Synectin-Dependent Regulation of Arterial Maturation

Abstract

Synectin, a ubiquitously expressed PDZ scaffold protein, has been shown to be a key regulator in the formation of arterial vasculature. Examination of the retinal vasculature in synectin^−/− mice demonstrated poor mural cell coverage of and attachment to the forming arterial tree, a defect reminiscent of retinal abnormalities observed in platelet derived growth factor (PDGF)-B^−/− mice. Primary cultures of synectin^−/− smooth muscle cells had normal expression of PDGFR-β and migrated normally in response to PDGF-BB. However, expression of PDGF-BB protein, but not mRNA, was reduced in lysates from arterial, but not venous, primary synectin^−/− endothelial cells (EC), that was restored by inhibition of proteosomal degradation. Transduction of synectin^−/− and ^+/+ EC with a bicistronic Pdgfb/gfp construct, resulted in comparable expression of green fluorescent protein in both EC populations while PDGF-BB expression was severely reduced in synectin^−/− EC. Finally, synectin expression in synectin^−/− arterial EC restored PDGF-BB protein levels. These results suggest that synectin deficiency results in increased degradation of PDGF-BB protein in arterial EC and, consequently, reduced recruitment of mural cells to newly forming arteries. This observation may explain the selective reduction in arterial morphogenesis observed in synectin knockout mice.

INTRODUCTION

Recruitment of mural cells is important for the stabilization and proper function of developing arterial vessels. Studies from several laboratories have implicated endothelial cell-produced platelet derived growth factor-B (PDGF-B) as a key player in mural cell recruitment to the forming arterial vasculature (Armulik et al., 2005). Knockout of PDGF-B or its primary receptor, PDGFR-β results in perinatal lethality and microvascular hemorrhaging due to extensive dilation caused by defective mural cell recruitment (Leveen et al., 1994; Soriano, 1994; Lindahl et al., 1997; Hellstrom et al., 1999).

Endothelial expression and retention of PDGF-B as well as N-sulphation of heparan sulfate proteoglycans are required for normal investment of mural cells into nascent blood vessels and remodeling arteries (Enge et al., 2002; Lindblom et al., 2003; Abramsson et al., 2007). Coincident with decreased pericyte density, retinas from mice with an endothelium-specific disruption of PDGF-B expression demonstrate increased regression of vessels in the outer retinal plexus (Enge et al., 2002). In addition to PDGF-B, disruption of the arterial specific gene ephrin-B2 is also associated with defective pericyte recruitment (Foo et al., 2006) with a phenotype similar to that of PDGF-B and PDGFR-β knockout mice. The precise mechanism of reduced mural coverage in ephrin-B2–deficient arteries is not clear, but this finding may point to a relationship between arterial specification and vessel stabilization.

Synectin, a ubiquitously expressed scaffold protein, has recently been shown to be a selective regulator of arterial specification and branching of small arterial vessels (Chittenden et al., 2006; Dedkov et al., 2007). Mice with a global disruption of synectin expression demonstrate reduced arterial branching and the arteries appear substantially smaller than corresponding arteries in wild-type mice. Furthermore, synectin-deficient arterial endothelial cells demonstrate reduced ephrin-B2 expression and decreased ability to respond to PDGF-BB stimulation, including defective migration, proliferation, lamellapodia formation, and branching. These findings suggest that mural cell recruitment may also be abnormal in synectin-null arteries.

To address this hypothesis, we used a retinal model similar to that was previously used for vascular morphogenesis studies in PDGF-B^−/− mice. We find that arteries in synectin null retinas show poor coverage by mural cells that is similar to PDGF-B^−/− mice. We further find that this defect is due to reduced PDGF-BB secretion by arterial, but not venous, synectin knockout endothelial cells, which in turn is caused by increased degradation of PDGF-BB protein. These findings identify synectin as an important posttranslational regulator of PDGF-BB expression in the arterial endothelium.

RESULTS

Synectin Is Not Required for Normal Primary Vascular Plexus Patterning but Plays a Role in Mural Cell Recruitment

To elucidate the role of synectin in the development of arterial vasculature, we used a previously characterized flat-mounted retinal model (Enge et al., 2002). Qualitative analysis of isolectin immunostaining of retinas derived from postnatal day (P) 5 synectin^−/− and ^+/+ mice revealed a similar extent of vascularization and comparable appearance of the primary vascular plexus in both groups (Fig. 1A,B). Quantification of isolectin-stained retinas further demonstrated similar mean distances of the retinal vascular plexus from the optic nerve, a measure of the extent of plexus formation, in synectin^−/− and control mice (Fig. 1C).

Fig. 1.

A,B: Flat-mounted retinas from synectin^+/+ (A) and synectin^−/− (B) mice immunostained for isolectin (green). C: Mean vessel length from the optic nerve was measured and quantified. Means ± SEM.

Although the plexus forms normally in synectin^−/− retinas, in comparison to synectin^+/+ (Fig. 2), synectin^−/− retinas demonstrated reduced expression of α-smooth muscle actin (SMA). Expression of NG2, a marker for both pericytes and vascular smooth muscle cells (SMC; Ozerdem et al., 2001), also appeared markedly reduced around synectin^−/− compared with synectin^+/+ arteries (arrows in Fig. 3A,B), while no difference in its expression was noted around veins and in the capillary plexus between synectin null and wild-type mice (Fig. 3B). Higher magnification revealed extensive gaps between mural cells covering the arterial endothelium (arrows in Fig. 3D, Figure 3F) compared with synectin^−/− (Fig. 3C,E), suggesting a defect in mural cell recruitment to arteries, including those oriented radially.

Fig. 2.

A–D: Isolectin (green) and smooth muscle actin (red) immunostaining of flat-mounted retinas from synectin^+/+ (A,C) and synectin^−/− (B,D) mice. Reduced expression of smooth muscle actin was seen in synectin^−/− retinas.

Fig. 3.

A–F: Isolectin (red) and NG2 (green) immunostaining of flat-mounted retinas from synectin^+/+ (A,C,E) and synectin^−/− (B,D,F) mice. Reduced expression of NG2 in arterial (arrows in B,D), but not venous, vessels and extensive gaps between mural cells (arrowheads, D) were seen in synectin^−/− vessels.

The observed mural cell coverage defect observed at this stage of retinal development (P5–P6) was not a transient event. Indeed, same defects were observed at P17 (Supp. Fig. S1, which is available online). Furthermore, the extent of arterial network and its branching appeared reduced at this time point in synectin null compared with control retinas. This is consistent with a previously observed decrease in arterial branching in adult synectin null mice (Chittenden et al., 2006) and suggests that the lack of mural cell coverage over time may lead to reduction in the size of arterial network.

Synectin Is Not Required for SMC Migration in Response to PDGF-BB

The observed defective mural cells recruitment in synectin^−/− retinas could be due to impaired responsiveness of these cells to PDGF-BB or impaired PDGF-BB production by synectin^−/− endothelial cells. To address the former possibility, we isolated SMCs from the abdominal aorta of synectin^−/− and synectin^+/+ mice and assessed expression of PDGFR-β and the ability to migrate in response to PDGF-BB in vitro. Western analysis revealed comparable levels of PDGFR-β in both synectin^−/− and ^+/+ SMC (Fig. 4A,B). Activation of Erk-1/2 in response to PDGF-BB was identical for both cell types at early time points and was actually significantly prolonged in synectin^−/− SMC (Fig. 4C,D), suggesting that synectin null SMC have the signaling machinery needed to activate migration in response to PDGF-BB. To test this directly, we measured migration of control and synectin null SMC in a scratch assay. There was no significant difference in migration of the two cell types in response to PDGF-BB or serum (Fig. 4E), demonstrating that SMC expression of synectin is not required for a normal migratory response to PDGF stimulation.

Fig. 4.

Effect of synectin on physiology of smooth muscle cells. A–D: Western blot and densitometry of basal platelet derived growth factor-B receptor (PDGFR-β) expression (A,B) and Erk induction (C,D, 50 ng/ml PDGF) in synectin^+/+ (black) and synectin^−/− (white) primary smooth muscle cells. E: Migration of synectin^+/+ and synectin^−/− in response to PDGF. Quantification of western blots and migration (n = 3) is based on results from three independent cell preparations. Means ± SEM.

Synectin Is Required for PDGF-B Stability in Arterial Endothelial Cells

Because the arterial defects in synectin^−/− mice appeared similar to those in Pdgfb^−/− mice and, because SMC function in these mice appeared normal, we next tested the hypothesis that PDGF-BB production was defective in cultured synectin null endothelial cells. Western analysis showed a dramatically reduced PDGF-BB expression in arterial, but not venous synectin^−/− endothelial cells (Fig. 5A). Restoration of synectin expression in synectin^−/− endothelial cells following an adenoviral transduction with a synectin-expressing construct, partially restored PDGF-BB expression, further linking synectin to PDGF-BB expression (Fig. 5A,B).

Fig. 5.

Effect of synectin on expression of platelet derived growth factor (PDGF)-BB) in arterial endothelial cells. A,B: Western blot (A) and densitometry (B) of PDGF-BB expression in arterial and venous synectin^−/− and synectin^+/+ endothelial cells following rescue of synectin expression by means of adenoviral transduction (n = 3). C: Quantification of pdgfb mRNA in synectin^−/− and synectin^+/+ arterial endothelial cells (n = 3). D: Western blot of PDGF-BB expression in arterial synectin^−/− endothelial cells following inhibition of proteosomal degradation with clasto-Lactacystin β-lactone (n = 3). *P < 0.05. E,F: Western blot (E) and densitometry (F) of HA-tagged-PDGF-BB and green fluorescent protein (GFP) in arterial synectin^−/− and synectin^+/+ endothelial cells following transduction with retroviral pdgfb-ha-IRES-gfp (n = 4, *P < 0.05).

To test whether reduced PDGF-BB expression in arterial cells was due to a transcriptional or translation defect, we used quantitative reverse transcriptase-polymerase chain reaction (RT-PCR) to measure Pdgfb mRNA levels in synectin^−/− and control endothelial cells. Surprisingly, mRNA levels were similar in both cell types (Fig. 5C), indicating that reduced PDGF-BB levels in synectin null endothelial cells are due to either a defect in translation or in protein stability. Treatment of synectin^−/− arterial endothelial cells with a proteosome degradation inhibitor, clasto-lactacystin β-lactone, led to a gradual increase in PDGF-BB protein levels (Fig. 5D) suggesting that increased degradation is responsible for the PDGF-BB production defect. We next assessed whether this was a global or specific defect by transducting arterial synectin^+/+ and synectin^−/− primary endothelial cells with a retroviral bicistronic construct encoding HA-tagged PDGF-BB and green fluorescent protein (GFP). Western blotting demonstrated equal GFP expression in both synectin^−/− and wild-type endothelial cells while PDGF-BB expression was dramatically reduced in synectin^−/− endothelial cells (Fig. 5E). Quantitative analysis of PDGF-BB vs. GFP protein expression shows dramatically reduced PDGF-BB protein levels in synectin null endothelial cells suggesting a specific-defect in PDGF-BB protein processing and/or degradation in these cells (Fig. 5F).

DISCUSSION

Disruption of synectin expression in mice resulted in decreased recruitment of mural cells to arterial vessels in the retina similar to what is observed after partial inactivation of PDGF-B or PDGFR-β (Enge et al., 2002; Tallquist et al., 2003). This in turn led to aberrant position of retinal arteries recapitulating in parts the PDGF-B loss of function phenotype. At the cellular level, the defect was traced to decreased PDGF-BB levels in arterial, but not venous endothelial cells in synectin^−/− mice, while SMC isolated from synectin^−/− mice expressed normal levels of PDGFR-β and migrated normally in response to PDGF-BB.

Mural cell coverage of arterial vessels governs vessel stability and PDGF-BB is a key factor promoting mural recruitment to the forming artery (Armulik et al., 2005). Indeed, absence of PDGF-BB is associated with the appearance of microaneurysms and hemorrhaging in endothelial-specific Pdgfb knockout mice (Enge et al., 2002). Cre-lox mosaic endothelial inactivation of the conditional Pdgfb allele due to cre-lox mediated recombination efficiency, illustrated that PDGF-B protein is a rate limiting factor for mural cell recruitment and expansion of the mural cell population in vascular development (Enge et al., 2002). Disruption of Pdgfrβ phenocopies Pdgfb knockouts, in particular the defective mural cell coverage accompanied by vessel dilation and microaneurysms (Hellstrom et al., 1999). Partial inactivation of PDGFR-β by replacement of one or several tyrosine residues in the intracellular domain of the receptor further illustrated that mural cell recruitment quantitatively depends on the cumulative phosphorylation of all receptor tyrosine residues upon PDGF-BB binding (Tallquist et al., 2003). Thus PDGF-BB levels and PDGFR-β signaling gage mural cell recruitment during vascular development.

Homozygous deletion of synectin results in reduced arterial density and synectin^−/− arterial endothelial cells demonstrated reduced responsiveness to PDGF-BB stimulation (Chittenden et al., 2006; Dedkov et al., 2007). Therefore, we considered the possibility that reduced PDGF-B responsiveness led to reduced arterial vasculature maturation, and possibly, involution of immature new vessels eventually accounting for the reduced arterial density seen in these mice. To evaluate this hypothesis, we used a retinal arterial maturation model that allows assessment of arterial morphogenesis in the early postnatal period. Surprisingly, SMC from synectin^−/− mice responded normally to PDGF-BB and demonstrated normal expression of PDGFR-β. However, we found a severe reduction in PDGF-BB levels selectively in arterial but not venous endothelial cells, when tested in vitro, suggesting that reduced PDGF-BB protein levels are the primary cause of defective arterial mural coverage. Unfortunately, low PDGF-BB expression prevented examining this question directly in flat-mounted retinas in vivo. Interestingly, this observation of reduced PDGF-BB expression specifically in arterial endothelial cells is in agreement with a selective defect in arterial but not venous morphogenesis in synectin^−/− mice (Chittenden et al., 2006). SMC in aorta (used for cell isolation) and retina (used for phenotypic analysis) originate from different embryonic compartments (reviewed in Yoshida and Owens, 2005) and could therefore possess different requirements for synectin during recruitment. However, with the exception of the liver, mural cells in all organs depend on PDGF-B signaling by means of PDGFR-β independent of their lineage origin (Andrae et al., 2008), suggesting that the observed deficiency in PDGF-BB protein levels in synectin^−/− cells may explain reduced arterial branching in various vascular beds in synectin^−/− mice.

The reduced PDGF-BB expression was traced to reduced protein levels despite normal transcription. This implies either reduced protein stability (increased degradation) or inefficient translation. Because PDGF-BB levels in synectin^−/− endothelial cells rose in the presence of a selective proteosomal inhibitor, clasto-lactacystin β-lactone, it is likely that increased protein degradation is the key event responsible for this phenotype.

While it not clear why synectin deficiency is associated with increased PDGF-BB degradation, synectin has recently been reported to be important for stability of IGF-IR and its protection from proteosomal degradation (Muders et al., 2006). Similar protection by synectin has also been shown for TGFβ type III receptor (Blobe et al., 2001) and the dopamine receptor D3R (Jeanneteau et al., 2004) suggesting that synectin may play a broader role in posttranslational regulation and processing of proteins. However, the necessity of synectin for protection from degradation is not global because following transduction with bicistronic GFP/PDGF-BB construct GFP but not PDGF-BB levels in synectin^−/− arterial endothelial cells are comparable to controls, suggesting the existence of a synectin-sensitive degradation pathway.

In summary, synectin deficiency results in increased degradation of PDGF-BB in arterial endothelial cells and reduced mural cell recruitment to newly forming arteries. This observation may explain the selective reduction in arterial morphogenesis observed in synectin knockout mice.

EXPERIMENTAL PROCEDURES

Synectin^−/− Mice

As described previously (Chittenden et al., 2006), these mice were generated from gene-trapped embryonic stem cells (129/SvEvBrd) from the Omnibank library (Lexicon Genetics) which have integrated a retroviral gene trap vector, VICTR25, into intron 1 of the synectin gene. After injection into C57BL/6J blastocysts and implantation into pseudopregnant females, resulting mice were backcrossed nine times. The tenth generation of C57BL/6-backcrossed mice was used for all studies.

Primary SMC and Endothelial Cell Isolation and Culture

Primary SMCs were isolated from adult mice using explant methods similar to those described previously (Fillinger et al., 1997) and used between passages 1 and 5. Briefly, after treatment of abdominal aortas with collagenase type II (Gibco) and removal of the adventitia, the medial layer was minced and cultured in growth media (DMEM, 20% fetal bovine serum, 1× nonessential amino acids, L-glutamine, penicillin/streptomycin, 50 μg/ml gentamicin sulfate, and 4 μg/ml Amphotericine B). Primary endothelial cells were isolated from the abdominal aorta and inferior vena cava of adult mice as described previously (Chittenden et al., 2006) and cultured between passages 1 and 5 in growth media (above) supplemented with heparin (Sigma) and Endothelial Mitogen (Biomedical Technologies, Stoughton MA). Rescue of synectin expression in synectin^−/− endothelial cells was achieved by transduction with an adenovirus encoding synectin and GFP or a control GFP virus as described previously(Chittenden et al., 2006).

PDGF-B Retrovirus

A plasmid encoding HA-tagged PDGF-B IRES GFP was generously provided by Dr. Peter Canoll, Columbia University, New York, NY (Assanah et al., 2006). This plasmid was co-packaged with a plasmid encoding vsv-G using GP293 cells. Endothelial cells were subjected to two rounds of transduction before being harvested for Western analysis.

Histological Analysis of Retinas

Retinas were isolated from synectin knockout and control neonatal mice as described previously (Enge et al., 2002). Briefly, eyes were dissected from neonatal mice and fixed overnight in 4% paraformaldehyde. After fixation, samples were washed with phosphate buffered saline (PBS). Anterior portions of the eye and any remaining hyaloid vessels were removed and the retina was collected, blocked, permeabilized overnight, and incubated with isolectin primary overnight. Retinas were then incubated overnight with SMA or NG2 primary overnight followed by incubation with fluorescent secondary antibodies for 2 hr at room temperature. Samples were post-fixed in 4% paraformaldehyde for 5 min, mounted, and imaged by laser scanning confocal microscopy.

Western Blotting

Cells were washed and collected in cold PBS with protease inhibitors, pelleted at 10,000 × g for 10 min and resuspended in RIPA with protease inhibitors. Equal amounts of protein were loading in 7.5% (PDGFR-β) or 12% (PDGF-BB) ReadyGels (Bio-Rad), separated and transferred at constant voltage. Membranes were blocked in 5% milk, TBS-T (Tris buffered saline, 0.05% Tween 20), incubated with primary antibody (1:1,000 anti-PDGFR-β, Cell Signaling; 1:200 anti PDGF-BB, Santa Cruz; 1:1,000 anti-HA, Roche; 1:1,000 anti-GFP, Santa Cruz) overnight at 4°C, followed by secondary antibody (1:2,000 anti-rabbit horseradish peroxidase [HRP] or anti-mouse HRP, Vector). Quantification was conducted using GeneTools (Syn-Gene, Cambridge, England).

Migration Assay

Primary SMC were plated in six-well culture plates and allowed to reach confluence. After overnight serum starvation in DMEM with 1% bovine serum albumin, the cell monolayer was wounded with a p20 pipette tip and stimulated with 50 ng/ml PDGF-BB. The subsequent gaps were imaged at baseline and after 8 hr. Migration distances were measured as previously described (Chittenden et al., 2006).

Quantitative RT-PCR

Total RNA was isolated from primary cells using the RNeasy Mini Kit (Qiagen) and reverse transcribed with High-Capacity cDNA Archive Kit (Applied Biosystems). Amplification of cDNA (20 ng per reaction) was conducted with gene-specific Taqman-based assays for PDGF-B (Mm00440678_m1) and b-actin (Mm00607939_s1) using Gene-Amp5700 sequence detection system.