Gene expression analysis of Tek/Tie2 signaling
Stephen H. Chen1,2,
Yael Babichev1,
Natalie Rodrigues1,
Daniel Voskas1,2,
Ling Ling1,2,
Vicky P. K. H. Nguyen1,2 and
Daniel J. Dumont1,2,3,4
1 Division of Molecular and Cellular Biology Research, Sunnybrook and Women's Research Institute, Toronto
2 Department of Medical Biophysics, University of Toronto, Toronto
3 Toronto-Sunnybrook Regional Cancer Centre, Toronto
4 Heart and Stroke/Richard Lewar Centre of Excellence, Faculty of Medicine, University of Toronto, Toronto, Ontario, Canada
 |
ABSTRACT
|
---|
The elaboration of the vasculature during embryonic development involves restructuring of the early vessels into a more complex vascular network. Of particular importance to this vascular remodeling process is the requirement of the Tek/Tie2 receptor tyrosine kinase. Mouse gene-targeting studies have shown that the Tie2-deficient embryos succumb to embryonic death at midgestation due to insufficient sprouting and remodeling of the primary capillary plexus. To identify the underlying genetic mechanisms regulating the process of vascular remodeling, transcriptomes modulated by Tie2 signaling were analyzed utilizing serial analysis of gene expression (SAGE). Two libraries were constructed and sequenced using embryonic day 8.5 yolk sac tissues from Tie2 wild-type and the Tie2-null littermates. After tag extraction, 45,689 and 45,275 SAGE tags were obtained for the Tie2 wild-type and Tie2-null libraries, respectively, yielding a total of 21,376 distinct tags. Close to 62% of the tags were uniquely annotated, whereas 10% of the total tags were unknown. Using semiquantitative PCR, the differential expression of eight genes was confirmed that included Elk3, an important angiogenic switch gene which was upregulated in the absence of Tie2 signaling. The results of this study provide valuable insight into the potential association between Tie2 signaling and other known angiogenic pathways as well as genes that might have novel functions in vascular remodeling.
receptor tyrosine kinase; angiogenesis; serial analysis of gene expression
 |
INTRODUCTION
|
---|
THE GROWTH AND THE EXPANSION of an early vascular system in developing embryos requires a remodeling step whereby a preexisting vasculature undergoes modification to ultimately produce a functional vascular network. This crucial developmental stage involves several discrete processes that may include intususceptive, intercalated and sprouting angiogenesis, vascular pruning and/or vasculogenesis (35, 47). In particular, the vasculature of the yolk sac is one of the first discrete vascular networks to form in the mouse embryo (36). By embryonic day 8.0 (E8.0) of mouse development, the vasculature forms a meshwork of vessels commonly referred to as a primary vascular plexus. This primitive vessel network is further modified to give rise to a more mature vascular system that consists of a hierarchy of larger vessels leading to smaller branches in a highly arborized and detailed pattern (47). Due to its relative simplicity, the yolk sac is an attractive tissue source for studying the processes of vascular remodeling (9).
To date, numerous signaling pathways have been shown to be indispensable in the formation of the vascular network. These pathways include two receptor tyrosine kinase (RTK)-mediated signaling pathways activated by the vascular endothelial growth factors (VEGF) (10) and the angiopoietins, which are the ligands for VEGFR13 and Tek/Tie2 receptors, respectively (10, 11, 14, 15, 20). Almost exclusively restricted to cells of the endothelial lineage, the expression of these RTKs is important for the normal development of a functional vascular system during embryogenesis (2, 8, 20, 43). Evidence of the involvement of these RTKs in vascular development is shown in mouse gene-targeting studies; mice engineered to carry a null mutation within the Tie2 or Tek gene die by E9.5 of embryonic development (17, 40). In addition to the vascular defects within the embryo proper, these Tie2-null (tek
sp/tek
sp) embryos also display defects in the remodeling of the yolk sac vasculature that remains as a primitive plexus. By E9.0, there are 75% fewer endothelial cells found within the yolk sac, suggesting that, in addition to a role in early vascular remodeling, the Tie2-signaling pathway may also provide critical survival cues (17). Evidently, conditional rescue of the vascular remodeling defects of the yolk sac and the embryo proper using a tetracycline-based transgenic approach demonstrated that Tie2 signaling indeed plays an important role in transducing cell survival signals in the midgestation embryo (22).
Despite intense effort in elucidating the various signaling pathways downstream of the Tie2 receptor, not much is currently known about the genes that are being modulated or how the process of vascular remodeling is being genetically regulated in general. Discoveries of novel angiogenic targets or regulators are continuing to being made that benefit the study of developmental as well as pathological angiogenesis; thus this underscores the importance of obtaining a better understanding of genetic mechanisms underlying vascular remodeling. To this end, a genomic comparison between the transcriptomes from Tie2 wild-type (WT) and Tie2-null (KO) yolk sacs was performed using serial analysis of gene expression (SAGE). SAGE technology (44) was developed as a high throughput and quantitative tool to assess polyadenylated transcripts by generating 10-bp sequence tags that are immediately downstream of the most 3' NlaIII recognition sequence site closest to the polyadenylated end. It provides a robust and unbiased analysis of the transcriptome within cells or tissues. In this study, two SAGE libraries were constructed from WT and KO yolk sacs at E8.5: a total of 21,376 distinct tags were sequenced and mapped to 13,609 known genes, while 2,158 tags were not annotated. Eight genes were selected after Gene Ontology analysis based on fold difference in expression as well as involvement in angiogenesis for further confirmation using semiquantitative PCR. Interestingly, Elk3/Net, a transcription repressor previously shown to control the angiogenic switch (50), was upregulated in the absence of the Tie2 receptor. The identification of this gene and others reveals potential pathways regulated by Tie2 and provides an entry point for the study of the genetic basis of vascular remodeling.
 |
MATERIALS AND METHODS
|
---|
Tissue isolation.
The generation of the Tie2-null (tek
sp/ tek
sp) double transgenic mice has been published previously (17). The morning of the vaginal plug was considered as day 0.5 of gestation. Embryos at approximately E8.5 were harvested before the manifestation of the tek
sp/tek
sp phenotype (17), and each embryo was separated into three parts, upper embryo, lower embryo and its yolk sac, which were immediately snap-frozen in a dry ice-ethanol bath and archived for future use. The upper embryos and the yolks were stored in 80°C freezer, while the lower embryo was used for genotyping as described (22). Embryos with ambiguous genotypes were excluded from these studies. RNA isolation was prepared by using TRIzol reagent (Invitrogen) according to manufacturer's instructions. The RNA concentration of the samples was spectrophotometrically measured at 260/280 nm, and 500 ng of the RNA from each pool were analyzed by agarose/ethidium bromide gel electrophoresis to assess for quality. The RNA used for the construction of the SAGE libraries in this study was isolated from E8.5 yolk sacs derived from the embryos of the tek
sp-heterozygote mating.
Construction of SAGE libraries.
Using the I-SAGE kit (Invitrogen), two SAGE libraries were constructed following the manufacturer's instructions with minor modifications. The starting amount of RNA used was 25 µg instead of the 510 µg suggested. Furthermore, an additional SphI restriction digest was performed before running the concatemers on a polyacrylamide gel to ensure optimal cloning efficiency in subsequent steps. The clones with tag inserts were sequenced on a CEQ8000 analysis system (Beckman) and an ABI 377 DNA sequencer (Applied Biosystems) using Dye Terminator (Beckman) and BigDye Terminator (Applied Biosystems), respectively.
SAGE data analysis.
SAGE tags were extracted, and duplicate ditags were removed from data analysis using SAGE 2000 software version 4.5 (http://www.invitrogen.com) (courtesy of Drs. Victor Velculescu and Kenneth Kinzler at The Johns Hopkins University School of Medicine). The tag-to-gene mappings were performed by matching tag sequences to the mouse reliable SAGEmap list (UniGene build no. 144) in Microsoft Access to obtain current annotations of UniGene clusters for the tags (26). Statistical comparison of the SAGE libraries was done using SAGEStat software (40), kindly provided by Dr. J. Ruijter of the University of Amsterdam (The Netherlands), which performed a comparative Z-test to calculate P values. Gene Ontology analysis of the genes was performed using the web-based Gene Ontology Tree Machine program (http://genereg.ornl.gov/gotm) (49). The complete SAGE data set can be obtained at http://www.angiogenesis.ca/sage.html.
Semiquantitative RT-PCR analysis.
Total RNA was prepared from yolk sac tissues of Tie2 KO and Tie2 WT embryos at E8.5 and quantified spectrophotometrically. cDNA was synthesized from 1 µg of total RNA with Thermoscript RT (Invitrogen) according to the manufacturer's instructions. PCR products were separated on a 1% agarose gel and stained with ethidium bromide. Specific primer sequences used for semiquantitative PCR analysis were designed using Primer3 program (38), and their respective annealing temperatures were empirically determined using gradient PCR: Ctsh (forward 5'-gagtacaaccacagactgcagat-3', reverse 5'-gcatgaactatccttgcctatg-3'), Psph (forward 5'-atctgactcctggcataagg-3', reverse 5'-cctaactgtgtggctttgag-3'), Sdf2 (forward 5'-gattagttctctcggatctagg-3', reverse 5'-accatctctaacccagtaagg-3'), Idb2 (forward 5'-aggtccgttaggaaaaacagc-3', reverse 5'-ccacagagtactttgctatcattcg-3'), Cul2 (forward 5'-ttgatgacaaggacgtctttcag-3', reverse 5'-gacagtttcactgttgttaaaggc-3'), Tm4sf8 (forward 5'-cgtgctattgactatgtgcag-3', reverse 5'- gtacgtgtgctttgactagatg-3'), Elk3 (forward 5'-ccaccaacgtcactgtcatc-3', reverse 5'-gtgtatgcaagcggagttca-3'), Dpp7 (forward 5'-tacccgtatcctactgactttc-3', reverse 5'-atgtactcaggttgctctgg-3'), and ß-actin (forward 5'-gcgctcggtgaggatcttca-3', reverse 5'-caaggccaaccgcgagaaga-3'). Semiquantitative PCR was performed using 1 µl of cDNA solution at 1, 1/5, and 1/10 dilutions with Taq polymerase (Sigma), according to manufacturer's instructions with the following modifications: initial incubation was set at 95°C for 3 min followed by 30 cycles of denaturation at 95°C for 30 s, annealing (at various temperatures according to specific genes) for 45 s, and extension at 72°C for 1 min and was finished with final extension at 72°C for 10 min.
 |
RESULTS
|
---|
Approximately 30 Tie2 WT and 30 Tie2 KO yolk sac tissues were pooled together from mouse littermates and used for total RNA extraction to generate the starting RNA samples for construction of the two SAGE libraries in this study. Once duplicate ditags were removed, a total of 90,964 tags for both libraries were sequenced and acquired. Specifically, 45,689 and 45,275 tags corresponding to WT and KO libraries, respectively, were obtained. Combined analysis of these two libraries yielded a total of 21,376 distinct tags, 13,465 tags from the WT library and 13,707 tags from the KO library. As shown in Fig. 1, over 55% of the tags were exclusive to each library (7,669 tags and 7,911 tags in the WT and KO libraries, respectively), but close to 90% of these tags were expressed only once (singletons) in each library. In contrast, 5,796 tags were observed in both libraries, and
75% of these tags were expressed at least twice in either library. In total,
29% (6,265 tags) of the total tag population had tag counts over 1 in either of the two libraries and thus are termed nonsingleton (NS) tags.

View larger version (19K):
[in this window]
[in a new window]
|
Fig. 1. Venn diagram comparison of 21,376 different tags from wild-type (WT) and knockout (KO) serial analysis of gene expression (SAGE) libraries. No. of nonsingleton (NS) tags in each category is indicated in the brackets.
|
|
More in-depth comparison between the two libraries indicates that there was a high degree of similarity in the distribution of tag abundance (Fig. 2). Almost all of the tags from both libraries had an abundance of <50 copies, and close to 90% of the tags had <5 copies. In fact, only
0.2% of the tags were observed >100 times, while over two-thirds of the tags were singletons. Overall, the tag distribution profiles of the two SAGE libraries presented in this study are similar to other SAGE libraries, suggesting that the expression of the majority of mammalian genes is usually at a low level (6, 41, 45).

View larger version (26K):
[in this window]
[in a new window]
|
Fig. 2. Distribution of tag frequencies for the WT and KO transcriptomes. Close to 2/3 of the tags were singletons.
|
|
After tag-to-gene mapping,
90% or 19,218 tags were annotated to known genes (Fig. 3), while close to 11% (2,158 tags) of the entire population of tags could not be mapped to any known genes, and only 6.8% or 147 tags out of these unannotated tags were NS. For tags with gene annotation, 62% or 11,830 tags were uniquely mapped to single UniGene clusters, while 38% (7,388 tags) of the tags were ambiguously mapped to multiple clusters. Approximately 4.3% or 264 annotated NS tags (151 were ambiguously mapped and 113 were uniquely mapped) had a P value less than 0.05. In contrast, there were only 9 out of 147 un-mapped NS tags (6.1%) had a P value less than 0.05. This finding indicates that only a small fraction of the total genes were differentially expressed at a statistically significant level which might be implicated in the cellular or physiological modification leading to vascular remodeling.

View larger version (34K):
[in this window]
[in a new window]
|
Fig. 3. Schema of the analysis of tag-to-gene mappings of WT and KO SAGE libraries. Approximately 90% of the tags were annotated to known transcripts.
|
|
The 20 most abundant tags are listed in Table 1 along with corresponding reliable gene annotation for these tags. Although the true identity of over one-half of these highly abundant tags could not be ascertained due to ambiguous mappings, a majority of these tags appeared to be involved in routine cellular processes such as energy metabolism and protein and lipid synthesis. This result is consistent with other SAGE libraries, reflecting the high demand for cell maintenance (6, 41). In addition, a few of these highly abundant tags were mapped to hemoglobin and upregulated in the KO library; this indicates that Tie2 signaling may act to suppress erythropoiesis, or the elevated levels may represent the presence of red blood cells accumulating in the KO yolk sacs. Moreover, there were a number of these highly abundant tags mapped ambiguously to transcribed/expressed sequences that have yet to be characterized. This ambiguity is possibly a result of spliced variants of the other gene annotations highly abundant in the yolk sacs.
To determine whether tags were expressed differentially, a statistical analysis based on Z-test using SAGEStat software (39) was performed on all tags to assign statistical significance for the difference between the two tag counts in the WT and KO libraries. A list of 20 tags with the smallest P values is presented in Table 2. Once again, tags associated with hemoglobin transcripts are among this list, suggesting that an accumulation of red blood cells was present in the KO yolk sacs. Two tags listed in Table 2 were unannotated, but, on further investigation using the SAGEmap database (26), both of these tags were also expressed at a medium-to-high level in other SAGE libraries, suggesting that these tags might have derived from transcripts common in normal cellular processes or from potential novel transcripts (S. H. Chen and D. J. Dumont, unpublished observations). Figure 4 represents the frequency of the NS tags that were differentially expressed by over 2-, 5-, and 10-fold. Out of 6,265 NS tags present in the two libraries, close to one-half (47% or 2,949 tags) were expressed differentially by over twofold, while about 9% (544 tags) of the NS tags were expressed by over fivefold. In addition, 0.43% (27 tags) of the NS tags were differentially expressed by over 10-fold. After tag-to-gene mapping, the identity of these NS tags was revealed. A list of 30 most downregulated tags in the KO yolk sacs is shown in Table 3, and a list of 30 most upregulated tags in the KO yolk sacs is presented in Table 4. All of these most differentially expressed tags are mapped except one tag which is upregulated in the KO yolk sac that is either lowly expressed or not expressed at all in most of the mouse SAGE libraries in the SAGEmap database (S. H. Chen and D. J. Dumont, unpublished observations).

View larger version (16K):
[in this window]
[in a new window]
|
Fig. 4. Graph of frequency of nonsingleton (NS) tags that were differentially expressed more than 2-, 5-, and 10-fold. Only tags with counts 2 in either library were chosen for calculation, and a count of 0.5 was used in place of 0 for fold difference calculation.
|
|
To gain further insight into the putative roles of differentially expressed genes underlying the WT and KO yolk sacs, a Gene Ontology approach using a web-based platform, Gene Ontology Tree Machine (GOTM), was utilized to assign Gene Ontology biological process terms based on the differentially expressed NS tags, each uniquely mapped to one UniGene cluster and differentially expressed by at least twofold (49). This list included UniGene clusters containing single NS tags as well as clusters with multiple NS tags. To minimize ambiguity in determining differential expression, UniGenes with multiple NS tags were selected only if all the NS tags mapping to that specific cluster were differentially expressed at least twofold in the same direction, either up- or downregulated in one library. A directed acyclic graph (DAG) representation of the Gene Ontology hierarchies based on biological processes using the stipulated criteria was generated, and an abridged version is shown in Fig. 5. Full DAG visualization of Gene Ontology based on biological process, molecular function, and cellular component for this list of differentially expressed genes can be found in the Supplementary Materials (available at the Physiological Genomics web site).1
The Gene Ontology categories shown in red contain significantly enriched gene numbers (P value < 0.01) compared with the mouse reference gene set of the program, whereas the black categories are nonenriched. To generate Fig. 5, 2, 079 differentially expressed NS tags mapping to 1,686 UniGene clusters were selected; 1,632 clusters had single NS tags (854 downregulated and 778 upregulated in the KO library), while 54 had multiple NS tag mappings that were all differentially expressed at least twofold in the same trend (29 downregulated and 25 upregulated in the KO library). About 33% or 557 of the differentially expressed genes are predicted to be involved in metabolism, and
22% or 367 genes are implicated in cellular physiological processes (Fig. 5). Highlighted in bold are Gene Ontology categories important in vascular development. Predictably, there were 45 genes that are involved in apoptosis, which is a branch of a signaling pathway known to be downstream of Tie2 signaling in endothelial cells, whereas 14 genes were found to be associated with blood vessel development, suggesting possible interaction with targets modulated by Tie2 and its downstream signaling (21, 25, 33).

View larger version (45K):
[in this window]
[in a new window]
|
Fig. 5. An abridged directed acyclic graph (DAG) representation of Gene Ontology (GO) assignments of differentially expressed NS tags based on biological processes. GO categories with significantly enriched gene numbers (P value < 0.01) are shown in red. Biological processes leading to death and blood vessel development are highlighted in bold. A full version of this graph is shown in Supplementary Fig. S1. The tag-to-gene mapping was based on UniGene build no. 144.
|
|
To further confirm the SAGE results, eight genes (Ctsh, Psph, Sdf2, Idb2, Cul2, Tm4sf8, Elk3, and Dpp7) were selected from the list of NS tags used to generate Fig. 5, and the differential expression of these genes was validated using semiquantitative PCR on E8.5 Tie2 WT and KO yolk sacs (Fig. 6). As observed, the semiquantitative PCR results recapitulated the SAGE data: seven of these eight genes (Ctsh, Psph, Sdf2, Idb2, Cul2, Tm4sf8, Dpp7) were downregulated in the KO yolk sacs, while one gene (Elk3) was upregulated in the KO yolk sacs.

View larger version (41K):
[in this window]
[in a new window]
|
Fig. 6. Semiquantitative PCR analysis showing the differential expression of 8 selected genes in the yolk sacs. A: for each gene, UniGene ID, SAGE tag sequence, tag count [both tag no. and normalized tag no., tags/million (TPM)], P values, and the GO biological process assignments are shown. B: semiquantitative PCR analysis on these 8 genes was performed on embryonic day 8.5 (E8.5) WT and KO yolk sac cDNA.
|
|
 |
DISCUSSION
|
---|
Genome-wide expression profiling of the Tie2 transcriptome.
In this study, the gene expression profiles of the developing mouse yolk sacs with or without Tie2 signaling were obtained and analyzed using SAGE. Compared with the embryo proper, yolk sac is a relatively simple tissue consisting of three cell layers: ectoderm, endoderm and mesoderm, part of which develops into the first blood vessel system in embryogenesis (34, 47). Yolk sac tissues were chosen as a source of endothelial cells, since yolk sacs can be removed quite easily from the embryos at that embryonic stage; therefore, there is no contamination with maternal or embryonic tissues. In addition, the isolation of endothelial cells from individual embryos before their genotyping was not feasible. Moreover, Tie2 WT and KO yolk sacs were strategically harvested at approximately E8.5 to ensure there was no significant loss of endothelial cells (17). SAGE was used as the choice of gene expression-profiling tool due to its high sensitivity and resolution for low-abundance genes such as transcription factors. In addition to the possibility of identifying novel transcripts that might be important in vascular remodeling, SAGE also facilitates the identification of known genes whose expression and function are not immediately downstream of Tie2. It is believed that the simple nature of yolk sac tissue and the specific time point at which embryos were harvested before the onset of visible phenotypes provide a strategy to enrich for Tie2-modulated endothelial cell-specific genes. In fact, a brief survey of the identified genes using the Gene Expression Omnibus database (7) illustrated that a large proportion of these genes are expressed in endothelial cells, although not restricted to endothelial cells (S. H. Chen and D. J. Dumont, unpublished observations). Specifically, Elk3, one of the eight genes validated in this study, has been shown to be expressed predominantly in endothelial cells and thought to act as an angiogenic switch in embryonic development (3, 50). Furthermore, a great majority of the tags were expressed at fewer than five copies in this study, illustrating the advantage of SAGE in capturing low-abundance genes. These observations further support the use of SAGE for examining gene expression profiles as well as the use of yolk sacs as a valid approach to identify genes that may play a role in vascular remodeling.
Initial analysis of the identity and distribution of the different tags from the two libraries reveals that there are a considerable number of differentially expressed genes potentially modulated by the Tie2-signaling pathway; 2,949 NS tags or 14% of the entire tag population were differently expressed by over twofold in the two libraries (Fig. 4). Although it is highly likely that some of these tags may map to the same genes and changes in the expression of a significant number of genes in the KO yolk sacs might be secondary to Tie2 signaling, the data obtained remain overwhelming and would require highly stringent data mining criteria to select for candidate genes that might be involved in vascular remodeling. From a pure systems biology perspective, however, it would be interesting to investigate further whether any kind of perturbation to the biological system would elicit a change in the transcriptome at a similar level. Collectively, these results illustrate the interdependence of the expression of numerous genes on the signaling pathways mediated by a single receptor such as Tie2.
Another intriguing observation is that the 15th most-abundant tag (GCCACTGTGC), ambiguously mapped to either Adamts7 or 2810433K01Rik, appeared to be highly expressed only in this library, whereas its expression was either low or nonexistent compared with other mouse SAGE libraries from other tissues archived in the SAGEmap database (S. H. Chen and D. J. Dumont, unpublished observations). This observation indicates that this tag might be involved in a cellular process unique to the yolk sac during development. In addition, over one-half of the highly abundant tags in Table 1 were not uniquely mapped, and a number of the ambiguous mappings observed correspond to similar genes, suggesting that either these tags were derived from other known highly abundant genes as spliced variants or more likely that the UniGene database contains redundant gene annotations. Interestingly, several of the most-abundant tags appeared to be derived from hemoglobin, which is likely a result of red blood cells present in the yolk sacs at the time of preparation. Expectedly, tags mapping to both adult and embryonic hemoglobin were present, since primitive red cells express both types of hemoglobin (24). In addition, there were more hemoglobin-associated tags in the KO yolk sacs, coinciding with the physiological observation that KO yolk sacs suffer from hemorrhage due to insufficient vasculature development, thus creating an accumulation of red blood cells. Alternatively, the higher abundance of hemoglobin in the KO yolk sacs might be attributed to the possible deregulation of hematopoiesis due to the absence of Tie2 signaling.
One of the advantages of SAGE is its ability to profile transcript expression regardless of any a priori knowledge with regards to identity. In this study, there were nine NS tags differentially expressed at statistically significant levels that were not annotated to the SAGEmap database and therefore are potential novel transcripts. In particular, one unmapped NS tag from Table 4 was among one of the most differentially expressed in the KO yolk sac, and it was virtually not found in any other libraries in the SAGEmap database. Further investigation is needed to ascertain whether this tag will lead to the identification of a novel gene and its possible involvement in angiogenesis.
Analysis of the unique tags based on Gene Ontology functions.
As shown in Fig. 5, a total of 2,079 NS tags were selected for further data mining by utilizing Gene Ontology according to their biological processes. These NS tags were uniquely annotated to single UniGene clusters and were differentially expressed by over twofold. By performing Gene Ontology analysis to these NS tags after gene annotations, potential biological themes that were affected as a result of the experiment can be distinguished and selected for further studies. As observed in other SAGE analyses, genes implicated in metabolism and cellular growth are most abundant (6, 41). This suggests that impaired signaling from the Tie2 receptor in KO yolk sacs may have a profound influence on normal cellular processes. In addition, Tie2 signaling may influence cell death, which is a known signaling branch downstream of Tie2. As shown in Fig. 5 and highlighted in bold, the programmed cell death category is significantly enriched with 47 genes; within these 47 genes, 11 are shown to be involved in the negative regulation of apoptosis. Three of these 11 genes were downregulated in the KO SAGE library, suggesting that there perhaps is a mixture of apoptotic and counteracting anti-apoptotic mechanisms in the KO yolk sac development where there is increased amount of death due to lack of Tie2. However, the precise mechanisms of these 11 genes are different, and each may be involved in diverse cellular pathways. Another interesting Gene Ontology term highlighted on the DAG is the blood vessel development category, which was significantly enriched with fourteen genes (Elk3, VEGFa, VEGFb, Col18a1/endostatin, ß-catenin, neuropilin-1, Rhob, Choroideremia, Ptk2, Hand1, Foxf1a, Hif1a, quaking, and Egfl7). One-half of these genes were upregulated in the KO SAGE libraries, suggesting that deficient Tie2 signaling may elicit gene expression changes relating to other pathways involved in vessel development. Although every effort was made to isolate Tie2 WT and KO yolk sacs before an observable phenotype, one cannot exclude the possibility that the changes in these biological processes might be secondary to vascular insufficiency and are not specific to Tie2 signaling; therefore, further characterization is required to confirm the roles of these genes in Tie2 signaling.
Validation of selected genes using semiquantitative PCR.
To confirm the validity of the SAGE data and to eliminate the selection of false-positive genes for further studies, semiquantitative PCR was performed on eight genes (Fig. 6). Initial literature research revealed that, while Elk3 is known to be an angiogenic switch and both Cul2 and Idb2 might have a possible association with angiogenesis through the hypoxia/HIF-1 pathway, the other five genes have either been poorly characterized, or their association with angiogenesis remains unknown (27, 29, 50). Further characterization studies are needed to explore the possible association of these genes downstream of Tie2 signaling.
Angiogenic switch gene, Elk3, is differentially expressed.
Of particular interest to this study is the identification of the transcription factor Elk3 (18), which was upregulated in the Tie2 KO yolk sacs. Elk3 is known to be expressed at sites of vasculogenesis and angiogenesis, and mice engineered to express a truncated form of Elk3 that do not contain the Ets domain (Net
/
) die postnatally of chylothorax, suggesting a later role for Elk3 in lymphatic development or function (3, 4). Interestingly, Net
/
animals that do survive present with decreased wound healing times and decreased levels of VEGF production, further indicating that Elk3 may also play a role in control of the angiogenic switch (50). In addition, Elk3 has been shown to repress the expression of inducible nitric oxide synthase (iNOS), a member of a family of enzymes responsible for the production of nitric oxide (NO), which is a short-lived signal transduction chemical known to play a pivotal role in vascular biology (13, 30). In fact, iNOS expression was downregulated in the KO library and was included as one of the four genes under the significantly enriched NO biosynthesis category in Fig. 5. This observation adds further evidence to the involvement of Elk3 and NO in this study. Specifically, NO is produced by a class of enzymes known as the NOSs, which includes iNOS, endothelial NOS (eNOS), and brain NOS (bNOS) (30, 31), and can act directly on caspase-3, providing an additional mode of anti-apoptotic signaling (15, 28, 37). Under pathological situations, iNOS expression has been colocalized within endothelial cells (19, 48), suggesting that iNOS expression in endothelial cells may also play a role in endothelial biology during times of chronic stress. However, a role for iNOS in angiogenic development remains unclear, since its expression is downregulated after E9.5 of development, although it does reappear upon activation of the endothelium (1, 19, 32). Nevertheless, both in vitro and in vivo experiments have suggested that the Tie2-signaling pathway plays an important role in endothelial cell survival (20, 46) and the angiogenic switch (42), both of which overlap in the functional roles of Elk3 (3, 4, 18, 50). Thus Elk3 could prove to be a key angiogenic factor regulated by Tie2 and is involved in embryonic development. Nevertheless, further studies are required to test this hypothesis.
Taken together, the results of this study provide insight into changes of the transcriptome in the presence or absence of Tie2 signaling. The confirmation of the differential expression of eight genes serves as a good starting point to further elucidate the Tie2-signaling pathways and potential cross-talk among other pathways. This study also suggests the involvement of Tie2 signaling in the regulation of Elk3 (23, 42, 50) and further supports a role for the production of NO as one of the mediators of angiopoietin-induced signaling (5, 12). Thus the identification and validation of Elk3 and seven other genes indicate that many of the unknown genes identified by this approach may also play integral roles in vascular remodeling. Future studies are needed to identify more genes that are differentially expressed and to ascertain the involvement of the validated genes in the vascular remodeling process via Tie2 signaling.
 |
GRANTS
|
---|
D. Voskas and V. P. K. H. Nguyen are supported by studentships from the Heart and Stroke Society of Canada. Y. Babichev is supported by a fellowship from the Heart and Stroke/Richard Lewar Centre of Excellence. D. J. Dumont is a Canadian Institute for Health Research Scientist and is a member of the Heart and Stroke/Richard Lewar Centre of Excellence and the McLaughlin Centre for Molecular Medicine, University of Toronto, Toronto, Canada. This work was supported by a grant from National Heart, Lung, and Blood Institute (HL-63224-01).
 |
DISCLOSURES
|
---|
We declare no competing financial interests.
 |
ACKNOWLEDGMENTS
|
---|
We thank Sue Farinaccio for excellent administrative assistance and Laura Kaufer and Ivana Koloper for help in the SAGE library construction. In addition, we thank Drs. Frank Baas, Fred van Ruissen, and Jan Ruijter (Academic Medical Centre, University of Amsterdam, The Netherlands) for invaluable advice on the analysis of SAGE data.
 |
FOOTNOTES
|
---|
Article published online before print. See web site for date of publication (http://physiolgenomics.physiology.org).
Address for reprint requests and other correspondence: D. J. Dumont, Sunnybrook and Women's Research Institute, 2075 Bayview Ave., Research Bldg., S-218, Toronto, Ontario, Canada M4N 3M5 (e-mail: dan.dumont{at}sw.ca).
10.1152/physiolgenomics.00063.2005
The Supplemental Material for this article (Supplemental Fig. S1) is available online at http://physiolgenomics.physiology.org/cgi/content/full/00063.2005/DC1.
 |
REFERENCES
|
---|
- Abu El-Asrar AM, Meersschaert A, Dralands L, Missotten L, and Geboes K. Inducible nitric oxide synthase and vascular endothelial growth factor are colocalized in the retinas of human subjects with diabetes. Eye 18: 306313, 2004.[CrossRef][ISI][Medline]
- Adams RH and Klein R. Eph receptors and ephrin ligands. essential mediators of vascular development. Trends Cardiovasc Med 10: 183188, 2001.[CrossRef][ISI]
- Ayadi A, Suelves M, Dolle P, and Wasylyk B. Net, an Ets ternary complex transcription factor, is expressed in sites of vasculogenesis, angiogenesis, and chondrogenesis during mouse development. Mech Dev 102: 205208, 2001.[CrossRef][ISI][Medline]
- Ayadi A, Zheng H, Sobieszczuk P, Buchwalter G, Moerman P, Alitalo K, and Wasylyk B. Net-targeted mutant mice develop a vascular phenotype and up-regulate egr-1. EMBO J 20: 51395152, 2001.[Abstract/Free Full Text]
- Babaei S, Teichert-Kuliszewska K, Zhang Q, Jones N, Dumont DJ, and Stewart DJ. Angiogenic actions of angiopoietin-1 require endothelium-derived nitric oxide. Am J Pathol 162: 19271936, 2003.[Abstract/Free Full Text]
- Baross A, Schertzer M, Zuyderduyn SD, Jones SJ, Marra MA, and Lansdorp PM. Effect of TERT and ATM on gene expression profiles in human fibroblasts. Genes Chromosomes Cancer 39: 298310, 2004.[CrossRef][ISI][Medline]
- Barrett T, Suzek TO, Troup DB, Wilhite SE, Ngau WC, Ledoux P, Rudnev D, Lash AE, Fujibuchi W, and Edgar R. NCBI GEO: mining millions of expression profilesdatabase and tools. Nucleic Acids Res 33 (Database issue): D562D566, 2005.[Abstract/Free Full Text]
- Betsholtz C, Karlsson L, and Lindahl P. Developmental roles of platelet-derived growth factors. Bioassays 23: 494507, 2001.[CrossRef][ISI][Medline]
- Breier G. Angiogenesis in embryonic developmenta review. Placenta 21, Suppl 1: S11S15, 2000.[CrossRef][ISI][Medline]
- Breier G, Albrecht U, Sterrer S, and Risau W. Expression of vascular endothelial growth factor during embryonic angiogenesis and endothelial cell differentiation. Development 114: 521532, 1992.[Abstract]
- Carmeliet P. Mechanisms of angiogenesis and arteriogenesis. Nat Med 6: 389395, 2000.[CrossRef][ISI][Medline]
- Chen JX, Lawrence ML, Cunningham G, Christman BW, and Meyrick B. HSP90 and Akt modulate Ang-1-induced angiogenesis via NO in coronary artery endothelium. J Appl Physiol 96: 612620, 2004.[Abstract/Free Full Text]
- Chen YH, Layne MD, Chung SW, Ejima K, Baron RM, Yet SF, and Perrella MA. Elk-3 is a transcriptional repressor of nitric-oxide synthase 2. J Biol Chem 278: 3957239577, 2003.[Abstract/Free Full Text]
- Davis S, Aldrich TH, Jones PF, Acheson A, Compton DL, Jain V, Ryan TE, Bruno J, Radziejewski C, Maisonpierre PC, and Yancopoulos GD. Isolation of angiopoietin-1, a ligand for the TIE2 receptor, by secretion-trap expression cloning. Cell 87: 11611169, 1996.[CrossRef][ISI][Medline]
- de Vries C, Escobedo JA, Ueno H, Houck K, Ferrara N, and Williams LT. The fms-like tyrosine kinase, a receptor for vascular endothelial growth factor. Science 255: 989991, 1992.[ISI][Medline]
- Dimmeler S, Haendeler J, Nehls M, and Zeiher AM. Suppression of apoptosis by nitric oxide via inhibition of interleukin-1ß-converting enzyme (ICE)-like and cysteine protease protein (CPP)-32-like proteases. J Exp Med 185: 601607, 1997.[Abstract/Free Full Text]
- Dumont DJ, Gradwohl G, Fong GH, Puri MC, Gertsenstein M, Auerbach A, and Breitman ML. Dominant-negative and targeted null mutations in the endothelial receptor tyrosine kinase, tek, reveal a critical role in vasculogenesis of the embryo. Genes Dev 8: 18971909, 1994.[Abstract]
- Giovane A, Pintzas A, Maira SM, Sobieszczuk P, and Wasylyk B. Net, a new ets transcription factor that is activated by Ras. Genes Dev 8: 15021513, 1994.[Abstract]
- Homma H, Takahashi T, Seki H, Ohtani M, Kondoh T, and Fukuda M. Immunohistochemical localization of inducible nitric oxide synthase in synovial tissue of human temporomandibular joints with internal derangement. Arch Oral Biol 46: 9397, 2001.[CrossRef][ISI][Medline]
- Jones N, Iljin K, Dumont DJ, and Alitalo K. Tie receptors: new modulators of angiogenic and lymphangiogenic responses. Nat Rev Mol Cell Biol 2: 257267, 2001.[CrossRef][ISI][Medline]
- Jones N, Master Z, Jones J, Bouchard D, Gunji Y, Sasaki H, Daly R, Alitalo K, and Dumont DJ. Identification of Tek/Tie2 binding partners. Binding to a multifunctional docking site mediates cell survival and migration. J Biol Chem 274: 3089630905, 1999.[Abstract/Free Full Text]
- Jones N, Voskas D, Master Z, Sarao R, Jones J, and Dumont DJ. Rescue of the early vascular defects in Tek/Tie2 null mice reveals an essential survival function. EMBO Rep 2: 438445, 2001.[Abstract/Free Full Text]
- Jones PF. Not just angiogenesiswider roles for the angiopoietins. J Pathol 201: 515527, 2003.[CrossRef][ISI][Medline]
- Kingsley PD, Malik J, Fantauzzo KA, and Palis J. Yolk sac-derived primitive erythroblasts enucleate during mammalian embryogenesis. Blood 104: 1925, 2004.[Abstract/Free Full Text]
- Kontos CD, Stauffer TP, Yang WP, York JD, Huang L, Blanar MA, Meyer T, and Peters KG. Tyrosine 1101 of Tie2 is the major site of association of p85 and is required for activation of phosphatidylinositol 3-kinase and Akt. Mol Cell Biol 18: 41314140, 1998.[Abstract/Free Full Text]
- Lash AE, Tolstoshev CM, Wagner L, Schuler GD, Strausberg RL, Riggins GJ, and Altschul SF. SAGEmap: a public gene expression resource. Genome Res 10: 10511060, 2000.[Abstract/Free Full Text]
- Leung SK and Ohh M. Playing Tag with HIF: The VHL Story. J Biomed Biotechnol 2: 131135, 2002.[CrossRef][Medline]
- Li J, Billiar TR, Talanian RV, and Kim YM. Nitric oxide reversibly inhibits seven members of the caspase family via S-nitrosylation. Biochem Biophys Res Commun 240: 419424, 1997.[CrossRef][ISI][Medline]
- Lofstedt T, Jogi A, Sigvardsson M, Gradin K, Poellinger L, Pahlman S, and Axelson H. Induction of ID2 expression by hypoxia-inducible factor-1: a role in dedifferentiation of hypoxic neuroblastoma cells. J Biol Chem 279: 3922339231, 2004.[Abstract/Free Full Text]
- Mungrue IN, Bredt DS, Stewart DJ, and Husain M. From molecules to mammals: what's NOS got to do with it? Acta Physiol Scand 179: 123135, 2003.[CrossRef][ISI][Medline]
- Mungrue IN, Husain M, and Stewart DJ. The role of NOS in heart failure: lessons from murine genetic models. Heart Fail Rev 7: 407422, 2002.[CrossRef][Medline]
- Nath AK, Enciso J, Kuniyasu M, Hao XY, Madri JA, and Pinter E. Nitric oxide modulates murine yolk sac vasculogenesis and rescues glucose induced vasculopathy. Development Disease 131: 24852496, 2004.
- Papapetropoulos A, Garcia-Cardena G, Dengler TJ, Maisonpierre PC, Yancopoulos GD, and Sessa WC. Direct actions of angiopoietin-1 on human endothelium: evidence for network stabilization, cell survival, and interaction with other angiogenic growth factors. Lab Invest 79: 213223, 1999.[ISI][Medline]
- Patan S. Vasculogenesis and angiogenesis as mechanisms of vascular network formation, growth and remodeling. J Neurooncol 50: 115, 2000.[CrossRef][ISI][Medline]
- Risau W. Mechanisms of angiogenesis. Nature 386: 671674, 1997.[CrossRef][ISI][Medline]
- Risau W and Flamme I. Vasculogenesis. Annu Rev Cell Dev Biol 11: 7391, 1995.[CrossRef][ISI][Medline]
- Rossig L, Fichtlscherer B, Breitschopf K, Haendeler J, Zeiher AM, Mulsch A, and Dimmeler S. Nitric oxide inhibits caspase-3 by S-nitrosation in vivo. J Biol Chem 274: 68236826, 1999.[Abstract/Free Full Text]
- Rozen S and Skaletsky H. Primer3 on the WWW for general users and for biologist programmers. In: Bioinformatics Methods and Protocols: Methods in Molecular Biology, edited by Krawetz S and Misener S. Totowa, NJ: Humana, 2000, p. 365386.
- Ruijter JM, Van Kampen AH, and Baas F. Statistical evaluation of SAGE libraries: consequences for experimental design. Physiol Genomics 11: 3744, 2002.[Abstract/Free Full Text]
- Sato TN, Tozawa Y, Deutsch U, Wolburg-Buchholz K, Fujiwara Y, Gendron-Maguire M, Gridley T, Wolburg H, Risau W, and Qin Y. Distinct roles of the receptor tyrosine kinases Tie-1 and Tie-2 in blood vessel formation. Nature 376: 7074, 1995.[CrossRef][ISI][Medline]
- Schwarz F, Duka A, Triantafyllidi E, Johns C, Duka I, Cui J, and Gavras H. Serial analysis of gene expression in mouse kidney following angiotensin II administration. Physiol Genomics 16: 9098, 2003.[Abstract/Free Full Text]
- Tait CR and Jones PF. Angiopoietins in tumours: the angiogenic switch. J Pathol 204: 110, 2004.[CrossRef][ISI]
- Veikkola T and Alitalo K. VEGFs, receptors and angiogenesis. Semin Cancer Biol 9: 211220, 1999.[CrossRef][ISI][Medline]
- Velculescu VE, Zhang L, Vogelstein B, and Kinzler KW. Serial analysis of gene expression. Science 270: 484487, 1995.[Abstract]
- Velculescu VE, Madden SL, Zhang L, Lash AE, Yu J, Rago C, Lal A, Wang CJ, Beaudry GA, Ciriello KM, Cook BP, Dufault MR, Ferguson AT, Gao Y, He TC, Hermeking H, Hiraldo SK, Hwang PM, Lopez MA, Luderer HF, Mathews B, Petroziello JM, Polyak K, Zawel L, Zang W, Zang X, Zhou W, Haluska FG, Jen J, Sukumar S, Landes GM, Riggins GR, Vogelstein B, and Kinzler KW. Analysis of human transcriptomes. Nat Genet 23: 387388, 1999.[CrossRef][ISI][Medline]
- Ward NL and Dumont DJ. The angiopoietins and Tie2/Tek: adding to the complexity of cardiovascular development. Semin Cell Dev Biol 13: 1927, 2002.[CrossRef][ISI][Medline]
- Wilting J and Christ B. Embryonic angiogenesis: a review. Naturwissenschaften 83: 153164, 1996.[CrossRef][ISI][Medline]
- Yamakawa H, Jezova M, Ando H, and Saavedra JM. Normalization of endothelial and inducible nitric oxide synthase expression in brain microvessels of spontaneously hypertensive rats by angiotensin II AT1 receptor inhibition. J Cereb Blood Flow Metab 23: 371380, 2003.[CrossRef][ISI][Medline]
- Zhang B, Schmoyer D, Kirov S, and Snoddy J. GOTree Machine (GOTM): a web-based platform for interpreting sets of interesting genes using Gene Ontology hierarchies. BMC Bioinformatics 5: 16, 2004.[CrossRef][Medline]
- Zheng H, Wasylyk C, Ayadi A, Abecassis J, Schalken JA, Rogatsch H, Wernert N, Maira SM, Multon MC, and Wasylyk B. The transcription factor Net regulates the angiogenic switch. Genes Dev 17: 22832297, 2003.[Abstract/Free Full Text]