IMMUNOLOGY OF DIABETES V

Persistence of Residual Beta Cells and IsletAutoimmunity during Increasing Durationof Diabetes in NOD Mice and ExperimentalApproaches toward Reversing New-Onset

Disease with Bioactive PeptidesShiva Reddy, Carlos Chun Ho Cheung, Ryan Chau Chia Chai,

and Jessica Astrid Rodrigues

School of Biological Sciences, The University of Auckland, Auckland, New Zealand

The precise fate of beta cells and the presence of islet infiltrates after onset of type 1 di-abetes have not yet been fully characterized. Recently we showed that in newly diabeticNOD mice an appreciable number of beta cells remain. This was also observed duringthe first 2 weeks of diabetes in NOD mice without treatment with insulin. However, themean number of beta cells per unit islet cross-sectional area decreased with increasingduration of disease. In contrast, glucagon and somatostatin cell numbers showed anincrease. The persistence of insulitis in several islets until 4 weeks of diabetes suggestsongoing beta cell autoimmunity over a protracted phase. Combined daily treatmentof newly diabetic NOD mice with epidermal growth factor (EGF) and gastrin for thefirst 14 days of diabetes resulted in temporary restoration of normoglycemia in 7 of 15mice. We speculate that the residual beta cells present soon after onset of diabetes mayrespond to experimental regeneration. Treatment of newly diabetic NOD mice withthe bioactive peptides EGF and gastrin resulted in partial and temporary reversal ofdiabetes. We propose that peptide therapies combined with other benign immunomod-ulatory approaches to rescue and preserve beta cells in the long term and to preventrecurring autoimmunity may be more effective than peptide therapy alone in reversingdiabetes in NOD mice.

Key words: beta cells; autoimmunity; peptide therapy; reversal of type 1 diabetes

Introduction

There has been renewed interest in deter-mining the number and fate of beta cells afteronset of type 1 diabetes and in the develop-ment of experimental approaches to stimulatebeta cell regeneration and its long-term preser-vation.1 Early pancreatic histopathologic stud-ies in humans showed that some beta cells re-main in many subjects with recent-onset type1 diabetes as well as in some with longstand-

Address for correspondence: Shiva Reddy, Ph.D., Senior Research Fel-low, School of Biological Sciences, The University of Auckland, PrivateBag 92019, Auckland, New Zealand. Voice: +64-9-3737599, ext. 82917;fax: +64-9-3737586. s.reddy@auckland.ac.nz

ing disease.2,3 The presence of beta cells hasalso been confirmed in a more recent studywhich examined the pancreas in recently di-agnosed human subjects.4 In an analysis in-volving 2,431 patients with long-standing dis-ease, approximately 15% and 33% of patientshad stimulated C-peptide levels greater than0.5 and 0.2–0.5 nmol/L, respectively, withinthe first 5 years of diagnosis.5 However, limitedaccess to suitable human pancreatic materialfrom subjects with recent-onset disease and atdefined time points during increasing durationof disease, and the scarcity of suitably preservedof pancreas, optimized for immunohistochem-ical and molecular analyses, have hamperedour understanding of the immunobiology of

Immunology of Diabetes V: Ann. N.Y. Acad. Sci. 1150: 171–176 (2008).doi: 10.1196/annals.1447.010 C© 2008 New York Academy of Sciences.

171

172 Annals of the New York Academy of Sciences

the disease during the clinical phase of type1 diabetes.

In the nonobese diabetic (NOD) mousemodel of human type 1 diabetes, preciseknowledge on residual beta cell mass, beta cellturnover, and the extent of beta cell autoimmu-nity after onset of disease is still lacking. Studiesby others have shown that in the NOD mousenot all beta cells have been killed at diagno-sis of disease.6 Recently, we analyzed pancre-ata from NOD mice at diagnosis, and at 1, 2,3, and 4 weeks thereafter using immunohisto-chemical tests, looking for the presence of in-sulin, glucagon, and somatostatin cells and glu-cose transporter-2 (glut2), and we correlatedour findings with the degree of insulitis andthe identity of islet immune cell phenotypes.7

In this recent study, insulitis, although variable,persisted after onset of diabetes but declinedas the duration of diabetes increased. Duringthis period, the number of beta cells also de-clined sharply, whereas glucagon and somato-statin cells increased. CD4 and CD8 T lym-phocytes and macrophages were seen withinthe islets after diabetes, irrespective of the pres-ence or absence of insulin immunoreactivecells.

Recent evidence suggests that a numberof bioactive peptides administered to newlydiabetic NOD mice or to normal mice ren-dered diabetic with alloxan leads to restora-tion of normoglycemia. For instance, short-term treatment of newly diabetic NOD micewith exendin-4, together with either anti-lymphocyte serum or lisofylline, results in long-term remission of type 1 diabetes.8,9 Restora-tion of normoglycemia was also achieved inalloxan-induced diabetic C57Bl6/J mice aftercontinuous release of combined gastrin andepidermal growth factor (EGF) for 1 week.10

Recent studies have shown that co-treatmentof newly diabetic NOD mice with gastrin andEGF for 14 days from diagnosis also leads tolong-term remission and significant resolutionof insulitis.11

Here, we have characterized the isletendocrine–immune cell pathologic axis in dia-

betic NOD mice at specific time points after on-set of disease by histochemical and immunohis-tochemical analysis. We then evaluated the ef-ficacy of combined treatment of newly diabeticNOD mice with EGF and gastrin in restoringlong-term normoglycemia.

Methods

Nonobese diabetic mice were available atthe University of Auckland’s Animal ResourcesUnit. The current rate of spontaneous diabetesis approximately 80% in females between theages of 80 and 250 days of our NOD mousecolony. Diabetes in NOD mice is defined asthe presence of a nonfasting blood glucosevalue >12 mM on three consecutive days.

After diagnosis of diabetes, mice were sac-rificed either immediately or after 1, 2, 3, and4 weeks of clinical disease (without insulin treat-ment; 5 mice per time point), and the pan-creas was prepared for histologic and immuno-histochemical testing.7 Sections were stainedby hemotoxylin and eosin to assess the de-gree of insulitis. Consecutive sections were ei-ther single or dual-immunostained for insulin,glucagon, somatostatin, and glut2, as reportedpreviously.7

Fifteen newly diagnosed NOD mice ofvarious ages (age range 83–207 days) wereco-injected intraperitoneally with human re-combinant EGF (1.2 μg/kg body weight;Roche Diagnostics, Mannheim, Germany) +synthetic rat gastrin (4 μg/kg body weight;Sigma Chemical Co., St. Louis, MO, USA)twice per day for the first 14 days after onsetof diabetes. Animals were monitored every al-ternate day for blood glucose levels by tail-veinpuncture and the samples read with the Accu-Check blood glucose meter. Animals were killedafter reversion to sustained hyperglycemia andthe pancreas was removed for histologic study.Three mice were killed from the treatmentgroup during normoglycemia (days 30, 33, and52), and the pancreas was studied immunohist-ochemically and by H&E staining.

Reddy et al.: Beta Cells after Onset of Diabetes in NOD Mice 173

Figure 1. Effects of combined treatment of newly diabetic NOD mice with gastrin andEGF, twice daily, on blood glucose levels in NOD mice: (A) four treated mice showing variableperiods of normoglycemia; (B) three remaining mice showing protracted periods of reversionto normoglycemia and in whom pancreatic histologic study was performed at days 30, 33and 52 of this phase. The vertical arrows indicate the frequency and duration of peptidetreatment.

Results

After administration of EGF + gastrin forthe first 14 days after diabetes, 7 of 15 miceshowed variable periods of suppression of bloodglucose, ranging from 14–67 days (Fig. 1Aand B). Peptide treatment in the remaining8 mice did not result in reversion to normo-glycemia (results not shown). The 7 mice withtemporary suppression of blood glucose also

showed intermittent episodes of hyperglycemia(Fig. 1). Three of the 7 mice were killed duringthe normoglycemic phase after treatment (days30, 33, and 52) for histochemical and immuno-histochemical assessment of pancreatic islets.The remaining 4 mice eventually reverted tohyperglycemia after variable periods of nor-moglycemia (Fig. 1A; 2 mice after 14 days,1 mouse after 22 days, and 1 mouse after67 days).

174 Annals of the New York Academy of Sciences

Figure 2. Photomicrographs from peptide-treated and untreated mice dual-labeled forinsulin (INS) and glucose transporter-2 (glut2): (A–C) peptide-treated animals showing isletimmunolabeled for insulin (A) and glut2 (B); (C) same islet as in (A,B) stained by H&E.Arrows in (A) and (B) point to insulin immunoreactive cells, which show an absence ofglut2 immunolabeling. (D-F) an islet from a peptide-untreated NOD mouse with 3 weeks ofdiabetes, stained for insulin (D) and glut2 (E); (F) the same islet as in (D,E) counterstained byH&E. Scale bar shown in (F) = 50 μm and applies to all photomicrographs.

Of the 7 treated mice that showed tem-porary remission from hyperglycemia, 5 micehad onset of diabetes ranging between 170and 207 days, with hyperglycemic values be-low 25 mM at onset. However, 7 of the 8 dia-betic NOD mice that failed to revert to normo-glycemia in the treatment group had an earlieronset of diabetes (83–157 days), with blood glu-cose values exceeding 25 mM at the beginningof peptide treatment.

Immunohistochemical analysis of pancreaticsections from the treated group of NOD micewith temporary remission showed the pres-ence of a considerable number of beta cellsin several islets, but with accompanying insuli-tis. There was an absence of glut2 in severalinsulin-positive cells (Fig. 2A–C). The islets ofuntreated NOD mice with 3 weeks of diabetesshowed a reduced number of beta cells, signifi-

cant insulitis, and an absence of glut2 immuno-labeling in almost all residual insulin-positivecells (Fig. 2D–F).

Discussion

We have recently shown that a significantnumber of beta cells remain in newly diabeticNOD mice.7 However, the number declineswith increasing duration of disease, with con-comitant increase in the number of glucagonand somatostatin cells. There has been an in-creased emphasis in developing and applyingnovel strategies to stimulate beta cell regen-eration at onset of type 1 diabetes. Success-ful beta cell regeneration and its protectionfrom autoimmune attack could lead to indepen-dence from exogenous insulin in human type

Reddy et al.: Beta Cells after Onset of Diabetes in NOD Mice 175

1 diabetic subjects. The presence of a “honey-moon phase,” which can last until 24 months,in about 43–56% of children with type 1 dia-betes suggests that at least some functional betacell mass remains during this phase.12,13 Thissubpopulation of beta cells may be amenableto experimental regeneration by suitable ther-apeutic ploys.

In the adult, most beta cells show only lowlevels of proliferation.14 However, in certainphysiological states, such as pregnancy associ-ated with insulin resistance or during obesity,there is an increase in beta cell mass.15,16 Themolecular mechanisms that underlie such betacell plasticity during these distinct physiologi-cal and pathophysiological states are unclear.Studies in rodents imply the role of certaintranscription factors in compensatory beta cellexpansion during insulin resistance.14

In type 1 diabetes, therapies aimed at betacell regeneration must take into account ongo-ing beta cell autoimmunity and some of theknown differences in the biochemical prop-erties between human and rodent beta cells.Previous studies have shown that the bioac-tive peptides, gastrin and EGF, when admin-istered to newly diabetic NOD mice, resultedin long-term reversal of disease. This was as-sociated with significant suppression of insulitisand preservation of beta cell mass.11 In ourpresent study, reversal of diabetes after admin-istration of combined peptide treatment wasobserved in approximately 50% of newly dia-betic NOD mice. However, this reversal was notlong-term. Several reasons may be attributed tothe differences in the studies of Suarez-Pinzonet al.11 and ours. In our study, we employed ratsynthetic gastrin, whereas Suarez-Pinzon et al.used human gastrin, which has a leucine substi-tution for methionine at position 15 to preventoxidation. Thus, differences in the source andpurity of gastrin and EGF employed in the twostudies, peptide stability, and biological potencymay be considered as likely reasons for the dis-crepant results. In addition, possible differencesin the ages of onset of diabetes in the two NODmouse colonies and the level of hyperglycemia

at onset of disease in mice in the treatmentstudies may have been contributing factorsas well.

In this study, the absence of glut2 expres-sion in several insulin-positive beta cells afteronset of diabetes implies that most survivingbeta cells fail to release insulin in response torising levels of blood glucose and, thus, theymay not be fully functional. Several studieshave suggested that hyperglycemia can impairbeta cell function on account of glucotoxicity,and it is likely that this dysfunction may bepresent in NOD mice during advanced au-toimmunity and hyperglycemia.17 Our previ-ous studies show an absence of glut2 expres-sion in several immunohistochemically positiveinsulin cells after onset of diabetes.7,18 This ob-servation is in contrast to a recent report thatshowed that in newly diabetic NOD mice, sev-eral beta cells express glut2, but are devoid ofinsulin.6 We ascribe this discordance in resultsto possible differences in antibody reactivityto insulin and/or glut2 and to the immuno-histochemical procedures employed in the twostudies.

In conclusion, the present studies show thata significant number of beta cells remain at di-agnosis of diabetes in the NOD mouse, butthey may not all be responsive to glucose.Combined treatment of newly diabetic NODmice with gastrin and EGF shows temporaryremission from diabetes. In the quest to re-verse diabetes in NOD mice, further stud-ies will be required to improve efficacy ofpeptide treatment and development of be-nign strategies to combat ongoing beta cellautoimmunity.

Acknowledgments

This study was supported by grants fromthe Auckland Medical Research Foundationand the New Zealand Child Health ResearchFoundation. We are grateful to Mr. VernonTintinger for maintaining the NOD mousecolony.

176 Annals of the New York Academy of Sciences

Conflicts of Interest

The authors declare no conflicts of interest.

References

1. Ablamunits, V., N.A. Sherry, J.A. Kushner & K.C.Herold. 2007. Autoimmunity and β cell regenerationin mouse and human type 1 diabetes: the peace isnot enough. Ann. N.Y. Acad. Sci. 1103: 19–32.

2. Foulis, A.K. & J.A. Stewart. 1984. The pancreas inrecent-onset type 1 (insulin-dependent) diabetes mel-litus: insulin content of islets, insulitis and associatedchanges in the exocrine acinar tissue. Diabetologia 26:456–461.

3. Foulis, A.K., C.N. Liddle, M.A. Farquharson, et al.

1986. The histopathology of the pancreas in type 1(insulin-dependent) diabetes mellitus: a 25-year re-view of deaths in patients under 20 years of age inthe United Kingdom. Diabetologia 29: 267–274.

4. Dotta, F., S. Censini, A.G. Van Halteren, et al. 2007.Coxackie B4 virus infection of β cells and naturalkiller cell insulitis in recent-onset type 1 diabetic pa-tients. Proc. Natl. Acad. Sci. USA 104: 5115–5120.

5. Palmer, J.P., G.A. Fleming , C.J. Greenbaum, et al.

2004. C-peptide is the appropriate outcome measurefor type 1 diabetes clinical trials to preserve β-cellfunction: report of an ADA Workshop, 21–22 Octo-ber 2001. Diabetes 53: 250–264.

6. Sherry, N.A., W. Chen, J.A. Kushner, et al. 2007.Exendin-4 improves reversal of diabetes in NODmice treated with anti-CD3 monoclonal antibodyby enhancing recovery of β-cells. Endocrinology 148:5136–5144.

7. Reddy, S., R.C.C. Chai, J.A. Rodrigues, et al. 2008.Presence of residual beta cells and co-existing islet au-toimmunity in the NOD mouse during long-standingdiabetes: a combined histochemical and immunohis-tochemical study. J. Mol. Hist. 39: 25–36.

8. Ogawa, N., J.F. List, J.F. Habener & T. Maki. 2004.Cure of overt diabetes in NOD mice by transienttreatment with anti-lymphocyte serum and exendin-4. Diabetes 53: 1700–1705.

9. Yang, Z., M. Chen, J.D. Carter, et al. 2006. Com-bined treatment with lisofylline and exendin-4 re-verses autoimmune diabetes. Biochem. Biophys. Res.

Commun. 344: 1017–1022.10. Rooman, I. & L. Bouwens. 2004. Combined gas-

trin and epidermal growth factor treatment inducesislet regeneration and restores normoglycaemia inC57Bl6/J mice treated with alloxan. Diabetologia 47:259–265.

11. Suarez-Pinzon, W.L., Y. Yan, R. Power, et al. 2005.Combination therapy with epidermal growth factorand gastrin increases β-cell mass and reverses hyper-glycaemia in diabetic NOD mice. Diabetes 54: 2596–2601.

12. Muhammad, B.J., P.G. Swift, N.T. Raymond & J.L.Botha. 1999. Partial remission phase of diabetes inchildren younger than age 10 years. Arch. Dis. Child.

80: 367–369.13. Bober, E., B. Dundar & A. Buyukgebiz. 2001. Partial

remission phase and metabolic control, in type 1 dia-betes mellitus in children and adolescents. J. Paediatr.

Endocrinol. Metab. 14: 435–441.14. Butler, P.C., J.J. Meier, A.E. Butler & A. Bhushan.

2007. The replication of β cells in normal physiology,in disease and for therapy. Nat. Clin. Prac. Endocrinol.

Metab. 3: 758–768.15. Green, I.C. & K.W. Taylor. 1972. Effects of preg-

nancy in the rat on the size and insulin secretoryresponse of the islets of Langerhans. J. Endocrinology

54: 317–325.16. Butler, A.E., J. Janson, S. Bonner-Weir, et al. 2003. β-

Cell deficit and increased β-cell apoptosis in humanswith type 2 diabetes. Diabetes 52: 102–110.

17. Thorens, B., Y-J. Wu, J.L. Leahey & G.C. Weir.1992. The loss of GLUT2 expression by glucose-unresponsive β cells of db/db mice is reversible andis induced by the diabetic environment. J. Clin. Invest.

90: 77–85.18. Reddy, S., M. Young, C.A. Poole & J.M. Ross.

1998. Loss of glucose transporter-2 precedes insulinloss in the nonobese diabetic and the low-dosestreptozotocin mouse models: a comparative im-munohistochemical study by light and confocal mi-croscopy. Gen. Comp. Endocrinol. 111: 9–19.