Placenta 35, Supplement A, Trophoblast Research, Vol. 28 (2014) S93eS99
Contents lists available at ScienceDirect
Placenta journal homepage: www.elsevier.com/locate/placenta
Review: The enigmatic role of endoglin in the placentaq A.L. Gregory a, b, G. Xu a, b,1, V. Sotov a, b, M. Letarte a, b, * a b
Molecular Structure and Function Program, Hospital for Sick Children, Department of Immunology, Canada Heart & Stroke Richard Lewar Centre of Excellence, University of Toronto, Canada
a r t i c l e i n f o
a b s t r a c t
Article history: Accepted 31 October 2013
The cellular expression, structure and function of endoglin, and its implication in several vascular disorders remain enigmatic, even 30 years after its discovery. Endoglin (CD105) is a homodimeric glycoprotein (180 kDa) constitutively expressed in the vascular endothelium. It is essential for cardiovascular development and mutations in the ENG gene lead to Hereditary Hemorrhagic Telangiectasia, a disorder characterized by arteriovenous malformations. Endoglin is also expressed in the syncytiotrophoblast throughout pregnancy, but transiently upregulated in the extravillous trophoblast of anchoring villi. Endoglin modulates responses to several TGF-b superfamily ligands and is essential for the negative regulation by TGF-b isoforms 1 and 3 of extravillous trophoblast differentiation. Membrane endoglin binds endothelial NO synthase and regulates its activation and vasomotor tone. There is also a circulating soluble form of endoglin (sEng; 65 kDa); its levels in the serum of women with preeclampsia are increased and correlated with disease severity. The exact sequence of sEng is still unresolved and the proposed mechanism of release from the syncytium by metalloproteases would not yield the expected size protein. The nature of the ligand sequestered by sEng is also an enigma. sEng is said to block the effects of TGF-b on NO-mediated vasorelaxation. However, sEng alone cannot scavenge these ligands for which it has very low affinity. sEng binds with high affinity to BMP9, which stimulates secretion from endothelial cells of the vascoconstrictor endothelin-1, also implicated in endothelial cell stabilization. It remains to be determined if scavenging of circulating BMP9 by sEng is important in preeclampsia and regulation of hypertension. Ó 2013 Published by IFPA and Elsevier Ltd.
Keywords: Vascular endothelium TGF-beta BMP9 Preeclampsia Endoglin Endothelin-1
1. Introduction There are now close to 1300 papers on endoglin listed on PubMed, thirty years after its discovery. Its constant versus transient expression pattern on a restricted number of cell types, the ability of membrane endoglin to interact with multiple ligands of the TGF-b superfamily, the exact sequence of the soluble protein (sEng), its mechanism of production from the placenta and its ligand specificity, are all fascinating but remain controversial aspects of endoglin biology. We will review endoglin distribution, structure, ligand specificity, and role in the vascular
q This research was funded by grant # NA 7093 from the Heart and Stroke Foundation of Canada. * Corresponding author. Hospital for Sick Children, 555 University Ave, Toronto M5G1X8, Canada. Tel.: þ1 416 813 6258; fax: þ1 416 813 7877. E-mail addresses: [email protected] (A.L. Gregory), guoxiongxu@ hotmail.com (G. Xu), [email protected] (V. Sotov), michelle.letarte@ sickkids.ca (M. Letarte). 1 Present address: Center Laboratory, Jinshan Hospital, Fudan University, Shanghai 201508, China. 0143-4004/$ e see front matter Ó 2013 Published by IFPA and Elsevier Ltd. http://dx.doi.org/10.1016/j.placenta.2013.10.020
endothelium, particularly in the context of normal pregnancy and preeclampsia. 2. Distribution and function of endoglin 2.1. Vascular endothelium We first identified endoglin with the monoclonal antibody 44G4, raised against cell surface proteins prepared from a childhood leukemia cell line [1]. It soon became apparent that this glycoprotein was found in blood vessels and it was later classified as the endothelial marker CD105 [2]. Activation of endothelial cells leads to increased expression of endoglin, in response to an angiogenic or inflammatory stimulus [3,4]. Endoglin is in fact widely used as a marker of tumor angiogenesis [3,5]. Although there is higher endoglin on dividing cells and none on senescent cells, it is not required for proliferation as endoglin-null endothelial cells divide faster than normal cells [6]. Endoglin is likely important for maintaining vascular homeostasis and quiescence of the vascular endothelium. In fact, mutations in the ENG gene lead to Hereditary Hemorrhagic Telangiectasia (HHT) type 1 [7], a vascular
S94
A.L. Gregory et al. / Placenta 35, Supplement A, Trophoblast Research, Vol. 28 (2014) S93eS99
dysplasia characterized by arteriovenous malformations and severe and frequent bleeding episodes. The disease is due to haploinsufficiency and reduced functional levels of endoglin [8]. We demonstrated that the underlying defect is at least in part due to endothelial dysfunction caused by disrupted association of endoglin with endothelial NO synthase (eNOS), leading to generation of superoxide and impaired vasomotor function [9]. 2.2. Mesenchymal and hematopoietic cells Endoglin is expressed in mesenchymal stem cells [10] and is required for hemangioblast and early hematopoietic development [11,12]. Endoglin is absent from most immune cell subsets but is induced by activation of monocytes into macrophages [13]. The role of endoglin may be to conserve stem cell potential, versus differentiation of precursor cells, via its modulation of TGF-b and BMP effects on lineage determination. TGF-b acts as a potent immunosuppressive agent by inhibiting cell growth, inducing apoptosis and contributing to the generation of T regulatory cells. Therefore the role of endoglin may be to tightly regulate cell differentiation and activation by TGF-b superfamily members. A striking example of transient expression of endoglin is during the endocardial-mesenchymal transformation that leads to heart septation and valve formation [14]. Endoglin-null embryos die at mid-gestation due to defects in vessel and heart development [15], indicating that endoglin is essential for both angiogenesis and heart development. Endoglin is also present on fibroblasts at the edge of a wound and on perivascular stromal cells involved in vascular remodeling [16]. Additionally, endoglin is found on vascular smooth muscle cells, particularly after arterial injury, and could mediate the effects of TGF-b on their migration and tissue repair [17]. 2.3. Placenta The human placenta is an abundant source of endoglin. We first purified endoglin from human term placenta, using monoclonal antibody 44G4; this led to cloning of the ENG gene from a placental library [18]. Immunostaining revealed expression of endoglin in the syncytiotrophoblast of term placenta [19], where it is actually present throughout pregnancy. Endoglin is also transiently expressed in extravillous trophoblast (EVT) in the proximal columns of anchoring villi [20]. We showed previously that monoclonal antibody 44G4, or antisense oligonucleotides to endoglin, stimulated budding and outgrowth of EVT from villous explants at 5e8 weeks of gestation [21], a process that is negatively regulated by TGF-b1/b3 [22]. Fibronectin synthesis and changes in integrin expression also occur during EVT differentiation. Furthermore, the distribution of endoglin on EVT parallels that of the a5b1 integrin [20], which was recently shown to interact with endoglin [23]. The interaction is mediated by the RGD peptide of human endoglin or the equivalent TDD motif found in pig and mouse endoglin [18,23]. We propose that the concurrent induction of endoglin and a5b1 integrin on budding EVT and their RGD e mediated interaction may contribute to the regulation of EVT differentiation. The subsequent decrease in these proteins in the distal columns may facilitate remodeling of spiral arteries, as reduced endoglin expression is associated with vasorelaxation [9]. 3. Structure of membrane and soluble endoglin 3.1. Membrane endoglin Human endoglin is an integral membrane glycoprotein, composed of an N-terminal signal peptide, an orphan domain, a zona pellucida (ZP) domain, a juxtamembrane region, a transmembrane
Fig. 1. Schematic diagram of membrane, recombinant and soluble endoglin. Membrane endoglin is an integral dimeric glycoprotein composed of two identical monomers of 90 kDa linked by disulphide bonds. The polypeptide is composed of an orphan domain with no known structural features (27e357; residues 1e26 representing the signal peptide) and contains five N-linked glycosylation sites (C). The orphan domain is followed by a zona pellucida domain (ZP; 358e586), and by juxtamembrane, transmembrane and short cytoplasmic domains. The recombinant soluble monomeric protein used in most studies corresponds to the complete extracellular domain of endoglin and has a molecular mass of 80 kDa (1e586). The soluble endoglin (sEng) monomer released from placenta into the maternal serum has a molecular mass of 65 kDa and its C terminal residue is likely shortly after Arg 406, within the ZP domain.
region and a short cytoplasmic domain [18,24] as shown in Fig.1. The protein is highly glycosylated with five potential N-linked sites within the orphan domain and O-glycan sites mostly in the ZP domain. Enzymatic removal of all the sugars leads to an observed reduction of about 25 kDa in molecular mass [25]. Membrane endoglin exists as a disulphide-linked homodimer of 180 kDa that contains 17 cysteine residues [18]. However the position of the interchain disulphide bond(s) has yet to be established. It was suggested that Cys582 is responsible for inter-chain bonding although this residue is not conserved in all species. Studies of a CD31-endoglin hybrid molecule indicated that cysteine residues within the fragment Cys330eCys412 were implicated [26]. Our own studies showed that Eng1e357 is expressed as a dimer, leaving Cys330 or Cys350 as the most likely residues involved in dimer formation (unpublished data).
3.2. Soluble endoglin Several studies reported the increased expression of a circulating form of endoglin in serum, plasma or other fluids from cancer patients, as summarized in a review by Bernabeu et al. [5]. This form is referred to as soluble endoglin (sEng), although its true solubility has not been established and it could be associated with other proteins [27]. A 80 kDa sEng was recently shown to be cleaved from the surface of human umbilical vein endothelial cells by MMP14 (or MT1-MMP); the cleavage site was shown to be at position 586, implying that the whole extracellular domain (Eng1e 586) was released [28]. However it remains to be proven that a similar mechanism occurs in tumors. For several years, we (and others) [24,29e31] have used the Eng1e586 construct (Fig. 1) to generate recombinant sEng; this form is soluble to some degree but has not proven satisfactory for structural studies, implying that it might not represent a functional physiological sEng, with a stable structure. A soluble form of endoglin is also present in increased amounts in the serum of women with preeclampsia (PE) [29]. It begins to rise
A.L. Gregory et al. / Placenta 35, Supplement A, Trophoblast Research, Vol. 28 (2014) S93eS99
8e10 weeks before PE onset and correlates well with disease severity, even better that the other soluble receptor associated with PE, namely sFlt-1 (or soluble VEGFR-1) [32]. Unlike sFlt-1, for which alternate splice variants have been described, none has so far been reported for sEng. MMP14 is expressed in preeclamptic placentae [33] and could cleave endoglin from the syncytiotrophoblast, yielding a protein of approximately 80 kDa. When we purified sEng form the serum of preeclamptic women, it was resolved by SDSPAGE as a 65 kDa band, as opposed to the 80 kDa mass of recombinant Eng1e586 [29] (Fig. 1). In our view, it is unlikely that the circulating soluble form of endoglin (sEng) observed in preeclampsia is produced by MMP14 cleavage at position 586. Mass spectrometry analysis of purified endoglin from sera of PE patients, revealed that sEng extended at least to Arginine 406; however we could not identify the C-terminal residue of sEng [29]. Taking carbohydrate side-chains into consideration, we would predict a 40e 45 kDa polypeptide for sEng produced by the placenta, indicating that the C-terminal is shortly after residue 406 (Fig. 1). The mechanism of generation of sEng from placenta, its exact sequence and structure, and functional testing of the purified physiological form must be accomplished to understand its role in placenta and PE.
4. Ligand specificity of membrane and soluble endoglin 4.1. Ligands interacting with membrane endoglin Membrane endoglin bound to radioactive TGF-b was first identified by chemical cross-linking in human umbilical vein endothelial cells, in complex with the types I and II serine kinase receptors [34]. Membrane endoglin interacts with TGF-b1 and TGFb3 but not TGF-b2 isoforms [34]. In most cell types, responses to TGF-b are mediated via the type II receptor (TbRII) and the type I receptor ALK5 [35] (Fig. 2). This pathway is known to inhibit cellular proliferation, mediate fibronectin and collagen production and induce plasminogen activator inhibitor-1 (PAI-1), a negative regulator of cell migration and angiogenesis. This signaling route is potentiated by endoglin in several cell types including cardiac fibroblasts [36]. In endothelial cells, there is an additional pathway mediating the effects of TGF-b, consisting of TbRII in complex with the type I receptor ALK1, and influenced by endoglin expression
Fig. 2. TGF-b and BMP9 signaling pathways in endothelial cells. In endothelial cells, TGF-b1 and eb3 bind to TbRII which then recruits and phosphorylates ALK5 or ALK1. These two type I receptors signal via Smad 2,3 and Smad 1,5,8 respectively. Smad4 is responsible for transfer of Smad complexes to the nucleus where they regulate gene transcription. BMP9 binds to ALK1 and BMPRII and signals via Smad 1,5,8 pathway. Endoglin can modulate all of these pathways in a cell and context-dependent manner.
S95
(Fig. 2). In fact, mutations in ALK1 lead to HHT type 2, implying that endoglin and ALK1 share pathways affected in these diseases [37]. Although the ALK5/Smad 2,3 pathway antagonizes the ALK1/Smad 1,5,8 pathway, ALK5 is essential for recruitment and activation of ALK1 [38]. Membrane endoglin was shown by co-immunoprecipitation studies to interact with several members of the TGF-b superfamily in association with the ligand binding receptor: TbRII for TGF-b1 and TGF-b3; ActRII for activin-A; ActRIIB for BMP7; and the type I receptors ALK3 and ALK6 in the case of BMP2 [39]. More recently, it was shown that overexpression of endoglin promotes responses to BMP9 and BMP10, which are the physiological ligands for ALK1, in a complex with either BMPRII or ActRIIA [40] (Fig. 2). Membrane endoglin can therefore modulate responses to several ligands in a cell- and context-dependent manner. Fig. 3 illustrates the effects of endoglin on Smad 1,5,8 phosphorylation induced by BMP2 and BMP9 in mouse embryonic endothelial cells with either no endoglin (Eng/ cells) or with normal levels (Engþ/þ cells). The response to BMP2 was higher in Eng/ than Engþ/þ cells, indicating that membrane endoglin inhibited BMP2 effects. In the receptor complex, endoglin associates with ALK3, which in turn can interact with ActRII, ActRIIB or BMPRII. ALK6 was absent from the Eng/ cells (as demonstrated by RT-PCR) and could not have mediated the potentiating effect of BMP2. Membrane endoglin potentiated the response to BMP9, as mentioned previously [40]. Here we show higher Smad 1,5,8 phosphorylation in Engþ/þ than Eng/ cells, particularly at low and physiologically relevant BMP9 concentrations. This response requires ALK1 as Alk1-deficient mouse endothelial cells do not bind BMP9 at low ligand concentrations, as reported [41], and as confirmed in our studies. The contribution of endoglin to TGF-b1 induced and TbRII/ALK1/ALK5 mediated Smad phosphorylation has been measured in these cells. We observed potentiation [6] while another group reported inhibition [42] of Smad 1,5,8 phosphorylation by endoglin using the same cells but at different passages. Both groups noted no effect of endoglin on TGFb1 induced Smad 2,3 phosphorylation (TbRII and ALK5 mediated). Thus endoglin can either stimulate or inhibit various responses to members of the TGF-b superfamily, in a ligand-, receptor- and celldependent manner. 4.2. Ligands binding to soluble endoglin It has been repeatedly stated in the literature that high levels of circulating sEng released from placenta during preeclampsia, may alter vascular homeostasis by scavenging TGF-b1 and preventing its binding to membrane receptors on endothelial cells. In an attempt to verify this point, we tested the ability of recombinant sEng (1e586) to block TGF-b1 induced Smad 1,5,8 phosphorylation and compared its activity to that of soluble TGF-b type II receptor (sTbRII). Fig. 4A shows that while sTbRII (extracellular domain only) blocked completely the ability of TGF-b1 to stimulate Smad 1,5,8 phosphorylation in endothelial cells, sEng (1e586) had no effect at the same concentration. Therefore, sTbRII can easily scavenge TGF-b1 while sEng (1e586) cannot. We then tested if sEng (1e586) could bind BMP9 on its own and compared its activity to that of a soluble form of ALK1 (sALK1). Fig. 4B shows that sEng (1e586) and sALK1 are equally able to block BMP9 stimulation of Smad 1,5,8 phosphorylation and can both scavenge circulating BMP9. The direct binding of ligands to immobilized purified recombinant soluble proteins was then assessed using surface plasmon resonance (SPR). Dissociation constants (KD) were estimated from the binding data, as a measure of the affinity of the interactions. Fig. 5A reveals that sEng (1e586) binds TGF-b1 with very low affinity (KD > 1.2 mM) while sTbRII binds with high affinity
A.L. Gregory et al. / Placenta 35, Supplement A, Trophoblast Research, Vol. 28 (2014) S93eS99
(B)
400
200
50
8
Phospho-Smad1/Total-Smad1
Phospho-Smad1/Total-Smad1
100
60kD 0
Smad1 BMP9 (pM)
25
60kD
400
PSmad1
8
6
4
2
0
Eng-/-
200
400
200
50
100
0
25
200
400
50
60kD 100
Smad1 BMP2 (pM)
0
60kD
25
PSmad1
Eng+/+
50
Eng-/-
100
Eng+/+
0
(A)
25
S96
0
100
200
300
400
[BMP2] (pM)
6
4
2
0
0
100
200
300
400
[BMP9] (pM)
Fig. 3. Endoglin differentially regulates BMP2 and BMP9 signaling in endothelial cells. Wild type mouse embryonic endothelial cells (Engþ/þ; C) and endoglin-null cells (Eng/; o) were serum-starved for 30 min and stimulated with increasing concentrations of either BMP2 (A) or BMP9 (B). Cell lysates were resolved by SDS-PAGE and analyzed by Western blot with antibodies specific for phosphorylated-Smad1 and total Smad1. The amount of P-Smad1 was quantified relative to total Smad1 levels. Data represent the mean SEM of 3 experiments. Endoglin inhibited BMP2 but promoted BMP9 effects on Smad1 signaling.
(KD ¼ 100 pM) (Fig. 5B). This implies that sEng (1e586) cannot bind this ligand under physiological conditions and could not scavenge TGF-b1, if present in circulation as a truly soluble protein; however it could do so if circulating in a complex with sTbRII. The binding of BMP9 to sEng and sALK1 was also analyzed by SPR. Fig. 5C and D
show that sEng (1e586) and sALK1 bind BMP9 with similar affinities (sEng, KD ¼ 5 nM; sALK1, KD ¼ 1.8 nM). Our data confirm similar findings by other groups [30,31]. Thus both sEng (1e586) and sALK1 could scavenge circulating BMP9 and block its effects, as suggested by Fig. 4B.
Fig. 4. Blocking of Smad1 phosphorylation by soluble ligand-binding receptors. Wild type mouse embryonic endothelial cells were starved for 2 h, and incubated for 1 h with increasing concentrations of (A) soluble human TGF-b receptor II (sTBRII) or recombinant sEng (1e586) or (B) soluble human ALK1-Fc (sALK1) or recombinant sEng (1e586), and stimulated for 30 min with (A) 100 pM TGF-b1 or (B) 25 pM BMP9. Cell lysates were resolved by SDS-PAGE and analyzed by Western blot with antibodies specific for phosphorylated-Smad1 and total Smad1. Phospho Smad1 was quantified relative to total Smad1. Data are the mean SEM of 3 experiments except for sTBRII where only 2 experiments were performed. sEng blocked BMP9 but not TGF-b1 binding to the cells while sTBRII and sALK1 had the expected ligand specificity.
A.L. Gregory et al. / Placenta 35, Supplement A, Trophoblast Research, Vol. 28 (2014) S93eS99
S97
Fig. 5. Kinetic analysis of the binding of TGF-b1 and BMP9 to their receptors. Purified receptors [recombinant sEng (1e586), sTBRII and sALK1] were amine-coupled to CM5 chips and different concentrations of TGF-b1 (25e800 nM; A and B) or BMP9 (1e500 nM; C and D) were injected over the captured receptors. The raw data (colored lines) are overlaid with a global fit 1:1 model (black lines) obtained by the BIA evaluation software. The resonance signal is plotted against ligand concentration. The estimated affinity constants (KD) are indicated for each ligandereceptor interaction.
As mentioned in Section 3.2, the 65 kDa estimated molecular mass for monomeric sEng does not correspond to the 80 kDa recombinant protein (1e586) used in the blocking and SPR experiments. Therefore, the functional properties of the true physiological form, including its ligand specificity will need to be re-assessed once the sequence of the protein is known and it can be expressed as a recombinant protein. Our data suggest that BMP9 is the more likely ligand to be scavenged by purified sEng associated with placenta and preeclampsia. However, one cannot rule out the possibility that sEng is present in circulation in a complex with sTbRII and even sTbRI that could scavenge TGF-b1 and -b3 and blocked their effects on the vascular system. 5. Regulation of endothelial function by membrane and soluble endoglin 5.1. eNOS and membrane endoglin We demonstrated previously that membrane endoglin, through its short cytoplasmic domain, associates with endothelial NO synthase (eNOS) and Hsp90 in caveolae [9]. ALK1 also associates with eNOS, implying a multimeric complex at the cell surface [43]. The importance of endoglin and ALK1 in eNOS activation is seen in endothelial cells from patients with HHT and in animal models of the disease, which show uncoupling of eNOS, reduced NO production and increased eNOS-derived superoxide production leading to impaired myogenic response [9,44]. In vivo treatment of Engþ/ or Alk1þ/ mice with the anti-oxidant Tempol was able to prevent the onset of age-dependent changes in the pulmonary vasculature, indicating the contribution of eNOS-derived reactive oxygen species to endothelial dysfunction [43,44]. Thus membrane
endoglin plays an important role in the regulation of eNOS activation, initiated by the association of its short cytoplasmic tail with intracellular eNOS in caveolae. 5.2. Endothelin-1 and soluble endoglin Human serum sEng cannot directly scavenge TGF-b, unless a complex of sEng-sTbRII is present. sEng alone will therefore not prevent the phosphorylation and activation of eNOS by TGF-b and the subsequent release of NO, as proposed [29]. We discussed above that sEng can scavenge circulating BMP9. However, in our studies, BMP9 did not induce eNOS phosphorylation and activation and therefore sEng could not act by preventing this process. Other mechanisms to explain how sEng may affect vasomotor function in preeclampsia, are based on the observations that BMP9 can induce production of the potent vasoconstrictor, endothelin-1 (ET-1) [45,46]. BMP9 stimulated ET-1 release from pulmonary endothelial cells under physiological concentrations and in a Smad1 (but not Smad4) dependent manner, involving the ALK1/ BMPRII receptors and the non-canonical p38 MAPK pathway [46]. It was suggested that BMP9 stimulated ET-1 production may be important for vascular stability. However, knocking down ALK1 or BMPRII not only attenuated the BMP9 effects, but also caused a significant increase in prepro ET-1 mRNA transcription and release of the mature peptide [46]. These findings imply that BMP9 and its receptors are important in regulating ET-1 production and consequently hypertension. sEng (1e586) can effectively scavenge BMP9 and would reduce ET-1 release and may lead to reduced vascular stability. Furthermore, ET-1 is elevated in preeclampsia and could increase maternal vasoconstriction; in that context, sEng would have a beneficial effect. However, scavenging BMP9 with sEng during preeclampsia could potentially reduce levels of functional
S98
A.L. Gregory et al. / Placenta 35, Supplement A, Trophoblast Research, Vol. 28 (2014) S93eS99
receptors and lead to a further rise in ET-1 production, and a sustained rise in arterial pressure. 6. Conclusions The enigmatic distribution of endoglin in terms of its persistent expression on vascular endothelium and the syncytiotrophoblast versus its transient expression in mesenchymal cells and the extravillous trophoblast likely will reveal cues as to the functions exerted in these different cell types during development. The sequence and 3D-structure of sEng produced by the placenta and elevated in preeclampsia remains to be established and its ligand specificity determined before its mode of action can be explained. The regulation of vasomotor function by membrane endoglin via its association with eNOS and by soluble endoglin via alteration of ET1 production add to the enigma concerning the essential physiological role of endoglin, as evidenced by altered levels leading to several vascular disorders. Conflict of interest statement The authors confirm that there is no conflict of interest in the above paper. Acknowledgments We wish to thank Zhijie Li and Dr James Rini of the Department of Biochemistry, University of Toronto, for their help and guidance with the surface Plasmon resonance experiments. References [1] Quackenbush EJ, Letarte M. Identification of several cell surface proteins of non-T, non-B acute lymphoblastic leukemia by using monoclonal antibodies. J Immunol 1985;134(2):1276e85. [2] Letarte M, Greaves A, Vera S. In: Schlossman SF, Boumsell L, Gilks W, Harlan J, Kishimoto T, Morimoto C, et al., editors. CD105 (endoglin) cluster report. Oxford: Oxford University Press; 1995. p. 1756e9. [3] Burrows FJ, Derbyshire EJ, Tazzari PL, Amlot P, Gazdar AF, King SW, et al. Upregulation of endoglin on vascular endothelial cells in human solid tumors: implications for diagnosis and therapy. Clin Cancer Res 1995;1(12):1623e34. [4] Torsney E, Charlton R, Parums D, Collis M, Arthur HM. Inducible expression of human endoglin during inflammation and wound healing in vivo. Inflamm Res 2002;51:464e70. [5] Bernabeu C, Lopez-Novoa JM, Quintanilla M. The emerging role of TGF-beta superfamily coreceptors in cancer. Biochim Biophys Acta 2009;1792(10): 954e73. [6] Pece-Barbara N, Vera S, Kathirkamathamby K, Liebner S, Di Guglielmo GM, Dejana E, et al. Endoglin null endothelial cells proliferate faster and are more responsive to transforming growth factor beta1 with higher affinity receptors and an activated Alk1 pathway. J Biol Chem 2005;280(30):27800e8. [7] McAllister KA, Lennon F, Bowles-Biesecker B, McKinnon WC, Helmbold EA, Markel DS, et al. Genetic heterogeneity in hereditary haemorrhagic telangiectasia: possible correlation with clinical phenotype. J Med Genet 1994;31: 927e32. [8] Pece N, Vera S, Cymerman U, White Jr RI, Wrana JL, Letarte M. Mutant endoglin in hereditary hemorrhagic telangiectasia type 1 is transiently expressed intracellularly and is not a dominant negative. J Clin Invest 1997;100(10):2568e79. [9] Toporsian M, Gros R, Kabir MG, Vera S, Govindaraju K, Eidelman DH, et al. A role for endoglin in coupling eNOS activity and regulating vascular tone revealed in hereditary hemorrhagic telangiectasia. Circ Res 2005;96(6):684e92. [10] Barry FP, Boynton RE, Haynesworth S, Murphy JM, Zaia J. The monoclonal antibody SH-2, raised against human mesenchymal stem cells, recognizes an epitope on endoglin (CD105). Biochem Biophys Res Commun 1999;265:134e9. [11] Perlingeiro RC. Endoglin is required for hemangioblast and early hematopoietic development. Development 2007;134(16):3041e8. [12] Borges L, Iacovino M, Mayerhofer T, Koyano-Nakagawa N, Baik J, Garry DJ, et al. A critical role for endoglin in the emergence of blood during embryonic development. Blood 2012;119(23):5417e28. [13] Lastres P, Bellon T, Cabanas C, Sanchez-Madrid F, Acevedo A, Gougos A, et al. Regulated expression on human macrophages of endoglin, an Arg-Gly-Aspcontaining surface antigen. Eur J Immunol 1992;22(2):393e7.
[14] Qu R, Silver MM, Letarte M. Distribution of endoglin in early human development reveals high levels on endocardial cushion tissue mesenchyme during valve formation. Cell Tissue Res 1998;292(2):333e43. [15] Bourdeau A, Dumont DJ, Letarte M. A murine model of hereditary hemorrhagic telangiectasia. J Clin Invest 1999;104(10):1343e51. [16] Matsubara S, Bourdeau A, terBrugge KG, Wallace C, Letarte M. Analysis of endoglin expression in normal brain tissue and in cerebral arteriovenous malformations. Stroke 2000;31(11):2653e60. [17] Ma X, Labinaz M, Goldstein J, Miller H, Keon WJ, Letarte M, et al. Endoglin is overexpressed after arterial injury and is required for transforming growth factor-beta-induced inhibition of smooth muscle cell migration. Arterioscler Thromb Vasc Biol 2000;20(12):2546e52. [18] Gougos A, Letarte M. Primary structure of endoglin, an RGD-containing glycoprotein of human endothelial cells. J Biol Chem 1990;265(15): 8361e4. [19] Gougos A, St. Jacques S, Greaves A, O’Connell PJ, d’Apice AJF, Buhring H-J, et al. Identification of distinct epitopes of endoglin, an RGD containing-glycoprotein of endothelial cells, leukemic cells and syncytiotrophoblasts. Int Immunol 1992;4(1):83e92. [20] St-Jacques S, Forte M, Lye SJ, Letarte M. Localization of endoglin, a transforming growth factor-beta binding protein, and of CD44 and integrins in placenta during the first trimester of pregnancy. Biol Reprod 1994;51(3): 405e13. [21] Caniggia I, Taylor CV, Ritchie JWK, Lye SJ, Letarte M. Endoglin regulates trophoblast differentiation along the invasive pathway in human placental villous explants. Endocrinology 1997;138:4977e88. [22] Caniggia I, Grisaru-Gravnosky S, Kuliszewsky M, Post M, Lye SJ. Inhibition of TGF-beta 3 restores the invasive capability of extravillous trophoblasts in preeclamptic pregnancies. J Clin Invest 1999;103(12):1641e50. [23] Rossi E, Sanz-Rodriguez F, Eleno N, Duwell A, Blanco FJ, Langa C, et al. Endothelial endoglin is involved in inflammation: role in leukocyte adhesion and transmigration. Blood 2013;121(2):403e15. [24] Llorca O, Trujillo A, Blanco FJ, Bernabeu C. Structural model of human endoglin, a transmembrane receptor responsible for hereditary hemorrhagic telangiectasia. J Mol Biol 2007;365(3):694e705. [25] Gougos A, Letarte M. Biochemical characterization of the 44G4 antigen from the HOON pre-B leukemic cell line. J Immunol 1988;141(6):1934e 40. [26] Raab U, Velasco B, Lastres P, Letamendia A, Cales C, Langa C, et al. Expression of normal and truncated forms of human endoglin. Biochem J 1999;339(Pt 3): 579e88. [27] Li CG, Wilson PB, Bernabeu C, Raab U, Wang JM, Kumar S. Immunodetection and characterization of soluble CD105-TGF-ß complexes. J Immunol Methods 1998;218:85e93. [28] Hawinkels LJ, Kuiper P, Wiercinska E, Verspaget HW, Liu Z, Pardali E, et al. Matrix metalloproteinase-14 (MT1-MMP)-mediated endoglin shedding inhibits tumor angiogenesis. Cancer Res 2010;70(10):4141e50. [29] Venkatesha S, Toporsian M, Lam C, Hanai J, Mammoto T, Kim YM, et al. Soluble endoglin contributes to the pathogenesis of preeclampsia. Nat Med 2006;12(6):642e9. [30] Castonguay R, Werner ED, Matthews RG, Presman E, Mulivor AW, Solban N, et al. Soluble endoglin specifically binds bone morphogenetic proteins 9 and 10 via its orphan domain, inhibits blood vessel formation, and suppresses tumor growth. J Biol Chem 2011;286(34):30034e46. [31] Alt A, Miguel-Romero L, Donderis J, Aristorena M, Blanco FJ, Round A, et al. Structural and functional insights into endoglin ligand recognition and binding. PLoS One 2012;7(2):e29948. [32] Levine RJ, Lam C, Qian C, Yu KF, Maynard SE, Sachs BP, et al. Soluble endoglin and other circulating antiangiogenic factors in preeclampsia. N Engl J Med 2006;355(10):992e1005. [33] Kaitu’u-Lino TJ, Palmer KR, Whitehead CL, Williams E, Lappas M, Tong S. MMP-14 is expressed in preeclamptic placentas and mediates release of soluble endoglin. Am J Pathol 2012;180(3):888e94. [34] Cheifetz S, Bellon T, Cales C, Vera S, Bernabeu C, Massague J, et al. Endoglin is a component of the transforming growth factor-beta receptor system in human endothelial cells. J Biol Chem 1992;267(27):19027e30. [35] Chen YG, Hata A, Lo RS, Wotton D, Shi Y, Pavletich N, et al. Determinants of specificity in TGF-beta signal transduction. Genes Dev 1998;12(14): 2144e52. [36] Kapur NK, Wilson S, Yunis AA, Qiao X, Mackey E, Paruchuri V, et al. Reduced endoglin activity limits cardiac fibrosis and improves survival in heart failure. Circulation 2012;125(22):2728e38. [37] Berg JN, Gallione CJ, Stenzel TT, Johnson DW, Allen WP, Schwartz CE, et al. The activin receptor-like kinase 1 gene: genomic structure and mutations in hereditary hemorrhagic telangiectasia type 2. Am J Hum Genet 1997;61(1): 60e7. [38] Lebrin F, Deckers M, Bertolino P, Ten Dijke P. TGF-beta receptor function in the endothelium. Cardiovasc Res 2005;65(3):599e608. [39] Barbara NP, Wrana JL, Letarte M. Endoglin is an accessory protein that interacts with the signaling receptor complex of multiple members of the transforming growth factor-beta superfamily. J Biol Chem 1999;274(2): 584e94. [40] David L, Mallet C, Mazerbourg S, Feige JJ, Bailly S. Identification of BMP9 and BMP10 as functional activators of the orphan activin receptor-like kinase 1 (ALK1) in endothelial cells. Blood 2007;109(5):1953e61.
A.L. Gregory et al. / Placenta 35, Supplement A, Trophoblast Research, Vol. 28 (2014) S93eS99 [41] Choi EJ, Kim YH, Choe SW, Tak YG, Garrido-Martin EM, Chang M, et al. Enhanced responses to angiogenic cues underlie the pathogenesis of hereditary hemorrhagic telangiectasia 2. PLoS One 2013;8(5):e63138. [42] Lee NY, Ray B, How T, Blobe GC. Endoglin promotes transforming growth factor beta-mediated Smad 1/5/8 signaling and inhibits endothelial cell migration through its association with GIPC. J Biol Chem 2008;283(47):32527e33. [43] Toporsian M, Jerkic M, Zhou YQ, Kabir MG, Yu LX, McIntyre BA, et al. Spontaneous adult-onset pulmonary arterial hypertension attributable to increased endothelial oxidative stress in a murine model of hereditary hemorrhagic telangiectasia. Arterioscler Thromb Vasc Biol 2010;30(3):509e17.
S99
[44] Jerkic M, Kabir MG, Davies A, Yu LX, McIntyre BA, Husain NW, et al. Pulmonary hypertension in adult Alk1 heterozygous mice due to oxidative stress. Cardiovasc Res 2011;92(3):375e84. [45] Star GP, Giovinazzo M, Langleben D. Bone morphogenic protein-9 stimulates endothelin-1 release from human pulmonary microvascular endothelial cells: a potential mechanism for elevated ET-1 levels in pulmonary arterial hypertension. Microvasc Res 2010;80(3):349e54. [46] Park JE, Shao D, Upton PD, Desouza P, Adcock IM, Davies RJ, et al. BMP-9 induced endothelial cell tubule formation and inhibition of migration involves Smad1 driven endothelin-1 production. PLoS One 2012;7(1):e30075.
Contents lists available at ScienceDirect
Placenta journal homepage: www.elsevier.com/locate/placenta
Review: The enigmatic role of endoglin in the placentaq A.L. Gregory a, b, G. Xu a, b,1, V. Sotov a, b, M. Letarte a, b, * a b
Molecular Structure and Function Program, Hospital for Sick Children, Department of Immunology, Canada Heart & Stroke Richard Lewar Centre of Excellence, University of Toronto, Canada
a r t i c l e i n f o
a b s t r a c t
Article history: Accepted 31 October 2013
The cellular expression, structure and function of endoglin, and its implication in several vascular disorders remain enigmatic, even 30 years after its discovery. Endoglin (CD105) is a homodimeric glycoprotein (180 kDa) constitutively expressed in the vascular endothelium. It is essential for cardiovascular development and mutations in the ENG gene lead to Hereditary Hemorrhagic Telangiectasia, a disorder characterized by arteriovenous malformations. Endoglin is also expressed in the syncytiotrophoblast throughout pregnancy, but transiently upregulated in the extravillous trophoblast of anchoring villi. Endoglin modulates responses to several TGF-b superfamily ligands and is essential for the negative regulation by TGF-b isoforms 1 and 3 of extravillous trophoblast differentiation. Membrane endoglin binds endothelial NO synthase and regulates its activation and vasomotor tone. There is also a circulating soluble form of endoglin (sEng; 65 kDa); its levels in the serum of women with preeclampsia are increased and correlated with disease severity. The exact sequence of sEng is still unresolved and the proposed mechanism of release from the syncytium by metalloproteases would not yield the expected size protein. The nature of the ligand sequestered by sEng is also an enigma. sEng is said to block the effects of TGF-b on NO-mediated vasorelaxation. However, sEng alone cannot scavenge these ligands for which it has very low affinity. sEng binds with high affinity to BMP9, which stimulates secretion from endothelial cells of the vascoconstrictor endothelin-1, also implicated in endothelial cell stabilization. It remains to be determined if scavenging of circulating BMP9 by sEng is important in preeclampsia and regulation of hypertension. Ó 2013 Published by IFPA and Elsevier Ltd.
Keywords: Vascular endothelium TGF-beta BMP9 Preeclampsia Endoglin Endothelin-1
1. Introduction There are now close to 1300 papers on endoglin listed on PubMed, thirty years after its discovery. Its constant versus transient expression pattern on a restricted number of cell types, the ability of membrane endoglin to interact with multiple ligands of the TGF-b superfamily, the exact sequence of the soluble protein (sEng), its mechanism of production from the placenta and its ligand specificity, are all fascinating but remain controversial aspects of endoglin biology. We will review endoglin distribution, structure, ligand specificity, and role in the vascular
q This research was funded by grant # NA 7093 from the Heart and Stroke Foundation of Canada. * Corresponding author. Hospital for Sick Children, 555 University Ave, Toronto M5G1X8, Canada. Tel.: þ1 416 813 6258; fax: þ1 416 813 7877. E-mail addresses: [email protected] (A.L. Gregory), guoxiongxu@ hotmail.com (G. Xu), [email protected] (V. Sotov), michelle.letarte@ sickkids.ca (M. Letarte). 1 Present address: Center Laboratory, Jinshan Hospital, Fudan University, Shanghai 201508, China. 0143-4004/$ e see front matter Ó 2013 Published by IFPA and Elsevier Ltd. http://dx.doi.org/10.1016/j.placenta.2013.10.020
endothelium, particularly in the context of normal pregnancy and preeclampsia. 2. Distribution and function of endoglin 2.1. Vascular endothelium We first identified endoglin with the monoclonal antibody 44G4, raised against cell surface proteins prepared from a childhood leukemia cell line [1]. It soon became apparent that this glycoprotein was found in blood vessels and it was later classified as the endothelial marker CD105 [2]. Activation of endothelial cells leads to increased expression of endoglin, in response to an angiogenic or inflammatory stimulus [3,4]. Endoglin is in fact widely used as a marker of tumor angiogenesis [3,5]. Although there is higher endoglin on dividing cells and none on senescent cells, it is not required for proliferation as endoglin-null endothelial cells divide faster than normal cells [6]. Endoglin is likely important for maintaining vascular homeostasis and quiescence of the vascular endothelium. In fact, mutations in the ENG gene lead to Hereditary Hemorrhagic Telangiectasia (HHT) type 1 [7], a vascular
S94
A.L. Gregory et al. / Placenta 35, Supplement A, Trophoblast Research, Vol. 28 (2014) S93eS99
dysplasia characterized by arteriovenous malformations and severe and frequent bleeding episodes. The disease is due to haploinsufficiency and reduced functional levels of endoglin [8]. We demonstrated that the underlying defect is at least in part due to endothelial dysfunction caused by disrupted association of endoglin with endothelial NO synthase (eNOS), leading to generation of superoxide and impaired vasomotor function [9]. 2.2. Mesenchymal and hematopoietic cells Endoglin is expressed in mesenchymal stem cells [10] and is required for hemangioblast and early hematopoietic development [11,12]. Endoglin is absent from most immune cell subsets but is induced by activation of monocytes into macrophages [13]. The role of endoglin may be to conserve stem cell potential, versus differentiation of precursor cells, via its modulation of TGF-b and BMP effects on lineage determination. TGF-b acts as a potent immunosuppressive agent by inhibiting cell growth, inducing apoptosis and contributing to the generation of T regulatory cells. Therefore the role of endoglin may be to tightly regulate cell differentiation and activation by TGF-b superfamily members. A striking example of transient expression of endoglin is during the endocardial-mesenchymal transformation that leads to heart septation and valve formation [14]. Endoglin-null embryos die at mid-gestation due to defects in vessel and heart development [15], indicating that endoglin is essential for both angiogenesis and heart development. Endoglin is also present on fibroblasts at the edge of a wound and on perivascular stromal cells involved in vascular remodeling [16]. Additionally, endoglin is found on vascular smooth muscle cells, particularly after arterial injury, and could mediate the effects of TGF-b on their migration and tissue repair [17]. 2.3. Placenta The human placenta is an abundant source of endoglin. We first purified endoglin from human term placenta, using monoclonal antibody 44G4; this led to cloning of the ENG gene from a placental library [18]. Immunostaining revealed expression of endoglin in the syncytiotrophoblast of term placenta [19], where it is actually present throughout pregnancy. Endoglin is also transiently expressed in extravillous trophoblast (EVT) in the proximal columns of anchoring villi [20]. We showed previously that monoclonal antibody 44G4, or antisense oligonucleotides to endoglin, stimulated budding and outgrowth of EVT from villous explants at 5e8 weeks of gestation [21], a process that is negatively regulated by TGF-b1/b3 [22]. Fibronectin synthesis and changes in integrin expression also occur during EVT differentiation. Furthermore, the distribution of endoglin on EVT parallels that of the a5b1 integrin [20], which was recently shown to interact with endoglin [23]. The interaction is mediated by the RGD peptide of human endoglin or the equivalent TDD motif found in pig and mouse endoglin [18,23]. We propose that the concurrent induction of endoglin and a5b1 integrin on budding EVT and their RGD e mediated interaction may contribute to the regulation of EVT differentiation. The subsequent decrease in these proteins in the distal columns may facilitate remodeling of spiral arteries, as reduced endoglin expression is associated with vasorelaxation [9]. 3. Structure of membrane and soluble endoglin 3.1. Membrane endoglin Human endoglin is an integral membrane glycoprotein, composed of an N-terminal signal peptide, an orphan domain, a zona pellucida (ZP) domain, a juxtamembrane region, a transmembrane
Fig. 1. Schematic diagram of membrane, recombinant and soluble endoglin. Membrane endoglin is an integral dimeric glycoprotein composed of two identical monomers of 90 kDa linked by disulphide bonds. The polypeptide is composed of an orphan domain with no known structural features (27e357; residues 1e26 representing the signal peptide) and contains five N-linked glycosylation sites (C). The orphan domain is followed by a zona pellucida domain (ZP; 358e586), and by juxtamembrane, transmembrane and short cytoplasmic domains. The recombinant soluble monomeric protein used in most studies corresponds to the complete extracellular domain of endoglin and has a molecular mass of 80 kDa (1e586). The soluble endoglin (sEng) monomer released from placenta into the maternal serum has a molecular mass of 65 kDa and its C terminal residue is likely shortly after Arg 406, within the ZP domain.
region and a short cytoplasmic domain [18,24] as shown in Fig.1. The protein is highly glycosylated with five potential N-linked sites within the orphan domain and O-glycan sites mostly in the ZP domain. Enzymatic removal of all the sugars leads to an observed reduction of about 25 kDa in molecular mass [25]. Membrane endoglin exists as a disulphide-linked homodimer of 180 kDa that contains 17 cysteine residues [18]. However the position of the interchain disulphide bond(s) has yet to be established. It was suggested that Cys582 is responsible for inter-chain bonding although this residue is not conserved in all species. Studies of a CD31-endoglin hybrid molecule indicated that cysteine residues within the fragment Cys330eCys412 were implicated [26]. Our own studies showed that Eng1e357 is expressed as a dimer, leaving Cys330 or Cys350 as the most likely residues involved in dimer formation (unpublished data).
3.2. Soluble endoglin Several studies reported the increased expression of a circulating form of endoglin in serum, plasma or other fluids from cancer patients, as summarized in a review by Bernabeu et al. [5]. This form is referred to as soluble endoglin (sEng), although its true solubility has not been established and it could be associated with other proteins [27]. A 80 kDa sEng was recently shown to be cleaved from the surface of human umbilical vein endothelial cells by MMP14 (or MT1-MMP); the cleavage site was shown to be at position 586, implying that the whole extracellular domain (Eng1e 586) was released [28]. However it remains to be proven that a similar mechanism occurs in tumors. For several years, we (and others) [24,29e31] have used the Eng1e586 construct (Fig. 1) to generate recombinant sEng; this form is soluble to some degree but has not proven satisfactory for structural studies, implying that it might not represent a functional physiological sEng, with a stable structure. A soluble form of endoglin is also present in increased amounts in the serum of women with preeclampsia (PE) [29]. It begins to rise
A.L. Gregory et al. / Placenta 35, Supplement A, Trophoblast Research, Vol. 28 (2014) S93eS99
8e10 weeks before PE onset and correlates well with disease severity, even better that the other soluble receptor associated with PE, namely sFlt-1 (or soluble VEGFR-1) [32]. Unlike sFlt-1, for which alternate splice variants have been described, none has so far been reported for sEng. MMP14 is expressed in preeclamptic placentae [33] and could cleave endoglin from the syncytiotrophoblast, yielding a protein of approximately 80 kDa. When we purified sEng form the serum of preeclamptic women, it was resolved by SDSPAGE as a 65 kDa band, as opposed to the 80 kDa mass of recombinant Eng1e586 [29] (Fig. 1). In our view, it is unlikely that the circulating soluble form of endoglin (sEng) observed in preeclampsia is produced by MMP14 cleavage at position 586. Mass spectrometry analysis of purified endoglin from sera of PE patients, revealed that sEng extended at least to Arginine 406; however we could not identify the C-terminal residue of sEng [29]. Taking carbohydrate side-chains into consideration, we would predict a 40e 45 kDa polypeptide for sEng produced by the placenta, indicating that the C-terminal is shortly after residue 406 (Fig. 1). The mechanism of generation of sEng from placenta, its exact sequence and structure, and functional testing of the purified physiological form must be accomplished to understand its role in placenta and PE.
4. Ligand specificity of membrane and soluble endoglin 4.1. Ligands interacting with membrane endoglin Membrane endoglin bound to radioactive TGF-b was first identified by chemical cross-linking in human umbilical vein endothelial cells, in complex with the types I and II serine kinase receptors [34]. Membrane endoglin interacts with TGF-b1 and TGFb3 but not TGF-b2 isoforms [34]. In most cell types, responses to TGF-b are mediated via the type II receptor (TbRII) and the type I receptor ALK5 [35] (Fig. 2). This pathway is known to inhibit cellular proliferation, mediate fibronectin and collagen production and induce plasminogen activator inhibitor-1 (PAI-1), a negative regulator of cell migration and angiogenesis. This signaling route is potentiated by endoglin in several cell types including cardiac fibroblasts [36]. In endothelial cells, there is an additional pathway mediating the effects of TGF-b, consisting of TbRII in complex with the type I receptor ALK1, and influenced by endoglin expression
Fig. 2. TGF-b and BMP9 signaling pathways in endothelial cells. In endothelial cells, TGF-b1 and eb3 bind to TbRII which then recruits and phosphorylates ALK5 or ALK1. These two type I receptors signal via Smad 2,3 and Smad 1,5,8 respectively. Smad4 is responsible for transfer of Smad complexes to the nucleus where they regulate gene transcription. BMP9 binds to ALK1 and BMPRII and signals via Smad 1,5,8 pathway. Endoglin can modulate all of these pathways in a cell and context-dependent manner.
S95
(Fig. 2). In fact, mutations in ALK1 lead to HHT type 2, implying that endoglin and ALK1 share pathways affected in these diseases [37]. Although the ALK5/Smad 2,3 pathway antagonizes the ALK1/Smad 1,5,8 pathway, ALK5 is essential for recruitment and activation of ALK1 [38]. Membrane endoglin was shown by co-immunoprecipitation studies to interact with several members of the TGF-b superfamily in association with the ligand binding receptor: TbRII for TGF-b1 and TGF-b3; ActRII for activin-A; ActRIIB for BMP7; and the type I receptors ALK3 and ALK6 in the case of BMP2 [39]. More recently, it was shown that overexpression of endoglin promotes responses to BMP9 and BMP10, which are the physiological ligands for ALK1, in a complex with either BMPRII or ActRIIA [40] (Fig. 2). Membrane endoglin can therefore modulate responses to several ligands in a cell- and context-dependent manner. Fig. 3 illustrates the effects of endoglin on Smad 1,5,8 phosphorylation induced by BMP2 and BMP9 in mouse embryonic endothelial cells with either no endoglin (Eng/ cells) or with normal levels (Engþ/þ cells). The response to BMP2 was higher in Eng/ than Engþ/þ cells, indicating that membrane endoglin inhibited BMP2 effects. In the receptor complex, endoglin associates with ALK3, which in turn can interact with ActRII, ActRIIB or BMPRII. ALK6 was absent from the Eng/ cells (as demonstrated by RT-PCR) and could not have mediated the potentiating effect of BMP2. Membrane endoglin potentiated the response to BMP9, as mentioned previously [40]. Here we show higher Smad 1,5,8 phosphorylation in Engþ/þ than Eng/ cells, particularly at low and physiologically relevant BMP9 concentrations. This response requires ALK1 as Alk1-deficient mouse endothelial cells do not bind BMP9 at low ligand concentrations, as reported [41], and as confirmed in our studies. The contribution of endoglin to TGF-b1 induced and TbRII/ALK1/ALK5 mediated Smad phosphorylation has been measured in these cells. We observed potentiation [6] while another group reported inhibition [42] of Smad 1,5,8 phosphorylation by endoglin using the same cells but at different passages. Both groups noted no effect of endoglin on TGFb1 induced Smad 2,3 phosphorylation (TbRII and ALK5 mediated). Thus endoglin can either stimulate or inhibit various responses to members of the TGF-b superfamily, in a ligand-, receptor- and celldependent manner. 4.2. Ligands binding to soluble endoglin It has been repeatedly stated in the literature that high levels of circulating sEng released from placenta during preeclampsia, may alter vascular homeostasis by scavenging TGF-b1 and preventing its binding to membrane receptors on endothelial cells. In an attempt to verify this point, we tested the ability of recombinant sEng (1e586) to block TGF-b1 induced Smad 1,5,8 phosphorylation and compared its activity to that of soluble TGF-b type II receptor (sTbRII). Fig. 4A shows that while sTbRII (extracellular domain only) blocked completely the ability of TGF-b1 to stimulate Smad 1,5,8 phosphorylation in endothelial cells, sEng (1e586) had no effect at the same concentration. Therefore, sTbRII can easily scavenge TGF-b1 while sEng (1e586) cannot. We then tested if sEng (1e586) could bind BMP9 on its own and compared its activity to that of a soluble form of ALK1 (sALK1). Fig. 4B shows that sEng (1e586) and sALK1 are equally able to block BMP9 stimulation of Smad 1,5,8 phosphorylation and can both scavenge circulating BMP9. The direct binding of ligands to immobilized purified recombinant soluble proteins was then assessed using surface plasmon resonance (SPR). Dissociation constants (KD) were estimated from the binding data, as a measure of the affinity of the interactions. Fig. 5A reveals that sEng (1e586) binds TGF-b1 with very low affinity (KD > 1.2 mM) while sTbRII binds with high affinity
A.L. Gregory et al. / Placenta 35, Supplement A, Trophoblast Research, Vol. 28 (2014) S93eS99
(B)
400
200
50
8
Phospho-Smad1/Total-Smad1
Phospho-Smad1/Total-Smad1
100
60kD 0
Smad1 BMP9 (pM)
25
60kD
400
PSmad1
8
6
4
2
0
Eng-/-
200
400
200
50
100
0
25
200
400
50
60kD 100
Smad1 BMP2 (pM)
0
60kD
25
PSmad1
Eng+/+
50
Eng-/-
100
Eng+/+
0
(A)
25
S96
0
100
200
300
400
[BMP2] (pM)
6
4
2
0
0
100
200
300
400
[BMP9] (pM)
Fig. 3. Endoglin differentially regulates BMP2 and BMP9 signaling in endothelial cells. Wild type mouse embryonic endothelial cells (Engþ/þ; C) and endoglin-null cells (Eng/; o) were serum-starved for 30 min and stimulated with increasing concentrations of either BMP2 (A) or BMP9 (B). Cell lysates were resolved by SDS-PAGE and analyzed by Western blot with antibodies specific for phosphorylated-Smad1 and total Smad1. The amount of P-Smad1 was quantified relative to total Smad1 levels. Data represent the mean SEM of 3 experiments. Endoglin inhibited BMP2 but promoted BMP9 effects on Smad1 signaling.
(KD ¼ 100 pM) (Fig. 5B). This implies that sEng (1e586) cannot bind this ligand under physiological conditions and could not scavenge TGF-b1, if present in circulation as a truly soluble protein; however it could do so if circulating in a complex with sTbRII. The binding of BMP9 to sEng and sALK1 was also analyzed by SPR. Fig. 5C and D
show that sEng (1e586) and sALK1 bind BMP9 with similar affinities (sEng, KD ¼ 5 nM; sALK1, KD ¼ 1.8 nM). Our data confirm similar findings by other groups [30,31]. Thus both sEng (1e586) and sALK1 could scavenge circulating BMP9 and block its effects, as suggested by Fig. 4B.
Fig. 4. Blocking of Smad1 phosphorylation by soluble ligand-binding receptors. Wild type mouse embryonic endothelial cells were starved for 2 h, and incubated for 1 h with increasing concentrations of (A) soluble human TGF-b receptor II (sTBRII) or recombinant sEng (1e586) or (B) soluble human ALK1-Fc (sALK1) or recombinant sEng (1e586), and stimulated for 30 min with (A) 100 pM TGF-b1 or (B) 25 pM BMP9. Cell lysates were resolved by SDS-PAGE and analyzed by Western blot with antibodies specific for phosphorylated-Smad1 and total Smad1. Phospho Smad1 was quantified relative to total Smad1. Data are the mean SEM of 3 experiments except for sTBRII where only 2 experiments were performed. sEng blocked BMP9 but not TGF-b1 binding to the cells while sTBRII and sALK1 had the expected ligand specificity.
A.L. Gregory et al. / Placenta 35, Supplement A, Trophoblast Research, Vol. 28 (2014) S93eS99
S97
Fig. 5. Kinetic analysis of the binding of TGF-b1 and BMP9 to their receptors. Purified receptors [recombinant sEng (1e586), sTBRII and sALK1] were amine-coupled to CM5 chips and different concentrations of TGF-b1 (25e800 nM; A and B) or BMP9 (1e500 nM; C and D) were injected over the captured receptors. The raw data (colored lines) are overlaid with a global fit 1:1 model (black lines) obtained by the BIA evaluation software. The resonance signal is plotted against ligand concentration. The estimated affinity constants (KD) are indicated for each ligandereceptor interaction.
As mentioned in Section 3.2, the 65 kDa estimated molecular mass for monomeric sEng does not correspond to the 80 kDa recombinant protein (1e586) used in the blocking and SPR experiments. Therefore, the functional properties of the true physiological form, including its ligand specificity will need to be re-assessed once the sequence of the protein is known and it can be expressed as a recombinant protein. Our data suggest that BMP9 is the more likely ligand to be scavenged by purified sEng associated with placenta and preeclampsia. However, one cannot rule out the possibility that sEng is present in circulation in a complex with sTbRII and even sTbRI that could scavenge TGF-b1 and -b3 and blocked their effects on the vascular system. 5. Regulation of endothelial function by membrane and soluble endoglin 5.1. eNOS and membrane endoglin We demonstrated previously that membrane endoglin, through its short cytoplasmic domain, associates with endothelial NO synthase (eNOS) and Hsp90 in caveolae [9]. ALK1 also associates with eNOS, implying a multimeric complex at the cell surface [43]. The importance of endoglin and ALK1 in eNOS activation is seen in endothelial cells from patients with HHT and in animal models of the disease, which show uncoupling of eNOS, reduced NO production and increased eNOS-derived superoxide production leading to impaired myogenic response [9,44]. In vivo treatment of Engþ/ or Alk1þ/ mice with the anti-oxidant Tempol was able to prevent the onset of age-dependent changes in the pulmonary vasculature, indicating the contribution of eNOS-derived reactive oxygen species to endothelial dysfunction [43,44]. Thus membrane
endoglin plays an important role in the regulation of eNOS activation, initiated by the association of its short cytoplasmic tail with intracellular eNOS in caveolae. 5.2. Endothelin-1 and soluble endoglin Human serum sEng cannot directly scavenge TGF-b, unless a complex of sEng-sTbRII is present. sEng alone will therefore not prevent the phosphorylation and activation of eNOS by TGF-b and the subsequent release of NO, as proposed [29]. We discussed above that sEng can scavenge circulating BMP9. However, in our studies, BMP9 did not induce eNOS phosphorylation and activation and therefore sEng could not act by preventing this process. Other mechanisms to explain how sEng may affect vasomotor function in preeclampsia, are based on the observations that BMP9 can induce production of the potent vasoconstrictor, endothelin-1 (ET-1) [45,46]. BMP9 stimulated ET-1 release from pulmonary endothelial cells under physiological concentrations and in a Smad1 (but not Smad4) dependent manner, involving the ALK1/ BMPRII receptors and the non-canonical p38 MAPK pathway [46]. It was suggested that BMP9 stimulated ET-1 production may be important for vascular stability. However, knocking down ALK1 or BMPRII not only attenuated the BMP9 effects, but also caused a significant increase in prepro ET-1 mRNA transcription and release of the mature peptide [46]. These findings imply that BMP9 and its receptors are important in regulating ET-1 production and consequently hypertension. sEng (1e586) can effectively scavenge BMP9 and would reduce ET-1 release and may lead to reduced vascular stability. Furthermore, ET-1 is elevated in preeclampsia and could increase maternal vasoconstriction; in that context, sEng would have a beneficial effect. However, scavenging BMP9 with sEng during preeclampsia could potentially reduce levels of functional
S98
A.L. Gregory et al. / Placenta 35, Supplement A, Trophoblast Research, Vol. 28 (2014) S93eS99
receptors and lead to a further rise in ET-1 production, and a sustained rise in arterial pressure. 6. Conclusions The enigmatic distribution of endoglin in terms of its persistent expression on vascular endothelium and the syncytiotrophoblast versus its transient expression in mesenchymal cells and the extravillous trophoblast likely will reveal cues as to the functions exerted in these different cell types during development. The sequence and 3D-structure of sEng produced by the placenta and elevated in preeclampsia remains to be established and its ligand specificity determined before its mode of action can be explained. The regulation of vasomotor function by membrane endoglin via its association with eNOS and by soluble endoglin via alteration of ET1 production add to the enigma concerning the essential physiological role of endoglin, as evidenced by altered levels leading to several vascular disorders. Conflict of interest statement The authors confirm that there is no conflict of interest in the above paper. Acknowledgments We wish to thank Zhijie Li and Dr James Rini of the Department of Biochemistry, University of Toronto, for their help and guidance with the surface Plasmon resonance experiments. References [1] Quackenbush EJ, Letarte M. Identification of several cell surface proteins of non-T, non-B acute lymphoblastic leukemia by using monoclonal antibodies. J Immunol 1985;134(2):1276e85. [2] Letarte M, Greaves A, Vera S. In: Schlossman SF, Boumsell L, Gilks W, Harlan J, Kishimoto T, Morimoto C, et al., editors. CD105 (endoglin) cluster report. Oxford: Oxford University Press; 1995. p. 1756e9. [3] Burrows FJ, Derbyshire EJ, Tazzari PL, Amlot P, Gazdar AF, King SW, et al. Upregulation of endoglin on vascular endothelial cells in human solid tumors: implications for diagnosis and therapy. Clin Cancer Res 1995;1(12):1623e34. [4] Torsney E, Charlton R, Parums D, Collis M, Arthur HM. Inducible expression of human endoglin during inflammation and wound healing in vivo. Inflamm Res 2002;51:464e70. [5] Bernabeu C, Lopez-Novoa JM, Quintanilla M. The emerging role of TGF-beta superfamily coreceptors in cancer. Biochim Biophys Acta 2009;1792(10): 954e73. [6] Pece-Barbara N, Vera S, Kathirkamathamby K, Liebner S, Di Guglielmo GM, Dejana E, et al. Endoglin null endothelial cells proliferate faster and are more responsive to transforming growth factor beta1 with higher affinity receptors and an activated Alk1 pathway. J Biol Chem 2005;280(30):27800e8. [7] McAllister KA, Lennon F, Bowles-Biesecker B, McKinnon WC, Helmbold EA, Markel DS, et al. Genetic heterogeneity in hereditary haemorrhagic telangiectasia: possible correlation with clinical phenotype. J Med Genet 1994;31: 927e32. [8] Pece N, Vera S, Cymerman U, White Jr RI, Wrana JL, Letarte M. Mutant endoglin in hereditary hemorrhagic telangiectasia type 1 is transiently expressed intracellularly and is not a dominant negative. J Clin Invest 1997;100(10):2568e79. [9] Toporsian M, Gros R, Kabir MG, Vera S, Govindaraju K, Eidelman DH, et al. A role for endoglin in coupling eNOS activity and regulating vascular tone revealed in hereditary hemorrhagic telangiectasia. Circ Res 2005;96(6):684e92. [10] Barry FP, Boynton RE, Haynesworth S, Murphy JM, Zaia J. The monoclonal antibody SH-2, raised against human mesenchymal stem cells, recognizes an epitope on endoglin (CD105). Biochem Biophys Res Commun 1999;265:134e9. [11] Perlingeiro RC. Endoglin is required for hemangioblast and early hematopoietic development. Development 2007;134(16):3041e8. [12] Borges L, Iacovino M, Mayerhofer T, Koyano-Nakagawa N, Baik J, Garry DJ, et al. A critical role for endoglin in the emergence of blood during embryonic development. Blood 2012;119(23):5417e28. [13] Lastres P, Bellon T, Cabanas C, Sanchez-Madrid F, Acevedo A, Gougos A, et al. Regulated expression on human macrophages of endoglin, an Arg-Gly-Aspcontaining surface antigen. Eur J Immunol 1992;22(2):393e7.
[14] Qu R, Silver MM, Letarte M. Distribution of endoglin in early human development reveals high levels on endocardial cushion tissue mesenchyme during valve formation. Cell Tissue Res 1998;292(2):333e43. [15] Bourdeau A, Dumont DJ, Letarte M. A murine model of hereditary hemorrhagic telangiectasia. J Clin Invest 1999;104(10):1343e51. [16] Matsubara S, Bourdeau A, terBrugge KG, Wallace C, Letarte M. Analysis of endoglin expression in normal brain tissue and in cerebral arteriovenous malformations. Stroke 2000;31(11):2653e60. [17] Ma X, Labinaz M, Goldstein J, Miller H, Keon WJ, Letarte M, et al. Endoglin is overexpressed after arterial injury and is required for transforming growth factor-beta-induced inhibition of smooth muscle cell migration. Arterioscler Thromb Vasc Biol 2000;20(12):2546e52. [18] Gougos A, Letarte M. Primary structure of endoglin, an RGD-containing glycoprotein of human endothelial cells. J Biol Chem 1990;265(15): 8361e4. [19] Gougos A, St. Jacques S, Greaves A, O’Connell PJ, d’Apice AJF, Buhring H-J, et al. Identification of distinct epitopes of endoglin, an RGD containing-glycoprotein of endothelial cells, leukemic cells and syncytiotrophoblasts. Int Immunol 1992;4(1):83e92. [20] St-Jacques S, Forte M, Lye SJ, Letarte M. Localization of endoglin, a transforming growth factor-beta binding protein, and of CD44 and integrins in placenta during the first trimester of pregnancy. Biol Reprod 1994;51(3): 405e13. [21] Caniggia I, Taylor CV, Ritchie JWK, Lye SJ, Letarte M. Endoglin regulates trophoblast differentiation along the invasive pathway in human placental villous explants. Endocrinology 1997;138:4977e88. [22] Caniggia I, Grisaru-Gravnosky S, Kuliszewsky M, Post M, Lye SJ. Inhibition of TGF-beta 3 restores the invasive capability of extravillous trophoblasts in preeclamptic pregnancies. J Clin Invest 1999;103(12):1641e50. [23] Rossi E, Sanz-Rodriguez F, Eleno N, Duwell A, Blanco FJ, Langa C, et al. Endothelial endoglin is involved in inflammation: role in leukocyte adhesion and transmigration. Blood 2013;121(2):403e15. [24] Llorca O, Trujillo A, Blanco FJ, Bernabeu C. Structural model of human endoglin, a transmembrane receptor responsible for hereditary hemorrhagic telangiectasia. J Mol Biol 2007;365(3):694e705. [25] Gougos A, Letarte M. Biochemical characterization of the 44G4 antigen from the HOON pre-B leukemic cell line. J Immunol 1988;141(6):1934e 40. [26] Raab U, Velasco B, Lastres P, Letamendia A, Cales C, Langa C, et al. Expression of normal and truncated forms of human endoglin. Biochem J 1999;339(Pt 3): 579e88. [27] Li CG, Wilson PB, Bernabeu C, Raab U, Wang JM, Kumar S. Immunodetection and characterization of soluble CD105-TGF-ß complexes. J Immunol Methods 1998;218:85e93. [28] Hawinkels LJ, Kuiper P, Wiercinska E, Verspaget HW, Liu Z, Pardali E, et al. Matrix metalloproteinase-14 (MT1-MMP)-mediated endoglin shedding inhibits tumor angiogenesis. Cancer Res 2010;70(10):4141e50. [29] Venkatesha S, Toporsian M, Lam C, Hanai J, Mammoto T, Kim YM, et al. Soluble endoglin contributes to the pathogenesis of preeclampsia. Nat Med 2006;12(6):642e9. [30] Castonguay R, Werner ED, Matthews RG, Presman E, Mulivor AW, Solban N, et al. Soluble endoglin specifically binds bone morphogenetic proteins 9 and 10 via its orphan domain, inhibits blood vessel formation, and suppresses tumor growth. J Biol Chem 2011;286(34):30034e46. [31] Alt A, Miguel-Romero L, Donderis J, Aristorena M, Blanco FJ, Round A, et al. Structural and functional insights into endoglin ligand recognition and binding. PLoS One 2012;7(2):e29948. [32] Levine RJ, Lam C, Qian C, Yu KF, Maynard SE, Sachs BP, et al. Soluble endoglin and other circulating antiangiogenic factors in preeclampsia. N Engl J Med 2006;355(10):992e1005. [33] Kaitu’u-Lino TJ, Palmer KR, Whitehead CL, Williams E, Lappas M, Tong S. MMP-14 is expressed in preeclamptic placentas and mediates release of soluble endoglin. Am J Pathol 2012;180(3):888e94. [34] Cheifetz S, Bellon T, Cales C, Vera S, Bernabeu C, Massague J, et al. Endoglin is a component of the transforming growth factor-beta receptor system in human endothelial cells. J Biol Chem 1992;267(27):19027e30. [35] Chen YG, Hata A, Lo RS, Wotton D, Shi Y, Pavletich N, et al. Determinants of specificity in TGF-beta signal transduction. Genes Dev 1998;12(14): 2144e52. [36] Kapur NK, Wilson S, Yunis AA, Qiao X, Mackey E, Paruchuri V, et al. Reduced endoglin activity limits cardiac fibrosis and improves survival in heart failure. Circulation 2012;125(22):2728e38. [37] Berg JN, Gallione CJ, Stenzel TT, Johnson DW, Allen WP, Schwartz CE, et al. The activin receptor-like kinase 1 gene: genomic structure and mutations in hereditary hemorrhagic telangiectasia type 2. Am J Hum Genet 1997;61(1): 60e7. [38] Lebrin F, Deckers M, Bertolino P, Ten Dijke P. TGF-beta receptor function in the endothelium. Cardiovasc Res 2005;65(3):599e608. [39] Barbara NP, Wrana JL, Letarte M. Endoglin is an accessory protein that interacts with the signaling receptor complex of multiple members of the transforming growth factor-beta superfamily. J Biol Chem 1999;274(2): 584e94. [40] David L, Mallet C, Mazerbourg S, Feige JJ, Bailly S. Identification of BMP9 and BMP10 as functional activators of the orphan activin receptor-like kinase 1 (ALK1) in endothelial cells. Blood 2007;109(5):1953e61.
A.L. Gregory et al. / Placenta 35, Supplement A, Trophoblast Research, Vol. 28 (2014) S93eS99 [41] Choi EJ, Kim YH, Choe SW, Tak YG, Garrido-Martin EM, Chang M, et al. Enhanced responses to angiogenic cues underlie the pathogenesis of hereditary hemorrhagic telangiectasia 2. PLoS One 2013;8(5):e63138. [42] Lee NY, Ray B, How T, Blobe GC. Endoglin promotes transforming growth factor beta-mediated Smad 1/5/8 signaling and inhibits endothelial cell migration through its association with GIPC. J Biol Chem 2008;283(47):32527e33. [43] Toporsian M, Jerkic M, Zhou YQ, Kabir MG, Yu LX, McIntyre BA, et al. Spontaneous adult-onset pulmonary arterial hypertension attributable to increased endothelial oxidative stress in a murine model of hereditary hemorrhagic telangiectasia. Arterioscler Thromb Vasc Biol 2010;30(3):509e17.
S99
[44] Jerkic M, Kabir MG, Davies A, Yu LX, McIntyre BA, Husain NW, et al. Pulmonary hypertension in adult Alk1 heterozygous mice due to oxidative stress. Cardiovasc Res 2011;92(3):375e84. [45] Star GP, Giovinazzo M, Langleben D. Bone morphogenic protein-9 stimulates endothelin-1 release from human pulmonary microvascular endothelial cells: a potential mechanism for elevated ET-1 levels in pulmonary arterial hypertension. Microvasc Res 2010;80(3):349e54. [46] Park JE, Shao D, Upton PD, Desouza P, Adcock IM, Davies RJ, et al. BMP-9 induced endothelial cell tubule formation and inhibition of migration involves Smad1 driven endothelin-1 production. PLoS One 2012;7(1):e30075.
