Skip to main content
  • Research article
  • Open access
  • Published:

A SAGE based approach to human glomerular endothelium: defining the transcriptome, finding a novel molecule and highlighting endothelial diversity

Abstract

Background

Large scale transcript analysis of human glomerular microvascular endothelial cells (HGMEC) has never been accomplished. We designed this study to define the transcriptome of HGMEC and facilitate a better characterization of these endothelial cells with unique features. Serial analysis of gene expression (SAGE) was used for its unbiased approach to quantitative acquisition of transcripts.

Results

We generated a HGMEC SAGE library consisting of 68,987 transcript tags. Then taking advantage of large public databases and advanced bioinformatics we compared the HGMEC SAGE library with a SAGE library of non-cultured ex vivo human glomeruli (44,334 tags) which contained endothelial cells. The 823 tags common to both which would have the potential to be expressed in vivo were subsequently checked against 822,008 tags from 16 non-glomerular endothelial SAGE libraries. This resulted in 268 transcript tags differentially overexpressed in HGMEC compared to non-glomerular endothelia. These tags were filtered using a set of criteria: never before shown in kidney or any type of endothelial cell, absent in all nephron regions except the glomerulus, more highly expressed than statistically expected in HGMEC. Neurogranin, a direct target of thyroid hormone action which had been thought to be brain specific and never shown in endothelial cells before, fulfilled these criteria. Its expression in glomerular endothelium in vitro and in vivo was then verified by real-time-PCR, sequencing and immunohistochemistry.

Conclusions

Our results represent an extensive molecular characterization of HGMEC beyond a mere database, underline the endothelial heterogeneity, and propose neurogranin as a potential link in the kidney-thyroid axis.

Background

Endothelial cells (EC) are frequently thought to be homogenous because of the multiple functions they share independent of the organ they serve, such as providing a non-thrombogenic surface, regulation of production or inhibition of vasoactive substances, haemostasis as well as leukocyte recruitment. Yet on closer examination, they show significant heterogeneity between similar vessels in different organ systems or in arterial versus venous endothelia [15]. Although known as highly specialized cells since the description of their fenestrated phenotype by F. Jorgensen more than 40 years ago [6] glomerular endothelial characteristics remain largely undefined.

Global gene expression studies added large amounts of valuable information to our knowledge on various EC and related pathologies, e.g. atherosclerosis [7]. Yet, in the case of human glomerular microvascular endothelial cells (HGMEC) developments did not have comparable pace. Due to challenges in obtaining, culturing and maintaining HGMEC studies employing primary cells in human glomerular research have been scarce, but very useful in obtaining new insights of human glomerular endothelium; most recently Amaral et al. investigated how Shiga toxin type-2 and Subtilase cytotoxin lead to damages characteristic for haemolytic uremic syndrome [8]. The ultimate aim to experiment with non-cultured glomerular EC has never been attained and the very first human glomerular endothelial cell line (GEnC) was not presented until 2006 [9]. This cell line has proved to be a useful tool in kidney research. Recently, glomerular endothelial barrier function and its regulation were finally studied in great detail using this tool while previously filtration barrier function research has mostly been on podocytes and the contribution of the glomerular endothelium had been relatively neglected [10]. These studies showed how reactive oxygen species present in common pathologies such as diabetes cause glomerular injury by directly disrupting glycocalyx and how chondroitin sulphate controlled by vascular growth factors A and C contributes to glomerular endothelial glycocalyx modulating the protein passage [8, 11]. Nevertheless, despite its usefulness in endothelial research an immortalized cell line is not suitable to define the cellular transcriptome with its predominant and specific transcripts. Thus, we hypothesized that such an investigation using genetically unmodified HGMEC would enhance our understanding regarding the source of the unique morphological characteristics, the behaviour in both culture and in disease and prepare the grounds for further studies of HGMEC. We used serial analysis of gene expression (SAGE), because it provides an unbiased approach to gene discovery and enables quantitative acquisition of most transcripts expressed [12]. Secondly, SAGE has become a powerful tool due to creation of large datasets holding more than two hundred million tags from a wide spectrum of tissues or cells including different EC which are publicly available as part of the Cancer Genome Anatomy Project [13].

The goals of this study were to establish extensive transcriptomic data as a step towards identification of the transcripts controlling the distinctive morphological and functional characteristics of glomerular endothelium and to underline endothelial diversity by comparisons between SAGE libraries from glomerular endothelial and uncultured glomeruli or from other non-glomerular EC. To reach these goals based on the known challenge in receiving useful data out of large transcript lists we describe a research strategy powerful enough to first confirm the endothelial origin of the transcript lists by bioinformatics, then to identify a low abundant transcript, neurogranin (NRGN) for the first time in glomerular endothelium using an in-silico analysis and finally to verify its expression in vitro and in vivo by means of sequencing and immunohistochemistry.

Results

Characterisation of HGMEC

Primary HGMEC formed monolayers and displayed typical cobblestone morphology (Figure 1A) in phase contrast microscopy. Immunofluorescence studies revealed distinct expression of von Willebrand Factor (vWF) and platelet/endothelial cell adhesion molecule 1 (PECAM1, CD31). Von Willebrand Factor staining demonstrated discrete, granular, perinuclear localisation (Figure 1B), whilst CD31 was expressed at the region of cell-to-cell contacts (Figure 1C). HGMEC retained functional characteristics of the microvasculature, expressing E-selectin and P-selectin (CD62E/P) in response to tumor necrosis factor (TNF) stimulation (Figure 1D), whereas unstimulated cells did not (Figure 1E).

Figure 1
figure 1

Characterisation of cultured human glomerular microvascular endothelial cells (HGMEC). A) Phase contrast micrograph of passage 3 purified HGMEC (magnification 200x), B and C) Immunofluorescence images of HGMEC probed for von Willebrand factor and PECAM1, respectively. D) and E) E selectin: HGMEC in panel D) were incubated for 12 hours with TNF alpha prior to fixation. No staining with E selectin was observed in unstimulated cells (E). Texas red conjugated secondary antibodies were used for detection and nuclei were counter stained with Hoechst dye (blue). Original magnification in B-E was 630x.

Transmission electron microscopy showed the presence of rod shaped microtubulated Weibel-Palade bodies (Figure 2A and B) which unambiguously identify the cells as endothelial [14]. Scanning electron microscopy demonstrated numerous fenestrae with a diameter of approx. 100 nm (Figure 2C). The presence of fenestrae as a hallmark of glomerular endothelial reflects the well-differentiated status of these cells [15].

Figure 2
figure 2

Electron microscopy (EM) of cultured human glomerular microvascular endothelial cells (HGMEC). A) HGMEC were cultured on nitrocellulose membranes and processed for EM. Transmission EM showing general cell structures such as lysosome (L), mitochondria (M), and endoplasmic reticulum (ER). Also the endothelial characteristic Weibel Palade bodies can be seen (circle). B) Magnification of a Weibel Palade body from panel A), showing rod shaped microtubules. C) Scanning EM of a HGMEC cell showing numerous fenestrae.

HGMEC SAGE library

The final HGMEC SAGE library which was constructed using the short SAGE protocol as it is superior to the long SAGE protocol in identifying differential expression of tags [16] contained 68,987 tags with 18,385 unique tags after electronic removal of contaminating linker sequences (Additional file 1: Table S1). It has been approved by the Gene Expression Omnibus (GEO) data depository (http://www.ncbi.nlm.nih.gov/geo) and assigned an accession number [GEO:GSM16892]. Key features of this library are shown in Table 1.

Table 1 Key features of the HGMEC SAGE library

Verification of endothelial origin of HGMEC SAGE library

The library was confirmed to be of endothelial origin with a classification approach as explained in detail in the Methods section. In short, we used the sum of the relative expression of 150 tags (Additional file 2: Table S2) as a test statistic: a value larger than the threshold 0.022 indicates an endothelial origin. In other words, if the sum of the total copy numbers of these tags account for 2.2% or more, that library qualifies as endothelial. With a sum of 0.070 (7%) for these 150 tags our HGMEC SAGE library is clearly classified to be of endothelial origin (Figure 3). Notably, two of the analyzed 18 endothelial cell SAGE libraries, “Vascular_endothelium_normal_breast_associated_P1H12 + _AP_1” [GEO: GSM384017] and “Normal corneal endothelium” [GEO: GSM1652], scored below the threshold of 0.022. Consequently, although this might be the reflection of a significant endothelial heterogeneity, only the remaining 16 endothelial cell SAGE libraries were merged to build a non-glomerular endothelial reference SAGE library for further analyses.

Figure 3
figure 3

Classification of glomerular microvascular endothelial cell (HGMEC) library as endothelial. A) Cross-validation using 63 non-endothelial and 18 endothelial SAGE-libraries not including HGMEC SAGE library. The minimum number of false classifications (false positive: a non-endothelial library classified as endothelial and false negative: an endothelial library classified as non-endothelial) was observed for n = 100 or 150 tags. B) Histogram showing the sum of relative expression for each of the endothelial and non-endothelial SAGE libraries as used for cross-validation in panel A). The sum of relative tag counts of 150 tags discriminates between endothelial and non-endothelial SAGE libraries. If the sum is larger than the threshold 0.022, the library is characterised as endothelial. This threshold is clearly exceeded by the HGMEC SAGE library (0.07).

Overlap analysis and mapping

After verification of the endothelial origin of this library, we took advantage of the existing public data (SAGE Genie) to reduce the high number of tags to a small group of relevant and differentially expressed tags in HGMEC. The algorithm we used as the research strategy is depicted in Figure 4. First, we determined overlapping transcripts from both the HGMEC SAGE library and the ex vivo glomerular library containing 44,334 tags [17]. This showed that 823 transcripts were shared by the cultured HGMEC and the non-cultured glomeruli that contain EC. This group represented transcripts in our HGMEC SAGE library with the potential of in vivo expression. Out of these 823 transcripts (Additional file 3: Table S3) 268 were differentially overexpressed in HGMEC using Chi-square test with p < 0.01 in comparison to the non-glomerular endothelial reference SAGE library (822,008 tags) and with the restriction that the observed count in HGMEC is higher than statistically expected (see Additional file 4: Table S4 for the complete list and Table 2 for the 50 most abundant tags as well as Additional file 5: Figure S1 for expression pattern by means of a volcano plot). To test whether our statistical analysis settings were adequate we applied multiple test correction to all p-values of the above described 823 transcripts by means of “false discovery rate (FDR)”. This p-value cut-off only included tags with a false discovery rate of < 0.05 (Additional file 4: Table S4) and was thus expected to result in less than 5% false positives. Surprisingly, this procedure led to the detection of five additional overexpressed transcripts and therefore was less stringent than our previous p-value cut-off of 0.01 with no correction (Additional file 3: Table S3, the additional five transcripts are labelled). Adding these transcripts to the set of overexpressed transcripts did not have an influence on our data or their interpretation. Nevertheless, we added the FDR corrected p-values to Table 2, Additional file 3: Table S3 and Additional file 4: Table S4.

Figure 4
figure 4

Short explanation of research strategy: definition of HGMEC transcriptome by generation of SAGE library and identification of HGMEC predominant transcripts based on 4 criteria. HGMEC: human glomerular microvascular endothelial cell.

Table 2 Expression of the 50 most abundant from 268 tags which were common to HGMEC and ex vivo glomerular SAGE libraries and overexpressed in HGMEC in comparison to the cumulative endothelial cell library

The expression of 20 randomly chosen tags from Table 2 was confirmed by quantitative RT-PCR (qRT-PCR). The cytogenetic locations of the 268 transcripts are noted in Additional file 6: Figure S2. According to this analysis chromosomes 1, 6, 11, 17 and 19 are carrying 43% of these genes.

Gene Ontology (GO) analysis

We identified the GO terms present in the 268 overexpressed genes by mapping them to the GO Biological Process and determining whether they occur more often in a category than expected. We defined a category to be redundant if its child term contained the same genes. Among the 60 detected categories 12 redundant child terms were deleted. The remaining most significant categories resulted in five connected components in the Gene Ontology graph that are listed in Table 3. The known interplay between different ubiquitination processes and apoptosis in different biological systems such as TNF receptor signaling [18] is reflected in HGMEC by the overexpression of many GO terms from these biological processes. Ubiquinol to cytochrome c (“Translation and energy metabolism” cluster) and nuclear migration (“Nuclear migration” cluster) are not displayed in Table 3 despite their significant overexpression due to the small numbers of molecules in the GO-terms mitochondrial electron transport. Nonetheless, in the case of “nuclear migration” (thought to be a common process of neuroepithelia in development and enabling the migration of nucleus between apical and basal surfaces) both of the molecules in this cluster were overexpressed in HGMEC and also in the regulatory microtubule associated molecule Tpx2 [19]. This is not a high abundance molecule and yet it was clearly detected in our HGMEC SAGE library (5.8 copies/200,000). Our GO analysis demonstrates that HGMEC have high expression of ribosomal proteins (23 out of 50 most abundant tags, Table 2). Also, molecules involved in interspecies interactions and cytoskeleton organization are enriched in HGMEC. One of two other molecules of interest in the glomerular endothelium, the von Willebrand cleaving protease ADAMTS13 [GenBank: NM_139025], tag CAGGCTGAAA, was not detected in HGMEC or in any other endothelial SAGE library except in the endothelium of the normal colon (12 copies/200,000). On the other hand caveolin-1 involved in endo- and transcytosis [20] (CAV1, [GenBank:NM_001753], tag TCCTGTAAAG, has a significant expression level in HGMEC (197 copies/200,000). Interestingly, the pathway analysis in HGMEC and dermal human microvascular EC (Vascular_normal_CS_control) SAGE libraries revealed novel differences regarding caveolae between these two microvascular endothelial cell types. In caveolae pathway caveolin-1, −2 and −3 were all present in HGMEC and absent in dermal microvascular EC. A complete list of molecules from this pathway present in HGMEC and absent in dermal microvascular EC is shown in Additional file 7: Table S5.

Table 3 Excerpt of gene ontology (GO) cluster classification

Glomerular and non-glomerular endothelial expression of neurogranin

To identify transcripts with the highest potential to be novel and HGMEC pre-dominant we applied the final group of transcripts of interest containing 268 tags (common to ex vivo glomeruli and differentially overexpressed in HGMEC when compared with non-glomerular endothelial cells) to a set of additional criteria: a) completely absent in all other nephron regions ex vivo[17] and b) more highly expressed than statistically expected in HGMEC. Ultimately, one transcript fulfilled all these criteria, namely NRGN. Its expression in HGMEC equalled to 192% of the statistically expected expression and it was enriched in HGMEC when compared to ex vivo glomerulus or to the fusion endothelial reference library with 52 versus 25 copies or 52 versus 21 copies per 200,000 tags, respectively. It was on position 174 when sorted from highest to the lowest expression level in the group of 268 tags (Additional file 4: Table S4). The corresponding tag for NRGN is TGACTGTGCT. Based on the SageGenie algorithm from Boon et al. [13] this tag has a high ranking of 95% and is classified as a reliable 3’ end of NRGN RefSeq transcripts [NM_001126181] and [NM_006176], which strongly supported the validity of the tag. Neurogranin transcript expression in GEnC under non-permissive conditions (5 days at 37°C) was shown by quantitative RT-PCR and the 116 bp long product was cloned, sequenced and matched the NRGN RefSeq transcript [NM_001126181], 237 bp, 100% (Figure 5). To confirm the expression of NRGN in the glomerular EC in vivo, frozen sections of human kidney were co-immunostained for NRGN (Figure 6C) and the EC marker vWF (Figure 6D). As shown in Figure 6E, both proteins show a significant co-localization confirming the transcriptional data that NRGN is expressed in the kidney’s glomerular EC. No reactivity was observed with secondary antibodies alone (Figure 6A and B). The endothelial expression of NRGN is further corroborated by the in silico analysis using each of the 18 publicly available endothelial libraries considered for the reference endothelial library. We found that 14 of them expressed the NRGN tag to some extent (Table 4), with a mean number of NRGN tags per 200,000 of 27.7 (standard deviation = 20.9).

Figure 5
figure 5

Amplification of a 116 bp fragment of the human neurogranin cDNA from cDNA isolated from immortalized human glomerular endothelial cell line GEnC. Agarose gel (1.5%) electrophoresis of 10 μL qPCR product using a template of cDNA generated by reverse transcription of total RNA isolated from GEnC (lane 1) or ddH2O as a negative control (lane 2).

Figure 6
figure 6

Immunofluorescence images of sections of human kidney showing glomerular expression of NRGN. Cryosections of human kidney were stained for neurogranin (NRGN) with anti-NRGN antibodies (C), von-Willebrand factor (vWF) with anti-vWF antibodies (D) and second antibodies alone (A and B). Signals of NRGN (red) and vWF (green) were merged in (E).

Table 4 List of all 18 non-glomerular endothelial SAGE libraries analyzed. 16 were merged to create the non-glomerular endothelial reference SAGE library

Two of the four libraries which did not contain the NRGN tag, were the ones which were below the “endothelial” threshold (Figure 3B) and thus not included into the fusion endothelial reference library. The other two were “Vascular normal_CS_control” from dermal microvascular EC and “Vascular_endothelium_normal_liver_associated_AP_NLEC1”. According to these data treatment of dermal microvascular EC with vascular growth factor (VEGF) known as a differentiation factor for EC [21] led to a detectable expression of NRGN, as shown in the “Vascular normal_CS_VEGF+” library (Table 4).

Discussion

Human glomerular endothelium is composed of highly specialized cells. Their molecular features leading to the unique phenotype remain largely unknown. In this study we define the transcriptome of HGMEC using SAGE [12]. Serial analysis of gene expression gives exceptionally unbiased transcriptional data because it is based on extraction of unique sequence tags for each distinct transcript, known or unknown, differentially expressed or not, from a population of transcripts. The frequency with which a given tag is present in a SAGE library reflects the abundance of its transcript resulting in quantification of obtained data. Thus, it does not involve error prone statistics resulting from comparison to laboratory controls and thus its data are readily comparable among different laboratories. Moreover, it obtains a broader snapshot of gene expression when compared to the commercial arrays [22]. These clearly distinguish SAGE from other methods used for similar purposes. Also, in endothelial cell research SAGE has proven to be both reliable and successful [2325]. It has to be noted that more recent and advanced high throughput sequencing methods such as deep RNA sequencing were considered for this study. This particular method displays similarities to SAGE regarding the linearity over a wide range of transcripts from low to high abundance as well as some technical advantages. It can work with significantly less starting material at a higher speed, with less work-load and cost for large sets of raw data from a single experiment to detect novel splice variants or alternative transcription sites [26, 27]. Yet, SAGE was the method of choice when designing the study due to the combination of its validity and the availability of unprecedented, large, well-maintained datasets and quality-controlled databases from SAGE libraries (e.g. Cancer Genome Anatomy Project including many endothelial cell types and glomeruli).

The main challenges of transcriptional studies remain a) the sequencing errors, which was reported to account for approximately 10% of tags in SAGE [28], b) the validation of detected transcripts, c) managing the vast number of transcripts and d) not “overseeing” novel transcripts of interest which are present, but not easily “visible” in the transcript library. In this study, these issues were addressed by incorporation of advanced bioinformatics and conventional methods such as real-time PCR and immunohistochemistry. Our bioinformatics strategy allowed verification of endothelial origin of the HGMEC as well as 16 out of the 18 SAGE libraries which were publicly available at the time (Table 4). The failure of two established endothelial SAGE libraries to qualify as endothelial using our methodology deserves to be mentioned. Although many possible explanations can be given for this phenomenon, we suggest that this is based on the significant differences between endothelia accounting for endothelial diversity. Nevertheless, we chose not to incorporate the transcript tags from these libraries into the non-glomerular endothelial reference library, although according to our calculations this would not change any statement or conclusion of this study. Furthermore, exclusion of singletons and comparison of HGMEC SAGE library first with ex vivo glomerular then with 16 endothelial SAGE libraries were fundamental. Only HGMEC-enriched tags with a significant overlap between HGMEC and ex vivo glomeruli under the simultaneous exclusion of common transcripts with non-glomerular endothelial cells were used for analyses, thus the probability of chasing tags based on sequencing errors was further minimized. Thus, expression of all the 20 randomly chosen transcripts in the resulting list of overexpressed HGMEC transcripts could be verified by qRT-PCR. Finally, robustness of our strategy was additionally validated by the demonstration of NRGN, a low-abundant molecule previously thought to be brain specific, in human glomerular endothelium in vitro and in vivo for the first time. Interestingly, although this transcript had never been known in kidney or any endothelial cell, our analysis showed that it was present in some previously generated endothelial SAGE libraries but was never recognized due to a lack of a strategy such as the one described in this study. The results from this stepwise procedure provide valuable large lists of transcripts expressed in human glomerular endothelium which have a high potential to be present in vivo and underline endothelial diversity.

Moreover, a genetic analysis of glomerular EC is especially intriguing: Not only because their uniqueness is evident in their morphology, behaviour in culture, and disease [2, 2937], but also because it is postulated that they have a distinct embryological origin and that the developing kidney generates its own endothelium [38]. Because the origin of distinct HGMEC features in vitro or in vivo is not fully understood, we defined the transcriptome of human glomerular endothelium by means of SAGE using HGMEC in culture which carried typical characteristics of differentiated glomerular endothelium including fenestrations and particularly focused on transcripts which are common to HGMEC and non-cultured ex vivo glomeruli. Only those transcripts were exposed to further analysis regarding their presence or absence in a fusion endothelial reference library. Thus, the vast numbers of transcripts were gradually reduced to manageable numbers and to relevant lists which are presented here. Next, these lists were explored in increasing depth. At first glance the remarkable abundance of ribosomal proteins in the HGMEC SAGE library was evident by the presence of 23 different ribosomal proteins among the most abundant 50 tags. We argue that this is not a mere consequence of cell culture and is a pattern found in other SAGE libraries from non-cultured cells, e.g. human monocytes (27 ribosomal proteins in the most abundant 50 transcripts) [39]. High translational activity possessed by HGMEC was reflected by the overexpression of “translation and energy metabolism” GO cluster. According to the GO analysis another cluster which was more abundant than statistically expected in HGMEC was “ubiquitination”. Ubiquitination is a posttranslational modification of proteins. Monoubiquitination is involved in various processes e.g. in endocytosis or transcriptional regulation. In contrast, polyubiquitination of a protein results in recognition and degradation of this protein [40]. In this context, it is noteworthy that internalization and transport of proteins, particularly albumin, is of special interest in physiology as well as pathology of glomerular endothelium, both of which involve caveolins. A major caveolin is caveolin-1 (CAV1) which is significantly related to albumin excretion [41]. Ubiquitination seems to be responsible also for the turnover for CAV1. Hence, the presence of CAV1 in HGMEC SAGE library and enrichment of ubiquitination GO terms in addition to high concentration of ribosomal proteins substantiate the previously suggested characteristics of glomerular endothelium. Novel differences between different microvascular EC appeared by comparing caveolae pathway: in contrast to HGMEC the dermal microvascular EC (Vascular_normal_CS_control) did not express any of the caveolin 1–3 in addition to some other qualitative differences (Additional file 7: Table S5). Our results regarding the absence of ADAMTS13 in HGMEC are conflicting. The presence of this von Willebrand cleaving protease had been demonstrated in immortalized glomerular as well as in dermal microvascular EC before [42]. In most of the previously established endothelial SAGE libraries including dermal microvascular EC (Vascular_normal_CS_control and Vascular_normal_CS_VEGF) this molecule was not detected. Neither a complete kidney (SAGE_Kidney_normal_B_1, GSM383901, data not shown) nor the liver endothelial cell SAGE library (Vascular_endothelium_normal_liver_associated_AP_NLEC1, GSM384018, Table 4) expressed any ADAMTS13 tags, although liver and kidney cells were used as positive controls in the study by Tati et al. Only the SAGE library from endothelium of normal colon expressed ADAMTS13. Probably, this is due to the low abundance of this molecule as it was demonstrated by its appearance only at high cycle numbers of RT-PCR experiments especially in human glomerular and microvascular EC [42]. It can be argued that in spite of remarkable tag numbers the sizes of the SAGE libraries were too small to detect this transcript.

Finally, when obtained transcript lists were further filtered using additional criteria (expression in HGMEC higher then statistically expected, overexpressed in HGMEC when compared to the non-glomerular endothelial reference SAGE library, never before shown in kidney or any endothelial cell type, higher expression in HGMEC than in other endothelial cell types) only one low abundance transcript, namely NRGN, stood out. It is reasonable to assume that a larger size of HGMEC SAGE library would probably lead to the detection of a higher number of NRGN tags and more transcripts fulfilling same criteria. Subsequently, we first confirmed the presence of NRGN transcript in HGMEC by sequencing and then the in vivo expression of NRGN protein in human glomerular endothelium by immunohistochemistry. Thus far, NRGN was believed to be brain specific and abundant in forebrain neurons whose interactions with calmodulin are suggested to play an important role in the regulation of synaptic responses and plasticity [43, 44]. In brain, NRGN expression is dependent on thyroid hormones [45] and it is a direct target of thyroid hormone action [4648]. There are numerous studies suggesting a link between thyroid and glomerular functions starting from embryogenesis [49] and involving chronic kidney disease as well as acute kidney injury [50], glomerular filtration rate [51] or resistant proteinuria [52]. Recent data in a rat model showed that proteinuria seen in hyperthyroidism is not due podocyte pathology and that hyper- as well as hypothyroidism lead to an increased capillary density when compared to control animals while hyperthyroidism resulted in an expansion of glomerular area [53]. Yet, molecules linking both organ systems have never been defined. We postulate that NRGN in glomerular endothelium represents a potential link between the kidney and the thyroid as this was described for brain and thyroid. Studies are underway to test the function of glomerular endothelial NRGN in the thyroid-kidney axis.

Conclusions

In this study we describe and analyze a HGMEC SAGE library as the first quantitative description of the human glomerular endothelial transcriptome. By using open-source SAGE data from uncultured glomeruli and from 16 other non-glomerular endothelial cell types which were merged to gain a reference endothelial library as well as by employing specific in-silico analyses an efficient research strategy was established. In this method we included an analysis to verify the declared origin of large transcript lists and strongly suggest this or similar analyses should be employed whenever using such transcript lists before including them in the studies, because 11% of the analyzed endothelial SAGE libraries failed to qualify as such. Also, integration of additional stringent filters into the bioinformatics to reduce the vast number of transcripts to manageable groups appeared to be indispensable. The described multi-step strategy which also involved GO and pathway analyses was capable of determining a list of genes potentially expressed in vivo, highlighting glomerular endothelial uniqueness as well as endothelial diversity and identifying the low abundant NRGN for the first time in glomerular endothelium and confirming its expression in vitro and in vivo. The demonstration of NRGN as a molecule potentially linking glomerulus to thyroid and the sufficiency of our transcriptomic data providing a novel insight in significant differences between the glomerular and other endothelia represent novel and exciting findings in glomerular endothelial research.

Methods

Culture of primary HGMEC

Human glomerular microvascular EC were isolated and cultured as previously described [54]. Briefly, human renal tissue was obtained, with informed consent, from macroscopically healthy pieces of nephrectomized kidneys. Written consent from patients was obtained before the surgery (Ethics Committee of the Medical University of Vienna protocol approval number EK 141/2002 047/05/2008 ensuring adherence to The Declaration of Helsinki). Isolated glomeruli from minced cortex were digested and plated on fibronectin coated cell culture dishes. After 7 days in cell culture medium including fetal calf serum (FCS), 50 μg/mL endothelial cell growth supplement, 30 U/mL heparin, 100 U/mL penicillin, 100 μg/mL streptomycin, and 2 mM alanine-glutamine HGMEC constituted approx. 10% of glomerular outgrowths. HGMEC were isolated from this mixed culture using CD31 immuno-magnetic sorting as per manufacturer’s instructions (Dynal, Hamburg, Germany). HGMEC were passaged 4 to 8 times.

Immortalized human glomerular endothelial cell line

Conditionally immortalized GEnC line was kindly provided by Dr. Satchell and Dr. Mathieson (University of Bristol, UK) and these cells were kept as instructed [9].

HGMEC characterisation

HGMEC were seeded onto fibronectin coated glass cover slips, and cultured until near confluence prior to fixation. For immunostaining of intracellular antigens cell monolayers were permeabilised by incubation in 1% Triton-X-100 for 20 min. Non-specific binding sites were blocked by incubating 1 h in blocking solution (5% (w/v) non-fat milk in phosphate-buffered saline (PBS). Primary antibody was incubated at 1 μg/mL for CD62P/E and CD31 and at 30 μg/mL for vWF for 1 hour. After probes were rinsed 3-times Texas red conjugated secondary antibody was incubated at 1:200 in antibody solution. To test the upregulation of CD62P/E cells were treated with human recombinant TNF alpha for 12 hours at 100U/mL.

Transmission electron microscopy of HGMEC

Human glomerular EC grown on microporous inserts were fixed with 1% glutaraldehyde, buffered with PBS, post fixed in 1% OsO4 buffered with Na-cacodylate (0.1 M), dehydrated in graded series of ethanol, and embedded in Polybed. Ultrathin sections were stained conventionally by uranyl acetate and lead citrate and were examined with a JOEL 100 C electron microscope at 100 KV.

Scanning electron microscopy of HGMEC

HGMEC were fixed with Karnowsky’s fixative, post fixed in 1% OsO4 buffered with Na-cacodylate (0.1 M), dehydrated in graded series of methanol, and submitted to critical point drying using CO2. Dried specimen were sputter coated with a 120 Å gold palladium layer and examined with a JOEL JSM 25-S scanning electron microscope.

Construction of the HGMEC SAGE library

SAGE libraries were constructed from cultured primary HGMEC according to the short SAGE protocol [12] as described at http://www.sagenet.org with modifications described previously [23, 24].

Confirmation of the endothelial origin of the HGMEC SAGE library

To confirm the endothelial origin of HGMEC SAGE library we applied a novel classification strategy as a learning process to distinguish endothelial libraries from others. Positive examples were eighteen endothelial SAGE libraries (Table 4) not including HGMEC. Negative examples were 63 SAGE libraries from non-endothelial normal tissues obtained from SAGE Genie (Additional file 8: Table S6). Only tag counts ≥2 were used and all tags which occurred in at least one of the libraries with a count ≥2 were considered. These were included in a tag-wise Wilcoxon rank sum test with the null hypothesis that the median expression of each tag is equal in both classes, and the alternative hypothesis that the median expression is larger in the endothelial libraries compared to the other libraries. The strategy was to select n tags highly expressed in endothelial libraries. In a subsequent step, the sum of the relative copy number of these tags was defined as an indicator for endothelial origin. The threshold was defined as the cut-off value to minimize wrong classifications based on the histogram consisting of sum of copy numbers of n tags. As shown in Figure 3B the chosen threshold 0.022 resulted in only 2 wrong classifications. To determine the optimal number of tags (n) to best distinguish endothelial from non-endothelial, we repeated the above described procedure using different numbers of n from 50 to 1000 tags (Figure 3A). The minimum number of false classifications was observed for n = 100 or 150 (Figure 3A). We chose 150 tags as depicted in Additional file 2: Table S2 for classifying the HGMEC SAGE library. Using these tags and the threshold of 0.022 the HGMEC SAGE library was clearly classified as endothelial (Figure 3B).

Construction of a non-glomerular endothelial reference SAGE library

To obtain an endothelial reference library, sixteen endothelial SAGE libraries (Table 4) were merged to form a cumulative endothelial SAGE library as a reference library consisting of 822,008 tags.

Analysis of differential expression of transcripts and tag-to-gene mapping

First, all tags with a copy number = 1 were eliminated from the HGMEC SAGE library. For determining the transcripts potentially present in vivo and specific for HGMEC we then combined the information present in the HGMEC SAGE library and in the ex vivo glomerular library by Chabardes-Garonne et al. [17]. We designated a transcript as being a potential HGMEC transcript if it occurred at least three times in the HGMEC SAGE library and in the ex vivo glomerular library. Each of these HGMEC transcripts was tested for overexpression compared with the cumulative non-glomerular endothelial reference SAGE library (Chi-Square-Test, p < 0.01).

Comparisons between libraries (HGMEC, glomerular, cumulative non-glomerular endothelial reference SAGE library) were performed in a Microsoft ACCESS database. Resulting tags (common or differentially expressed among libraries) or entire tag lists from libraries were automatically identified over the web by means of “IdenTAG”. IdenTAG was written in PERL specifically for this study and implemented a web-client tailored for the SAGE Genie site (http://cgap.nci.nih.gov/SAGE/AnatomicViewer)[13] of the Cancer Genome Anatomy Project at the US National Cancer Institute.

Gene Ontology and pathway analysis

For GO analysis of the overexpressed transcripts the UniGene-IDs were mapped via SAGE Genie to the corresponding Gene Symbols. Overrepresentation of genes in a GO category was determined by the Fisher test (p < 0.01). Subsequently, we identified the main functional categories present in these overexpressed genes by mapping them to GO Biological Process and determining whether they occur more often in a category than expected. All statistical analyses including the false discovery rate correction (p < 0.05) and the annotation with GO Terms were performed using R (http://www.r-project.org) and Bioconductor (http://www.bioconductor.org). For pathway analysis the software PathwayArchitect from Stratagene (La Jolla, CA, USA) was used.

Quantitative real time PCR

An ABI Prism 7700 Sequence Detector (Applied Biosystems, California, USA) was used for mRNA quantification via real-time PCR. Complementary DNA from HGMEC, GEnC cultured at 37°C or from human brain was generated. For human NRGN (assay ID details: Hs00183469) GEnC cDNA was used. Confirmation RT-PCR reactions for the 20 randomly selected transcripts were performed using HGMEC cDNA (List of transcripts and assay ID in Table 2). Beta-actin was used as endogenous control gene (Applied Biosystems, Hs99999903_m1). Normalization of Ct values of each gene (ß-actin and NRGN) and determination of NRGN expression in GEnC cells grown at 37°C was calculated according to the 2 ΔΔCt method [55].

Selection, cloning and sequencing of NRGN

We searched for novel transcripts with the highest potential to be expressed in HGMEC in vivo based on 4 criteria: 1) common to ex vivo glomerulum (which consequently contains non-cultured glomerular endothelial cells) and HGMEC SAGE library; 2) complete absence in the remaining seven nephron segment SAGE libraries namely proximal convoluted tubule, proximal straight tubule, medullary thick ascending limb of Henle’s loop, cortical thick ascending limb of Henle’s loop, cortical collecting duct, outer medullary collecting duct and distal convoluted tubule containing approximately 350,000 tags [17]; 3) never before shown in EC, 4) enrichment of the transcript in HGMEC: observed expression higher in HGMEC when compared to the non-glomerular endothelial reference library and to the statistically expected count in HGMEC. After NRGN was recognized as the only tag fulfilling these criteria its qRT-PCR product (116 bp) was purified and cloned in a pDrive cloning vector. Minipreps were prepared using the Qiaprep spin Miniprep Kit (Qiagen, Valencia, CA, USA) and vectors containing the insert were selected by EcoRI digestion and sequenced using the M13 reverse primer.

Neurogranin immunofluorescence staining in human kidney sections

For immunofluorescence, 15 μm cryostat sections of human kidney tissue from the fully anonymized archive at the Department of Pathology of the Medical University of Vienna ensuring compliance to the guidelines from the Medical University of Vienna and adherence to The Declaration of Helsinki (see section “Culture of primary HGMEC” above) were fixated and blocked. They were then incubated with 1:500 rabbit anti- NRGN polyclonal or anti-vWF monoclonal antibody for 1 hour. Primary antibodies were visualized using the TSA Plus Kit (PerkinElmer, Wellesley, MA, USA) and photographed with a fluorescence microscope.

Abbreviations

Bp:

Base pair

CAV1:

Caveolin-1

EC:

Endothelial cell

EM:

Electron microscopy

GEnC:

Glomerular endothelial cell

GEO:

Gene expression omnibus

GO:

Gene ontology

HGMEC:

Human glomerular microvascular endothelial cells

NRGN:

Neurogranin

PBS:

Phosphate-buffered saline

PCR:

Polymerase chain reaction

PECAM1:

Platelet/endothelial cell adhesion molecule 1

RT-PCR:

Real time polymerase chain reaction

SAGE:

Serial analysis of gene expression

SEM:

Scanning electron microscopy

TNF:

Tumor necrosis factor

VEGF:

Vascular growth factor

vWF:

Von Willebrand factor.

References

  1. Auerbach R, Alby L, Morrissey LW, Tu M, Joseph J: Expression of organ-specific antigens on capillary endothelial cells. Microvasc Res. 1985, 29: 401-411. 10.1016/0026-2862(85)90028-7.

    Article  CAS  PubMed  Google Scholar 

  2. Craig LE, Spelman JP, Strandberg JD, Zink MC: Endothelial cells from diverse tissues exhibit differences in growth and morphology. Microvasc Res. 1998, 55: 65-76. 10.1006/mvre.1997.2045.

    Article  CAS  PubMed  Google Scholar 

  3. Stan RV, Kubitza M, Palade GE: PV-1 is a component of the fenestral and stomatal diaphragms in fenestrated endothelia. Proc Natl Acad Sci U S A. 1999, 96: 13203-13207. 10.1073/pnas.96.23.13203.

    Article  CAS  PubMed Central  PubMed  Google Scholar 

  4. Betz J, Bauwens A, Kunsmann L, Bielaszewska M, Mormann M, Humpf HU, Karch H, Friedrich AW, Muthing J: Uncommon membrane distribution of Shiga toxin glycosphingolipid receptors in toxin-sensitive human glomerular microvascular endothelial cells. Biol Chem. 2012, 393: 133-147.

    Article  CAS  PubMed  Google Scholar 

  5. Chi JT, 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. 10.1073/pnas.1434429100.

    Article  CAS  PubMed Central  PubMed  Google Scholar 

  6. Jorgensen F: The Ultrastructure of the Normal Human Glomerulus. 1966, Copenhagen: Munksgaard

    Google Scholar 

  7. de Waard V, van den Berg BM, Veken J, Schultz-Heienbrok R, Pannekoek H, van Zonneveld AJ: Serial analysis of gene expression to assess the endothelial cell response to an atherogenic stimulus. Gene. 1999, 226: 1-8. 10.1016/S0378-1119(98)00577-0.

    Article  CAS  PubMed  Google Scholar 

  8. Amaral MM, Sacerdoti F, Jancic C, Repetto HA, Paton AW, Paton JC, Ibarra C: Action of shiga toxin type-2 and subtilase cytotoxin on human microvascular endothelial cells. PLoS One. 2013, 8: e70431-10.1371/journal.pone.0070431.

    Article  CAS  PubMed Central  PubMed  Google Scholar 

  9. Satchell SC, Tasman CH, Singh A, Ni L, Geelen J, von Ruhland CJ, O'Hare MJ, Saleem MA, van den Heuvel LP, Mathieson PW: Conditionally immortalized human glomerular endothelial cells expressing fenestrations in response to VEGF. Kidney Int. 2006, 69: 1633-1640. 10.1038/sj.ki.5000277.

    Article  CAS  PubMed  Google Scholar 

  10. Satchell SC, Braet F: Glomerular Endothelial Cell Fenestrations: an Integral Component of the Glomerular Filtration Barrier. Am J Physiol Renal Physiol. 2009, 296: 947-956. 10.1152/ajprenal.90601.2008.

    Article  Google Scholar 

  11. Foster RR, Armstrong L, Baker S, Wong DW, Wylie EC, Ramnath R, Jenkins R, Singh A, Steadman R, Welsh GI, Mathieson PW, Satchell SC: Glycosaminoglycan regulation by VEGFA and VEGFC of the glomerular microvascular endothelial cell glycocalyx in vitro. Am J Pathol. 2013, 183: 604-616. 10.1016/j.ajpath.2013.04.019.

    Article  CAS  PubMed Central  PubMed  Google Scholar 

  12. Velculescu VE, Zhang L, Vogelstein B, Kinzler KW: Serial analysis of gene expression. Science. 1995, 270: 484-487. 10.1126/science.270.5235.484.

    Article  CAS  PubMed  Google Scholar 

  13. Boon K, Osorio EC, Greenhut SF, Schaefer CF, Shoemaker J, Polyak K, Morin PJ, Buetow KH, Strausberg RL, De Souza SJ, Riggins GJ: An anatomy of normal and malignant gene expression. Proc Natl Acad Sci U S A. 2002, 99: 11287-11292. 10.1073/pnas.152324199.

    Article  CAS  PubMed Central  PubMed  Google Scholar 

  14. Weibel ER, Palade GE: New Cytoplasmic Components In Arterial Endothelia. J Cell Biol. 1964, 23: 101-112. 10.1083/jcb.23.1.101.

    Article  CAS  PubMed Central  PubMed  Google Scholar 

  15. van Setten PA, van Hinsbergh VW, van der Velden TJ, van de Kar NC, Vermeer M, Mahan JD, Assmann KJ, van den Heuvel LP, Monnens LA: Effects of TNF alpha on verocytotoxin cytotoxicity in purified human glomerular microvascular endothelial cells. Kidney Int. 1997, 51: 1245-1256. 10.1038/ki.1997.170.

    Article  CAS  PubMed  Google Scholar 

  16. Lu J, Lal A, Merriman B, Nelson S, Riggins G: A comparison of gene expression profiles produced by SAGE, long SAGE, and oligonucleotide chips. Genomics. 2004, 84: 631-636. 10.1016/j.ygeno.2004.06.014.

    Article  CAS  PubMed  Google Scholar 

  17. Chabardes-Garonne D, Mejean A, Aude JC, Cheval L, Di Stefano A, Gaillard MC, Imbert-Teboul M, Wittner M, Balian C, Anthouard V, Robert C, Ségurens B, Wincker P, Weissenbach J, Doucet A, Elalouf JM: A panoramic view of gene expression in the human kidney. Proc Natl Acad Sci U S A. 2003, 100: 13710-13715. 10.1073/pnas.2234604100.

    Article  CAS  PubMed Central  PubMed  Google Scholar 

  18. Karin M, Gallagher E: TNFR signaling: ubiquitin-conjugated TRAFfic signals control stop-and-go for MAPK signaling complexes. Immunol Rev. 2009, 228: 225-240. 10.1111/j.1600-065X.2008.00755.x.

    Article  CAS  PubMed  Google Scholar 

  19. Kosodo Y, Suetsugu T, Suda M, Mimori-Kiyosue Y, Toida K, Baba SA, Kimura A, Matsuzaki F: Regulation of interkinetic nuclear migration by cell cycle-coupled active and passive mechanisms in the developing brain. Embo J. 2011, 30: 1690-1704. 10.1038/emboj.2011.81.

    Article  CAS  PubMed Central  PubMed  Google Scholar 

  20. Hayer A, Stoeber M, Ritz D, Engel S, Meyer HH, Helenius A: Caveolin-1 is ubiquitinated and targeted to intralumenal vesicles in endolysosomes for degradation. J Cell Biol. 2010, 191: 615-629. 10.1083/jcb.201003086.

    Article  CAS  PubMed Central  PubMed  Google Scholar 

  21. Nourse MB, Halpin DE, Scatena M, Mortisen DJ, Tulloch NL, Hauch KD, Torok-Storb B, Ratner BD, Pabon L, Murry CE: VEGF induces differentiation of functional endothelium from human embryonic stem cells: implications for tissue engineering. Arterioscler Thromb Vasc Biol. 2010, 30: 80-89. 10.1161/ATVBAHA.109.194233.

    Article  CAS  PubMed Central  PubMed  Google Scholar 

  22. Ibrahim AF, Hedley PE, Cardle L, Kruger W, Marshall DF, Muehlbauer GJ, Waugh R: A comparative analysis of transcript abundance using SAGE and Affymetrix arrays. Funct Integr Genomics. 2005, 5: 163-174. 10.1007/s10142-005-0135-4.

    Article  CAS  PubMed  Google Scholar 

  23. Sengoelge G, Luo W, Fine D, Perschl AM, Fierlbeck W, Haririan A, Sorensson J, Rehman TU, Hauser P, Trevick JS, Kulak SC, Wegner B, Ballermann BJ: A SAGE-based comparison between glomerular and aortic endothelial cells. Am J Physiol Renal Physiol. 2005, 288: F1290-F1300. 10.1152/ajprenal.00076.2004.

    Article  CAS  PubMed  Google Scholar 

  24. St Croix B, Rago C, Velculescu V, Traverso G, Romans KE, Montgomery E, Lal A, Riggins GJ, Lengauer C, Vogelstein B, Kinzler KW: Genes expressed in human tumor endothelium. Science. 2000, 289: 1197-1202. 10.1126/science.289.5482.1197.

    Article  CAS  PubMed  Google Scholar 

  25. Wang CH, Su PT, Du XY, Kuo MW, Lin CY, Yang CC, Chan HS, Chang SJ, Kuo C, Seo K, Leung LL, Chuang YJ: Thrombospondin type I domain containing 7A (THSD7A) mediates endothelial cell migration and tube formation. J Cell Physiol. 2010, 222: 685-694.

    CAS  PubMed  Google Scholar 

  26. Costa V, Aprile M, Esposito R, Ciccodicola A: RNA-Seq and human complex diseases: recent accomplishments and future perspectives. Eur J Hum Genet. 2013, 21: 134-142. 10.1038/ejhg.2012.129.

    Article  CAS  PubMed Central  PubMed  Google Scholar 

  27. Wang Z, Gerstein M, Snyder M: RNA-Seq: a revolutionary tool for transcriptomics. Nat Rev Genet. 2009, 10: 57-63. 10.1038/nrg2484.

    Article  CAS  PubMed Central  PubMed  Google Scholar 

  28. Pleasance ED, Marra MA, Jones SJ: Assessment of SAGE in transcript identification. Genome Res. 2003, 13: 1203-1215. 10.1101/gr.873003.

    Article  CAS  PubMed Central  PubMed  Google Scholar 

  29. Green DF, Hwang KH, Ryan US, Bourgoignie JJ: Culture of endothelial cells from baboon and human glomeruli. Kidney Int. 1992, 41: 1506-1516. 10.1038/ki.1992.220.

    Article  CAS  PubMed  Google Scholar 

  30. Holmen C, Elsheikh E, Christensson M, Liu J, Johansson AS, Qureshi AR, Jalkanen S, Sumitran-Holgersson S: Anti endothelial cell autoantibodies selectively activate SAPK/JNK signalling in Wegener's granulomatosis. J Am Soc Nephrol. 2007, 18: 2497-2508. 10.1681/ASN.2006111286.

    Article  CAS  PubMed  Google Scholar 

  31. Kniesel U, Wolburg H: Tight junctions of the blood–brain barrier. Cell Mol Neurobiol. 2000, 20: 57-76. 10.1023/A:1006995910836.

    Article  CAS  PubMed  Google Scholar 

  32. Mitra D, Jaffe EA, Weksler B, Hajjar KA, Soderland C, Laurence J: Thrombotic thrombocytopenic purpura and sporadic hemolytic-uremic syndrome plasmas induce apoptosis in restricted lineages of human microvascular endothelial cells. Blood. 1997, 89: 1224-1234.

    CAS  PubMed  Google Scholar 

  33. Nitta K, Uchida K, Yumura W, Nihei H: Characterization of the glomerular endothelial cell in culture. Nippon Jinzo Gakkai Shi. 1993, 35: 887-891.

    CAS  PubMed  Google Scholar 

  34. Obrig TG, Louise CB, Lingwood CA, Boyd B, Barley-Maloney L, Daniel TO: Endothelial heterogeneity in Shiga toxin receptors and responses. J Biol Chem. 1993, 268: 15484-15488.

    CAS  PubMed  Google Scholar 

  35. Ott MJ, Olson JL, Ballermann B: Chronic in vitro flow promotes ultrastructural differentiation of endothelial cells. Endothelium. 1995, 3: 21-30. 10.3109/10623329509024655.

    Article  Google Scholar 

  36. Roberts JM, Taylor RN, Musci TJ, Rodgers GM, Hubel CA, McLaughlin MK: Preeclampsia: an endothelial cell disorder. Am J Obstet Gynecol. 1989, 161: 1200-1204. 10.1016/0002-9378(89)90665-0.

    Article  CAS  PubMed  Google Scholar 

  37. Ryan GB: The glomerular sieve and the mechanisms of proteinuria. Aust N Z J Med. 1981, 11: 197-206. 10.1111/j.1445-5994.1981.tb04233.x.

    Article  CAS  PubMed  Google Scholar 

  38. Hyink DP, Tucker DC, St John PL, Leardkamolkarn V, Accavitti MA, Abrass CK, Abrahamson DR: Endogenous origin of glomerular endothelial and mesangial cells in grafts of embryonic kidneys. Am J Physiol. 1996, 270: F886-F899.

    CAS  PubMed  Google Scholar 

  39. Ishii M, Hashimoto S, Tsutsumi S, Wada Y, Matsushima K, Kodama T, Aburatani H: Direct comparison of GeneChip and SAGE on the quantitative accuracy in transcript profiling analysis. Genomics. 2000, 68: 136-143. 10.1006/geno.2000.6284.

    Article  CAS  PubMed  Google Scholar 

  40. Hicke L: Protein regulation by monoubiquitin. Nat Rev Mol Cell Biol. 2001, 2: 195-201. 10.1038/35056583.

    Article  CAS  PubMed  Google Scholar 

  41. Moriyama T, Tsuruta Y, Shimizu A, Itabashi M, Takei T, Horita S, Uchida K, Nitta K: The significance of caveolae in the glomeruli in glomerular disease. J Clin Pathol. 2011, 64: 504-509. 10.1136/jcp.2010.087023.

    Article  PubMed  Google Scholar 

  42. Tati R, Kristoffersson AC, Stahl AL, Morgelin M, Motto D, Satchell S, Mathieson P, Manea-Hedstrom M, Karpman D: Phenotypic expression of ADAMTS13 in glomerular endothelial cells. PLoS One. 2011, 6: e21587-10.1371/journal.pone.0021587.

    Article  CAS  PubMed Central  PubMed  Google Scholar 

  43. Huang KP, Huang FL: Calcium-sensitive translocation of Calmodulin and Neurogranin between Soma and Dendrites of Mouse Hippocampal CA1 Neurons. ACS Chem Neurosci. 2011, 2: 223-230. 10.1021/cn200003f.

    Article  CAS  PubMed Central  PubMed  Google Scholar 

  44. Vallortigara J, Alfos S, Micheau J, Higueret P, Enderlin V: T3 administration in adult hypothyroid mice modulates expression of proteins involved in striatal synaptic plasticity and improves motor behavior. Neurobiol Dis. 2008, 31: 378-385. 10.1016/j.nbd.2008.05.015.

    Article  CAS  PubMed  Google Scholar 

  45. Dong J, Liu W, Wang Y, Xi Q, Chen J: Hypothyroidism following developmental iodine deficiency reduces hippocampal neurogranin, CaMK II and calmodulin and elevates calcineurin in lactational rats. Int J Dev Neurosci. 2010, 28: 589-596. 10.1016/j.ijdevneu.2010.07.230.

    Article  CAS  PubMed  Google Scholar 

  46. Iniguez MA, Rodriguez-Pena A, Ibarrola N, Aguilera M, Munoz A, Bernal J: Thyroid hormone regulation of RC3, a brain-specific gene encoding a protein kinase-C substrate. Endocrinology. 1993, 133: 467-473.

    CAS  PubMed  Google Scholar 

  47. Iniguez MA, Rodriguez-Pena A, Ibarrola N, Morreale de Escobar G, Bernal J: Adult rat brain is sensitive to thyroid hormone. Regulation of RC3/neurogranin mRNA. J Clin Invest. 1992, 90: 554-558. 10.1172/JCI115894.

    Article  CAS  PubMed Central  PubMed  Google Scholar 

  48. Sharlin DS, Gilbert ME, Taylor MA, Ferguson DC, Zoeller RT: The nature of the compensatory response to low thyroid hormone in the developing brain. J Neuroendocrinol. 2010, 22: 153-165. 10.1111/j.1365-2826.2009.01947.x.

    Article  CAS  PubMed  Google Scholar 

  49. Kumar J, Gordillo R, Kaskel FJ, Druschel CM, Woroniecki RP: Increased prevalence of renal and urinary tract anomalies in children with congenital hypothyroidism. J Pediatr. 2009, 154: 263-266. 10.1016/j.jpeds.2008.08.023.

    Article  CAS  PubMed Central  PubMed  Google Scholar 

  50. Iglesias P, Diez JJ: Thyroid dysfunction and kidney disease. Eur J Endocrinol. 2009, 160: 503-515. 10.1530/EJE-08-0837.

    Article  CAS  PubMed  Google Scholar 

  51. Suher M, Koc E, Ata N, Ensari C: Relation of thyroid disfunction, thyroid autoantibodies, and renal function. Ren Fail. 2005, 27: 739-742. 10.1080/08860220500243338.

    Article  CAS  PubMed  Google Scholar 

  52. Rodriguez-Gomez I, Sainz J, Wangensteen R, Moreno JM, Duarte J, Osuna A, Vargas F: Increased pressor sensitivity to chronic nitric oxide deficiency in hyperthyroid rats. Hypertension. 2003, 42: 220-225. 10.1161/01.HYP.0000081944.47230.69.

    Article  CAS  PubMed  Google Scholar 

  53. Rodriguez-Gomez I, Banegas I, Wangensteen R, Quesada A, Jimenez R, Gomez-Morales M, O'Valle F, Duarte J, Vargas F: Influence of thyroid state on cardiac and renal capillary density and glomerular morphology in rats. J Endocrinol. 2013, 216: 43-51. 10.1530/JOE-12-0208.

    Article  CAS  PubMed  Google Scholar 

  54. Aydin S, Signorelli S, Lechleitner T, Joannidis M, Pleban C, Perco P, Pfaller W, Jennings P: Influence of microvascular endothelial cells on transcriptional regulation of proximal tubular epithelial cells. Am J Physiol Cell Physiol. 2008, 294: C543-C554.

    Article  CAS  PubMed  Google Scholar 

  55. Livak KJ, Schmittgen TD: Analysis of relative gene expression data using real-time quantitative PCR and the 2(−Delta Delta C(T)) Method. Methods. 2001, 25: 402-408. 10.1006/meth.2001.1262.

    Article  CAS  PubMed  Google Scholar 

Download references

Acknowledgements

We gratefully acknowledge skilful technical assistance of Ms. Agnes Perschl and thank Ms. Mathilde Sengoelge and Mr. Andrew Rees for critical reading. GEnC cell line was generously provided by Dr. Simon Satchell and Dr. Peter Mathieson from the University of Bristol. This study was supported by the Austrian Science Fund (FWF) (P16129-B08) as well as the Medical Scientific Fund of the Mayor of the City of Vienna (project number 09038) granted to Guerkan Sengoelge. Arndt von Haeseler acknowledges financial support from the WWTF (Vienna Science Chair in Bioinformatics). Sophia Blake was funded by Long Term Fellowships from EMBO and FEBS.

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Guerkan Sengoelge.

Additional information

Competing interests

The authors declare that they have no competing interests.

Authors’ contributions

GS conceived the study, carried out SAGE and finalized the manuscript. WW performed RT-PCR, cloning, sequencing and drafted the manuscript. AK carried out bioinformatics and most statistical analyses. AvH participated in design and interpretation of bioinformatics data as well as in drafting the manuscript. MS wrote the script IdenTag for matching tags from SAGE libraries with public databases. WP and PJ isolated and cultured human glomerular endothelial cells and characterized them by means of immunohistochemistry as well as electron microscopy. AW had set up the initial Access database which included all SAGE libraries and in which merging of tags from different SAGE libraries was accomplished. SB carried out immunohistochemistry for von Willebrand Factor and neurogranin in human kidney sections. GSP participated in the design of the study and drafting the manuscript. All authors read and approved the final manuscript.

Guerkan Sengoelge, Wolfgang Winnicki contributed equally to this work.

Electronic supplementary material

Additional file 1: Table S1: Complete HGMEC SAGE library after exclusion of linker sequences. (XLS 4 MB)

12864_2014_6406_MOESM2_ESM.xls

Additional file 2: Table S2: List of 150 tags used for endothelial classification as described in Figure 3. (XLS 34 KB)

12864_2014_6406_MOESM3_ESM.xls

Additional file 3: Table S3: Complete list of 823 tags common to SAGE libraries from HGMEC and ex vivo glomeruli. FDR: multiple test correction by false discovery rate. (XLS 159 KB)

12864_2014_6406_MOESM4_ESM.xls

Additional file 4: Table S4: Complete list of 268 tags which are common to SAGE libraries from HGMEC and ex vivo glomeruli and overexpressed in HGMEC compared to the 16 non-glomerular endothelial cell SAGE libraries including information on cytogenetic location of each tag. FDR: multiple test correction by false discovery rate. (XLS 94 KB)

12864_2014_6406_MOESM5_ESM.pdf

Additional file 5: Figure S1: Volcano plot showing p-values against log fold change for the 823 transcripts from Additional file 3: Table S3. Each red dot symbolizes one transcript which is common to HGMEC and ex vivo glomeruli and is enriched in HGMEC when compared to non-glomerular endothelial cells (268 transcripts from Additional file 4: Table S4). The plot shows a typical appearance except that because of the high p-values some of the sample points are accumulated on both over- (positive values on the x-axis) and underrepresentation (negative values on the x-axis) sides. Note the logarithmic scaling of the y-axis. (PDF 15 KB)

12864_2014_6406_MOESM6_ESM.doc

Additional file 6: Figure S2: Chromosome distribution of 268 HGMEC enriched genes as listed in Additional file 4: Table S4. (DOC 33 KB)

12864_2014_6406_MOESM7_ESM.xls

Additional file 7: Table S5: Lists of caveolae pathway molecules present in HGMEC and absent in dermal microvascular endothelial cells. Connectivity: high degree of connectivity represents proteins which are more likely to be essential for survival. (XLS 25 KB)

12864_2014_6406_MOESM8_ESM.xls

Additional file 8: Table S6: List of the 63 non-endothelial SAGE libraries which were used for the algorithm to verify the endothelial origin of a given SAGE library. (XLS 28 KB)

Authors’ original submitted files for images

Rights and permissions

This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Sengoelge, G., Winnicki, W., Kupczok, A. et al. A SAGE based approach to human glomerular endothelium: defining the transcriptome, finding a novel molecule and highlighting endothelial diversity. BMC Genomics 15, 725 (2014). https://doi.org/10.1186/1471-2164-15-725

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1186/1471-2164-15-725

Keywords