Human Molecular Genetics, 2001, Vol. 10, No. 13 1347-1357
© 2001 Oxford University Press
Analysis of several endoglin mutants reveals no endogenous mature or secreted protein capable of interfering with normal endoglin function
Blood and Cancer Research Program, The Hospital for Sick Children and Department of Immunology, University of Toronto, 555 University Avenue, Toronto M5G 1X8, Canada and 1Imperial College School of Medicine, National Heart and Lung Institute, Hammersmith Hospital, London, UK
Received February 2, 2001; Revised and Accepted April 26, 2001.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
Hereditary hemorrhagic telangiectasia type 1 (HHT1) is associated with mutations in the ENDOGLIN gene which normally codes for a polypeptide of 653 amino acids expressed at the cell surface as a dimeric glycoprotein. To maximize the detection of potential mutant proteins, we analyzed by pulse-chase experiments the expression of large truncation mutants in endothelial cells from newborns with HHT1. A mutant truncated at residue 490 (
490) and the 517 mutant, previously suggested to act as dominant negative, were undetectable. Proteins 471 and 571 were barely detectable as transient monomers of 62 and 72 kDa. A de novo 13 bp deletion in exon 11 encoded a monomeric protein of 70 kDa (557), present at low levels in activated monocytes. Six novel missense mutants and S411 were expressed only as the 80 kDa intracellular precursor of surface endoglin, suggesting impaired processing. All nine novel mutations reported failed to be expressed other than intracellularly. Several constructs of endoglin were expressed in COS-1 cells; only the full-length protein was processed to the cell surface. Recombinant 586, corresponding to the complete extracellular domain, was secreted as monomeric and dimeric glycosylated species. Our studies show that all HHT1 mutants analyzed, although expressed to various degrees in COS-1 cells, are either undetectable, present at low levels as transient intracellular forms, or expressed as partially glycosylated precursors in endogenous cells. These mutants do not form heterodimers with normal endoglin and do not interfere with its normal trafficking to the cell surface, further supporting the haploinsufficiency model. |
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
Endoglin is a homodimeric membrane glycoprotein with a molecular mass of 180 kDa which is expressed predominantly on endothelial cells (1). The polypeptide has an extracellular domain of 561 amino acids (from N-terminal E26 to G586, M1 being the initiation codon), a single transmembrane region (L587–W611) and a cytoplasmic tail of 47 amino acids terminating in Ala658 (2,3). The cytoplasmic tail is rich in serine and threonine residues, three of which are constitutively phosphorylated (2,4,5). There are four potential N-linked glycosylation sites and a potential region of O-linked glycosylation (6). Cloning and sequencing of betaglycan, also known as the TGF-ß type III receptor and capable of binding all three isoforms of TGF-ß, revealed that its transmembrane and cytoplasmic domains were 70% similar to those of endoglin (7,8). This led to the identification of endoglin as a protein which could be cross-linked to TGF-ß1 and -ß3 isoforms (9), and which co-precipitated with the TGF-ß serine/threonine kinase receptors type II and type I (5,10,11). More recently, we have demonstrated that endoglin is capable of binding not only TGF-ß1/ß3 but also activin, BMP-2 and BMP-7, via its association with their respective ligand binding receptors (12). Endoglin is thus an accessory protein of the receptor complex for several members of the TGF-ß superfamily.
While endoglin is a glycoprotein primarily expressed on endothelial cells, betaglycan is a membrane proteoglycan expressed in a variety of cells but generally at low levels on endothelial cells (13). Betaglycan can also be enzymatically cleaved from the cell surface and has been detected in serum, extracellular matrices and cell culture fluids (7,14). Although low levels of endoglin have been observed in serum, using an ELISA assay (15–17), no secreted form, or enzymatically cleaved product, has been observed in the media of endothelial cells in culture. Serum endoglin is likely contained in membrane particles shed from the vascular endothelium.
Mutations in the Endoglin gene result in hereditary hemorrhagic telangiectasia type 1 (HHT1) (18), whereas mutations in the ALK-1 gene, a serine/threonine kinase type I receptor of the TGF-ß superfamily, lead to HHT2 (19). HHT is an autosomal dominant disorder affecting 1/8000 individuals and characterized by heterogeneity of its clinical manifestations. Although a higher incidence of pulmonary arteriovenous malformations (PAVMs) is present in HHT1 than in HHT2, clinical heterogeneity is observed with both HHT1 and HHT2 genotypes. In some families, the disease is associated with nose bleeds and telangiectases whereas in others, several members also manifest cerebral, pulmonary and/or hepatic arteriovenous malformations (AVMs). To date, 41 mutations associated with HHT1 have been reported. These are found in the first 12 exons of the gene and include substitutions, deletions and insertions (18,20–30). A detailed phenotypic analysis of eight families with characterized HHT1 mutations showed that the severity of manifestations was independent on the type and position of the mutation (23).
Analysis of endoglin expression in multiple families has provided evidence for haploinsufficiency as the predominant model for HHT1. Null alleles have been observed (23,24,26) and most mutant proteins are only expressed as transient intracellular species, if expressed at all (22,26,28,31). We also reported that four missense endoglin mutations were expressed as intracellular precursors and were never processed into mature surface glycoproteins (25). A recent paper, looking exclusively at expression in COS-1 cells, has suggested the presence of dominant negative mutants of endoglin that could be detected at the cell surface as heterodimers or, in one case, secreted (29). In the current study, we performed pulse-chase studies with several large truncation mutants in endothelial cells from HHT1 newborns to optimize detection of potential mutants in endogenous cells. We also analyzed several novel missense mutations. None of the mutant proteins could be detected at the cell surface or secreted. The expression of different endoglin fragments in COS-1 cells was also studied, particularly to identify potential soluble forms of the protein. The only secreted recombinant fragment generated corresponded to the complete extracellular domain of endoglin. The data presented in this paper and the extensive protein expression studies performed in about 200 patients do not provide any evidence for dominant negative mutations associated with HHT1.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
Pulse-chase analysis of normal endoglin in human umbilical vein endothelial cells (HUVEC)
The rate of processing and turnover of normal endoglin in endothelial cells was analyzed prior to undertaking the analysis of truncation mutations associated with HHT1. HUVEC were metabolically labeled with a 20 min pulse of [35S]methionine, followed by chase periods ranging from 0 to 24 h and immunoprecipitated with monoclonal antibody (mAb) P4A4, specific for endoglin. Figure 1A shows a representative experiment. Endoglin is synthesized as a precursor (P) of 80 kDa, first detectable at time zero and maximum at 30–60 min of chase. This precursor is processed into the mature glycoprotein (E), first seen after 30 min and maximally expressed at the cell surface after 3 h of chase. Analysis under reducing and non-reducing conditions revealed that partial dimerization of the precursor can be seen as early as time zero. After 3 h of chase, no residual monomer is seen and the mature glycoprotein is completely dimerized (Fig. 1A). The relative amount of mature endoglin was estimated at each time point relative to the maximum level observed in each of several experiments (Fig. 1B). The expression of mature endoglin peaked after 2–3 h and began to decline thereafter. The half-life of endoglin at the cell surface was estimated at 17 h, indicating that it is a relatively stable glycoprotein.
|
Pulse-chase analysis of endoglin truncation mutants
Metabolic labeling of HUVEC and peripheral blood-activated monocytes, using a standard steady state labeling period of 3.5 h, has revealed reduced levels of endoglin in more than 30 newborns and 180 adults from HHT1 families (22,25,26,28; Table 1 and data not shown). However, in the vast majority of cases, no mutant protein could be visualized. To optimize the detection of potential transient protein species, some of the larger truncation mutations were analyzed by pulse-chase studies in HUVEC. These cells can readily be expanded in vitro in primary cultures and express much higher levels of endoglin than activated monocytes.
|
All mutations described in this paper are shown diagrammatically in Figure 2. The H129 and H304 HUVEC both expressed 45% of normal endoglin levels, under steady state conditions (Table 1). They were shown to carry mutations in exon 11 which lead to interruption of the normal protein sequence at residues 490 (
490) and 517 (517) (26). The 517 mutant, also known as GC, was described previously as a dominant negative (29). Neither 490 nor 517 proteins could be detected by pulse-chase studies, even as transient precursors at early time points (Fig. 3A). When the data is expressed as percentage of the maximum level of surface endoglin (E) present in the control, it can be seen that the rate of processing of the normal allele is similar in both HHT1 HUVEC, although at lower levels than the control (Fig. 3B). Processing of the normal precursor (P) to the mature form (E) of endoglin, as measured by the ratio E/E+P, was identical in these mutants and the control HUVEC, indicating no interference by the mutated allele with the expression of the normal allele (data not shown).
|
|
Two other HUVEC, H317 with a novel nonsense mutation in exon 12, which leads to a stop (
571), and H628 with a previously reported deletion in exon 10 (26), which causes termination of the normal sequence at residue 471 (471), are illustrated in Figure 3C. Traces of 571 were detected at 0.5–2 h by immunoprecipitation with either mAb P3D1 or mAb P4A4. A polyclonal antibody to endoglin also immunoprecipitated this mutant, but less effectively than the monoclonal antibodies (data not shown). The mutant is seen as a monomer, estimated at 72 kDa. The detection of the mutant is limited because of its low level of labeling and its rapid turnover relative to the normal allele and control sample (Fig. 3D). The 571 mutant was also visible in the standard 3.5 h labeling assay, albeit at low levels. The 471 mutant appears as a 62 kDa monomer, immunoprecipitated with both mAb P3D1 and mAb P4A4 but only at 0.5 and1 h chase periods. This mutant was not detectable in the standard labeling assay and represents a very minor species that turns over very quickly (Fig. 3D). So despite their low level of expression as intracellular proteins, 571 and 471 did not interfere with processing of the corresponding normal allele to a mature surface glycoprotein (Fig. 3D). The presence of these mutants in the culture media was also tested; none were detected, indicating that these truncation mutants could not be secreted at a significant rate in endogenous cells.
We report a de novo mutation in a child (H89) with clinical manifestations of HHT including several small cerebral AVMs (CAVMs) (Materials and Methods). The 13 bp deletion (1672–1684) at the end of exon 11 leads to termination of the normal sequence at codon 557, followed by 10 additional amino acids (Table 1). This mutant protein was visible by standard labeling of the activated monocytes derived from patient H89 as a weak monomeric band of 70 kDa, which was absent from samples derived from the parents and sister who did not carry the mutated allele (Fig. 4A). The 557 mutant protein was detectable with both mAb P3D1 and P4A4 and represented 
10% of the total endoglin present in the sample of the affected individual, which itself was 43% of the control. The 557 mutant was not detectable in the culture media, indicating that it was not secreted by activated monocytes.
|
A novel nonsense mutation in exon 6 (G782A) terminating at codon 260 is also reported. This mutant was not visible by standard metabolic labeling either in HUVEC of newborn H455 or in activated monocytes from the affected parent. The normal allele of endoglin was expressed as the mature form (E) at levels corresponding to 39% and 55% (Table 1).
Missense mutants are visible as intracellular precursor proteins
We report seven missense mutations (including six new ones and one independently reported) as well as one new indel mutation which accumulated as the intracellular endoglin precursor of 80 kDa. Figure 4B illustrates a sample from patient H134, with a C479A mutation in exon 4 previously described in a Japanese family, and leading to A160D substitution. The steady state metabolic labeling of peripheral blood-activated monocytes revealed a precursor band (P) of higher intensity than the mature form (E). The percentage of mature endoglin in this sample was reduced to 44% of the normal level (Table 1) whereas the percentage of precursor was 104% of the normal level, suggesting that the missense mutant protein is not fully glycosylated and cannot reach the cell surface. For the vast majority of mutations analyzed, the ratio (mutant/control) of precursor endoglin (P) is generally similar to that of the mature forms (E), i.e. on average 50%. A higher ratio of precursor was also seen for patient H269 with a novel mutation in exon 8 (G1088A) leading to C363Y substitution (Fig. 4B). The percentage of precursor (P) in this case was 122% of the normal level whereas that of mature endoglin (E) was reduced to 46% of the normal level (Table 1). A third case is illustrated in Figure 4C, where a G1220A mutation in exon 9a, causing an S407Q substitution, was associated with the accumulation of the intracellular precursor, readily detectable with either mAb P4A4 or mAb P3D1. The levels of precursor (P) in patients H703, H705 and H706 with this mutation were 104, 113 and 146%, respectively, estimated relative to the normal precursor band intensity. Levels of mature endoglin (E) were reduced in these samples to 41, 45 and 59% of the normal level (Table 1).
Figure 4D illustrates two distinct mutations identified in the same experiment, both of which led to increased precursor level in activated monocytes of the patients. A T95G mutation in exon 2 giving rise to a L32R substitution, gave a precursor ratio of 105% in patient H17 relative to normal and a reduced surface endoglin (58%). The deletion of AGC (1231–1233) in exon 9a gave rise to a mutant lacking serine 411 (S411). This mutant was also expressed as an intracellular protein, detectable with both mAb in sample H175, and migrating in the precursor position (88% of the normal P value) with only the normal allele being expressed at the surface (53% of control).
Two additional missense mutants were also expressed as intracellular precursors in HUVEC. In newborn H729, a G155A mutation in exon 2 led to a G52D substitution and to accumulation of precursor (138% of control) and reduction of mature form to 37% (Fig. 4E and Table 1). Another new T374A mutation in exon 2, yielding a V125D substitution, was also seen in newborn H507 as an intracellular 80 kDa precursor (160% of the normal level) and a mature surface protein reduced to 40% (Fig. 4E and Table 1). These data further support our previous observations with four missense mutants (G52V, C53R, W149C and L221P) which were only expressed as intracellular precursors in HHT1 patient’s cells or when engineered in COS-1 cells (25).
Expression of endoglin fragments in COS-1 cells
Since higher levels of expression can be achieved by transient transfection of COS-1 cells, and since evidence of dominant negative activity had been observed in that system, we engineered several mutants and variants of endoglin (Fig. 5A). Newly synthesized proteins were metabolically labeled with [35S]methionine and immunoprecipitated with mAb P3D1 as it recognizes an epitope located within the first 200 amino acids of endoglin and should bind to all recombinant proteins used in this study. Lysates from cells transfected with vector alone and immunoprecipitated with mAb P3D1, or lysates of cells transfected with full-length endoglin and immunoprecipitated with a control non-immune IgG, showed no reactivity (Fig. 5B and C). Under reducing conditions, normal endoglin was resolved as the partially glycosylated precursor (P) of 80 kDa and the fully processed glycoprotein (E) of 90 kDa, as seen in endogenous cells. Under non-reducing conditions, endoglin was fully dimerized at 180 kDa.
|
Fragment
230 represents a translation product derived from an abnormally spliced mRNA and isolated with a polyclonal antibody to human endoglin from a 
gt11 cDNA expression library from HUVEC (2). The corresponding cDNA retained introns 5 and 6, and a stop codon was reached after 54 nucleotides of intron 5, giving rise to a fragment containing the first 230 amino acids of endoglin followed by 18 amino acids encoded by intron 5 (32). When transfected into COS-1 cells, 230 was detected in total cell lysates as a band of 36 kDa under both reducing and non-reducing conditions (Fig. 5B and C), suggesting that this short form of endoglin cannot dimerize.
Fragment 462/632 corresponds to an unusual in-frame variant, detected by RT–PCR in different human cell types, and apparently spliced using unconventional sites at positions 1386 in exon 10 and 1896 in exon 14. Although the cDNA is readily amplified, no corresponding protein has been observed in human samples. Such a protein would lack 124 amino acids of the extracellular and transmembrane domains but would retain a portion of the cytoplasmic tail in-frame with the rest of the polypeptide, as shown in Figure 5A. The 462/632 construct was expressed in total lysate from COS-1 transfectants mostly as a monomer of 68 kDa (Fig. 5B and C), indicating that it could potentially exist in vivo. Dimers and potential oligomers of higher molecular mass were observed under non-reducing conditions.
517, the putative dominant negative mutant, was expressed in COS-1 cells as a readily detectable monomer of 60 kDa with visible traces of homodimers (Fig. 5B and C). We previously demonstrated by metabolic and surface labeling that this mutant, when co-expressed with normal endoglin in COS-1 cells, did not form heterodimers and was not detected at the cell surface (22).
558, which lacks 28 amino acids of the extracellular domain, and is of a size similar to the mutant 557 described in endogenous cells (Fig. 4), was expressed in the total cell extract of COS-1 cells. It was resolved as a band of 68 kDa, which showed partial dimerization (Fig. 5B and C). 586, which represents the entire ectodomain of endoglin, was expressed in COS-1 cells as a single band of 72 kDa (reduced); partial dimerization was observed under non-reducing conditions (Fig. 5C).
586 is the only engineered construct secreted in the culture media of COS-1 cells
The presence of secreted recombinant endoglin fragments in the culture media was tested by immunoprecipitation from radiolabeled transfected COS-1 cells. When full-length endoglin was expressed, no proteolytic fragment was observed in the culture media (Fig. 5D, lane 1). However, 586 was secreted (lane 2) while 558, 517, 462/632 and 230 were not (lanes 3–6).
The rate of secretion of 586, its stability and extent of dimerization were examined by pulse-chase studies. At time zero, a major band of 72 kDa was detected in the total cell extract, with 20% in a dimeric form of 130 kDa (Fig. 6, top panels). At 1 h, minor bands representing higher glycosylation intermediates of 586 were seen migrating above the major monomer and dimer. The level of intracellular dimerization reached a plateau of 50% after 4 h and remained stable until the end of the chase period. 586 was readily detectable in the culture media after 2 h where it accumulated as a fully glycosylated species of 78 kDa, reaching 90% secretion by 8 h (Fig. 6, bottom panels). Soluble 586 was stable for the 8 h chase period in this experiment and for more than 24 h in additional experiments. The secreted mononers and dimers corresponded to the glycosylated forms of higher molecular mass, seen as minor components in the cell extract. Thus 586 can be produced as a glycosylated soluble protein, capable of 50% dimerization.
|
586 does not form heterodimers with normal endoglinPotential heterodimer formation between
586 and normal endoglin monomer was tested by co-transfection of COS-1 cells and surface biotinylation. Figure 7A shows that endoglin (lanes 2 and 6) was found at the cell surface whereas 586 (lanes 3 and 7), which lacks a transmembrane region, was not. When both proteins were co-expressed, only homodimers of normal endoglin were detected at the cell surface; no traces of heterodimers were observed (lanes 4 and 8). Metabolic labeling of the co-transfected COS-1 cells was performed in the same experiment and revealed that endoglin and 586 were independently expressed and not inhibiting each other’s expression (Fig. 7B). Dimerization of each protein remained unchanged and no trace of heterodimer was seen (Fig. 7B, lane 4). Furthermore, 586 was still secreted in the medium of COS-1 cells co-transfected with normal endoglin (data not shown). Thus soluble 586 cannot form heterodimers with normal endoglin.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
In this study, we report nine novel endoglin mutations associated with HHT1 and analyze by pulse-chase studies some large truncation mutants. We demonstrate that mutations leading to premature termination are either not expressed or are expressed at very low levels in endogenous HUVEC and activated monocytes. Missense mutations are retained intracellularly as partially glycosylated precursors. Thus these mutants are likely misfolded and consequently unable to form heterodimers with normal endoglin and reach the cell surface. Several constructs corresponding to HHT mutations and to natural or engineered variants of endoglin were expressed to various degrees in COS-1 cells, but only as intracellular species.
586, corresponding to the complete extracellular domain of endoglin, was the only recombinant protein secreted in the culture medium; it was glycosylated and able to dimerize to some extent, suggesting that it should retain biological activity. The presented data are in accord with our previous observations and publications, suggesting that HHT1 results from haploinsufficiency rather than from dominant negative effects. It should also be noted that each family analyzed in the current study has at least one member with a CAVM or PAVM, representing a severe clinical presentation of HHT. A dominant negative model predicts that mutants capable of interfering with normal endoglin function should be associated with a more serious clinical outcome. Patients with proven null mutations or lack of detectable mutant proteins would then be expected to have a milder disease, which is not the case in our current series, nor in earlier reports (23,26).
To summarize our data supporting the haploinsufficiency model, we have observed reduced levels of surface endoglin in endothelial cells of newborns and/or peripheral blood-activated monocytes of patients from 98 families with HHT. Endoglin mutations have been confirmed in 54 of these families (22,25,26,28; and M. Letarte and U. Cymerman, unpublished data). In the vast majority of these cases, no mutant protein was observed by standard steady state metabolic labeling. However, that method can detect some intracellular mutants in endogenous cells as demonstrated for an in-frame exon 3 skip (22), the duplication of exons 3–8 (28) and the newly described truncated proteins 557 and 571. In the current study, we also demonstrate that seven missense mutants and S411 can be detected by standard metabolic labeling in endogenous cells as the 80 kDa precursor of endoglin (Fig. 4). This confirms our previous observations with four other missense mutant proteins (G52V, C53R, W149C and L221P) (25). Thus, the missense mutants studied to date are expressed intracellularly and as such, cannot interfere with the normal function of a cell surface glycoprotein such as endoglin.
In the current study, we also used pulse-chase studies to monitor the expression of potential transient species in HUVEC of HHT1 newborns, focusing on truncations in the second part of the extracellular domain. We first estimated that processing of normal endoglin to the cell surface (dimer formation and glycosylation) required 1–3 h. The half-life of endoglin was estimated at 17 h in HUVEC, representing a relatively stable glycoprotein. The processing of mutants 471 and 571 was very different. Although they could be detected at early time points with both mAb P3D1 and P4A4, they were no longer detectable after 2 h or less. Other truncated mutants (417 and 517) were not visible even at early time points, suggesting very unstable species. Processing of the normal allele was unchanged in these HUVEC, indicating that the mutant allele did not interfere with trafficking of normal endoglin to the cell surface.
The 517 mutant, reported to act as dominant negative (29), was not detectable in HUVEC, even when using pulse-chase studies, suggesting that it is most unstable in endogenous cells. However, it was readily detectable in COS-1 cell extracts but was not secreted (Fig. 5). In our previous COS-1 cell studies, 517 could not be labeled by surface biotinylation, when co-expressed with normal endoglin, indicating that it could not form a heterodimer (22). Furthermore, expression of normal endoglin was unaltered in these cells, indicating that the 517 mutant could not interfere with normal trafficking (29). Thus our previous co-expression and cell surface labeling data in COS-1 cells (22) as well as the current quantitative pulse-chase studies in HUVEC (Fig. 3) and further studies in COS-1 cells (Fig. 5) do not support the proposition that 517 can act as a dominant negative mutant. A nonsense mutant 350 expressed in COS-1 cells was also suggested to potentially interfere with normal trafficking when co-expressed with standard endoglin (29). We were unable to demonstrate expression of this nonsense mutant (22) as well as nonsense 276 (26) and 260 (Table 1) in endogenous cells, further confirming that these mutants cannot act as dominant negative. We also previously expressed the 276 in COS-1 cells and demonstrated that it could not form heterodimers or interfere with normal processing (22). The larger extracellular fragment of endoglin tested, 586, was also incapable of heterodimer formation (Fig. 7). Thus, although HHT truncation mutants are detectable when overexpressed in COS-1 cells, they are generally not found at significant levels in endogenous cells, and do not form heterodimers that could reach the cell surface and interfere with normal endoglin function.
Heterodimerization would also imply that the residues involved in inter-chain bonds are still present in the mutants. When endoglin mutants and variants were expressed in COS-1 cells, we observed that 230 could not dimerize whereas 462/632, 558 and 586 could. Previous studies had shown that 276 was also unable to form dimers while 517 could (22), suggesting that inter-chain disulfide bonds were located between residues 276 and 462. This is substantiated by results obtained with PECAM-1/endoglin chimeric constructs which imply residues between C330 and C412 in these inter-chain bonds (33). Furthermore, the alignment of the primary sequences of endoglin and betaglycan reveals that three cysteine residues (C330, C350 and C382) are present in dimeric endoglin and absent in monomeric betaglycan. Such residues are the likely candidates for the inter-chain disulfide bond(s). The suggestion that C350 could still form heterodimers with normal endoglin (29) would require that endoglin carry a single inter-chain bond at position C330. Although this is possible, the above considerations on instability of truncated mutants do not favor the formation of this heterodimer. Site-specific directed mutagenesis of the three putative cysteine residues in a stable recombinant form of endoglin is required to determine which one(s) are implicated in formation of the normal homodimer.
The identification of soluble truncated forms of endoglin is of great interest as they could indeed interfere with normal surface endoglin or be active as recombinant proteins. Endoglin is homologous to betaglycan, which exists as a monomeric protein which can be cleaved from the cell surface and secreted as a soluble form. However, no secreted product could be found in the supernatant of COS-1 cells transfected with normal endoglin (Fig. 5). We tested several constructs in this study as well as 276 in a previous report (22). None were secreted except for 586, which was found exclusively in the medium after 8 h, in a partially dimeric and glycosylated form, stable for at least 24 h (Fig. 6). Two additional forms of recombinant endoglin, 431 and 437, were secreted by COS-1 cells as dimers (33). It was suggested that such constructs might represent natural protease cleavage products which would utilize an arginine, lysine, lysine (RKK) site at position 437, similar to that of betaglycan. However, we have never observed a soluble form of normal endoglin in the culture media of HUVEC analyzed by pulse-chase studies or in COS-1 cells transfected with full-length endoglin (Fig. 5), and our data do not support the existence of a normally secreted form of endoglin. In the current study, we have also observed that the large truncated mutants of endoglin, which were detectable by pulse-chase studies, were not present in the culture media. Since the vast majority of mutants are very transient species not even detectable by metabolic labeling, soluble forms of endoglin must be rare, if present at all in HHT patients.
In summary, we have further confirmed by analysis of several mutants of endoglin that haploinsufficiency is to date the only demonstrated model of HHT1. Although we cannot rule out the potential existence of dominant negative mutants of endoglin, there is currently no evidence for their association with HHT1.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
Patient samples and mutation analysis
Informed consent was obtained from all individuals participating in the study. Positive clinical diagnosis was provided by physicians, according to the presence of at least three established criteria (nose bleeds, telangiectases, organ involvement such as pulmonary, cerebral or hepatic AVMs, and a suitable family history) (34). Due to our clinical referral base, many probands in the HHT families have PAVMs or CAVMs. Families and their members are given numbers and patients are referred to with the prefix H (for HHT). Data is presented on 13 HHT1 families in the current study, nine of which have previously unreported mutations (Table 1). One of these is also a rare example of a de novo mutation in HHT: a child (H89) diagnosed with HHT and with several small CAVMs since birth was shown by quantitative multiplex-polymerase chain reaction (QM-PCR) analysis and sequencing to carry a 13 bp deletion in exon 11 of endoglin (Table 1). Neither mother nor father expressed this mutation or showed clinical manifestations of disease. DNA fingerprinting with eight different markers confirmed that individual H89 and his parents shared the expected alleles and established that this mutation arose spontaneously in the child.
Genomic DNA was isolated from HUVEC, placenta and blood lymphocytes. All mutations were analyzed by QM-PCR and/or sequencing using the MicroGene Blaster sequencer (Visible Genetics, Toronto, Canada) and as described previously (26).
Cell culture and transfections
All reagents were from Gibco BRL Canadian Life Technologies (Burlington, Canada) unless otherwise specified. HUVEC were prepared from newborns from HHT1 families and control babies and used at equivalent passage number and cellular density as previously reported (1,22). Activated monocytes, which express endoglin optimally after 16–24 h in culture, were prepared from peripheral blood samples of patients with HHT1 and control unaffected individuals by published procedures (25).
COS-1 cells were maintained and transiently transfected with expression constructs by the DEAE-dextran-chloroquine method, as described previously (35,36).
Preparation of expression constructs
The EcoRI fragment of human full-length endoglin cDNA subcloned into the pcEXV-1 vector (pcEXV-L Endo) (3) or the pCMV5 vector was used. Fragment 230 is the translation product of a cDNA clone (2) (Fig. 5A). The construct was engineered by generation of a PCR fragment for its 3' end and ligation to the 5' end of normal endoglin via a unique XhoI site (32). Forward primer, 5'-GGC CGC ACG CTC GAG TGG CGG CCG CGT ACT CCA-3'; reverse primer, 5'-TCT TGG ATC CTA TCA GGG GGG TGG TCT CTC GGG GTG GGG ACT-3'. Restriction sites are underlined.
Fragment 462/632 corresponds to an endoglin mRNA variant detected by RT–PCR in human cells and lacking from bp 1386 in exon 10 to bp 1896 in exon 14 (Fig. 5A, numbering from the ATG codon) (18). The construct was prepared by PCR amplification of the 3' end of this variant using forward primer, 5'-AGG CGG TGG TCA ATA TCC TGT CGA-3', bp 1271–1294; and reverse primer, 5'-TCT TGG ATC CGT GGA GGG ACC CCA AGG TGT TCC AA-3', corresponding to bp 2175–2151. A BamHI restriction site (underlined) was included in the reverse primer. The 391 bp PCR product was ligated with the 5' end of normal endoglin cDNA using a unique SacI site at bp 1295 and cloned into Bluescript (Stratagene, La Jolla, CA). The HindIII/BamHI fragment was subcloned into pCMV5 vector.
The construct for fragment 558 was prepared by introducing a valine codon at position 558 followed by a stop codon (TAG) making use of a unique endoglin Tth111I restriction site (Fig. 5A). Full length endoglin cDNA was subcloned into Bluescript using HindIII linkers and synthetic, complementary oligomers containing an in-frame stop codon and a BamH1 overhang (5'-TCTAGACG, 3'-AGATCTGCCTAG) were blunt-ligated to the Tth111I linearized plasmid. After digestion with HindIII, the HindIII–BamH1 1.7 kb insert was ligated to the pcDNAI/Neo vector (Invitrogen, Carlsbad, CA) and subcloned into pcEXV-1 vector by blunt ligation into the EcoRI site.
Fragment 586, which represents the complete extracellular domain of endoglin (Fig. 5A) was engineered by long PCR (Dr S. Pichuantes, Chiron Corporation, Emeryville, CA) using as template the original endoglin clone 18A (2). A HindIII site was introduced in the forward primer while a BamHI site and two stop codons were introduced in the reverse primer. Forward primer, 5'-CAC GCC AAG CTT ATG GAC GCG GGC ACG CTC CCT CTG GCT GTT GCC CTG CTG CTG GCC AGC TGC AGC CTC AGC CCC ACA AGT CTT-3'; reverse primer, 5'-TCT TGG ATC CTA TTA GCC TTT GCT TGT GCA ACC AGA CAG GTC AGG GCT-3'. Restriction sites are underlined and start and stop codons are in bold type. The PCR fragment was cloned into Bluescript and subcloned into the EcoRI sites of pcEXV-1 by blunt ligation. 586 was also subcloned into pCMV5 as a HindIII/BamHI fragment. All constructs were sequenced entirely using the T7 sequencing kit (Pharmacia Biotech, Québec, Canada) except for 462/632 and 558, which were sequenced at the junctions.
Metabolic labeling and pulse-chase studies
Endoglin expression in HUVEC, activated monocytes and transfected COS-1 cells was analyzed by metabolic labeling according to published procedures (22,25,26). Confluent monolayers of HUVEC (1–2 x 106 cells per 100 mm dish), activated monocytes derived from 20 ml of blood, or COS-1 cells, 2 days post-transfection, at an average density of 6 x 105 cells/60 mm dish, were washed and incubated in methionine-free high glucose Dulbecco’s modified Eagle’s medium (DMEM) for 30 min and labeled for 3.5 h (referred to as standard procedure) with 1 ml of met-free DMEM containing 100 µCi/ml of [35S]methionine (Trans-label, ICN Pharmaceuticals, Montréal, Canada). Culture media were collected in experiments where potential secreted forms were analyzed. Cells were solubilized in 1 ml of 0.01 M Tris buffer pH 7.5 plus 0.128 M NaCl, 0.001 M EDTA, 1% Triton X-100 and protease inhibitors. The amount of total labeled protein in each cell extract was determined by trichloroacetic acid precipitation.
Total cell lysates and culture media were immunoprecipitated using mAb P3D1 (4 µg) or mAb P4A4 (1 µg) and protein G–Sepharose CL-4B (Pharmacia Biotech). Both antibodies have been characterized and used extensively; they are specific for human endoglin and recognize epitopes mapping within the first 230 terminal amino acids (P3D1) or between residues 277 and 331 (P4A4) (32). Equivalent amounts of radioactivity were fractionated by SDS–PAGE (4–12% Tris-Glycine, Novex Experimental Technology, San Diego, CA) and analyzed by fluorography or autoradiography with BioMax MS films and LE Transcreen intensifying screen system (Eastman Kodak Canada, Rochester, NY). Radioactivity in each band was quantified by PhosphorImager and ImageQuant software (Molecular Dynamics, Sunnyvale, CA).
For pulse-chase experiments of HUVEC or transfected COS-1 cells, equivalent cell numbers were pulsed for 20 min with 250 µCi/ml [35S]methionine followed by two washes and a further incubation in serum-free DMEM for chase periods ranging from 0 to 24 h. Cells were lysed, immunoprecipitated with mAb P3D1 or P4A4 and analyzed as described above.
Cell surface biotinylation
Transfected COS-1 cells were surface-labeled with biotin, lysed, immunoprecipitated with mAb P3D1 and analyzed as described previously (22).
|
ACKNOWLEDGEMENTS |
|---|
We thank Dr S. Pichuantes for help with generation and sequencing of endoglin constructs and Dr E.A. Wayner for mAbs P3D1 and P4A4. This work was supported by grant HHT-FY99-677 from the March of Dimes and by the Canadian Institutes of Health Research. M.-E.P. was a recipient of a FRSQ-FCAR Santé Scholarship, Québec, N.P.-B. is a post-doctoral fellow of the Canadian Institutes of Health Research and M.L. is a Terry Fox Research Scientist of the National Cancer Institute of Canada. C.S. is supported by the Wellcome Trust.
|
FOOTNOTES |
|---|
+ To whom correspondence should be addressed. Tel: +1 416 813 6258; Fax: +1 416 813 6255; Email: mablab@sickkids.on.caPresent address: Nadia Pece-Barbara, Program in Molecular Biology and Cancer, Samuel Lunenfeld Research Institute, Mount Sinai Hospital, Toronto, M5G 1X5, Canada
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
1 Gougos, A. and Letarte, M. (1988) Identification of a human endothelial cell antigen with monoclonal antibody 44G4 produced against a pre-B leukemic cell line. J. Immunol., 141, 1925–1933.[Abstract]
2 Gougos, A. and Letarte, M. (1990) Primary structure of endoglin, an RGD-containing glycoprotein of human endothelial cells. J. Biol. Chem., 265, 8361–8364.
3 Bellón, T., Corbi, A., Lastres, P., Calés, C., Cebrián, M., Vera, S., Chefeitz, M., Massagué, J., Letarte, M. and Bernabeu, C. (1993) Identification and expression of two forms of the human transforming growth _factor-ß-binding protein endoglin with distinct cytoplasmic regions. Eur. J. Immunol., 23, 2340–2345.[Web of Science][Medline]
4 Lastres, P., Martin-Perez, J., Langa, C. and Bernabeu, C. (1994) Phosphorylation of the human-transforming-growth-factor-ß-binding protein endoglin. Biochem. J., 301, 765–768.
5 Yamashita, H., Ichijo, H., Grimsby, S., Morén, A., ten Dijke, P. and Miyazono, K. (1994) Endoglin forms a heteromeric complex with the signaling receptors for transforming growth factor-ß. J. Biol. Chem., 269, 1995–2001.
6 Gougos, A. and Letarte, M. (1988) Biochemical characterization of the 44G4 antigen from the HOON pre-B leukemic cell line. J. Immunol., 141, 1934–1940.[Abstract]
7 López-Casillas, F., Cheifetz, S., Doody, J., Andres, J.L., Lane, W.S. and Massagué, J. (1991) Structure and expression of the membrane proteoglycan betaglycan, a component of the TGF-ß receptor. Cell, 67, 785–795.[Web of Science][Medline]
8 Wang, X.-F., Lin, H.Y., Ng-Eaton, E., Downward, J., Lodish, H.F. and Weinberg, R.A. (1991) Expression cloning and characterization of the TGF-ß type III receptor. Cell, 67, 797–805.[Web of Science][Medline]
9 Cheifetz, S., Bellón, T., Calés, C., Vera, S., Bernabeu, C., Massagué, J. and Letarte, M. (1992) Endoglin is a component of the transforming growth factor-ß receptor system in human endothelial cells. J. Biol. Chem., 267, 19027–19030.
10 Zhang, H.W., Shaw, A.R.E., Mak, A. and Letarte, M. (1996) Endoglin is a component of the TGF-ß receptor complex of human pre-B leukemic cells. J. Immunol., 156, 565–573.
11 Lastres, P., Letamendia, A., Zhang, H., Rius, C., Almendro, N., Lopez, L.A., Langa, C., Fabra, A., Letarte, M. and Bernabeu, C. (1996) Endoglin modulates cellular responses to TGF-ß1. J. Cell Biol., 133, 1109–1121.
12 Barbara, N., Wrana, J.L. and Letarte, M. (1999) 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., 274, 584–594.
13 Massague, J. (1990) The transforming growth factor-beta family. Annu. Rev. Cell. Biol., 6, 597–641.[Web of Science]
14 Andres, J.L., Stanley, K., Cheifetz, S. and Massagué, J. (1989) Membrane-anchored and soluble forms of betaglycan, a polymorphic proteoglycan that binds transforming growth factor-ß. J. Cell Biol., 109, 3137–3145.
15 Wang, J.M., Wilson, P.B., Kumar, S., Pye, D. and Hunter, R.D. (1994) Quantitation of endothelial cell specific protein E-9 employing a single monoclonal antibody in an indirect sandwich ELISA. J. Immunol. Methods, 171, 55–64.[Web of Science][Medline]
16 Blann, A.D., Wang, J.M., Wilson, P.B. and Kumar, S. (1996) Serum levels of the TGF-beta receptor are increased in atherosclerosis. Atherosclerosis, 120, 221–226.[Web of Science][Medline]
17 Li, C.G., Wilson, P.B., Bernabeu, C., Raab, U., Wang, J.M. and Kumar, S. (1998) Immunodetection and characterization of soluble CD105-TGF-ß complexes. J. Immunol. Methods, 218, 85–93.[Web of Science][Medline]
18 McAllister, K.A., Grogg, K.M., Johnson, D.W., Gallione, C.J., Baldwin, M.A., Jackson, C.E., Helmbold, E.A., Markel, D.S., McKinnon, W.C., Murrell, J. et al. (1994) Endoglin, a TGF-ß binding protein of endothelial cells is the gene for hereditary haemorrhagic telangiectasia type 1. Nat. Genet., 8, 345–351.[Web of Science][Medline]
19 Berg, J.N., Gallione, C.J., Stenzel, T.T., Johnson, D.W., Allen, W.P., Schwartz, C.E., Jackson, C.E., Porteous, M.E.M. and Marchuk, D.A. (1997) The activin receptor-like kinase 1 gene: genetic structure and mutations in hereditary hemorrhagic telangiectasia type 2. Am. J. Hum. Genet., 6, 60–67.
20 McAllister, K.A., Baldwin, M.A., Thukkani, A.K., Gallione, C.J., Berg, J.N., Porteus, M.E., Guttmacher, A.E. and Marchuk, D.A. (1995) Six novel mutations in the endoglin gene in hereditary hemorrhagic telangiectasia tupe 1 suggest a dominant-negative effect of receptor function. Hum. Mol. Genet., 4, 1983–1985.
21 Yamaguchi, H., Azuma, H., Shigekiyo, T. and Inoue, H. (1997) A novel missense mutation in the endoglin gene in hereditary hemorrhagic telangiectasia. Thromb. Haemost., 77, 243–247.[Web of Science][Medline]
22 Pece, N., Vera, S., Cymerman, U., White, R.I.J., Wrana, J.L. and Letarte, M. (1997) Mutant endoglin in Hereditary Hemorrhagic Telangectasia type I is transiently expressed intracellularly and is not a dominant negative. J. Clin. Invest., 100, 2568–2579.[Web of Science][Medline]
23 Shovlin, C.L., Hughes, J.M.B., Scott, J., Seidman, C.E. and Seidman, J.G. (1997) Characterization of endoglin and identification of novel mutations in hereditary hemorrhagic telangiectasia. Am. J. Hum. Genet., 61, 68–79.[Web of Science][Medline]
24 Gallione, C.J., Klaus, D.J., Yeh, E.Y., Stenzel, T.T., Xue, Y., Anthony, K.B., McAllister, K.A., Baldwin, M.A., Berg, J.N., Lux, A. et al. (1998) Mutation and expression analysis of the endoglin gene in hereditary hemorrhagic telangiectasia reveals null alleles. Hum. Mutat., 11, 286–294.[Web of Science][Medline]
25 Pece-Barbara, N., Cymerman, U., Vera, S., Marchuk, D.A. and Letarte, M. (1999) Expression analysis of four endoglin missense mutations suggests that haploinsufficiency is the predominant mechanism for hereditary hemorrhagic telangiectasia type 1. Hum. Mol. Genet., 8, 2171–2181.
26 Cymerman, U., Vera, S., Pece-Barbara, N., Bourdeau, A., White, R.I.,Jr, Dunn, J. and Letarte, M. (2000) Identification of hereditary hemorrhagic telangiectasia type 1 in newborns by protein expression and mutation analysis of endoglin. Pediatr. Res., 47, 24–35.[Web of Science][Medline]
27 Gallione, C.J., Scheessele, E.A., Reinhardt, D., Duits, A.J., Berg, J.N., Westermann, C.J.J. and Marchuk, D.A. (2000) Two common endoglin mutations in families with hereditary hemorrhagic telangiectasia in the Netherlands Antilles: evidence for a founder effect. Hum. Genet., 107, 40–44.[Web of Science][Medline]
28 Bourdeau, A., Cymerman, U., Paquet, M.E., Meschino, W., McKinnon, W.C., Guttmacher, A.E., Becker, L. and Letarte, M. (2000) Endoglin expression is reduced in normal vessels but still detectable in arteriovenous malformations of patients with hereditary hemorrhagic telangiectasia type 1. Am. J. Pathol., 156, 911–923.
29 Lux, A., Gallione, C.J. and Marchuk, D.A. (2000) Expression analysis of endoglin missense and truncation mutations: insights into protein structure and disease mechanisms. Hum. Mol. Genet., 9, 745–755.
30 Bourdeau, A., Faughnan, M.E. and Letarte, M. (2000) Endoglin-deficient mice, a unique model to study Hereditary Hemorrhagic Telangiectasia. Trends Cardiovasc. Med., 10, 279–285.[Web of Science][Medline]
31 Shovlin, C. and Letarte, M. (1999) Hereditary hemorrhagic telangiectasia and pulmonary arteriovenous malformations: Issues in clinical management and review of pathogenic mechanisms. Thorax, 54, 714–729.
32 Pichuantes, S., Vera, S., Bourdeau, A., Pece, N., Kumar, S., Wayner, E.A. and Letarte, M. (1997) Mapping epitopes to distinct regions of the extracellular domain of endoglin using bacterially expressed recombinant fragments. Tissue Antigens, 50, 265–276.[Web of Science][Medline]
33 Raab, U., Velasco, B., Lastres, P., Letamendia, A., Cales, C., Langa, C., Tapia, E., Lopez-Bote, J.P., Paez, E. and Bernabeu, C. (1999) Expression of normal and truncated forms of human endoglin. Biochem. J., 339, 579–588.
34 Shovlin, C.L., Guttmacher, A.E., Buscarini, E., Faughnan, M.E., Hyland, R.H., Westermann, C.J., Kjeldsen, A.D. and Plauchu, H. (2000) Diagnostic criteria for hereditary hemorrhagic telangiectasia (Rendu-Osler-Weber syndrome). Am. J. Med. Genet., 91, 66–67.[Web of Science][Medline]
35 Attisano, L., Wrana, J.L., Montalvo, E. and Massague, J. (1996) Activation of signalling by the activin receptor complex. Mol. Cell. Biol., 16, 1066–1073.
36 Hoodless, P.A., Haerry, T., Abdollah, S., Stapleton, M., O’Connor, M.B., Attisano, L. and Wrana, J.L. (1996) MADR1, a MAD-related protein that functions in BMP2 signaling pathways. Cell, 85, 489–500.[Web of Science][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  J. Wipff, A. Kahan, E. Hachulla, J. Sibilia, J. Cabane, O. Meyer, L. Mouthon, L. Guillevin, C. Junien, C. Boileau, et al. Association between an endoglin gene polymorphism and systemic sclerosis-related pulmonary arterial hypertension Rheumatology, April 1, 2007; 46(4): 622 - 625. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. W. van Laake, S. van den Driesche, S. Post, A. Feijen, M. A. Jansen, M. H. Driessens, J. J. Mager, R. J. Snijder, C. J. J. Westermann, P. A. Doevendans, et al. Endoglin Has a Crucial Role in Blood Cell-Mediated Vascular Repair Circulation, November 21, 2006; 114(21): 2288 - 2297. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. I. Koleva, B. A. Conley, D. Romero, K. S. Riley, J. A. Marto, A. Lux, and C. P. H. Vary Endoglin Structure and Function: DETERMINANTS OF ENDOGLIN PHOSPHORYLATION BY TRANSFORMING GROWTH FACTOR-beta RECEPTORS J. Biol. Chem., September 1, 2006; 281(35): 25110 - 25123. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N L Prigoda, S Savas, S A Abdalla, B Piovesan, D Rushlow, K Vandezande, E Zhang, H Ozcelik, B L Gallie, and M Letarte Hereditary haemorrhagic telangiectasia: mutation detection, test sensitivity and novel mutations J. Med. Genet., September 1, 2006; 43(9): 722 - 728. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Jerkic, J. V. Rivas-Elena, J. F. Santibanez, M. Prieto, A. Rodriguez-Barbero, F. Perez-Barriocanal, M. Pericacho, M. Arevalo, C. P.H. Vary, M. Letarte, et al. Endoglin Regulates Cyclooxygenase-2 Expression and Activity Circ. Res., August 4, 2006; 99(3): 248 - 256. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Sabba, G. Pasculli, P. Suppressa, F. D'Ovidio, G. M. Lenato, F. Resta, G. Assennato, and G. Guanti Life expectancy in patients with hereditary haemorrhagic telangiectasia QJM, May 1, 2006; 99(5): 327 - 334. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Fernandez-L, F. Sanz-Rodriguez, F. J. Blanco, C. Bernabeu, and L. M. Botella Hereditary Hemorrhagic Telangiectasia, a Vascular Dysplasia Affecting the TGF-{beta} Signaling Pathway. Clin. Med. Res., March 1, 2006; 4(1): 66 - 78. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S A Abdalla and M Letarte Hereditary haemorrhagic telangiectasia: current views on genetics and mechanisms of disease J. Med. Genet., February 1, 2006; 43(2): 97 - 110. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Letarte, M.-L. McDonald, C. Li, K. Kathirkamathamby, S. Vera, N. Pece-Barbara, and S. Kumar Reduced endothelial secretion and plasma levels of transforming growth factor-{beta}1 in patients with hereditary hemorrhagic telangiectasia type 1 Cardiovasc Res, October 1, 2005; 68(1): 155 - 164. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. Lebrin, M. Deckers, P. Bertolino, and P. ten Dijke TGF-{beta} receptor function in the endothelium Cardiovasc Res, February 15, 2005; 65(3): 599 - 608. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. E. Harrison, R. Berger, S. G. Haworth, R. Tulloh, C. J. Mache, N. W. Morrell, M. A. Aldred, and R. C. Trembath Transforming Growth Factor-{beta} Receptor Mutations and Pulmonary Arterial Hypertension in Childhood Circulation, February 1, 2005; 111(4): 435 - 441. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A Karabegovic, M Shinawi, U Cymerman, and M Letarte No live individual homozygous for a novel endoglin mutation was found in a consanguineous Arab family with hereditary haemorrhagic telangiectasia J. Med. Genet., November 1, 2004; 41(11): e119 - e119. [Full Text] [PDF]
|
|
|
|
 |
|
 J.-C. Tille and M.S. Pepper Hereditary Vascular Anomalies: New Insights Into Their Pathogenesis Arterioscler Thromb Vasc Biol, September 1, 2004; 24(9): 1578 - 1590. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Li, R. Issa, P. Kumar, I. N. Hampson, J. M. Lopez-Novoa, C. Bernabeu, and S. Kumar CD105 prevents apoptosis in hypoxic endothelial cells J. Cell Sci., July 1, 2003; 116(13): 2677 - 2685. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. E. DUFF, C. LI, J. M. GARLAND, and S. KUMAR CD105 is important for angiogenesis: evidence and potential applications FASEB J, June 1, 2003; 17(9): 984 - 992. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. A. Marchuk, S. Srinivasan, T. L. Squire, and J. S. Zawistowski Vascular morphogenesis: tales of two syndromes Hum. Mol. Genet., April 2, 2003; 12(90001): R97 - 112. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. van den Driesche, C. L. Mummery, and C. J.J. Westermann Hereditary hemorrhagic telangiectasia: an update on transforming growth factor {beta} signaling in vasculogenesis and angiogenesis Cardiovasc Res, April 1, 2003; 58(1): 20 - 31. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Satomi, R. J. Mount;, M. Toporsian, A. D. Paterson, M. C. Wallace, R. V. Harrison, and M. Letarte Cerebral Vascular Abnormalities in a Murine Model of Hereditary Hemorrhagic Telangiectasia Stroke, March 1, 2003; 34(3): 783 - 789. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. M. Botella, T. Sanchez-Elsner, F. Sanz-Rodriguez, S. Kojima, J. Shimada, M. Guerrero-Esteo, M. P. Cooreman, V. Ratziu, C. Langa, C. P. H. Vary, et al. Transcriptional activation of endoglin and transforming growth factor-beta signaling components by cooperative interaction between Sp1 and KLF6: their potential role in the response to vascular injury Blood, December 1, 2002; 100(12): 4001 - 4010. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Sanchez-Elsner, L. M. Botella, B. Velasco, C. Langa, and C. Bernabeu Endoglin Expression Is Regulated by Transcriptional Cooperation between the Hypoxia and Transforming Growth Factor-beta Pathways J. Biol. Chem., November 8, 2002; 277(46): 43799 - 43808. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Guerrero-Esteo, T. Sanchez-Elsner, A. Letamendia, and C. Bernabeu Extracellular and Cytoplasmic Domains of Endoglin Interact with the Transforming Growth Factor-beta Receptors I and II J. Biol. Chem., August 2, 2002; 277(32): 29197 - 29209. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||