Gap Junction Communication Mediates Transforming Growth Factor-β Activation and Endothelial-Induced Mural Cell Differentiation
During blood vessel assembly, endothelial cells recruit mesenchymal progenitors and induce their differentiation into mural cells via contact-dependent transforming growth factor-β (TGF-β) activation. We investigated whether gap junction channels are formed between endothelial cells and recruited mesenchymal progenitors and whether intercellular communication is necessary for endothelial-induced mural cell differentiation. Mesenchymal progenitors from Cx43−/− murine embryos and Cx43+/+ littermates were cocultured with prelabeled endothelial cells. Intracellular dye injection and dual whole-cell voltage clamp revealed that endothelial cells formed gap junction channels with Cx43+/+ but not Cx43−/− progenitors. In coculture with endothelial cells, Cx43−/− progenitors did not undergo mural cell differentiation as did Cx43+/+ cells. Stable reexpression of Cx43 in Cx43−/− cells (reCx43) restored their ability to form gap junctions with endothelial cells and undergo endothelial-induced mural cell differentiation. Cocultures of endothelial cells and either Cx43+/+ or reCx43 mesenchymal cells produced activated TGF-β; endothelial-Cx43−/− cocultures did not. However, Cx43−/− cells did produce latent TGF-β and undergo mural cell differentiation in response to exogenous TGF-β1. These studies indicate that gap junction communication between endothelial and mesenchymal cells mediates TGF-β activation and subsequent mural cell differentiation.
- blood vessel formation
- gap junctions
- endothelial cells
- mural cell differentiation
- transforming growth factor-β
Embryonic blood vessel formation begins with the coalescence of mesodermal progenitors and their differentiation into endothelial and blood cells, forming blood islands1 that fuse and define a primitive circulatory plexus.2 During branching and remodeling of the initial plexus that leads to a well-defined vascular network,3 endothelial tubes acquire a surrounding vessel wall of mural cells (pericytes or smooth muscle cells) via the secretion of platelet-derived growth factor-B that acts as a chemoattractant and mitogen for mural cell precursors.4–6 On contact with endothelial cells, the recruited progenitors are induced toward a mural cell fate,7 in a process mediated by the activation of transforming growth factor-beta (TGF-β).
Although TGF-β activation is not completely understood,8,9 it is clear that activated TGF-β plays a critical role in the induction of mural cell differentiation. Via TGF-β control elements10 or upregulation of serum response factor (SRF),11 TGF-β promotes coordinated transcriptional activation of cytoskeletal and contractile genes, including smooth muscle (SM)-α-actin, SM-γ-actin, SM22α, calponin, and SM-myosin heavy chain, needed for differentiated function.12 Mural cells not only modulate blood flow but also stabilize vessels6,13 and may sustain endothelial survival through the production of Ang-114 or VEGF-A.15
The exact interactions between endothelial and mesenchymal cells during vessel assembly that lead to TGF-β activation and mural cell differentiation are not well defined but are presumably complex given the number of genes implicated by mutation studies to regulate these processes during vascular development. Our studies aim to elucidate the role of one component of endothelial-mesenchymal cell interactions in vessel assembly—heterocellular communication via gap junctions.
Gap junctions are aggregates of intercellular channels16 that connect the cytoplasms of adjoining cells and allow the passage of second messengers, ions, and metabolites.17 Gap junction channels are composed of connexin (Cx) proteins, of which there are at least 20.18,19 The presence of gap junctions has been documented among and between vascular cells.20–22 Conflicting reports of their exact Cx composition suggest that Cx distribution varies among vascular beds and vessel types and is differentially regulated in response to injury and growth.
In adults, quiescent endothelial cells predominantly express Cx37 and Cx4023–27; proliferating or activated endothelial cells also express Cx43,27,28 as do endothelial cells during vascular development.29 Mural cell gap junctions are predominantly composed of Cx4326 but may also contain Cx40,24,30,31 Cx37, or Cx45. Although Cx45 is not a major component of gap junctions in adult vasculature, its expression is detected in vascular smooth muscle during development.32 Furthermore, Cx45-deficient mice die midgestation and exhibit abnormal vessel structures; endothelial tubes form but are not invested with mural cells.32 These studies suggest that gap junctions are needed for blood vessel formation and, more specifically, for endothelial-induced mural cell differentiation. However, a role for gap junctions in the induction of mural cell differentiation has not been demonstrated.
Herein, we demonstrate that gap junction–dependent communication occurs between endothelial cells and the mesenchymal cells they recruit; in the absence of such communication, endothelial-induced mural cell differentiation is impaired. Furthermore, the mechanism underlying gap junction–mediated endothelial-induced mural cell differentiation requires activation of TGF-β, which induces the expression of SRF and mural-specific genes in mesenchymal progenitors.
Materials and Methods
Endothelial and smooth muscle cells were isolated from bovine aortas33 and cultured,5 as described. Mural cell progenitors, phenotypically34 similar to 10T1/2 cells,5 were isolated as embryonic fibroblasts from undifferentiated mesenchyme of E14 mice deficient for Cx43 (Cx43−/−) or wild-type littermates (Cx43+/+), kindly provided by Dr Alan Lau, University of Hawaii, Honolulu, Hawaii) and cultured in DMEM with 10% FCS. Cx43−/− cells that were stably transfected with Cx43 cDNA (reCx43) were maintained in 6 μg/mL puromycin. All experiments were performed in DMEM/2% CS and repeated at least three times.
Cocultures of Endothelial Cells and Mural Cell Progenitors
Cocultures of mesenchymal cells and endothelial cells, which were labeled with PKH2635 that is incorporated into plasma and intracellular membranes, were established in an under-agarose system5 or via 1:1 simultaneous plating. In some experiments, neutralizing anti–TGF-β antibodies (10 μg/mL; kindly provided by Genzyme) were incorporated into the agarose or culture media before cell plating.
Intracellular Microinjection of Lucifer Yellow Dye
PKH26-labeled endothelial cells were cocultured 1:1 with mural cell progenitors and microinjected36 using glass micropipettes (Eppendorf) containing 10% (wt/vol) Lucifer Yellow dye in 300 mmol/L LiCl. The total number of fluorescent cells per injection was counted 2 minutes after injection and photographed. Microinjection was performed on 10 to 20 cell pairs of each type of coculture in each of four to five separate experiments.
Dual Whole-Cell Voltage Clamp
PKH26-labeled endothelial cells and mesenchymal cells were coplated onto glass coverslips, and heterocellular and homocellular pairs were subjected to dual whole-cell voltage clamp. Patch-type microelectrodes were fabricated from 1.2-mm filament glass and filled with (in mmol/L) KCl 135, TEACl 10, CaCl2 0.5, MgCl2 3, glucose 5, HEPES 10, EGTA 10, and Na2ATP 5 (320 mOsm; pH 7.2). After the dual whole-cell voltage clamp configuration was achieved, cells were alternately stepped between 0 and −10 mV to determine macroscopic junctional conductance (gj).37,38 Pairs with no obvious macroscopic coupling (gj<0.2 nS) were evaluated for individual channels by holding one cell at 0 mV and the other at either −40 or +40 mV, thereby establishing a transjunctional voltage difference of +40 or −40 mV, respectively, relative to the cell at 0 mV.
PKH-labeled endothelial cells and mural cell progenitors were cocultured in the underagarose assay,5 sometimes in the presence of anti–TGF-β (10 μg/mL), and then fixed and immunostained for SM-α-actin, SM-myosin heavy chain, calponin, SM22α, as described,5 or Cx proteins using anti-Cx43 (1:1000) generously provided by Dr Elliott Hertzberg (Albert Einstein College of Medicine, New York, NY) or anti-Cx45 (Chemicon; 1:1000). Antibody-antigen complexes were visualized using the Vectastain Elite ABC Kit.
Western Blot Analysis
Endothelial cells and mural cell progenitors were cultured alone or in 1:1 coculture with or without anti–TGF-β (10 μg/mL) or with 0 to 2.5 ng/mL TGF-β1. Total protein was isolated,39 and 10 μg of protein from solo cultures and 20 μg from cocultures were subjected to Western analysis. The following primary antibodies were used: SM-α-actin, 1:1000; SM-myosin heavy chain, 1:200; SM22α, 1:750; calponin, 1:500; Cx43, 1:1000; Cx45, 1:1000; SRF, 1:1000, followed by HRP-linked secondary antibodies (1:10 000). Antibody-antigen complexes were revealed using ECL-Plus, and quantification was performed using a digital imaging system.
Total RNA was isolated from endothelial and mesenchymal cells. First-strand cDNA synthesis was performed using 1 μg of total RNA, Cx-specific primers (detailed in the online data supplement, available at http://www.circresaha.org), and SUPERSCRIPT II RNase H− reverse transcriptase (RT) (1 μL=200 U; GibcoBRL).
Cell-Cell Adhesion Assay
Cx43+/+ and Cx43−/− mesenchymal cells were cultured overnight and rinsed with PBS, and then endothelial cells were plated on top and incubated for 0 to 240 minutes; unattached cells were aspirated off, and wells were rinsed. Remaining adherent mesenchymal and endothelial cells were trypsinized and counted using a Coulter counter. For each time point, mesenchymal cells were cultured without endothelial cells and similarly trypsinized and counted; this value was subtracted from the total number of cells per experimental well to determine the number of adherent endothelial cells. Experimental values represent mean±SD of triplicates per time point from one of five experiments performed.
ELISA Assay for TGF-β1
An ELISA assay kit (DRG International) was used to measure the levels of TGF-β protein (latent and activated) produced by endothelial cells, mesenchymal cells, and 48-hour cocultures of endothelial and Cx43+/+, Cx43−/−, or reCx43 mesenchymal cells. Conditioned media were collected and assayed according to the manufacturer’s instructions. The absorbance of each sample was read at 450 nm within 10 minutes.
Bioassay for Activated TGF-β
A bioassay40 was used to measure levels of endogenously activated TGF-β. Mink lung epithelial cells (MLECs), expressing a truncated PAI-1 promoter linked to the luciferase reporter (kindly provided by Dr Daniel Rifkin, New York University, New York, NY), were incubated for 60 minutes with conditioned media from 48-hour cocultures of endothelial and mesenchymal cells. Unconditioned media containing 0 to 100 pg/mL TGF-β1 was similarly assayed to create a linear standard curve. In control experiments, conditioned media from cocultures was preincubated with neutralizing anti–TGF-β (10 μg/mL) before addition to MLEC. Cells were processed for the luciferase reporter assay (Promega).
An expanded Materials and Methods section can be found in the online data supplement, available at http://www.circresaha.org.
To determine whether gap junction communication is required for endothelial-induced mural cell differentiation, we needed similarly derived mesenchymal progenitors that form or fail to form functional gap junctions with endothelial cells on heterocellular contact. We used mesenchymal cells derived from Cx43−/− or Cx43+/+ murine embryos34; assessment of their ability to form gap junctions with endothelial cells is detailed below.
Endothelial Cells Communicate With Cx43+/+, but not Cx43−/−, Progenitors
Intracellular Microinjection of Lucifer Yellow
PKH26-labeled endothelial cells were cocultured with mesenchymal cells and then injected with Lucifer Yellow. In endothelial-Cx43+/+ cell pairs, dye transferred readily to neighboring Cx43+/+ cells (Figure 1A) and other endothelial cells. Lucifer Yellow transferred equally well to endothelial cells when adjacent Cx43+/+ cells were injected, and Cx43+/+ cells exhibited homocellular dye transfer (not shown). In contrast, in endothelial-Cx43−/− cell pairs, injected Lucifer Yellow dye did not transfer between cells regardless of which cell type was injected (Figure 1B). Homocellular dye transfer was not observed between Cx43−/− cells.
Dual Whole-Cell Voltage Patch Clamp
Because gap junction channels can display charge and size selectivity41,42 against large negatively charged molecules (ie, Lucifer Yellow ≈450 Da, 2-charge), cell pairs that did not transfer dye could still transfer smaller or positively charged molecules. Thus, we used dual whole-cell voltage clamp to measure the extent of electrical coupling between endothelial and Cx43−/− cells. No channel activity was observed in endothelial-Cx43−/− cell pairs (n=8) (Figure 1D). In contrast, all of the endothelial-Cx43+/+ cell pairs were coupled (n=4), with junctional conductances ranging from 0.2 to 6 nS. Single-channel events were observed in two pairs; most of the events had amplitudes of 50 to 60 pS, although events of ≈100 pS were also seen (Figure 1C). Both channel amplitudes are exhibited by cells expressing Cx43.38,43 No voltage-dependent uncoupling was observed at this transjunctional driving force, suggesting that Cx45-comprised channels were not present.34
Cx43 Expression Between Endothelial and Mesenchymal Cells
To determine which Cx genes were expressed in the endothelial and mesenchymal cells, among those reportedly expressed in the vasculature, we performed RT–polymerase chain reaction (PCR) analyses using Cx-specific primers. Endothelial cells expressed mRNA for Cx37, Cx40, Cx43, and Cx45. Cx43+/+ mesenchymal cells expressed Cx43 and Cx45. Cx43−/− mesenchymal cells expressed only Cx45 (Figure 1E), which is consistent with electrophysiological studies demonstrating the presence of Cx45-like gap junction channels34 in Cx43−/− cells. Such channels were not evident in the present study, and expression of Cx45 in Cx43−/−, Cx43+/+ or endothelial cells was not detectable. However, immunocytochemical analyses did reveal the presence of Cx43 among and between endothelial and Cx43+/+ mesenchymal cells (not shown), which is consistent with electrophysiological studies, in which endothelial and mesenchymal cell pairs displayed single-channel properties characteristic of Cx43-containing channels.
Gap Junctions Mediate Endothelial-Induced Mural Cell Differentiation
Because endothelial cells formed functional gap junction channels with Cx43+/+ mesenchymal cells but not with similarly derived cells that lacked Cx43 (Cx43−/−), we used these two cell types to evaluate whether heterocellular gap junction communication per se is necessary for endothelial-induced mural cell differentiation. Mesenchymal cells were cocultured with PKH26-labeled endothelial cells and then immunostained for mural cell markers.
As expected,5 wild-type mesenchymal cells exhibited increased expression of SM-α-actin and induction of SM-myosin heavy chain on contact with endothelial cells (Figure 2A). Mural cell differentiation was not observed in mesenchymal cells not contacting endothelial cells (Figure 2A, left) nor in mesenchymal cells cultured in conditioned media from endothelial cells (not shown). Data shown were generated using BAECs, but similar results were obtained using rodent microvascular endothelial cells and HUVECs (not shown). Western analyses confirmed an upregulation of SM-α-actin, ranging from 5- to 10-fold among four experiments, and induction of calponin (not shown), SM22α, and SM-myosin heavy chain (Figure 2B) in Cx43+/+ cells cocultured with endothelial cells. In contrast, Cx43−/− cells exhibited no increase in SM-α-actin expression on contact with endothelial cells and no induction of SM22α, SM-myosin heavy chain, or calponin (not shown), as evidenced by immunocytochemistry (Figure 3A) and Western analyses (Figure 3B).
To determine whether the observed lack of endothelial-induced mural cell differentiation of Cx43−/− progenitors was attributable to their inability to adhere to endothelial cells, we measured cell-cell adhesion in endothelial-mesenchymal cell cocultures. The ability of endothelial cells to adhere to Cx43+/+ versus Cx43−/− mesenchymal cells was equivalent and linear for up to 90 minutes, at which time the total number of adherent cells was maximal (mean±SD from three experiments was 27 520±226 and 26 133±1150 adherent endothelial cells, respectively).
Reexpression of Cx43 in Cx43−/− Mesenchymal Cells
To rule out the possibility that a factor other than a deficiency of Cx43 in Cx43−/− mesenchymal cells accounted for their lack of mural cell differentiation in response to endothelial cells, we performed similar experiments with Cx43−/− cells stably transfected with full-length Cx43 cDNA (reCx43; provided by Dr Alan Lau).
The expression level and cellular distribution of Cx43 protein in reCx43 cells was examined. Western analyses revealed that total Cx43 level in reCx43 cells was similar to that of Cx43+/+ cells. Wild-type cells exhibited a predominant 42-kDa band, representative of unphosphorylated Cx43 protein, and an equal distribution of less-abundant 44- and 46-kDa bands that reflect different phosphorylation states of Cx4344 (Figure 4A). reCx43 cells, in contrast, exhibited predominantly 42- and 44-kDa forms of Cx43. This phenomenon was observed in other cells constitutively expressing Cx43,45 and its functional significance, if any, is not known.
Immunocytochemical analyses demonstrated appropriate localization of Cx43 in reCx43 cells in regions of cell-cell contact (Figure 4B). A typical perinuclear distribution of Cx43 was also observed in Cx43+/+ and reCx43 mesenchymal cells, although higher in reCx43 cells, which is characteristic of constitutive expression of gap junction proteins and not known to affect function.45
Restoration of Gap Junction Communication
To determine whether the stable reexpression of Cx43 in Cx43−/− mesenchymal cells would restore the ability to form functional gap junctions with endothelial cells, cocultures were established and subjected to Lucifer Yellow microinjection. In each of three experiments, injected endothelial cells exhibited dye transfer to reCx43 cells (Figure 4C) and vice versa. Furthermore, Cx43 protein was localized to sites of endothelial-reCx43 cell contact (not shown).
Restoration of Endothelial-Induced Mural Cell Differentiation
To determine whether the reestablishment of gap junction communication restores endothelial-induced mural cell differentiation, reCx43 cells and endothelial cells were cocultured and examined for mural cell markers. reCx43 cells exhibited an upregulation of SM-α-actin expression (5- to 10-fold) and induction of calponin (not shown), SM22α, and SM-myosin heavy chain proteins, as revealed by immunocytochemical (Figure 5A) and Western analyses (Figure 5B).
Collectively, these experiments demonstrate that heterocellular gap junction communication is necessary for endothelial-induced mural cell differentiation. Our previous studies demonstrated that TGF-β is activated upon endothelial-mesenchymal cell contact and mediates mural cell differentiation. Thus, we aimed to determine whether TGF-β similarly mediates endothelial-induced mural cell differentiation in these studies.
TGF-β Mediates Endothelial-Induced Mural Cell Differentiation
Endothelial cells were cocultured with Cx43+/+ or reCx43 mesenchymal cells in the presence of neutralizing anti–TGF-β. Total protein was isolated and subjected to Western analyses, which revealed that neutralization of endogenously activated TGF-β suppressed the upregulation of SM-α-actin, SM22α, and calponin by ≈85% to 95% (Figure 6).
Cx43−/− Cells Can Differentiate Into Mural Cells
Because TGF-β mediates endothelial-induced mural cell differentiation and Cx43−/− cells were not induced to become mural cells in response to endothelial cells, we investigated whether Cx43−/− mesenchymal cells were responsive to exogenous, activated TGF-β. We found that TGF-β upregulated SM-α-actin and induced other mural cell–specific proteins including SM22α (Figure 7A). Via Western blot analysis, we determined that this response was linear up to 1 ng/mL TGF-β1 (Figure 7B) and equal in magnitude to that of Cx43-containing mesenchymal cells cocultured with endothelial cells (Figures 2 and 5⇑) (≈5-fold upregulation of SM-α-actin).
Gap Junction Communication Mediates TGF-β Activation
Because Cx43−/− mesenchymal cells underwent mural cell differentiation in response to exogenous TGF-β but not in coculture with endothelial cells, we aimed to determine whether the production or activation of TGF-β was suppressed in endothelial-Cx43−/− cocultures. We found similar total levels of TGF-β (latent plus activated) in endothelial-Cx43+/+, endothelial-Cx43−/−, and endothelial-reCx43 cocultures of 1.47±0.03, 1.51±0.28, and 1.52±0.19 ng/mL, respectively (Figure 8A). Similar measurements of media from solo cultures revealed that all cell types produced similar amounts of TGF-β on a per-cell basis (≈1.5 ng/mL).
We measured levels of endogenously activated TGF-β in conditioned media from endothelial-Cx43+/+ and -reCx43 cocultures and found that both induced transcriptional activation of the PAI promoter and luciferase activity, indicating the presence of endogenously activated TGF-β (4.0±0.3 and 4.2±0.4 pg/mL, respectively; Figure 8B); conditioned media from endothelial-Cx43−/− cocultures did not. Incubation of coculture-conditioned media with neutralizing anti–TGF-β before exposure to MLEC suppressed activation of the reporter gene, demonstrating that luciferase activity was induced via endogenously activated TGF-β.
Based on previous studies,11 we proposed that TGF-β activation mediated mural cell differentiation via the upregulation of SRF. To test this, we examined the expression of SRF in endothelial-mesenchymal cocultures and found, as expected, that SRF was induced only in cocultures that enabled heterocellular communication, TGF-β activation, and mural cell differentiation (Figure 8C).
During blood vessel assembly, mural cell differentiation is initiated in response to contact-dependent interactions with endothelial cells.5,7 In the present studies, we demonstrate for the first time that functional gap junction channels are established between endothelial cells and the mural cell progenitors that they recruit and are necessary for endothelial-induced mural cell differentiation. We additionally determined that the mechanism by which gap junctions mediate endothelial-induced mural cell differentiation involves the activation of TGF-β, which then induces SRF and mural cell–specific gene expression.
The mesenchymal cells that did not form gap junctions with endothelial cells (Cx43−/−) did not undergo endothelial-induced differentiation toward a mural cell phenotype. Of note, the Cx43−/− cells express low levels of Cx45, although this expression was not evident as channel activity and not adequate to support the formation of functional heterocellular gap junctions with endothelial cells. Importantly, these cellular processes were restored by stable reexpression of Cx43. Furthermore, Cx43−/− mesenchymal cells exhibited similar ability to adhere to endothelial cells compared with Cx43+/+ cells and could be induced toward a mural cell fate in response to exogenous, activated TGF-β. Thus, it was the specific lack of gap junction proteins in Cx43−/− cells and the consequent inability to establish functional heterocellular gap junctions with endothelial cells that prevented activation of TGF-β and subsequent mural cell differentiation.
We demonstrated that gap junctions composed of Cx43 were sufficient to mediate endothelial-induced mural cell differentiation. Whether other connexins are capable of supporting such induction or are, indeed, more relevant to neovascularization in specific circumstances remains to be determined. Cx43 is not likely to be the only gap junction protein capable of mediating these cellular processes. In fact, embryos that lack Cx43 survive through gestation and exhibit some degree of mural cell differentiation46; thus, other Cx proteins expressed in vascular cells (Cx45, Cx40, and Cx37) must be capable of serving a similar function. The phenotype of Cx45-deficient mice suggests that this Cx plays a critical role in early vascular development.32 Thus, it is possible that Cx43 and Cx45 fulfill similar roles at different stages of vascular development and perhaps under different conditions postnatally.
Interestingly, of the four connexins expressed in vascular cells, Cx43 is the most versatile. The homomeric/homotypic channels formed by Cx43 display the least selectivity (allowing passage of a broad spectrum of small molecules), the greatest permeability, and the highest degree of regulation of these properties by intracellular signaling cascades. Our model system is conducive to sorting out which connexins and mutants thereof will support induction of mural cell differentiation, but clearly extensive experimentation will be required. Nonetheless, we have demonstrated that gap junction communication per se is necessary to support endothelial-induced differentiation of mural cells via signaling that leads to TGF-β activation. It is conceivable that continued heterocellular communication between endothelial cells and new mural cells enables the transfer of soluble signals from the endothelium or blood circulation to mediate the subsequent differentiation of additional mural cell layers in large-caliber vessels. Such signals may include retinoic acid, which is generated by endothelial cells and can upregulate mural cell–specific proteins in mesenchymal cells (Hirschi laboratory, unpublished data, 2003).
It is not yet clear how gap junction communication regulates TGF-β activation. Gap junctions are known to mediate intercellular exchange of ions and second messenger molecules. Such communication may be necessary for exchange of the signal leading to intracellular cleavage of TGF-β or extracellular cleavage of the latency-associated peptide. Importantly, because Cx43−/− and Cx43+/+ cells adhere equally well to progenitor cells, it is not likely that gap junctions merely function to bring the two distinct cell types close enough together for a physical interaction, such as that needed for the binding of mannose-6-phosphate residues of latency-associated peptide and their receptors. Furthermore, despite recent reports that Cx43 can modulate cell behavior in the absence of channel function,47 we found that expression of Cx43 without contact and junction formation is insufficient for TGF-β activation and endothelial-induced mural cell differentiation.
In summary, our studies demonstrate that functional gap junction channels are established between endothelial and mesenchymal cells that they recruit and that heterocellular gap junction communication is necessary for contact-dependent activation of TGF-β, which mediates endothelial-induced mural cell differentiation. Thus, we have defined a critical step in the process of blood vessel assembly, which may provide an additional target for the control of neovascularization in vascular therapies. These findings not only extend our present knowledge of the process of blood vessel formation but may also provide insights into biological processes involving heterocellular interactions in other organ systems.
This work was supported by AHA-National SDG 9930054N, USDA 6250-51000-033, and NIH R01-HL61408 grants to K.K.H. K.D.H. is supported by USDA grant 6250-21520-041 and NIH grant R01-GM57427. J.M.B. is supported by NIH RO1-HL58732-1 and AHA-National GIA 0050020N.
Original received October 2, 2002; resubmission received July 8, 2003; revised resubmission received August 5, 2003; accepted August 5, 2003.
Doetschman TA, Gossler A, Kemler R. Blastocyst-derived embryonic stem cells as a model for embryogenesis. In: Feichtingen W, Kemeter P, eds. Future Aspects in Human In Vitro Fertilization. Berlin, Germany: Springer-Verlag; 1987: 187–195.
Risau W, Sariola H, Zerwes HG, Sasse J, Ekblom P, Kemler R, Doetschman T. Vasculogenesis and angiogenesis in embryonic stem cell-derived embryoid bodies. Development. 1988; 102: 471–478.
Lindahl P, Johansson B, Leveen P, Betsholtz C. Pericyte loss and micro-aneurysm formation in platelet-derived growth factor B-chain-deficient mice. Science. 1997; 277: 242–245.
Hirschi KK, Rohovsky SA, D’Amore PA. PDGF, TGF-β and heterotypic cell-cell interactions mediate the recruitment and differentiation of 10T1/2 cells to a smooth muscle cell fate. J Cell Biol. 1998; 141: 805–814.
Hirschi KK, Rohovsky SA, Smith SR, Beck LH, D’Amore PA. Endothelial cells modulate the proliferation of mural cell precursors via PDGF-BB and heterotypic cell contact. Circ Res. 1999; 84: 298–305.
Dennis PA, Rifkin DB. Cellular activation of transforming growth factor β requires binding to the cation-independent mannose 6-phosphate/insulin-like growth factor type II receptor. Proc Natl Acad Sci U S A. 1991; 88: 580–584.
Kojima S, Nara K, Rifkin DB. Requirement for transglutaminase in the activation of latent transforming growth factor-β in bovine endothelial cells. J Cell Biol. 1993; 121: 439–448.
Hautmann MB, Madsen CS, Owens GK. A transforming growth factor-β (TGF-β) control element drives TGF-β-induced stimulation of smooth-muscle alpha-actin gene expression in concert with 2 CARG elements. J Biol Chem. 1997; 272: 948–956.
Landerholm TE, Dong X-R, Lu J, Belaguli NS, Schwartz RJ, Majesky MW. A role for serum response factor in coronary smooth muscle differentiation from proepicardial cells. Development. 1999; 126: 2053–2062.
Zachary I. Signaling mechanisms mediating vascular protective actions of vascular endothelial growth factor. Am J Physiol. 2001; 280: C1375–C1386.
Revel J-P, Karnovsky M. Hexagonal array of subunits in intercellular junctions of the mouse heart and liver. J Cell Biol. 1967; 33: C7–C12.
Sweet E, Abraham EH, D’Amore PA. Functional evidence of gap junctions between capillary endothelial cells and pericytes in vitro. Invest Ophthalmol Vis Sci. 1988; 29: 109a.Abstract.
Little TL, Xia J, Duling BR. Dye tracers define differential endothelial and smooth muscle coupling patterns within the arteriolar wall. Circ Res. 1995; 76: 498–504.
Bastide B, Neyses L, Ganten D, Paul M, Willecke K, Traub O. Gap junction protein connexin40 is preferentially expressed in vascular endothelium and conductive bundles of rat myocardium and is increased under hypertensive conditions. Circ Res. 1993; 73: 1138–1149.
Hennemann H, Suchyna T, Lichtenberg-Frate H, Jungbluth L, Dahl E, Schwarz J, Nicholson BJ, Willecke K. Molecular cloning and functional expression of mouse connexin40, a second gap junction gene preferentially expressed in lung. J Cell Biol. 1992; 117: 1299–1310.
Bruzzone R, Haefliger J-A, Gimlich RL, Paul DL. Connexin40, a component of gap junctions in vascular endothelium, is restricted in its ability to interact with other connexins. Mol Biol Cell. 1993; 4: 7–20.
Gabriels JE, Paul DL. Connexin43 is highly localized to sites of disturbed flow in rat aortic endothelium but connexin37 and connexin40 are more uniformly distributed. Circ Res. 1998; 83: 636–643.
Kruger O, Plum A, Kim J-S, Winterhager E, Maxeiner S, Hallas G, Kirchhoff S, Traub O, Lamers WH, Willecke K. Defective vascular development in connexin45-deficient mice. Development. 2000; 127: 4179–4193.
Martyn KD, Kurata WE, Warn-Cramer BJ, Burt JM, TenBroek E, Lau AL. Immortalized connexin43 knockout cell lines display a subset of biological properties associated with the transformed phenotype. Cell Growth Differ. 1997; 8: 1015–1027.
Cottrell GT, Burt JM. Heterotypic gap junction channel formation between heteromeric and homomeric Cx40 and Cx43 connexons. Am J Physiol. 2001; 281: C1559–C1567.
He DS, Jiang JX, Taffet SM, Burt JM. Formation of heteromeric gap junction channels by connexins 40 and 43 in vascular smooth muscle cells. Proc Natl Acad Sci U S A. 1999; 96: 6494–6500.
Ruch RJ, Bonney WJ, Sigler K, Guan X, Matesic D, Schafer LD, Dupont E, Trosko JE. Loss of gap junctions from DDT-treated rat liver epithelial cells. Carcinogenesis. 1994; 15: 301–306.
Veenstra RD, Wang HZ, Beblo DA, Chilton MG, Harris AL, Beyer EC, Brink PR. Selectivity of connexin-specific gap junctions does not correlate with channel conductance. Circ Res. 1995; 77: 1156–1165.
Lampe PD, TenBroek EM, Burt JM, Kurata WE, Johnson RG, Lau AF. Phosphorylation on connexin43 on serine368 by protein kinase c regulates gap junctional communication. J Cell Biol. 2000; 149: 1503–1512.
Musil LS, Cunningham BA, Edelman GM, Goodenough DA. Differential phosphorylation of the gap junction protein connexin43 in junctional communication-competent and -deficient cell lines. J Cell Biol. 1990; 111: 2077–2088.
Hirschi KK, Xu C, Tsukamoto T, Sager R. Gap junction genes Cx26 and Cx43 individually suppress the cancer phenotype of human mammary carcinoma cells and restore differentiation potential. Cell Growth Differ. 1996; 7: 861–870.
Reaume AG, de Sousa PA, Kulkarni S, Langille BL, Zhu D, Davies TC, Juneja SC, Kidder GM, Rossant J. Cardiac malformation in neonatal mice lacking connexin 43. Science. 1995; 267: 1831–1834.