Advertisement
Molecular and Cellular Pathobiology

Twist1 Is a Key Regulator of Cancer-Associated Fibroblasts

Keun-Woo Lee, So-Young Yeo, Chang Ohk Sung and Seok-Hyung Kim
DOI: 10.1158/0008-5472.CAN-14-0350 Published January 2015

Abstract

Cancer-associated fibroblasts (CAF) are key contributors to malignant progression, but their critical regulators remain largely unknown. In this study, we examined the role of Twist1, a central regulator of epithelial–mesenchymal transition in carcinoma cells, in the transdifferentiation of normal quiescent fibroblasts to CAF and we defined its upstream controls and downstream effectors. Primary human gastric fibroblast and CAF cultures were established from gastrectomy specimens and validated as nontumor cells by somatic mutation analyses. In these cultures, exposure to the proinflammatory cytokine IL6 commonly expressed in tumors was sufficient to induce Twist1 expression in normal fibroblasts and transdifferentiate them into CAFs via STAT3 phosphorylation. In xenograft models, tumor infiltration of Twist1-expressing CAFs was enhanced strongly by ectopic IL6 expression in gastric or breast cancer cells. We found that Twist1 expression was necessary and sufficient for CAF transdifferentiation. Enforced expression of Twist1 in normal fibroblasts was also sufficient to drive CAF marker expression and malignant character in gastric cancer cells both in vitro and in vivo. Conversely, silencing the expression of Twist1 in CAFs abrogated their tumor-promoting properties. Downstream of Twist1, we defined the chemokine CXCL12 as a transcriptional target. Clinically, CXCL12 and Twist1 expression were correlated in CAFs present in gastric tumor specimens. Finally, ectopic expression of Twist1 in normal fibroblasts suppressed premature senescence, whereas Twist1 attenuation accelerated senescence in CAFs. Our findings define Twist1 as a compelling target to deprogram the tumor-supporting features of the cancer microenvironment. Cancer Res; 75(1); 73–85. ©2014 AACR.

Introduction

Tumor growth and progression are markedly influenced by the tumor microenvironment, which comprises the extracellular matrix (ECM) and a variety of stromal cells, such as fibroblast, monocytes, and endothelial cells (1, 2). Fibroblasts are the most common cells in connective tissue, and their main function is to maintain the ECM and regulate epithelial differentiation by stromal–epithelial crosstalk (3, 4). Fibroblasts display a wide spectrum of activation states: normal fibroblasts show thin, wavy, small spindle-cell morphology, whereas activated fibroblasts are large, plump, spindle-shaped mesenchymal cells with stress fibers (1, 3). Among activated fibroblasts, those found in association with cancer cells are known as cancer-associated fibroblasts (CAF; ref. 3).

CAFs promote cancer progression through remodeling of the ECM, induction of angiogenesis, and recruitment of inflammatory cells and directly stimulate cancer cell proliferation via the secretion of growth factors and mesenchymal–epithelial cell interactions (5). Although the biological and clinical significance of CAFs in tumor microenvironment has been widely recognized, the mechanisms by which normal fibroblasts are transformed into CAFs remain vastly unknown.

In our previous study, we showed that Twist1 is expressed in stromal fibroblasts in gastric cancer tissues and that Twist1-expressing fibroblasts possess CAF-like properties (6). Twist1 is a basic helix–loop–helix transcription factor that is essential for the development of mesodermally derived tissues (7, 8). During mouse embryogenesis, Twist1 is expressed in various mesodermal tissues including neural crest-derived mesenchyme, branchial arch, muscle, bone, and fibroblasts (9, 10). After birth, Twist1 expression is barely detectable in normal mesenchymal cells of adult tissues and limited to mesenchymal stem cells (11, 12) and human white adipocytes (13). On the other hand, Twist1 is known to be an important inducer of epithelial–mesenchymal transition (EMT) and Twist1 overexpression has been reported in a variety of epithelial cancer cells with clinical correlation with poor prognosis (14–16). However, despite the abundance of Twist 1-expressing stromal fibroblasts within cancer tissue, its clinical significance and regulating mechanism remain almost completely unknown.

In the current study, we investigated molecular mechanism by which Twist1 is induced in CAFs and the role of Twist1 in driving the transdifferentiation of quiescent fibroblasts into CAFs. In addition, we showed that Twist1 expression is necessary and sufficient for CAF transdifferentiation.

Materials and Methods

Cell lines

The gastric normal fibroblast cell line Hs738 was obtained from the ATCC. Hs738 and the stomach cancer cell lines MKN1, MKN28, MKN74, NCI-N87, SNU638, and SNU668, and the breast cancer cell line MCF7 were used. Cell lines were purchased from the Korean Cell Line Bank that performs cell line characterizations using DNA fingerprinting analysis and passaged in our laboratory for fewer than 6 months. Maintenance procedures are described in Supplementary Materials and Methods.

Isolation and culture of fibroblasts

Human stomach tumor specimens were obtained from patients undergoing surgery at Samsung Medical Center of SungKyunKwan University of Medicine (Seoul, Korea). An experienced pathologist grossly examined and obtained representative samples of the tumor tissues (CAF, human gastric CAF) and distal normal tissues (NF, human gastric normal fibroblast). Detailed procedures are described in Supplementary Materials and Methods.

Incubation of fibroblasts with conditioned media from gastric cancer cells

Gastric cancer cells were seeded in 10 cm dishes, and after overnight attachment and growth, cells were washed twice with PBS and grown in serum-free D/F12 media. Conditioned medium (CM) was collected after 24 hours (cells reached 70%–80% confluence) and centrifuged at 1,300 rpm for 5 minutes to remove pellet debris. Cells were incubated with CM for 2 days (48 hours).Conditioned media were freshly changed each day. NF cells were treated with recombinant IL6 (Cat#CTP0063, Invitrogen, Life Technologies) for 2 days (48 hours). For inhibition of STAT3 activity, cells were treated with STAT3 inhibitor (cucurbitacin I, Cucumi's sativus L., Cat#238590; Calbiochem) for 24 hours.

Tissue specimens and construction of tissue microarrays

A total of 332 formalin-fixed, paraffin-embedded gastric tissue samples were included in this study. All patients provided written informed consent according to institutional guidelines. Detailed procedures are described in Supplementary Materials and Methods.

Immunohistochemical analysis

Detailed methods and list of antibodies used are described in Supplementary Materials and Methods.

Senescence-associated β-galactosidase activity assay

Detailed procedures are described in Supplementary Materials and Methods.

ELISA for CXCL12/SDF1a

The concentration of CXCL12/SDF1a was estimated in CM from NF#14 (GFP/Twist1), NF#32 (GFP/Twist1), CaF#14 (ShNS/Shtwist1), and CaF#32 (ShNS/Shtwist1). Detailed procedures are described in Supplementary Materials and Methods.

Cell viability/proliferation assay

Stomach cancer cells were cultured with CM from normal and CAFs for 2 days. Cells were detached, seeded in 96 well plates, and cultured for 24 hours. Detailed procedures are described in Supplementary Materials and Methods.

Xenograft model

Stomach cancer cells subcutaneously implanted in BALB/c-nude mice as the tumor model. To examine the effect of Twist1-positive fibroblasts on xenograft tumor growth, 5 × 105 stomach cancer cells were coinjected with 1.5 × 106 Twist1-positive fibroblasts [Hs738(Twist1)]. As controls, the same number of cancer cells were injected alone or with Twist1-negative fibroblast [Hs738(GFP)]. To investigate the effect of IL6 secreted by cancer cells on surrounding CAFs in vivo, SNU638+GFP and MCF7+GFP cells were infected with IL6-expressing lentivirus. Tumor cells (3 × 106; SNU638 and 2 × 106; MCF7) were implanted subcutaneously in the right flank in 0.1 mL of serum-free medium and growth factor–reduced Matrigel (1:1). Tumor volume was determined every 2 weeks.

Lentivirus transduction

Detailed procedures are described in Supplementary Materials and Methods.

RNA extraction and quantitative real-time reverse transcription PCR

Total RNA was extracted using the RNeasy Kit (#74104, Qiagen). The cDNA was synthesized from 1 μg of total RNA using High Capacity cDNA Reverse Transcription Kits (#4368813, Applied Biosystems), according to the manufacturer's protocol. Real-time reverse transcription (RT)-PCR was performed using the ABI 7900 HT Fast Real-Time PCR System (Applied Biosystems). GAPDH was used as an internal loading control. PCR was performed using SYBR Green PCR Master Mix (Applied Biosystems). Thermal cycling conditions were as follows: 10 minutes at 95°C followed by 40 cycles at 95°C for 15 seconds and 60°C for 1 minute. Each measurement was performed in duplicate. PCR product quality was monitored using post-PCR melt-curve analysis. The cycle threshold (Ct), the fractional cycle number at which the amount of amplified target reaches a fixed threshold, was determined. Primer sequences are described in Supplementary Materials and Methods.

SDS-PAGE

Detailed procedures are described in Supplementary Materials and Methods.

Invasion/migration assay

Detailed procedures are described in Supplementary Materials and Methods.

Chromatin immunoprecipitation assay

CAF#14 cells were grown to 70% to 80% confluency and fixed with 1% formaldehyde. Nuclear extracts were sonicated to shear the DNA into fragments smaller than 500 bp. Twist1–DNA complexes were immunoprecipitated using Protein G Dynabeads (Cat. 10003D, Invitrogen) conjugated with anti-Tiwst1 IgG or control mouse IgG. After the precipitated DNA was eluted, crosslinks were reversed by 0.3 mol/L NaCl and purified with the QIAquick PCR Purification Kit (28104, Qiagen). PCR was performed on the purified DNA using primer sets as described in Supplementary Materials and Methods.

Luciferase assay

The pGL3.CXCL12 promoter (−1,442 to +28) luciferase vector was generated by inserting cxcl12 promoter sequence. The 293T cells (5 × 104 cells/well) were plated into 24-well plates and incubated for 24 hours. Promoter vectors were cotransfected with either pSG5.htwist1.HA or pSG5.HA vector into cells using Lipofectamine 2000 (Invitrogen). Luciferase reporter gene assays were performed using the Luciferase Assay System (Promega). Sequences were available in Supplementary Materials and Methods.

Site-directed mutagenesis

The Twist1-binding E-BOX-1 region (CAGGTG) and nonbinding E-BOX-2 region (CATCTG) on the CXCL12 promoter were mutated using the pGL3.CXCL12 (−1,442 to +28) luciferase vector as the template DNA. The PCR amplification for site-directed mutagenesis was performed using the corresponding mutant primers. The sequence changes for the E-BOX-1 (−514 bp to −509 bp) htwist1 binding site were as follows: WT, GAGGTG and mutant E-BOX-1, TTGGTG. The sequence changes for the E-BOX-2 (−1,042 to −1,037) were as follows: WT, CATCTG; and mutant E-BOX-2, TTTCTG. Sequences were available in Supplementary Materials and Methods.

Immunofluorescence assay

Paraffin sections from MKN74, SNU638, and MCF7 xenograft models were used and detailed procedures are described in Supplementary Materials and Methods.

The Cancer Genome Atlas data

mRNA expression data were obtained for 9 tumors of The Cancer Genome Atlas (TCGA) carcinoma project through the cBio cancer genomics portal (http://www.cbioportal.org/public-portal).

Gene expression microarray

Data have been deposited in Gene Expression Omnibus (accession number GSE62740). Expression profiling analysis was performed using the Agilent Oligo Microarray Kit 6.0K accordingly. The detailed procedures are described in Supplementary Materials and Methods. All experiments were performed in duplicate.

High-throughput sequencing

Genomic DNA was isolated from two patients for each cancer cell, matched CAFs, and matched normal fibroblasts. Library construction and targeted exome capture (505 genes) using the Agilent SureSelect Target Enrichment System were performed. Sequencing of exome-enriched libraries was performed using Illumina's MiSeq System. Average target coverage was 138.3 for the six tested samples and more than 90% of target bases achieving 30 times coverage were obtained in all six samples. Detail methods and sequencing metrics are available in Supplementary Materials and Methods.

Statistical analysis

Correlations were examined using Pearson χ2, Fisher exact test, Mann–Whitney U test, or Spearman test, as appropriate. Overall survival (OS) was determined using the Kaplan–Meier method and survival curves were compared using the log-rank test. Statistical analysis was performed using Stata/IC statistical software (version 12; StataCorp) or the R program.

Results

Twist1 upregulation by the IL6/STAT3 axis in quiescent fibroblasts with the expression of CAF markers

In a previous study, we showed that Twist1 is expressed in CAFs derived from gastric cancer tissues with significant associations with CAF markers expression and poor clinical outcome (6). Despite these results suggesting the importance of Twist1 in CAFs, the mechanism by which Twist1 is activated in fibroblasts remains unknown. To address this issue, two primary normal fibroblast cultures (NF#19 and NF#32) were established from gastric nontumor tissues. When we cultured these normal fibroblasts with CM from four gastric cancer cell lines (MKN28, 74, NCI-N87, and SNU668), only CM from SNU668 significantly upregulated mRNA levels of CAF markers, including FAPα, FSP1, PDGFRα, and Twist1 (Fig. 1A and B). Next, we searched for cytokines selectively secreted by SNU668 because cytokines are important soluble factors in CM for CAF activation (17). The mRNA levels of cytokine genes obtained from the BioCarta data set for gastric cancer cell lines were validated using gene expression data from TCGA project. The results showed that IL6 mRNA levels were distinctly high in SNU668 cells compared with those of other cell lines (Fig. 1C). IL6 protein levels were also markedly increased in the CM of SNU668 (Fig. 1D), suggesting that IL6 is a major inducer of CAF transdifferentiation. To test this hypothesis, the effect of IL6 blocking on CAF transdifferentiation was evaluated by adding IL6-blocking antibody to CM from SNU668 cells, which abolished the enhanced expression of Twist1 and CAF markers induced by CM of SNU668 (Fig. 1E). Furthermore, acquired CAF phenotype in fibroblast by IL6 was maintained even after removal of IL6 (Supplementary Fig. S1).

Figure 1.
Figure 1.

Twist1 is required for IL6-induced differentiation of quiescent fibroblasts into CAFs. A, experimental design for incubation of NFs with CM from four gastric cancer cells. B, the mRNA of Twist1 and CAF markers (FAPa, FSP1, PDGFRa/b, and CXCL12) in gastric NFs cultured with CM derived from four gastric cancer cell lines. C, the mRNA levels of cytokines from TCGA data in gastric cancer cell lines. D, Western blot analysis of IL6 in the CM from gastric cancer cells. E, the effect of IL6 blocking on mRNA of CAF markers and Twist1 in gastric NF. IL6 blocking antibody was added to CM from SNU668. F, NF#5, #19, and #32 primary gastric NFs were incubated for 48 hours with recombinant IL6 and the expression of CAF markers and Twist1 was determined. G, NF#15 was transduced with either nontargeting shRNA (shNS) or shRNA targeting Twist1 (shTw1) using lentivirus. Then it was treated with IL6. H, pSTAT3 and Twist1 were upregulated by IL6 in gastric NFs (left). Twist1 upregulation by IL6 was reduced by a STAT3 inhibitor (JSI-124) in a dose-dependent manner (right). *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Next, gastric normal fibroblasts were treated with IL6 at 50 ng/mL for 2 days and its impact on the expression of CAF markers was assessed by Western blot analyses. Twist1 and CAF markers such as FAPα, PDGFRα, PDGFRβ, α-SMA, and FSP1 were significantly upregulated by IL6 in normal fibroblasts (Fig. 1F and Supplementary Fig. S2A). This IL6-induced upregulation of CAF markers was repressed by silencing of Twist1 expression using shRNA (Fig. 1G and Supplementary Fig. S2B). Because IL6 signaling is mediated by STAT3 phosphorylation in other types of cells (18), we examined whether STAT3 is phosphorylated in gastric fibroblasts in response to IL6 signaling. The results was that IL6 increased pSTAT3 levels as well as Twist1 in two gastric fibroblast lines in a dose-dependent manner, and IL6-induced upregulation of Twist1 was abrogated by a STAT3 inhibitor (JSI-124; Fig. 1H and Supplementary Fig. S2C). These results indicated that Twist1 expression is activated in fibroblast by the IL6/STAT3 axis.

The effect of IL6 on generation of Twist1-expressing CAFs was further investigated in in vivo models. Ectopic IL6-expressing SNU638-GFP cancer cells (SNU638GFP/IL6) and MCF7-GFP cells (MCF7GFP/IL6), which were originally negative for IL6 (Fig. 2A and B) were generated and injected subcutaneously into immunocompromised nude mice (Fig. 2C). Tumorigenicity of ectopic IL6-expressing SNU638 and MCF7 was remarkably superior to respective controls (Fig. 2C). Furthermore, Twist1-expressing stromal cells were much more frequent in xenograft tumors formed by these IL6-expressing SNU638GFP/IL6 cells and MCF7GFP/IL6 cells compared with controls (P < 0.001; Fig. 2D and E). These Twist1-expressing stromal cells were negative for GFP but positive for Tenascin-C (Tn-C), indicating that these cells were CAFs not originated from cancer cells (Fig. 2D). This experiment was repeated using another gastric cancer cell, MKN74, and identical result was obtained (Supplementary Fig. S3).

Figure 2.
Figure 2.

The IL6 cytokine induces the generation of Twist1-expressing CAFs in vivo. A, gastric cancer cell SNU638 and breast cancer cell MCF7 were transduced with GFP or GFP/IL6 genes by lentivirus. B, IL6 expression was confirmed by Western blot analysis. C, xenograft of ectopic IL6-expressing cells (SNU638GFP/IL6, MCF7GFP/IL6) and respective controls (SNU638GFP and MCF7GFP) to BALB/c nude mice. D, immunofluorescence staining of Twist1 and Tenascin-c (Tn-c) in xenograft tumors. Nuclei were visualized by staining with DAPI (blue). Scale bars, 50 μm (magnification, ×400). Scatter plot (right) shows a strong correlation between Twist1 (red) and Tn-c (green) expression in CAFs. E, immunohistochemical analysis of Twist1 in xenograft tumors. F, immunohistochemical analysis of IL6, pSTAT3, and Twist1 in gastric cancer patients' tissues.

The expressions of IL6, pSTAT3, and Twist1 in gastric cancer tissues were studied using immunohistochemistry and IL6 expression was prevalent in the cancer stroma in all cases (n = 46). pSTAT3 expression was detected in both cancer cells and stromal fibroblasts, whereas Twist1 expression was mainly limited to CAFs. There was a significant correlation between pSTAT3 and Twist1 expression in CAFs (P < 0.001; Fig. 2F).

Twist1 expression is sufficient for transdifferentiation of quiescent fibroblasts to CAFs

To determine whether Twist1 alone confers normal fibroblast with CAF-like properties, gastric normal fibroblasts (Hs738, NF#19, and NF#32) were induced to express Twist1 using lentivirus, and we found that the protein levels of CAF markers, such as FAPα, PDGFRα, PDGFRβ, α-SMA, and FSP1, were enhanced by Twist1 in these normal fibroblasts (Fig. 3A and Supplementary Fig. S2D). Next, we collected CM from control fibroblasts (NF#14GFP and NF#32GFP), ectopic Twist1-expressing fibroblasts (NF#14Twist1, NF#32Twist1), and patient matched CAFs (CAF#14, #32), then cultured gastric cancer cells with these CM (Fig. 3B). CM from Twist1-expressing NF#14Twist1 and NF#32Twist1 significantly enhanced the proliferation of MKN28 and MKN74 cells to the same extent as CM from CAF#14, 32 did (Fig. 3B and Supplementary Fig. S4), and remarkably increased the migratory and invasive abilities of gastric cancer cells (Fig. 3C). Finally, the effect of Twist1-exprssing fibroblasts on tumorigenicity of gastric cancer cells was examined in vivo using a xenograft model. Specifically, human SNU638 and SNU668 gastric cancer cells (5 × 105 cells each) were coinjected subcutaneously into nude mice in a suspension with either Twist1-expressing fibroblast or control fibroblasts (1.5 × 106 cells each). Xenografts containing Twist1-expressing CAFs grew significantly larger and faster than both of xenografts infused with Twist1-negative fibroblasts and xenografts by cancer cells alone (no fibroblasts; P < 0.001; Fig. 3D). Microvessel density and cancer cells' proliferation rate within xenograft tumor were also significantly increased when Twist1-expressing fibroblasts were added (Fig. 3E and F). These data indicated that ectopic Twist1-expressing normal fibroblasts acquired both the functional and expressional attributes of CAFs, strongly suggesting that Twist1 is a sufficient condition for CAF transdifferentiation.

Figure 3.
Figure 3.

Twist1 is sufficient for transition of normal fibroblast to CAF. A, the expression of CAF markers in response to Twist1 overexpression in gastric NFs (Hs738, NF#19, NF#32) was evaluated by Western blot analysis. Proliferation (B) and migration/invasion (C) of stomach cancer cells, MKN28, MKN74, and SNU668 after incubation with CM derived from GFP (control)/Twist1-expressing normal fibroblasts (NF#14,32) and CAFs (CAF#14,32). D, SNU638 (left) and SNU668 (right), were coinjected with GFP or Twist1-overexpressing gastric NFs (Hs738) into the flanks of nude mice. Xenograft tumors containing Twist1-expressing Hs738 grew much faster and larger than controls did. **, P < 0.01. E, microvessel density determined by counting number of microvessel per high-power field in the section stained with an antibody to CD34. F, Proliferative indices of gastric cancer cells in xenograft tumors. The number of Ki67 positive cells per 1,000 total cells was plotted for each tumor. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Twist1 expression in CAFs is essential for the maintenance of the CAF phenotype

To investigate Twist1′s role in CAF, primary cultures of CAFs (CAF#14 and 32) were established from stomach cancer patients' tissues and both CAFs expressed Twist1. Then, both CAFs' somatic mutational profiles were compared with those of matched cancer cells using high-throughput sequencing technology to exclude contamination by cancer cells. The results was that cancer cells and CAFs had distinctive mutational patterns in both patients, with cancer cells showing frequent mutations, whereas no mutations, except for one synonymous mutation, were identified in CAFs (Fig. 4A). The detailed somatic mutation profiles of the isolated cancer cells and CAFs in the two patients were exported to an Excel table. The different somatic mutation patterns in CAFs and cancer cells suggested that these CAFs were neither contaminated cancer cells nor CAFs originated from cancer cells. Silencing Twist1 expression in both gastric CAFs, CAF#14 and CAF#32, using shRNA markedly reduced the expression of CAF markers, including FAPα, PDGFRα, PDGFRβ, α-SMA, and FSP1 (Fig. 4B and Supplementary Fig. S2E). Furthermore, the enhanced proliferation of MKN28 and MKN74 gastric cancer cells induced by CM derived from endogenously Twist1-expressing CAF#14 and CAF#32 was not observed when these cancer cells were treated with CM from Twist1-depleted CaF#14 and CaF#32 cells (Fig. 4C). The ability of CAFs to facilitate the migration and invasion of gastric cancer cells was also remarkably reduced by knockdown of Twist1 in gastric CAFs (Fig. 4D). These results indicate that Twist1 is necessary for CAF transdifferentiation and essential for the maintenance of the CAF phenotype.

Figure 4.
Figure 4.

Knockdown of Twist1-inhibited CAF features. A, somatic mutation profiles of CAFs (CAF#14, 32) and matched cancer cells using high-throughput sequencing technology. B, the reduced expression of CAF markers in response to Twist1 knockdown by shRNA. Proliferation (C) and migration/invasion (D) of stomach cancer cells, MKN28, MKN74, and SNU668 after incubation with various CM for 48 hours. CM were prepared from either Twist1-expressing or Twist1 knockdown CAFs (CAF#14, 32). E, volcano plots of mRNA expression microarray data demonstrated upregulated genes in Twist1 overexpressing normal fibroblast (NF#14, 32) and downregulated genes in Twist1 knockdown CAFs (CAF#14, 32). F, the gene list whose expression was increased in Twist1-induced fibroblasts and simultaneously reduced in Twist1 deleted CAFs (FC > 1.5, P < 0.05).

Genome-wide analyses of the effect of Twist1 on mRNA expressional changes in fibroblasts identified CAF phenotype-related Twist1 target genes

To elucidate the functional mechanism of Twist1 in CAFs, changes in mRNA expression profiles induced by Twist1 were analyzed using mRNA expression microarrays. To improve specificity of this analysis, we designed two experimentations; (i) gain-of-function effect of twist1 in normal fibroblasts (NF#14 and NF#32) and (ii) loss-of-function effect of twist1 in patient matched CAFs (CAF#14 and CAF#32). These experimentations were repeated two times. Genes upregulated in the two Twist1-induced fibroblast lines (NF#14 and #32) and those downregulated in the two Twist1-deleted CAFs (CaF#14 and #32) are shown using a volcano plot (Fig. 4E). Among approximately 60,000 probes annotating 20,000 genes, 221 genes were significantly upregulated in Twist1-induced fibroblasts and 889 genes were significantly downregulated in Twist1-deleted CAF (cut-off point: fold change > 1.5 and P < 0.05). Among them, 23 genes whose expression was increased in Twist1-induced fibroblasts and simultaneously reduced in Twist1-deleted CAFs were considered candidate target genes upregulated by Twist1. As shown in Fig. 4F, these genes include many CAF phenotype-related genes involved in remodeling of cancer microenvironment, such as SULF2, lysyl oxidase (LOX), CXCL12, Tn-C (TNC), delta-sarcoglycan (SGCD), and HAPLN1.

CXCL12, a potent regulator of the CAF phenotype, is directly upregulated by Twist1 in fibroblasts with clinical relevance

On the basis of the results of our gene-chip study and functional significance of CXCL12 as a strong tumor-promoting chemokine, we suspect that CXCL12 is a key target of Twist1 in CAFs (19, 20). First, we validated our mRNA microarray results in both mRNA and protein levels. We confirmed that the secretion of CXCL12 was markedly enhanced by Twist1 alone in normal fibroblast (NF#32) and Twist1 is also essential for CXCL12 secretion in CAF (CaF#32) using ELISA assay (Fig. 5A). Besides, Twist1 overexpression increased the mRNA level of CXCL12 in normal fibroblasts (NF#14 and NF#32), whereas knockdown of Twist1 decreased the mRNA level of CXCL12 in CAFs (CAF#14 and CAF#32; Fig. 5B). The association between Twist1 and CXCL12 was analyzed in other types of cancer using public mRNA expression data from TCGA, which showed a significant correlation between Twist1 and CXCL12 expression in various cancers, such as breast, colorectal, ovarian, endometrial, lung, thyroid, and renal cancers, and glioblastoma (Supplementary Fig. S5). These results suggest that Twist1 regulates CXCL12 expression at the transcriptional level.

Figure 5.
Figure 5.

Twist1 directly upregulated the expression of CXCL12 with clinical relevance. A, detection of secreted protein of CXCL12/SDF1a by ELISA. B, mRNA level of CXCL12 in ectopic Twist1-overexpressing NFs (NF#14 and NF#32; left) and in Twist1 knockdown patient-matched CAFs (CAF#14 and CAF#32; right). C, CXCL12 promoter constructs. D, Luciferase assay showed changes in the CXCL12 promoter activity in response to transient induction of Twist1 in 293T cells. This effect was abolished by mutation in E-box1. E, ChIP assay results with schematic illustration of PCR primer binding sites in CXCL12 promoter region. F, Twist1 and CXCL12 expression in the tumor stroma of gastric cancer tissues (adenocarcinoma). G, Overall survival rates of patients stratified by stromal CXCL12 and Twist1 expression pattern in CAFs. H, invasion/migration (left) and proliferation (right) of gastric cancer cells. These cancer cells were incubated with CM from GFP-expressing normal fibroblast (NF-GFP, control), ectopic Twist1-expressing NF (NF-Tw1), and ectopic Twist1 overexpressing and CXCL12 knockdown fibroblast (NF-Tw1+shCXCL12) *, P < 0.05; **, P < 0.01; ***, P < 0.001.

To further examine the regulation of CXCL12 by Twist1, we performed luciferase reporter assays using the pGL3.CXCL12 vector carrying the CXCL12 promoter region (Fig. 5C). The pGL3.CXCL12 and pSG5.hTwist1.HA vectors were transiently cotransfected into NIH3T3 cells. As shown in Fig. 5D, the CXCL12 promoter-reporter activity was significantly affected when Twist1 was introduced into the cells. Two E-box sequences located at 509 bp and at 1,037 bp upstream of the transcription start site were identified in the promoter region of CXCL12. To confirm the role of these E-boxes, mutations were introduced [CAGGTG (−509) to TTGGTG and CATCTG (−1,037) to TTTCTG] and the mutant vectors were transfected into NIH3T3 cells. Mutation of the proximal E-box (−509) abrogated the effect of Twist1 on CXCL12 promoter activity, whereas mutation of the distal E-box (−1,037) made no difference, indicating that the proximal E-box is crucial for Twist1 function. Next, we performed chromatin immunoprecipitation (ChIP) assays in endogenous Twist1-expressing CAFs (CaF#14). Specific primer sets for the known Twist1-binding E-box CACATG (PDGFRα; −1,839 to −1,834 bp) in the PDGFRα promoter were used as a positive control (17). Twist1 bound specifically to the positive control (PDGFRα; −1,839 to −1,834 bp) and the proximal E-box CAGGTG (Fig. 5E), whereas no amplification was detected in the distal E-box (E-box2) or in the NC site containing no E-box sequence (Fig. 5E). These results indicate that Twist1 binds directly to the endogenous CXCL12 promoter.

In gastric cancer tissues, CXCL12 is expressed in both cancer cells and stromal cells and there is a significant correlation in stromal expressions between CXCL12 and Twist1 (P < 0.001; Fig. 5F). In addition, stromal expression of CXCL12 was associated with an unfavorable clinical outcome (P = 0.0232; Fig. 5G). Patients' group with both Twist1- and CXCL12-expressing CAFs had the worst prognosis, whereas the other group in which CAFs were negative for both Twist1 and CXCL12 had the best prognosis, and the difference was statistically significant (P < 0.001; Fig. 5G). Additional results for clinical relevance of CXCL12 in gastric cancer were described in Supplementary Results and Tables S1–S4 and Supplementary Fig. S6.

Furthermore, knockdown of CXCL12 in fibroblasts (NF#14, 32) significantly impaired tumor-promoting ability conferred by Twist1 as shown Fig. 5H and Supplementary Fig. S7. CM from CXCL12-deleted fibroblasts (NF#14, 32) failed to enhance migration and invasion of gastric cancer cells even after enforced Twist1 overexpression.

Twist1 contributes to perpetual CAF activation by repressing the senescence of fibroblasts and CAFs

One of the CAF-specific properties that distinguishes them from normal activated fibroblasts such as myofibroblasts is their constitutive activation (3). In contrast to CAFs, myofibroblasts activated during conventional inflammation regress to a normal state through senescence or apoptosis after the source of inflammation is removed (3). Because Twist1 suppresses cellular senescence in various types of epithelial cancer cells (21), we hypothesized that Twist1 induced the constitutive activation of CAFs through repression of senescence. To validate this hypothesis, the effect of Twist1 on cellular senescence was examined in both normal fibroblasts and CAFs. Twist1 overexpression significantly reduced SA staining and downregulated the expressions of CDKN2B(p15) and CDKN2A(p16), representative inducers of senescence, in gastric normal fibroblasts (NF#5 and NF#15; Fig. 6A). On the other hand, silencing Twist1 expression by shRNA accelerated senescence in CAFs, as detected by increased SA staining in CAF#13 and CAF#36 (Fig. 6B), and upregulated p15 and p16 expressions in all three gastric CAFs (CAF#13, CAF#14, and CAF#32). Our results strongly suggest that Twist1 contributes to the generation of CAFs by repressing cellular senescence. Together, our findings support a model in which Twist1 is a key regulator of CAFs (Fig. 6C).

Figure 6.
Figure 6.

Twist1 repressed cellular senescence of CAFs. A, β-galactosidase staining showed that the number of senescent cells (arrowhead) was decreased in Twist1-overexpressing NFs (#5, #15, and #32; bottom) compared with the GFP-expressing controls (top;magnification, ×100). The mRNA level of the senescence markers, p15 and p16, assayed by qRT-PCR. B, the number of β-galactosidase–stained cells was increased by silencing Twist1 expression (bottom) compared with nontargeting shRNA controls (top). The mRNA levels of p15 and p16 were elevated in shTwist1 cells compared with ShNS controls. C, schematic illustration showing the role of the IL6/pSTAT3/Twist1 axis in the CAFs. *, P < 0.05; **, P < 0.01.

Discussion

In our previous study, we showed that Twist1 is frequently overexpressed in stromal fibroblasts surrounding gastric cancer cells (6). In additional, these Twist1-expressing stromal fibroblasts expressed CAF markers such as FSP1 and PDGFRα with association with poor prognosis (6). Despite these results, there was no direct evidence for functional role of Twist1 as a key regulator of CAFs. Therefore, in the current study, we established fibroblasts and CAFs models from gastrectomy specimens and confirmed that Twist1 expression is necessary and sufficient for CAF transdifferentiation. In addition, IL6/STAT3 axis was discovered to be a key upstream control of Twist1. Microarray analysis of the effect of Twist1 on mRNA expression in fibroblasts identified CXCL12 as a key Twist1′s target in CAFs. Moreover, Twist1 was revealed to suppress cellular senescence of normal fibroblasts and CAFs.

As Yang and colleagues first reported that Twist1 is a critical player in cancer metastasis (16), most studies focused on Twist1′s role in cancer cells (22–24), with little attention paid to its possible role in the tumor microenvironment, including CAFs. In contrast to rare expression of Twist1 in adult normal fibroblasts, Twist1 is frequently and strongly expressed in embryonic fibroblasts (9, 11, 25) and pathologic fibroblasts in idiopathic pulmonary fibrosis (26, 27) and desmoid tumor (28). Bridges and colleagues (26) reported that Twist1 enhanced survival and accumulation of fibroblasts in fibrotic lung disease, which is consistent with results of our current study. Bacac and colleagues (28) reported that Twist1 is highly expressed in desmoid tumors (neoplastic myofibroblastic lesions) compared with nodular fasciitis (nontumorous proliferation of fibroblasts). Furthermore, Spaeth and colleagues reported that CD44-induced acquisition of CAF phenotype in mesenchymal stem cell is mediated by Twist activation. This result is highly consistent with our result and also highlights the importance of Twist1 in CAF (29).

The role of Twist1 as a major EMT inducer implicates that certain Twist1-expressing CAFs may be derived from malignant epithelial cells. However, EMT of cancer cells probably accounts for only a fraction of CAFs (3, 30). In this current study, CAFs isolated from tumors showed no mutation in contrast to their matched cancer cells displaying more than 9 mutations (Fig. 4A). Despite a few previous studies claiming that CAFs are genetically abnormal, recent evidences showed CAFs to be genetically normal (31, 32). No somatic mutations were detected in CAF cultures generated from primary pancreatic adenocarcinomas (32, 33). In addition, we also found that Twist1-expressing stromal cells in xenograft formed by ectopic GFP/IL6-expressing cancer cells were negative for GFP (Fig. 2). Taken together, our results indicate that Twist1-expressing CAFs are more likely to be derived from noncancer cells.

Recent studies showed that IL6 facilitates CAF transdifferentiation in several types of cancers, including prostate cancer (34–37). The enhancement of actomyosin contractility by the IL6/STAT3 pathway was recently suggested as a mechanism of IL6-induced CAF transdifferentiation (35). However, transcription factors mediating this IL6-induced CAF transdifferentiation remain unknown. Meanwhile, Twist1 was suggested as a downstream transcription factor of the IL6/STAT3 pathway in some types of epithelial cancer cells (38, 39). Twist1 is transcriptionally induced by STAT3 in breast epithelial cancer cells (38) and IL6 stabilizes Twist1 and enhances cancer cell motility in head and neck epithelial cancer cells (39). These results raise the possibility that Twist1 is also a downstream effecter of the IL6/STAT3 pathway in fibroblasts and CAFs. Indeed, Twist1 was revealed to be induced by the IL6/STAT3 axis in gastric normal fibroblasts and was indispensable for IL6-induced CAF transdifferentiation (Fig. 1). Furthermore, IL6 was abundant in the cancer stroma and a significant positive correlation between pSTAT3 and Twist1 was observed in cancer tissues (Fig. 2F).

In the current study, we searched for Twist1 target genes in CAFs using mRNA microarray and most of identified candidate genes, such as CXCL12 and Tn-C were previously reported to be CAF phenotype-related (19, 20, 40). Orimo and colleagues reported that CXCL12 secreted by CAFs promotes angiogenesis and recruitment of additional inflammatory cells (20). Kojima and colleagues reported that CXCL12 plays a role in the maintenance of the CAF phenotype through formation of an autostimulatory loop (19). Furthermore, CXCL12 has recently emerged as a therapeutic target in various cancers (41–43). However, the transcription factor that regulates CXCL12 remains unknown. In the current study, we showed that CXCL12 was transcriptionally regulated by Twist1 and the expressions of CXCL12 and Twist1 were significantly correlated to each other in cancer tissues with significant clinical relevance. In our previous study, we examined the genome-wide transcriptional regulation by Twist1 in synovial sarcoma combining both ChIP-SEQ and mRNA array (23). In that study, we found that Twist1 upregulated CXCL12 expression in synovial sarcoma and bound to a similar sequence in the proximal E-box site (−509 bp), which strongly supports the results of the present study (23).

Although the identity of CAFs remains obscure, one of the characteristics that distinguish CAFs from conventional activated fibroblasts is their constitutive activation. During normal inflammation, once a wound is repaired, the number of activated fibroblasts decreases and resting phenotype is restored (3). Unlike wound healing, CAFs at tumor site remain perpetually activated (3). In the current study, we showed that Twist1 suppressed cellular senescence in fibroblasts. In fact, the suppression of senescence by Twist1 was previously reported in various types of cells. Maestro and colleagues reported that Twist1 inhibits apoptosis and bypasses p53-induced growth arrest by direct and indirect modulation of the ARF/MDM2/p53 pathway in mouse embryonic fibroblasts (25). Ansieau and colleagues reported that Twist1 represses premature senescence by abrogating regulators of p53- and Rb-dependent pathways in breast cancer cells (21). Our results are highly consistent with those of previous studies and suggest that Twist1 is likely to promote CAF transdifferentiation through the suppression of senescence.

Fibroblast is a highly heterogeneous entity. Chang and colleagues showed that fibroblasts from different anatomical regions display characteristic different phenotypes (44). In addition to the differences between fibroblasts from different anatomical sites, fibroblasts separated from a single tissue are not composed of a homogeneous population (45). Likewise, the expressional profiles of CAFs may show considerable heterogeneity depending on the anatomical region. Therefore, fibroblasts and CAFs used in study should be derived from the same organ to minimize unexpected bias (3). In the current study, we established primary fibroblast and CAF cultures from gastric cancer tissues to study the interaction between gastric cancer cells and CAFs.

In conclusion, we identified Twist1 as a novel and key transcription factor that regulates CAFs. On the basis of our results, Twist1 is suggested as a potent therapeutic target to reverse unfavorable reprogramming of cancer microenvironment.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

Authors' Contributions

Conception and design: K.-W. Lee, S.-H. Kim

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): K.-W. Lee, S.-Y. Yeo, C.O. Sung, C.O. Sung, S.-H. Kim

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): C.O. Sung

Writing, review, and/or revision of the manuscript: K.-W. Lee, S.-Y. Yeo, C.O. Sung, S.-H. Kim

Study supervision: S.-H. Kim

Grant Support

This study was supported by Samsung Biomedical Research Institute grant GL1B30411. And this work was also supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MEST; No. 2013R1A1A2063277).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Footnotes

  • Note: Supplementary data for this article are available at Cancer Research Online (http://cancerres.aacrjournals.org/).

  • Received February 5, 2014.
  • Revision received October 20, 2014.
  • Accepted October 20, 2014.

References

  1. 1.↵
    1. Mueller MM,
    2. Fusenig NE
    . Friends or foes - bipolar effects of the tumour stroma in cancer. Nat Rev Cancer 2004;4:839–49.
    OpenUrlCrossRefPubMed
  2. 2.↵
    1. Hanahan D,
    2. Coussens LM
    . Accessories to the crime: functions of cells recruited to the tumor microenvironment. Cancer Cell 2012;21:309–22.
    OpenUrlCrossRefPubMed
  3. 3.↵
    1. Kalluri R,
    2. Zeisberg M
    . Fibroblasts in cancer. Nat Rev Cancer 2006;6:392–401.
    OpenUrlCrossRefPubMed
  4. 4.↵
    1. Wiseman BS,
    2. Werb Z
    . Stromal effects on mammary gland development and breast cancer. Science 2002;296:1046–9.
    OpenUrlAbstract/FREE Full Text
  5. 5.↵
    1. Shimoda M,
    2. Mellody KT,
    3. Orimo A
    . Carcinoma-associated fibroblasts are a rate-limiting determinant for tumour progression. Semin Cell Dev Biol 2010;21:19–25.
    OpenUrlCrossRefPubMed
  6. 6.↵
    1. Sung CO,
    2. Lee KW,
    3. Han S,
    4. Kim SH
    . Twist1 is up-regulated in gastric cancer-associated fibroblasts with poor clinical outcomes. Am J Pathol 2011;179:1827–38.
    OpenUrlCrossRefPubMed
  7. 7.↵
    1. Hebrok M,
    2. Wertz K,
    3. Fuchtbauer EM
    . M-twist is an inhibitor of muscle differentiation. Dev Biol 1994;165:537–44.
    OpenUrlCrossRefPubMed
  8. 8.↵
    1. Lee MS,
    2. Lowe GN,
    3. Strong DD,
    4. Wergedal JE,
    5. Glackin CA
    . TWIST, a basic helix-loop-helix transcription factor, can regulate the human osteogenic lineage. J Cell Biochem 1999;75:566–77.
    OpenUrlCrossRefPubMed
  9. 9.↵
    1. Fuchtbauer EM
    . Expression of M-twist during postimplantation development of the mouse. Dev Dyn 1995;204:316–22.
    OpenUrlCrossRefPubMed
  10. 10.↵
    1. Stoetzel C,
    2. Weber B,
    3. Bourgeois P,
    4. Bolcato-Bellemin AL,
    5. Perrin-Schmitt F
    . Dorso-ventral and rostro-caudal sequential expression of M-twist in the postimplantation murine embryo. Mech Dev 1995;51:251–63.
    OpenUrlCrossRefPubMed
  11. 11.↵
    1. Wang SM,
    2. Coljee VW,
    3. Pignolo RJ,
    4. Rotenberg MO,
    5. Cristofalo VJ,
    6. Sierra F
    . Cloning of the human twist gene: its expression is retained in adult mesodermally-derived tissues. Gene 1997;187:83–92.
    OpenUrlCrossRefPubMed
  12. 12.↵
    1. Isenmann S,
    2. Arthur A,
    3. Zannettino AC,
    4. Turner JL,
    5. Shi S,
    6. Glackin CA,
    7. et al.
    TWIST family of basic helix-loop-helix transcription factors mediate human mesenchymal stem cell growth and commitment. Stem Cells 2009;27:2457–68.
    OpenUrlCrossRefPubMed
  13. 13.↵
    1. Pettersson AT,
    2. Laurencikiene J,
    3. Mejhert N,
    4. Näslund E,
    5. Bouloumié A,
    6. Dahlman I,
    7. et al.
    A possible inflammatory role of twist1 in human white adipocytes. Diabetes 2010;59:564–71.
    OpenUrlAbstract/FREE Full Text
  14. 14.↵
    1. Ansieau S,
    2. Morel AP,
    3. Hinkal G,
    4. Bastid J,
    5. Puisieux A
    . TWISTing an embryonic transcription factor into an oncoprotein. Oncogene 2010;29:3173–84.
    OpenUrlCrossRefPubMed
  15. 15.↵
    1. Puisieux A,
    2. Valsesia-Wittmann S,
    3. Ansieau S
    . A twist for survival and cancer progression. Br J Cancer 2006;94:13–7.
    OpenUrlCrossRefPubMed
  16. 16.↵
    1. Yang J,
    2. Mani SA,
    3. Donaher JL,
    4. Ramaswamy S,
    5. Itzykson RA,
    6. Come C,
    7. et al.
    Twist, a master regulator of morphogenesis, plays an essential role in tumor metastasis. Cell 2004;117:927–39.
    OpenUrlCrossRefPubMed
  17. 17.↵
    1. Eckert MA,
    2. Lwin TM,
    3. Chang AT,
    4. Kim J,
    5. Danis E,
    6. Ohno-Machado L,
    7. et al.
    Twist1-induced invadopodia formation promotes tumor metastasis. Cancer Cell 2011;19:372–86.
    OpenUrlCrossRefPubMed
  18. 18.↵
    1. Hsu KW,
    2. Hsieh RH,
    3. Huang KH,
    4. Fen-Yau Li A,
    5. Chi CW,
    6. Wang TY,
    7. et al.
    Activation of the Notch1/STAT3/Twist signaling axis promotes gastric cancer progression. Carcinogenesis 2012;33:1459–67.
    OpenUrlAbstract/FREE Full Text
  19. 19.↵
    1. Kojima Y,
    2. Acar A,
    3. Eaton EN,
    4. Mellody KT,
    5. Scheel C,
    6. Ben-Porath I,
    7. et al.
    Autocrine TGF-beta and stromal cell-derived factor-1 (SDF-1) signaling drives the evolution of tumor-promoting mammary stromal myofibroblasts. Proc Natl Acad Sci U S A 2010;107:20009–14.
    OpenUrlAbstract/FREE Full Text
  20. 20.↵
    1. Orimo A,
    2. Gupta PB,
    3. Sgroi DC,
    4. Arenzana-Seisdedos F,
    5. Delaunay T,
    6. Naeem R,
    7. et al.
    Stromal fibroblasts present in invasive human breast carcinomas promote tumor growth and angiogenesis through elevated SDF-1/CXCL12 secretion. Cell 2005;121:335–48.
    OpenUrlCrossRefPubMed
  21. 21.↵
    1. Ansieau S,
    2. Bastid J,
    3. Doreau A,
    4. Morel AP,
    5. Bouchet BP,
    6. Thomas C,
    7. et al.
    Induction of EMT by twist proteins as a collateral effect of tumor-promoting inactivation of premature senescence. Cancer Cell 2008;14:79–89.
    OpenUrlCrossRefPubMed
  22. 22.↵
    1. Lee KW,
    2. Kim JH,
    3. Han S,
    4. Sung CO,
    5. Do IG,
    6. Ko YH,
    7. et al.
    Twist1 is an independent prognostic factor of esophageal squamous cell carcinoma and associated with its epithelial-mesenchymal transition. Ann Surg Oncol 2012;19:326–35.
    OpenUrlCrossRefPubMed
  23. 23.↵
    1. Lee KW,
    2. Lee NK,
    3. Ham S,
    4. Roh TY,
    5. Kim SH
    . Twist1 is essential in maintaining mesenchymal state and tumor-initiating properties in synovial sarcoma. Cancer Lett 2014;343:62–73.
    OpenUrlCrossRefPubMed
  24. 24.↵
    1. Lee KW,
    2. Lee NK,
    3. Kim JH,
    4. Kang MS,
    5. Yoo HY,
    6. Kim HH,
    7. et al.
    Twist1 causes the transcriptional repression of claudin-4 with prognostic significance in esophageal cancer. Biochem Biophys Res Commun 2012;423:454–60.
    OpenUrlCrossRefPubMed
  25. 25.↵
    1. Maestro R,
    2. Dei Tos AP,
    3. Hamamori Y,
    4. Krasnokutsky S,
    5. Sartorelli V,
    6. Kedes L,
    7. et al.
    Twist is a potential oncogene that inhibits apoptosis. Genes Dev 1999;13:2207–17.
    OpenUrlAbstract/FREE Full Text
  26. 26.↵
    1. Bridges RS,
    2. Kass D,
    3. Loh K,
    4. Glackin C,
    5. Borczuk AC,
    6. Greenberg S
    . Gene expression profiling of pulmonary fibrosis identifies Twist1 as an antiapoptotic molecular “rectifier” of growth factor signaling. Am J Pathol 2009;175:2351–61.
    OpenUrlCrossRefPubMed
  27. 27.↵
    1. Pozharskaya V,
    2. Torres-Gonzalez E,
    3. Rojas M,
    4. Gal A,
    5. Amin M,
    6. Dollard S,
    7. et al.
    Twist: a regulator of epithelial-mesenchymal transition in lung fibrosis. PLoS ONE 2009;4:e7559.
    OpenUrlCrossRefPubMed
  28. 28.↵
    1. Bacac M,
    2. Migliavacca E,
    3. Stehle JC,
    4. McKee T,
    5. Delorenzi M,
    6. Coindre JM,
    7. et al.
    A gene expression signature that distinguishes desmoid tumours from nodular fasciitis. J Pathol 2006;208:543–53.
    OpenUrlCrossRefPubMed
  29. 29.↵
    1. Spaeth EL,
    2. Labaff AM,
    3. Toole BP,
    4. Klopp A,
    5. Andreeff M,
    6. Marini FC
    . Mesenchymal CD44 expression contributes to the acquisition of an activated fibroblast phenotype via TWIST activation in the tumor microenvironment. Cancer Res 2013;73:5347–59.
    OpenUrlCrossRefPubMed
  30. 30.↵
    1. Petersen OW,
    2. Nielsen HL,
    3. Gudjonsson T,
    4. Villadsen R,
    5. Rank F,
    6. Niebuhr E,
    7. et al.
    Epithelial to mesenchymal transition in human breast cancer can provide a nonmalignant stroma. Am J Pathol 2003;162:391–402.
    OpenUrlCrossRefPubMed
  31. 31.↵
    1. Campbell I,
    2. Qiu W,
    3. Haviv I
    . Genetic changes in tumour microenvironments. J Pathol 2011;223:450–8.
    OpenUrlCrossRefPubMed
  32. 32.↵
    1. Takebe N,
    2. Khan N,
    3. Ivy P,
    4. Schulz T,
    5. Timmer W,
    6. Harris PJ
    . Review of cancer – associated fibroblasts and therapies that interfere with their activity. Tumor Microenvironment Ther 2013;1:19–36.
    OpenUrl
  33. 33.↵
    1. Walter K,
    2. Omura N,
    3. Hong SM,
    4. Griffith M,
    5. Goggins M
    . Pancreatic cancer associated fibroblasts display normal allelotypes. Cancer Biol Ther 2008;7:882–8.
    OpenUrlCrossRefPubMed
  34. 34.↵
    1. Giannoni E,
    2. Bianchini F,
    3. Masieri L,
    4. Serni S,
    5. Torre E,
    6. Calorini L,
    7. et al.
    Reciprocal activation of prostate cancer cells and cancer-associated fibroblasts stimulates epithelial-mesenchymal transition and cancer stemness. Cancer Res 2010;70:6945–56.
    OpenUrlAbstract/FREE Full Text
  35. 35.↵
    1. Sanz-Moreno V,
    2. Gaggioli C,
    3. Yeo M,
    4. Albrengues J,
    5. Wallberg F,
    6. Viros A,
    7. et al.
    ROCK and JAK1 signaling cooperate to control actomyosin contractility in tumor cells and stroma. Cancer Cell 2011;20:229–45.
    OpenUrlCrossRefPubMed
  36. 36.↵
    1. Cirri P,
    2. Chiarugi P
    . Cancer-associated-fibroblasts and tumour cells: a diabolic liaison driving cancer progression. Cancer Metastasis Rev 2012;31:195–208.
    OpenUrlCrossRefPubMed
  37. 37.↵
    1. Paland N,
    2. Kamer I,
    3. Kogan-Sakin I,
    4. Madar S,
    5. Goldfinger N,
    6. Rotter V
    . Differential influence of normal and cancer-associated fibroblasts on the growth of human epithelial cells in an in vitro cocultivation model of prostate cancer. Mol Cancer Res 2009;7:1212–23.
    OpenUrlAbstract/FREE Full Text
  38. 38.↵
    1. Cheng GZ,
    2. Zhang WZ,
    3. Sun M,
    4. Wang Q,
    5. Coppola D,
    6. Mansour M,
    7. et al.
    Twist is transcriptionally induced by activation of STAT3 and mediates STAT3 oncogenic function. J Biol Chem 2008;283:14665–73.
    OpenUrlAbstract/FREE Full Text
  39. 39.↵
    1. Su YW,
    2. Xie TX,
    3. Sano D,
    4. Myers JN
    . IL-6 stabilizes Twist and enhances tumor cell motility in head and neck cancer cells through activation of casein kinase 2. PLoS One 2011;6:e19412.
    OpenUrlCrossRefPubMed
  40. 40.↵
    1. Midwood KS,
    2. Orend G
    . The role of tenascin-C in tissue injury and tumorigenesis. J Cell Commun Signal 2009;3:287–310.
    OpenUrlCrossRefPubMed
  41. 41.↵
    1. Duda DG,
    2. Kozin SV,
    3. Kirkpatrick ND,
    4. Xu L,
    5. Fukumura D,
    6. Jain RK
    . CXCL12 (SDF1alpha)-CXCR4/CXCR7 pathway inhibition: an emerging sensitizer for anticancer therapies? Clin Cancer Res 2011;17:2074–80.
    OpenUrlAbstract/FREE Full Text
  42. 42.↵
    1. Domanska UM,
    2. Kruizinga RC,
    3. Nagengast WB,
    4. Timmer-Bosscha H,
    5. Huls G,
    6. de Vries EG,
    7. et al.
    A review on CXCR4/CXCL12 axis in oncology: no place to hide. Eur J Cancer 2013;49:219–30.
    OpenUrlCrossRefPubMed
  43. 43.↵
    1. Feig C,
    2. Jones JO,
    3. Kraman M,
    4. Wells RJ,
    5. Deonarine A,
    6. Chan DS,
    7. et al.
    Targeting CXCL12 from FAP-expressing carcinoma-associated fibroblasts synergizes with anti-PD-L1 immunotherapy in pancreatic cancer. Proc Natl Acad Sci U S A 2013;110:20212–7.
    OpenUrlAbstract/FREE Full Text
  44. 44.↵
    1. Chang HY,
    2. Chi JT,
    3. Dudoit S,
    4. Bondre C,
    5. van de Rijn M,
    6. Botstein D,
    7. et al.
    Diversity, topographic differentiation, and positional memory in human fibroblasts. Proc Natl Acad Sci U S A 2002;99:12877–82.
    OpenUrlAbstract/FREE Full Text
  45. 45.↵
    1. Sorrell JM,
    2. Caplan AI
    . Fibroblast heterogeneity: more than skin deep. J Cell Sci 2004;117:667–75.
    OpenUrlAbstract/FREE Full Text
Back to top
Cancer Research: 75 (1)
January 2015
Volume 75, Issue 1

Sign up for alerts

View this article with LENS

Open full page PDF
Article Alerts
Email Article
Citation Tools
Twist1 Is a Key Regulator of Cancer-Associated Fibroblasts
Keun-Woo Lee, So-Young Yeo, Chang Ohk Sung and Seok-Hyung Kim
Cancer Res January 1 2015 (75) (1) 73-85; DOI: 10.1158/0008-5472.CAN-14-0350
del.icio.us logo Digg logo Reddit logo Twitter logo CiteULike logo Facebook logo Google logo Mendeley logo

Jump to section

Advertisement