The Diabetes-Prone BB Rat Carries a Frameshift Mutation in Ian4, a Positional Candidate of Iddm1
Diabetes-prone (DP) BB rats spontaneously develop insulin-dependent diabetes resembling human type 1 diabetes. They also exhibit lifelong T-cell lymphopenia. Functional and genetic data support the hypothesis that the gene responsible for the lymphopenia, Lyp, is also a diabetes susceptibility gene, named Iddm1. We constructed a 550-kb P1-derived artificial chromosome contig of the region. Here, we present a corrected genetic map reducing the genetic interval to 0.2 cM and the physical interval to 150–290 kb. A total of 13 genes and six GenomeScan models are assigned to the homologous human DNA segment on HSA7q36.1, 8 of which belong to the family of immune-associated nucleotides (Ian genes). Two of these are orthologous to mouse Ian1 and -4, both excellent candidates for Iddm1. In normal rats, they are expressed in the thymus and T-cell regions of the spleen. In the thymus of lymphopenic rats, Ian1 exhibits wild-type expression patterns, whereas Ian4 expression is reduced. Mutational screening of their coding sequences revealed a frameshift mutation in Ian4 among lymphopenic rats. The mutation results in a truncated protein in which the COOH-terminal 215 amino acids—including the anchor localizing the protein to the outer mitochondrial membrane—are replaced by 19 other amino acids. We propose that Ian4 is identical to Iddm1.
Diabetes in the spontaneously diabetic and lymphopenic BB rat resembles type 1 diabetes in humans. In both diseases, the β-cells are selectively destroyed in an autoimmune process (1). Their lifelong and profound T-cell lymphopenia (2) is characterized by a reduction in CD4+ peripheral T-cells, a severe reduction in CD8+ peripheral T-cells (3), and an almost total absence of RT6+ T-cells (4). The absence of RT6+ T-cells is caused by elevated apoptosis of peripheral T-cells—especially recent thymic migrant cells—and possibly a reduced thymic output of naive T-cells (5–10). This survival defect manifests itself already during the latest stages of intrathymic development (11). In agreement with this, accelerated apoptotic death has been observed among diabetes-prone (DP)-BB mature single-positive thymocytes (12,13), and numbers of TcrhiCD4−8+ and TcrhiCD4+8+ thymocytes are reduced in the DP-BB rat (6). Taken together, we conclude that the lymphopenia gene product is involved in the regulation of apoptosis in the T-cell lineage and is already expressed during intrathymic development.
Genetically, diabetes and lymphopenia cosegregate in experimental crosses inasmuch as spontaneous diabetes is never observed without lymphopenia (14–16). Several studies have shown that the diabetogenic effect of T-cell lymphopenia is caused by the absence of RT6+ regulatory T-cells (17). RT6+ T-cell depletion of the nonlymphopenic and nondiabetic diabetes-resistant (DR) BB rat induces diabetes in 50% of the animals (18), whereas transfusion of CD4+ T-cells from DR-BB rats to young DP-BB rats reduces the diabetes incidence of recipients in a manner dependent on the number of RT6+ T-cells transferred (19).
The gene responsible for the lymphopenic phenotype (Iddm1 or Lyp) is positioned on rat chromosome 4 (15,20). We have mapped the gene to a 0.3-cM region between markers D4Got59 and Abp1 and constructed a 550-kb P1-derived artificial chromosome (PAC) contig covering the region (21). The region is homologous to a segment on human chromosome 7q35–36 (22). In the present study, we refined the genetic interval to a 0.2-cM region between markers D4Rat214 and Abp1 corresponding to a physical interval of 150–290 kb. The aim of the present study was to determine which genes are harbored within this interval and ascertain whether any of these fulfill the candidate gene profile and have mutations in lymphopenic rats that could explain the phenotype. Here, we present data to support the hypothesis that the gene Ian4 fulfills these requirements.
RESEARCH DESIGN AND METHODS
BB rats (DP-BB and DR-BB) were from the colony kept at the Hagedorn Research Institute (14). Brown Norway (BN), New England Deaconess Hospital (NEDH), and Fischer 344 (F344) rats were obtained from M&B (Lille Skensved, Denmark). Wistar-Furth (WF) rats were obtained from Iffa Credo (Lyon, France).
We used a panel of 23 animals known to have recombinations in the vicinity of Iddm1 (21). The lymphopenic state of the animals was determined from spleen samples using quantitative RT-PCR, which was performed as described below. The primers used for evaluating the Cd8b expression level were 5′-TGTCTCAAAGAGCGCCAAGAT-3′ and 5′-CCAGGAACTCAAAGTGCTTGT-3′ (derived from Cd8b; GenBank accession no. X04310). Typing of microsatellite markers and Abp1 was performed as described previously (20,21,22). Rat Ian4 was positioned using PCR. The primers 5′-CTGTGACCAGGACCAGTCAGG-3′and 5′-TGGTTCTGGATCTTTGACTCG-3′ gave rise to a 100-bp product from Ian4wt and a 99-bp product from Ian4bb, as expected. The PCRs were added [32P]-dCTP and run on standard sequencing gels (Gel-Mix 6; Life Technologies, Carlsbad, CA).
Cloning of Ian1 and -4.
PCR primers from human and mouse sequences of Ian1 and -4 used to amplify the corresponding rat genes were 5-TGGGGCGTTATACTGTGGAAG-3′ and 5′-ACTCCCGTCTTATCCGTGCTT-3′ (for Ian1) and 5′-GTCAAAGGCCCAGAACCAAG-3′ and 5′-AGAGGTCATTGCTGTGGAAGG-3′ (for Ian4). 5′- and 3′-RACE were performed using the 5′- and 3′-RACE kits from Life Technologies on thymic RNA from DR-BB rats. Total RNA was isolated from thymocytes with Trizol reagent (Life Technologies). Gene-specific primers used for Ian1 5′-RACE were 5′-CTCTCCAGTTCCTCTCTG-3′ and 5′-AGGTTTGTTTCTGAATCACAC-3′. The gene-specific primer used for Ian1 3′-RACE was 5′-GAGACGGCTCCTGAAGTCCTT-3′. Gene-specific primers used for Ian4 5′-RACE were 5′-TTCTCACCAGGGCCATGA-3′, 5′-GGGTGAAGAGGACAATCATGT-3′, and 5′-AGGTAGCAGTTCCCAATGTCC-3′. The gene-specific primer used for Ian4 3′-RACE was 5′-GGACATTGGGAACTGCTACCT-3′. PCR products were cloned in pCR2.1-TOPO or pCRII-TOPO with a TOPO TA cloning kit (Invitrogen, Gronningen, the Netherlands).
Sequencing and sequence analysis.
Overlapping fragments of 300–500 bp from Ian1 and -4 were PCR-amplified, and PCR products were cloned for sequencing of the coding sequences of the genes. For Ian1 the following primer pairs were used: 5′-CCCAAATTTTCAGGGATCTGA-3′/5′-ACCAGAAGCAGAGCATGAGGC-3′, 5′-AGATTACTCGCTGTGTTGCTC-3′/5′-TCTGCTTCATACTCCCGTCTT-3, and 5′-AAGTGCTGAGGGTGTGATTCA-3′/5′-TTGGCCATGGTGTTTGTTATG-3′. For Ian4 the following primer pairs were used: 5′-TTTTTCCTGCTTCGGACTCAG-3′/5′-GCCTGGACTCGAACGCTGGTC-3′ (specific for Ian4S), 5′-TTATAGCCAAACAGTTAGGAG-3′/5′-TTTCCGCAGCCAGATTTACCC-3′ (specific for Ian4 L), 5′-GGATCCTCCTGGTGGGTAAAT-3′/5′-GGGTGAAGAGGACAATCATGT-3′, 5′-AACTGGGACGCTACACAGTCG-3′/5′-TTCACTGTAGCCACCTCTAAG-3′, and 5′-GGCCCTGGTGAGAAGACTGGA-3′/5′-GGCAGGATTCACCCAAAGAGA-3′. At least two independently cloned PCR products from each primer set were sequenced to avoid PCR artifacts. Sequencing was performed by MWG-Biotech. Protein sequences were aligned using BLASTp (available online at www.ncbi.nlm.nih.gov/BLAST), followed by fine adjustment using visual inspection. Transmembrane regions were defined in silico using the TMHMM server (available online at www.cbs.dtu.dk/services/TMHMM-2.0).
Tissue-specific expression patterns were determined by PCR using Rat Multiple Choice cDNAs (OriGene Technologies). Quantitative differences in expression levels in thymus were determined using quantitative RT-PCR on cDNA made with SuperScript (Life Technologies) from DR- and DP-BB total thymic RNA (made with TRIZOL Reagent; Life Technologies). [32P]-dCTP was added to the PCRs, which were then run on standard sequencing gels (Gel-Mix 6; Life Technologies), and primer-specific bands were quantitated using a Typhoon 8600 Variable Mode Imager (Amersham-Pharmacia Biotech, Little Chalfont, U.K.). Primers specific for Ian1 and -4 are described above. Primers specific for the housekeeping gene G6pd (5′-GACCTGCAGAGCTCCAATCAAC-3′ and 5′-CACGACCCTCAGTACCAAAGGG-3′ from GenBank accession no. NM_017006) were used as an internal control in all reactions.
In situ hybridization.
In situ hybridization was performed on paraffin sections that were heated to 60°C for 30 min, deparaffinized in xylene, and rehydrated. The slides were acid-treated in 0.2 mol/l HCl, treated with proteinase K (5 μg/ml), and subsequently fixed in 4% paraformaldehyde, followed by dehydration before the 35S-UTP–labeled antisense or sense probes were added. The sections were then incubated overnight at 47°C in a hybridization solution containing radiolabeled RNA probe (80 pg/μl) in a solution of 50% deionized formamide, 10% dextran sulfate, tRNA (1 μg/ml), ficoll 400 (0.02% wt/vol), 0.02% polyvinylpyrrolidone (wt/vol), 0.2% BSA fraction V (wt/vol), 10 mmol/l dithiothreitol, 0.3 mol/l NaCl, 0.5 mmol/l EDTA, 10 mmol/l Tris-Cl, and 10 mmol/l sodium phosphate (pH 6.8). After hybridization, slides were washed at 57 and 62°C. The sections were then treated with RNase A (20 μg/ml), dehydrated, and air-dried. Autoradiographic film and subsequent emulsion was applied, and the sections were developed after 2–3 weeks of exposure. The 35S-UTP–labeled probe was made from the same rat-specific PCR products used for RT-PCR cloned in the pCRII-TOPO vector (see above). Antisense probe was made by in vitro transcription using the SP6 promoter, whereas sense (control) probe was made by in vitro transcription of the same plasmid from the T7 promoter. All probe preparations, including both sense and antisense probes, were adjusted to 2 × 106 cpm/μl. Probes were stored at −20°C until use.
Refined mapping of Iddm1 to a 0.2-cM region.
Our previous genetic mapping of Iddm1 was dependent on a few recombinations, each influencing the fine positioning of the gene within the region between D4Got59 and Abp1 (22). Indeed, only three rats from our experimental crosses between the DP-BB rat and three different strains of nonlymphopenic rats exhibited recombinations within this interval. Phenotyping of their lymphopenic state (i.e., homozygosity versus heterozygosity for the mutated Iddm1) had been performed on blood samples taken from live animals by flowcytometric evaluations of the frequency of TCRαβ-labeled cells among white blood cells (described in 16), whereas the genotyping of microsatellite markers (and Abp1) was based on the PCR of genomic DNA from organ samples procured at the time of killing. This procedure cannot rule out the possibility of a sample mismatch. We therefore decided to verify the lymphopenic phenotype and marker genotyping on the same frozen samples. For lymphopenia detection, we chose to perform quantitative RT-PCR for Cd8b levels in the spleen because the number of CD8-positive T-cells is severely reduced in lymphopenic rats (3). Spleen samples from the panel of backcross animals harboring recombinations in the region (including the three aforementioned animals) were thus used. All previous data were verified, except for one case where the quantitative RT-PCR unambiguously typed the splenic sample as lymphopenic, although our previous records of blood samples had typed it as nonlymphopenic (Fig. 1). This results in a refined position of Iddm1 to the 0.2-cM interval between markers D4Rat214 and Abp1, corresponding to a maximum of 290 kb (Fig. 2). This position relative to any other is supported by a log-likelihood of 5.3.
A gene map of the Iddm1-region.
As indicated in Fig. 2, four of the rat PAC ends exhibit homology to mouse and human sequences. This enabled us to align the rat Iddm1 region with a region on human chromosome 7 that has been completely sequenced (GenBank accession no. NT_007704). A total of 13 genes and six GenomeScan models have been assigned a position between RIP60 and ABP1 on human chromosome 7 according to the National Center for Biotechnology Information (NCBI) Homo sapiens map view of HSA7 (available online at www.ncbi.nlm.nih.gov/cgi-bin/Entrez/map_search) (Table 1). As can be seen from the Fig. 2, this region encompasses the entire homologous Iddm1 region in the rat.
Because only one of the genes has been characterized (LR8; a gene expressed by lung fibroblasts) (23), we searched the GenBank for similarity to known genes. Five genes (LOC136361, LOC136362, LOC136364, Hs7_7861_27_13_2, and Hs7_7861_27_18_1) exhibit no similarity to any known gene. The protein encoded by LOC136359 exhibits low similarity to human zinc finger protein 41 (24). LOC90512 is 89% identical to the protein encoded by ALS2CR2, a probable protein kinase (25). The protein encoded by LOC136363 is similar to the COOH-terminal end of rat transient receptor protein 6, a Ca transport protein encoded by Trrp6 (26). The protein encoded by LOC115262 is 99% identical to the protein encoded by HCA112 (locus link no. 55365), a gene with unknown function, but some similarity to LR8 (23). The GenomeScan model gene Hs7_7861_26_7_2 exhibits similarity to mouse zinc finger protein 29 (27). The remaining eight genes all exhibit similarity to members of a family of GTP-binding proteins, the immune-associated nucleotides (Ian genes). Three members of this gene family have been described: Ian1 is a thymic selection marker expressed at various stages of thymocyte development (28). Ian2 is induced in B-lymphocytes by infections of Plasmodium chabaudi malaria (29,30). Ian3 (28) has not been functionally characterized. Ian4 encodes a GTP-binding protein localized in the outer mitochondrial membrane, and it is expressed in both spleen and a mouse T-cell line (31). The gene is overexpressed in hematopoietic precursor 32D cells transfected with a Bcr/Abl oncogene, which is involved in several leukemias.
Another gene with high similarity to the Ian genes is the Arabidopsis AIG1 gene, which is induced during a bacterial resistance response leading to apoptosis (37). The orthologues of both mouse Ian1 and Ian4 seem to be included in our interval: FLJ11110 encodes a 329–amino acid protein that exhibits 61% identity to the protein encoded by mouse Ian1 (a 328–amino acid protein), whereas FLJ11296 encodes a 307–amino acid protein that exhibits 61% identity to the protein encoded by mouse Ian4 (a 295–amino acid protein). The remaining seven genes exhibit a lower similarity to either of the two mouse genes. The Ian genes seem to be predominantly expressed in lymphoid tissues, and a role in apoptosis can be inferred by the function of AIG1. Therefore, we believed these genes to be good candidates for Iddm1. Because the functions of Ian1 and -4 make them obvious candidates, we decided to characterize them.
Cloning of rat Ian1 and -4.
We designed primers specific for the rat orthologue of Ian1 and Ian4 by placing primers in a region of homology between human and mouse genes. The resulting PCR products were sequenced, and the sequence was used to isolate the entire coding region using the rapid amplification of cDNA end (RACE) procedure. For rat Ian1, a 1,423-bp transcript was discovered (GenBank accession no. AY070268). An open reading frame of 310 bp was observed. The alignment of rat, mouse, and human sequences is shown in Fig. 3A. For rat Ian4, two different 5′ ends were discovered, resulting in transcripts of 1,245 bp and 1,310 bp, respectively. The two transcripts contain two open reading frames of, respectively, 334 and 308 codons (Ian4 L and Ian4S; GenBank accession nos. AY055776 and AY055777), the last 293 codons being identical. Both are highly homologous to both mouse Ian4 and to human IAN4 proteins. The alignment of rat, mouse, and human sequences shown in Fig. 3B suggests that IAN4 is indeed the human orthologue of murine Ian4, albeit Dahéron et al. (31) named the human gene IAN5. Although the sequences differ in their COOH-terminal ends, they all contain a putative transmembrane tail.
Expression pattern of Ian1 and -4.
Using a panel of tissue-specific cDNAs, we determined the expression patterns of Ian1 and -4. Ian1 is primarily expressed in spleen, thymus, heart, lung, and intestine, and it is expressed, to a lesser degree, in liver, kidney, stomach, and muscle. Ian4 (both isoforms) is primarily expressed in spleen, thymus, heart, lung, and intestine, and it is expressed, to a lesser degree, in kidney, stomach, and muscle (Fig. 4). We then performed in situ hybridization of the two genes on thymus and spleen. In the DR-BB rat, Ian1 mRNA is expressed in the periarteriolar lymphatic sheets of the spleen (Fig. 5Aa, Ae, and Af) and in thymic medulla (Fig. 5Ba, Be, and Bf). In the DP-BB rat, Ian1 seems to be expressed at reduced levels in both spleen and thymus (Fig. 5Ac, Ag, Ah, Bc, Bg, and Bh). For Ian4, the DR-BB expression pattern is as follows: in the spleen, Ian4 is expressed primarily in the periarteriolar lymphatic sheets (Fig. 6Aa, Ae, and Af), whereas in the thymus, Ian4 mRNA is evenly distributed all over the thymus (Fig. 6Ba, Be, and Bf). In the DP-BB rat, very low amounts of Ian4 are expressed in both spleen and thymus (Fig. 6Ac, Ag, Ah, Bc, Bg, and Bh). Because expression level differences in the thymus might be caused by a promoter mutation in the gene (that might be the primary gene defect leading to lymphopenia), differences in thymic expression levels were also evaluated by quantitative RT-PCR on thymus cDNA of DP- and DR-BB origin. No significant difference between DP- and DR-BB rat levels was found for Ian1, whereas for Ian4 we detected a 3.4-fold higher thymic expression level in the DR-BB rat relative to the DP-BB rat (Fig. 7).
A frameshift mutation in the BB allele of the Ian4 gene leads to a truncated protein.
The coding region of Ian1 and -4 were screened for mutations in the DP-BB rat relative to the nonlymphopenic DR-BB rat. For Ian1, two allele differences were detected: one silent and one conservative (Ala55Val) mutation. For Ian4, five allele differences were detected: two silent mutations, two conservative missense mutations (Ala156Val and Met203Ile), and one frameshift mutation (450delC). The frameshift mutation results in a mutated DP-BB allele (Ian4bb) encoding a truncated protein, where the 215 last amino acids of wild-type Ian4 (Ian4wt) are replaced by 19 other amino acids (Fig. 3B). The COOH-terminal transmembrane region is missing altogether, and since the 20-most COOH-terminal amino acids of mouse Ian4 are necessary for localization of the gene product in the outer mitochondrial membrane (31), the Ian4bb gene product is probably not correctly localized. The Ian4wt sequence was confirmed in four other nonlymphopenic rat strains (BN, WF, NEDH, and F344), whereas the Ian4bb sequence was confirmed in two other lymphopenic rat strains, namely a WF rat congenic for the lymphopenia locus of DP-BB origin and a F344 rat congenic for the lymphopenia locus and the major histocompatibility complex of DP-BB origin (32,33). The single-nucleotide deletion in Ian4bb gives rise to a polymorphism that we used for genetic mapping of Ian4. We positioned Ian4 to the same position as Iddm1, between D4Rat214 and Abp1 (data not shown).
We have refined the genetic positioning of rat Iddm1 to a 0.2-cM (or 150- to 290-bp) region on rat chromosome 4 and produced a transcript map of this region by comparative mapping to humans. Two candidate genes were chosen, and the rat homologues were cloned. Both candidate genes were subjected to expression profiling, which showed both of them to be expressed in the thymus, where Iddm1 is expected to be expressed. Dahéron et al. (31) detected low levels of mouse Ian4 in the spleen, but not in the thymus. This discrepancy between our data and theirs might be due to differences between rat and mouse expression patterns and perhaps the differently applied techniques, since Dahéron et al. used a Northern blotting technique. However, we confirmed our observations using in situ hybridizations. The confinement to the periarteriolar lymphatic sheets in the spleen confirms a role of both genes in the T-cell lineage, as suggested already by the expression in the thymus. The in situ hybridizations showed Ian4 to be expressed all over the thymus, suggesting that the expression is important at all time points during thymic development. In contrast, Ian1 is only expressed in the thymic medulla, where the most mature thymocytes reside (34). The in situ hybridizations also suggested a reduced expression level of the two genes in both thymus and spleen. Because of the T-cell lymphopenia in the DP-BB rat, a reduced expression in spleen is expected for all T-cell–specific transcripts in the spleen, whereas a reduced thymic expression should only be expected from a subset of such genes, including Iddm1 itself.
Quantitative RT-PCR confirmed the lower expression of Ian4—but not Ian1—among DP-BB thymocytes as compared with DR-BB thymocytes. This could be caused by a secondary effect of the lymphopenia or a mutation in Ian4 itself, either as a result of a promoter mutation or as a result of nonsense-mediated mRNA decay, which is seen as a result of both nonsense and frameshift mutations (35). We screened the coding sequences of the genes by RT-PCR and found a frameshift mutation in Ian4. This mutation results in a truncated gene product that lacks the 205-most COOH-terminal codons (or two-thirds of the total protein), including the COOH-terminal transmembrane region sequence that localizes Ian4 to the outer mitochondrial membrane in the mouse. Apart from this mutation, only silent or conservative mutations were discovered in either gene.
Considering the impact of the frameshift mutation on the protein, it is unlikely that the Ian4bb gene product is able to maintain the function of the wild-type gene product. We have thus detected a deleterious mutation in a gene present in the narrow Iddm1 region. Several observations support the hypothesis that Ian4 is involved in the regulation of T-cell apoptosis, making it an excellent candidate for Iddm1. First, high expression of Ian4 is associated with leukemia in the mouse (31), whereas absent/low expression appears to be associated with lymphopenia. Second, the gene product is anchored to the outer mitochondrial membrane, and this organelle is an important site for the regulation of cell death (36). Third, the highly homologous Ian1 gene is differentially expressed during thymocyte development, a process known to involve apoptotic events, and the related plant gene, AIG1, is involved in hypersensitivity response against bacterial pathogens, a process leading to apoptosis (37), both suggesting the involvement of Ian family members in the control of apoptosis. Finally, Ian4 is a GTP-binding protein, and small GTP-binding proteins such as Drosophilae Ras and rice Rac have been shown to be involved in apoptosis (36). In addition, our data show that Ian4 is expressed both in thymus and spleen, and that it is confined to the T-cell–rich areas of the latter. Altogether, this suggests that Ian4 protects against apoptosis among T-cells. We believe the circumstantial evidence presented here establishes Ian4 as being identical to rat Iddm1.
Whether the human homologue of rat Iddm1 plays a role in human type 1 diabetes is unknown, but we have performed an association study using a closely positioned microsatellite marker on a Danish patient population (38). This study did not show any association to human type 1 diabetes. Although the marker was positioned only 14 kb from human IAN4, we cannot, however, rule out an association with the gene itself. Moreover, other genes involved in the generation of regulatory T-cells may be type 1 diabetes susceptibility genes in humans, as is indeed CTLA4 (39). Under all circumstances, the identification of rat Iddm1 is of importance for the elucidation of the precise mechanisms for T-cell survival and the generation of regulatory T-cells.
This work was supported by Research Grant 1-1999-744 (to H.M.), Postdoctoral Fellowship Award 398330 (to L.H.) from the Juvenile Diabetes Foundation International, and a Focused Research Grant from EFSD/JDRF/Novo Nordisk A/S (to H.M.).
We thank Stine Bisgaard and Vibeke Ladas for technical assistance.
Address correspondence and reprint requests to Helle Markholst, MD, Hagedorn Research Institute, Niels Steensens Vej 6, DK-2820 Gentofte, Denmark. E-mail:.
Received for publication 21 December 2001 and accepted in revised form 21 February 2001.
The Hagedorn Research Institute is an independent basic research component of Novo Nordisk. H.M. and L.H. are employees of Novo Nordisk.
DP, diabetes-prone; DR, diabetes-resistant; PAC, P1-derived artificial chromosome.