By continuing to browse this site you agree to us using cookies as described in About Cookies
Notice: Wiley Online Library will be unavailable on Saturday 7th Oct from 03.00 EDT / 08:00 BST / 12:30 IST / 15.00 SGT to 08.00 EDT / 13.00 BST / 17:30 IST / 20.00 SGT and Sunday 8th Oct from 03.00 EDT / 08:00 BST / 12:30 IST / 15.00 SGT to 06.00 EDT / 11.00 BST / 15:30 IST / 18.00 SGT for essential maintenance. Apologies for the inconvenience.
In humans and mice, the family of platelet-derived growth factors (PDGFs) consists of four different PDGF ligands, PDGFA–D, and has been shown to drive cellular responses including cell proliferation, survival, and migration. They function as secreted, disulfide-bonded homodimers, but PDGFA and B can also form functional heterodimers. Together with their cognate tyrosine kinase receptors PDGFRα and PDGFRβ, which form dimers upon ligand binding, the functional PDGF signaling network is composed (Heldin and Westermark,1999). PDGFRα knockout studies showed that the receptor is essential for embryonic development. Homozygotes die during embryonic development, exhibiting incomplete cephalic closure, subepidermal blebs, alterations in vertebrae, ribs, and sternum (Soriano,1997).
Furthermore, PDGF signaling is necessary for vascular development (Magnusson et al.,2007). During vascular development, contractile mesenchymal cells which regulate vessel diameter and consequently blood flow, known as mural cells, are recruited to coat nascent vessels. The two major classes of mural cells are vascular smooth muscle cells (vSMCs) and pericytes. The development of vascular mural cells requires PDGFB/PDGFRβ signaling. Mice deficient for either PDGFB or PDGFRβ die perinatally owing to extensive hemorrhaging and lack numerous blood vessels or the vessels are incompletely covered by mural cells (Leveen et al.,1994; Soriano,1994). Recently, Looman and colleagues showed, that expression of a mutated, autoactivating form of PDGFRβ leads to defective blood vessel formation in the placenta (Looman et al.,2007), indicating that PDGF signaling needs to be closely controlled.
In humans, autoactivating mutations of PDGFRα have been found in a large number of gastrointestinal stromal tumors (GISTs; Heinrich et al.,2003). To test the impact of such a mutation for embryonic development, mice carrying the human PDGFRα D842V mutant in the endogenous ROSA26 locus (R26hPDGFRαPM; MGI accession no. 3814559, Gt(ROSA)26Sortm1(PDGFRA*)Hsc; Moenning et al.,2009) were mated with Sox2Cre mice, inducing Cre-mediated recombination in all cells of the epiblast (Hayashi et al.,2002). Embryos expressing the mutated PDGFRα die during midgestation. Some embryos display failure of embryonic turning and a general growth retardation. Other mutant animals suffer from multiple hemorrhages. We show that mutant animals have enlarged blood vessels and lack pericytes around the aorta. In general, proliferation in these embryos is reduced and apoptosis is enhanced. This is the first report showing that a subtle alteration in PDGFRα leads to a dominant lethal phenotype.
Mice With Ubiquitous Expression of an Activated PDGFRα Die at Midgestation
We speculated that ubiquitous expression of a constitutive active PDGFR molecule would be detrimental for embryonic development. This was further substantiated by the fact that Looman et al. failed in an attempt to establish a transgenic mouse line expressing an autoactivated PDGFRβ allele. There, chimeric embryos died owing to an early placental failure resulting from insufficient labyrinth formation and disorganized fetal blood vessels in the placenta (Looman et al.,2007). So we decided to breed the transgenic line R26hPDGFRaPM developed by us to the Sox2Cre (Hayashi et al.,2002) transgenic mice to achieve epiblast specific expression of the R26hPDGFRaPM transgene.
No Sox2Cre;R26hPDGFRaPM double transgenic animals were born, indicating that ubiquitous expression of mutant PDGFRα is not compatible with full embryonic development. Analysis of embryonic stages revealed that embryogenesis proceeds normal up to embryonic day (E) 8.5 (n = 12) in Sox2Cre;R26hPDGFRaPM double transgenic animals. However, afterward, overall developmental retardation becomes apparent. A total of 22% (n = 9) of Sox2Cre;R26hPDGFRaPM embryos fail to turn and display a balloon like pericardium (Fig. 1C compare to 1A). In these embryos, at E10.5 the yolk sac retained the typical immature honeycomb pattern (Fig. 1F compare to 1D,E) and failed to reorganize into the normal hierarchical array of large and small vessels (Fig. 1I compare to G,H). At E10.5, the remaining embryos exhibit growth retardation and massive hemorrhage in the trunk and the pericardial cavity (Fig. 1B) as well as an undulated neural tube (shown later). At E12.5, only dead embryos being in the process of resorption could be detected (not shown). In situ hybridization using a human PDGFRα probe showed a uniform expression in tissues of mice carrying the Sox2Cre;R26hPDGFRaPM allele (Fig. 1K,L). Western Blot analysis showed that the transgene resulted in very moderate overexpression of the transgenic PDGFRα protein (Fig. 1M, arrow). The smaller than expected size (130 kDa vs. 185 kDa) suggests that the mutant receptor might not be glycosylated correctly (Bejcek et al.,1993).
Cell Proliferation and Cell Death in Mutant Embryos
PDGFs were identified as mitogenic factors but have also been shown to act as survival factors (Soriano,1997). Hence, it is possible that the defects observed in the mutant embryos might be owing to abnormal cell proliferation, cell death, or both. First, apoptotic cell death was examined at E9.5. Here, considerable amount of apoptosis was detected in Sox2Cre;R26hPDGFRaPM embryos when compared to littermate controls (Fig. 2B compare to 2A). Specifically, the mesenchyme around the dorsal aorta (Fig. 2B, arrows) was affected in Sox2Cre;R26hPDGFRaPM animals. Next, embryos were labeled with 5-bromodeoxyuridine (BrdU) to detect proliferating cells (Fig. 2C,D). Sections of the grossly retarded Sox2Cre;R26hPDGFRaPM embryos which had not turned displayed the most significant decrease of proliferating cells in all tissues (Fig. 2D). As expected, the difference in the rate of proliferation was not so pronounced in Sox2Cre;R26hPDGFRaPM embryos which appeared only slightly retarded at E10.5 (data not shown). Taken together, expression of the mutant PDGFRα seems to impact on both, proliferation and apoptosis.
Neural Tube Defects
Because the embryos displayed neural tube defects, whole-mount in situ analyses detecting sonic hedgehog (shh) were performed. Shh signaling regulates dorsoventral patterning of the neural tube by repressing genes that are associated with dorsal and lateral fates and by activating genes associated with ventral fates (Briscoe and Ericson,2001). Here, no differences in spatiotemporal expression between wild-type and Sox2Cre;R26hPDGFRaPM embryos could be detected (Fig. 3A,B). Morphologically, the mutants displayed an undulating and kinked neural tube (Fig. 3C,D). However, the neuroepithelium itself seemed to be regular. Vibratome sections through the mid-trunk region revealed that the neural folds failed to fold up and oppose each other in the dorsal aspect of the tube. Of note, the curvature of the adjacent epithelium of the mutants is concave compared to wild-type controls, which display a convex form (Fig. 3E,F, arrows), suggesting a lack of mesenchymal cells in this area. Whole-mount analysis using brachyury (T) as marker for mesenchyme showed no difference in mesenchyme formation (Supp. Fig. S1, which is available online). The lack in cell number in the mesenchyme can be explained by the altered proliferation and apoptosis kinetics shown before.
Next, we decided to analyze the genesis of somites in more detail. Whole-mount in situ analysis using uncx4.1 a marker for posterior somites (Fig. 4A,B) and Tbx18 (Fig. 4C,D) a marker for anterior somites did not reveal any differences between control and Sox2Cre;R26hPDGFRaPM embryos, suggesting that there was normal development of the anterior–posterior axis of somites. However, when we used HeyL as general marker for somites, Sox2Cre;R26hPDGFRaPM embryos exhibited a reduced, disorganized expression in the anterior, older somites (Fig. 4F) whereas wild-types showed a strong and broad expression (Fig. 4E). In addition, at E9.5, HeyL is also expressed in the vSMCs surrounding the arteries (Fig. 4G, arrows; Leimeister et al.,2000). In contrast, no HeyL expression could be detected in Sox2Cre;R26hPDGFRaPM in area of vSMCs (Fig. 4H, arrows). So, HeyL expression was disturbed in older somites and lost in the area where blood vessels should be forming, indicating a defect in blood vessel formation.
Vascular Defects in Sox2Cre;R26hPDGFRaPM Embryos
Because the lack of HeyL gene expression together with the macroscopic analyses, which showed enlarged pericardium, partially enlarged ventricles, and hemorrhage pointed to a defect in the blood vessels, we decided to analyze the vasculogenesis and the angiogenesis in mutant animals. Histological sections at E9.5 showed, that the dorsal aorta in mutant was enlarged (Fig. 5A,B, da). To confirm this, whole-mount antibody staining was performed using anti–platelet endothelial cell adhesion molecule (PECAM) -1 antibody. PECAM is a major constituent of the endothelial cell intercellular junction (Albelda et al.,1990) where up to 106 PECAM-1 molecules are concentrated (Newman,1994). While whole-mount staining revealed successful vasculogenesis and angiogenesis throughout the mutant embryo with a dense network of superficial and deep blood vessels (Fig. 5C,D), the dorsal aorta appeared enlarged and exhibited bulges (Fig. 5D, arrows). Further analysis using Flk-1 (fetal liver kinase 1) tyrosine kinase receptor, which is expressed on the surface of endothelial cells at all stages of development (Millauer et al.,1993), was performed. Here, mutants exhibit disorganized, truncated vessels that do not pass the dorsal somites (Fig. 5E,F, arrows). This result indicated that aberrant expression of PDGFRα leads to a disorganized vasculature.
Next, we addressed the question whether the aortic vessels begin proper maturation using SM22 expression as a marker. At E9.5, SM22 is expressed in the SMCs around the dorsal aorta, in the heart, and in the myotome. In situ hybridization showed that SMCs are associated with the aorta along its length in controls (Fig. 5G) and mutants (Fig. 5H). The staining intensity in mutant embryos is reduced to a faint stripe (Fig. 5H, arrows), whereas the wild-type showed a very strong and broad SM22 expression (Fig. 5G, arrows). Vibratome sections were used to visualize SMCs around the dorsal aorta in E9.5 embryos after in situ hybridization with SM22 probe (Fig. 5I,K). Here, the mutant embryos showed a dramatic reduction in SM22 staining intensity. In some cases, the staining indicative for SM22 expression failed to form a complete circle, suggesting that the process of wall formation is impaired or that the cell number might be reduced after successful wall formation. This finding suggested that development of vascular mural cells consisting of SMCs and pericytes were impaired. This might account for the aortic wall aneurisms observed.
In our study, we showed that ubiquitous expression in the embryo proper of autoactivated PDGFRα during embryogenesis leads to embryonic lethality up to E12.5, suggesting that misexpression causes fundamental consequences after implantation. We further demonstrated that proliferation is reduced and apoptosis is enhanced in the mutant embryos. The consequences are a dysmorphic neural tube most likely originating from lack of mesenchymal tissue. Although somitogenesis appears to be mildly affected, vasculogenesis is clearly affected. The number of SM22-positive vSMCs is reduced in mutants and the dorsal aorta is found to be enlarged leading to aneurisms and consequently death of the animals.
Analysis of the Sox2Cre;R26hPDGFRaPM revealed that some embryos fail to initiate turning and display a defect in yolk sac blood vessel development. A phenocopy of this phenotype has been reported after sm22Cre-mediated conditional deletion of both PDGFRα and PDGFRβ. In vitro allantois cultures demonstrated a requirement for PDGF signaling in blood vessel maturation (French et al.,2008). We speculate that Sox2Cre-mediated expression of the autoactivated PDGFRαPM transgene in all cells of the yolk sac leads to disturbance of the local signaling in these embryos.
The second group of Sox2Cre;R26hPDGFRaPM while developing further, display a size reduction, hemorrhages, and an undulated neural tube. In mice, neural tube closure is initiated around E8.5 at the border between cervical and hindbrain, the border between midbrain and forebrain and the rostral limit of the forebrain (Sakai,1989) and is completed by E9.5. At E9.5 in Sox2Cre;R26hPDGFRaPM embryos, a proper closure occurred in the craniofacial region only, while the tube in the trunk region is kinked, neural folds splayed from the midline. Interestingly, PDGFRα knockouts show a spina bifida caused by a missing fusion of the neural folds (Soriano,1997). To examine if misexpression and autoactivation of PDGFRα affects the neural tube development, in situ hybridizations with Msx1 were carried out. Dorsoventral patterning of the neural tube is regulated by bone morphogenic proteins (BMPs) and Msx1 is a target gene of these proteins involved in neural tube development (Monsoro-Burq et al.,1996; Timmer et al.,2002). The result of this analysis showed that there are no differences in expression pattern, indicating that the neural tube patterning is not impaired. However, according to a theory by Schoenwolf (Schoenwolf,1984; Smith and Schoenwolf,1997), neural tube closure is dependent on morphogenic movements of the neural tube itself and on paraxial mesenchyme. The physical pressure from paraxial mesenchyme on the neural tube is required for neural tube closure and is compromised by low cell proliferation of paraxial mesenchyme. We showed generally increased apoptosis and reduced proliferation in paraxial mesenchymal cells leading to a deficit in physical pressure, which might be responsible for perturbed neural fold fusion.
To characterize the development of the somites in the Sox2Cre;R26hPDGFRaPM embryos, we analyzed the expression of Uncx4.1, Tbx18, and HeyL. Uncx4.1 is required for maintenance of posterior somite characteristics (Leitges et al.,2000; Mansouri et al.,2000) and Tbx18 maintains the integrity of the anterior somite compartment (Bussen et al.,2004). In control and mutant embryos, Uncx4.1 is expressed in the entire posterior half of the somite and the expression of Tbx18 is restricted to the anterior part of the somite, indicating that the anterior–posterior somite polarity is not affected in Sox2Cre;R26hPDGFRaPM embryos. Hence, the defects seen in neural tube, somites and mesenchyme are most likely a secondary defect.
To examine defects in the vascular system of the mutants, anti-PECAM antibody staining was performed and revealed a strongly enlarged dorsal aorta in the mutant embryos. Due to the increase of apoptotic cells around the dorsal aorta in the mutants the aneurysm could be caused by a lack of SMCs. Controls and mutants showed SM22 expression in the SMCs around the dorsal aorta and in the myotome. However, the staining in the mutants was much weaker and transverse sections exhibited a decrease number of SMCs. The role of SMCs is contraction and regulation of blood vessel tone–diameter, blood pressure, and blood flow distribution, essential functions for the organism. Reduced numbers of SMCs lead to weakened aortic vessel walls and might be the reason of the massive hemorrhages in the trunk and the pericardial cavity of mutant embryos. However, we cannot unequivocally rule out that expression of the R26hPDGFRaPM transgene in the endothelial cells disrupts endothelial signaling or overactivation of mesoderm might indirectly lead to the vascular abnormalities observed here. Cell type-specific activation of the R26hPDGFRaPM transgene using available Cre-transgenic mice will help in addressing these issues. The question remains, which molecular cascade might be deregulated by misexpression of the autoactivated PDGFRα in the embryos?
In many aspects, the Sox2Cre;R26hPDGFRaPM mice resemble the phenotypic spectrum observed in mice knockout for Hey1/2 double mutants (Fischer et al.,2002), Jag1 (Xue et al.,1999), and Notch1 (Krebs et al.,2000). Promotor studies and analysis of Notch1 transgenic mice suggest that the HeyL gene is a potential target of the Notch signaling cascade (Nakagawa et al.,1999; Leimeister et al.,2000; Maier and Gessler,2000). Also, HeyL was implicated as a Notch1 effector because Notch1 and Delta-like1 knockout mutants display a loss of HeyL expression (Leimeister et al.,2000) and HeyL is found up-regulated in a transgenic “gain of function” approach of Notch1 (Lin et al.,2000). Notch signaling has been demonstrated to be crucial for development and maintenance of the vasculature (Conlon et al.,1995; Krebs et al.,2000). Promoter studies suggest a potential regulation of HeyL transcription by Notch3 (Maier and Gessler,2000). Campos et al. showed that in cultured vSMCs platelet-derived growth factor (PDGF) was able to down-regulate Notch3 and Jagged-1 through ERK-dependent signaling mechanisms (Campos et al.,2002). Our in situ studies using HeyL probes showed a reduction and disorganized expression in the somites of mutants and lack of expression in vSMCs where it is found coexpressed with Notch3 (Joutel et al.,2000). So we speculate that expression of the mutant PDGFRα in SMCs might lead to ERK-mediated Notch3 down-regulation, which results in HeyL down-regulation, causing the defect in the vSMCs.
Recently, transgenic mice have been established that express the mutated PDGFR D842V from the endogenous PDGFR promoter (Olson and Soriano,2009). These animals display perinatal lethality due to a defective lung development. Differences to our study consist in the choice of a promoter conferring a more restricted expression pattern than the ubiquitously active ROSA 26 gene promoter. Moreover, the transgene was activated using the Meox2cre mouse strain known to confer a mosaic pattern of cre-mediated recombination within the epiblast (Hayashi et al.,2002; Delgado et al.,2008). The resulting differences in spatiotemporal expression may explain the earlier and more deleterious consequences of our R26hPDGFRaPM transgene. Of note, however, Olson and Soriano reported a similar activation of downstream pathways as we had observed when analyzing osteoblast cultures derived from R26hPDGFRaPM transgenic mice (Moenning et al.,2009). Phosphorylation of Erk 1/2 and Akt remained unchanged while phospholipase C-γ was activated. Hence, we speculate that this mutation leading to autoactivation of PDFGRα might affect downstream signaling in a highly selective manner.
The defects observed in the Sox2Cre;R26hPDGFRaPM embryos are reminiscent of the human CADASIL syndrome (Cerebral Autosomal Dominant Arteriopathy with Subcortical Infarcts and Leukoencephalopathy, OMIM #125310). This disease is characterized by vSMC degeneration caused by mutations in the extracellular domain of the Notch3 receptor molecule resulting in a disruption of the native Notch3 structure causing misfolding of the protein (Joutel et al.,2000; Peters et al.,2004). Knockout animals demonstrate, that Notch3 is required for maturation of vSMCs (Domenga et al.,2004). There too, the arteries were found to be enlarged exhibiting a thinner, disorganized tunica media composed of discontinuous layers of noncohesive smooth muscle. Thus, the Sox2Cre;R26hPDGFRaPM mouse might serve as a model for this disease.
Derivation and Genotyping of Transgenic Mice
Derivation of transgenic mice and the detection of the R26hPDGFRaPM allele (MGI accession no. 3814559, Gt(ROSA)26Sortm1(PDGFRA*)Hsc) was described elsewhere (Moenning et al.,2009). For genotyping of the R26hPDGFRaPM transgene, the following primers were used: WT1AS 5′-ctcccaaagtcgctctgagtt-3′, WT1S 5′-cccattttccttatttgcccct-3′, and SA1AS 5′-gacatcatcaaggaaaccctg-3′. Polymerase chain reaction (PCR) was run for 35 cycles (45 sec 94°C, 30 sec 60°C, and 30 sec 72°C). For genotyping the Sox2Cre transgene, the following primers were used: Sox2Cre up 5′-atttgcctgcattaccggtc-3′, Sox2Cre down 5′-atcaacgttttgttttcgga-3′. The PCR was run for 35 cycles (45 sec 94°C, 30 sec 57°C, and 45 sec 72°C).
The day the vaginal plug was observed was considered as embryonic day 0.5 (E0.5). Pregnant mice were killed at E9.5; the embryos were dissected free from the extraembryonic membranes, fixed in 4% neutral buffered formalin for 12 hr, and embedded in paraffin. The infiltrated Embryos were sectioned at 2–5 μm and stained with hematoxylin and eosin.
Whole-mount in situ hybridizations were performed essentially as described (Schorle et al.,1996). Sections of SM22 and Msx1 whole-mount in situ hybridizations were performed on agarose-embedded embryos with a Vibratome (VT 1000, Leica Bensheim, Germany) as described (Maldonado-Saldivia et al.,2000).
Western Blot embryos were lysed in RIPA buffer (10 mM Tris-HCl pH 7,2, 150 mM NaCl, 5 mM EDTA, 0,1% sodium dodecyl sulfate [SDS], 1% Na deoxycholate, 1% Triton X-100) containing protease inhibitors (Complete Mini; Roche, Penzberg, Germany). Protein concentration of cell lysates was determined with the BCA protein assay kit according to the manufacturer's protocol (Pierce, Rockford, IL). Equal amounts of protein (15 μg) were resolved by SDS-polyacrylamide gel electrophoresis and transferred to a polyvinylidene fluoride membrane. Nonspecific binding sites were blocked, and membranes were incubated with antibodies to detect PDGFRα (1:200; Spring Bioscience, Fremont, CA) followed by incubation with horseradish peroxidase-conjugated secondary antibodies. Immunoreactivity was detected using the enhanced chemiluminescence kit (Pierce).
Whole-Mount Antibody Staining
Whole-mount antibody staining of embryos was performed as described (Schorle et al.,1996). Briefly, embryos were fixed in methanol:dimethyl sulfoxide (DMSO; 4: 1) for 12 hr at 4°C with gentle agitation. The embryos were then bleached in methanol:DMSO:30% H2O2 (4 :1:1) for 5 hr and blocked for 2 hr in phosphate-buffered saline containing 2% milk powder and 0.1% Triton X-100. The first antibody (NF160, 1:300; Sigma or anti-PECAM/CD31, 1:200; Biosciences) was incubated overnight at 4°C, thereafter the embryos were treated with horseradish peroxidase-conjugated secondary antibodies (1: 200). The embryos were then incubated with 0.3 mg/ml diaminobenzidine for 20 min. Color reaction was induced by adding H2O2 to a final concentration of 0.03%. The embryos were then cleared in benzyl alcohol/benzyl benzoate (1:2).
Pregnant mice were injected intraperitoneally with 0.25 mg/g BrdU in phosphate-buffered saline as described (Jager et al.,2003). Animals were killed 1 hr after BrdU administration. BrdU incorporation was detected in sections of embryos using a BrdU immunohistochemistry kit (Merck, Darmstadt, Germany).
TUNEL (terminal deoxynucleotidyl transferase–mediated deoxyuridinetriphosphate nick end-labeling) staining was performed as described (Jager et al.,2003) using the Apop Tag In Situ Detection kit (Merck) and was used to examine apoptosis in sections of E9.5 embryos.
We thank Mathilde Hau-Liersch and Inge Heim for excellent technical assistance and Olaf Babczynski for animal care. We thank the following individuals for supplying probes for in situ hybridization: Flk1, Tbx18, Uncx4.1 (A. Kispert, Hannover); Flk1 (M. Breier, Dresden); HeyL, SM22, Notch1, Notch3 (M. Gessler, Wurzburg); Msx1 (R. Maas, Boston); and Shh (M. Blum, Hohenheim). H.S. was funded by the Deutsche Krebshilfe and the Stem Cell Network NRW.