Histological analysis of arteriovenous anastomosis-like vessels established in the corpus luteum of cows during luteolysis
The mechanisms regulating the function and regression of the corpus luteum (CL) have not yet been elucidated in detail. The regressed CL of cows was previously reported to be filled with unusual vessels like arteriovenous anastomosis (AVA); however how these vessels are being established during luteolysis remains unknown.
The bovine CL at different luteal stages and regressing bovine CL induced by prostaglandin F2α (PGF) were histologically analyzed using light and electron microscopic levels. The changes in mRNA expression of genes encoding α-smooth muscle actin (SMA; Acta2) and transforming growth factor β1 (Tgfb1) in luteal tissues were analyzed by quantitative RT-PCR.
AVA-like vessels appeared in the regressed CL with a diameter less than 1.5 cm in which no functional luteal cells and macrophages were observed. Epithelioid cells in the AVA-like vessel wall were immunoreactive for SMA, and the lumen of the vessels were narrow. Immunoreaction for SMA was found in the tunica media of typical arteries and arterioles, and pericytes around capillary vessel. Cells with elongated cytoplasmic processes ―resident fibroblasts expressing vimentin― distributed in the CL parenchyma without any association with blood vessels are also immunoreactive for SMA, and accumulated around arteries and arterioles during the late-luteal stage. In the regressed CL, walls of arteries and arterioles consisted of more than two layers of epithelioid cells positive for both SMA and desmin, suggesting that they are myofibroblasts transformed from fibroblasts. The percentage of the area positive for SMA and the mRNA expression of Acta2 were significantly increased in the regressed CL; however, they did not alter when a luteolytic dose of PGF was injected in vivo and collected within 24 h after the injection. On the other hand, Tgfb1, a known regulator for myofibroblast transformation, was significantly increased in PGF-induced regressing CL as well as in the CL during the late-luteal stage.
SMA-positive myofibroblasts accumulates around the arteries and arterioles to form AVA-like vessels during luteolysis in cows. PGF indirectly regulates myofibroblast transformation through enhancing the expression of TGFβ1. These peculiar AVA-like vessels may be involved in the regulation of blood flow in the bovine CL during luteolysis.
KeywordsLuteolysis Arteriovenous anastomosis Myofibroblasts PGF2α TGFβ1
α-smooth muscle actin
Transmission electron microscope
Transforming growth factor β1
The corpus luteum (CL) is formed from an ovulated dominant follicle in the ovary, and secretes progesterone, which is essential for the establishment and maintenance of pregnancy. In mammals, the CL undergoes luteolysis at the end of luteal phase . There are two phases in luteolysis: functional and structural luteolysis. Functional luteolysis is characterized by a decrease in serum progesterone concentrations and is followed by structural luteolysis, in which the volume of the CL decreases due to apoptotic death of luteal cells. It is generally accepted that spontaneous luteolysis in cows is initiated by uterus-derived prostaglandin F2α (PGF) if pregnancy does not occur . In cows, serum concentration of progesterone markedly decreases approximately 19 days after ovulation and this is accompanied by an increase in plasma PGF metabolites, which represents the production of PGF in the uterus . An intramuscular injection of PGF into cows during the mid-luteal stage ―but not the early-luteal stage― has been shown to induce luteolysis, resulting in a rapid decrease in serum progesterone levels eight hours after the injection [3, 4]. Although PGF is a major luteolysin in cows, the molecular mechanisms by which PGF regulates functional and structural luteolysis currently remain unclear; however, its effects are known to be mediated by various intra-ovarian factors, such as endothelin 1, nitric oxide, angiotensin II, and interferon γ [5, 6, 7, 8, 9, 10, 11].
In cows, ovarian blood flow decreases during luteal regression [4, 12] and the resulting hypoxic condition affects functional and structural luteolysis  although a temporal increase in blood flow has been reported in spontaneous and PGF-induced luteolysis . Previous studies noted morphological changes in the vasculature during luteolysis in ewes  and cows . Sawyer et al.  found that vascular occlusion occurred following the sloughing of endothelial cells into the lumina of small blood vessels during luteolysis, suggesting that vascular occlusion causes a decreased blood supply, resulting in hypoxic conditions in the CL. Hojo et al.  demonstrated that capillaries disappeared within the CL at the late and regressed luteal stages, whereas larger blood vessels that stained for α-smooth muscle actin (SMA) increased in number. Changes in the vascular structure have been suggested to contribute to a decrease in blood flow and luteolytic process in the CL.
Although remodeling of the vasculature in the CL appears to be essential in the regulation of luteal function during luteolysis, the mechanisms responsible for controlling changes in vascular structures in the bovine CL during luteolysis have yet to be clarified. Höflinger  noted the existence of arteriovenous anastomosis (AVA)-like vessels in the regressed CL of cows. AVA is characterized by precapillary connections between small arteries and veins. AVA called Hoyer-Grosser’s organs are distributed in the subcutaneous tissue of the tips of the fingers and toes of humans, and have also been reported in the ears of some animals including rabbits. The corpora cavernosa penis of various animals possesses a similar structure in the helicine arteries [17, 18, 19]. AVA is a vascular shunt and is involved in the regulation of blood flow in the skin or corpora cavernosa penis. Without stimulations by vasodilative molecules such as nitric oxide secreted from peripheral nerves, AVA contracts and blood flow is remarkably obstructed. AVA shows a characteristic morphology that is markedly different from those of normal arteries and veins: unique round epithelioid cells form the walls of AVA. Epithelioid cells are contractive cells containing SMA in the cytoplasm, and react to both vasoconstrictive and vasodilative signals such as endothelin and nitric oxide .
Because the blood flow is known to decrease during luteolysis and the appearance of AVA-like vessels is likely to contribute to a decrease in blood flow in the CL of cows during luteolysis; however, the role of AVA-like vessels in regulating the function of CL and how this structure is constructed in the CL remain unclear. The objective of this study is to clarify the contribution of AVA-like vessels to the regulation of luteal function in cows, we herein investigated how they are formed in the bovine CL during luteolysis using histological techniques.
Ovaries with the CL of Holstein cows were obtained at a local abattoir, kept in normal saline at 25 − 30 °C, and delivered to the laboratory within 6 h of being collected. Luteal stages were classified as early-luteal (days 2–3 after ovulation), mid-luteal (days 8–12), late-luteal (days 15–17), and regressed (over day 19) by macroscopic observations of the CL according to a previous study . The stages of the bovine CL were classified into the early-luteal (n = 7), mid-luteal (n = 13), late-luteal (n = 14), and the regressed stages (n = 31). Collection of the CL tissues was approved by the local ethical committee of Hokkaido University (Approval no. 13–0052).
Information on the bovine corpus luteum used for the present study
A: Normal cycling CL
Number of samples
For the histological analysis
For the qPCR analysis
Days after ovulation
Ocher, Brown, White
1.9 ± 0.19
2.5 ± 0.07
2.1 ± 0.14
0.9 ± 0.06
B: PGF-induced regressing CL
Hours after the PGF injection
Number of samples
For the histological analysis
For the qPCR analysis
CL tissues dissected out from the ovaries were halved, and either immersed in fixatives as described below for the histological analysis or frozen in liquid nitrogen for RNA purification. Information on the CL used for this study was summarized in Table 1.
CL tissues were dissected from the ovaries and immersed in 4 % paraformaldehyde for 24 h at 4 °C. Tissues were dehydrated through graded series of ethanol and immersed into xylene, then embedded in paraffin according to the standard methodology, and 5-μm-thick sections were obtained. Sections were de-waxed, subjected to Hematoxylin and Eosin staining, and observed under a light microscope (BX51; Olympus corporation, Tokyo, Japan). The middle CL parenchyma tissues were dissected into small pieces and fixed with 2.5 % glutaraldehyde for ultrastructural observations under a transmission electron microscope (TEM, H-7100; Hitachi, Tokyo, Japan). Tissue pieces were post-fixed with 1 % OsO4 for 1.5 h, dehydrated through a graded series of ethanol, and embedded into Epon resin (Quetol 812). Ultra-thin sections were obtained, stained with uranyl acetate lead citrate, and observed under a TEM.
All procedures were performed at room temperature. After deparaffinization, endogenous peroxidase was blocked by 3 % H2O2 in distilled water for 20 min. Sections were then incubated with Avidin/Biotin blocking solution (Vector Laboratories Inc., Burlingame, CA, USA) for 15 min each. After washing with phosphate-buffered saline (PBS) twice, sections were pre-incubated for non-immune blocking with Block Ace (DS Pharma Biomedical Co., Ltd., Osaka, Japan) for 60 min. Sections were then incubated with a mouse anti-SMA antibody (1:1,000; clone 1A4, #A2574; Sigma-Aldrich Corporation, St. Louis, MO, USA) in PBS overnight. The sections were incubated with biotinylated anti-mouse IgG (1:500; Vector Laboratories Inc.) for 1 h, followed by an incubation using a Vectastain ABC Elite kit (Vector Laboratories Inc.) for 1 h. The reaction site of the primary antibody was visualized using an ImmPACT™ DAB Peroxidase Substrate Kit (Vector Laboratories Inc.) for 3 min. Sections were counterstained with hematoxylin and observed under a light microscope. The total CL area in the section and the SMA-immunoreactive area in the CL were measured using Image J (http://imagej.nih.gov/ij/), and the percentage of SMA-positive area per the total CL area was calculated. Control sections were incubated with PBS instead of the primary antibody and the disappearance of the signals was confirmed.
For luteal steroidogenic cell, macrophage, and proliferating cell staining, antigen retrieval was performed for 1 min in boiling 0.01 M citrate buffer (pH 6.0) using a pressure cooker. Endogenous peroxidase and Avidin/Biotin blocking were then performed as described above. After a pre-incubation with Block Ace for 1 h, the sections were incubated with either goat anti-3β-hydroxysteroid dehydrogenase (3β-HSD) (1:2,000; #sc-30820; Santa Cruz Biotechnology, Inc., Dallas, TX, USA), mouse anti-MAC387 (1:1,000; #ab22506; Abcam Japan Ltd., Tokyo, Japan), which recognizes the L1 or Calprotectin molecule, an intracytoplasmic antigen expressed by tissue macrophages, or rabbit anti-Ki-67 antibody (1:200; ThermoFisher Scientific Inc., Waltham, MA, USA) overnight. Sections incubated with the 3β-HSD antibody were then incubated with a peroxidase-labeled anti-goat IgG (1:200; Vector Laboratories Inc.) for 1 h. On the other hand, MAC387- and Ki-67-stained sections were incubated with biotinylated anti-mouse or rabbit IgG for 1 h and Vecterstain ABC Elite kit as described above. Reaction sites for all three antibodies were visualized using an ImmPACT™ DAB Peroxidase Substrate Kit (Vector Laboratories Inc.) for 3 min and observed under a light microscope as described above. The total CL area in the section and the number of 3β-HSD- or MAC387- immunoreactive cells in the CL was measured using Image J, and the positive cell number in 1 mm2 was calculated.
For double immunostaining for SMA and Ki-67, the deparafinized sections were rinsed in distilled water and antigen retrieval was performed by boiling the sections in 0.01 M citrate buffer (pH 6.0) for 1 min using a pressure cocker. The sections were incubated with Block Ace for 1 h and rabbit anti-Ki-67 antibody overnight as described above. After washing with PBS three times, the sections were incubated with Alexa Fluor 488-conjugated anti-rabbit IgG (1:200; Life Technologies Japan, Tokyo, Japan) for 2 h. The sections were again incubated with Block Ace for 1 h and then incubated with mouse anti-SMA antibody as described above. The binding site for SMA antibody was visualized using AlexaFluor 594-labeled anti-mouse IgG (1:200; Life Technologies Japan) and observed under a confocal laser microscope (FV300; Olympus, Tokyo, Japan).
For triple staining, sections were incubated with a mouse anti-vimentin antibody (1:500; Dako Japan ltd., Tokyo, Japan) overnight, and then with biotinylated anti-mouse IgG for 1 h as described above. The antibody-binding site for the 1st primary antibody against vimentin visualized using the TSA™ Fluorescein System (PerkinElmer Inc., Waltham, MA, USA). The 1st primary antibody was washed out in boiling 0.01 M citrate buffer (pH 6.0) using a microwave for 2.5 min, and sections were then blocked with Block Ace for 1 h. They were incubated with the 2nd primary antibody, a rabbit anti-desmin antibody (1:500; ThermoFisher Scientific Inc.) overnight and with a Cy5-labeled anti-rabbit antibody (1:200; Jackson ImmunoResearch Laboratories, Inc., West Grove, PA, USA) for 2 h. Sections were again blocked with Block Ace for 60 min and then incubated with a 3rd primary antibody, the mouse anti-SMA antibody (1:500). The binding site for the 3rd primary antibody was visualized by AlexaFluor 594-labeled anti-mouse IgG (1:200; Life Technologies Japan). Sections were observed under a confocal laser microscope as described above. Lasers used in this study were Ar (488 nm), HeNe (G) (543 nm), and HeNe (R) (633 nm).
Quantitative PCR (qPCR)
Primers used for the present study
After testing for normality, all statistical analyses were performed using one-way ANOVA with Tukey’s multiple comparisons test using GraphPad Prism 6 software (GraphPad Software Inc., San Diego, CA, USA), and P < 0.05 was regarded as significant. Values in the graphs represent the mean ± SEM.
Characteristic AVA-like vessels appear in the regressed CL of cows
SMA-positive cells increase in number during luteolysis
SMA-positive cells are myofibroblasts originating intraluteal fibroblast
PGF-induced TGFβ1 may contribute to myofibroblast transformation
This study revealed the detailed morphology of AVA-like vessels appearing in the regressed bovine CL and how they are constructed during luteolysis. The regressed CL of cows contained abundant atypical arteries with thick walls of epithelioid cells that were strongly immunoreactive for SMA. These SMA-positive cells forming the walls of AVA-like vessels are myofibroblasts transformed from resident fibroblasts during luteolysis: SMA-expressing fibroblastic cells with elongated cytoplasmic processes increased in number during the late-luteal stage and accumulated around the arteries to form AVA-like vessels. Luteolytic PGF is unlikely to be a direct cause of this phenomenon, but indirectly regulate the transformation of fibroblasts into myofibroblasts by enhancing the expression of TGFβ1, a potent regulator for myofibroblast transformation.
It is well known that the CL development is associated with angiogenesis and angiogenic factors such as vascular endothelial growth factor A and fibroblast growth factor 2 are involved in this process . Endothelial cells proliferate in the developing CL and a fine network of capillary vessels is established in the CL at the mid-luteal stages accompanied by an increased serum progesterone level [28, 29, 30]. During luteal regression, apoptosis of luteal cells as well as endothelial cells occur and the number of capillary vessels decreases [31, 32], suggesting a dynamic vascular remodeling is involved in luteolytic process. We confirmed an existence of apoptotic cells in some CL at the late-luteal and early regressed stage, and certain numbers of apoptotic cells were endothelial cells (data not shown). However, capillaries were more visible when pericytes were immunostained for SMA in the CL at the late-luteal stage. We still do not know the role of SMA-positive pericytes in the regulation of luteolysis. On the other hand, some researchers noted an appearance of non-capillary vessels with thickened wall in the regressing CL of various animals including cows [15, 23, 33], that show similar morphology to AVA-like vessels observed in this study.
The appearance of AVA-like vessels in the regressed bovine CL was previously reported by Höfliger ; in stages more than 25 days after ovulation, the CL consisted of crooked arteries with the characteristics of AVA. Bauer et al.  also noted an increased number of non-capillary vessels with thickened vessel walls in the bovine CL. Hirschberg et al.  more recently reported the existence of AVA-like arteries with thickened medial layers showing SMA immunoreactivity during the later luteal stage in cows; however, they reported that “the source of this abundance of vascular smooth muscle-like cells is presently not clear”. We revealed here that SMA-positive fibroblastic cells increased in number in the CL during the late-luteal stage, and accumulated around the arteries distributed in the periphery of the CL and intraluteal connective tissue. It is important to note that a certain population of SMA-immunoreactive cells was positive for vimentin in the CL during the late-luteal stage and proliferation marker, Ki-67, was not associated with an increase of SMA immunoreaction, suggesting that they are resident fibroblasts lying between luteal steroidogenic cells. As luteolysis progressed, the immunoreaction for SMA was restricted to the vascular epithelioid cells with a strong immunoreaction for desmin, a smooth muscle cell-specific intermediate filament. The arterial walls at this stage became thicker and the lumen became narrower, indicating a contractive condition. It is reasonable to assume that the origin of SMA-positive epithelioid cells in AVA-like vessels is myofibroblasts transformed from resident fibroblasts in the CL parenchyma. The AVA-like vessels appearing in association with luteolysis may assist luteolysis by decreasing luteal blood flow in cows.
Ovarian blood flow is known to decrease during luteolysis [4, 12]. Miyamoto et al.  clearly demonstrated this change in blood flow during spontaneous and PGF-induced luteolysis in cows. According to their findings, the concentrations of plasma PGF metabolites, indicating the production of uterine PGF, significantly increased 17 days after ovulation, and luteal blood flow concomitantly increased. They also showed that an intramuscular injection of PGF significantly decreased luteal blood flow after 8 h, although a temporal increase of luteal blood flow was observed after 1 h. In the spontaneously regressing CL, luteal blood flow decreases two days after a peak in plasma PGF metabolite concentrations . The accumulation of SMA-positive myofibroblasts around the arteries in the CL during the late-luteal stage and following the establishment of AVA-like vessels in the regressed CL may be involved in decreasing blood flow during spontaneous luteolysis in cows. However, we failed to detect AVA-like vessels in the present PGF-induced regressing CL collected 2–24 h later. We consider that the AVA-like vessels are gradually established in the CL over the repeated estrous cycle because the CL with AVA-like vessels we classified into regressed CL contain the old regressed CL formed more than two cycle before and remained in the ovary. The formation of AVA-like vessels may be observed in the PGF-induced regressing CL collected more than 48 h after a PGF injection.
PGF is a major luteolysin in cows, but does not appear to directly regulate the transformation of fibroblasts to myofibroblasts because the present study failed to show the accumulation of SMA-positive cells or the up-regulation of Acta2 mRNA expression with PGF-induced luteolysis. TGFβ1 is known to affect the transformation of fibroblasts into myofibroblasts. The results of the present study revealed that the administration of a luteolytic dose of PGF significantly enhanced the expression of Tgfb1 in the bovine CL. This is consistent with a previous study showing that PGF up-regulated the expression of TGFβ1 . TGFβ1 has been suggested to activate fibroblasts in order to enhance the expression of contractile proteins such as SMA, vimentin, and desmin. Maroni and Davis  demonstrated that the treatment of luteal fibroblasts with TGFβ1 induced the expression of laminin, collagen type 1, and matrix metalloproteinase 1, suggesting that TGFβ1 stimulated by PGF is involved in fibrotic changes in the CL during luteolysis. TGFβ1 may influence steroidogenic actions in the bovine CL similar to other TGFβ superfamily members such as activin A and bone morphogenetic proteins [37, 38, 39]. Nevertheless, TGFβ1 appears to contribute to the transformation of myofibroblasts during luteolysis in cows in order to establish AVA-like vessels.
In conclusion, SMA-positive myofibroblasts accumulate around arteries to form AVA-like vessels in the regressing CL of cows. This morphological change in the vasculature appears to be regulated by PGF-induced TGFβ1, and may be involved in vascular occlusion and a reduction in blood flow during luteolysis in cows.
The authors appreciate to Prof. Seiji Katagiri, Associate Prof. Masahi Nagano, and the members of the Laboratory of Theriogenology, Graduate School of Veterinary Medicine, Hokkaido University, as well as Dr. Dariusz J. Skarzynski, Institute of Animal Reproduction and Food Research, Polish Academy of Science, for their kind help with collecting bovine CL tissues.
This work was supported by a Grant-in-Aid for Scientific Research (C) (General) (#26460262) to JNK.
Availability of data and materials
Data supporting the findings of this manuscript has been included in the manuscript.
JNK conceived and designed the experiments. JNK and KM carried out the experiments, analyzed the data. KH and KO provided the PGF-treated bovine CL samples and technical support. JNK and TI wrote the manuscript and all authors read and approved the final manuscript.
The authors declare that they have no competing interests.
Consent for publication
Ethics approval and consent to participate
- 11.Shirasuna K, Asaoka H, Acosta TJ, Wijayaqunawardane MP, Matsui M, Ohtani M, Miyamoto A. Endothelin-1 within the corpus luteum during spontaneous luteolysis in the cow: local interaction with prostaglandin F2a and angiotensin II. J Cardiovasc Pharmacol. 2004;44 Suppl 1:S252–5.CrossRefPubMedGoogle Scholar
- 16.Höflinger H. Das Ovar des Rindes in den verschiedenen Lebensperioden unter besonderer Berücksichtigung seiner funktionellen Feinstruktur (The functional fine structure of the bovine ovary in its different cyclic conditions). Acta Anat. 1947;3 suppl 5:1–196.Google Scholar
- 35.Hou X, Arvisais EW, Jiang C, Chen DB, Roy SK, Pate JL, Hansen TR, Rueda BR, Davis JS. Prostaglandin F2alpha stimulates the expression and secretion of transforming growth factor B1 via induction of the early growth response 1 gene (EGR1) in the bovine corpus luteum. Mol Endocrinol. 2008;22:403–14.CrossRefPubMedGoogle Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.