Immunohistochemical study of autoimmune pancreatitis using anti-IgG4 antibody and patients’ sera S Aoki, T Nakazawa, H Ohara, H Sano, H Nakao, T Joh, T Murase, 1 T Eimoto 1 & M Itoh Department of Internal Medicine and Bioregulation and 1 Department of Clinical Pathology, Nagoya City University Graduate School of Medical Sciences, Nagoya, Japan Date of submission 24 November 2004 Accepted for publication 29 December 2004 Aoki S, Nakazawa T, Ohara H, Sano H, Nakao H, Joh T, Murase T, Eimoto T & Itoh M (2005) Histopathology 47, 147–158 Immunohistochemical study of autoimmune pancreatitis using anti-IgG4 antibody and patients’ sera Aims: Autoimmune pancreatitis (AIP), characterized by raised serum IgG4 levels, is frequently complicated by disorders of extrapancreatic organs. The aim of the present study was to examine immunohistochemically which extrapancreatic organs are affected, and whe- ther an autoantibody to such organs is present in the serum of AIP patients. Methods: Various tissues ⁄ organs obtained from AIP patients were studied immunohistochemically with an anti-IgG4 antibody. To examine the presence of an autoantibody in the serum of AIP patients, sera were incubated with various normal organs ⁄ tissues extrac- ted for other diseases, followed by detection with an anti-IgG4 antibody. Sera were also examined before and after glucocorticoid therapy. Results: Marked infiltration of IgG4+ plasma cells was observed in the pancreas, liver, bile duct and salivary gland of many of the AIP patients examined. The normal epithelia of the pancreatic ducts, bile ducts, gallbladder and salivary gland ducts reacting with the patients’ sera were detectable by the anti-IgG4 anti- body. Following glucocorticoid therapy the IgG4 anti- body from the patients’ sera showed decreased reactivity with these tissues. Conclusions: AIP may also affect extrapancreatic organs, the serum of AIP patients may contain an IgG4 autoantibody to various organs and glucocorti- coid therapy may improve such disorders. Keywords: anti-IgG4 antibody, autoantibody, autoimmune pancreatitis, IgG4-positive plasma cells, immunohisto- chemistry Abbreviations: AIP, autoimmune pancreatitis; CA, carbonic anhydrase; HPF, high-power field; PSC, primary sclerosing cholangitis Introduction Autoimmune pancreatitis (AIP) is a chronic pancrea- titis that has recently been increasingly reported. Sarles et al. 1 first reported the phenomenon of pancreatitis with hypergammaglobulinaemia and its possible rela- tion to the phenomena of self-immunization. Since the term autoimmune pancreatitis was coined in 1995, 2 a number of reports have been published. 3–7 In 2001 Etemad et al. 8 categorized the major predisposing factors to chronic pancreatitis as (i) toxic-metabolic, (ii) idiopathic, (iii) genetic, (iv) autoimmune, (v) recur- rent and severe acute pancreatitis, and (vi) obstructive (TIGAR-O system), including AIP as one of six risk factors. The pathogenesis of AIP remains unknown. An autoimmune mechanism has been postulated, and pancreatic duct epithelium-derived carbonic anhydrase II and lactoferrin have been reported to be potential target antigens. 4,9,10 Pathological studies of AIP tissue have demonstrated infiltration of CD4+ T lymphocytes Address for correspondence: Shigeru Aoki MD, Department of Internal Medicine and Bioregulation, Nagoya City University Gradu- ate School of Medical Sciences, 1 Kawasumi, Mizuho-cho, Mizuho-ku, Nagoya, 467–8601, Japan. e-mail: [email protected]Ó 2005 Blackwell Publishing Limited. Histopathology 2005, 47, 147–158. DOI: 10.1111/j.1365-2559.2005.02204.x
Transcript
Immunohistochemical study of autoimmune pancreatitis usinganti-IgG4 antibody and patients’ sera
S Aoki, T Nakazawa, H Ohara, H Sano, H Nakao, T Joh, T Murase,1 T Eimoto1 & M ItohDepartment of Internal Medicine and Bioregulation and 1Department of Clinical Pathology, Nagoya City University
Graduate School of Medical Sciences, Nagoya, Japan
Date of submission 24 November 2004Accepted for publication 29 December 2004
Aoki S, Nakazawa T, Ohara H, Sano H, Nakao H, Joh T, Murase T, Eimoto T & Itoh M
(2005) Histopathology 47, 147–158
Immunohistochemical study of autoimmune pancreatitis using anti-IgG4 antibody andpatients’ sera
Aims: Autoimmune pancreatitis (AIP), characterizedby raised serum IgG4 levels, is frequently complicatedby disorders of extrapancreatic organs. The aim of thepresent study was to examine immunohistochemicallywhich extrapancreatic organs are affected, and whe-ther an autoantibody to such organs is present in theserum of AIP patients.Methods: Various tissues ⁄ organs obtained from AIPpatients were studied immunohistochemically with ananti-IgG4 antibody. To examine the presence of anautoantibody in the serum of AIP patients, sera wereincubated with various normal organs ⁄ tissues extrac-ted for other diseases, followed by detection with ananti-IgG4 antibody. Sera were also examined beforeand after glucocorticoid therapy.
Results: Marked infiltration of IgG4+ plasma cells wasobserved in the pancreas, liver, bile duct and salivarygland of many of the AIP patients examined. Thenormal epithelia of the pancreatic ducts, bile ducts,gallbladder and salivary gland ducts reacting with thepatients’ sera were detectable by the anti-IgG4 anti-body. Following glucocorticoid therapy the IgG4 anti-body from the patients’ sera showed decreasedreactivity with these tissues.Conclusions: AIP may also affect extrapancreaticorgans, the serum of AIP patients may contain anIgG4 autoantibody to various organs and glucocorti-coid therapy may improve such disorders.
Autoimmune pancreatitis (AIP) is a chronic pancrea-titis that has recently been increasingly reported. Sarleset al.1 first reported the phenomenon of pancreatitiswith hypergammaglobulinaemia and its possible rela-tion to the phenomena of self-immunization. Since theterm autoimmune pancreatitis was coined in 1995,2
a number of reports have been published.3–7 In 2001Etemad et al.8 categorized the major predisposingfactors to chronic pancreatitis as (i) toxic-metabolic,(ii) idiopathic, (iii) genetic, (iv) autoimmune, (v) recur-rent and severe acute pancreatitis, and (vi) obstructive(TIGAR-O system), including AIP as one of six riskfactors.
The pathogenesis of AIP remains unknown. Anautoimmune mechanism has been postulated, andpancreatic duct epithelium-derived carbonic anhydraseII and lactoferrin have been reported to be potentialtarget antigens.4,9,10 Pathological studies of AIP tissuehave demonstrated infiltration of CD4+ T lymphocytes
Address for correspondence: Shigeru Aoki MD, Department of
Internal Medicine and Bioregulation, Nagoya City University Gradu-
ate School of Medical Sciences, 1 Kawasumi, Mizuho-cho, Mizuho-ku,
around the main pancreatic duct,4 expression of HLA-DR antigen in the pancreatic duct cells4 and thepresence of anti-pancreatic duct antibodies in theserum of AIP patients.11 Hamano et al.12 found thatAIP patients had significantly higher serum IgG4 levelsthan patients with other pancreatic or biliary systemdiseases. In addition, they reported patients with AIPcomplicated by retroperitoneal fibrosis.13 Their histo-logical study showed infiltration of IgG4+ plasma cellsin ureteral and pancreatic lesions.
Kamisawa et al.14 reported diffuse infiltration ofCD4+ or CD8+ T lymphocytes and IgG4+ plasma cellsin peripancreatic tissue, extrahepatic bile ducts andgallbladder as well as in the pancreas. We have alsoreported a high frequency of bile duct lesions,15 andgeneralized lesions in AIP patients,16 suggesting thatAIP affects not only the pancreas but also extrapancre-atic organs. From these observations, we postulated thepresence of autoantibody to various tissues in the serumof AIP patients. However, there has been no histologicalproof as to the presence of such autoantibodies in AIPpatients’ serum. In this pathological study, we investi-gated which extrapancreatic organs are affected inAIP, and whether an autoantibody to such organswas present in the serum. First, we studied not onlythe pancreas but also extrapancreatic tissues ⁄ organsderived from AIP patients by immunohistochemistry.Second, we analysed with which components of normaltissue the patients’ serum specifically reacted, usingorgans extracted for other diseases. In addition, sinceglucocorticoid therapy has been reported to improvedramatically pancreatic and bile duct lesions, theimmunoreactivity of the patients’ serum to these tissueswas examined before and after glucocorticoid therapy.
Materials and methods
case selection and clinical summary of
aip patients
Twenty cases of AIP, representing patients seen from1993 to 2004, were retrieved from the file of theDepartment of Internal Medicine and Bioregulation,Nagoya City University Graduate School of MedicalSciences. Cases were selected based on the diagnosticcriteria for AIP proposed by the Japan PancreaticSociety.17 The criteria are summarized as follows.(1) Pancreatic imaging studies show diffuse narrowingof the main pancreatic duct with irregular wall (morethan 1 ⁄ 3 length of the entire pancreas) and enlarge-ment of the pancreas. (2) Laboratory data demonstrateabnormally elevated levels of serum c-globulin and ⁄ orIgG, or the presence of autoantibodies. (3) Histopatho-
logical examination of the pancreas shows fibroticchanges with lymphocyte and plasma cell infiltration.For diagnosis, criterion 1 must be present, togetherwith at least one of criteria 2 and 3. Four cases fulfilledall the diagnostic criteria of 1, 2 and 3, 15 cases only 1and 2, one case only 1 and 3, respectively.
Table 1 summarizes the clinical data for the 20patients in this study group. All of them had enlarge-ment of the pancreas and narrowing of the mainpancreatic duct. Three portions of the biliary tree wereevaluated separately. They were the intrapancreaticbile duct defined as the bile duct in the pancreas, theextrahepatic bile duct defined as the common bile ductoutside the pancreas, common hepatic duct, and distal10 mm of the left and right hepatic ducts, and theintrahepatic bile duct consisting of the remainder of theproximal biliary tree. Cholangiography was performedin a standard fashion by endoscopic retrograde andpercutaneous transhepatic techniques. Seventeen cases(85%) had a stricture of the intrapancreatic bile duct,seven cases (35%) had a stricture of the extrahepaticbile duct and six cases (30%) had a stricture ofthe intrahepatic bile duct. Four cases (20%) showedan abnormality of the salivary gland. Mean serumc-globulin was 2.14 g ⁄ dl (1.15 < normal < 1.25 g ⁄ dl),IgG concentration was 2232 mg ⁄ dl (870 < normal< 1700 mg ⁄ dl) and IgG4 concentration was 519.5mg ⁄ dl (normal < 135 mg ⁄ dl12). Antinuclear antibodywas detected in 15 of 19 cases (78.9%).
immunohistochemical analysis of patients’
t issues
Tissues from various organs (pancreas, liver, bile duct,stomach, duodenum, large intestine and salivary gland)obtained from 20 AIP patients were examined forinfiltration of IgG4+ plasma cells. Immunohistochem-istry was performed on formalin-fixed paraffin-embed-ded specimens. A control group included six patientswith chronic calcifying pancreatitis, 10 patients withprimary sclerosing cholangitis (PSC), six patients withchronic gastritis and five patients with Sjogren’s syn-drome (Table 2). The diagnosis of PSC was made on thebasis of the radiological findings of diffuse and irregularnarrowing of the intra- and extrahepatic bile ductson endoscopic retrograde cholangiopancreatography,the complication of ulcerative colitis, and histologicalexamination of the liver. The 3 lm-thick sections wereimmunostained for mouse anti-human IgG4 monoclo-nal antibody (MC011, dilution 1 : 100; Birmingham,UK), using an automated immunostainer (OptiMaxPlus; BioGenex, San Ramon, CA, USA) by the standardstreptavidin–biotin–peroxidase method (Histofine Kit;
Nichirei, Tokyo, Japan). The reactivity of the anti-human IgG4 antibody was abolished by the absorptiontest, i.e. by preincubating the antibody with IgG4(Calbiochem, San Diego, CA, USA).
Infiltration of IgG4+ plasma cells was evaluated bycounting the number of positive plasma cells in the 10high-power fields (HPFs, ·400) and dividing the valueby 10. All slides were reviewed in a blinded manner bytwo of the authors (S.A. and T.M.) and the degrees ofinfiltration were categorized by the following fourscores: score 0 ¼ 0 per 1 HPF, score 1 £ 20 per1 HPF, score 2 ¼ 20–50 per 1 HPF, and score 3 ‡ 50per 1 HPF. When the assessment was different betweenthe two observers, agreement was reached using adouble-headed microscope.
immunohistochemical analysis using
patients’ sera
Sera obtained from six AIP patients (cases 1–6, fivemale, one female, mean age 64 years, range 48–76 years) and six healthy volunteers (five male, one
female, mean age 64.4 years, range 60–67 years) wereused. Surgically removed, formalin-fixed paraffin-embedded tissue sections obtained from patients withother diseases, but with no autoimmune disease, wereimmunostained. Normal tissues (three cases each fromthe pancreas, extrahepatic biliary duct, salivary gland,oesophagus, stomach, small intestine, large intestineand lung) were obtained from patients having under-gone pancreatoduodenectomy for pancreatic cancer,hepatectomy for gallbladder cancer, sialoadenectomyfor adenoma, oesophagectomy for oesophageal cancer,gastrectomy for gastric cancer, enterectomy or colec-tomy for caecal cancer and pneumonectomy for lungcancer, respectively. Sections (3-lm thick) of threedifferent samples from one organ were immunoreactedovernight with sera from the AIP patients and controlpatients diluted 1 : 10 by 0.01 m phosphate-bufferedsaline, respectively. Sections were then incubated withmouse anti-human IgG4 monoclonal antibody, fol-lowed by the standard streptavidin–biotin–peroxidasemethod. The anti-human IgG4 monoclonal antibodylost its immunoreactivity following the absorption test
Table 2. Immunohistochemical analysis of tissues of autoimmune pancreatitis (AIP) patients and control groups
with IgG4. Eighteen slides for each tissue were reviewedin a blinded manner by two of the authors (S.A. andT.M.) and the extent of IgG4+ epithelial cells wascategorized as follows: score 0 ¼ 0–5%, score 1 ¼ 6–20%, score 2 ¼ 21–50%, and score 3 ‡ 50%. Whenthe assessment was different between the two observers,agreement was also reached using a double-headedmicroscope. Sera obtained from three AIP patients(cases 3, 4 and 5) before and after glucocorticoidtherapy were also used for the immunostaining ofnormal tissues to examine the influence of glucocorti-coid therapy and the IgG4+ cells were similarly scored.Prednisolone was initiated at 30 mg ⁄ day in all threecases and reduced by 5 mg every 2 weeks. Serum fromcases 3, 4 and 5 was obtained 8 weeks, 24 weeks and32 weeks, respectively, after initiation of the therapy.
All subjects provided written informed consent forlaboratory tests and invasive treatments such asresection, biopsy, endoscopic retrograde cholangiopan-creatography and percutaneous transhepatic cholagio-graphy.
statistical analysis
For statistical analysis of immunohistochemical resultsbetween the patients’ and control materials, the Mann–Whitney test was used to calculate two-sided P-values.For analysis of glucocorticoid therapy effect, the Wilc-oxon’s rank sum test for paired samples was used.P < 0.05 was considered statistically significant.
Results
immunohistochemical analysis of patients’
t issues
The results of immunohistochemistry are summarized inTable 2. Infiltration of IgG4+ plasma cells in and aroundthe pancreatic duct walls was observed in four of the fiveAIP patients (Figure 1A). Three patients had a score of 2or higher. One case scored 0. Because the specimen ofthis case was obtained by endoscopic ultrasound-guidedfine-needle aspiration biopsy, the specimen was very
Figure 1. IgG4 immunoreactivity of tissues obtained from autoimmune pancreatitis patients: IgG4+ plasma cell infiltration was observed.
A, In and around the pancreatic duct wall. B, Around the intrahepatic bile duct in the portal areas. C, Around the extrahepatic bile duct wall.
small. No damage to pancreatic duct epithelium in AIPwas noted. The scores of six cases with chronic calcifyingpancreatitis were significantly low.
In the liver of seven of the 11 patients, the limitingplates were destroyed and bridging fibrosis was noted infour patients. In all patients, infiltration of lymphocytesand plasma cells was seen around the intrahepatic bileduct in the portal areas, with no damage to bile ductepithelial cells. Infiltration of IgG4+ plasma cells wasobserved in all patients (Figure 1B). The scores of 10patients with PSC were significantly low.
IgG4+ plasma cell infiltration was observed in andaround the extrahepatic and intrapancreatic bile ductwalls of seven of the eight patients (Figure 1C), andfour of them had a score of 2 or higher, with nodamage to bile duct cells. One other patient had a scoreof 0, but insufficient tissue was obtained from thesepatients since biopsy was performed after endoscopicbiliary drainage for obstructive jaundice. In contrast,the scores of four cases with PSC were significantly low.
In the salivary gland, three patients showed abun-dant IgG4+ plasma cell infiltration with a score of 2
or 3. However, five patients with Sjogren’s syndromeshowed no infiltration by IgG4+ plasma cells.
IgG4+ plasma cells infiltrated the gastric mucosallayer of two of the seven patients, but none of themscored 3. Similar plasma cells infiltrated the duodenalmucosal layer of four of the six patients, but mostof them had a low score. IgG4+ plasma cells infiltra-ted the colonic mucosal layer of four of the sixpatients (Figure 1D) and two of them had a score of3. However, no significant difference in the immuno-reactivity between AIP and control groups in thegastrointestinal tract was apparent.
immunohistochemical analysis us ing
patients’ sera
Table 3 shows the reactivity of the patients’ serumautoantibody to various tissues. Table 4 shows theeffect of glucocorticoid therapy on the immunohisto-chemical analysis. In the pancreas, the duct epitheliumreacted positively with serum IgG4 from all of thepatients (Figure 2A). No islets or acinar cells stained
Table 3. Immunohistochemical scoring of tissues reactive with patients’ sera
positively. Imaging of all the patients with positiveserum showed an enlarged pancreas with diffusenarrowing of the main pancreatic duct. Stainabilityby the sera was significantly decreased after gluco-corticoid therapy compared with that before therapy.
In the liver, hepatocytes were weakly and diffuselystained. The intrahepatic bile duct epithelium was alsopositively stained with sera from five of the six patients(Figure 2B). Imaging showed intrahepatic bile ductstricture in two of the five positive-serum patients, butno abnormalities in the remaining three patients.Glucocorticoid therapy significantly decreased theimmunoreactivity.
The extrahepatic and intrapancreatic bile duct epi-thelia reacted with sera from all of the patients(Figure 2C). Imaging showed extrahepatic bile ductstricture in four of these six patients, but no abnor-malities in one patient. Glucocorticoid therapy signifi-cantly decreased the immunoreactivity.
In the gallbladder, the epithelial cells were positivelystained with sera from all of the patients (Figure 2D).Imaging showed gallbladder wall thickening in three
of them. Glucocorticoid therapy significantly decreasedthe immunoreactivity.
The duct epithelium in the salivary gland waspositive (Figure 3A) with the serum from each patient.The acinar cells were negative. Glucocorticoid therapysignificantly decreased the immunoreactivity.
In the oesophagus, the squamous epithelial cells andthe accessory gland duct epithelium were positive(Figure 3B) with the sera from two patients. Theaccessory gland acinar cells were negative. The stomach(Figure 3C) and lung were negligibly immunoreactive,and the small intestine and large intestine (Figure 3D)were unreactive. Decreased immunoreactivity followingglucocorticoid therapy was not apparent.
Immunostaining similarly performed with normalserum from the six controls was negative.
Discussion
The characteristic histological findings in AIP areperipancreatic duct inflammatory cell infiltration, pan-creatic acinar cell atrophy, pancreatic parenchymal
Table 4. Effect of glucocorticoid therapy on tissues reactive with patients’ sera
Before glucocorticoid therapy After glucocorticoid therapy
fibrosis, and obstructive phlebitis.18 Since extrapancre-atic organs have also been reported to be affected inAIP patients,13–15,19,20 we histologically and immuno-histochemically examined the extrapancreatic organsfor the development of lesions. Many IgG4+ plasmacells infiltrated in the pancreas, portal areas, extra-hepatic bile duct and salivary gland. In addition, theserum from AIP patients contained an IgG4 autoanti-body reactive with the epithelium of the normalpancreatic duct, bile ducts, gallbladder and salivarygland, which correlated relatively well with the clinicalfeatures of the patients. The autoantibody showedweak or no reactivity with the oesophagus, stomach,small and large intestines or lung.
As has been previously reported, hypergammaglob-ulinaemia or high serum IgG levels are observed in AIPpatients. Anti-nuclear antibodies were positive in78.9% of the AIP patients, which strongly suggeststhat they have an autoimmune disease. In addition,Hamano et al.12 found that serum IgG4 levels are
specifically high, as has been confirmed in this study.Kamisawa et al.21 examined IgG4+ plasma cells invarious organs of eight AIP patients, and reported thatthe pancreas, bile duct, gallbladder, liver and stomachwere infiltrated by IgG4+ plasma cells. Our results werelargely similar to theirs except for the finding of fewerIgG4+ plasma cells in the gastric and duodenalmucosae.
These findings suggest that AIP is a systemic diseasein which IgG4 autoantibodies to various organ anti-gens are present in the serum of patients. Kitagawaet al.11 investigated the presence of anti-pancreaticduct antibodies using an anti-IgG antibody. However,according to present knowledge, the anti-IgG4 anti-body is more specific. Furthermore, IgG4+ plasma cellinfiltration may not indicate that the infiltrated organ istargeted. In this study, in order to prove the presence ofan autoantibody to pancreas and ⁄ or extrapancreaticorgans, the patients’ serum was reacted with normaltissue, and the positive reaction was detected by
Figure 2. IgG4 immunoreactivity of normal tissues incubated with the serum of autoimmune pancreatitis patients in pancreas, liver,
extrahepatic bile duct and gallbladder: the duct epithelium in the pancreas (A), the intrahepatic bile duct epithelium in the liver (B),
the extrahepatic bile duct epithelium (C) and the epithelium in the gallbladder (D) were immunoreactive.
Immunohistochemical study of autoimmune pancreatitis 155
staining with an anti-IgG4 antibody. As a result, theepithelia of pancreatic duct, bile duct, gallbladder andsalivary gland duct in particular showed a strongreaction, suggesting that the patients’ sera contain anautoantibody to the respective epithelial cells, inducingan inflammatory cell infiltration.
It has been reported that AIP is frequently compli-cated by lower bile duct stricture, posing difficulty inthe distinction between AIP and pancreatic cancer.5,22
This stricture has been generally considered to be aconstriction due to pancreatic inflammation. On theother hand, intrahepatic bile duct stricture has alsobeen reported to complicate PSC.23 However, somedifferences between PSC and AIP exist. PSC is anintractable disease, whereas AIP-associated bile ductlesions are improved by glucocorticoid therapy.24
Shimanski et al.25 analysed the features of the pancre-atic duct in PSC patients, and reported that theprevalence of chronic pancreatitis in PSC patients was
low. The main histological feature of bile duct lesions inAIP is lymphoplasmacytic infiltration of the bile ductwall. In AIP, the bile duct epithelium is well preserved,with no onion-skinning of the hepatic interlobular bileducts as seen in PSC. These observations suggest thatalthough PSC- and AIP-associated bile duct lesionsmay show very similar features, they are differentdiseases. We have used the term ‘atypical PSC’ for AIPcomplicated by intrahepatic bile duct stricture, andhave proposed that it is a disease state different fromPSC.15 In the present study, the degree of infiltration ofIgG4+ plasma cells around the bile duct in the portalareas with AIP was significantly higher than with PSC.IgG4+ plasma cell infiltration around the extrahepaticbile duct as well as the intrapancreatic bile duct wasalso significantly higher. Moreover, the serum from fiveof six patients was positive for IgG4 antibody in the bileduct epithelium. These results also suggest that AIP is adisease state clearly different from PSC.
Figure 3. IgG4 immunoreactivity of normal tissues incubated with the serum of autoimmune pancreatitis patients in salivary gland,
oesophagus, stomach and large intestine. The duct epithelium in the salivary gland was immunoreactive (A). The squamous epithelium and the
accessory gland duct epithelium in the oesophagus were weakly reactive (B). The mucosal epithelium in the stomach was negligibly stained (C).
The mucosal epithelium in the large intestine was unreactive (D).
It has been reported that AIP is frequently associatedwith thickening of the gallbladder wall. Abrahamet al.26 studied the gallbladder of patients with lympho-plasmacytic sclerosing pancreatitis (AIP), PSC, chroniccholelithiasis and benign pancreatic diseases, andreported that diffuse lymphoplasmacytic chronic chole-cystitis was present in 50% of PSC cases and 25% ofAIP cases, but in only 5% of chronic cholelithiasis andnone of non-AIP benign pancreatic diseases. In thepresent study, the successful detection of IgG4 auto-antibody reacting with the gallbladder epitheliumstrongly suggests that the gallbladder is affected bythe disease.
Previous studies have reported on salivary glandlesions in Sjogren’s syndrome and chronic pancreati-tis.27,28 However, 90% of patients with Sjogren’ssyndrome are middle-aged women, whereas AIP iscommon in elderly men. Hamano et al.12 have reportedthat serum IgG4 levels in Sjogren’s syndrome patientsare significantly lower than those in AIP patients,and Kamisawa et al.14 stained the salivary gland ofSjogren’s syndrome patients with an anti-IgG4 anti-body, to find that plasma cells were negative for IgG4.We have also stained the salivary glands of Sjogren’ssyndrome patients with an anti-IgG4 antibody and theresults were the same as those described by Kamisawaet al.14 In the present study, IgG4+ plasma cellinfiltration in the salivary gland was prominent inthree patients and the sera from AIP patients stronglyreacted with the normal salivary gland epithelium.Thus, AIP may affect the salivary gland by a mechan-ism different from that of Sjogren’s syndrome.
Kamisawa et al.21 performed IgG4 staining of thelarge intestine in two AIP patients, and observed strongor moderate staining. In the present study, we similarlyobserved score 3 infiltration of IgG4+ plasma cells intwo of the six patients. However, immunostaining withthe AIP patients’ serum was negative in the normalsmall and large intestines. More case studies areneeded, since a study has reported AIP complicatedby inflammatory bowel disease.29 On the other hand,there have been a few clinical reports of stomachlesions complicating AIP. Shinji et al.30 reported thatgastric ulcer was found significantly more frequently inpatients with autoimmune pancreatitis compared withcontrol patients. In the present study, we observed low-grade infiltration of IgG4+ plasma cells in the patients’tissue and a marginal reaction of the patients’ serumwith the normal stomach. Nevertheless, the develop-ment of lesions in the stomach or large intestine, even ifthey coexist with AIP, may produce scanty symptomsbecause of the wider lumen compared with that of thebile duct.
Many studies have reported that glucocorticoidtherapy is effective for AIP patients.6,20,27,31,32
Although the number of AIP patients receiving gluco-corticoid therapy was small in the present study, thereaction of the patients’ sera with normal epithelia ofthe pancreatic duct, bile duct, gallbladder and salivarygland duct significantly became weak after glucocorti-coid therapy. This finding suggests a decrease ordisappearance of autoantibody, and correlates wellwith the improvement in clinical symptoms of AIP afterglucocorticoid therapy.
The reactivity of the patients’ serum IgG4 in thepresent study with pancreatic and extrapancreatictissues ⁄ organs suggests the presence of a specificantigen at these sites. Spicer et al.33 examined thetissue distribution of carbonic anhydrase (CA), andreported that the bile duct in the liver, epithelial cells inthe gallbladder, duct epithelial cells in the pancreas,and epithelial cells in the colon and appendix werestrongly stained for CA. Kitagawa et al.11 investigatedthe presence of anti-pancreatic duct antibody in theserum of an AIP patient by Western blotting using ananti-CAII antibody, and reported the detection of aband at the same position as the CAII band. However,since the anti-CAII antibody itself has also beenreported to be detected in approximately 30% ofpatients with Sjogren’s syndrome or systemic lupuserythematosus,9 the antigen is not necessarily specificfor AIP patients. AIP-specific antigens remain to beidentified.
In conclusion, the present immunohistochemicalstudy has examined not only the pancreas but alsothe liver, biliary system, salivary gland, gastrointestinaltract and lung tissues, using patients’ tissues or serum,and has shown the presence in the serum of anepithelial cell-specific autoantibody that belongs toIgG4. The use of the sera from AIP patients may help topredict whether the pancreas alone will remain affec-ted, or what extrapancreatic organs will also beaffected. Studies of more patients will contribute tothe elucidation of the pathogenesis of AIP.
Acknowledgements
We are indebted to Mr Seizoh Nagaya of the Depart-ment of Clinical Pathology for technical assistance withhistological analysis.
References
1. Sarles H, Sarles JC, Camatte R et al. Observations on 205
confirmed cases of acute pancreatitis, recurring pancreatitis, and
chronic pancreatitis. Gut 1965; 6; 545–559.
Immunohistochemical study of autoimmune pancreatitis 157
