Overexpression of hypoxia-inducible-factor 1α(HIF-1α) in oesophageal squamous cell carcinoma correlates with lymph node metastasis and pathologic stage

The purpose of this study is to investigate the clinical and histopathologic significance of hypoxia-inducible-factor 1α (HIF-1α) expression in oesophageal squamous cell carcinoma. One hundred and thirty surgically resected specimens of OSCC were immunohistochemically assessed for HIF-1α expression with monoclonal antibody. High HIF-1α immunostaining was detected in 40 specimens. The percentage of high HIF-1α expression cases increased with tumour stage according to pTNM system. High HIF-1α expression correlated with pTNM stage, depth of tumour invasion, lymph node metastasis, distant metastasis, lymphatic invasion and positive surgical margin. The overall survival rate was worse in patients with high HIF-1α pattern than in patients with low-expression pattern. Univariate analyses identified high HIF-1α positivity, depth of tumour invasion, lymph node metastasis, distant metastasis, lymphatic invasion, and a positive surgical margin as risk factors. Multivariate analyses indicated that depth of tumour invasion, lymph node metastasis and positive surgical margin, but not HIF-1α, were independent prognostic factors. Survival in patients with a high HIF-1α expression was significantly worse than in those with low expression in patient treated with adjuvant therapy.

Oesophageal cancer is an aggressive tumour with a poor prognosis. Although diagnostic and surgical techniques have been advanced, survival rates have not been improved in the last decade, and the 5year survival rate of patients with surgically treated oesophageal cancer remains less than 50% in spite of three-field lymph node dissection and combination chemotherapy and radiotherapy (Torres et al, 1999;Adham et al, 2000;Ando et al, 2000;Collard et al, 2001).
The choice of therapeutic strategy is based primarily on whether lymph node metastasis has occurred. The presence of lymph node metastasis is the most important determinant of outcome (Ando et al, 2000;Kusumi et al, 2000). Patients with oesophageal cancer without nodal metastasis have a lower rate of recurrence after operation than those with nodal metastasis (Kato et al, 1996;Matsubara et al, 1996). However, it is difficult to determine whether lymph node metastases have occurred preoperatively in spite of new imaging techniques. Thus, the identification of a marker that predicts lymph node metastasis, and hence prognosis, is highly desirable.
A hypoxic microenvironment is characteristic of many solid tumours. In the absence of neovascularisation, tumours cannot grow beyond several cubic millimeters, because the diffusion of oxygen, glucose, and other nutrients from blood vessels is limited (Dang and Semenza, 1999). Cancer cell proliferation may outpace the rate of angiogenesis, resulting in tissue hypoxia. To surmount these limitations, the tumour needs to acquire abilities that allow it to adapt to a hypoxic microenvironment (Maxwell et al, 1999;Zhong et al, 1999).
Hypoxia-inducible-factor 1a(HIF-1a)is a 120 kDa nuclear protein. Hypoxia-inducible-factor 1 is a heterodimer, consisting of an a and a b subunit, both belonging to the basic -helix -loop -helix Per-aryl hydrocarbon receptor nuclear translocator-Sim (PAS) family of transcription factors (Wang et al, 1995). Hypoxiainducible-factor-1 is an important component of a widely operative transcriptional response, activated by hypoxia, cobaltous ions, and iron chelation. Hypoxia-inducible-factor 1 activates transcription of hypoxia-inducible genes, including those encoding erythropoietin (Wang and Semenza 1993), vascular endothelial growth factor (VEGF) (Forsythe et al, 1996), heme oxygenase-1 (Hoetzel et al, 2001), inducible nitric oxide synthase (Jung et al, 2000), and the glycolytic enzymes aldolase A, enolase 1, lactate dehydrogenase A , phosphofructokinase I (Minchenko et al, 2002), and phosphoglycerate kinase I (Li et al, 1996). The C-terminal of HIF-1a binds to p300, and p300/CBP-HIF complexes participate in the induction of hypoxia-responsive genes, including VEGF (Arany et al, 1996). Induction of HIF-1a in response to hypoxia is instantaneous, and it can be expressed very early in carcinogenesis, before histologic evidence of angiogenesis or invasion exists (Ryan et al, 2000).
It has been shown that HIF-1a is a key player in the cancer cells response to low-oxygen tension in a variety of physiologic processes including embryogenesis (Ryan et al, 2000), angiogenesis, tumorigenesis (Maxwell et al, 1997) and metastases (Zhong et al, 1999). Moreover, hypoxic regions have been shown to be both chemo-and radiation resistant (Teicher, 1994;Aebersold et al, 2001;Koukourakis et al, 2001).
In the current study, we examined archival material from 130 surgical specimens of oesophageal squamous cell carcinoma (OSCC) for HIF-1a immunoreactivity. The purpose was to determine the correlation between HIF-1a immunoreactivity and clinical and histopathologic factors.

Patients and tissue samples
Surgical specimens were collected from 130 patients with primary OSCC, who underwent radical total oesophagectomy and threefield lymph node dissection from 1989 to 1999 at the Department of Surgical Oncology of Hokkaido University Hospital, Hokkaido Gastroenterology Hospital, or Teine Keijinkai Hospital. Cases of in-hospital death were excluded. The clinicopathologic stage was determined according to the TNM classification system of the International Union Against Cancer (UICC) (Sobin and Wittekind, 1997).

Immunohistochemistry
The expression of HIF-1a was determined immunohistochemically in paraffin-embedded specimens fixed in 10% formalin. Histologic slides, 4 mm in thickness, were deparaffined in xylene and rehydrated through a series of graded ethanol. Endogenous peroxidase activity was blocked by incubation in 3% hydrogen peroxide in methanol for 10 min. The sections were washed twice in phosphate-buffered saline (PBS) and incubated with 10% normal goat serum (Histofine SAB-PO kit, Nichirei Corporation, Tokyo, Japan) for 30 min. The slides were then exposed overnight to a monoclonal antibody against HIF-1a(HIF-1aAb-4, NEO MARKERS, Fremont, CA, USA) at a dilution of 1:400 at 41C. After washing in PBS, a biotinylated goat antibody to mouse immunoglobulin (Histofine SAB-PO kit, Nichirei Corporation) was applied, followed by incubation at room temperature for 60 min. The immunohistochemical reactions were developed in freshly prepared 3,3 0 -diamino-benzidine tetrahydrochloride (Histofine SAB-PO kit, Nichirei Corporation). Slides were counterstained in haematoxylin and coverslipped in a systemic mounting medium. Tissue samples incubated with nonimmune serum served as negative controls. Immunostaining was evaluated in three visual fields at a power of Â 200 under an Olympus microscope (Olympus Optical, Tokyo, Japan).
Tumour cell immunoreactivity to HIF-1a protein was scored based on the number of cells exhibiting the nuclear or cytoplasmic staining using the following classification system: À, no staining; 1 þ , nuclear staining in less than 1% of cells; 2 þ , nuclear staining in 1% -10% of cells and/or with weak cytoplasmic staining; 3 þ , nuclear staining in 10% to 50% of cells and/or with distinct cytoplasmic staining; 4 þ , nuclear staining in more than 50% of cells and/or with strong cytoplasmic staining. Hypoxia-induciblefactor 1a 3 þ and 4 þ were considered high expression patterns while the remaining cases were considered to be low expression.
All specimens were evaluated by three investigators who were blinded to the patients' clinical information.

Statistical analysis
Either the w 2 test or Fisher's exact test was used to analyse the correlation between HIF-1a expression and clinicopathologic features. The cumulative survival rate was calculated by the Kaplan -Meier method, and the significance of differences in survival was analysed by the log-rank test. The univariate and multivariate analyses were performed using the Cox proportional hazard regression model; Po 0.05 was considered significant in all analyses. Computations were performed using the Statview J version 4.5 (SAS Institute, Inc., Cary, NC, USA) software package.

Patient factor
Specimens from 130 patients were included in the current study (113 male and 17 female patients). The median patient age was 63 years (range, 38 -82 years). A relatively large number of patients had early-stage disease (81 patients, 62%). Sixty-six patients (51%) had lymph node metastases and 22 patients (17%) had distant nodal metastases. No patient had distant organ metastasis at the time of operation. The study population had the following performance status (PS): PS0, 114 patients; PS1, 15 patients; and PS2, one patient. Following radical operation, adjuvant therapy was administered to 52 patients (Stage I; seven cases, Stage II; 23 cases; Stage III; 11 cases, and Stage IV; 11 cases). Chemotherapy, radiotherapy, and chemoradiotherapy were treated in 12, 18, and 22 patients, respectively (Table 1). The median follow-up period was 29 months (range, 2 -114 months).
HIF-1a immunoreactivity had already been identified in carcinoma in situ of oesophagus ( Figure 1F).

Kaplan-Meier survival analysis
The overall 5-year survival rate was 50.4%. The survival curve of patients with a high HIF-1a expression tumours was worse than that of patients with low-expression tumours (log-rank test, P ¼ 0.0007; Figure 2).

Multivariate survival analysis
Cox regression multivariate analysis identified depth of tumour invasion (P ¼ 0.0091), lymph node metastasis (P ¼ 0.0003), and positive surgical margin (P ¼ 0.0375), as independent unfavorable factors. High HIF-1a expression was not an independent prognostic factor (Table 4).

Kaplan-Meier survival analysis of the patient treated with adjuvant therapy
Survival in patients with a high HIF-1a expression pattern was significantly worse than in those with a low-expression pattern in the patient treated with adjuvant therapy (P ¼ 0.0464; Figure 3).

DISCUSSION
The current results show that (1) high HIF-1a expression correlates with depth of tumour invasion, lymph node metastasis, distant metastasis, pTNM stage, lymphatic invasion and a positive surgical margin, and (2) survival in patients with a high HIF-1a pattern was worse than in those with low-expression pattern.
Although HIF-1a was not an independent unfavourable prognostic factor, its expression may strongly influence both tumour proliferation and lymph node metastasis in OSCC. However, it has been reported that HIF-1a overexpression was not significantly correlated with pathological parameter in other cancers, including head and neck cancer (Hockel et al, 1993;Fyles et al, 1998), and oropharyngeal cancer (Aebersold et al, 2001). Thus, HIF-1a expression seems to behave in a tissue-dependent manner.
Hypoxia has been shown to compromise the beneficial effects of chemotherapeutic drugs (Teicher, 1994) and interfere with the response of tumours to radiation (Moulder and Rockwell, 1987). Pretreatment oxygenation levels have been found to be predictive of the radiation response and survival of patients with cancer of the uterine cervix (Hockel et al, 1993;Fyles et al, 1998), head and neck (Gatenby et al, 1988;Nordsmark et al, 1996), oropharyngeal (Aebersold et al, 2001), and early oesophageal cancer (Koukourakis et al, 2001). In the current study, of all pathological stages, overexpression of HIF-1a in OSCC significantly correlates with an unfavourable prognosis in the patients treated with adjuvant therapies.
Preoperative studies on biopsy specimens obtained by endoscopy might allow clinicians to make better-informed therapeutic decisions in conjunction with this marker.
In conclusion, we have suggested that (1) high HIF-1a expression may be a marker for lymph node metastasis; and (2) high HIF-1a expression may predict an unfavourable prognosis in the patient treated with OSCC. HIF-1a in oesophageal cancer T Kurokawa et al