CD44+ subpopulation of colorectal cancer cells

Stem Cell Research (2013) 11, 820–833

Available online at www.sciencedirect.com

www.elsevier.com/locate/scr

Fibronectin extra domain A (EDA) sustains CD133 + /CD44 + subpopulation of colorectal cancer cells Juanjuan Ou a , Jia Deng a , Xing Wei a , Ganfeng Xie a , Rongbin Zhou a , Liqing Yu b,⁎, Houjie Liang a,⁎⁎ a

Department of Oncology and Southwest Cancer Center, Southwest Hospital, Third Military Medical University, Chongqing, China b Department of Animal and Avian Sciences, University of Maryland, College Park, MD 20742, USA

Received 6 November 2012; received in revised form 24 April 2013; accepted 19 May 2013

Abstract Fibronectin is a major extracellular matrix glycoprotein with several alternatively spliced variants, including extra domain A (EDA), which was demonstrated to promote tumorigenesis via stimulating angiogenesis and lymphangiogenesis. Given that CD133+/CD44+ cancer cells are critical in tumorigenesis of colorectal cancer (CRC), we hypothesize that fibronectin EDA may promote tumorigenesis by sustaining the properties of CD133+/CD44+ colon cancer cells. We found that tumor tissue and serum EDA levels are substantially higher in advanced versus early stage human CRC. Additionally we showed that tumor tissue EDA levels are positively correlated with differentiation status and chemoresistance, and correlated with a poor prognosis of CRC patients. We also showed that in colon cancer cells SW480, CD133+/CD44+ versus CD133−/CD44− cells express significantly elevated EDA receptor integrin α9β1. Silencing EDA in SW480 cells reduces spheroid formation and cells positive for CD133 or CD44, which is associated with reduced expressions of embryonic stem cell markers and increased expressions of differentiation markers. Blocking integrin α9β1 function strongly reversed the effect of EDA overexpression. We also provided evidence suggesting that EDA sustains Wnt/β-catenin signaling activity via activating integrin/FAK/ERK pathway. In xenograft models, EDA-silenced SW480 cells exhibit reduced tumorigenic and metastatic capacity. In conclusion, EDA is essential for the maintenance of the properties of CD133+/CD44+ colon cancer cells. © 2013 Elsevier B.V. All rights reserved.

Introduction New cases for colorectal cancer (CRC) are about 1 million per year worldwide in 2002 (Parkin et al., 2005). In the United States, the number is estimated to reach 1.5 million in 2009 and the CRC is the second leading cause of cancer

⁎ Corresponding author. Fax: + 1 301 405 0223. ⁎⁎ Corresponding author. Fax: + 86 23 65425219. E-mail addresses: [email protected] (L. Yu), [email protected] (H. Liang). 1873-5061/$ - see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.scr.2013.05.009

death (Jemal et al., 2009). A growing body of evidence suggests that some subpopulations of cancer cells within a tumor are capable of self-renewing, differentiating into a phenotypically heterogeneous progeny, and surviving harsh microenvironment (Al-Hajj et al., 2003; Bonnet and Dick, 1997; Lapidot et al., 1994; Singh et al., 2003). These subset of tumor cells, though aberrant, are operationally referred to as tumor progenitor cells (Chaffer and Weinberg, 2011; Clarke et al., 2006; Gupta et al., 2009; Reya et al., 2001). Within certain tumors as many as 25% of the cancer cells have been suggested to possess the properties of tumor progenitor cells (Kelly et al., 2007; Quintana et al., 2008).

EDA-integrin α9β1 sustains CD133+/44+ colorectal cancer cell The existence of tumor progenitor cells in a tumor suggests that efficient cancer therapy may require simultaneous targeting of both the bulk of non-progenitor cells as well progenitor cells (Chaffer and Weinberg, 2011). Using cell surface markers coupled with fluorescenceactivated cell sorting (FACS), several groups have isolated from human CRC specimens several subpopulations of cancer cells that have the properties of tumor progenitor cells, including CD133 positive (CD133+) subpopulation (O'Brien et al., 2007; Ricci-Vitiani et al., 2007), cells positive for both CD44 and epithelial cell adhesion molecule (epCAM) (epCAMhigh/ CD44+) (Dalerba et al., 2007), and CD26+ cells (Pang et al., 2010). In cultured human CRC cell lines HT29 and SW1222, CD44 and CD24 double positive cells have been reported to have tumor progenitor cell properties, and intriguingly, a single CD44+/CD24+ cell from SW1222 cell line, when grown in matrigel, is capable of forming large megacolonies that differentiate into enterocyte, enteroendocrine, and goblet cell lineages (Yeung et al., 2010). Clinicopathological studies have suggested that CD133 is an independent prognostic marker for low survival in CRC patients (Choi et al., 2009; Horst et al., 2009; Li et al., 2009a; Wang et al., 2009). Taken together, the aforementioned findings imply that targeting CD133+ and CD44+ cancer cells or pathways sustaining CD133+/ CD44+ subpopulation might be an efficient strategy for CRC therapy. Tumor microenvironment plays critical roles in regulating tumor evolution, growth and metastasis (Allen and Louise Jones, 2010; Kenny et al., 2007). Cancer cells interact with their extracellular matrix (ECM) during proliferation and migration. Fibronectin is an important ECM glycoprotein with several alternatively spliced variants (ffrench-Constant, 1995). We and others have shown that the alternative extra domain A (EDA) of fibronectin promotes angiogenesis, lymphangiogenesis and metastasis of colorectal tumors (Ou et al., 2011, 2010; Rybak et al., 2007). The expression levels of EDA are linked to migration capacity of primary glioma cells (Ohnishi et al., 1998). EDA modulates cell behaviors via binding to its cell surface integrin receptors consisting of α and β subunits (Cabodi et al., 2010; Cox et al., 2010; Desgrosellier and Cheresh, 2010; Hynes, 2002). Recently, integrins were used for selection of human neural progenitor cells (Hall et al., 2006) and for enrichment of tumor progenitor cells in glioblastoma (Lathia et al., 2010), breast cancer (Allen et al., 2011; Vaillant et al., 2008), and prostate cancer (Patrawala et al., 2007). Integrin a9β1 is a receptor for fibronectin EDA (Shinde et al., 2008), and it was recently implicated in promotion of tumor cell invasion in breast cancer (Allen et al., 2011). Collectively, these previous studies led us to the hypothesis that EDA may act as a pivotal factor for maintaining the properties of CD133+/CD44+ subpopulation in CRCs. In this study, we tested this hypothesis and found that EDA levels are substantially higher in tumor tissues and serum of advanced versus early stage CRC patients, and correlated with a poor prognosis of CRC patients. We demonstrated that EDA expression is required for maintaining CD133+ and CD44+ subpopulations and tumor colony forming capacity in SW480 human CRC cells. In vivo, we found that inhibiting EDA expression in SW480 cells reduces their growth and metastasis. Additionally we provided evidence suggesting that the focal adhesion kinase (FAK)/ERK/β-catenin signaling pathway is involved in EDA function. Our finding for the first time links EDA to CD133+/

821 CD44+ subpopulation properties and Wnt/β-catenin activity of colorectal cancer cells.

Materials and methods Tissue and blood samples A tissue chip consisting of 105 human CRC specimens and 10 human normal colorectal tissues for statistical analysis of clinicopathological features was purchased from Xian-Yi-Lina Biological Technology (China). According to the information from the tissue chip company, the ten normal human colon tissues were obtained from tissue biopsies of non-cancer subjects who had undergone colonoscopy examination. A tissue chip consisting of 56 human CRC biopsy specimens collected at our hospital was used for analysis of chemotherapy resistance parameters. Another tissue chip consisting of 86 human CRC biopsy specimens collected at our hospital was used for analysis of the correlation between EDA expression levels and disease free survival (DFS) of CRC patients. A set of chips consisting of 176 human CRC specimens with overall survival follow-up information were purchased from Shanghai Outdo Biotech (China) and used specifically for the analysis of the correlation between EDA expression levels and overall survival (OS) of CRC patients. A separate set of 40 human CRC tissue specimens for immunofluorescence microscopy studies and 77 blood samples used for measuring blood EDA concentrations were collected from CRC patients who had undergone surgeries or biopsy at our hospital. Tumors were staged by anatomic pathologists in the Department of Pathology of our hospital, according to the Union for International Cancer Control classification system. All experiments were approved by the ethics committee of Southwest Hospital, the Third Military Medical University.

Cell culture The human SW480 CRC cell line was obtained from the American Type Culture Collection (Manassas, VA), and maintained in Leibovitz's L-15 Medium (Invitrogen Corp.) supplemented with 10% fetal bovine serum at 37 °C under 5% CO2.

Antibodies A monoclonal anti-focal adhesion kinase (FAK) (Catalog#: 20431) antibody was obtained from Upstate Biotechnology (Upstate, Waltham, MA). A polyclonal anti-phospho-FAK (phospho-Tyr397) (Catalog#: KAP-TK131D) antibody and a CD44/-PE conjugated antibody (Catalog#: 561858, Isotype IgG2b, κ) were from BD Biosciences (San Diego, CA). A CD133/-APC conjugated antibody (Catalog#: 130-090-854) was obtained from Miltenyi Biotech. A rabbit monoclonal anti-phospho-Akt (Catalog#: 4058L), a rabbit polyclonal antiAkt antibody (Catalog#: 9272S), a rabbit monoclonal antiPhospho-p44/42 MAPK (Erk1/2) (Catalog#: 4370) and a mouse monoclonal anti-Erk antibody (Catalog# 4696S) were from Cell Signaling (Beverly, MA). Mouse monoclonal antibodies against human fibronectin EDA domain (IST-9) (Catalog#: ab6328) and integrin α9β1 (Y9A2) (Catalog#: ab27947), an isotype control antibody (mouse IgG1) for integrin α9β1, a rabbit monoclonal

822 anti-CD44 antibody (Catalog#: ab51037), a mouse monoclonal anti-CD133 antibody (Catalog#: ab27699) and a goat polyclonal anti-OCT4 antibody (Catalog#: ab27985) were purchased from Abcam Inc. A goat polyclonal antibody raised against cleaved caspase-3 (h176) (Catalog# sc-22171), a goat polyclonal antibody against cleaved caspase-9 (h331) (Catalog#: sc-22182), a rabbit polyclonal antibody against cleaved poly ADP-ribose polymerase (PARP) (h215)-R (Catalog#: sc-23461R) and a cyclin D1 (H-295) (Catalog#: sc-753), and a mouse monoclonal antibody against glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (D-6) (Catalog#: sc-166545), were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). A mouse monoclonal antibody against active β-catenin (Catalog#: 05-665), a rabbit polyclonal antibody against anti-βcatenin (Catalog#: AB19022) and a rabbit polyclonal antiGlut1 antibody (Catalog#: 07-1401) were from Millipore Corp.

ELISA EDA concentrations in the sera of patients with CRC were quantified using a human fibronectin EDA-specific ELISA Kit (BPE301106H, RB, Shanghai Hushang Biotechnology Co., Ltd.).

Immunohistochemical and immunofluorescence microscopy studies

J. Ou et al.

FAK and control shRNA plasmids and establishment of stable cell lines One pair of oligonucleotides with a short hairpin was designed based on the human FAK cDNA sequences (GenBank accession no. L13616) to target human FAK (forward: 5′-GATCCCCCCACCT GGGCCAGTATTATTTCAAGAGAATAATACTGGCCCAGGTGGTTTT TGGAAA-3′ and reverse: 3′-GGGGGTGGACCCGGTCATAAT AAAGTTCTCTTATTATGACCGGGTCCACCAAAAACCTTTTCGA-5′). A pair of control primers was also synthesized to generate a control shRNA construct (forward: 5′-GATCCCCGA CGTGGG ACTGAAGGGGTTTCAAGAGAACCCCTTCAGTCCCACGTCTTTTTG GAAA-3′ and reverse: 3′-GGGCTGCA CCCTGACTTCCCCAAA GTTCTCTTGGGGAAGTCAGGGTGCAGAAAAACCTTTTCGA-5′), which had no homology to relevant human genes. These shRNAs were then inserted into pGenesil-l vector. All resultant constructs were verified by DNA sequencing, and then transfected into SW480-derived CD133+ /CD44+ cells stably overexpressing EDA with lipofectamineTM 2000 transfection reagent (Invitrogen, Carlsbad, CA, USA). Transfected cells were enriched by selection for 1 week with 1 mg/ml G418. Three stable cell lines were established, including an untransfected group, a line expressing a control shRNA, and a cell line expressing FAK shRNA.

Flow cytometry analysis of cell cycle distribution All tissue chip slides were routinely dewaxed, rehydrated, and prepared for immunohistochemistry. The slides were incubated in 0.3% H2O2 in methanol for 30 min to block endogenous peroxidase. Antigens were retrieved with 10 mmol/l sodium citrate (pH 6) for 15 min in a microwave oven. The slides were then incubated with the selected antibody at 37 °C for 1 h and at 4 °C overnight. Slides without treatment with the primary antibody served as negative controls. The slides were developed with an EnVisionTM method (DAKO, Capinteria, CA). The slides were visualized using the diaminobenzidine solution (DAKO, Capinteria, CA), and then lightly counterstained with hematoxylin. Evaluation of immunohistochemical staining reaction was performed in accordance with the Immunoreactive Score (IRS). IRS = SI (staining intensity) X PP (percentage of positive cells). Negative SI = 0; Weak SI = 1; Moderate SI = 2; Strong SI = 3. Negative PP = 0; 10% PP = 1; 11–50% PP = 2; 51–80% PP = 3; and N 80% PP = 4. Ten microscopic fields (100×) from different areas of each tissue section on a tissue chip were used for the IRS evaluation. Slides were examined and scored independently by 3 pathologists blinded to the patient information. Frozen tissue samples for immunofluorescence staining were cut into 4 μm serial sections. All the frozen sections were fixed in ice-acetone for 20 min, washed with PBS for 3 times for 5 min each, and incubated for 30 min at room temperature in a protein-blocking solution. The sections were incubated with the primary antibodies for 1 h at 37 °C and then at 4 °C overnight. After wash, the sections were incubated at 37 °C for 1 h with appropriate secondary antibodies, including FITCconjugated goat anti-rabbit IgG (1:50, Santa Cruz), FITCconjugated goat anti-mouse IgG, (1:50, Santa Cruz), or TRITCconjugated goat anti-mouse IgG, (1:50, Beyotime, China). The sections were counterstained with Hoechst 33258 to reveal cell nuclei. The specificity of a primary antibody was verified by omitting that antibody in the reaction.

Cells were trypsinized and washed with PBS for 3 times. The cells were resuspended in 0.5 ml PBS, fixed for 24 h in 75% alcohol, and then resuspended in PBS containing 50 μg/ml of propridium iodode (PI) and 10 μg/ml of RNasaA, The cells were incubated at room temperature for 20 min in the dark prior to flow cytometry analysis using FCM (FACSCan, Becton Dickinson, San Jose, CA). The program used for the analysis was ModFit Version 5.2 from Verity Software House (Topsham, ME).

Protein extraction and Western blotting Cell lysates were prepared with M-PER Mammalian Protein Extraction Reagent (PIERCE, PA, USA). A total of 30 μg of lysate proteins were separated by SDS-PAGE after heat denature, transferred onto PVDF membranes, and incubated with 5% non-fat milk dissolved in PBS-Tween 20 solution for 1 h, followed by incubation with a primary antibody overnight at 4 °C. After wash, the membranes were incubated with an appropriate HRP-conjugated secondary antibody, and then developed with enhanced chemiluminescence (ECL) detection reagents (Amersham Pharmacia Biosciences).

Co-immunoprecipitation assay Total protein lysates (500 μg) from each sample were immunoprecipitated in 400 μl lysate buffer containing 2 μl anti-EDA mouse monoclonal antibody and inhibitors of various proteases, phosphotases and kinases at 4 °C for 4 h with rotation. Protein A-conjugated agarose beads (25 μl) were then added into the immunoprecipitation reaction with an additional 5 h of rotation at 4 °C. The antigen-antibody complexes were precipitated by a quick

EDA-integrin α9β1 sustains CD133+/44+ colorectal cancer cell centrifugation, followed by 4 times of wash with cold PBS. Controls included an aliquot of rabbit serum to replace the EDA antibody, in the immunoprecipitation reaction. The pellets were suspended in 20 μl of 2 × SDS reducing Western blot loading buffer and boiled for 10 min, followed by SDS-PAGE. The EDA-immunoprecipitates were subjected to Western blot assay to detect EDA and integrin α9β1 in the immunoprecipitates.

Lentiviral particles and establishment of stable cell lines The lentiviral particles were purchased from GeneChem Co., Ltd. (Shanghai). Stably transfected cells were established following the manufacturer's instructions.

Isolation of CD133+/CD44+ SW480 cells

subpopulation from

SW480 cells were cultured in serum free Leibovitz's L-15 Medium with bFGF (10 ng/ml), and EGF (10 ng/ml) (stem cell culture medium) for at least 24 h to enhance expression of surface antigens. Cells were then labeled with CD133/-APC conjugated antibody and a CD44/-PE conjugated antibody and sorted by fluorescence-activated cell sorting (FACS).

Sphere formation of tumor cells SW480 cells were cultured overnight. After trypsinization, the cells were plated in 96-well plates and cultured in the stem cell culture medium described above. The medium was refreshed every 3 days to sustain the tumor cell sphere formation. One week later, the spheres were photographed and the size and number of spheres were analyzed by ImageJ (NIH).

Cell apoptosis Cell apoptosis was measured using the Annexin V-PE Apoptosis Detection Kit I (BD PharMingen, San Diego, CA, USA) via the manufacturer's instructions. Briefly, cells were harvested with mild trypsinization and resuspended in binding buffer. Samples were stained with Annexin V-PE and 7AAD. Fluorescence was measured using an 11-color LSR II (BD Biosciences). Ten thousand events were collected per sample. Data were analyzed using FlowJo software (Tree Star, Ashland, OR, USA). Gates were set based on single-stain controls. Percentage death was measured as cells that stained positive for either or both 7AAD and Annexin V.

Clonogenic assay For clonogenic assays, EDA-overexpressing CD133+/CD44+ cells were plated into 24-well plates and cultured overnight. The cells were then grown in the same medium containing the integrin α9β1 antibody (20 μg/l), IST (20 μg/l) or an isotype IgG (20 μg/l). Ten days later after plating, colonies were stained using the crystal violet solution and the number

823 and area of colonies of N 50 cells were calculated using ImageJ software.

In vivo tumor models Four-to-six week-old Balb/c nude mice (body weight: 16 to 20 g) were purchased from the Experimental Animal Center, Institute of Laboratory Animal Sciences (China). The mice were subcutaneously injected with control or EDA silenced SW480 cells (1 × 106 cells per mouse) at the both thighs. The mice were sacrificed 6 weeks after injection and the xenografts were collected. Metastases were induced by tail vein injection of 1 × 106 control or EDA silenced SW480 cells. Animals were sacrificed 8 weeks after injection. All of our animal studies have been approved by the Institutional Animal Care and Use Committee of the Third Military Medical University.

Statistical analysis Data are expressed as mean ± SEM (standard error of the mean). The pathologic scoring data of human specimens were analyzed by 2 biostatisticians in the Department of Statistics, The Third Military Medical University, China. The statistical analysis was performed by one-way ANOVA (when N 3 groups) or Students t-test (between two groups) using Graph Pad Prism software. For cancer survival analysis, Kaplan–Meier Survival Curves of Overall Survival were used. For the analysis of correlation between EDA expression levels and clinical parameters, Fisher's exact test was used. Differences between the values were considered statistically significant when P b 0.05.

Results Fibronectin EDA levels are increased in the tumor tissues and blood samples of patients with advanced CRCs It has been reported that EDA expression levels are significantly higher in malignant tumors than benign tumors and normal tissues (Rybak et al., 2007). We have previously shown that EDA stimulates lymphangiogesis and lymphatic metastasis of CRC cells (Ou et al., 2010). Based on these observations, we speculated that EDA levels may be higher in advanced CRC as well as correlated to clinicopathological features. To examine this speculation, we used tissue chips to perform immunohistochemistry staining for the correlation analysis between EDA and clinicopathological features. Consistently, EDA levels were substantially higher in CRC in comparison to that in normal colon tissue and were significantly higher in CRC of clinically advanced stages (III and IV) relative to early stages (I and II) (Fig. 1a). Additionally, we assayed EDA concentrations in blood samples of 77 patients with CRC, and found that EDA concentrations were significantly increased in patients with advanced CRC than those with early stage CRC (Fig. 1b). As shown in Table 1, EDA levels were also correlated with poor differentiation and metastasis of CRC. To identify the relationship between tumor tissue EDA levels and patients'

824 responses to chemotherapy, we performed immunohistochemistry with another tissue chip from tumor biopsies of 56 stage III/IV CRC patients without surgery. We found that tumor tissue EDA levels were negatively correlated with the Objective Response Rate, an indicator of chemosensitivity (Table 2). More impressively, the patients with higher EDA expression levels had poorer disease free survival (DFS) (Fig. 1c) and overall survival (OS) than those with lower EDA expression (Fig. 1d). These findings indicate that CGI-58 deficiency promotes CRC progression.

Increased expression of EDA receptor integrin α9β1 in CD133+/CD44+ cancer cells It has been shown that CD44+ and CD133+ cells play an important role in tumor initiation and progression (Chaffer and Weinberg, 2011). Given the increased expression of EDA in advanced carcinomas (Fig. 1), we hypothesized that EDA pathway may sustain the CD133+/CD44+ cell subpopulation. To test this hypothesis, we first examined the levels of EDA, EDA receptor integrin α9β1, an embryonic stem cell marker OCT3/4 (Nichols et al., 1998; Takahashi et al., 2007) and a progenitor cell marker CD133 (Miraglia et al., 1997; Singh et al., 2004) in human CRC surgical specimens by immunofluorescence microscopy. We collected 40 tissue specimens consisting of 17 advanced staged CRC (III & IV) and 23 early staged CRC (I & II). As expected, OCT3/4-positive cells were frequently seen

J. Ou et al. in the EDA-enriched region (Figs. 2a–b). Integrin α9β1 also colocalized with CD133 (Figs. 2a–b). To determine if integrin α9β1 expression levels differ between CD133+/CD44+ and CD133-/CD44− cells, we separated these two subsets of cells in cultured SW480 human CRC cells by the fluorescence-activated cell sorter, and measured integrin α9β1 mRNA as well as protein levels respectively. Interestingly, the EDA receptor integrin α9β1 mRNA levels were N 2.5-fold higher in cells positive for CD133 and CD44 than those negative for these two cell surface markers (Fig. 2c). Consistently, the integrin α9β1 protein was more abundant in CD133+/CD44+ cells than CD133−/CD44− cells (Fig. 2d). Despite increased expression of integrin α9β1, the expression levels of mRNA and protein for its ligand EDA were not different between CD133+/CD44+ and CD133-/CD44− cells (Figs. 2e–g).

EDA is required for the sphere formation capacity of CRC cells To determine if EDA autocrine and/or paracrine contribute to the sphere formation capacity of CRC cells, we used a lentiviral shRNA-based approach to silence EDA expression in SW480 cells cultured in the serum-free stem cell culture medium. As shown in Fig. 3, while the tumor cell sphere formation didn't differ between control shRNA lentivirusinfected and uninfected SW480 cells, EDA silencing significantly reduced the number (Fig. 3a) and the size (Fig. 3b) of

Figure 1 EDA levels in tumor tissues and blood samples are positively correlated with clinical stages of CRC patients. (a) Immunohistochemistry of the tissue chip containing human CRC specimens and normal tissues using an anti-EDA antibody. (b) The plasma EDA concentrations in CRC patients with different stages. n = 32 for stages I/II (male = 15, female = 17; b 60 years old = 16, ≥60 years old = 16). n = 45 for stages III/IV (male = 25, female = 20; b 60 years old = 23, ≥60 years old = 22). *, P b 0.01. (c) Statistical analysis of the correlation between EDA expression levels and the Disease Free Survival of CRC patients (stages I & II, n = 61; stage III, n = 25). P b 0.01 (Kaplan–Meier Survival Curves; IRS ≤ 6 included in EDA low; IRS N 6 included in EDA high). (d) Statistical analysis of the correlation between EDA expression levels and the Overall Survival of CRC patients (stages I & II, n = 105; stage III, n = 71). P b 0.01 (Kaplan–Meier Survival Curves; IRS ≤ 6 included in EDA low; IRS N 6 included in EDA high).

EDA-integrin α9β1 sustains CD133+/44+ colorectal cancer cell Table 1

The correlation between EDA expression levels and clinicopathological features in colorectal carcinomas. No. Pts a

Variables

Site Colon adenocarcinoma Adjacent colon tissue Age (years) b 60 ≥ 60 Gender male female Organ Colon rectal Differentiation well moderate poor Tumor status T2 T3 T4 Nodal status N0 N1–2 Metastasis status M0 M1 Clinical stage I + II III + IV a

EDA expression index (IRS)

Fisher's (P)

b3

3–7

≥8

105 10

45 (44.8) 9 (90.0)

33 (31.4) 1 (10.0)

25 (23.8) 0 (0)

43 62

15 (34.9) 32 (51.6)

17 (39.5) 16 (25.8)

11 (25.6) 14 (22.6)

61 44

27 (44.3) 20 (45.5)

21 (34.4) 12 (27.3)

13 (21.3) 12 (27.3)

74 31

32 (43.2) 15 (48.4)

25 (33.8) 8 (25.8)

17 (23.0) 8 (25.8)

31 48 26

17 (54.8) 22 (45.8) 8 (30.8)

13 (41.9) 14 (29.2) 6 (23.1)

1 (3.2) 12 (25.0) 12 (46.2)

15 50 40

5 (33.3) 27 (54.0) 15 (37.5)

6 (40.0) 12 (24.0) 15 (37.5)

4 (26.7) 11 (22.0) 10 (25.0)

75 30

39 (52.0) 8 (26.7)

23 (30.7) 10 (33.3)

13 (17.3) 12 (40.0)

90 15

44 (48.9) 3 (20.0)

28 (31.1) 5 (33.3)

18 (20.0) 7 (46.7)

61 44

38 (62.3) 9 (20.5)

16 (26.2) 17 (38.6)

7 (11.5) 18 (40.9)

0.03

0.201

0.668

0.724

0.003

0.435

0.021

0.048

0.000

Pts, Patients.

tumor cell spheres by ~ 40% and ~ 65%, respectively, when compared to the control shRNA-infected cells. Since EDA promoted the sphere formation of SW480 cells, we speculated that EDA silencing may reduce the relative abundance of CD133 and CD44 positive cell subpopulation in SW480 cells. Consistently, CD133+ cells were reduced from 90% in the uninfected SW480 cells and 88% in the control shRNA lentivirus-infected SW480 cells to 56% in the EDA shRNA lentivirus-infected SW480 cells. Similarly, CD44+ cells fell

Table 2

from 58% and 54% in the two control groups to 10% in the EDA silencing group (Fig. 3c). In agreement with reduced CD133+/ CD44+ cells, EDA shRNA virus-infected SW480 cells expressed decreased amounts of many embryonic stem cell markers including SOX2 (Avilion et al., 2003; Wang et al., 2006), NANOG (Chambers et al., 2003), and OCT3/4 (Nichols et al., 1998), and increased amounts of protein markers for differentiated cells such as cytokeratin 7 (CK7) and mucin 2 (MUC2) (Shin et al., 2010) (Fig. 3d).

The correlation between EDA expression and clinical objective response rate of advanced CRC patients.

Variables

Objective response rate Complete response Partial response Stable disease a

825

Pts, Patients.

No. Pts a

EDA expression index (IRS)

Fisher's (P)

b3

3–7

≥8

12 (60.0) 9 (45.0) 3 (18.8)

6 (30.0) 4 (20.0) 4 (25.0)

2 (10.0) 7 (35.0) 9 (56.3)

0.034 20 20 16

826

J. Ou et al.

Figure 2 Frequent co-existence of EDA or integrin α9β1 with stem cell markers in CRC tissues and increased expression of integrin α9β1 in CD133+/CD44+ cancer cells. (a) Immunofluorescence histochemistry of CRC specimens demonstrating OCT3/4 (red) localization in the region with abundant EDA expression (green), co-localization (purple, marked with white arrows) of OCT3/4 (red) with nucleus (DAPI blue), and co-localization (yellow, marked with white arrows) of integrin α9β1 (red) with CD133 (green). (b) Quantification of the positive rates of OCT3/4+, CD133+, integrin α9β1+ and CD133+/integrin α9β1+ cells based on the immunofluorescence staining. (e–f) The mRNA and protein levels of integrin α9β1 and EDA in CD133/CD44 double positive and double negative subpopulations of SW480 cells. *, P b 0.01.

EDA receptor integrin α9β1 is required for clonogenicity of CD133 + /CD44 + cells and for proliferation of CRC cells Since EDA has been shown to function through binding to its receptor integrin α9β1 in SW480 cancer cells (Shinde et al., 2008), we examined the effect of EDA silencing on the interaction of EDA and integrin α9β1 in SW480 cells by immunoprecipitation. As shown in Fig. 4a, EDA silencing dramatically reduced immunoprecipitable integrin α9β1, indicating the existence of EDA and integrin α9β1 interactions in these cells. To determine if EDA requires integrin α9β1 to promote proliferation of CRC cells, we stably overexpressed EDA in CD133+/CD44+ cells sorted from SW480 cells in order to amplify the interaction between EDA and integrin α9β1, and then treated these CD133+/CD44+ cells with an integrin α9β1 blocking antibody or an isotope control antibody. Inhibition of integrin α9β1 function in EDA overexpressing CD133+/ CD44+ cells significantly decreased tumor cell sphere formation (data not shown). Strikingly, integrin α9β1 blocking antibody treatment reduced the number and the size of CD133+/CD44+ cell colonies by ~ 94 and 97%, respectively, when compared to isotype control antibody treatment, which didn't significantly alter CD133+/CD44+ cell colony formation relative to un-treated group (Fig. 4b). To show that EDA plays a predominant role in promoting CRC cell proliferation, CD133+/CD44+cells were infected with lentiviruses overexpressing EDB, another major alternatively

spliced form of fibronectin. As shown in Fig. 4c, EDA antibody still substantially inhibited cancer cell colony formation even EDB was overexpressed. Collectively, our data suggest that the EDA-integrin α9β1 interaction plays a critical role for the proliferation of CRC cells. To determine how integrin α9β1 inhibition influences SW480 cell proliferation, we performed cell cycle analysis and found that the G2 + S phase fraction in the integrin α9β1 antibody-treated SW480 cells fell to ~ 50% of two control groups (Fig. 4d), and consistently, the expression of cyclin D1, a protein known to promote cell cycle progression through G1–S phase (Fu et al., 2004) was suppressed (Fig. 4f). We also found that the apoptosis in integrin α9β1 antibody-treated SW480 cells significantly increased (Fig. 4e), and three apoptotic protein markers, including cleaved caspase 3, cleaved caspase 9, and cleaved poly (ADP-ribose) polymerase (PARP), were up-regulated in the integrin α9β1 antibody-treated SW480 cells (Fig. 4f).

EDA silencing reduces activation of FAK/ERK/βcatenin signaling pathway in CD133+/CD44+ colon cancer cells The Wnt/β-catenin pathway is critical in the development and progression of CRCs (Kinzler and Vogelstein, 1996). We therefore measured beta-catenin protein levels in these cell subpopulations and in the entire cell population of SW480. As expected, CD133+/CD44+ versus CD133−/CD44− cells express

EDA-integrin α9β1 sustains CD133+/44+ colorectal cancer cell

827

Figure 3 EDA silencing reduces sphere formation capacity and CD133+ and CD44+ subpopulations of SW480 cells. (a, b) Images and quantification of the number and size of spheres formed from SW480 cells without lentiviral infection (uninfected) or infected with lentiviruses expressing control shRNA or EDA shRNA under serum-free stem cell culture conditions as described in Methods. (c) The percentage of CD133+ and CD44+ subpopulations determined by flow cytometry in the same cell groups. (d) Western blots of embryonic stem cell markers (SOX2, NANOG, and OCT3/4) and differentiation markers (CK7 and MUC2) in the same cell groups. *, P b 0.05 (vs. control shRNA).

more abundant beta-catenin protein and the expression level of beta-catenin in the entire cell population is in the middle of two subpopulations (Fig. 5a). To determine if inhibiting EDA suppresses Wnt/β-catenin pathway, we measured levels of active β-catenin in EDA shRNA-treated SW480 cells and found that the active β-catenin was substantially reduced in EDA shRNA-treated SW480 cells relative to control groups (Fig. 5B). Since FAK is a downstream mediator of integrin signaling (Cabodi et al., 2010) and plays an important role in maintaining the properties of embryonic and cancer stem cells (Ginestier et al., 2010; Lee et al., 2011), we then explored whether FAK is involved in the suppression of Wnt signaling by EDA silencing. Indeed, EDA silencing in SW480 cells dramatically reduced phosphorylation of FAK as well as the two FAK downstream kinases Akt and Erk1/2 (Cabodi et al., 2010) (Fig. 5b). We have previously generated a plasmid expressing FAK shRNA that can efficiently silence FAK in SW480 cells (Chen et al., 2010). To determine if EDA requires FAK to activate β-catenin, this FAK shRNA plasmid was transiently transfected into SW480-derived CD133+/CD44+ cells stably overexpressing EDA. In these EDA-overexpressing CD133+/CD44+ cells, there was abundant expression of active β-catenin protein, and FAK shRNA dramatically attenuated the expression of active β-catenin protein in EDAoverexpressing CD133+/CD44+ cells (Fig. 5c). Since both PI3K-Akt pathway and MAPK-Erk pathway are downstream targets of FAK and both phospho-Akt and phospho-Erk were reduced in EDA shRNA-treated SW480 cells (Fig. 5b), to determine the relative contribution of PI3K-Akt and MAPK-Erk in EDA-mediated activation of β-catenin, we treated EDA-overexpressing CD133+ /CD44+ cells with a

PI3K-Akt pathway inhibitor LY294002 and a MAPK-Erk pathway inhibitor PD98059. As shown in Fig. 5e, only the MAPK-Erk inhibitor PD98059 can reduce expression of active β-catenin in EDA-overexpressing CD133+/CD44+ cells, suggesting a predominant role of MAPK-Erk pathway in mediating EDA-induced activation of β-catenin. On the contrary, FAK knockdown, PI3K/Akt inhibitor or MAPK/Erk inhibitor treatment has no effect on β-catenin in CD133−/CD44− cells (Figs. 5 d, f). Based on these observations, we proposed a signaling pathway for EDA regulation of β-catenin activity in Fig. 5g.

EDA silenced SW480 cells show reduced tumorigenic capacity To determine the in vivo significance of our in vitro observations, we subcutaneously injected both thighs of each nude mouse (3 mice per group) with equal numbers (1.0 × 105 per mouse) of control or EDA knockdown SW480 cells. The xenografts in EDA knockdown group generated significantly slower than control groups (Fig. 6a). Additionally, EDA knockdown SW480 cells showed significantly reduced tumor diameters (Fig. 6b) and a 60% reduction in tumor weight (Fig. 6c).

Xenografts derived from EDA-silenced SW480 cells express reduced proliferation maker and show increased apoptosis The volume of a tumor is balanced by proliferation and apoptosis of tumor cells. To address why EDA-silenced

828

J. Ou et al.

Figure 4 Blocking EDA and integrin α9β1 function reduces clonogenicity of CD133+/CD44+ cells and promotes apoptosis in SW480 cells. (a) Co-immunoprecipitation assay showing interactions of EDA with integrin a9β1 in SW480 cells. (b) Images and quantification of the number and size of colonies formed from SW480-derived CD133+/CD44+ cells that were stably infected with EDA-expressing lentiviruses and then treated without (Untreated) or with isotype control antibody (Ab) (20 ng/ml) or integrin a9β1 blocking antibody (20 ng/ml). *, P b 0.05 (vs. isotype antibody). (c) Images and quantification of the number and size of colonies formed from SW480-derived CD133+/CD44+ cells that were stably infected with EDB-expressing lentiviruses and then treated without (Untreated) or with isotype control antibody (Ab) (20 ng/ml) or EDA blocking antibody (IST-9) (20 ng/ml). *, P b 0.05 (vs. isotype antibody). (d) Cell cycle distributions determined by FACS analysis of whole SW480 cells treated with the antibodies. (e) Annexin V-7AAD FACS assay of cell apoptosis. P b 0.0001. (f) Western blots of a cell cycle progression-promoting protein cyclin D1 and apoptosis-related proteins (cleaved caspase 3, caspase 9, and PARP) in whole SW480 cells treated with the antibodies.

SW480 cells grew slower in xenografts, we did immunohistochemistry using the EDA antibody and an antibody against a proliferation marker Ki67, and performed TUNEL assays in xenografts collected from three groups of nude mice (Fig. 6d). As expected, EDA was almost undetectable in the xenografts derived from EDA shRNA lentivirus-infected SW480 cells. The expression of Ki67 was substantially lower in these xenografts relative to those from the two control groups. TUNEL assays revealed a dramatic increase in apoptosis in EDA shRNA xenografts. Immunofluorescence staining showed that the frequency of CD133+ /CD44+ cells was reduced in EDA shRNA xenografts. Meanwhile, there were decreased expression levels of progenitor cell markers, including OCT3/4, GLUT-1 (Li et al., 2009b),

and the activated form of β-catenin in these xenografts (Fig. 6e).

EDA-silenced SW480 cells show reduced metastasis To determine if EDA expression influences tumor metastasis, we injected nude mice by tail vein either SW480 cells without infection of any lentiviruses, or with lentiviruses expressing either control shRNA or EDA shRNA, and then compared the incidence of liver and lung metastases 8 weeks after injection. The gross appearance and H&E staining of tissues showed a significant decrease in the number and the size of liver metastases in the mice injected with EDA silenced SW480 cells

EDA-integrin α9β1 sustains CD133+/44+ colorectal cancer cell

829

Figure 5 FAK/ERK/β-catenin pathway is involved in EDA-integrin α9β1 signal transduction. (a) Western blots of active β-catenin in CD133+/CD44+, CD133−/CD44− and the whole bulk of SW480 cells. (b) Western blots of active β-catenin as well as phosphorylated and total FAK, ATK and ERK1/2 in whole SW480 cells without lentiviral infection (Uninfected) or infected with lentiviruses expressing control shRNA or EDA shRNA. (c) Western blots of active β-catenin in SW480-derived CD133+/CD44+ that were stably infected with EDA-expressing lentiviruses and then transiently transfected with a control shRNA plasmid or FAK shRNA plasmid. (d) Western blots of active β-catenin in SW480-derived CD133−/CD44− transiently transfected with a control shRNA plasmid or FAK shRNA plasmid. (e–f) Western blots of active β-catenin in EDA-overexpressing CD133+/CD44+ cells or in CD133−/CD44− cells treated with a PI3K/AKT inhibitor LY294002 (25 μmol/ml) and a MAPK/ERK inhibitor PD98059 (25 μmol/ml) for 24 h. (g) Proposed EDA signaling pathway in CRC cells.

as compared with those injected with control cells (Fig. 6f). Similar changes were observed for lung metastases (Fig. 6g).

Discussion The major finding of this study was that EDA-integrin α9β1 pathway is required for maintaining CD133 and CD44 positive

subpopulations and tumorigenic capacity of CRC cells. We provided several pieces of evidence supporting involvement of FAK/ERK/β-catenin pathway in EDA-α9β1 function. To our knowledge, this study is the first to link EDA to CD133+/ CD44+ subpopulation properties and Wnt/β-catenin activity of cancer cells. It is interesting to find that EDA level in tumor tissues and patient blood samples is positively correlated to the clinical

830

J. Ou et al.

Figure 6 EDA silencing reduces tumorigenic and metastatic capacity of SW480 cells. Equal numbers (1.0 × 105 per mouse) of SW480 cells without lentiviral infection or with lentiviruses expressing control shRNA or EDA shRNA were injected subcutaneously into the both thighs of Balb/c nude mice. (a) The in vivo imaging of xenografts. Note: mice in the uninfected group showed no fluorescence due to the absence of lentiviral infection and thus normal instead of fluorescent images are included here. A photograph of a representative xenograft in each group is also shown. (b, c) The maximum diameters and weight of xenografts produced by SW480 cells without lentiviral infection or infected with lentiviruses expressing control shRNA or EDA shRNA. * P, b 0.05 (vs. control shRNA). (d) Immunohistochemistry of expression levels of EDA and a proliferation protein marker Ki67, and TUNEL assays for apoptosis in xenografts formed 6 weeks after injection of SW480 cells without lentiviral infection or infected with lentiviruses expressing control shRNA or EDA shRNA. (e) Immunofluorescence staining of CD133, CD44, OCT3/4, GLUT-1 and activated β-catenin in the same tumor xenografts. (f) Equal numbers (1.0 × 106 per mouse) of SW480 cells without lentiviral infection or with lentiviruses expressing control shRNA or EDA shRNA were injected via the tail vein into Balb/c nude mice. The gross appearance of livers and lungs 8 weeks post injection and H&E staining of liver and lung sections.

EDA-integrin α9β1 sustains CD133+/44+ colorectal cancer cell stages and chemoresistance of CRC patients (Fig. 1, Table 1). Advanced stage cancer cells may secrete increased amounts of EDA into the circulation. Alternatively, stromal tissues such as tumor-associated fibroblasts and blood/ lymphatic vessels in patients bearing advanced CRCs may synthesize and secrete more EDA in response to aggressiveness of tumors. Increased immunohistochemical staining of EDA in the tissues of advanced CRCs may reflect either increased expression of EDA by cancer cells and/or retention of EDA from other sources as a result of increased expression of EDA receptors in cancer cells. Although we didn't focus on how EDA is enriched in advanced human CRCs in this study, we believe future studies specifically addressing this issue would be of clinical importance. CRC, like other cancers, displays remarkable heterogeneity. Several subpopulations of CRC cells from human surgical specimens and cultured cell lines have been identified to display more significant proliferation and metastatic properties, including cells positive for CD133 and for CD44 (Dalerba et al., 2007; O'Brien et al., 2007; Ricci-Vitiani et al., 2007; Yeung et al., 2010). Although our study found no differential expression of EDA between CD133+/CD44+ and CD133−/CD44− cells sorted from SW480 CRC cells, we did see a significantly increased expression of an EDA receptor integrin α9β1 in CD133+/CD44+ cells (Figs. 2b–e), which may cause enrichment of both autocrine and paracrine EDA in CD133+/CD44+ cells via ligand-receptor interactions, thereby amplifying EDA signaling in this subset of cancer cells. This observation has an important application because it suggests that integrin α9β1 may be a potential cell surface marker for enrichment of CSCs and for prediction of disease progression in CRCs. Indeed, some integrins have been used for selection of normal neural stem/progenitor cells (α6 and β1 integrins) (Hall et al., 2006; Lathia et al., 2010) and cancer progenitor cells from breast (β3 integrin) (Vaillant et al., 2008) and prostate (α2and β1 integrins) (Patrawala et al., 2007). Future studies are required to determine if integrin α9β1-enriched CRC cells are true cancer stem/progenitor cells. In this study, CD133−/CD44− CRC cells are considered important in tumor progression because they secrete EDA and can sustain EDA-integrin α9β1 pathway in a paracrine manner. CD133+/CD44+ cells may take advantage of CD133−/ CD44− cells-secreted EDA via increasing expression of EDA receptor integrin α9β1. Interestingly, EDA silencing in SW480 cells changes relative abundance of CD133+/CD44+ and CD133−/CD44− subpopulations, indicating an EDA-dependent dynamic equilibrium between these two subsets of cells. A recent report found that CD133− relative to CD133+ human CRC cells are more resistant to 5-fluorouracil (FU) (Hongo et al., 2011). Collectively, these findings highlight the importance of cancer therapy targeting non-CSCs and tumor microenvironment in addition to CSCs. Silencing EDA strikingly reduces the formation of spheroids in SW480 cells, which is associated with reduced CD133+ and CD44+ subpopulations and expression of several embryonic stem cell markers including SOX2, NANOG, and OCT3/4 (Fig. 3). In contrast, the expression of differentiation-related marker CK7 and MUC2 are increased (Fig. 3). These observations demonstrated that EDA signaling is required for sustaining CD133+/CD44+ subpopulation and sphere formation of SW480 cells, and EDA may act through maintaining expression levels of stemness-associated proteins. We showed

831 that EDA interacts with its receptor integrin α9β1 in SW480 cells (Fig. 4a). When integrin α9β1 function is blocked by an monoclonal antibody, the colony formation of EDAoverexpressing CD133+ /CD44+ cells and the cell cycle progression of bulk SW480 cells are substantially suppressed, which is associated with increased apoptosis (Figs. 4b–f). These findings support a central role of integrin α9β1 in EDA signaling, at least, in SW480 CRC cells. It is well established that integrins alone or their interactions with growth factors and chemokines in a paracrine and autocrine manner promotes the evolution, progression and metastasis of tumors (Desgrosellier and Cheresh, 2010). Given this, integrins are considered as potential therapeutic targets in cancer (Cox et al., 2010). What might be signal transduction pathways downstream of EDA-integrin α9β1 signaling? Wnt signaling plays important roles in sustaining stem cells in normal tissues and cancers (Logan and Nusse, 2004) and in the development of CRC (Kinzler and Vogelstein, 1996; Sancho et al., 2004). We observed a substantial downregulation of active β-catenin protein in fibronectin EDA silenced SW480 cells, indicating suppression of Wnt signaling by EDA silencing in these cells. To the best of our knowledge, this study is the first to show that EDA is required for maintaining Wnt/β-catenin activity. Our observation raised an interesting question of how EDA is linked to the regulation of β-catenin. We found that the phosphorylation of FAK, Akt and Erk is down-regulated in the EDA-silenced SW480 cells (Fig. 5a). FAK is an immediate downstream target of integrins. Interestingly, silencing of FAK also reduces activation of β-catenin (Fig. 5b). FAK may exert effects through Akt and/or Erk pathways (Cabodi et al., 2010). Additionally, Akt was shown to link FAK to β-catenin activation in breast cancer progenitor cells exposed to the interleukin 8 receptor CXCR1, playing a crucial role in maintaining the properties of these cells (Ginestier et al., 2010). It was also reported that the Erk pathway can regulate β-catenin activation (Kim et al., 2007). Using specific inhibitors of MAPK/Erk and PI3K/Akt pathways, we showed that activation of β-catenin requires MAPK/Erk pathway, but not PI3K/Akt pathway in our cells (Fig. 5c). Thus, we propose that EDA may sustain CD133+/ CD44+ subpopulation in CRC cells via activating Wnt signaling through integrin α9β1-FAK-MAPK/ERK pathway (Fig. 5g). More than 20 years ago, Gradl et al. identified fibronectin as a direct target of Wnt/β-catenin pathway (Gradl et al., 1999). Together with our finding that fibronectin EDA sustains β-catenin activity, it appears there is a feed-forward regulatory loop between fibronectin secretion and β-catenin activity. This loop may represent a vicious cycle for promoting cell growth. Thus, targeting this cycle in cancer cells might be effective in controlling cancer growth. While the in vivo significance of our in vitro findings was examined in xenograft and metastasis studies, we observed a dramatic suppression of tumor growth and metastasis of SW480 CRC cells by EDA silencing, which is associated with increased apoptosis and reduced expression of a cell proliferation marker Ki67, progenitor cell/ embryonic stem cell markers CD133, CD44, OCT3/4 and GLUT-1, and active form of β-catenin (Fig. 6). This finding is important because it uncovered a crucial role of EDA in sustaining tumorigenic and metastatic capacity of CRC cells in animal models.

832

J. Ou et al.

Targeting the extra domain B (EDB) of fibronectin or using EDB for drug delivery has proved very successful in cancer treatment (Borsi et al., 2003; Huang et al., 1997; Kaspar et al., 2006; Santimaria et al., 2003; Sauer et al., 2009). EDA has not been selected as a target for cancer therapy. Our observations that EDA is enriched in advanced human CRCs and that there is an increased expression of an EDA receptor integrin α9β1 in CD133+/CD44+ subpopulation of CRC cells support a potential benefit of EDA/integrin α9β1-based targeting approach for inhibiting cancer progression and for limiting colon cancer progenitor cells.

Abbreviations

CRC EDA FAK shRNA

colorectal cancer extra domain A of fibronectin focal adhesion kinase short hairpin RNA

Acknowledgments This work was supported in part by grant number 81000965 from the National Natural Science Foundation of China (to J. O.); by 973 Program No. 2010cb529403 from the National Basic Research Program of China (to. J. O. & H. L.); and by Award Number R01DK085176 from the National Institute of Diabetes and Digestive and Kidney Diseases (to L.Y.). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institute of Diabetes and Digestive and Kidney Diseases or the National Institute of Health.

No conflicts of interest were declared. Author contributions The Authors have made the following declarations about their contributions: Conceived and designed the experiments: HJL LQY. Performed the experiments: JJO JD RBZ GFX. Analyzed the data: XW RBZ. Contributed reagents/ materials/analysis tools: JJO XW. Wrote the paper: HJL LQY.

References Al-Hajj, M., Wicha, M.S., Benito-Hernandez, A., Morrison, S.J., Clarke, M.F., 2003. Prospective identification of tumorigenic breast cancer cells. Proc. Natl. Acad. Sci. U. S. A. 100, 3983–3988. Allen, M., Louise Jones, J., 2010. Jekyll and Hyde: the role of the microenvironment on the progression of cancer. J. Pathol. 223, 162–176. Allen, M.D., Vaziri, R., Green, M., Chelala, C., Brentnall, A.R., Dreger, S., Vallath, S., Nitch-Smith, H., Hayward, J., Carpenter, R., et al., 2011. Clinical and functional significance of alpha9beta1 integrin expression in breast cancer: a novel cellsurface marker of the basal phenotype that promotes tumour cell invasion. J. Pathol. 223, 646–658. Avilion, A.A., Nicolis, S.K., Pevny, L.H., Perez, L., Vivian, N., LovellBadge, R., 2003. Multipotent cell lineages in early mouse development depend on SOX2 function. Genes Dev. 17, 126–140. Bonnet, D., Dick, J.E., 1997. Human acute myeloid leukemia is organized as a hierarchy that originates from a primitive hematopoietic cell. Nat. Med. 3, 730–737.

Borsi, L., Balza, E., Carnemolla, B., Sassi, F., Castellani, P., Berndt, A., Kosmehl, H., Biro, A., Siri, A., Orecchia, P., et al., 2003. Selective targeted delivery of TNFalpha to tumor blood vessels. Blood 102, 4384–4392. Cabodi, S., del Pilar Camacho-Leal, M., Di Stefano, P., Defilippi, P., 2010. Integrin signalling adaptors: not only figurants in the cancer story. Nat. Rev. Cancer 10, 858–870. Chaffer, C.L., Weinberg, R.A., 2011. A perspective on cancer cell metastasis. Science 331, 1559–1564. Chambers, I., Colby, D., Robertson, M., Nichols, J., Lee, S., Tweedie, S., Smith, A., 2003. Functional expression cloning of Nanog, a pluripotency sustaining factor in embryonic stem cells. Cell 113, 643–655. Chen, Y., Wang, Z., Chang, P., Xiang, L., Pan, F., Li, J., Jiang, J., Zou, L., Yang, L., Bian, Z., et al., 2010. The effect of focal adhesion kinase gene silencing on 5-fluorouracil chemosensitivity involves an Akt/NF-kappaB signaling pathway in colorectal carcinomas. Int. J. Cancer 127, 195–206. Choi, D., Lee, H.W., Hur, K.Y., Kim, J.J., Park, G.S., Jang, S.H., Song, Y.S., Jang, K.S., Paik, S.S., 2009. Cancer stem cell markers CD133 and CD24 correlate with invasiveness and differentiation in colorectal adenocarcinoma. World J. Gastroenterol. 15, 2258–2264. Clarke, M.F., Dick, J.E., Dirks, P.B., Eaves, C.J., Jamieson, C.H., Jones, D.L., Visvader, J., Weissman, I.L., Wahl, G.M., 2006. Cancer stem cells — perspectives on current status and future directions: AACR Workshop on cancer stem cells. Cancer Res. 66, 9339–9344. Cox, D., Brennan, M., Moran, N., 2010. Integrins as therapeutic targets: lessons and opportunities. Nat. Rev. Drug Discov. 9, 804–820. Dalerba, P., Dylla, S.J., Park, I.K., Liu, R., Wang, X., Cho, R.W., Hoey, T., Gurney, A., Huang, E.H., Simeone, D.M., et al., 2007. Phenotypic characterization of human colorectal cancer stem cells. Proc. Natl. Acad. Sci. U. S. A. 104, 10158–10163. Desgrosellier, J.S., Cheresh, D.A., 2010. Integrins in cancer: biological implications and therapeutic opportunities. Nat. Rev. Cancer 10, 9–22. ffrench-Constant, C., 1995. Alternative splicing of fibronectin — many different proteins but few different functions. Exp. Cell Res. 221, 261–271. Fu, M., Wang, C., Li, Z., Sakamaki, T., Pestell, R.G., 2004. Minireview: Cyclin D1: normal and abnormal functions. Endocrinology 145, 5439–5447. Ginestier, C., Liu, S., Diebel, M.E., Korkaya, H., Luo, M., Brown, M., Wicinski, J., Cabaud, O., Charafe-Jauffret, E., Birnbaum, D., et al., 2010. CXCR1 blockade selectively targets human breast cancer stem cells in vitro and in xenografts. J. Clin. Invest. 120, 485–497. Gradl, D., Kuhl, M., Wedlich, D., 1999. The Wnt/Wg signal transducer beta-catenin controls fibronectin expression. Mol. Cell. Biol. 19, 5576–5587. Gupta, P.B., Chaffer, C.L., Weinberg, R.A., 2009. Cancer stem cells: mirage or reality? Nat. Med. 15, 1010–1012. Hall, P.E., Lathia, J.D., Miller, N.G., Caldwell, M.A., ffrenchConstant, C., 2006. Integrins are markers of human neural stem cells. Stem Cells 24, 2078–2084. Hongo, K., Tanaka, J., Tsuno, N.H., Kawai, K., Nishikawa, T., Shuno, Y., Sasaki, K., Kaneko, M., Hiyoshi, M., Sunami, E., et al., 2011. CD133(−) Cells, Derived From a Single Human Colon Cancer Cell Line, Are More Resistant to 5-Fluorouracil (FU) Than CD133(+) Cells, Dependent on the beta1-Integrin Signaling. J. Surg. Res. 175, 278–288. Horst, D., Kriegl, L., Engel, J., Kirchner, T., Jung, A., 2009. Prognostic significance of the cancer stem cell markers CD133, CD44, and CD166 in colorectal cancer. Cancer Invest. 27, 844–850. Huang, X., Molema, G., King, S., Watkins, L., Edgington, T.S., Thorpe, P.E., 1997. Tumor infarction in mice by antibody-

EDA-integrin α9β1 sustains CD133+/44+ colorectal cancer cell directed targeting of tissue factor to tumor vasculature. Science 275, 547–550. Hynes, R.O., 2002. Integrins: bidirectional, allosteric signaling machines. Cell 110, 673–687. Jemal, A., Siegel, R., Ward, E., Hao, Y., Xu, J., Thun, M.J., 2009. Cancer statistics, 2009. CA Cancer J. Clin. 59, 225–249. Kaspar, M., Zardi, L., Neri, D., 2006. Fibronectin as target for tumor therapy. Int. J. Cancer 118, 1331–1339. Kelly, P.N., Dakic, A., Adams, J.M., Nutt, S.L., Strasser, A., 2007. Tumor growth need not be driven by rare cancer stem cells. Science 317, 337. Kenny, P.A., Lee, G.Y., Bissell, M.J., 2007. Targeting the tumor microenvironment. Front. Biosci. 12, 3468–3474. Kim, D., Rath, O., Kolch, W., Cho, K.H., 2007. A hidden oncogenic positive feedback loop caused by crosstalk between Wnt and ERK pathways. Oncogene 26, 4571–4579. Kinzler, K.W., Vogelstein, B., 1996. Lessons from hereditary colorectal cancer. Cell 87, 159–170. Lapidot, T., Sirard, C., Vormoor, J., Murdoch, B., Hoang, T., CaceresCortes, J., Minden, M., Paterson, B., Caligiuri, M.A., Dick, J.E., 1994. A cell progenitor human acute myeloid leukaemia after transplantation into SCID mice. Nature 367, 645–648. Lathia, J.D., Gallagher, J., Heddleston, J.M., Wang, J., Eyler, C.E., Macswords, J., Wu, Q., Vasanji, A., McLendon, R.E., Hjelmeland, A.B., et al., 2010. Integrin alpha 6 regulates glioblastoma stem cells. Cell Stem Cell 6, 421–432. Lee, S.H., Lee, Y.J., Han, H.J., 2011. Role of hypoxia-induced fibronectin-integrin beta1 expression in embryonic stem cell proliferation and migration: Involvement of PI3K/Akt and FAK. J. Cell. Physiol. 226, 484–493. Li, C.Y., Li, B.X., Liang, Y., Peng, R.Q., Ding, Y., Xu, D.Z., Zhang, X., Pan, Z.Z., Wan, D.S., Zeng, Y.X., et al., 2009a. Higher percentage of CD133 + cells is associated with poor prognosis in colon carcinoma patients with stage IIIB. J. Transl. Med. 7, 56. Li, Z., Bao, S., Wu, Q., Wang, H., Eyler, C., Sathornsumetee, S., Shi, Q., Cao, Y., Lathia, J., McLendon, R.E., et al., 2009b. Hypoxiainducible factors regulate tumorigenic capacity of glioma stem cells. Cancer Cell 15, 501–513. Logan, C.Y., Nusse, R., 2004. The Wnt signaling pathway in development and disease. Annu. Rev. Cell Dev. Biol. 20, 781–810. Miraglia, S., Godfrey, W., Yin, A.H., Atkins, K., Warnke, R., Holden, J.T., Bray, R.A., Waller, E.K., Buck, D.W., 1997. A novel fivetransmembrane hematopoietic stem cell antigen: isolation, characterization, and molecular cloning. Blood 90, 5013–5021. Nichols, J., Zevnik, B., Anastassiadis, K., Niwa, H., Klewe-Nebenius, D., Chambers, I., Scholer, H., Smith, A., 1998. Formation of pluripotent stem cells in the mammalian embryo depends on the POU transcription factor Oct4. Cell 95, 379–391. O'Brien, C.A., Pollett, A., Gallinger, S., Dick, J.E., 2007. A human colon cancer cell capable of progenitor tumour growth in immunodeficient mice. Nature 445, 106–110. Ohnishi, T., Hiraga, S., Izumoto, S., Matsumura, H., Kanemura, Y., Arita, N., Hayakawa, T., 1998. Role of fibronectin-stimulated tumor cell migration in glioma invasion in vivo: clinical significance of fibronectin and fibronectin receptor expressed in human glioma tissues. Clin. Exp. Metastasis 16, 729–741. Ou, J.J., Wu, F., Liang, H.J., 2010. Colorectal tumor derived fibronectin alternatively spliced EDA domain exserts lymphangiogenic effect on human lymphatic endothelial cells. Cancer Biol. Ther. 9, 186–191. Ou, J., Li, J., Pan, F., Xie, G., Zhou, Q., Huang, H., Liang, H., 2011. Endostatin suppresses colorectal tumor-induced lymphangiogenesis by inhibiting expression of fibronectin extra domain A and integrin alpha9. J. Cell. Biochem. 112, 2106–2114. Pang, R., Law, W.L., Chu, A.C., Poon, J.T., Lam, C.S., Chow, A.K., Ng, L., Cheung, L.W., Lan, X.R., Lan, H.Y., et al., 2010. A

833 subpopulation of CD26+ cancer stem cells with metastatic capacity in human colorectal cancer. Cell Stem Cell 6, 603–615. Parkin, D.M., Bray, F., Ferlay, J., Pisani, P., 2005. Global cancer statistics, 2002. CA Cancer J. Clin. 55, 74–108. Patrawala, L., Calhoun-Davis, T., Schneider-Broussard, R., Tang, D.G., 2007. Hierarchical organization of prostate cancer cells in xenograft tumors: the CD44 + alpha2beta1 + cell population is enriched in tumor-progenitor cells. Cancer Res. 67, 6796–6805. Quintana, E., Shackleton, M., Sabel, M.S., Fullen, D.R., Johnson, T.M., Morrison, S.J., 2008. Efficient tumour formation by single human melanoma cells. Nature 456, 593–598. Reya, T., Morrison, S.J., Clarke, M.F., Weissman, I.L., 2001. Stem cells, cancer, and cancer stem cells. Nature 414, 105–111. Ricci-Vitiani, L., Lombardi, D.G., Pilozzi, E., Biffoni, M., Todaro, M., Peschle, C., De Maria, R., 2007. Identification and expansion of human colon-cancer-progenitor cells. Nature 445, 111–115. Rybak, J.N., Roesli, C., Kaspar, M., Villa, A., Neri, D., 2007. The extra-domain A of fibronectin is a vascular marker of solid tumors and metastases. Cancer Res. 67, 10948–10957. Sancho, E., Batlle, E., Clevers, H., 2004. Signaling pathways in intestinal development and cancer. Annu. Rev. Cell Dev. Biol. 20, 695–723. Santimaria, M., Moscatelli, G., Viale, G.L., Giovannoni, L., Neri, G., Viti, F., Leprini, A., Borsi, L., Castellani, P., Zardi, L., et al., 2003. Immunoscintigraphic detection of the ED-B domain of fibronectin, a marker of angiogenesis, in patients with cancer. Clin. Cancer Res. 9, 571–579. Sauer, S., Erba, P.A., Petrini, M., Menrad, A., Giovannoni, L., Grana, C., Hirsch, B., Zardi, L., Paganelli, G., Mariani, G., et al., 2009. Expression of the oncofetal ED-B-containing fibronectin isoform in hematologic tumors enables ED-B-targeted 131IL19SIP radioimmunotherapy in Hodgkin lymphoma patients. Blood 113, 2265–2274. Shin, J.H., Bae, J.H., Lee, A., Jung, C.K., Yim, H.W., Park, J.S., Lee, K.Y., 2010. CK7, CK20, CDX2 and MUC2 Immunohistochemical staining used to distinguish metastatic colorectal carcinoma involving ovary from primary ovarian mucinous adenocarcinoma. Jpn. J. Clin. Oncol. 40, 208–213. Shinde, A.V., Bystroff, C., Wang, C., Vogelezang, M.G., Vincent, P.A., Hynes, R.O., Van De Water, L., 2008. Identification of the peptide sequences within the EIIIA (EDA) segment of fibronectin that mediate integrin alpha9beta1-dependent cellular activities. J. Biol. Chem. 283, 2858–2870. Singh, S.K., Clarke, I.D., Terasaki, M., Bonn, V.E., Hawkins, C., Squire, J., Dirks, P.B., 2003. Identification of a cancer stem cell in human brain tumors. Cancer Res. 63, 5821–5828. Singh, S.K., Hawkins, C., Clarke, I.D., Squire, J.A., Bayani, J., Hide, T., Henkelman, R.M., Cusimano, M.D., Dirks, P.B., 2004. Identification of human brain tumour progenitor cells. Nature 432, 396–401. Takahashi, K., Tanabe, K., Ohnuki, M., Narita, M., Ichisaka, T., Tomoda, K., Yamanaka, S., 2007. Induction of pluripotent stem cells from adult human fibroblasts by defined factors. Cell 131, 861–872. Vaillant, F., Asselin-Labat, M.L., Shackleton, M., Forrest, N.C., Lindeman, G.J., Visvader, J.E., 2008. The mammary progenitor marker CD61/beta3 integrin identifies cancer stem cells in mouse models of mammary tumorigenesis. Cancer Res. 68, 7711–7717. Wang, J., Rao, S., Chu, J., Shen, X., Levasseur, D.N., Theunissen, T.W., Orkin, S.H., 2006. A protein interaction network for pluripotency of embryonic stem cells. Nature 444, 364–368. Wang, Q., Chen, Z.G., Du, C.Z., Wang, H.W., Yan, L., Gu, J., 2009. Cancer stem cell marker CD133 + tumour cells and clinical outcome in rectal cancer. Histopathology 55, 284–293. Yeung, T.M., Gandhi, S.C., Wilding, J.L., Muschel, R., Bodmer, W.F., 2010. Cancer stem cells from colorectal cancer-derived cell lines. Proc. Natl. Acad. Sci. U. S. A. 107, 3722–3727.