Cell proliferation and vascular morphology in the marmoset corpus luteum

F.M. Young1,3, F.E. Rodger1, P.J. Illingworth2 and H.M. Fraser1

1 MRC Reproductive Biology Unit, Centre for Reproductive Biology, 37 Chalmers Street, Edinburgh EH3 9EW, UK, and 2 Department of Obstetrics and Gynaecology, Westmead Hospital, University of Sydney, NSW, Australia


    Abstract
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Luteal formation is associated with angiogenesis and low progesterone production. Maximal mid-luteal phase progesterone production is concurrent with extensive vascularization, and luteolysis occurs when steroidogenesis decreases. Angiogenic cell proliferation and vascular changes have not been examined in the marmoset. The aim of this study was to examine vascular morphology throughout the luteal phase by identifying: (i) von Willebrand factor VIII antigen (vW)-immunopositive endothelial cells; (ii) Ki67-positive proliferating cells; and (iii) bromodeoxyuridine-positive proliferating cells. Marmoset corpora lutea were examined thoughout the cycle, and natural regression was compared with induced luteolysis after administration of a prostaglandin F2{alpha} analogue or gonadotrophin-releasing hormone (GnRH) antagonist. Steroidogenic and endothelial cells were positive for proliferation markers. Endothelial cell proliferation was highest during luteal formation, then decreased and remained low during the luteal phase and functional regression, however endothelial cell proliferation increased during structural regression. Endothelial cell proliferation was unchanged by induced regression. The area of vW immunostaining was highest during luteal formation, decreased thereafter and remained constant during the luteal phase and regression. Distribution of immunostaining indicated the presence of an extensive capillary network, but during structural regression the numbers of capillaries decreased and numbers of microvessels increased. These results suggest that vascular changes are concurrent with changes in the functional status of the marmoset corpus luteum.

Key words: angiogenesis/BrdU/corpus luteum/endothelial cells/Ki67


    Introduction
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
The corpus luteum of the ovary secretes progesterone which is essential for the maintenance of early pregnancy. If fertilization does not occur, progesterone production declines resulting in functional luteal regression and menstruation. Removal of luteal tissue, or structural luteal regression, occurs after progesterone production has ceased. Luteal function is related to vascular status: maximum progesterone production does not occur until the vasculature is fully formed (McClellan et al., 1975Go), and it is possible that changes in the vasculature may also instigate or regulate luteal regression (Fraser et al., 1998Go). Therefore, it is important to examine the vasculature in order to fully understand the mechanisms that control luteal function.

The high rates of endothelial cell proliferation which occur during formation of the corpus luteum have been well characterized in ruminants (Jablonka-Shariff et al., 1993Go; Zheng et al., 1994Go), macaques (Christenson and Stouffer, 1996Go) and women (Rodger et al., 1997Go). These studies also demonstrated that proliferation decreases as the luteal phase proceeds, and remains low during functional luteal regression. In ruminant corpora lutea the network of fine capillaries required for optimal production of steroid hormones during the luteal phase (Redmer and Reynolds, 1996Go) is replaced by larger microvessels (Fraser et al., 1998Go) during luteal regression (Modlich et al., 1996Go). However, luteolysis is regulated by different mechanisms in primates and ruminants, and there are no data regarding cell proliferation or vascular status during structural luteal regression in primates.

Induction of luteolysis with pharmacological agents has a number of advantages over the collection of spontaneously regressing luteal tissue; tissue can be collected at known time points after induction of luteolysis, and different luteolytic agents can be used in order to gain insight into the mechanisms controlling luteal involution. Luteal regression can be induced in marmosets in two ways; by administration of a gonadotrophin-releasing hormone (GnRH) antagonist, which deprives the corpus luteum of luteinizing hormone (LH) support (Hodges et al., 1988Go), or by administration of a prostaglandin (PG) F2{alpha} analogue, which is thought to act directly at the level of the corpus luteum (Summers et al., 1985Go)

This study examined cell proliferation and vascular morphology in a commonly used non-human primate model, the marmoset monkey, and focused on cell proliferation and vascular changes during spontaneous and induced luteal regression. Specific aims were to: (i) identify cell proliferation using Ki67 and bromodeoxyuridine (BrdU); (ii) investigate changes in vascular morphology by determining the temporal and morphological distribution of endothelial cells using von Willebrand factor VIII antigen (vW) as a marker; and (iii) identify proliferating cells by co-labelling with BrdU and the steroidogenic cell marker 3ß-hydroxysteroid dehydrogenase (3ßHSD), or the endothelial cell marker vW.


    Materials and methods
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Animals and treatments
Female marmoset monkeys (Callithrix jacchus) were housed at the MRC Reproductive Biology Unit Primate Centre and experiments were carried out in accordance with the Animals (Scientific Procedures) Act 1986. Blood samples taken on alternate days by femoral venepuncture and ovarian cycles were monitored by measuring the plasma progesterone concentration (Smith et al., 1990Go). The follicular phase was defined as that period during which progesterone concentrations were <=12 nmol/l. Day 1 of the luteal phase was assumed to be the day on which the progesterone concentration rose to >32 nmol/l and was followed by a sustained elevation in progesterone. Ovaries containing corpora lutea were collected for a previously described study (Young et al., 1997Go) during the early luteal phase on days 2–4 (n = 3 animals), during the mid-luteal phase on day 10 (n = 4 animals), during functional luteal regression on day 18 (n = 4 animals) and during structural luteal regression on days 22–25 (n = 3 animals). Ovaries were also collected from two animals late in the follicular phase when corpora lutea were 26–28 days old.

Additional animals were administered i.v. BrdU (Boehringer Mannheim, Lewes, Sussex, UK) by slow infusion at a dose of 20 mg/ml saline and ovaries were collected 1 h later. Ovaries containing corpora lutea were collected on luteal days 2–4 (n = 3), during functional regression on luteal day 18 (n = 3) and during structural regression on luteal days 22–25 (n = 3).

Marmosets were also treated on day 9 of the luteal phase with either a GnRH antagonist, Antarelix® (N-Ac-D-Nal1,D-pCl-Phe2, D-Pal3,D-(hic)6,-Lys(iPr)8,D-Ala10; Deghenghi et al., 1993) 1 mg/kg s.c (Young et al., 1997Go), or a PGF2{alpha} analogue cloprostenol (Planate®, Coopers Animal Health Ltd, Crewe, Cheshire, UK) 1 µg i.m.

Ovaries containing corpora lutea were collected 12 h (n = 3 per treatment) or 24 h (n = 4 per treatment) after administration of luteolytic agents. PGF2{alpha} does not induce luteal regression during the early luteal phase, and is only luteolytic after luteal days 7/8. GnRH antagonist is anti-steroidogenic during the early luteal phase and was therefore used as a comparison for induction of luteolysis. The mid-luteal phase was chosen to avoid superimposing induced luteolysis on spontaneous, naturally occurring luteal regression. Corpora lutea collected 48 h after administration of PGF2{alpha} on luteal days 8 and 9, looked very similar to corpora lutea collected 24 h after treatment with PGF2{alpha}.

Ovaries were fixed in 4% (w/v) buffered paraformaldehyde (BDH Laboratories Ltd, Poole, Dorset, UK) for 24 h prior to embedding in paraffin according to standard procedures. Sections (4 µm) were mounted on poly-L-lysine coated (Sigma, Poole, Dorset, UK) slides, air-dried at 56°C and subjected to immunohistochemistry.

Immunohistochemistry
Endothelial cells were identified as previously described (Rodger et al., 1997Go) but with the following modifications. Primary antibody (von Willebrand factor (vW), polyclonal rabbit anti-human immunoglobulin G (IgG) (Dako, High Wycombe, Bucks, UK) was diluted 1:250 in Tris-buffered saline (TBS, 0.05 mol/l Tris, NaCl 9 g/l) and negative control sections were treated with normal rabbit serum (NRS) (SAPU, Carluke, UK) also diluted 1:250 in TBS before being incubated for 18 h at 4°C. Visualization was with avidin–biotin complex conjugated to alkaline phosphatase (ABC–AP, Vector Laboratories, Peterborough, UK) and Vector Red (Vector Laboratories) according to the manufacturer's instructions.

Analysis of proliferation depends on assessment of the number of cells undergoing mitosis in a given population, but the likelihood of observing mitotic spreads in histological sections is very low (O'Shea et al., 1986Go). More sensitive techniques for evaluating proliferation exploit characteristics of the cell cycle. Ki67 antigen is a nuclear, non-histone protein routinely used to identify proliferating cells (Gerdes et al., 1984Go; Bruno and Darzynkiewicz, 1992Go). Ki67-positive cells were identified as described previously (Rodger et al., 1997Go). Primary antibody (Ki67 antigen, monoclonal mouse anti-recombinant protein, Novocastra, Peterborough, UK) was diluted 1:50 in TBS and negative control sections were treated with mouse IgG (Vector Laboratories) also diluted 1:50 in TBS before incubation for 18 h at 4°C. Detection of the primary antibody was carried out in the same manner as for vW.

DNA synthesis occurs before mitosis, and BrdU can be incorporated into new DNA in place of thymidine and visualized using standard immunocytochemical techniques (Boulton and Hodgson, 1995Go). BrdU incorporation in vivo was used to confirm and extend data about proliferation rates obtained by Ki67 immunohistochemistry. Sections were deparaffinized in xylene and hydrated in 100, 95 and 70% ethanol. Endogenous peroxidase activity was blocked by incubating in 3% (v/v) hydrogen peroxide in methanol (BDH Laboratories Ltd) for 20 min at room temperature. BrdU antigen retrieval was by exposure to two 5 min cycles of microwave irradiation at 700 W in glycine–HCl buffer pH 7.4 (0.2 mol/l glycine, 0.2 mol/l HCl). Buffer was replenished to the original volume after each 5 min cycle of microwaving. Primary antibody (mouse monoclonal antibody clone BMC 9318 to BrdU, Boehringer Mannheim) was diluted 1:30 in TBS. Negative control sections were treated with mouse immunoglobulin G (Vector Laboratories) which was also diluted 1:30 in TBS before all sections were incubated for 18 h at 4°C. After washing with TBS, rabbit anti-mouse IgG (Dako) was applied at a dilution of 1:60 in blocking medium (20% NRS, 5% bovine serum albumin w/v in TBS) for 30 min at room temperature. This was followed by three washes with TBS and incubation with mouse peroxidase anti-peroxidase (mouse PAP; Dako) diluted 1:100 in TBS for 30 min at room temperature. Visualization was by application of 3,3'-diaminobenzidine tetrahydrochloride (DAB; Sigma Chemical Co Ltd, 0.04% w/v and 0.01% hydrogen peroxide in 0.05 mol/l Tris).

Proliferating endothelial cells were identified by vW and BrdU dual immunostaining. Von Willebrand immunocytochemistry was as described above, sections were then washed in water before proceeding with BrdU detection as described above.

The nature of BrdU-positive non-endothelial cells was investigated by dual immunostaining with 3ßHSD primary antibody (polyclonal rabbit anti-human 3ßHSD, gift of Prof Van Luu-The, the CHUL Research Centre, Quebec, Canada) in order to determine whether these cells were steroidogenic. BrdU detection proceeded as described above. After visualization, sections were washed with water before application of the 3ßHSD primary antibody which was diluted 1:300 in TBS. Negative control sections were treated with NRS (SAPU) also diluted 1:300 in TBS. Primary antibody and negative control sections were then incubated for 18 h at 4°C. Visualization was with ABC-AP and Vector Red (Vector Laboratories) made up according to the manufacturer's instructions. Visualization substrates can sometimes mask the colour of a second dye used in co-localization studies. Vector Red is fluorescent and not masked by DAB.

Analysis of data
Area of vW immunostaining was measured using NIH Image 1.57 software which converted the colour image to black and white. The optical density value that corresponded to the lightest vW immunostaining was determined for each section, then all areas above that threshold were measured. Negative control sections were measured in the same way, and areas with optical density values equal to or higher than the threshold for vW immunoreactivity were subtracted from the mean value for the corresponding experimental section. Four fields of view of luteal tissue were randomly selected and measured for each ovarian section, then their mean taken as being representative for that animal. Measuring more than four unit areas, from different ovarian sections, did not significantly alter the mean. The area of vW immunoreactivity (µm2) was calculated for each animal and expressed as mean area ± SEM for each experimental group.

Ki67 immunocytochemistry was quantified using a highly optimized microscope environment (Brugal et al., 1992Go) which was used previously and allowed comparison with data from an earlier study (Young et al., 1997Go). Four fields of view of luteal tissue were randomly selected and the mean of the four unit areas was taken as being representative for that animal. Measuring more than four unit areas from serial sections did not alter the calculated mean. Cells were classified as being either steroidogenic or non-steroidogenic, and also Ki67-positive or negative. Cells were classified as being steroidogenic or non-steroidogenic by morphological, not functional, criteria. Steroidogenic cells were >12 µm in diameter with regular circular outlines, abundant cytoplasm and circular nuclei. Cells with these morphological characteristics were also 3ßHSD-positive in this study. Non-steroidogenic cells were elongated with a long axis of 4–8 µm, large nuclei and little cytoplasm. Cells with this morphology were also vW-positive in this study, so the majority of these were endothelial cells, and a minority were other cell types; fibroblasts, pericytes and immigrant leukocytes (Lei et al., 1991Go).

The numbers of Ki67-positive steroidogenic and non-steroidogenic nuclei per unit area were expressed as a mean ± SEM for each experimental group. The mean percentages ± SEM of Ki67 positive steroidogenic and non-steroidogenic cells were also calculated for each experimental group.

BrdU immunohistochemistry was quantified by scoring every cell in the field of view of a x40 objective lens as being either steroidogenic or non-steroidogenic (using the morphological criteria described above) and either BrdU-positive or -negative. Four fields of view were scored for each corpus luteum, and the mean scores of the four unit areas were taken as being representative for that animal. Data were expressed as mean percentages ± SEM of BrdU-positive steroidogenic and non-steroidogenic cells for each experimental group.

BrdU and vW co-labelling was quantified by scoring every BrdU-positive cell in the field of view of a x40 lens, and classifying it as being either vW-positive or -negative. Four fields of view were scored for each corpus luteum, and the mean scores of four unit areas were taken as being representative for that animal. The mean ± SEM was then determined for each experimental group. BrdU and 3ßHSD quantification was obtained in the same way.

Progesterone assay
Progesterone concentrations were determined by radioimmunoassay as described previously (Smith et al., 1990Go). The means of each group were expressed ± SEM.

Statistical analysis
Data from untreated mid-luteal phase corpora lutea were compared with corpora lutea after induced regression with either PGF2{alpha} or GnRH antagonist in a one-way analysis of variance using treatment as a between subject variable. Differences in numbers of immunopositive cells in early, mid-, late and follicular phase corpora lutea were also investigated by one-way analysis of variance using stage of the luteal phase as a between subject variable. Confidence intervals for the differences between means were evaluated by using Fisher's least significant difference (LSD) procedure. P < 0.05 was considered to be significant.

Percentages of proliferating cells in corpora lutea from the early, mid-, late and follicular phases were also examined by one-way analysis of variance using stage of the luteal phase as a between subject variable. Percentage data were log-transformed before conducting one-way analysis of variance. Differences were evaluated by using Fisher's LSD procedure; P < 0.05 was considered to be significant. Mean differences in BrdU and vW co-localization were determined by paired t-tests, and P < 0.01 was considered to be significant.


    Results
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Progesterone concentrations
Table IGo shows the plasma progesterone values of animals in the early luteal phase.


View this table:
[in this window]
[in a new window]
 
Table I. Plasma progesterone values (nmol/l) in marmosets through the luteal phase and after luteolytic treatments
 
Von Willebrand factor VIII immunocytochemistry
Specific vW immunoreactivity was found in microvessels containing red blood cells (Figure 1BGo, luteal day 24) and in capillaries which did not contain erythrocytes (Figure 1AGo, luteal day 10). There was no vW immunoreactivity in either the steroidogenic cells or in the negative control sections.



View larger version (134K):
[in this window]
[in a new window]
 
Figure 1. Von Willebrand factor VIII antigen (vW) immunohistochemistry in marmoset corpora lutea. (A) Luteal day 10, (B) luteal day 24 (structural luteal regression). Specific vW immunostaining is black. (A) Staining localized between rounded steroidogenic cells (arrows). (B) Number of capillaries is decreased (arrows). Scale bars = 20 µm.

 
The pattern of vW staining in the early and mid luteal phases was consistent with an extensive capillary network in which every steroidogenic cell was in contact with the vasculature (Figure 1AGo). This was not the case in structurally regressing corpora lutea; immunostaining characteristic of small capillaries was decreased, but staining characteristic of microvessels, including arterioles and venules, was increased (Figure 1BGo). The area of immunoreactive vW staining in the early luteal phase (Figure 2Go) was significantly higher than at all other stages of the cycle. Luteal vasculature was further examined by determining the number of discrete areas of vW immunostaining at each stage of the luteal phase, and using the assumption that each discrete area of vW immunostaining represented one blood vessel. The number of blood vessels per unit area was highest during the early luteal phase then decreased to a baseline level during the mid luteal phase and functional regression (Figure 3Go). The number of individual blood vessels then decreased markedly as structural luteal regression progressed, although the area of vW immunostaining remained constant during both functional and structural regression.



View larger version (16K):
[in this window]
[in a new window]
 
Figure 2. Area of von Willebrand factor VIII antigen (vW) immunostaining in marmoset corpora lutea. Ovaries containing corpora lutea were collected on luteal days 2–4 (early, n = 3), luteal day 10 (mid, n = 4), luteal day 18 (functional luteal regression, n = 4), luteal days 22–25 (structural luteal regression, n = 3) and luteal days 26–28 (late structural regression, n = 2). The total area of vW specific immunostaining per unit area of luteal tissue was determined for each corpus luteum, and shown as mean ± SEM for each stage of the luteal phase. Early luteal phase staining was significantly higher than at any other time (**P < 0.001).

 


View larger version (11K):
[in this window]
[in a new window]
 
Figure 3. The number of blood vessels per unit area in marmoset corpora lutea. Each data point ({blacktriangleup}) represents the number of distinct areas of von Willebrand immunostaining per unit area of luteal tissue from one animal. Ovaries containing corpora lutea were collected on luteal days 2–4 (early, n = 3), luteal day 10 (mid, n = 4), luteal day 18 (functional luteal regression, n = 4), luteal days 22–25 (structural luteal regression, n = 3) and luteal days 25–28 (late structural regression, n = 2).

 
Induction of luteolysis by GnRH antagonist resulted in an increase in the area of vW immunoreactivity at 12 and 24 h while PGF2{alpha} was without a significant increase in vW-immunoreactivity at either 12 or 24 h after luteolytic treatment (Figure 4Go). The pattern of vW immunostaining after induced regression was slightly different from that seen in spontaneously regressing corpora lutea. There was a greater range of blood vessel size, with some sections having predominantly small capillaries similar to those seen in untreated mid luteal phase corpora lutea, and others having a combination of capillaries and microvessels.



View larger version (19K):
[in this window]
[in a new window]
 
Figure 4. Area of von Willebrand factor VIII antigen (vW) immunostaining in marmoset corpora lutea after induced luteal regression. Ovaries containing corpora lutea were collected on luteal day 10 (control, mid-luteal, n = 3), 12 h (PG12, n = 3) and 24 h (PG24, n = 4) after PGF2{alpha}, and 12 h (Gn12, n = 3) and 24 h (Gn24, n = 4) after gonadotrophin-releasing hormone (GnRH) antagonist. The area (mean ± SEM) of vW immunostaining 12 and 24 h after induction of luteolysis with GnRH antagonist was significantly higher than in controls (*P < 0.05).

 
Ki67 immunohistochemistry
Non-steroidogenic cells were elongated with a long axis of 4–8 µm, large nuclei and little cytoplasm. Ki67-positive non-steroidogenic cells had intense nuclear staining. Steroidogenic cells were >12 µm in diameter with regular circular outlines, abundant cytoplasm and circular nuclei. Ki67 immunostaining was also nuclear, but generally less intense than in non-steroidogenic cells (Figure 5Go). No staining was observed in negative control sections (Figure 5Go inset). The number of Ki67-positive non-steroidogenic cells per unit area was highest during the early luteal phase (13.9 ± 8.7) but decreased as the luteal phase proceeded and was lowest during functional luteal regression (4 ± 0.4) on luteal day 18. However, this increased significantly (P < 0.05) to 6 ± 0.6 positive non-steroidogenic cells per unit area during structural regression (Figure 6Go).



View larger version (104K):
[in this window]
[in a new window]
 
Figure 5. Ki67 immunocytochemistry in mid-luteal phase day 10 marmoset corpus luteum. Marmoset ovarian section shows luteal tissue and a follicle. Ki67 immunopositive cell nuclei are black, and negative cell nuclei are light grey. Open arrows = non-steroidogenic cells; solid arrow = steroidogenic cell; scale bar = 20 µm. Inset shows Ki67 antibody negative control.

 


View larger version (19K):
[in this window]
[in a new window]
 
Figure 6. Numbers of Ki67 immunopositive cells in spontaneously regressing marmoset corpora lutea. Data from luteal days 2–4 (early, n = 3), luteal day 10 (mid, n = 4), luteal day 18 (functional luteal regression, n = 4) and luteal days 22–25 (structural luteal regression, n =3). The total number of Ki67-positive steroidogenic (black bars) and non-steroidogenic (white bars) cells per unit area were determined for each corpus luteum, and shown as mean ± SEM for each stage of the luteal phase. Number of Ki67-positive non-steroidogenic cells on luteal day 18 significantly lower (*P < 0.05) than during luteal days 22–24. Ki67-positive steroidogenic cells significantly lower (**P < 0.01) on luteal day 18 than on luteal days 2–4.

 
The numbers of Ki67-positive steroidogenic cells per unit area was 18.6 ± 3.2 cells in control mid-luteal phase day 10 corpora lutea and fell to 2.42 ± 1.1 cells (P < 0.01) 24 h after administration of PGF2{alpha}, and 2.4 ± 0.7 cells (P < 0.01) 24 h post-GnRH antagonist. The number of Ki67-positive non-steroidogenic cells remained constant before and after induced luteal regression.

The percentage of steroidogenic cells which were Ki67 immunopositive did not change significantly through the luteal phase nor during spontaneous luteal regression (Figure 10Go). The percentage of Ki67-positive non-steroidogenic cells decreased significantly from 8.4 ± 0.8% during the early luteal phase, to 0.6 ± 0.3% (P < 0.01) during functional luteal regression and 2.2 ± 0.5% (n = 3, P < 0.01) during structural luteal regression.



View larger version (28K):
[in this window]
[in a new window]
 
Figure 10. Percentages (mean ± SEM) of Ki67 and BrdU-positive cells in spontaneously regressing marmoset corpora lutea. Data collected on luteal days 2–4 (early, n = 3), luteal day 18 (functional luteal regression, n = 3) and luteal days 22–25 (structural luteal regression, n = 3). The total number of Ki67-positive non-steroidogenic (black bars), BrdU non-steroidogenic (white bars), Ki67 steroidogenic (checked bars) and BrdU steroidogenic (striped bars) cells per unit area were determined for each corpus luteum, and expressed as a percentage of the total number of non-steroidogenic or steroidogenic cells for that unit area. BrdU **P < 0.001; Ki67 *P < 0.05.

 
BrdU immunocytochemistry
BrdU immunopositive steroidogenic cell numbers were low during the early luteal phase (4.3 ± 2.3, mean ± SEM), absent in functionally regressing corpora lutea (Figure 7Go), and high in some structurally regressing corpora lutea (43.2 ± 40). Non-steroidogenic cell numbers were 29 ± 6 in the early luteal phase, falling to 2.6 ± 0.8 during functional regression but increasing to 38 ± 28 during structural regression. The percentage of BrdU-positive non-steroidogenic cells was 8.36 ± 0.8 cells per unit area (n = 3, Figure 10Go) in the early luteal phase, 0.55 ± 0.3 (n = 3, P < 0.001) during functional luteal regression and 2.2 ± 0.5% (n = 3) during structural luteal regression.



View larger version (177K):
[in this window]
[in a new window]
 
Figure 7. Bromodeoxyuridine (BrdU) immunocytochemistry in functionally regressing marmoset corpus luteum. (a) Developing, antral and atretic follicles with BrdU-labelled granulosa cells; (b) primary antibody-negative control serial section of (a); and (c) luteal tissue from the same ovary. BrdU-immunopositive cell nuclei are black and immunonegative nuclei are grey. Scale bars = 20 µm.

 
BrdU co-localization with 3ßHSD
A proportion of cells with the morphological appearance of steroidogenic cells had pink 3ßHSD cytoplasmic immunostaining (Figure 8aGo). Brown BrdU-positive cells were not 3ßHSD immunopositive (Figure 8aGo). Figure 8bGo shows that BrdU immunopositive cells are not 3ßHSD immunofluorescent, and also that the nuclei of steroidogenic cells are 3ßHSD immunonegative as would be expected.



View larger version (138K):
[in this window]
[in a new window]
 
Figure 8. Identification of bromodeoxyuridine (BrdU) immunopositive proliferating cells in marmoset corpora lutea. Panels (a) and (b) are the light and fluorescence micrographs respectively of same field of view of a corpus luteum collected on luteal day 18 during functional luteal regression. (a) BrdU immunopositive nuclei are brown, and 3ß hydroxysteroid dehydrogenase (3ßHSD) immunopositive cells are pink–red. Inset is the antibody negative control for this section. (b) 3ßHSD-positive immunostaining also fluoresces red, but BrdU immunostaining is not visible. Arrows in (a) and (b) indicate BrdU-positive cells that are 3ßHSD immunonegative. (cf) BrdU co-localized with the endothelial cell marker, von Willebrand factor VIII antigen (vW). (c) Ovary collected in the early luteal phase with a follicle to the left and luteal tissue to the right. The avascular granulosa layer of the follicle contains BrdU immunopositive proliferating cells, but no vW immunostaining. Both BrdU and vW immunostaining are apparent in the theca layer of the follicle and in the corpus luteum. Arrow indicates a BrdU-positive nucleus which is vW immunonegative. (d) Antibody-negative control serial section of (c). (e) and (f) Light and fluorescence micrographs respectively of same field of view. (e) Brown BrdU specific immunostaining, and pink–red vW specific immunostaining. (f) vW-positive immunostaining also fluoresces red, but BrdU immunostaining is not visible. (e) and (f) Arrows in BrdU-positive cells are vW immunonegative. Scale bars = 20 µm.

 
BrdU co-localization ICC with von Willebrand factor VIII antigen
Pink vW immunopositivity was predominantly cytoplasmic and localized to the edges of lumina which contained red blood cells (Figure 8c,eGo), as well as to luteal cells which had the morphological appearance of endothelial cells but which did not clearly enclose a lumen. Positive vW immunostaining was also apparent in the vascular theca layer, but not the avascular granulosa layers of follicles (Figure 8cGo). Brown BrdU immunostaining was localized to cell nuclei, and was found in granulosa, theca and luteal cells. BrdU immunopositive cells were both vW immunofluorescent and vW immunonegative (Figure 8fGo), indicating that only a proportion of the BrdU-positive cells were endothelial cells.

In the early luteal phase, 64.6% of BrdU-positive cells were also vW immunopositive (17 ± 3.5, mean ± SEM, Figure 9Go). This decreased significantly to 0.42 ± 0.2 (P < 0.001) during functional regression, then returned to values similar to those seen in the early luteal phase (15.7 ± 2.5).



View larger version (16K):
[in this window]
[in a new window]
 
Figure 9. Numbers of proliferating cells after co-localization immunohistochemistry with bromodeoxyuridine (BrdU) and von Willebrand factor VIII antigen (vW). Data from luteal days 2–4 (early, n = 3), luteal day 18 (functional luteal regression, n = 3) and luteal days 22–25 (structural luteal regression, n = 3). Cells positive for both BrdU and vW (black bars) or immunopositive for BrdU only (white bars). Mean ± SEM number of BrdU-positive cells at each stage of the luteal phase. *P < 0.001.

 
Comparison of Ki67 and BrdU
In order to compare assessments of proliferation by Ki67 and BrdU immunocytochemistry, data were expressed as percentages (Figure 10Go). The percentage of BrdU-positive cells was always lower than the percentage of Ki67-positive cells, except in structurally regressing corpora lutea. The numbers of Ki67 and BrdU immunopositive non-steroidogenic cells paralleled each other reasonably well, but percentages of Ki67-positive steroidogenic cells were higher than percentages of BrdU-positive steroidogenic cells. The percentage of proliferating non-steroidogenic cells was highest during the early luteal phase, significantly decreased during functional luteal regression, and increased during structural luteal regression.


    Discussion
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
This study has demonstrated that there is intense angiogenesis during formation of the marmoset corpus luteum, that 3ßHSD-positive steroidogenic cells do not proliferate; that there is a population of proliferating cells within the corpus luteum which are not endothelial cells; and that the number of proliferating cells decreases concomitant with functional luteal regression. In addition we have shown that the process of natural luteolysis is associated with formation of large blood vessels, a phenomenon that does not occur following induced luteolysis.

Immunocytochemistry for vW indicated that during the early and mid-luteal phase vascular morphology was consistent with an extensive network of fine capillaries with a high surface area to volume ratio. During structural luteal regression, however, the area of vW immunostaining remained constant whilst the number of individual blood vessels per unit area decreased. This was based on the assumption that each distinct area of vW immunopositive staining represented one blood vessel, even though it was noted that the same blood vessel may intersect the histological plane of section any number of times, and that some endothelial cells may have been vW immunonegative, thus leading to an underestimation of the amount of vasculature present. Despite these limitations, these data support the hypothesis that there is an initial period of intense angiogenesis during formation of the corpus luteum, then a period of vascular regression during functional luteolysis. Endothelial cell proliferation increases during structural regression, suggesting a second period of angiogenesis in order to form larger blood vessels suitable for channelling away the cellular debris of structural regression. Macrophages do not proliferate in the human corpus luteum (Gaytán et al., 1998Go) therefore it is likely that the proliferating non-endothelial, non-steroidogenic cells are fibroblasts. In human corpora albicans, the endothelial area of each vessel increases and the number of blood vessels decreases in comparison with mid-luteal phase values (Suzuki et al., 1998Go), indicating a change in human luteal vasculature during structural regression consistent with our data.

The administration of luteolytic agents resulted in functional luteal regression, unchanged proliferation and the maintenance of an extensive capillary network. In contrast, untreated corpora lutea had functional regression accompanied by decreased proliferation, followed by structural regression characterized by a decrease in the number of capillaries and an increase in the number of larger microvessels. These differences may be related to the level of PGF2{alpha} receptor expression. Human PGF2{alpha} receptor mRNA was lowest during the mid-luteal phase, and increased as the corpus luteum aged (Ottander et al., 1999Go). This suggests that PGF2{alpha} receptor levels during luteal day 10 in marmosets may have been sufficient to facilitate the inhibition of steroidogenesis, but not high enough to cause regressive structural changes characterized by decreased proliferation and vascular remodelling.

Oxytocin and oxytocin receptor numbers peak in primate corpora lutea during the mid-luteal phase (Khan-Dawood and Dawood, 1999). Oxytocin stimulates the production of PGF2{alpha}, and also stimulates endothelial cell production of endothelin-1 (ET-1) (Girsch et al., 1996). ET-1 inhibits steroidogenic enzymes and progesterone production in rat granulosa cells (Tedeschi et al., 1992) and inhibits progesterone production by human luteal cells in vitro (Apa et al., 1998Go). ET-1 has been demonstrated in porcine (Flores et al., 1995Go) and human (Apa et al., 1998Go) corpora lutea and values in bovine corpora lutea were highest during regression (Girsh et al., 1996Go). It is possible that in primates, oxytocin stimulates PGF2{alpha} and ET-1 production during the mid-luteal phase. PGF2{alpha} generates the oxygen free radical, hydrogen peroxide (Riley and Behrman, 1991Go) which may trigger a cascade of oxygen free radical reactions capable of inhibiting LH-stimulated steroidogenesis in a very short period of time. Mid-luteal phase production of ET-1 might also inhibit progesterone production, and augment the anti-steroidogenic effects of PGF2{alpha}. Both PGF2{alpha} and ET-1 concentrations increase, but oxytocin production decreases, as the corpus luteum ages, suggesting that although oxytocin might be an instigating factor in functional regression, it is not required to maintain the process.

The number of proliferating non-steroidogenic cells did not change at any time after induction of luteolysis, therefore the significant increase in the area of vW immunoreactive staining after administration of GnRH antagonist can probably be attributed to the dynamics of shrinking tissue. In this scenario the steroidogenic cells die as a result of GnRH antagonist induced LH-depletion (Young et al., 1997Go), and the remaining blood vessels move closer together, effectively increasing the number of blood vessels in each unit area.

A proportion of cells with the morphological appearance of steroidogenic cells were both Ki67 and BrdU-immunopositive. This was unexpected, since a number of studies have demonstrated that steroidogenic cell numbers do not increase in bovine (O'Shea et al., 1989Go), ovine (O'Shea et al., 1986Go) or human (Lei et al., 1991Go) corpora lutea.

This suggests that steroidogenic cells may be in the cell cycle, and thus expressing the cell cycle antigen Ki67, but that they may be arrested in one part of the cell cycle and do not proceed to DNA synthesis or mitosis. Others have described cells with the morphological features of steroidogenic cells positive for proliferation markers in the bovine (Zheng et al., 1994Go), ovine (Jablonka-Shariff et al., 1993Go), macaque (Christenson and Stouffer, 1996Go) and human (Rodger et al., 1997Go; Gaytán et al., 1998Go) corpora lutea, and that these cells did not co-label with the steroidogenic cell marker 3ßHSD. This suggests that cells with the morphological appearance of steroidogenic cells may not produce steroid hormones and express cell cycle antigens simultaneously. Our data support this hypothesis: BrdU-positive cells with the morphological appearance of steroidogenic cells were 3ßHSD-negative. Similarly, the majority of Ki67-positive granulosa–lutein and theca–lutein cells in human corpora lutea did not co-label with 3ßHSD (Rodger et al., 1997Go). In the present study, the number of Ki67-positive steroidogenic cells was decreased after induction of luteolysis with either GnRH antagonist or PGF2{alpha}, whereas the number of Ki67-positive non-steroidogenic cells remained unchanged. This differential regulation of cell types by luteolytic agents requires further investigation.

High variability within experimental groups may be attributed to the lack of differentiation between the younger (ovulated more recently) and older corpora lutea. In marmosets, ovulation can occur as much as 24 h apart in the same cycle and, therefore, the value derived from one ovary encompassed variablity in luteal age. Our cycle tracking regime was accurate ± 24 h; therefore an ovary collected on day 4 of the luteal phase may in fact have been 3 or 5 days post-ovulation/LH. The Ki67 value derived from each experimental group therefore encompassed variability attributable to cycle tracking. In addition, the large SEM in the early luteal phase group probably reflected the degree of differentiation/luteinization in individual cells, and variable maturation of the corpora lutea. The large SEM in 20–25 day old corpora lutea probably reflected variability in the length of the luteal phase. Marmosets had a luteal phase of 20 ± 10 days (mean ± SD, n = 87, range 6–63 days). The median length of the luteal phase was 18 days, but functional regression could start from day 12 onwards, and progression of functional and structural regression was highly variable. Ovaries collected on luteal day 22 could be beginning structural regression in one animal whilst another animal might have completed structural regression by luteal day 22.

Low numbers of steroidogenic cells were BrdU-positive in the early luteal phase, when granulosa or theca–lutein might conceivably be proliferating during the process of transformation from follicle to corpus luteum, but there were no BrdU-labelled steroidogenic cells in the late luteal phase, indicating that steroidogenic cells were not synthesising DNA at this time. The number of BrdU-labelled steroidogenic cells increased during structural luteal regression, but the number of Ki67-positive steroidogenic cells remained constant. One of the first stages of apoptosis can be activation of endogenous endonucleases and fragmentation of DNA, and it is possible that the very early stages of DNA fragmentation might also be accompanied by futile DNA repair processes. Apoptotic glandular prostate cells incorporate BrdU during futile DNA repair when the cells are in the G0 phase (Berges et al., 1993Go), therefore BrdU-labelled non-proliferating cells might be observed in tissues in which there are also high rates of apoptosis. It is possible that a proportion of the BrdU-labelled steroidogenic and non-steroidogenic cells observed in structurally regressing corpora lutea are not actually proliferating, but are in the initial stages of cell death.

In summary, it is possible to draw three main conclusions: (i) morphological changes in luteal vasculature do not precede functional luteal regression in marmoset corpora lutea. Vascular changes occur after functional regression is completed and progesterone concentrations have fallen to follicular phase values; (ii) the vasculature changes from an extensive network of fine capillaries during the luteal phase to a system of larger blood vessels during luteal regression; and (iii) one component of functional luteal regression is a dramatic reduction in the rate of proliferation.


    Acknowledgments
 
The authors thank K.D.Morris and staff for animal care, I.Swanston and staff for progesterone assays, M.Miller and S.McPherson for skilled technical support, Dr R.Deghenghi (Europeptides) for the gift of Antarelix, and the Department of Pathology, University of Edinburgh, for the use of HOME analysis facilities.


    Notes
 
3 To whom correspondence should be addressed at: Department of Biotechnology, Flinders University of South Australia, Flinders Medical Centre, Bedford Park, Adelaide, South Australia Back


    References
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Apa, R., Miceli, F., de Feo, D. et al. (1998) Endothelin-1: expression and role in human corpus luteum. J. Reprod. Immunol. (Copenhagen), 40, 370–376.

Berges, R.R, Furuya, Y., Remington, L. et al. (1993) Cell proliferation, DNA repair, and P53 function are not required for programmed death of prostatic glandular cells induced by androgen ablation. Proc. Nat. Acad. Sci. USA, 90, 8910–8914.[Abstract]

Boulton, R.A, and Hodgson, H.J.F (1995) Assessing cell proliferation: a methodological review. Clin. Sci., 88, 119–130.[ISI][Medline]

Bruce, N.W. and Hillier, K. (1974) The effect of prostaglandin F2{alpha} on ovarian blood flow and corpora lutea regression in the rabbit. Nature, 249, 176–177.[ISI][Medline]

Brugal, G. Dye, R. Krief, B. et al. (1992) HOME: The highly optimized microscope environment. Cytometry, 13, 109–116.[ISI][Medline]

Bruno, S. and Darzynkiewicz, Z. (1992) Cell cycle dependent expression and stability of the nuclear protein detected by Ki67 antibody in HL-60 cells. Cell Prolif., 25, 31–40.[ISI][Medline]

Clement, P.B. (1987) Histology of the ovary. Am. J. Surg. Pathol., 11, 277–303.[ISI][Medline]

Christenson, L.K. and Stouffer, R.L. (1996) Proliferation of microvascular endothelial cells in the primate corpus luteum during the menstrual cycle and simulated early pregnancy. Endocrinology, 137, 367–374.[Abstract]

Deghengi, R., Boutignon, F., Wuthrich, P. et al. (1993) Anterelix (EP 24332) a novel water soluble LHRH antagonist. Biomed. Pharmacother., 47, 107–110.[ISI][Medline]

Flores, J.A., Winter, T.A., Knight, J.W. and Veldhuis, J.D. (1995) Nature of endothelin binding in the procine ovary. Endocrinology, 136, 5014–5019.[Abstract]

Fraser, H.M., Lunn, S.F., Kim, H. et al. (1998) Insulin-like growth factor binding protein-3 mRNA expression in endothelial cells of the primate corpus luteum. Hum. Reprod., 13, 2180–2185.[Abstract]

Gaytán, F., Morales, C., García-Pardo, C. et al. (1998) Macrophages, cell proliferation and cell death in the human menstrual corpus luteum. Biol. Reprod., 59, 417–425.[Abstract/Free Full Text]

Gerdes, J., Lemke, H., Baisch, H. et al. (1984) Cell cycle analysis of a cell proliferation-associated human nuclear antigen defined by the monoclonal antibody Ki67. J. Immunol., 133, 1710–1715.[Abstract/Free Full Text]

Girsh, E., Wang, W., Mamluk, R. et al. (1996) Regulation of endothelin-1 expression in the bovine corpus luteum: elevation by prostaglandin F2 alpha. Endocrinology, 137, 5189–90.[ISI][Medline]

Hodges, J.K., Green, D.I., Cottingham, P.G. et al. (1988) Induction of luteal regression in the marmoset monkey (Callithrix jacchus) by a gonadotrophin-releasing hormone antagonist and the effects on subsequent follicular development. J. Reprod. Fertil., 82, 743–752.[Abstract]

Jablonka-Shariff, A., Grazul-Bilska, A.T., Redmer, D.A. et al. (1993) Growth and cellular proliferation of ovine corpora lutea throughout the estrous cycle. Endocrinology, 138, 1871–1879.[Abstract/Free Full Text]

Khan-Dawood, F.S. and Dawood, M.Y. (1998) Comparative aspects of oxytocin in baboon (Papio hamadryus anubis) and human corpora lutea. Hum. Reprod. Update, 4, 371–382.[Abstract/Free Full Text]

Lei, Z.M., Chegnini, N. and Rao, C.V. (1991) Quantitative cell composition of human and bovine corpora lutea from various reproductive states. Biol. Reprod., 44, 1148–1156.[Abstract]

McClellan, M.C., Diekman M.A., Abel, J.H. et al. (1975) Luteinising hormone, progesterone and the morphological development of normal and superovulated corpora lutea in sheep. Cell Tiss. Res., 164, 291–307.[ISI][Medline]

Modlich, U., Kaup, F.-J. and Augustin, H. (1996) Cyclic angiogenesis and blood vessel regression in the ovary: blood vessel regression during luteolysis involves endothelial cell detachment and vessel occlusion. Lab. Invest., 74, 771–780.[ISI][Medline]

O'Shea, J.D., Rodgers, R.J. and Wright, P.J. (1986) Cellular composition of the sheep corpus luteum in the mid and late luteal phases of the oestrous cycle. J. Reprod. Fertil., 76 685–691.

O'Shea, J.D., Rodgers, R.J. and D'Occhio, M.J.D. (1989) Cellular composition of the cyclic corpus luteum of the cow. J. Reprod. Fertil., 85, 483–487.[Abstract]

Ottander, U., Leung, C. and Olofsson, J.I. (1999) Functional evidence for divergent receptor activation mechanisms of luteotrophic and luteolytic events in the human corpus luteum. Mol. Hum. Reprod., 5, 391–395.[Abstract/Free Full Text]

Redmer, D.A. and Reynolds, L.P. (1996) Angiogenesis in the ovary. Rev. Reprod., 1, 182–192.[Abstract/Free Full Text]

Riley, J. and Behrman, H. (1991) In vivo generation of hydrogen peroxide in the rat corpus luteum during luteolysis. Endocrinology, 128, 1749–1753.[Abstract]

Rodger, F.E., Young, F.M., Fraser, H.M. and Illingworth, P.J. (1997) Endothelial cell proliferation follows the mid-cycle luteinising surge, but not human chorionic gonadotrophin rescue, in the human corpus luteum. Hum. Reprod., 12, 1723–1729.[Abstract]

Smith, K.B, Lunn, S.F and Fraser, H.M. (1990) Inhibin secretion during the ovulatory cycle and pregnancy in the common marmoset monkey. J. Endocrinol., 126, 489–495.[Abstract]

Summers, P.M., Wennink, C.J. and Hodges, J.K. (1985) Cloprostenol induced luteolysis in the marmoset monkey (Callithrix jacchus). J. Reprod. Fertil., 73, 133–138.[Abstract]

Suzuki, T., Sasano, H., Takaya, R. et al. (1998) Cyclic changes of vasculature and vascular phenotypes in normal human ovaries. Hum. Reprod., 13, 953–959.[Abstract]

Tedischi, C., Hazum, E., Kokia, E. et al. (1992) Endothelin-1 as a luteinisation inhibitor: inhibition of rat granulosa cell progesterone accumulation via selective modulation of key steroidogenic steps affecting both progesterone formation and degradation. Endocrinology, 131, 2476–2478.[Abstract]

Young, F.M., Illingworth, P.J., Lunn, S.F. et al. (1997) Cell death during luteal regression in the marmoset monkey (Callithrix jacchus). J. Reprod. Fertil., 111, 109–119.[Abstract]

Zheng, J., Fricke, P., Reynolds, L.P. et al. (1994) Evaluation of growth, cell proliferation and cell death in bovine corpora lutea throughout the estrous cycle. Biol. Reprod., 51, 623–632.[Abstract]

Submitted on July 1, 1999; accepted on November 19, 1999.