Analysis of the IDDM Candidate Gene Prss16 in NOD and NON Mice

The thymus-specific serine protease Prss16 is highly expressed by the epithelial cells in the thymic cortex. It has been suggested to play an important role in the positive selection of T cells through the antigen presention pathway of the cortical antigen presenting cells. Recently, the gene ecoding Prss16 has been linked to insulin dependent diabetes mellitus (IDDM) susceptibility independent of HLA-DR3 suggesting the Prss16 may be involved in the development of autoimmune disease. Due to the similarities of the gene structure and expression pattern between the human and mouse genes, we compared Prss16 between non-obese diabetic (NOD) and non-obese non-diabetic (NON) mice. Analysis of the Prss16 coding region failed to identify any differences in sequence. Northern analysis and semi-quantitative reverse transcriptase polymerase chain reaction showed that the mRNA was equal in size and abundance in the two strains. In situ hybridization showed similar patterns of staining. Therefore, our data suggests that there is no significant different in the gene structure, transcription level, and expression pattern of Prss16 gene between NOD and NON mice.

Type 1 or insulin-dependent diabetes mellitus (IDDM) is a polygenic autoimmune disease caused by T cell-mediated destruction of insulin-producing pancreatic b cells (Birk and Cohen, 1993;Vyse and Todd, 1996). IDDM susceptibility is most tightly associated with genes linked to the HLA region of the Major Histocompatability Complex (MHC) on human chromosome 6p21. The strongest associations reside within the HLA-DRB1, HLA-DQA1, and HLA-DQB1 (class II) regions (Todd et al., 1987;Nepom, 1995). However, variants in this region do not fully explain the association between the MHC and IDDM (Robinson et al., 1993;Lie et al., 1999a). Using transmission disequilibrium testing, Lie et al. have shown evidence for an IDDM susceptibility locus independent of MHC class II located telomeric of the MHC region near the D6S2223. The strongest candidate gene for this susceptibility locus is Prss16, also referred to as thymus specific serine protease (Bowlus et al., 1999).
Prss16 encodes a protein with homology to prolylcarboxypeptidase (PCP) and is specifically expressed by cortical epithelial cells in the thymus. The pattern of expression around the thymic capsule and vessels is similar to that reported for cathepsin L, a protease involved in antigen presentation during the positive selection of T cells (Nakagawa et al., 1998). Similarly, Prss16 is believed to function in the positive selection of T cells. Defects in positive selection have been implicated in the development of autoimmunity and IDDM, in particular (Luhder et al., 1998).
The purpose of this study was to evaluate Prss16 as a diabetes susceptibility locus in the non-obese diabetic (NOD) mouse by comparison to the non-obese non diabetic (NON) mouse. Mutations that might effect the coding region, mRNA size and expression were screened for by direct sequencing, Northern hybridization, semiquantitative real-time PCR, and in situ hybridization.

DNA Sequencing
NON and NOD mice genomic DNA was purchased from the Jackson Laboratory (Maine, USA). Seven pairs of primers were designed using Primer Express software (Applied Biosystems, Foster City, CA) for amplification and sequencing of all 12 exons in both strains of mice (Table I). PCR reactions were performed under the following conditions: 4 ng/ml template DNA, 1 pmol/ml each primer, 1 mM dNTPs, 1.5 mM MgCl 2 , 1X PCR Buffer, 0.02 U/ml Taq polymerase in a final volume of 50 ml. Cycling conditions were: 948C for 3 min, 35 cycles X (948C for 45 s, 568C for 45 s, and 728C for 1 min), 728C for 4 min. All PCR products were analyzed on a 1% agarose gel and sequenced with the same primers on an ABI 377 Sequencer using Big Dye Terminator Cycle Sequencing kit (Applied Biosystems, Foster City, CA). Sequences were aligned using Sequencher software (Genecodes, Ann Arbor, MI).

Total RNA Isolation, Probe Labeling, and Northern Blot Analysis
Six-week-old male C57BL/6, NOD/LtJ and NON/LtJ mice were purchased from the Jackson Laboratory (Maine, USA). Total RNA was prepared from the mouse thymi using TriZol w Reagent (Life Technologies, Rockville, MD. USA) according to the manufacturer's instructions. Total RNA (10 mg) was separated on 0.8% formaldehyde agarose gel and transferred onto Hybond-Nþ nylon transfer membrane (Amersham Pharmacia Biotech, Piscataway, NJ, USA). A mouse cDNA clone (Genbank Accession AI587874) was radiolabeled and hybridized as previously described. After two washes each in 2 £ SSC, 0.1% SDS for 20 min at room temperature and 0.2 £ SSC, 0.1% SDS for 20 min at 658C, the membrane was exposed to the phosphor screen for 72 h at room temperature, then analyzed using the PhosphorImager (Molecular Dynamics, Sunnyvale, CA, USA).

Thymic Stromal Cell Population Enrichment and Real Time RT-PCR
Stromal cells were isolated from the thymi of 3 weeksold male NOD ðn ¼ 4Þ and NON ðn ¼ 4Þ mice as previously described (French et al., 1997). Briefly, fresh thymus was cut into small pieces, suspended in RPMI 1640 (Life Technologies, Grand Island, NY, USA), and homogenized in a loosely fitting glass homogenizer. The cell suspension was filtered through a 100-mm pore cell strainer (FALCON) and left to sediment at 48C for 10 min. The supernatant was separated from the sediment. The material retained by the strainer and the sediment were resuspended in cold (48C) PBS and left to sediment at 48C for 10 min, and the supernatant was separated from the sediment and the procedure was repeated three times. The final sediment was briefly pelleted at 800g and RNA was extracted using TriZol w reagent as described above.
Semi-quantitative real time PCR was performed on an ABI 7700 Sequence Detector using SYBR w Green Assays (PE Biosystems, Foster City, CA, USA.). Briefly, 500 mg of stromal RNA was reverse transcribed in a 20 ml reaction using SuperScript Reverse Transcriptase (Life Technologies) with 2.5 mM random hexamer primer (Life Technologies) according to the manufacturer's instructions. 5 ml of 1:50 dilution of cDNA reaction was used for template with primers TQ 295F (CAGCCACTGGATCC-CTTCA) and TQ 262R (TGATCATTCACCCAGTA-CCGC) in a 50 ml reaction according to the manufacturer's recommendations. Cycling consisted of: 1 £ 508C for 2 min, 1 £ 958C for 10 min, 40 £ (958C for 15 s, 608C for 1 min). Similarly, the 18S ribosomal RNA was amplified with primers 18S.F (CTACCACATCCAAGGAAGGCA) and 18S.R (CAGACTTGCCCTCCAATGGA). All samples were performed in triplicate. PCR products were separated on a 2% agarose gel to confirm specific amplification. Quantitation was performed using Sequence Detector V1.7. A standard cDNA derived from a C57BL6 mouse thymus was used as a calibrator in each assay.

In Situ Hybridization
Mouse thymi from the NOD and NON mice were sectioned at 8 mm and fixed in 4% paraformaldehyde. Sense and antisense cRNA probes were generated from the same cDNA clone used for Northern hybridization with incorporation of digoxigenin-UTP (Roche Molecular Biochemicals, Indianopolis, IN, USA). Probes were hybridized overnight at 588C to thymic sections and washed. Following incubation with alkaline phosphataseconjugated anti-digoxigenin antibody, slides were incubated in BCIP/NBT at room temperature for 36 h. The section were counterstained with Nuclear Fast Red (Vector Laboratories, Burlingame, CA, USA) and analyzed by light microscopy.

Genomic Structure
To determine the genomic structure of the mouse Prss16 gene in both NOD and NON mice, genomic DNA was sequenced with seven different pairs of primers. No differences in sequence were identified between NOD and NON.

Prss16 Expression in NOD and NON Mice
Since NOD mice have abnormal thymic structure and develop the autoimmune phenotype, we investigated the thymic expression of Prss16 in NOD and NON mice. Northern blots of NOD and NON thymus RNA were used to screen for Prss16 mRNA level and size (Fig. 1a). No difference in size or band intensity could be detected. We also performed semi-quantitative PCR on thymic stromal cells to further investigate differences in Prss16 mRNA (Fig. 1b). When normalized to 18S rRNA, no difference in the relative amount of Prss16 in the thymic stromal cells from NOD and NON mice could be detected ð p . 0:70Þ: We have previously shown by in situ hybridization that Prss16 is much more prevalent in C57BL6 and BALB/C than NOD thymus (Cheunsuk et al., 2002). However, comparison of NOD to NON thymus failed to reveal any notable difference in the pattern or density of Prss16 staining (Fig. 2).

DISCUSSION
Identification of genetic variants that confer susceptibility to or protection from IDDM is important to identify high-risk individuals and potential targets for preventive therapies. Several susceptibility loci have been identified including the MHC class II, the INS VNTR and most recently CTLA4 (Eaves et al., 1999;Ueda et al., 2003). Similar influences of these loci have also been reported in NOD mice supporting its role as a model of IDDM.
In total, more than 20 non-MHC linked genes have been suggested to influence diabetes susceptibility with the MHC contributing the greatest influence (Griffiths et al., 1999;Grattan et al., 2002). Most of this effect has been attributed to the unusual sequence of the HLA-DQ allele in humans and H2 g7 allele in NOD mice which both encode non-aspartic acid residues at position 57 of the b chain. However, other FIGURE 1 (a) Northern hybridization of Prss16 probe to total thymus RNA from C57BL/6, NOD ðn ¼ 4Þ and NON ðn ¼ 4Þ mice. (b) Prss16 mRNA relative to 18S ribosomal RNA in thymic stromal cells from the thymi of three weeks old male NOD ðn ¼ 4Þ and NON ðn ¼ 4Þ mice. Shown is the average and SEM for Prss16 mRNA relative to 18S rRNA.

FIGURE 2
In situ hybridization of Prss16 to thymi from NOD and NON mice. genetic influences linked to the MHC have been suggested by several studies. Dissecting out the precise locus has been difficult due to the strong linkage disequilibrium in the region. However, using transmission disequilibrium tests, a clear association has been established in the region telomeric of the MHC class I region on chromosome . Notably, the rate of transmission of the D6S2223 allele 3 was 0.2 in probands compared to the expected 0.5. In addition, this same allele is associated with susceptibility to celiac disease (Lie et al., 1999b).
Several polymorphisms in Prss16 have recently been identified including a missense mutation and a 15 base pair deletion (7548_7563del) predicted to result in a loss of 5 amino acids (Lie et al., 2002). Interestingly, the 7548_7563del has an allele frequency of 0.17 and is in linkage disequilibrium with the DQA1*05 -DQB1*02 haplotype. The potential association between this and other polymorphisms with IDDM is currently under investigation.
In mouse, the orthologue of Prss16 is highly conserved with expression similarly restricted to thymic cortical epithelial cells (Cheunsuk et al., 2002). In addition, Prss16 is expressed at lower levels in autoimmune prone mice such as NOD and New Zealand Black (NZB) mice compared to C57BL6 and BALB/C mice. However, Prss16 is not syntenic to the MHC on mouse chromosome 17. Rather, it is encoded in the comparative region centromeric of the satin (sa) locus on mouse chromosome 13. Idd14 is a diabetes susceptibility locus that has been mapped to chromosome 13 linked to D13Mit61 (McAleer et al., 1995). This locus was identified by an outcross of NOD with congenic NON.H2 g7 followed by an F1 intercross. In the F2 generation, susceptibility was associated with increased NOD homozygosity and decreased NON homozygosity at D13Mit61.
For these reasons, we investigated Prss16 as a possible susceptibility gene for the development of diabetes in NOD mice. Analysis of Prss16 gene of NOD and NON mice demonstrated no mutation in the coding sequence of the NOD mice compare to that of the NON mice. Previously, we have shown by in situ hybridization that Prss16 expression in NOD mice is lower compared to C57BL6 and Balb/c mice. In addition, lower expression is also noted in NZB mice which are prone to autoimmune disease as well. However, we found no difference by in situ hybridization between NOD and NON mice. In all strains gene expression is specific to cortical epithelial cells. Furthermore, Northern analysis and the semiquantitative PCR did not demonstrate any significant different in transcript size or abundance between the strains.
In conclusion, we have evaluated Prss16 as a candidate gene for diabetes susceptibility in NOD mice. We were unable to detect any differences in the coding sequence, transcript size or transcript abundance. Although we cannot conclusively rule out Prss16 as the Idd14 diabetes susceptibility gene in the NOD mouse, our results make it highly unlikely. Interestingly, more closely linked with D13Mit61 is cathepsin L which when deleted in mice has been shown to result in abnormal selection of CD4þ T cells and development of natural killer T cells (Honey et al., 2002). Further investigation of cathepsin L as the Idd14 susceptibility gene is warranted. It is important to note that our results do not exclude Prss16 as a susceptibility gene in human IDDM.