DIABETES/METABOLISM RESEARCH AND REVIEWS RESEARCH Diabetes Metab Res Rev 2011; 27: 609–613. Published online in Wiley Online Library (wileyonlinelibrary.com) DOI: 10.1002/dmrr.1214
ARTICLE
Blue eyes as a risk factor for type 1 diabetes
Enrico Di Stasio1† Daria Maggi2† Enzo Berardesca3 Georgiana Clare Marulli3 Carla Bizzarri4 Angelo Lauria2 Rosalba Portuesi2 Maria Gisella Cavallo5 Francesco Costantino6 Raffaella Buzzetti5 Elisa Astorri2 Dario Pitocco7 Marco Songini8 Paolo Pozzilli2,9∗ The IMDIAB Group‡ 1 Institute of Biochemistry and
Clinical Biochemistry, Catholic University of Sacred Hearth, Rome, Italy 2 Department of Endocrinology and Diabetes, University Campus Bio-Medico, Rome, Italy 3 Department of Clinical Dermatology, ‘San Gallicano’ IRCCS, Rome, Italy 4 Endocrinology and Diabetes, ‘Bambino Ges` u’ Hospital, Rome, Italy 5 Department of Internal Medicine and Medical Specialities Polo Pontino, ‘Sapienza’ University, Rome, Italy 6 Department of Pediatrics, ‘Sapienza’ University, Rome, Italy 7 Diabetology, Catholic University of Sacred Hearth, Rome, Italy 8 Diabetology, ‘Brotzu’ Hospital, Cagliari, Italy 9 Centre for Diabetes, St Bartholomew’s and the Royal London School of Medicine, Queen Mary, University of London, London, UK
Abstract Background A high frequency of blue eyes and fair skin are reported in northern European Caucasians with type 1 diabetes (T1D). Also there is an inverse relationship between latitude and T1D incidence. We determined whether iris colour and skin pigmentation are risk factors in a Caucasian population living in two Mediterranean regions located at the same latitude with higher ultraviolet B irradiance, but with different T1D incidence. Methods We studied iris colour in 281 consecutive subjects with T1D and 298 controls. Skin type was evaluated by melanin quantification. Results In Lazio, blue eyes and fair skin type are significantly more common in T1D subjects than in controls (21 versus 9%, p = 0.002; 50 versus 35%, p < 0.001, respectively). In Sardinia, the frequency of blue eyes in T1D subjects is twice that in controls (5.8 versus 2.6% and significantly higher when compared to the expected calculated frequency in the entire population). By logistic regression analysis, only blue eyes are independent and significant predictors of T1D [odds ratio for blue eyes = 2.2; 95% confidence interval (1.1–4.4), p = 0.019]. Conclusions As previously shown in a Caucasian population from northern Europe, blue eyes and a trend for fair skin increase the risk for T1D also in a Caucasian population born and residing in a Mediterranean region (Continental Italy). This finding may be relevant for explaining different T1D incidence as prevalence of blue eyes differ substantially between northern and southern European Caucasians. Copyright 2011 John Wiley & Sons, Ltd. Keywords risk
type 1 diabetes; latitude; vitamin D; skin type; type 1 diabetes
Introduction
∗
Correspondence to: Paolo Pozzilli, Department of Endocrinology and Diabetes, University Campus Bio-Medico, 00128 Rome, Italy E-mail:
[email protected] † Dr
Di Stasio and Dr Maggi contributed equally to the preparation of this article. ‡ See Appendix for members of IMDIAB group. Received : 27 December 2010 Revised: 15 March 2011 Accepted: 26 April 2011
Copyright 2011 John Wiley & Sons, Ltd.
Type 1 diabetes (T1D) is a polygenic, autoimmune and multifactorial disease characterized by the destruction of pancreatic β cells. The genetic component of T1D is largely associated with human leucocyte antigen system (HLA) genes, with other minor susceptibility genes also being involved [1]. Predisposing environmental factors include, amongst the most relevant ones, early introduction of cow’s milk, reduced sunlight exposure and vitamin D intake [2–4]. Significant geographic differences have been reported in T1D prevalence, with higher prevalence in Nordic countries where ultraviolet B (UVB) irradiance is less pronounced than in southern countries [3]. Also, a seasonal variation in the months of T1D diagnosis has been described [5], with more cases diagnosed during winter than in summer, suggesting an important role for sunlight and vitamin D as immunomodulatory agents by virtue of the natural suppression of the autoimmune response against β cells exerted by this agent [6–8].
610
Furthermore, an increased risk for T1D is associated with a lower average yearly temperature, increasing distance from the Equator and Caucasian ethnicity [9]. It was in 1930 that a study carried out in a T1D population from the region that is today the Czeck Republic reported that subjects with T1D had a particular phenotype characterized by the presence of a blue iris [10]. Nearly 60 years later, Ziegler et al. evaluated the phototype and the sensitivity to UVB irradiance in German T1D subjects confirming the higher prevalence of lowpigment skin type in affected individuals compared with that in healthy subjects born and resident in the same geographical area [11]. On the basis of these premises, the aim of this study was to determine whether eye colour and skin type are risk factors for T1D also in a Caucasian population born and resident in a Mediterranean area including continental Italy and Sardinia where disease incidence is fourfold different [12] despite the fact that these regions are located at the same latitude.
Materials and methods We studied a Caucasian population of 178 consecutive unselected T1D (80 females, 98 males, mean age 33 ± 14 years) from the Lazio region of continental Italy and 103 T1D (32 females, 71 males, mean age 33 ± 5 years) from Sardinia. As control group we studied 181 (93 females, 88 males, mean age 36 ± 16 years) and 117 (67 females 50 males, mean age 37 ± 9 years) normal subjects with no family history of diabetes living and resident since birth in Lazio and Sardinia, respectively. In all participating centres, dedicated specific outpatient clinics are available for TID patients only, thus allowing a very easy unbiased and unselected participation of patients in this study. For control subjects, consecutive cases (excluding diabetic patients) admitted to hospital for routine blood tests were included in the analysis. Patients and controls were homogeneous in terms of genetic origin in that they were all southern Europeans. This study was approved by the Ethical Committee of the University Campus Bio-Medico within the framework of the observational studies of the IMDIAB group and informed consent was signed by participants. Individual iris colour was evaluated in daylight and classified into blue iris versus all other iris colour types. Categories of iris colour were differentiated on the basis of predominant colour (blue, green or brown). We have also evaluated individual skin type by developing a short and modified version of a standardized questionnaire [11] to examine iris colour, hair colour (blond, red, brown, and black) and burning history (ability to tan, risk of burns). At least two observers evaluated iris colour types, and concordance in judgement was obtained in nearly all cases (κ index = 0.88). A portable reflectance instrument, Mexameter 18 (Courage & Khazaka, Cologne, Germany), was used for the Copyright 2011 John Wiley & Sons, Ltd.
E. Di Stasio et al.
Table 1. Different skin types Skin type I II III IV V VI
Characteristic Celtic type: always burns easily and never tans White Caucasian: always burns easily and tans minimally with difficulty Mixed: burns minimally and tans gradually and uniformly Mediterranean: burns minimally and tans profusely Asiatic-Indian skin Black skin
Melanin index 0–150 50–250 100–350 150–500 150–650 600–999
quantification of skin melanin only in subjects from the Lazio region. The instrument is characterized by a digital readout narrow-band spectrophotometer. The light sources are three light-emitting diodes with narrow bands of emitted wavelengths: green λ = 568 ± 3 nm; red λ = 660 ± 3 nm; ultra-red λ = 870 ± 10 nm. A photo detector measures the light by the skin, and the reflectance in the bands is transformed into erythema and melanin indexes and expressed in numeric values from 0 to 999. Measurements were performed on the ventral aspect of the forearm. We then correlated phenotype characteristics with the melanin index (using the Mexameter 18 for the quantification of skin melanin). Accordingly, patients were classified as fair (Celtic or Caucasian) and tanned skin type (Mediterranean or mixed) (Table 1). For statistical data analysis, continuous variables were expressed as mean ± standard deviation, categorical variables displayed as frequencies and the appropriate parametric or nonparametric test was used to assess significance of the differences between subgroups. For sample size calculation we used data from a pilot study on 25 controls and patients. A sample size of 157 subjects in each group was estimated for 80% power and α 0.05. A significant difference in the frequency of blue eyes was defined as the primary endpoint. Moreover, occurrence of eye colour, age and gender on T1D incidence was analysed by multiple logistic regression to estimate odds ratio (OR) and confidence intervals (CIs) for correlation coefficient.
Results In the Lazio region, blue eyes compared with all other eyes colour types are significantly more common in T1D subjects than in controls (21.3 versus 9.4%, p = 0.002), and the relative risk for a subject with blue iris to belong to the T1D group is 2.6 (95% CI 1.4–4.8). In Sardinia, despite the presence of blue eyes being rare, their frequency in T1D subjects is twice that of controls (5.8 versus 2.6%). On the basis of data on the frequency of blue irises in subjects with T1D versus controls and on T1D prevalence in Sardinia (0.3% data from conscription physical examination), the presence of blue eyes in T1D Diabetes Metab Res Rev 2011; 27: 609–613. DOI: 10.1002/dmrr
611
Blue Eyes and Type 1 Diabetes
subjects, although a rare finding in this population in general, is significantly higher compared to the expected calculated frequency in the entire population (p = 0.037). Finally, considering the entire population under study (n = 579), the presence of blue eyes is strongly associated with T1D (15.7 versus 6.7%, p = 0.001, OR = 2.6, 95% CI 1.5–4.5). Table 2 gives the frequency distribution of skin type by eye colour in subjects from Lazio. The most common skin type is the mixed one (52.1%), whereby mixed skin type and brown eyes shows the highest frequency (48.2%). As expected, the presence of blue eyes is strongly associated with fair phenotype. In Lazio, fair skin is significantly more prevalent in T1D subjects than in controls (50 versus 35%, p < 0.001; Figure 1). The OR for susceptibility to T1D in patients with fair skin is 1.8 (95% CI 1.2–2.8, p = 0.007). Interestingly, introducing gender and age as potential confounding factors in the multivariate logistic regression model (Table 3), we confirmed the presence of blue eyes, excluding fair skin, as an independent and significant predictor of T1D (OR = 2.25, 95% CI 1.15–4.43, p = 0.019).
Table 2. Frequency distribution of phototype stratified by eye colour in subjects from the Lazio (n = 359) Skin type Iris colour
Celtic (%)
Caucasian (%)
Mixed (%)
Mediterranean (%)
Brown Green Blue
0.6 0.8 3.1
20.1 7.5 10.6
48.2 2.5 1.4
5.0 0.0 0.3
Total
4.5
38.2
52.1
5.3
Figure 1. Skin type distribution for type 1 diabetic subjects and controls from Lazio region. Odds ratio for fair skin 1.8 (95% confidence interval 1.2–2.82, p = 0.007) Copyright 2011 John Wiley & Sons, Ltd.
Finally, the calculated population attributable risk on the entire population under study is 23% for blue eyes, 14% for fair skin and 35% for blue eyes plus fair skin.
Discussion This study showed that in subjects with T1D living either in continental Italy (Lazio) or in Sardinia the presence of blue eyes is significantly more common in T1D than in non-diabetic subjects. Moreover in Lazio, the fair skin fenotype is also more frequent in T1D patients. However, correcting the results for age and gender, only blue eyes are associated with T1D with a relative risk of 2.25. These figures are comparable in terms of relative risk to those reported for genes associated with T1D [1]. Genetic susceptibility to T1D is mainly determined by HLA class II genes but also by a number of other genes outside this region that have been shown to influence the expression of the disease [1]. In the light of our results, the question arises why blue eyes and fair skin contribute to increased risk for T1D. The significantly increased prevalence of these traits in T1D may be related to genes involved in the control of their transmission that may also affect disease susceptibility. Variable iris pigmentation in healthy humans is the result of differential melanin deposition in iris melanocytes [13]. Inheritance has been shown to be polygenic, with at least eight genes involved in the transmission of the trait. Similarly, skin pigmentation is a polygenic quantitative trait and different genetic determinants of normal skin pigmentation variation have been reported in recent years. The majority of these genes, including encoding melanocortin 1 receptor, oculocutaneous albinism II protein, thyrosinase, membrane-associated transporter protein and agouti signalling protein, are involved in the control of melanin production and distribution, and some of them contribute to the control of both eye colour and skin pigmentation variation [14]. With regard to environmental factors, exposure to UVB irradiance is certainly the most important factor influencing evolutionary selection on human pigmentation. It may be possible that melanin itself may predispose to autoimmunity besides genetic inheritance. Melanocytes are immunocompetent cells acting as important components of the skin immune system. Upon stimulation, melanocytes can express major histocompatibility complex (MHC) molecules and other surface markers such as inter-cellular adhesion molecule 1 (ICAM-1) and CD40, and produce different cytokines [15]. In addition, melanocytes are capable of processing and presenting antigens to T cells [16] and the cooperation between different skin-immune competent cell types is known to play an important role in the protection against infectious pathogens. It is interesting to note that differences in skin pigmentation and sunlight exposure in different races and countries may explain the geographical distribution of T1D characterized by a north–south gradient in disease incidence [9]. Diabetes Metab Res Rev 2011; 27: 609–613. DOI: 10.1002/dmrr
612
E. Di Stasio et al.
Table 3. Multivariate logistic regression analysis. Data refer to type 1 diabetic subjects and controls from Lazio (n = 359)
Variable Eye colour (blue versus green/brown) Age (years) Gender (male versus female) Skin type (Celtic/Caucasian versus Mediterranean/Mixed)
β
Standard error
p
0.813 −0.024 0.281 0.420
0.345 0.008 0.219 0.239
0.019 0.002 0.198 0.078
Odds ratio (95% confidence interval) 2.25 (1.15–4.43) 0.98 (0.96–0.99) 1.32 (0.86–2.03) 1.52 (0.95–2.43)
Statistically significant (p = 0.019; p = 0.002) are in bold.
A similar distribution with increased incidence at higher latitudes is observed also for other autoimmune diseases such as multiple sclerosis, Crohn disease, rheumatoid arthritis and lupus erythematosus [17]. A relevant role in this context is played by vitamin D as an immunomodulatory agent. It has been reported that subjects with fair skin type, despite an increased production of vitamin D, may have reduced circulating levels compared with subjects having dark skin due to their sun exposure habits and geographical location [18]. The active form of vitamin D is produced in the skin when 7-dehydrocholesterol reacts with UVB irradiance; therefore, subjects living at high latitudes take less advantage of the protective action of vitamin D. Vitamin D reduces lymphocyte proliferation, Th1 and Th17 cytokine production and promotes Th2 response and the activity of CD4+ CD25+ immunoregulatory cells, thus favouring protection against the autoimmune response towards β cells [19]. In this respect, it should be noted that we and others have reported low levels of 25(OH)D3 and 1,25(OH)2 D3 in the plasma of subjects with newly diagnosed T1D compared with those in the controls [20]. The relevance of blue eyes as a risk factor for T1D is shown by the observation that in a Caucasian population where these characters are less frequent, blue eyes are more frequent in subjects with T1D than in controls. The Mediterranean region is known and appreciated for its climate, with significantly more hours of sunshine compared with northern Europe (average yearly sunshine is 2700 h in Italy versus 1900 h in Finland) and is populated by subjects where the category 3 phototype is more prevalent compared with northern Europe where phototypes 1 and 2 are the most common. In the Mediterranean area, between 80 and 90% of the population have brown colour eyes compared with northern Europe where 30–40% of the population shows such character. The question remains why in Sardinia despite the rare presence of blue eyes – although its frequency is double in T1D subjects than in controls – the disease incidence is fourfold higher than in continental Italy. Another interesting finding of our study is that blue eyes and fair skin type are risk factors for T1D independent of latitude and then of UVB irradiance. Sardinia and Lazio have the same latitude; therefore, we can exclude that, at least in the Mediterranean region, the UVB irradiance is relevant in determining the different incidence rate Copyright 2011 John Wiley & Sons, Ltd.
as suggested by Mohr et al. in a worldwide study on the relationship between T1D incidence rates and latitude [3]. We have previously reported that Sardinians who migrated to the Lazio region retained the same incidence rate as the ancestry population [12]. Therefore, taken together, all these data clearly indicate that it is the genetic predisposition that determines the prevalence of T1D in response to potentially causative environmental factors. In conclusion, prediction of T1D remains a major goal as preventive measures are now under way. To identify subjects at increased risk for T1D, the simple evaluation of eye colour may be a useful parameter to consider in calculating a weighted risk of developing T1D when other parameters such as genetic susceptibility and presence of autoantibodies to islet cells are taken into account. In particular, it should be of interest to evaluate the risk of blue eyes associated with HLA alleles of susceptibility or other genetic predisposition markers. In literature, there are no studies showing that the loci linked with different eye colour are risk loci for T1D. However, in a preliminary report [21] we have evaluated a possible association between different single nucleotide polymorphisms of the oculocutaneous albinism II protein, melanocortin 1 receptor (MC1R) and membrane-associated transporter protein genes and the eye colour and the fair skin in T1D subjects. This work revealed that some single nucleotide polymorphisms of the oculocutaneous albinism II protein gene was significantly more common in T1D subjects with green/brown eye colour compared T1D subjects with blue eyes. No statistically significant association was observed between MC1R, membraneassociated transporter protein genes and eye colour or skin. These preliminary observations merit further investigation.
Acknowledgements This work was supported by grants from University Campus BioMedico (UCBM), ‘Centro Internazionale Studi Diabete (CISD)’, and Rome and ASRIS Onlus, Italy.
Conflict of interest The authors have no conflicts of interest. Diabetes Metab Res Rev 2011; 27: 609–613. DOI: 10.1002/dmrr
613
Blue Eyes and Type 1 Diabetes
Appendix Members of IMDIAB group Altomare M, Astorri E, Barchetta I, Benevento D, Beretta Anguissola G, Bizzarri C, Buzzetti R, Capizzi M, Cappa M, Cassone Faldetta MR, Cavallo MG, Cipolloni L, Cipponeri
E, Costantino F, Crin` o A, Di Stasio E, Fioriti E, Ghirlanda G, Guglielmi C, Lauria A, Maddaloni E, Maggi D, Manfrini S, Maurizi AR, Moretti C, Morviducci L, Napoli N, Nardone MR, Palermo A, Patera P, Pitocco D, Portuesi R, Pozzilli P, Scrocca R, Spera S, Strollo R, Suraci C, Tubili C, Tuccinardi D, Valente L, Visalli N.
References 1. Pociot F, Akolkar B, Concannon P, et al. Genetics of type 1 diabetes: what’s next? Diabetes 2010; 59(7): 1561–1571. 2. Savilahti E, Saarinen KM. Early infant feeding and type 1 diabetes. Eur J Nutr 2009; 48(4): 243–249. 3. Mohr SB, Garland CF, Gorham ED, Garland FC. The association between ultraviolet B irradiance, vitamin D status and incidence rates of type 1 diabetes in 51 regions worldwide. Diabetologia 2008; 51(8): 1391–1398. 4. Hypponen E, Laara E, Reunanen A, Jarvelin MR, Virtanen SM. Intake of vitamin D and risk of type 1 diabetes: a birth-cohort study. Lancet 2001; 358(9292): 1500–1503. 5. Moltchanova EV, Schreier N, Lammi N, Karvonen M. Seasonal variation of diagnosis of type 1 diabetes mellitus in children worldwide. Diabet Med 2009; 26(7): 673–678. 6. Deluca HF, Cantorna MT. Vitamin D: its role and uses in immunology. FASEB J 2001; 15(14): 2579–2585. 7. Gregori S, Giarratana N, Smiroldo S, Uskokovic M, Adorini L. A 1alpha, 25-dihydroxyvitamin D(3) analog enhances regulatory T-cells and arrests autoimmune diabetes in NOD mice. Diabetes 2002; 51(5): 1367–1374. 8. Marjam¨ aki L, Niinist¨ o S, Kenward MG, et al. Maternal intake of vitamin D during pregnancy and risk of advanced beta cell autoimmunity and type 1 diabetes
Copyright 2011 John Wiley & Sons, Ltd.
9.
10.
11.
12.
13.
14.
15.
in offspring. Diabetologia 2010; 53(8): 1545–1548. Karvonen M, Viik-Kajander M, Moltchanova E, Libman I, LaPorte R, Tuomilehto J. Incidence of childhood type 1 diabetes worldwide. Diabetes Mondiale (DiaMond) Project Group. Diabetes Care 2000; 23(10): 1516–1526. Schmidt R. Sthenic and asthenic diabetes. Klin Wochenschr 1930; 42: 1969–1974. Ziegler AG, Baumgartl HJ, Ede G, et al. Low-pigment skin type and predisposition for development of type I diabetes. Diabetes Care 1990; 13(5): 529–531. Muntoni S, Fonte MT, Stoduto S, et al. Incidence of insulin-dependent diabetes mellitus among Sardinian-heritage children born in Lazio region, Italy. Lancet 1997; 349(9046): 160–162. Imesch PD, Wallow IH, Albert DM. The color of the human eye: a review of morphological correlates and some conditions that affect iridial pigmentation. Surv Ophthalmol 1997; 41(Suppl. 2): 117–123. Sulem P, Gudbjartsson DF, Stacey SN, et al. Genetic determinants of hair, eye and skin pigmentation in Europeans. Nat Genet 2007; 39(12): 1443–1452. Derr´e L, Corvaisier M, Charreau B, et al. Expression and release of HLA-E by melanoma cells and melanocytes:
16.
17.
18.
19.
20.
21.
potential impact on the response of cytotoxic effector cells. J Immunol 2006; 177(5): 3100–3107. Schuler P, Contassot E, Irla M, et al. Direct presentation of a melanocyteassociated antigen in peripheral lymph nodes induces cytotoxic CD8+ T cells. Cancer Res 2008; 68(20): 8410–8418. Borradale D, Kimlin M. Vitamin D in health and disease: an insight into traditional functions and new roles for the ‘sunshine vitamin’. Nutr Res Rev 2009; 22(2): 118–136. Malvy DJ, Guinot C, Preziosi P, et al. Relationship between vitamin D status and skin phototype in general adult population. Photochem Photobiol 2000; 71(4): 466–469. Rausch-Fan X, Leutmezer F, Willheim M, et al. Regulation of cytokine production in human peripheral blood mononuclear cells and allergenspecific Th cell clones by 1alpha, 25-dihydroxyvitamin D3. Int Arch Allergy Immunol 2002; 128(1): 33–41. Pozzilli P, Manfrini S, Crin` o A, et al. Low levels of 25-hydroxyvitamin D3 and 1,25-dihydroxyvitamin D3 in patients with newly diagnosed type 1 diabetes. Horm Metab Res 2005; 37(11): 680–683. Spoletini M, Maggi D, Zampetti S, et al. OCA2 gene polymorphisms are associated with specific eye colour and skin phenotype in type 1 diabetes. Diabetes 2010; 59: A330–A331.
Diabetes Metab Res Rev 2011; 27: 609–613. DOI: 10.1002/dmrr
