Enrichment of Genes in the Aortic Intima That Are Associated With Stratified Epithelium
http://www.100md.com
《循环学杂志》
the Department of Pathology (P.P.Y., A.A.T.), Vanderbilt University Medical Center, Nashville, Tenn
the Department of Pathology and Immunology (V.M., J.H.L.), Washington University School of Medicine, St Louis, Mo.
Abstract
Background— Arteries and veins are exposed to different pressures and are easily distinguished by morphology. Although several recent studies have focused on differential gene expression between the arterial and venous endothelium, the molecular distinctions that give rise to the dramatic structural distinctions between arteries and veins, such as in the organization of the intima, are not known.
Methods and Results— We used high-density oligonucleotide arrays to analyze the transcriptional profile of the mouse aorta and inferior vena cava (IVC), not restricting our analysis to the endothelium, to identify genes whose expression was enriched in aorta over other tissues and the IVC. By quantitative reverse transcription–polymerase chain reaction analysis, these genes have been shown to be highly expressed in the mouse aorta and were either expressed at low levels or were undetectable in the murine IVC. By immunofluorescence analysis of human tissue, we determined that a subset of these aorta-enriched proteins exhibited a primarily intima-restricted expression. Intimal expression of at least a subset of these genes, plakoglobin, galectin 7, sciellin, and SPRR3, was also detected in other types of arteries but not in veins. Furthermore, SPRR3 expression in the intima was primarily associated with atheromas. The proteins identified are functionally related in that they are known to also be enriched in stratified epithelia, where they play an important role in stress-bearing and barrier properties.
Conclusions— Vascular expression of these genes has not been reported previously. Our observations suggest that they may play a significant role in the mechanisms by which large arteries may adapt to biomechanical stress.
Key Words: hemodynamics ; arteries ; veins ; plaque ; blood flow
Introduction
Large-diameter arterial vessels such as the aorta are exposed to high-pressure flow and, in contrast to veins, characteristically are surrounded by a thick intimal and medial layer (Figure 1).1,2 The vena cava and other veins have comparatively thin intima and media and sustain dramatically less pressure.1 In fact, the predominant vascular feature when a vein is exposed to arterial flow is smooth muscle cell proliferation and matrix deposition that results in intimal thickening.3,4 It is widely believed that at least some aspects of the differences in the functional phenotype of vascular components, including the vascular endothelium and intimal smooth muscle cells in arteries and veins, are related to these hemodynamic considerations.5,6 Furthermore, many vascular diseases are limited to either arteries or veins. For example, the earliest atherosclerotic lesions are limited to large elastic and muscular arteries.1,7,8 Consequently, one of the major initiatives of vascular biology is to understand this functional and anatomic heterogeneity at a molecular level.7,8 Because so little is known about the molecular differences between arteries and veins, we used high-density oligonucleotide arrays to analyze the transcriptional profile of the mouse aorta and inferior vena cava (IVC) to identify genes whose expression was enriched in aorta over other tissues and the IVC.
Methods
RNA Isolation and cDNA Synthesis
Vascular tissue was obtained from 6 C57BL/6 mice (Jackson Labs), aged 4 to 8 weeks, by careful dissection of the aorta (aortic arch to thoracic aorta) and IVC (the segment between the right atrium to just superior to the liver). The vascular samples were pooled before isolation of RNA. Various other tissues, including liver, spleen, kidney, brain, skeletal muscle, lung, pancreas, heart, and skin, were also obtained from 2 to 6 separate sets of C57BL/6 mice, pooled, flash-frozen, and processed to isolate RNA. Total RNA was isolated with TRIzol LS reagent (Invitrogen). Integrity of the RNA was confirmed by RNA with an A260/A280 ratio >1.9 and 28S/18S ratio of >1.4 (Agilent 2100 Bioanalyzer RNA 6000 LabChip kit).
Microarray Analysis
the total RNA, biotinylated cRNA probes were generated, fragmented, and applied to Mouse MU74A (version 2) GeneChip arrays for aorta, skin, IVC, and heart; version 1 was used for other tissues, including heart (Affymetrix). The overall fluorescence intensity across each chip was scaled to 1500, and comparisons of the levels of transcripts were performed with Affymetrix analysis software, Microarray Suite 5.0 (MAS 5.0). A recent analysis found the same level of discordance across other algorithms used for data analysis as MAS 5.0.9–12 Hence, the present data analysis used only MAS 5.0 software, which has implemented algorithms for data analysis that are based on nonparametric statistical techniques. The detection probability value cutoffs to define "present" (P), "absent" (A), and "marginal" (M) were the default settings (any probability value falling below 0.04 is assigned "P"; above 0.06, "A"; and those between, "M").11 With the exception of heart tissue, which was run both on version 1 and version 2 chips, all other samples were analyzed by single microarray chip. The data obtained with the MAS 5.0 algorithms/statistical tests were imported into Spotfire DecisionSite software to examine expression across various tissues. Absolute calls (restricted to P, M, and A) were used to examine the frequency of present calls and further filter the data for the absolute calls with query devices.13 Spotfire DecisionSite does not distinguish between origins of data; in the present analysis, this tool was used to filter the data on the basis of the call metric assigned by MAS 5.0.13 No further normalization of data was performed. Genes for which expression was present in aorta and absent in other tissues queried (including IVC) were the main focus of the present analysis.
Quantitative Reverse Transcription–Polymerase Chain Reaction Analysis
cDNA was prepared from pooled samples of total RNA from aorta, IVC, esophagus, and skin. Each pooled sample represents tissue from 2 to 6 mice, and quantitative reverse transcription–polymerase chain reaction (qRT-PCR) analysis was performed in triplicate for each pool and averaged (n=1). qRT-PCR data for aorta and IVC represent the mean value of qRT-PCR data from 6 distinct pools of RNA (n=6) from 6 different groups of mice. qRT-PCR data from esophagus and skin represent the mean from 3 distinct pools of RNA from 3 different groups of mice (n=3). Each pooled sample was prepared and assayed independently. For aorta and IVC, 3 of 6 pools were obtained from 2- to 3-month-old mice and 3 pools were from 1-, 6-, and 12-month-old mice. For esophagus and skin, all pooled samples (n=3) were obtained from 2- to 3-month-old mice. We observed no significant age-dependent changes in gene expression. Therefore, data are expressed as n=6 for aorta and IVC and n=3 for esophagus and skin. Total RNA was purified on RNeasy mini column (Qiagen) per the manufacturer’s protocol. cDNA synthesis was performed with random primers and the Superscript First-Strand Synthesis System for RT-PCR (Invitrogen). SYBR green qRT-PCR was performed with 1 μL of cDNA reverse transcribed, 100 nmol/L of gene-specific primers, and SYBR green master mix (Applied Biosystems) in an Applied Biosystems ABI Prism 7700 sequence detector according to the manufacturer’s protocol. Four or 5 serial dilutions of cDNA were used to create a standard curve from cycle thresholds to allow comparisons between samples. A standard curve of cycle thresholds for 18S ribosomal RNA transcript was used to normalize data between groups. All primer sets were analyzed with RT+cDNA, RT-cDNA, and no template control before experimental use. The average cycle threshold of qRT-PCR reactions for an experiment was used to quantify the amount of mRNA relative to the standard curve by previously described methods.14 The average fold induction and standard error for the distinct genes were calculated for each condition.14 Data are presented as mean of the data obtained from separate experiments plus or minus SEM. Student t tests were performed to compare different data sets. A probability value of <0.05 was interpreted to denote statistical significance. Gene-specific primers were designed with Primer Express 1.0 (Applied Biosystems) and spanned an intron for all genes studied. Amplification of a single polymerase chain reaction (PCR) product of the expected size was confirmed by polyacrylamide gel electrophoresis and with the dissociation curve. Sequences for the primer sets are available on request.
Immunofluorescence
Segments of human large arteries and veins were obtained from 5 autopsies from both male and female subjects who did not have a vascular cause of death. Five-micrometer sections were prepared from formalin-fixed, paraffin-embedded tissue blocks or 10-μm sections were prepared from frozen tissue embedded in OCT. Immunofluorescence and immunohistochemistry on deparaffinized sections were performed after antigen retrieval in citrate buffer. The primary antibodies used recognized cytokeratin 13 (1:25, Dako), SPRR3 (1:200, Alexis Biochemicals), plakoglobin (1:200, Sigma), galectin 7 (1:200, a kind gift from Dr Thierry Magnaldo, Laboratory of Genetic Instability and Cancer, CNRS UPR2169, Institut Gustave Roussy, Villejuif, France), sciellin (1:300, kind gift from Dr Howard Baden, Department of Dermatology, Cutaneous Biology Research Center, Harvard Medical School, Massachusetts General Hospital, Boston, Mass), and desmocollin 1 (1:50, Research Diagnostics, Inc). Slides were washed and then incubated with appropriate secondary antibodies conjugated to fluorochromes (Molecular Probes). Slides were sealed with Vectashield mounting media (Vector Laboratories) with DAPI. SPRR3 and keratin 13 were also analyzed by immunohistochemical staining on a Ventana Gen System (Ventana Medical Systems) counterstained with hematoxylin. Control slides for mouse monoclonal antibodies were incubated with irrelevant mouse antibody of the same IgG subclass, whereas control slides for rabbit polyclonals were incubated with normal rabbit serum (Dako) before incubation with secondary antibodies. Slides were viewed on a Nikon Microphot-SA fluorescence microscope, and images were captured with a Colorview camera (Olympus) and AnalySIS software (Soft Imaging System).
Results and Discussion
Gene Expression Analysis Shows Overlap Between Skin and Aorta Expressed Genes
Of the genes queried (12 000 probe sets, 8619 unigene clusters), mRNAs represented by 271 probe sets were upregulated in the aorta over the IVC by 2-fold (raw data available in the Data Supplement). To determine which genes specifically discriminated for the aorta, we used Affymetrix and Spotfire software to identify a set of genes that were determined to be present in the aorta but absent in murine IVC, spleen, liver, kidney, brain, skeletal muscle, lung, pancreas, intestine, and heart. Transcripts represented by 93 probe sets were identified as present in the aorta and absent in the IVC and all other tissues used in the comparison. A subset of the probe sets (n=50), however, were absent from the Mu74A (version 1) used to analyze the other tissues (with the exception of the IVC, skin, and heart), and hence, their expression in these tissues could not be determined. In addition, other modifications of the probe sets present in the new version may have resulted in altered hybridization efficiency and, consequently, different absolute calls.15 To evaluate the systemic variability resulting from the use of 2 different versions of GeneChip arrays, we reanalyzed the labeled murine heart cRNA used for analysis on version 1 on the version 2 chip. This analysis demonstrated that most probe sets (>88%) were assigned the same absolute calls on both versions (data not shown). When only the number of "P" calls for every probe set was analyzed, there was better correlation (>94%). 2D scatterplots of signal intensity (obtained from MAS 5.0 analysis) from the 2 chips were plotted against each other. The degree of deviation of the spots from the main diagonal was interpreted as a measure of systemic variation of intensity of expression. A Pearson correlation coefficient of 0.82 was calculated when the comparison was restricted to probe sets with signal intensities >1000 (data not shown), which supports reports from other laboratories that have validated the use of microarray data from different platforms.15 Nevertheless, we realized that limiting our analysis to 1 chip for each set of pooled tissue (12 chips in all) limited the statistical confidence of the data; however, we used the microarray strategy only as the initial mechanism to tentatively identify genes of interest. All of the genes subsequently were analyzed by a more accurate experimental method on independently obtained, separate samples for quantifying differences in gene expression.
To our surprise, a striking number (22/93, 24%) of the genes expressed specifically in the aorta were previously reported to be enriched in skin and other stratified epithelia and represented 18 of the 40 highest-ranking signature genes in the experiment (Table 1).16,17 The various tissues used in our initial analysis to select for aorta-enriched genes did not contain stratified epithelia. To rule out the possibility that the significant overlap between the genes expressed in the aorta and skin/stratified epithelia was not merely the result of our analysis strategy, we performed microarray analysis on pooled mouse skin RNA. We additionally examined the subset of genes coexpressed in the aorta with each of the other tissues, eg, heart, lung, and spleen, but which were absent in all other remaining tissues. This analysis was expected to identify a list of organ-specific (such as cardiac or liver) signature genes that were also expressed in the aorta. The liver, spleen, intestine, brain, heart, and lung contained an overlap of only 2 or 3 genes with the aorta. Furthermore, none of these genes represented tissue-enriched signature genes. The pancreas and skeletal muscle contained no genes in common with the aorta by this analysis. By contrast, 15% (22/147) of the genes present in the skin but absent in IVC, spleen, liver, kidney, brain, skeletal muscle, lung, pancreas, and intestine were also present in the aorta, which confirms the unexpected prevalence discerned by our initial analysis (data not shown).
Common Features of Cornified Cell Envelope in Skin and Aorta
We chose to further pursue this novel aorta/stratified epithelium common gene association (Table 1). These genes fell into 2 broad but overlapping categories: (1) a group highly enriched in stratified epithelia (ie, skin), many of them mapped to the epidermal differentiation complex, a region that contains many genes vital to epidermal differentiation17,18; and (2) a group of adhesive junction–related genes. Interestingly, many of the identified genes that fell within the first category, such as cytokeratin 13, SPRRs, sciellin, and members of the S-100 calcium binding protein family, are known to be involved in the formation of the cornified cell envelope (CE).19 Terminally differentiating stratified squamous epithelial cells assemble a specialized structure on their periphery, termed "CE," that is composed of numerous structural proteins that become cross-linked by several transglutaminase enzymes into a macromolecular assembly that is critical to the mechanical and barrier properties of that tissue.20,21 Interestingly, aside from the enduring significantly greater biomechanical stress, the aorta also exhibits enhanced barrier function, as demonstrated by Evans blue dye exclusion experiments, compared with the IVC.22 Some CE precursors have been described in nonstratified epithelia that have barrier properties.23,24 For example, the SPRR family members also serve as cross-bridging proteins linked to desmosomal desmoplakin, and sciellin has been identified in simple epithelia with barrier properties, like the amnion, that lack a CE.25,26 In general, however, expression of many of the CE precursor proteins outside of stratified epithelia, specifically vascular tissue, has not been described.
The expression of several CE precursor genes in the aorta led us to further query the GeneChip data (or perform qRT-PCR if transcripts were not represented on chip) to identify the vascular expression profile of other known components of the CE (Table 2). Assembly of the CE is believed to initiate with the multimerization of periplakin, envoplakin, and involucrin, which are crosslinked by N(-glutamyl)lysine isopeptide cross-links by transglutaminase I (TG1).21,27 The formation of this scaffold is an early step in the assembly of CE structures of many different tissues that contain stratified squamous epithelia. Both periplakin and envoplakin were dramatically overexpressed in the aorta relative to the IVC (17- and 15-fold, respectively; Table 2). TG1 expression was not detected in IVC; expression in aorta was 2-fold lower than that detected in mouse skin or esophagus. Transglutaminase 3 (TG3), which also is involved in cross-linking during later steps of CE assembly,20 was detected in the aorta at approximately the same level as in the skin and 6-fold higher than levels in the IVC. Cross-linking of cytokeratins to the components of the CE (eg, SPRRs and envoplakin) is critical in the coordination and maintenance of epidermal structure.20,21 Cytokeratin 13 (Tables 1 and 2) was found at high levels only in murine aorta. Other known CE precursor proteins, such as involucrin, filaggrin, elafin, annexin I, repetin, cystatin A, trichohyalin, and loricrin,28 were not detected in murine aorta or IVC with the microarray analysis or by qRT-PCR (data not shown).
Cell Adhesive Junction–Related Genes Are Enriched in the Aorta Over IVC
A second subset of the skin-enriched genes identified in our microarray analysis represented components of adhesive junctions involved in cell-cell adhesion (Tables 1 and 2). We were especially surprised by the identification of desmocollin I in the aorta, because desmosomes have not been detected in ultrastructural analysis of large vessels.29 Expression of other desmosomal cadherins, desmocollin 2 and desmoglein 2, was also detected in low levels in both aorta and IVC, with similar signal intensity (data not shown), whereas desmocollin 3 was undetectable in both murine aorta and IVC. Plakophilin I is a member of the armadillo family of arm proteins that directly interact with cadherins located in both adherens junctions and desmosomes (Table 1).29 These junctions play an important role in providing strength to tissues that experience mechanical stress, such as heart and epidermis, and are highly expressed in these tissues.29–31 Plakoglobin belongs to the same family as plakophilin I, has a similar role in adhesive junctions, and has been shown to be important in interendothelial adhesion during shear stress.32 We determined whether expression of plakoglobin was also enriched in the aorta. Plakoglobin expression in the murine aorta by microarray analysis was only modestly higher than in the IVC (2-fold; Table 2).
qRT-PCR Confirms Microarray Results
For a subset of the genes, qRT-PCR was used as an independent method to confirm the gene expression differences between the artery and IVC and to compare the relative expression in the murine aorta to other tissues that contain cornified stratified epithelia, such as the esophagus and skin (Table 2). For the majority of the genes tested that were called "absent" in the IVC by microarray analysis, the transcript was undetectable by qRT-PCR with a separate set of experimental samples (Table 2); however, low levels of galectin 7 expression (27-fold less than that in the aorta) were detected in the IVC by qRT-PCR. The precise function of galectin 7 is not yet known; however, its expression is linked to differentiation of skin.33 The relative expression of many of these skin-enriched genes in the aorta relative to mouse skin by qRT-PCR analysis varied (Table 2), with generally higher expression in murine skin; however, mRNA expression of SPRR3 was higher in aorta than in skin. Interestingly, the level of expression of many of the genes involved in CE in the murine aorta was comparable to expression levels detected in the murine esophagus, a tissue with a defined CE (Table 2).
The magnitude of overexpression determined by qRT-PCR was similar to that determined from the signal intensity on the microarray for most of the genes, except plakoglobin. Relative expression of plakoglobin in the aorta was 8-fold higher than in the IVC by qRT-PCR, whereas the difference was only 2-fold by microarray analysis (Table 1). Desmocollin I was detected in the aorta only at a relatively low level (signal intensity of 439). This is supported by the finding that expression in the aorta was >200-fold less than in murine skin. The difference of expression of each gene examined by qRT-PCR between the artery and vein was statistically significant.
This high level of correlation observed between the array and qRT-PCR data was anticipated given that the majority of aorta-enriched genes identified by our analysis represented transcripts detected with very high signal intensity. Studies from other laboratories have shown very strong reproducibility among this set.12,34 We also tested for a number of other functionally related transcripts that were not identified in our analysis (either because they were absent in the chip, eg, envoplakin, or because they were also expressed in other tissues, eg, plakoglobin transcripts were also detected in murine heart). In each case, we were able to confirm aorta-enriched expression, which provides further support for the association between the aorta and this group of skin-enriched proteins.
Protein Expression of Selected Murine Aorta-Enriched Genes in Human Aorta and Other Arteries
Immunofluorescence and immunohistochemistry analysis with vascular tissue obtained from human autopsy samples was used to confirm whether protein expression was restricted only to the aorta and also to determine the vascular localization of a subset of these proteins. In each analysis (except for cytokeratin 13), the same slide was also stained with the nucleic acid dye DAPI to determine cellular localization within the tissue. Expression of sciellin, plakoglobin, galectin 7, SPRR3 (Figures 1 and 2), and cytokeratin 13 (Data Supplement Figure I) was restricted to the human aorta, not IVC (Data Supplement Figure II). Expression of sciellin (Figures 1c and 1e) and cytokeratin 13 (Data Supplement Figure I) was restricted to the intima of the aorta, with the majority of staining in the portion of the intima closest to the lumen. An antibody specific for plakoglobin and galectin 7 stained cells throughout the intima (Figures 1g, 2a, and 2c). No specific staining was seen in the media or adventitia of the aorta for sciellin, cytokeratin 13, or galectin 7 (data not shown). Staining was absent in the human IVC with antibodies against sciellin, galectin 7, and SPRR3 (Data Supplement Figure II). Although staining in the IVC was primarily absent with antibody to plakoglobin (data not shown), several isolated areas of punctate immunoreactivity were detected, particularly in regions in which the intima was thickened (Figure 1i). This sporadic staining pattern was confirmed on 3 independent experiments; nonspecific staining due to autofluorescence was ruled out by careful examination of the other channels for fluorescence (data not shown). We noted that SPRR3 showed discontinuous staining throughout the aortic intima. To study the localization of SPRR3 further, we performed immunohistochemistry on various human arteries and veins. Interestingly, localization of SPRR3 was limited to atheromatous plaques in several artery-derived vessels and was not detectable in veins (Figure 3).
Genes Present in IVC but Not Aorta or Other Tissues
Complementary analysis was performed to identify genes expressed in the IVC but absent in the aorta and all other tissues (excluding skin) used in the comparison presented for the aorta study. Transcripts represented by 17 probe sets were identified as present in the IVC and absent in the aorta and in murine spleen, liver, kidney, brain, skeletal muscle, lung, pancreas, intestine, and heart. Of these, all but 2 had signal intensities below 1000 in the IVC. This is noteworthy because "P" calls made by MAS 5.0 perform well at high concentrations but are much less reliable at the low concentrations reflected by the low signal intensity.12,34 We could not discern any functional relationships among these genes (Table 3).
Comparison of Adult Aorta Gene Expression With Reported Artery/Vein Specified Genes
There are only a few known artery- or vein-specific genes.7,8 EphrinB2 is expressed in developing arteries, whereas EphB4 is expressed in veins.35,36 However, analysis of the vena cava of adult mice in whole mount revealed patchy staining of the lacZ reporter targeted to the ephrinB2 locus.35,36 Expression of EphB4, which is considered specific for venous endothelium, was detected in lower levels in the wall of the adult dorsal aorta.35 More recently, Lawson and colleagues37,38 and Zhong et al39 have described expression of the Notch and Gridlock family of genes in precursors of endothelial cells, in which they promote arterial fate and repress venous differentiation. Unlike the ephrin genes, these and other differentially expressed genes have pronounced differences that exist only during early development.38 The expression profile of some of these genes in our microarray analysis is provided in Table 4 and did not demonstrate restricted expression.
Several recent reports have also used microarray analysis to identify the molecular basis for the structural and functional diversity in vessels, particularly in endothelial cells.40–43 Many of these studies have been aimed at the identification of endothelial genes that are regulated by arterial levels of shear stress using in vitro models.40–43 Because the present study used whole-vessel homogenates, not just endothelium, we suspect that our platform lacks the sensitivity to identify the distinctions between the arterial and venous endothelium reported by others. Hence, the message detection levels of many of these genes were detected both in the aorta and IVC at similar levels in our analysis (Table 4).
Using gene array to compare the gene expression profile of the aorta and IVC, we have identified several genes that are expressed specifically in the arterial but not venous intima. The tunica intima is composed of a single layer of endothelium that lines the lumen, smooth muscle cells, and connective tissue; the major cell type of the intima is the intimal smooth muscle cell.1 A surprising number of these arterial intima–enriched genes are also enriched in stratified epithelia, such as the epidermis.17 A number of these represented precursors of the CE and proteins involved in adhesive junctions.17,21 The identification of such a large number of known CE precursor proteins and proteins involved in cell-cell adhesion in the arterial intima suggests that they may play a role in the long-recognized enhanced barrier and stress-bearing properties of arteries compared with veins.22 Expression of SPRR3, specifically, appears to be associated with atherosclerotic plaques. The biomechanical disruptions around atherosclerotic plaques have long been described but are not well understood.44
Although the aorta has a number of CE precursors, it lacks some of the important components, such as involucrin, which is suggested to be a ubiquitous component of CEs in all stratified epithelia.20 Together with the fact that a CE has not been morphologically demonstrated in the aorta or large arteries, it is unlikely that the aorta contains a classic CE; however, these data raise the possibility that it may contain a less well-developed CE. Alternatively, these precursors may be participating in the cytoplasmic cross-linking of structural proteins. The human phenotype of known mutations in a few CE precursors usually manifests as various skin deformities, such as ichthyosis and hyperkeratosis.20 Although vascular defects in such conditions have not been appreciated and linked yet, it will be interesting to study such patients to determine whether there is greater predisposition to arterial pathology, such as atherosclerosis or aneurysms.
Interestingly, the aortic intima at birth is limited to a lamina of endothelial cells that closely abuts the internal elastic lamina.1 During growth and development, there is proliferation and development of the intima in the large arteries, and in the aorta, it expands to 0.2 mm in adults.1 It is hypothesized that biomechanical stress represents an important stimulus for differentiation of not just the endothelium but the entire intima.1,5 Future studies to examine the role of arterial intima–enriched proteins in intimal smooth muscle cells during vascular adaptation to biomechanical stress are under way.
Acknowledgments
This work was supported by a Vanderbilt physician scientist development grant and institutional funds from Washington University School of Medicine.
Footnotes
The online-only Data Supplement for this article can be found with this article at http://circ.ahajournals.org.
References
Stehbens WE. Structural and architectural changes during arterial development and the role of hemodynamics. Acta Anat. 1996; 157: 261–274.
Mills I, Sumpio BE. Vascular Smooth Muscle Cells. Armonk, NY: Futura Publishing Co; 1997.
Glagov S, Zarins CK, Masawa N, Xu CP, Bassiouny H, Giddens DP. Mechanical functional role of non-atherosclerotic intimal thickening. Front Med Biol Eng. 1993; 5: 37–43.
Zubileqicz T, Wronski J, Bourriez A, Terlecki P, Guinault AM, Muscatelli-Groux B, Michalak J, Melliere D, Becquemin JP, Allaire E. Injury in vascular surgery: the intimal hyperplastic response. Med Sci Monit. 2001; 7: 316–324.
Fry DL. Arterial intimal-medial permeability and coevolving structural responses to defined shear-stress exposures. Am J Physiol Heart Circ Physiol. 2002; 283: H2341–H2355.
Schwartz SM, Majesky MW, Murry CE. The intima: development and monoclonal responses to injury. Atherosclerosis. 1995; 118: S125–S140.
Cleaver O, Melton DA. Endothelial signaling during development. Nat Med. 2003; 9: 661–668.
Carmeliet P. Angiogenesis in health and disease. Nat Med. 2003; 9: 653–660.
Tan PK, Downey TJ, Spitznagel EL, Xu P, Fu D, Dimitrov DS, Lempicki RA, Raaka BM, Cam MC. Evaluation of gene expression measurements from commercial microarray platforms. Nucleic Acids Res. 2003; 31: 5676–5684.
Li C, Wong W. Model-based analysis of oligonucleotide arrays: expression index computation and outlier detection. Proc Natl Acad Sci U S A. 2001; 98: 31–36.
Affymetrix: Microarray Suite User Guide V. Affymetrix; 2001. Available at: http://www.affymetrix.com/support/technical/manuals/.affx.
Shippy R, Sendera TJ, Lockner RL, Chockalingam P, Kaysser-Kranich T, Watts G, Alsobrook J. Performance evaluation of commercial short-oligonucleotide microarrays and the impact of noise in making cross-platform correlations. BMC Genomics. 2004; 5: 61–81.
Dressen IM, Husing J, Kruse E, Boes T, Jockel JH. Software packages for quantitative microarray-based gene expression analysis. Curr Pharm Biotechnol. 2003; 4: 417–437.
Pfaffl MW. A new mathematical model for relative quantification in real-time RT-PCR. Nucleic Acids Res. 2001; 29: 2004–2007.
Nimgaonkar A, Sanoudou D, Butte A, Haslett JN, Kunkel LM, Beggs AH, Kohane IS. Reproducibility of gene expression across generations of Affymetrix microarrays. BMC Bioinformatics. 2003; 4: 27–39.
Curto EV, Lambert GW, Davis RL, Wilborn TW, Dooley TP. Biomarkers of human skin cells identified using dermarray DNA arrays and new bioinformatics methods. Biochem Biophys Res Commun. 2002; 291: 1052–1064.
Marenholz I, Zirra M, Fischer DF, Backendorf C, Ziegler A, Mischke D. Identification of human epidermal differentiation complex (EDC)-encoded genes by subtractive hybridization of entire YACs to a gridded keratinocyte cDNA library. Genome Res. 2001; 11: 341–355.
Gibbs S, Funeman R, Wiegant J, Geurts van Kessel A, Putte P, Backendorf C. Molecular characterization and evolution of the SPRR family of keratinocyte differentiation markers encoding small proline-rich proteins. Genomics. 1993; 16: 630–637.
Jarnik M, Kartasova T, Steinert PM, Lichti U, Steven AC. Differential expression and cell envelope incorporation of small proline-rich protein 1 in different cornified epithelia. J Cell Science. 1996; 109: 1381–1391.
Nemes Z, Steinert PM. Bricks and mortar of the epidermal barrier. Exp Mol Med. 1999; 31: 5–19.
Kalinin AE, Kajava AV, Steinert PM. Epithelial barrier function: assembly and structural features of the cornified cell envelope. Bioessays. 2002; 24: 789–800.
Kwei S, Stavrakis G, Takahas M, Taylor G, Folkman MJ, Gimbrone MA, Garcia-Cardena G. Early adaptive responses of the vascular wall during venous arterialization in mice. Am J Pathol. 2004; 164: 81–89.
Bonilla IE, Tanabe K, Strittmatter SM. Small proline-rich repeat protein 1A is expressed by axotomized neurons and promotes axonal outgrowth. J Neurosci. 2002; 22: 1303–1315.
Cabral A, Voskamp P, Cleton-Jansen A, South A, Nizetic D, Backendorf C. Structural organization and regulation of the small proline-rich family of cornified envelope precursors suggest a role in adaptive barrier function. J Biol Chem. 2001; 26: 19231–19237.
Champliaud M, Baden HP, Koch M, Jin W, Burgeson RE, Viel A. Gene characterization of sciellin (SCEL) and protein localization in vertebrate epithelia displaying barrier properties. Genomics. 2000; 70: 264–268.
Baden HP, Kvedar JC. Epithelial cornified envelope precursors are in the hair follicle. J Invest Dermatol. 1993; 101: 72S–74S.
Karashima T, Watt FM. Interaction of periplakin and envoplakin with intermediate filaments. J Cell Sci. 2002; 115: 5027–5037.
Marshall D, Hardman MJ, Nield KM, Byrne C. Differentially expressed late constituents of the epidermal cornified envelope. Proc Natl Acad Sci U S A. 2001; 98: 13031–13036.
Vestweber D. Molecular mechanisms that control endothelial cell contacts. J Pathol. 2000; 190: 281–291.
Schnittler HJ, Puschel B, Drenchkhahn D. Role of cadherins and plakoglobin in interendothelial adhesion under resting conditions and shear stress. Am J Physiol Heart Circ Physiol. 1997; 273: H2396–H2405.
Troyanovsky RB, Chitaev NA, Troyanovsky SM. Cadherin binding sites of plakoglobin: localization, specificity and role in targeting to adhering junctions. J Cell Sci. 1996; 109: 3069–3078.
Hoffmann I, Mertens C, Brettel M, Nimmrich V, Schnolzer M, Herrmann H. Interaction of plakophilins with desmoplakin and intermediate filament proteins: an in vitro analysis. J Cell Sci. 2000; 113: 2471–2483.
Timmons PM, Colnot C, Cail I, Poirier F, Magnaldo T. Expression of galectin-7 during epithelial development coincides with the onset of stratification. Int J Dev Biol. 1999; 43: 229–235.
McShane LM, Shih JH, Michalowska AM. Statistical issues in the design and analysis of gene expression microarray studies of animal models. J Mammary Gland Biol Neoplasia. 2003; 8: 359–374.
Gale NW, Baluk P, Pan L, Kwan M, Holash J, DeChiara TM, McDonald DM, Yancopoulos GD. Ephrin-B2 selectively marks arterial vessels and neovascularization sites in the adult, with expression in both endothelial and smooth muscle cells. Develop Biol. 2001; 230: 151–160.
Shin D, Garcia-Cardena G, Hayashi SI, Gerety S, Asahara T, Stavrakis G, Isner J, Folkman J, Gimbrone MA, Anderson DJ. Expression of EphrinB2 identifies a stable genetic difference between arterial and venous vascular smooth muscle as well as endothelial cells, and marks subsets of microvessels at sites of adult neovascularization. Develop Biol. 2001; 230: 139–150.
Lawson ND, Vogel AM, Weinstein BM. Sonic hedgehog and vascular endothelial growth factor acts upstream of the notch pathway during arterial endothelial differentiation. Develop Cell. 2002; 3: 127–136.
Lawson ND, Weinstein BM. Arteries and veins: making a difference with zebrafish. Nat Rev. 2002; 3: 674–682.
Zhong TP, Childs S, Leu JP, Fishman MC. Gridlock signalling pathway fashions the first embryonic artery. Nature. 2001; 414: 216–220.
Garcia-Cardena G, Comander J, Anderson DJ, Blackman BR, Gimbrone MA. Biomechanical activation of vascular endothelium as a determinant of its functional phenotype. Proc Natl Acad Sci U S A. 2001; 98: 4478–4485.
Dekker RJ, van Soest S, Fontijn RD, Salamanca S, de Groot PG, VanBavel E, Pannekoek H, Horrevoets AJ. Prolonged fluid shear stress induces a distinct set of endothelial cell genes, most specifically lung Kruppel-like factor (KLF2). Blood. 2002; 100: 1689–1698.
Chen BP, Li Y, Zhao Y, Chen K, Li S, Lao J, Yuan S, Shyy JY, Chien S. DNA microarray analysis of gene expression in endothelial cells in response to 24-h shear stress. Physiol Genomics. 2001; 7: 55–63.
Chi J, Chang HY, Haraldsen G, Jahnsen FL, Troyanskaya OG, Chang DS, Wang Z, Rockson SG, van de Rijn M, Botstein D, Brown PO. Endothelial cell diversity revealed by global expression profiling. Proc Natl Acad Sci U S A. 2003; 100: 10623–10628.
Salunke NV, Topoleski LDT, Humphrey JD, Mergner WJ. Compressive stress-relaxation of human atherosclerotic plaque. J Biomed Mater Res. 2001; 55: 236–241.(Pampee P. Young, MD, PhD;)
the Department of Pathology and Immunology (V.M., J.H.L.), Washington University School of Medicine, St Louis, Mo.
Abstract
Background— Arteries and veins are exposed to different pressures and are easily distinguished by morphology. Although several recent studies have focused on differential gene expression between the arterial and venous endothelium, the molecular distinctions that give rise to the dramatic structural distinctions between arteries and veins, such as in the organization of the intima, are not known.
Methods and Results— We used high-density oligonucleotide arrays to analyze the transcriptional profile of the mouse aorta and inferior vena cava (IVC), not restricting our analysis to the endothelium, to identify genes whose expression was enriched in aorta over other tissues and the IVC. By quantitative reverse transcription–polymerase chain reaction analysis, these genes have been shown to be highly expressed in the mouse aorta and were either expressed at low levels or were undetectable in the murine IVC. By immunofluorescence analysis of human tissue, we determined that a subset of these aorta-enriched proteins exhibited a primarily intima-restricted expression. Intimal expression of at least a subset of these genes, plakoglobin, galectin 7, sciellin, and SPRR3, was also detected in other types of arteries but not in veins. Furthermore, SPRR3 expression in the intima was primarily associated with atheromas. The proteins identified are functionally related in that they are known to also be enriched in stratified epithelia, where they play an important role in stress-bearing and barrier properties.
Conclusions— Vascular expression of these genes has not been reported previously. Our observations suggest that they may play a significant role in the mechanisms by which large arteries may adapt to biomechanical stress.
Key Words: hemodynamics ; arteries ; veins ; plaque ; blood flow
Introduction
Large-diameter arterial vessels such as the aorta are exposed to high-pressure flow and, in contrast to veins, characteristically are surrounded by a thick intimal and medial layer (Figure 1).1,2 The vena cava and other veins have comparatively thin intima and media and sustain dramatically less pressure.1 In fact, the predominant vascular feature when a vein is exposed to arterial flow is smooth muscle cell proliferation and matrix deposition that results in intimal thickening.3,4 It is widely believed that at least some aspects of the differences in the functional phenotype of vascular components, including the vascular endothelium and intimal smooth muscle cells in arteries and veins, are related to these hemodynamic considerations.5,6 Furthermore, many vascular diseases are limited to either arteries or veins. For example, the earliest atherosclerotic lesions are limited to large elastic and muscular arteries.1,7,8 Consequently, one of the major initiatives of vascular biology is to understand this functional and anatomic heterogeneity at a molecular level.7,8 Because so little is known about the molecular differences between arteries and veins, we used high-density oligonucleotide arrays to analyze the transcriptional profile of the mouse aorta and inferior vena cava (IVC) to identify genes whose expression was enriched in aorta over other tissues and the IVC.
Methods
RNA Isolation and cDNA Synthesis
Vascular tissue was obtained from 6 C57BL/6 mice (Jackson Labs), aged 4 to 8 weeks, by careful dissection of the aorta (aortic arch to thoracic aorta) and IVC (the segment between the right atrium to just superior to the liver). The vascular samples were pooled before isolation of RNA. Various other tissues, including liver, spleen, kidney, brain, skeletal muscle, lung, pancreas, heart, and skin, were also obtained from 2 to 6 separate sets of C57BL/6 mice, pooled, flash-frozen, and processed to isolate RNA. Total RNA was isolated with TRIzol LS reagent (Invitrogen). Integrity of the RNA was confirmed by RNA with an A260/A280 ratio >1.9 and 28S/18S ratio of >1.4 (Agilent 2100 Bioanalyzer RNA 6000 LabChip kit).
Microarray Analysis
the total RNA, biotinylated cRNA probes were generated, fragmented, and applied to Mouse MU74A (version 2) GeneChip arrays for aorta, skin, IVC, and heart; version 1 was used for other tissues, including heart (Affymetrix). The overall fluorescence intensity across each chip was scaled to 1500, and comparisons of the levels of transcripts were performed with Affymetrix analysis software, Microarray Suite 5.0 (MAS 5.0). A recent analysis found the same level of discordance across other algorithms used for data analysis as MAS 5.0.9–12 Hence, the present data analysis used only MAS 5.0 software, which has implemented algorithms for data analysis that are based on nonparametric statistical techniques. The detection probability value cutoffs to define "present" (P), "absent" (A), and "marginal" (M) were the default settings (any probability value falling below 0.04 is assigned "P"; above 0.06, "A"; and those between, "M").11 With the exception of heart tissue, which was run both on version 1 and version 2 chips, all other samples were analyzed by single microarray chip. The data obtained with the MAS 5.0 algorithms/statistical tests were imported into Spotfire DecisionSite software to examine expression across various tissues. Absolute calls (restricted to P, M, and A) were used to examine the frequency of present calls and further filter the data for the absolute calls with query devices.13 Spotfire DecisionSite does not distinguish between origins of data; in the present analysis, this tool was used to filter the data on the basis of the call metric assigned by MAS 5.0.13 No further normalization of data was performed. Genes for which expression was present in aorta and absent in other tissues queried (including IVC) were the main focus of the present analysis.
Quantitative Reverse Transcription–Polymerase Chain Reaction Analysis
cDNA was prepared from pooled samples of total RNA from aorta, IVC, esophagus, and skin. Each pooled sample represents tissue from 2 to 6 mice, and quantitative reverse transcription–polymerase chain reaction (qRT-PCR) analysis was performed in triplicate for each pool and averaged (n=1). qRT-PCR data for aorta and IVC represent the mean value of qRT-PCR data from 6 distinct pools of RNA (n=6) from 6 different groups of mice. qRT-PCR data from esophagus and skin represent the mean from 3 distinct pools of RNA from 3 different groups of mice (n=3). Each pooled sample was prepared and assayed independently. For aorta and IVC, 3 of 6 pools were obtained from 2- to 3-month-old mice and 3 pools were from 1-, 6-, and 12-month-old mice. For esophagus and skin, all pooled samples (n=3) were obtained from 2- to 3-month-old mice. We observed no significant age-dependent changes in gene expression. Therefore, data are expressed as n=6 for aorta and IVC and n=3 for esophagus and skin. Total RNA was purified on RNeasy mini column (Qiagen) per the manufacturer’s protocol. cDNA synthesis was performed with random primers and the Superscript First-Strand Synthesis System for RT-PCR (Invitrogen). SYBR green qRT-PCR was performed with 1 μL of cDNA reverse transcribed, 100 nmol/L of gene-specific primers, and SYBR green master mix (Applied Biosystems) in an Applied Biosystems ABI Prism 7700 sequence detector according to the manufacturer’s protocol. Four or 5 serial dilutions of cDNA were used to create a standard curve from cycle thresholds to allow comparisons between samples. A standard curve of cycle thresholds for 18S ribosomal RNA transcript was used to normalize data between groups. All primer sets were analyzed with RT+cDNA, RT-cDNA, and no template control before experimental use. The average cycle threshold of qRT-PCR reactions for an experiment was used to quantify the amount of mRNA relative to the standard curve by previously described methods.14 The average fold induction and standard error for the distinct genes were calculated for each condition.14 Data are presented as mean of the data obtained from separate experiments plus or minus SEM. Student t tests were performed to compare different data sets. A probability value of <0.05 was interpreted to denote statistical significance. Gene-specific primers were designed with Primer Express 1.0 (Applied Biosystems) and spanned an intron for all genes studied. Amplification of a single polymerase chain reaction (PCR) product of the expected size was confirmed by polyacrylamide gel electrophoresis and with the dissociation curve. Sequences for the primer sets are available on request.
Immunofluorescence
Segments of human large arteries and veins were obtained from 5 autopsies from both male and female subjects who did not have a vascular cause of death. Five-micrometer sections were prepared from formalin-fixed, paraffin-embedded tissue blocks or 10-μm sections were prepared from frozen tissue embedded in OCT. Immunofluorescence and immunohistochemistry on deparaffinized sections were performed after antigen retrieval in citrate buffer. The primary antibodies used recognized cytokeratin 13 (1:25, Dako), SPRR3 (1:200, Alexis Biochemicals), plakoglobin (1:200, Sigma), galectin 7 (1:200, a kind gift from Dr Thierry Magnaldo, Laboratory of Genetic Instability and Cancer, CNRS UPR2169, Institut Gustave Roussy, Villejuif, France), sciellin (1:300, kind gift from Dr Howard Baden, Department of Dermatology, Cutaneous Biology Research Center, Harvard Medical School, Massachusetts General Hospital, Boston, Mass), and desmocollin 1 (1:50, Research Diagnostics, Inc). Slides were washed and then incubated with appropriate secondary antibodies conjugated to fluorochromes (Molecular Probes). Slides were sealed with Vectashield mounting media (Vector Laboratories) with DAPI. SPRR3 and keratin 13 were also analyzed by immunohistochemical staining on a Ventana Gen System (Ventana Medical Systems) counterstained with hematoxylin. Control slides for mouse monoclonal antibodies were incubated with irrelevant mouse antibody of the same IgG subclass, whereas control slides for rabbit polyclonals were incubated with normal rabbit serum (Dako) before incubation with secondary antibodies. Slides were viewed on a Nikon Microphot-SA fluorescence microscope, and images were captured with a Colorview camera (Olympus) and AnalySIS software (Soft Imaging System).
Results and Discussion
Gene Expression Analysis Shows Overlap Between Skin and Aorta Expressed Genes
Of the genes queried (12 000 probe sets, 8619 unigene clusters), mRNAs represented by 271 probe sets were upregulated in the aorta over the IVC by 2-fold (raw data available in the Data Supplement). To determine which genes specifically discriminated for the aorta, we used Affymetrix and Spotfire software to identify a set of genes that were determined to be present in the aorta but absent in murine IVC, spleen, liver, kidney, brain, skeletal muscle, lung, pancreas, intestine, and heart. Transcripts represented by 93 probe sets were identified as present in the aorta and absent in the IVC and all other tissues used in the comparison. A subset of the probe sets (n=50), however, were absent from the Mu74A (version 1) used to analyze the other tissues (with the exception of the IVC, skin, and heart), and hence, their expression in these tissues could not be determined. In addition, other modifications of the probe sets present in the new version may have resulted in altered hybridization efficiency and, consequently, different absolute calls.15 To evaluate the systemic variability resulting from the use of 2 different versions of GeneChip arrays, we reanalyzed the labeled murine heart cRNA used for analysis on version 1 on the version 2 chip. This analysis demonstrated that most probe sets (>88%) were assigned the same absolute calls on both versions (data not shown). When only the number of "P" calls for every probe set was analyzed, there was better correlation (>94%). 2D scatterplots of signal intensity (obtained from MAS 5.0 analysis) from the 2 chips were plotted against each other. The degree of deviation of the spots from the main diagonal was interpreted as a measure of systemic variation of intensity of expression. A Pearson correlation coefficient of 0.82 was calculated when the comparison was restricted to probe sets with signal intensities >1000 (data not shown), which supports reports from other laboratories that have validated the use of microarray data from different platforms.15 Nevertheless, we realized that limiting our analysis to 1 chip for each set of pooled tissue (12 chips in all) limited the statistical confidence of the data; however, we used the microarray strategy only as the initial mechanism to tentatively identify genes of interest. All of the genes subsequently were analyzed by a more accurate experimental method on independently obtained, separate samples for quantifying differences in gene expression.
To our surprise, a striking number (22/93, 24%) of the genes expressed specifically in the aorta were previously reported to be enriched in skin and other stratified epithelia and represented 18 of the 40 highest-ranking signature genes in the experiment (Table 1).16,17 The various tissues used in our initial analysis to select for aorta-enriched genes did not contain stratified epithelia. To rule out the possibility that the significant overlap between the genes expressed in the aorta and skin/stratified epithelia was not merely the result of our analysis strategy, we performed microarray analysis on pooled mouse skin RNA. We additionally examined the subset of genes coexpressed in the aorta with each of the other tissues, eg, heart, lung, and spleen, but which were absent in all other remaining tissues. This analysis was expected to identify a list of organ-specific (such as cardiac or liver) signature genes that were also expressed in the aorta. The liver, spleen, intestine, brain, heart, and lung contained an overlap of only 2 or 3 genes with the aorta. Furthermore, none of these genes represented tissue-enriched signature genes. The pancreas and skeletal muscle contained no genes in common with the aorta by this analysis. By contrast, 15% (22/147) of the genes present in the skin but absent in IVC, spleen, liver, kidney, brain, skeletal muscle, lung, pancreas, and intestine were also present in the aorta, which confirms the unexpected prevalence discerned by our initial analysis (data not shown).
Common Features of Cornified Cell Envelope in Skin and Aorta
We chose to further pursue this novel aorta/stratified epithelium common gene association (Table 1). These genes fell into 2 broad but overlapping categories: (1) a group highly enriched in stratified epithelia (ie, skin), many of them mapped to the epidermal differentiation complex, a region that contains many genes vital to epidermal differentiation17,18; and (2) a group of adhesive junction–related genes. Interestingly, many of the identified genes that fell within the first category, such as cytokeratin 13, SPRRs, sciellin, and members of the S-100 calcium binding protein family, are known to be involved in the formation of the cornified cell envelope (CE).19 Terminally differentiating stratified squamous epithelial cells assemble a specialized structure on their periphery, termed "CE," that is composed of numerous structural proteins that become cross-linked by several transglutaminase enzymes into a macromolecular assembly that is critical to the mechanical and barrier properties of that tissue.20,21 Interestingly, aside from the enduring significantly greater biomechanical stress, the aorta also exhibits enhanced barrier function, as demonstrated by Evans blue dye exclusion experiments, compared with the IVC.22 Some CE precursors have been described in nonstratified epithelia that have barrier properties.23,24 For example, the SPRR family members also serve as cross-bridging proteins linked to desmosomal desmoplakin, and sciellin has been identified in simple epithelia with barrier properties, like the amnion, that lack a CE.25,26 In general, however, expression of many of the CE precursor proteins outside of stratified epithelia, specifically vascular tissue, has not been described.
The expression of several CE precursor genes in the aorta led us to further query the GeneChip data (or perform qRT-PCR if transcripts were not represented on chip) to identify the vascular expression profile of other known components of the CE (Table 2). Assembly of the CE is believed to initiate with the multimerization of periplakin, envoplakin, and involucrin, which are crosslinked by N(-glutamyl)lysine isopeptide cross-links by transglutaminase I (TG1).21,27 The formation of this scaffold is an early step in the assembly of CE structures of many different tissues that contain stratified squamous epithelia. Both periplakin and envoplakin were dramatically overexpressed in the aorta relative to the IVC (17- and 15-fold, respectively; Table 2). TG1 expression was not detected in IVC; expression in aorta was 2-fold lower than that detected in mouse skin or esophagus. Transglutaminase 3 (TG3), which also is involved in cross-linking during later steps of CE assembly,20 was detected in the aorta at approximately the same level as in the skin and 6-fold higher than levels in the IVC. Cross-linking of cytokeratins to the components of the CE (eg, SPRRs and envoplakin) is critical in the coordination and maintenance of epidermal structure.20,21 Cytokeratin 13 (Tables 1 and 2) was found at high levels only in murine aorta. Other known CE precursor proteins, such as involucrin, filaggrin, elafin, annexin I, repetin, cystatin A, trichohyalin, and loricrin,28 were not detected in murine aorta or IVC with the microarray analysis or by qRT-PCR (data not shown).
Cell Adhesive Junction–Related Genes Are Enriched in the Aorta Over IVC
A second subset of the skin-enriched genes identified in our microarray analysis represented components of adhesive junctions involved in cell-cell adhesion (Tables 1 and 2). We were especially surprised by the identification of desmocollin I in the aorta, because desmosomes have not been detected in ultrastructural analysis of large vessels.29 Expression of other desmosomal cadherins, desmocollin 2 and desmoglein 2, was also detected in low levels in both aorta and IVC, with similar signal intensity (data not shown), whereas desmocollin 3 was undetectable in both murine aorta and IVC. Plakophilin I is a member of the armadillo family of arm proteins that directly interact with cadherins located in both adherens junctions and desmosomes (Table 1).29 These junctions play an important role in providing strength to tissues that experience mechanical stress, such as heart and epidermis, and are highly expressed in these tissues.29–31 Plakoglobin belongs to the same family as plakophilin I, has a similar role in adhesive junctions, and has been shown to be important in interendothelial adhesion during shear stress.32 We determined whether expression of plakoglobin was also enriched in the aorta. Plakoglobin expression in the murine aorta by microarray analysis was only modestly higher than in the IVC (2-fold; Table 2).
qRT-PCR Confirms Microarray Results
For a subset of the genes, qRT-PCR was used as an independent method to confirm the gene expression differences between the artery and IVC and to compare the relative expression in the murine aorta to other tissues that contain cornified stratified epithelia, such as the esophagus and skin (Table 2). For the majority of the genes tested that were called "absent" in the IVC by microarray analysis, the transcript was undetectable by qRT-PCR with a separate set of experimental samples (Table 2); however, low levels of galectin 7 expression (27-fold less than that in the aorta) were detected in the IVC by qRT-PCR. The precise function of galectin 7 is not yet known; however, its expression is linked to differentiation of skin.33 The relative expression of many of these skin-enriched genes in the aorta relative to mouse skin by qRT-PCR analysis varied (Table 2), with generally higher expression in murine skin; however, mRNA expression of SPRR3 was higher in aorta than in skin. Interestingly, the level of expression of many of the genes involved in CE in the murine aorta was comparable to expression levels detected in the murine esophagus, a tissue with a defined CE (Table 2).
The magnitude of overexpression determined by qRT-PCR was similar to that determined from the signal intensity on the microarray for most of the genes, except plakoglobin. Relative expression of plakoglobin in the aorta was 8-fold higher than in the IVC by qRT-PCR, whereas the difference was only 2-fold by microarray analysis (Table 1). Desmocollin I was detected in the aorta only at a relatively low level (signal intensity of 439). This is supported by the finding that expression in the aorta was >200-fold less than in murine skin. The difference of expression of each gene examined by qRT-PCR between the artery and vein was statistically significant.
This high level of correlation observed between the array and qRT-PCR data was anticipated given that the majority of aorta-enriched genes identified by our analysis represented transcripts detected with very high signal intensity. Studies from other laboratories have shown very strong reproducibility among this set.12,34 We also tested for a number of other functionally related transcripts that were not identified in our analysis (either because they were absent in the chip, eg, envoplakin, or because they were also expressed in other tissues, eg, plakoglobin transcripts were also detected in murine heart). In each case, we were able to confirm aorta-enriched expression, which provides further support for the association between the aorta and this group of skin-enriched proteins.
Protein Expression of Selected Murine Aorta-Enriched Genes in Human Aorta and Other Arteries
Immunofluorescence and immunohistochemistry analysis with vascular tissue obtained from human autopsy samples was used to confirm whether protein expression was restricted only to the aorta and also to determine the vascular localization of a subset of these proteins. In each analysis (except for cytokeratin 13), the same slide was also stained with the nucleic acid dye DAPI to determine cellular localization within the tissue. Expression of sciellin, plakoglobin, galectin 7, SPRR3 (Figures 1 and 2), and cytokeratin 13 (Data Supplement Figure I) was restricted to the human aorta, not IVC (Data Supplement Figure II). Expression of sciellin (Figures 1c and 1e) and cytokeratin 13 (Data Supplement Figure I) was restricted to the intima of the aorta, with the majority of staining in the portion of the intima closest to the lumen. An antibody specific for plakoglobin and galectin 7 stained cells throughout the intima (Figures 1g, 2a, and 2c). No specific staining was seen in the media or adventitia of the aorta for sciellin, cytokeratin 13, or galectin 7 (data not shown). Staining was absent in the human IVC with antibodies against sciellin, galectin 7, and SPRR3 (Data Supplement Figure II). Although staining in the IVC was primarily absent with antibody to plakoglobin (data not shown), several isolated areas of punctate immunoreactivity were detected, particularly in regions in which the intima was thickened (Figure 1i). This sporadic staining pattern was confirmed on 3 independent experiments; nonspecific staining due to autofluorescence was ruled out by careful examination of the other channels for fluorescence (data not shown). We noted that SPRR3 showed discontinuous staining throughout the aortic intima. To study the localization of SPRR3 further, we performed immunohistochemistry on various human arteries and veins. Interestingly, localization of SPRR3 was limited to atheromatous plaques in several artery-derived vessels and was not detectable in veins (Figure 3).
Genes Present in IVC but Not Aorta or Other Tissues
Complementary analysis was performed to identify genes expressed in the IVC but absent in the aorta and all other tissues (excluding skin) used in the comparison presented for the aorta study. Transcripts represented by 17 probe sets were identified as present in the IVC and absent in the aorta and in murine spleen, liver, kidney, brain, skeletal muscle, lung, pancreas, intestine, and heart. Of these, all but 2 had signal intensities below 1000 in the IVC. This is noteworthy because "P" calls made by MAS 5.0 perform well at high concentrations but are much less reliable at the low concentrations reflected by the low signal intensity.12,34 We could not discern any functional relationships among these genes (Table 3).
Comparison of Adult Aorta Gene Expression With Reported Artery/Vein Specified Genes
There are only a few known artery- or vein-specific genes.7,8 EphrinB2 is expressed in developing arteries, whereas EphB4 is expressed in veins.35,36 However, analysis of the vena cava of adult mice in whole mount revealed patchy staining of the lacZ reporter targeted to the ephrinB2 locus.35,36 Expression of EphB4, which is considered specific for venous endothelium, was detected in lower levels in the wall of the adult dorsal aorta.35 More recently, Lawson and colleagues37,38 and Zhong et al39 have described expression of the Notch and Gridlock family of genes in precursors of endothelial cells, in which they promote arterial fate and repress venous differentiation. Unlike the ephrin genes, these and other differentially expressed genes have pronounced differences that exist only during early development.38 The expression profile of some of these genes in our microarray analysis is provided in Table 4 and did not demonstrate restricted expression.
Several recent reports have also used microarray analysis to identify the molecular basis for the structural and functional diversity in vessels, particularly in endothelial cells.40–43 Many of these studies have been aimed at the identification of endothelial genes that are regulated by arterial levels of shear stress using in vitro models.40–43 Because the present study used whole-vessel homogenates, not just endothelium, we suspect that our platform lacks the sensitivity to identify the distinctions between the arterial and venous endothelium reported by others. Hence, the message detection levels of many of these genes were detected both in the aorta and IVC at similar levels in our analysis (Table 4).
Using gene array to compare the gene expression profile of the aorta and IVC, we have identified several genes that are expressed specifically in the arterial but not venous intima. The tunica intima is composed of a single layer of endothelium that lines the lumen, smooth muscle cells, and connective tissue; the major cell type of the intima is the intimal smooth muscle cell.1 A surprising number of these arterial intima–enriched genes are also enriched in stratified epithelia, such as the epidermis.17 A number of these represented precursors of the CE and proteins involved in adhesive junctions.17,21 The identification of such a large number of known CE precursor proteins and proteins involved in cell-cell adhesion in the arterial intima suggests that they may play a role in the long-recognized enhanced barrier and stress-bearing properties of arteries compared with veins.22 Expression of SPRR3, specifically, appears to be associated with atherosclerotic plaques. The biomechanical disruptions around atherosclerotic plaques have long been described but are not well understood.44
Although the aorta has a number of CE precursors, it lacks some of the important components, such as involucrin, which is suggested to be a ubiquitous component of CEs in all stratified epithelia.20 Together with the fact that a CE has not been morphologically demonstrated in the aorta or large arteries, it is unlikely that the aorta contains a classic CE; however, these data raise the possibility that it may contain a less well-developed CE. Alternatively, these precursors may be participating in the cytoplasmic cross-linking of structural proteins. The human phenotype of known mutations in a few CE precursors usually manifests as various skin deformities, such as ichthyosis and hyperkeratosis.20 Although vascular defects in such conditions have not been appreciated and linked yet, it will be interesting to study such patients to determine whether there is greater predisposition to arterial pathology, such as atherosclerosis or aneurysms.
Interestingly, the aortic intima at birth is limited to a lamina of endothelial cells that closely abuts the internal elastic lamina.1 During growth and development, there is proliferation and development of the intima in the large arteries, and in the aorta, it expands to 0.2 mm in adults.1 It is hypothesized that biomechanical stress represents an important stimulus for differentiation of not just the endothelium but the entire intima.1,5 Future studies to examine the role of arterial intima–enriched proteins in intimal smooth muscle cells during vascular adaptation to biomechanical stress are under way.
Acknowledgments
This work was supported by a Vanderbilt physician scientist development grant and institutional funds from Washington University School of Medicine.
Footnotes
The online-only Data Supplement for this article can be found with this article at http://circ.ahajournals.org.
References
Stehbens WE. Structural and architectural changes during arterial development and the role of hemodynamics. Acta Anat. 1996; 157: 261–274.
Mills I, Sumpio BE. Vascular Smooth Muscle Cells. Armonk, NY: Futura Publishing Co; 1997.
Glagov S, Zarins CK, Masawa N, Xu CP, Bassiouny H, Giddens DP. Mechanical functional role of non-atherosclerotic intimal thickening. Front Med Biol Eng. 1993; 5: 37–43.
Zubileqicz T, Wronski J, Bourriez A, Terlecki P, Guinault AM, Muscatelli-Groux B, Michalak J, Melliere D, Becquemin JP, Allaire E. Injury in vascular surgery: the intimal hyperplastic response. Med Sci Monit. 2001; 7: 316–324.
Fry DL. Arterial intimal-medial permeability and coevolving structural responses to defined shear-stress exposures. Am J Physiol Heart Circ Physiol. 2002; 283: H2341–H2355.
Schwartz SM, Majesky MW, Murry CE. The intima: development and monoclonal responses to injury. Atherosclerosis. 1995; 118: S125–S140.
Cleaver O, Melton DA. Endothelial signaling during development. Nat Med. 2003; 9: 661–668.
Carmeliet P. Angiogenesis in health and disease. Nat Med. 2003; 9: 653–660.
Tan PK, Downey TJ, Spitznagel EL, Xu P, Fu D, Dimitrov DS, Lempicki RA, Raaka BM, Cam MC. Evaluation of gene expression measurements from commercial microarray platforms. Nucleic Acids Res. 2003; 31: 5676–5684.
Li C, Wong W. Model-based analysis of oligonucleotide arrays: expression index computation and outlier detection. Proc Natl Acad Sci U S A. 2001; 98: 31–36.
Affymetrix: Microarray Suite User Guide V. Affymetrix; 2001. Available at: http://www.affymetrix.com/support/technical/manuals/.affx.
Shippy R, Sendera TJ, Lockner RL, Chockalingam P, Kaysser-Kranich T, Watts G, Alsobrook J. Performance evaluation of commercial short-oligonucleotide microarrays and the impact of noise in making cross-platform correlations. BMC Genomics. 2004; 5: 61–81.
Dressen IM, Husing J, Kruse E, Boes T, Jockel JH. Software packages for quantitative microarray-based gene expression analysis. Curr Pharm Biotechnol. 2003; 4: 417–437.
Pfaffl MW. A new mathematical model for relative quantification in real-time RT-PCR. Nucleic Acids Res. 2001; 29: 2004–2007.
Nimgaonkar A, Sanoudou D, Butte A, Haslett JN, Kunkel LM, Beggs AH, Kohane IS. Reproducibility of gene expression across generations of Affymetrix microarrays. BMC Bioinformatics. 2003; 4: 27–39.
Curto EV, Lambert GW, Davis RL, Wilborn TW, Dooley TP. Biomarkers of human skin cells identified using dermarray DNA arrays and new bioinformatics methods. Biochem Biophys Res Commun. 2002; 291: 1052–1064.
Marenholz I, Zirra M, Fischer DF, Backendorf C, Ziegler A, Mischke D. Identification of human epidermal differentiation complex (EDC)-encoded genes by subtractive hybridization of entire YACs to a gridded keratinocyte cDNA library. Genome Res. 2001; 11: 341–355.
Gibbs S, Funeman R, Wiegant J, Geurts van Kessel A, Putte P, Backendorf C. Molecular characterization and evolution of the SPRR family of keratinocyte differentiation markers encoding small proline-rich proteins. Genomics. 1993; 16: 630–637.
Jarnik M, Kartasova T, Steinert PM, Lichti U, Steven AC. Differential expression and cell envelope incorporation of small proline-rich protein 1 in different cornified epithelia. J Cell Science. 1996; 109: 1381–1391.
Nemes Z, Steinert PM. Bricks and mortar of the epidermal barrier. Exp Mol Med. 1999; 31: 5–19.
Kalinin AE, Kajava AV, Steinert PM. Epithelial barrier function: assembly and structural features of the cornified cell envelope. Bioessays. 2002; 24: 789–800.
Kwei S, Stavrakis G, Takahas M, Taylor G, Folkman MJ, Gimbrone MA, Garcia-Cardena G. Early adaptive responses of the vascular wall during venous arterialization in mice. Am J Pathol. 2004; 164: 81–89.
Bonilla IE, Tanabe K, Strittmatter SM. Small proline-rich repeat protein 1A is expressed by axotomized neurons and promotes axonal outgrowth. J Neurosci. 2002; 22: 1303–1315.
Cabral A, Voskamp P, Cleton-Jansen A, South A, Nizetic D, Backendorf C. Structural organization and regulation of the small proline-rich family of cornified envelope precursors suggest a role in adaptive barrier function. J Biol Chem. 2001; 26: 19231–19237.
Champliaud M, Baden HP, Koch M, Jin W, Burgeson RE, Viel A. Gene characterization of sciellin (SCEL) and protein localization in vertebrate epithelia displaying barrier properties. Genomics. 2000; 70: 264–268.
Baden HP, Kvedar JC. Epithelial cornified envelope precursors are in the hair follicle. J Invest Dermatol. 1993; 101: 72S–74S.
Karashima T, Watt FM. Interaction of periplakin and envoplakin with intermediate filaments. J Cell Sci. 2002; 115: 5027–5037.
Marshall D, Hardman MJ, Nield KM, Byrne C. Differentially expressed late constituents of the epidermal cornified envelope. Proc Natl Acad Sci U S A. 2001; 98: 13031–13036.
Vestweber D. Molecular mechanisms that control endothelial cell contacts. J Pathol. 2000; 190: 281–291.
Schnittler HJ, Puschel B, Drenchkhahn D. Role of cadherins and plakoglobin in interendothelial adhesion under resting conditions and shear stress. Am J Physiol Heart Circ Physiol. 1997; 273: H2396–H2405.
Troyanovsky RB, Chitaev NA, Troyanovsky SM. Cadherin binding sites of plakoglobin: localization, specificity and role in targeting to adhering junctions. J Cell Sci. 1996; 109: 3069–3078.
Hoffmann I, Mertens C, Brettel M, Nimmrich V, Schnolzer M, Herrmann H. Interaction of plakophilins with desmoplakin and intermediate filament proteins: an in vitro analysis. J Cell Sci. 2000; 113: 2471–2483.
Timmons PM, Colnot C, Cail I, Poirier F, Magnaldo T. Expression of galectin-7 during epithelial development coincides with the onset of stratification. Int J Dev Biol. 1999; 43: 229–235.
McShane LM, Shih JH, Michalowska AM. Statistical issues in the design and analysis of gene expression microarray studies of animal models. J Mammary Gland Biol Neoplasia. 2003; 8: 359–374.
Gale NW, Baluk P, Pan L, Kwan M, Holash J, DeChiara TM, McDonald DM, Yancopoulos GD. Ephrin-B2 selectively marks arterial vessels and neovascularization sites in the adult, with expression in both endothelial and smooth muscle cells. Develop Biol. 2001; 230: 151–160.
Shin D, Garcia-Cardena G, Hayashi SI, Gerety S, Asahara T, Stavrakis G, Isner J, Folkman J, Gimbrone MA, Anderson DJ. Expression of EphrinB2 identifies a stable genetic difference between arterial and venous vascular smooth muscle as well as endothelial cells, and marks subsets of microvessels at sites of adult neovascularization. Develop Biol. 2001; 230: 139–150.
Lawson ND, Vogel AM, Weinstein BM. Sonic hedgehog and vascular endothelial growth factor acts upstream of the notch pathway during arterial endothelial differentiation. Develop Cell. 2002; 3: 127–136.
Lawson ND, Weinstein BM. Arteries and veins: making a difference with zebrafish. Nat Rev. 2002; 3: 674–682.
Zhong TP, Childs S, Leu JP, Fishman MC. Gridlock signalling pathway fashions the first embryonic artery. Nature. 2001; 414: 216–220.
Garcia-Cardena G, Comander J, Anderson DJ, Blackman BR, Gimbrone MA. Biomechanical activation of vascular endothelium as a determinant of its functional phenotype. Proc Natl Acad Sci U S A. 2001; 98: 4478–4485.
Dekker RJ, van Soest S, Fontijn RD, Salamanca S, de Groot PG, VanBavel E, Pannekoek H, Horrevoets AJ. Prolonged fluid shear stress induces a distinct set of endothelial cell genes, most specifically lung Kruppel-like factor (KLF2). Blood. 2002; 100: 1689–1698.
Chen BP, Li Y, Zhao Y, Chen K, Li S, Lao J, Yuan S, Shyy JY, Chien S. DNA microarray analysis of gene expression in endothelial cells in response to 24-h shear stress. Physiol Genomics. 2001; 7: 55–63.
Chi J, Chang HY, Haraldsen G, Jahnsen FL, Troyanskaya OG, Chang DS, Wang Z, Rockson SG, van de Rijn M, Botstein D, Brown PO. Endothelial cell diversity revealed by global expression profiling. Proc Natl Acad Sci U S A. 2003; 100: 10623–10628.
Salunke NV, Topoleski LDT, Humphrey JD, Mergner WJ. Compressive stress-relaxation of human atherosclerotic plaque. J Biomed Mater Res. 2001; 55: 236–241.(Pampee P. Young, MD, PhD;)