| Human Molecular Genetics | Pages |
©1999 Oxford University Press |
Expression analysis of four endoglin missense mutations suggests that haploinsufficiency is the predominant mechanism for hereditary hemorrhagic telangiectasia type 1
Introduction
Results
Endoglin expression is induced on activated monocytes within 5 h in culture
Endoglin expression is reduced in HHT1 patients from four families with missense mutations
HHT1 missense mutants expressed in COS-1 cells are not fully processed and not detectable at the surface
Discussion
Materials And Methods
Cell culture and transfections
Antibodies
Patient samples and mutation analysis
Preparation of expression constructs
Flow cytometry
Metabolic labeling and pulse-chase experiments
Cell surface biotinylation
Acknowledgements
References
Expression analysis of four endoglin missense mutations suggests that haploinsufficiency is the predominant mechanism for hereditary hemorrhagic telangiectasia type 1
Received June 3, 1999; Revised and Accepted August 16, 1999
ENDOGLIN codes for a homodimeric membrane glycoprotein that interacts with receptors for members of the TGF-[beta] superfamily and is the gene mutated in the autosomal dominant vascular disorder hereditary hemorrhagic telangiectasia type 1 (HHT1). We recently demonstrated that functional endoglin was expressed at half levels on human umbilical vein endothelial cells (HUVECs) and peripheral blood activated monocytes from HHT1 patients. Two types of mutant protein were previously analyzed, the product of an exon 3 skip which was expressed as a transient intracellular species and prematurely truncated proteins that were undetectable in patient samples. Here we report the analysis of four proteins resulting from point mutations, with missense codons G52V and C53R in exon 2, W149C in exon 4 and L221P in exon 5. Metabolic labeling of activated monocytes from confirmed, clinically affected patients revealed reduced expression of fully processed normal endoglin in all cases. Pulse-chase analysis with HUVECs from a newborn with the C53R substitution indicated that mutant endoglin remained intracellular as a precursor form and did not impair processing of the normal protein. Biotinylation of cell surface proteins, metabolic labeling and pulse-chase analysis revealed that none of the engineered missense mutants was significantly expressed at the surface of COS-1 transfectants. Thus, these four HHT1 missense mutations lead to transient intracellular species which cannot interfere with normal endoglin function. These data suggest that haploinsufficiency, leading to reduced levels of one of the major surface glyco-proteins of vascular endothelium, is the predominant mechanism underlying the HHT1 phenotype.
INTRODUCTION
Hereditary hemorrhagic telangiectasia (HHT) is inherited as an autosomal dominant trait at a frequency of 1 in 10 000 and exhibits age-related penetrance with variable expressivity (1,2). The most common clinical manifestations involve the development of vascular abnormalities seen as telangiectases on skin and lesions in nasal mucosa that bleed readily. Such lesions are characterized by direct arteriovenous connections that may develop in several organs and can lead to large arteriovenous malformations (AVM) in brain, lung and liver. These may cause serious complications such as stroke, brain abscess or hemorrhage (reviewed in ref. 3).
Genetic linkage studies have revealed that HHT is a heterogeneous disorder. The first locus [hereditary hemorrhagic telangiectasia type 1 (HHT1)] was mapped to chromosome 9q33-34 (4,5), where ENDOGLIN was defined as the affected gene (6). We have recently defined endoglin (CD105) as an accessory membrane glycoprotein that interacts with signalling receptor complexes for several members of the transforming growth factor-[beta] (TGF-[beta]) superfamily (7). It is expressed at high levels on vascular endothelial cells (8); however, other sites of expression include activated monocytes, where it is observed at lower levels (9). A second locus [hereditary hemorrhagic telangiectasia type 2 (HHT2)] mapping to chromosome 12q was shown to be ALK1 (activin receptor-like kinase 1) (10-12). It is also expressed on endothelial cells and is a type I receptor of the TGF-[beta] superfamily, but its true physiological ligand has not been identified (13). TGF-[beta] and other members of its superfamily signal through related serine/threonine kinase receptors, types I and II (14,15). Generally, the type II receptor binds ligand, recruits a type I receptor and activates it by transphosphorylation, thus the type I receptor is the primary transducer of signals to downstream components known as Smads (16-18). These findings that endoglin and ALK1 are mutated in HHT suggest that both proteins are involved in a common pathway controlling vascular development and/or homeostasis. A third variant of HHT was reported in a single large family with serious liver complications (19). Identification of other loci might reveal one or more components of a common signalling pathway involving both endoglin and ALK1.
To date, 29 different mutations have been reported in HHT1 patients (6,20-24) and 18 distinct mutations have been described for ALK1 (12,25,26). HHT1 families appear to have a higher prevalence of pulmonary AVM while HHT2 families in general have a milder phenotype and a later onset of disease (6,23,27,28). Furthermore, clinical heterogeneity within families is high, suggesting that additional factors are contributing to severity of disease.
Although a dominant-negative model of endoglin dysfunction was initially proposed for HHT1, we recently demonstrated that a mutant protein missing 47 amino acids corresponding to an exon 3 skip, but with an intact transmembrane region, was transiently expressed intracellularly both in monocytes from an HHT1 patient and in human umbilical vein endothelial cells (HUVECs) from the child of this patient (22). As the cell surface expressed protein was still able to associate with the TGF-[beta] receptor complex, this indicated that it is the reduction in the level of surface endoglin, rather than interference by a mutant protein, that is involved in the generation of HHT1. The 29 mutations reported to date include various deletions, an insertion and point mutations leading either to premature stop codons or to frameshifts that result in predicted truncated proteins (6,21,23,24). Metabolic labeling of cells expressing some of these mutations showed that the corresponding proteins were not expressed at the cell surface nor were they secreted (22). In addition, three null mutations have been described where mRNA transcripts were undetectable (23,24). These data suggest that endoglin haploinsufficiency is the molecular basis for this genetic disorder.
The observations that every HHT1 family so far has a distinct mutation and that mutations of all types are distributed throughout the gene are consistent with a haploinsufficiency model (29). However, missense mutations often result in dominant-negative mechanisms, constitutive protein activation or new protein functions (30). Here we study four ENDOGLIN missense mutations to determine whether a mechanism other than halploinsufficiency could be associated with HHT1. To quantitate endoglin expression in adult HHT1 patient samples we have further characterized the use of activated monocytes and report that activation of monocytes for at least 21 h in culture is needed to reach steady-state levels of endoglin. We used these conditions for analyzing missense mutations from four HHT1 families. Two missense mutations, G155T and T157C in exon 2, resulting in G52V and C53R codons, and one in exon 4, G447C, creating a W149C codon, were previously characterized as disease associated (24). We report a novel missense mutation in exon 5, a T662C substitution, leading to L221P conversion. These missense mutations result in misfolded proteins that are retained intracellularly and expressed minimally at the surface. Thus, only the normal ENDOGLIN allele gets expressed at the cell surface in these HHT1 patients, correlating with previous data and demonstrating that haploinsufficiency is the predominant mechanism for this genetic disorder.
RESULTS
Endoglin expression is induced on activated monocytes within 5 h in culture
Adherence of monocytes induces their differentiation to the macrophage lineage, which is accompanied by the induction of expression of several genes. It was shown that adherence of peripheral blood monocytes to autologous plasma-coated plastic dishes enhanced their expression of endoglin, which was then stable for up to 7 days (9). By single color flow cytometry we also found that 94% of activated monocytes expressed endoglin (CD105) with a mean relative fluorescence intensity (RFI) of 77 after attachment to plastic and incubation in culture for 21 h (Fig. 1A). In contrast to published results (9), there was no further induction with time in culture; in fact expression was slightly reduced when tested at 45 and 114 h. Endoglin expression correlated with that of the monocyte activation marker CD11b, present on 91% of cells at a mean RFI of 102 after 21 h in culture, but also decreased with time. We also observed that monocyte monolayers were often unstable with longer times in culture and in particular with an autologous plasma coating. For these reasons, we routinely tested endoglin expression on activated monocytes after adherence to uncoated plastic for a period not exceeding 1 day in culture.
Figure 1. Kinetics of appearance of endoglin on monocytes. (A) Normal PBMCs from 60 ml of blood were equally divided for three analyses. Monocytes were separated from lymphocytes by adherence to plastic, then further activated by incubation in culture. Total time in culture is indicated. Single color flow cytometry analysis of endoglin (CD105, solid black) and the monocyte activation marker (CD11b, grey line) was performed as described in Materials and Methods. A gate (<2% positive) was defined at each time point with IgG control (black line) and the percentages of CD105+ and CD11b+ cells within this gate are shown. The mean RFI values with increasing time in culture were 77, 41 and 44 for CD105, and 102, 61 and 69 for CD11b. Note the log scale used when expressing fluorescence intensity. (B) Normal PBMCs from 60 ml of blood, equally divided for three time points, were analyzed by two color flow cytometry. Freshly isolated cells (time 0) and monocytes isolated by adherence to plastic for 1 h or further activated for 20 h were stained with MY4-RD1 (anti-CD14) followed by IgG1 or mAb P4A4 (anti-CD105) then FITC-conjugated F(ab[prime])2 goat anti-mouse IgG. Forward and side scatter analysis and quadrant statistics are shown (left). The upper right quadrant representing monocytes were gated (R1) and analyzed with the FL2 detector (right). A gate was defined with IgG-RD1 as a negative control (<2% +cells). Over 99% cells were CD14+ (solid black) within the R1 gate, with mean RFI values of 1860, 2020 and 1740 at 0, 1 and 21 h, respectively. A second gate was set on all CD14+ cells (R2) and analyzed with the FL1 detector. A gate was established with the IgG control (black line; <2% positive); the percentage of CD105+ cells (solid black) within this gate is shown and the mean RFI values with increasing time were 91, 77 and 200. (C) Normal PBMCs from 20 ml of blood were used for each time point. Monocytes were enriched by adherence to plastic for 1 h and further activated for 4 and 16 h in culture. The total time in culture indicated includes a 4 h period of metabolic labeling with [35S]methionine. Cells were solubilized in Triton X-100 and lysates were immunoprecipitated with mAb P3D1 and fractionated by 4-12% gradient SDS-PAGE under non-reducing (lanes 1 and 2) or reducing (lanes 3 and 4) conditions. Dimeric (180 kDa) and monomeric (90 kDa) fully glycosylated endoglin (E) and partially glycosylated precursor (P, 160 and 80 kDa) are indicated. The E band was quantitated by PhosphorImager and the average pixel values were 13 130 and 18 670 for 5 and 21 h, respectively.
Measuring endoglin levels on peripheral blood monocytes (PBMCs) by selecting the monocyte population using two color flow cytometry would facilitate routine testing and overcome the need for activation. Since endoglin expression has not been reported on PBMCs, we determined the basal level on freshly isolated monocytes and on cells obtained after 1 and 21 h adherence. Forward and side scatter analysis of PBMCs at time 0 shows four populations of cells: based on size and granularity, the upper right quadrant contains 17% of the total cells representing monocytes and contaminating granulocytes, the lower right quadrant (61% of total cells) is mostly lymphocytes and the lower left (17%) contains mostly contaminating platelets and red blood cells (Fig. 1B). Analysis of these populations for the myelomonocytic marker CD14 (with the FL2 detector using rhodamine) revealed that the upper right quadrant was >95% positive, confirming that these cells are of the myeloid lineage. The upper left quadrant (5% of total cells) was also CD14+, while the lower right quadrant contained 22% CD14+ cells. Thus, by far the cleanest monocyte population was in the upper right quadrant and a tight region R1 was defined which contained 99% CD14+ cells with a mean RFI of 1860. A tighter population of CD14+ cells was defined with a second gate (R2) and then analyzed for CD105 expression (with the FL1 detector, using FITC). Only 22% of the CD14+ cells expressed CD105 at time 0, with a mean RFI of 91.
Attachment of PBMCs to plastic for 1 h, followed by washing and release of the adherent layer, yielded an enriched monocyte fraction (49% of total cells; upper right quadrant). All cells in the monocyte gate, defined by R1, were CD14+ with a mean RFI of 2020. There was no increase in CD105 expression on these cells at this time. When PBMCs were plated for 1 h, washed, incubated in culture for a further 20 h and then released, samples then consisted of 70% monocytes (upper right quadrant). There was no significant change in CD14 expression on this activated monocyte population (R1); however, 99% of these cells were now CD105+ with a mean RFI of 200 (R2). At all time points, the lower right quadrant was at most 22% CD14+ and <20% CD105+, confirming that endoglin is not present on lymphocytes (data not shown). Together these data show that the monocyte population is enriched by adherence to plastic and that this attachment is required to induce endoglin expression, whereby complete induction can be observed after 21 h in culture. Thus, it would be feasible to use one or two color flow cytometry to quantify endoglin levels on activated monocytes, but not on PBMCs as they do not express sufficient levels.
We previously used metabolic labeling followed by immunoprecipitation analysis as this permits the detection of potential mutant proteins, as well as enabling the quantification of the expression of fully glycosylated endoglin. We thus show the kinetics of induction of newly synthesized endoglin by metabolic labeling (Fig. 1C). Monocytes were isolated from PBMCs by attachment to plastic for 1 h, washed in methionine-free medium and metabolically labeled for 4 h. The earliest activation time (time in culture) that was tested by metabolic labeling was 5 h, since we established previously that 4 h labeling is needed to reach steady-state levels. Both fully glycosylated endoglin (E) and the partially glycosylated precursor (P) were seen on activated monocytes after 5 h in culture (Fig. 1C). These species were identified by previous surface labeling and metabolic studies (22). When compared with 21 h in culture (including 4 h labeling), where maximum expression is reached, the expression of glycosylated endoglin (E) was 70% compared with the maximum. Interestingly, expression of the precursor (P) was higher at 5 h, suggesting that new endoglin synthesis had been recently initiated. This difference was more prominent under reducing conditions, suggesting that these species (E and P) are better resolved under these conditions. We did not detect a further increase at later time points (data not shown). However, as the monocyte layer often becomes more fragile and cells lift off with longer time in culture, reproducible results are more readily obtained by culturing the monocytes for 16-24 h followed by labeling for 4 h.
Endoglin expression is reduced in HHT1 patients from four families with missense mutations
Families 5, 85 and 89 were previously characterized and the missense codons, W149C, C53R and G52V, in these families reported (24). The individuals from these families used in this study, H150, H278 and H295, are clinically affected and mutations were confirmed by sequencing of appropriate exons (Fig. 2). The HUVEC sample H319, belonging to family 85, was also confirmed as carrying the mutation leading to the missense codon C53R. We identified a novel missense mutation in patient H277 that was found by sequencing exon 5 (Fig. 2); this point mutation causes conversion of L221 to P.
Figure 2. Sequence of four missense mutations. Exons 2, 4 and 5 were amplified by PCR from genomic DNA isolated from normal peripheral blood lymphocytes, lymphocytes of clinically affected HHT1 patients H295, H278, H150 and H277 and HUVECs from H319. Purified products were sequenced using a cycle sequencing protocol and resolved using a MicroGeneBlaster Sequencer (Visible Genetics). The profiles revealing mixed sequence in patients H295, H278, H150 and H319 resulting in the indicated substitutions confirm that these patients carry the previously described missense mutations (K = G or T; Y = T or C; S = G or C). The sequence of exon 5 from patient H277 shows the novel T->C substitution (S) creating the L->P missense codon at position 221.
To determine whether missense mutations in ENDOGLIN can also lead to haploinsufficiency, we quantified endoglin expression in activated monocytes from the four clinically affected adults described above. To maximize detection of a mutated protein, expression of endoglin was analyzed with monoclonal antibodies (mAbs) which recognize epitopes mapping to three distinct areas of the extracellular domain (31), as shown in Figure 3A. In all samples, fully glycosylated endoglin expression was reduced relative to controls (Fig. 4A). HUVECs from patient H319 were also reduced compared with normal HUVECs (Fig. 4B). Quantitation of these gels is shown in Table 1 and indicates that endoglin expression is ~50% or less, with the exception of patient H278, which had 70% relative to normal with the P3D1 antibody. However, expression levels in HUVEC sample H319, from the same family and confirmed as having the same mutation (C53R), were comparable to those seen with the other missense mutations.
Figure 3. Schematic diagram of missense mutations in endoglin cDNA and protein. (A) The intron-exon boundaries are indicated on the cDNA sequence, as is the initiation codon ATG corresponding to base pair 1 and to M1 of the leader peptide. The positions of the four missense mutations relevant to this study are depicted. The polypeptide structure of endoglin is illustrated with the approximate positions of cysteine residues (black circles) and the potential N-linked glycosylation sites (tridents). The N-terminal residue of the mature protein is E26. mAbs recognizing epitopes located in three distinct areas of the extracellular domain of endoglin are depicted: P3D1 binding to the N-terminal region (E26-G230); P4A4 and 44G4 binding to the Y277-G331 region; and RMAC8 binding to the C-terminal region (G331-G586). (B) PCR strategy and primers used to generate constructs with the corresponding missense mutations. The expression construct pcEXV-EndoL was used as a template; the positions of primers are shown relative to the cDNA insert and the size of each product is indicated above. For the first PCR, two fragments were made for each mutation that were complementary to the forward (MX5) and reverse (MX3) mutagenic primers. For the second PCR, complementary fragments were amplified in four separate reactions with outer primers AX5 and AX3 to generate four mutagenized products that were all 827 bp in length. These fragments were digested with SacII and SbfI as indicated and the 740 bp fragment was subcloned back into the expression construct.
Figure 4. Analysis of endoglin expression on activated monocytes from four HHT1 families. (A) PBMCs from 30 ml of blood from clinically diagnosed HHT1 patients H295, H278, H150 and H277, with the missense mutations indicated, or from normal controls (N) were activated by adherence to plastic and incubation in culture for at least 21 h, which includes labeling with [35S]methionine for 4 h. Metabolically labeled activated monocytes were processed as described in Figure 1A except that both mAbs P3D1 and P4A4 were used to immunoprecipitate endoglin. Equivalent c.p.m. (estimated by trichloroacetic acid precipitation of aliquots of total lysates prior to immunoprecipitation) were loaded for patients and matched controls in each experiment. (B) H319 and normal HUVECs were similarly labeled and analyzed as in (A). Duplicate samples from both H319 and normal HUVECs were loaded. Total radioactivity in the normal fully glycosylated endoglin (E; 90 kDa) band for each mAb P3D1 and P4A4 immunoprecipitate were quantitated by PhosphorImager and the pixel values for each patient relative to the matched normal controls are reported in Table 1.
Table 1. Clinical and molecular data for four HHT1 families with missense mutations analyzed
| Family no. | Clinical manifestations in family members | Patient | Mutation | Missense codon | Fully glycosylated (%) (HHT/normal) | ||
| P3D1 | P4A4 | Mean | |||||
| 89 | PAVM, CAVM | 295 adult | G155->T | G52->V | 37 | 57 | 47 |
| 85 | PAVM, CAVM | 278 adult | T157->C | C53->R | 71 | 50 | 61 |
| 85 | PAVM, CAVM | 319 newborn | T157->C | C53->R | 40 | 35 | 38 |
| 5 | PAVM, CAVM | 150 adult | G447->C | W149->C | 26 | 29 | 28 |
| 84 | PAVM | 277 adult | T662->C | L221->P | 41 | 29 | 35 |
In activated monocytes from all patients except H278, the band migrating as the partially glycosylated precursor (P) was stronger than normal (Fig. 4A). An elevated precursor was also seen in the HUVEC sample H319 (Fig. 4B), which has the C53R mutation. This suggested that mutant proteins were migrating in that position or could be interfering with the synthesis of normal protein. As it was difficult to obtain >30 ml of blood from all adult patients, it was not feasible to use monocytes for pulse-chase experiments. We thus tested the synthesis and stability of both fully processed endoglin (E) and the partially processed form (P) by pulse-chase experiments with HUVEC H319, as this was the only cord sample available from the families in this study (Fig. 5). In both normal and HUVEC sample H319, the maximum expression of fully glycosylated endoglin is reached after 2 h. From the analysis of multiple experiments, maximum expression of normal endoglin generally persists up to 4 h. However, as shown in Figure 5, expression at 3.5 h was 90% of that observed at 2 h, in both normal and H319 samples, and probably reflects experimental variation. Although total endoglin (E) was reduced in the patient (H319) relative to normal, the rate of appearance of fully processed endoglin (E) when expressed relative to the total (E + P) was not significantly different between these two samples. This was observed when mAbs recognizing epitopes in all three regions of the extracellular domain were used (Fig. 5). Total expression of fully processed endoglin (H319) relative to normal was similarly reduced at each time point, indicating that processing of the normal protein is not affected by the presence of a mutant protein (Fig. 5B). However, the band migrating as the precursor was more stable in the patient's cells and was still visible after a 3.5 h chase time. In contrast, precursor levels were reduced to 10% in normal cells by 2 h (Fig. 5C). Similar results were obtained using a polyclonal antibody (pAb) to human endoglin (data not shown). Interestingly, mAb RMAC8, recognizing an epitope between G331 and G586 of the extracellular domain of endoglin, was slightly more reactive with the precusor form of HUVEC H319. These data suggest that the intracellular precursor pool is a combination of mutant and normal at time 0 and 1 h wheareas it is mostly mutant after 2 h of chase. Thus, the mutant precursor is misfolded and retained intracellularly, where it is eventually degraded.
Figure 5. Pulse-chase analysis of endoglin in HUVECs expressing the C53R missense mutation. (A) H319 and normal HUVECs were pulsed with [35S]methionine for 20 min and chased with medium containing Met. At each time point, extracts were prepared, immunoprecipitated with mAbs P3D1, P4A4 and RMAC8 to endoglin and fractionated as described in Figure 1A. Total radioactivity in the normal fully glycosylated endoglin (E; 90 kDa) and precursor (P; 80 kDa) was quantitated by PhosphorImager with ImageQuant software. (B) Rate of appearance of fully glycosylated endoglin (E) for H319 (closed symbols) and normal HUVECs (open symbols); total pixel values for E relative to the total (E + P) at each time point and for each antibody was plotted versus chase time. Circles, P3D1; squares, P4A4; triangles, RMAC8. (C) Stability of partially glycosylated precursor (P) for H319 (closed symbols) and normal HUVECs (open symbols); total pixel values for P relative to total (E + P) at each time point and for each antibody was plotted versus chase time. Circles, P3D1; squares, P4A4; triangles, RMAC8.
HHT1 missense mutants expressed in COS-1 cells are not fully processed and not detectable at the surface
To establish whether missense proteins were expressed at the cell surface, we generated each mutation by overlap PCR (see Materials and Methods; Fig. 3B). COS-1 cells transiently transfected with each mutant were compared with those transfected with normal endoglin by both metabolic labeling and cell surface biotinylation. As shown in Figure 3A, all four missense codons fall within the first 200 amino acids, where the P3D1 epitope was mapped, whereas the P4A4 epitope falls outside this region. Thus, we expected that at least one of these two antibodies would react with each mutant protein. After metabolic labeling for 4 h (Fig. 6A), the mutants G52V and C53R were barely detectable (compare lanes 1 and 2 with lanes 7 and 8). The total expression in COS-1 cells of these mutants relative to normal endoglin was 3 and 8% with mAbs P3D1 and P4A4, respectively (Fig. 6B). Mutant W149C was only detectable with mAb P4A4 (lanes 3 and 9) at 16% of normal (Fig. 6B). Mutant L221P exhibited the highest expression level, 24% compared with normal, but mAb P3D1 was more reactive with this mutant than mAb P4A4. The metabolic labeling also showed that the mutants co-migrated with the partially processed precursor (P) of normal endoglin and were thus not fully processed, and would be degraded intracellularly. The highest levels of mutant detected relative to normal precursor (P) (Fig. 6C) were 9 (G52V), 17 (C53R), 45 (W149C) and 110% (L221P). Cell surface biotinylation (Fig. 6A and D) revealed that only proteins with the mutations W149C and L221P were detectable at the surface of COS-1 transfectants, but at levels of 3 and 9% of normal endoglin. In transfection systems, expression levels are very high, thus a small percentage of misfolded proteins might escape degradation and be detectable at the cell surface. These data suggest that the four missense mutants are not processed fully and are present as transient species degraded intracellularly.
Figure 6. Expression of endoglin with missense codons in COS-1 cells. COS-1 cells were transiently transfected with pcEXV-EndoL (normal endoglin) and pcEXV constructs containing the generated missense codons in endoglin G52V, C53R, W149C and L221P (as illustrated in Fig. 3). (A) Transfected cells were labeled with [35S]methionine for 4 h, solubilized in Triton X-100, immunoprecipitated with mAbs P3D1 (lanes 1-6) and P4A4 (lanes 7-12), analyzed and quantitated as in Figure 1A. Lanes 13 (P3D1) and 14 (P4A4) represent images of normal endoglin as seen with the PhosphorImager with ImageQuant software and were used to identify both fully glycosylated endoglin (E) and partially glycosylated precursor (P) for quantitation. Cells were also surface labeled with biotin and extracts containing equivalent total protein from surface labeled transfectants were immunoprecipitated with mAbs P3D1 (lanes 1-6) and P4A4 (lanes 7-12). Eluates were fractionated, gels were transferred to PVDF nylon membranes, probed with streptavidin-horseradish peroxidase and detected by enhanced chemiluminescence. Multiple exposures were obtained and gels were quantitated using a densitometer and ImageQuant software. (B) The total pixel values for all bands in each lane were calculated and background from corresponding regions in the empty vector (pcEXV-1) control was subtracted. The values for each missense mutant transfectant were plotted relative to normal endoglin transfectant. (C) As metabolically labeled mutant proteins co-migrated with the partially glycosylated normal precursor, total pixel values for each precursor minus background from empty vector control were plotted as a percentage of normal precursor. (D) The total amount of mutant protein measured by surface labeling from all bands in each transfectant was expressed relative to that obtained for normal, after subtracting background (vector alone) from each. Note the difference in scales of the histograms.
Since pulse-chase experiments of HUVEC sample H319 (C53R) suggested that the precursor was more stable and accumulated intracellularly, we studied this mutant by pulse-chase experiments in COS-1 cells (Fig. 7). At time 0, only precursor was observed for both the C53R mutant and normal endoglin. Levels of expression relative to normal were 40% with RMAC8 and 13% with P4A4 (Fig. 7 compare lanes 2 and 3). As the mAb RMAC8 reacts with the C-terminal portion of the extracellular domain, whereas the C53R substitution lies within the N-terminal portion, it is likely to be more efficient at detecting this mutant. This also indicates that the C53R mutant is fully synthesized. A pAb to human endoglin also precipitated the mutant equally well as P4A4 (data not shown). Thus, it is critical to use mAbs reacting with different antigenic regions of endoglin (Fig. 3A) when analyzing expression of mutant forms. After a 3.5 h chase no fully glycosylated form (E) of the mutant C53R is expressed at the cell surface, where normal endoglin is almost exclusively found (Fig. 7, compare lanes 6 and 8 with lane 7). The C53R mutant is still detectable, but only as a precursor, intracellular form (Fig. 7, P, lane 7). The level of this mutant protein was higher when detected with RMAC8 mAb than with P4A4 and similar to levels of mutant W149C shown in Figure 6. Thus, the COS cell data confirm that missense mutants are expressed, but only as intracellular forms.
Figure 7. Pulse-chase analysis in COS-1 cells expressing recombinant endoglin and the mutant C53R. COS-1 cells were transiently transfected with pcEXV-LEndo (normal endoglin), the pcEXV construct containing the generated C53R missense codons in endoglin and a combination of both. Transfected cells were pulsed with [35S]methionine for 20 min and chased with medium containing Met for 0 and 3.5 h. At each time point, extracts were prepared, immunoprecipitated with mAbs P4A4 and RMAC8 and fractionated as described in Figure 1A. Total radioactivity in the normal fully glycosylated endoglin (E; 90 kDa) and precursor (P; 80 kDa) was quantitated by PhosphorImager with ImageQuant software.
DISCUSSION
ENDOGLIN is the gene mutated in HHT1. It is expressed at high levels on endothelial cells, but at lower levels on activated monocytes. We previously used both cell types to analyze expression of normal and mutant endoglin in samples from HHT1 patients (22). Since endothelial cells from adult patients are not readily available, we are limited to the use of peripheral blood to analyze endoglin expression in these HHT1 patients. By flow cytometry, we now confirm that maximum endoglin expression is reached on activated monocytes within 1 day in culture and is not further increased even after 5 days. We also demonstrate by two color flow cytometry that in peripheral blood only cells of the myeloid lineage (CD14+) express endoglin, and at low levels. These endoglin-positive cells represent only 4% of total PBMCs. Enrichment of the monocyte population by attachment to plastic and incubation for 1 day in culture are required for maximal endoglin expression. After 21 h in culture, 99% of activated monocytes expressed twice as much endoglin as monocytes at time 0. By metabolic labeling, expression of fully processed endoglin was activated within 5 h in culture and was at 70% of the maximum observed after 21 h in culture. Thus, in order to measure levels of endoglin on monocytes, either by flow cytometry or metabolic labeling, activation for 21 h in culture is optimal.
Mutations resulting in reduced protein expression as seen for HHT1 may arise from a number of mechanisms, such as deletions and truncations caused by nonsense and frameshift mutations, as well as some promoter and splice site mutations (30). Missense mutations are more likely to be dominant-negatives or they might enhance or disrupt protein function. If expressed at normal levels, such mutants could reveal key functional residues. To test this possibility, we analyzed mutant proteins corresponding to four missense mutations. Fully glycosylated endoglin in activated monocytes from clinically affected patients having missense mutations in exons 2, 4 and 5 was reduced to mean values ranging from 28 to 61% of normal (Table 1). As multiple patient samples are difficult to obtain, in particular from elderly patients, as is the case with H278 (where a value of 71% was observed with mAb P3D1), experimental variation must be considered. In an independent study, three separate experiments performed with a confirmed HHT1 patient gave endoglin levels of 35 ± 11, 42 ± 8 and 60 ± 10% of normal. Furthermore, the variation observed does not reflect the mutation per se, as evidenced by the differences observed for H319, with a mean level of 38% relative to normal, compared with 61% for H278 carrying the same missense mutation. Recently, we have obtained data for two patients from another family carrying a missense mutation at codon 383. Analysis of monocyte cultures from both these patients revealed 63 ± 7 and 28 ± 5% of normal levels. Thus, there is experimental variation in the level estimated for a single patient as well as for different family members with the same mutation. The range of all mean values obtained for affected individuals in all HHT1 families analyzed to date is ~30-70% (32). There is also variation in the level of endoglin estimated for normal individuals with mean values ranging from 73 to 140% with an average of 100 ± 35% (33). Along with experimental variation, the range of expression could indicate that endoglin gene transcription is not stably induced and might reflect differences in initiation and deactivation of transcription. In a recent publication, where a stochastic model of gene expression was used to estimate product levels from one and two alleles, it was demonstrated that fluctuations from the expected values of 50 or 100% product were greater when only one allele is expressed (34). Thus, even in an ideal representation of haploinsufficiency, such as a null mutation, the expression level from the normal allele will deviate from the theoretical 50%.
Serious complications of HHT, such as pulmonary arteriovenous malformation (PAVM), were reported in all four families, while cerebral arteriovenous malformation (CAVM) was found in three of them, indicating that missense mutations can cause as severe a phenotype as other types of HHT1 mutation. This is consistent with a previous study where 30 individuals from eight different HHT families were analyzed with respect to age of presentation and severity of clinical symptoms. There was a lack of correlation between genotype and clinical manifestations, supporting a haploinsufficiency model (23). So far, >60 families with reduced endoglin expression have been identified and only four mutant proteins have been detected, along with the present four detected in this study; in all cases mutant proteins were transiently expressed, intracellularly (22; unpublished data).
As proteins with missense codons are not a priori distinct from normal ones in patient samples, we expressed and analyzed these mutants in COS-1 cells. By metabolic labeling, we found mutant proteins expressed in COS-1 transfectants co-migrating with partially processed normal endoglin precursor. None of these mutant proteins was significantly expressed at the cell surface, indicating that they are retained intracellularly as transient species. This is in agreement with patient samples where elevated precursor levels were noted, suggesting a mixture of mutant and normal precursors. Pulse-chase experiments with HUVEC H319 demonstrated that precursor accumulated in particular after 2 h of chase. Yet, the rate of processing of endoglin produced from the normal allele was not affected. In addition, pulse-chase experiments in COS-1 cells confirmed that mutant proteins are indeed expressed, albeit intracellulary and not at the cell surface.
Mutations leading to a stable cell surface expressed protein associated with HHT1 have yet to be described. Such mutants would be beneficial in structure-function analyses. However, from the analysis of expression of the missense mutants and considering effects of these specific mutations on structure, one can begin to ascertain key residues in protein folding. The use of mAbs reactive with epitopes mapping to the different regions of the extracellular domain of endoglin is critical in the analysis of mutant proteins (31). Mutations in exons 2 and 4 were more disruptive than that in exon 5, as judged from the level of precursor protein produced. The mutation created at C53 is located 13 residues downstream from C40; both cysteines are conserved in endoglin and betaglycan amongst species (35,36) and likely form an intracellular disulfide bond. Replacing C53 with R, a positively charged residue, did not disrupt dimer formation (data not shown), but altered folding such that the P3D1 epitope was completely destroyed, P4A4 reactivity was reduced and maximum reactivity was observed with RMAC8 (Fig. 3). The mutation creating G52V is equally disruptive and this mutant protein also dimerizes. This region must be required for proper folding of the N-terminal domain, early in the synthesis of endoglin. As two or more consecutive domains are thought to fold co-translationally, disruption of this domain likely results in a completely misfolded protein that would be degraded in the endoplasmic reticulum (37). Introduction of a C at position W149 could also be disrupting formation of a disulfide bond. The fourth mutation in exon 5, converting L221 to P, likely disrupts folding, as it occurs in a predicted [alpha]-helix (38,39). P4A4 reactivity, which mapped previously to a region downstream of this position (Fig. 3), is completely abolished. This indicates structural alterations beyond the mutation at L221 and implies that the P4A4 epitope, although still detectable in a synthetic denatured protein (31), requires proper folding for maximum detection.
We conclude from our studies that these missense mutations lead to transient intracellular species and are not significantly expressed at the cell surface. To date, all HHT1 patients analysed at the protein level have revealed reduced levels of fully glycosylated normal endoglin at the cell surface. Endothelial cells express ~106 endoglin molecules on the cell surface, where it is known to function and interact with the TGF-[beta] receptor complex (7,40). Cell surface endoglin produced from the normal allele in HUVECs expressing a transient intracellular mutant with an intact transmembrane region was previously found to function normally and thus interact with the TGF-[beta]1 receptor complex (22). We demonstrate that in HHT1 patients the missense mutants are also intracellular, misfolded proteins that by virtue of their localization would not alter the function of the normal allele expressed at the cell surface. These data are consistent with haploinsufficiency being the predominant mechanism associated with this genetic vascular disorder.
MATERIALS AND METHODS
Cell culture and transfections
All materials were from Gibco BRL, Canadian Life Technologies (Burlington, Ontario, Canada), unless otherwise specified. PBMCs were obtained from whole venous blood as described (22). Briefly, blood samples were depleted of erythrocytes by gravity sedimentation through Dextran T-500 and mononuclear cells recovered by Ficoll-Paque density gradient centrifugation. Separation of monocytes from lymphocytes was achieved by adherence to plastic for 1 h at 37°C. This adherence step initiates the activation of monocytes into a macrophage-like phenotype. Lymphocytes were collected and used for DNA isolation. Monocyte layers were then washed and incubated in culture for periods of time varying from 4 h to 5 days, to determine optimal conditions of endoglin expression.
HUVECs were derived from newborns of HHT1 families and control babies by previously published procedures (41). For assays, equivalent cell densities and passage numbers were employed with patient and matched controls as published (22).
COS-1 cells were maintained and transiently transfected with expression constructs using the DEAE/Dextran/chloroquine method as reported (42,43). Assays were performed 2 days post-transfection.
Antibodies
P3D1 and P4A4 mAbs were provided by Dr E.A. Wayner (Seattle, WA). RMAC8 was obtained via the 5th International Leukocyte Differentiation Workshop (44). The mAbs were shown to react with epitopes located between the N-terminal residues E26 and G230 for P3D1, between residues Y277 and G331 for P4A4 and 44G4 (38,45) and between G331 and G586 for RMAC8 (31). These epitopes are depicted in Figure 3 and are shown relative to the four missense mutations studied in this paper. FITC-conjugated F(ab[prime])2 goat anti-mouse IgG was purchased from Tago (Burlingame, CA). Rhodamine (RD1)-conjugated MY4 (IgG2b) reactive with CD14, and mAb MO-1, reactive with CD11b, were used as markers of the myelomonocytic lineage and were obtained from Coulter Electronics (Hialeah, FL), as were IgG2b and IgG1 isotype controls.
Patient samples and mutation analysis
Informed consent was obtained from all individuals participating in the study, including clinically diagnosed HHT parents or their spouses, in the case of newborn umbilical cord samples. Positive diagnoses were based on established criteria (1,46). All patients and families were given a number and patients are referred to with the prefix H (for HHT). Patients H295 and H278 (as well as baby H319) belong to large families whose clinical history and linkage to 9q33 were previously reported to have substitutions G155T and T157C in exon 2 of the ENDOGLIN gene, leading to the missense codons G52V and C53R (24). In this paper, these two families are designated families 89 and 85, respectively. Patient H150, also included in this study, belongs to a family designated here family 5. This family was previously described as having the missense mutation G447C that creates a W149->C conversion (24). Patient H277, with the novel missense mutation T662C in exon 5 creating an L->P substitution at codon 221, belongs to family 84.
All mutations in individuals tested in the study were confirmed by sequencing as well as in two additional clinically affected members from family 5. Genomic DNA was isolated from peripheral blood lymphocytes using DNAzol reagent (Gibco BRL). Appropriate exons were amplified by PCR and sequenced using a cycle sequencing protocol described (22) and primers reported previously (6). Products were resolved on a MicroGeneBlaster Sequencer (Visible Genetics, Toronto, Ontario, Canada). Positions of these mutations are shown in Figure 3A.
Preparation of expression constructs
Full-length endoglin cDNA in the mammalian expression construct pcEXV-EndoL was provided by Dr C. Bernabeu (Centro de Investigaciones Biológicas, CSIC, Madrid, Spain) (47). Four missense mutations in endoglin cDNA were generated by an overlap PCR strategy (Fig. 3B). The unique restriction sites SacII, at position 32 within the insert of pcEXV-EndoL, and SbfI, at position 772, were used to subclone final mutagenized fragments. Outer primers (forward, AX5, 5[prime]-CTGCAGGGAATTCCGTGGACAGCAT-3[prime]; reverse, AX3, 5[prime]-AGATCTGCATGTTGTGGTTGGCGTCGAT-3[prime]) were thus designed to flank these two restriction sites and encompass all four missense mutations. Complementary mutagenic primers were designed for each mutation:
M1X3, 5[prime]-TGAGCCACGCAGACCTTCGAAACCTGGCTAGT-3[prime]);
T157C (M2X5, 5[prime]-ACTAGCCAGGTTTCGAAGGGCCGCGTGGCTCA-3[prime], and
M2X3, 5[prime]-TGAGCCACGCGGCCCTTCGAAACCTGGCTAGT-3[prime]);
G447C (M3X5, 5[prime]-AGACCCAGATCCTTGAGTGCGCAGCTGAGAG-3[prime], and
M3X3, 5[prime]-CTCTCAGCTGCGCACTCAAGGATCTGGGTCCTGCCG-3[prime]);
T662C (M4X5, 5[prime]-ACAAGGAGGCGCACATCCCGAGGGTCCTGCCG-3[prime], and
M4X3, 5[prime]-CGGCAGGACCCTCGGGATGTGCGCCTCCTTGT-3[prime]).
The mutations generated with each set of primers are in bold. For M1 and M2, a T was introduced instead of a C, shown underlined. This destroys the BsaI site, but the codon remains unaffected. For M3 the mutation creates an FspI site and for M4 the mutation destroys an SauI site. The first PCR consisted of eight reactions using pcEXV-EndoL as template and AX5 with each mutagenic reverse primer and AX3 with each mutagenic forward primer, as shown in Figure 3B. An aliquot of 20 ng of template was amplified in a 100 µl volume containing Perkin-Elmer reagents (Perkin-Elmer, Roche Molecular Systems, Branchburg, NJ) in PCR reaction buffer: 2.0 mM MgCl2, 200 nM each dNTP, 200 ng of primers, 2 U AmpliTaq (Perkin-Elmer, Roche Molecular Systems) and 1 U VentR (New England Biolabs, Mississauga, Ontario, Canada) DNA polymerases. Cycling conditions were: two cycles of 5 min at 94°C, 3 min at 58°C, 3 min at 72°C, then 23 cycles of 30 s at 94°C, 30 s at 58°C, 60 s at 72°C, and a final extension of 5 min at 72°C. PCR products were fractionated by agarose gel electrophoresis, excised and purified with Qiaex (Qiagen, Chatsworth, CA). The second PCR consisted of four reactions using purified fragments complementary to the mutagenic primers and AX5 with AX3 outer primers to create the full-length products containing the mutations. Cycling conditions were the same. All four products of 827 bp were subjected to restriction analysis to confirm the presence of mutations, then purified as above. Mutagenized fragments were extracted with chloroform:isoamyl alcohol (24:1), then precipitated prior to digestion with SacII and SbfI to produce 740 bp fragments that were subcloned into pcEXV-EndoL. PCR was performed on picked bacterial colonies, followed by restriction analysis. Positive clones were sequenced in both forward and reverse directions with primers just upstream (forward, 5[prime]-AAGAACTGCTCCTCAGTGAT-3[prime]) and downstream (reverse, 5[prime]-AAGGAGTATTCTCCAGT-3[prime]) of AX5 and AX3, respectively, by ACGT (Toronto, Canada).
Flow cytometry
Normal PBMCs were obtained and monocytes were enriched by adherence, activated and incubated in culture as described above. Adherent cells were released with 10 mM EDTA in phosphate-buffered saline (PBS) (Ca2+, Mg2+-free), washed and resuspended in PBS containing 2% fetal bovine serum. All samples were incubated with 2.0% autologous plasma at each staining step to prevent Fc receptor-mediated binding. For single color flow cytometry, PBMCs from 60 ml of blood were equally divided and harvested after 21, 45 and 114 h. Activated monocytes were incubated with saturating amounts of mAb 44G4 or the IgG1 isotype control followed by FITC-conjugated F(ab[prime])2 goat anti-mouse IgG. For two color flow cytometry, PBMCs from 60 ml of blood were also equally divided and used at time 0 and after 1 and 21 h adherence. PBMCs or activated monocytes were then incubated with saturating amounts of IgG-RD1 and MY4-RD1, followed by extensive washing and incubation with mAb P4A4 or IgG1 plus FITC-conjugated F(ab[prime])2 goat anti-mouse IgG. A gate was set on CD14+ cells and these were analyzed for CD105 expression. Percentage of positive cells and mean fluorescence intensity were determined relative to negative controls. All samples were analyzed on the FACScan (Becton Dickinson, Mountain View, CA) and in each experiment an unstained sample was used for compensation and calibration of the FACScan. FITC was detected with the FL1 detector and RD1 with the FL2 detector.
Metabolic labeling and pulse-chase experiments
Endoglin expression in activated monocytes from 20-30 ml of blood, equivalent numbers of HUVECs or transfected COS-1 cells was quantitated by metabolic labeling according to published procedures (22). Briefly, cells were incubated with 100 µCi/ml [35S]methionine (Trans-label; ICN Pharmaceuticals Canada, Montreal, Quebec, Canada) in methionine-free Dulbecco's modified Eagle's medium (DMEM) (low glucose; Gibco BRL) for 4 h, solubilized in lysis solution containing 1% Triton X-100, equivalent counts per minute (c.p.m.) values were estimated by trichloroacetic acid precipitation and lysates immunoprecipitated with saturating amounts of antibodies (P3D1, 4 µg; P4A4, 1.0 µg). Immune complexes were collected with bovine serum albumin-adsorbed Gamma-bindG Sepharose (Pharmacia Biotech Canada, Montreal, Quebec, Canada), washed, eluted, equivalent c.p.m. values were fractionated by SDS-PAGE (4-12% gradient gels; Novex Experimental Technology, San Diego, CA) and gels were either subjected to fluorography or autoradiography with BioMax MS film and the BioMax TranScreen LE intensifying screen system (Eastman Kodak, Rochester, NY). Multiple exposures of each experiment were obtained. Gels were then quantitated using a PhosphorImager and ImageQuant Software (Molecular Dynamics, Sunnyvale, CA) as described. For pulse-chase experiments using HUVECs, equivalent cell numbers were pulse labeled for 20 min with 250 µCi/ml [35S]methionine followed by two washes, then incubated further in serum-free DMEM for a chase period of between 0 and 3.5 h. Cells were lysed and analyzed as described above.
Cell surface biotinylation
Equivalent numbers of transiently transfected COS-1 cells were surface labeled with biotin as reported (22). Cells were lysed and immunoprecipitated as described under Metabolic labeling, except that lysates were precleared for at least 1 h with protein A-Sepharose CL-4B (Pharmacia Biotech Canada). Eluates were fractionated on 4-12% SDS-PAGE gels, electrotransferred to PVDF nylon membranes for 1 h at 45 V, blocked for 1 h in TBS-T (0.02 M Tris, pH 7.5, 0.137 M NaCl, 0.1% Tween-20) containing 5% dry skimmed milk powder and probed with streptavidin-horseradish peroxidase (400-fold dilution in TBS-T; Amersham Life Sciences, Oakville, Ontario, Canada) for 20 min. Biotinylated proteins were detected using enhanced chemiluminescence (Amersham Life Sciences); multiple exposures were obtained. Bands on autoradiographs were quantitated using a densitometer and ImageQuant Software (Molecular Dynamics).
ACKNOWLEDGEMENTS
We gratefully acknowledge all the patients who participated in the study and the support of the HHT Foundation International. We are also grateful to J. MacDonald and Dr A.E. Guttmacher for providing clinical information on the patients. We thank Dr C. Bernabeu (Madrid, Spain) for providing the expression construct END and the pAb to endoglin, Dr E.A. Wayner (Seattle, WA) for mAbs P3D1 and P4A4, L. Gunaratnam for assisting on establishing conditions for the macrophage cultures, K.A. Galley and Dr G. Boileau for advice on site-directed mutagenesis, and Dr J. Wrana for insightful criticism and helpful discussions. We thank Visible Genetics Inc. for financial support and use of the MicroGeneBlaster Sequencer. This work was supported by grant no. NA3434 from the Heart and Stroke Foundation of Ontario and from a grant from the Medical Research Council of Canada.
REFERENCES
+To whom correspondence should be addressed. Tel: +1 416 813 6258; Fax: +1 416 813 6255; Email: mablab{at}sickkids.on.ca
This page is run by Oxford University Press, Great Clarendon Street, Oxford OX2 6DP, as part of the OUP Journals
Comments and feedback: jnl.info{at}oup.co.uk
Last modification:
Copyright© Oxford University Press, 1999.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  F. Gedge, J. McDonald, A. Phansalkar, L.-S. Chou, F. Calderon, R. Mao, E. Lyon, and P. Bayrak-Toydemir Clinical and Analytical Sensitivities in Hereditary Hemorrhagic Telangiectasia Testing and a Report of de Novo Mutations J. Mol. Diagn., April 1, 2007; 9(2): 258 - 265. [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]
|
|
|
|
|
|
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]
|
|
|
|
 |
|
 C. B. Schmidt-Weber, M. Letarte, S. Kunzmann, B. Ruckert, C. Bernabeu, and K. Blaser TGF-{beta} signaling of human T cells is modulated by the ancillary TGF-{beta} receptor endoglin Int. Immunol., July 1, 2005; 17(7): 921 - 930. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Toporsian, R. Gros, M. G. Kabir, S. Vera, K. Govindaraju, D. H. Eidelman, M. Husain, and M. Letarte A Role for Endoglin in Coupling eNOS Activity and Regulating Vascular Tone Revealed in Hereditary Hemorrhagic Telangiectasia Circ. Res., April 1, 2005; 96(6): 684 - 692. [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]
|
|
|
|
 |
|
 F. Sanz-Rodriguez, A. Fernandez-L., R. Zarrabeitia, A. Perez-Molino, J. R. Ramirez, E. Coto, C. Bernabeu, and L. M. Botella Mutation Analysis in Spanish Patients with Hereditary Hemorrhagic Telangiectasia: Deficient Endoglin Up-regulation in Activated Monocytes Clin. Chem., November 1, 2004; 50(11): 2003 - 2011. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A Chaouat, F Coulet, C Favre, G Simonneau, E Weitzenblum, F Soubrier, and M Humbert Endoglin germline mutation in a patient with hereditary haemorrhagic telangiectasia and dexfenfluramine associated pulmonary arterial hypertension Thorax, May 1, 2004; 59(5): 446 - 448. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Seki, J. Yun, and S. P. Oh Arterial Endothelium-Specific Activin Receptor-Like Kinase 1 Expression Suggests Its Role in Arterialization and Vascular Remodeling Circ. Res., October 3, 2003; 93(7): 682 - 689. [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]
|
|
|
|
 |
|
 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.-E. Paquet, N. Pece-Barbara, S. Vera, U. Cymerman, A. Karabegovic, C. Shovlin, and M. Letarte Analysis of several endoglin mutants reveals no endogenous mature or secreted protein capable of interfering with normal endoglin function Hum. Mol. Genet., June 1, 2001; 10(13): 1347 - 1357. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Bourdeau, M. E. Faughnan, M.-L. McDonald, A. D. Paterson, I. R. Wanless, and M. Letarte Potential Role of Modifier Genes Influencing Transforming Growth Factor-{beta}1 Levels in the Development of Vascular Defects in Endoglin Heterozygous Mice with Hereditary Hemorrhagic Telangiectasia Am. J. Pathol., June 1, 2001; 158(6): 2011 - 2020. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G M F Wallace and C L Shovlin A hereditary haemorrhagic telangiectasia family with pulmonary involvement is unlinked to the known HHT genes, endoglin and ALK-1 Thorax, August 1, 2000; 55(8): 685 - 690. [Abstract] [Full Text]
|
|
|
|
|
|
S. A. Abdalla, N. Pece-Barbara, S. Vera, E. Tapia, E. Paez, C. Bernabeu, and M. Letarte Analysis of ALK-1 and endoglin in newborns from families with hereditary hemorrhagic telangiectasia type 2 Hum. Mol. Genet., May 1, 2000; 9(8): 1227 - 1237. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Lux, C. J. Gallione, and D. A. Marchuk Expression analysis of endoglin missense and truncation mutations: insights into protein structure and disease mechanisms Hum. Mol. Genet., March 22, 2000; 9(5): 745 - 755. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Massagué and Y.-G. Chen Controlling TGF-beta signaling Genes & Dev., March 15, 2000; 14(6): 627 - 644. [Full Text]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||