Introduction

Type 1 diabetes (T1D) is an immune-mediated disease causing the destruction of pancreatic β-cell.1, 2 The extent of β-cell destruction and the resulting residual β-cell function are variable. Indeed, C-peptide secretion could be found at onset, during the so-called remission phase.3, 4 This endogenous insulin secretion could completely disappear soon after diagnosis or persist over a long period of time.3, 4

These findings suggest that in T1D, natural course of β-cell destruction may vary markedly, and it may be influenced by different genetic backgrounds.

A novel gene related to T1D has recently been described, namely the protein tyrosine phosphatase non-receptor type 22 (PTPN22).5 PTPN22 is located outside the HLA region, it maps to human chromosome 1p13.3-p13.1 and encodes for the lymphoid-specific tyrosine phosphatase (Lyp), which is expressed in immature and mature lymphocytes of type B and T, as well as in natural killer cells. This protein can be found in three isoforms, Lyp1–3, being Lyp1 the most abundant.6 Although the exact role of this protein remains unclear, studies have shown that Lyp may negatively regulate the proximal signaling pathway after T-cell receptor activation, thus affecting maturation and proliferation.7, 8 Even if data are lacking, it is likely that Lyp may have a similar role in B cells.9, 10 This single-nucleotide polymorphism PTPN22 C1858T, encoding Arg620Trp, was implicated in the disruption of the mechanism of T-cell deactivation, and it has been suggested to modify the thresholds of thymic selection, leading to an expansion of autoreactive T cells. This may explain recent reports about the association of PTPN22 and many autoimmune diseases, like T1D,5 rheumatoid arthritis,11 systemic lupus erythematosus,12 Graves’ disease13, 14 and Hashimoto thyroiditis (HT).15 The association between T1D and PTPN22 C1858T was confirmed in United States,5, 16, 17 German,18 Dutch,19 British,20 Finnish,21 Danish,22 Spanish,23 Italian,24, 25 Czech,26 Azeri,26 Ukrainian Caucasian27 and Colombian28 populations, but not in African-American,11 Asian29 or Japanese30 populations, likely due to the low frequency of the 1858T allele in these three populations.

In recent years, immunologists have tried to exploit immunotherapeutic strategies for treating and/or preventing autoimmune diseases.31 The goal of these strategies should be to eliminate autoimmune processes before the complete loss of ability of cells, that is, pancreatic β-cells in T1D, in order to ensure adequate blood glucose levels.

The aims of this study were (1) to evaluate the prevalence of the C1858T polymorphism in PTPN22 gene in a population of children and adolescents with T1D and (2) to investigate the association between the PTPN22 C1858T variant, gender distribution, age at onset, autoimmunity, residual β-cell function and insulin requirement (IR) at 6 months after diagnosis.

Materials and methods

This study was performed in a cohort of Caucasian children and adolescents (n=113 patients, 58 males and 55 females) with T1D, consecutively diagnosed from January 2003 to November 2013. Patients were recruited and followed up at the Pediatric Diabetes Outpatient Unit, Department of Pediatrics, Faculty of Medicine, ‘G. D’Annunzio’ University of Chieti.

Only patients with an initial diagnosis of T1D according to the ISPAD (International Society for Pediatric and Adolescent Diabetes) criteria were enrolled in the study while patients with suspected non-T1D (type 2 diabetes, maturity-onset diabetes of the young or secondary diabetes) were excluded.

For each subject, the following data were collected: gender, age at study entry, age at diagnosis, severity of onset (presence or absence of ketoacidosis, in particular pH, base excess and bicarbonate levels), evaluation of pancreatic autoimmunity, glycated hemoglobin (HbA1c), body mass index (BMI), personal history of HT and celiac disease (CD), and family history of T1D of at least one first-degree relative. Diabetes-related autoimmunity was assessed at diagnosis of the disease, while CD and thyroid-related autoimmunity were determined at diabetes onset and at annual intervals.

All the subjects were treated with an intensive insulin therapy consisting of three injections of rapid insulin (lispro or aspart) plus glargine insulin once a day. Periodical adjustments were made every 3 or 6 months.

To estimate the residual β-cell function, the C-peptide dosage was measured by immunoassay. The evaluation of pancreatic autoimmunity through the presence of glutamic acid decarboxylase-65 (GAD), insulinoma-associated 2 molecule (IA2) and islet cell antibodies (ICA) was determined at the time of disease onset using commercial kit radioimmunoassay (Bio-Rad, Milan, Italy).

Capillary HbA1C was determined by column assay at diagnosis and after 6 months. Moreover, daily IR 6 months after diagnosis was detected.

Additional information on the presence of other specific autoantibodies (thyroperoxidase autoantibodies and anti-gliadin and anti-transglutaminase IgA antibodies) and familial history of T1D was evaluated at diagnosis.

The C1858T variant (rs2476601) was investigated in a group of 113 patients by PCR amplification followed by PCR-RFLP and agarose gel electrophoresis.

Genomic DNA was extracted from Buccal Swabs using Chelex 100 Resin according to the manufacturer’s instructions (Bio-Rad Laboratories, Hercules, CA, USA) and stored at −20 °C.

PCR was performed in a final reaction volume of 25 μl containing 20–50 ng of genomic DNA template, standard PCR Buffer, 100 μm dNTPs, 1,5 mm MgCl2, 100 nm primers and 1.75U of AB Taq (AB Analitica, Padova, Italy). The amplification conditions consisted of an initial denaturation at 95 °C for 10 min, followed by 30 cycles of 95 °C for 30 s, 63 °C for 60 s, 72 °C for 60 s, and a final extension at 72 °C for 7 min.

The upstream primer 5′-ACCTCCTGGGTTTGTACCTT-3′ and the downstream primer 5′ACTTCTCAGGTCCTTTCAATGT-3′ were used to generate a 678-bp amplicon. The PCR-product size and purity were evaluated through agarose gel electrophoresis.

The PTPN22 genotypes were identified by CViQI restriction endonuclease digestion (Thermo Fisher Scientific, Waltham, MA, USA) which recognizes the G^TAC sites by making a blunt cut. Enzymatic digestion was performed at 37 °C for 2 h according to the manufacturer’s instructions, and the digested products were separated on a 3% agarose gel. The enzymatic digestion created two fragments of 391 bp and 232 bp for the CC variant, two fragments of 391 bp and 283 bp for the TT variant and three fragments of 391 bp, 283 bp and 232 bp for the CT variant.

Written informed consent was obtained from all parents or legal guardians of participating subjects. Statistical analysis was performed by using SPSS (Statistical Package for the Social Science), version 17.0 software for Windows (SPSS Inc, Chicago, IL, USA). A P-value of ⩽0.05 was considered as significant. The distribution of the study population was investigated by Kolmgorov–Smirnov Test.

Normal distribution variables (glycemia, HbA1c, BMI and age at onset; HbA1c and IR after 6 months) were analyzed by comparing the means of patients bearing the tested polymorphism with the means of patients with the normal nucleotide sequence using the Independent Samples T-Test.

Non-normal distribution variables (C-peptide at diagnosis) were analyzed using the Mann–Whitney test.

The prevalence of the studied polymorphism for dichotomous variables such as gender, first grade familiarity for T1D, presence of concomitant autoimmune diseases and the positivity for autoimmune antibodies (ICA, GAD and IA2) were studied using the Fisher’s exact test.

Results

The mean age at diagnosis was 98.5±49.6 s.d. months. BMI was 18.06±3.40 kg m−2 s.d. and HbA1c was 11.38±2.17% s.d. at the time of diagnosis (Table 1).

Table 1 Distribution of CC and CT+TT genotypes in relation to gender, age at onset, BMI, glycemia, HbA1c (mean±s.d.) and C-peptide at first evaluation (median and IQR)
Full size table

We found an association of the T allele with T1D, being this allele present in 17.7% of patients, frequency higher than the one previously reported in the general population.24 The C/T genotype was significantly more frequent in patients (14.2%) than T/T (3.5%). Because of the low frequency of 1858T allele, the C/T and T/T genotypes were grouped for statistical analysis.

We did not find statistically significant association of C1858T polymorphism with a positive family history of T1D; however, nearly 50% of C/T and T/T genotype carriers had a family history of T1D. The C1858T PTPN22 polymorphism was not associated with the age of diabetes onset (100.3±46.1 s.d. months for CT/TT versus 96.8±53.1 s.d. months for CC; P=0.77) (Table 1).

The frequency of C/T and T/T genotypes was 15.5% in males versus 20% in females (Table 1). The distribution of C/T and T/T genotypes carrying the predisposing allele did not differ significantly between males and females.

There were no significant differences with respect to gender distribution, age, anthropometric data, HbA1c at diagnosis, other metabolic parameters, family history of diabetes between patients with wild-type genotype and patient carriers of the C/T and T/T genotypes.

The C-peptide level was significantly higher for the T-allele carriers compared with C-allele (wild-type) carriers at disease onset (genotype C/T-T/T:C-peptide=0.60 ng ml−1 (0.32–0.84) versus genotype C/C:C-peptide=0.40 ng ml−1 (0.20–0.50); P=0.001) (Table 1 and Figure 1).

Figure 1
figure 1

Children bearing the mutated allele of the investigated gene (PTPN22 C1858T) presented significantly higher levels of C-peptide at the diabetes onset. *P=0.001.

PowerPoint slide

Full size image

At 6 months, IR was lower in subjects carrying the PTPN22 polymorphism than in subjects homozygous for C allele. Means±s.d. for IR at 6 months were 0.59±0.23 and 0.48±0.19, respectively, in CC subjects and in CT/TT subjects, P=0.04 (Table 2 and Figure 2).

Table 2 Distribution of CC and CT+TT genotypes in relation to HbA1c and insulin requirement at 6 months from diagnosis (mean± s.d.)
Full size table
Figure 2
figure 2

Children bearing the mutated allele of the investigated gene (PTPN22 C1858T) had a significant lower insulin requirement 6 months after the diabetes diagnosis. *P=0.04.

PowerPoint slide

Full size image

Metabolic control did not differ between the two patient groups during the first year after diabetes onset: genotype C/T-T/T:HbA1c=6.96±1.26% (s.d.) versus genotype C/C:HbA1c=7.19±1.27% (s.d.), n.s. (P=0.32).

The presence of GAD, IA2 and ICA autoantibodies was determined at the time of disease onset. Overall, the C1858T genotype was associated with GAD positivity (63% of CC subjects versus 80% of C/T-T/T subjects; P=0.19), although the result was not statistically significant (Table 3).

Table 3 Distribution of CC and CT+TT genotypes in relation to the positivity for pancreatic autoantibodies or other autoimmune diseases
Full size table

In contrast, no effect of C1858T genotype on ICA (32% of C/C subjects versus 21% of C/T-T/T subjects, P=0.32) and IA2 (77% of C/C subjects versus 85% of C/T-T/T subjects, P=0.53) positivity was found (Table 3).

Regarding other autoimmune diseases, among C/T and T/T genotype carriers, HT and CD were more frequent than wild-type carriers, but P-values were not statistical significant (T1D+HT: 17% of C/C subjects versus 25% of C/T-T/T subjects; P=0.21; T1D + CD: 8% of C/C subjects versus 15% of C/T-T/T subjects; P=0.30) (Table 3).

These results remain preliminary, due to the small number of patients.

Discussion

Recent researches have showed that PTPN22 gene is associated with the development of T1D and other autoimmune diseases.

Bottini et al.32 suggested that the risk-carrying allele 1858T suppresses T-cell receptor signaling more efficiently during thymic development, which results in the survival of autoreactive T cells.32

The first report by Bottini et al.5 of the association of PTPN22 C1858T polymorphism with T1D was confirmed by several other association studies.21, 33

This finding was reinforced by recently published meta-analyses, which suggest that the PTPN22 C1858T polymorphism may contribute to the predisposition of T1D, especially in populations of Europe and America.34

Our study provides another replication of the association of PTPN22 1858T allele with T1D. We confirmed the association of the 1858T allele with T1D, with a carrier frequency of 17.7%, compared with the frequency in healthy subjects as previously reported in other studies carried out in the same region (5.6% in the study of Saccucci et al.,24 6.7% in the study of Gloria-Bottini et al.35).

These findings support previous results. Gianchecchi et al.36 showed 17% heterozygosity for PTPN22 C1858T in T1D patients with a higher frequency than in the control group (8.75%).

Mainardi-Novo et al.37 have demonstrated a higher frequency of the C1858T PTPN22 gene polymorphism in T1D patients: C/T-T/T genotypes in 18.7 versus 10.6% of controls.

Also in the series by Kordonouri et al.,38 a higher frequency of the PTPN22 C1858T polymorphism in young T1D patients was found.

A recent study performed by Gloria-Bottini et al.,35 involving patients from continental Italy and Sardinia, demonstrated an association between PTPN22 and T1D (13.6% in T1D versus 6.7%) in patients from continental Italy.

The interaction of the PTPN22 C1858T polymorphism with gender in T1D was suggested.22, 39

Our series did not find a different distribution of the C/T and T/T genotypes between males and females, in agreement with the majority of reported studies.21, 24, 37, 39, 40

Genetic factors have been also reported to have a certain influence on age of diagnosis. However, in the present study the C1858T polymorphism was not associated with the age of diabetes onset, in agreement with a recent Italian study of Gloria-Bottini et al.35 and as previously reported by other researchers.16, 18, 20, 30, 40, 41

Regarding autoimmunity, in the present study we have shown an increased frequency of GAD positivity among PTPN22 polymorphism carriers, although our results are not statistically significant, most likely owing to the sample size. On the other hand, no effect of C1858T genotype on IA2 and ICA positivity (P=0.53 and P=0.32, respectively) was recorded.

An explanation for these findings could be that PTPN22 1858T is involved in two independent control processes from onset and during the following 12 months.42 In our series, T1D autoimmunity was detected only at diabetes onset.

These results are in agreement with the findings of Maziarz et al.,43 which found that the association between PTPN22 (C/T+T/T) and GAD-positive T1D was much stronger than the association between PTPN22 (C/T+T/T) and GAD-negative T1D. Also in the study of Mainardi-Novo et al.,37 the polymorphism was associated with a higher frequency of GAD.

The increased association we have found between 1858T allele in T1D patients with CG and familial history of T1D supports the concept that this genotype confers a general susceptibility to autoimmune diseases, which are known to occur with increased frequency in T1D patients. However, Rueda et al.44 did not observe any statistically significant deviation after comparing allele and genotypic frequencies of PTPN22 C1858T between patients with CD and controls.

In our series, C-peptide levels at onset were significantly higher in T1D subjects carrying the 1858T variant of the PTPN22 gene. According to this result, we also found lower IR after 6 months of disease in diabetic patients carrying the PTPN22 variant. This finding provides a further evidence of better residual β-cell function in C/T and T/T genotype carriers, mainly taking into account similar HbA1c values in the two patient groups. In our series, HbA1C values for the first year after diabetes onset were also similar in the two groups regardless of C-peptide values at diagnosis, probably due to our strict educational program for newly diagnosed patients. These results conflict with other studies. Indeed, Petrone et al.33 found an association between the T allele of the C1858T variant, low fasting C-peptide (as a surrogate marker of residual β-cell mass) and poorer glycemic control at diagnosis, while no differences were shown in IR. These findings were independent of age at onset, sex and HLA risk groups.

Andersen et al.45 showed an association between genetic pattern, including multiple numbers of risk alleles for T1D-associated nucleotide polymorphisms (INS and PTPN22 genes), β-cell function and glycemic control during the first year after diagnosis. They also demonstrated a similar trend between stimulated C-peptide and proinsulin.

Nielsen et al.42 suggested an association between carriers of the T allele and high proinsulin throughout the first year after disease onset. Interestingly, in this study of Hvidoere Study Group on Childhood Diabetes, C/T and T/T carriers presented significantly higher proinsulin (30%) levels at onset and over the 12-month study period compared with the C/C genotype group. Kaas et al.46 reported that proinsulin and C-peptide were positively associated with diabetes onset. Our series did not measure proinsulin levels, but we speculate that the results from the Hvidoere Study Group on Childhood Diabetes are in line with our findings.

In our patients, we found a relationship among C-peptide levels, GAD positivity at diabetes onset, IR at 6 months and PTPN22 1858T variant. These results indicate that the extension of β-cell destruction in T1D could be controlled in part by the PTPN22 gene. Indeed, this mutation is thought to allow T cells to remain activated for a longer period of time.43

Prevention of β-cell loss in T1D patients is expected to improve insulin secretion and disease management.47

The aim of tertiary prevention in T1D and one of the major goals of current research48 is delaying or even stopping β-cell destruction.49 A number of studies tried to develop and identify compounds able to counteract immunological response against pancreatic islets.48, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60

Identification of subset of T1D patients who will be eligible for tertiary prevention trials could encompass genetic markers as well as metabolic parameters. PTPN22 1858T is one of the most recognized T1D-associated allele and shows, in our patients, an association with a residual β-cell function.

Recentlty, a class of Lyp inhibitors able to counteract the gain-of-function mutation generated by the C1858T polymorphism have been identified.61, 62, 63, 64

Xie et al.62 reported a class of thiazolidine-2,4-diones and 2-thioxothiazolidin-4-ones acting as potent inhibitors of Lyp, suggesting a new approach to treatment of autoimmune diseases by inhibition of Lyp. Moreover, Vang et al.63 studied inhibition of Lyp by benzofuran salicylic acids.

This new approach may be a future benefit for tertiary prevention trials, aimed toward preventing the progression of β-cell loss. Our results provide further evidence of the heterogeneity of T1D.

Undoubtedly, a weakness of our study is represented by the lack of data on C-peptide at 6 months. However, our results show an IR of ⩽0.5 units per kg per day associated with HbA1C ⩽7.5 % in CT+TT carriers: these parameters have been used to define the partial remission period, that is obviously associated with residual β-cell function.65, 66

Moreover, our CT+TT carriers also showed a mean value of ‘insulin dose-adjusted A1C’ (IDDA1C, a qualitative measure of partial remission) of 5.22. This value fulfills the threshold (⩽9) established by Mortensen et al.67

A strength of our study is the evidence of higher C-peptide levels at onset, lower IR after 6 months of disease, in C/T and T/T carriers, suggesting that PTPN22 C1858T could be associated with slow disease progression.

Our results, once confirmed by further studies, may be useful in identifying a subset of patients who can potentially benefit from novel therapeutic approaches in the coming years.