MEF were acquired from PKD2 TG embryos or Pkd2 KO embryos during development (13.5 days) [11]. Heads and limes were removed from embryos. The remaining embryonic tissues were minced and dispersed in 0.05% trypsin prior to incubation at 37℃ for 15 min. Cells were plated in Dulbecco's Modified Eagle's Medium of Defined Minimal Essential medium (Welgene Biotech, Taipei, Taiwan) supplemented with 10% fetal bovine serum (Welgene Biotech) and were cultured at 37℃ in an atmosphere of 5% CO₂ until confluent growth was achieved. MEF cells were frozen as stocks at the second passage and were used for the subsequent studies at the third passage.

Total RNA was prepared using a commercial kit (Qiagen, Valancia, CA, USA) according to the manufacturer's instructions. Gene expression profiles were obtained using the CodeLink Uniset Mouse Expression Bioarray (Amersham, Buckinghamshire, UK). The 30,000 gene probes contained in the bioarray allows the detection of differences in gene expression as low as 1.3-fold with 95% confidence. Ten micrograms of total RNA was amplified, and labeled cDNA was produced. Biotin-labeled cDNA was hybridized to the array overnight in a shaking incubator at 37℃, and excess target sequences were washed away using a series of saline sodium citrate washes. The array was stained by treatment with streptavidin-Alexaflour 647 (Molecular Probes, Eugene, OR, USA), the excess was washed away, and the array was scanned at an excitation wavelength of 632 nm using a GenePix scanner (Molecular Devices, Sunnyvale, CA, USA). The resulting image was quantified, and the intensity of each spot was divided by the median spot intensity to provide a scaled and comparable number across multiple arrays.

Microarrays were scanned using an Arraywox scanner (Applied Precision, Seattle, WA, USA), analyzed using ImaGene version 5.1 software (Biodiscovery, Segundo, CA, USA), and normalized using Genesight version 3.2 software (Biodiscovery). Normalization was performed by subtracting the means of all genes. The data normalized by Genesight were compared using an M/A plot. Genes differentially expressed were identified by intensity differences, after subtracting the background intensity. Genes showing expression changes of at least 2-fold were selected, and these selected genes were clustered by the hierarchical method.

The normalized log ratio corresponding to each time point was exported to the software for clustering algorithm. Acuity version 3.1 (Molecular Devices) was used for the 'gene shaving' algorithm. To estimate the number of clusters in a dataset, the 'Gap Statistics' attribute of the Acuity software package was used. The cluster number, estimated by gap statistics, was used for an input parameter in gene shaving. After gene shaving, a single linkage was used for hierarchical clustering.

Total RNA was prepared using a commercial kit (Qiagen) according to the manufacturer's instructions. Single-strand cDNA was synthesized by incubating 5 µg total RNA with 200 units AMV, 100 nM oligo(dT)12-18, 1 mM dNTP mixture, and 40 units RNase inhibitor at 42℃ for 1 h in a final volume of 25 µL. The reaction was terminated by incubation at 70℃ for 15 min. The initial amount of mRNA and reaction conditions were optimized to obtain linearity for mouse 18s rRNA. For semi-quantitative reverse transcription (RT)-PCR, the used primers were: human PKD2, forward 5'-CGTGCCCCAGCCCAGTC-3' and reverse 5'-TTCCAGTACAGCCCATCCAATAAG-3'; mouse Pkd2, forward 5'-TGCGAGGGCTGCGAGGTC-3' and reverse 5'-TGTCAGCTTGCGTGTGGTTGC-3'; mouse 18s rRNA, forward 5'-GTAACCCGTTGCACCCCATT-3' and reverse 5'-CCATCCAATCGGTAGTAGCG-3'. RT-PCR cycling conditions were as follows: 10 min at 95℃, 25 cycles of 50 s at 94℃, 50 s at 57℃, 50 s at 72℃, and 5 min at 72℃. The amplified products were separated on a 1% agarose gel. For real-time PCR, the used primers were: mouse β-actin, (forward 5'-GACGATGCTCCCCGGGCTGTATTC-3' and reverse 5'-TCTCTTGCTCTGGGCCTCGTCACC-3') used as a positive control; mouse peroxisome proliferator-activated receptor γ (PPAR-γ), forward 5'-TCTTAACTGCCGGATCCACAAAAA-3' and reverse 5'-ATCTCCGCCAACAGCTTCTCCTTC-3'; mouse protein kinase α (PKCα), forward 5'-GGGCAGCCTCCGTTTGATGGT-3' and reverse 5'-CGCTTGGCAGGGTGTTTGGTC-3'; mouse Wnt4, forward 5'-GCCATCGAGGAGTGCCAATACC-3' and reverse 5'-GGCCACACCTGCTGAAGAGATG-3'; mouse integrin α4 (ITGα4), forward 5'-GTAGCCCCAGTGGAGAGCCTTGTG-3' and reverse 5'-ATGCCAGTGGGGAGTTTGTTATCG-3'; mouse IL6ST, forward 5'-TGAATCGGACCCACTTGAGAGG-3' and reverse 5'-CAGGAGCGGCTTGTTTGAGGTA-3'; mouse aquaporin 1 (AQP1), forward 5'-GGAGGCGCCGAGACTTAGGT-3' and reverse 5'-GCGGGTGAGCACAGCAGAGC-3'; mouse transforming growth factor-β2 (TGF-β2), forward 5'-TCATCCCGAATAAAAGCGAAGAGC-3' and reverse 5'-AGGGCAACAACATTAGCAGGAGAT-3'. Real-time RT-PCR was performed using the real-time SensiMixPlus SYBR kit as described by the manufacturer's instructions (Quantance, London, UK).

Proteins were extracted using radio-immunoprecipitation assay buffer in MEF cells. Proteins were separated by 12% sodium dodecyl sulfate-polyacryamide gel electrophoresis and were transferred to a polyvinylidene fluoride membrane (Millipore, Bellerica, MA, USA). Membranes were blocked with 5% non-fat dry milk and were incubated with various antibodies. Antibodies to PC2 (anti-human and anti-mouse) and AQP1 were obtained from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Antibodies to β-actin were obtained from Sigma-Aldrich (St. Louis, MO, USA). Anti-Wnt4 was obtained from R&D Systems (Minneapolis, MN, USA). Each membrane was washed with phosphate-buffered saline-Tween, and immunocomplexes were detected by enhanced chemiluminescence (Amersham).

Mouse Pkd2 mRNA was revealed in KO wild-type (KOWT) MEF cells but not in KO MEF cells. Also, Pkd2 mRNA was expressed in TG wild-type (TGWT) MEF cells and PKD2 TG MEF cells. Human PKD2 mRNA was only expressed in PKD2 TG MEF cells obtained from PKD2 TG embryos following injection of the human PKD2 transgene (Fig. 1A). Consistently, the expression level of polycystin-2 was lower in KO MEFs than KOWT MEFs as well as higher in TG MEFs than TGWT MEFs (Fig. 1B).

We used a mouse 30 K whole gene oligonucleotide microarray to identify mRNAs whose expression was altered by the Pkd2 KO and PKD2 TG MEF cells. The normalized log ratios corresponding to each time point were exported to the software for clustering algorithm. We used Axon's Acuity 3.1 (Axon Instruments, Union City, CA, USA) for using the 'gene shaving' algorithm. To estimate the number of clusters in a dataset, we used 'Gap Statistics' in Acuity 3.1. The cluster number estimated by gap statistics was used for an input parameter in gene shaving. After gene shaving, we used a single linkage hierarchical clustering. The single linkage hierarchical clustering method divides 8 'gene shaving' clusters (Fig. 2). The majority of genes whose expression was appreciably altered encoded proteins involved in critical biological processes, such as metabolism, transcription, cell adhesion, cell cycle, and signal transcription. Forty-five genes whose expression was changed by 2-fold or greater were identified (Table 1).

To verify the results obtained by cDNA microarray analysis, expression patterns for selected genes were confirmed at the RNA and protein levels. We selected seven genes among the 45 genes (Fig. 3A). As result of realtime RT-PCR, the expression patterns of PKCα, ITGα4, and AQP1 were identical to microarray analysis in Pkd2 KO MEF cells (Fig. 3B). In contrast, 6 genes, except for TGF-β2, were identically expressed in PKD2 TG MEF cells (Fig. 3C). Especially, the AQP1 gene may be significantly regulated by differential Pkd2 gene expression level. Furthermore, to confirm the AQP1 protein expression level, western blot analysis was performed using MEF cells as well as kidney tissues obtained from Pkd2 KO (or heterozygote) and PKD2 TG mice. As result, AQP1 protein level was consistent with mRNA levels in both MEFs and tissues (Fig. 4).