IL13 and periostin in active fibrogenic areas of the extrahepatic bile ducts in biliary atresia patients

The leading pathology of biliary atresia (BA) is inflammatory and fibrous obstruction of extrahepatic bile duct, but the pathogenesis remains unclear. IL13 is a cytokine associated with allergies and inflammatory fibrosis, and periostin induces fibrogenesis by stimulation with IL13. We analyzed the involvement of IL13 and periostin in inflammatory fibrosis in the extrahepatic bile duct of BA patients. Surgically resected tissues from the hepatic hilar area of BA patients were immunostained with CD45, α-SMA, IL13 and periostin and statistically analyzed. Fibroblasts from the resected tissue were cultured with recombinant IL13, and periostin production was analyzed by quantitative polymerase chain reaction and Western blotting. IL13 was stained in 93% of large and micro bile ducts, and 92.1% matched with the CD45 location (p = 0.006) around the large bile ducts. Periostin staining correlated with the localization of IL13 and αSMA (p < 0.001) around the large bile ducts. Periostin mRNA and protein were upregulated by IL13 stimulation in cultured fibroblasts. IL13 was associated with induced periostin expression by fibroblasts, playing a vital role in the pathogenesis of fibrogenesis around the extrahepatic bile duct in BA.


Introduction
Biliary atresia (BA) is a progressive cholestatic hepatopathy in infants with the pathology of inflammatory and fibrous obstruction of the extrahepatic bile ducts by an unknown etiology. If untreated, the liver develops fatal cirrhosis, and even after primary surgery, about half of patients need liver transplantation because of the progression of cirrhosis.
The standard primary surgery for BA is Kasai portoenterostomy (KPE), wherein the remnants of obstructed extrahepatic bile ducts are resected and replaced by intestine. According to histological analyses, resected tissues from the hilar area of the liver show active inflammatory fibrosis around the remnant of bile ducts, including infiltration of inflammatory cells, such as neutrophils, eosinophils and macrophages. A histological examination of the remnant is crucial for clarifying the pathogenesis, and many histological studies for BA have reported correlations between the histological features and clinical prognoses [1][2][3][4][5][6][7][8]. However, few reports have focused on the immune responses that precede fibrosis, especially around the extrahepatic bile ducts in BA.
IL13 is a type 2 immune response cytokine that causes fibrosis at the site of inflammation in various allergic diseases and inflammatory gastrointestinal diseases [9]. The type 2 immune response plays roles in tissue repairs and homeostasis. However, its chronic activity and overreaction can cause tissue damage or fibrosis [10]. IL13 is secreted from Helper T2 cells or other innate immunity cells in response to epithelial injury and promotes mucus secretion by inducing the proliferation of glandular epithelium. IL13 also participates in the activation of fibroblasts and progression of tissue fibrosis [11]. Although some studies have demonstrated the involvement of abnormal immunological activities in BA, most were type 1 immune responses [12,13]. Regarding IL13, few reports have suggested its contribution to BA progression, except for its presence in fibrotic liver or elevated serum levels [14,15]. Only the involvement of IL13 in the pathogenesis of extrahepatic bile duct has been reported, and even then, only in rhesus rotavirus type A (RRV) BA model mice [16,17], not human BA.
As another factor related to fibrosis, periostin is a cell adhesion molecule secreted from fibroblasts by IL13 stimulation. It causes fibrogenesis by binding to extracellular matrix, such as type 1 collagen. As a matricellular protein, periostin promotes cell migration and activation of inflammatory cells by binding integrin molecules, eventually causing chronicity and exacerbation of inflammation [18]. Periostin is highly expressed in allergic diseases or inflammatory fibrosis, such as asthma [19] or idiopathic pulmonary fibrosis [20]. Since the serum levels well correspond to the status of diseases, periostin has been reported as a possible serum biomarker that may aid in deciding on treatment for asthma or other allergic diseases [21]. Regarding BA, periostin has been reported as a possible serum marker of liver fibrosis that reflects disease progression [22]. However, its presence in tissues of BA patients has never been examined, especially around the extrahepatic bile ducts.
In this study, we focused on the immunoreaction at the hepatic hilar area, the primary damaged region in BA patients. We analyzed the contribution of IL13 and its downstream factor periostin to the inflammation and fibrogenesis around the bile ducts by histological and cellular approaches.

Ethics
Tissue samples and clinical data were obtained from 32 BA patients who received surgery in our institution from January 2008 to March 2020. All patients were informed of the study and consented to their inclusion. The analyses of this study were approved by the ethics committee of our institution.

Fibroblast culture from a BA patient
Small fragments of tissues were obtained from the hepatic hilar area of a BA patient at surgery. The tissues were dissected to further smaller pieces and incubated in DMEM low-glucose medium (Nacalai Tesque) containing 5% fetal bovine serum (FBS) (BioWest, Nuaillé, France) and Penicillin/Streptomycin (Nacalai Tesque). After cells had adhered to the culture dish and spread, they were passaged and used for experiments.

Immunofluorescence (IF)
IF was performed to confirm that the cultured cells were active fibroblasts by staining with α-SMA. Cells were fixed for 15 min with 4% paraformaldehyde and permeabilized for 1 h with 0.3% Triton-X and 10% FBS in PBS. Cells were incubated overnight at 4 °C in primary antibody solution containing 1:200 of anti-α-SMA antibody (1395-1-AP; Proteintech). Following PBS washes, cells were incubated for 30 min at room temperature in PBS containing 1:400 of secondary antibody labeled with Alexa Fluor 546 (Goat anti-Rabbit IgG(H + L); Thermo Fisher Scientific, Hudson, NH, USA). Finally, cells were encapsulated with DAPI for nuclear staining. Cells were observed by a fluorescence microscope.

Quantitative real-time polymerase chain reaction (qPCR)
Cultured fibroblasts were stimulated with Recombinant Human IL-13 (Pepro Tech, Cranbury, NJ, USA) at 1, 10, and 100 ng/ml for 72 h, and total RNA was extracted using a FastGeen RNA Basic Kit (NIPPON Genetics, Tokyo, Japan) following the manufacturer's protocol. cDNA was synthesized using PrimeScript RT Master MIX (TaKaRa Bio, Shiga, Japan). Real-time qPCR was performed by StepOnePlus (Thermo Fisher Scientific) using THUNDERBIRD SYBR qPCR Mix (TOYOBO, Osaka, Japan). Periostin gene expression was measured and normalized by Glyceraldehyde-3-Phosphate Dehydrogenase (GAPDH). Primers were manufactured by Thermo Fisher Scientific. The primer sequences were as follows (F for forward, R for reverse): Periostin: F: CAG CAA ACC ACC TTC ACG GATC, R: TTA AGG AGG CGC TGA ACC ATGC, GAPDH: F: GTC TCC TCT GAC TTC AAC AGCG, R: ACC ACC CTG TTG CTG TAG CCAA. Data were analyzed by the ΔΔC T method, and the values were calculated as relative values to the non-treated control sample.

Statistical analyses
Categorical data were compared using the chi-squared test (with Yates' correction when appropriate). All statistical analyses were performed using the StatMate software program (Version 5 for Win&Mac Hybrid; Atms, Chiba, Japan).

Histological analyses of tissues resected from the hepatic hilar area of BA patients
A total of 50 samples (right side: 23 samples, left side: 19 samples, unknown region: 8 samples) from 32 BA patients were analyzed. The total number of bile ducts counted in the tissues was 745. We distinguished the bile ducts as large and micro ducts according to the diameter (≥150 or <150 µm, respectively). This was because the large ducts show more damaged epithelium and seemed to be original ducts, whereas micro ducts were round and rather well-preserved, presumed to be newly generated ducts or peribiliary glands (Fig. 1a-d). The total number of large ducts was 150, and the total number of micro ducts was 595.
To analyze the inflammatory cell infiltration around the bile ducts, immunostaining for CD45 was performed. CD45 was stained as a marker of leucocytes in small round cells forming clusters around the bile ducts (Fig. 1a). CD45 was positive around 659 of the 745 bile ducts (88.5%) and was particularly remarkable around the large ducts (140 of 150 ducts, 93.3%). IL13 staining was also performed, and the pattern was compared with the staining of CD45. IL13 staining was found among infiltrating cells as well as in the bile duct epithelium (Fig. 1c). IL13 was positive around 135 large ducts (90%) and 558 micro ducts (93%), with no marked preference of staining noted among duct sizes (Table 1). Regarding large ducts, IL13 was positive in 92.1% (p = 0.006) of CD45-positive ducts ( Table 2).
In summary, IL13 staining was found around large bile ducts with inflammatory cell infiltration as well as in the epithelium of micro bile ducts. Fibrotic markers α-SMA and periostin were significantly positive around large bile ducts, corresponding to IL13-positive regions.

Induction of periostin by IL13 stimulation in fibroblasts derived from BA tissue
The cultured cells, originally obtained from the hepatic hilar area tissues of a BA patient, were spindle-shaped and α-SMA positive by immunofluorescence and confirmed to be active fibroblasts (Fig. 2a). Both periostin mRNA expression and protein production from cells by recombinant IL13 Fig. 1 Examples of the bile duct sizes and immunohistochemical staining patterns in continuous sections of tissue. A large bile duct (≥150 µm) is located in the center, and micro ducts (<150 µm) surround the large duct. a CD45, b α-SMA,  stimulation were analyzed using qPCR and WB. Periostin mRNA expression was dose-dependently elevated by 72 h of IL13 stimulation, and the value with 100 ng/ml stimulation was about fourfold that in untreated cells (Fig. 2b). The protein analysis showed the different forms of periostin, including a 90-kDa secretory glycoprotein and 84/74-kDa isoform. Bands of these forms were stronger in a sample of cells with IL13 treatment than in untreated cells on WB (Fig. 2c).
In summary, we cultured fibroblasts obtained from the hepatic portal area of BA patients. Periostin production was induced by stimulation with recombinant IL13, and the upregulation was confirmed for both mRNA and protein.

Discussion
In this study, we found that IL13 was located at the region of inflammatory cell infiltration around the bile ducts of BA patients according to histological analyses of tissues from the hilar area of the liver. Furthermore, a downstream factor of IL13, periostin, was found around the remains of large bile ducts matched to the location of IL13 and α-SMA, at the site of active fibroblasts. The periostin production by IL13 was also confirmed by in vitro analyses with cultured fibroblasts from BA patients. These results show the actual process of fibrogenesis at the extrahepatic bile ducts of BA: the immunocytes around the bile ducts produce IL13, which has marked effects on fibroblasts and the production of periostin.
While several immunological studies have reported type 1 immune responses in BA patients [12,13], type 2 responses have been rarely reported. However, Li et al. reported that IL33, a member of the IL1 family of cytokines that is known to activate the type 2 immune response, was released from the epithelium of bile ducts by cytotoxic injury, causing cholangiocyte proliferation and rapid enlargement of the extrahepatic bile ducts in mice. They further showed that this response was mediated by IL-13 released from increased type 2 innate lymphoid cells (ILC2s) induced by IL33 stimulation [17]. Their findings correspond well to ours in tissues of BA patients. IL13 reportedly plays roles in cell proliferation in allergic diseases, such as asthma [23]. In our analyses, IL13 was detected not only among fibroblasts but also at the epithelium of micro bile ducts, which were found in the fibrotic remnants from the hepatic hilar area. It is suggested that IL13 plays roles in not only the fibrogenesis but also the proliferation of cholangiocytes and production of micro bile ducts.
Periostin has been well studied for its involvement in allergic inflammation and fibrosis disease, and it is known to be expressed in epithelium and fibroblasts [19,20]. Mitamura et al. [24] reported the involvement of oxidative stress in the expression of periostin, which is mediated by TGF-β1 and IL-13 in normal human dermal fibroblasts. The correlation between liver fibrosis and periostin has also been reported. A previous report revealed that periostin expression was significantly upregulated in a mouse model of chronic liver fibrosis induced by carbon tetrachloride and bile duct ligation [25]. Regarding BA, periostin levels in post-KPE serum samples have been assessed, and the association with ultrasonography hepatic fibrosis scores has been evaluated [22]. Our findings also suggest the association of periostin expression with active fibrogenesis around bile ducts at KPE. As a biomarker, periostin is detected in very high concentrations in serum and is easily measured by an immunoassay. Recently, many assay methods for periostin detection have been developed [21,26]. Serum periostin levels may thus be useful as a biomarker in BA before primary surgery. Further investigations and follow-up of the periostin expression, including serum, in BA patients are thus needed.
In conclusion, we found that IL13 was associated with cholangiocyte proliferation and induced periostin expression by fibroblasts, playing a vital role in the pathogenesis of fibrogenesis around the extrahepatic bile duct in BA. In the future, it will be necessary to accumulate data and conduct more detailed investigations into the involvement of periostin and IL13 in pathological conditions.

Competing interests
The authors declare no competing interests.