Experiments were approved by the Animal Experimental Ethical Inspection of Laboratory Animal Centre, Wenzhou Medical University, Wenzhou, China (permit no. wydw2015-0128). A total of 45 male C57/BL6 mice were supplied by Laboratory Animal Centre of Wenzhou Medical University, housed in specific pathogen-free condition with constant temperature of 24°C and 12-h light/dark cycle. The mice had free access to food and water until 24 h before experiments. When mice were 6 weeks old and about 25 g, acute pancreatitis was induced by 2 sets of 6 hourly intra-peritoneal injections of cearulein (50 µg/kg; Sigma-Aldrich; Merck Millipore, Darmstadt, Germany) on alternating days. Luteolin (Chengdu Must Bio-Technology, Chengdu, China) was dissolved by DMSO at a concentration of 40 mg/ml, and then diluted by normal saline to 0.05 mg/ml. Mice were intra-peritoneally injected with luteolin daily at a dosage of 2 mg/kg. The first injection of luteolin was performed immediately subsequent to the completion of cearulein injections on the first day. The last injection of luteolin was performed on the day prior to sacrifice. The next day after finishing induction of pancreatitis was defined as day 1. Mice were sacrificed on days 1, 3 and 5 in the 9 different groups (day 1-CTRL, day 1-AP and day 1-AP+Luteolin, and so on for days 3 and 5). Each group comprised 5 mice.

Pancreas tissues were fixed overnight in 4% paraformaldehyde (Sigma-Aldrich; Merck Millipore), embedded in paraffin, cut into 4-µm-thick sections, and placed on slides. Sections were subjected to H&E and immunohistochemical staining as previously described (19,29). For H&E staining, sections were stained with hematoxylin for 4 min and with eosin for 1.5 min. For immunohistochemistry, 3–3′-diaminobenzidine tetrahydrochloride was used as a chromogen. Primary antibodies were diluted according to the instructions and were incubated overnight at 4°C (Table I). Bright-field images were acquired using a DM4000B microscope (Leica Microsystems GmbH, Wetzlar, Germany).

Pancreas tissue was preserved in liquid nitrogen. Total RNA was extracted by tissue dissociation in TRIzol reagent (Invitrogen; Thermo Fisher Scientific, Inc., Waltham, MA, USA) according to the manufacturer's protocol. cDNA was synthesized using 1 µg of total RNA in a 20 µl final volume by reverse transcription with RevertAid RT Reverse Transcription kit (Thermo Fisher Scientific, Inc., Waltham, MA, USA). qPCR was performed with 0.3 µl of cDNA, 1.5 µl forward and reverse primers (2 µM), 1.7 µl double distilled water, and 5 µl SYBR Green Mix (Applied Biosystems; Thermo Fisher Scientific, Inc.). Expression levels were normalized using a custom primer set for GAPDH. cDNA was amplified using ABI Prism 7500 Sequence Detection System (Applied Biosystems; Thermo Fisher Scientific, Inc.). Primer sequences are listed in Table II. The thermocycling conditions were: Stage 1: 95°C, 5 min; Stage 2: 95°C, 30 sec, 60°C, 1 min, 40 cycles; melt curve: 95°C, 5 min; and, cooling at 1.6°C/sec to 60°C.

Comparisons among multiple groups were made with a one-way analysis of variance followed by least significant difference or Dunnett's t-test. P<0.05 was considered to indicate a statistically significant difference. Results are represented as the mean ± standard error. All data presented represent at least 3 mice and represent at least three 4-µM sections unless otherwise noted.

H&E staining was performed and an immunohistochemistry assay was conducted to determine whether luteolin has effects on ADM formation induced by acute pancreatitis. By H&E staining, it was identified that ADM was most evident on day 3, compared with days and day 5. A prominent change of ADM induced by acute pancreatitis was the formation of tubular complexes, which was characterized by circular arrangement of acinar cells that resembled ducts, and the tubular complexes of each field were counted (magnification, ×200) (30,31). The H&E staining slides exhibited no tubular complex in the control group (Fig. 1A), and 3 days subsequent to acute pancreatitis, this number increased to 32.27±5.58 (Fig. 1B). However, with the treatment with luteolin, 22±3.79 tubular complexes were observed per field (P<0.05) (Fig. 1C and D). This indicates that acute pancreatitis-induced the formation of tubular complexes, which was then significantly inhibited by luteolin.

To confirm luteolin's effects on ADM further, an immunohistochemistry assay was performed in order to detect ectopic expression of CK19 in acinar cells. CK19-positive acini were then counted to determine the extent of ADM. As the immunohistochemistry assay indicated, unlike tubular complexes which were clearer on day 3, CK19-positive acini rapidly appeared on day 1 and could last up to day 5. The number of CK19-positive acini were counted on days 1 and 3 to assess whether luteolin has effects on ADM. It was identified that on day 1, the number of CK19-positive acini increased from 1.13±0.45 to 93.67±13.54 per field with the injection of cearulein (P<0.05) (Fig. 1E and F), and this number decreased to 63.47±11.48 with the treatment of luteolin (P<0.05) (Fig. 1G and H). A total of 3 days subsequent to the induction of acute pancreatitis, mice which were not treated with luteolin had 34.73±6.69 CK19-positive acini per field (Fig. 1J). In mice treated with luteolin, this number decreased to 22.07±5.04 however this was not significant (P=0.072) (Fig. 1K and L). The results of H&E staining and immunohistochemistry assay both indicated that luteolin inhibits ADM induced by acute pancreatitis.

Numerous kinds of transcriptional factors, such as SOX9 and HNF6, have been confirmed to be essential for ADM formation (9). In order to investigate how luteolin inhibits pancreatitis-induced ADM formation, immunohistochemistry was conducted to determine the effects of luteolin on SOX9 expression (Fig. 2). SOX9 was expressed in ductal cells and centroacinar cells alone in the normal pancreas tissues, while it was induced in acinar cells following acute pancreatitis. SOX9-positive cells increased significantly one day subsequent to pancreatitis (Fig. 2A and B). On day 1, for the control group, 93.67±11.83 of pancreatic cells per field were detected to exhibit expression of SOX9. However, for mice with acute pancreatitis, this number increased to 213.9 ±17.98 per field (P<0.05). Furthermore, when treated with luteolin, this number decreased to 151.4±16.99 (P<0.05) (Fig. 2C and J). Acute pancreatitis-induced ectopic SOX9 expression lasted up to day 3. On day 3, for mice injected with cearulein and treated with luteolin, SOX9-positive cells decreased significantly from 232.45±16.02 to 184.07±21.22 per field (P<0.05) (Fig. 2D-F and J). Luteolin also caused the decrease of SOX9-positive cells on day 5, however not significantly (P=0.112; Fig. 2G-I and J), and this is suggested to be due to the fact that ectopic SOX9 expression was almost decreased to normal levels. With the results of immunohistochemistry assay, it was concluded that luteolin inhibits acute pancreatitis-induced SOX9 expression in acinar cells, and as SOX9 is essential for ADM formation, luteolin's effects of inhibiting ADM may partly depend on the inhibition of SOX9 expression.

A number of signaling pathways have been confirmed to be essential for ADM and pancreatic carcinogenesis, including the interleukin (IL)-STAT3 pathway, EGFR-ras pathway and canonical Wnt pathway (11–13,15–18). An immunohistochemistry assay was conducted to determine the effects of luteolin on these pathways. Consistent with previous studies (15,16), p-STAT3 positive acinar cells were not observed in the exocrine part of the pancreas in the control group, however is widely observed in mice with pancreatitis on day 1 (Fig. 3A and B). However, with the treatment of luteolin, p-STAT3-positive cells decreased significantly during pancreatitis (Fig. 3C). It was also identified that pancreatitis significantly induced phosphorylation of EGFR in acinar cells (Fig. 3D and E) and with the treatment of luteolin, phosphorylation of EGFR was decreased significantly on day 1 (Fig. 3F). Therefore, not only the STAT3 pathway, however additionally the EGFR pathway were inhibited by luteolin during pancreatitis. The expression of β-catenin was determined following the treatment of luteolin. However, it seems that luteolin has no significant effect on the expression and membrane-translocation of β-catenin (data not shown). As STAT3 and EGFR signaling pathways serve a role in ADM, and the inhibition of them may contribute to luteolin's effects of inhibiting ADM formation.

It has been reported that pancreatitis-induced ADM and pancreatic carcinogenesis were accompanied with persistent proliferation of pancreatic cells (17). Cell cycle activation marker Ki67 was used to determine proliferative level of acinar cells. Few acinar cells were identified to be Ki67-positive on the control group (Fig. 4A). And frequency of Ki67-positive cells increased to 46.62±7.45 per field one day after the induction of pancreatitis (Fig. 4B). However, with the treatment of luteolin, Ki67-positive cells decreased to 19.4±6.45 per field (P<0.05; Fig. 4C). On day 3, mice on the acute pancreatitis group had 68±9.54 Ki67-positive cells per field (Fig. 4D and E). Whilst with the treatment of luteolin, this number decreased to 47.67±6.32 (P<0.05; Fig. 4F). On day 5, the number of Ki67-positive cells per field decreased from 30.8±3 to 27.43±4.52 with the treatment of luteolin (Fig. 4G-I), however it was not significant (Fig. 4J). Therefore, it was concluded that pancreatitis-induced proliferation of exocrine cells was inhibited by luteolin.

Pancreatic carcinogenesis was accompanied by EMT of acinar cells, which significantly promoted dissemination of acinar cells under malignant transition (22). Thus, it was investigated whether acute pancreatitis promotes EMT of acinar cells and if luteolin has effects on it. CDH2, which was an adhesion molecule in the membrane and a marker of mesenchyma, was chosen to indicate EMT change of acinar cells during pancreatitis. By immunohistochemistry, it was observed that CDH2 was highly expressed in islets and marginally expressed in acinar cells of normal pancreas (Fig. 5A). However, during pancreatitis, expression of CDH2 increased significantly in acinar cells on day 1 (Fig. 5B). Subsequently, the expression of CDH2 on mice treated with luteolin was established. The results indicated that pancreatitis-induced expression of CDH2 was effectively inhibited by luteolin (Fig. 5C).

RT-qPCR was conducted to further confirm the results. Certain EMT markers, including CDH1, CDH2, EpCAM, Snail, Slug, Twist, Vimentin, Zeb1 and ZO-1, were selected to indicate EMT change of acinar cells. As compared with the control group, the results of PCR indicated that CDH1, CDH2, Vimentin and Zeb1 changed significantly 1 day after acute pancreatitis (Fig. 5D). Changes of the four markers all indicated EMT changes to acinar cells. No marker was identified to change significantly against EMT (Fig. 5D) and compared with the acute pancreatitis group, CDH1, CDH2, Snail and Zeb1 changed significantly with the treatment of luteolin (Fig. 5D). Changes of the four markers all indicated that the EMT changes of acinar cells were reversed. No marker suggested that luteolin would promote EMT (Fig. 5D). Therefore, with the results of the immunohistochemistry assay and RT-qPCR, it was confirmed that acute pancreatitis promoted EMT of acinar cells, and this was reversed by luteolin.