Article Text

Download PDFPDF

Prognostic significance of FOXP1 as an oncogene in hepatocellular carcinoma
  1. Yixin Zhang1,
  2. Suqin Zhang1,
  3. Xudong Wang2,
  4. Jibin Liu3,
  5. Liping Yang3,
  6. Song He4,
  7. Li Chen4,
  8. Jianfei Huang5
  1. 1Department of Hepatobiliary Surgery, Affiliated Tumor Hospital of Nantong University, Nantong, Jiangsu Province, China
  2. 2Department of Laboratory Medicine, Affiliated Hospital of Nantong University, Nantong, Jiangsu Province, China
  3. 3Department of Laboratory Medicine, Affiliated Tumor Hospital of Nantong University, Nantong, Jiangsu Province, China
  4. 4Department of Pathology, Affiliated Tumor Hospital of Nantong University, Nantong, Jiangsu Province, China
  5. 5Department of Pathology, Affiliated Hospital of Nantong University, Nantong, Jiangsu Province, China
  1. Correspondence to Dr Jianfei Huang, Department of Pathology, Affiliated Hospital of Nantong University, Nantong 226001, China; ntyydoctor{at}126.com

Abstract

Background Forkhead Box P1 (FOXP1) has been described as both a tumour suppressor candidate and a potential oncogene. The aim of this study is to identify new prognostic biomarkers and therapeutic target structures for the diagnosis and treatment of hepatocellular carcinoma (HCC).

Methods The expression of FOXP1 mRNA in HCC was characterised using real-time PCR and 20 pairs of fresh frozen HCC tissues and corresponding non-cancerous tissues. FOXP1 protein expression in HCC was confirmed using immunohistochemistry on a tissue microarray chip. Finally, FOXP1 expression was correlated with conventional clinicopathological features of HCC and patient outcome.

Results The expression of FOXP1 mRNA and protein in HCC cells was much higher than in normal hepatic cells (Z=2.315, p=0.021 and χ2=28.071, 95% CI 0.233 to 0.480, p<0.001, individually). The comparison of clinicopathological characteristics and immunohistochemistry by χ2 test analysis showed that the high expression of FOXP1 in HCC was related to large tumour diameter (χ2=6.210, p=0.013), high serum α-fetoprotein levels (χ2=6.920, p=0.031) and later stage grouping with tumour node metastasis classification (χ2=6.714, p=0.035). Kaplan–Meier survival and Cox regression analysis showed that high FOXP1 expression (HR=2.182, 95% CI 1.146 to 4.154, p=0.018) and regional lymph node metastasis (HR=2.326, 95% CI 1.037 to 5.217, p=0.041) were independent prognosis factors.

Conclusions From this investigation the authors elucidated for the first time that the correlation of high FOXP1 expression correlates with an aggressive malignant phenotype and may constitute a novel prognostic factor for HCC. These results also support a role for FOXP1 as an oncogene in HCC.

  • FOXP1
  • HCC
  • immunohistochemistry
  • TMA
  • prognostic factor
  • liver cancer
  • oncogenes
  • antibodies
  • cancer research
View Full Text

Statistics from Altmetric.com

Introduction

Liver cancer, especially hepatocellular carcinoma (HCC), is the fifth most prevalent malignancy worldwide and the third most common cancer related to mortality globally.1 HCC is a complex disease involving many factors and its staging systems can be very complicated. Important risk factors for developing HCC are cirrhosis and hepatitis B virus.2–4 Many studies have indicated a number of molecular markers that are involved in the proliferation, invasion, differentiation, metastasis and survival of HCC.5 For most HCC patients who present with an advanced tumour stage,6 the novel molecular therapies have had a limited therapeutic effect on malignant tumours5 7 8. Thus, there is an urgent need for new prognostic biomarkers and therapeutic target structures in the diagnosis and treatment of HCC.

Forkhead Box P1 (FOXP1), one of four members of the FOX subfamily of forkhead transcription factors that have a broad range of functions, is widely expressed and has been reported to play an important role in normal human tissues and in several types of solid tumours.9–14 Studies on human FOXP1 have focused primarily on its key role in human malignancy and in cellular differentiation.9 Because of its observed overexpression in diverse tumour types, FOXP1 has been described both as a tumour suppressor candidate and as a potential oncogene. Loss of FOXP1 was seen in endometrial cancer,15 prostate cancer,16 17 renal cell carcinoma,18 and has been shown to confer poor prognosis in breast cancer.19 20 By contrast, FOXP1 as an oncogene with high expression levels in many kinds of B cell lymphomas was shown to correlate with a worse outcome.21–25 In a recent report, it was shown, using western blot and immunohistochemical methods, that the expression of FOXP1 increased in human primary HCC samples compared with corresponding normal liver tissues.26 Previous research has failed to study the relationship between FOXP1 expression in HCC and clinicopathological characteristics. Whether FOXP1 expression has prognostic implications and whether its status can be used as a potential guide to therapy in HCC remains to be investigated.

In the present study, we investigated FOXP1 gene and protein expression in a number of HCC samples with tumour-adjacent normal tissues. Moreover, we have carefully analysed the relationship between FOXP1 expression and clinicopathology to determine its clinicopathological significance in a selected group of patients with HCC. Finally, we evaluated the prognostic significance of the expression of the FOXP1 protein in HCC.

Materials and methods

Patients and tissue samples

A panel of formalin-fixed, paraffin-embedded HCC tissues (n=114) and tumour-adjacent normal tissues (n=107) from patients undergoing surgical therapy was obtained from the Affiliated Tumor Hospital of Nantong University, China, between 2003 and 2006. The average age of the group was 49.28 years (range 21–75 years). Before surgical therapy, none of the patients had received neoadjuvant chemotherapy, radiation therapy or immunotherapy. The tissue microarray (TMA) chip containing the HCC and the corresponding peritumoural tissues was produced as duplicate 0.6 mm tissue cores by Shanxi Chaoying Biotechnology Co., Ltd (Shanxi Province, China). A panel of 20 fresh frozen HCC tissues and corresponding adjacent non-cancerous tissues, obtained from the tissue bank of the same hospital, were also included in this study. Survival was calculated from the date of surgery to the date of death or last follow-up. Medical details (including age, tumour size, serum level of α-fetoprotein (AFP), 5-year follow-up survival records and other information) were collected from the medical records of each patient (see table 1). Tumour staging was in accordance with the WHO standards.27 Ethics approval to perform this study was obtained from the Human Research Ethics Committee of the Affiliated Tumor Hospital of Nantong University, Jiangsu Province, China.

Table 1

Correlation of high FOXP1 expression with clinicopathological characteristics in patients with hepatocellular carcinoma

One-step quantitative real-time PCR

Total RNA was extracted from a subset of the frozen tissues mentioned above using the Trizol reagent (Invitrogen, Carlsbad, California, USA). Total RNA extraction, quality control and one-step quantitative real-time PCR were performed as previously reported.28 FOXP1-specific oligonucleotide primers, forward 5′-TCTCATTGATTGGAGCATT-3′ and reverse 5′-GGCAGTGGTAGGATAAAC-3′, were designed to give a 148 bp PCR product. The glyceraldehyde 3-phosphate dehydrogenase (GAPDH) mRNA level was used to standardise the mRNA levels of the FOXP1 gene and melt curves were analysed to estimate the accuracy of the real-time PCR results. The GAPDH primers were obtained from Invitrogen (Shanghai, China). Amplification conditions consisted of 30 min at 42°C for reverse transcription, 2 min at 94°C for Taq activation followed by 35 cycles of 94°C for 20 s, one cycle of 58°C for 20 s and elongation at 72°C for 30 s.

Relative quantification was defined as the amount of specific mRNA normalised to non-cancerous liver tissues as determined using the comparative cycle threshold (Ct) method. Relative quantification was defined as the amount of specific mRNA normalised to liver tissues as determined using the comparative Ct method. The FOXP1 gene expression was defined as 2–ΔΔCt, where ΔΔCt=ΔCtliver cancer tissues−ΔCtmatching liver tissues. The matching liver tissues sample having the median level of expression for FOXP1 gene was chosen as the calibrator sample and used to normalise expression of liver tissues samples (FOXP1 gene expression in matching liver tissues=1). ΔCt is defined as CtFOXP1−CtGAPDH.

Immunohistochemistry

The TMA sections used for immunohistochemistry (IHC) were deparaffinised, and peroxidase was quenched with methanol and 3% H2O2 for 15 min. For antigen retrieval, the sections were boiled under pressure in citrate buffer, pH 6.0, for 3 min. Then the tissues were incubated for 1 h with primary mouse anti-FOXP1 (ab32010; Abcam, Cambridge, UK) which was diluted 1:300 in 1% bovine serum albumin. Following washing with phosphate buffered saline (PBS), the sections were incubated with horse radish peroxidase-conjugated goat anti-mouse antibody (Dako Cytomation, USA) for 15 min and washed again with PBS. The colour was developed by 15-minute incubation with diaminobenzidine solution (Kem-En-Tec Diagnostics, Denmark), and sections were weakly counterstained with haematoxylin. For the negative control reactions, PBS was used instead of the primary antibody.

Blinded evaluations of the immunostaining of the above FOXP1 and independent observation were carried out simultaneously. Positive cell staining percentages were scored into four categories: 0 for 0%, 1 for 1–33%, 2 for 34–66% and 3 for 67–100% staining. The IHC staining intensities were also scored into 4 grades: 0, 1, 2 and 3. The sum of the percentage and intensity scores was used as the final FOXP1 staining score as described earlier.29 The staining scores were defined as low expression for scores of 0–2 and high expression for scores from 3 to 6.

Statistical analysis

The expression of FOXP1 mRNA in fresh frozen HCC tissues and in the corresponding peritumoural tissues normalised to GAPDH was analysed with the Wilcoxon non-parametric signed-rank test. The relationships between FOXP1 expression and the relevant patient and tumour characteristics were assessed using the χ2 test. For the TMA slides, the following clinical data were evaluated: gender, age, tumour size, preoperative serum AFP level and other clinicopathological information, as well as the patients' outcome. Survival curves were calculated using the Kaplan–Meier method and compared by the log-rank test. Factors shown to be of prognostic significance in the univariate models were evaluated in a multivariable Cox regression model. For all analyses, a p value <0.05 was regarded as statistically significant. Data were analysed using STATA V.9.0 (Stata Corporation).

Results

Measurement of FOXP1 mRNA expression in HCC by one-step quantitative PCR

Total RNA was extracted from the 20 fresh frozen HCC tissues and subjected to one-step quantitative PCR to investigate the expression of FOXP1 mRNA. To compare the expression of the mRNA, we also investigated samples from the matched tumour adjacent tissues. When normalised to GAPDH, the means of FOXP1 mRNA in the HCC and corresponding non-cancerous tissues were 1.55±0.78 and 0.99±0.46 (Z=2.315, p=0.021), respectively. Thus, the average FOXP1 expression was 1.58-fold higher in the cancer samples than in the non-malignant tissues (figure 1).

Figure 1

One-step quantitative PCR FOXP1 mRNA levels. FOXP1 expression was normalised against the GAPDH mRNA level. The FOXP1 mRNA levels were higher in hepatocellular carcinoma (cancer) tissues than in hepatic (non-cancer) cells (p<0.05). FOXP1, Forkhead Box P1; GAPDH, glyceraldehyde 3-phosphate dehydrogenase.

Immunohistochemistry of FOXP1 protein in HCC and peritumoural tissues

The expression of FOXP1 protein in HCC was studied by immunohistochemical analysis. Typically observed immunohistochemical staining patterns for FOXP1 are shown in figure 2. Although positive staining was mainly localised in the cytoplasm of HCC and hepatic cells, a combination of nuclear and cytoplasmic positive staining was observed in a few samples (figure 2). No FOXP1 immunolabelling was observed in the stromal fibroblasts of these tissues.

Figure 2

Expression pattern of FOXP1 in tumour tissue and in tumour-adjacent tissue of HCC TMA sections revealed by immunohistochemical staining. (FOXP1 is stained brown). (A1 and A2) The expression pattern of FOXP1 in poorly differentiated HCC tissue. (B1 and B2) The expression pattern of FOXP1 in well-differentiated HCC tissue and in the tumour-adjacent tissue. (C1 and C2) The expression pattern of FOXP1 in peritumoural tissues of HCC showing a negative immunohistochemical reaction. (Original magnification ×40 in A1, B1 and C1, ×400 in A2, B2 and C2). FOXP1, Forkhead Box P1; HCC, hepatocellular carcinoma; TMA, tissue microarray.

High FOXP1 expression was detected in 80 (70.18%) of the 114 HCC tissues and in 37 (34.58%) of the 107 matched tumour-adjacent tissues. The data showed statistical significance (χ2=28.071, 95% CI 0.233 to 0.480, p<0.001) using χ2 test analysis and was consistent with the FOXP1 mRNA level that was found in the HCC samples.

Correlation between FOXP1 expression and clinicopathological parameters

The relationship between levels of FOXP1 protein and the clinicopathological parameters of 114 HCC patients is shown in (table 1). High FOXP1 expression was associated with large tumour diameter (p=0.013), high serum AFP (p=0.031) and later tumour node metastasis (TNM) stages (p=0.035). By contrast, no statistically significant correlation was found between high expression of FOXP1 and the following factors: gender, age, differentiation, liver cirrhosis, portal vein invasion, regional lymph node metastasis and other information (table 1).

FOXP1 expression and patient survival

As expected, FOXP1 protein overexpression showed a significant association with 5-year survival (p=0.046). In addition, known HCC clinical prognostic factors such as differentiation (p=0.019), portal vein invasion (p=0.027), regional lymph node metastasis (p=0.001), multifocal gross classification (p=0.006) and TNM stage (p=0.007) showed a statistically significant association with 5-year survival in Cox regression univariate analysis (table 2). All these factors were included in a multivariable analysis. High FOXP1 expression (p=0.018) and regional lymph node metastasis (p=0.041) were identified as independent predictive factors for poor outcome of HCC. Kaplan–Meier survival curves demonstrated that patients with high FOXP1 expression and regional lymph node metastasis had a significantly shorter survival time compared with those with no or low FOXP1 protein expression (figure 3).

Table 2

Univariate and multivariable analysis of prognostic factors in hepatocellular carcinoma for 5-year survival

Figure 3

Survival curves after surgical therapy in patients with hepatocellular carcinoma. (The FOXP1 expression curves were calculated by the Kaplan–Meier method.) (A) Survival curves calculated by FOXP1 expression. FOXP1=1 is the high FOXP1 expression group (red line); FOXP1=0 is the low and no expression of FOXP1 group (blue line). (B) Survival curves calculated by regional lymph node (LN) metastasis. LN=1 is the group with regional lymph node metastasis (red line); LN=0 is the group without regional lymph node metastasis (blue line). FOXP1, Forkhead Box P1.

Discussion

FOXP1, initially named glutamine-rich factor 1 (QRF1), was discovered in a screen to detect glutamine-rich transcription factors in B cells.30 The full-length human FOXP1 gene was first cloned with a murine monoclonal antibody (JC12) which recognised a dissimilarly expressed protein in normal and malignant B cells.11 As a member of the broadly expressed FOXP subset of forkhead (FOX) transcription factors, FOXP1 has a diverse repertoire of functions ranging from the regulation of B-cell development,13 cardiac development,31 lung and oesophagus development14 to monocyte and macrophage differentiation.32

Importantly, FOXP1 is considered to be an oncogene or a tumour suppressor in different human malignancies.30 In the present study, we described the relationship between FOXP1 expression and clinicopathological parameters in HCC patients. To compare the expression levels of FOXP1 mRNA, 20 fresh frozen HCC tissue samples and matched tumour-adjacent tissues were used. When normalised to GAPDH, the quantitative PCR results showed that the mean FOXP1 mRNA expression was much higher in HCC than in the corresponding non-cancerous tissue. We also showed that the FOXP1 protein was overexpressed in HCC and this was confirmed by IHC analysis using the TMA slide and χ2 test. However, in an earlier study, the expression pattern of FOXP1 in tumour cells was not correlated with variable intensity IHC in different types of cancers.18 In our study, immunohistochemically positive staining of FOXP1 was localised mainly in the cytoplasmic and nuclear compartments of the normal and neoplastic liver tissues. The positive pattern agrees with previous reports which showed that FOXP1 was overexpressed in tumour cells compared with non-tumour cells.15–20 26 Furthermore, cytoplasmic expression was found more frequently in HCC cells; a similar expression pattern was reported earlier for endometrial adenocarcinoma.15

In a previous review, the cytoplasmic expression of FOXP1 in malignant endometrium was linked only with deep myometrial invasion and poor differentiation.15 In prostate cancer, nuclear FOXP1 expression was positively correlated with androgen receptor expression and with a post-operative Gleason score; cytoplasmic FOXP1 expression was not significantly correlated with any of the clinicopathological variables that were studied.16 17 In clear cell renal cell carcinomas, nuclear expression or a combination of nuclear, membranous and cytoplasmic expression of FOXP1 was noticed. In patients with renal cancer, the expression of FOXP1 was negatively correlated with tumour grading and Ki67 status in tumours, while no significant correlation was found between FOXP1 expression and survival.18 The loss of FOXP1 nuclear protein expression in breast carcinoma cells was reported to be associated with a significantly poor patient outcome.19 20 Taken together, these findings suggested that FOXP1 might act as a tumour suppressor gene in human cancer.

In the present investigation, high FOXP1 expression was correlated significantly with large tumour diameter, high serum AFP and later TNM stage in HCC patients. We found that FOXP1 expression in was higher in liver cancer cells than in normal hepatic cells, indicating that FOXP1 could play an oncogene role in HCC. Furthermore, high FOXP1 expression along with regional lymph node metastasis was identified to be an independent predictive factor for poor outcome in HCC.

Our data clearly showed that high cytoplasmic expression of FOXP1 was associated with a significantly poor survival. This result is in agreement with current understanding of the role of FOXP1 as an oncogene in lymphoma.21–25 Although some studies seem to contradict these findings, the majority of studies on FOXP1 suggest its protective role in tumour progression.15–20 A recent paper reported evidence that describes FOXP1 as an oncogene that can induce aggressive tumour growth.26

In conclusion, to the best of our knowledge, this is the first report that correlates high FOXP1 expression with an aggressive malignant phenotype and indicates that FOXP1 may constitute a novel prognostic marker for HCC. Additionally, these results support a role for FOXP1 as an oncogene in HCC.

Take-home messages

  • Forkhead Box P1 (FOXP1) has been described as both a tumour suppressor candidate and a potential oncogene. The relationship between FOXP1 expression and the clinicopathological characteristics of patients with hepatocellular carcinoma (HCC) has not yet been satisfactorily defined.

  • FOXP1 expression was higher in HCC cells than in normal hepatic cells. The expression pattern of FOXP1 was mainly localised in the cytoplasm of HCC cells and normal hepatic cells.

  • High expression of FOXP1, as an oncogene, in HCC was related to progressive pathological factors, including large-tumour diameter, high serum α-fetoprotein levels and later stage grouping with tumour node metastasis (TNM) classification.

  • High FOXP1 expression and regional lymph node metastasis were independent poor prognostic factors for HCC patients.

References

View Abstract

Footnotes

  • YZ and SZ contributed equally to this study.

  • Funding This study was supported by grants from the Social Development and Applied Research Projects (K2010012) of Nantong, Jiangsu Province, China, and the International Cooperation and Exchanges (2011) from the Department of Health of Jiangsu Province, China.

  • Competing interests None.

  • Ethics approval This study was conducted with the approval of the Human Research Ethics Committee of the Affiliated Tumor Hospital of Nantong University.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.