Sage Journals: Discover world-class research

Abstract

Cancer-associated fibroblasts are abundant in the desmoplastic stroma of pancreatic ductal adenocarcinomas and are considered to play important roles in tumor progression. In this study, we investigated the expression status of secreted protein acidic and rich in cysteine, periostin, fibroblast-activated protein, and the newly developed proCOL11A1 antibody in the stroma of surgically resected pancreatic ductal adenocarcinomas and their prognostic implications. Tissue microarrays were constructed from 155 surgically resected pancreatic ductal adenocarcinomas and paired non-neoplastic pancreata and from another independent set of 48 normal/benign pancreata, and immunohistochemical stains were performed for proCOL11A1, fibroblast-activated protein, secreted protein acidic and rich in cysteine, and periostin. The immunohistochemical stain results were correlated with clinicopathological features and survival data. proCOL11A1, fibroblast-activated protein, secreted protein acidic and rich in cysteine, and periostin expression was significantly increased in the intratumoral stroma of pancreatic ductal adenocarcinomas compared to paired non-neoplastic pancreata (proCOL11A1: 145/155 (93.5%) vs 26/154 (16.9%); fibroblast-activated protein: 139/143 (97.2%) vs 82/132 (62.1%); secreted protein acidic and rich in cysteine: 113/150 (75.3%) vs 49/132 (37.1%); periostin: 135/151 (89.4%) vs 45/135 (33.3%); p < 0.001, all). While the four markers were expressed at lower levels in normal/benign pancreata, there were no significant differences in the expression frequencies among normal pancreas, acute pancreatitis, and chronic pancreatitis. Interestingly, on survival analysis, low intratumoral fibroblast-activated protein⁺ cancer-associated fibroblast counts (<100/high-power field) were associated with a significantly reduced overall survival compared to those with high fibroblast-activated protein⁺ cancer-associated fibroblast counts (p = 0.010; hazard ratio 5.2 (95% confidence interval 1.3–21.3)). Similar patterns were seen for proCOL11A and secreted protein acidic and rich in cysteine and overall and disease-free survival, although not statistically significant. In conclusion, we demonstrate that the presence of cancer-associated fibroblasts in the tumor stroma may not always be associated with a poor prognosis as suggested in many studies; on the contrary, it may even be associated with prolonged survival, supporting the recent experimental findings that tumor stroma may have a protective role rather than enhance aggressive behavior.

Keywords

Pancreatic ductal adenocarcinoma cancer-associated fibroblast fibroblast-activated protein proCOL11A1 secreted protein acidic and rich in cysteine periostin prognosis

Introduction

The tumor microenvironment is composed of a diverse population of cells, including tumor cells, inflammatory cells, fibroblasts, endothelial cells and pericytes, and extracellular matrix.^1,2 This milieu is receiving increasing attention as a target for cancer treatment, as it is becoming increasingly evident that there is cross-talk between tumor epithelial cells and the surrounding stroma and that the tumor microenvironment may play important roles in tumor initiation, progression, metastasis, and resistance to chemotherapy. Cancer-associated fibroblasts (CAFs), which are fibroblasts that are irreversibly activated due to persistent injurious stimuli, constitute a large proportion of the tumor microenvironment and are easily recognizable in the desmoplastic reaction of cancers.³ Pancreatic ductal adenocarcinoma (PDAC) is a prototypical example of a cancer with extensive desmoplastic reaction; the fibrous component often exceeds 80% of the entire tumor volume. As PDAC is one of the most lethal solid malignancies with a 5-year survival rate of less than 5% and mean survival time from diagnosis of 6 months,^4,5 and the current chemotherapeutic options have limited impacts on the natural course of PDAC, targeting the stromal components of the tumor is becoming a field of extensive research for this cancer.

While abundant desmoplastic reaction is a hallmark of PDAC, the stroma itself is heterogeneous, with an increasing list of CAF-related markers for PDAC in the recent literature. Recently, Moffitt et al.⁶ have described tumor- and stroma-specific subtypes of PDAC through virtual microdissection of tumor, stromal, and normal gene expression signatures, and they defined two prognostically different stromal subtypes—“normal stroma” and “activated stroma”—of PDAC, supporting the complexity of the tumor stroma. The “activated stroma” subtype PDACs demonstrated overexpression of genes such as SPARC, POSTN, FAP, COL11A1, and THBS2, while the “normal stroma” subtype expressed genes encoding known markers for pancreatic stellate cells such as vimentin, desmin, and smooth muscle actin.⁶

As there is a limited number of human clinicopathological studies on the tumor stroma of PDAC, and also a limited number of available antibodies that label CAFs, we investigated the expression status of secreted protein acidic and rich in cysteine (SPARC), periostin (POSTN), fibroblast-activated protein (FAP), and the newly developed proCOL11A1 antibody⁷ in our cohort of surgically resected PDACs and correlated their expression statuses with survival.

Materials and methods

Case selection and tissue microarray construction

This study was approved by the Institutional Review Board of Seoul National University Bundang Hospital (IRB# B-1702-383-302). In all, 155 surgically resected PDACs, retrieved from the surgical pathology files of Seoul National University Bundang Hospital from 2004 to 2012, were included. None of the cases received neoadjuvant chemotherapy or radiotherapy. The electronic medical records, surgical pathology reports, and hematoxylin- and eosin-stained slides were reviewed for each case, and the clinicopathological variables noted included age, sex, tumor size, tumor differentiation, angiolymphatic and venous invasion, and the pathologic tumor and nodal stages by both American Joint Committee on Cancer (AJCC) tumor–node–metastasis (TNM) 7th and 8th editions. The clinicopathological characteristics are summarized in Table 1.

Table 1.

Clinicopathological summary of the pancreatic ductal adenocarcinomas (n = 155).

Age (year, mean ± SD)	64.56 ± 9.95
Sex (n (%))
Male	91 (58.7)
Female	64 (41.3)
Tumor size^a (cm, mean ± SD)	3.53 ± 1.54
Differentiation (n (%))
Well differentiated	7 (4.5)
Moderately differentiated	138 (89.0)
Poorly differentiated	10 (6.5)
Angiolymphatic invasion (n (%))
Absent	71 (45.8)
Present	84 (54.2)
Venous invasion (n (%))
Absent	87 (56.1)
Present	68 (43.9)
Perineural invasion (n (%))
Absent	23 (14.8)
Present	132 (85.2)
Follow-up
Recurrence^b (n (%))	97 (62.6)
Death due to PDAC (n (%))	82 (52.9)

SD: standard deviation; PDAC: pancreatic ductal adenocarcinoma.

Largest dimension of tumor.

Local recurrence or distant metastasis.

Follow-up data were also obtained from the medical records, including recurrences (including local recurrence and distant metastasis) and death. The median follow-up period was 25.4 months (range: 0.9–109.7 months). Overall survival was defined as the interval from initial surgical treatment for PDAC to the date of death, and disease-free survival was defined as the interval from initial surgical treatment to local or distant recurrence.

Tissue cores measuring 2 mm in diameter were sampled from formalin-fixed paraffin-embedded PDAC tissues and arranged in recipient tissue microarray blocks using a trephine apparatus (Superbiochips Laboratories, Seoul, Korea). Two or three cores were sampled from each PDAC, depending on the amount of histological heterogeneity present in the tumor. One core was sampled from the paired non-neoplastic pancreas for each case for comparison.

In addition, a separate tissue microarray was constructed from an independent set of 48 benign pancreata, surgically resected between 2007 and 2012 at Seoul National University Bundang Hospital. These cases were resected for reasons other than PDAC, including benign cysts, traumatic lacerations, and arteriovenous malformations, and representative tissue cores from histologically normal pancreas (n = 27), acute pancreatitis (n = 6), and chronic pancreatitis (n = 15) were sampled for tissue microarray construction.

Immunohistochemistry

Immunohistochemical stains for proCOL11A1 (1:400, mouse monoclonal, clone 1E8.33; Oncomatryx, Derio, Spain), POSTN (1:100, rabbit polyclonal; Abcam, Cambridge, UK), SPARC (1:100, rabbit polyclonal; Santa Cruz Biotechnology, Dallas, TX, USA), FAP (1:50, rabbit polyclonal; Abcam), S100A4 (1:1500, rabbit polyclonal; Dako, Glostrup, Denmark), and uPAR (1:40, mouse monoclonal; Abcam) were performed on 4-μm-thick sections obtained from tissue microarray blocks. Briefly, tissue sections were deparaffinized in xylene and rehydrated in graded alcohol, and antigen retrieval was performed using citrate buffer (pH 6.0) for 20 min. Sections were incubated with primary antibodies at room temperature for 60 min and then with secondary antibodies (EnVision Detection System; Dako) for 30 min. Counterstaining was performed using Mayer’s hematoxylin.

proCOL11A1, SPARC, POSTN, and FAP were expressed in the fibroblast cytoplasm (Figure 1). proCOL11A1- and FAP-positive fibroblasts were manually counted under the microscope at 400× magnification (high-power field (HPF)) for each case, and a total of 10 HPFs were counted for each case. PDACs showing staining for these markers in at least one fibroblast/HPF were deemed “positive.” As the staining intensity varied for SPARC and POSTN, staining for these two markers was analyzed in a semiquantitative manner, as previously described.^8,9 Immunolabeling was scored as negative, weak, moderate, or strong positive staining in the stroma, and SPARC or POSTN positivity was defined as moderate or strong staining. Positivity for S100A4 and uPAR was defined as nuclear and/or cytoplasmic expression in at least 10% of tumor epithelial cells.

Figure 1.

Immunohistochemical stain results. The cytoplasmic staining in intratumoral fibroblasts is seen for (a and b) proCOL11A1, (c) FAP, (d–f) SPARC ((d) weak, (e) moderate, and (f) strong positive), and (g–i) POSTN ((g) weak, (h) moderate, and (i) strong positive). Focal weak cytoplasmic expression in the tumor epithelial cells was seen for (f) SPARC and (h) POSTN. ((a) and (c–i) 200× magnification; (b) 400× magnification).

Statistical analysis

All statistical analyses were conducted using IBM SPSS Statistics version 19–23 (IBM Corp., Armonk, NY, USA). Chi-square and t-tests were performed as deemed appropriate. Univariate analyses for overall survival and disease-free survival were performed using the Kaplan–Meier method and log-rank tests. Statistical significance was defined as p < 0.05. Statistically significant variables from the univariate analysis and those that met the proportional hazard assumptions were further analyzed by the Cox proportional hazard method.

Results

Expression of proCOL11A1, FAP, SPARC, and POSTN in PDAC and paired non-neoplastic pancreata

The results are summarized in Figures 1 and 2 and Table 2. proCOL11A1-positive fibroblasts were seen in at least small amounts in 145/155 (93.5%) PDACs and in 26/154 (16.9%) matched non-neoplastic pancreata. proCOL11A1-positive fibroblasts were significantly more abundant in PDACs (48.3 ± 127.8/HPF, range: 0–926/HPF) compared to non-neoplastic tissues (1.8 ± 6.9/HPF, range: 0–44/HPF; p < 0.001, t-test). Similarly, FAP-positive fibroblasts were more frequently present in PDACs (139/143, 97.2%)compared to matched non-neoplastic pancreata (82/132, 62.1%), with a significant difference in FAP-positive fibroblast count (tumor: 49.4 ± 39.4/HPF, range: 0–183/HPF vs non-tumor: 11.4 ± 15.4/HPF, range: 0–61/HPF; p < 0.001, t-test). There were no significant differencesin the staining intensity for COL11A1 and FAP, andthese markers were not expressed in the tumor epithelial cells.

Figure 2.

Box plots and bar graphs demonstrating the differences in (a) proCOL11A1, (b) FAP, (c) SPARC, and (d) POSTN expression between PDAC and paired non-neoplastic pancreata.

Table 2.

The expression status of proCOL11A1, SPARC, POSTN, and FAP in pancreatic ductal adenocarcinomas and matched non-neoplastic pancreas.

	Pancreatic ductal adenocarcinoma (n (%))	Matched non-neoplastic pancreas (n (%))	p value
proCOL11A1	145/155 (93.5%)	26/154^a (16.9%)	p < 0.001
FAP	139/143^a (97.2%)	82/132^a (62.1%)	p < 0.001
SPARC	113/150^a (75.3%)	49/132^a (37.1%)	p < 0.001
POSTN	135/151^a (89.4%)	90/135^a (66.7%)	p < 0.001

Missing tissue cores were present after performing immunohistochemical stains for multiple antibodies.

SPARC and POSTN expression varied in intensity from weak, moderate, to strong. 113/150 (75.3%) cases showed either moderate or strong SPARC expression in the intratumoral fibroblasts (moderate: 76/150 (50.7%), strong: 37/150 (24.7%)). In contrast, SPARC positivity was less frequent in non-neoplastic pancreata (49/132 (37.1%)), with only 9 (5.8%) cases showing strong expression (p < 0.001). POSTN expression was seen in 135/151 (89.4%) cases (moderate: 95/151 (62.9%), strong: 40/151 (26.5%)). POSTN expression in the non-neoplastic pancreata was less frequent (45/135 (33.3%), p < 0.001, Chi-square test), and strong expression was seen in only 10 (7.4%) cases. There were no significant correlations between the expression statuses of the four markers.

Expression of proCOL11A1, FAP, SPARC, and POSTN in normal and benign pancreata

As the four markers were expressed in paired non-neoplastic pancreata, albeit at low levels, we investigated whether they were also expressed in histologically normal pancreata or in acute or chronic pancreatitis by staining an additional tissue microarray containing a separate cohort of 48 normal and benign pancreata. Overall, proCOL11A1 and FAP-positive fibroblasts were seen in 68.8% (81.6 ± 87.0/HPF, range: 0–360) and 62.5% (43.2 ± 75.5/HPF, range: 0–472) of the cases. 21/48 (43.8%) of cases expressed POSTN (moderate: 20 (41.7%), strong: 1 (2.1%)) and 25/48 (52.1%) expressed SPARC (moderate: 13 (27.1%), strong: 12 (25.0%)) (Table 3, Supplementary Figure 1).

Table 3.

The expression status of proCOL11A1, SPARC, POSTN, and FAP in an independent non-neoplastic pancreas cohort (n = 48).

	Normal pancreas (n (%))	Acute pancreatitis (n (%))	Chronic pancreatitis^a (n (%))	p value
proCOL11A1	7/27 (25.9%)	1/6 (16.7%)	7/15 (46.2%)	p = 0.607
FAP	8/27 (29.6%)	2/6 (33.3%)	8/15 (53.8%)	p = 0.537
SPARC	13/27 (48.1%)	3/6 (50%)	7/15 (46.2%)	p = 0.990
POSTN	12/27 (44.4%)	4/6 (66.7%)	11/15 (76.9%)	p = 0.168

Two cases of autoimmune pancreatitis were included in the chronic pancreatitis group.

When the expression status of proCOL11A1, FAP, SPARC, and POSTN was compared according to the histological classification (normal, acute pancreatitis, and chronic pancreatitis groups), no significant differences were seen between the three groups. Notably, although two of the chronic pancreatitis cases were autoimmune pancreatitis, characterized by an abundant number of fibroblasts arranged in a storiform pattern, there were no differences in the expression status of the four markers between autoimmune pancreatitis and the other chronic pancreatitis cases.

Correlation with clinicopathological factors and survival analysis

The expression statuses of proCOL11A1, FAP, SPARC, and POSTN in the tumoral stroma did not correlate with any of the clinicopathological factors studied. However, some differences in survival were seen according to their expression status (Figure 3). On univariate analysis, low FAP-positive intratumoral CAF counts (<100/HPF) were associated with a significantly reduced overall survival compared to those with high FAP-positive CAF counts (p = 0.010; hazard ratio (HR) = 5.2 (95% confidence interval (CI) = 1.3–21.3); median survival 23.7 months in FAP-low group, not reached in FAP-high group). A similar trend was seen for disease-free survival although not statistically significant (p = 0.185; median survival 10.9 months in FAP-low group, 38.1 months in FAP-high group). This pattern was also seen for proCOL11A1 and SPARC. Although not statistically significant, the absence of proCOL11A1-positive CAFs was associated with a decreased overall survival (p = 0.175; HR = 1.7 (95% CI = 0.8–3.7); median survival 13.1 months in proCOL11A1-negative group, 30.6 months in proCOL11A1-positive group) and disease-free survival (p = 0.200; HR = 2.9 (95% CI = 1.3–6.2); median survival 3.8 months in proCOL11A1-negative group, 11.6 months in proCOL11A1-positive group). SPARC-negative tumors showed decreased overall survival (p = 0.143; median survival 20.6 months in SPARC-negative group, 30.7 months in SPARC-positive group), but disease-free survival was similar between the two groups.

Figure 3.

Kaplan–Meier survival curves demonstrating differences in overall and disease-free survival according to the expression status of (a and b) FAP, (c and d) proCOL11A1, (e and f) SPARC, and (g and h) POSTN in PDACs.

The opposite findings were seen for POSTN. POSTN-positive PDACs were associated with reduced overall survival (p = 0.154; median survival = 102.1 months in POSTN-negative group, 24.2 months in POSTN-positive group) and disease-free survival (p = 0.079; median survival not reached in POSTN-negative group, 11.1 months in the POSTN-positive group), although statistical significance was not reached, possibly due to the small number of POSTN-negative PDACs.

Correlation between epithelial–mesenchymal transition–related marker and CAF marker expression status

In order to investigate whether there were any correlations between the expression status of epithelial–mesenchymal transition (EMT)-related markers and CAF-related markers, immunohistochemical stains were performed for S100A4 and uPAR. S100A4 and uPAR positivity was seen in 113/155 (72.9%) and 50/148 (32.1%) PDACs, respectively. A positive correlation was seen between proCOL11A1 and S100A4-positivity: S100A4-positive PDACs showed significantly more abundant proCOL11A1-positive CAFs in the tumor stroma compared to S100A4-negative tumors (58.0 ± 147.3/HPF in S100A4-positive tumors vs 22.2 ± 32.4/HPF in S100A4-negative tumors; p = 0.016, t-test). No significant correlations were seen for the other CAF- or EMT-related markers.

Discussion

Due to the limited benefit of chemotherapy targeting the cancer cells in PDAC, there has been increasing interest in the tumoral stroma as a potential therapeutic target, especially as PDAC is characterized by an extensive desmoplastic reaction with a relatively sparse tumor cellularity.¹⁰ The desmoplastic stromal reaction contains CAFs—a generic term for all fibroblasts associated with tumor—and these CAFs comprise a functionally diverse population, contributing to the complexity of the tumor microenvironment. CAFs express markers, such as FAP, that are not expressed in resting fibroblasts or those associated with wound healing.³ As there is a long list of such markers, we selected four proteins—proCOL11A, FAP, POSTN, and SPARC—for which there are commercially available antibodies and which have also been described to be overexpressed in the “activated stroma” subtype of PDACs according to the recent study by Moffitt et al.⁶

FAP is a 95-kDa cell surface glycoprotein with collagenolytic activity and currently considered as the main marker for CAFs in various solid malignancies, including cancers of the colorectum, breast, lung, and pancreas.¹¹ The presence of intratumoral FAP-positive fibroblasts has been associated with a poor prognosis in PDAC,¹² and it has been suggested that FAP may modify the extracellular matrix, providing a more permissive environment for tumor cell invasion and metastasis.^13,14

SPARC (osteonectin) is another glycoprotein that is transiently secreted into the extracellular matrix and becomes rapidly degraded by several proteases.¹⁵ Recently, SPARC has received a lot of attention in the treatment of PDAC: SPARC, which has a high affinity for albumin, binds to albumin in nab-paclitaxel, increasing the delivery of paclitaxel to the tumor.¹⁶ SPARC is expressed in CAFs in the stroma immediately surrounding the tumor epithelial cells and also in some other stromal components such as endothelial cells and infiltrating inflammatory cells, and SPARC expression has also been reported in the tumor cells. In this study, we found strong SPARC expression in the intratumoral CAFs and endothelial cells and focal weak SPARC labeling in the tumor epithelial cells. SPARC expression was more frequently present in the tumor, compared to the non-neoplastic pancreatic parenchyma as expected. SPARC expression in PDACs and most other solid tumors have been associated with poor prognosis.^8,15,17

POSTN, a disulfide-linked 90-kDa secretory protein, is a regulator of bone formation and maintenance. It is produced in part by activated fibroblasts³ and has been shown to play roles in facilitating tumor progression and metastasis in various human cancers.^18,19 In the pancreas, POSTN has been shown to be expressed in high levels in high-grade pancreatic intraepithelial neoplasia (PanIN) and PDACs, while early-stage PanINs were negative for POSTN.²⁰

proCOL11A1 has recently been demonstrated to be a promising new marker for CAFs in cancers of breast,²¹ colorectum,²² ovaries,²³ and also in PDACs.⁷ Garcia-Pravia et al.⁷ demonstrated that proCOL11A1 labeled CAFs of PDAC but not the stromal cells of benign inflammatory lesions such as chronic pancreatitis, suggesting that proCOL11A1 is a specific marker for CAFs in PDAC and may even aid in the distinction between PDAC and chronic pancreatitis in diagnostically challenging cases. Therefore, we had initially aimed to validate the utility of the novel proCOL11A1 antibody in the identification of CAFs in pancreatic cancer in this study. However, although there were significant differences in the amount of proCOL11A1-positive fibroblasts between tumor tissue and non-neoplastic pancreata, proCOL11A1 expression was identified in non-neoplastic pancreata and also in the normal/benign cohort. As the previous immunohistochemical studies have mostly been reported from a single group in Spain, the specificity of proCOL11A1 as a CAF marker needs to be further validated in independent cohorts.

In this study, we examined the expression status of proCOL11A1, SPARC, FAP, and POSTN in PDAC and non-neoplastic pancreatic parenchyma and found, as expected, significantly increased expression of these four proteins in the tumoral stroma of PDACs compared to paired non-neoplastic pancreata. This supports the existing literature on CAFs. In addition, there were no significant correlations between the expression statuses of the different CAF-associated proteins in this study, supporting the notion that CAFs are a heterogeneous population of activated fibroblasts with distinct functions.³ However, our data also demonstrate that the presence of CAFs in the tumor stroma of PDAC is not associated with a poor prognosis as previously suggested in many studies; on the contrary, it may even be associated with prolonged survival. This is an interesting finding and merits further discussion and future validation in larger and independent cohorts.

The notion that tumor stroma may in fact restrain tumor growth rather than support its aggressive behavior has been demonstrated in recent experimental models.^24–26 Ozdemir et al.²⁴ demonstrated that depletion of alpha-smooth muscle actin-positive myofibroblasts accelerated PDAC growth, and Rhim et al.²⁵ showed that reducing the amount of desmoplastic stroma in PDAC by deleting sonic hedgehog (shh) resulted in more aggressive behavior, undifferentiated histology, and increased angiogenesis. This may explain in part the lack of consistency in the previous literature on the clinicopathological associations between CAFs and cancer prognosis. For example, an association between FAP expression and favorable survival has been recently demonstrated in lung cancer.²⁷ While statistical significance was reached only for FAP and overall survival in our study, there were clear differences in survival as seen from the survival curves and the differences in median survival. The lack of statistical significance could be ascribed to the very small number of PDACs that do not express CAF-related markers, and additional analysis of a larger independent cohort of PDACs would be required to validate these results.

On examining the expression status of EMT-related markers, we found that high proCOL11A1⁺ CAF counts were significantly associated with increased EMT-related marker expression. Therefore, we could not provide evidence that the anti-tumor effect of proCOL11A1⁺ CAFs involves the suppression of EMT in this cohort. However, this study is limited by the fact that it is an immunohistochemical analysis on formalin-fixed paraffin-embedded tissues, and further in vivo and in vitro functional studies using fresh frozen tissues would be necessary to elucidate the mechanisms behind the possible anti-tumor role of CAFs in PDACs.

In conclusion, we demonstrate that (1) FAP, proCOL11A, SPARC, and POSTN expression is higher in the intratumoral stroma of PDAC compared to non-neoplastic pancreata; (2) there are no significant correlations between the expression statuses of these markers suggesting that CAFs comprise a heterogeneous population of activated fibroblasts; and (3) the presence of CAFs in the tumor stroma is not always associated with a poor prognosis in PDACs, supporting the recent findings that tumor stroma may restrain tumor progression rather than enhance aggressive behavior.

Footnotes

Acknowledgements

H.P., Y.L., H.L., and H.K. performed the research; H.P. and H.K. designed the research study; J.W.K., J.H.H., J.K., Y.S.Y., and H.S.H. provided the study materials; H.P., Y.L., J.W.K., and H.K. analyzed the data; H.P. and H.K. wrote the paper; and H.K. finally approved the manuscript.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Ethical approval

This study was approved by the Institutional Review Board of Seoul National University Bundang Hospital (IRB no. B-1702-383-302).

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was supported by the SNUBH Research Fund (grant no. 02-2015-005) and the Basic Science Research Program through NRF funded by the Ministry of Education (NRF-2013R1A1A2062320 and NRF-2016R1D1A1A09919042).

References

Quail

Joyce

. Microenvironmental regulation of tumor progression and metastasis. Nat Med 2013; 19: 1423–1437.

Feig

Gopinathan

Neesse

. The pancreas cancer microenvironment. Clin Cancer Res 2012; 18: 4266–4276.

Kalluri

. The biology and function of fibroblasts in cancer. Nat Rev Cancer 2016; 16: 582–598.

Siegel

Naishadham

Jemal

. Cancer statistics, 2013. CA Cancer J Clin 2013; 63: 11–30.

Hidalgo

. Pancreatic cancer. N Engl J Med 2010; 362: 1605–1617.

Moffitt

Marayati

Flate

. Virtual microdissection identifies distinct tumor- and stroma-specific subtypes of pancreatic ductal adenocarcinoma. Nat Genet 2015; 47: 1168–1178.

Garcia-Pravia

Galvan

Gutierrez-Corral

. Overexpression of COL11A1 by cancer-associated fibroblasts: clinical relevance of a stromal marker in pancreatic cancer. PLoS One 2013; 8: e78327.

Infante

Matsubayashi

Sato

. Peritumoral fibroblast SPARC expression and patient outcome with resectable pancreatic adenocarcinoma. J Clin Oncol 2007; 25: 319–325.

Bae

Wen

. Overexpression of POSTN in tumor stroma is a poor prognostic indicator of colorectal cancer. J Pathol Transl Med 2017; 51: 306–313.

10.

Lunardi

Muschel

Brunner

. The stromal compartments in pancreatic cancer: are there any therapeutic targets? Cancer Lett 2014; 343: 147–155.

11.

Park

Lenter

Zimmermann

. Fibroblast activation protein, a dual specificity serine protease expressed in reactive human tumor stromal fibroblasts. J Biol Chem 1999; 274: 36505–36512.

12.

Shi

Chen

. Expression of fibroblast activation protein in human pancreatic adenocarcinoma and its clinicopathological significance. World J Gastroenterol 2012; 18: 840–846.

13.

Cohen

Alpaugh

Palazzo

. Fibroblast activation protein and its relationship to clinical outcome in pancreatic adenocarcinoma. Pancreas 2008; 37: 154–158.

14.

Lee

Mullins

Franco-Barraza

. FAP-overexpressing fibroblasts produce an extracellular matrix that enhances invasive velocity and directionality of pancreatic cancer cells. BMC Cancer 2011; 11: 245.

15.

Vaz

Ansari

Sasor

. SPARC: a potential prognostic and therapeutic target in pancreatic cancer. Pancreas 2015; 44: 1024–1035.

16.

Garber

. Stromal depletion goes on trial in pancreatic cancer. J Natl Cancer Inst 2010; 102: 448–450.

17.

Mantoni

Schendel

Rodel

. Stromal SPARC expression and patient survival after chemoradiation for non-resectable pancreatic adenocarcinoma. Cancer Biol Ther 2008; 7: 1806–1815.

18.

Kyutoku

Taniyama

Katsuragi

. Role of periostin in cancer progression and metastasis: inhibition of breast cancer progression and metastasis by anti-periostin antibody in a murine model. Int J Mol Med 2011; 28: 181–186.

19.

Erkan

Kleeff

Gorbachevski

. Periostin creates a tumor-supportive microenvironment in the pancreas by sustaining fibrogenic stellate cell activity. Gastroenterology 2007; 132: 1447–1464.

20.

Shi

Washington

Chaturvedi

. Fibrogenesis in pancreatic cancer is a dynamic process regulated by macrophage-stellate cell interaction. Lab Invest 2014; 94: 409–421.

21.

Freire

Dominguez-Hormaetxe

Pereda

. Collagen, type XI, alpha 1: an accurate marker for differential diagnosis of breast carcinoma invasiveness in core needle biopsies. Pathol Res Pract 2014; 210: 879–884.

22.

Galvan

Garcia-Martinez

Vazquez-Villa

. Validation of COL11A1/procollagen 11A1 expression in TGF-beta1-activated immortalised human mesenchymal cells and in stromal cells of human colon adenocarcinoma. BMC Cancer 2014; 14: 867.

23.

Cheon

Tong

Sim

. A collagen-remodeling gene signature regulated by TGF-beta signaling is associated with metastasis and poor survival in serous ovarian cancer. Clin Cancer Res 2014; 20: 711–723.

24.

Ozdemir

Pentcheva-Hoang

Carstens

. Depletion of carcinoma-associated fibroblasts and fibrosis induces immunosuppression and accelerates pancreas cancer with reduced survival. Cancer Cell 2014; 25: 719–734.

25.

Rhim

Oberstein

Thomas

. Stromal elements act to restrain, rather than support, pancreatic ductal adenocarcinoma. Cancer Cell 2014; 25: 735–747.

26.

Gore

Korc

. Pancreatic cancer stroma: friend or foe? Cancer Cell 2014; 25: 711–712.

27.

Kilvaer

Khanehkenari

Hellevik

. Cancer associated fibroblasts in stage I-IIIA NSCLC: prognostic Impact and their correlations with tumor molecular markers. PLoS One 2015; 10: e0134965.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.21 MB

The prognostic significance of cancer-associated fibroblasts in pancreatic ductal adenocarcinoma