A Novel Approach to Investigation of the Pathogenesis of Active Minimal-Change Nephrotic Syndrome Using Subtracted cDNA Library Screening

Abstract

Clinical and experimental observations suggest that minimal-change nephrotic syndrome (MCNS) results from T cell dysfunction, via unknown mechanisms. For the identification of genes that are potentially involved in MCNS, a subtractive cDNA library was constructed from cDNA from T cell-enriched peripheral blood mononuclear cells obtained from the same patient during relapse versus remission (“relapse minus remission”). This library was screened by differential hybridization with forward (“relapse minus remission”) and reverse (“remission minus relapse”) subtractive cDNAs probes, as well as unsubtracted probes from relapse and remission, and irrelevant nephrotic syndrome (membranous nephropathy). A total of 84 transcripts were isolated, of which 12 matched proteins of unknown function and 30 were unknown clones. Among the 42 known transcripts, at least 18 are closely involved in the T cell receptor-mediated complex signaling cascade, including genes encoding components of the T cell receptor and proteins associated with the cytoskeletal scaffold, as well as transcription factors. In particular, it was demonstrated that the expression levels of Fyb/Slap, L-plastin, and grancalcin were increased during relapse, suggesting that the integration of proximal signaling after T cell engagement involves the preferential recruitment of these cytoskeleton-associated proteins in MCNS. Because very low levels of interleukin-12 receptor β2 mRNA were detected in relapse samples, the interleukin-12 signaling pathway might be defective, suggesting that, in MCNS, T cell activation evolves toward a T helper 2 phenotype. Therefore, the combination of subtractive cloning and differential screening constitutes an efficient approach to the identification of genes that are likely to be involved in the pathophysiologic processes of MCNS.

Minimal-change nephrotic syndrome (MCNS) is a glomerular disease that is characterized by heavy proteinuria and a relapsing/remitting course, without histologic evidence of classic immune mechanism-mediated injury (¹). Among these patients, persistent immunogenic stimuli (such as viral infections, immunizations, or allergens) might trigger nephrotic relapses. During relapses, CD4⁺ and CD8⁺ T cell subsets are expanded and levels of cytokines such as tumor necrosis factor-α, interleukin-8 (IL-8), and IL-13 are increased (^2–4). Convincing evidence for an immune origin of MCNS came from its sensitivity to immunosuppressive therapy, but the molecular link between the immune system and kidney disease is still unknown. Recently, we demonstrated that nuclear extracts of peripheral blood mononuclear cells (PBMC) from patients undergoing relapses displayed persistently high levels of nuclear factor-κB (NF-κB) DNA binding activity, which might account for the increased levels of cytokines (⁵). In contrast, remissions were characterized by upregulation of IκBα and downregulation of most of those cytokines.

It is thought that mechanisms initiating MCNS operate in the context of immune alterations that affect particular peripheral T cells. To identify the molecular events underlying T cell dysfunction during MCNS relapse, we undertook subtractive and differential cloning of transcripts that are selectively induced or upregulated in lymphocytes of patients with MCNS relapse. To this end, we prepared a subtractive cDNA library from T cell-enriched PBMC obtained from a patient during relapse and remission. Differential screening of this library led to the identification of 84 clones. At least 18 clones encode parts of genes involved in tightly coordinated steps of T cell activation, supporting the hypothesis that MCNS is a T cell-mediated disease. The expression of Fyb/Slap, L-plastin, and grancalcin (proteins involved in cytoskeletal rearrangement and integrin activation) was increased during relapse. Furthermore, we demonstrated that this T cell response was associated with downregulation of IL-12 receptor (IL-12R) β2 mRNA levels, suggesting that a T helper 2 (Th2) phenotype develops early in the course of this disease.

Materials and Methods

Patients

The cohort of patients analyzed in this study has already been described (⁵). For children, the criteria of the International Study of Kidney Disease in Children were used for the diagnosis and management of MCNS (⁶). For adults, the diagnosis of MCNS or membranous nephropathy (MN) was confirmed by renal biopsy before inclusion. Blood samples were obtained from patients with relapses before any treatment. Remission samples were collected during periods of inactive disease without steroid treatment. Informed consent was obtained from the parents and whenever possible from the pediatric patients, as well as from adult patients and normal volunteers.

The patient selected for construction of the subtractive library was a 24-yr-old Caucasian man who was referred to us at the onset of nephrotic syndrome. The clinical examination demonstrated edema, but no other anomalies were noted. BP was 130/70 mmHg. Laboratory assessments demonstrated the following findings: serum creatinine concentration, 63 μM; hemoglobin level, 1.4 g/dl; white blood cell count, 5500/ml; platelet count, 250,000/ml; serum albumin concentration, 2.3 g/dl; proteinuria, 30 g/d (selective type); urinary white and red blood cell counts, <5000/ml (leukocyturia, negative). The serum complement concentration was normal, the serum IgG level was low (4 dg/dl), and the serum IgE level was increased (400 KUI; normal, 150). Tests for serum antinuclear antibodies, anti-streptolysin O, anti-double-stranded DNA, and anti-neutrophil cytoplasm antibodies, and cryoglobulinemia yielded negative results. Serologic assays were negative for Epstein-Barr virus, Yersinia, Lyme disease, syphilis, and hepatitis B and C and positive for cytomegalovirus (IgG type). The renal histologic examination demonstrated minimal changes in 25 glomeruli, without mesangial cell proliferation, interstitial cell infiltration, or focal glomerulohyalinosis. Immunofluorescence staining for IgA, IgG, and IgM, complement 3 and 4, and fibrinogen yielded negative results.

The patient was treated with prednisone (1 mg/kg per d) for 4 wk, and the dosage was then progressively tapered. Complete remission was obtained within 10 d after the initiation of treatment. Blood samples were obtained from this patient during relapse (before the initiation of steroid therapy) and 2 mo after remission (while the patient was receiving 20 mg/d prednisone therapy).

Construction of Forward (“Relapse minus Remission”) and Reverse (“Remission minus Relapse”) Subtracted cDNA

Peripheral blood samples were obtained during the relapse and remission phases, and the mononuclear cell fraction (PBMC) was isolated through a Ficoll/Hypaque gradient. A T cell-enriched population was obtained by filtering the PBMC suspension through a nylon wool column. Total RNA was prepared by using a Qiagen kit (Qiagen, Chatsworth, CA), according to the instructions provided by the manufacturer. The integrity of the RNA was confirmed by analysis of a 1-μg sample on a 0.8% agarose gel, followed by ethidium bromide staining. Poly(A) RNA was purified by using an Oligotex mRNA kit (Qiagen). “Relapse” and “remission” cDNA syntheses were performed in parallel, with 1 μg of poly(A) RNA each, using the reverse transcriptase SII (Life Technologies BRL, Eragny, France) and primers provided in the PCR-select cDNA subtraction kit (Clontech, Palo Alto, CA). Both cDNA populations were digested with the restriction enzyme RsaI, to yield shorter, blunt-end molecules (⁷).

The forward subtracted cDNA was prepared as follows: the relapse cDNA was subdivided into two parts, and each was ligated to a different adaptor. Fractions of each adaptor-linked relapse cDNA were mixed, and the mixture was used as an unsubtracted cDNA control sample. Subtractive hybridization was performed in two rounds. First, each adaptor-linked cDNA population was separately mixed with a 60-fold excess of remission cDNA. After denaturation for 90 s at 98°C, subtractive hybridization was performed for 8 h at 68°C. This first hybridization round enriches for cDNA sequences specifically expressed during relapse. Then, the two reaction products were mixed and a fivefold excess of denatured remission cDNA was added. A second hybridization was performed for 16 h at 68°C. During this step, single-stranded cDNA specific for the relapse phase, bearing different adaptors, formed hybrids that were subsequently amplified by two rounds of PCR. In the first PCR (24 cycles), only hybrid cDNA were exponentially amplified. The second nested PCR (12 cycles) enriched these specific sequences while reducing background levels.

The reverse subtracted cDNA was prepared by using the same protocol but switching the relapse and remission cDNA. For assessment of the efficiency of the subtraction, aliquots of both subtracted and unsubtracted cDNA were blotted and probed with a 550-bp, PstI, β₂-microglobulin fragment, using standard protocols (⁸).

Cloning of the Forward Subtracted cDNA into pBluescript II SK(+) Phagemid Vector

The amplified forward subtracted cDNA was blunt-ended, ligated to XhoI-NotI adaptors (Stratagene), and inserted into the NotI-digested pBluescript II SK(+) phagemid (Stratagene) (⁸). Supercompetent Escherichia coli XL-1 blue cells (Stratagene) were transformed with an aliquot of the ligation mixture, by heat shock (⁸).

One fraction of the library (10,000 colonies) was plated onto nitrocellulose filters. Each master filter was duplicated twice and stored at −20°C while the duplicates were treated for differential screening (⁸).

Preparation of Probes and Differential Screening

Two micrograms of subtracted cDNA (forward and reverse) were sequentially digested with RsaI, SmaI, and EagI restriction enzymes, to remove all adaptors. After purification on a Chromaspin-100 column (Clontech), 80 ng of the digested cDNA were labeled with 5 μCi of [³²P]dCTP (3000 Ci/mmol; Amersham, Paisley, UK). Multiplex probes from unsubtracted, relapse, and remission MCNS samples, as well as MN samples, were prepared from 10 μg of PBMC total RNA as described (⁹).

Each duplicate filter with the subtracted cDNA library was incubated with the forward or reverse subtracted probe (2 × 10⁶ cpm/ml) at 72°C for 16 h and was washed for 4 × 30 min in 2× SSC (1× SSC is 150 mM NaCl, 15 mM sodium citrate, pH 7)/0.5% sodium dodecyl sulfate (SDS) and for 3 × 30 min in 0.2× SSC/0.5% SDS at 68°C. After exposure, filters were dehybridized by boiling in 0.5% SDS for 10 min and were then rehybridized with unsubtracted multiplex probes as described above.

RNA Dot-Blot Analysis

PBMC RNA isolated from patients experiencing relapses without steroid therapy and from those in remission were pooled separately and denatured at 65°C. Then, 2 μg of each mixture were dot-blotted on Hybond-N membranes (Amersham). The membranes were washed twice in 6× SSC and cross-linked by exposure to ultraviolet light. Hybridization was performed in 5× SSPE (1× SSPE is 150 mM NaCl, 10 M NaH₂PO₄-H₂O, 1 mM ethylenediaminetetraacetate), 50% deionized formamide, 2× Denhardt’s reagent [1× Denhardt’s reagent is 0.1% Ficoll type 400 (Pharmacia, St. Albans, Herts, UK), 0.1% polyvinylpyrrolidone, 0.1% bovine serum albumin], 0.5% SDS, 100 μg/ml salmon sperm DNA, with 2 × 10⁶ cpm/ml levels of selected clones. In parallel, the sensitivity of hybridization was checked by using a glyceraldehyde phosphate dehydrogenase (GAPDH) probe.

DNA Sequencing

The preparation and sequencing of double-stranded plasmid DNA templates were performed as described previously (¹⁰). Nucleic acid and protein database searches were performed by using resources of the National Center for Biotechnology Information.

Semiquantitative Reverse Transcription-PCR

The primer sequences and the main characteristics of PCR are indicated in Table 1. The expression levels of several transcripts listed in Table 1 were analyzed by semiquantitative reverse transcription (RT)-PCR, as described previously (⁵). Southern blots of amplified products were detected with specific internal oligonucleotides. PCR results were normalized to GAPDH expression.

Table 1: Sets of primers used for semiquantitative RT-PCR^a

Western Blotting

Cytoplasmic extracts of PBMC were prepared and protein concentrations were determined as described previously (⁵). A polyclonal antibody raised against grancalcin was a gift from Lollike Karsten (Novo Nordisk, Bagsvaerd, Denmark) and was used at 1/5000. Monoclonal anti-human L-plastin antibody (anti-LPLA4.1) was a gift from Eric J. Brown and Hua Shen (University of California, San Francisco, San Francisco, CA) and was used at 1 μg/ml. Monoclonal anti-human Fyb/Slap antibody, a mouse IgG1 reacting with the SH3 domain of Fyb/Slap (BD Transduction Laboratories, Los Angeles, CA), was used at 1 μg/ml. Polyclonal anti-actin antibody (A2066) was purchased from Sigma Chemical Co. (St. Louis, MO) and was used at 1/2500. Cytoplasmic protein extracts (50 to 60 μg) were resolved by 10% SDS-polyacrylamide gel electrophoresis and were transferred to nitrocellulose membranes by electroblotting. Immunoblotting and detection (ECL detection kit; Amersham) were performed according to the instructions provided by the manufacturer.

Results

Construction and Differential Screening of the Forward Subtracted cDNA Library

T cell dysfunction in MCNS involves the turning on and off of a number of genes that are likely to play key roles in the development of the disease. To identify these genes, we isolated cDNA corresponding to upregulated mRNA in T cells from a patient with MCNS who was experiencing nephrotic relapse, using the strategy outlined in Figure 1. As a first step, we performed subtractive hybridization of cDNA from T cell-enriched PBMC from the same patient in relapse versus remission, to avoid differences among individuals. The forward subtracted cDNA population, enriched in relapse-induced cDNA, was cloned in pBluescript II SK(+) phagemid, and 10,000 clones were analyzed by differential hybridization with the following probes: for positive screening, forward subtracted and first-strand relapse cDNA probes; for negative screening, reverse subtracted cDNA, first-strand remission cDNA, and MN probes. The efficiency of the subtraction was assessed by analysis of β₂-microglobulin expression in subtracted and unsubtracted samples. As presented in Figure 2, the β₂-microglobulin probe demonstrated a strong signal with the unsubtracted double-stranded cDNA, but no signal was detectable with the forward subtracted templates (Figure 2). The results of differential screening are presented in Figure 3. Preliminary screening of the library with relapse and remission unsubtracted cDNA probes identified approximately 1000 clones. These clones were arrayed and hybridized with three types of probes. The forward subtracted probe hybridized to all clones, with variable intensity (Figure 3A). In contrast, approximately 40 or 45% of the clones did not hybridize to the reverse subtracted probe or corresponded to downregulated mRNA, respectively (Figure 3B). Hybridization of the library with the MN probe demonstrated that 75% of the clones exhibited significant signals, indicating that these clones corresponded to genes likely to be upregulated in response to the nephrotic state, independently of MCNS (Figure 3C). Finally, 127 clones appeared to be selectively upregulated and/or exclusively expressed during MCNS relapse; these were considered to be relevant to the disease and were retained for further analysis.

Figure 1.: Strategy for the detection of genes differentially expressed in minimal-change nephrotic syndrome (MCNS). Double-stranded cDNA was synthesized from 1 μg of peripheral blood mononuclear cell (PBMC) poly(A) RNA obtained from the same patient during relapse and remission phases. Subtractive hybridizations (forward and reverse) and PCR amplifications were performed as described in Materials and Methods. Forward subtracted cDNA was cloned in pBluescript II SK(+) phagemid. Ten thousand clones from the library were screened with two types of probes, i.e., positive probes, consisting of forward subtracted cDNA and unsubtracted relapse cDNA, and negative probes, including reverse subtracted cDNA, unsubtracted remission cDNA, and membranous nephropathy (MN) cDNA.

Figure 2.: Analysis of subtracted cDNA for β₂-microglobulin expression. Subtracted (lanes a) and unsubtracted (lanes b) cDNA were amplified with two rounds of PCR amplification. One-fifth of the first and second PCR products were electrophoresed on a 1.5% agarose gel, transferred to a nylon filter, and hybridized with a β₂-microglobulin probe. DNA molecular weight markers are indicated at the left.

Figure 3.: Differential screening of subtracted clones. Autoradiograms of identical filters containing sets of cDNA clones selected from the cDNA library enriched in genes expressed in MCNS are presented. The filters were hybridized with three [α-³²P]dCTP-labeled multiplex probes. (A) Forward subtracted probe, corresponding to cDNA enriched in genes expressed in MCNS relapse. (B) Reverse subtracted probe, corresponding to cDNA enriched in genes expressed in remission. (C) First-strand cDNA MN probe synthesized from 10 μg of PBMC total RNA. Arrowheads indicate clones corresponding to transcripts upregulated in MCNS relapse (black), compared with reverse MCNS and MN findings (white).

DNA Sequence and Expression Analyses of the Subtracted Clones

Partial sequences of these 127 cDNA were determined and compared with sequences available in databases (Table 2). Sequences obtained for 97 clones corresponded to 54 known human genes. The very low redundancy of these sequences, including sequences related to abundant transcripts such as β-actin and Elf-1α, emphasized the efficiency of the subtraction strategy. Among these transcripts, 12 corresponded to proteins with no assigned function. Thirty sequences did not match any sequences in the databases and thus correspond to currently unidentified genes.

Table 2: Transcripts upregulated in PBMC from patients experiencing MCNS relapses^a

Table 2A: Continued

We analyzed the expression of eight genes involved in the T cell receptor (TCR) signaling pathway. In parallel, we analyzed the expression of seven transcripts with no database matches. Dot blots loaded with pools of PBMC RNA from patients experiencing nephrotic relapses without steroid therapy and patients in remission were hybridized with radiolabeled cDNA inserts corresponding to the selected transcripts. As presented in Figure 4A, the signal intensities detected with the L-plastin, grancalcin, TCR δ chain, nuclear factor of activated T cells 5 (NFAT5), macropain (proteasome α2 subunit), IL-7R, IgE-dependent histamine-releasing factor, and Jak1 probes were increased during relapse and were low or virtually absent during remission. In comparison, similar signal intensities were observed for GAPDH.

Figure 4.: Expression analysis of corresponding subtracted transcripts in MCNS. (A) Transcripts corresponding to known genes. PBMC total RNA pools from four patients in relapse and four patients in remission were applied at 2 μg/slot onto a Hybond N membrane and were probed with subtracted cDNA corresponding to L-plastin, grancalcin, T cell receptor (TCR) δ chain, NFAT5, proteasome α2 subunit, interleukin-7 (IL-7) receptor (IL-7R), IgE-dependent histamine-releasing factor (HRF), Jak1 transcripts, and glyceraldehyde phosphate dehydrogenase (GAPDH) (as probe control). (B) Transcripts corresponding to seven unknown clones. PBMC total RNA pools from seven patients in relapse and eight patients in remission were applied at 2 μg/slot onto a Hybond N membrane and were probed with cDNA corresponding to clones SC1 to SC7.

Transcripts corresponding to seven unknown clones displayed high signal intensities in PBMC from patients with nephrotic relapse (Figure 4B). For three clones (SC1, SC4, and SC5), no signal was detected in samples from patients in remission. For the SC7 clone, a significant signal was detected in remission samples but to a lesser extent than in relapse samples. The strong signals obtained in relapse samples but not remission samples suggest that the selected clones correspond to transcripts that are likely recruited during active MCNS. The complete sequencing of these clones is now in progress. The unknown genes observed among the selected clones might encode new proteins that play key roles in MCNS. Taken together, these results indicate that the forward subtracted cDNA library is highly enriched in sequences that are specifically upregulated in MCNS.

Many genes identified in this work are involved in cell division and DNA synthesis, including thiopurine methyltransferase, deoxyguanosine kinase, tyrosinase-related protein, elongation factor 1α, initiation factor 4B, and DNA-binding protein A. Such expression supports the concept of T cell activation during MCNS relapse.

mRNA coding for some proteins involved in the rearrangement of the actin cytoskeleton, i.e., L-plastin and Fyb/Slap, as well as grancalcin, seem to be recruited during MCNS relapse. To obtain further insight into these observations, we analyzed the expression levels of the corresponding transcripts and proteins by using RT-PCR and immunoblotting, respectively. The expression of L-plastin and grancalcin mRNA was simultaneously analyzed for six patients experiencing MCNS relapses and four normal subjects (Figure 5A). The expression of GAPDH was monitored in parallel, to control for variations in RT-PCR. L-plastin and grancalcin mRNA levels were increased among patients experiencing MCNS relapses, compared with control subjects. Grancalcin mRNA was detected at a lower level than L-plastin among normal subjects. We further investigated whether there were discrepancies between the mRNA and protein expressions of L-plastin and grancalcin. Cytosolic extracts of PBMC were analyzed by immunoblotting for 15 patients with MCNS relapse (Figure 5B) and 15 control subjects (Figure 5C), including subjects previously tested for mRNA expression. The blots were then reprobed with anti-actin antibody, to assess variations among the samples. The increase in L-plastin protein levels, compared with control values, was confirmed during the relapse phase. The grancalcin protein was increased during relapse, whereas its expression was hardly detectable in most normal subjects. Taken together, these results demonstrated a good correlation between the mRNA and protein levels, suggesting active transcription of these genes during relapse.

Figure 5.: Expression of L-plastin and grancalcin in MCNS relapse. (A) Expression of L-plastin and grancalcin mRNA among patients experiencing MCNS relapses (n = 6) and normal subjects (n = 4). Reverse transcription (RT) was performed with 2 μg of RNA, and 1/40th of the reaction product was used for PCR, as described in Materials and Methods. One-fifth of the PCR product was analyzed. The image shows the negative staining of the electrophoresis gel with ethidium bromide. The expression of GAPDH was monitored in parallel. (B and C) Immunodetection of grancalcin and L-plastin proteins in PBMC from patients experiencing MCNS relapses (n = 15) (B) and normal subjects (n = 15) (C). Cytosolic proteins (50 to 60 μg) were analyzed by Western blotting using an anti-L-plastin (1 μg/ml) monoclonal antibody and an anti-grancalcin polyclonal antibody (1/5000). Variations among samples were assessed by incubation with anti-actin polyclonal antibody (1/2500). The blots were hybridized with anti-L-plastin, dehybridized, and successively reprobed with anti-grancalcin and anti-actin. The sizes of molecular markers are indicated on the right. The asterisk corresponds to the Jurkat extract sample that was used as a positive control for Fyb/Slap expression.

We analyzed, using semiquantitative RT-PCR, the expression of Fyb/Slap mRNA during the relapse and remission phases among seven patients (Figure 6). Fyb/Slap mRNA levels were significantly increased during relapse, compared with remission and control values. This result cannot be explained by immunologic perturbations induced by the nephrotic state itself, because Fyb/Slap expression levels were similar among patients with MN-linked nephrotic syndrome and normal control subjects. At the protein level, the expression of Fyb/Slap was easily detected for 12 of 15 patients with MCNS relapse (Figure 7). Patients 5, 8, and 10 exhibited lower levels. However, Fyb/Slap was hardly or not detected among normal subjects, except for patient 6. The positive control samples consisted of protein extracts from the Jurkat cell line, which is known to express high levels of Fyb/Slap (¹¹).

Figure 6.: Expression of Fyb/Slap mRNA in MCNS relapse. The relative expression of Fyb/Slap mRNA among patients with MCNS (n = 7), during relapse without steroids and during remission (A), patients with MN (n = 5) (B), and control subjects (n = 5) (C) was assessed. RT was performed with 2 μg of total RNA, and 1/40th of the reaction product was used for PCR, as described in Materials and Methods. The expression of GAPDH was monitored in parallel, for assessment of variations among samples.

Figure 7.: Immunodetection of Fyb/Slap protein in PBMC from patients experiencing MCNS relapses (n = 15) and normal subjects (n = 15). Proteins (50 to 60 μg) from cytosolic extracts were analyzed by Western blotting using anti-Fyb/Slap (1 μg/ml). To assess variations among samples, the blots were dehybridized and reprobed with anti-actin. The asterisk corresponds to the extract from the Jurkat cell line that was used as a positive control for Fyb/Slap expression.

We sought to determine whether T cell activation in MCNS relapse was associated with a particular Th cell phenotype. The IL-12R complex consists of β1 and β2 subunits and mediates the IL-12 signaling pathway, which is turned off during the differentiation of naive T cells into Th2 but not Th1 populations (^12,13). Human Th1 cells, but not Th2 cells, selectively express the IL-12R β2 transcript, which is the signaling component of IL-12R. The expression of IL-12R transcripts was analyzed in PBMC samples obtained from seven pediatric patients with MCNS during relapse without steroid therapy and during remission. The levels of IL-12R β1 transcripts were similar among patients with MCNS, during both relapse and remission, patients with MN, and normal control subjects (Figure 8). In contrast, IL-12R β2 transcripts were expressed at very low levels during relapse, whereas the patients with MN exhibited higher levels of IL-12R β2 transcripts, compared with normal subjects. These results suggest that MCNS relapses without steroids are associated with downregulation of the IL-12 signaling pathway, which is usually encountered in stable Th2 cells.

Figure 8.: Downregulation of IL-12R β2 subunit mRNA expression in MCNS relapse. The relative expression of IL-12R β1 and β2 subunit mRNA among patients with MCNS (n = 7), during relapse without steroids and during remission (A), patients with MN (n = 5) (B), and control subjects (n = 5) (C) was assessed. Semiquantitative RT-PCR was performed with 2 μg of total RNA, as described in Materials and Methods. The expression of GAPDH was monitored in parallel, as a control.

Discussion

Almost 30 yr ago, it was postulated that MCNS results from abnormal T cell activation, via unknown mechanisms (¹⁴). In this work, we chose a strategy able to provide some insights into the immunologic alterations that occur in this disease. Using a subtractive cDNA library derived from T lymphocyte-enriched PBMC obtained from a patient during relapse and remission, combined with differential screening, we isolated 84 transcripts that seemed to be upregulated during the relapse phase. Of this series, 42 corresponded to genes with known function, 12 matched proteins of no known function, and 30 corresponded to new genes.

At least 18 genes identified in the screening of this library are involved in T cell signaling. The TCR is a multimeric complex composed of α and β (or δ and γ) chains, which form a variable antigen/MHC binding site, and of CD3 chains (γ, δ, ε, and ζ), which link signals from the cell surface to downstream effectors (¹⁵). Activation of T cells is a multistep process that begins with ligation of the T cell antigen receptor to its cognate peptide ligand, coupled to class I MHC antigen presented by antigen-presenting cells. Additional interactions between T cell surface coreceptors, including CD2, CD4, CD8, and CD28, and their counterparts on the antigen-presenting cells, i.e., CD48, class II MHC, class I MHC, and B7, respectively, are required for an effective immune response. Concurrent coreceptor activation and TCR stimulation induce tyrosine phosphorylation of CD3 by the proximal kinases p56^lck and ZAP 70, which link the TCR to cytoplasmic adaptor proteins such as Fyb/Slap (which exhibited increased levels during the active phase of the disease). Fyb/Slap binds to the SH2 domain of the Src kinases Fyn-T and SLP-76 (¹¹). Therefore, the recruitment of Fyb/Slap may represent an important step in the propagation of specific TCR signaling in MCNS.

The engagement of coreceptors initiates recruitment of the guanine nucleotide exchange factor Vav, which activates small GTPases of the Cdc42/Rac/Rho family (¹⁶), including the Rho A transcript that is upregulated during relapse, according to our subtractive library. Rho proteins stimulate phosphatidylinositol-4-phosphate 5-kinase, resulting in local accumulation of phosphatidylinositol-4,5-bisphosphate. Phosphatidylinositol-4,5-bisphosphate is a precursor of inositol-3,4,5-trisphosphate, which is required for Ca²⁺ mobilization and dissociation of actin-binding proteins. On resting T cells, integrins such as leukocyte-function antigen-1(LFA-1) are confined within the actin cytoskeleton by inhibitory interactions with actin-binding proteins. Local increases in inositol-3,4,5-trisphosphate levels disrupt these interactions and induce remodeling of the cytoskeleton, resulting in the release of integrins. Activated integrins undergo conformational changes and form molecular clusters that migrate to the cell surface, where they interact with the apposing membranes of antigen-presenting cells. This process may be mediated in part by Fyb/Slap and L-plastin, through interactions with other components such as actin and grancalcin (^17,18). L-plastin, once recruited after engagement of the TCR/MHC complex, mediates costimulatory signals through associated receptors such as CD2 and is functionally associated with cytokine secretion (¹⁹). Grancalcin is a member of the EF-hand Ca²⁺-binding protein family. The release of Ca²⁺ from intracellular stores induces conformational changes and translocation of grancalcin from the cytosol to the cell membrane (¹⁸). The recent demonstration of interactions with L-plastin suggests the possible involvement of grancalcin in integrin regulation (¹⁸).

T cell activation promotes the recruitment of IL-7R and Jak1; for both, mRNA expression was increased during relapse, compared with remission. The absence of IL-7R signaling results in the accumulation of functionally inactive T cells in the periphery (²⁰). The IL-7R signaling pathway depends on Jak1, as indicated by the lack of IL-7-induced proliferation in thymocytes from Jak1-deficient mice (²¹). Both IL-7R and Jak1 are recruited in multiple signaling pathways, involving different cytokines and growth factors (²²).

The effector arm of TCR signaling involves the recruitment of specific transcription factors (according to the type of immune challenge), with subsequent activation of downstream target genes and initiation of particular immune responses. In the screening described here, at least four factors were identified, including NFAT5, c-Maf, AP-2β, and NF-κB. Most of the target genes of these factors remain to be determined.

In a previous study, we demonstrated strong NF-κB activity in MCNS relapse, partly attributable to an increase in proteasome activity (⁵). We have now identified several subtracted transcripts, such as IL-1β, RANTES, p38 mitogen-activated protein kinase, Traf6, and macropain (proteasome α2 subunit), that are implicated in signal regulation, leading to the activation of NF-κB.

IL-12R β2 transcripts were expressed at very low levels among patients with MCNS during relapse, compared with normal subjects and patients with MN, whereas IL-12R β1 transcripts were detected at similar levels in all samples. In addition, we recently observed that the transcription factor c-Maf was selectively recruited during the relapse phase (le Gouvello S, Valanciuté A, Pawlak A, Solhonne B, Grimbert P, Roudot-Thoraval F, Salomon R, Lang P, Guellae[Combining Diaeresis]n G, Sahali-D; manuscript in preparation). Both findings strongly suggest that, in MCNS, T cells evolve toward a Th2 profile. Moreover, histamine-releasing factor, which is known to enhance the production of IL-13 by basophils (²³), may also contribute to the development of Th2 cells in MCNS.

Our results demonstrating the upregulation of many genes that are closely involved in T cell responses support previous studies suggesting that MCNS is a T cell-mediated disease (^14,24). Furthermore, they identify, for the first time, important signaling pathways that shape T cell responses during MCNS relapse. It is likely that the genes isolated in this library will contribute to our knowledge of the pathophysiologic processes of MCNS.

We thank L. Lyonnet for technical assistance. We are grateful to Dr. Eric J. Brown and Dr. Hua Shen for providing the anti-L-plastin monoclonal antibody (anti-LPA4-1) and to Dr. Lollike Karsten for providing the anti-grancalcin polyclonal antibody. We thank Dr. Y. Laperche and Dr. P. M. Ronco for critical review of the manuscript. We are indebted to Dr. P. Niaudet, Dr. M. Broyer, Dr. F. Bouissou, Dr. B. Boudailliez, and numerous other colleagues for providing us with blood samples and clinical information, as well as for their support and advice. This work was supported in part by a National Academy of Medicine grant (to Dr. Sahali), by the Association pour l’Utilization du Rein Artificiel (Paris, France), by Novartis SA, and by the University of Paris XII.