Spatially regulated differentiation of endometrial vascular smooth ...

Human Reproduction vol.15 no.2 pp.284–292, 2000

Spatially regulated differentiation of endometrial vascular smooth muscle cells

Gaby Kohnen1,2,3, Steven Campbell1,5, Michael D.Jeffers2,4 and Iain T.Cameron1 1Department

of Obstetrics and Gynaecology, 2Departments of Pathology, Glasgow Royal Infirmary, 10 Alexandra Parade, Glasgow G31 2ER, 3Leeds General Infirmary, Leeds, UK and 4The Adelaide and Meath Hospital, Dublin, Ireland 5To

whom correspondence should be addressed

Angiogenesis within the human endometrium involves the development of arterioles and elaboration of a capillary network. It was postulated that maturation of these arterioles involves a spatially regulated process of vascular smooth muscle cell (VSMC) differentiation. The endometrial vascular tree was therefore examined immunohistochemically for evidence of longitudinal and radial gradients of VSMC phenotype. Twenty-three hysterectomy specimens and 15 first trimester decidual tissues were studied. Five cytoskeletal markers (α and γ-smooth muscle (sm) actin, sm myosin, desmin, vimentin), three endothelial markers (CD31, CD34, factor VIII related antigen) and two steroid receptors (oestrogen and progesterone) were detected immunohistochemically. α-sm actin was present throughout the wall of basal arterial segments and extended longitudinally towards the endometrial surface. Sm myosin expression was more restricted longitudinally and radially within in the vascular tree. The expression of γ-sm actin was even more restricted than myosin. In first trimester decidua, however, γ-sm actin was widely distributed within the wall of spiral arteries that were not invaded by trophoblast. Oestrogen and progesterone receptors were present in peri-vascular stromal cells but absent from vascular smooth muscle and endothelium. Endometrial VSMC differentiation involves a progressive increase in cytoskeletal complexity and occurs in a spatially regulated fashion. Key words: angiogenesis/differentiation/endometrium/smooth muscle/vascular

Introduction Growth of the endometrial vascular tree after menstruation involves elongation and muscularization of the arterioles and expansion of the capillary network (Boyd and Hamilton, 1970; Ramsey, 1977; Ramsey and Donner, 1980). The endometrium is therefore a convenient tissue in which to study a variety of vascular phenomena occurring during the process of physiological angiogenesis (for reviews see Folkman and D’Amore, 1996; Folkman, 1997; Hanahan, 1997; Risau, 1997; Rogers et al., 1998; Smith, 1998; Yancopoulos et al., 1998). 284

Although considerable effort has been made to describe the growth of the microvascular component of the endometrial vascular tree either during the normal menstrual cycle or after exposure to synthetic steroids (Rogers et al., 1993; Goodger and Rogers, 1994, 1995; Wingfield et al., 1995; Kooy et al., 1996; Palmer et al., 1996; Rogers et al., 1998), less is known about the elaboration of the larger diameter muscularized vessels and the accompanying differentiation of the perivascular cells in this tissue (Sheppard and Bonnar, 1979; Kaiserman-Abramof and Padykula, 1989; Abberton et al., 1996, 1999; Rogers, 1998). The endometrial arterioles originate as branches of the radial arteries within the inner myometrium (Markee, 1940; Ramsey, 1977; Rogers, 1998). As the endometrium thickens during the course of the menstrual cycle, the arterioles extend longitudinally through the endometrial stroma in a highly oriented fashion towards the luminal surface of the uterus. These arterioles develop a spiral configuration. A new cycle of vessel growth commences at menstruation when the more luminal portion of the endometrium is shed (Kooy et al., 1996; Rogers et al., 1998). The remaining portion, basal endometrium, is a densely cellular stroma containing the stumps of tubular epithelial glands and remnants of the vascular tree (Noyes et al., 1950). A new surface epithelium regenerates by migration of cells from the glandular stumps (Ferenczy et al., 1976a,b). The tissue then begins to thicken as a consequence of oestrogen-driven cellular proliferation within the tubular epithelial glands and stromal fibroblasts during the pre-ovulatory phase (Noyes et al., 1950; Jansen and Johannisson, 1985; Li et al., 1988). Proliferation, particularly within the epithelial compartment, declines rapidly following exposure of the tissue to progesterone. In contrast, the proliferation of vascular endothelium is not limited to a particular stage of the menstrual cycle and is therefore not thought to be under the direct control of either oestrogen or progesterone (Rogers et al., 1998). Nevertheless, convolution and thickening of the arterioles becomes more obvious in tissue section during the post-ovulatory secretory phase of the endometrial cycle (Markee, 1940; Ramsey, 1977; Sheppard and Bonnar, 1979). The muscularization of arterioles involves either differentiation from as yet unidentified precursor cells within the endometrial stroma or outgrowth from pre-existing vessels within the basal part of the tissue. Molecular examination of the smooth muscle cell phenotype throughout the cycle and in early pregnancy should therefore provide clues about the nature of the differentiation pathway. Within the embryo and within in-vitro models, the differentiation of vascular smooth muscle cells (VSMC) from their precursors is correlated with distinct temporal expression © European Society of Human Reproduction and Embryology

Smooth muscle cell differentiation

patterns of smooth muscle-specific contractile proteins and extracellular matrix (Kocher et al., 1985; Thyberg et al., 1991; Burke et al., 1994; Hungerford et al., 1996; Hirschi et al., 1998; Landerholm et al., 1999; for review of contractile proteins see Owens, 1995). As the smooth muscle cells differentiate they undergo a progressive transformation that results in the expression of more muscle-specific proteins. These temporal patterns of expression have also been observed during experimentally induced intimal thickening (Kocher et al., 1986, 1991), and in cultures of rat aortic smooth muscle cells (Owens et al., 1986). The earliest known marker of differentiation in VSMC is α-smooth muscle (sm) actin, the induction of which is coincident with the recruitment of presumptive smooth muscle cell precursors into avian and quail embryonic vessel walls (Duband et al., 1993; Landerholm et al., 1999) and vessels of rat lung and primitive airway tubules (Mitchell et al., 1990). Other components such as calponin, h-caldesmon, smooth muscle αtropomyosin and γ-sm actin appear later during differentiation (Owens, 1995; Landerholm et al., 1999). The aim of the present study was to assess the differentiation of VSMC within the human endometrium during the course of the menstrual cycle. By examining the distribution of oestrogen and progesterone receptors, we also sought to obtain evidence about whether differentiation is directly driven by ovarian steroids. We postulated that VSMC differentiation occurs by a progressive specialization of the cytoskeleton, similar to that observed in early embryonic development (Owens, 1995). It was further predicted that because of the highly orientated nature of endometrial growth, spatially distinctive patterns of expression would arise in which early markers of differentiation would be more extensively distributed than later markers. In order to test this hypothesis, an immunocytochemical study was undertaken in which the spatial distribution of cytoskeletal proteins, endothelial markers and steroid receptors was examined in tissue sections. Marker distribution was assessed qualitatively along the vessel axis and radially within the vessel wall. Materials and methods Specimens Uterine biopsies, consisting of endometrium and attached myometrium, were obtained from twenty-three healthy women undergoing hysterectomy for benign disease at the West Glasgow Hospitals University NHS Trust and Glasgow Royal Infirmary NHS Trust, with the approval of local ethics committees. Informed consent was obtained in each case where an additional biopsy was obtained for the purpose of research. Histological assessment of endometrial tissue was carried out according to standard morphological criteria (More, 1987) and the stage of the menstrual cycle was noted. No specimen was found to exhibit evidence of neoplasia, hyperplasia or infection. In addition, we collected decidual tissue from 15 healthy women undergoing surgical termination of normal pregnancy during the 7th–11th week of gestation. Specimens were collected within 10 min following suction curettage, frozen in liquid nitrogen and stored at –70°C. Immunocytochemistry Antibodies to five cytoskeletal markers, three endothelial markers, and oestrogen and progesterone receptors were used in serial paraffin

sections of endometrium and myometrium (Table I). The hysterectomy specimens were fixed in 4% formaldehyde in phosphate buffer and embedded in paraffin. Serial sections (4 µm) were mounted on glass slides coated with 3-aminopropyltriethoxysilane (Sigma, Poole, UK) and deparaffinized using xylene and a graded series of alcohol concentrations. For detection of von Willebrand Factor sections were pretreated in 0.1% trypsin at 37°C for 15 min. In cases where high temperature antigen retrieval was used (Table I) the sections were microwave irradiated for 25 min (5⫻5 min) in a 600 Watt oven within a glass beaker containing approximately 1 l of 10 mmol/l citrate-buffer (pH 6.0). As detection of sm myosin was highly variable after 25 minutes of irradiation, the period of pre-treatment was increased to 40 min (8⫻5 min) for this antigen. The sections were then washed in Dulbecco’s phosphate buffered saline, pH 7.4 (PBS; ICN, UK) supplemented with 0.1% Triton⫻100 (Sigma) for 3 min followed by PBS without Triton for 3⫻5 min. The sections were pre-incubated for 20 min in PBS supplemented with 1.5% (w/vol) bovine serum albumin (BSA; Sigma). Primary and secondary antibodies were diluted in PBS supplemented with 1.5% BSA. After incubation with the primary antibodies, sections were washed in PBS-Triton (3 min) and PBS (3⫻5 min), and the appropriate biotinylated secondary antibodies, either rabbit anti-mouse (1:500, Dako, Ely, Cambridge, UK) or swine anti-rabbit (1:500, Dako), were applied for a further 30 min. Sections were then washed in PBS and incubated in 1% (vol/vol) H2O2 in absolute methanol for 10 min to inactivate endogenous peroxidase. The biotinylated bridging antibody was detected after incubation with streptavidin-peroxidase conjugate (StreptABComplex/HRP, Dako) for 20 min followed by 2.1 mg/ml 3,3 diaminobenzidine tetrahydrochloride (Sigma) and 0.3% (w/vol) H2O2 in Tris buffered saline pH 7.6. Sections were washed in distilled water, counterstained with Harris’ haematoxylin and mounted in DPX (BDH, Poole, Dorset, UK). Two negative controls were used in which the primary antibody was replaced with either PBS supplemented with 1.5% BSA or irrelevant antibodies (serum from non-immunized rabbits or a mouse monoclonal IgG1 antibody to Aspergillus niger glucose oxidase). Positive controls included sections from tonsil and large bowel for the endothelial markers and cytoskeletal proteins, and breast tissue for progesterone and oestrogen receptors. All immunolocalizations were characterized by intense staining and the negative controls were totally free of reaction product within both endometrium and myometrium for every marker. Cryostat sections (10 µm) of decidual tissue were fixed in acetone at 4°C for 10 min, air-dried and rinsed in PBS for 10 min. They were then pre-incubated for 20 min in PBS supplemented with 1.5% (w/v) BSA. Primary antibodies (vimentin, γ-sm actin, cytokeratin, and α2 laminin) were applied for 60 min and detected as described for paraffin sections.

Results Morphological examination of the uterine tissues revealed that four specimens were obtained during the proliferative phase of the cycle, seven were early secretory, four were late secretory and eight menstrual. Tissue distribution of cytoskeletal markers The distribution of each cytoskeletal marker was examined throughout the full thickness of the endometrium in order to determine vascular specificity. Sm myosin and γ-sm actin expression were restricted to vascular smooth muscle within the endometrium. α-sm actin was localized in VSMC and 285

G.Kohnen et al.

Table I. List of antibodies and utilization Antigen

Clone

Dilution

Pretreatment

Supplier

Reference

α-smooth muscle actin γ-smooth muscle actin Smooth muscle myosin (heavy chain) Desmin Vimentin Cytokeratin CD34 CD31 von Willebrand polyclonal Factor (factor VIII related antigen) Progesterone receptor Oestrogen receptor Laminin α2 light chain (anti-human merosin)

1A4 B4 HSM-V Polyclonal Polyclonal MNF116 QB-END/10 JC/70A

1:500 1:500 1:150 1:100 1:100 1:400 1:20 1:20 1:1000

n/a n/a Microwave n/a n/a n/a Microwave Microwave Trypsin

Sigma ICN Sigma Euro-patha Euro-patha Dako Novocastrab Dako Dako

Skalli et al., 1986 Lessard, 1988

1A6 1D5 5H2

1:20 1:50 1:900

Microwave Microwave n/a

Novocastra Dako Gibco BRLc

Ramani et al., 1990 Parums et al., 1990

Al Saati et al., 1993 Leivo et al., 1989

n/a ⫽ not applicable. aEuro-path, Bude, UK. bNovacastra, Newcastle-upon-Tyne, UK. cGibco BRL, Paisley, UK.

pericytes of capillaries and was found within stromal cells of the basal endometrium. Although on occasions this actin isoform was present in abundance around some cystically dilated glands, it was widely distributed in the stromal cells of the basal endometrium. Desmin was present within a few stromal cells in the basal region of the tissue and had a restricted distribution within vessels. Desmin was only present within a few centrally positioned VSMC in basal endometrium in the secretory phase. Vimentin was present in stroma, epithelium and VSMC throughout the menstrual cycle. Smooth muscle myosin and γ-sm actin were therefore the most specific vascular markers of the cytoskeletal proteins examined. Smooth muscle and vascular smooth muscle of the myometrium served as internal controls in every tissue section. These muscle types expressed each of the cytoskeletal markers without variation across the menstrual cycle, with the exception of desmin, which was present in myometrium but only in some of the outer muscle fibres of the myometrial vessels. CD31 and von Willebrand Factor were expressed only by endothelial cells of the endometrium and myometrium. The staining pattern of von Willebrand Factor was more intense than CD31 under the experimental conditions used. CD34 was found in endothelial cells throughout the vascular tree in endometrium and myometrium and was also expressed in endometrial stromal cells of the basal endometrium during the secretory and menstrual phases of the cycle. CD34 was more widely distributed in the stroma than γ-sm actin. In myometrium, CD34 reactivity was limited to perivascular stromal cells and spindle-shaped cells surrounding muscle cell bundles. The distribution of von Willebrand Factor was therefore used to assess endothelial cell distribution in relation to VSMC. The spatial distribution of cytoskeletal markers within endometrial blood vessels Three cytoskeletal proteins, α- and γ-sm actin and sm myosin, exhibited spatially distinctive patterns of expression within the endometrial vascular tree (Figures 1, 2 and 3). 286

α-sm actin was present throughout the vascular tree. It was thus present in large muscularized basal vessels close to the myometrium (Figure 3A) and in much smaller vessels including pericytes surrounding capillaries close to the luminal epithelium (Figure 2A). In large muscularized vessels α-sm actin was distributed throughout the entire thickness of the vessel wall. Qualitative assessment of α-sm actin distribution suggested that the muscularized layer of the spiral arteries of basal and superficial endometrium became thicker during the secretory phase. Sm myosin had a slightly more restricted distribution than α-sm actin (Figures 2A, B and 3A, B). Though VSMC of larger vessels usually co-expressed α-sm actin and sm myosin, the latter was not expressed around capillaries, where pericytes expressed only α-sm actin (Figure 2B). In larger arteries of the basal endometrium, the radial distribution of myosin within the vessel wall was also more restricted than that of α-sm actin (Figure 3B). Myosin distribution was more limited in the outer vessel wall than alpha actin. This situation was most apparent during the secretory phase of the cycle when the basal vessels became more muscularized. The vascular distribution of γ-sm actin was more basally restricted than either α-sm actin or sm myosin (Figures 1C, 2C and 3C). Although patchy in the outer vessel wall, the γ isoform was present throughout the entire thickness of the wall in basal vessels in the proliferative phase and was weakly expressed in muscularized vessels which were surrounded by decidualized stroma. γ-sm actin was absent from smaller muscularized vessels and capillaries throughout the stroma (Figures 1C and 2C). Steroid receptor distribution Oestrogen and progesterone receptor distribution was examined within the walls of large muscularized vessels where it was possible to compare the distribution of smooth muscle markers in consecutive or semi-consecutive sections (Figure 4). Neither of the receptors was found in VSMC or endothelial cells of endometrium and myometrium by the methods employed in

Smooth muscle cell differentiation

expressed in the stroma and glands comparable to the proliferative phase of the cycle. Decidual vessels in first trimester Highly decidualized areas of stroma were identified by laminin α2 distribution (Kohnen et al., 1998). These areas contained prominent muscularized spiral arterioles. These vessels did not contain extravillous trophoblast as assessed by cytokeratin expression. Cytokeratin expression was restricted to remaining endometrial glandular epithelial cells. The VSMC of these vessels expressed γ-sm actin throughout the entire thickness of the vessel wall (Figure 5).

Figure 1. Longitudinal distribution of cytoskeletal markers within myometrium, basal endometrium and inner functional endometrium in the proliferative phase of the cycle. (A) α-sm actin is expressed in myometrial smooth muscle cells, myometrial vascular smooth muscle, stromal cells of the basal endometrium (small arrows) and the vascular smooth muscle cells of both basal and superficial functional endometrium (large arrows). (B) Sm myosin is present in the myometrial smooth muscle (arrow) and extends into the smaller calibre vessels of the superficial functional endometrium. It is more widely distributed within the vascular tree than γ-sm muscle actin. Unlike α-sm actin, myosin is not expressed in the stromal cells of the basal endometrium. (C) γ-sm actin is more restricted than the α- isoform. It is absent from the stromal cells of the basal layer and is not expressed in smaller calibre vessels (arrows). (D) Vimentin is weakly expressed in myometrial smooth muscle cells and is widely distributed in the endometrium. It is present in the glandular epithelium, stroma and vascular smooth muscle. Bar ⫽ 110 µm.

this study. Both receptors were, however, present in myometrial smooth muscle. Comparison of α-sm actin and steroid receptor distribution in serial sections of thick-walled arteries demonstrated that stromal cells directly adjacent to the vascular smooth muscle expressed both nuclear receptors. Expression of nuclear receptors within stromal and epithelial cells varied during the course of the menstrual cycle as previously reported (Bergeron et al., 1988; Press et al., 1988). The exception was that in menstrual phase tissue of five specimens, oestrogen and progesterone receptors were

Discussion This study provides evidence that VSMC differentiation within the endometrial vascular tree occurs in a spatially organized fashion. As the endometrial vessel wall develops, α-smooth muscle actin is probably expressed in pericytes surrounding microvessels (Abberton, 1999) and presumptive smooth muscle cells. The relationship between endometrial pericytes and cells of VSMC phenotype remains unclear. Within larger muscularized vessels, sm myosin and γ-sm actin are also present. The differences in longitudinal and radial distribution, particularly between α- and γ-sm actin, imply that maturation of endometrial arterioles occurs in an outwardly radial direction and is propagated in a longitudinal manner towards the endometrial surface as the tissue grows. The process probably continues after implantation until the smooth muscle cells throughout the thickness of these small arterioles become fully differentiated. Whether or not other markers appear after γ-sm actin has yet to be established. The radial and longitudinal patterns that have been observed in this study are also present during embryonic arteriogenesis (Burke et al., 1994; Owens, 1995; Hungerford et al., 1996). The situation within the embryo, however, is likely to be dependent on vessel size and structure (Selmin et al., 1991; Burke et al., 1994; Hungerford et al., 1996). Large elastic vessels, such as the aorta, which have a more elaborate structure than the spiral arterioles, exhibit greater developmental complexity. Nevertheless, longitudinal proximo-distal spread of maturation markers within the aorta is analogous to the basal-to-superficial gradient described here within the endometrium. Radially propagated maturation in larger diameter muscular and elastic arteries occurs in both outward and inward directions. The direction of these radial gradients appears to be related to vessel size and type (Selmin et al., 1991). Within the avian outflow tract, which gives rise to the aorta and pulmonary arteries, gradients of cytoskeletal and extracellular matrix components have been examined (Burke et al., 1994; Hungerford et al., 1996). Although myosin and actin spread in an outwardly radial fashion, elastin, collagen I and collagen VI appear first in the outer vessel wall and spread inwards (Burke et al., 1994). These results suggest that where more than one clearly demarcated layer exists within the vessel wall, such as the adventia and media, there is the potential for a more developmentally complex sequence. In the case of the endometrium, where the muscularized vessels are small in 287

G.Kohnen et al.

Figure 2. The distribution of α-sm muscle actin, sm myosin and γ-sm actin in the immediate subluminal area of the superficial functional endometrium in the early secretory phase demonstrates a spatial restriction of γ-sm actin expression. (A) α-sm actin extends into the smallest calibre vessels (arrows). (B) Sm myosin shows a similar distribution to α-sm actin. (C) In contrast, γ-sm actin is absent from the smaller vessels. Bar ⫽ 45 µm.

diameter and relatively simple in wall structure, outwardly spreading radial change might be the dominant feature. The establishment of developmental axes and the existence of complex migratory patterns within the early embryo would seem to produce other differences from the pattern of angiogenesis observed here. The maturation of the quail aortic wall occurs in both a ventro-dorsal and a radial direction (Hungerford et al., 1996). It seems unlikely that physiological adult arteriogenesis and maturation would exhibit this developmental polarity, for the progenitor cells in the adult are probably present around the entire circumference of the maturing vessel. The curvilinear nature of the spiral arterioles raises the possibility that maturation does not occur in a radially even way. Despite this, preferential vascular smooth muscle differentiation, and deposition of matrix components at any given point along the length of the arteriole, might induce the type of kinking or spiralling observed in these vessels (S.Campbell and V.Mariatou, unpublished data). In addition to these spatial aspects of vascular maturation, development within the embryo and endometrium have temporal components (Owens, 1995; Landerholm et al., 1999). Quail pro-epicardial cells that originate on the surface of the heart as an epithelium undergo a mesenchymal transition, invade the matrix of the heart and differentiate into coronary artery smooth muscle cells (Landerholm et al., 1999). Isolated quail pro-epicardial cells undergo a series of temporally coordinated changes and thus progressively acquire a smooth muscle phenotype in culture. Changes in morphology, cytoskeletal organization and appearance of cytoskeletal markers occur in a regulated way. α-sm actin, the earliest marker to appear in vitro, is followed by γ-sm actin, calponin and then sm myosin. In-vivo and in-vitro evidence suggests that control of this expression pattern may be regulated by transcriptionally active serum response factor. This DNA binding protein may be involved in transcriptional activation during muscle differentiation and may be important in maintaining the expression of smooth muscle specific genes in the developing vessel (Landerholm et al., 1999). The origins of endometrial VSMC are at present unclear. 288

They could emerge from the walls of the basal portions of spiral arterioles that remain behind after menstruation. If this pattern of development occurred, the progenitor cells would be expected to migrate along the length of endothelial tubes as they elongate during the course of endometrial thickening. Alternatively, the presumptive smooth muscle cells may differentiate from surrounding stromal cells. One obvious possibility is that the α-sm actin-expressing stromal cells of the basal layer could be the source of such precursors. Regardless of the source of these cells, it is clear from the relative radial distribution of the cytoskeletal markers that the differentiation process must occur in situ in the vicinity of the endothelium. Although steroid receptors were not localized to vessels during the present study, initial formation and sprouting of endothelial tubes, recruitment of cells to the maturing vessel wall and their subsequent differentiation into smooth muscle cells is almost certainly regulated by a large variety of locally acting factors. The angiogenic growth factors include epidermal growth factor, transforming growth factor α (TGF-α), transforming growth factor β (TGF-β), tumour necrosis factor α (TNF-α), angiogenin, platelet-derived endothelial cell growth factor (PDGF), and vascular endothelial growth factor (VEGF; for review see Smith, 1998). In addition, the synthesis of angiopoeitin-1, angiopoeitin-2 and their tyrosine kinase receptor TIE2 could facilitate the maturation of vessels, the initial formation of which might have been influenced by VEGF and its receptors VEGF-R1 and VEGF-R2 (Folkman and D’Amore, 1996; Davis et al., 1996; Suri et al., 1996; Vikkula et al., 1996; Hanahan, 1997; Yancopoulos et al., 1998). It has been postulated, for example, that expression of the TIE2 ligand in mesenchymal cells that will be recruited to the wall of developing vessels is controlled by negative feedback from endothelial cells (Vikkula et al., 1996). The recruitment of presumptive smooth muscle cells to the vessel wall around endothelial tubes appears to be influenced by factors that promote cell migration, proliferation and differentiation. TGFα, which may be in this category, has been shown to promote the in-vitro smooth muscle differentiation of multipotent embryonic cells co-cultured with endothelium (Hirschi et al.,

Smooth muscle cell differentiation

Figure 3. Radial distribution of cytoskeletal markers within basal vessels of early secretory endometrium. (A) α-sm muscle actin is localized throughout the entire thickness of the vascular smooth muscle cell layer of the largest vessel. Smaller vessels and some stromal cells of the basal endometrium express this actin isoform. (B) Sm myosin is present throughout the entire wall of the largest vessel but is weaker in the outer longitudinal layer. Sm myosin is not expressed in endometrial stromal cells. (C) γ-sm actin is coexpressed in the inner muscle layer of the largest vessel and has a more patchy distribution in the outer longitudinal layer. This actin isoform has a more restricted distribution in small vessels (large arrows) and is absent from smallest diameter vessels (small arrows) and the endometrial stromal cells. (D) In the largest vessel, desmin is primarily expressed in the outer longitudinal muscle layer. Stromal cells in the late secretory phase of the endometrial cycle also express desmin. Bar ⫽ 60 µm.

1998). Similarly, PDGF-B antibodies prevent endothelial induced mesenchymal cell migration (Hirschi et al., 1998). Decidual spiral arteries play an important role in pregnancy when invading trophoblast transforms the muscular walls into large tortuous channels (Boyd and Hamilton, 1970; Pijnenborg, 1980, 1981, 1998; Benirschke and Kaufmann, 1995). This process is pronounced at the implantation site whereas the paraplacental or mural arteries elsewhere in the decidual tissue retain their structure (Ramsey and Donner, 1980). A recent study showed that decidual spiral artery remodelling, including muscular hypertrophy, begins before cellular interaction with cytotrophoblast (Craven et al., 1998). The authors found that decidual vessels exhibited more α-sm actin stain compared to non-pregnant secretory endometrial vessels. By contrast, we

Figure 4. Sm myosin, progesterone and oestrogen receptor distribution in large muscularized vessels of the basal endometrium. (A) Sm myosin is expressed in abundance in vascular smooth muscle of these basal vessels but is absent from the stroma. (B) Progesterone receptor is expressed in glandular epithelium and stromal cells but is absent from the vascular smooth muscle and endothelium. The oestrogen receptor was also absent from the vessel and had a comparable distribution to the progesterone receptor (not shown). Bar ⫽ 50 µm.

observed more pronounced immunostaining for γ-sm actin in decidual arteries (that were not yet invaded by trophoblast) than arteries of secretory endometrium. The present findings suggest that there is an ongoing process of vascular maturation that occurs as part of decidualization in addition to the remodelling promoted by trophoblast invasion. The absence of oestrogen and progesterone receptors from VSMC suggests that these ovarian steroids may not play a direct role in VSMC differentiation. Although a non-nuclear receptor-based mechanism cannot be excluded, and possible limits in the sensitivity of immunocytochemical detection cannot be ignored, it is notable that receptors can be detected in epithelial and stromal cells, which are known to be oestrogen and progesterone sensitive in vivo. Similarly, the absence of these nuclear receptors in the vascular endothelium supports the view of Rogers and colleagues (1998) that these cells, which proliferate throughout the course of the menstrual cycle, are not directly regulated by ovarian steroids. By contrast, both actin and progesterone receptors have been identified in the peripheral cells of small muscularizing vessels (Rogers et al., 1998). The apparent discrepancy between the receptor distribution described here and that described by Rogers (1998) may have an interesting biological basis. If smaller diameter thin walled superficial vessels such as those sampled by Pipelle biopsy (Abberton et al., 1996, 1999) contain VSMC at an earlier stage of differentiation they might still retain aspects of the precursor cell phenotype. The presence of progesterone 289

G.Kohnen et al.

Figure 5. Expression of γ-sm actin and laminin α2 in semi-parallel cryostat sections of first trimester decidua (10 weeks of gestation). (A) The prominent spiral arterioles express γ-sm actin throughout the entire vessel wall. (B) Fully decidualized tissue shows abundant laminin α2 light chain expression in both stromal cells and vascular smooth muscle cells. Bar ⫽ 20 µm.

receptors in thin walled vessels demonstrated by double staining with actin (Rogers 1998) raises the intriguing possibility that the smooth muscle cells arise by recruitment and differentiation of perivascular stromal cells that express this steroid receptor. The type of analysis carried out in the present study was not sufficiently precise to examine receptor distribution in such small vessels. The findings of Rogers nevertheless suggest that pericytes or VSMC during earlier stages of differentiation might be responsive to progesterone. This finding might suggest that in the early stage of precursor cell differentiation the cells might be responsive to progesterone and that the precursor cells may be stromal cells expressing the receptor. Paradoxically, progesterone might act on larger vessels outwith the uterus, for the receptor has been found in the medial layer of human and rat aorta and is co-localized with α-sm actin (Ingegno et al., 1988; Lee et al., 1997). Exposure of cultured rat aortic cells to progesterone inhibited thymidine incorporation and cell proliferation, suggesting that the receptor is functionally active (Lee et al., 1997). In addition to the vascular patterns of the smooth muscle markers, endometrial stromal cells of basal endometrium had immunocytochemically detectable vimentin and α-sm actin. In the present study, α-sm actin-containing cells were widely distributed, rather than specifically associated with glands as has been previously reported (Czernobilsky et al., 1993). Both 290

normal and neoplastic endometrial stromal cells may exhibit features of myofibroblasts or smooth muscle cells (Farhood et al., 1991; Franquemont et al., 1991; Lillemoe et al., 1991; Czernobilsky et al., 1993). A previous report suggested that expression of α-sm actin within the endometrial stroma was restricted to areas around cystically dilated glands of the basal endometrium, perhaps as a consequence of mechanical stress (Czernobilsky et al., 1993). By contrast, Franquemont et al. (1991), who investigated curettings obtained mostly from the superficial endometrium, found widespread distribution of αsm actin. We report here that numerous endometrial stromal cells express α-sm actin and vimentin throughout the menstrual cycle and that these cells are mostly limited to basal endometrium. The cytoskeletal co-expression of α-sm actin and vimentin in these cells is comparable to the pattern of expression seen in wound myofibroblasts (Darby et al., 1990). Myofibroblasts are not only responsible for synthesis and accumulation of extracellular matrix, but they are also thought to play an important role in wound contraction (Majno et al., 1971; Sappino et al., 1990; Schu¨ rch et al., 1992). Since these myofibroblasts appear restricted to the basal endometrium, they may have a mechanical or biochemical function that preserves the integrity of the basal endometrium at menstruation. Some basal cells also expressed CD34, which is found on endothelial cells of benign and malignant vascular lesions (Ramani et al., 1990; Anthony and Ramani, 1991; Cohen et al., 1993; Weiss and Nickoloff, 1993). CD34 has also been found on non-vascular soft tissue tumours (Ramani et al., 1990; Cohen et al., 1993), nerve cell tumours (Weiss and Nickoloff, 1993) and lipomatous tumours (Suster and Fisher, 1997). The precise nature of these non-vascular CD34-positive cells that resemble fibroblasts has not been completely elucidated although it has been suggested that they are capable of forming basement membrane (Weiss and Nickoloff, 1993). Recently it was postulated that these cells play a supportive role in the maturation or proliferation of adjacent stem cells (Weiss and Nickoloff, 1993; Suster and Fisher, 1997). The role of CD34 within the endometrial stroma is not clear but may be related to differentiation and maturation of endometrial glands and vessels. In summary, endometrial VSMC differentiate in a spatially regulated fashion analogous to the process that occurs in the vertebrate embryo. The marked difference in the distribution of the α- and γ-sm actin isoforms within the endometrial arterioles suggests that differentiation extends longitudinally from the basal to the superficial endometrium and proceeds in an outwardly radial direction within the vessel wall. Differentiation appears to occur in the absence of immunohistochemically identifiable oestrogen and progesterone receptors in the vascular cells and continues into early pregnancy, where the vessels undergo further maturation even in the absence of trophoblast invasion.

Acknowledgements The authors are grateful to gynaecology and midwifery colleagues at Glasgow Royal and Western Infirmaries for help in collecting speci-

Smooth muscle cell differentiation

mens. The excellent technical assistance of Mrs Anne Young and the technical staff in the Department of Pathology at the Royal Infirmary is acknowledged. G.K. was in receipt of a European Community Human Capital and Mobility Fellowship (no. 9030364).

References Abberton, K.M, Taylor, N.H., Healy, D.L. and Rogers, P.A.W. (1996) Vascular smooth muscle α-actin distribution around endometrial arterioles during the menstrual cycle: increased expression during the perimenopause and lack of correlation with menorrhagia. Hum. Reprod., 11, 204–211. Abberton, K.M, Taylor, N.H., Healy, D.L. and Rogers, P.A.W. (1999) Vascular smooth muscle cell proliferation in arterioles of the human endometrium Hum. Reprod., 14, 1072–1079. Al Saati, T., Clamens, S., Cohen-Knafo, E. et al. (1993) Production of monoclonal antibodies to human estrogen-receptor protein (ER) using recombinant ER (RER). Int. J. Cancer, 55, 651–654. Anthony, P.P. and Ramani, P. (1991) Endothelial markers in malignant vascular tumours of the liver: superiority of QB-END/10 over von Willebrand factor and Ulex europaeus agglutinin 1. J. Clin. Pathol., 44, 29–32. Benirschke, K. and Kaufmann, P. (1995) Pathology of the Human Placenta, 3rd edn. Springer, New York. Bergeron, C., Ferenczy, A., Toft, D.O. et al. (1988) Immunocytochemical study of progesterone receptors in the human endometrium during the menstrual cycle. Lab. Invest., 59, 862–869. Boyd, J.D., and Hamilton, W.J. (1970) The Human Placenta. Heffer, Cambridge. Burke, R.D., Wang, D. and Jones, V.M. (1994) Ontogeny of vessel wall components in the outflow tract of the chick. Anat. Embryol., 189, 447–456. Cohen, P.R., Rapini, R.P., and Farhood, A.I. (1993) Expression of the human hematopoietic progenitor cell antigen CD34 in vascular and spindle cell tumors. J. Cutan. Pathol., 20, 15–20. Craven, C.M., Morgan T., and Ward, K. (1998) Decidual spiral artery remodelling begins before cellular interaction with cytotrophoblasts. Placenta, 19, 241–252. Czernobilsky, B., Remadi, S. and Gabbiani, G. (1993) Alpha-smooth muscle actin and other stromal markers in endometrial mucosa. Vir. Arch. A. Pathol. Anat., 422, 313–317. Darby, I., Skalli, O, and Gabbiani, G. (1990) α-Smooth muscle actin is transiently expressed by myofibroblasts during experimental wound healing. Lab. Invest., 63, 21–29. Davis, S., Aldrich, T.H., Jones, P.F. et al. (1996) Isolation of angiopoietin-1, a ligand for the TIE2 receptor, by secretion-trap expression cloning. Cell, 87, 1161–1169. Duband, J.L., Gimona, M., Scatena, M. et al. (1993) Calponin and SM22 as differentiation markers of smooth muscle: spatiotemporal distribution during avian embryonic development. Differentiation, 55, 1–11. Farhood, A.I. and Abrams, J. (1991) Immunohistochemistry of endometrial stromal sarcoma. Hum. Pathol., 22, 224–230. Ferenczy, A. (1976a) Studies on the cytodynamics of human endometrial regeneration. I. Scanning electron microscopy. Am. J. Obstet. Gynecol., 124, 64–74. Ferenczy, A. (1976b) Studies on the cytodynamics of human endometrial regeneration. II. Transmission electron microscopy and histochemistry. Am. J. Obstet. Gynecol., 124, 582–595. Folkman, I. (1997) Addressing tumor blood vessels. Nature Biotechnol., 15, 510. Folkman, I. and D’Amore, P.A. (1996) Blood vessel formation: what is its molecular basis? Cell, 87, 1153–1155. Franquemont, D.W., Frierson, H.F.J. and Mills, S.E. (1991) An immunohistochemical study of normal endometrial stroma and endometrial stromal neoplasms. Evidence for smooth muscle differentiation. Am. J. Surg. Pathol., 15, 861–870. Goodger, A.M. and Rogers, P.A.W. (1994) Endometrial endothelial cell proliferation during the menstrual cycle. Hum. Reprod., 9, 399–405. Goodger, A.M. and Rogers, P.A. (1995) Blood vessel growth in the endometrium. Microcirculation, 2, 329–343. Hanahan, D. (1997) Signaling vascular morphogenesis and maintenance. Science, 277, 48–50. Hirschi, K.K., Rohovsky, S.A. and D’Amore, P.A. (1998) PDGF, TGF-α, and heterotypic cell-cell interactions mediate endothelial cell-induced recruitment of 10T1/2 cells and their differentiation to a smooth muscle fate. J. Cell Biol., 141, 805–814.

Hungerford, J.E., Owens, G.K., Argraves, W.S. and Little, C.D. (1996) Development of the aortic vessel wall as defined by vascular smooth muscle and extracellular matrix markers. Develop. Biol., 178, 375–392. Ingegno, M.D., Money, S.R., Thelmo, W. et al. (1988) Progesterone receptors in the human heart and great vessels. Lab. Invest., 59, 353–356. Jansen, R.P.S. and Johannisson, E. (1985) Endocrine response in the female. In Shearman, R.P. (ed.), Clinical Reproductive Endocrinology. Churchill Livingstone, Edinburgh, pp. 109–164. Kaiserman-Abramof, I.R. and Padykula, H.A. (1989) Angiogenesis in the postovulatory primate endometrium: the coiled arteriolar system. Anat. Rec., 224, 479–489. Kocher, O. and Gabbiani, G. (1986) Expression of actin mRNAs in rat aortic smooth muscle cells during development, experimental intimal thickening, and culture. Differentiation, 32, 245–251. Kocher O., Skalli O., Cerutti D. et al. (1985) Cytoskeletal features of rat aortic cells during development. An electron microscopic, immunohistochemical, and biochemical study. Circul. Res., 56, 829–838. Kocher, O., Gabbiani, F., Gabbiani, G. et al. (1991) Phenotypic features of smooth muscle cells during the evolution of experimental carotid artery intimal thickening. Biochemical and morphologic studies. Lab. Invest., 65, 459–470. Kohnen, G., Campbell, S., Irvine, G.A. et al. (1998) Endothelin receptor expression in human decidua. Mol. Hum. Reprod., 4, 185–193. Kooy, J., Taylor, N.H., Healy D.L. and Rogers, P.A.W. (1996) Endothelial cell proliferation in the endometrium of women with menorrhagia and in women following endometrial ablation. Hum. Reprod., 11, 1067–1072. Landerholm, T.E., Dong, X-R., Lu, J. et al. (1999) A role for serum response factor in coronary smooth muscle differentiation from proepicardial cells. Development, 126, 2053–2062. Lee, W-S., Harder, J.A., Yoshizumi, M. et al. (1997) Progesterone inhibits arterial smooth muscle cell proliferation. Nature Med., 9, 1005–1008. Leivo, I., Laurila, P., Wahlstrom, T. and Engvall, E. (1989) Expression of merosin, a tissue-specific basement membrane protein, in the intermediate trophoblast cells of choriocarcinoma and placenta. Lab. Invest., 60, 783–790. Leggard, J.L. (1988) Two monoclonal antibodies to actin: one muscle selective and one generally reactive. Cell Motil. Cytoskel., 10, 349–362. Li, T.C., Rogers, A.W., Dockery, P. et al. (1988) A new method of histological dating of human endometrium in the luteal phase. Fertil. Steril., 50, 52–60. Lillemoe, T.J., Perrone, T., Norris, H.J. and Dehner, L.P. (1991) Myogenous phenotype of epithelial-like areas in endometrial stromal sarcomas. Arch. Pathol. Lab. Med., 115, 215–219. Majno, G., Gabbiani, G., Hirschel, B.J. et al. (1971) Contraction of granulation tissue in vitro: similarity to smooth muscle cells. Science, 173, 548–550. Markee, J.I. (1940) Menstruation in intraocular endometrial transplants in the rhesus monkey. Contrib. Embryol., 177, 223–305. Mitchell, J.J., Reynolds, S.E., Leslie, K.O. et al. (1990) Smooth muscle cell markers in developing rat lung. Am. J. Resp. Cell Mol. Biol., 3, 515–523. More, I.A.R. (1987) The normal human endometrium. In Fox, H. (ed.), Obstetrical and Gynaecological Pathology, 3rd edn. Churchill Livingstone, Edinburgh, pp. 302–319. Noyes, R.W., Hertig, A.T. and Rock, J. (1950) Dating the endometrial biopsy. Fertil. Steril., 1, 3–25. Owens, G.K., Loeb, A., Gordon, D. and Thompson M.M. (1986) Expression of smooth muscle-specific alpha-isoactin in cultured vascular smooth muscle cells: relationship between growth and cytodifferentiation. J. Cell Biol., 102, 343–352. Owens, G.K. (1995) Regulation of differentiation of vascular smooth muscle cells. Physiol. Rev., 75, 487–517. Palmer, J.A., Lau, T.M., Hickey, M. et al. (1996) Immunohistochemical study of endometrial microvascular basement membrane components in women using Norplant. Hum. Reprod., 11, 2142–2150. Parums, D.V., Cordell, J.L., Micklem, K. et al. (1990) JC70: a new monoclonal antibody that detects vascular endothelium associated antigen on routinely processed tissue sections. J. Clin. Pathol., 43, 752–757. Pijnenborg, R. (1998) The human decidua as a passage-way for trophoblast invasion. Troph. Res., 11, 229–241. Pijnenborg, R., Dixon, G., Robertson, W.B. and Brosens, I. (1980) Trophoblastic invasion of human decidua from 8 to 18 weeks of pregnancy. Placenta, 1, 3–19. Pijnenborg, R., Robertson, W.B., Brosens, I. and Dixon, G., (1981) Review article: trophoblast invasion and the establishment of haemochorial placentation in man and laboratory animals. Placenta, 2, 71–92. Press, M.F., Udove, J.A., and Greene, G.L. (1988) Progesterone receptor

291

G.Kohnen et al. distribution in the human endometrium. Analysis using monoclonal antibodies to the human progesterone receptor. Am. J. Pathol., 131, 112–124. Ramani, P., Bradley, N.J. and Fletcher, C.D. (1990) QBEND/10, a new monoclonal antibody to endothelium: assessment of its diagnostic utility in paraffin sections. Histopathology, 17, 237–242. Ramsey, E.M. (1977) Vascular anatomy. In Wynn, R.M. (ed.), Biology of the Uterus, 2nd edn. Plenum Press, London, pp. 59–76. Ramsey, E.M. and Donner, M.W. (1980) Placental Vasculature and Circulation. Georg Thieme, Stuttgart. Risau, W. (1997) Mechanisms of angiogenesis. Nature, 386, 671–674. Rogers, P. (1998) The endometrial vascular bed. In Cameron, I.T., Fraser, I.S. and Smith, S.K. (eds), Clinical Disorders of the Endometrium and Menstrual Cycle. Oxford University Press, Oxford, pp. 31–45. Rogers, P.A., Au, C.L. and Affandi, B. (1993) Endometrial microvascular density during the normal menstrual cycle and following exposure to longterm levonorgestrel. Hum. Reprod., 8, 1396–1404. Rogers, P.A.W., Lederman, F. and Taylor, N. (1998) Endometrial microvascular growth in normal and dysfunctional states. Hum. Reprod. Update, 4, 503–508. Sappino, A.P., Schu¨ rch, W. and Gabbiani, G. (1990) Biology of disease. Differentiation repertoire of fibroblastic cells: expression of cytoskeletal proteins as marker of phenotypic modulations. Lab. Invest., 63, 144–161. Schu¨ rch, W., Seemayer, T.A. and Gabbiani, G. (1992) Myofibroblasts. In Sternberg, S.S. (ed.), Histology for Pathologists. Raven Press, New York, pp. 109–143. Selmin, O., Volpin, D. and Bressan, G.M. (1991) Changes of cellular expression of mRNA for tropoelastin in the intraembryonic arterial vessels of developing chick revealed by in situ hybridization. Matrix, 11, 347–358. Sheppard, B.L. and Bonnar, J. (1979) The development of vessels of the endometrium during the menstrual cycle. In Diczfalusy, E., Fraser, I.S. and Webb, F.T.G. (eds), Endometrial Bleeding and Steroidal Contraception, Pitman Press, Bath, pp. 65–85. Skalli, O., Ropraz, P., Trzeciak, A. et al. (1986) A monoclonal antibody against alpha-smooth muscle actin: a new probe for smooth muscle differentiation. J. Cell Biol., 103, 2787–2796. Smith, S.K. (1998) Angiogenesis, vascular endothelial growth factor and the endometrium. Hum. Reprod. Update, 4, 509–519. Suri, C., Jones, P.F., Patan, S. et al. (1996) Requisite role of angiopoietin-1, a ligand for the TIE2 receptor, during embryonic angiogenesis. Cell, 87, 1171–1180. Suster, S. and Fisher, C. (1997) Immunoreactivity for the human hematopoietic progenitor cell antigen (CD34) in lipomatous tumors. Am. J. Surg. Pathol., 21, 195–200. Thyberg, J., Hedin, U., Sjo¨ lund, M. et al. (1991) Regulation of differentiated properties and proliferation of arterial smooth muscle cell. Atherosskleros, 10, 966–990. Vikkula, M., Boon, L.M., Carraway, K.L. et al. (1996) Vascular dysmorphogenesis caused by an activating mutation in the receptor tyrosine kinase TIE2. Cell, 87, 1181–1190. Weiss, S.W. and Nickoloff, B.J. (1993) CD-34 is expressed by a distinctive cell population in peripheral nerve, nerve sheath tumors, and related lesions. Am. J. Surg. Pathol., 17, 1039–1045. Wingfield, M., Macpherson, A., Healy, D.L. and Rogers, P.A.W. (1995) Cell proliferation is increased in the endometrium of women with endometriosis. Fertil. Steril., 64, 340–346. Yancopoulos, G.D., Klagsbrun, M. and Folkman, J. (1998) Vasculogenesis, angiogenesis, and growth factors: ephrins enter the fray at the border. Cell, 93, 661–664. Received on July 19, 1999; accepted on October 28, 1999

292