Endometrial integrins and the establishment of uterine receptivity

Endometrial integrins and the establishment of uterine receptivity B.A.Lessey Department of Obstetrics and Gynecology, University of North Carolina, Chapel Hill, NC 27599-7570, USA

Our understanding of the factors responsible for the initial interaction between maternal and embryonic epithelium leading to successful implantation remains incomplete. Temporal and spatial expression of specific endometrial peptides may contribute to the establishment of a period of uterine receptivity, whereby the endometrium becomes hospitable to the implanting blastocyst. The failure to establish receptivity may account for a significant number of cases of infertility in the female, especially affecting women with luteal phase deficiency, endometriosis, hydrosalpinges, recurrent pregnancy loss and unexplained infertility. Integrins are a family of cell adhesion molecules that have now been largely accepted as markers of uterine receptivity. Their pattern of expression during the menstrual cycle suggests that integrins may also provide a means to study the factors that regulate the establishment of uterine receptivity in general. The vitronectin receptor is an Ovp3 integrin that appears during the opening of the putative window of implantation. Its expression appears to be associated with the downregulation of epithelial oestrogen and progesterone receptors (PR). We postulate that the selective loss of PR on endometrial epithelium is critical to the establishment of uterine receptivity and allows the emergence of paracrine influence by underlying stroma through specific growth factors. Disruption of the normal pattern of integrins found in certain infertility states may also reflect a shift in the paracrine milieu, possibly due to the influence of inflammatory cytokines associated with endometriosis or tubal disease. Understanding what regulates and dysregulates endometrial integrins may lead to better treatment and diagnostic strategies for infertility patients as well as novel approaches to contraception. Key words: endometrial receptivity/implantation/infertility/integrins/markers

Introduction The extracellular matrix (ECM) and its receptors mediate many fundamental physiological duties of the cell, including embryogenesis and development, cell attachment and migration, apoptosis and cell survival. From this perspective Human Reproduction Volume 13 Supplement 3 1998 © European Society for Human Reproduction and Embryology Downloaded from https://academic.oup.com/humrep/article-abstract/13/suppl_3/247/799170 by guest on 23 November 2017

247

B.A.Lessey

FB

ax

Oil col/LM a

ICAM 1,2 a L

2 col/LM

« 3 col/LM/FN on P

CX4FN

Fx, FB, ICAM i a m

a5FN

C3bi

a6

FN.VCAMaddressin

(34 LM

CX7LM

a 8 Tenascin a 9 Tenascin

VN,FN,FB,vWF, TSP,OP,BSP-1

VN

av

(38

Figure 1. Schematic representation of the pairing patterns of integrin subunits. Note that Pi pairs with many different a subunits, while ocv forms the central pairing partner for may different p subunits. Col = collagen; LM = laminin, FN = fibronectin, VN = vitronectin, FB = fibrinogen, vWF = von Willibrand factor, OP = osteopontin, BSP = bone sialoprotein 1, ICAM 1 and,2 = intercellular adhesion molecules, FX = factor X, C3bi = complement component C3bi.

there is perhaps no cellular process more interesting and complex than that of implantation of the embryo and subsequent placental-endometrial interactions. Implantation has been viewed as a 'receptor-mediated' phenomenon (Yoshinaga, 1989). It may not be surprising that the cell-cell interactions occurring during implantation depend on the expression of specific cell adhesion receptors and their ECM ligands (Aplin, 1996). The growing complexity of the cycle-specific changes in expression of these molecules is astonishing and continues to lead us on a path of discovery. The integrin family of cell adhesion molecules are a major class of receptors for the ECM and participate in cell-cell and cell-substratum interaction (Albelda and Buck, 1990). The specific participation of integrins in implantation is only beginning to be understood, but investigation has identified roles in signal transduction (Werb et al, 1989), maintenance of epithelial polarity (Thie et al, 1995; Aplin et al, 1996), and a developmental progression of placental cytotrophoblast to an invasive phenotype (Damsky et al, 1993; 1994). These receptors are present on the plasma membrane as heterodimeric a and |3 glycoprotein subunits. This expanding family of proteins is shown schematically in Figure 1 along with the known ligands for each. The distribution of integrins on different tissues is predictably segregated between epithelial and stromal cells, mirroring the localization of their ECM ligands. Integrins in the endometrium The role of integrins in reproduction has recently been reviewed (Bronson and Fusi, 1996; Sueoka et al, 1997). Both ourselves (Lessey et al 1992, 1994a; 248 Downloaded from https://academic.oup.com/humrep/article-abstract/13/suppl_3/247/799170 by guest on 23 November 2017

Integrins and implantation

1994b; 1995; Meyer et al, 1997) and others (Nishida et al, 1991; Tabibzadeh, 1992; Bischof et al, 1993; Bruess et al, 1993; Klentzeris et al, 1993, Bridges et al, 1994; Ruck et al, 1994; van der Linden, et al, 1994; Beliard et. al., 1996; Taylor et al, 1996) have investigated integrins in the endometrium. The number of such reports are increasing and demonstrate the interest in endometrial integrins and their potential role in health and disease. A major focus on integrins in the endometrium has been the potential involvement of these cell adhesion molecules in embryo-endometrial interaction at the time of implantation. A window of implantation was perhaps first suggested by Finn (1977) and has been studied in animal models (McLaren and Michie, 1956; Hodgen, 1983). In the human, the timing of this window was suggested by Hertig et al (1956), who performed secretory phase hysterectomies on pregnant women and demonstrated that embryos originated from uteri obtained prior to cycle day 20 were free floating and embryos from day 21 and beyond were all attached to the uterine wall. These data and more recent studies utilizing donor/recipient cycles (Navot et al, 1991; Bergh and Navot, 1992) suggest that attachment occurs around cycle day 20 of an idealized 28 day cycle and that there are temporal limits placed on the embryo's ability to implant. There have been many molecular markers proposed to identify this period of receptivity (for review see Ilesanmi et al, 1993). We and others have focused on the integrins as potential markers of uterine receptivity (Lessey et al, 1992, 1994a; Tabibzadeh, 1992) and noted that several integrins undergo alterations in the epithelium and decidua during implantation (Lessey et al, 1994a; Ruck et al, 1994). The co-expression of three integrins, 0^(3], ot4(3i and av(33, frames this putative window of implantation (Figure 2). The endometrial epithelial integrin av(33 appears on the surface of the endometrial epithelium coincident with the opening of this window (cycle days 19-20). This pattern of ocv|33 has subsequently been verified by Mardon and co-workers (Rai et al, 1996). This particular integrin recognizes the three amino acid sequence arg-gly-asp (RDG) which has been implicated in trophoblast attachment and outgrowth (Armant et al, 1986; Yelian et al, 1995). Decreased ocvp3 expression during the window of implantation has been observed with luteal phase deficiency (LPD) (Lessey et al, 1992), endometriosis (Lessey et al, 1994b), unexplained infertility (Lessey et al, 1995) and hydrosalpinges (Meyer et al, 1997). Recent exciting data suggests that the av(33 integrin binds and activates matrix metalloproteinases and plaminogen activators (Brooks et al, 1996; Yebra et al, 1996). Since avf33 expression has been documented on the apical portion of the epithelium on both the luminal surface of the endometrium (Aplin et al, 1996; Lessey et al, 1996a) and the surface of the embryo (Campbell et al, 1995), these molecules could serve to activate and position enzymes that are thought to participate in implantation. Coupled with recent data which supports a role for proteolytic fragments of basement membrane as a signal for migration, and thus trophoblast invasion (Giannelli et al, 1997), these patterns of integrin expression have proven to be quite interesting and continue to provide clues to the factors that favour the 249 Downloaded from https://academic.oup.com/humrep/article-abstract/13/suppl_3/247/799170 by guest on 23 November 2017

B.A.Lessey

UJ

DC

o o X

/lenses

Ovulation Implantation day 14 days 20-24

Pregnancy

Figure 2. The pattern of cycle specific endometrial integrin expression during the menstrual cycle and early pregnancy. Note the co-expression of all three integrins (oqpj, a$h and av(33) during the putative window of implantation. Reproduced with permission from the American Society of Reproductive Medicine (Lessey et al, 1994a).

establishment of receptivity and potential mechanisms of endometrial mediated trophoblast invasion. Other integrins have been evaluated in the endometrium of humans and in animal and cell line models. The av(33 integrin has been described on the baboon endometrium, though its expression is delayed for ~2 weeks compared with the human (Fazleabas et al, 1996). This delay corresponds to the delayed invasion noted for the placenta in this primate and suggests that av(33 may be involved in the establishment of the invasive phenotype. The integrin avf35, similar to ocv(33, is present on both the endometrial and embryonic epithelium at the time of implantation (Aplin et al, 1996; Campbell et al, 1995). These two integrins may, therefore, serve different functions, as suggested by Tran and Illsley (1997) who showed that av{35 but not av(33 was involved in trophoblast attachment to RL95 endometrial cells. Using these cells, Thie et al (1997), have expanded their work on the role of apical integrins and shown that the trophoblast adhesion is dependent on the polarization of the epithelium and an intact cytoskeleton. Pytela and co-workers have described the ocvp6 fibronectin and tenascin receptor (Busk et al, 1993; Yokosaki, et al, 1996) as having a role in epithelial function (Breuss et al, 1995) and is present on the endometrial luminal epithelium of the primate uterus (Bruess et al, 1993). The a4(37 integrin is primarily found on leukocytes and endothelium (Brezinschek et al, 1996) and Campbell described this integrin on the human oocyte (Campbell et al, 1995). The otv|38 integrin, 250 Downloaded from https://academic.oup.com/humrep/article-abstract/13/suppl_3/247/799170 by guest on 23 November 2017


described as a vitronectin receptor, has not yet been extensively studied and its distribution in reproductive tissues is not known. Two new integrins, agf^ and a9fih recognize fibronectin and vitronectin (Schnapp et al, 1995) and tenascin respectively (Palmer et al, 1993), and have been examined in the endometrium (Lessey et al, 1996a). While a 8 Pi was not seen, the ocgp! was present on the luminal epithelium throughout the menstrual cycle and on the glandular epithelium only during the luteal phase. This pattern of integrin expression follows somewhat the cyclic nature of tenascin distribution in the cycling endometrium (Harrington et al, 1997). The oc9Pi tenascin receptor has also been described on the mouse embryo and is thought to be involved development and differentiation of tissues (Wang et al, 1995).

Integrins and the regulation of cellular function in vitro It has long been recognized that cell shape and polarity are important to the maintenance of cell function (Folkman and Moscona, 1978; Gospodarowicz et al, 1978). The role of integrins in this process have been elegantly demonstrated by Bissell and Streuli, in mammary epithelium from mice (Streuli and Bissell, 1991; Streuli et al, 1991). Integrin expression has been studied in endometrial cells in vitro. In human stroma, integrin expression appears to be regulated, independent of sex steroid effects (Grosskinsky et al, 1996). Recent studies comparing epithelial and stromal cells in culture also failed to find a significant effect of steroid hormones on integrin expression, but clear differences between cell types were noted (Sillem et al, 1997). There have been several studies which demonstrate that basement membrane (matrigel) alters morphology and establishes a polarized phenotype in endometrial cells in vitro (Bentin-Ley et al, 1994; Classen-Linke et al, 1997). Matrixmediated changes in gene expression have been demonstrated (Hopfer et al, 1996; Strunck et al, 1996). Using the endometrial cell line, HEC IB (Behrens et al, 1996) demonstrated that laminin, but not collagen mediated these changes, likely through the oc6(34 integrin (Strunck et al, 1996).

Regulation of endometrial integrin expression The regulated expression of secretory phase integrins suggested that steroid hormones likely play a role in their appearance. It is well known that oestrogen and progesterone prepare the endometrium for implantation. With the discovery of the o^P] collagen/laminin receptor (VLA-1) and its expression on secretory phase endometrial epithelium (Lessey et al, 1992; Tabibzadeh, 1992), it seemed likely that progesterone directly up-regulates this integrin in endometrium. Since that original observation, we have used the well-differentiated cell line, Ishikawa cells, to demonstrate that progesterone stimulates expression of the a{ integrin subunit (Lessey et al, 1996b; Castelbaum et al, 1997). This is the first integrin 251 Downloaded from https://academic.oup.com/humrep/article-abstract/13/suppl_3/247/799170 by guest on 23 November 2017

B.A.Lessey

that has been shown to be clearly hormonally regulated. While yet to be demonstrated, we believe that OC4PJ will also turn out to be regulated by progesterone in the endometrium. It is more difficult to understand what regulates the expression of epithelial av(33 integrin 6-8 days after ovulation, since the timing of expression occurs well after the rise in serum progesterone (Lessey et al, 1992). In 1988, we (Lessey et al, 1988) and others (Garcia et al, 1988; Press et al, 1988), discovered that secretory phase endometrial epithelium selectively lose the progesterone receptor (PR) while stromal cells maintain expression of this receptor. The loss of PR results from the rise in serum progesterone concentrations which effectively reduces PR as well as the oestrogen receptor (ER). These shifts in receptor content occur in close proximity to the onset of uterine receptivity on cycle days 19-20 (post-LH surge days 5-6) concomitant with the appearance of ccvp3. From this observation, the hypothesis was developed that av(33 should be inhibited by the sex steroids and that loss of steroid receptors would enhance the expression of this integrin. This hypothesis was supported in a study we published in 1996, which correlated the expression of av(33 with endometrial PR and histology, in women with luteal phase deficiency (Lessey et al, 1996c). Histological delay in these patients was associated with the absence of ccv(33 integrin and a persistence of the epithelial PR levels. Successful medical treatment that restored normal histology also resulted in a return of both av(33 integrin expression and loss of epithelial PR. The mechanism of steroid action may be indirect, through elaboration of specific growth factors. We have used the Ishikawa cell line to demonstrate that transforming growth factor a (TGFa) and epidermal growth factor (EGF) significantly enhance the expression of the av(33 integrin (Castelbaum et al, 1997; Somkuti et al, 1997). These principles have been corroborated using the P3 promoter sequences in reporter gene constructs in collaboration with Dr Paul F.Bray (Johns Hopkins University, Baltimore, MD, USA) (Yuan et al, 1996). Based on these and ongoing studies we now believe that the initiation of uterine receptivity depends on the local withdrawal of steroid action in the endometrial epithelium and steroid-mediated paracrine effects from the stroma. As shown in Figure 3, this specific down-regulation of epithelial steroid receptors in the presence of high concentrations of circulating oestrogen and progesterone, results in a functional partitioning of the secretory phase into two segments. As shown in Figure 3A,C, when progesterone and PR concentrations are high both epithelial and stromal cells respond to the effects of ovarian steroids. When PR (and ER) are selectively down-regulated in the epithelial cells (Figure 3B,C), the sex steroids act primarily on the stromal cell which may then influence epithelial cells directly through elaboration of specific paracrine factors. We postulate that receptivity is regulated by this shift in cellular influence, indirectly through the action of progesterone. 252 Downloaded from https://academic.oup.com/humrep/article-abstract/13/suppl_3/247/799170 by guest on 23 November 2017


i:

? ^ " / r' '*>' J*?4

f

' • •