Role of eosinophils in the initiation and progression of pancreatitis pathogenesis

Abstract

Eosinophilic pancreatitis (EP) is reported in humans; however, the etiology and role of eosinophils in EP pathogenesis are poorly understood and not well explored. Therefore, it is interesting to examine the role of eosinophils in the initiation and progression of pancreatitis pathogenesis. Accordingly, we performed anti-major basic protein immunostaining, chloroacetate esterase, and Masson’s trichrome analyses to detect eosinophils, mast cells, and collagen in the tissue sections of mouse and human pancreas. Induced eosinophils accumulation and degranulation were observed in the tissue sections of human pancreatitis, compared with no eosinophils in the normal pancreatic tissue sections. Similarly, we observed induced tissue eosinophilia along with mast cells and acinar cells atrophy in cerulein-induced mouse model of chronic pancreatitis. Additionally, qPCR and ELISA analyses detected induced transcript and protein levels of proinflammatory and profibrotic cytokines, chemokines like IL 5, IL-18, eotaxin-1, eotaxin-2, TGF-β1, collagen-1, collagen-3, fibronectin, and α-SMA in experimental pancreatitis. Mechanistically, we show that eosinophil-deficient GATA1 and endogenous IL-5-deficient mice were protected from the induction of proinflammatory and profibrotic cytokines, chemokines, tissue eosinophilia, and mast cells in a cerulein-induced murine model of pancreatitis. These human and experimental data indicate that eosinophil accumulation and degranulation may have a critical role in promoting pancreatitis pathogenesis including fibrosis. Taken together, eosinophil tissue accumulation needs appropriate attention to understand and restrict the progression of pancreatitis pathogenesis in humans.

NEW & NOTEWORTHY The present study for the first time shows that eosinophils accumulate in the pancreas and promote disease pathogenesis, including fibrosis in earlier reported cerulein-induced experimental models of pancreatitis. Importantly, we show that GATA-1 and IL-5 deficiency protects mice form the induction of eosinophil active chemokines, and profibrotic cytokines, including accumulation of tissue collagen in an experimental model of pancreatitis. Additionally, we state that cerulein-induced chronic pancreatitis is independent of blood eosinophilia.

INTRODUCTION

Pancreatitis is defined as an acute or chronic inflammatory process of the pancreas and characterized by the induction of proinflammatory cytokines, chemokines, and tissue recruitment of inflammatory cells, including eosinophils (22–24, 60). Several clinical reports have indicated induction of eosinophils in the pancreatic biopsies of patients, and the disease is called eosinophilic pancreatitis (EP) (1, 4, 18, 20, 44, 50). Induced IL-18 is reported in the blood of both acute (45) and chronic pancreatitis (43, 59) and implicated in the development of pancreatic remodeling and fibrosis (43).

Recently, we and others implicated tissue-specific overexpression of IL-18 in the induction of tissue eosinophilia and fibrosis (8, 16, 31). Some clinical reports suggest that EP mimics pancreatic neoplasia, as a number of eosinophils are detected in pancreatic tumor (4, 13, 40). Most recently, a microscopic examination of a pancreatic biopsy revealed that eosinophil infiltration occurs into the pancreatic duct, acini, and interstitium and shows fibrous connective tissue hyperplasia and accumulation of collagen (44).

Additionally, a clinical report also implicates EP in eosinophilic gastroenteritis (21). However, no attempt has been made to understand the mechanistic role of IL-18 and eosinophil induction in experimental and human pancreatitis, including pancreatic malignancy. Accordingly, we tested the hypothesis that IL-18 synergies with endogenous IL-5 may have an important role in promoting pancreatic eosinophilia and its associated remodeling and fibrosis that may promote pancreatic malignancy. Herein, we report the accumulation of tissue eosinophils and highly induced levels of tissue IL-18 with mild induction of IL-5 in the cerulein-induced experimental pancreatitis in mice. The intensity of eosinophilia and degranulation in the tissue was associated with the severity of the disease in cerulein-induced chronic pancreatitis.

Several investigators have reported EP in patients, but it is still referred to as a rare disease. The possibility of eosinophil accumulation in the pancreas has not been studied, as most pancreatic biopsies are performed using endoscopic ultrasound with fine-needle aspirate, with the purpose to detect whether patients have developed pancreatic malignancy. These procedures do not provide sufficient tissue for pathological examination; therefore, the EP in patients may not be detected in most hospitals and clinics, meaning it is ignored and termed as a rare disease. Taken together, our present experimental approach for the first time shows that EP may not be a rare disease and that it needs appropriate attention to understand the mechanistic role of eosinophils in promoting pancreatitis pathogenesis, including fibrosis. Notably, pancreatic fibrosis is the major concern for failed therapies in chronic pancreatitis and pancreatic malignancy.

MATERIALS AND METHODS

Patient tissue samples.

Human pancreatic tissue samples of chronic pancreatitis (n = 3) and normal pancreas (n = 2) were obtained from the Biospecimen Core Facility, Louisiana Cancer Research Consortium. The specimens from normal portions of human nonmalignant pancreatic patients showed no histopathological abnormalities, and the samples from patients with chronic pancreatitis analyzed showed no visible tumor. Patients’ pancreatic tissues were collected during surgical resection of pancreatic tumors and surgical procedures performed in chronic pancreatitis. Approximately 100-mg segments of pancreatic tissue were taken and immediately frozen in liquid nitrogen or 4% formaldehyde. All fixed tissues were used for paraffin embedding, sectioning, and processing for immunostaining and routine light microscopy. Written, informed consent was obtained from each patient per the Institutional Review Board approval for the study. The details of each patient’s clinical characteristics are provided in Table 1.

Table 1.

Details of human pancreatic biopsies

S. No.	Age	Sex	Lymphovascular Invasion	Perineural Invasion	% Tumor	Pathological Status
1	45	M	—	—	—	Normal tissue
2	63	F	—	—	—	Normal tissue
3	50	F	—	—	—	Chronic pancreatitis
4	67	F	—	—	—	Chronic pancreatitis
5	63	M	—	ND	—	Chronic pancreatitis

M, male; F, female; ND, not defined; —, absent.

Mice.

Specific pathogen-free Balb/c mice (wild-type) and GATA1 gene-deficient mice were obtained from Jackson Laboratory (Bar Harbor, ME). IL-5 gene-deficient Balb/c background mice were obtained from the laboratory of Marc Rothenberg, MD, PhD (Cincinnati Children’s Hospital Medical Center, Cincinnati, OH). The mice were maintained in a pathogen-free barrier facility. All experimental mice were age (6–8 wk) and sex matched. The Institutional Animal Care and Use Committee (IACUC) approved the animal protocol in accordance with the National Institutes of Health guidelines. We performed all experiments according to animal ethics rules and regulations.

Experimental pancreatitis.

Chronic pancreatitis was induced by repetitive cerulein injections per previously described protocol (52, 53). In brief, cerulein (Sigma-Aldrich, St. Louis, MO) was given by repetitive intraperitoneal (i.p.) injections (50 μg/kg, 6 hourly injections/day; 3 days/wk) for up to 4 wk; control mice received 100 μl saline. Mice were then euthanized 3 days after the last cerulein injection, and the pancreas was collected for further experiments. Characteristic features that mimic human and mouse chronic pancreatitis include acinar cell atrophy, increased infiltration of inflammatory cells, increased number of tissue mast cells, macrophages, fibrosis, and induced IL-18, including proinflammatory Th2 cytokines (14, 24, 37, 43, 53).

Histopathological analysis.

Mouse pancreatic tissue specimens were fixed with 4% paraformaldehyde and embedded in paraffin using standard techniques. The paraffin-embedded sections (5 μm) were stained with hematoxylin-eosin to determine the histopathological characteristic features in pancreatitis compared with control tissue sections of mice. Tissue section slides were coded, and four to five randomly chosen microscopic fields were graded blindly on the scale of 0 (absent) to 4 (severe). The parameters included were acinar cell damage and the accumulation of inflammatory cells in the tissue, as described earlier (7).

Tissue collagen analysis.

Collagen staining was performed on tissue sections by Masson's trichrome staining (Poly Scientific Research, Bay Shore, NY) method for the detection of collagen fibers according to the manufacturer's recommendations (31), and collagen tissue thickness was measured using a video-assistant integrated computer software program ImagePro software analyzer (Media Cybernetics, Warrendale, PA); collagen thickness is expressed in square microns.

Tissue eosinophil analysis.

Tissue sections were immunostained with antiserum against mouse eosinophil major basic protein (MBP, i.e., eosinophil-specific granule), a kind gift of Drs. James and Nancy Lee (Mayo Clinic, Scottsdale, AZ), as described (25, 28). In brief, endogenous peroxidase in the tissues was quenched with 0.3% hydrogen peroxide in methanol followed by nonspecific protein blocking with normal goat serum. Tissue sections were then incubated with rat anti-MBP (1:6,000) overnight at 4°C, followed by 1:250 dilution of biotinylated goat anti-rat IgG secondary antibody and avidin-peroxidase complex (Vector, Burlingame, CA) for 30 min each. These slides were further developed with nickel diaminobenzidine-cobalt chloride solution to form a black precipitate and counterstained with nuclear fast red. Negative controls include replacing the primary antibody with normal goat serum to check endogenous biotin and peroxidase activity. Quantification of the eosinophils was performed using a video-assistant integrated computer software program ImagePro software analyzer (Media Cybernetics). The eosinophil numbers were expressed as number of eosinophils per square millimeter. A total of four to five high-power fields in each pancreatic section were evaluated for eosinophil counts.

Mast cell analysis.

The 5-μm sections of paraffin-embedded or frozen pancreas tissue sections were deparaffinized and stained with chloroesterase staining (Sigma-Aldrich) and detected by light microscopy. The pink-stained mast cells were quantified by counting the stained cells in each pancreatic tissue section of mice with the assistance of digital morphometry using the ImagePro software analyzer (Media Cybernetics) and expressed as mast cells per square millimeter, as described earlier (29, 32). A total of four to five high-power fields in each pancreatic section were evaluated for mast cell counts.

Immunofluorescence analysis.

Paraffin-embedded pancreatic tissue sections were deparaffinized, blocked with normal goat serum to reduce nonspecific binding, and incubated overnight at 4°C with primary antibodies like anti-α-smooth muscle actin (α-SMA) for smooth muscle actin, anti-MBP for eosinophils, anti-tryptase for mast cells, followed by 2-h incubation with secondary antibodies at room temperature like anti-mouse IgG-PE and IgG-FITC. The α-SMA+, tryptase+, and MBP+ immunostained sections were counterstained and mounted using DAPI fluorescent material. The images were captured using an Olympus BX51 microscope with appropriate filters, and photomicrographs are presented as original magnification ×400. α-SMA+ cell numbers were expressed as α-SMA+ cells per square millimeter.

Real-time PCR analysis.

The pancreas of mice was lysed with TRIzol reagent (Life Technologies, Carlsbad, CA) for total RNA preparation according to the manufacturer’s instructions. Real-time PCR analysis for quantitation of mRNA was performed following the method described earlier (26). In brief, the RNA samples (500 ng) were subjected to reverse transcription analysis using iScript cDNA synthesis kit (Bio-Rad, Hercules, CA) according to the manufacturer’s protocol. IL-5, IL-18, eotaxin-1, eotaxin-2, TGF-β1, collagen 1, collagen 3, α-SMA, and fibronectin were quantified by real-time PCR using the CFX connect Real-Time System (Bio-Rad) and SsoAdvanced Universal SYBR^R Green Supermix (Bio-Rad). Results were then normalized with 18S rRNA amplified from the same cDNA mix and expressed as relative expression compared with the controls. cDNA was amplified using the primers listed in Table 2.

Table 2.

Primer list used for qPCR analysis

Name of Primers	Forward Primer	Reverse Primer
m18 s rRNA	GCAATTATTCCCCATGAACG	GGCCTCACTAAACCATCCAA
mIL-5	TCCCATGAGCACAGTGGTGAAAG	CACAGTACCCCCACGGACAGTTT
mIL-18	ACTTTGGCCGACTTCACTGT	GGGTTCACTGGCACTTTGAT
mEotaxin-1	GGCTCACCCAGGCTCCATCC	TTTTGGTCCAGGTGCTTTGTGG
mEotaxin-2	CTCCTTCTCCTGGTAGCCTGC	GTGATGAAGATGACCCCTGCCTT
mCollagen-1	TGTTCAGCTTTGTGGACCTC	GGTTTCCACGTCTCACCATT
mCollagen-3	CAGGATCTGTCCTTTGCGAT	CCCACTCCAGACTTGACATC
mFibronectin	CGAAGAGCCCTTACAGTTCC	CCGTGTAAGGGTCAAAGCAT
mα-SMA	CCTGGAGAAGAGCTACGAAC	CCCCTGACAGGACGTTGTTA
mTGF-β1	GCAACAATTCCTGGCGTTAC	GCTGAATCGAAAGCCCTGTA

ELISA.

TGF-β1, IL-18, eotaxin-1, and eotaxin-2 concentrations in the saline-treated and cerulein-treated mouse pancreatic tissue homogenates were quantified by ELISA analysis using human/mouse TGF-β1 (2nd generation) ELISA Ready-Set-Go (eBiosciences, San Diego, CA), mouse IL-18 Platinum ELISA kit (Affymetrix, eBiosciences), and eotaxin-1 and eotaxin-2 ELISA kits (Duo Set ELISA; R & D Systems, Minneapolis, MN) per manufacturers’ protocols.

Flow cytometric analysis for mouse blood eosinophils.

Blood eosinophil analysis in an experimental pancreatitis mouse model was performed by flow cytometry. The blood cells were stained with different florescence-tagged anti-CCR3 (BioLegend, San Diego, CA) and anti-Siglec-F (BioLegend) antibodies. Respective isotype-matched anti-IgGs were used as control antibodies. The analysis was performed using a FACSCalibur (BD Biosciences, Franklin Lakes, NJ) and analyzed by FlowJo software (48, 49).

Statistical analysis.

The nonparametric Mann–Whitney U-test was performed for comparison of data between two groups and Krustal-Wallis for comparison of more than two groups. Parametric data were compared using t-tests or ANOVA. Values are reported as means ± SD; P values < 0.05 were considered statistically significant.

RESULTS

Increased accumulation of eosinophils in human pancreatitis.

EP has been reported as a rare entity in pancreatitis (18, 44); therefore, not much effort has been made to understand the mechanism and role of eosinophils in disease pathogenesis. To establish the role of eosinophils in human pancreatitis, we first examined eosinophils in the pancreas of a normal individual and patients with pancreatitis. Herein, we report that eosinophil accumulation occurs in human pancreatitis compared with no eosinophils in the pancreatic tissue section of the normal individual (Fig. 1, A and B). The black arrows indicate intact eosinophils in the pancreatitis sections (Fig. 1B). A higher-magnification photomicrograph shows degranulation of eosinophils by green arrows to signify MBP+ eosinophilic extracellular granules in the pancreatitis tissue sections (Fig. 1D); no anti-MBP-positive extracellular eosinophil granules were observed in the normal pancreatic tissue section (Fig. 1C).

Fig. 1.

Analysis of eosinophils in human pancreatic biopsies. A representative photomicrograph shows an anti-major basic protein (MBP)-stained tissue section of normal individual pancreatic biopsies (A) and biopsies of a patient with pancreatitis showing accumulated eosinophils, marked by black arrows (B). Anti-MBP-positive extracellular granules were shown at high (×1,000) original magnification and marked with green arrows in the pancreatitis tissue section (D) compared with no granules in the tissue section of normal human pancreas (C). Black arrows indicate intact eosinophils, and green arrows show extracellular MBP+ eosinophilic granules. Quantitative data for eosinophils are presented as eosinophils/mm² (E). Photomicrographs are presented in both ×400 and ×1,000 of original magnification. Normal tissue, n = 2; chronic pancreatitis, n = 3.

Quantification of eosinophils was performed by morphometric analysis in tissue sections of normal individuals and patients with chronic pancreatitis, and data indicate ~200 eosinophils/mm² area for patients with chronic pancreatitis (Fig. 1E). Because EP is considered a rare disease and not much attention is given to the role of eosinophils in the induction and progression of pancreatitis pathogenesis, we further explored a critical role of eosinophils in the progression of pancreatitis pathogenesis using a cerulein-induced experimental pancreatitis mouse model.

Eosinophil accumulation in cerulein-induced acute and chronic pancreatitis in wild-type mice.

The cerulein-induced mouse model is an established experimental model of pancreatitis, but the existence of eosinophils is never examined (51–53). Therefore, we focused our studies to examine whether the role of eosinophils in cerulein-induced pancreatitis is also ignored and not explored like in humans. The mouse model of pancreatitis was induced following intraperitoneal injection of cerulein using the protocol reported earlier with slight modifications for chronic pancreatitis, as per the protocol of 4 wk of saline or cerulein intraperitoneal injections (Fig. 2A). The pancreas of these wild-type mice was analyzed for eosinophils, mast cells, eosinophil active cytokines, and chemokines. We examined mouse pancreatic tissue sections by light microscopy and detected a number of inflammatory cells, marked acinar cell injury, and atrophy in cerulein-induced chronic pancreatitis compared with saline-treated mice (Fig. 2, B and C). Because H and E staining was not conclusive to detect eosinophil infiltration and accumulation, anti-MBP immunostaining was performed to confirm the infiltration and accumulation of eosinophils in the pancreas of the cerulein-induced model of pancreatitis.

Fig. 2.

Eosinophil detection in the cerulein-induced mouse model of chronic pancreatitis. Cerulein-induced chronic pancreatitis is induced in wild-type and IL-5 gene-deficient mice following the protocol shown (A). A representative photomicrograph of hematoxylin and eosin-stained saline-treated (B) and cerulein-treated (C) wild-type mice pancreas tissue sections detected several inflammatory cells, as indicated by the black arrows. The eosinophils in the pancreatic tissue sections of saline-treated and cerulein-treated wild-type mice were further confirmed by performing anti-major basic protein (MBP) immunostaining. The anti-MBP immunostaining detected no eosinophils in saline (D), but a number of eosinophils in cerulein-treated wild-type mice are indicted by black arrows; degranulated extracellular MBP granules are indicated by green arrows (E). Real-time PCR and ELISA analysis indicate induced mRNA and protein levels of IL-18, eotaxin-1, and eotaxin-2 in cerulein-treated wild-type mice compared with saline-treated mice (F and G). However, we detect only induced levels of IL-5 mRNA in cerulein-treated wild-type mice compared with saline-treated mice (F). Data represent means ± SD, n = 8 mice /group. All photomicrographs shown are the original magnification ×400.

Interestingly, induced accumulation of intact anti-MBP-stained eosinophils and degranulated eosinophil extracellular granules was detected in a pancreatic tissue section of cerulein-treated wild-type mice compared with no eosinophils in saline-treated mice (Fig. 2, D and E). Furthermore, induced transcript and protein levels of eosinophil active cytokines and chemokines were found in cerulein-induced experimental pancreatitis in mice. The quantitative real-time PCR and ELISA analyses indicated significantly induced levels of IL-5, IL-18, and chemokines eotaxin-1 and eotaxin-2 transcript (Fig. 2F) and protein (Fig. 2G) levels in the pancreas of cerulein-induced pancreatitis in mice compared with saline-treated mice. Furthermore, we also analyzed the peripheral blood eosinophilia using anti-CCR3 and anti-Siglec-F antibodies by flow cytometric analysis. Our analysis indicated no significant change in peripheral blood eosinophilia in cerulein-treated compared with saline-treated wild-type mice (data not shown). These data indicate that EP is an independent to peripheral blood eosinophilia, at least in experimental pancreatitis, which needs to be carefully examined in patients with pancreatitis to establish the disease entity.

IL-5 gene-deficient and eosinophil-deficient GATA1 mice have improved pancreatic pathogenesis.

Hematoxylin-eosin staining was performed to analyze histopathological changes, acinar cell injury, infiltratory inflammatory cells, and the pancreas/mouse body weight (mg/gm) ratio of IL-5 gene-deficient mice, eosinophil-deficient GATA1 mice, and wild-type mice following the induction of chronic pancreatitis. We observed severe atrophy in cerulein-treated wild-type mice compared with IL-5 gene-deficient mice and eosinophil-deficient GATA1 mice; however, saline-treated mice showed normal healthy acinar cells (Fig. 3, A–H). The semiquantitative analysis on the scale of 0–4 indicated that cerulein-treated IL-5 gene-deficient mice and cerulein-treated GATA1 mice have significantly reduced acinar cell damage compared with cerulein-treated wild-type mice (Fig. 3G). Similarly, we saw significantly decreased inflammatory cell accumulation in pancreatic tissue sections of cerulein-treated IL-5 gene-deficient mice and cerulein-treated GATA1 mice compared with the cerulein-treated wild-type mice (Fig. 3H). Furthermore, histologically some interlobular edema was observed in the pancreatic tissue sections of some cerulein-treated wild-type mice, but per the pancreas/mouse body wt (mg/g) ratio, edema was not confirmed. A significantly reduced pancreas/mouse body wt (mg/g) ratio in cerulein-induced chronic pancreatitis was found compared with saline-treated mice. The saline-treated pancreas/body wt ratio of wild-type mice was 11.98 ± 1.4 mg/g compared with 4.13 ± 0.48 mg/g in cerulein-treated wild-type mice (Fig. 3I). In contrast, a comparable pancreas/mouse body wt (mg/g) ratio was observed in between saline-treated IL-5 gene-deficient mice, saline-treated eosinophil-deficient GATA1 mice, cerulein-treated IL-5 gene-deficient mice, and cerulein- treated eosinophil-deficient GATA1 mice (Fig. 3I). The intralobular edema was more prominent in acute pancreatitis compared with presented chronic pancreatitis (data not shown).

Fig. 3.

Pancreatic pathogenesis in IL-5 gene-deficient and eosinophil-deficient GATA1 mice. A representative photomicrograph shows hematoxylin and eosin-stained saline-treated and cerulein-treated wild-type mice (A and B); saline-treated and cerulein-treated IL-5 gene-deficient mice are shown (C and D); saline-treated and cerulein-treated eosinophil-deficient GATA1 mice are also shown (E and F). Acinar cell damage score (G), inflammatory infiltrate score (H), and pancreas weight/mouse body wt ratio (I) of saline- and cerulein-treated wild-type mice, IL-5 gene-deficient mice, and eosinophil-deficient GATA1 mice have been shown. Data are presented as means ± SD, n = 8 mice/group. All photomicrographs shown are the original magnification ×400.

Endogenous IL-5-deficient mice show no pancreatic eosinophilia and reduced levels of eosinophil active cytokines and chemokines.

Because IL-5 is a well-known growth, differentiation, and survival factor for eosinophils (5, 41), to test the role of eosinophils in pancreatitis pathogenesis, we further induced chronic pancreatitis in wild-type and IL-5 gene-deficient mice following the cerulein injection protocol as described (Fig. 2A). Interestingly, no eosinophils were detected in pancreas of saline-treated as well as cerulein-treated IL-5 gene-deficient mice. The morphometric analysis for eosinophil quantitation was performed in saline- and cerulein-treated wild-type and IL-5 gene-deficient mice (Fig. 4A). In addition, we also observed improvement in pancreatic pathogenesis in cerulein-treated IL-5 gene-deficient mice along with reduced levels of eosinophil active cytokines and chemokine mRNA and protein levels. The quantitative real-time PCR and ELISA analyses have indicated that cerulein-treated wild-type mice show induction in IL-5 mRNA compared with saline-treated wild-type mice, whereas no IL-5 mRNA transcripts were detected in IL-5 gene-deficient mice (Fig. 4B). Furthermore, IL-5 gene-deficient mice detected comparable mRNA levels of IL-18 and eotaxin-2 with the cerulein-treated wild-type mice (Fig. 4, C and E). Importantly, we observed a significant decrease in the protein levels of IL-18 and eotaxin-2 in cerulein-treated IL-5 gene-deficient mice compared with wild-type mice (Fig. 4, F and H). The levels of mRNA and protein of eotaxin-1 were comparable in wild-type and IL-5 gene-deficient mice (Fig. 4, D and G). The data indicated that in the absence of endogenous IL-5, eotaxin-2 has a major role compared with eotaxin-1 in promoting cerulein-induced pancreatic eosinophilia in mice.

Fig. 4.

Eosinophil active cytokine and chemokine levels in cerulein-treated wild-type and IL-5 gene-deficient mice. Quantitative eosinophil numbers of saline- and cerulein-treated wild-type as well as IL-5 gene-deficient mice were presented as eosinophils/mm² (A). The real-time PCR and ELISA analysis was performed to detect mRNA and protein levels of cytokines (IL-5, IL-18) and chemokines (eotaxin-1, eotaxin-2) in saline- and cerulein-treated mouse model of chronic pancreatitis of wild-type mice and IL-5 gene-deficient mice. The significantly increased mRNA level of IL-5 (B), IL-18 (C), and eotaxin-2 (E) has been observed in the pancreas of cerulein-treated wild-type mice compared with saline-treated wild-type mice. However, reduced but not significant changes in mRNA levels of IL-18 (C) and eotaxin-2 (E) were found in cerulein-treated IL-5 gene-deficient mice compared with cerulein-treated wild-type mice. Induced protein levels of IL-18 (F) and eotaxin-2 (H) were quantitated in the pancreas of cerulein-treated wild-type mice compared with saline-treated wild-type mice, and the protein levels of IL-18 (F) and eotaxin-2 (H) were significantly reduced in cerulein-treated IL-5 gene-deficient mice compared with cerulein-treated wild-type mice. The mRNA and protein level of eotaxin-1 was comparable in cerulein-treated wild-type and cerulein-treated IL-5 gene-deficient mice (D and G). Data are presented as means ± SD, n = 8 mice/group.

IL-5 gene-deficient mice and eosinophil-deficient GATA1 mice show reduced mast cell accumulation in experimental chronic pancreatitis.

Induced mast cells in human pancreatitis and in experimental models of pancreatitis have been earlier reported (9, 12, 15, 46, 55); therefore, we were interested to know whether mast cells are also decreased in IL-5 gene-deficient mice. Earlier, it has been shown that eosinophil granular MBP has a role in mast cell accumulation, activation, and degranulation (3, 30). We detected induced mast cells and their degranulation in the cerulein-induced chronic pancreatitis mouse model compared with saline-treated mice. The intact mast cells are marked by black arrows, and degranulated mast cells are marked by yellow arrows (Fig. 5, A and B). Interestingly, mast cell numbers were also found decreased in cerulein-induced IL-5 gene-deficient mice compared with the wild-type mice (Fig. 5, C and D). Furthermore, a high-magnification (×1,000) photomicrograph clearly showed that these are the mast cells, not red blood cells (Fig. 5, E and F). The morphometric quantitation of mast cells in wild-type and IL-5 gene-deficient mice following saline and cerulein-induced chronic pancreatitis was performed and presented (Fig. 5G). The pancreas of eosinophil-deficient GATA1 mice was further analyzed for mast cell accumulation and degranulation in chronic pancreatitis. The morphometric analysis indicates reduced numbers of mast cells in cerulein-induced eosinophil-deficient GATA1 mice compared with the cerulein-treated wild-type mice (Fig. 5H). Additionally, the double immunofluorescence staining using anti-MBP and anti-tryptase antibody detected degranulation of mast cells nearby eosinophils in the pancreatic tissue section of saline-treated (Fig. 5, I–K) and cerulein-treated wild-type mice (Fig. 5, L–N). The tryptase+ intact mast cells are indicated by white arrows, MBP+ intact eosinophils by green arrows, and tryptase+ degranulation product of mast cells by red arrows in the photomicrograph of pancreatic tissue sections (Fig. 5N). Several overlapped double-stained cells were detected in the photomicrograph that may be the overlapped degranulation product on each of the cell types (Fig. 5N).

Fig. 5.

Mast cell analysis in cerulein-induced chronic pancreatitis. Chloroacetate esterase staining on paraffin-embedded pancreatic tissue sections for mast cells was performed in cerulein-induced chronic pancreatitis model of wild-type mice and IL-5 gene-deficient mice. The cerulein-induced mast cells were detected in the pancreas of both wild-type and IL-5 gene-deficient mice compared with very few in saline-treated mice. A pink granule-stained mast cell (black arrow) and its degranulation (yellow arrow) have been shown in a representative photomicrograph of wild-type saline (A), cerulein-treated mice (B), IL-5 gene-deficient saline-treated mice (C), and cerulein-treated mice (D). High-magnification (×1,000) photomicrograph shows clear mast cells not red blood cells (E and F). The quantitative mast cell numbers have been presented in saline- and cerulein-treated wild-type mice and IL-5 gene-deficient mice as mast cells/mm² (G). Quantitation of mast cells in saline- and cerulein-treated eosinophil-deficient GATA1 mice and wild-type mice (H) is shown. Double immunofluorescence staining for major basic protein (MBP) and tryptase in the pancreatic tissues section of saline (I–K) and cerulein-treated wild-type mice (L–N) is shown. White arrow marks tryptase+ intact mast cells, green arrows mark MBP+ intact eosinophils, and red arrow marks tryptase+ degranulation product of mast cells (N). All photomicrographs shown are of original magnification ×400, ×1,000. Data are presented as means ± SD, n = 8 mice/group.

Cerulein-induced collagen is reduced in IL-5 gene-deficient mice and eosinophil-deficient GATA1 mice.

Chronic pancreatitis is documented with induced fibrosis (43); therefore, we examined whether eosinophils have a role in the induced accumulation of tissue collagen in chronic pancreatitis. We show that Masson’s trichrome-stained tissue sections of cerulein-treated mice have increased collagen in the pancreatic tissue sections compared with the saline-treated wild-type mice (Fig. 6, A and B); however, a significantly reduced collagen deposition was observed in cerulein-treated IL-5 gene-deficient mice (Fig. 6D) compared with the cerulein-treated wild-type mice (Fig. 6B). The semiquantitative collagen accumulation in the tissue was performed using morphometric analysis to determine the thickness of perivascular collagen in saline-treated and cerulein-treated wild-type mice and IL-5 gene-deficient mice (Fig. 6E). Additionally, to establish the role of eosinophils in pancreatitis pathogenesis, we examined collagen accumulation in the eosinophil-deficient GATA1 mice following the induction of cerulein-induced chronic pancreatitis. The morphometric analysis shows significantly reduced accumulation of tissue collagen (Fig. 6F) in the pancreas of cerulein-treated eosinophil-deficient GATA1 mice compared with the cerulein-treated wild-type mice.

Fig. 6.

Analysis of collagen deposition in cerulein-induced chronic pancreatitis Masson’s trichrome staining on paraffin-embedded pancreatic tissue sections for collagen accumulation was performed in a cerulein-induced chronic pancreatitis model of wild-type mice and IL-5 gene-deficient mice. A representative photomicrograph shows Masson’s trichrome-stained pancreatic tissue section of saline-treated (A) and cerulein-treated (B) wild-type mice and saline-treated (C) and cerulein-treated (D) IL-5 gene-deficient mice. The perivascular collagen thickness in the saline- and cerulein-treated pancreatic tissue sections of wild-type and IL-5 gene-deficient mice has been measured and shown as collagen thickness as µm² (E). Quantitation of collagen thickness in eosinophil-deficient GATA1 mice and wild-type mice is presented (F). All photomicrographs of original magnification ×400 have been shown here. Data are presented as means ± SD, n = 8 mice/group.

Eosinophil deficiency in IL-5 gene-deficient mice downregulates fibrosis-associated genes in cerulein-induced pancreatitis.

Because eosinophils are the source of TGF-β1, which has an important role in the initiation of tissue fibrosis (27, 34), we further examined mRNA expression of profibrosis genes (TGF-β1, collagen-1, collagen-3, fibronectin, and α-SMA) in the pancreas of saline-treated and cerulein-treated wild-type mice and IL-5 gene-deficient mice. The real-time PCR analysis has shown a significantly increased mRNA expression of TGF-β1, collagen-1, collagen-3, fibronectin, and α-SMA in cerulein-treated mice compared with the saline-treated wild-type mice (Fig. 7, A–E); however, the levels of TGF-β1, collagen-1, collagen-3, fibronectin, and α-SMA have been found decreased in cerulein-treated IL-5 gene-deficient mice compared with the cerulein-treated wild-type mice (Fig. 7, A–E). The ELISA analysis further confirmed that the protein level of TGF-β1 in cerulein-treated mice was induced compared with the saline-treated wild-type mice and significantly decreased in cerulein-treated IL-5 gene-deficient mice (Fig. 7F).

Fig. 7.

Analysis of fibrosis-associated proteins in the pancreas following saline- and cerulein-treated mice. mRNA and protein levels were analyzed in cerulein-induced chronic pancreatitis. Real-time PCR examined induced mRNA levels of TGF-β1 (A), collagen-1 (B), collagen-3 (C), fibronectin (D), and α-smooth muscle actin (α-SMA) (E) in saline- and cerulein-treated wild-type mice and IL-5 gene-deficient mice. ELISA measured protein levels of TGF-β1 in saline- and cerulein-treated wild-type mice and IL-5 gene-deficient mice (F). The mRNA and protein levels are expressed as means ± SD (n = 8 mice/group).

Furthermore, we show that the α-SMA-positive cells were increased in cerulein-treated wild-type mice compared with the cerulein-treated IL-5 gene-deficient mice by performing immunofluorescence analysis (Fig. 8, A–H). The α-SMA-positive cells were quantified by morphometric analysis in saline- and cerulein-treated pancreatic tissue section of wild-type and IL-5 gene-deficient mice, and the average number of α-SMA-positive cells from multiple mouse sections has been presented as means ± SD (Fig. 8I).

Fig. 8.

Immunofluorescence analysis of α-smooth muscle actin (α-SMA) proteins in the pancreas following saline- or cerulein-treated mice. Immunofluorescence analysis revealed anti-α-SMA-positive cells in saline- and cerulein-treated wild-type mice (A, B, E, and F) and saline- and cerulein-treated IL-5 gene-deficient mice (C, D, G, and H). Morphometric analysis was performed for quantitation of α-SMA-positive cells in the pancreatic sections of saline- and cerulein-treated wild-type mice and IL-5 gene-deficient mice (I). All photomicrographs are presented as ×400 of original magnification, n = 8 mice/group.

DISCUSSION

The etiology of EP is poorly understood, and the role of eosinophils in the initiation and progression of pancreatitis pathogenesis has not been explored. The clinical symptoms and characteristics for patients with EP are reported as levels of peripheral eosinophils >1.5 × 10⁹ for >6 mo, with no history of rhinitis, bronchial asthma or other allergic diseases, no eosinophilic infiltration in heart, skin, and gastrointestinal tract, including exclusion of leukemia, parasitic infection, and a diagnosis of hypereosinophilic syndrome, per the international standards (44, 47). However, several patients with EP have peripheral blood eosinophilia, but the experimental pancreatitis indicates that EP is independent to peripheral eosinophilia because no blood eosinophilia is observed in the experimental model of cerulein-induced pancreatitis. Similarly, several patients with eosinophilic gastrointestinal disease showed peripheral eosinophilia, but still most eosinophilic gastrointestinal disorders are recognized as independent disease entities (33, 36, 42). Similarly, EP may also be dependent on or independent from peripheral blood eosinophilia. EP is recognized as a rare disease (10); however, our presented data indicate that EP may not be a rare but not well-diagnosed and ignored disease entity.

In the present study, we provided evidence that eosinophils indeed accumulated in the pancreas of human and experimental pancreatitis. Interestingly, the cerulein-induced pancreatitis model is widely used to study disease pathogenesis (19), but no efforts have been made to examine the eosinophil accumulation in the pancreas. Our analysis shows increased numbers of eosinophil accumulation, including induced levels of eosinophil-active chemokines such as eotaxin-2 and eotaxin-1 in the cerulein-induced chronic pancreatitis models. Eotaxins are well-known eosinophil-specific chemokines and are responsible for the recruitment of eosinophils (17, 35, 38, 61). GATA1 transcription factor is required for eosinophil lineage development, whereas IL-5 is a differentiation growth and survival factor for eosinophils (39, 41, 54, 58). We show that cerulein-treated IL-5 gene-deficient mice and eosinophil-deficient GATA1 mice showed less acinar cell damage, decreased infiltration of inflammatory cells, reduced collagen, and less mast cell accumulation. Of note, cerulein-treated IL-5 gene-deficient mice did not show accumulation of eosinophils in the pancreas. These mouse data support the fact that induced eosinophils may be a significant factor in promoting pancreatitis pathogenesis. Additionally, we found that both mRNA and protein levels of IL-18 were highly increased, but we fail to detect induced IL-5 protein levels in our model. This may be due to the low sensitivity of our ELISA analysis. Notably, a recent report showed induced IL-5 in the same pancreatitis model using ultrasensitive Luminex assay (53). This indicates that IL-5 indeed is induced but in low levels, which is sufficient to provide eosinophil differentiation, growth, and survival. Our findings suggest that the absence of endogenous IL-5 reduces eosinophils, and induced IL-18 indicates the contributory role of IL-18 in promoting eosinophil accumulation and degranulation during pancreatitis pathogenesis. These observations are consistent with earlier reported studies that IL-18 overexpression promotes eosinophil accumulation in the lung and gastrointestinal tract (8, 16, 31). Furthermore, IL-18 is also implicated as a mast cell differentiation and maturation factor (56, 57), and we have shown that mast cells were also induced in the cerulein-induced mouse model of eosinophil-associated pancreatitis. Mast cells are well-known effector cells, and eosinophils-derived MBP is capable of activating and degranulating the mast cells (3, 30). Tryptase, a serine protease stored in mast cell granules, induces the synthesis of type I collagen in human fibroblasts (11) and stimulates fibroblast proliferation (2, 6, 43) and chemotaxis (11). Therefore, the induced eosinophils, mast cells, and collagen in wild-type mice compared with IL-5 gene-deficient mice and eosinophil-deficient GATA1 mice indicate an important role of eosinophils and IL-18 in pancreatitis pathogenesis. These experimental findings suggest that eosinophil accumulation has a role in chronic pancreatitis, including remodeling and fibrosis.

Taken together, we provide the first evidence that eosinophil accumulation in the pancreas is important in promoting experimental and human pancreatitis. Secondly, we show that eosinophil-active cytokine IL-5, IL-18, chemokine eotaxin-1, eotaxin-2, including profibrotic protein TGF-β1, and α-SMA are induced in a cerulein-induced experimental model of pancreatitis. Thirdly, we show induced collagen accumulation in wild-type mice compared with the eosinophil-deficient GATA1 and endogenous IL-5-null mice. Notably, eosinophils are a source of TGF-β and other profibrotic cytokines that promote tissue remodeling and fibrosis. In conclusion, the present study highlights the role of eosinophils in promoting pancreatitis pathogenesis and indicates that EP may be an independent disease entity that needs further attention for early detection and for providing treatment strategy.

GRANTS

This work was partially supported by National Institutes of Health Grant R01-AI-080581 (to A. Mishra).

DISCLOSURES

There are no conflicts of interest to disclose for all authors except Dr. Mishra, who discloses past and present consultancy relationships with Axcan Pharma, Aptalis, Elite Biosciences, Calypso Biotech SA, and Enumeral Biomedical.

AUTHOR CONTRIBUTIONS

A.M. conceived and designed research; M.M. and A.K.V. performed experiments; M.M. and S.U.V. analyzed data; A.M. interpreted results of experiments; M.M. prepared figures; M.M. drafted manuscript; A.M. edited and revised manuscript.