Basophils activated via TLR signaling may contribute to pathophysiology of type 1 autoimmune pancreatitis

Background Pathophysiology of type 1 autoimmune pancreatitis (AIP) is still unclear. We previously reported that M2 macrophages might play an important role in type 1 AIP. Recently, it has been reported that basophils regulate differentiation to M2 macrophages. In this study, we investigated basophils from the pancreatic tissue and peripheral blood of individuals with type 1 AIP. Methods By using immunohistochemistry, we investigated basophils in pancreatic tissue from 13 patients with type 1 AIP and examined expression of toll-like receptors (TLRs) by these cells. Additionally, we obtained peripheral blood samples from 27 healthy subjects, 40 patients with type 1 AIP, 8 patients with alcoholic chronic pancreatitis, 10 patients with bronchial asthma, and 10 patients with atopic dermatitis, and analyzed activation of basophils by stimulating them with ligands of TLR1–9. We also compared TLR expression in basophils from the tissue and blood samples. Results Basophils were detected in pancreatic tissues from 10 of 13 patients with type 1 AIP. Flow cytometric analysis revealed that the ratios of basophils activated by TLR4 stimulation in type 1 AIP (9.875 ± 1.148%) and atopic dermatitis (11.768 ± 1.899%) were significantly higher than those in healthy subjects (5.051 ± 0.730%; P < 0.05). Levels of basophils activated by TLR2 stimulation were higher in seven type 1 AIP cases. Furthermore, stimulation of TLR2 and/or TLR4, which were expressed by basophils in pancreas, activated basophils in peripheral blood. Conclusions Basophils activated via TLR signaling may play an important role in the pathophysiology of type 1 AIP.

Results Basophils were detected in pancreatic tissues from 10 of 13 patients with type 1 AIP. Flow cytometric analysis revealed that the ratios of basophils activated by TLR4 stimulation in type 1 AIP (9.875 ± 1.148%) and atopic dermatitis (11.768 ± 1.899%) were significantly higher than those in healthy subjects (5.051 ± 0.730%; P \ 0.05). Levels of basophils activated by TLR2 stimulation were higher in seven type 1 AIP cases. Furthermore, stimulation of TLR2 and/or TLR4, which were expressed by basophils in pancreas, activated basophils in peripheral blood. Conclusions Basophils activated via TLR signaling may play an important role in the pathophysiology of type 1 AIP.

Introduction
Autoimmune pancreatitis (AIP) is characterized by chronic pancreatic inflammation caused by immune mechanisms [1,2]. Steroid treatment is remarkably effective for AIP, which may be sometimes difficult to be differentiated from pancreatic carcinoma [3]. The first report of AIP was a case of chronic pancreatitis with hypergammaglobulinemia that was published in 1961 [4]. Thirty years later, this pathological syndrome was defined as lymphoplasmacytic sclerosing pancreatitis (LPSP) [5], and proposed to be referred to as ''autoimmune pancreatitis'' by Yoshida et al. in 1995 [6]. Hamano et al. revealed that serum IgG4 levels were elevated in patients with AIP [7], and this condition is currently classified as type 1 AIP. On the other hand, Notohara et al. reported a different pathological condition, which was defined as idiopathic duct-centric pancreatitis (IDCP), which is classified as type 2 AIP [8]. In 2011, the international consensus diagnostic criteria (ICDC) for AIP classified two subtypes of AIP, type 1 and type 2 [9]. Type 1 AIP is recognized as the pancreatic manifestation of IgG4-related disease (IgG4-RD), which involves various organs, including the pancreas, salivary glands, thyroid glands, lungs, bile ducts, kidneys, retroperitoneum, lymph nodes, and others [10][11][12][13]. In serology, most patients with type 1 AIP exhibit elevated serum IgG4 levels [7], and some of them show elevated serum IgE [14,15]. Histological features of type 1 AIP include diffuse lymphoplasmacytic and eosinophilic infiltrates, storiform fibrosis, obliterative phlebitis, and the infiltration of a large number of IgG4-positive plasma cells [9]. Regulatory T cells and T helper 2 (Th2) cytokines, such as IL-4, IL-5, and IL-13, are considered to contribute to pathophysiology of type 1 AIP [16,17]. We have previously reported that the number of regulatory T cells is increased in peripheral blood and infiltrated target organs of type 1 AIP [18]. This particularly concerns inducible T cell co-stimulator (ICOS)-positive regulatory T cells, which promote IgG4 production via IL-10, and ICOS-negative regulatory T cells, which induce fibrosis via transforming growth factor (TGF)-b [18]. Furthermore, the pathophysiology of type 1 AIP is characterized by Th2-dominant reaction, but its mechanism has not yet been elucidated [19]. Our latest report suggested that M2 macrophages play an important role in the development of type 1 AIP [20]. IL-33 produced by M2 macrophages also promotes Th2 cytokine production conferring Th2-dominant pathophysiology in IgG4-RD [21]. Additionally, numerous macrophages were identified in AIP-affected tissue [22]. On the other hand, basophils were found to be required for the transformation of inflammatory monocytes into M2 macrophages in allergic disease [23]. However, the role of basophils in type 1 AIP is not clear.
In this study, to clarify the involvement of basophils activated via the toll-like receptor (TLR) signaling pathway in type 1 AIP, we investigated infiltration of TLR-expressing basophils in pancreatic tissue and their reactivity upon TLR stimulation in patients with type 1 AIP.

Methods
In our experiments, we initially sought to confirm the presence of basophils in pancreatic tissue and then investigated whether they expressed TLRs. Furthermore, we examined the responses of basophils to TLR stimulation using peripheral blood mononuclear cells from patients with type 1 AIP.

Subjects
For immunohistochemical analysis, we examined pancreatic tissue samples from 13 patients with type 1 AIP (6 men and 7 women; mean age 66 years; range 56-76 years).
We also conducted immunofluorescence analysis in 10 of 13 cases. All cases were surgically treated at the Kansai Medical University Hospital between 2001 and 2013, and in all cases, pancreatic ductal adenocarcinoma was suspected before the operation. The patients were histopathologically diagnosed with type 1 AIP. For comparison, we also examined pancreatic tissue samples from 10 patients with alcoholic chronic pancreatitis (ACP). All the 10 ACP patients underwent surgical operation.
For flow cytometry analysis, we recruited 27 healthy subjects (17 men and 10 women; mean age 66 years; range 42-83 years), 40 patients with type 1 AIP, who had not received corticosteroid treatment (29 men and 11 women; mean age 65 years; range 44-83 years), 8 patients with ACP (6 men and 2 women; mean age 57 years; range 39-69 years), 10 patients with bronchial asthma, who did not receive oral medication (6 men and 4 women; mean age 52 years; range 39-74 years), and 10 patients with atopic dermatitis, who did not receive oral medication (8 men and 2 women; mean age 49 years; range 27-77 years). Patient profiles are listed in Table 1. There was no difference in gender distribution between type 1 AIP and healthy subjects. The mean ages of bronchial asthma patients and atopic dermatitis patients were lower than those of patients with type 1 AIP and healthy subjects (P \ 0.05). However, ages of patients with type 1 AIP, ACP, and healthy subjects were not significantly different. The mean values of serum IgG4 levels in type 1 AIP patients were significantly higher than in other groups (P \ 0.05; Table 1). The numbers of basophils in the peripheral blood were not different in patients with type 1 AIP, ACP, and healthy subjects. All patients with type 1 AIP were diagnosed with AIP by the ICDC [9].
This study was approved by the ethics committee of the Kansai Medical University.

Immunohistochemistry and antibodies
We used mouse antibodies specific to human basophils (2D7; Abcam, Cambridge, UK) in order to identify basophils by immunohistochemical staining. The fields of view with the highest density of basophils were evaluated, and the numbers of basophils in three different fields of 1 mm 2 each were summed. Next, we examined the expression of TLRs on infiltrated basophils using double-immunofluorescence staining of pancreatic tissue from patients with type 1 AIP with mouse antibodies specific to human basophils and rabbit antibodies specific to human TLR1-9.
For immunohistochemical staining, 4-lm-thick formalin-fixed and paraffin-embedded sections were prepared, deparaffinized, and rehydrated using xylene and solutions with gradually decreasing concentrations of alcohol. Endogenous peroxidase activity was blocked for immunohistochemical sections using a treatment with 3% H 2 O 2 in methanol for 10 min at room temperature. After washing in distilled water, the slides were treated with 0.05% trypsin (Abcam) in PBS (pH 7.3) for 1 h at room temperature to enhance antigenicity. All slides were incubated for 10 min in Protein Block Serum-Free (Dako, Kyoto, Japan). After protein blocking, the slides were incubated overnight at 4°C with mouse primary antibodies specific to human basophils (2D7). The primary antibodies were dissolved in Signal Enhancer HIKARI (Nacalai Tesque, Kyoto, Japan). The slides were then incubated with secondary antibodies using a Chem Envision kit/HRP (Dako) following the manufacturer's instructions. Finally, antibody binding was detected using 3,3 0 -diaminobenzidine (Dojindo, Kumamoto, Japan). Sections were counterstained with hematoxylin. Negative controls were evaluated by replacing the primary antibodies with similarly diluted non-immunized serum. Images were obtained with a light microscope (DP73; Olympus, Tokyo, Japan).
In the double-immunofluorescence method, the slides were incubated with mouse antibodies specific to human basophils (2D7) and rabbit antibodies specific to human TLR1-9 (TLR1: AbFrontier, Seoul, Korea; TLR2: Novus, Littleton, CO, USA; TLR3: Novus; TLR4: StressMarq Biosciences, Victoria, Canada; TLR5: AbFrontier; TLR6: AbFrontier; TLR7: AbFrontier; TLR8: AbFrontier; TLR9: GeneTex, Irvine, CA, USA) as the primary antibodies overnight at 4°C. The slides were then incubated with Alexa Fluor 546 anti-mouse immunoglobulins and Alexa Fluor 488 anti-rabbit immunoglobulins (Molecular Probes, Eugene, OR, USA) as secondary antibodies for 1 h at room temperature, protected from light. After incubation with the primary and secondary antibodies, slides were mounted with VECTASHIELD Mounting Medium and 4 0 ,6-diamidino-2-phenylindole (DAPI; Vector Laboratories, Burlingame, CA, USA) to counterstain nuclei and preserve fluorescence. Negative controls were evaluated by replacing the primary antibodies with similarly diluted non-immunized serum. Images were obtained with an immunofluorescence microscope (DP73; Olympus). Additionally, we were able to obtain both the pancreatic tissues and blood from 5 out of the 10 patients, whose samples were used for immunofluorescence analysis, so we examined the relationship of each TLR.

Flow cytometry and antibodies
We analyzed the activation of basophils using Allergenicity Kits (Beckman Coulter Company, Brea, CA, USA). The There was no difference in gender distribution between type 1 AIP and healthy subjects. The mean ages of bronchial asthma patients and atopic dermatitis patients were lower than those of patients with type 1 AIP and healthy subjects (P \ 0.05). However, ages of patients with type 1 AIP, ACP, and healthy subjects were not significantly different. The mean values of serum IgG4 levels in type 1 AIP patients were significantly higher than in other groups (P \ 0.05) NS no significant differences compared with healthy or other disease controls * P \ 0.05 significant differences compared with healthy controls J Gastroenterol (2018) 53:449-460 451 analysis of whole-blood specimens collected with EDTA as an anticoagulant was performed according to the manufacturer's instructions. Whole-blood specimens were stimulated with TLR1-9 ligands Pam 3 CSK4 (400 ng/mL, InvivoGen, San Diego, CA, USA), HKLM (10 8 cells/mL, InvivoGen), poly:IC (50 lg/mL, Sigma, St. Louis, MO, USA), LPS (1 lg/mL, Sigma), FLA-ST (10 lg/mL, Invi-voGen), FSL-1 (5 lg/mL, InvivoGen), Imiquimod (5 lg/ mL, InvivoGen), ssRNA40/LyoVec (5 lg/mL, InvivoGen) or ODN2006 (10 lg/mL, AdipoGen, San Diego, CA, USA), respectively. First, we mixed 100 lL of wholeblood specimens, activation solution, and three color reagents (CRTH2-FITC: BM16, CD203c-PE: 97A6, and CD3-PC7: UCHT1) with 10 lL of a TLR ligand or PBS in each tube. Activation solution is an optimized calciumenriched buffer to activate basophils in vitro. Then, we incubated each tube for 15 min at 37°C in the atmosphere of 95% air/5% carbon dioxide, protected from light. After staining, we added 2 mL of Fix-and-Lyse solution, which included the Allergenicity Kit, to remove erythrocytes interference. We incubated the tubes for another 10 min at room temperature, protected from light. We then centrifuged the tubes for 5 min at 200 g and aspirated the supernatant twice. Finally, we re-suspended the cells in 0.3 mL of PBS with 0.1% formaldehyde and analyzed them using flow cytometry. Flow cytometric analysis was performed by FACS Calibur II (BD Biosciences, Franklin Lakes, NJ, USA). We counted the ratio of activated basophils in accordance with the manufacturer's instructions for the Allergenicity Kit (Fig. 1).
First, lymphocytes and monocytes were isolated, whereas CD3-positive events were excluded. Next, basophils were isolated as CD294 (CRTH2)-positive events from lymphocytes and monocytes. Finally, we drew line A so that 2-5% of basophils were included above it, and defined them as CD203c-positive basophils, because 2-5% of basophils are activated basophils in normal conditions. CD203c has been demonstrated to be a specific activation marker of basophils. Additionally, we compared the ratio of CD203c-positive basophils before and after TLR stimulation in each group.

Statistical analysis
In immunohistochemistry and flow cytometry experiments, quantitative data are presented as the mean ± standard error of the mean (SE). The Tukey-Kramer test was used to compare quantitative values in flow cytometry experiments. Differences were considered to be statistically significant when the value of P was less than 0.05 (* P \ 0.05).

Results
Identification of basophils in pancreatic tissue from patients with type 1 AIP Basophils (2D7-positive cells) were detected in the pancreatic tissue samples from 10 of 13 patients with type 1 AIP cases by immunohistochemistry (Fig. 2). In contrast, 2D7-positive cells were not detected in the pancreatic tissue from patients with ACP cases (Fig. 3). The average number of 2D7-positive cells was 8.615 ± 2.528, measured in three different 1-mm 2 fields of view in samples from each of the 13 cases (Fig. 4).
Relationship between TLR expression by basophils from pancreatic tissue and activation of peripheral blood basophils upon stimulation of TLR1-9 We examined the relationship between TLR expression by basophils from pancreatic tissue and activation of basophils in the peripheral blood by stimulating them with TLR1-9 ligands. We found that the levels of activated basophils in the peripheral blood were elevated by TLR2 stimulation in the two patients (case 1 and 2) that exhibited TLR2-positive basophils in the pancreatic tissue. In one case (case 3), which exhibited both TLR2-and TLR4-positive basophils in the pancreatic tissue, levels of activated basophils in the peripheral blood of this patient were elevated by only both TLR2 and TLR4 stimulation. In the two cases (case 5 and 6) that exhibited TLR4-positive basophils in the pancreatic tissue, levels of activated basophils in peripheral blood were elevated by TLR4 stimulation, but not other types of TLR (Table 3). Thus, in patients with type 1 AIP, stimulation of the TLR that is expressed on the basophils in the pancreatic tissue seems to activate circulating basophils.

Discussion
In this study, we identified the presence of basophils in pancreatic tissue of patients with type 1 AIP. We also found that the ratio of basophils activated by TLR4 signaling in the peripheral blood of patients with type 1 AIP was significantly higher than that in healthy subjects. In addition, the ratio of basophils activated by TLR2 signaling was elevated in 7 of 40 cases of type 1 AIP (Fig. 6). Remarkably, we revealed that stimulation of the type of TLR that was expressed on infiltrated basophils in type 1 AIP-affected pancreatic tissue also activated circulating basophils (Table 3). On the other hands, basophils were not detected in the pancreatic tissue with ACP, and the ratio of basophils activated by TLR signaling in the peripheral blood of patients with ACP were not significantly different as compared with healthy subjects. Furthermore, the ratio Fig. 1 Identification of resting and activated basophils from peripheral blood by flow cytometry. All leukocytes were stained with CD3, CD294, and CD203c. All leukocytes formed discrete FSC/SSC populations after fixation and lysis (a). CD3positive events (R1) were removed due to the isolation of lymphocytes and monocytes (b). Basophils were defined as CD294 (CRTH2)-positive events (R2) (c). CD203c is a specific marker of activated basophils. Line A was drawn so that 2-5% of basophils were included above it, and we defined them as activated basophils. We compared the percentage of basophils above line A before and after the stimulation (d). These experiments were performed using the Allergenicity Kit of TLR4-activated basophils in patients with type 1 AIP tended to be elevated compared to ACP. Therefore, we speculated that basophils activated by TLR signaling were important in the development of type 1 AIP. Basophils comprise less than 1% of human peripheral blood leukocytes and can live only for a few days in the non-activated condition [24]. However, it has been shown that 10-times higher levels of Th2 cytokines such as IL4 and IL13 are immediately produced compared with those in lymphocytes, even when a small percentage of basophils are activated [25,26]. They are generally not present in normal tissue and become recruited to affected tissue sites only under certain conditions, for example, during allergic reactions [27]. It has been shown that basophils act as initiators of inflammatory cell recruitment during the progression of IgE-mediated chronic allergic inflammation [28]. It has also been reported that basophils in skin lesions from atopic dermatitis were detected in approximately 60% of patients; however, the cell density was low as compared with patch-tested lesions [29]. Patients with type 1 AIP have frequent complications of allergic conditions, such as bronchial asthma or allergic dermatitis, characterized by elevated serum IgE and eosinophilia [21,30]. In this study, we identified basophils in pancreatic tissue samples from 10 of 13 patients with type 1 AIP (Fig. 4), but the cell density was low. This finding indicates that pathophysiology of type 1 AIP may be similar to that of an allergic disease, such as atopic dermatitis. We, therefore, speculate that basophils infiltrating in the tissues might play an important role in the Th2-dominant condition of type 1 AIP.
Basophil-derived IL-4 is considered to be involved in the recruitment of monocytes and their differentiation to M2 macrophages in allergic skin [23]. Macrophages are composed of at least two distinct groups of M1 and M2 phenotypes. M1 macrophages elicit an inflammatory responses and play a central role in host defense against bacterial and viral infections, whereas M2 macrophages play roles in anti-inflammatory responses, reactions to parasitic infections, as well as in tissue repair and remodeling [31]. M2 macrophages induced by IL-4 are involved in fibrosis as they produce IL-10 and CCL18 [32]. Notohara et al. reported that CD163-positive spindle-shaped macrophages contribute to LPSP manifestations, such as storiform fibrosis in type 1 AIP [33]. We previously reported that macrophages (especially CD163-positive cells), stimulated via TLR signaling pathways, might play an important role in pancreatic tissue from patients with type 1 AIP [20]. Furthermore, the number of M2 macrophages in salivary glands of individuals with IgG4-RD was significantly higher than that in healthy subjects, and the distribution of IL-10 and CCL18 closely paralleled that of M2 macrophages [34]. These findings suggest that basophils induce M2 macrophages, which, in turn, may contribute to fibrosis in the pancreatic tissue affected by type 1 AIP. It is known that Th2 cytokine production is elevated in pancreatic tissue of type 1 AIP patients, but the mechanism of this phenomenon has not yet been clarified [35]. It was recently demonstrated that endogenous IL-33 promoted Th2 cytokine production and induced allergic inflammation during allergic airway inflammation [36]. IL-33 produced by M2 macrophages promoted Th2 cytokine production via IL-33 receptor activation and contributed to Th2-dominant pathophysiology in IgG4-RD [37]. Another study revealed that basophils rapidly produced large amounts of IL-4 in response to various stimuli, including activation by ligands of TLRs, in innate-type allergy, and could induce differentiation of Th2 cells without cross-linkage of FceRI [38]. On the basis of these reports and our present data, we speculate that basophils are important cells in Th2-dominant pathophysiology of type 1 AIP.
Additionally, we identified the presence of TLR2-and/ or TLR4-positive basophils in pancreatic tissue from  (Table 2). We could confirm that basophils express TLR2 or TLR4 in this study, but some researchers reported that antigen presenting cells such as dendritic cells, macrophages, B cells, and T cells express TLR2 or TLR4 [39][40][41]. The presence of basophils expressing different TLRs (TLR2 and/or TLR4) in the pancreatic tissue suggests that pathophysiology of type 1 AIP may be caused by a heterogeneous inflammatory condition.
There was no difference of the clinical data (blood test results, location of pancreatic swelling, or extent of involvement of other organs) between the TLR2-positive group and TLR4-positive group in AIP patients. There was also no correlation between the ratio of basophils activated by TLR signaling and serum IgG4 and IgE levels (data not shown). However, Watanabe et al. reported that basophils activated via TLR2 or TLR4 signaling enhanced the production of IgG4 through a BAFF-mediated signaling pathway in IgG4-RD [42]. On the other hand, there was no correlation between serum IgG4 levels and the number of IgG4-positive cells in AIP patients (data not shown). Paik et al. also reported that there was no correlation between serum and tissue IgG4 levels [43]. Basophils that infiltrated into pancreatic tissue of patients with type 1 AIP may also contribute to IgG4 production.
TLRs are pattern recognition receptors that recognize pathogen-associated molecular patterns (PAMPs) or damage-associated molecular patterns (DAMPs). PAMPs are associated with microbial pathogens, whereas DAMPs are linked to cell components, which are released during cell damage or death, to induce the innate immune response [43]. It is well known that PAMPs derived from abundant gut bacteria are transferred to the serum and bone marrow, where they promote systemic innate immunity [44]. It is also possible that DAMPs activate basophils in pancreatic tissue from patients with type 1 AIP. DAMPs of TLR2 and TLR4 include biglycan, high-mobility group box protein 1 (HMGB1), heat shock protein 70, fibronectin, and others [45]. They have been shown to play roles in the induction of experimental pancreatitis [46]. HMGB1, released by necrotic acinar cells, induces tissue injury and inflammation via TLR4 activation [47,48]. As a result of acinar cell necrosis, HMGB1 may activate basophils via TLR4 in pancreatic tissue of type 1 AIP patients. HMGB1 and selfnucleic acid complexes are also involved in the development of chronic inflammatory disease [49,50]. Thus,  Gender TLR1 TLR2 TLR3 TLR4 TLR5 TLR6 TLR7 TLR8 TLR9   1  6 3  M  -?  -------2  7 1  F  -?  -------3  7 8  F  -?  -?  -----4  6 7  M  -?  -? ----- Case no. is the same as those in Table 2. No. 4, 7-10 have not been tested by flow cytometry In case 1 and case 2, which exhibited TLR2-positive basophils in the pancreatic tissue, levels of activated basophils in the peripheral blood of these patients were elevated by TLR2 stimulation. In case 3, which exhibited both TLR2-and TLR4-positive basophils in the pancreatic tissue, levels of activated basophils in peripheral blood were elevated by only both TLR2 and TLR4 stimulation. In case 5 and case 6, that exhibited TLR4-positive basophils in the pancreatic tissue, levels of activated basophils in peripheral blood were elevated by TLR4 stimulation, but not other types of TLR  [51,52,53]. In the future, it will be necessary to clarify the differences in the properties of basophils activated via different TLRs in order to better understand pancreatitis pathophysiology.
In conclusion, our results demonstrate that together with M2 macrophages, basophils activated via TLR signaling may also play an important role in pathophysiology of type 1 AIP. Acknowledgements This study was partially supported by a Grantin-Aid for Scientific Research of the Ministry of Culture and Science of Japan (17K09468, 15K09052), by the Research Program on Intractable Diseases from the Ministry of Labor and Welfare of Japan, and by Grants-in-Aid from the Ministry of Education, Culture, Sports, Science and Technology of Japan from the CREST Japan Science and Technology Agency. We thank professor A-Hon Kwon for pancreatic surgery. We warmly thank Mr. Gonda for technical advice about flow cytometry data analysis.

Compliance with ethical standards
Conflict of interest The authors declare that they have no conflicts of interest.