The similarity of Type 1 autoimmune pancreatitis to pancreatic ductal adenocarcinoma with significant IgG4-positive plasma cell infiltration
- First Online:
- Cite this article as:
- Fukui, Y., Uchida, K., Sumimoto, K. et al. J Gastroenterol (2013) 48: 751. doi:10.1007/s00535-012-0677-x
High serum immunoglobulin G4 (IgG4) levels and infiltration of IgG4-positive cells are characteristic of Type 1 autoimmune pancreatitis (AIP). We previously reported that increased regulatory T cells (Tregs) may regulate IgG4 production in AIP. Although an increased serum IgG4 concentration is observed in some patients with pancreatic ductal adenocarcinoma (PDA), clarification is still necessary. We have therefore studied the correlations between IgG4-positive cells and Tregs in patients with PDA.
Subjects and methods
A total of 21 PDA and nine AIP patients were enrolled in our study. The numbers and ratios of Tregs, IgG4-positive, and IgG-positive cells immunohistochemically stained with anti-Foxp3, IgG4, and IgG antibodies, respectively, were counted in three areas of resected pancreata in PDA, peritumoral pancreatitis (PT), and obstructive pancreatitis (OP).
In PDA, PT, OP area, the number of IgG4-Positive cells (5.183 ± 1.061, 2.250 ± 0.431, 4.033 ± 1.018, respectively; p < 0.05) and the ratio of IgG4/IgG (0.391 ± 0.045, 0.259 ± 0.054, 0.210 ± 0.048, respectively; p < 0.05) were significantly lower than those in AIP (21.667 ± 2.436 and 0.306 ± 0.052, respectively). The numbers of IgG4-positive cells did not differ significantly among the three areas of resected pancreata examined. However, the IgG4/IgG (0.391 ± 0.045) and Foxp3/monocyte (0.051 ± 0.008) ratios in PDA area were significantly (p < 0.05) higher than those in OP area (IgG4/IgG: 0.210 ± 0.048; oxp3/monocyte: 0.0332 ± 0.005), but not in PT area. Of the 21 cases of PDA, the ratio of IgG4/IgG was >40 % in nine (43 %), six (29 %) and three (14 %) cases in PDA, PT and OP area, respectively. Foxp3 and IgG4 were positively correlated in OP area, but not in PDA and PT area.
Clinicians should be careful when basing a differential diagnosis of PDA and AIP on the numbers of IgG4-positive cells and the ratio of IgG4/IgG, especially when determined using a small biopsied sample.
KeywordsAutoimmune pancreatitisIgG4Regulatory T cellsForkhead box P3Pancreatic ductal adenocarcinoma
In 1961, Sarles et al.  first observed a case of idiopathic chronic pancreatitis with hypergammaglobulinemia, in which an autoimmune mechanism was supposedly involved. In 1991, Kawaguchi et al.  defined the pathologic feature as lymphoplasmacytic sclerosing pancreatitis (LPSP), and in 1995, Yoshida et al.  proposed the concept of “autoimmune pancreatitis (AIP).” In 2001, Hamano et al.  reported that elevated serum immunoglobulin G4 (IgG4) levels were highly specific and sensitive for the diagnosis of AIP. Kamisawa et al.  subsequently suggested that AIP is a systemic disease, based on the findings that the pancreas and other involved organs showed an abundant infiltration of IgG4-positive plasma cells. Thereafter, many AIP cases have been reported by Japanese investigators, and AIP has been accepted as a new clinical entity [6–9].
Reports from Europe  and the USA  describe unique histological patterns in the resected pancreata of patients with mass-forming chronic non-alcoholic pancreatitis with epithelial destruction by granulocytes. This disease has been called idiopathic duct centric pancreatitis (IDCP) , AIP with granulocyte epithelial lesions (AIP with GEL), or Type 2 AIP . In 2011, the International Consensus Diagnostic Criteria for Autoimmune Pancreatitis (ICDC) was published. According to ICDC, AIP can be either a Type 1 AIP (LPSP) or a Type 2 AIP (IDCP) . Most of the Japanese AIP cases reported to date are LPSP, with very few reports of IDCP .
Patients with AIP commonly present with a mass in the head of the pancreas and may also have jaundice from associated bile duct strictures. Thus, they are often presumed to have a pancreatic ductal adenocarcinoma (PDA) and may undergo pancreatic resection without a definitive preoperative diagnosis. High serum IgG4 concentrations (135 mg/dL) may be helpful in distinguishing AIP from PDA preoperatively . However, about 10 % of the cases with elevated serum IgG4 levels have been reported as PDA . Moreover, infiltration of IgG4-positive cells into the tissue is characteristic of Type 1 AIP [5, 16–22]. It has clearly been shown that cases of Type 1 AIP may have numerous IgG4-positive plasma cells within the periductal inflammatory infiltrate, leading to various density thresholds have been suggested as useful markers in distinguishing Type 1 AIP from chronic pancreatitis. There have been a few recent reports of IgG4-positive cells in PDA [19, 22, 23]. However, the ratio of IgG4-positive cells to infiltrated IgG-positive cells (IgG4/IgG) in PDA remains unclear. The use of the IgG4/IgG ratio has proved to be a more valuable marker than the absolute counts of IgG4-positive cells.
Great attention has been focused on the relation between various autoimmune disease and regulatory T cells (Tregs) [24–30]. Recently, a few reports have suggested that Forkhead box P3 (Foxp3)-positive cells have been seen in PDA [31–33]. We previously reported that increased quantities of Tregs may influence IgG4 production in Type 1 AIP [34–36]. However, the relationship between IgG4 and Foxp3 in PDA is still obscure. We compared the relationship between IgG4-positive cells and Tregs in PDA and Type 1 AIP.
Clinical profile and characteristics of the patient and control groups
67 ± 2 (51–78)
Type 1 AIP
65 ± 2 (56–75)
Histopathology and immunohistochemistry
Formalin-fixed and paraffin-embedded specimens were prepared and used for histopathologic and immunohistochemical studies. Sections 4-μm thick were cut for hematoxylin and eosin (H&E), Elastica van Gieson (EvG), and immunohistochemical staining. Formalin-fixed paraffin embedded pancreatic sections were deparaffinized and rehydrated using xylene and a graded descending series of alcohol. Endogenous peroxidase activity was blocked for all sections in 3 % H2O2/methanol for 10 min. After washing in distilled water, the slides for IgG and IgG4 were treated by proteinase (Sigma, St. Louis, MO) for 15 min at room temperature, and the slides for Foxp3 were exposed to microwave pretreatment in a target retrieval solution (Dako Japan, Kyoto, Japan) at 100 °C for 20 min to enhance antigenicity. All slides were incubated for 10 min in protein blocking reagent without serum (ProTaqs Biocyc GmbH & Co., Berlin, Germany). The slides were then incubated at 4 °C overnight with primary antibodies. The primary antibodies used were a goat polyclonal antibody against human IgG (Vector laboratories, Burlingame, CA), a mouse monoclonal antibody against human IgG4 (Zymed Laboratories, San Francisco, CA), and a biotin labeled rat polyclonal antibody against human Foxp3 (eBioscience, San Diego, CA). The slides for IgG and IgG4 were then incubated with secondary antibodies using the Elite ABC goat IgG kit (Vector Laboratories) and Chem Envision kit/HRP (Dako Japan), following the manufacturers’ instructions. The slides for Foxp3 were treated with avidin-biotinylated peroxidase complex (Vector Laboratories). Finally, antibody binding was detected using 3, 3′-diaminobenzidine (DAB) (Dojindo, Kumamoto, Japan). Sections were counterstained with hematoxylin . Negative controls were evaluated by replacing the primary antibody with similarly diluted nonimmunized serum.
Immunohistochemical findings on IgG and IgG4-positive cells
Number of cases with more than ten IgG4-positive cells/high power field and a more than 40 % IgG4/IgG ratio in PDA and Type 1 AIP
IgG4/IgG > 40 %
IgG4 > 10/hpf and IgG4/IgG > 40 %
Type 1 AIP
9/9 (100 %)
8/9 (89 %)
8/9 (89 %)
1/21 (5 %)*
9/21 (43 %)**
1/21 (5 %)*
0/21 (0 %)*
6/21 (29 %)**
0/21 (0 %)*
2/21 (10 %)*
3/21 (14 %)*,†
1/21 (5 %)*
Immunohistochemical findings on Foxp3-positive cells
Correlation between the Foxp3/Mono and IgG4/Mono ratio in OP
In an attempt to clarify the pathophysiological differences between PDA and AIP, in our study we focused on the number of IgG4-positive cells and the ratio of IgG4/IgG. The ratio of cases with >10 positive cells/hpf of IgG4-positive cells was 5 % (1/21) in the PDA area and 10 % (2/21) in the OP area (Table 2). Our study of distribution of IgG4-positive cells in PDA revealed scattered or focal findings (Figs. 3, 4, 5). To date, only a few studies have evaluated IgG4 staining in PDA. In a study by Deshpande et al. , IgG4-positive cells (>1 positive cells/×20) were identified in 11 of 19 PDA cases, and documentation of increased numbers of tissue IgG4 positive plasma cells, although not an entirely specific marker for AIP, may provide ancillary evidence for the diagnosis of an IgG4-related systemic disease. It is not possible to compare our results with previous ones due to the different methods used to count positive cells, but it is clear that an increased number of tissue IgG4-positive cells is not a specific marker for Type 1 AIP.
According to the report of Zhang et al. , three of 25 resected PDA showed moderate to marked numbers(>10 positive cells/hpf) of IgG4-positive cells, and the distribution of IgG4-positive cells was patchy; moreover, there was no increased staining in areas of chronic pancreatitis adjacent to the tumor. This result relating to the IgG4-positive cell count of the PT area depended on each specimen. Therefore, in the future, it would seem necessary to require a more accurate assessment of the number of samples collected. Dhall et al.  reported that diffuse and dense staining (>50 positive cells/hpf) for IgG4 is specifically seen in LPSP and that very little and scattered or focal staining for IgG4 was seen in PDA. These results depended on each stage of the specimen. In our study, the distribution of IgG4-positive cells produced a similar finding, but the number of IgG4-positive cells was less obvious in Type 1 AIP. This difference in distribution might be useful for distinguishing Type 1 AIP from PDA.
We investigated the ratio of IgG4/IgG and found it to range from 0 to 0.71 (mean 0.39) in the PDA area, from 0 to 0.67 (mean 0.26) in the PT area, and from 0 to 0.68 (mean 0.21) in the OP area (Fig. 6c). Many investigators have examined the ratio of IgG4/IgG in inflammatory diseases [40–44], but few studies have evaluated the ratio of IgG4/IgG in PDA [45, 46]. Strehl et al.  reported that the ratio of IgG4/IgG in ten adenocarcinomas (4 pancreas, 2 colon, 2 lung, 2 breast) ranged between 0 and 0.48 (mean 0.21). Sepehr et al.  reported that the IgG4/IgG ratio of the ampulla in 30 invasive ductal adenocarcinomas ranged between 0 and 0.16 (mean 0.03). To date, there has been no detailed report of IgG4/IgG exceeding 40 % in PDA. In the 2011 comprehensive diagnostic criteria for IgG4-related disease (IgG4-RD), pathological findings of marked IgG4-positive cell infiltration (>10 cells/hpf) and an IgG4/IgG ratio of >40 % are considered to be diagnostic of IgG4-RD . In our study, the ratio of IgG4/IgG exceeding 40 % in the PDA area was much higher than expected. Although the ratio decreased with distance from the PDA area; 14 % (3/21) of the cases showed dense staining in the OP area (>40 % IgG4/IgG) (Table 2). Since only 5 % (1/21) of cases fulfilled IgG4 >10/hpf and IgG4/IgG ratio >40 % in the OP area, clinicians should be careful when basing their differential diagnosis of PDA and AIP on the number of IgG4-positive cells and the ratio of IgG4/IgG.
In the OP area, Foxp3/Mono and IgG4/Mono were positively correlated. Apart from pathogenesis, in terms of disease activity suppression, Tregs have been reported to be increased in the periphery or inflammatory sites in patients with such chronic inflammatory diseases as viral hepatitis B/C [48, 49] or Helicobacter pylori gastritis . Our previous studies showed positive correlations between IgG4 and Tregs in patients with AIP using peripheral blood , pancreas tissues , and extra pancreatic lesions , suggesting that Tregs may promote the production of IgG4 via interleukin-10 (IL-10) secreted from Tregs. In the present study, there was a significant relationship between IgG4-positive cells and Tregs, although only a few Tregs and IgG4-positive cells infiltrated into the OP area of PDA compared with AIP. Taken together, our findings suggest a similar mechanism producing IgG4 in AIP and OP of PDA. However, further studies are required to clarify this mechanism. In the PT and PDA area, Foxp3/Mono and IgG4/Mono were not correlated. IgG4 may be produced in both the PT and PDA area by a mechanism different from that in Type 1 AIP. There have been several reports of Foxp3-positive cells in PDA [31–33]. In a study by Hiraoka et al. , the expression of Foxp3 was increased by the presence of a poorly differentiated tumor for the purpose of providing protection against tumoral immunity. In our study, all of the tumors were moderately differentiated tumors; hence, we could not investigate the difference of degree of differentiation. In an immunohistochemistry using an anti-Foxp3 antibody (PCH101; eBioscience) that detected the n-terminal residues of Foxp3, Hinz et al.  reported that they detected Foxp3 expression in tumor cells of 24/39 cases of PDA. In this study, with same anti-Foxp3 antibody (PCH101; eBioscience), tumor-infiltrating lymphocytes were detectable, but malignant epithelial cells were not (data not shown). Foxp3 may, therefore, develop against tumor immunity, unlike Type 1 AIP in PDA.
In conclusion, our findings suggest that the infiltration of IgG4-positive cells is also found in obstructive pancreatitis along with PDA. Therefore, clinicians should be very careful making a differential diagnosis of PDA and AIP based on the number of IgG4-positive cells and the ratio of IgG4/IgG, especially with a small biopsied sample taken by fine needle aspiration.
This study was partially supported by (1) Grant-in-Aid for Scientific Research (C) of the Ministry of Culture and Science of Japan (20590810, 24591020), (2) the Research Program on Intractable Diseases, from the Ministry of Labor and Welfare of Japan, and (3) grants-in-aid from the Ministry of Education, Culture, Sports, Science and Technology of Japan, from CREST Japan Science and Technology Agency.
Conflict of interest
The authors declare that they have no conflict of interest.