Expression of desmin and dimentin intermediate filaments in human decidual cells during first trimester pregnancy

Placenta

(1995),

16, 261-275

Expression of Desmin and Vimentin Intermediate Filaments in Human Decidual Cells During First Trimester Pregnancy A. CAN’, M. TEKELiOGLU A. BALTACI”

&

Department of Histology and Embryology, Ankara University School of Medicine, Sihhiye, 06339, Ankara, Turkey “SSK

Maternity

Hospital,

Etlik,

Ankara,

Turkey

bTo whom all correspondence should be addressed at: Department of Anatomy and Cell Biology, Tufts University School of Medicine, 136 Harrison Avenue, Boston, MA 02111, USA Paper

accepted

1611.1994

SUMMARY

Human endometrial stromal cells (decidual cells) display dramatic alterations in cell shape and size during decidualization. The present study was designed to demonstrate the expression of two major cytoskeletal elements, desmin and vimentin, in human pregnant endometrial decidual cells. Additionally, stage-dependent variations of those intermediate filaments (IFS) among gestational weeks were also evaluated with regard to the support and maintenance of decidualization. Materials were obtained from legal suction terminations of pregnancies of 3-10 weeks gestation. Tissue specimens were either blocked in parafin or enzymatically dissociated for isolation of decidual cells which subsequently were cultured as monolayers. Immunoperoxidase and immunoJuorescence staining methods were applied by using anti-desmin and antivimentin monoclonal antibodies. Both desmin and vimentin expression were observed during the early weeks of pregnancy (‘3-6 weeks). These two types of IFS were also detected in short-term cultures in a jilamentous fashion either within the cell body or at cellular attachment plaques. When decidual cells were cultured for longer periods (4MO days), the expression of desmin dramatically declined while vimentin expression was maintained in a rather d@use and more abundant fashion. The in situ expression of desmin and vimentin in later weeks ofgestation (7-10 weeks) correlated with immunojluorescencestaining of long-term cultured cells in that desmin staining was very weak and mostly undetectable where vimentin expression persisted and was evenly distributed throughout the entire stroma. The results demonstrate the diflerential expressionof two major IFS, desmin and vimentin, in human endometrial stromal cells during decidualization and subsequent placentation. The persistence of 0143-4004/95/030261

+ 12 $08.00/O

0

1995 W. B. Saunders

Company

Ltd

262

Placenta

(1995),

Vol. 16

vimentin in all stages examined suggests that this IF is probably involved in cell morphology and nucleocytoplasmic integrity. The temporal pattern of desmin expression suggests a role for this IF during the rapid onset of the decidualization process.

INTRODUCTION Human endometrial stromal cells undergo a differentiation process termed decidualization which begins just prior to blastocyst implantation and proceeds throughout pregnancy (for reviews, see Verma, 1983; Cornillie, Lauweryns and Brosen, 1985; Anderson, 1991). Previous studies have suggested that decidualized stromal cells (decidual cells) may play a crucial role in implantation and subsequent growth of the embryo (Mori et al, 1991; Can, Tekelioglu and Biberoglu, 1991). Although our knowledge about the cellular pathways of decidualization and its physiological role remains limited, recent data shows that decidualization in which many synthetic processes take place such as DNA, RNA and protein synthesis (Bell, 1989; Tabibzadeh, 1992). These synthetic processescoincide with several cell-cell interactions in an autocrine, paracrine and juxtacrine manner within a highly vascularized endometrial microenvironment (Leake, Carr and Rinaldi, 1991; Seppala et al, 1992). During decidualization, stromal cells display remarkable alterations in their size, shape and organelle distribution. An increase in total cell volume is observed as well as a change in shape from fusiform to round-ovoid with large euchromatic nuclei. In addition, cells contain prominent free ribosomes, Golgi and dilated cisternae of endoplasmic reticulum (Can, Tekelioglu and Biberoglu, 1991; Tekelioglu and Can, 1992) characteristics of cells which are indicators of a higher synthetic and secretory activity. This characteristic change occurs during the early days of pregnancy and is maintained to some extent throughout the first trimester. Therefore, given the role microtubules, microfilaments and intermediate filaments (IFS) play in determining cellular morphology and function, this suggests a role for IFS in the decidualization process. Desmin and vimentin, as two distinctive types of IFS are recognized as the most stable components of the cytoskeleton of many cells and differ significantly from other cytoskeletal organelles such as microtubules and microfilaments, in several important respects (for review, see Steinert and Roop, 1988; Albers and Fuchs, 1992). Desmin and vimentin are assembled from single subunits with molecular masses of about 57 000 and 55 000 Da, respectively. Both display cell-specific localization and are frequently expressed in a specific pathway of differentiation (Fuchs et al, 1987). Vimentin is the major IF of widely distributed cells derived from the embryonic mesoderm including those of connective tissues whereas desmin expression is restricted to all types of muscle cells. Thus, endometrial stromal cells, which are mesodermal in origin, display vimentin. However, recent studies show positive desmin immunoreactivity in both rat (Glasser and Julian, 1986; Glasser et al, 1987) and human decidualized endometrium (Halperin et al, 1991). In the former study (Glasser and Julian, 1986; Glasser et al, 1987), an overlapping pattern of more than one type of IF was detected in rat endometrial decidualized stromal cells. A similar observation has been made in human fetal tissues (van Muijen, Ruiter and Warnaar, 1987) as well as during the development of arterial wall smooth muscle cells (Gard and Lazarides, 1980). Additionally, stage-dependent expression variations of desmin and/or vimentin were also noted in some epithelial cells such as adenocarcinoma cells (Sato et al, 1986) and keratinocytes (van Muijen and Ponec, 1986). In some instances desmin was successfully used as a highly specific tool to investigate the differentiation of muscle cells (Schultheiss et al, 1991). Therefore, varying expression of the

Can, Tekelioglu,

Baltaci:

Desmin and Vimentin

in Human

Decidual

Cells

263

IF proteins may occur during the decidualization of endometrial stromal cells which undergo a series of structural and functional alterations throughout the early pregnancy. The present study was designed to investigate the expression of desmin and vimentin, in human decidual cells throughout the first trimester of pregnancy and in spontaneously decidualized cultures in vitro.

MATERIALS Sources

AND

METHODS

of materials

All chemicals were purchased from Sigma Chemical Co., St Louis, MO, USA unless otherwise stated. Endometrial tissues (- 10 g wet weight of decidua) were obtained by legal suction terminations (n=46) of 3-10 weeks of gestations under paracervical anaesthesia. Dating of gestational week was monitored by either serum P-human chorionic gonadotrophin (P-hCG) levels (for 3rd week of gestation 25 mlU/ml or higher was accepted as positive) or ultrasonographic (USG) examination. Thus, beginning from the 4th week of gestation, transvaginal USG was used for evaluation of the diameter of the amniotic sac, followed by crown-rump length assessment of the embryo by transabdominal USG after the 5th week. Study protocols were examined and approved by the IRB (Institutional Review Board) of the authors’ university and a written consent was obtained from each pregnant woman after an oral explanation of the procedure. Samples were placed in transport medium [Hank’s buffered salt solution containing penicillinstreptomycin (20 ul/ml)] and transferred to the laboratory in an ice-cold container. A small block of decidual tissue was dissected out from whole material and fixed in phosphate buffered 2.5 per cent (v/v) glutaraldehyde and 1.5 per cent (w/v) paraformaldehyde for 4 h followed by washing in 0.1 M phosphate buffer, dehydrated in ascending grades of ethanol and embedded in Histoplast (Shandon, UK). The remaining clumps of decidual tissue (approximately 250 mg wet weight) were checked under a dissecting microscope to remove any blood and mucus debris, washed in transport medium, and then taken through an enzymatic digestion procedure for the isolation of decidual cells (see below).

Cultures

Different protocols were employed to minimize the potential experimental artifacts which might affect the decidual cell cytoskeleton in vitro. Thus different types of substrate, periods of culture and composition of culture medium were examined. Initial cell separation was achieved by the enzymatic digestion method summarized below. Successive short-term and long-term cultures, either on glass coverslips or polystyrene flasks, were set up and fed by complete medium or serum-free medium. The latter was used to test whether there is any effect of fetal calf serum (FCS) on decidual cell differentiation that might cause an artifactual cytoskeletal phenotype. Decidual cells were separated after collagenase digestion as previously described by Osteen et al (1989) and modified as follows; the tissues were placed in phenol red-free Dulbecco’s modified Eagle’s medium (DMEM) containing 2 mg/ml collagenase type IA and incubated in a 37°C shaking water bath for 3 h in a capped, conical glass tube. Every 60 min the tube was centrifuged at low speed (300 g,,,) for 10 min, and the supernatant was transferred into another tube. After the final centrifugation, supernatant was filtered sequentially through two

264

Placenta (1995), Vol. 16

metal mesh membrane filters of 104 pm and 73.7 pm, respectively. The filtered fraction containing decidual cells was separated into two equal parts, the first part was diluted with complete medium [DMEM supplemented with 10 per cent fetal calf serum (Seromed, Germany), 2.4 @/ml nystatin suspension, 10 @/ml penicillin streptomycin suspension, 0.2 U/ml insulin (NPH), 146 mg/l L-glutamine] and the second part was diluted with serum-free medium (DMEM and supplemented with the same reagents as in the complete medium except FCS and insulin), counted with a Neubauer haemacytometer and then plated out either onto poly-L lysine (1 mg/ml) coated glass coverslips placed in 6-well polystyrene culture plates (Greiner, Austria) or 25 cm2 culture flasks (Greiner, Austria) at a density of approximately lo3 cells/cm2 of substrate. Short-term cultures were incubated for 7 days at 37°C in 95 per cent air and 5 per cent CO, in complete medium or in serum-free medium while the long-term cultures were incubated for much longer periods (40-60 days) in the same culture media. In both culture systems, the media was changed twice a week. The morphology of cells in both culture systems was examined by an inverted microscope (Nikon TMS, Japan) during every medium change and cytoskeletal assembly was evaluated as evidenced by the presence of numerous attachment plaques (lamellopodia, filopodia) and stress fibres. Short-term cultures at days 4-7 and long-term cultures at days 40-60 were terminated, fixed and stained for fluorescence microscopy using anti-desmin and anti-vimentin antibodies.

Immunocytochemistry Tissue sections. Endometrial tissue sections (5 pm thickness) were deparaffinized and rehy-

drated in descending grades of ethanol and then washed with phosphate buffered saline (PBS). The standard peroxidase anti-peroxidase method was employed with minor modifications. All samples were treated prior to the primary antibody with 3 per cent H,O, for 5 min to block the endogenous peroxidase activity followed by a brief rinse in PBS and protease treatment for 30 min at 37°C (see details in Rangdaeng and Troung, 1991). Background staining was minimized by pre-incubation with normal goat serum (1:20 diluted with PBS) for 20 min at room temperature. Sections were incubated with primary antibodies for 90 min at 37°C. The desmin monoclonal antibody (clone no:DE-U-10) was produced using desmin from pig stomach as an immunogen (Debus, Weber and Osborn, 1983) and the vimentin monoclonal antibody (clone no:VIM-13.2) was produced using human foreskin fibroblasts as an immunogen (Adams and Watt, 1988). This was followed by the treatment of goat anti-mouse peroxidase conjugated secondary antibody (1:50 dilution in PBS) for 90 min at 37°C and then peroxidase anti-peroxidase complex (1:SOO dilution) for 30 min at room temperature, Diaminobenzidine tetrahydrochloride (DAB) was used as a chromogen (5 min). Sections were counterstained with Mayer’s haematoxylin (5 min) and then dehydrated in ethanols, cleared in xylene and mounted in Entallan (Merck, Germany). Cultures. Short-term and long-term monolayer cultures either on glass coverslips or in flasks were rinsed in PBS-Ca2+-Mg2+, fixed for 10 min in - 20°C methanol, dried in air and then rehydrated in PBS for 5 min. Cells were incubated with blocking buffer [S per cent (v/v) normal goat serum, 0.5 per cent (v/v) Triton X-100 in PBS] for 30 min at room temperature. Following this treatment, cells were incubated with mouse primary antibodies for 90 min at 37°C in a humidified chamber. Several attempts have been made to find out the optimum dilution titres of primary antibodies to obtain the maximal signal to noise ratio. Optimal dilutions were found within the range of the manufacturer’s recommendations. Thus,

Can, Tekelioglu,

Baltaci:

Desmin and Vmentin

in Humun

Decidual

Cells

265

anti-desmin and anti-vimentin antibodies were diluted 1: 10 and 1: 100, respectively, in PBS. As a secondary layer, the cells were treated with fluorescein isothiocyanate (FITC) conjugated goat-anti mouse antibody. After rinsing three times for 5 min in PBS, coverslips were mounted in 90 per cent glycerol, 10 per cent PBS, pH 9.5. Control staining of sections and cultures was carried out by the substitution of the primary antibodies with PBS using the same staining protocols. No staining was observed under these conditions. Specimens were photographed using a Zeiss Axioskop (Germany) microscope equipped with fluorescence 490-520 nm filter set and Neofluar Oil 100 X objective on Kodak Gold II (400 ASA) and Kodak Ektar (1000 ASA) films. All negatives were enlarged so that the final magnification is equal for all cultured cells.

RESULTS Expression

of desmin

and vimentin

in tissue

sections

The staining of desmin varied in decidual cells during first trimester decidual tissue. Abundant staining was observed in decidual cells from tissue samples of 3-6 weeks of pregnancy [Figure l(A-C)]. In contrast, very little staining was observed in tissue samples of 9-10 weeks of pregnancy (Figure 2). In these instances, faint staining was restricted to a few decidual cells where occasional extracelluar matrix staining was also detected. Three to six gestational week decidua labelled with anti-desmin antibody showed a diffuse and abundant pattern of staining, often with focal condensations of intracytoplasmic immunopositivity [Figure l(A)]. This type of staining was observed in large, round to oval stromal cells with euchromatin-rich nuclei, a characteristic of decidualized stromal cells [Figure l(A) and (B)]. In the early weeks of gestation, desmin reactivity was evenly distributed throughout the decidua whereas in vascular compartments, glandular cells, glands and luminal epithelial cells staining was absent [Figure l(B) and (C)l. Small punctate foci stained positively with anti-desmin antibody in myometrial and periarteriolar smooth muscle cells which correspond to the cytoplasmic dense bodies in muscle cells [Figure l(D)]. Even though some fluctuations in number of decidualized cells were noted between different individuals in a given section of their endometria, the immunostaining of desmin was compatible with that of decidual cell frequency differences; thus positive staining was significantly higher in sections where decidual cells predominated. The pattern of staining in early gestational decidua (3-6 weeks) with anti-vimentin antibody was similar to that of desmin in 3-6 weeks decidua (Figure 3). Additionally, weak staining was observed in glandular cells, while in smooth muscle cells no staining was apparent [compare perivascular smooth muscle cells in Figure l(B) with Figure 31. In contrast to desmin, vimentin expression is maintained within the decidua during entire first trimester in a similar manner to that in early weeks.

Effects

of variables

in culturing

methods

and morphology

of cultured

cells

The growth, division and differentiation patterns of cells in different culture systems were evaluated by phase contrast microscopy to assessany effects of culture protocols (see Materials and Methods, Cultures) on morphology of cells grown in vitro. No morphological changes were observed due to the type of substrate; cells appeared morphologically similar whether grown on glass coverslips or in polystyrene flasks. However, a slight confluency retardation was noted in cultures grown in flasks which was not apparent after the 3rd culture day. In

Placenta

(1995),

Vol. 16

Figure 1. Four week gestational endometrium section incubated with anti-desmin antibody. (A) Mature decidual cells show positive intracytoplasmic staining as seen by the deposits of brown pigmentation. Note that blood-derived cells (i.e. leucocytes and red blood cells) are negatively stained (arrows). (B) Desmin staining was evenly distributed throughout the stroma, even in elongated perivascular stromal cell (arrows) whereas weaker staining was noted in the arteriolar smooth muscle cells. (C) Absence of desmin staining in glandular and luminal epithelial cells is observed in 4 week gestational endometrium. (D) Small punctate foci are stained with desmin antibody (arrows) in irregular myometrial smooth muscle outgrowths in some endometrial sections corresponding to desmin filaments within cytoplasmic dense bodies of smooth muscle cells. Scale bar in (A) (B) and (D) 20 pm; in (C) 200 pm,

addition, no significant differences were noted between cells fed by complete medium and serum-free medium. Both displayed similar morphology, growth and differentiation features [Figure 4(B) and (C)] such as confluency time, doubling time and density. In contrast, several morphological differences were noted between short- and long-term cultured cells. The morphology of short-term cultured stromal cells was similar to that

Can, Tekelioglu,

Baltaci:

Desmin and Vimentin

in Human

Fipre 2. A section of 9 week gestational endometrial in decidual cells. Occasional faint staining is found

D&dual

Cells

stroma demonstrating the remarkable in a few decidual cells (arrowheads).

267

decline of desmin Scale bar 50 pm.

staining

previously described (Osteen et al, 1989). Specifically, short-term cultured decidual cells maintained a slight globular cell shape, were occasionally binucleated, and bore several long and flattened attachment plaques [Figure 4(A)]. During the first week of culture, no organized cytoskeletal network was observed as evidenced by the lack of stress fibres by phase contrast microscopy. A gradual decrease in the nucleocytoplasmic ratio was noted that corresponded with a decreasing mitotic rate. As the number of days in culture progressed, several morphological changes were seen. Thus, during the later weeks of culture, cells became flattened and increased in size. A decrease in number of long slender cellular processes was also noted which coincided with a substantial increase in intracytoplasmic stress fibres [Figure 4(B) and (C)l.

Expression

of desmin

and vimentin

in cultured

cells

Specimens obtained from different culture protocols were stained with either anti-desmin or anti-vimentin antibodies. The results showed no differences in staining patterns between cultures grown on glass coverslips and those grown on polystyrene flasks. Additionally, no differences were noted between desmin and vimentin staining patterns of cultures grown in varying culture medium compositions. The staining patterns of short-term (up to 7 days) and long-term (6-9 weeks) cultures with desmin and vimentin antibodies showed significant differences. Short-term cultures labelled with anti-desmin antibody displayed a homogeneous staining pattern of thin and filamentous fibres found throughout the cytoplasm that extended into the cellular protrusions such as lamellopodia and filopodia [Figure S(A)]. Th ese filaments were organized in parallel array extending from one end of the cell to another and overlapping the nucleus. There was a slight staining intensity difference among individual cells within the same culture dish which can be

268

Placenta

(1995),

Vol. 16

Figure 3. Vimentin staining in 4 week gestational endometrial stroma. Note large nucleated mature decidual cells stain intensely with anti-vimentin antibody (arrows). Staining pattern is similar to desmin in that immunopositive cells are evenly distributed among blood vessels (asterisks). This staining pattern is restricted to decidual cells but also is found rarely in glandular epithelial cells (not shown). Scale bar 20 pm.

interpreted as a result of cell shape differences. Younger cells having relatively globular shape stained more intensely as compared to their flattened neighbours [compare Figure 5(A) with (B)]. Multiple slender and flattened protrusions also displayed positive staining with anti-desmin antibody. No significant differences in desmin staining were observed over the course of 2-7 days of culture. Staining of short-term cultures with anti-vimentin antibody exhibited a similar pattern to that of desmin-stained cells, however, vimentin-positive filaments showed a somewhat reticular pattern of distribution. Bundles of vimentin filaments were observed overlapping one another in cytoplasm and coalescing as a perinuclear filamentous mantle [Figure 6(A)]. This reticular network was replaced by a parallel array of filaments along cellular protrusions [Figure 6(A) and (B) arrows]. Similar to desmin, no variations in staining patterns were observed due to length of culture period between 2-7 days. In contrast, when decidual cells were cultured for longer periods (40-60 days) significant differences in the staining of cultures labelled with desmin and vimentin antibodies were observed. A dramatic decline in the level of desmin immunoreactivity was evidenced by the almost complete absence of immunofluorescence staining with anti-desmin antibody. Occasionally a pale and weak punctate staining of desmin was found in perinuclear cytoplasm of a few cells [Figure 7(A)]. Control stainings showed no positivity. Unlike desmin, novel distribution and intensity pattern of vimentin staining was noted. In contrast to the remarkable decline of desmin expression, vimentin displayed a more diffuse, broad filamentous staining as observed in flattened, extremely large long-term cultured decidual cells [Figure 7(B)]. Their end-to-end and side-to-side protrusions extending between cells were also brightly stained with anti-vimentin antibody. Although the long-term cells had extremely

Can, Tekelioglu,

Baltaci:

Figwe

(7 days) and long-term

4. Short

Desmin and Vimentin

in Human

(7 weeks) cultures

Decidual

Cells

of decidual

269

cells grown

on glass cover: slips. Short-term

CUltWe :d cells (A) display a globular cell shape with several slender long fdopodia and lamellopc bdia. In contrast, medium display long-te :rm cultured cells in (B) complete medium (see Materials and Methods) or (C) serum-free extrem ely flat cell bodies bearing well organized cytoskeletal components. Note stress fibres along the long axis of intercellular processes found between neighbouring cells. Scale bar in (A) 210 pm; in (B) and cells a*Id in extensive (C)

50

Pm.

Placenta

270

(1995),

Vol. 16

Figure 5. Desmin staining in short-term (7 days) cultures of decidual cells. (A) Fine filamentous fibres are observed in the cell body or at cellular attachment plaques in a relatively flattened decidual cell. (B) In a more ellipsoid cell [younger than the one in (A)] staining intensity was increased, possibly due to arising the superimposition of desmin-positive filaments within the cytoplasm. Scale bar 20 pm.

flattened cytoplasms which were decorated by fine, straight cytoskeletal elements recognized by phase contrast microscopy, immunofluorescence signal was quite high and homogeneously distributed throughout the cytoplasm [Figure 7(B)] suggesting a significant increase in expression of vimentin in long-term cultured decidual cells.

DISCUSSION In recent years, many secretory and structural proteins derived from human decidualized stromal tissue have been reported (for reviews, see Fay and Grudzinskas, 1991; Seppala et al, 1992). However, conflicting evidence exists as to whether these peptides and proteins originate from decidual cells, epithelial cells or the intermingling lymphoid cells (i.e. macrophages, granulocytes and lymphocytes). The search for a cellular marker of decidualization has led to the identification and expression of structural proteins in decidual stromal cells. Importantly, certain types of IFS have been reported in rat (Glasser and Julian, 1986) and human (Kisalus, Herr and Little, 1987; Halperin et al, 1991) decidualized stromal cells. However, no extensive

Can, Tekeliogb,

B&xi:

Desmin and Vimentin

zn Human

Decidual

Cells

271

Figure 6. Vimentin staining in short-term cultures of decidual cells. (A) Similar to desmin staining (see Figure 5), line filamentous fibres arc observed, however these vimentin-positive filaments (see Figure 5) show a reticular pattern of distribution. Bundles of vimentin filaments were observed overlapping one another in the cytoplasm presenting a perinuclear filamentous mantle. (B) Parallel arrays of vimentin-positive filaments (arrows) are observed within filopodia which display close proximity to neighbouring cells. Scale bar 20 pm.

and correlative report has been performed concerning the IF proteins of human pregnant decidual cells in both in vivo and in vitro conditions. The present study sought to identify the expression of two major IF proteins, desmin and vimentin, in human endometrial stromal cells when they undergo decidualization both in vivo and in vitro. It has been shown that stromal cells in vitro exhibit a higher degree of desmin and/or vimentin synthesis during decidualization in rats (Glasser and Julian, 1986), mouse (Zorn, De Oliveira and Abrahamsohn, 1990) and human (Halperin et al, 1991). Glasser and Julian (1986) reported that, during decidualization, desmin and vimentin expression undergoes a 3.76 times increase in its rate of synthesis. The ratio of desmin to vimentin synthesis peaks at 96 h of culture. Vimentin synthesis increases at a rate proportional to total cell protein whereas desmin synthesis increases at a rate more rapid than vimentin in short-term cultured decidual cells. In contrast, while vimentin was present in non-decidualized stromal cells desmin was not detectable. This finding is similar to the present study in that desmin expression could be found only during the early days of culture in vitro and in tissue sections of 3-6 weeks of gestation. When cells were cultured for longer periods, desmin expression was markedly diminished whereas vimentin expression was maintained or increased. These findings are further supported by the absence of desmin staining and presence of vimentin staining in situ. Taken together, these results suggest a dynamic regulation of both the synthesis and distribution of IF proteins during the decidualization process. It is possible that the temporal expression of desmin may reflect the structural plasticity of stromal cells during decidualization. Thus, synthesis of desmin is switched on only during decidualization then gradually decreases and is switched off in later stages. Whereas, the persistence of vimentin in either non-decidual or decidual stages may reflect a non-specific cell integrity such as the centration

272

Placenla

(1995),

Vol. 16

Figure 7. Long-term cultures (8 weeks) of decidual cells stained with anti-desmin (A) and anti-vimentin (B) antibodies, (A) Dramatic decline of anti-desmin staining is observed in long-term cultures of decidual cells [compare with Figure 5(A)] in which perinuclear cytoplasms were faintly stained in a punctate manner (arrows show the margins of stained perinuclear cytoplasm). (B) In contrast, the intensity of anti-vimentin staining was increased [compare with Figure 6(B)] with a novel distribution pattern in long-term cultures of decidual cells. Scale bar 50 utn.

of nuclei and abutment of nuclear membrane (Osborn, 1993) rather than playing role in decidualization. Its preferential perinuclear localization supports this hypothesis since close proximity of 9 nm filaments to nuclear lamina were clearly demonstrated in mature mouse decidual cells (Zorn, De Oliveira and Abrahamsohn, 1990). A similar on/off phenomenon was found by other authors (van Muijen and Ponec, 1986) in cultured keratinocytes in that they found a marked expression of vimentin apart from keratin in most of the proliferating cells where the former was lacking in differentiating keratinocytes. Thus, the authors assumed that multiple expression of IFS may be a common phenomenon in normal proliferating tissues such as fetal tissue as well as in tumours. The overlapping expression of IF proteins during decidualization and proliferation of endometrium found in this study seems to partially support that hypothesis. However, the functional significance of the decline of desmin expression after 67 weeks of gestation and its regulation remains unclear. The differential expression of desmin during the onset of decidualization could be due to rapid transformation of cell shape and organelle distribution as a consequence of increased synthetic activity that has been proposed by Zorn, De Oliveira and Abrahamsohn (1990). In the present study we observed a striking difference in vimentin expression during the early and later stages of first trimester pregnancy in cultured cells. Early stage cells display perinuclear localization accompanied by a reticular network throughout the cytoplasm whereas neither a perinuclear localization nor a reticular network could be detected in the highly intense cytoplasmic staining of later stage cells. Although vimentin filaments are known to act primarily to position the nucleus within the protoplasm, their peculiar features such as being regulated by growth (Ferrari et al, 1986) or transient changes during mitosis (Chou et al, 1990) leads to the suggestion of vimentin playing a role during early pregnancy. Thus, their

Can, Tekeliogh,

Baltaci:

Desmin and Vimentin

in Human

Decidual

Cells

273

desmin-like expression in early stages may be due to rapid growth and differentiation process whereas the increase in later stages could be an indicator of cellular regression, a condition where most IF types are known to accumulate (Ghadially, 1982). If this is true, coincident down-regulation of desmin with an increase in vimentin expression could be interpreted as a result of cell regression or apoptosis in later stages of pregnancy. However, it is important to note that even in severe conditions, vimentin and/or desmin expression can be altered without any conspicuous effects on cell appearance or function (Franke, 1993). The experiments in this study were designed to minimize the possibility of cellular artifacts arising from the type of culturing methods used. For this purpose, cells were grown on either glass coverslips or in polystyrene flasks. In addition, two different culture media were used where the first consisted of a defined serum-free medium while the second contained base medium supplemented with FCS and insulin. The latter reagents were reported to have a potential growth effect on cultured decidual cells (Irwin, Utian and Eckert, 1991). Unlike the results of Osteen et al (1989), no delay in the attachment of cells to substrate or any IF expression variation were noted between FCS-supplemented and serum-free medium conditions. However, since serum contains many substances which never come into contact with most body cells (Clark, 1983) and may therefore modify the phenotypic behaviour of the cells in vitro (Medrano et al, 1990), the use of serum-free medium in this study was preferred. Phenol red was omitted from the culture media, in view of the known weak oestrogenic activity of this compound (Berthois, Katzenellenbogen and Katzenellenbogen, 1986). Although the culture conditions used in this study were not generally designed to test the differentiated cellular functions of endometrial cells, some reports (Glasser and Julian, 1986; Osteen et al, 1989) indicate that the expression and maintenance of the differentiative phenotype is mostly independent of the conditions of culture used. For instance, Glasser and Julian (1986) showed there is no need for hormone and/or additional growth factor supplementation during decidualization in vitro. Osteen et al (1989) used similar culture methods and finally concluded that the in vitro culture conditions allow endometrial stromal cells to exhibit several morphological, immunological and functional similarities to their in vivo correlates. Our in vivo findings strongly support this idea in that a similar trend in the expression of desmin and vimentin was observed in both in vitro and in vivo conditions. In conclusion, we believe that isolation and characterization of decidual cells on the basis of expression of their cytoskeletal components may provide a suitable model for studies of the cell-cell interactions that influence the growth, differentiation and to a lesser extent gene expression of decidual cells at various stages of pregnancy. Furthermore, the documentation of stage-dependent changes in the expression and distribution of cytoskeletal components by high resolution immunofluorescence staining in this study provide a useful model upon which to study the regulation of motility in this cell type. Ultimately, the temporal and spatial remodelling and the cross-linkage of the remaining cytoskeletal components (microfilaments, microtubules and associated proteins) must be elucidated in order to provide a reliable and reproducible database which could then serve as a model of normal growth and differentiation in human decidua.

ACKNOWLEDGEMENTS The authors would like to thank Susan Messinger, PhD for her constructive criticisms. This work was financially supported by TUBITAK grants TAG-1017 to A. Can and M. Tekelioglu. We thank Erhan Color Photo Laboratories for their excellent work in printing coloured micrographs.

Placen1a

274

(199.i),

Vol. I6

REFERENCES Adams, J. C. & Watt, F. M. (1988) An unusual strain of human keratinocytes which do not stratify or undergo terminal differentiation in culture. 3oburnal of Cell Biology, 107, 1027-1038. Albers, K. & Fuchs, E. (1992) The molecular biology of intermediate filament proteins. International Review of Cytology, 134, 243-279. Anderson, M. C. (1991) The normal uterus. In Female Reproductive System. Systemzc Pathology, vol. 6,3rd edn (Ed.) Anderson, M. C. pp. 132-133. Edinburgh: Churchill Livingstone. Bell, S. C. (1989) Decidualization and insulin-like growth factor (IGF) binding protein: implications for its role in stromal cell differentiation and the decidual cell in haemochorial placentation. Human Reproduction, 4, 125-130. Berthois, Y., Katzenellenbogen, J. A. & Katzenellenbogen, B. S. (1986) Phenol red in tissue culture media is a weak estrogen: implications concerning in the study of estrogen-responsive cells in culture. Proceedings of National Academy of Sciences, USA, 83, 2496-2501 Can, A., Tekelioglu, M. & Biberoglu, K. (1991) Structure of premenstrual endometrium in HMG+HCG induced anovulatory women. European 3ournal of Obstetrics Gynecology znd Reproductive Biology, 42, 119-129. Chou, Y. H., Bischoff, J. R., Beach, D. & Goldman, R. D. (1990) Intermediate filament organization during mitosis is mediated by p34cdc2 phosphorylation of vimentin. Ce& 62, 1063-1071. Clark, J. (1983) Reducing the requirements for serum supplements in high yield microcarrier cell culture. In Hormonally Defined Media (Ed.) Fisher, G. & Weiser, R. J. pp. 6-15. Berlin: Springer-Verlag. Cornillie, F. J., Lauweryns, J. M. & Brosen, I. A. (1985) Normal human endometrium. An ultrastuctural survey. Gynecologic and Obstetric Investigation, 20, 113-129. Debus, E., Weber, K. & Osborn, M. (1983) Monoclonal antibodies to desmin, the muscle specific intermediate filament protein. EIZlBO 3ouma1, 2, 2305-2312. Fay, T. N. & Grudzinskas, J. G. (1991) Human endometrial peptides: a review of their potential role in implantation and placentation. Human Reproduction, 6, 1311-1326. Ferrari, S., Battini, R., Kaczmarek, L., Rittling, S., Calabretta, B., de Riel, J. K., Philiponis, V., Wei, J. F. & Baserga, R. (1986) Coding sequence and growth regulation of the human vimentin gene. Molecular and Cellular Biology, 6, 3614-3620. Franke, W. W. (1993) The intermediate filaments. In Guidebook to the Cytoskeletal and Motor Proteins (Ed.) Kreis, T. & Vale, R. pp. 137-143. Oxford: Oxford University Press. Fuchs, E., Tyner, A. L., Giudice, G. J., Marchuk, D., RayChaudhury, A. & Rosenberg, M. (1987) The human keratin genes and their differential expression. Current Topics in Developmental Biology, 22, 5-34. Gard, D. L. & Lazarides, E. (1980) The synthesis and distribution of desmin and vimentin during neogenesis in vitro. Cell, 19, 263-275. Ghadially, F. N. (1982) Intracytoplasmic filaments. In Ultrastructural Pathology of the Cell and iMatrix. (Ed.) Ghadially, F. N. pp. 629-686. London: Butterworths. Glasser, S. R. &Julian, J. (1986) Intermediate filament protein as a marker of uterine stromal cell decidualization. Biology of Reproduction, 35, 463474. Glasser, S. R., Lampelo, S., Munir, M. I. & Julian, J. (1987) Expression of desmin, laminin and fibronectin during in situ differentiation (decidualization) of rat stromal cells. Differentiation, 35, 132-142. Halperin, R., Fleminger, G., Kracier, P. F. & Hadas, E. (1991) Desmin as an immunochemical marker of human decidual cells and its expression in menstrual fluid. Human Reproduction, 6, 186189. Irwin, J. C., Utian. W. H. & Eckert, R. L. (1991) S ex steroids and growth factors differentially regulate the growth and differentiation of cultured human endometrial stromal cells. Endocrinology, 129, 2385-2392. Khong, T. Y., Lane, E. B. & Robertson, W. B. (1986) An immunocytochemical study of fetal cells at the maternal-placental interface using monoclonal antibodies to keratins, vimentin and desmin. C&i and Tissue Research, 246, 189-195. Kisalus, L. L., Herr, J. C. & Little, C. D. (1987) Immunolocalization of extracellular matrix proteins and collagen synthesis in first-trimester human decidua. Anatomical Record, 218, 402-415. Leake, R., Carr, L. & Rinaldi, F. (1991) Autocrine and paracrine effects in the endometrium. Annuls of the New York Academy of Sciences, 622, 145-148. Medrano, E. E., Resnicoff, M., Cafferata, E. G. A., Larcher, F., Podhajcer, O., Bover, L. & Molinari, B. (1990) Increased secretory activity and estradiol receptor expression are among other relevant of MCF-7 human breast tumor cell growth which are expressed only in the absence of serum. Experimental Cell Resewch, 188, 2-9. Mori, T., Takakura, K., Narimoto, K., Kariya, M., Imai, K., Fujiwara, H., Okamoto, N., Kariya, Y., Shiotani, M., Umaoka, Y., Kanzaki, H., Noda, Y. & Uchida, A. (1991) Endocrine and immune implications of human endometrial decidualization in implantation. Annals of the New York Academy of Sciences, 326, 321-330. Osborn, M. (1993) Vimentin. In Guidebook to the Cytoskeleial and Motor Proteins (Ed.) Kreis, T. & Vale, R. pp. 169-171. Oxford: Oxford University Press. Osteen, K. G., Hill, G. A., Hargrove, J. T. & Gorstein, F. (1989) Development of a method to isolate and culture highly purified populations of stromal and epithelial cells from human endomerrial biopsy specimens. Fertility and Sterility,

52, 965-972.

Randgaeng, S & Troung, actin. Amer2can 3oumal

L. D. (1991) Comparative of Clinical

Pathology,

immunohistochemical

96, 3245.

staining

for desmin

and muscle-specific

Can,

Tekelioglu,

Baltaci:

Desmin

and Vimentin

in Human

D&dual

Cells

Sam, M., Hayashi, Y., Yanagawa, T., Yoshida, H., Yura, Y., Azuma, filaments and specific markers in a human salivary gland adenocarcinoma Research,

275 M. & Ueno, A. (1985) Intermediate-sized cell line and its nude mice tumors. Cancer

45, 3878-3884.

Schultheiss, T., Zhanghiang, L., Ishikawa, H., Zamir, I., Stoeckert, C. J. & Holtzer, H. (1991) Desmin/ vimentin IFS are dispensable for many aspects of myogenesis. Journal of Cell Biology, 114, 953-966. Seppala, M., Angervo, M., Riittinen, L., Yajima, M., Julkunen, M. & Koistinen, R. (1992) Peptides and proteins in the human endometrium. Reproductive Medicine Reviews, 1, 37-55. Steinert, P. M. & Roop D. R. (1988) Molecular and cellular biology of intermediate filaments, Annual Rezrzew of Biochemistry,

57, 593-625.

Tabibzadeh, S. (1992) Patterns of expression of integrin molecules in human endometrium throughout the menstrual cycle. Human Reproduction, 6, 876-882. Tekelioglu, M. & Can, A. (1992) Stromal environment of normal secretory endometrium during the time of implantation up to the end of the menstrual cycle. Human Reproduction/Abstracts of the 8th Annual Meeting of the ESHRE, The Hague, no. 315. van Muijen, G. N. P., Ruiter, D. J. & Warnaar, S. 0. (1987) Coexpression of intermediate filament polypeptides in human fetal and adult tissues. Laboratory Investigation, 57, 359-369. van Muijen, G. N. P. & Ponec, M. (1986) Differentiation induced expression of cytokeratins in cultured human keratinocytes. Journal of Investrgatioe Dermatology, 87, 173-179. Verma, V. (1983) Ultrastructural changes in human endometrium at different phases of the menstrual cycle and their functional significance. Gynecologic and Obstetric Investigation, 15, 193-212. Zorn, T. M. T., De Oliveira, S. & Abrahamsohn, P. A. (1990) Organization of intermediate filaments and their association with collagen-containing phagosomes in mouse decidual cells. Journal of Structural Biology, 103, 23-33.