- Email: [email protected]
Insulin as a T cell antigen in type 1 diabetes supported by the evidence from the insulin knockout NOD mice
Diabetes Research and Clinical Practice 77S (2007) S155–S160 www.elsevier.com/locate/diabres
Insulin as a T cell antigen in type 1 diabetes supported by the evidence from the insulin knockout NOD mice§,§§ Hiroaki Moriyama *, Masao Nagata, Takashi Arai, Yasuyo Okumachi, Katsumi Yamada, Reiko Kotani, Hisafumi Yasuda, Kenta Hara, Koich Yokono Department of Internal and Geriatric Medicine, Kobe University Graduate School of Medicine, 7-5-1 Kusunoki-cho, Chuo-ku, Kobe 650-0017, Japan Accepted 29 January 2007 Available online 24 April 2007
Abstract Rodents have two functional preproinsulin genes named insulin 1 and insulin 2 on different chromosome and have two amino acid differences in insulin B chain. We have established insulin 1 or insulin 2 knockout (KO) non-obese diabetic (NOD) colonies in the animal institute of Kobe University and evaluated anti-insulin autoimmunity. Similar to the previous report, insulin 1-KO provides strong protection from insulitis (islet-infiltration of mononuclear cells) and diabetes, whereas the insulin 2-KO markedly accelerated insulitis and development of diabetes even at further backcross breeding with NOD/Shi/Kbe mice (P < 0.0001). Expression of serum anti-insulin autoantibodies (IAA) was enhanced in insulin 2-KO mice at a time between 10 and 15 weeks of age (P < 0.005) while the expression of insulin 1-KO NOD mice was rather reduced. Furthermore, T cell reactivity in splenocytes of insulin 2-KO NOD mice to insulin 1 B:9-23 peptide was increased (P < 0.05), suggesting that expanding insulin-reactive T cells may contribute to the acceleration of diabetes in insulin 2-KO mice. Based on those observations, we hypothesize that insulin 1 is a crucial T cell antigen in murine autoimmune diabetes and modification of anti-insulin autoimmunity can be applicable to antigenbased therapy for human type 1 diabetic patients. # 2007 Elsevier Ireland Ltd. All rights reserved. Keywords: Type 1 diabetes; Insulin; Autoantigen; Autoimmune; ELISPOT
1. Introduction
§ This research was supported by a grant for 21st century COE program, ‘‘Center of Excellence for Signal Transduction Disease: Diabetes Mellitus as Model’’ and Grants-in-aid for Scientific Research (1759031, 17590932 and 18590990) from Ministry of Education, Culture, Sports, Science and Technology of Japan. §§ This study was presented at 13th Korea–Japan Symposium on Diabetes Mellitus which was held on 11–12 November 2005 in Seoul, Korea. * Corresponding author. Tel.: +81 78 382 5901; fax: +81 78 382 5919. E-mail address: [email protected] (H. Moriyama).
There have been several lines of evidence for the importance of insulin as a T cell antigen of type 1 diabetes. First of all, insulin is a pancreatic b cell specific antigen among type 1 diabetes-related antigens. Human genetic analysis demonstrated that the insulin gene variable tandem repeat (VNTR) allele (IDDM2) is associated with the protection from the development of human type 1 diabetes [1]. In addition, insulin autoantibody (IAA) is an excellent predictive maker for the development of type 1 diabetes in both human and the non-obese diabetes (NOD) mouse, which is a well-established animal model for type 1 diabetes [2,3].
0168-8227/$ – see front matter # 2007 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.diabres.2007.01.050
S156
H. Moriyama et al. / Diabetes Research and Clinical Practice 77S (2007) S155–S160
Majority of T cells in islet-infiltrates of NOD mice recognize insulin B:9-23 peptide [4] and human T cell reactivity to B:9-23 peptide are detected in type 1 diabetes patients [5]. Administration of insulin or insulin peptide through various routes (subcutaneous, oral, nasal and intravenous) to NOD mice prevents diabetes [6–9]. Those results indicate that insulin and the peptide exhibit either pathogenic or protective character depending on the situation probably through induction or regulation of autoreactive T cells. Rodents have two functional preproinsulin genes named insulin 1 and insulin 2 on different chromosome and have two amino acid differences in insulin B chain. Insulin 1 lacks intron present in insulin 2, suggesting that insulin 1 is a retroposon of insulin 2 gene generated by an RNA-mediated duplicationtransposition event [10]. We have created insulin 1 or insulin 2-KO NOD congenic mice and reported that dramatic opposite effects of KO genes were unexpectedly observed [11]. In this study, the KO effects were confirmed at further backcross generation with the breeding of NOD/ Shi/Kbe mice. Analyses for the mechanism of autoreactive T cells have been also performed.
2.3. Diagnoses of diabetes Homozygous KO mice were produced by the breeding of heterozygous mice at BC9 generation for insulin 1-KO and at BC12 for insulin 2-KO mice and the development of diabetes have been followed up. The glucose level for the homozygous KO mice was measured weekly with FreeStyle (TheraSense, Alameda, CA) or Glutest Ace (Sanwakagaku, Nagoya, Japan) blood glucose monitoring system. The mice were considered diabetic after two consecutive blood glucose values were greater than 250 mg/dl. After development of diabetes, the mice were sacrificed immediately and the pancreata of some mice were fixed in 10% buffered formalin for histological analysis. 2.4. Histology The pancreata obtained from the mice were fixed in 10% buffered formalin and then embedded in paraffin. Paraffin sections were stained with hematoxylin/eosin. Pancreatic sections were microscopically examined for the degree of insulitis. 2.5. Anti-insulin autoantibody assay
2. Materials and methods
Mice were bled for the measurement of serum anti-insulin autoantibody (IAA) levels. IAA was measured with a 96-well filtration plate micro-IAA assay as previously reported [2] and expressed as an index: index = (sample D (cpm) negative control D (cpm))/(positive D (cpm) negative control D (cpm)). A value of 0.01 or greater is considered positive.
2.1. Mice
2.6. Antigens
The original knockouts were produced in 129S1/SvImJ embryonal cell lines by Jami and coworkers [12]. The knockout cell lines were microinjected into C57BL/6 blastocytes and lines developed with C57BL/6. Insulin 1 and insulin 2 KO congenic NOD mice were established by breeding the original insulin KO mice using speed congenic method fixing NOD diabetogenic loci (idd 1–14) at Barbara Davis Center [13]. Both strains have been backcrossed onto NOD/Bdc mice and non-NOD genomic regions flanking each insulin gene knockout were then less than 10 cM. We have further been breeding the knockouts onto NOD/ Shi/Kbe mice in the Institute for Experimental Animals, Kobe University School of Medicine. We are at the 10th backcross (BC) generation for insulin 1-KO mice and the 13th BC for insulin 2-KO mice. All animals were handled under the Guidelines for Animal Experimentation of Kobe University School of Medicine.
HPLC-purified insulin 1B:9-23 peptide (PHLVEALYLVCGERG) and insulin 2B:9-23 peptide (SHLVEALYLVCGERG) were purchased from Invitrogen Japan (Tokyo, Japan). Purified recombinant human GAD65 expressed in yeast (Saccharomyces cerevisiae) was obtained from RSR Ltd. (Cardiff, UK). These peptides and the protein were used as antigens in vitro ELISPOT assay.
2.2. Genotype analysis Genomic DNA was extracted from mouse tails. Insulin KO mice were genotyped for KO genes and wild-type insulin genes by using PCR as previously described [13]. The PCR products were electrophoresed on 2% agarose gels and visualized by ethidium bromide staining.
2.7. ELISPOT assay The murine ELISPOT assay was performed with modification of human ELISPOT assay as previously described [14]. In brief, splenocytes (2 105) were cultured in nitro-cellulosebottomed 96-well microtiter plates (Millititer, Millipore Corp., Bedford, MA) in 200 ml of RPMI 1640 with 1% FCS. ELISPOT assay for mouse IFN-g or IL-4 was performed as manufacturer’s instruction (Mabtech AB, Stockholm, Sweden) and spots were analyzed with an ImmunoSpot Analyzer (Cellular Technology, Cleveland, OH). 2.8. Statistics Survival curves were analyzed with the log rank test at Kaplan–Meier method. Mann–Whitney U-test was used to compare the values of IAA expression. Wilcoxon matched
H. Moriyama et al. / Diabetes Research and Clinical Practice 77S (2007) S155–S160
S157
pair tests was applied to compare the number of ELISPOT counts.
3. Results Similar to the previous report [11], dramatic opposite effects of KO genes were observed by survival curve analysis (Fig. 1, P < 0.0001). The cumulative diabetic incidence of insulin 2-KO NOD mice at 30 weeks of age was 100% and most mice (25/29, 86%) developed overt diabetes by 17 weeks of age. Four mice became overt diabetes by 10 weeks of age with severe insulitis examined by histological analysis (data not shown). In contrast, no insulin 1-KO NOD mice developed overt diabetes although significant but rather mild insulitis were observed at later weeks of age. To further investigate anti-insulin autoimmuity in insulin-KO NOD mice, we examined the expression of serum insulin autoantibodies (IAA) of insulin 1-KO and insulin 2-KO NOD mice at a time between 10 and 15 weeks of age. As shown in Fig. 2, all of insulin 2-KO NOD mice (5/5) showed enhanced IAA values and the IAA level were much higher than that of wild-type NOD mice (P < 0.005). IAA of insulin 1-KO NOD mice seemed to be lower than that of wild-type NOD mice although two groups were not statistically different. We hypothesized that insulin 2 knockout might accelerate anti-insulin autoimmunity because of decreased tolerance to insulin peptides, especially to the important peptide, namely insulin B:9-23 peptide. We therefore evaluated T cell reactivity in splenocytes to islet-related autoantigens with ELISPOT assay. As shown in Fig. 3, the number of IFN-g spots to hGAD65 protein, insulin 1 B:9-23 peptide, and insulin 2 B:9-23 peptide were elevated in comparison to unstimualeted
Fig. 1. Life table analysis of insulin-KO NOD mice. Filled squares indicate insulin homozygous insulin 2-KO NOD mice at BC 12 generation (n = 29). Filled circles indicate homozygous insulin 1KO NOD at BC9 (n = 30) and filled triangles indicate wild-type NOD/ Shi/Kbe mice (n = 37). The three curves are dramatically different with log rank test (P < 0.0001).
Fig. 2. Serum anti-insulin autoantibody levels. IAA of homozygous insulin 2-KO, insulin 1-KO and wild-type NOD mice measured at a time between 10 and 15 weeks of age were plotted. Filled squares, filled circles and filled triangle indicate insulin 2-KO, insulin 1-KO, and wild-type NOD mice, respectively. IAA level of homozygous insulin 2-KO NOD mice are significantly higher than that of wild-type NOD mice with Mann–Whitney U-test (P < 0.005).
control wells. Statistic analysis with Wilcoxon matched pair test indicated that the number of IFN-g spots to insulin 1 B:9-23 peptide, beside other two antigens, were significantly higher than that of unstimulated control wells. On the other hand, any number of IL-4 spot to these three antigens was not increased by the antigenic stimulation. 4. Discussion In a previous study, we created insulin 1-KO and insulin 2-KO NOD congenic mice [13]. The insulin 1 knockout strongly prevented the progression of autoimmune diabetes in NOD mice whereas the insulin 2 knockout remarkably accelerated type 1 diabetes of NOD mice [11]. In this study, we have established insulin 1-KO and insulin 2-KO NOD colonies with NOD/Shi/Kbe mice. Further backcross onto NOD/Shi/ Kbe mice did not affect either the protective effect of insulin1-KO genes at BC9 or disease-enhancing effect of insulin 2-KO gene at BC12. Our results indicated that the effect of knockout genes was consistent even bred with the NOD mice in different subcolony. In other aspect, the environmental change (From Barbara Davis Center in Colorado to Kobe University in Japan) did not affect on the knockout genes effect. The insulin autoantibody (IAA) is an excellent predictive maker for the development of type 1 diabetes in both human and NOD mice and the presence of antiinsulin autoantibodies in NOD mice usually correlate with insulitis [15]. We have launched the measurement for mIAA in our laboratory with the same protocol at Babara Davis Center. Our mIAA assay indicated that insulin 2 knockout gene strongly enhanced the
S158
H. Moriyama et al. / Diabetes Research and Clinical Practice 77S (2007) S155–S160
Fig. 3. ELISPOT assay for IFN-g and IL-4 responses to GAD65 protein, insulin 1 B:9-23 peptide, and insulin 2 B:9-23 peptide in insulin 2-KO NOD mice. Individual spot formation of splenocytes from insulin 2-KO NOD mice is shown with the stimulation by GAD65 antigen or insulin peptides in the figure. Insulin 1 B:9-23-stimulated IFN-g responses are significantly observed compared with those in the absence of antigens (P < 0.05 by Wilcoxon matched pair test).
expression of serum anti-insulin autoantibodies, corresponding to the aggressive insulitis and acceleration of diabetes in insulin 2 KO-NOD mice. Polychronakos and Chentoufi demonstrated, using original insulin-KO mice with 129/C57BL/6 background genes, that thymic expression of insulin was related to the number of copies of the insulin 2 gene [16]. Similarly, insulin 2 KO-NOD mice also showed no insulin in thymus while insulin 1-KO NOD mice showed significant thymic insulin expression when detected using real-time PCR analysis [17]. We therefore, hypothesize that insulin 2-KO gene may accelerate type 1 diabetes because of diminished expression of insulin within the thymus and decreased tolerance to insulin epitopes. T cell reactivity in splenocytes of insulin 2-KO NOD mice to insulin B:9-23 peptide in this study is consistent with this hypothesis. In addition, protection from diabetes in the insulin 1-KO NOD mice is hypothesized to be due to removal of the more pathogenic insulin peptide sequence including B9-23 (B9 Proline v.s. Serine). Recently, the double insulin knockout mice lacking both the native insulin 1 and native insulin 2 genes have been established by the combination of insulin 1-KO, insulin 2-KO, and another NOD transgenic mouse with a muted preproinsulin gene to have alanine at position B16 rather than tyrosine of the B:9-23 sequence [18]. The double insulin-KO NOD mice that were metabolically rescued with a muted gene encoding preproinsulin, totally abrogated autoimmunity of NOD mice with prevention of anti-insulin autoantibodies, insulitis, and diabetes. The result from double insulin-KO NOD
mice suggests that the native insulin B:9-23 sequence is essential for anti-islet autoimmunity of the NOD mouse and the difference between insulin 1 B9-23 and insulin 2 B9-23 sequence might be crucial for the opposite outcomes of single insulin-KO NOD mice. Insulin B:9-23 peptide was discovered when Daniel et al. cloned islet-reactive CD4 T cells directly from islets of NOD mice by using whole islets as initial antigens [4]. Insulin B:9-23 has been identified that binds to both mouse I-Ag7 and human DQ8 [19]. Insulin B:15-23 recognized by CD8 T cells with Kd molecules is within insulin B:9-23 [20]. The B:9-23 reactive T cell clones and the B:15-23 reactive CD8 T cell clone (clone G9C8) are pathogenic in that those clones are able to transfer diabetes. In contrast to those pathogenic T cell clones, Zekzer et al. described that an T cell clone (clone 2H6) reacting insulin B:9-23 and B:12-25 could mediate disease protection [21]. Furthermore, Du et al. have demonstrated that 2H6 TCR transgenic NOD possessed insulin specific immunoregulatory T cells by virtue of their production of TGF-b1 similar to the parent clone [22]. Further evidence supporting a more pathogenic role of insulin 1 over insulin 2 peptide comes from experiments immunizing mice models with both forms of B:9-23 peptides. Subcutaneous injection of the insulin 2 B:9-23 but not the insulin 1 peptide significantly protects NOD mice from diabetes [23]. Treatment with B:9-23 immunization and the TLR3 ligand, poly:IC, induces diabetes in BALB/c mice having transgenic expression of costimulatory B7.1 molecules on pancreatic b cells, termed experimental
H. Moriyama et al. / Diabetes Research and Clinical Practice 77S (2007) S155–S160
autoimmune diabetes (EAD) [24]. In the EAD model, insulin 1 B:9-23 peptide immunization accelerated diabetes in the B7.1 mice compared with insulin 2 peptide [23]. One issue of obvious interest is the applicability of our observation to human subjects. It is possible that the most appropriate strategy for inducing tolerance in autoimmunity may be the combination of important autoantigen peptide with potent immunomodulatory agent. Recently, combination treatment of anti-CD3 and nasal proinsulin peptide can reverse recent-onset diabetes in two murine models with much higher efficacy than monotherapy with anti-CD3 or antigen alone [25]. However, immune response to key antigenic dominant shows complexity between the balance of induction of pathogenic cells and regulatory cells, often referred to as the ‘‘double-edged sword’’ [26]. Much effort is needed in studying basic pathogenesis with animal models as in carrying out human clinical trial to increase the probability that nontoxic antigen-specific therapies will be performed in the course of human type 1 diabetes. Acknowledgements We thank Ms. Atsumi Katsuta and Mr. Takeshi Hamada for the laboratory assistance. References [1] D. Hanahan, Peripheral-antigen-expressing cells in thymic medulla: factors in self-tolerance and autoimmunity, Curr. Opin. Immunol. 10 (1998) 656–662. [2] L. Yu, D.T. Robles, N. Abiru, P. Kaur, M. Rewers, K. Kelemen, et al., Early expression of antiinsulin autoantibodies of humans and the NOD mouse: evidence for early determination of subsequent diabetes, PNAS 97 (2000) 1701–1706. [3] E. Bonifacio, M. Atkinson, G. Eisenbarth, D. Serreze, T.W. Kay, E. Lee-Chan, et al., International Workshop on Lessons From Animal Models for Human Type 1 Diabetes: identification of insulin but not glutamic acid decarboxylase or IA-2 as specific autoantigens of humoral autoimmunity in nonobese diabetic mice, Diabetes 50 (2001) 2451–2458. [4] D. Daniel, R.G. Gill, N. Schloot, D. Wegmann, Epitope specificity, cytokine production profile and diabetogenic activity of insulin-specific T cell clones isolated from NOD mice, Eur. J. Immunol. 25 (1995) 1056–1062. [5] D.G. Alleva, P.D. Crowe, L. Jin, W.W. Kwok, N. Ling, M. Gottschalk, et al., A disease-associated cellular immune response in type 1 diabetics to an immunodominant epitope of insulin, J. Clin. Invest. 107 (2001) 173–180. [6] Z.J. Zhang, L. Davidson, G. Eisenbarth, H.L. Weiner, Suppression of diabetes in nonobese diabetic mice by oral administration of porcine insulin, Proc. Natl. Acad. Sci. USA 88 (1991) 10252–10256.
S159
[7] A. Muir, A. Peck, M. Clare-Salzler, Y.H. Song, J. Cornelius, R. Luchetta, et al., Insulin immunization of nonobese diabetic mice induces a protective insulitis characterized by diminished intraislet interferon-gamma transcription, J. Clin. Invest. 95 (1995) 628–634. [8] D. Daniel, D.R. Wegmann, Protection of nonobese diabetic mice from diabetes by intranasal or subcutaneous administration of insulin peptide B-(9-23), PNAS 93 (1996) 956–960. [9] P.R. Hutchings, A. Cooke, Comparative study of the protective effect afforded by intravenous administration of bovine or ovine insulin to young NOD mice, Diabetes 44 (1995) 906–910. [10] M.B. Soares, E. Schon, A. Henderson, S.K. Karathanasis, R. Cate, S. Zeitlin, et al., RNA-mediated gene duplication: the rat preproinsulin I gene is a functional retroposon, Mol. Cell Biol. 5 (1985) 2090–2103. [11] H. Moriyama, N. Abiru, J. Paronen, K. Sikora, E. Liu, D. Miao, et al., Evidence for a primary islet autoantigen (preproinsulin 1) for insulitis and diabetes in the nonobese diabetic mouse, Proc. Natl. Acad. Sci. USA 100 (2003) 10376–10381. [12] L. Deltour, P. Leduque, N. Blume, O. Madsen, P. Dubois, J. Jami, et al., Differential expression of the two nonallelic proinsulin genes in the developing mouse embryo, PNAS 90 (1993) 527– 531. [13] J. Paronen, H. Moriyama, N. Abiru, K. Sikora, E. Melanitou, S. Babu, et al., Establishing insulin 1 and insulin 2 knockout congenic strains on NOD genetic background, Ann. NY Acad. Sci. 1005 (2003) 205–210. [14] R. Kotani, M. Nagata, H. Moriyama, M. Nakayama, K. Yamada, S.A., et al. Detection of GAD65-reactive T-Cells in type 1 diabetes by immunoglobulin-free ELISPOT assays, Diabetes Care 25 (2002) 1390–1397. [15] D.T. Robles, G.S. Eisenbarth, N.J.M. Dailey, L.B. Peterson, L.S. Wicker, Insulin autoantibodies are associated with islet inflammation but not always related to diabetes progression in NOD congenic mice, Diabetes 52 (2003) 882–886. [16] A.A. Chentoufi, C. Polychronakos, Insulin expression levels in the thymus modulate insulin-specific autoreactive T-cell tolerance: the mechanism by which the IDDM2 locus may predispose to diabetes, Diabetes 51 (2002) 1383–1390. [17] M. Nakayama, N. Babaya, D. Miao, K. Sikora, J.F. Elliott, G.S. Eisenbarth, Thymic expression of mutated B16:A preproinsulin messenger RNA does not reverse acceleration of NOD diabetes associated with insulin 2 (thymic expressed insulin) knockout, J. Autoimmunol. 25 (2005) 193–198. [18] M. Nakayama, N. Abiru, H. Moriyama, N. Babaya, E. Liu, D. Miao, et al., Prime role for an insulin epitope in the development of type 1 diabetes in NOD mice, Nature 435 (2005) 220– 223. [19] K.H. Lee, K.W. Wucherpfennig, D.C. Wiley, Structure of a human insulin peptide-HLA-DQ8 complex and susceptibility to type 1 diabetes, Nat. Immunol. 2 (2001) 501–507. [20] F.S. Wong, J. Karttunen, C. Dumont, L. Wen, I. Visintin, I.M. Pilip, et al., Identification of an MHC class I-restricted autoantigen in type 1 diabetes by screening an organ-specific cDNA library, Nat. Med. 5 (1999) 1026–1031. [21] D. Zekzer, F.S. Wong, L. Wen, M. Altieri, T. Gurlo, H. von Grafenstein, et al., Inhibition of diabetes by an insulin-reactive CD4 T-cell clone in the nonobese diabetic mouse, Diabetes 46 (1997) 1124–1132. [22] W. Du, F.S. Wong, M.O. Li, J. Peng, H. Qi, R.A. Flavell, et al., TGF-b signaling is required for the function of insulin-reactive T regulatory cells, J. Clin. Invest. 116 (2006) 1360–1370.
S160
H. Moriyama et al. / Diabetes Research and Clinical Practice 77S (2007) S155–S160
[23] D. Devendra, J. Paronen, H. Moriyama, D. Miao, G.S. Eisenbarth, E. Liu, Differential immune response to B:9-23 insulin 1 and insulin 2 peptides in animal models of type 1 diabetes, J. Autoimmun. 23 (2004) 17–26. [24] H. Moriyama, L. Wen, N. Abiru, E. Liu, L. Yu, D. Miao, et al., Induction and acceleration of insulitis/diabetes in mice with a viral mimic (polyinosinic-polycytidylic acid) and an insulin selfpeptide, Proc. Natl. Acad. Sci. USA 99 (2002) 5539–5544.
[25] D. Bresson, L. Togher, E. Rodrigo, Y. Chen, J.A. Bluestone, K.C. Herold, et al., Anti-CD3 and nasal proinsulin combination therapy enhances remission from recent-onset autoimmune diabetes by inducing Tregs, J. Clin. Invest. 116 (2006) 1371– 1381. [26] A. Pugliese, Peptide-based treatment for autoimmune diseases: learning how to handle a double-edged sword, J. Clin. Invest. 111 (2003) 1280–1282.
Insulin as a T cell antigen in type 1 diabetes supported by the evidence from the insulin knockout NOD mice§,§§ Hiroaki Moriyama *, Masao Nagata, Takashi Arai, Yasuyo Okumachi, Katsumi Yamada, Reiko Kotani, Hisafumi Yasuda, Kenta Hara, Koich Yokono Department of Internal and Geriatric Medicine, Kobe University Graduate School of Medicine, 7-5-1 Kusunoki-cho, Chuo-ku, Kobe 650-0017, Japan Accepted 29 January 2007 Available online 24 April 2007
Abstract Rodents have two functional preproinsulin genes named insulin 1 and insulin 2 on different chromosome and have two amino acid differences in insulin B chain. We have established insulin 1 or insulin 2 knockout (KO) non-obese diabetic (NOD) colonies in the animal institute of Kobe University and evaluated anti-insulin autoimmunity. Similar to the previous report, insulin 1-KO provides strong protection from insulitis (islet-infiltration of mononuclear cells) and diabetes, whereas the insulin 2-KO markedly accelerated insulitis and development of diabetes even at further backcross breeding with NOD/Shi/Kbe mice (P < 0.0001). Expression of serum anti-insulin autoantibodies (IAA) was enhanced in insulin 2-KO mice at a time between 10 and 15 weeks of age (P < 0.005) while the expression of insulin 1-KO NOD mice was rather reduced. Furthermore, T cell reactivity in splenocytes of insulin 2-KO NOD mice to insulin 1 B:9-23 peptide was increased (P < 0.05), suggesting that expanding insulin-reactive T cells may contribute to the acceleration of diabetes in insulin 2-KO mice. Based on those observations, we hypothesize that insulin 1 is a crucial T cell antigen in murine autoimmune diabetes and modification of anti-insulin autoimmunity can be applicable to antigenbased therapy for human type 1 diabetic patients. # 2007 Elsevier Ireland Ltd. All rights reserved. Keywords: Type 1 diabetes; Insulin; Autoantigen; Autoimmune; ELISPOT
1. Introduction
§ This research was supported by a grant for 21st century COE program, ‘‘Center of Excellence for Signal Transduction Disease: Diabetes Mellitus as Model’’ and Grants-in-aid for Scientific Research (1759031, 17590932 and 18590990) from Ministry of Education, Culture, Sports, Science and Technology of Japan. §§ This study was presented at 13th Korea–Japan Symposium on Diabetes Mellitus which was held on 11–12 November 2005 in Seoul, Korea. * Corresponding author. Tel.: +81 78 382 5901; fax: +81 78 382 5919. E-mail address: [email protected] (H. Moriyama).
There have been several lines of evidence for the importance of insulin as a T cell antigen of type 1 diabetes. First of all, insulin is a pancreatic b cell specific antigen among type 1 diabetes-related antigens. Human genetic analysis demonstrated that the insulin gene variable tandem repeat (VNTR) allele (IDDM2) is associated with the protection from the development of human type 1 diabetes [1]. In addition, insulin autoantibody (IAA) is an excellent predictive maker for the development of type 1 diabetes in both human and the non-obese diabetes (NOD) mouse, which is a well-established animal model for type 1 diabetes [2,3].
0168-8227/$ – see front matter # 2007 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.diabres.2007.01.050
S156
H. Moriyama et al. / Diabetes Research and Clinical Practice 77S (2007) S155–S160
Majority of T cells in islet-infiltrates of NOD mice recognize insulin B:9-23 peptide [4] and human T cell reactivity to B:9-23 peptide are detected in type 1 diabetes patients [5]. Administration of insulin or insulin peptide through various routes (subcutaneous, oral, nasal and intravenous) to NOD mice prevents diabetes [6–9]. Those results indicate that insulin and the peptide exhibit either pathogenic or protective character depending on the situation probably through induction or regulation of autoreactive T cells. Rodents have two functional preproinsulin genes named insulin 1 and insulin 2 on different chromosome and have two amino acid differences in insulin B chain. Insulin 1 lacks intron present in insulin 2, suggesting that insulin 1 is a retroposon of insulin 2 gene generated by an RNA-mediated duplicationtransposition event [10]. We have created insulin 1 or insulin 2-KO NOD congenic mice and reported that dramatic opposite effects of KO genes were unexpectedly observed [11]. In this study, the KO effects were confirmed at further backcross generation with the breeding of NOD/ Shi/Kbe mice. Analyses for the mechanism of autoreactive T cells have been also performed.
2.3. Diagnoses of diabetes Homozygous KO mice were produced by the breeding of heterozygous mice at BC9 generation for insulin 1-KO and at BC12 for insulin 2-KO mice and the development of diabetes have been followed up. The glucose level for the homozygous KO mice was measured weekly with FreeStyle (TheraSense, Alameda, CA) or Glutest Ace (Sanwakagaku, Nagoya, Japan) blood glucose monitoring system. The mice were considered diabetic after two consecutive blood glucose values were greater than 250 mg/dl. After development of diabetes, the mice were sacrificed immediately and the pancreata of some mice were fixed in 10% buffered formalin for histological analysis. 2.4. Histology The pancreata obtained from the mice were fixed in 10% buffered formalin and then embedded in paraffin. Paraffin sections were stained with hematoxylin/eosin. Pancreatic sections were microscopically examined for the degree of insulitis. 2.5. Anti-insulin autoantibody assay
2. Materials and methods
Mice were bled for the measurement of serum anti-insulin autoantibody (IAA) levels. IAA was measured with a 96-well filtration plate micro-IAA assay as previously reported [2] and expressed as an index: index = (sample D (cpm) negative control D (cpm))/(positive D (cpm) negative control D (cpm)). A value of 0.01 or greater is considered positive.
2.1. Mice
2.6. Antigens
The original knockouts were produced in 129S1/SvImJ embryonal cell lines by Jami and coworkers [12]. The knockout cell lines were microinjected into C57BL/6 blastocytes and lines developed with C57BL/6. Insulin 1 and insulin 2 KO congenic NOD mice were established by breeding the original insulin KO mice using speed congenic method fixing NOD diabetogenic loci (idd 1–14) at Barbara Davis Center [13]. Both strains have been backcrossed onto NOD/Bdc mice and non-NOD genomic regions flanking each insulin gene knockout were then less than 10 cM. We have further been breeding the knockouts onto NOD/ Shi/Kbe mice in the Institute for Experimental Animals, Kobe University School of Medicine. We are at the 10th backcross (BC) generation for insulin 1-KO mice and the 13th BC for insulin 2-KO mice. All animals were handled under the Guidelines for Animal Experimentation of Kobe University School of Medicine.
HPLC-purified insulin 1B:9-23 peptide (PHLVEALYLVCGERG) and insulin 2B:9-23 peptide (SHLVEALYLVCGERG) were purchased from Invitrogen Japan (Tokyo, Japan). Purified recombinant human GAD65 expressed in yeast (Saccharomyces cerevisiae) was obtained from RSR Ltd. (Cardiff, UK). These peptides and the protein were used as antigens in vitro ELISPOT assay.
2.2. Genotype analysis Genomic DNA was extracted from mouse tails. Insulin KO mice were genotyped for KO genes and wild-type insulin genes by using PCR as previously described [13]. The PCR products were electrophoresed on 2% agarose gels and visualized by ethidium bromide staining.
2.7. ELISPOT assay The murine ELISPOT assay was performed with modification of human ELISPOT assay as previously described [14]. In brief, splenocytes (2 105) were cultured in nitro-cellulosebottomed 96-well microtiter plates (Millititer, Millipore Corp., Bedford, MA) in 200 ml of RPMI 1640 with 1% FCS. ELISPOT assay for mouse IFN-g or IL-4 was performed as manufacturer’s instruction (Mabtech AB, Stockholm, Sweden) and spots were analyzed with an ImmunoSpot Analyzer (Cellular Technology, Cleveland, OH). 2.8. Statistics Survival curves were analyzed with the log rank test at Kaplan–Meier method. Mann–Whitney U-test was used to compare the values of IAA expression. Wilcoxon matched
H. Moriyama et al. / Diabetes Research and Clinical Practice 77S (2007) S155–S160
S157
pair tests was applied to compare the number of ELISPOT counts.
3. Results Similar to the previous report [11], dramatic opposite effects of KO genes were observed by survival curve analysis (Fig. 1, P < 0.0001). The cumulative diabetic incidence of insulin 2-KO NOD mice at 30 weeks of age was 100% and most mice (25/29, 86%) developed overt diabetes by 17 weeks of age. Four mice became overt diabetes by 10 weeks of age with severe insulitis examined by histological analysis (data not shown). In contrast, no insulin 1-KO NOD mice developed overt diabetes although significant but rather mild insulitis were observed at later weeks of age. To further investigate anti-insulin autoimmuity in insulin-KO NOD mice, we examined the expression of serum insulin autoantibodies (IAA) of insulin 1-KO and insulin 2-KO NOD mice at a time between 10 and 15 weeks of age. As shown in Fig. 2, all of insulin 2-KO NOD mice (5/5) showed enhanced IAA values and the IAA level were much higher than that of wild-type NOD mice (P < 0.005). IAA of insulin 1-KO NOD mice seemed to be lower than that of wild-type NOD mice although two groups were not statistically different. We hypothesized that insulin 2 knockout might accelerate anti-insulin autoimmunity because of decreased tolerance to insulin peptides, especially to the important peptide, namely insulin B:9-23 peptide. We therefore evaluated T cell reactivity in splenocytes to islet-related autoantigens with ELISPOT assay. As shown in Fig. 3, the number of IFN-g spots to hGAD65 protein, insulin 1 B:9-23 peptide, and insulin 2 B:9-23 peptide were elevated in comparison to unstimualeted
Fig. 1. Life table analysis of insulin-KO NOD mice. Filled squares indicate insulin homozygous insulin 2-KO NOD mice at BC 12 generation (n = 29). Filled circles indicate homozygous insulin 1KO NOD at BC9 (n = 30) and filled triangles indicate wild-type NOD/ Shi/Kbe mice (n = 37). The three curves are dramatically different with log rank test (P < 0.0001).
Fig. 2. Serum anti-insulin autoantibody levels. IAA of homozygous insulin 2-KO, insulin 1-KO and wild-type NOD mice measured at a time between 10 and 15 weeks of age were plotted. Filled squares, filled circles and filled triangle indicate insulin 2-KO, insulin 1-KO, and wild-type NOD mice, respectively. IAA level of homozygous insulin 2-KO NOD mice are significantly higher than that of wild-type NOD mice with Mann–Whitney U-test (P < 0.005).
control wells. Statistic analysis with Wilcoxon matched pair test indicated that the number of IFN-g spots to insulin 1 B:9-23 peptide, beside other two antigens, were significantly higher than that of unstimulated control wells. On the other hand, any number of IL-4 spot to these three antigens was not increased by the antigenic stimulation. 4. Discussion In a previous study, we created insulin 1-KO and insulin 2-KO NOD congenic mice [13]. The insulin 1 knockout strongly prevented the progression of autoimmune diabetes in NOD mice whereas the insulin 2 knockout remarkably accelerated type 1 diabetes of NOD mice [11]. In this study, we have established insulin 1-KO and insulin 2-KO NOD colonies with NOD/Shi/Kbe mice. Further backcross onto NOD/Shi/ Kbe mice did not affect either the protective effect of insulin1-KO genes at BC9 or disease-enhancing effect of insulin 2-KO gene at BC12. Our results indicated that the effect of knockout genes was consistent even bred with the NOD mice in different subcolony. In other aspect, the environmental change (From Barbara Davis Center in Colorado to Kobe University in Japan) did not affect on the knockout genes effect. The insulin autoantibody (IAA) is an excellent predictive maker for the development of type 1 diabetes in both human and NOD mice and the presence of antiinsulin autoantibodies in NOD mice usually correlate with insulitis [15]. We have launched the measurement for mIAA in our laboratory with the same protocol at Babara Davis Center. Our mIAA assay indicated that insulin 2 knockout gene strongly enhanced the
S158
H. Moriyama et al. / Diabetes Research and Clinical Practice 77S (2007) S155–S160
Fig. 3. ELISPOT assay for IFN-g and IL-4 responses to GAD65 protein, insulin 1 B:9-23 peptide, and insulin 2 B:9-23 peptide in insulin 2-KO NOD mice. Individual spot formation of splenocytes from insulin 2-KO NOD mice is shown with the stimulation by GAD65 antigen or insulin peptides in the figure. Insulin 1 B:9-23-stimulated IFN-g responses are significantly observed compared with those in the absence of antigens (P < 0.05 by Wilcoxon matched pair test).
expression of serum anti-insulin autoantibodies, corresponding to the aggressive insulitis and acceleration of diabetes in insulin 2 KO-NOD mice. Polychronakos and Chentoufi demonstrated, using original insulin-KO mice with 129/C57BL/6 background genes, that thymic expression of insulin was related to the number of copies of the insulin 2 gene [16]. Similarly, insulin 2 KO-NOD mice also showed no insulin in thymus while insulin 1-KO NOD mice showed significant thymic insulin expression when detected using real-time PCR analysis [17]. We therefore, hypothesize that insulin 2-KO gene may accelerate type 1 diabetes because of diminished expression of insulin within the thymus and decreased tolerance to insulin epitopes. T cell reactivity in splenocytes of insulin 2-KO NOD mice to insulin B:9-23 peptide in this study is consistent with this hypothesis. In addition, protection from diabetes in the insulin 1-KO NOD mice is hypothesized to be due to removal of the more pathogenic insulin peptide sequence including B9-23 (B9 Proline v.s. Serine). Recently, the double insulin knockout mice lacking both the native insulin 1 and native insulin 2 genes have been established by the combination of insulin 1-KO, insulin 2-KO, and another NOD transgenic mouse with a muted preproinsulin gene to have alanine at position B16 rather than tyrosine of the B:9-23 sequence [18]. The double insulin-KO NOD mice that were metabolically rescued with a muted gene encoding preproinsulin, totally abrogated autoimmunity of NOD mice with prevention of anti-insulin autoantibodies, insulitis, and diabetes. The result from double insulin-KO NOD
mice suggests that the native insulin B:9-23 sequence is essential for anti-islet autoimmunity of the NOD mouse and the difference between insulin 1 B9-23 and insulin 2 B9-23 sequence might be crucial for the opposite outcomes of single insulin-KO NOD mice. Insulin B:9-23 peptide was discovered when Daniel et al. cloned islet-reactive CD4 T cells directly from islets of NOD mice by using whole islets as initial antigens [4]. Insulin B:9-23 has been identified that binds to both mouse I-Ag7 and human DQ8 [19]. Insulin B:15-23 recognized by CD8 T cells with Kd molecules is within insulin B:9-23 [20]. The B:9-23 reactive T cell clones and the B:15-23 reactive CD8 T cell clone (clone G9C8) are pathogenic in that those clones are able to transfer diabetes. In contrast to those pathogenic T cell clones, Zekzer et al. described that an T cell clone (clone 2H6) reacting insulin B:9-23 and B:12-25 could mediate disease protection [21]. Furthermore, Du et al. have demonstrated that 2H6 TCR transgenic NOD possessed insulin specific immunoregulatory T cells by virtue of their production of TGF-b1 similar to the parent clone [22]. Further evidence supporting a more pathogenic role of insulin 1 over insulin 2 peptide comes from experiments immunizing mice models with both forms of B:9-23 peptides. Subcutaneous injection of the insulin 2 B:9-23 but not the insulin 1 peptide significantly protects NOD mice from diabetes [23]. Treatment with B:9-23 immunization and the TLR3 ligand, poly:IC, induces diabetes in BALB/c mice having transgenic expression of costimulatory B7.1 molecules on pancreatic b cells, termed experimental
H. Moriyama et al. / Diabetes Research and Clinical Practice 77S (2007) S155–S160
autoimmune diabetes (EAD) [24]. In the EAD model, insulin 1 B:9-23 peptide immunization accelerated diabetes in the B7.1 mice compared with insulin 2 peptide [23]. One issue of obvious interest is the applicability of our observation to human subjects. It is possible that the most appropriate strategy for inducing tolerance in autoimmunity may be the combination of important autoantigen peptide with potent immunomodulatory agent. Recently, combination treatment of anti-CD3 and nasal proinsulin peptide can reverse recent-onset diabetes in two murine models with much higher efficacy than monotherapy with anti-CD3 or antigen alone [25]. However, immune response to key antigenic dominant shows complexity between the balance of induction of pathogenic cells and regulatory cells, often referred to as the ‘‘double-edged sword’’ [26]. Much effort is needed in studying basic pathogenesis with animal models as in carrying out human clinical trial to increase the probability that nontoxic antigen-specific therapies will be performed in the course of human type 1 diabetes. Acknowledgements We thank Ms. Atsumi Katsuta and Mr. Takeshi Hamada for the laboratory assistance. References [1] D. Hanahan, Peripheral-antigen-expressing cells in thymic medulla: factors in self-tolerance and autoimmunity, Curr. Opin. Immunol. 10 (1998) 656–662. [2] L. Yu, D.T. Robles, N. Abiru, P. Kaur, M. Rewers, K. Kelemen, et al., Early expression of antiinsulin autoantibodies of humans and the NOD mouse: evidence for early determination of subsequent diabetes, PNAS 97 (2000) 1701–1706. [3] E. Bonifacio, M. Atkinson, G. Eisenbarth, D. Serreze, T.W. Kay, E. Lee-Chan, et al., International Workshop on Lessons From Animal Models for Human Type 1 Diabetes: identification of insulin but not glutamic acid decarboxylase or IA-2 as specific autoantigens of humoral autoimmunity in nonobese diabetic mice, Diabetes 50 (2001) 2451–2458. [4] D. Daniel, R.G. Gill, N. Schloot, D. Wegmann, Epitope specificity, cytokine production profile and diabetogenic activity of insulin-specific T cell clones isolated from NOD mice, Eur. J. Immunol. 25 (1995) 1056–1062. [5] D.G. Alleva, P.D. Crowe, L. Jin, W.W. Kwok, N. Ling, M. Gottschalk, et al., A disease-associated cellular immune response in type 1 diabetics to an immunodominant epitope of insulin, J. Clin. Invest. 107 (2001) 173–180. [6] Z.J. Zhang, L. Davidson, G. Eisenbarth, H.L. Weiner, Suppression of diabetes in nonobese diabetic mice by oral administration of porcine insulin, Proc. Natl. Acad. Sci. USA 88 (1991) 10252–10256.
S159
[7] A. Muir, A. Peck, M. Clare-Salzler, Y.H. Song, J. Cornelius, R. Luchetta, et al., Insulin immunization of nonobese diabetic mice induces a protective insulitis characterized by diminished intraislet interferon-gamma transcription, J. Clin. Invest. 95 (1995) 628–634. [8] D. Daniel, D.R. Wegmann, Protection of nonobese diabetic mice from diabetes by intranasal or subcutaneous administration of insulin peptide B-(9-23), PNAS 93 (1996) 956–960. [9] P.R. Hutchings, A. Cooke, Comparative study of the protective effect afforded by intravenous administration of bovine or ovine insulin to young NOD mice, Diabetes 44 (1995) 906–910. [10] M.B. Soares, E. Schon, A. Henderson, S.K. Karathanasis, R. Cate, S. Zeitlin, et al., RNA-mediated gene duplication: the rat preproinsulin I gene is a functional retroposon, Mol. Cell Biol. 5 (1985) 2090–2103. [11] H. Moriyama, N. Abiru, J. Paronen, K. Sikora, E. Liu, D. Miao, et al., Evidence for a primary islet autoantigen (preproinsulin 1) for insulitis and diabetes in the nonobese diabetic mouse, Proc. Natl. Acad. Sci. USA 100 (2003) 10376–10381. [12] L. Deltour, P. Leduque, N. Blume, O. Madsen, P. Dubois, J. Jami, et al., Differential expression of the two nonallelic proinsulin genes in the developing mouse embryo, PNAS 90 (1993) 527– 531. [13] J. Paronen, H. Moriyama, N. Abiru, K. Sikora, E. Melanitou, S. Babu, et al., Establishing insulin 1 and insulin 2 knockout congenic strains on NOD genetic background, Ann. NY Acad. Sci. 1005 (2003) 205–210. [14] R. Kotani, M. Nagata, H. Moriyama, M. Nakayama, K. Yamada, S.A., et al. Detection of GAD65-reactive T-Cells in type 1 diabetes by immunoglobulin-free ELISPOT assays, Diabetes Care 25 (2002) 1390–1397. [15] D.T. Robles, G.S. Eisenbarth, N.J.M. Dailey, L.B. Peterson, L.S. Wicker, Insulin autoantibodies are associated with islet inflammation but not always related to diabetes progression in NOD congenic mice, Diabetes 52 (2003) 882–886. [16] A.A. Chentoufi, C. Polychronakos, Insulin expression levels in the thymus modulate insulin-specific autoreactive T-cell tolerance: the mechanism by which the IDDM2 locus may predispose to diabetes, Diabetes 51 (2002) 1383–1390. [17] M. Nakayama, N. Babaya, D. Miao, K. Sikora, J.F. Elliott, G.S. Eisenbarth, Thymic expression of mutated B16:A preproinsulin messenger RNA does not reverse acceleration of NOD diabetes associated with insulin 2 (thymic expressed insulin) knockout, J. Autoimmunol. 25 (2005) 193–198. [18] M. Nakayama, N. Abiru, H. Moriyama, N. Babaya, E. Liu, D. Miao, et al., Prime role for an insulin epitope in the development of type 1 diabetes in NOD mice, Nature 435 (2005) 220– 223. [19] K.H. Lee, K.W. Wucherpfennig, D.C. Wiley, Structure of a human insulin peptide-HLA-DQ8 complex and susceptibility to type 1 diabetes, Nat. Immunol. 2 (2001) 501–507. [20] F.S. Wong, J. Karttunen, C. Dumont, L. Wen, I. Visintin, I.M. Pilip, et al., Identification of an MHC class I-restricted autoantigen in type 1 diabetes by screening an organ-specific cDNA library, Nat. Med. 5 (1999) 1026–1031. [21] D. Zekzer, F.S. Wong, L. Wen, M. Altieri, T. Gurlo, H. von Grafenstein, et al., Inhibition of diabetes by an insulin-reactive CD4 T-cell clone in the nonobese diabetic mouse, Diabetes 46 (1997) 1124–1132. [22] W. Du, F.S. Wong, M.O. Li, J. Peng, H. Qi, R.A. Flavell, et al., TGF-b signaling is required for the function of insulin-reactive T regulatory cells, J. Clin. Invest. 116 (2006) 1360–1370.
S160
H. Moriyama et al. / Diabetes Research and Clinical Practice 77S (2007) S155–S160
[23] D. Devendra, J. Paronen, H. Moriyama, D. Miao, G.S. Eisenbarth, E. Liu, Differential immune response to B:9-23 insulin 1 and insulin 2 peptides in animal models of type 1 diabetes, J. Autoimmun. 23 (2004) 17–26. [24] H. Moriyama, L. Wen, N. Abiru, E. Liu, L. Yu, D. Miao, et al., Induction and acceleration of insulitis/diabetes in mice with a viral mimic (polyinosinic-polycytidylic acid) and an insulin selfpeptide, Proc. Natl. Acad. Sci. USA 99 (2002) 5539–5544.
[25] D. Bresson, L. Togher, E. Rodrigo, Y. Chen, J.A. Bluestone, K.C. Herold, et al., Anti-CD3 and nasal proinsulin combination therapy enhances remission from recent-onset autoimmune diabetes by inducing Tregs, J. Clin. Invest. 116 (2006) 1371– 1381. [26] A. Pugliese, Peptide-based treatment for autoimmune diseases: learning how to handle a double-edged sword, J. Clin. Invest. 111 (2003) 1280–1282.