Amylase alpha-2A autoantibodies: novel marker of autoimmune pancreatitis and fulminant type 1 diabetes

Abstract

Objective: The pathogenesis of autoimmune pancreatitis (AIP) and fulminant type 1 diabetes remains unclear, although it is known that immune-mediated processes severely compromise the endocrine and exocrine functions in both diseases.

Research design and methods: We have screened a lambdaTriplEx2 human pancreas cDNA library with serum from a patient with AIP and obtained positive clones. Sequence analysis revealed that 7 of 10 clones were identical to human amylase alpha-2A. Using a recombinant COOH-terminal amylase alpha-2A protein, we developed an enzyme-linked immunosorbent assay system to detect autoantibodies against human amylase alpha-2A.

Results: All 15 serum samples from patients with AIP recognized the recombinant protein, whereas sera from 25 patients with chronic alcoholic pancreatitis and sera from 25 patients with a pancreas tumor did not. Interestingly, 88% (15/17) of patients with fulminant type 1 diabetes were positive for an autoantibody against amylase alpha-2A. These antibodies were detected in 21% of patients with acute-onset type 1 diabetes (9 of 42) and 6% of type 2 diabetic patients (4 of 67).

Conclusions: These results suggest that an autoantibody against amylase alpha-2A is a novel diagnostic marker for both AIP and fulminant type 1 diabetes and that, clinically and immunologically, AIP and fulminant type 1 diabetes are closely related.

FIG. 1.

Cloning of human amylase α-2A cDNAs from λTriplEx2 human pancreas cDNA library. Seven clones of human amylase α-2A cDNAs. Their lengths and 5′-ends are shown (A in ATG is designated as 1). The top bar indicates human amylase α-2A cDNA as reported by Wise et al. (17), and the common regions shared by all seven clones, from codons 299 to 512, are shown in the bottom bar.

FIG. 2.

Western blot analysis and ELISA for detecting anti-human AMY-2A. A: Western blot analysis. Recombinant human AMY-2A (50 ng) from codons 299 to 512 (AMY-2A/299–512) was electrophoresed in 0.1% SDS–15% polyacrylamide and transferred onto a nitrocellulose filter. The filters reacted with serum (×1,000) from an AIP patient (line 1) and normal control sera (lines 3 and 4). Line 2, AIP patient's serum preincubated with 1 μg/ml AMY-2A/299–512. MW, molecular weight markers. B: Immunoprecipitation of ³⁵S–AMY-2A with antibodies. ³⁵S–AMY-2A was incubated with goat anti-amylase (line 1), normal goat IgG (line 2), sera from AIP patients (lines 3 and 4), and sera from healthy volunteers (lines 5 and 6) and then precipitated with protein G–sepharose. The pellets were electrophoresed in 0.1% SDS–15% polyacrylamide and analyzed with Bas 2000 image analyzer (Fujix, Tokyo). C: Correlation between the result of ELISA and that of immunoprecipitation. By coating the recombinant human AMY-2A/299–512, we developed an ELISA system for detecting anti-human AMY-2A. Sera from 11 patients with AIP (•) and two normal control subjects (○) were assayed by ELISA and immunoprecipitation for detecting the autoantibody. D: Absorption of positive ELISA signal with recombinant AMY-2A. One milliliter of a patient's serum (1:500) was preincubated with the recombinant protein at the indicated dose overnight at 4°C, and then the serum was used as the first antibody. The data are the mean of triplicate values. OD, optical density. E: Serum dilution experiment in ELISA assay. Positive serum from patient A.O. was diluted as indicated, and ELISA assay was carried out. The data are the mean of triplicate values. F: ROC analysis of the healthy volunteers and fulminant type 1 diabetic patients. We carried out ROC analysis of the healthy volunteers (n = 100) and fulminant type 1 diabetic patients (n = 17) with MedCalc.

FIG. 3.

Prevalence of autoantibody against human AMY-2A in patients with various pancreatic diseases. A: Prevalence of autoantibody against human AMY-2A in patients with AIP (n = 15), chronic alcoholic pancreatitis (CP, n = 25), pancreatic tumor (PT, n = 25), control subjects from healthy volunteers (C, n = 100), and Hashimoto's thyroiditis (H, n = 47) was examined by ELISA, as described in research design and methods. The data are the mean of triplicate values. The dotted line shows a cutoff value. Fisher's exact test was carried out between AIP and control groups. *P < 0.001. B: Time course of anti-AMY antibody and IgG4 of AIP patients. AIP patient (A.O.), whose IgG was used to screen λTriplEx2 human pancreas cDNA library, was treated with prednisolone (arrow). Before and after the treatment, anti-AMY antibody (•-•) and IgG4 (○-○) were measured. In the other AIP patient (T.M.), titer of the anti-AMY antibody (▴-▴) was also measured before and after prednisolone treatment (arrow).

FIG. 4.

Prevalence of autoantibody against human AMY-2A in patients with various types of diabetes. Prevalence of autoantibody against human AMY-2A (•) in patients with fulminant type 1 diabetes (FT1DM, n = 17), acute-onset type 1 diabetes (AT1DM, n = 42), type 2 diabetes (T2DM, n = 67), and control subjects from healthy volunteers (C, n = 100) was studied by ELISA, as described in research design and methods. The data are the mean of triplicate values. The dotted line shows a cutoff value. Fisher's exact test was carried out between fulminant type 1 diabetic and control groups. *P < 0.001.