Perlecan is a large multidomain heparan sulfate proteoglycan of the extracellular matrix. Expression of this proteoglycan changes dynamically during embryo implantation and placentation. Perlecan is expressed by various cells of the embryo including trophectoderm and trophoblast as well as the maternal compartment, including basal lamina underlying uterine epithelia and endothelia and, most dynamically, in developing decidua. Perlecan supports various biological functions, including cell adhesion, growth factor binding, and modulation of apoptosis. Moreover, studies in other systems demonstrate that perlecan expression and activity can be controlled at many levels, including transcription, alternative splicing, and extracellular proteolysis. This review will discuss changes in perlecan expression that occur during embryo implantation and placentation. Furthermore, we propose a model in which perlecan represents an extracellular scaffold protein that supports complex, distinct functions in its full-length form or smaller forms generated by alternative mRNA splicing, extracellular proteolysis, or glycosidase action.
Heparan sulfate proteoglycans (HSPGs) are defined as proteins containing one or more covalently attached heparan sulfate (HS) chains. HSPGs can be divided into three major classes: (1) lipid-anchored, e.g. the glypicans; (2) transmembrane, e.g. the syndecans; and (3) extracellular or secreted. The latter class includes several unrelated proteoglycan cores, including collagen type XVIII, agrin, and perlecan. Perlecan is a large multifunctional HSPG found in virtually all basal lamina as well as in the interstitial matrix of certain tissues, including cartilage (Farach-Carson et al. 2005), bone stroma (Schofield et al. 1999), and uterine decidua (French et al. 1999). In addition to its structural role, both the HS constituents and the protein core of perlecan support multiple biological activities relevant to many processes that occur during normal tissue homeostasis and embryonic development as well as pathologies. The normal processes of embryo implantation and placentation are quite complex and require the full spectrum of these activities. Therefore, this system provides an excellent platform to examine perlecan functions. This review will discuss the key events of embryo implantation and placentation, the complexities of perlecan biology, and the potential functions of perlecan as an extracellular scaffold protein modulating events occurring in early mammalian pregnancy and development.
Embryo implantation and placentation
Embryo implantation is a well-coordinated process in which the uterus matures to a state in which it can support embryo attachment. This maturation is controlled by the actions of ovarian steroids produced by corpora lutea. This endocrine mechanism links preimplantation embryonic development to uterine development. The mammalian egg is released from the mature follicle, leaving behind the hormone-producing corpus luteum. Following fertilization, the single-cell zygote undergoes a series of cell divisions within the glycoprotein coat of the zona pellucida, ultimately giving rise to the blastocyst, a fluid-filled structure containing the cells of the inner cell mass and surrounded by a layer of trophectoderm. After hatching from the zona pellucida, the trophectoderm will mediate embryo attachment to and invasion of the endometrium, and subsequently develop into hormone-producing trophoblast and form the placenta. The placenta itself is a complex organ consisting of several trophoblastic cell types, which provides hormonal signals and support fetal nutrition and waste exchange by establishing contact with the maternal blood supply. Reviews of these processes, including differences that occur among species, are available (Carson et al. 2000; Chaddha et al. 2004; Enders and Carter 2006).
The endometrium undergoes a cyclic process of differentiation in response to the actions of ovarian steroids. This involves proliferation of both epithelial and stromal elements leading to endometrial growth. Toward the latter half of this cyclical process, the endometrium becomes transiently receptive with regard to embryo attachment, after which point it transitions to a refractory state wherein implantation cannot occur until hormone levels fall and the endometrial cycle can again start. Thus, the attachment-competent blastocyst must arrive in the uterus during this “window” of uterine receptivity. Minimally, the receptive uterine state requires that the epithelial cells lining the luminal surface be capable of binding the attachment-competent blastocyst. A variety of adhesion-promoting molecules have been described at the surfaces of attachment-competent blastocysts and receptive phase uterine epithelia, including HSPGs and their binding proteins [reviewed in Carson et al. (2000)]. Nonetheless, gene knockout studies in mice indicate that none of these candidates are absolutely required to support the initial events of embryo attachment. This indicates that there is considerable redundancy in the systems involved in early embryo implantation. The attachment systems which embryos can utilize in various species have been reviewed (Carson et al. 2000; Aplin and Kimber 2004). In response to embryo attachment, the uterine stroma undergoes a differentiation process resulting in the formation of decidua. Decidual tissue formation includes both extracellular matrix (ECM) remodeling and de novo expression of ECM components, not usually found in interstitial ECM, including laminin, collagen type IV, and perlecan [reviewed in Murray and Lessey (1999) and Schatz et al. (1999)].
Perlecan expression in embryo implantation and placentation
Most of the available data on perlecan expression during early mammalian embryonic development have been derived from studies of mice. Perlecan is first detected at the four- to eight-cell stage (Dziadek et al. 1985); however, this expression is transient because other studies indicate that perlecan is not expressed in blastocysts until after hatching from the zonae pellucidae (Smith et al. 1997). This timing parallels increases in HSPG synthesis (Farach et al. 1987). At this point, perlecan is found on the external surface of the trophectoderm, i.e. at a position where it could participate in blastocyst attachment. In this regard, HS-dependent interactions support embryo attachment to various substrates, including laminin, fibronectin, and primary cultures of uterine epithelial cells (Farach et al. 1987). It is not clear which cell surface components retain perlecan at the trophectodermal cell surface; however, β1- and β3-containing integrins have been identified as perlecan receptors in other systems (Hayashi et al. 1992; Brown et al. 1997) and these integrin complexes have been identified on the trophectoderm surface (Armant 2005). As mentioned above, the initial interactions with the uterine epithelial surface appear to be HS dependent. In this regard, the transmembrane form of heparin binding-epidermal growth factor (HB-EGF) is induced locally at embryo implantation sites (Das et al. 1994). Other candidate HS-binding molecules expressed by uterine epithelial cell surfaces include amphiregulin (Das et al. 1995) and HIP/RPL29 (Rohde et al. 1996). Null mutants have been generated to all of these proteins and none display an overt implantation defect, again probably due to redundancy in function (Luetteke et al. 1999; Iwamoto et al. 2003; Kirn-Safran, Oristian, et al. 2007; Kirn-Safran, Oristian, Focht, et al. 2007). It should be noted that all of these proteins may participate in implantation. Loss of any one (or two) of these proteins may still leave enough HS-binding activity to support this process due to redundancy of function. It also is possible that expression of other HS-binding proteins is elevated when one is lost. Finally, HS-binding proteins that have not yet been identified or studied in the context of implantation may play key roles. Thus, considerably more work will need to be done to assess fully the role of HS-dependent interactions during early stages of embryo implantation.
Perlecan is induced locally in decidual tissue surrounding the implantation site, first in the primary decidua and later in the secondary decidual zone (Figure 1; French et al. 1999). In situ hybridization results indicate that increases in perlecan mRNA expression are transient; however, immunostaining data indicates that perlecan protein expression is robust and persistent. Together, these observations suggest that perlecan protein is quite stable in this environment, in spite of the extensive metalloprotease-dependent remodeling that occurs (Murray and Lessey 1999; Schatz et al. 1999). As indicated in Figure 2, perlecan is cleaved at discrete sites by certain matrix metalloproteases (MMP), including MMP3 (stromelysin), MMP1 (collagenase), and plasmin (Whitelock et al. 1996), as well as by the cell surface proteases, membrane-type MMPs (d'Ortho et al. 1997). More recently, the C-terminal portion of perlecan has been shown to be a substrate for bone morphogenetic protein (BMP)-1/Tolloid-like metalloproteases (Gonzalez et al. 2005). It is not clear if the perlecan present in endometrial tissues is primarily in an intact form, is cleaved or if cleavage only occurs in select regions of decidua. Experiments should be performed to test these possibilities, especially because certain perlecan fragments may have activities distinct from those of intact protein. For example, a C-terminal fragment released from endothelial cell-derived perlecan protects fibroblasts from apoptosis (Laplante et al. 2006) and may play a similar role in decidua. A similar fragment (endorepellin) has been shown to be antiangiogenic (Gonzalez et al. 2005) and could contribute to the lack of vasculature observed in primary decidual tissue (Parr et al. 1986).
The high level of pericellular perlecan expression in decidua may impact implantation and placentation at several levels. Through its ability to bind growth factors and cytokines, perlecan can control delivery of these proteins by limiting their diffusion from local sites of expression and concentrating them at the site of rapid embryo development, trophoblast differentiation, and placental formation. Perlecan also is likely to play an important role in neo-vascularization associated with embryo implantation and placental development. Perlecan is well known to bind and enhance the activities of angiogenic growth factors such as vascular endothelial growth factor (VEGF) and fibroblast growth factors (FGFs) (Jiang and Couchman 2003) and, thus, is likely to promote the massive tissue vascularization that occurs during placentation. Moreover, endothelial cells directly bind to the perlecan protein core (Hayashi et al. 1992). This latter activity is modulated by the presence of glycosaminoglycan chains. Thus, ECM remodeling enzymes, heparanases or metalloproteases, that also are abundantly expressed in decidua can convert perlecan from a form that may stimulate endothelial cell growth and migration to forms that better support stable endothelial cell adhesion and inhibit further endothelial cell migration and proliferation, e.g. by the production of endorepellin (Gonzalez et al. 2005).
Perlecan expression has been examined in developing human placentae (Rohde et al. 1998; Yang et al. 2005). As expected, it is found in all basal laminae of these structures; however, perlecan also accumulates in Nitabuch's membrane (Rohde et al. 1998), an area of attachment at the maternal–fetal interface placental villi and the decidua (Figure 1). Curiously, perlecan expression is elevated and the glycosaminoglycan composition is altered in placentae in women with diabetes mellitus (Chen et al. 2007). In the latter case, there is an increase in the chondroitin/dermatan sulfate content in the diabetic placentae. The functional significance of these observations is not clear, but it has been suggested that these changes may be associated with changes in placental structure. Another report, based entirely on immunostaining of mouse embryos, indicates that perlecan expression is decreased by interferon-γ (Fontana et al. 2004). In spite of the nonquantitative nature of these studies, they nonetheless are consistent with observations in cell lines indicating that interferon-γ potently inhibits perlecan promoter activity (Sharma and Iozzo 1998). A prediction of these studies is that inflammation that can occur in response to infection during pregnancy would reduce perlecan expression and contribute to spontaneous abortion and recurrent miscarriage associated with this state (Raghupathy 2001; Laird et al. 2006). Relatively, little is known about transcriptional regulation of perlecan. Such studies are needed in the context of endometrial and placental biology since they might identify pathways and potential therapeutic approaches to maintain perlecan expression during inflammatory challenge.
Physiological consequences of perlecan mutations
Perlecan mutations have been created or identified in a number of species ranging from fruit flies to humans. Although implantation and placentation does not occur in the nonmammalian species, the defects observed in these cases may be instructive when thinking about the potential impact on mammalian reproductive events. In Drosophila, perlecan mutations occur in a gene called terribly reduced optic lobes or trol (Voigt et al. 2002). The trol mutations appear to be manifest primarily in the nervous system, acting downstream of an antiproliferative gene, Ana, a secreted protein with antiproliferative activity. The trol phenotype results in failure of quiescent neuroblasts to initiate proliferation resulting in small eyes and brains; however, this defect can be overcome by increasing the expression of string, a Cdc25 ortholog (Park et al. 2003). While the placenta is not innervated, the endometrial vasculature is (Akerlund 1994; Sastry 1997). Thus, proper uterine innervation would be required not only for myometrial contractile activity, but also for controlling uterine blood flood in support of the developing embryo and fetus. Defects in controlling proper uterine blood flow underlie preeclampsia, although no clear relationship between preeclampsia and perlecan expression, function, or processing has been determined.
Other studies indicate that Drosophila perlecan defects can be rescued by exogenous FGF-2 and Indian hedgehog; however, binding studies indicate that hedgehog binds to perlecan in a HS-independent manner, presumably through interactions with the protein core (Park et al. 2003). Hedgehog proteins also bind to perlecan and initiate signals in mammalian cells (Datta et al. 2006). Collectively, these observations indicate that perlecan plays a role in both HS-dependent and -independent signaling by controlling gradients of secreted factors regulating cell proliferation. From another perspective, Indian hedgehog is expressed in uterine epithelium and is a progesterone-responsive gene (Takamoto et al. 2002). While Indian hedgehog expression appears to decrease shortly after embryo attachment, a series of other HS-binding proteins, particularly those of the BMP/BMP-antagonist family are induced robustly in decidua where perlecan probably plays a role in restricting their diffusion (Paria et al. 2001). Thus, features of perlecan function determined in studies of lower animals have parallels in uterine physiology.
In Caenorhabditis elegans, perlecan has been identified as the product of the UNC-52 gene (Rogalski et al. 2001). Lack of perlecan/UNC-52 in nematodes results in severe body-wall muscle defects and appears to be related to abnormal formation of integrin-containing complexes, although it is not clear if perlecan interactions in this regard are direct or indirect [reviewed in Rogalski et al. (2001)]. Perlecan is highly expressed in basal lamina underlying uterine longitudinal and circular myometrial cells as well as uterine vasculature. Therefore, perlecan is likely to participate in neuromuscular junction organization and control of uterine blood flow, as mentioned above. In addition, UNC-52 defects impact the developmental program of gonads apparently through impaired growth factor signaling (Merz et al. 2003). UNC-52 gene products can occur as one of three major size variants generated by alternative splicing due to the presence of polyadenylation sites downstream of exons (Dziadek et al. 1985; Merz et al. 2003; Yang et al. 2005). The shortest form (S) includes domains I–III, the intermediate-size form (M) contains domains I–IV, while the longest form (L) contains all five domains. Additional minor splice variants have been detected that yield differences in domains III and IV as well as a potential for > 50 splice variants [reviewed in Mullen et al. (1999); and Rogalski et al. (2001)]. In this system, several RNA-binding proteins have been identified by genetic screens, which appear to regulate UNC-52 alternative splicing, including Smu-1, Smu-2, and Mec-8 (Spike et al. 2002; Spartz et al. 2004). C. elegans mutants have been generated that prevent formation of the L and M isoforms and demonstrate that domain V is not essential for normal muscle formation and function; however, mutations that eliminate the M and L isoforms phenocopy UNC-52/perlecan nulls, demonstrating that the S-form cannot support the associated biological activity (Merz et al. 2003). Because the S-isoform would carry as much glycosaminoglycan as the M isoform, these observations demonstrate that interactions with domain IV are essential.
Mammalian perlecan also is a large complex gene with 94 exons (Cohen et al. 1993) and, therefore, has great potential to undergo alternative splicing; however, only one report of alternative splicing of perlecan in mammalian systems has appeared (Joseph et al. 1996). It is not clear how widespread alternative splicing is for perlecan, although this seems likely. Recent studies indicate that perlecan variants occur in cartilage although it is not clear if this results from proteolysis, alternative splicing or both (Melrose et al. 2006). Moreover, mammalian mutations in Smu-1 also result in increased expression of perlecan splice variants suggesting that this process is conserved throughout evolution (Sugaya et al. 2006). As mentioned above, given the different functions associated with different perlecan domains, a systematic examination of perlecan splice variants as well as proteolytic processing in fetal–placental tissues is warranted.
Mouse null mutations in perlecan have been created, but display massive developmental abnormalities affecting the heart, brain, kidney, and skeletal tissues, among others (Arikawa-Hirasawa et al. 1999; Costell et al. 1999). Most nulls die around gestational day 11.5, although a small percentage of pups go to term and die shortly afterwards. Surprisingly, placental development appears normal in these animals, indicating that perlecan function is dispensible for development of this tissue. The fact that the few nulls that go to term die shortly after birth makes it impossible to assess the role of perlecan in uterine physiology. Nonetheless, these observations demonstrate that blastocysts do not need to express perlecan to implant. This is similar to observations made with many other null mutations of ECM components or cell-adhesion molecules which collectively indicate that there is considerable redundancy in embryo attachment systems [reviewed in Carson et al. (2000)]. A mouse perlecan mutant has been created that lacks the glycosaminoglycan attachment sites in domain I (Rossi et al. 2003). These animals have defects in lens and kidney development; however, no detailed information about placental or uterine development or function has been reported. Careful studies of the fertility, implantation success, and decidua formation in these mice is warranted. Nonetheless, because these mutants have been successfully bred with collagen XVIII null mice, it appears that they are at least partially fertile (Rossi et al. 2003). Two human mutations in perlecan are known, namely dyssegmental dysplasia, Silverman-Handmaker type [DDSH; MIM#224410; Arikawa-Hirasawa et al. (2001)] and Schwartz-Jampel syndrome type I [SJSI; MIM#255800; Stum et al. (2006)]. Both are rare, autosomal recessive diseases resulting in perlecan truncation mutants. DDSH newborns display massive skeletal abnormalities and die shortly after birth. SJSI is represented by a spectrum of mutations with most patients reaching adulthood with a spectrum of disorders related to the degree of haploinsufficiency, notably severe muscle stiffness (Stum et al. 2005). No systematic studies on fertility of SJSI patients or the placentae for either DDSH or SJSI are available. Recently, the mouse knockin model of SJSI was created with a musculoskeletal phenotype similar to that of humans; however, no information on the impact of this mutation on implantation or placentation is yet available (Rodgers et al. 2007).
Perlecan contains three or more HS chains in unique domain I, and an additional glycosaminoglycan attachment site in domain V, approximately 4000 amino acids away (Figure 3). Individual perlecan domains have unique abilities to interact with various heparin-binding (HB) growth factors such as VEGF, HB-EGF, and FGF-2; hedgehog, ECM, and basement membrane components; and the extracellular domains of cell surface components, together defining the complex perlecan “interactome” (Figure 4). Some interactions are heparin/HS-dependent, some are heparin-influenced, and still others involve core protein interactions that occur independently of heparin/HS. The function of perlecan as a co-receptor mediating heparin-binding growth factor delivery and receptor signaling has been well discussed (Jiang and Couchman 2003; Fjeldstad and Kolset 2005), but these functions involve only a subset of the perlecan domains, and a relatively small portion of the total molecule. The exact functions of the other domains are not well understood, although they offer a wide variety of conserved structural motifs with broad functional potential. A number of authors, including ourselves, have attempted to develop an integrated model that globally defines the function of perlecan in the various tissues where it plays a crucial role in essential processes such as organogenesis, wound healing, and angiogenesis (Farach-Carson et al. 2005; Iozzo 2005). Recently, Knox and collaborators (Knox and Whitelock 2006) suggested that the context of the local extracellular environment defines the function and downstream effects of perlecan on tissue structure and function. It also has been suggested that degradation of intact perlecan generates bioactive fragments such as endorepellin with unique activities distinct from those of the parent HS proteoglycan (Bix et al. 2004). Degradation of perlecan may also play a dynamic role in reproductive processes. For example, selective destruction of perlecan occurs during ovulation in the focal intraepithelial matrix that develops between granulosa cells (focimatrix) and the follicular basal lamina in ovarian follicles (Irving-Rodgers et al. 2006). Amniotic fluid also has been shown to contain a C-terminal fragment of perlecan that may provide a useful prognostic indicator for patients at risk of premature rupture of the fetal membrane (Thadikkaran et al. 2005). Dynamic changes in perlecan expression also occur in the region of the uterus near the implanting embryo in the peri-implantation period with loss of perlecan reported in the endometrial stroma (San Martin et al. 2004) and in the outer surface of the embryonic trophectoderm after the initial stage of embryo attachment (Carson et al. 1993). It is not clear whether the dramatic reductions in perlecan in all of these cases occur in order to generate active functional fragments of perlecan, to remove a function of the intact proteoglycan, or a combination of both effects that may accompany tissue remodeling.
Perlecan—an extracellular scaffold
A major unanswered question is why nature designed the large perlecan molecule as a composite of close to 50 connected, but independently folding, protein modules rather than as a series of smaller proteins. What advantage does linking these functional units confer? As discussed elsewhere in this article, it is not clear to what extent mammalian tissues produce perlecan isoform variants. Conditions that favor increased transcription and translation, or stabilize the functional lifetime of perlecan typically will generate a very large secreted protein with the potential to support many complex activities. While the large domain IV of perlecan has been suggested to function as a “complex cluster of heterotypic interaction sites” supporting ECM assembly (Hopf et al. 2001), the concept of the complete perlecan structure as an extracellular scaffolding protein modulating signaling pathways in target cells has not been explored. This is true despite the provocative hint of this function from two independent observations. First, recent data indicate that perlecan contributes to the assembly of acetylcholine receptor clusters and influences development and signaling within the neuromuscular junction (Smirnov et al. 2005). Second, human mutations producing perlecan haploinsufficiency in the ECM manifest as SJSI, which is characterized by disruptions in cellular electrical activity (Cao et al. 1978). Targeted reduction of perlecan expression alters intracellular signaling (Savore et al. 2005), including signaling pathways known to originate in “signalosomes” held within lipid rafts (Chu et al. 2004). Thus, we undertook a structure and function analysis using published literature and public databases to explore the concept that perlecan has the necessary properties to function as an extracellular scaffold that might directly link the ECM to the cell surface, particularly at sites at which cellular signals are initiated.
Figure 2 shows a representation of the domain structures of perlecan in addition to the proteolytic sites. Domains and subdomains are color-coded and descriptions may be found in Pfam (http://www.sanger.ac.uk/Software/Pfam/). Domain I contains the three glycosaminoglycan attachment sites and the sea urchin sperm protein-enterokinase-agrin (SEA) module (Dolan et al. 1997). Domain II contains a low-density lipoprotein (LDL) receptor domain class A motif which is characterized by six disulfide-bound cysteines and a highly conserved cluster of negatively charged amino acids, of which many are clustered on one face of the module; this domain is known to be both LDL and calcium binding (Costell et al. 1996). Interestingly, such domains are thought to be modulators of Wnt/β- catenin and Wnt/calcium signaling, which play key roles in many biological processes (Yang 2003). Perlecan modulation of Wnt signaling has been reported in Drosophila and in C. elegans (Merz et al. 2003; Nobuo 2003), but has not been examined in mammals. Domain III of Pln contains both laminin B and laminin EGF domains, the latter of which contain repeats of about 60 amino acids in length that include four conserved disulfide bonds and form inflexible rod-like structures (Chakravarti et al. 1995). Domain IV of Pln contains three types of immunoglobulin (Ig) domains all of which have a fold that consists of a beta-sandwich formed of seven strands in two sheets with a Greek-key topology (Hopf et al. 1999). Individual Ig domains are stabilized by disulfide bonds. Such domains typically are found in cell-adhesion domains of a variety of proteins including various cell adhession molecules, and are found in the extracellular domains of sodium channel-β subunits, and are highly interactive in both cell–cell and cell-substratum-binding events (Crossin and Krushel 2000). Pln domain V, cleaved to produce endorepellin (Bix et al. 2004), contains a fourth alternate glycosaminoglycan attachment site (Friedrich et al. 1999) and consists of laminin G and EGF-like domains. The laminin G domain contains a “jellyroll fold” with hydrophobic core residues located in central beta strands, whereas the EGF-like domain is formed from two-stranded β-sheets followed by a loop to a C-terminal short-two-stranded sheet; such domains usually contain three disulfides. Endorepellin has been reported to potently inhibit four aspects of angiogenesis, including endothelial cell migration, collagen-induced endothelial tube morphogenesis, and blood vessel growth in the chorioallantoic membrane and in Matrigel® plug assays (Mongiat et al. 2003). The function of this domain in preventing angiogenic invasion is an intriguing possibility.
Unlike previous linear depictions, Figure 3 shows a new composite model for intact perlecan based on available images obtained using rotary shadowing of individually expressed domains and atomic force microscopy (Chakravarti et al. 1995; Costell et al. 1996; Brown et al. 1997; Dolan et al. 1997; Hopf et al. 1999; Chen and Hansma 2000). As shown, the predicted dimensions of the intact perlecan molecule span a distance of some 100–200 nm depending on the degree of twisting, especially of the long-Ig repeat modules in domain IV that span some 60–80 nm. Domain II spans an average distance of 18 nm and domains III and V each span approximately 20 nm. Domain I is the smallest, but is made larger by the presence of the extended glycosaminoglycan chains. To place these dimensions into the context of perlecan as a putative extracellular scaffolding protein, a typical lipid raft signaling microdomain has an average diameter of 6–20 nm. Of interest, the size of these rafts is thought to be able to increase to 100–200 nm by coalescence and stabilization of smaller rafts that are cross linked (Edidin 2003). The size of these rafts has been proposed to influence complex signaling cascades that are scaffolded in plasma membrane microdomains (Nicolau et al. 2006). Furthermore, functional receptors in myocytes have been shown to be organized into multiprotein domains of approximately 140 nm average diameter (Ianoul et al. 2005). Hence, it is very intriguing to speculate that a complex ECM molecule such as perlecan with a diameter of 100–200 nm can serve to cluster extracellular domains of transmembrane proteins, stabilize their interactions, and hence create stable “signalosomes” that can modulate cell function. The structure and location of perlecan in the territorial matrix of various cells make it an ideal candidate to serve as an extracellular scaffold and may provide a novel explanation for why nature created perlecan as a long complex heterofunctional-binding protein.
Summary and future directions
Perlecan expression changes in a dynamic fashion during pre and peri-implantation stage embryo development as well as in decidual and placental tissues at the fetal–maternal interface. These changes correspond with key implantation-related events and are consistent with the proposed roles for perlecan in promoting cell adhesion, growth factor binding, ECM organization and modulation of signaling pathways affecting cell proliferation and apoptosis. Perlecan's large size along with the distinct physical, biochemical, and biological activities of its individual modular domains are consistent with a major function as a macromolecule that provides a support or scaffold for coordination of the bioactivity of multiple, complex cellular, and tissue morphogenetic events. Alternative mRNA splicing and/or proteolytic processing would give rise to perlecan variants or fragments that either are devoid of certain activities or gain new ones, e.g. endorepellin. Future studies in the fetal–placental unit should be aimed at determining to what extent alternative splicing and proteolytic processing of perlecan occurs in these tissues as well as the factors that control perlecan transcription. The critical biological importance of implantation and placentation along with the accessibility of various models of these processes make this an ideal system to study novel aspects of perlecan biology.
The authors appreciate the careful reading and helpful comments of Dr Catherine Kirn-Safran and Ms Sonia D'Souza, who also performed some of the staining and imaging shown in Figure 1. We are also grateful for the excellent secretarial assistance of Mrs Doreen Anderson and Ms Sharron Kingston. The authors were supported by NIH grants HD25235, NCI P01 CA09891.2 and, COBRE P20-RR16458.
Conflict of interest statement
bone morphogenetic protein
dyssegmental dysplasia, Silverman-Handmaker type
fibroblast growth factors
heparan sulfate proteoglycans
Schwartz-Jampel syndrome type I
vascular endothelial growth factor.