Expression pattern of cancer-associated fibroblast and its clinical relevance in intrahepatic cholangiocarcinoma

    Expression pattern of cancer-associated fibroblast and its clinical relevance in intrahepatic cholangiocarcinoma Xiao-fang Zhang, Min Dong, Yu-hang Pan, Jian-ning Chen, Xiang-qi Huang, Yi Jin, Chun-kui Shao PII: DOI: Reference:

S0046-8177(17)30132-6 doi: 10.1016/j.humpath.2017.04.014 YHUPA 4201

To appear in:

Human Pathology

Received date: Revised date: Accepted date:

10 January 2017 31 March 2017 19 April 2017

Please cite this article as: Zhang Xiao-fang, Dong Min, Pan Yu-hang, Chen Jian-ning, Huang Xiang-qi, Jin Yi, Shao Chun-kui, Expression pattern of cancer-associated fibroblast and its clinical relevance in intrahepatic cholangiocarcinoma, Human Pathology (2017), doi: 10.1016/j.humpath.2017.04.014

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT Expression pattern of cancer-associated fibroblast and its clinical relevance in intrahepatic cholangiocarcinoma

T

Xiao-fang Zhang1,3*, Min Dong2,3*, Yu-hang Pan1,3, Jian-ning Chen1,3, Xiang-qi Huang1,3, Yi Jin1,3, Chun-kui Shao 1,3† 1

MA NU

SC

RI P

Department of Pathology, The Third Affiliated Hospital, Sun Yat-sen University, Guangzhou, 510630, China 2 Department of Medical Oncology, The Third Affiliated Hospital, Sun Yat-sen University, Guangzhou, 510630, China 3 Guangdong Provincial Key Laboratory of Liver Disease Research, Guangzhou, China * Xiao-fang Zhang and Ming Dong contributed equally to this work. †

Corresponding author: Chun-kui Shao, MD, PhD. Department of Pathology, the Third Affiliated Hospital, Sun Yat-sen University, 600 Tianhe Rd, Guangzhou, 510630, China. Fax: +86-20-85253130. E-mail: [email protected].

AC

CE

PT

ED

Running head: Cancer-associated fibroblast in ICC

ACCEPTED MANUSCRIPT

ED

MA NU

SC

RI P

T

Abstract Intrahepatic cholangiocarcinoma (ICC) is a highly malignant neoplasm and lacks of effective treatment, characterized by dense desmoplastic stroma rich in cancer-associated fibroblasts (CAFs), which have been indicated to facilitate tumor progression in several types of human cancer. However, the clinical relevance of CAFs in ICC has not been fully characterized. Here, we evaluated the histological phenotype of CAFs and immunohistochemical expressions of α-SMA, FSP-1, and PDGFRβ in 71 ICC cases, and found that immature CAF phenotype was significantly associated with lymph node metastasis (P = 0.045), advance TNM stage (P = 0.025) and poor 5-year overall survival (OS) (38.5% vs. 78.6%, P = 0.015). In addition, α-SMA, FSP-1, and PDGFRβ were positively expressed in stromal fibroblasts in 63.4% (45/71), 84.5% (60/71), and 78.9% (56/71) of patients, respectively. Positive expression of α-SMA was correlated with poor differentiation (P = 0.032); FSP-1 expression in stromal fibroblasts was linked with lymph node metastasis (P = 0.022) and immature phenotype (P = 0.048). What’s more, positive expression of FSP-1 in cancer cells was observed in 22.5% (16/71) of cases and was correlated with worse 5-year OS (36.4% vs.76.7 %, P = 0.014). Importantly, in multivariate analysis, histological CAF phenotype was an independent prognostic factor for OS in ICC. Our findings demonstrated histological categorization of CAFs was a useful predictor for prognosis, providing new evidence that CAFs play crucial role in tumor progression and can serve as potential therapeutic targets in ICC.

AC

CE

PT

Keywords: Intrahepatic cholangiocarcinoma; cancer-associated fibroblast; α-SMA; FSP-1; PDGFRβ

ACCEPTED MANUSCRIPT

AC

CE

PT

ED

MA NU

SC

RI P

T

1. Introduction Intrahepatic cholangiocarcinoma (ICC) is the second commonest primary hepatic tumor [1]. Although it has been viewed as a relatively rare malignancy, interest in this disease is rising as the incidence of ICC increased markedly worldwide and its mortality rate remains high, with a 5-year overall survival (OS) rate of less than 5% [2]. Lacking specific symptoms, the majority of patients with ICC are diagnosed at advanced stages. Surgical resection offers the only strategy for curative treatment, but the recurrence and metastasis rates of advanced ICC were still high even after surgical resection, a previous study had reported a recurrence of more than 50% following resection of ICC [3]. Its resistance to chemo- and radio-treatment and lack of effective molecular targeted therapy also contribute to the poor outcome of patients [4]. Therefore, identification of novel predictor for patient outcome and possible therapeutic targets for individualized therapy in ICC is urgently required. In recent years, the importance of tumor microenvironment in tumor progression has been increasingly recognized. Cancer-associated fibroblasts (CAFs), which are defined as the fibroblastic cells found in a tumor [5], are the most prominent cell type within the tumor microenvironment and have been demonstrated to be a facilitator of tumor growth and progression in a variety of human cancers, including colorectal cancer [6], breast cancer [6], pancreatic cancer [7], and esophageal squamous cell carcinoma (ESCC) [8]. Although CAFs are sometimes referred to as myofibroblasts, they are actually a highly heterogeneous cell population. Morphologically, CAFs are comprised of at least two cell types: myofibroblasts and normal fibroblasts-like cells [9]. Myofibroblasts are characterized by large spindle-shaped cells with indented nuclei, while the latter cell type has a thin and small spindle shape of normal fibroblast. Both subtypes of CAFs have been reported to have tumor-promoting activities in mice models [10]. However, few studies have focused on the clinical significance of CAFs based on the histological morphology. Apart from the spindle-like morphology, several markers have also been suggested to define CAFs, such as α-smooth muscle actin (α-SMA) [11], fibroblast specific protein 1 (FSP-1) [12], and platelet-derived growth factor receptor β (PDGFRβ) [13]. α-SMA is a widely used marker to identify myofibroblasts [14]. FSP-1, also known as S100A4, is another common marker of CAFs and has been indicated to identify a unique population of CAFs distinct from α-SMA positive myofibroblasts [15]. Besides, PDGFR, including PDGFRα and PDGFRβ, is a kind of tyrosine kinase receptor which is mainly expressed in fibroblasts. PDGFRβ expression seems to be more common in general [16]. Despite all the markers, molecular definition of CAFs remains a challenge due to the lack of specific and exclusive markers. However, increasing studies have demonstrated that these markers have promising prognostic value. For example, in colorectal cancer, breast cancer, ESCC, and prostate cancer, expressions of α-SMA, FSP-1 or PDGFRβ in CAFs are associated with aggressive feature and poor prognosis [8, 17-19]. ICC is a highly malignant neoplasm with excessive stromal desmoplastic reaction. In the desmoplastic stroma of ICC, there are abundant ECM proteins and

ACCEPTED MANUSCRIPT

SC

RI P

T

α-SMA-positive CAFs surrounding the malignant ducts, glands or nests. Given the abundance of CAFs and the tumor-promoting role CAFs played in other cancer types, the impact of CAFs on ICC might be also considerable. However, studies with regard to histological morphology and molecular markers of CAFs in ICC are still limited, and their clinical relevance has not been fully characterized. Thus, this study aimed to evaluate a histological classification of CAFs and the expression of three CAF markers, α-SMA, FSP-1, and PDGFRβ, and their relationship with clinicopathological factors and prognosis in ICC.

AC

CE

PT

ED

MA NU

2. Materials and methods 2.1. Patients Formalin-fixed, paraffin-embedded tumor samples were collected from patients who underwent surgical resection for ICC at the Third Affiliated Hospital of Sun Yat-Sen University from 2010 to 2015. Clinicopathological data were obtained from the archives of the Department of Pathology at the Third Affiliated Hospital. All patients were staged according to seventh American Joint Committee on Cancer TNM staging system for ICC [20]. Tumor histological types and differentiation were categorized based on the grading system described by the World Health Organization classification [21]. Follow-up information was obtained by telephone to confirm the status of the patients by the time of December 31, 2015, and the date of death during the follow-up period. This study was approved by the Ethics Committee of the Third Affiliated Hospital of Sun Yat-Sen University. Written informed consent was collected from all patients before surgery. 2.2. Classification of CAFs by Histology CAF classification was evaluated by two pathologists who have no prior knowledge of clinicopathological data. According to their morphology on Hematoxylin and Eosin (HE) slides, fibroblasts were classified into mature and immature groups. Thin, wavy, and small spindle-shaped fibroblasts were defined as mature fibroblasts; fat, plump spindle-shaped fibroblasts with one or two nucleoli were regarded as immature fibroblasts (Figure 1A-B). When the immature fibroblasts accounted for more than 50% of the total fibroblasts within the tumor stoma, the case was considered as immature CAF phenotype, otherwise it was regarded as mature CAF phenotype [8]. 2.3. Immunohistochemical (IHC) staining The HE slides were reviewed to choose the representative tumor area for IHC. Then 4μm thick sections were cut from the corresponding paraffin-embedded tissue blocks and dried at 65℃ for 2 hours for IHC staining. We performed the IHC assay using the two-step EnVision procedure (Dako, Glostrup, Denmark). Sections on slides were dewaxed in xylene, rehydrated in gradient ethanol solutions, and heated with EDTA (pH = 8.0) for antigen retrieval in an autoclave. Hydrogen peroxide was applied to block the endogenous peroxidase activity. Sections were then incubated with primary antibodies against α-SMA (dilution 1:200, mouse monoclonal, PA0943; Novocastra, Leica Biosystems, NT, UK), FSP-1(dilution 1:100, rabbit polyclonal, ab27957; Abcam, Cambridge, MA, USA), or PDGFRβ (dilution 1:100, rabbit monoclonal,

ACCEPTED MANUSCRIPT

CE

PT

ED

MA NU

SC

RI P

T

ab32570; Abcam, Cambridge, MA, USA) for an hour at room temperature, followed by incubation with secondary antibody for 30 min. Diaminobenzidine (DAB) was used to visualize the immunostaining and hematoxylin was used to counterstain the sections. The known positive colon cancer tissues were used as positive controls, and slides stained with normal nonimmune mouse or rabbit serum rather than primary antibodies were used as negative controls. 2.4. Evaluation of the IHC staining The IHC results were evaluated by two pathologists who had no prior knowledge of clinicopathological data. On the basis of the staining intensity as well as the percentage of positive cells, the scores were defined as follows: 1, when the staining intensity was weak in < 50% or moderate in < 20% of stromal cells; 2, when the staining was weak in ≥ 50%, moderate in 20–50% or strong in < 20% of stromal cells; 3, when the staining was moderate in > 50% or strong in > 20% of the stromal cells. The representative images of different staining intensity were shown in Supplementary Figure 1. Sections with score ≥ 2 were determined as positive for each protein expression [8]. 2.5. Statistical analysis The relationship between protein expression and clinicopathological factors was analyzed using Chi-square test, Fisher’s exact test or Kruskal Wallis test. OS was defined as the period from the date of surgery to the date of death or the last day of follow-up. Survival curves were analyzed with the Kaplan-Meier method and compared using the log-rank test. Multivariate analysis of putative prognostic factors was evaluated in a Cox proportional hazards model. Two-sided P < 0.05 was considered statistically significant. Statistical analyses were performed using SPSS 20.0 statistics software (SPSS Inc., Chicago, IL, USA).

AC

3. Results 3.1. Patient Characteristics Patient clinicopathological features were summarized in Table 1. Tumors were classified as good differentiation in 2 (2.8%), moderate differentiation in 44 (62.0%), and poor differentiation in 25 (35.2%). Representative cases of tumor differentiation were shown in Figure 2. Among the 71 patients, 41 were available for follow-up information. During a median of 15.7 (range, 1.3–63.2) months of follow-up, 14 of the 41 patients (34.1%) died. 3.2. Histological classification of CAFs Sixteen of the 71 cases (22.5%) were classified as immature CAF phenotype and the rest 55 cases (77.5%) were mature phenotype (Figure 1A-B). The immature CAF phenotype was significantly associated with lymph node metastasis (P = 0.045) and advanced TNM stage (P = 0.025) (Table 1). Compared to the subgroup with mature CAF phenotype, patients with immature CAF phenotype had a remarkably poor 5-year OS (38.5% vs. 78.6%, P = 0.015) (Figure 1C). 3.3. Expression of α-SMA, FSP-1 and PDGFRβ and its correlation with clinicopathological factors As shown in Figure 3, in stromal fibroblasts, α-SMA was located in the cytoplasm,

ACCEPTED MANUSCRIPT

AC

CE

PT

ED

MA NU

SC

RI P

T

FSP-1 was expressed in both the cytoplasm and nucleus, and PDGFRβ was stained on the membrane. In addition, cytoplasmic and nuclear staining of FSP-1 in cancer cells was also observed. α-SMA and PDGFRβ were positively expressed in stromal fibroblasts in 63.4% (45/71) and 78.9% (56/71) of patients with ICC, respectively. Positive expression of FSP-1 in stromal fibroblasts and cancer cells was observed in 84.5% (60/71) and 22.5% (16/71) of cases, respectively. As shown in Table 2, α-SMA expression in stromal fibroblasts was observed more frequently in tumors with poor differentiation than those with good and moderate differentiation (P = 0.032). FSP-1 expression in stromal fibroblasts was associated with lymph node metastasis (P = 0.022) and immature CAF phenotype (P = 0.048). However, PDGFRβ expression in stromal fibroblasts and FSP-1 expression in cancer cells were not correlated with any clinicopathological parameters. Besides, relevance analysis showed that the expression of three markers in stromal fibroblasts was not significantly correlated with one another (Table 3). 3.4. α-SMA, FSP-1 and PDGFRβ expression and survival As shown in Figure 4, the expression of α-SMA, FSP-1, or PDGFRβ in stromal fibroblasts was not significantly associated with patient outcome (all P > 0.05, Figure 4A-C). However, positive expression of FSP-1 in cancer cells was significantly associated with poorer 5-year OS (positive vs. negative FSP-1 subgroups: 36.4% vs.76.7 %, P = 0.014, Figure 4D). 3.5. Univariate and Multivariate analysis In addition to histological CAF phenotype and FSP-1 expression in cancer cells, univariate analysis revealed that T stage (P = 0.017), lymph node metastasis (P = 0.029) and TNM stage (P = 0.037) were also prognostic factors for OS in ICC (Table 4). In multivariate analysis, only histological CAF phenotype was an independent prognostic factor for OS (hazard ratio [HR], 4.437; 95% confidence interval [CI]: 1.223–16.093, P = 0.023). 4. Discussion In the present study, we classified ICC cases into immature and mature CAF phenotypes according to the histological morphology of stromal fibroblasts and examined the expression of three CAF markers, including α-SMA, FSP-1, and PDGFRβ using IHC in 71 ICC patients. We found that immature CAF phenotype was significantly associated with lymph node metastasis and advanced TNM stage and independently predicted poor survival in ICC. In addition, positive α-SMA expression was correlated with poor differentiation, and FSP-1 expression in stromal fibroblasts was more frequently observed in cases with lymph node metastasis and immature CAF phenotype. Furthermore, we found that cases with FSP-1 expression in cancer cells had a worse 5-year OS. In our histological categorization of fibroblasts, the mature fibroblasts had a normal fibroblasts-like morphology, while the immature fibroblasts had a shape similar to myofibroblasts. In this study, we found the ICC patients with immature CAF phenotype had a more aggressive feature and significantly poorer OS than those with

ACCEPTED MANUSCRIPT

AC

CE

PT

ED

MA NU

SC

RI P

T

mature phenotype, consistent with previous studies in ESCC [8] and rectal cancer [22], indicating the two subtypes of CAFs have distinct activity in facilitating tumor progression. Immature fibroblasts, which might represent myofibroblasts, have a large oval nucleus with highlighted nucleoli, containing a rough endoplasmatic reticulum as well as a prominent Golgi apparatus, which might be the morphological reflection of their increased proliferative function and enhanced secretion of ECM constituents such as type I collagen, tenascin C, and fibronectin [23]. Increased tenascin-C expression in the stroma was correlated with increased invasiveness and poor prognosis in breast cancer [24] and bladder cancer [25]. During breast cancer progression, excess ECM deposition is correlated with a worse outcome [26]. Therefore, our results suggest that the histological CAF phenotype is a useful predictor for aggressive behavior and poor survival in ICC. Of all the CAF markers, α-SMA is the most well accepted one [11] and has been reported to be associated with shorter survival time in several human cancers, including colon [6], breast [6], and cholangiocarcinoma [27]. In our study, we found α-SMA was positively expressed in CAFs in 63.4% of patients with ICC, consistent with the high α-SMA expression rate of 60% in a previous study in cholangiocarcinoma [27]. However, although positive α-SMA expression was associated with poor differentiation in the present study, we could not detect its prognostic value reported in a previous study performed in ICC [27]. This discrepancy may due to the relatively smaller sample size and different IHC scoring system in our study. FSP-1, a member of the calcium-binding protein family [28], is produced mainly by fibroblasts. The intracellular FSP-1 increases cell mobility by interacting with cytoskeleton such as nonmuscle myosin heavy chain (NMMHC) IIA and actin, thus stimulates the movement of fibroblasts and enlarges their range interacting with tumor cells. In addition, intracellular FSP-1 expression also facilitates the production of MMP-13, which plays a key role in tumor metastasis [29]. In the present study, we found FSP-1 was abundantly expressed in stromal fibroblasts and its expression was correlated with lymph node metastasis, consistent with its property of promoting tumor metastasis. It is noteworthy that FSP-1 can be expressed not only in stromal fibroblasts, but also in tumor cells. Positive FSP-1 expression in tumor cells was found in 22.5% of cases with ICC in our study, and it was significantly associated with poor survival. It has been demonstrated that epithelial cells undergoing EMT also express FSP-1 and it can be used as an EMT marker [30]. Thus, FSP-1-positive tumor cells might represent a group of cells which are undergoing EMT, in line with the function of FSP-1 as a negative prognostic factor in ICC in our study. PDGFRβ is another extensively expressed CAF marker. Although stromal PDGFRβ positivity has been reported to be associated with shorter survival time in prostate cancer [18], breast cancer [31], and papillary thyroid cancer [32], its correlation with clinicopathological factors and prognosis could not be detected in ICC in the present study. Considering the relatively small sample size in this study, the prognostic value of PDGFRβ in ICC should be further investigated in large scale study.

ACCEPTED MANUSCRIPT

RI P

T

In conclusion, our study demonstrated that histological immature CAF phenotype and FSP-1 expression in cancer cells were correlated with poor clinical outcome in patients with ICC. Cox regression multivariate analysis confirmed that histological CAF phenotype was an independent prognostic factor in ICC. With the increasing evidence that CAFs play a crucial role in tumor-promoting, CAFs may serve as novel targets of cancer therapy in the future. Our findings might provide new evidence for CAFs in facilitating tumor progression and potential therapeutic targets for ICC.

MA NU

SC

Acknowledgements This work was supported by National Natural Science Foundation of China (No. 81272553 & No.81572309) and Natural Science Foundation of Guangdong Province (No. S2012020010898 & No. 2014A030313034). Conflict of Interest All the authors declare that they have no conflict of interest.

AC

CE

PT

ED

References Khan SA, Taylor-Robinson SD, Toledano MB, Beck A, Elliott P, Thomas HC. Changing international trends in mortality rates for liver, biliary and pancreatic tumours. Journal of hepatology 2002; 37, 806-813. Shaib Y, El-Serag HB. The epidemiology of cholangiocarcinoma. Seminars in liver disease 2004; 24, 115-125. Endo I, Gonen M, Yopp AC, Dalal KM, Zhou Q, Klimstra D, D'Angelica M, DeMatteo RP, Fong Y, Schwartz L, Kemeny N, O'Reilly E, Abou-Alfa GK, Shimada H, Blumgart LH, Jarnagin WR. Intrahepatic cholangiocarcinoma: rising frequency, improved survival, and determinants of outcome after resection. Annals of surgery 2008; 248, 84-96. Sirica AE, Dumur CI, Campbell DJ, Almenara JA, Ogunwobi OO, Dewitt JL. Intrahepatic cholangiocarcinoma progression: prognostic factors and basic mechanisms. Clinical gastroenterology and hepatology : the official clinical practice journal of the American Gastroenterological Association 2009; 7, S68-78. Ohlund D, Elyada E, Tuveson D. Fibroblast heterogeneity in the cancer wound. The Journal of experimental medicine 2014; 211, 1503-1523. Hanley CJ, Noble F, Ward M, Bullock M, Drifka C, Mellone M, Manousopoulou A, Johnston HE, Hayden A, Thirdborough S, Liu Y, Smith DM, Mellows T, Kao WJ, Garbis SD, Mirnezami A, Underwood TJ, Eliceiri KW, Thomas GJ. A subset of myofibroblastic cancer-associated fibroblasts regulate collagen fiber elongation, which is prognostic in multiple cancers. Oncotarget 2016; 7, 6159-6174. Ino Y, Yamazaki-Itoh R, Oguro S, Shimada K, Kosuge T, Zavada J, Kanai Y, Hiraoka N. Arginase II expressed in cancer-associated fibroblasts indicates tissue hypoxia and predicts poor outcome in patients with pancreatic cancer. PloS one 2013; 8, e55146. Ha SY, Yeo SY, Xuan YH, Kim SH. The prognostic significance of cancer-associated

ACCEPTED MANUSCRIPT

AC

CE

PT

ED

MA NU

SC

RI P

T

fibroblasts in esophageal squamous cell carcinoma. PloS one 2014; 9, e99955. Hanahan D, Weinberg RA. Hallmarks of cancer: the next generation. Cell 2011; 144, 646-674. Kuperwasser C, Chavarria T, Wu M, Magrane G, Gray JW, Carey L, Richardson A, Weinberg RA. Reconstruction of functionally normal and malignant human breast tissues in mice. Proceedings of the National Academy of Sciences of the United States of America 2004; 101, 4966-4971. Tomasek JJ, Gabbiani G, Hinz B, Chaponnier C, Brown RA. Myofibroblasts and mechano-regulation of connective tissue remodelling. Nature reviews Molecular cell biology 2002; 3, 349-363. Strutz F, Okada H, Lo CW, Danoff T, Carone RL, Tomaszewski JE, Neilson EG. Identification and characterization of a fibroblast marker: FSP1. The Journal of cell biology 1995; 130, 393-405. Pietras K, Sjoblom T, Rubin K, Heldin CH, Ostman A. PDGF receptors as cancer drug targets. Cancer cell 2003; 3, 439-443. Serini G, Gabbiani G. Mechanisms of myofibroblast activity and phenotypic modulation. Experimental cell research 1999; 250, 273-283. Sugimoto H, Mundel TM, Kieran MW, Kalluri R. Identification of fibroblast heterogeneity in the tumor microenvironment. Cancer biology & therapy 2006; 5, 1640-1646. Ostman A, Heldin CH. PDGF receptors as targets in tumor treatment. Advances in cancer research 2007; 97, 247-274. Surowiak P, Suchocki S, Gyorffy B, Gansukh T, Wojnar A, Maciejczyk A, Pudelko M, Zabel M. Stromal myofibroblasts in breast cancer: relations between their occurrence, tumor grade and expression of some tumour markers. Folia histochemica et cytobiologica / Polish Academy of Sciences, Polish Histochemical and Cytochemical Society 2006; 44, 111-116. Hagglof C, Hammarsten P, Josefsson A, Stattin P, Paulsson J, Bergh A, Ostman A. Stromal PDGFRbeta expression in prostate tumors and non-malignant prostate tissue predicts prostate cancer survival. PloS one 2010; 5, e10747. Tsujino T, Seshimo I, Yamamoto H, Ngan CY, Ezumi K, Takemasa I, Ikeda M, Sekimoto M, Matsuura N, Monden M. Stromal myofibroblasts predict disease recurrence for colorectal cancer. Clinical cancer research : an official journal of the American Association for Cancer Research 2007; 13, 2082-2090. Edge SB, Compton CC. The American Joint Committee on Cancer: the 7th edition of the AJCC cancer staging manual and the future of TNM. Annals of surgical oncology 2010; 17, 1471-1474. Nakanuma Y, Sato Y, Harada K, Sasaki M, Xu J, Ikeda H. Pathological classification of intrahepatic cholangiocarcinoma based on a new concept. World journal of hepatology 2010; 2, 419-427. Ueno H, Jones AM, Wilkinson KH, Jass JR, Talbot IC. Histological categorisation of fibrotic cancer stroma in advanced rectal cancer. Gut 2004; 53, 581-586. Kalluri R, Zeisberg M. Fibroblasts in cancer. Nature reviews Cancer 2006; 6, 392-401.

ACCEPTED MANUSCRIPT

AC

CE

PT

ED

MA NU

SC

RI P

T

Ishihara A, Yoshida T, Tamaki H, Sakakura T. Tenascin expression in cancer cells and stroma of human breast cancer and its prognostic significance. Clinical cancer research : an official journal of the American Association for Cancer Research 1995; 1, 1035-1041. Brunner A, Mayerl C, Tzankov A, Verdorfer I, Tschorner I, Rogatsch H, Mikuz G. Prognostic significance of tenascin-C expression in superficial and invasive bladder cancer. Journal of clinical pathology 2004; 57, 927-931. Walker RA. The complexities of breast cancer desmoplasia. Breast cancer research : BCR 2001; 3, 143-145. Chuaysri C, Thuwajit P, Paupairoj A, Chau-In S, Suthiphongchai T, Thuwajit C. Alpha-smooth muscle actin-positive fibroblasts promote biliary cell proliferation and correlate with poor survival in cholangiocarcinoma. Oncol Rep 2009; 21, 957-969. Garrett SC, Varney KM, Weber DJ, Bresnick AR. S100A4, a mediator of metastasis. The Journal of biological chemistry 2006; 281, 677-680. Donato R, Cannon BR, Sorci G, Riuzzi F, Hsu K, Weber DJ, Geczy CL. Functions of S100 proteins. Current molecular medicine 2013; 13, 24-57. Okada H, Danoff TM, Kalluri R, Neilson EG. Early role of Fsp1 in epithelial-mesenchymal transformation. The American journal of physiology 1997; 273, F563-574. Paulsson J, Sjoblom T, Micke P, Ponten F, Landberg G, Heldin CH, Bergh J, Brennan DJ, Jirstrom K, Ostman A. Prognostic significance of stromal platelet-derived growth factor beta-receptor expression in human breast cancer. The American journal of pathology 2009; 175, 334-341. Sun WY, Jung WH, Koo JS. Expression of cancer-associated fibroblast-related proteins in thyroid papillary carcinoma. Tumour biology : the journal of the International Society for Oncodevelopmental Biology and Medicine 2016; 37, 8197-8207.

ACCEPTED MANUSCRIPT

AC

CE

PT

ED

MA NU

SC

RI P

T

Figure legends Fig.1 Histological classification of stromal fibroblast on hematoxylin and eosin slides. (A) Mature type, fibroblasts show thin and small spindle cell morphology. (B) Immature type, fibroblasts show large and plump spindle cell morphology. (C) Kaplan-Meier survival curves for overall survival according to histologic subtype of CAF. Patients with immature CAF phenotype had a worse overall survival than those with mature phenotype. Fig.2 Representative area of intrahepatic cholangiocarcinoma with (A) good differentiation (original magnification, ×100), (B) moderate differentiation (original magnification, ×100), and (C) poor differentiation (original magnification, ×100). Fig.3 α-SMA, FSP-1 and PDGFRβ expression in ICC. Representative cases showed HE and positive IHC staining of α-SMA, FSP-1, and PDGFRβ in ICC; the left column (HE staining, original magnification, ×100), the middle column (positive IHC staining, original magnification, ×100), the right column (positive IHC staining, original magnification, ×400). Fig.4 Kaplan-Meier survival curves for overall survival in ICC according to α-SMA, FSP-1, and PDGFRβ expression. No significant difference was observed between positive and negative expression of α-SMA (A), FSP-1 in fibroblasts (B), or PDGFRβ (C). Patients with positive FSP-1 expression in cancer cells had a worse OS than those with negative FSP-1 expression (D).

AC

CE

PT

ED

MA NU

SC

RI P

T

ACCEPTED MANUSCRIPT

Figure 1

T

ACCEPTED MANUSCRIPT

AC

CE

PT

ED

MA NU

SC

RI P

Figure 2

MA NU

SC

RI P

T

ACCEPTED MANUSCRIPT

AC

CE

PT

ED

Figure 3

AC

Figure 4

CE

PT

ED

MA NU

SC

RI P

T

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIPT

AC

CE

PT

ED

MA NU

SC

RI P

T

Table 1. Comparison of clinicopathologic characteristics according to histological subtypes of cancer associated fibroblast. Subtypes of cancer associated fibroblast P value N(total) Mature Immature 71 55 16 Age (years) < 60 43 36 7 0.118 ≥ 60 28 19 9 Gender Male 33 24 9 0.373 Female 38 31 7 Tumor size < 5cm 32 22 10 0.111 ≥ 5cm 39 33 6 Tumor location Perihilar 9 8 1 0.652 Peripheral 62 47 15 Histological subtypes Papillary 32 26 6 0.798 adenocarcinoma Tubular 28 21 7 adenocarcinoma Others 11 8 3 Differentiation Well/moderate 46 34 12 0.388 Poor 25 21 4 Perineural invasion Positive 11 10 1 0.246 Negative 60 45 15 Vascular invasion Positive 5 5 0 0.211 Negative 66 50 16 a T stage T1＋T2 61 48 13 0.684 T3＋T4 10 7 3 a N stage N0 42 36 6 0.045* N1 29 19 10 a TNM stage I + II 36 32 4 0.025*

ACCEPTED MANUSCRIPT III + IV 35 23 th The 7 AJCC TNM staging system (20). * Indicates statistical significance.

12

AC

CE

PT

ED

MA NU

SC

RI P

T

a

ACCEPTED MANUSCRIPT

AC

CE

PT

ED

MA NU

SC

RI P

T

Table 2. α-SMA, FSP-1, and PDGFRβ expression in relation to clinicopathological characteristics in intrahepatic cholangiocarcinoma. α-SMA in FSP-1 in PDGFRβ in FSP-1 in fibroblasts fibroblasts fibroblasts cancer cells Characteris P P P tics positi negati P positi negati posit negati posi negati valu valu valu ve ve value ve ve ive ve tive ve e e e Age(years) 1.00 0.1 < 60 28 15 0.707 36 7 34 9 0.96 7 36 0 18 ≥ 60 17 11 24 4 22 6 9 19 Gender 0.32 0.54 0.8 Male 25 13 0.651 34 4 31 7 9 29 5 9 04 Female 20 13 26 7 25 8 7 26 Tumor size 0.97 0.30 0.6 < 5cm 18 14 0.259 27 5 27 5 8 24 8 4 53 ≥ 5cm 27 12 33 7 29 10 8 31 Tumor location 0.91 1.00 1.0 Perihilar 3 6 0.103 7 2 7 2 2 7 7 0 00 Peripheral 42 20 53 9 49 13 14 48 Histologica l subtypes Papillary 0.05 0.79 0.9 22 10 0.674 30 2 26 6 8 24 carcinoma 1 7 32 Tubular 16 12 23 5 22 6 6 22 carcinoma Others 7 4 7 4 8 3 2 9 Differentia tion Well/mode 25 rate Poor 20 Perineural invasion Positive

6

Negative 39

21

0.032

36

10

5

24

1

5

0.508 10

1

21

*

50

10

0.08 37 3

9

0.66 9 2

37

19

6

7

18

0.52 9 3 47

2 13

0.79 2 9 5 14 46

0.4 17

1.0 00

ACCEPTED MANUSCRIPT Vascular invasion

T1＋T2

40

T3 + T4 5 N stagea

25

55

21

0.343 52

5

0 11

9

8

N0

24

18

N1 TNM stagea

21

8

I + II

22

14

0.687 28

III + IV 23 CAF phenotype

12

32

2

0.189 32

ED

28

10

0.32 4 1 52 0.64 48 8 8

1

8 3

0.02 32 2* 24

0.18 27 9 29

0.04 44 8* Immature 10 6 11 5 12 Abbreviations: CAF, cancer associated fibroblast a The 7th AJCC TNM staging system (20). * Indicates statistical significance. 20

AC

0.934 49

PT

35

CE

Mature

6

1 14

1.00 2 3 0 14 52

13 2

10 5

9 6

11 4

0.6 79

T

Negative 41 T stsgea

0.646 5

RI P

1

SC

4

MA NU

Positive

1.00 12 49 0 4 6

0.2 17

0.50 7 5 9

0.1 54

35 20

0.41 5 31 7 11 24

0.0 77

0.73 11 44 1 5 11

0.4 97

ACCEPTED MANUSCRIPT

AC

CE

PT

ED

MA NU

SC

RI P

T

Table 3. Correlation analysis of α-SMA, FSP-1, and PDGFRβ expression. FSP-1 PDGFRβ P value P value negativ r negativ r positive positive e e α-SMA positive 39 6 0.079 0.515 37 8 0.108 0.370 negative 21 5 19 7 FSP-1 positive 49 11 0.160 0.183 negative 7 4

ACCEPTED MANUSCRIPT

AC

CE

PT

ED

MA NU

SC

RI P

T

Table 4. Cox proportional Hazard regression analysis of patients’ overall survival in intrahepatic cholangiocarcinoma. Univariate Multivariate analysis analysis Variables Categories P value HR 95% CI P value Age (years) ≥ 60 vs < 60 0.525 NA Gender male vs female 0.304 NA Tumor size ≥ 5cm vs < 5cm 0.086 2.679 0.596-12.034 0.199 perihilar vs Tumor location 0.128 NA peripheral papillary vs Histological tubular vs 0.142 NA subtypes of ICC others poor vs Differentiation 0.394 NA well/moderate Perineural positive vs 0.180 NA invasion negative Vascular positive vs 0.457 NA invasion negative T3+T4 vs 0.074 T stagea 0.017* 4.283 0.867-21.150 T1+T2 a N stage N1 vs N0 0.029* 7.193 0.926-55.856 0.557 a * TNM stage III + IV vs I+II 0.037 7.829 0.488-125.674 0.146 Histological 0.023* immature vs * subtypes of 0.015 4.437 1.223-16.093 mature CAFs α-SMA positive vs expression in 0.094 2.272 0.383-13.463 0.366 negative CAFs FSP-1 positive vs expression in 0.396 NA negative CAFs PDGFRβ positive vs expression in 0.264 NA negative CAFs FSP-1 0.218 positive vs * expression in 0.014 2.461 0.587-10.327 negative tumor cells Abbreviations: CAF, cancer associated fibroblast; CI, confidence interval; HR, hazard ratio; NA, not applicable. a The 7th AJCC TNM staging system (20). * Indicates statistical significance.

ACCEPTED MANUSCRIPT

AC

CE

PT

ED

MA NU

SC

RI P

T

Highlights Histological CAF phenotype was an independent prognostic factor for OS in ICC. FSP-1 expression in cancer cells was correlated with worse 5-year OS in ICC. CAFs can be potential therapeutic targets in ICC.