Forkhead box Q1 (FOXQ1) is a member of the forkhead transcription factor family, and it has recently been proposed to participate in gastric acid secretion and mucin gene expression in mice. However, the role of FOXQ1 in humans and especially in cancer cells remains unknown. We found that FOXQ1 mRNA is overexpressed in clinical specimens of colorectal cancer (CRC; 28-fold/colonic mucosa). A microarray analysis revealed that the knockdown of FOXQ1 using small interfering RNA resulted in a decrease in p21CIP1/WAF1 expression, and a reporter assay and a chromatin immunoprecipitation assay showed that p21 was one of the target genes of FOXQ1. Stable FOXQ1-overexpressing cells (H1299/FOXQ1) exhibited elevated levels of p21 expression and inhibition of apoptosis induced by doxorubicin or camptothecin. Although cellular proliferation was decreased in H1299/FOXQ1 cells in vitro, H1299/FOXQ1 cells significantly increased tumorigenicity [enhanced green fluorescent protein (EGFP): 2/15, FOXQ1: 7/15] and enhanced tumor growth (437 ± 301 versus 1735 ± 769 mm3, P < 0.001) in vivo. Meanwhile, stable p21 knockdown of H1299/FOXQ1 cells increased tumor growth, suggesting that FOXQ1 promotes tumor growth independent of p21. Microarray analysis of H1299/EGFP and H1299/FOXQ1 revealed that FOXQ1 overexpression upregulated several genes that have positive roles for tumor growth, including VEGFA, WNT3A, RSPO2, and BCL11A. CD31 and terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling staining of the tumor specimens showed that FOXQ1 overexpression mediated the angiogenic and antiapoptotic effect in vivo. In conclusion, FOXQ1 is overexpressed in CRC and enhances tumorigenicity and tumor growth presumably through its angiogenic and antiapoptotic effects. Our findings show that FOXQ1 is a new member of the cancer-related FOX family. Cancer Res; 70(5); 2053–63

The forkhead box (Fox) gene family is a large and diverse group of transcription factors that share certain characteristics of a conserved, ∼100 amino acid DNA-binding motif known as the forkhead or winged helix domain; over 100 proteins with forkhead domains have been identified, comprising at least 17 subclasses to date (1). The Fox gene family plays various important roles, not only in biological processes including development, metabolism, immunology, and senescence but also in cancer development (2, 3).

Forkhead box Q1 (FOXQ1, also known as HFH1) is a member of the FOX gene family and contains the core DNA binding domain, whereas the flanking wings of FOXQ1 contribute to its sequence specificity (4). As a transcription factor, FOXQ1 is known to repress the promoter activity of smooth muscle–specific genes, such as telokin and SM22α, in A10 vascular muscle cells (5), and FOXQ1 expression is regulated by Hoxa1 in embryonic stem cells (6). The biological function of Foxq1 has been clearly identified in hair follicle differentiation in satin (sa) homozygous mice (7); interestingly, satin mice also exhibit suppressed natural killer cell function and T-cell function, suggesting a relation with immunology. Satin mice have provided evidence that Hoxc13 regulates foxq1 expression and that “cross-talk” occurs between Homeobox and Fox (8). Foxq1 mRNA is widely expressed in murine tissues, with particularly high expression levels in the stomach and bladder (5). Recently, two important findings have been reported regarding its involvement in stomach surface cells. Foxq1-deficient mice exhibit a lack of gastric acid secretion in response to various secretagogue stimuli (9). On the other hand, Foxq1 regulates gastric MUC5AC synthesis, providing clues as to the lineage-specific cell differentiation in gastric surface epithelia (10). Despite accumulating evidence supporting the biological function of the murine foxq1 gene in hair follicle morphogenesis and gastric epithelial cells, no data regarding the cellular and biological functions of human FOXQ1, especially in cancer cells, are available.

p21CIP1/WAF1 (hereafter called p21) is a member of the cip/kip family of cyclin kinase inhibitors, and initial reports have shown that p21 functions as a G1 cyclin kinase inhibitor (11, 12) and a downstream molecule of p53 (13). p21 possesses a variety of cellular functions, including the negative modulation of cell cycle progression (14), cellular differentiation (15), and the regulation of p53-dependent antiapoptosis (reviewed in ref. 16). The expression of p21 is regulated by both p53-dependent and p53-independent mechanisms at the transcriptional level. Other regulatory mechanisms of p21 expression involve proteasome-mediated degradation, mRNA stability, alterations in the epigenetic silencing of the p21 promoter, and secondary decreases resulting from viral activity targeting p53, such as the activities of human papilloma virus and hepatitis C virus (17). However, its expression is considered to be regulated mainly at the transcriptional level (18). Accumulating data indicate that many molecules from diverse signaling pathways can activate or repress the p21 promoter, including p53, transforming growth factor-β (TGF-β), c-jun, Myc, Sp1/Sp3, signal transducers and activators of transcriptions, CAAT/enhancer binding protein-α (C/EBP-α), C/EBP-β, basic helix-loop-helix proteins, and myogenic differentiation 1 (reviewed in ref. 19). Thus, p21 is integrally involved in both cell cycle and apoptosis; therefore, identifying its regulatory molecules is of great importance.

We performed a microarray analysis of clinical samples of paired colorectal cancer (CRC) specimens and normal colonic mucosa specimens to identify genes that were overexpressed in CRC. Our results revealed that FOXQ1 gene expression was ∼28-fold higher in CRC than in normal colonic mucosa, and we hypothesized that FOXQ1 may play a role in CRC. In the present study, we investigated the biological function of FOXQ1.

Antibodies

The following antibodies were used: anti-p21, anti-p53, anti-cdk2, anti-cdk4, anti-cyclin D, anti-phosphorylated Rb, anti-poly(ADP-ribose) polymerase (PARP), anti-cleaved PARP, anti-caspase-3, anti-cleaved caspase-3, secondary antibodies, and Myc-tag mouse antibody (Cell Signaling), as well as anti-β-actin (Santa Cruz Biotechnology). A mouse anti-CD31 monoclonal antibody was purchased from BD Biosciences.

Cell lines and cultures

The DLD-1, MKN74, H1299, SBC3, and U251 cell lines were cultured in RPMI 1640 (Sigma). The WiDr, CoLo320DM, and human embryonic kidney cell line 293 (HEK293) cell lines were cultured in DMEM (Sigma), and the LoVo cell line was cultured in Ham/F12 medium [Life Technologies Bethesda Research Laboratories (BRL)]. All media were supplemented with 10% heat-inactivated fetal bovine serum (Life Technologies BRL), and the cell lines were maintained in a 5% CO2-humidified atmosphere at 37°C.

Patients and samples

Paired CRC and noncancerous colonic mucosa samples were evaluated using a microarray analysis in the first consecutive 10 patients. These samples and another 36 CRC samples were analyzed using real-time reverse transcription–PCR (RT-PCR). The RNA extraction method and the quality check protocol have been previously described (20). This study was approved by the institutional review board of the National Cancer Center Hospital, and written informed consent was obtained from all the patients.

Plasmid construction, viral production, and stable transfectants

The cDNA fragment encoding human full-length FOXQ1 was isolated using PCR and Prime STAR HS DNA polymerase (TaKaRa) with 5′-GGG AAT TCG CGG CCA TGA AGT TGG AGG TCT TCG TC-3′ and 5′-CCC TCG AGC GCT ACT CAG GCT AGG AGC GTC TCC AC-3′ sense and antisense primers, respectively. The methods used in this section have been previously described (21). Short hairpin RNA (shRNA) targeting p21 was constructed using oligonucleotides encoding small interfering RNA (siRNA) directed against p21 and a nonspecific target as follows: 5′-CTA AGA GTG CTG GGC ATT TTT-3′ for p21 shRNA and 5′-TGT TCG CAG TAC GGT AAT GTT-3′ for control shRNA. They were cloned into an RNAi-Ready pSIREN-RetroQZsGreen vector (Clontech) according to manufacturer's protocol. The stable transfectants expressing enhanced green fluorescent protein (EGFP) or FOXQ1 or FOXQ1 with shRNA targeting p21 for each cell line were designated as HEK293/EGFP, HEK293/FOXQ1, CoLo320/EGFP, CoLo320/FOXQ1, H1299/EGFP, H1299/FOXQ1, H1299/FOXQ1/sh-control, and H1299/FOXQ1/sh-p21. The FOXQ1 human cDNA was tagged at the NH2 terminus with the myc epitope using the pCMV-Myc vector (Clontech) for chromatin immunoprecipitation (ChIP) assay.

SiRNA transfection

Two different sequences of siRNA targeting human FOXQ1 and negative control siRNA were purchased from QIAGEN. The sequences of FOXQ1 and control siRNA were as follows: FOXQ1#1 sense, 5′-CCA UCA AAC GUG CCU UAA A-3′ and antisense, 5′-UUU AAG GCA CGU UUG AUG G-3′; FOXQ1#4 sense, 5′-CGC GGA CUU UGC ACU UUG A-3′ and antisense, 5′-UCA AAG UGC AAA GUC CGC G-3′; control siRNA (scramble) sense, 5′-UUC UCC GAA CGU GUC ACG U-3′ and antisense, 5′-ACG UGA CAC GUU CGG AGA A-3′; control siRNA (GFP) sense, 5′-GCA AGC UGA CCC UGA AGU UCA U-3′ and antisense, 5′-GAA CUU CAG GGU CAG CUU GCC G-3′. The methods of transfection have been previously described (22).

Real-time RT-PCR and Western blot analysis

The methods used in this section have been previously described (21). The primers used for real-time RT-PCR were purchased from Takara as follows: FOXQ1 forward, 5′-CGC GGA CTT TGC ACT TTG AA-3′ and reverse, 5′-AGC TTT AAG GCA CGT TTG ATG GAG-3′; p21 forward, 5′-TCC AGC GAC CTT CCT CAT CCA C-3′ and reverse, 5′-TCC ATA GCC TCT ACT GCC ACC ATC-3′; glyceraldehyde-3-phosphate dehydrogenase (GAPD) forward, 5′-GCA CCG TCA AGG CTG AGA AC-3′ and reverse, 5′-ATG GTG GTG AAG ACG CCA GT-3′. The experiment was performed in triplicate.

Luciferase reporter assay

The human p21 promoter reporter vector was constructed according to a previously described method (13). The p21 promoter fragment was cut between the KpnI and XhoI restriction sites and was transferred into the luciferase reporter vector pGL4.14 (Promega). All sequences were verified using DNA sequencing. The empty and p21 promoter-containing reporter vectors were designated as pGL4.14-mock and pGL4.14-p21, respectively. All the samples were examined in triplicate.

ChIP

ChIP was carried out using the ChIP-IT Express Enzymatic kit (Active Motif) according to manufacturer's protocol. HEK293 cells were transfected with empty vector (Myc) or Myc-tagged FOXQ1 vector. The putative region of the p21 promoter (−2264 to −1971) was amplified with the following primers: 5′-TTG AGC TCT GGC ATA GAA GA-3′ (forward) and 5′-TAC CCA GAC ACA CTC TAA GG-3′ (reverse). As a negative control, the glyceraldehyde-3-phosphate dehydrogenase (GAPDH) second intron promoter was amplified with the following primers: 5′-AAT GAA TGG GCA GCC GTT AG-3′ (forward) and 5′-AGC TAG CCT CGC TCC ACCTGA C-3′ (reverse).

Xenograft studies

Two separate xenograft studies were performed independently. Nude mice (BALB/c nu/nu; 6-week-old females; CLEA Japan, Inc.) were used for the in vivo studies and were cared for in accordance with the recommendations for the Handling of Laboratory Animals for Biomedical Research compiled by the Committee on Safety and Ethical Handling Regulations for Laboratory Animals Experiments, Kinki University. The ethical procedures followed and met the requirements of the United Kingdom Coordinating Committee on Cancer Research guidelines (23). To assess tumorigenicity, suspensions of 1 × 106 H1299/EGFP or H1299/FOXQ1 cells (in 0.1 mL PBS) were s.c. injected into the left or right flanks of nude mice (n = 15), respectively. To evaluate tumor growth, a suspension of 6 × 106 H1299/EGFP, H1299/FOXQ1, H1299/FOXQ1/sh-control, and H1299/FOXQ1/sh-p21 cells (in 0.1 mL PBS) were s.c. inoculated (n = 10) into nude mice. The tumor volume was calculated as length × width2 × 0.5. The tumor formation was assessed every 2 to 3 d. At the end of the experiment, the mice were sacrificed and the xenografts were resected, fixed in 10% buffered formalin for 6 to 10 h, and processed for histologic analysis.

Immunohistochemical and immunofluorescence staining

The methods used in this section have been previously described (24, 25).

Microarray analysis

The microarray procedure and analysis were performed according to the Affymetrix protocols and BRB Array Tools software, Ver. 3.3.0,4

developed by Dr. Richard Simon and Dr. Amy Peng, as reported previously (21, 26).

Statistical analysis

The statistical analyses were performed using Microsoft Excel (Microsoft) to calculate the SD and to test for statistically significant differences between the samples using a Student's t test. A P value of <0.05 was considered statistically significant.

FOXQ1 mRNA was overexpressed in CRCs

A microarray analysis for 10 paired CRC samples identified 30 genes as being significantly upregulated by >10-fold in CRC (P < 0.001; Supplementary Table S1). FOXQ1, an uncharacterized transcription factor, was upregulated by 28-fold in the CRC specimens (Fig. 1A), exhibiting the fourth highest level of upregulation [after interleukin-8, matrix metalloproteinase-1 (MMP), and MMP-3]. Real-time RT-PCR for the 10 paired samples and an additional 36 CRC samples showed that FOXQ1 mRNA was markedly overexpressed in the CRC samples but was only expressed at a very low level in noncancerous colonic mucosa (P < 0.001; Fig. 1B). The average levels of FOXQ1 expression were 299 ± 326 and 4.0 ± 5.0 (×104/GAPD), respectively.

Figure 1.

FOXQ1 expression in CRC. A, mRNA expression of FOXQ1 obtained from a microarray analysis of 10 CRC and paired normal mucosa specimens. The values indicate the normalized signal intensity. B, the mRNA expression levels of FOXQ1 were determined using real-time RT-PCR for 10 paired and an additional 36 CRC samples. C, the mRNA expression levels of FOXQ1 were determined using a real-time RT-PCR analysis of human normal tissue (left) and 30 human cancer cell lines, HEK293, and human umbilical vascular endothelial cell (HUVEC) cell lines (right). GC, gastric cancer; LC, lung cancer; Rel mRNA, normalized mRNA expression levels (FOXQ1/GAPD × 104).

Figure 1.

FOXQ1 expression in CRC. A, mRNA expression of FOXQ1 obtained from a microarray analysis of 10 CRC and paired normal mucosa specimens. The values indicate the normalized signal intensity. B, the mRNA expression levels of FOXQ1 were determined using real-time RT-PCR for 10 paired and an additional 36 CRC samples. C, the mRNA expression levels of FOXQ1 were determined using a real-time RT-PCR analysis of human normal tissue (left) and 30 human cancer cell lines, HEK293, and human umbilical vascular endothelial cell (HUVEC) cell lines (right). GC, gastric cancer; LC, lung cancer; Rel mRNA, normalized mRNA expression levels (FOXQ1/GAPD × 104).

Close modal

FOXQ1 expression in normal tissues and cancer cell lines

To investigate the expression of FOXQ1, we analyzed the mRNA expression levels of FOXQ1 in panels of human normal tissues and cancer cell lines using real-time RT-PCR. High levels of FOXQ1 expression were observed in the stomach, salivary gland, prostate, trachea, and fetal liver among the 24 normal tissues that were examined (Fig. 1C, left). Relatively weak expression levels were detected in brain-derived tissues, kidney, lung, placenta, and thyroid gland. These results were consistent with those of a previous report (27).

In the cancer cell line panel, the mRNA expression levels of FOXQ1 were higher in gastric cancer, CRC, and lung cancer cell lines than in the other cancer cell lines, indicating that the expression of FOXQ1 varies among specific cancers (Fig. 1C, right). Interestingly, the overexpression of FOXQ1 in CRC arose from normal colonic mucosa with very low expression levels during carcinogenesis.

p21 is a target gene of FOXQ1

To examine the function of FOXQ1 as a transcription factor and to explore its target genes, we performed a microarray analysis using a CRC cell line, DLD-1, transfected with FOXQ1-targeting siRNA or control siRNA. Two sequences of FOXQ1-siRNA, FQ#1 and FQ#4, were used to exclude the off-target effect of siRNA. Real-time RT-PCR showed that both sequences of FOXQ1-siRNA suppressed FOXQ1 mRNA expression by ∼80% in DLD-1 cells (Fig. 2A); thus, FQ#4 was used as the FOXQ1-siRNA in the following experiments. A microarray analysis showed that 19 genes were downregulated by FOXQ1-siRNA (Fig. 2B; Supplementary Table S2); p21 was the fifth most-downregulated gene. Because p21 is a key regulator of cell cycle and apoptosis, we focused on p21 as a target molecule of FOXQ1.

Figure 2.

FOXQ directly regulates p21 transcription. A, FOXQ1-targeting siRNA (FQ#1 and FQ#4) suppressed FOXQ1 expression in DLD-1 cells. The mRNA expression levels of FOXQ1 were determined using real-time RT-PCR. B, microarray analysis of DLD-1 cells transfected with control-siRNA or FOXQ1-siRNA. The longitudinal axis indicates the mRNA expression of FOXQ1-siRNA transfected cells and the horizontal axis indicates that of control-siRNA. Arrow, FOXQ1 or p21 expression. Each point indicates the normalized and log base 2 transformed microarray data. C, induction of p21 promoter activity by FOXQ1. Luciferase vectors with either an empty or p21 promoter (pGL4.14-mock or pGL4.14-p21) were transiently cotransfected with a mock or FOXQ1 expression plasmid (pcDNA3.1-mock or pcDNA3.1-FOXQ1) expressing β-galactosidase as an internal control. The results were normalized to β-galactosidase activity and are representative of at least three independent experiments. D, ChIP of FOXQ1 on the promoter of p21. HEK293 cells were transfected with empty vector (Myc) or Myc-tagged FOXQ1 vector. Agarose gel shows PCR amplification (35 cycles) of the p21 promoter using inputs (1% of chromatin used for ChIP) or ChIPs as templates. Primers to the GAPDH promoter were used as the negative control.

Figure 2.

FOXQ directly regulates p21 transcription. A, FOXQ1-targeting siRNA (FQ#1 and FQ#4) suppressed FOXQ1 expression in DLD-1 cells. The mRNA expression levels of FOXQ1 were determined using real-time RT-PCR. B, microarray analysis of DLD-1 cells transfected with control-siRNA or FOXQ1-siRNA. The longitudinal axis indicates the mRNA expression of FOXQ1-siRNA transfected cells and the horizontal axis indicates that of control-siRNA. Arrow, FOXQ1 or p21 expression. Each point indicates the normalized and log base 2 transformed microarray data. C, induction of p21 promoter activity by FOXQ1. Luciferase vectors with either an empty or p21 promoter (pGL4.14-mock or pGL4.14-p21) were transiently cotransfected with a mock or FOXQ1 expression plasmid (pcDNA3.1-mock or pcDNA3.1-FOXQ1) expressing β-galactosidase as an internal control. The results were normalized to β-galactosidase activity and are representative of at least three independent experiments. D, ChIP of FOXQ1 on the promoter of p21. HEK293 cells were transfected with empty vector (Myc) or Myc-tagged FOXQ1 vector. Agarose gel shows PCR amplification (35 cycles) of the p21 promoter using inputs (1% of chromatin used for ChIP) or ChIPs as templates. Primers to the GAPDH promoter were used as the negative control.

Close modal

To confirm the microarray data, p21 downregulation by FOXQ1-siRNA was examined using real-time RT-PCR and a Western blot analysis in DLD-1 cells. The results indicated that both sequences of FOXQ1-siRNA (FQ#1 and FQ#4) downregulated p21 expression at both the mRNA and protein levels. In addition, we confirmed the downregulation of p21 by FOXQ1-siRNA in other cell lines (WiDr and HEK293), obtaining similar results (Supplementary Fig. S1).

FOXQ1 directly increases the transcription activity of p21

We performed a luciferase reporter assay to determine whether FOXQ1 regulates p21 expression at the transcriptional level. A 2.4-kb section of the p21 promoter region was subcloned into a luciferase vector according to a previously described method (13, 28). The p21 promoter activity was increased by >8-fold when cotransfected with a FOXQ1 expression vector, compared with an empty vector (Fig. 2C). To determine whether FOXQ1 directly binds to p21 promoter, we transfected Myc or Myc-tagged FOXQ1 vectors into HEK293 cells and then conducted ChIP experiments. A segment of the p21 promoter containing putative FOXQ1 binding site (−2264 to −1971) is precipitated with specific antibody, only if, FOXQ1 was induced (Fig. 2D). The result indicates that FOXQ1 binds to the p21 promoter and upregulates p21 transcriptional activity.

p53-independent p21 induction by FOXQ1 in cancer cells

Because p53 is the most important regulatory molecule of p21, we examined the downregulation of p21 by FOXQ1-siRNA in several cell lines with p53–wild type, p53-mutant, or p53-null statuses. These cell lines were transfected with control-siRNA or FOXQ1-siRNA, and p21 induction was enhanced by doxorubicin (2931). The experiments were performed using three p53–wild type cell lines, three p53-mutation cell lines, and one p53-null cell line (Fig. 3A–C). Without doxorubicin exposure, all seven cell lines showed that p21 expression was downregulated by FOXQ1-siRNA. Notably, with doxorubicin exposure, considerable p21 downregulation by FOXQ1-siRNA was observed in the p53-mutation and p53-null cell lines, compared with in the p53–wild type cell lines. In the p53-null H1299 cell line, FOXQ1-siRNA completely suppressed p21 expression. These results suggest that p21 induction by FOXQ1 is p53 independent. An immunofluorescence study of p21 in H1299 cells also showed that p21 was completely downregulated by FOXQ1-siRNA (Fig. 3D).

Figure 3.

p21 induction by FOXQ1 and p53 status in cancer cells. The seven cell lines were transfected with control-siRNA or FOXQ1-siRNA for 24 h, and the cells were exposed to doxorubicin at a final concentration of 0.5 or 1 μmol/L for a further 24 h to enhance p21 induction. Western blot analyses for p21 and p53 were performed in three p53–wild type cell lines (A), three p53-mutant cell lines (B), and one p53-null cell line (C). The experiment was performed in duplicate. D, immunofluorescence p21 staining and 4′,6-diamidino-2-phenylindole (DAPI) staining for H1299 cells transfected with control-siRNA (top) or FOXQ1-siRNA (bottom) for 48 h. Scr, scramble-siRNA (control); FQ#4, FOXQ1-targeting siRNA. β-Actin was used as an internal control.

Figure 3.

p21 induction by FOXQ1 and p53 status in cancer cells. The seven cell lines were transfected with control-siRNA or FOXQ1-siRNA for 24 h, and the cells were exposed to doxorubicin at a final concentration of 0.5 or 1 μmol/L for a further 24 h to enhance p21 induction. Western blot analyses for p21 and p53 were performed in three p53–wild type cell lines (A), three p53-mutant cell lines (B), and one p53-null cell line (C). The experiment was performed in duplicate. D, immunofluorescence p21 staining and 4′,6-diamidino-2-phenylindole (DAPI) staining for H1299 cells transfected with control-siRNA (top) or FOXQ1-siRNA (bottom) for 48 h. Scr, scramble-siRNA (control); FQ#4, FOXQ1-targeting siRNA. β-Actin was used as an internal control.

Close modal

Overexpression of FOXQ1 increases p21 expression and exhibits an antiapoptotic effect in cancer cells

Next, we established a stable FOXQ1-overexpressing cell line to confirm the induction of p21 expression by FOXQ1 and to detect any changes in the cellular phenotype of the cancer cells. FOXQ1 overexpression induced p21 expression (both mRNA and protein) in HEK293 and CoLo320 cells (Supplementary Fig. S1). Notably, p21 protein expression was markedly induced by >10-fold in the H1299/FOXQ1 cells (Supplementary Fig. S1). These results indicated that FOXQ1 robustly induces p21 expression, consistent with the findings of the siRNA study.

p21 induces an antiapoptotic effect and exerts a protective role against apoptosis induced by DNA damage. To elucidate the role of apoptosis induced by FOXQ1 in cancer cells, we examined the apoptotic effect in H1299/EGFP and H1299/FOXQ1 cells using anticancer drugs. The overexpression of FOXQ1 inhibited the apoptosis induced by doxorubicin (H1299/EGFP: 7.9 ± 1.9%, H1299/FOXQ1: 2.7 ± 0.7%; Fig. 4A). Similarly, camptothecin-induced apoptosis was also inhibited in FOXQ1-overexpressing cells (H1299/EGFP: 7.4 ± 2.1%, H1299/FOXQ1: 2.5 ± 1.0%; Fig. 4B). Western blotting revealed that FOXQ1 overexpression decreased the levels of cleaved caspase-3 and cleaved PARP induced by doxorubicin (Fig. 4C). These results are consistent with those obtained using flow cytometry.

Figure 4.

Overexpression of FOXQ1 promotes an antiapoptotic effect. Stable H1299 cell lines expressing EGFP or FOXQ1 (H1299/EGFP, H1299/FOXQ1) were exposed to doxorubicin (A) or camptothecin (B) at a final concentration of 3 μmol/L. Apoptotic cells were detected by Annexin V and propidium iodide (PI) using flow cytometry. C, Western blot analysis for apoptosis-related molecules. EGFP- or FOXQ1-expressing cells were exposed to doxorubicin at the indicated doses (0–5 μmol/L) for 24 h. β-Actin was used as an internal control. Dox, doxorubicin; CPT, camptothecin; EGFP, H1299/EGFP; FOXQ1, H1299/FOXQ1. *, P < 0.05.

Figure 4.

Overexpression of FOXQ1 promotes an antiapoptotic effect. Stable H1299 cell lines expressing EGFP or FOXQ1 (H1299/EGFP, H1299/FOXQ1) were exposed to doxorubicin (A) or camptothecin (B) at a final concentration of 3 μmol/L. Apoptotic cells were detected by Annexin V and propidium iodide (PI) using flow cytometry. C, Western blot analysis for apoptosis-related molecules. EGFP- or FOXQ1-expressing cells were exposed to doxorubicin at the indicated doses (0–5 μmol/L) for 24 h. β-Actin was used as an internal control. Dox, doxorubicin; CPT, camptothecin; EGFP, H1299/EGFP; FOXQ1, H1299/FOXQ1. *, P < 0.05.

Close modal

Overexpression of FOXQ1 decreases cellular proliferation but enhances tumorigenicity and tumor growth in vivo

Stable H1299/FOXQ1 cells showed decreased cellular proliferation compared with control cells in vitro (Fig. 5A). Expressions of Cdk4, cyclin D1, and Cdk2 were decreased by FOXQ1 expression in H1299/FOXQ1 cells and resulted in a decrease of phosphorylated Rb expression (Fig. 5A). To examine the biological functions of FOXQ1 overexpression in vivo, we evaluated tumorigenicity and tumor growth using H1299/EGFP or H1299/FOXQ1 cells. H1299/FOXQ1 cells exhibited a significantly elevated level of tumorigenesis in vivo (GFP 2/15, FOXQ1 7/15, P < 0.05; Fig. 5B). In addition, the tumor volume was markedly larger in H1299/FOXQ1 cells than in H1299/EGFP cells (EGFP: 437 ± 301, FOXQ1: 1735 ± 769 mm3, P < 0.001; Fig. 5B) on day 25.

Figure 5.

Overexpression of FOXQ1 enhances tumorigenicity and tumor growth in vivo. A, cellular growth and immunoblotting analysis of H1299 cell lines stably expressing EGFP or FOXQ1 (H1299/EGFP, H1299/FOXQ1). A total of 2 × 103 cells of each cell line were seeded in 96-well plates and evaluated after 0, 24, 48, and 72 h using MTT assay. Error bars, SD. Protein levels of H1299/EGFP and H1299/FOXQ1 cells were examined by Western blotting using specific antibody to p21, Cdk2, Cdk4, cyclin D, and phosphorylated Rb (pRb) protein. β-Actin was used as an internal control. EGFP, stable EGFP-overexpressing cells; FOXQ1, stable FOXQ1-overexpressing cells. B, H1299/EGFP and H1299/FOXQ1 cells were evaluated for their tumorigenicity in vivo. Mice (n = 15) were s.c. inoculated with a total of 1 × 106 cells. The numerical data indicate the number of mice. A total of 6 × 106 H1299/EGFP or H1299/FOXQ1 cells were s.c. inoculated into the right flank of each mouse to evaluate the tumor growth in vivo (n = 12). Representative H&E staining of tumor specimens was also shown. C, stable p21 knockdown or control cells obtained from H1299/FOXQ1 cells (H1299/FOXQ1/sh-control and H1299/FOXQ1/sh-p21) were evaluated for cellular growth and immunoblotting analysis. D, a total of 6 × 106 H1299/FOXQ1/sh-control or H1299/FOXQ1/sh-p21 cells were s.c. inoculated into the right flank of each mouse to evaluate the tumor growth (n = 10). *, P < 0.05.

Figure 5.

Overexpression of FOXQ1 enhances tumorigenicity and tumor growth in vivo. A, cellular growth and immunoblotting analysis of H1299 cell lines stably expressing EGFP or FOXQ1 (H1299/EGFP, H1299/FOXQ1). A total of 2 × 103 cells of each cell line were seeded in 96-well plates and evaluated after 0, 24, 48, and 72 h using MTT assay. Error bars, SD. Protein levels of H1299/EGFP and H1299/FOXQ1 cells were examined by Western blotting using specific antibody to p21, Cdk2, Cdk4, cyclin D, and phosphorylated Rb (pRb) protein. β-Actin was used as an internal control. EGFP, stable EGFP-overexpressing cells; FOXQ1, stable FOXQ1-overexpressing cells. B, H1299/EGFP and H1299/FOXQ1 cells were evaluated for their tumorigenicity in vivo. Mice (n = 15) were s.c. inoculated with a total of 1 × 106 cells. The numerical data indicate the number of mice. A total of 6 × 106 H1299/EGFP or H1299/FOXQ1 cells were s.c. inoculated into the right flank of each mouse to evaluate the tumor growth in vivo (n = 12). Representative H&E staining of tumor specimens was also shown. C, stable p21 knockdown or control cells obtained from H1299/FOXQ1 cells (H1299/FOXQ1/sh-control and H1299/FOXQ1/sh-p21) were evaluated for cellular growth and immunoblotting analysis. D, a total of 6 × 106 H1299/FOXQ1/sh-control or H1299/FOXQ1/sh-p21 cells were s.c. inoculated into the right flank of each mouse to evaluate the tumor growth (n = 10). *, P < 0.05.

Close modal

p21 does not contribute to FOXQ1-mediated tumor growth in vivo

Because emerging evidence has indicated that p21 may have dual functions with regard to tumor progression and the suppression of cancer cells (32, 33), the shRNA targeting p21 or shRNA control viral vectors were further introduced into the H1299/FOXQ1 cells to elucidate the involvement of p21 in increased FOXQ1-mediated tumorigenicity and tumor growth in vivo. Stable H1299/FOXQ1/sh-p21 cells were slightly increased in cellular proliferation in vitro (Fig. 5C). In addition, tumor growth of H1299/FOXQ1/sh-p21 cells was increased compared with control cells in vivo (Fig. 5D). The results clearly indicate that p21 has negative roles for cellular proliferation and tumor growth in FOXQ1-overexpressing cells, suggesting that p21 does not contribute to FOXQ1-mediated tumor growth in FOXQ1-overexpressing cells in vivo.

Overexpression of FOXQ1 promotes angiogenesis and antiapoptosis in vivo

To gain an insight into the mechanism by which FOXQ1 enhances tumor growth in vivo, we performed the microarray analysis on H1299/EGFP and H1299/FOXQ1 cells. Fifty-two genes were upregulated over 4-fold by overexpression of FOXQ1 including several genes that have positive roles for tumor growth, such as VEGFA, WNT3A, RSPO2, and BCL11A (Fig. 6A). Overexpression of FOXQ1 upregulated the VEGFA expression for 4.4-fold, suggesting the possibility of enhanced angiogenesis. Real-time RT-PCR for these cells and vascular endothelial growth factor (VEGF) staining of tumor specimens confirmed the result (Fig. 6B). Furthermore, CD31 staining of the tumor specimens showed that FOXQ1 overexpression significantly increased the angiogenesis in vivo.

Figure 6.

FOXQ1 promotes angiogenic and antiapoptotic effects in vivo. A, microarray analysis for H1299/EGFP or H1299/FOXQ1 cells. The upregulated genes over 4-fold by FOXQ1 were shown in the list. B, the mRNA expression levels of VEGFA were determined using a real-time RT-PCR analysis. Rel mRNA, normalized mRNA expression levels (VEGFA/GAPD × 104). VEGF, CD31, TUNEL, and p21 staining of tumor specimens inoculated with H1299/EGFP or H1299/FOXQ1 cells. Microvessel density (MVD) was determined by CD31-positive endothelial cells in tumor specimens using computer-assisted image analysis (Image J software package). C, diagram of a proposed mechanism of FOXQ1 for tumorigenicity and tumor growth. *, P < 0.05.

Figure 6.

FOXQ1 promotes angiogenic and antiapoptotic effects in vivo. A, microarray analysis for H1299/EGFP or H1299/FOXQ1 cells. The upregulated genes over 4-fold by FOXQ1 were shown in the list. B, the mRNA expression levels of VEGFA were determined using a real-time RT-PCR analysis. Rel mRNA, normalized mRNA expression levels (VEGFA/GAPD × 104). VEGF, CD31, TUNEL, and p21 staining of tumor specimens inoculated with H1299/EGFP or H1299/FOXQ1 cells. Microvessel density (MVD) was determined by CD31-positive endothelial cells in tumor specimens using computer-assisted image analysis (Image J software package). C, diagram of a proposed mechanism of FOXQ1 for tumorigenicity and tumor growth. *, P < 0.05.

Close modal

Terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling (TUNEL) and p21 immunostaining of the tumor specimens showed that p21 expression was increased and apoptosis was inhibited in H1299/FOXQ1 cells (Fig. 6B). These results strongly suggest that FOXQ1 promotes tumorigenicity and tumor growth with its angiogenic and antiapoptotic properties in vivo (Fig. 6C).

FOX transcription factors are an evolutionarily conserved superfamily that control a wide spectrum of biological processes. Several Fox gene family members are involved in the etiology of cancer. Only the FOXO family has been regarded as bona fide tumor suppressors that promote apoptosis and cell cycle arrest at G1 (34, 35). The loss of FOXO function observed in alveolar rhabdomyosarcoma through chromosomal translocation was first identified in relation to cancer. Many target genes of FOXO have been reported to date, including p21, cyclin D, Bim, TRAIL, and ER-α (36). On the other hand, the overexpression of FOXM is observed in head and neck cancer, breast cancer, and cervical cancer, and it enhances proliferation and tumor growth in vitro (37), suggesting that FOXM may be an oncogene. Although the available evidence is not conclusive, FOXP, FOXC, and FOXA have been linked to tumorigenesis and progression of certain cancers (36). Thus, the FOX family is thought to act as either an oncogene or a tumor suppressor. In the present study, we showed that the overexpression of FOXQ1 played a tumor-promoting role in CRC.

The p21 promoter region contains several definitive DNA regulatory elements, such as the p53-binding domain, E-box, Smad binding element, and TGF-β response elements. In the case of the other FOX family member FOXO, a recent report showed that the p21 promoter contains a consensus forkhead binding element (GGATCC) immediately upstream of the first Smad binding element and that the FOXO and Smad complexes activate p21 expression, whereas the FOXG1 protein binds to FOXO and blocks p21 induction (38). On the other hand, the consensus binding sequence (5′-NA(A/T)TGTTTA(G/T)(A/T)T-3′) has been defined for human FOXQ1 (4). The p21 promoter region contains several putative FOXQ1 binding sites according to its consensus binding sequence. Indeed, we have shown that FOXQ1 binds to a segment of the p21 promoter, indicating that FOXQ1 directly transactivates the p21 gene expression.

The initial descriptions of p21 were thought to indicate a tumor suppressor–like role, and p21 was almost solely regarded as a modulator with the principal function of inhibiting a cyclin-dependent kinase activity and, hence, cell cycle progression, because it was originally identified as a mediator of p53-induced growth arrest. However, emerging evidence has indicated that p21 may have dual functions with regard to tumor progression and the suppression of cancer cells, with examples of other genes with dual functions including TGF-β, Notch, Runx3, E2F, and p21 (32). Besides its growth inhibitory role, p21 is known to have a positive effect on cell proliferation (3941). A more recent study on leukemic stem cells showed a p21-dependent cellular response that leads to reversible cell cycle arrest and DNA repair; such data clearly illustrate the oncogenic potential of p21 (33). We have shown that p21 has negative roles for tumor growth using FOXQ1-overexpressing cells with knockdown of p21 (Fig. 5D).

Recently, accumulating evidence has shown that FOX transcriptional factors are involved in VEGF regulation and angiogenesis. For example, forkhead has exhibited a positive role in mediating induction of VEGF (4244). In the present study, we identified VEGFA as a candidate target gene of FOXQ1 by microarray analysis and showed that FOXQ1 increased angiogenesis in vivo. Interestingly, although overexpression of FOXQ1 decreases cellular proliferation in vitro, it enhances tumorigenicity and tumor growth in vivo. We consider that this discrepancy can be explained by these angiogenic and antiapoptotic effects of FOXQ1 contribute to enhanced tumor growth in vivo, although p21 negatively functions.

We showed that the overexpression of FOXQ1 inhibited doxorubicin-induced and camptothecin-induced apoptosis in p53-inactivated cancer cells. Therefore, we speculated that FOXQ1 might be a new determinant factor of resistance to drug-induced apoptosis and might represent a poor prognostic factor for CRC patients.

In conclusion, FOXQ1 is markedly overexpressed in CRC and enhances tumorigenicity and tumor growth in vivo. We have elucidated a biological function of FOXQ1, which directly upregulates p21 transcription and promotes angiogenesis and antiapoptosis. Our findings support FOXQ1 as a new member of the cancer-related FOX family in cancer cells.

No potential conflicts of interest were disclosed.

We thank Dr. Richard Simon and Dr. Amy Peng for providing us with the BRB ArrayTools software. This free software was very useful and has been developed for user-friendly applications. We also thank Eiko Honda and Shinji Kurashimo for technical assistance.

Grant Support: Third-Term Comprehensive 2nd term of the 10-Year Strategy for Cancer Control, the program for the promotion of Fundamental Studies in Health Sciences of the National Institute of Biomedical Innovation, Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology of Japan grant-in-aid, and Research Resident Fellowship from the Foundation of Promotion of Cancer Research in Japan (H. Kaneda).

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.

1
Jonsson
H
,
Peng
SL
. 
Forkhead transcription factors in immunology
.
Cell Mol Life Sci
2005
;
62
:
397
409
.
2
Carlsson
P
,
Mahlapuu
M
. 
Forkhead transcription factors: key players in development and metabolism
.
Dev Biol
2002
;
250
:
1
23
.
3
Tran
H
,
Brunet
A
,
Griffith
EC
,
Greenberg
ME
. 
The many forks in FOXO's road
.
Sci STKE
2003
;
2003
:
RE5
.
4
Overdier
DG
,
Porcella
A
,
Costa
RH
. 
The DNA-binding specificity of the hepatocyte nuclear factor 3/forkhead domain is influenced by amino-acid residues adjacent to the recognition helix
.
Mol Cell Biol
1994
;
14
:
2755
66
.
5
Hoggatt
AM
,
Kriegel
AM
,
Smith
AF
,
Herring
BP
. 
Hepatocyte nuclear factor-3 homologue 1 (HFH-1) represses transcription of smooth muscle-specific genes
.
J Biol Chem
2000
;
275
:
31162
70
.
6
Martinez-Ceballos
E
,
Chambon
P
,
Gudas
LJ
. 
Differences in gene expression between wild type and Hoxa1 knockout embryonic stem cells after retinoic acid treatment or leukemia inhibitory factor (LIF) removal
.
J Biol Chem
2005
;
280
:
16484
98
.
7
Hong
HK
,
Noveroske
JK
,
Headon
DJ
, et al
. 
The winged helix/forkhead transcription factor Foxq1 regulates differentiation of hair in satin mice
.
Genesis
2001
;
29
:
163
71
.
8
Potter
CS
,
Peterson
RL
,
Barth
JL
, et al
. 
Evidence that the satin hair mutant gene Foxq1 is among multiple and functionally diverse regulatory targets for Hoxc13 during hair follicle differentiation
.
J Biol Chem
2006
;
281
:
29245
55
.
9
Goering
W
,
Adham
IM
,
Pasche
B
, et al
. 
Impairment of gastric acid secretion and increase of embryonic lethality in Foxq1-deficient mice
.
Cytogenet Genome Res
2008
;
121
:
88
95
.
10
Verzi
MP
,
Khan
AH
,
Ito
S
,
Shivdasani
RA
. 
Transcription factor foxq1 controls mucin gene expression and granule content in mouse stomach surface mucous cells
.
Gastroenterology
2008
;
135
:
591
600
.
11
Harper
JW
,
Adami
GR
,
Wei
N
,
Keyomarsi
K
,
Elledge
SJ
. 
The p21 Cdk-interacting protein Cip1 is a potent inhibitor of G1 cyclin-dependent kinases
.
Cell
1993
;
75
:
805
16
.
12
Xiong
Y
,
Hannon
GJ
,
Zhang
H
,
Casso
D
,
Kobayashi
R
,
Beach
D
. 
p21 is a universal inhibitor of cyclin kinases
.
Nature
1993
;
366
:
701
4
.
13
el-Deiry
WS
,
Tokino
T
,
Velculescu
VE
, et al
. 
WAF1, a potential mediator of p53 tumor suppression
.
Cell
1993
;
75
:
817
25
.
14
Brugarolas
J
,
Moberg
K
,
Boyd
SD
,
Taya
Y
,
Jacks
T
,
Lees
JA
. 
Inhibition of cyclin-dependent kinase 2 by p21 is necessary for retinoblastoma protein-mediated G1 arrest after γ-irradiation
.
Proc Natl Acad Sci U S A
1999
;
96
:
1002
7
.
15
Sherr
CJ
,
Roberts
JM
. 
CDK inhibitors: positive and negative regulators of G1-phase progression
.
Genes Dev
1999
;
13
:
1501
12
.
16
Garner
E
,
Raj
K
. 
Protective mechanisms of p53-21-pRb proteins against DNA damage-induced cell death
.
Cell Cycle
2008
;
7
:
277
82
.
17
Maki
CG
,
Howley
PM
. 
Ubiquitination of p53 and p21 is differentially affected by ionizing and UV radiation
.
Mol Cell Biol
1997
;
17
:
355
63
.
18
Gartel
AL
,
Tyner
AL
. 
Transcriptional regulation of the p21((WAF1/CIP1)) gene
.
Exp Cell Res
1999
;
246
:
280
9
.
19
Gartel
AL
,
Tyner
AL
. 
The role of the cyclin-dependent kinase inhibitor p21 in apoptosis
.
Mol Cancer Ther
2002
;
1
:
639
49
.
20
Yamanaka
R
,
Arao
T
,
Yajima
N
, et al
. 
Identification of expressed genes characterizing long-term survival in malignant glioma patients
.
Oncogene
2006
;
25
:
5994
6002
.
21
Tanaka
K
,
Arao
T
,
Maegawa
M
, et al
. 
SRPX2 is overexpressed in gastric cancer and promotes cellular migration and adhesion
.
Int J Cancer
2009
;
124
:
1072
80
.
22
Takeda
M
,
Arao
T
,
Yokote
H
, et al
. 
AZD2171 shows potent antitumor activity against gastric cancer over-expressing fibroblast growth factor receptor 2/keratinocyte growth factor receptor
.
Clin Cancer Res
2007
;
13
:
3051
7
.
23
United Kingdom Co-ordinating Committee on Cancer Research (UKCCCR)
. 
Guidelines for the Welfare of Animals in Experimental Neoplasia (Second Edition)
.
Br J Cancer
1998
;
77
:
1
10
.
24
Iwasa
T
,
Okamoto
I
,
Suzuki
M
, et al
. 
Inhibition of insulin-like growth factor 1 receptor by CP-751,871 radiosensitizes non-small cell lung cancer cells
.
Clin Cancer Res
2009
;
15
:
5117
25
.
25
Shimada
K
,
Nakamura
M
,
Anai
S
, et al
. 
A novel human AlkB homologue, ALKBH8, contributes to human bladder cancer progression
.
Cancer Res
2009
;
69
:
3157
64
.
26
Igarashi
T
,
Izumi
H
,
Uchiumi
T
, et al
. 
Clock and ATF4 transcription system regulates drug resistance in human cancer cell lines
.
Oncogene
2007
;
26
:
4749
60
.
27
Bieller
A
,
Pasche
B
,
Frank
S
, et al
. 
Isolation and characterization of the human forkhead gene FOXQ1
.
DNA Cell Biol
2001
;
20
:
555
61
.
28
Datto
MB
,
Yu
Y
,
Wang
XF
. 
Functional analysis of the transforming growth factor β responsive elements in the WAF1/Cip1/p21 promoter
.
J Biol Chem
1995
;
270
:
28623
8
.
29
Mahyar-Roemer
M
,
Roemer
K
. 
p21 Waf1/Cip1 can protect human colon carcinoma cells against p53-dependent and p53-independent apoptosis induced by natural chemopreventive and therapeutic agents
.
Oncogene
2001
;
20
:
3387
98
.
30
Mukherjee
S
,
Conrad
SE
. 
c-Myc suppresses p21WAF1/CIP1 expression during estrogen signaling and antiestrogen resistance in human breast cancer cells
.
J Biol Chem
2005
;
280
:
17617
25
.
31
Seoane
J
,
Le
HV
,
Massague
J
. 
Myc suppression of the p21(Cip1) Cdk inhibitor influences the outcome of the p53 response to DNA damage
.
Nature
2002
;
419
:
729
34
.
32
Rowland
BD
,
Peeper
DS
. 
KLF4, p21 and context-dependent opposing forces in cancer
.
Nat Rev Cancer
2006
;
6
:
11
23
.
33
Viale
A
,
De Franco
F
,
Orleth
A
, et al
. 
Cell-cycle restriction limits DNA damage and maintains self-renewal of leukaemia stem cells
.
Nature
2009
;
457
:
51
6
.
34
Brunet
A
,
Bonni
A
,
Zigmond
MJ
, et al
. 
Akt promotes cell survival by phosphorylating and inhibiting a Forkhead transcription factor
.
Cell
1999
;
96
:
857
68
.
35
Paik
JH
,
Kollipara
R
,
Chu
G
, et al
. 
FoxOs are lineage-restricted redundant tumor suppressors and regulate endothelial cell homeostasis
.
Cell
2007
;
128
:
309
23
.
36
Myatt
SS
,
Lam
EW
. 
The emerging roles of forkhead box (Fox) proteins in cancer
.
Nat Rev Cancer
2007
;
7
:
847
59
.
37
Kalin
TV
,
Wang
IC
,
Ackerson
TJ
, et al
. 
Increased levels of the FoxM1 transcription factor accelerate development and progression of prostate carcinomas in both TRAMP and LADY transgenic mice
.
Cancer Res
2006
;
66
:
1712
20
.
38
Seoane
J
,
Le
HV
,
Shen
L
,
Anderson
SA
,
Massague
J
. 
Integration of Smad and forkhead pathways in the control of neuroepithelial and glioblastoma cell proliferation
.
Cell
2004
;
117
:
211
23
.
39
Dong
Y
,
Chi
SL
,
Borowsky
AD
,
Fan
Y
,
Weiss
RH
. 
Cytosolic p21Waf1/Cip1 increases cell cycle transit in vascular smooth muscle cells
.
Cell Signal
2004
;
16
:
263
9
.
40
Dupont
J
,
Karas
M
,
LeRoith
D
. 
The cyclin-dependent kinase inhibitor p21CIP/WAF is a positive regulator of insulin-like growth factor I-induced cell proliferation in MCF-7 human breast cancer cells
.
J Biol Chem
2003
;
278
:
37256
64
.
41
Zhang
C
,
Kavurma
MM
,
Lai
A
,
Khachigian
LM
. 
Ets-1 protects vascular smooth muscle cells from undergoing apoptosis by activating p21WAF1/Cip1: ETS-1 regulates basal and inducible p21WAF1/Cip: ETS-1 regulates basal and inducible p21WAF1/Cip1 transcription via distinct cis-acting elements in the p21WAF/Cip1 promoter
.
J Biol Chem
2003
;
278
:
27903
9
.
42
Banham
AH
,
Boddy
J
,
Launchbury
R
, et al
. 
Expression of the forkhead transcription factor FOXP1 is associated both with hypoxia inducible factors (HIFs) and the androgen receptor in prostate cancer but is not directly regulated by androgens or hypoxia
.
Prostate
2007
;
67
:
1091
8
.
43
Furuyama
T
,
Kitayama
K
,
Shimoda
Y
, et al
. 
Abnormal angiogenesis in Foxo1 (Fkhr)-deficient mice
.
J Biol Chem
2004
;
279
:
34741
9
.
44
Gupta
S
,
Joshi
K
,
Wig
JD
,
Arora
SK
. 
Intratumoral FOXP3 expression in infiltrating breast carcinoma: its association with clinicopathologic parameters and angiogenesis
.
Acta Oncol
2007
;
46
:
792
7
.

Supplementary data