Cell Biology

Sam68-Like Mammalian Protein 2, Identified by Digital Differential Display as Expressed by Podocytes, Is Induced in Proteinuria and Involved in Splice Site Selection of Vascular Endothelial Growth Factor

Clemens D. Cohen, Peter P. Doran, Simone M. Blattner, Monika Merkle, Guo Q. Wang, Holger Schmid, Peter W. Mathieson, Moin A. Saleem, Anna Henger, Maria P. Rastaldi and Matthias Kretzler
JASN July 2005, 16 (7) 1958-1965; DOI: https://doi.org/10.1681/ASN.2005020204

Abstract

Podocytes, the glomerular epithelial cells of the kidney, share important features with neuronal cells. In addition to phenotypical and functional similarities, a number of gene products have been found to be expressed exclusively or predominantly by both cell types. With the hypothesis of a common transcriptome shared by podocytes and neurons, digital differential display was used to identify novel podocyte-expressed gene products. Comparison of brain and kidney cDNA libraries with those of other organs identified Sam68-like mammalian protein 2 (SLM-2), a member of the STAR family of RNA processing proteins, as expressed by podocytes. SLM-2 expression was found to be restricted in the kidney to podocytes. In proteinuric diseases, SLM-2, a known regulator of neuronal mRNA splice site selection, was found significantly upregulated on mRNA and protein levels. Knockdown of SLM-2 by short interfering RNA in podocytes was performed to evaluate its biologic role. RNA splicing of vascular endothelial growth factor (VEGF), a key regulator of the filtration barrier and expressed as functionally distinct splice isoforms, was evaluated. VEGF165 expression was found to be reduced by 25% after SLM-2 knockdown. In vivo, the glomerular expression of SLM-2 correlated with the mRNA levels of VEGF165. This study demonstrates the power of digital differential display to predict cell type–specific gene expression by hypothesis-driven analysis of tissue cDNA libraries. SLM-2-dependent VEGF splicing indicates the importance of mRNA splice site selection for glomerular filtration barrier function.

Podocytes represent a highly specialized cell type with complex cell function. As the final barrier of the glomerular filtration unit, podocytes build a highly complex network by interdigitating so-called foot processes (1). Recently, several studies led to the detection of specific gene products that are crucial for the physiology of this cell type. These proteins are especially involved in the organization of the slit diaphragm (the cell–cell contact between the foot processes), the specialized cytoskeleton of podocytes, and cell adhesion to the glomerular basement membrane. Many of these proteins are reported to be expressed primarily or exclusively in neurons and podocytes. Besides nephrin, an essential protein of the slit diaphragm, densin; synaptopodin; glomerular epithelial protein 1; the synaptic vesicle molecule rab3A and its effector rabphilin-3a; and the amino acid transporters CAT3 and EAAT3 show preferential or restricted expression in both of these cell types (2–8). This may reflect physiologic similarities, as both cells share important characteristics such as arborized cell process formation, highly specialized cell–cell contacts, and a complex cytoskeletal organization (9).

These intriguing similarities of glomerular epithelial and neuronal cells prompted us to take advantage of the hypothesis of a shared transcriptome to predict new podocyte-expressed genes. The digital differential display (DDD) method was used to compare cDNA libraries of brain and kidney with those of all other tissues. This technique, part of the Cancer Genome Anatomy Project and available at the National Center for Biotechnology Information, exploits the large number of publicly available cDNA libraries corresponding to different tissues (10). It permits selection of (1) cDNA libraries to be compared, (2) comparison of the constituent sequences, and (3) output of a list of differentially expressed sequences. DDD experiments have been performed successfully to compare cancerous and healthy tissue (e.g., reference 11) and to isolate cDNA specific for a given library (12). This technique has not previously been used to predict cell type–specific gene expression by a hypothesis-driven comparison of two organs with other tissues. We searched for mRNA with high expression in brain and kidney but no or significantly lower expression in other tissues. This should allow the detection of mRNA templates preferentially expressed by renal cells with a gene expression resembling the one of brain cells. According to the shared transcriptome of podocytes and neurons, the encoded proteins should be found preferentially expressed in podocytes.

This approach allowed the prediction of Sam68-like mammalian protein 2 (SLM-2) to be expressed in podocytes, which was confirmed in vitro and in vivo. SLM-2 (13) is the gene product of the human gene KHDRBS3 on chromosome 8. The gene name is an acronym for the known characteristics of SLM-2: KH domain containing, RNA binding, signal transduction associated. Here, we demonstrate that SLM-2 is regulated in glomerular diseases and is involved in alternative mRNA splicing of the vascular endothelial growth factor (VEGF-A).

Materials and Methods

DDD

DDD, a bioinformatic tool available at the National Center for Biotechnology Information (www.ncbi.nlm.nih.gov/UniGene/info_ddd.html), analyzes the frequencies of cDNA and expressed sequence tag (EST) in expression libraries. Briefly, this technique compares the constituent sequences of different libraries to determine the relative frequency of each transcript in the libraries analyzed. Fisher exact test is used to determine the statistical significance of the number of times that sequences from the selected libraries are assigned to a specific UniGene cluster. In this manner, the relative abundance of transcripts in each of the libraries can be determined and thus differential expression of genes identified. A recently developed program, Digital Extractor (14), was used to compare human cDNA libraries of normal adult kidney and brain with other organ cDNA libraries.

Cell Culture

Three glomerular visceral epithelial cell lines were used and cultured according to previously published protocols: A conditionally immortalized human podocyte cell line (15), a stable T-SV40 immortalized human glomerular visceral epithelial cell line (16), and a conditionally immortalized murine podocyte cell line (17).

Reverse Transcriptase–PCR and Real-Time Reverse Transcriptase–PCR

Total RNA was isolated using silica-gel columns following the respective protocol (RNeasy Mini; Qiagen, Hilden, Germany). Control mRNA from whole human kidney and microdissected human glomeruli were isolated according to a reported protocol (18). Reverse transcriptase–PCR (RT-PCR) was performed using heat-activated TaqDNA polymerase (Amplitaq Gold; Applied Biosystems, Darmstadt, Germany). The amplified cDNA was analyzed on a nondenaturing 5% polyacrylamide gel, stained with Vistra Green (Amersham, Braunschweig, Germany), and visualized with ImageQuant Software on a Storm Fluorophosphorimager (both Molecular Dynamics, Krefeld, Germany). Real-time RT-PCR was performed on an ABI PRISM 7700 Sequence Detection System (Applied Biosystems) as described previously (18). All renal biopsies were performed according to the local ethical committees’ directives, and samples were processed according to the protocol of the European Renal cDNA bank (18).

The following oligonucleotide primers (300 nmol/L) and probes (100 nmol/L) were used: Human SLM-2 (NM 006558), sense primer 5′-CCG TAA AAC AGT TCC CTA AGT TCA AC-3′ (exon 2/3, i.e., spanning exons 2 to 3), antisense primer 5′-CTT CCT TGG CCT TGT CTC TCA T-3′ (exon 3/4); fluorescence labeled probe (FAM) 5′-CCA CGT GGC AAT TCT CTG AAG CGT TTA-3′; human Ankyrin repeat domain 7 (ANKRD7; NM 019644), sense primer 5′-CTG CAT AGA TAC CCA CAA TTC ACT G-3′ (exon 5/6), antisense primer 5′-TTG TAG AAA TGG GTT CCA TAT CTT CC-3′ (exon 6/7); and human hypothetical protein XP_172449 (XM172449), sense primer 5′-ACT CGC AGG CCC TGT GG-3′, antisense primer 5′-CTC GGT GTC AAG GAG CAA GAG-3′. Commercially available predeveloped TaqMan reagents were used for human glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and 18S rRNA (Applied Biosystems). The primers spanning at least one exon-intron boundary did not show any amplification signal tested on genomic DNA. All primers and probes were obtained from Applied Biosystems.

Western Blot

Cultured glomerular epithelial cells were harvested with lysis buffer (50 mM Tris-HCl [pH 7.4], 1% Nonidet P-40, 0.25% sodium deoxycholate, 150 mM NaCl, 1 mM EGTA, 1 mM Na3VO4, and Complete Protease Inhibitor Cocktail [Roche, Mannheim, Germany]). Extracted proteins were boiled in loading buffer for 10 min, resolved by 10% SDS-PAGE under reducing conditions, and transferred to an Immobilon-P membrane (Millipore, Eschborn, Germany). The membrane was blocked with 5% skim milk, incubated in a 1:100 dilution of a purified polyclonal goat anti–SLM-2 antibody (M20; Santa Cruz Biotechnology, Heidelberg, Germany) overnight, and rinsed with PBS that contained 0.1% Tween 20. Immune complexes were visualized using enhanced chemoluminescence (ECL; Amersham Biosciences, Freiburg, Germany). Commercially available murine brain protein extract served as positive control for SLM-2 detection (ImmunoGlobe, Himmelstadt, Germany).

Immunofluorescence

Glomerular epithelial cells were grown on collagen-coated glass coverslips for 24 h, fixed with formaldehyde (2% formaldehyde and 4% sucrose in PBS) for 5 min, incubated for 10 min with 0.3% Triton in PBS, and washed in PBS. After 30 min of incubation with blocking solution (2% FCS, 2% BSA, and 0.2% fish gelatin in PBS), the fixed cells were incubated with 1:10 to 1:50 polyclonal goat anti–SLM-2 antibody (Santa Cruz Biotechnology) for 60 min. After washing, the cells were incubated with 1:200 Cy2-labeled anti-goat antibody (Dianova, Hamburg, Germany) for 60 min. Coverslips were fixed with mowiol solution (Calbiochem-Novabiochem, Bad Soden, Germany), and fluorescence microscopy was performed.

Immunohistochemistry

Brain and kidney specimens from 3-mo-old Balb/C mice or human biopsies were fixed in 4% buffered paraformaldehyde and embedded in paraffin. Immunohistochemistry was performed on 8-μm-thick brain sections and 5-μm-thick kidney sections that were sequentially deparaffinized, rehydrated, and microwave-treated in 0.01 M citrate buffer. After washing, sections were incubated with 0.5% avidin and 0.01% biotin to suppress endogenous avidin-binding activity, then postfixed in a methanol-H2O2 solution to block endogenous peroxidase. The specimens then were incubated sequentially with a blocking solution (Zymed, Milan, Italy), the primary polyclonal goat anti-mouse SLM-2 antibody (Santa Cruz Biotechnology), followed by the secondary rabbit anti-goat biotinylated antibody (Zymed) and the peroxidase-labeled streptavidin (Zymed). Peroxidase activity was detected with 3,5-diaminobenzidine (Sigma, St. Louis, MO). Sections then were counterstained by Cresyl Blue (DDK, Milan, Italy), dehydrated, and mounted in Permount (DDK). Nuclear staining in the brain was taken as positive control. Specificity of antibody labeling was demonstrated by the lack of staining after substituting PBS and proper control immunoglobulins (Zymed) for the primary antibody. Positive and negative controls were run together.

Short Interfering RNA Experiments and VEGF Isoform Analysis

The conditionally immortalized human podocyte cell line (15) was transiently transfected under nonpermissive culture conditions after differentiation with sequence-specific short interfering RNA (siRNA; 200 nM) using Lipofectamine 2000 (Invitrogen, Karlsruhe, Germany). Target sequences were KHDRBS3–2 (SLM-2 siRNA2) sense 5′-GGC AAG GAU GAA GAA AAG Utt-3′, antisense 5′-ACU UUU CUU CAU CCU UGC Ctt-3′; and KHDRBS3–3 (SLM-2 siRNA3) sense 5′-GGA UGA AGA AAA GUA CAU C-3′, antisense 5′-GAU GUA CUU UUC UUC AUC C-3′, corresponding to regions in the exon 2 of SLM-2. All siRNA, including negative control of nonspecific nucleotide sequence (negative control siRNA 1), were obtained from Ambion (Silencer siRNA; Ambion, Austin, TX).

Primers for analysis of VEGF isoforms (NM 003376) were as follows: For RT-PCR separated by gel electrophoresis, sense 5′-CAA ATG TGA ATG CAG ACC AAA-3′ (exon 4/5), antisense 5′-GGG AGG CTC CTT CCT CCT-3′ (3′-UTR), expected product size 92 bp (VEGF121) and 224 bp (VEGF165), respectively. The primers were designed specifically for the detection of stimulatory VEGF isoforms; recently described inhibitory VEGF isoforms were not amplified (19). Real-time RT-PCR VEGF165-specific assay: Sense primer 5′-CAA ATG TGA ATG CAG ACC AAA GA-3′ (exon 4/5), antisense primer 5′-CAG GAA CAT TTA CAC GTC TGC G-3′ (exon 7), and fluorescence-labeled probe (FAM) 5′-AGC AAG ACA AGA AAA TCC CTG TGG GCC-3′ (exon 5/7). Assay for all human VEGF-A splice isoforms: Sense 5′-GCC TTG CTG CTC TAC CTC CAC-3′ (exon 1), antisense 5′-ATG ATT CTG CCC TCC TCC TTC T-3′ (exon 2), internal fluorescence-labeled probe (FAM) 5′-AAG TGG TCC CAG GCT GCA CCC AT-3′ (exon 1/2).

Statistical Analyses

Data are given as mean ± SD. Statistical analysis was performed using Kruskall-Wallis, Mann-Whitney tests, and Pearson correlations (SPSS 12.0, SPSS Inc., Chicago, IL). P < 0.05 was considered to indicate statistically significant differences.

Results

DDD Predicts Brain/Kidney-Specific Templates

A DDD experiment was conducted to compare normal adult kidney and brain cDNA libraries (total of 63,268 sequences) with all other available normal adult tissue libraries except for testis (204,515 sequences; Figure 1). Testis libraries were excluded because of the significant similarity of its transcriptome to the one of brain (20). DDD takes advantage of the UniGene database by comparing the number of times that sequences from different libraries are assigned to a particular UniGene cluster. This analysis produced 55 UniGene clusters that were specific to the brain and/or kidney pool. Two further DDD experiments were performed to isolate brain-specific and kidney-specific clusters. The subtraction of these brain-specific and kidney-specific clusters from the brain and/or kidney results identified 11 putative brain- and kidney-specific clusters. Detailed analysis of the chromosomal regions was performed. Thereby, four clusters had to be eliminated as they mapped to gene loci of previously deleted templates. From the remaining seven clusters, an automated literature search led to removal of an additional four sequences because of their known significant expression in other tissue than brain and kidney. The remaining three sequences, Ankyrin repeat domain 7 (ANKRD7), hypothetical protein XP_172449, and SLM-2 (KHDRBS3), were assumed to be potentially restricted to brain and kidney.

Figure 1.
Figure 1.

Schematic view of the digital differential display (DDD) experiment. Normal adult kidney and brain cDNA libraries (total of 63,268 sequences) were compared with all other available normal adult tissue libraries (204,515 sequences, testis excluded). Fifty-five UniGene clusters that were potentially specific for brain and/or kidney were retrieved. Further DDD experiments resulted in 11 clusters that were putatively specific for brain and kidney. Chromosomal localization and automated literature search led to removal of eight sequences. The expression of the remaining three sequences was predicted to be restricted to brain and kidney and therefore potentially specific for podocytes in the kidney.

SLM-2 Shows a Restricted Expression in Podocytes

RT-PCR experiments were carried out with specific primers on cDNA from two human glomerular epithelial cell lines. No ANKRD7 or hypothetical protein XP_172449 mRNA templates could be amplified. RT-PCR for SLM-2 was positive in both glomerular epithelial cell lines and could also be detected on isolated human glomeruli and whole human kidney cDNA (Figure 2A).

Figure 2.
Figure 2.

Sam68-like mammalian protein 2 (SLM-2) expression in cultured podocytes. (A) Reverse transcriptase–PCR (RT-PCR) for SLM-2 on renal tissue. RT-PCR for SLM-2 resulted in a product of the corresponding size (146 bp) in whole human kidney, isolated human glomeruli, and two human podocyte cell lines (shown Podoc. II [15]); human brain cDNA served as positive control. Non–reverse transcribed samples (RT−) and no-template controls (ddH2O) remained negative in all experiments. (B) Western blot analysis of SLM-2. Protein extracts of two human (Podoc. I [16] and II [15]) and one murine (Podoc. III [17]) podocyte cell lines showed a single band of the expected size for SLM-2 (approximately 50 kD). Control murine brain protein extracts gave a band of identical size. Negative control experiments without primary antibody resulted in no positive signal. (C) Immunofluorescence for SLM-2 on cultured human podocytes. A nuclear signal was demonstrated on all glomerular epithelial cell lines, the human conditionally immortalized cell line shown here (15). Control experiments resulted in no nuclear positivity (insert).

The protein expression of SLM-2 in glomerular epithelial cell lines was first demonstrated by Western blot. A single band of the expected size was found in two human and one murine podocyte cell line with an anti–SLM-2 antiserum; murine brain protein extract served as positive control (Figure 2B). Immunofluorescence experiments were performed to demonstrate the intracellular localization of SLM-2 in human podocytes. In both cell lines, a predominantly nuclear signal was evident as reported in previous studies for SLM-2 (21) (Figure 2C).

Furthermore, expression of SLM-2 in vivo was studied by immunohistochemistry. First, murine and human brain sections were studied to establish the antiserum for SLM-2. A nuclear signal could be demonstrated in both species as reported previously for SLM-2 (22) (Figure 3A). In murine kidney sections, the nuclear staining was restricted to glomerular epithelial cells (Figure 3B). No other glomerular cells were positive for SLM-2. Only in a few individual tubular epithelial cells could discrete nuclear signals also be detected. In human renal sections, the staining was less intense but again restricted to glomerular epithelial cells (Figure 4B). Summarized, these findings confirm the podocyte-specific expression of SLM-2 predicted by the computational approach.

Figure 3.
Figure 3.

Immunohistochemistry of SLM-2 on brain and kidney. (A) SLM-2 staining on murine brain. A positive nuclear signal was found in nearly all neurons of murine brain sections. Similar patterns were found on human brain sections (data not shown). Negative controls that used only the secondary antibody gave no positive signal (insert). (B) SLM-2 staining on murine kidney. In murine kidney, podocytes demonstrated a positive nuclear staining for SLM-2 (arrows). No other glomerular cell type was stained positive for SLM-2. In the tubulointerstitial compartment, a few tubular epithelial cells and some intravascular cells gave slight positive signals (data not shown). Negative controls gave no positive signal (insert).

Figure 4.
Figure 4.

SLM-2 expression in microdissected human glomeruli. (A) SLM-2 mRNA expression. SLM-2 mRNA was quantified in microdissected renal biopsies from control tissue and biopsies from various proteinuric diseases. SLM-2 was significantly induced in all proteinuric groups compared with control samples, independent of the housekeeper gene used for normalization (18S rRNA [shown], glyceraldehyde-3-phosphate dehydrogenase [GAPDH[; *P < 0.05, **P < 0.01 versus controls; CON, controls; MCD, minimal-change disease; BNS, benign nephrosclerosis; MGN, membranous glomerulonephritis; FSGS, focal segmental glomerulosclerosis). (B) Immunohistochemistry for SLM-2. Immunohistochemistry for SLM-2 on human tissue showed low positivity in healthy control subjects but pronounced staining for SLM-2 in glomerular epithelial cells. Shown here are a healthy control subject (A) and a biopsy of a patient with FSGS (B; nonsclerosed part of the glomerular tuft). (C) Quantification of SLM-2 immunohistochemistry. Quantification of the immunohistochemical staining for SLM-2 demonstrated a significantly higher number of positive nuclei in both human proteinuric diseases (MGN and FSGS) compared with CON (n = 5 for each condition; **P < 0.01 versus CON).

SLM-2 Is Induced in Human Proteinuric Diseases

To study a potential regulation of SLM-2 in podocytes in vivo, we investigated the expression of SLM-2 mRNA in microdissected glomeruli from patients with proteinuria and nonproteinuric control subjects. Glomeruli from 77 patients with proteinuria and nonproteinuric control subjects were investigated (minimal-change disease [MCD], n = 13; benign nephrosclerosis [BNS], n = 16; membranous glomerulonephritis [MGN], n = 31; focal segmental glomerulosclerosis [FSGS], n = 9; controls, n = 8 [four tumor nephrectomies and four pretransplant biopsies of living-related donors]). SLM-2 mRNA was significantly induced in the disease groups, independent from the housekeeper used for normalization (18S rRNA, P = 0.01; GAPDH, P = 0.02; SLM-2 expression normalized for 18S rRNA: control 1.45 ± 0.62, MCD 3.23 ± 1.38, BNS 2.92 ± 1.17, MGN 2.79 ± 2.26, FSGS 3.08 ± 1.82; normalized to GAPDH: control 0.06 ± 0.04, MCD 0.23 ± 0.15, BNS 0.14 ± 0.10, MGN 0.10 ± 0.07, FSGS 0.14 ± 0.12; Figure 4A).

To confirm a corresponding induction on protein level, we performed SLM-2 staining of fixed sections from routine renal biopsies. In control tissue (n = 5), only a minority of glomeruli showed positive nuclear staining. In contrast, sections from two different proteinuric diseases (MGN, n = 5; FSGS, n = 5) showed a significantly increased number of SLM-2–positive nuclei in glomerular epithelial cells (control 0.2 ± 0.45, MGN 2.0 ± 0.7, FSGS 2.6 ± 1.14; P < 0.01 for both disease groups versus control subjects; Figure 4, B and C).

Knockdown of SLM-2 in Podocytes Leads to Alternative Splicing of VEGF

These data may reflect a role of SLM-2 in the biology of podocytes during proteinuria. To study the role of SLM-2 in podocytes in vitro, we established a knockdown model. Conditionally immortalized human podocytes were transfected with two different siRNA directed against human SLM-2. Both siRNA led to a significant reduction of SLM-2 mRNA below 20% of control conditions (normalized to GAPDH: time control set as 1.0, negative control 0.99 ± 0.35, SLM-2 siRNA2 0.16 ± 0.04, SLM-2 siRNA3 0.13 ± 0.04; normalized to 18S rRNA: 1.0, 1.19 ± 0.59, 0.14 ± 0.04, 0.11 ± 0.06; n = 4; P < 0.05 for each SLM-2 siRNA compared with control conditions; Figure 5A).

Figure 5.
Figure 5.

Alternative splicing of vascular endothelial growth factor (VEGF-A) after SLM-2 knockdown. (A) Knockdown efficiency by RNA interference. Human conditionally immortalized podocytes were transfected with negative control siRNA1, SLM-2 siRNA2, and SLM-2 siRNA3, respectively. SLM-2 expression was determined 48 h after transfection by real-time RT-PCR. Transfection of SLM-2 siRNA led to a significant decrease of SLM-2 templates below 20% of control conditions. Data shown are normalized to GAPDH; n = 4; *P < 0.05 for SLM-2 siRNA versus control conditions. (B) VEGF RT-PCR with gel electrophoresis. RT-PCR with primers amplifying all VEGF-A isoforms showed a reduction of the VEGF165 isoform (top band) compared with the VEGF121 isoform (bottom band). The order of control and siRNA groups corresponds to C. (C) Real-time RT-PCR for VEGF isoforms. Real-time RT-PCR with isoform-specific primers was performed to re-confirm the above results and showed a 25% reduction of the VEGF165 isoform normalized to all VEGF isoforms in the SLM-2 knockdowns (*P < 0.05 versus control). (D) Sequences of potential SLM-2 binding sites. Exon 7 of the human VEGF gene, included in the VEGF165 isoform, shows an element with high similarity to the potential exonic splicing enhancers binding SLM-2 in other known genes: The sequences in the human transformer-2β (tra) exon 2, human tau exon 2, and murine CD44 exon v5 were described by Stoss et al. (22). The human CD44 exon v5 and VEGF exon 7 were found by alignment. Identical nucleotides to the human VEGF exon 7 sequence are shown in bold.

SLM-2 has been shown to influence alternative splice site selection of different genes. Two of these genes are CD44 and tau; SLM-2 promotes inclusion of the variable exon v5 in CD44 and exons 2 and 3 in tau (22–24). Although mRNA templates for CD44 and tau could be detected in cultured human podocytes and microdissected glomeruli, only SLM-2–independent isoforms were expressed in vitro and in vivo (CD44-v5− and tau-2−3−; data not shown). Next, VEGF-A, which is known to be expressed as alternatively spliced isoforms in podocytes, was studied and endogenous VEGF was found to be alternatively spliced after knockdown of SLM-2. SLM-2 knockdown led to a 25% reduction of the isoform VEGF165, which contains exon 7, compared with control conditions (VEGF165/all VEGF isoforms: time control set as 1.0, negative control siRNA 0.96 ± 0.08, SLM-2 siRNA2 0.74 ± 0.22, SLM-2 siRNA3 0.76 ± 0.16; n = 4; P < 0.05 for both SLM-2 siRNA; Figure 5, B and C). This substantiates a role of SLM-2 for the inclusion of VEGF exon 7 and therefore promoting expression of VEGF165 over VEGF121. As Stoss et al. (22) were able to characterize the potential binding element of SLM-2 in exonic splicing enhancers of different genes, we searched for homologies in the exon 7 of human VEGF. As shown in Figure 5D, a corresponding purin-rich binding site can also be found in exon 7 of the human VEGF gene. Nephrin, another podocyte-expressed gene with reported splice isoforms, did not show a potential SLM-2 binding element (25,26). As SLM-2 is significantly upregulated in proteinuric diseases, the expression of VEGF165 was analyzed in the 77 human renal biopsies specified above. The expression of SLM-2 showed a highly significant correlation with the expression of VEGF165 (r = 0.63, P < 0.01; Figure 6), potentially indicating a role of SLM-2 in mRNA splice site selection in podocytes in vivo.

Figure 6.
Figure 6.

Correlation of VEGF165 isoform and SLM-2 in human biopsies. Expression of SLM-2 mRNA correlated positively with the expression of VEGF165 in microdissected human glomeruli (r = 0.63, P < 0.01). The expression levels of both genes were determined on isolated glomeruli of the 77 patients specified above; each patient’s sample is represented by one dot.

Discussion

The computational approach of DDD allowed the prediction of a novel podocyte expressed gene, SLM-2. Its expression is limited to glomerular epithelial cells in the kidney, it is upregulated in proteinuric diseases, and it is involved in mRNA splice site selection. SLM-2 represents a mammalian member of the STAR family; STAR is an acronym for signal transduction and activation of RNA. Its members share an extended hnRNP K homology (KH) RNA binding domain and are involved in mitosis and cell-cycle regulation, signal transduction, and alternative splice site selection (reviewed in reference 27). The best characterized member of these RNA-binding proteins is Sam68 (Src-associated substrate during mitosis of 68 kD), first described as a target of the tyrosin kinase Src during mitosis (28,29). SLM-2 that is closely related to Sam68 was first described by three independent groups in 1999. Di Fruscio et al. (13) detected SLM-2 by screening a λ ZAP II mouse brain cDNA library and demonstrated its RNA-binding potential. This RNA binding is inhibited after phosphorylation by the epithelial-specific tyrosine kinase BRK/sik (30), whereas other kinases, such as Src and Fyn, do not phosphorylate SLM-2 (13,31). Venables et al. first showed SLM-2’s nuclear localization (21); later, this group could demonstrate its regulation by proteasomic degradation (32). A significant expression of SLM-2 in the kidney was first detected by Lee’s group (33), who was also showing that expression of SLM-2 can lead to cell growth arrest.

As presented in our study, SLM-2 is expressed in glomerular epithelial cells and influences the splice site selection of an important glomerular signal transductor, VEGF. The role of VEGF in development and function of the renal glomerulus has been studied comprehensively and was reviewed recently (34). It appears in three major isoforms, named for the number of amino acids VEGF121, VEGF165, and VEGF189, corresponding on mRNA levels to exons 1 to 5 and 8 in all isoforms and additionally exon 7 (VEGF165) or 6 and 7 (VEGF189) (35). In the glomerulus, all major VEGF isoforms are produced by the podocyte (36). The VEGF receptors are found mainly on endothelial cells, but a weaker expression on podocytes suggests an additional paracrine signaling pattern. Individual VEGF isoforms play specific roles in glomerular function. Podocyte-specific deletion of VEGF in mice leads to death at birth and absence of a glomerular filtration barrier (37). Mice that express only VEGF122 have a significant reduction in the number of vascularized glomeruli compared with controls (38). Thus, VEGF122 provides partial but incomplete rescue compared with the complete podocyte-specific knockout. In contrast, podocyte-specific overexpression of VEGF165 leads to a striking collapsing glomerulopathy (37). The systemic administration of VEGF165 can accelerate glomerular repair in necrotizing glomerulonephritis (39). Differences between VEGF isoforms are potentially due to differences in binding to and release from the glomerular basement membrane (40,41). However, little is known about the molecular regulation of alternative splice site selection of VEGF mRNA. The data presented in this study suggest a significant role of SLM-2 on isoform-specific expression of VEGF, hereby potentially influencing the functional regulation of the glomerular filter.

It also demonstrates the importance of RNA splicing for cell type–specific regulation of gene expression. Besides transcriptional regulation and RNA interference, this biologic mechanism enables the cell to expand the limited amount of encoded genes toward the plethora of cell-specific proteins (42,43). RNA splice site selection has not previously been studied in podocytes but could add a novel level to understanding the complex biology of this highly specialized cell type.

The computational approach of DDD used in this study allowed for the first time the prediction of a cell type–specific gene expression by comparison of two organs with other cDNA libraries. This was possible because of the shared transcriptome of neuronal cells and podocytes. Although this successful proof of concept led to the identification of SLM-2, three limitations have to be discussed. First, SLM-2 has its most prominent expression in testis and therefore was even named testis-STAR (21). The reason that our DDD results revealed this gene is the exclusion of testis in the list of organs included in the analysis. This was necessary because of the exhausting number of transcripts found in human testis, making a testis-included DDD screen almost impossible. Furthermore, for our approach, brain cDNA were especially important, and it was shown recently by Guo et al. (20) that the transcriptomes of brain and testis are the most similar ones. This made the exclusion of testis central for searching cDNA predominantly expressed in brain and kidney. The second limitation of our approach is the lack of suborgan-specific cDNA libraries. The availability of glomerular cDNA libraries (instead of whole kidney) possibly would have increased the predictive impact of the analysis. More specific, the reason that no known podocyte-specific gene was found by the DDD experiment may be the low quantitative representation of podocytes in whole-kidney libraries. As third weakness of the DDD approach, one may call the low predictive value of 33% (one of three predicted genes confirmed). This predictive power is comparable with the ones of previous studies that used DDD, in which 20 to 30% of predicted templates could be confirmed (11,44,45). It underlines the importance of confirmation experiments after computational analysis, but as with all novel techniques, the number of false-positive calls should decrease when DDD becomes a more frequently used tool for gene expression prediction.

This study stresses the potential that DDD has to predict cell type–specific gene expression. It allowed prediction of the podocyte-restricted expression of SLM-2 in the glomerulus. Along with the confirmation of this finding, SLM-2 was shown to be upregulated in proteinuric diseases and involved in splice site selection of VEGF. Studying the regulation of SLM-2 activity in vitro and in vivo may further elucidate alternative splicing as a central mechanism of podocyte biology.

Acknowledgments

This study was supported by the State of Bavaria “Bayerischer Habilitationsförderpreis” and Friedrich-Baur-Stiftung to C.D.C., and EU-FPV: Chronic Kidney Disease (QLRT-2001-01215) to P.P.D., M.P.R., and M.K.

We thank S. Irrgang, K. Frach, and I. Edenhofer for excellent technical assistance; B. Luckow for support; and D. Schlondorff for suggestions, support, and critical revision of the manuscript.

We thank all members of the European Renal cDNA Bank (ERCB) and their patients for support and cooperation. The members of the ERCB at the time of this study were as follows: J.P. Rougier, P. Ronco, Paris, France; M.P. Rastaldi, G. D'Amico, Milan, Italy; F. Mampaso, Madrid, Spain; P. Doran, H.R. Brady, Dublin, Ireland; D. Monks, C. Wanner, Würzburg, Germany; A.J. Rees, Aberdeen, UK; F. Strutz, G. Muller, Gottingen, Germany; P. Mertens, J. Floege, Aachen, Germany; T. Risler, Tubingen, Germany; L. Gesualdo, F.P. Schena, Bari, Italy; J. Gerth, U. Ott, G. Wolf, Jena, Germany; R. Oberbauer, D. Kerjaschki, Vienna, Austria; B. Banas, B. Kramer, Regensburg, Germany; W. Samtleben, Munich, Germany; H. Peters, H.H. Neumayer, Berlin, Germany; K. Ivens, B. Grabensee, Dusseldorf, Germany; M. Zeier, H.J. Gröne, Heidelberg, Germany; M. Merta, V. Tesar, Prague, Czech Republic; K. Blouch, K. Lemley, Stanford, California; R. Nelson, Phoenix, Arizona; C.D. Cohen, M. Kretzler, D. Schlondorff, Munich, Germany.

Footnotes

  • Published online ahead of print. Publication date available at www.jasn.org.

References

  1. ↵
    Pavenstadt H, Kriz W, Kretzler M: Cell biology of the glomerular podocyte. Physiol Rev 83 : 253 –307, 2003
    OpenUrlCrossRefPubMed
  2. ↵
    Beltcheva O, Kontusaari S, Fetissov S, Putaala H, Kilpelainen P, Hokfelt T, Tryggvason K: Alternatively used promoters and distinct elements direct tissue-specific expression of nephrin. J Am Soc Nephrol 14 : 352 –358, 2003
    OpenUrlAbstract/FREE Full Text
  3. Ahola H, Heikkila E, Astrom E, Inagaki M, Izawa I, Pavenstadt H, Kerjaschki D, Holthofer H: A novel protein, densin, expressed by glomerular podocytes. J Am Soc Nephrol 14 : 1731 –1737, 2003
    OpenUrlAbstract/FREE Full Text
  4. Mundel P, Heid HW, Mundel TM, Kruger M, Reiser J, Kriz W: Synaptopodin: An actin-associated protein in telencephalic dendrites and renal podocytes. J Cell Biol 139 : 193 –204, 1997
    OpenUrlAbstract/FREE Full Text
  5. Thomas PE, Wharram BL, Goyal M, Wiggins JE, Holzman LB, Wiggins RC: GLEPP1, a renal glomerular epithelial cell (podocyte) membrane protein-tyrosine phosphatase. Identification, molecular cloning, and characterization in rabbit. J Biol Chem 269 : 19953 –19962, 1994
    OpenUrlAbstract/FREE Full Text
  6. Rastaldi MP, Armelloni S, Berra S, Li M, Pesaresi M, Poczewski H, Langer B, Kerjaschki D, Henger A, Blattner SM, Kretzler M, Wanke R, D’Amico G: Glomerular podocytes possess the synaptic vesicle molecule Rab3A and its specific effector rabphilin-3a. Am J Pathol 163 : 889 –899, 2003
    OpenUrlCrossRefPubMed
  7. Tagawa M, Shirasawa T, Yahagi Y, Tomoda T, Kuroyanagi H, Fujimura S, Sakiyama S, Maruyama N: Identification of a receptor-type protein tyrosine phosphatase expressed in postmitotic maturing neurons: Its structure and expression in the central nervous system. Biochem J 321 : 865 –871, 1997
  8. ↵
    Gloy J, Reitinger S, Fischer KG, Schreiber R, Boucherot A, Kunzelmann K, Mundel P, Pavenstadt H: Amino acid transport in podocytes. Am J Physiol Renal Physiol 278 : F999 –F1005, 2000
    OpenUrlPubMed
  9. ↵
    Kobayashi N: Mechanism of the process formation; podocytes vs. neurons. Microsc Res Tech 57 : 217 –223, 2002
    OpenUrlCrossRefPubMed
  10. ↵
    Strausberg RL, Buetow KH, Emmert-Buck MR, Klausner RD: The cancer genome anatomy project: Building an annotated gene index. Trends Genet 16 : 103 –106, 2000
    OpenUrlCrossRefPubMed
  11. ↵
    Scheurle D, DeYoung MP, Binninger DM, Page H, Jahanzeb M, Narayanan R: Cancer gene discovery using digital differential display. Cancer Res 60 : 4037 –4043, 2000
    OpenUrlAbstract/FREE Full Text
  12. ↵
    Lord-Grignon J, Tetreault N, Mears AJ, Swaroop A, Bernier G: Characterization of new transcripts enriched in the mouse retina and identification of candidate retinal disease genes. Invest Ophthalmol Vis Sci 45 : 3313 –3319, 2004
    OpenUrlAbstract/FREE Full Text
  13. ↵
    Di Fruscio M, Chen T, Richard S: Characterization of Sam68-like mammalian proteins SLM-1 and SLM-2: SLM-1 is a Src substrate during mitosis. Proc Natl Acad Sci U S A 96 : 2710 –2715, 1999
    OpenUrlAbstract/FREE Full Text
  14. ↵
    Madden SF, O’Donovan B, Furney SJ, Brady HR, Silvestre G, Doran PP: Digital extractor: Analysis of digital differential display output. Bioinformatics 19 : 1594 –1595, 2003
    OpenUrlCrossRefPubMed
  15. ↵
    Saleem MA, O’Hare MJ, Reiser J, Coward RJ, Inward CD, Farren T, Xing CY, Ni L, Mathieson PW, Mundel P: A conditionally immortalized human podocyte cell line demonstrating nephrin and podocin expression. J Am Soc Nephrol 13 : 630 –638, 2002
    OpenUrlAbstract/FREE Full Text
  16. ↵
    Delarue F, Virone A, Hagege J, Lacave R, Peraldi MN, Adida C, Rondeau E, Feunteun J, Sraer JD: Stable cell line of T-SV40 immortalized human glomerular visceral epithelial cells. Kidney Int 40 : 906 –912, 1991
    OpenUrlCrossRefPubMed
  17. ↵
    Mundel P, Reiser J, Zuniga Mejia Borja A, Pavenstadt H, Davidson GR, Kriz W, Zeller R: Rearrangements of the cytoskeleton and cell contacts induce process formation during differentiation of conditionally immortalized mouse podocyte cell lines. Exp Cell Res 236 : 248 –258, 1997
    OpenUrlCrossRefPubMed
  18. ↵
    Cohen CD, Frach K, Schlondorff D, Kretzler M: Quantitative gene expression analysis in renal biopsies: A novel protocol for a high-throughput multicenter application. Kidney Int 61 : 133 –140, 2002
    OpenUrlCrossRefPubMed
  19. ↵
    Cui TG, Foster RR, Saleem M, Mathieson PW, Gillatt DA, Bates DO, Harper SJ: Differentiated human podocytes endogenously express an inhibitory isoform of vascular endothelial growth factor (VEGF165b) mRNA and protein. Am J Physiol Renal Physiol 286 : F767 –F773, 2004
    OpenUrlCrossRefPubMed
  20. ↵
    Guo J, Zhu P, Wu C, Yu L, Zhao S, Gu X: In silico analysis indicates a similar gene expression pattern between human brain and testis. Cytogenet Genome Res 103 : 58 –62, 2003
    OpenUrlCrossRefPubMed
  21. ↵
    Venables JP, Vernet C, Chew SL, Elliott DJ, Cowmeadow RB, Wu J, Cooke HJ, Artzt K, Eperon IC: T-STAR/ETOILE: A novel relative of SAM68 that interacts with an RNA-binding protein implicated in spermatogenesis. Hum Mol Genet 8 : 959 –969, 1999
    OpenUrlCrossRefPubMed
  22. ↵
    Stoss O, Olbrich M, Hartmann AM, Konig H, Memmott J, Andreadis A, Stamm S: The STAR/GSG family protein rSLM-2 regulates the selection of alternative splice sites. J Biol Chem 276 : 8665 –8673, 2001
    OpenUrlAbstract/FREE Full Text
  23. Arikan MC, Memmott J, Broderick JA, Lafyatis R, Screaton G, Stamm S, Andreadis A: Modulation of the membrane-binding projection domain of tau protein: Splicing regulation of exon 3. Brain Res Mol Brain Res 101 : 109 –121, 2002
    OpenUrlPubMed
  24. ↵
    Li K, Arikan MC, Andreadis A: Modulation of the membrane-binding domain of tau protein: Splicing regulation of exon 2. Brain Res Mol Brain Res 116 : 94 –105, 2003
    OpenUrlPubMed
  25. ↵
    Ahola H, Wang SX, Luimula P, Solin ML, Holzman LB, Holthofer H: Cloning and expression of the rat nephrin homolog. Am J Pathol 155 : 907 –913, 1999
    OpenUrlPubMed
  26. ↵
    Holthofer H, Ahola H, Solin ML, Wang S, Palmen T, Luimula P, Miettinen A, Kerjaschki D: Nephrin localizes at the podocyte filtration slit area and is characteristically spliced in the human kidney. Am J Pathol 155 : 1681 –1687, 1999
    OpenUrlCrossRefPubMed
  27. ↵
    Lukong KE, Richard S: Sam68, the KH domain-containing superSTAR. Biochim Biophys Acta 1653 : 73 –86, 2003
    OpenUrlCrossRefPubMed
  28. ↵
    Taylor SJ, Shalloway D: An RNA-binding protein associated with Src through its SH2 and SH3 domains in mitosis. Nature 368 : 867 –871, 1994
    OpenUrlCrossRefPubMed
  29. ↵
    Fumagalli S, Totty NF, Hsuan JJ, Courtneidge SA: A target for Src in mitosis. Nature 368 : 871 –874, 1994
    OpenUrlCrossRefPubMed
  30. ↵
    Haegebarth A, Heap D, Bie W, Derry JJ, Richard S, Tyner AL: The nuclear tyrosine kinase BRK/Sik phosphorylates and inhibits the RNA-binding activities of the Sam68-like mammalian proteins SLM-1 and SLM-2. J Biol Chem 279 : 54398 –54404, 2004
    OpenUrlAbstract/FREE Full Text
  31. ↵
    Stoss O, Novoyatleva T, Gencheva M, Olbrich M, Benderska N, Stamm S: p59(fyn)-mediated phosphorylation regulates the activity of the tissue-specific splicing factor rSLM-1. Mol Cell Neurosci 27 : 8 –21, 2004
    OpenUrlCrossRefPubMed
  32. ↵
    Venables JP, Dalgliesh C, Paronetto MP, Skitt L, Thornton JK, Saunders PT, Sette C, Jones KT, Elliott DJ: SIAH1 targets the alternative splicing factor T-STAR for degradation by the proteasome. Hum Mol Genet 13 : 1525 –1534, 2004
    OpenUrlCrossRefPubMed
  33. ↵
    Lee J, Burr JG: Salpalpha and Salpbeta, growth-arresting homologs of Sam68. Gene 240 : 133 –147, 1999
    OpenUrlCrossRefPubMed
  34. ↵
    Eremina V, Quaggin SE: The role of VEGF-A in glomerular development and function. Curr Opin Nephrol Hypertens 13 : 9 –15, 2004
    OpenUrlCrossRefPubMed
  35. ↵
    Tischer E, Mitchell R, Hartman T, Silva M, Gospodarowicz D, Fiddes JC, Abraham JA: The human gene for vascular endothelial growth factor. Multiple protein forms are encoded through alternative exon splicing. J Biol Chem 266 : 11947 –11954, 1991
    OpenUrlAbstract/FREE Full Text
  36. ↵
    Kretzler M, Schroppel B, Merkle M, Huber S, Mundel P, Horster M, Schlondorff D: Detection of multiple vascular endothelial growth factor splice isoforms in single glomerular podocytes. Kidney Int Suppl 67 : S159 –S161, 1998
    OpenUrlCrossRefPubMed
  37. ↵
    Eremina V, Sood M, Haigh J, Nagy A, Lajoie G, Ferrara N, Gerber HP, Kikkawa Y, Miner JH, Quaggin SE: Glomerular-specific alterations of VEGF-A expression lead to distinct congenital and acquired renal diseases. J Clin Invest 111 : 707 –716, 2003
    OpenUrlCrossRefPubMed
  38. ↵
    Mattot V, Moons L, Lupu F, Chernavvsky D, Gomez RA, Collen D, Carmeliet P: Loss of the VEGF(164) and VEGF(188) isoforms impairs postnatal glomerular angiogenesis and renal arteriogenesis in mice. J Am Soc Nephrol 13 : 1548 –1560, 2002
    OpenUrlAbstract/FREE Full Text
  39. ↵
    Shimizu A, Masuda Y, Mori T, Kitamura H, Ishizaki M, Sugisaki Y, Fukuda Y: Vascular endothelial growth factor165 resolves glomerular inflammation and accelerates glomerular capillary repair in rat anti-glomerular basement membrane glomerulonephritis. J Am Soc Nephrol 15 : 2655 –2665, 2004
    OpenUrlAbstract/FREE Full Text
  40. ↵
    Gengrinovitch S, Berman B, David G, Witte L, Neufeld G, Ron D: Glypican-1 is a VEGF165 binding proteoglycan that acts as an extracellular chaperone for VEGF165. J Biol Chem 274 : 10816 –10822, 1999
    OpenUrlAbstract/FREE Full Text
  41. ↵
    Iozzo RV, San Antonio JD: Heparan sulfate proteoglycans: Heavy hitters in the angiogenesis arena. J Clin Invest 108 : 349 –355, 2001
    OpenUrlCrossRefPubMed
  42. ↵
    Black DL: Mechanisms of alternative pre-messenger RNA splicing. Annu Rev Biochem 72 : 291 –336, 2003
    OpenUrlCrossRefPubMed
  43. ↵
    Stamm S, Ben-Ari S, Rafalska I, Tang Y, Zhang Z, Toiber D, Thanaraj TA, Soreq H: Function of alternative splicing. Gene 344C : 1 –20, 2005
    OpenUrl
  44. ↵
    Ghadersohi A, Odunsi K, Lele S, Collins Y, Greco WR, Winston J, Liang P, Sood AK: Prostate derived Ets transcription factor shows better tumor-association than other cancer-associated molecules. Oncol Rep 11 : 453 –458, 2004
    OpenUrlPubMed
  45. ↵
    Dennis JL, Vass JK, Wit EC, Keith WN, Oien KA: Identification from public data of molecular markers of adenocarcinoma characteristic of the site of origin. Cancer Res 62 : 5999 –6005, 2002
    OpenUrlAbstract/FREE Full Text
Back to top

In this issue

View Selected Citations (0)
Print
Download PDF
Email Article
Citation Tools
Request Permissions
Sam68-Like Mammalian Protein 2, Identified by Digital Differential Display as Expressed by Podocytes, Is Induced in Proteinuria and Involved in Splice Site Selection of Vascular Endothelial Growth Factor
Clemens D. Cohen, Peter P. Doran, Simone M. Blattner, Monika Merkle, Guo Q. Wang, Holger Schmid, Peter W. Mathieson, Moin A. Saleem, Anna Henger, Maria P. Rastaldi, Matthias Kretzler
JASN Jul 2005, 16 (7) 1958-1965; DOI: 10.1681/ASN.2005020204
del.icio.us logo Digg logo Reddit logo Twitter logo CiteULike logo Facebook logo Google logo Mendeley logo

Jump to section