, cysteine-rich (osteonectin) (SPARC) during the follicular-luteal transition and CL development in the cow. Luteal VEGFA concentrations increased as the CL developed but were lower in the regressing CL. Conversely, luteal FGF2 concentrations were highest immediately postovulation in the collapsed follicle and declined as the CL developed. Furthermore, three FGF2 isoforms were present in the collapsed follicle, but only one isoform was detected in older CL. Interestingly, FGF2 concentrations increased in the regressing CL. Western blot analysis for SPARC showed the presence of two isoforms, which were constitutively expressed throughout CL development. Further studies investigated the regulation of FGF2 by LH, which showed that FGF2 concentrations in preovulatory follicular fluid were higher in those animals that had experienced an LH surge. Moreover, LH stimulated FGF2 production in dispersed luteal cells. Conversely, the LH surge had no effect on follicular fluid VEGFA concentrations. In conclusion, FGF2 was more dynamic than VEGFA and SPARC during the follicular-luteal transition, which suggests that FGF2 plays a key role in the initiation of angiogenesis at this time. Furthermore, it is likely that this is stimulated by the LH surge. The results also suggest that VEGFA and SPARC have a more constitutive, but essential, role in the development of the CL vasculature.
angiogenesis, corpus luteum, cow, FGF2, follicle, growth factors, luteinizing hormone, ovulation, SPARC, VEGFA
Follicular-luteal transition is a dynamic process, which involves a series of biochemical and morphological changes in the preovulatory follicle following the LH surge [1, 2]. These include the differentiation of theca and granulosa cells into luteal cells, tissue remodeling and growth, a switch in steroidogenesis, and increasing progesterone production. In order to meet these demands, the growth of blood vessels and establishment of a blood supply (angiogenesis) is essential [1, 2]. The rates of luteal growth and angiogenesis are such that they are only equaled by the fastest growing tumors. Indeed, the mature corpus luteum (CL) is so vascular that the majority of luteal cells are adjacent to one or more capillaries [2–4]. Furthermore, the CL has one of the greatest rates of blood flow per unit of tissue. Moreover, ovarian blood flow is highly correlated with the rate of progesterone secretion [2–5]. This is of particular importance in cows, since inadequate postovulation progesterone is associated with poor embryo development and reduced fertility [6–8].
Ovarian angiogenesis is controlled by a plethora of angiogenic factors, including vascular endothelial growth factor A (VEGFA), fibroblast growth factor 2 (FGF2) (basic) and secreted protein, acidic, cysteine-rich (osteonectin) (SPARC) [9–12]. VEGFA is not only a highly specific mitogen for vascular endothelial cells but also has been shown to regulate their migration and tube formation. It is now well established that there are five different human VEGFA isoforms consisting of 121, 145, 165, 189, or 206 amino acids [3, 13]. These all arise from alternative splicing of a single gene, and bovine VEGFA is one amino acid shorter than human VEGFA [14]. The different isoforms have different biochemical properties. Namely, VEGFA121 is a soluble, secreted form; VEGFA145 and VEGFA165 are also secreted but can bind to heparin and the ECM, while VEGFA189 and VEGFA206 are cell associated. VEGFA exerts its action through two tyrosine kinase receptors, namely fms-related tyrosine kinase 1 (FLT1) and kinase insert domain receptor (KDR) [13]. The importance of VEGFA in luteal angiogenesis and function was demonstrated by Fraser et al. [15] where they showed the immunoneutralization of VEGFA during the early luteal phase reduced endothelial cell number and decreased progesterone production in marmoset monkeys. Recently, a similar observation has been reported in the cow [16].
FGF2 is a heparin-binding growth factor, which occurs in an 18-kDa cytoplasmic and four large molecular weight nuclear isoforms (22, 22.5, 24, and 34 kDa) [17, 18]. FGF2 is a mitogen for a plethora of cell types including fibroblasts, osteoblasts, and endothelial cells and is also involved in cell migration and differentiation [17, 18]. While it has long been known that FGF2 is present and active in the ovary, remarkably little is known about its exact role in ovarian angiogenesis. Recently, preliminary evidence has shown that immunoneutralization of FGF2 suppressed CL growth and progesterone production in the cow in a similar manner to that of VEGFA immunoneutralization (Yamashita H et al., unpublished observations).
SPARC is a matricellular protein, which is secreted and interacts with cell surface receptors, extracellular matrix (ECM), and growth factors. SPARC is typically expressed in tissue undergoing morphogenesis and reorganization such as during development or in response to injury. The mature SPARC protein has four functional domains (I to IV). The functionality of SPARC protein is influenced by proteolytic action, with the different peptide fragments having a differential influence on biological action (e.g., endothelial cell behavior) [19]. For example, SPARC itself is contra-adhesive, inhibiting endothelial cell proliferation and migration, but has been shown to promote endothelial tube formation [20]. In contrast, the basic proteolytic SPARC peptide 2.3 (amino acids 113–130 and contains the Cu2+ binding site) stimulates endothelial cell proliferation and angiogenesis [21]. Moreover, Smith et al. [12] detected SPARC mRNA and protein in the ovine CL, but SPARC has not been investigated to date in the bovine ovary.
Hence, the objective of this study was to investigate the temporal changes in VEGFA, FGF2, and SPARC concentrations during follicular-luteal transition and CL development and regression in the cow. This knowledge is essential to our understanding of how angiogenesis is regulated at these critical time points. This is necessary since the CL requires a very high blood flow rate to meet its metabolic demands and this is dependent on rapid and intense angiogenesis. This whole process is essential, since luteal inadequacy (the failure to produce progesterone) is a major contributor to poor embryo development and fertility.
The animal studies (experiments 2, 3, and 4) were conducted under the Animals (Scientific Procedures) Act 1986. All experiments were performed on mature Holstein-Friesian cows, which had calved at least once and were obtained from local commercial farms.
Experiment 1: Angiogenic Factors in the Developing CL
Ovaries were collected from a local abattoir and transported back to the laboratory in 1x PBS at room temperature. The CLs were classified in the following groups based on criteria outlined by Ireland et al. [22]; Arosh et al. [23]: Days 1–2 (collapsed follicle, n = 6); Days 3–4 (very early, n = 5); Days 5–6 (early, n = 7); and Days 8–12 (mid, n = 3). The CLs were weighed and dissected into 1-cm segments and then stored at –80°C until analysis. The mean ± SEM CL weight for each group was as follows: 0.5 ± 0.2 g (Days 1–2); 1.2 ± 0.3 g (Days 3–4); 2.3 ± 0.4 g (Days 5–6), and 3.8 ± 0.6 g (Days 8–12). One segment was fixed in Bouins fixative (Sigma, Poole, UK) for 4 h for immunohistochemical analysis. Following fixation, tissue sections were dehydrated in a graded series of ethanol, cleared in toluene, and embedded in paraffin wax using standard procedures.
Experiment 2: Angiogenic Factors in Timed Luteal Tissue
This experiment was performed on 15 mature Holstein-Friesian cows at the end of lactation. The estrous cycles were synchronized with two i.m. injections of 50 µg cloprostenol (Estrumate, Schering Plough Animal Health, Welwyn Garden City, UK) 12 days apart. Estrus was detected by three times daily observation and the day on which estrus was first observed defined as Day 0. The cows (n = 3 per group) were then slaughtered on Days 2, 5, 12, 18 and during the follicular phase (regressing CL). The follicular phase cows were slaughtered 48 h following the second injection of prostaglandin F2
. Immediately prior to slaughter, a single blood sample was collected by jugular venepuncture and plasma samples stored at –20°C for progesterone analysis. The CLs were processed as described in Experiment 1.
Experiment 3: Effect of LH Surge on Preovulatory Follicle
This experiment was performed on 11 nonlactating cows. Estrous cycles were synchronized by inserting a controlled internal drug release (CIDR) device (Pfizer Animal Health, Sandwich, UK) for 10 days and injecting i.m. 50 µg cloprostenol (Estrumate, Schering Plough Animal Health, Welwyn Garden City, UK) on the day of CIDR withdrawal. During the midluteal phase of the subsequent cycle, luteolysis was induced by injecting i.m. 50 µg cloprostenol. Blood samples were collected every 4 h from the second cloprostenol injection until slaughter via an indwelling jugular vein catheter. All samples were collected into heparinized tubes, centrifuged at 1500 x g for 10 min and plasma stored at –20°C until analysis. In order to determine follicular growth, the ovaries were scanned up to twice daily by transrectal ultrasonography using Sonovet 6000 with a 7.5-MHz linear array transducer (BCF Technology, Livingstone, UK) from the second cloprostenol injection until slaughter. The animals were then slaughtered either 48 h (n = 6, anticipated pre-LH surge) or 84 h after second cloprostenol injection (n = 5, post-LH surge, prior to ovulation). All follicles >8 mm were dissected out, follicular fluid was aspirated using a 21-ga needle and 2-ml syringe. The follicular fluid was then centrifuged at 2000 x g for 5 min, and the supernatant was collected and stored at –80°C until analysis. Plasma LH concentrations were analyzed by RIA as described by Price et al. [24]. The sensitivity of the assay was 0.2 ng/ml, and the intra-assay coefficient of variation was 9.6% with all samples analyzed in a single assay.
Experiment 4: Effect of LH on Luteal FGF2 Production
This experiment was performed on five mature nonlactating cows. The estrous cycles were synchronized by inserting a CIDR device (Pfizer Animal Health, Sandwich, UK) for 10 days and injecting i.m. 50 µg cloprostenol (Estrumate, Schering Plough Animal Health, Welwyn Garden City, UK) on the day of CIDR withdrawal. During the mid-luteal phase of the subsequent cycle, luteolysis was induced with a second cloprostenol injection. At this point, the ovaries were scanned daily by transrectal ultrasonography as previously described. Once the cows had to show estrus signs, this was increased to twice daily. Ovulation was defined as the day the preovulatory follicle disappeared. The corpora lutea were collected 4 days after ovulation. The luteal cells were dispersed as described later.
CL Content Analysis of VEGFA and FGF2 (Experiments 1 and 2)
For FGF2 and VEGFA analyses, 200 mg of CL was homogenized on ice in 10 volumes of 1x PBS containing 1 mg/ml dithiothreitol (DTT, Sigma, Poole, UK) and protease inhibitor cocktail (Complete, Roche, Lewes, UK). The samples were then centrifuged at 2000 x g for 15 min and aliquots stored at –80°C until analysis. The FGF2 concentration was determined using an FGF2 ELISA kit (R&D Systems, Abingdon, UK) as per instructions, while the concentration of VEGFA was determined by RIA as previously described and validated by Berisha et al. [25]. Samples were measured in a single assay, and the intra-assay CVs were 7.5% and 7.4% for FGF2 and VEGFA, respectively.
Western Blot Analysis for VEGFA, FGF2, and SPARC (Experiments 1, 2, and 3)
Western blot analysis was performed to detect and compare the levels of FGF2, VEGFA, and SPARC during CL development as previously described by Robinson et al. [26]. Briefly, 100 mg of CL sample was homogenized on ice in 1x PBS containing 1 mg/ml dithiothreitol (DTT, Sigma, Poole, UK) and protease inhibitor cocktail (Complete, Roche, Lewes, UK). The samples were then centrifuged at 2000 x g for 15 min and aliquots stored at –80°C until analysis. The protein concentration was determined using the Bradford assay (Bio Rad, Hemel Hempstead, UK). Protein samples (25 µg total protein) were subjected to SDS-PAGE using a 12% resolving gel (gel size 10.2 x 8.3 mm) according to the method described by Laemmli [27]. All the samples were loaded randomly onto two gels with a positive control (midcycle bovine CL) on each gel. The separated proteins were electrotransferred to PDVF membrane (BioRad, Hemel Hempstead, UK) and blocked with 5% milk in PBS-Tween (PBS-T) for 2 h. The membranes were then incubated with either 1 µg/ml rabbit anti-human FGF2 (Autogen Bioclear, Calne, Wiltshire, UK), 1 µg/ml rabbit anti-human VEGFA (Autogen Bioclear), or 2 µg/ml mouse anti-human SPARC (Zymed, Paisley, UK) overnight at 4°C. Control membranes were incubated with appropriate concentration of control sera. On the next day, the membranes were washed in PBS-T for 5 x 10 min and then incubated with the secondary antibody, either goat anti-rabbit IgG-peroxidase conjugate (1:5000, Sigma, Poole, UK) or sheep anti-mouse IgG-peroxidase (1:50 000, Amersham Biosciences, Amersham, UK) for 1 h at room temperature. The membranes were washed again as before and visualized with an enhanced chemiluminescence system (Amersham Biosciences). Immunoreactive bands (VEGFA, FGF2, SPARC, and tubulin) were detected using the Molecular Imager FX (Bio Rad, Hemel Hempstead, UK) and semi-quantified by densitometry using Quantity One Version 4.2 (BioRad). The results are presented as a ratio of VEGFA, FGF2, or SPARC to tubulin.
Localization of SPARC Protein by Immunohistochemistry (Experiment 1)
Tissue sections (5 µm; three sections per slide) were placed onto Super-frost Plus slides (BDH, Poole, UK) and then dewaxed in xylene, rehydrated through a series of descending concentrations of ethanol and washed in PBS. Antigen retrieval was performed by boiling sections in 10 mM citrate buffer, pH 6.0 in a microwave for 10 min, which was followed by blocking endogenous peroxidase activity with 0.03% (v/v) hydrogen peroxide in methanol. Nonspecific binding was blocked by incubating sections with 50 mg/ml BSA for 30 min. The primary antibody used was 2.5 µg/ml monoclonal mouse anti-human SPARC (Zymed, Paisley, UK). For the control sections, equivalent concentrations of mouse IgG were used. The slides were incubated in a humidified chamber at 4°C overnight. On the following day, the primary antibody was detected using the Dako EnVision system as per instructions (Dako, Ely, Cambridgeshire, UK) and visualization was performed with diaminobenzidine (Dako).
Follicular Fluid Analysis (Experiment 3)
The progesterone concentration in follicular fluid was determined using RIA as previously validated by Corrie et al. [28] and Law et al. [29]. The samples were diluted 200-fold into RIA buffer, and the intra-assay CV was 7.0%. Follicular fluid concentrations of estradiol were determined using RIA as previously described by Grant et al. [30]. The samples were diluted 2000-fold into RIA buffer, and the intra-assay CV was 9.9%. In both cases, all samples were analyzed in a single assay.
FGF2 Production by Luteal Cells (Experiment 4)
The culture of dispersed luteal cells was based on the method previously described by Robinson et al. [26]. Briefly, the corpora lutea recovered on Day 5 were sliced using scissors, and all connective tissue was trimmed away. The luteal cells were dissociated by incubating the luteal tissue in DMEM/F12 1:1 media (Gibco, Paisley, UK) containing 2 mg/ml collagenase I (Lorne Laboratories, Reading, UK) and 25 µg/ml DNase I (Sigma, Poole, UK) for 90 min in a shaking water-bath at 37°C. The dissociated cells were filtered through 100 µm gauze and washed three times. The cells were then plated out at 2 x 105 viable cells (as determined by trypan blue) per well in serum free DMEM/F12 1:1 media supplemented with 100 units/ml pencillin, 10 µg/ml streptomycin, 10 µg/ml insulin, 5.5 µg/ml transferrin, 5 ng/ml selenium and 5 mg/ml BSA (all Sigma, Poole, UK) into 96 well plates. The cells were challenged with either control or 100 ng/ml LH (AFP11743B, biopotency 1.06 x oLH NIDDK-I-2; a gift from Dr. A.F. Parlow, NIDDK, California) with four replicates per treatment. This concentration of LH has been previously shown to produce maximum production of progesterone by bovine luteal cells [26]. The luteal cells were incubated at 38°C in a humidified incubator in 5% CO2/95% air for 18 h, then the media was collected and stored at –20°C. FGF2 concentrations were analyzed by ELISA kit (R&D Systems, Abingdon, UK) as per manufacturer instructions.
All the data were checked for normality and heterogeneity of variance. If these parameters were not satisfied then the data were log transformed. For experiments 1 and 2, the effect of CL developmental stage (slaughterhouse or timed material) on VEGFA, FGF2, and SPARC concentrations were analyzed by one-way ANOVA. For the Western blot densitometry analysis the VEGFA, FGF2, and SPARC levels were expressed as a ratio to tubulin OD. If there was a significant time of CL stage effect, then Bonferroni's multiple comparison was used to determine where the differences lay. For experiment 3, the effect of the LH surge on follicular fluid concentrations of estradiol, progesterone, FGF2, and VEGFA was compared using an unpaired Student t-test. Simple linear regression was performed between all variables investigated. For experiment 4, the production of FGF2 by dispersed luteal cells was normalized to percentage of control wells and then analyzed by one-way ANOVA in a randomized block design with LH as the factor.
For all experiments, P < 0.05 was considered significant. All data are quoted as mean ± SEM unless stated otherwise.
Experiment 1: Angiogenic Factors in the Developing CL
In experiment 1, changes in VEGFA, FGF2, and SPARC protein concentrations during the early growth and development of the bovine CL were initially analyzed by Western blot. No bands were detected in membranes treated with control sera, thus all bands present were specific for VEGFA, FGF2, and SPARC. VEGFA analysis revealed the presence of a single band at 17 kDa (Fig. 1A). Densitometric analysis of this band showed that VEGFA protein levels increased as the CL developed (P < 0.05; Fig. 1B). Western blot analysis for FGF2 showed the presence of a band (13 kDa) in all samples. There were two further bands at 17 and 20 kDa, albeit at lower intensities. However, these bands were only detectable in the Day 1–2 CL (collapsed follicle) samples (Fig. 2A). Densitometric analysis for the 13-kDa band revealed that the collapsed follicle had the highest FGF2 levels. After Day 2, the luteal content of FGF2 decreased and was then maintained at constant levels for the remainder of the growth and development period of the bovine CL (P < 0.001; Fig. 2B). These observations were confirmed by determining the luteal concentration of FGF2 by ELISA in the same samples. Again, FGF2 concentrations were highest in the collapsed follicle (232 ± 18 ng/g) and declined thereafter as the CL developed (P < 0.001; Fig. 2C).
|
|
Immunoblots for SPARC protein identified two bands at 38 and 46 kDa (Fig. 3A). The 46 kDa band is likely to represent the mature form of SPARC, while identity of the second band at 38 kDa is unknown but may represent a proteolytic fragment of the mature SPARC. Similar observations have been reported for SPARC by Yan et al. [31]. The intensity of both the 38- and 46-kDa bands did not change throughout the growth and development of the bovine CL (P > 0.15; Figs. 3B and 3C). A third band at 36 kDa was present in one sample (Days 5–6), but its physiological significance is unclear. Immunolocalization revealed that SPARC in the developing CL was localized to the endothelial cells in all CLs from Days 5 to 8. There was also evidence of SPARC immunostaining in both small and, to a lesser extent, large luteal cells, although the intensity of staining varied between CLs with no obvious pattern (Fig. 4).
|
|
Experiment 2: Angiogenic Factors in Timed Luteal Tissue
These results were further verified and extended by collecting CL samples at precise times throughout the estrous cycle. VEGFA concentrations tended to increase as CL developed with highest concentrations being observed on Day 12. During luteolysis, luteal VEGFA concentrations decreased (P < 0.001; Fig. 5A). Conversely, luteal FGF2 concentrations were much more dynamic throughout the estrous cycle, with 6- and 14-fold higher concentrations on Day 2 compared with Days 5 and 12, respectively (P < 0.001; Fig. 5B). Interestingly, FGF2 concentrations increased on Day 18 (prior to luteolysis) and were maintained at these concentrations in the regressing CL (P < 0.001; Fig. 5B).
|
Experiment 3: Effect of LH Surge on Preovulatory Follicle
None of the pre-LH surge animals had experienced an LH surge (data not shown). However, all animals in the post-LH surge group had or were experiencing an LH surge at the time of slaughter (range, 6–22 h since the LH surge was initiated). The size of the preovulatory follicle at slaughter was not different between the two groups (15.3 ± 1.5 mm vs. 14.2 ± 0.8 mm pre- vs. post-LH surge, respectively; P > 0.15). There were no differences in the follicular fluid concentration of estradiol between pre- and post-LH surge cows (P > 0.15; Fig. 6A). However, progesterone concentrations in follicular fluid were 4-fold greater in post-LH surge cows (P < 0.001; Fig. 6B). Similarly, the estradiol:progesterone ratio in follicular fluid was lower in the post-LH surge group (P < 0.001, data not shown). The levels of angiogenic factors, VEGFA and FGF2 in follicular fluid were also determined. Western blot analysis of VEGFA revealed the presence of two bands at 17 and 20 kDa. Neither of the two VEGFA isoforms was affected by the LH surge (P > 0.15; Fig. 7A and B). Conversely, FGF2 concentrations in follicular fluid were 7-fold higher in cows, which had experienced an LH surge (P < 0.001; Fig. 7C).
|
|
Experiment 4: Effect of LH on Luteal FGF2 Production
In this experiment, the effects of LH on FGF2 production by dispersed luteal cells were investigated. The concentration of FGF2 in the control wells was 21.7 ± 5.7 pg/ml. FGF2 production was increased by almost 50% with treatment with 100 ng/ml LH (P < 0.001; Fig. 8).
|
In the current study, we have found that FGF2 was up-regulated during the follicle-luteal transition, whereas the pattern of VEGFA protein expression was less dynamic during this period. Furthermore, LH increased FGF2 production both in vivo and in vitro. Collectively, these data suggest that the largely overlooked angiogenic factor, FGF2, plays a more important role in the initiation of angiogenesis postovulation, while VEGFA plays a more constitutive role in the maintenance of the developing capillaries/blood vessels. We have also shown for the first time that SPARC protein is present constantly during the growth and development of the bovine CL.
This is the first study, to our knowledge, that has examined the protein expression patterns for both FGF2 and VEGFA in frequently collected tissue samples during the dynamic follicle-luteal transition period and early CL development. While FGF2 and VEGFA are likely to be the principal angiogenic factors controlling follicular angiogenesis, other angiogenic factors such as SPARC are likely to play a modulatory role. FGF2 and VEGFA are potent stimulators of microvascular endothelial cell proliferation and migration, while VEGFA also promotes vascular permeability. Berisha et al. [25] found that in the cow FGF2 was primarily located on the endothelial cells and pericytes of the theca cell layer, with higher levels of FGF2 mRNA expression detected in more developed follicles. In contrast, granulosa cells are the major source of VEGFA in the follicle [25, 32] and again expression increases as the follicle develops [25, 33]. Consequently, there is an accumulation of VEGFA in the follicular fluid, which is likely to direct blood vessel extension towards the granulosa cell layer. VEGFA is also expressed in the theca layer [25, 32]. In the present study, we have shown that in the bovine preovulatory follicle, FGF2 concentrations in the follicular fluid were increased following an endogenous LH surge. This is in agreement with Berisha et al. [34], who showed that FGF2 mRNA in follicles was up-regulated after an induced LH peak. They also reported that the LH surge induced a translocation of FGF2 protein in the preovulatory follicle. Namely, that in the pre-LH surge follicle FGF2 was localized to the cytoplasm of blood vessels in the theca, whereas after the LH surge FGF2 was immunolocalized only to nucleolus of the granulosa cells. Further support for the up-regulation of FGF2 mRNA by LH was shown by Stirling et al. [35], where LH stimulated a dose-dependent increase in FGF2 mRNA in bovine luteal cells. Moreover, in the present study, LH increased FGF2 protein production by 50% in the dispersed bovine luteal cells. Collectively, these data from both in vivo and in vitro studies would suggest that LH stimulates follicular FGF2, which in turns initiates angiogenesis and tissue remodelling processes. Conversely, follicular fluid VEGFA was unaffected by the LH surge in cows. This is in contrast to studies in macaques [36] and rats [37] where follicular fluid VEGFA was increased by gonadotropins. The reasons for these discrepancies are unclear, but it is feasible that there are species differences. However, Schams et al. [38] showed that LH stimulated VEGFA protein production by cultured bovine granulosa cells. The other difference between these studies is that in the present one, follicular fluid or granulosa cells were collected from cows undergoing a natural rather than induced LH surge.
The analysis of luteal FGF2 concentrations during early CL development in the cow revealed three exciting observations. First, FGF2 concentrations remained very high in the collapsed follicle (and to lesser extent in Day 2 CL) following its up-regulation in follicular fluid by the LH surge, at a time when the proliferation index is highest and intense angiogenesis is occurring [26, 39]. Thereafter, luteal FGF2 concentrations decreased but were still maintained at high levels for the remainder of CL development. Schams and Berisha [4] also observed that luteal FGF2 concentrations were higher in the early CL (Days 1–4), but this study has dissected out the changes that occur during this early period. We have shown that it is the collapsed follicle that has high luteal FGF2 concentrations. Conversely, Stirling et al. [35] reported no changes in luteal FGF2 mRNA in the early to late bovine CL. However, they did not distinguish between the different days in the early CL and only measured FGF2 mRNA. Our results suggest that in the cow, the LH surge initiates luteal angiogenesis and remodelling by stimulating a transient up-regulation of FGF2 protein. Then as the CL grows and develops, the requirement for angiogenesis diminishes and FGF2 production starts to switch off. Interestingly, in support of this observation, there was a negative correlation between plasma progesterone and luteal FGF2 concentrations on Day 5 (data not shown). This could suggest that as the CL begins to secrete progesterone, its requirement for angiogenesis (and thus FGF2) decreases. Thus on Day 5, the more developed CL has lower luteal FGF2 concentrations. In addition, there was a shift in dynamics of the FGF2 isoforms with two additional isoforms (at 17 and 20 kDa) being detected in the collapsed follicle. Schams and Berisha [4] also reported the disappearance of a higher molecular weight FGF2 isoform (18 kDa) during CL development, while the lower molecular weight isoform was present throughout. They also observed dynamism in the FGF2 immunostaining with a shift away from endothelial cell localisation (early CL) to luteal cell cytoplasm (mid-CL) [4], although it remains to be determined whether this reflects the changes in the FGF2 isoforms. It is feasible that the different cell type in the CL express different FGF2 isoforms. Third, luteal FGF2 concentrations increased in the late and regressing CL. This is in agreement with previous studies, which showed increases in FGF2 mRNA [35] and protein [40] in the bovine CL during luteolysis. The up-regulation of FGF2 protein at a time when little or no angiogenesis is occurring is surprising. However, it is possible that FGF2 is required for the maintenance or growth of larger blood vessels in the CL, which are necessary for the removal of necrotic tissue during structural luteolysis.
In the current study, we have demonstrated that luteal VEGFA concentrations were constantly high throughout CL development with the highest levels being observed in the mid-CL and a pronounced down-regulation in the regressing CL. The current findings are consistent with previous reports in the cow [25, 26, 41]. However, it appears that luteal VEGFA production during CL development is different between species with pigs [42] and humans [43] similar to the cow, while luteal VEGFA in sheep [44], macaque [45], and horses [46] was higher in the early CL. The reasons for these differences are unclear. However, in all cases VEGFA is present in CL at high concentrations and has been shown to be important both in the early CL development [15] and maintenance of mid-CL [47]. There is increasing evidence that VEGFA plays a fundamental role in the maintenance of the vasculature in the CL, when it is no longer undergoing rapid angiogenesis. These functions could include the regulation of vascular permeability or survival of endothelial cells and it is only when the CL starts to regress that VEGFA expression is down-regulated.
Another interesting observation was that two VEGFA isoforms (at 17 and 20 kDa) were present in the preovulatory follicular fluid, but only one (17-kDa) isoform was evident in the developing CL. Similarly, Redmer et al. [44] only found a single VEGFA band in the ovine CL. The VEGFA isoforms in the follicular fluid are likely to be secreted isoforms. Therefore, the 17-kDa band is likely to represent VEGFA120, and the 20-kDa band represents VEGFA164. Furthermore, this suggests that VEGFA120 and VEGFA164 are involved in promoting angiogenesis in the bovine follicle, while only VEGFA120 is active in the bovine CL. Contrary to these results, VEGFA gene analyses have shown that both the VEGFA120 and VEGFA164 mRNA transcripts are present in the ovine and bovine CL [25, 44, 48]. Moreover, VEGFA164 mRNA was the more abundant isoform and represented two thirds of the signal. The reasons for these discrepancies remain to be elucidated. However, it is feasible that the VEGFA genes are translationally regulated. Immunohistochemical studies have shown that VEGFA is localized to both the small and large luteal cells in the cow [25, 41, Robinson et al., unpublished observations]. Collectively, these results suggest that VEGFA may act as a chemoattractant for spouting endothelial cells, on which the FLT1 and KDR have been localized [41, Robinson et al., unpublished observations]. This hypothesis is supported by the fact that it is the secreted VEGFA120 that is present in the bovine CL.
This is the first paper, to our knowledge, that reports presence of the angiogenic factor SPARC in the bovine CL. The mature SPARC form was present at constant levels throughout CL development. Conversely, Smith et al. [12] observed an increase in SPARC mRNA expression in the ovine CL between Days 3 and 7. The likely explanations for these differences are either differential regulation during the translation/posttranslation steps or a species difference. However, the immunolocalization of SPARC to endothelial and luteal cells was in strong agreement with previous observations in the rat and ovine CL [12, 49]. SPARC is often expressed in tissue that is undergoing structural reorganization, thus the presence of SPARC in the bovine CL would suggest a role for SPARC in the remodelling of the CL. The contra-adhesive properties of SPARC might facilitate this by reducing the adhesion of luteal cells to the ECM thereby enabling their spatial redistribution. Previously, the presence of a second SPARC band has not been reported. It is feasible that this might represent a proteolytic fragment of the mature SPARC. Moreover, peptide fragments of SPARC are known to be angiogenic. Thus the action of SPARC and its proteolytic fragments on luteal angiogenesis warrant further investigation.
In conclusion, the present study has clearly demonstrated that FGF2, VEGFA, and SPARC have different roles in regulating angiogenesis in the bovine ovary. In particular, it has highlighted the dynamic role that FGF2 plays during the remodelling transition periods (e.g., follicular-luteal transition and CL regression). Furthermore, FGF2 was up-regulated by LH in vivo and in vitro and thus generating novel hypotheses of how luteal angiogenesis is regulated. Conversely, VEGFA and SPARC were more constitutively expressed at all stages of vasculature development of the preovulatory follicle and CL but still play critical roles in regulating luteal angiogenesis. Collectively, this suggests the angiogenic process is initiated by a burst of FGF2 activity induced by the LH surge, while the development and maintenance of this vasculature is more controlled by VEGFA and SPARC.
ACKNOWLEDGMENTS
The authors wish to thank Jim Craigon for his statistical guidance. We are also grateful to the staff of Nottingham University for their assistance with sample collection and analysis and to the staff of the Animal Research Unit for the care of the animals.
FOOTNOTES
1Supported by the BBSRC grant 42/S/16622. 
Correspondence: 2Correspondence and current address: R.S. Robinson, School of Veterinary Medicine and Science, University of Nottingham, Sutton Bonington Campus, Loughborough, LEICS LE12 5RD, United Kingdom. FAX: 44 115 951 6415; e-mail: Bob.Robinson{at}nottingham.ac.uk
Received: 10 July 2006.
First decision: 14 August 2006.
Accepted: 13 February 2007.
REFERENCES
This article has been cited by other articles:
|
|
 |
|
  R S Robinson, A J Hammond, G E Mann, and M G Hunter A novel physiological culture system that mimics luteal angiogenesis Reproduction, March 1, 2008; 135(3): 405 - 413. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
