ORIGINAL ARTICLE

Negativation of type 1 diabetes-associated autoantibodiesto glutamic acid decarboxylase and insulin in childrentreated with oral calcitriolDimitrios T. PAPADIMITRIOU,1,2 Chrisanthi MARAKAKI,1 Andreas FRETZAYAS,1

Polyxeni NICOLAIDOU1 and Anastasios PAPADIMITRIOU1

13rd Department of Pediatrics, Attikon University Hospital and 2Division of Pediatric Endocrinology, Athens Medical Center, Athens, Greece

Correspondence

Dimitris T. Papadimitriou, av. Kifissias 58,15125, Maroussi, Athens, Greece.Tel: +30 2103638536Fax: +30 2106106406Email: info@pedoendo.gr

Received: 28 October 2012; revised 16December 2012; accepted 19 December2012.

doi: 10.1111/1753-0407.12023

Abstract

Background: Based on recent knowledge of the possible involvement of1,25-dihydroxyvitamin D in the pathogenesis of type 1 diabetes (T1D) andthe results of its administration in animal models, we conducted a clinicaltrial by treating high-risk children, positive for T1D autoantibodies, with oralcalcitriol.Methods: The present prospective trial was performed on 12 children(1.5–13 years old) who were investigated for the potential risk of T1D becauseof an already diagnosed association of celiac disease and autoimmune thy-roiditis (four girls), autoimmune thyroiditis at a very young age (two girls, twoboys), a diagnosis of T1D in siblings (two boys), and impaired glucose toler-ance (IGT; one boy, one girl). Serum autoantibody levels, including islet cellautoantibodies, anti-glutamic acid decarboxylase (GAD) 65, insulin autoan-tibodies (IAA), and anti-tyrosine phosphatase, and markers of calciummetabolism were evaluated prior to and at 6-monthly intervals after theinitiation of 0.25 mg/day calcitriol for 1–3 years.Results: In all children, persistent negativation of the anti-GAD65 antibod-ies and IAA was observed within 0.4–2.1 years. Of the two children with IGT,the boy proved to have maturity onset diabetes of the young (MODY) 2,whereas the glycemic profile was normalized in the girl.Conclusions: Despite the small number of subjects and the absence of acontrol group in the present study, 0.25 mg/day calcitriol effectively negati-vates anti-GAD65 antibodies and IAA after a median time of 6 months. Thissimple, safe, and low-cost strategy may prove effective in the prevention ofT1D in the future.

Keywords: anti-glutamic acid decarboxylase 65 antibody, calcitriol, insulinautoantibodies, prevention trial, type 1 diabetes.

Significant findings of the study: Negativation of type 1 diabetes-associated autoantibodies to glutamic aciddecarboxylase and insulin in children treated with oral calcitriol.What this study adds: The study suggests the possibility of using vitamin D analogues in the prevention of type 1diabetes. To our knowledge, nothing similar has been tried so far in children.

bs_bs_banner

Journal of Diabetes 5 (2013) 344–348

344 © 2013 Wiley Publishing Asia Pty Ltd and Ruijin Hospital, Shanghai Jiaotong University School of Medicine

Introduction

Beyond its role in calcium metabolism, vitamin D haspleiotropic effects in a number of physiological func-tions, especially as a biological inhibitor in inflammatoryhyperactivity. Experimental studies have highlighted thepossibility of using vitamin D analogues in the preven-tion and treatment of several autoimmune diseases, suchas multiple sclerosis, type 1 diabetes (T1D), inflamma-tory bowel disease, rheumatoid arthritis, autoimmunethyroiditis and systemic lupus erythematosus.1

Approximately 80%–90% of circulating vitamin Doriginates from ultraviolet B-induced vitamin D produc-tion in the skin, with dietary vitamin D intake contribut-ing only a minor amount. Vitamin D from both sources ishydroxylated to 25-hydroxyvitamin D (25(OH)D3) in theliver and is then further hydroxylated by 1a-hydroxylaseto the most active vitamin D metabolite, namely 1,25-dihydroxyvitamin D3 (calcitriol) in the kidneys.2

Vitamin D is involved (directly and indirectly) in theregulation and differentiation of immune system cells.Both the adaptive (T and B lymphocyte function) andinnate immune (antimicrobial activity and antigen pre-sentation) systems are affected by calcitriol and its recep-tor.3 The global campaign for reducing exposure tosunlight has led to substantial reductions in the rates ofskin cancer, but has also inadvertently resulted in a pan-demic of vitamin D deficiency. Reduced total 25(OH)D3

levels, which are used to assess vitamin D status, havebeen noted in several autoimmune diseases, with vitaminD supplementation have therapeutic effects in animalmodels of these diseases.2

Vitamin D seems to have a significant role in T1D. Thepancreatic insulin-producing b-cells, as well as numerouscell types in the immune system, express the vitamin Dreceptor (VDR) and vitamin D-binding protein (DBP).In addition, individuals with T1D seem to have a higherincidence of hypovitaminosis D.4 Pharmacologic dosesof calcitriol have been used to prevent insulitis and T1Din non-obese diabetic (NOD) mice, as well as in othermodels of T1D, possibly via immune modulation as wellas direct effects on b-cell function.5

Because of this recent scientific data implying calcitriolin the pathogenesis of T1D and the positive results ofcalcitriol supplementation in animal models in vivo, weconducted a pilot clinical trial of daily supplementationwith 0.25 mg calcitriol, p.o., in children at high risk ofT1D who were positive for T1D autoantibodies.

Methods

In all, 12 children (seven girls, five boys; aged 1.5–13years at the time of presentation) at high risk of T1D

were investigated in the 3rd Department of Pediatrics,Attikon University Hospital, between 2006 and 2011.These children were deemed to be high risk because of analready diagnosed association of celiac disease andautoimmune thyroiditis (four girls), the onset of autoim-mune thyroiditis at a very young age (<3 years of age intwo girls and one boy; <5 years of age in one boy), adiagnosis of T1D in siblings (two boys), and impairedglucose tolerance (IGT; one boy, one girl).

Serum levels of islet cell autoantibodies (ICA), autoan-tibodies to glutamic acid decarboxylase (Gad 65) antigen(anti-GAD65), anti-insulin autoantibodies (IAA), andautoantibodies against protein tyrosine phosphatase IA2(anti-IA2) were determined prior to and at 6-monthlyintervals after the initiation of 0.25 mg/day calcitriol, p.o.Calcitriol was started immediately after the discoverythat the children were positive to the autoantibodies andwas continued for a period of 1–3 years. All measure-ments were performed using commercially availableELISA kits (EIA-1594, EIA-1910, and EIA-1593 forICA, anti-GAD65, and IAA respectively [DRG Diag-nostics, Marburg, Germany]; Medizym 3803 for anti-IA2 [Medipan, Berlin, Germany]). In children who werepositive for the autoantibodies, measurements of anti-body titers were verified within 1 month, prior to theinclusion to the study.

All children underwent detailed clinical examinationand laboratory investigation every 6 months. Further-more, calcium metabolism was evaluated before cal-citriol initiation and during follow-up; this involveddetermination of Ca, P, parathyroid hormone (PTH),alkaline phosphatase (ALP) and total 25(OH)D3 levels,as well as the Ca/creatinine (Cr) ratio in a 2-h morningurine sample.

Written informed consent was obtained from allparents. This trial was approved by the Ethics ResearchCommittee of Attikon University Hospital.

Results

For all children (aged 1.5–13 years at the time of inclu-sion in the study; median age 6.6 years) persistent nega-tivation of anti-GAD65 antibodies (n = 7) and IAA(n = 6) was observed within 0.4–2.1 years (median 0.6years) of the initiation of calcitriol treatment. The char-acteristics of the children in the present study are given inTable 1. One child was found to be positive for bothanti-GAD65 antibodies and IAA, but was negative forboth after calcitriol supplementation. None of the chil-dren were positive for ICA or IA2 autoantibodies. Of thetwo children with IGT, the boy finally proved to havematurity onset diabetes of the young (MODY) 2 (Case11), whereas the glycemic profile of the girl was normal-

D.T. PAPADIMITRIOU et al. Calcitriol and type 1 diabetes

© 2013 Wiley Publishing Asia Pty Ltd and Ruijin Hospital, Shanghai Jiaotong University School of Medicine 345

ized within 1 year. Calcitriol was discontinued 1 yearafter negativation of the autoantibodies had beenachieved. All 12 children are still being followed-up andnone has presented positive for autoantibodies or devel-oped any side effects from the calcitriol supplementation.Calcium metabolism remained within the normal rangefor age in all children. To diagnose hypercalciuria,special attention was paid to the calcium/Cr ratio inmorning 2-h urine sample.

In addition, children with Hashimoto’s disease orceliac disease were evaluated for autoantibody titersagainst thyroglobulin and thyroid peroxidase or anti-transglutaminase, respectively. Although there was nosignificant change in the thyroid autoantibodies, changesin anti-transglutaminase antibodies could not beassessed because the children were initiated on a gluten-free diet as soon as the diagnosis had been established.

Discussion

In the present pilot clinical trial, children positive forT1D-associated autoantibodies were treated with oralcalcitriol. To our knowledge, there has been no similarstudy thus far. There are certainly important limitationsof the present study, the major being the small number ofsubjects and the absence of a control group. However,our results indicate that the simple, safe, and low-costtreatment with calcitriol may be quite effective in theprevention of T1D, because it resulted in the negativa-tion of the T1D-associated autoantibodies in all high-risk children studied.

Autoantibodies against islet and b-cell antigens canappear early in life, although their pathogenetic role

remains unclear. Often, IAA are the first autoantibodiesto be detected in childhood. Some, but not all, childrenwho are positive to IAA develop anti-GAD65 and IA2autoantibodies.6 The risk of T1D is strongly associatedwith the number of antibodies that are positive. It hasbeen proposed that the age of islet autoantibodyappearance and mean levels of IAA, but not GAD65 orIA2 autoantibodies, can predict the age of onset ofT1D.7 In our cohort, most of the children included(except one) were positive to just one type of autoanti-body and they may not have developed diabetes at alater stage. Nevertheless, this does not reduce the sig-nificance of the observation of the negativation of theantibodies to which the children were positive. Tofurther assess this, we compared the antibody titers inthe children in the present study with those in childrennewly diagnosed with T1D during the same period inour department. Of 27 children (16 girls and 11 boys;1.8–16.6 years of age) with newly diagnosed T1D, eightwere not positive to any antibodies, three were positiveto only one antibody, nine were positive to two anti-bodies, and seven were positive to three antibodies.Eight children were positive for anti-GAD65 antibodies,10 children were positive for IAA, 11 children werepositive for ICA, and 13 children were positive for IA2antibodies. The median titer of anti-GAD65 antibodiesand IAA was 1.4 and 1.3, respectively, which did notdiffer significantly that in the children receiving cal-citriol supplementation.

In the present study, we followed the recommenda-tions issued in 2001 by the Immunology of DiabetesSociety,8 according to which strategies for full risk evalu-ation should include determination of at least three of

Table 1 Subject characteristics, inclusion criteria and results of the calcitriol prevention trial

Caseno. Gender Inclusion criteria

PositiveT1D Abs

Antibodytiter

Age at inclusion(years)

Time ofnegativation aftercalcitriol initiation(years)

Antibody titerafter calcitrioltreatment

Follow-up afterdiagnosis (years)

1 F Hashimoto’s and celiacdisease

IAA 1.3 9.3 0.9 0.80 2.82 F Anti-GAD 1.4 13.0 1.5 0.15 4.83 F Anti-GAD 2.0 7.8 2.1 0.80 4.04 F IAA 1.2 12.5 0.5 0.90 1.35 M Hashimoto’s <5 years old IAA 1.3 1.9 0.4 0.70 0.9

Anti-GAD 1.2 0.806 M IAA 1.6 5.4 0.5 0.60 4.57 F Anti-GAD 1.6 1.5 0.5 0.80 4.58 F Anti-GAD 1.4 3.5 0.5 0.55 4.59 M Sibling with T1D IAA 1.4 9.5 1.5 0.70 2.3

10 M IAA 1.4 3.5 0.7 1.00 1.711 M IGT Anti-GAD 1.6 5.1 1.7 0.90 4.912 F Anti-GAD 1.2 11.6 0.5 0.22 1.9

Normal levels for anti-glutamic acid decarboxylase (GAD) 65 and insulin autoantibodies (IAA) are <1 and <1.1 IU/mL, respectively.T1D, type 1 diabetes; Abs, antibodies; IGT, impaired glucose tolerance.

Calcitriol and type 1 diabetes D.T. PAPADIMITRIOU et al.

346 © 2013 Wiley Publishing Asia Pty Ltd and Ruijin Hospital, Shanghai Jiaotong University School of Medicine

the four best-established markers, namely IAA, ICA,and antibodies to GAD65 and IA-2. An additionalmajor autoantigen that was identified recently againstthe zinc transporter ZnT8 was not available for commer-cial use in Greece at the time of the study.9

The high predictive value of T1D autoantibodies infirst-degree relatives in a number of large prospectivestudies decreases in the general population, in which theincidence of T1D is lower. The Karlsburg T1D risk studyon schoolchildren,10 aiming to evaluate the predictivediagnostic value of diabetes-associated autoantibodies inthe general population, found that of a total of 11 840schoolchildren tested for all four autoantibodies, 6.9%were positive for a single autoantibody and that only0.7% were positive to multiple autoantibodies. The find-ings suggest that combined autoantibody screening inthe general population clearly identifies those at risk forinsulin-dependent diabetes.10 Nevertheless, autoantibod-ies can both appear and disappear in follow-upsamples,11 whereas a substantial number of patients whoare positive for autoantibodies do not develop thedisease.

Several studies suggest that vitamin D supplementa-tion in children reduces the risk of T1D. For example, theDiabetes Autoimmunity Study in the Young (DAISY)reported that the maternal dietary intake of vitamin Dwas significantly associated with a decreased risk of theappearance of islet autoantibodies in offspring, indepen-dent of the human leukocyte antigen (HLA) genotype,family history of T1D, the presence of gestational diabe-tes mellitus, and ethnicity.12 Increasing vitamin D intakeduring pregnancy reduces the incidence of islet autoanti-bodies in offspring13 and cod liver oil taken during thefirst year of life reportedly reduced the risk of childhood-onset T1D.14 A multicentre case-control study alsoshowed an association between vitamin D supplementa-tion in infancy and a decreased risk of T1D.15

Together, these findings suggest the possibility ofusing vitamin D analogues for the prevention of T1D,either by preventing vitamin D deficiency or via theimmunomodulatory effects of vitamin D. Published dataindicate that 1,25(OH)2D3 has potent direct effects on Bcell responses, inhibiting plasma cell differentiation andinducing the apoptosis of activated B cells in culture,16

and that there is an inverse correlation between serumtotal IgG and vitamin D levels.17 However, other studiesindicate that vitamin D does not significantly affect theantibody response to infections and is not immunosup-pressive.18 Supplementation with calcitriol after the onsetof T1D did not induce any sustainable beneficial effect inthe IMDIAB XIII trial.19

Importantly, despite all the beneficial effects ofvitamin D supplementation, it can have deleterious

effects if overdosed. Hypercalcemia, metastatic calcifica-tion, including tumoral calcinosis around the joints andvascular calcification, and renal impairment secondaryto hypercalcemia have all been reported.20 Acute LD50

studies have been performed on both mice and rats andindicate that the approximate oral lethal dose of cal-citriol ranges between 1.35 and 3.9 mg/kg (see http://www.rochecanada.com/fmfiles/re7234008/Research/ClinicalTrialsForms/Products/ConsumerInformation/MonographsandPublicAdvisories/Rocaltrol/rocaltrolpmE.pdf, accessed 7 February 2013). Thesevalues are several orders of magnitude higher thanthe proposed clinical dose of 0.25 mg once or twice daily(approximately 8–10 ng/kg per day).20

Antibody testing in early childhood can potentiallyidentify individuals destined to develop T1D, but data onthe predictive value of positive antibodies in the generalpopulation, from which 90% of new cases are derived,remain poor. Despite the abundance of clinical studies,no agent has yet been shown to be effective in preventingT1D. At the doses used in the present study, calcitriolwas effective in the negativation of T1D autoantibodies,although larger and preferably double-blind studies areneeded to confirm the results. With the increasing avail-ability (and at lower cost) of autoantibody screening forT1D and the promising ongoing research into the immu-nomodulatory effects of vitamin D analogues, we feelthat the prevention of T1D in the general populationmay be feasible in the future.

Disclosure

The authors declare no conflicts of interest.

References

1. Brown AJ, Slatopolsky E. Vitamin D analogs: Therapeu-tic applications and mechanisms for selectivity. MolAspects Med. 2008; 29: 433–52.

2. Holick MF. Vitamin D deficiency. N Engl J Med. 2007;357: 266–81.

3. Bikle DD. Vitamin D regulation of immune function.Vitam Horm. 2011; 86: 1–21.

4. Takiishi T, Gysemans C, Bouillon R, Mathieu C.Vitamin D and diabetes. Endocrinol Metab Clin NorthAm. 2010; 39: 419–46.

5. Mathieu C, Waer M, Laureys J, Rutgeerts O, Bouillon R.Prevention of autoimmune diabetes in NOD mice by1,25-dihydroxyvitamin D3. Diabetologia. 1994; 37: 552–8.

6. Ziegler AG, Hummel M, Schenker M, Bonifacio E.Autoantibody appearance and risk for development ofchildhood diabetes in offspring of parents with type 1diabetes: The 2-year analysis of the German BABYDIABstudy. Diabetes. 1999; 48: 460–8.

D.T. PAPADIMITRIOU et al. Calcitriol and type 1 diabetes

© 2013 Wiley Publishing Asia Pty Ltd and Ruijin Hospital, Shanghai Jiaotong University School of Medicine 347

7. Steck AK, Johnson K, Barriga KJ et al. Age of isletautoantibody appearance and mean levels of insulin, butnot GAD or IA-2 autoantibodies, predict age of diagno-sis of type 1 diabetes: Diabetes autoimmunity study in theyoung. Diabetes Care. 2011; 34: 1397–99.

8. Bingley PJ, Bonifacio E, Ziegler AG, Schatz DA,Atkinson MA, Eisenbarth GS. Immunology of DiabetesSociety Proposed guidelines on screening for risk of type1 diabetes. Diabetes Care. 2001; 24: 398.

9. Wenzlau JM, Juhl K, Yu L et al. The cation efflux trans-porter ZnT8 (Slc30A8) is a major autoantigen in humantype 1 diabetes. Proc Natl Acad Sci USA. 2007; 104:17040–45.

10. Schlosser M, Strebelow M, Rjasanowski I, Kerner W,Wassmuth R, Ziegler M. Prevalence of diabetes-associated autoantibodies in schoolchildren: The Karls-burg Type 1 Diabetes Risk Study. Ann N Y Acad Sci.2004; 1037: 114–17.

11. Christie MR, Roll U, Payton MA, Hatfield EC, ZieglerAG. Validity of screening for individuals at risk fortype i diabetes by combined analysis of antibodies torecombinant proteins. Diabetes Care. 1997; 20: 965–70.

12. Fronczak CM, Baron AE, Chase HP et al. In uterodietary exposures and risk of islet autoimmunity in chil-dren. Diabetes Care. 2003; 26: 3237–42.

13. Chiu KC, Chu A, Go VLW, Saad MF. HypovitaminosisD is associated with insulin resistance and b cell dysfunc-tion. Am J Clin Nutr. 2004; 79: 820–5.

14. Stene LC, Joner G. Use of cod liver oil during thefirst year of life is associated with lower risk ofchildhood-onset type 1 diabetes: A large, population-based, case-control study. Am J Clin Nutr. 2003; 78:1128–34.

15. The EURODIAB Substudy 2 Study Group. Vitamin Dsupplement in early childhood and risk for type I (insulin-dependent) diabetes mellitus. Diabetologia. 1999; 42:51–4.

16. Chen S, Sims GP, Chen XX, Gu YY, Chen S, Lipsky PE.Modulatory effects of 1,25-dihydroxyvitamin D3 onhuman B cell differentiation. J Immunol. 2007; 179: 1634–47.

17. Pincikova T, Nilsson K, Moen IE et al. Inverse relationbetween vitamin D and serum total immunoglobulin G inthe Scandinavian Cystic Fibrosis Nutritional Study. EurJ Clin Nutr. 2011; 65: 102–9.

18. Peelen E, Rijkers G, Meerveld-Eggink A et al. Relativelyhigh serum vitamin D levels do not impair the antibodyresponse to encapsulated bacteria. Eur J Clin MicrobiolInfect Dis. 2013; 32: 61–9.

19. Bizzarri C, Pitocco D, Napoli N et al. No protectiveeffect of calcitriol on beta-cell function in recent-onsettype 1 diabetes: The IMDIAB XIII trial. Diabetes Care.2010; 33: 1962–3.

20. Rukshana S, Knott C, Rees L. The virtues of vitamin D:But how much is too much? Pediatr Nephrol. 2010; 25:1607–20.

Calcitriol and type 1 diabetes D.T. PAPADIMITRIOU et al.

348 © 2013 Wiley Publishing Asia Pty Ltd and Ruijin Hospital, Shanghai Jiaotong University School of Medicine