www.elsevier.com/locate/jcfJournal of Cystic Fibrosis 12 (2013) 803–805

Short Communication

ZnT8 antibodies in patients with cystic fibrosis: An expression of secondarybeta-cell damage?

Carla Bizzarri a,⁎, Germana Giannone b, Danila Benevento a, Enza Montemitro c,Federico Alghisi c, Marco Cappa a, Vincenzina Lucidi c

a Unit of Endocrinology and Diabetes, Bambino Gesù Children's Hospital, Rome, Italyb Department of Chemistry, Bambino Gesù Children's Hospital, Rome, Italyc Unit of Cystic Fibrosis, Bambino Gesù Children's Hospital, Rome, Italy

Received 10 December 2012; receive in revised 23 December 2012; accepted 4 March 2013Available online 25 March 2013

Abstract

Cystic Fibrosis-Related Diabetes (CFRD) is caused by a severe insulin deficiency with associated different degrees of insulin resistance. Dataconcerning the potential impact of autoimmunity are conflicting. Ninety subjects with cystic fibrosis (CF) were tested for glucose tolerance andautoantibodies against insulin (IAA), glutamic acid decarboxylase (GADA), protein tyrosine phosphatase (IA2) and zinc transporter 8 (Znt8A).Eighty-three subjects showed a normal glucose tolerance (92.2%), 6 subjects (6.6%) impaired glucose tolerance and 1 subject (1.1%) newlydiagnosed CFRD. Four subjects were found positive for both IAA and GADA (4.4%), one subject (1.1%) for both IA2 and GADA, and one subject(1.1%) for isolated GADA. Three subjects (3.3%) showed isolated ZnT8A positivity. ZnT8A positivity in CF patients is uncommon and notassociated with other autoantibodies. ZnT8A may not represent a specific indicator of a primary autoimmune beta-cell destruction, but possibly theexpression of a secondary damage of the pancreatic islets with autoantigen release.© 2013 European Cystic Fibrosis Society. Published by Elsevier B.V. All rights reserved.

Keywords: ZnT8; Cystic fibrosis; Diabetes; Autoimmunity

1. Introduction

Cystic Fibrosis-Related Diabetes (CFRD) is the most commonchronic complication associated with cystic fibrosis (CF). Theprevalence is low in children younger than 10 years and increasesprogressively during puberty, up to 50% in 30 year-old adults.The abnormal function of the chloride channels results in viscouspancreatic secretion, with obstruction of the exocrine pancreas,progressive fibrosis and fat infiltration, leading to disruption ofthe islets and derangement of insulin secretion [1,2]. CFRDshares common features with type 1 diabetes (T1D) and type 2diabetes (T2D). The primary defect is the progressive impairment

⁎ Corresponding author at: Bambino Gesù Children's Hospital, Unit oEndocrinology and Diabetes, P.zza S. Onofrio 4 00165, Rome, ItalyTel.: +39 06 68592508; fax: +39 06 68593374.

E-mail address: carla.bizzarri@opbg.net (C. Bizzarri).

1569-1993/$ -see front matter © 2013 European Cystic Fibrosis Society. Pubhttp://dx.doi.org/10.1016/j.jcf.2013.03.001

f.

lished

of insulin secretion, but different degrees of insulin resistancemay be associated [3,4]. The insensitivity to insulin has beenrelated to a primary hepatic insulin resistance with elevatedfasting and postprandial glucose production and to the pulmonaryinfections, with frequent use of glucocorticoids [2–4]. Insulinsecretion is linked to zinc transport mediated by the islet-specificzinc transporter ZnT8, a granule membrane protein encoded bythe SLC30A8 gene. Altered ZnT8 activity is supposed to beassociated with impaired glucose-induced insulin response [5–7].Binding of zinc to insulin controls the crystallization of thehormone, insulin storing in beta-cells and release after meals.ZnT8 has been identified both as a major autoantigen of T1D andas a genetic marker of T2D. Autoantibodies against ZnT8(ZnT8A) are positive in most T1D patients [5–7] and in a smallproportion of T2D patients. ZnT8A prevalence has been reportedto be 1.4%, in an Italian population of T2D patients negative forother T1D-related autoantibodies [8]. We analyzed ZnT8A in CF

by Elsevier B.V. All rights reserved.

Table 1Clinical and metabolic characteristics of antibody-positive and antibody-negative

804 C. Bizzarri et al. / Journal of Cystic Fibrosis 12 (2013) 803–805

patients, correlating ZnT8A positivity with clinical characteristicsand other T1D-related autoantibodies.

CF patients.

Antibodypositive

Antibodynegative

p

Number of patients 9 81Males/females 3/6 31/50 0.27Mean age (years) 15.9±3.3 18.07±7.9 0.3Body mass index[weight in kg/(height in m)2]

20.02±4.3 19.7±2.4 0.79

%FEV1 (percent forced expiratoryvolume in 1 s)

78.3±12.74 69.05±24.6 0.31

Fasting glucose (mg/dl) 87.3±12.9 84.9±8.6 0.52120 min glucose (mg/dl) 101.8±30.9 111.7±28.6 0.45Fasting insulin (mIU/ml) 7.6±4.5 8.2±1.58 0.56120 min insulin (mIU/ml) 34.9±22.2 35.7±21.4 0.93

Fisher's exact Test was used to analyze differences in gender distribution, differencesbetween means were analyzed by unpaired t test two-tailed.

2. Methods

Ninety consecutive CF subjects with pancreatic exocrineinsufficiency were tested for glucose tolerance and T1D-relatedautoantibodies, namely antibodies against insulin (IAA), glutamicacid decarboxylase (GADA), protein tyrosine phosphatase (IA2)and Znt8A. A standard oral glucose tolerance test (OGTT,1.75 g/kg body weight up to a maximum of 75 g) was performedwith flavored glucose (Glucosio Sclavo Diagnostics) after8–10 h overnight fast. Blood samples were obtained every30 min for 120 min for glucose and insulin measurements. Allpatients were tested during a period of clinical stability (definedas free from pulmonary exacerbations in the preceding 6 weeks).Glucose tolerance was classified according to the criteria of theAmerican Diabetes Association [9]. Subjects with already knownCFRD were excluded. All patients were Caucasians of Italianancestry. Written informed consent was obtained from parentsand/or patients. The Local Scientific Committee approved theprotocol.

GADA were analyzed by direct radioligand assay (Centak®anti-GAD65, Medipan, Dahlewitz, Germany). Lower detectionlimit was 0.6 U/ml. The cut off for positivity was 1 U/ml. IAAwere analyzed by a semi quantitative radioimmunoassay (AIA-100, DIAsource, Nivelles, Belgium), normal reference interval wasestablished by measurement in 100 healthy subjects that neverreceived insulin therapy. A binding percentage higher than meanvalue + 3 SD (8.2%) was considered as the cut off for positivity.IA2 were investigated by a direct radioligand assay (Centak®anti-IA2,Medipan, Selchow,Germany). Lower detection limit was0.7 U/ml. The cut off for positivity was 1.1 U/ml. Quantitativedetermination of ZnT8A in serum was performed by ELISAcommercial kits (RSR limited – Cardiff – United Kingdom). Theupper limit for ZnT8A positivity was 15 U/ml, that was the 99thpercentile of 100 healthy subjects. In the 2010 Diabetes Auto-antibody Standardization ProgramWorkshop, assay sensitivity andspecificity were 68% and 99%, respectively. Blood glucose wasmeasured immediately after sampling bymodified glucose oxidasemethod (Beckman Coulter, Brea, CA, USA). Insulin wasdetermined by chemiluminescent immunometric assay using acommercial kit (Immulite 2000, DPC, Los Angeles, CA, USA).

Table 2Characteristics of CF patients with positive ZnT8A.

Patient 1 Patient 2 Patient 3

Age (years) 20 19 13Gender Male Male FemaleCF mutation DF508/DF508 N1303K/N1303K DF508/DF508GADA (U/ml) 0.2 0.8 0.2IAA (%) 6 7 7IA2 (U/ml) 0.6 0.3 0.1ZnT8 (U/ml) 292.3 55.8 20.4

3. Results

Table 1 shows the characteristics of our study group. Eighty-three subjects (92.2%) showed a normal glucose tolerance, 6subjects (6.6%) impaired glucose tolerance and 1 subject (1.1%)newly diagnosed CFRD. Four subjects were found positive forboth IAA and GADA (4.4%), 1 subject (1.1%) for both IA2 andGADA, 1 subject (1.1%) for isolated GADA, while 3 subjects(3.3%) showed isolated ZnT8A positivity. Family history of T1Dand T2D was negative in all antibody-positive patients andglucose tolerance was normal in 8/9 (88.8%). ZnT8A positivesubjects (Table 2) showed a normal glucose tolerance.

4. Discussion

In CFRD, the correlation between the degree of islet damageand clinical diabetes is poor, leading to the hypothesis that thereare other factors causing diabetes [2]. Previous data concerning thepotential impact of autoimmunity are conflicting. The firstpublished study [10] analyzed GADA and IA2 in 30 CF patients,compared to 30 newly diagnosed T1D patients and 30 normalsubjects. GADA were found positive in 50% of the CF patients,93% of the T1D patients and 0% of the controls. IA2 antibodieswere detected in 40% of the CF group, 93% of the T1D group and0% of the controls. Among the fifteen CF patients with positiveGADA and IA2, four were affected by overt CRFD and fiveshowed a prediabetic state. A few years later, an Italian study [11]did not confirm the described high prevalence, reporting IA2positivity in 0% of CF patients and GADA positivity in 12.5%. Nodifferences were observed between subjects with or withoutCFRD. Autoreactive T lymphocytes are considered crucial forT1D pathogenesis. Cellular immune responses to the beta-cellautoantigen glutamic acid decarboxylase (GAD) was analyzed inCF patients with or without CFRD, compared to T1D patients andhealthy subjects [12]. Response to GAD was observed in 7/8 CFpatients and in all T1D patients. Post-stimulation production ofIFN-gamma and TNF-beta was observed in 2/4 CFRD, 1/4 CFpatients without CFRD and in 7/8 T1D patients. All these patientscarried anHLA-DQ genotype associated with T1D. A recent study[13] analyzed autoantibodies and HLA haplotypes associated withT1D in 76 patients with overt CFRD. GADA and/or IA2 were

805C. Bizzarri et al. / Journal of Cystic Fibrosis 12 (2013) 803–805

found positive in 5% of patients, IAA in 32% (probably inducedby insulin administration). No patient was found specificallypositive for ZnT8A. T1D-associated HLA haplotypes were foundin 36% of CF patients, compared to 47% of the general populationand more than 90% of T1D patients.

Considering that islet-autoantibodies appear during the pre-symptomatic phase of T1D, before the onset of hyperglycemia,we studied CF patients without a previous diagnosis of CFRD.The three patients positive for isolated ZnT8A showed a normalglucose tolerance. This finding may suggest that ZnT8Arepresent early markers of beta-cell damage, disappearing withthe progression of islet disruption. On the other hand, ZnT8Apositivity appears uncommon and nonspecific, i.e. not associatedwith other T1D autoantibodies or particular clinical characteris-tics. Nearly 2% of the general population has been foundpositive for one of IA2, GADA, IAA, or ZnT8A [14], whilethe specific prevalence of ZnT8A is already unknown. So, wecannot assert that the proportion of Znt8A positive CF patientsis really significantly elevated. ZnT8 is a secretory granulemembrane protein that is probably released when beta-cellsare being damaged. As supposed in T2D patients negativefor other autoantibodies [8], ZnT8A in CF patients may notrepresent a specific indicator of the activation of a primaryautoimmune beta-cell destruction, but possibly the expressionof a secondary damage of the pancreatic islets with autoantigenrelease.

Conflict of interest

Nothing to be declared.

Acknowledgments

We greatly appreciate the secretarial assistance of Mrs. LucianaLuciani.

References

[1] Waugh N, Royle P, Craigie I, Ho V, Pandit L, Ewings P, et al. Screeningfor cystic fibrosis-related diabetes: a systematic review. Health TechnolAssess 2012;16:1–180.

[2] Moran A, Becker D, Casella SJ, Gottlieb PA, Kirkman MS, Marshall BC,et al. Epidemiology, pathophysiology, and prognostic implications of cysticfibrosis-related diabetes: a technical review. Diabetes Care 2010;33:2677–83.

[3] Brennan AL, Geddes DM, Gyi KM, Baker EH. Clinical importance ofcystic fibrosis-related diabetes. J Cyst Fibros 2004;3:209–22.

[4] Bremer AA. The use of insulin in cystic fibrosis-related diabetes: aphilosophical difference from the use of insulin in Type 1 and Type 2diabetes. Pediatr Pulmonol 2012;47:1039–141.

[5] Rungby J. Zinc, zinc transporters and diabetes. Diabetologia 2010;53:1549–51.[6] Lefebvre B, Vandewalle B, Balavoine AS, Queniat G, Moerman E,

Vantyghem MC, et al. Regulation and functional effects of ZNT8 inhuman pancreatic islets. J Endocrinol 2012;214:225–32.

[7] Wijesekara N, Chimienti F, Wheeler MB. Zinc, a regulator of islet functionand glucose homeostasis. Diabetes Obes Metab 2009;11(Suppl. 4):202–14.

[8] Lampasona V, Petrone A, Tiberti C, Capizzi M, Spoletini M, di Pietro S, et al.Non Insulin Requiring Autoimmune Diabetes (NIRAD) Study Group. Zinctransporter 8 antibodies complement GAD and IA-2 antibodies in theidentification and characterization of adult-onset autoimmune diabetes: NonInsulin Requiring Autoimmune Diabetes (NIRAD) 4. Diabetes Care 2010;33:104–8.

[9] Expert Committee on the Diagnosis, Classification of Diabetes Mellitus.Report of the expert committee on the diagnosis and classification ofdiabetes mellitus. Diabetes Care 2003;26:S5–S20.

[10] Nousia-Arvanitakis S, Galli-Tsinopoulou A, Dracoulacos D, KaramouzisM, Demitriadou A. Islet autoantibodies and insulin dependent diabetesmellitus in cystic fibrosis. J Pediatr Endocrinol Metab 2000;13:319–24.

[11] Minicucci L, Cotellessa M, Pittaluga L, Minuto N, d'Annunzio G,Avanzini MA, et al. Beta-cell autoantibodies and diabetes mellitus familyhistory in cystic fibrosis. J Pediatr Endocrinol Metab 2005;18:755–60.

[12] Stechova K, Kolouskova S, Sumnik Z, CinekO, KverkaM, FaresjoMK, et al.Anti-GAD65 reactive peripheral blood mononuclear cells in the pathogenesisof cystic fibrosis related diabetes mellitus. Autoimmunity 2005;38:319–23.

[13] Gottlieb PA, Yu L, Babu S, Wenzlau J, Bellin M, Frohnert BI, et al. Norelation between cystic fibrosis-related diabetes and type 1 diabetesautoimmunity. Diabetes Care 2012;35:e57.

[14] Yu L, Herold K, Krause-Steinrauf H, McGee PL, Bundy B, Pugliese A, et al.Rituximab selectively suppresses specific islet antibodies. Diabetes 2011;60:2560–5.