Amino acid polymorphisms altering the glycosylation of IL-2 do not protect from type 1 diabetes in the NOD mouse Masahito Kamanakaa, Dan Rainbowb, Karin Schuster-Gosslerc, Elizabeth E. Eynona,d, Alexander V. Chervonskyc,e, Linda S. Wickerb, and Richard A. Flavella,d,1 aDepartment
of Immunobiology and dHoward Hughes Medical Institute, Yale University School of Medicine, New Haven, CT 06520; bJuvenile Diabetes Research Foundation/Wellcome Trust Diabetes and Inflammation, Laboratory, Department of Medical Genetics, Cambridge Institute for Medical Research, University of Cambridge, Cambridge CB2 0XY, United Kingdom; cThe Jackson Laboratory, Bar Harbor, ME 04609; and eDepartment of Pathology, University of Chicago, Chicago, IL 60637 Contributed by Richard A. Flavell, April 30, 2009 (sent for review February 20, 2009)
Idd3 is one of many gene regions that affect the development of type 1 diabetes (T1D) in the nonobese diabetic (NOD) mouse. Idd3 has been localized to a 650-kb region on chromosome 3 containing the IL-2 gene. Exon 1 of the IL-2 gene is polymorphic between the susceptible NOD and the protective C57BL/6 (B6) alleles, causing multiple amino acid changes that have been proposed to be responsible for the differing glycosylation status. To address whether this coding polymorphism recapitulates the disease suppression mediated by the B6 Idd3 allele, we generated knockin mice in which exon 1 of the B6 IL-2 allele replaces the homologous region in the NOD allele. We generated these mice by targeting the NOD allele of NOD/129 F1 ES cells. IL-2 protein from the knockin mice showed the glycosylation pattern of the B6 IL-2 isoform, confirming that the amino acid differences encoded within exon 1 affect the glycosylation of the IL-2 protein. However, unlike NOD.B6 Idd3 congenic mice, the knockin mice were not protected from T1D. Furthermore, the difference in amino acid sequence in the IL-2 protein did not affect the level of expression of IL-2. This approach provides a general method for the determination of a functional role of a given genomic sequence in a disease process. Further, our result demonstrates that the variants in exon 1 of the IL-2 gene are not responsible for T1D suppression in NOD.B6 Idd3 mice, thereby supporting the hypothesis that variants in the regulatory region affecting expression levels are causative. Idd3 兩 NOD ES cells 兩 flanking gene problem
T
1D is a multifactorial disease to which multiple genetic and environmental factors contribute. The genetic control of T1D has been well characterized by linkage analyses of T1D-prone NOD mice and nondiabetic strains (1). Regions influencing disease are termed Idd for insulin-dependent diabetes and have been extensively studied for the responsible genes (2). Among them, Idd3 on mouse chromosome 3 has a significant effect on T1D development and lymphocyte infiltration into the pancreatic islets (3, 4). NOD.B6 Idd3 congenic mice that have the Idd3 region from the B6 strain are protected from T1D and insulitis. By the study of NOD.B6 Idd3 congenic strains, the Idd3 region has been mapped to a 650-kb region, which includes the genes encoding IL-2, IL-21, and FGF-2 (5) (Fig. 1). Among them, IL-2 and IL-21 have notable roles in the immune response. IL-2 promotes the proliferation of T cells and is required for the development and maintenance of naturally occurring regulatory cells (nTreg), which negatively control immune responses. Importantly, the NOD and B6 IL-2 alleles have multiple variants in exon 1 that alter the amino acid sequence (3). In the case of NOD, position 6 and 10 of the mature protein of IL-2 are prolines, whereas, in B6, they are serines. In addition, amino acid 12–15 of the mature IL-2 protein in NOD is deleted in B6; the 8 glutamine repeat of amino acid 19–26 in NOD is 4 residues longer in B6. Electrophoresis of IL-2 molecules revealed differential mobility, thus indicating different glycosylation patterns 11236 –11240 兩 PNAS 兩 July 7, 2009 兩 vol. 106 兩 no. 27
for the NOD and B6 allotypes (6). IL-2 secreted from NOD cells exhibited one major homogenous band of a larger apparent molecular weight than that of the major B6 IL-2 product; and for B6 IL-2, additional lower molecular weight bands were also present. In a comparison of several mouse strains, proline at amino acid position 6 correlated with a NOD-like electrophoresis pattern, whereas serine at position 6 associated with a B6-like pattern (6). Considering the importance of IL-2 in immunity, a difference in the structure and/or glycosylation of IL-2 may have a significant effect on T1D development. Indeed, IL-2 can bind to heparan sulfate through its glycosylated residues and deposit itself on tissue matrix via this interaction (7, 8). Thus, it was proposed that differences in glycosylation might affect the binding ability of IL-2 to the tissue matrix, which may subsequently lead to abnormal homeostasis of T cells observed in NOD mice (9). These facts suggest that the structural difference in IL-2 might affect the immunity of NOD mice and may explain the reduced incidence of diabetes in NOD.B6 Idd3 congenic mice. The most definitive way to address this problem is to generate a mouse in which the IL-2 gene of NOD mouse is replaced with that of B6. However, it has been difficult to generate a knockin mouse on the NOD background because of the difficulty in obtaining good ES cell lines capable of efficient germ-line transmission. On the other hand, backcrossing to the NOD strain of genetically modified mice typically developed from 129 or B6 ES cells have the problem that linked, non-NOD-derived flanking genes are retained during back-crossing (10). Knockout and transgenic mice generated in different genetic backgrounds have been backcrossed to NOD mice to assess the role of proteins in the T1D development. Even with extensive backcrossing, flanking regions from the original strain remain. These genes or sequences can have a significant effect on the phenotype of the mice. For example, IFN-␥ receptor knockout mice have shown opposite phenotypes dependent on the number of backcross generations because of this flanking gene problem (11). For a similar reason, analysis of congenic mice has not been able to precisely pinpoint the region responsible for disease susceptibility. Although NOD ES cells are ideal as a solution to the ‘‘linked genes’’ problem, these cells are not yet efficient in germ-line transmission (12). To overcome the poor transmission rate of pure NOD ES cells, F1 ES cells of NOD and 129 have been generated by us (between the 129S1/SvImJ and NOD/ShiLtJ strains) and others (13). All genes of these ES cells contain 1 copy each of the Author contributions: M.K., L.S.W., and R.A.F. designed research; M.K. and D.R. performed research; K.S.-G., E.E.E., and A.V.C. contributed new reagents/analytic tools; M.K. and D.R. analyzed data; and M.K., D.R., L.S.W., and R.A.F. wrote the paper. The authors declare no conflict of interest. Freely available online through the PNAS open access option. 1To
whom correspondence should be addressed. E-mail:
[email protected].
This article contains supporting information online at www.pnas.org/cgi/content/full/ 0904780106/DCSupplemental.
www.pnas.org兾cgi兾doi兾10.1073兾pnas.0904780106
D3Nds76
D3Nds36
Il2
3Bm291A3 20Bm291A3
Chromosome 3
Il21 Fgf2
Tenr
Centrin4
Kiaa1109
Kiaa1371
NOD
Ala Pro Thr Ser Ser Pro Thr Ser Ser Pro Thr Ser Ser Ser Thr Ala Glu Ala(Gln)8
C57BL/6
Ala Pro Thr Ser Ser Ser Thr Ser Ser Ser Thr --- --- --- --- Ala Glu Ala(Gln)12
NOD and 129 alleles. By homologous recombination, we can generate knockout or knockin NOD alleles without altering any of the flanking DNA. Here, we used these F1 ES cells to generate knockin mice in which the IL-2 gene of the NOD genome was replaced with that of B6. Our results demonstrate that the structural change in the IL-2 molecule is not responsible for the disease suppression found in NOD.B6 Idd3 congenic mice. Results Establishment of F1 ES Cells for Knockout Experiment. We generated
NOD/129 F1 ES cells from embryos obtained after crossing the NOD/ShiLtJ with 129/SvImJ mice as described in the materials and methods. Several ES cell lines were established and tested for their ability to produce chimeras capable of transmitting the NOD/129 F1 genome from ES cells. We chose several ES cell lines and demonstrated that these cells were capable of good germ-line transmission; we then used them to generate knockin mice. Generation of IL-2 Knockin Mice. To elucidate how structural differences in IL-2 contribute to the development of T1D, we generated an IL-2 knockin mouse in which exons I and II as well as intron I
A
and a portion of intron II were replaced with that of B6. We cloned the IL-2 gene from a NOD BAC library and made a construct as shown in Fig. 2A. We exchanged the Sbf1-ClaI fragment containing exons I and II cloned from the B6 genome to the corresponding region in the NOD-derived DNA. We transfected this construct into the ES cells and picked 90 colonies after selection with G418 and ganciclovir. We prepared DNA from these colonies and identified 10 clones having successful homologous recombination by Southern blotting using the 5⬘ and 3⬘ probes shown in Fig. 2B. At this point, as the ES cells were heterozygous for 129 and NOD, targeting could have occurred in either the 129 or NOD allele. It was difficult to determine which allele was targeted because sequences from NOD and 129 are almost identical within and surrounding the IL-2 gene. One of the ways to identify the targeted allele is by generating homozygous knockout ES cells followed by the analysis of microsatellite markers. By using higher levels of G418, it is possible to select clones in which duplication of allele containing neomycin resistance gene has occurred (14). We were unable to find clones that survived in high concentrations (2 mg/mL) of G418 in our F1 ES cells. Instead we generated chimeric mice from each clone and determining the targeted allele in a
SbfI
Portion of the IL-2 genomic region of NOD mice
PsiI
BclI
ClaI
I II
RV
III
SbfI RV
Targeting Vector
neo
SbfI
RV
PsiI
After R emoval of Neo Gene
I II
B
Homologous Recombination + -
+
C Neo removed
-
-
+
ClaI
RV
III
1kb
+
~20kb ~20kb 6kb
TK
6kb 4.5kb
Gene from C 57BL /6 Mice LoxP Southern blotting probe
Fig. 2. Generation of IL-2 knockin mice. (A) The construct for IL-2 knockin mice (NOD.IL-2B6ex1). The DNA for the region surrounding IL-2 Exon I was cloned from a genomic library of the NOD mouse. The IL-2 gene of B6 mice (shaded box) was cloned by PCR and replaced the corresponding NOD sequence. The construct was completed by cloning the long- and short-arm fragments into a pEasyflox plasmid containing a neomycin resistance gene and thymidine kinase gene. (Top) Genome of NOD mice. (Middle) Targeting vector. (Bottom) Genome after successful homologous recombination and removal of the neomycin resistance gene. (B) Southern blotting of ES DNA to identify targeted clones. The ⬇20-kb band is of NOD genome, and the 6-kb band is of clones successfully targeted. (C) Southern blotting of ES DNA to identify clones in which the neomycin gene has been removed. The 6-kb band is of the targeted clone, and the 4.5-kb band identifies clones in which the neomycin gene has been removed. The Southern blots shown were performed using the 3⬘ end probe.
Kamanaka et al.
PNAS 兩 July 7, 2009 兩 vol. 106 兩 no. 27 兩 11237
IMMUNOLOGY
Fig. 1. The genes located in the Idd3 region. The genes located in the Idd3 region, which is defined by recombination points of congenic mouse strains having been phenotyped at Idd3, are between the defined marker in Fgf2 and 20Bm291A3. Idd3 contains the genes encoding IL-2 and IL-21, as well as several other genes. Comparison of the sequences of Exon I of the IL-2 gene predicts that the amino acid sequence of mature proteins produced from the NOD and B6 are different. The location of micro satellite markers D3Nds76 and D3Nds36 are shown. The diagram is not proportional in regard to gene length. The left side is telomeric.
NOD allele targeted case
Chimera
129 allele targeted case
NOD
Chimera
NOD
NOD
A
NOD.IL-2B6ex1
IL-2
A
IL-10 D3Nds36
NOD
NOD/129 B6 (mutant)
B
6 D. B NO
D. IL
NO
ni) (T .
- + + + + - - +
2
-
IL -
Targeted allele
NO
Clone #24.9
B
D. IL -2 N
O
129
IL-2 Exon I
Id d3
129 Chr. 3
D co nt -2 ro B6 l NO ex 1 D
NOD Chr. 3
NOD ( 230 bp) 129 (222 bp)
Targeted allele
+
Clone #6.31 + + + + - B6 NOD ( 230 bp) 129 (222 bp) C57B L /6 (212 bp)
Fig. 3. Identification of ES clones that targeted the NOD allele in F1 ES cells. (A) Retrospective identification of alleles targeted in F1 ES cells. Chimeric mice from ES cells were crossed with NOD mice. The ES clone could have had either the NOD or 129 allele targeted. In the former, pups having the NOD allele targeted should have the NOD/NOD genotype at D3Nds36, whereas pups not having a targeted allele should have a NOD/129 genotype at D3Nds36. In the latter case, the opposite pattern is expected. (B) The electrophoresis of the D3Nds36 microsatellite PCR products from the pups derived from ES clones. (Upper) Result of clone no. 24.9. (Lower) Result of clone no. 6.31. On the top of each lane is indicated the presence (⫹) or absence (⫺) of the targeted gene. B6 indicates mouse DNA (i.e., not of ES cell origin). NOD and 129 alleles at D3Nds36 were distinguished by the 230 and 222 bp bands, respectively.
retrospective fashion. Although the Idd3 region from NOD and 129 is almost identical around the IL-2 gene, there is a microsatellite marker, D3Nds36, which can distinguish between the NOD and 129 alleles. By the combined analysis of the mutated gene and the D3Nds36 allele, we could retrospectively determine which allele was targeted in the ES cells (Fig. 3A). We removed the neo resistance gene by transient transfection with a Cre recombinase expressing plasmid into these ES cells. We then picked colonies and identified several colonies from 3 parental clones that were successfully deleted for the neomycin gene (Fig. 2C). We produced chimeras from 3 ES clones and crossed these mice with NOD mice. Fig. 3B shows the results of the analysis of microsatellite markers. The pups derived from clone no. 24.9 showed the pattern corresponding to targeting of the NOD allele. In this case, mutant allele positive mice showed a single D3Nds36 band corresponding to the NOD allele, whereas pups not having the mutation showed double bands corresponding to the 129 and NOD alleles. On the other hand, clone no. 6.9 and another clone (not shown) showed the opposite pattern, indicating that targeting took place in the 129 allele. Thus, we further backcrossed the pups derived from clone no. 24.9 to the NOD background using the expedited approach (‘‘speed backcrossing’’), selecting at each generation for the greatest number of NOD alleles at Idd markers for 12 generations (15). In summary, we have established a method to generate knockin/knockout mice in NOD/ 129 F1 ES cells and generated NOD mice in which a portion of the NOD IL-2 gene is replaced with that of B6. Analysis of the IL-2 Knockin Mice. First, we confirmed the expression
of IL-2 in the knockin mice (NOD.IL-2B6ex1), as the genetic 11238 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0904780106
Fig. 4. Intracellular and western analysis of IL-2 from NOD.IL-2B6ex1 mouse. (A) Intracellular IL-2 in CD4⫹ T cells detected by FACS. Splenocytes were stimulated with PMA and ionomycin for 5 h and intracellular IL-2 and IL-10 were detected by FACS. The plot shows CD4⫹ gated cells. (B) IL-2 secreted from NOD.IL-2B6ex1 cells detected by Western blotting. Culture supernatant from NOD.IL-2NOD control, NOD.IL-2B6ex1, NOD, and NOD.B6 Idd3 splenocytes were collected after stimulation with PHA for 12 h. Recombinant IL-2 derived from insect cells and the supernatants were subjected to electrophoresis and blotted with an IL-2-specific antibody. Please note that the right side migrated more slowly than the left because of the particular gel electrophoresis performed.
manipulation could have affected its transcription. Specifically, although the neomycin resistance gene was removed from the ES cells, 1 loxP site remains in intron II, which could affect the gene regulation. As shown in Fig. 4A, the protein level of IL-2 at the single cell level was not altered in activated CD4 T cells. We analyzed the frequencies and the numbers of nTreg cells detected as CD4⫹CD25⫹ cells, but found no significant difference between NOD.IL-2B6ex1 and NOD controls (P ⫽ 0.59). As the glycosylation pattern of IL-2 has been reported to correlate with the IL-2 aa sequence, we performed Western blotting of IL-2 in the supernatant collected from splenocytes stimulated with PHA (Fig. 4B). The mobility of IL-2 protein in polyacrylamide gels from knockin mice was the same as that of IL-2 produced by cells from NOD.B6 Idd3 congenic mice, and both proteins migrated more quickly than did NOD IL-2 protein. These results demonstrate that the amino acid sequence variation encoded by exon 1 is responsible for the different glycosylation patterns of the B6 and NOD IL-2 allotypes. Development of T1D. We next tested whether the structural variation
of the IL-2 gene affected T1D development. Mice were backcrossed 12 times to the NOD parental strain and the mice derived from interbreeding of IL-2 knockin heterozygous mice were used to assess the frequency of diabetes and insulitis. As shown in Fig. 5A, the progeny homozygous for the knockin allele (NOD.IL-2B6ex1), developed T1D with a similar overall frequency as the mice homozygous for the NOD IL-2 allele although there was a slight delay in the development of disease (P ⫽ 0.044). Both the NOD and the NOD.IL-2B6ex1 progeny developed more T1D than the NOD.B6 Idd3 congenic mice examined contemporaneously (P ⫽ 0.0000043 and P ⫽ 0.00004769, respectively). As the Idd3 region derived from the B6 strain is known to inhibit insulitis in NOD.B6 Kamanaka et al.
A
A
100
NOD x NOD.IL-2B6ex1 F1
NOD x NOD.B6 Idd3 F1 Product size (bp)
80
60
B6 40 NOD (n=52) NOD.IL-2B6ex1 NOD.IL-2 B6ex1 (n=39) NOD.B6 Idd3 (n=21)
20
50
100
B6
NOD
50
0 0
NOD 60
B % of PCR product
Disease free mice (%)
Product size (bp)
150
200
40
30
20
10
Age of mice (days) 0
NOD x NOD.IL-2B6ex1 F1
NOD x NOD.B6 Idd3 F1
100
Fig. 6. Allele-specific expression (ASE) of IL-2. (A) Example of the results obtained from an ASE assay. The IL-2 gene was amplified with fluorescencelabeled primers and digested with restriction enzyme. PCR products from the B6 and NOD alleles were quantified by measuring the fluorescence peaks. (B) Comparison of allele specific expression of IL-2 gene. The expression levels of each allele from NOD ⫻ NOD.IL-2B6ex1 and NOD ⫻ NOD.B6 Idd3 F1 mice are presented. The error bars indicate standard deviations. The open bars represent the NOD allele and the filled bars represent the B6 allele. *, P ⫽ 0.029, Mann–Whitney U test.
80 60 40 20 0
NOD
NOD.IL-2B6ex1 B6ex1 NOD.B6 Idd3 NOD.IL-2
Fig. 5. Development of T1D in NOD.IL-2B6ex1 mice. (A) Frequency of diabetes in NOD.IL-2B6ex1 mice. NOD.IL-2NOD control, NOD.IL-2B6ex1, and NOD.B6 Idd3 congenic mice were analyzed for the development of T1D. (B) The frequency and severity of insulitis in NOD.IL-2B6ex1 mice. The severity of insulitis was evaluated in female NOD.IL-2NOD (n ⫽ 11), NOD.IL-2B6ex1 (n ⫽ 10), and NOD.B6 Idd3 mice (n ⫽ 9) at 8 weeks of age. Each animal received one of the following scores: none (white bar, all islets observed are free of insulitis), mild (hatched bar, ⬍ 10% of the islets are infiltrated), moderate (gray bar, 10 –50% of the islets are affected), or extensive (black bar, ⬎ 50% of the islets have insulitis).
Idd3 mice, we examined the pancreatic islets of NOD.IL-2B6ex1 mice. As shown in the Fig. 5B, there was no significant difference between NOD.IL-2B6ex1 and littermate NOD.IL-2NOD controls (P ⫽ 0.67). Compared with these 2 groups of mice, NOD.B6 Idd3 mice had much reduced infiltration of lymphocytes into the islets (NOD.IL-2NOD controls vs. NOD.B6 Idd3, P ⫽ 0.002; NOD.IL2B6ex1 vs. NOD.B6 Idd3, P ⫽ 0.016). These results of T1D development and insulitis indicate that the structural difference of the IL-2 protein contributes only slightly to the frequency of T1D and the disease suppression mediated by the B6 Idd3 allele. Therefore the large protective effect of the B6 Idd3 region must be caused by the IL-2 polymorphisms found in the regulatory regions of the gene (5) rather than in exon 1. Allele Specific Expression (ASE) Assay. Yamanouchi et al. showed that autoimmune disease susceptibility and resistance alleles at Idd3 correlates with differential expression of IL-2 (5). They reported that IL-2 mRNA expression from the B6 allele is higher than from the NOD allele. Therefore, we tested whether the structural change in IL-2 affected the expression. We generated 2 groups of F1 mice: NOD ⫻ NOD.IL-2B6ex1 and NOD ⫻ NOD.B6 Idd3. We quantified the allele specific expression level of IL-2 by RT-PCR Kamanaka et al.
followed by cleavage with a restriction enzyme that cuts only the B6 PCR product. As shown in Fig. 6, IL-2 mRNA expression from the knockin allele is equivalent to that from the NOD allele. As expected, the B6 allele was expressed more than NOD allele in cells from NOD ⫻ NOD.B6 Idd3 F1 mice. Therefore, from these results, we conclude that the structural difference in the IL-2 molecule does not affect its expression level. Recently, Yamanouchi et al. (5) showed that polymorphisms in the IL-2 gene locus other than those determining the amino acid sequence difference are likely important in disease suppression by the B6 Idd3 allele. They also showed that transcriptional activity of IL-2 in cells from NOD.B6 Idd3 mice is higher than that from NOD cells. Their finding is compatible with our observation that the structural variation in the IL-2 molecule does not contribute substantially to disease development. These observations, therefore, place emphasis on the importance of the region(s) of the IL-2 gene that control IL-2 expression. The region that is responsible for the IL-2 gene transcriptional difference might be located in the extended 5⬘ region of IL-2. Del Rio et al. (16) reported that the SNP at ⫺1010C is responsible for the quantitative difference of IL-2 transcription because the variant alters AP-1 binding at this site. This hypothesis has yet to be tested in the disease model in vivo. To conclude, we found that disease inhibition by the B6 Idd3 allele is mostly caused, not through amino acid differences in the IL-2 structural gene but elsewhere, perhaps the sequences regulating IL-2 expression. In this report, we have established a system to generate knockout/knockin directly into the NOD genome obviating the concerns related to flanking regions. Materials and Methods Generation of NOD/129 F1 ES Cell Lines. 129/SvImJ females (agouti) were bred to NOD/ShiLtJ (albino) males and the d3.5 embryos were isolated and kept in a drop of KSOM medium (Millipore) covered with oil for 24 h. Subsequently, embryos were transferred into 96-well plates (1 embryo per well) containing an embryonic fibroblast cell layer and DMEM supplemented with 15% FCS (selected lot from Sigma) and LIF. After 5 days, cells were trypsinized and kept for 5– 6 days with the daily medium changes. Growing colonies were exPNAS 兩 July 7, 2009 兩 vol. 106 兩 no. 27 兩 11239
IMMUNOLOGY
Percentage of Animals
B
panded into larger plates and frozen or tested for germ-line transmission. For that, F1 ES cells were injected into d3.5 blastocysts from matings of B6 mice. All ES cell lines could generate ⬇100% chimerism in the mice produced (Table S1). Male chimeras from 5 ES lines (line nos. 5, 6, 7, 8, and 9) when mated with NOD female mice produced white offspring, indicating that the NOD allele from NOD/129 F1 ES cells was transmitted ( Table S2). Generation of Knockin ES Cells. To generate the knockin construct targeting the IL-2 locus, we first cloned the IL-2 gene from NOD by screening a NOD genomic library. The segment of the gene to be substituted was obtained from B6 DNA by PCR and confirmed by sequencing; a targeting construct to knock in the B6 IL-2 gene was generated as follows. The Psil-Clal fragment from the cloned NOD IL-2 gene was cloned into the pBluescript vector. Then, the Sbfl-Bcll fragment was cut out and substituted with the corresponding segment from B6 mice. This substituted Psi-ClaI fragment was cut by Bcll and linked to a fragment encoding the neomycin resistance gene flanked by loxP sequences. The fragment was then removed using Notl and Xhol and cloned into our standard thymidine kinase-expressing vector to yield the construct. DNA of this targeting construct was linearized by cleavage with Notl and electroporated into one of the 129/NOD ES cell lines (no. 7). Ten million cells were plated in selection media that are routinely used for our 129 ES cells (DMEM, 15% FCS, 103 units/mL LIF, 2 mM L-glutamine, 300 g/mL G418, and 1 M ganciclovir). Ninety colonies were obtained and transferred to duplicate 24 well plates. DNA was prepared from these cells and cleaved by restriction enzymes. Southern blotting analysis was performed to identify those clones in which appropriate gene targeting had occurred. For the 5⬘ side of the integration event, DNA was cleaved with EcoRV and BamHI and hybridized with a probe as illustrated in Fig. 2. Southern blot showed that all 10 clones exhibited the correct DNA pattern (for example, see Fig. 2). To analyze the 3⬘ side of the integration event, DNA was cleaved with EcoRV and Southern blots generated. Hybridization with the second probe (see Fig. 2) yielded the appropriate hybridization pattern for the same 10 clones. We were encouraged by the relatively high frequency (10/90 cell lines or 11%) of homologous recombination that we have observed with these ES cells. Karyotypic analysis of these 10 ES cell lines was performed and it was confirmed they had no obvious chromosomal abnormality. Identification of Clones Targeted to NOD Allele. To identify clones whose NOD allele was targeted, the microsatellite marker D3nds36 was used in a retrospective manner as described in the result section. PCR was performed with primer pair 5⬘-ATGAGTTGGGAAGCTTGTGC-3⬘ and 5⬘-GTAAAGGCCAAGGGAAAAGG-3⬘. Then polyacrylamide gel electrophoresis was performed, and the gel stained with ethidium bromide. The targeted clones were identified by digestion of PCR product amplified by cattgacacttgtgctccttgt and cagcatcaaaatgcaatcatct. The B6 allele has an EcoRV site within this product, whereas the NOD allele does not. PCR product of the NOD allele is not cleaved with EcoRV and produces a 569 bp band, whereas the B6 product is cleaved into 2 bands of 443 bp and 126 bp. The mice were backcrossed to NOD/ShiLtJ with the same screening method described above.
g/mL streptomycin, and 50 M 2-mercaptoethanol) with 50 ng/mL phorbol 12-myristate 13-acetate (PMA) and 250 ng/mL ionomycin for 5 h with an addition of GolgiStop during the last 2 h. Cells were stained with anti-CD4 antibody and fixed using BD Cytofix/Cytoperm kit. Then cells were stained with antibodies against IL-2 and IL-10 and were analyzed using a FACSCalibur (BD Biosciences). Western Blotting. SDS/PAGE was performed on recombinant mouse IL-2 samples (the B6 allele expressed in baculovirus-infected Trichoplusia ni insect cells, BD) and IL-2-containing culture supernatants from NOD.IL-2NOD control, NOD.IL-2B6ex1, NOD, and NOD.B6 Idd3 according to the procedure of Laemmli. Proteins were transferred onto nylon membranes using standard procedures. The membrane was incubated with anti-IL-2 antibody (JES6 –1A12, BD), was washed twice, and was incubated with HRP-conjugated sheep antirat Ig antibody (F(ab⬘)2 fragment) (Amersham). The membrane was developed using West Pico Luminol/Enhancer solutions (Pierce). Analysis of T1D Development. Diabetes was monitored by measuring urine glucose with Diastix (Bayer). Animals were considered diabetic after 2 consecutive readings of ⱖ500 mg/dL. The frequency of diabetes was compared between strains with the Kaplan–Meier log-rank test. Histology of Islets. The presence of insulitis was assessed after fixation of pancreata in buffered 10% formalin and paraffin sectioning. Tissue sections (5 m) were stained with hematoxylin and eosin and examined for the presence of mononuclear cell infiltration. Two noncontiguous sections of pancreas were examined for each mouse. Comparisons of groups were performed using the Mann–Whitney U test. Allele-Specific Expression (ASE) assay of IL-2 gene. NOD ⫻ NOD.IL-2B6ex1 F1 and NOD ⫻ NOD.B6 Idd3 F1 mice were injected with 5 g of anti-CD3 and were killed after 1 h. CD4 T cells were isolated by MACS (Miltenyi) using negative selection with anti-CD8 (TIB105, ATCC), anti-NK1.1(HB191, ATCC), antiIA(M5114, ATCC). Total RNA was purified from these cells by TRIzol (Invitrogen) according to the manufacturer’s instructions, and was DNase treated using the RNase-free DNase kit (Qiagen). 1 g of DNase treated total RNA was used as the template for cDNA synthesis using SuperScript II reverse transcriptase (Invitrogen). We performed RT-PCR on the IL-2 gene using primers labeled with the fluorescent dye Fluorescein (Fam) (forward Fam-CATTCCCTTTCCCAACACAT, reverse AACGTGTAAATAACATGTGACCAGA). PCR product (154 bp) has (EcoRV, New England Biolabs) cleavage site in B6 allele generating an 88bp fragment that is not in NOD. After digestion of the PCR product, it was analyzed by capillary electrophoresis (Applied Biosystems 3730xl) and analyzed by Genemapper software to quantitate the PCR product from each allele. Allele specific expression was calculated as a percentage in added value of B6 and NOD allele.
Intracellular Staining of Cytokines. 5 ⫻ 106 splenocytes from NOD.IL-2B6ex1 knockin and their control NOD littermates were isolated and cultured in RPMI medium 1640 (10% FBS, 100 g/mL L-glutamine, 100 units/mL penicillin, 100
ACKNOWLEDGMENTS. We thank Linda Evangelisti and Cindy Hughes for generating ES cells and chimeric mice, respectively; Judy Stein for screening; Jonathan Alderman for managing the mouse program; Fran Manzo for help with manuscript preparation; and Li Wen for providing NOD mice with the support of National Institutes of Health, Yale University, New Haven, CT, Grant P30 DK 045735-17. R.A.F. is an investigator of the Howard Hughes Medical Institute. L.S.W. is a Juvenile Diabetes Research Foundation/Wellcome Trust Principal Research Fellow. This work was supported by National Institutes of Health Grants DK064312 and DK04735.
1. Serreze DV, Leiter EH (1994) Genetic and pathogenic basis of autoimmune diabetes in NOD mice. Curr Opin Immunol 6:900 –906. 2. Maier LM, Wicker LS (2005) Genetic susceptibility to type 1 diabetes. Curr Opin Immunol 17:601– 608. 3. Ghosh S, et al. (1993) Polygenic control of autoimmune diabetes in nonobese diabetic mice. Nat Genet 4:404 – 409. 4. Wicker LS, et al. (1994) Resistance alleles at two non-major histocompatibility complexlinked insulin-dependent diabetes loci on chromosome 3, Idd3 and Idd10, protect nonobese diabetic mice from diabetes. J Exp Med 180:1705–1713. 5. Yamanouchi J, et al. (2007) Interleukin-2 gene variation impairs regulatory T cell function and causes autoimmunity. Nat Genet 39:329 –337. 6. Podolin PL, et al. (2000) Differential glycosylation of interleukin 2, the molecular basis for the NOD Idd3 type 1 diabetes gene? Cytokine 12:477– 482. 7. Wrenshall LE, Platt JL (1999) Regulation of T cell homeostasis by heparan sulfate-bound IL-2. J Immunol 163:3793–3800. 8. Wrenshall LE, Platt JL, Stevens ET, Wight TN, Miller JD (2003) Propagation and control of T cell responses by heparan sulfate-bound IL-2. J Immunol 170:5470 –5474. 9. King C, Ilic A, Koelsch K, Sarvetnick N (2004) Homeostatic expansion of T cells during immune insufficiency generates autoimmunity. Cell 117:265–277.
10. Ridgway WM, Healy B, Smink LJ, Rainbow D, Wicker LS (2007) New tools for defining the ‘genetic background’ of inbred mouse strains. Nat Immunol 8:669 – 673. 11. Kanagawa O, Xu G, Tevaarwerk A, Vaupel BA (2000) Protection of nonobese diabetic mice from diabetes by gene(s) closely linked to IFN-gamma receptor loci. J Immunol 164:3919–3923. 12. Nagafuchi S, et al. Establishment of an embryonic stem (ES) cell line derived from a non-obese diabetic (NOD) mouse: in vivo differentiation into lymphocytes and potential for germ line transmission. FEBS Lett 455:101–104, 1999. 13. Brook FA, et al. (2003) The derivation of highly germline-competent embryonic stem cells containing NOD-derived genome. Diabetes 52:205–208. 14. Mortensen RM, Conner DA, Chao S, Geisterfer-Lowrance AA, Seidman JG (1992) Production of homozygous mutant ES cells with a single targeting construct. Mol Cell Biol 12:2391–2395. 15. Serreze DV, et al. (1996) B lymphocytes are essential for the initiation of T cell-mediated autoimmune diabetes: Analysis of a new ‘‘speed congenic’’ stock of NOD. Ig mu null mice. J Exp Med 184:2049 –2053. 16. del Rio R, et al. (2008) SNPs upstream of the minimal promoter control IL-2 expression and are candidates for the autoimmune disease-susceptibility locus Aod2/Idd3/Eae3. Genes Immun 9:115–121.
11240 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0904780106
Kamanaka et al.
