BNIP3 is a proapoptotic protein regulated by hypoxia-inducible factor 1. We analyzed BNIP3 expression in 105 tumor samples from early operable, non-small lung cancer and the relationship of expression to hypoxia-inducible factor 1α, other hypoxia-regulated pathways, and prognosis. There was strong cytoplasmic expression in >10% of cells in 40 of 105 cases. BNIP3 expression was associated significantly with high hypoxia-inducible factor 1α (P = 0.003), carbonic anhydrase 9 (P = 0.04), and was inversely associated with bcl-2 expression (P = 0.009). High BNIP3 expression was a major independent factor for overall survival. Thus, high expression of a hypoxia regulated proapoptotic pathway was associated with a selection of an aggressive phenotype in vivo.

BNIP3 (19 kDa interacting protein) is a proapoptotic mitochondrial protein isolated through its interaction with bcl-2 and the adenovirus E1B19kDa proteins (1, 2). BNIP3 activates caspase-independent necrosis-like cell death by opening the mitochondrial permeability transition pore (3, 4). Experimental studies show that BNIP3 mRNA levels increase in response to hypoxia, and this effect is mediated through the hypoxiainducible factor 1α pathway activation (5, 6, 7). It is suggested that BNIP3 is involved in the induction of hypoxic necrosis in tumors (8, 9). Necrosis is linked to tumor aggressiveness, presumably due to selection of apoptosis resistance tumor clones within the hypoxic-necrotic environment and induction of genes involved in tumor growth by hypoxia-inducible factor α. BNIP3 activation, therefore, could contribute to the development of an aggressive tumor phenotype.

In the present study, we examine the immunohistochemical patterns of BNIP3 expression in non-small cell lung cancer and assessed the association of BNIP3 with hypoxia-inducible factor α overexpression, the expression of other hypoxia regulated molecules (CA9, DEC1, and LDH-5) and of angiogenic factors (vascular endothelial growth factor, platelet-derived endothelial cell growth factor, and basic fibroblast growth factor), as well as the expression of the bcl-2 protein. Furthermore, we examine the prognostic relevance of BNIP3 expression in patients with operable non-small cell lung cancer.

We examined 105 tumor samples from patients with early operable (T1, 2-N0, 1-M0 staged) non-small cell lung cancer (71 squamous and 34 adenocarcinomas). Paraffin-embedded material was obtained from the archives of the Department of Pathology, University of Oxford (Oxford, United Kingdom). The same material has been used in previous studies of ours on the prognostic role of various oncoproteins and angiogenesis-related parameters (10, 11, 12, 13, 14, 15, 16, 17, 18). Forty-four patients had T1-stage and 61 T2-stage disease. Node involvement (N1-stage) was present in 34 of 105 patients. Histological grade 3 (poorly differentiated neoplasms) was noted in 59 of 105, whereas the remaining cases were grouped in one category of grade 1/2 (well/moderate degree of differentiation). There were 85 male patients and 20 female, their ages ranging from 35 to 74 years (median, 63). All of the patients were treated with surgery alone without postoperative radiotherapy or chemotherapy. Patients dying within 60 days after operation were excluded, so as to avoid bias from perioperative death. Survival analysis (overall survival) was performed in 96 of 105 patients. Of 96 patients, 46 were alive at the time of the last follow-up. The follow-up of patients at the time of analysis ranged from 70 to 2500 days (median, 1233 days).

Assessment of BNIP3 Protein Expression.

The BNIP3 protein was assessed using a rabbit polyclonal antibody. The antibodies were made from a glutathione S-transferase fusion protein containing amino acids 1–149 of the BNIP3 protein. The rabbit serum was tested by Western blotting with recombinant protein for BNIP3, and the reactive band was removed by incubating the rabbit antiserum with the glutathione S-transferase-fusion protein. The antibody was also tested on skeletal muscle that showed high levels of BNIP3 mRNA. Details on the anti-BNIP3 antibody used have been published previously (7, 19).

Immunohistochemistry was performed at the Department of Pathology, Democritus University of Thrace. Sections were deparaffinized, and peroxidase was quenched with methanol and H2O2 3% for 15 min. Thereafter, slides were placed in antigen unmasking buffer (pH 6.0; code: TAR001; ILEM, Italy) and microwaving followed (3 × 4min). The primary antibody (1:400) was applied overnight at room temperature. After washing with Tris-buffered saline, sections were incubated with a secondary mouse antirabbit antibody (Kwik Biotinylated Secondary, 0.69A; Shandon-Upshaw, Pittsburgh, PA) for 15 min and washed in Tris-buffered saline. Kwik Streptavidin peroxidase reagent (039A Shandon-Upshaw) was applied for 15 min, and sections were again washed in Tris-buffered saline. The color was developed by 15-min incubation with 3,3′-diaminobenzidine solution, and sections were weakly counterstained with hematoxylin. Alveolar macrophages and chondrocytes were used a positive internal control. Normal rabbit immunoglobulin-G was substituted for the primary antibody as the negative control, at a concentration where immunostaining of control slides gave a faint cytoplasmic staining.

The percentage of cancer cells with cytoplasmic or nuclear BNIP3 reactivity was recorded after inspection of all of the fields in the tissue sample (A. G., E. S., K. C. G.). The percentage of positive cells was recorded in each individual field, and the final score for each case was the median value obtained.

Other Immunohistochemistry.

Table 1 shows the antibodies and details of the immunohistochemical procedures used to detect the expression of various oncoproteins and growth factors/receptors. Extensive report of the methods used has been published previously in referenced papers (10, 11, 12, 13, 14, 15, 16, 17, 18).

The percentage of cells with strong cytoplasmic and/or nuclear expression of hypoxia-inducible factor 1α and hypoxia-inducible factor 2α, of the lactate dehydrogenase LDH-5, and of the DEC1/STRA13 protein was recorded, and median value was used to define two groups of low versus high reactivity, as described previously (15, 17, 18). Similarly, the median percentage of cells with membrane and/or strong cytoplasmic CA9 reactivity was used to define two groups of low versus high CA9 expression (median 25%; Ref. 16). The median value of the percentage of cells with cytoplasmic reactivity was used to define two groups of low versus high vascular endothelial growth factor (median 70%), thymidine phosphorylase (median 50%), and basic fibroblast growth factor (median 80%) expression (11, 12, 13, 14). A 10% cancer cell positivity was required to score a case as positive for bcl-2 protein cytoplasmic expression and for p53 protein nuclear accumulation, which is the general accepted cutoff points for these antibodies (10).

Microvessel counting was used for angiogenesis assessment. For eye appraisal, sections were scanned at low power (×40 and ×100) and afterward at ×200 field so as to group cases in three vascular grade categories (low, medium, and high). The areas of the highest vascularization were chosen at low power (×100) and microvessel counting followed on three chosen ×200 fields of the highest density. The microvessel density was the sum of the vessel counts obtained in these three fields. Microvessels adjacent to normal lung were excluded from the appraisal. Vessels with a clearly defined lumen or well-defined linear vessel shape but not single endothelial cells were taken into account for microvessel counting. The 33rd and 66th percentile (25 and 45 vessels per ×200 optical field, respectively) was used to define three groups of low, medium, and high microvessel density (10).

Assessment of Necrosis.

The percentage of optical fields (×250) with necrosis was recorded by three observers separately (A. G., E. S., K. C. G.). Necrotic areas in >50% of the number of examined fields, which was the mean value of fields with necrosis given by the observers, was scored as extensive and, in <50%, as limited.

Statistical Analysis.

Statistical analysis and graphic presentation were performed using the GraphPad Prism 2.01 package (GraphPad, San Diego, CA). The Fisher’s exact test and the χ2-t test was used for testing relationships between categorical variables as appropriate. Spearman analysis was used to assess correlation between continues variables (percentage of positive cells in a sample). Survival curves were plotted using the method of Kaplan-Meier, and the log-rank test was used to determine statistical differences between life tables. A Cox proportional hazard model was used to assess the effects of patient and tumor variables on overall survival. A P ≤ 0.05 was considered significant.

BNIP3 Expression.

BNIP3 protein was expressed by the bronchial epithelium adjacent to tumors, exhibiting a predominantly cytoplasmic pattern of expression (Fig. 1,A). Nuclear expression of the BNIP3 was noted in chondrocytes of the bronchial cartilage (Fig. 1 A). Alveolar cells were negative, whereas alveolar macrophages exhibited cytoplasmic reactivity. The alveolar and peribroncheal vessels were negative.

The patterns of BNIP3 expression by tumoral cells were predominantly cytoplasmic (Fig. 1,B), although a varying extent of nuclear staining was noted in 16 cases (ranging from 10% to 50% of cells; median 10%; Fig. 1 C). In 6 of 16 cases the BNIP3 localization was exclusively nuclear, whereas in the remaining cases, nuclear expression coexisted with strong cytoplasmic reactivity. In 41 of 105 cases no cancer cell BNIP3 reactivity was noted. Cytoplasmic reactivity was noted in 64 of 105 cases, but strong cytoplasmic staining was noted in only 40 of 105 cases (ranging from 10% to 100% of cells; median 70%).

A scoring system was developed to classify the cases examined in two groups of low versus high BNIP3 reactivity. Briefly, lack of BNIP3 expression (41 of 105 cases) or weak cytoplasmic expression (24 of 105 cases) defined the group with “low” BNIP3 expression. Cases with strong cytoplasmic BNIP3 reactivity in >10% of cancer cells without (40 of 105 cases) or with nuclear (10 of 105 cases) expression or cases with pure nuclear reactivity (6 of 105 cases) were grouped as bearing “high” BNIP3 reactivity.

Strong BNIP3 cytoplasmic reactivity by tumor stroma fibroblasts was noted in 31of 105 cases (Fig. 1 D). Stroma BNIP3 reactivity was always associated with extensive BNIP3 expression by cancer cells (P < 0.0001). BNIP3 reactivity by the intratumoral vessels was noted in 27 of 105 cases. Tumor-associated lymphocytes were negative.

BNIP3 Association with Histological Variables.

There was no association of BNIP3 expression with tumor necrosis stage, histology type, and grade or with the degree of necrosis (Table 2).

BNIP3 Association with Hypoxia and Angiogenesis Markers.

Table 2 shows the association of BNIP3 expression by cancer cells with the expression of hypoxia-regulated proteins. A significant association of BNIP3 expression with hypoxia-inducible factor 1α, CA9, and LDH-5 reactivity was noted.

No association of BNIP3 with vascular endothelial growth factor, thymidine phosphorylase, basic fibroblast growth factor expression, or the vascular density was noted (Table 2).

BNIP3 Association with Proliferation and bcl-2.

BNIP3 expression by cancer cells was inversely related to the bcl-2 expression. Bcl-2 overexpression in BNIP3-expressing cells was rare (4 of 44 cases). BNIP3-expressing tumors also exhibited a significantly lower ki67 proliferation index (Table 2).

Survival.

In univariate analysis high BNIP3 expression was linked with poor overall survival (Fig. 2,A). Fig. 2,B shows analysis according to the nuclear/cytoplasmic staining subcellular patterns. Double stratification analysis with bcl-2 expression revealed a strong differential effect of these two proteins in the postoperative outcome of non-small cell lung cancer, the BNIP3 being strongly related to poor survival and the bcl-2 with unusually high rates of 5-year survival (30% versus 80%, respectively; Fig. 2,C). In multivariate analysis, BNIP3 expression was an independent marker of prognosis (Table 3).

Several experimental data provide evidence that BNIP3 is a proapoptotic protein induced under hypoxic stress (5, 6, 7). In situ hybridization analysis of RNA expression in human tumors showed that BNIP3 is expressed in human tumors in perinecrotic regions, although BNIP3 overexpression is also noted in some well-vascularized tumors, suggesting that stimuli other than hypoxia may be also involved in BNIP3 gene regulation (7).

The patterns of BNIP3 expression by normal and tumoral tissues and the role of BNIP3 in the prognosis of human malignancies are unknown. In the present study we investigated the expression of BNIP3 in a series of non-small cell lung carcinomas. Overexpression of BNIP3 in the cancer cell cytoplasm was evident in 40% of cases examined (whether of squamous or adenocarcinoma histotype), whereas nuclear localization of BNIP3 was also noted in a subset of carcinomas examined. Although no association of BNIP3 expression with the extent of tumor necrosis or the degree of vascularization was noted, a significant correlation of BNIP3 with several hypoxia-regulated proteins (hypoxia-inducible factor 1α, CA9, and LDH-5) was confirmed. This finding goes along with experimental evidence that BNIP3 is a hypoxia-inducible factor 1α target gene (7) and shows coordinate regulation of the hypoxia transcriptome.

Although to execute its function BNIP3 needs to interact with mitochondria, it is notable that there were two conditions in which it was nuclear in location. The first was in chondrocytes, which are in areas without blood supply and are hypoxic and indeed express hypoxia-inducible factor 1α. Nuclear location would be a mechanism to escape the apoptotic effects of this protein on these cells. In a small subset of the tumors it was either substantially or entirely nuclear. This could represent a selective mechanism to escape the apoptotic pathway. The number of patients is too small for a separate survival analysis to be significant, but their survival was slightly worse than the overall BNIP3-positive cases.

The cytoplasmic expression in normal tissues and cells that are not hypoxic shows that there must be additional pathways regulating BNIP3 and that there may be additional steps necessary for its activation as opposed to expression. Work on cardiac myocytes in hypoxia has shown that it is acidosis in addition to up-regulation of BNIP3 that can trigger apoptosis (20). It is of interest to note that LDH5 was coexpressed frequently with BNIP3 (62%) and with CA9 (55%). Because CA9 may protect from acidosis, tumors coexpressing CA9 and BNIP3 are expected to be very aggressive. This was indeed confirmed in survival analysis where tumors expressing both proteins (LDH5 and BNIP3 or CA9 and BNIP3) had poorer prognosis (although not significantly due to the low number of cases) with respect to tumors expressing only one of these proteins (P = 0.12 and 0.10, respectively).

Stromal expression of BNIP3 was also noted and was coordinately expressed with tumor BNIP3. Although the stroma contains the vasculature, this suggests it may be hypoxic in those cases, and several other hypoxia markers such as CAIX and hypoxia-inducible factor αs are reported to be expressed in tumor stroma.

BNIP3 belongs to the bcl-2 family of proteins and shares an apoptotic function presumably antagonizing the antiapoptotic function of bcl-2. BNIP3 is an activator of caspase-independent cell death pathway through mitochondrial permeability transition pore opening (3, 4). In the present study a significant inverse association between the expression of bcl-2 and BNIP3 proteins was noted, suggestive of the existence of two different lung cancer phenotypes, tumors with bcl-2 protein overexpression and tumors with a preferential up-regulation of BNIP3. These two groups of tumors define distinct clinical behaviors as, quite paradoxically, when tumors expressing the antiapoptotic protein bcl-2 have an indolent clinical course and when tumors with activated proapoptotic pathways (BNIP3 overexpression) are clinically aggressive. We expected that induction of apoptosis by this pathway would reduce the tumor bulk and aggressiveness. However, what is recognized from the work of Graeber et al.(21) is that hypoxia can act to select tumor cell populations that are better able to survive under an unfavorable microenvironment, to migrate, and contribute to greater aggression, which contributes to poor prognosis. However, it is unknown how hypoxia becomes a selection factor. We propose the BNIP3 pathway as being one of those downstream pathways by which hypoxia selects an aggressive phenotype.

In addition, tumors may escape the negative effects of hypoxia by compartmentalizing BNIP3 into an area where it is no longer able to induce apoptosis; for example, we found the nuclear localization in a subset of cases that had a particularly bad prognosis. This would represent a different mechanism whereby cells could escape the hypoxia-driven cell death and continue to proliferate. Another possibility to consider is that the more hypoxic a tumor, the greater the induction of hypoxia-inducible factor and that other pathways affecting tumor growth such as vascular endothelial growth factor and glucose transport are also more strongly switched on, and these are the contributing factors to the aggressive behavior. Therefore, strong BNIP3 induction may be a mark of other pathways in addition to itself and indicate strong hypoxia-inducible factor signaling.

BNIP3-expressing tumors have a significantly lower KI67 proliferation index, in keeping with hypoxia slowing down proliferation. Despite this slower proliferation there was an adverse outcome. This would suggest that being able to survive hypoxia, many of the pathways regulated by hypoxia-inducible factor would favor invasion and metastasis by mechanisms independent of the proliferation rate. Indeed, there is recent evidence that hypoxia-inducible factor mediates the suppression of proliferation, although it is associated with an aggressive phenotype and the effects of hypoxia-inducible factor transfection in experimental models show a more aggressive behavior (22, 23, 24).

One would initially have expected necrosis to correlate with BNIP3, but this was not the case. A possibility is that different mechanisms determine the development of large areas of cells dying versus individual cells, e.g., vascular occlusion or failure to develop a protective bypass pathway, e.g., phosphatidylinositol 3′-kinase activation. Thus, BNIP3 expression may define an evolutionary selection process for tumor growth to overcome the effects of this pathway. Although hypoxiainducible factor activation induces many genes of likely benefit for tumor growth (25), a proapoptotic gene is not one of them, so bypass pathways may be selected to overcome the detrimental effect of this gene expression allowing continued hypoxia-inducible factor activation of the other pathways rather than selecting against hypoxia-inducible factor itself. Understanding such pathways may allow therapeutic exploitation of endogenous BNIP3 (8).

Finally, the persistence of BNIP3 overexpression as an independent indicator of poor prognosis in different multivariate models certainly suggests that BNIP3 deserves additional investigation as a prognostic marker.

Grant support: Cancer Research UK and the Tumor and Angiogenesis Research Group.

Requests for reprints: Alexandra Giatromanolaki, Tumor and Angiogenesis Research Group, P.O. Box 12, Alexandroupolis 68100, Greece. Phone: 30-6932-480808; Fax: 30-25510-30349; E-mail: targ@her.forthnet.gr

Fig. 1.

A, strong cytoplasmic expression of BNIP3 by bronchial epithelium adjacent to BNIP3-positive lung adenocarcinoma; B, lung adenocarcinoma with mixed cytoplasmic/nuclear expression; C, nuclear BNIP3 expression in a squamous cell lung carcinoma; and D, strong BNIP3 expression by stroma fibroblasts adjacent to BNIP3-positive squamous cell lung cancer.

Fig. 1.

A, strong cytoplasmic expression of BNIP3 by bronchial epithelium adjacent to BNIP3-positive lung adenocarcinoma; B, lung adenocarcinoma with mixed cytoplasmic/nuclear expression; C, nuclear BNIP3 expression in a squamous cell lung carcinoma; and D, strong BNIP3 expression by stroma fibroblasts adjacent to BNIP3-positive squamous cell lung cancer.

Close modal
Fig. 2.

Kaplan-Meier overall survival curves according to the patterns of BNIP3 expression (A), according to the subcellular patterns of BNIP3 expression (low expression versus nuclear expression versus strong cytoplasmic expression; B), and after double stratification for BNIP3 and bcl-2 expression (C). A, solid line, BNIP3low (55pts); dashed line, BNIP3high (41pts); P = 0.004. B, solid line A, negative (53pts); thick solid line B, nuclear (14pts); dotted line C, cytoplasmic (29pts); A vs. B, P = 0.0007; A vs. C, P = 0.01. C, dashed line A, BNIP3neg/bcl-2neg (41pts); solid line B, BNIP3neg/bcl-2pos (14pts); dotted line C, BNIP3pos/bcl-2neg (41pts); A vs. B, P = 0.07; A vs. C, P = 0.002; B vs. C, P = 0.05.

Fig. 2.

Kaplan-Meier overall survival curves according to the patterns of BNIP3 expression (A), according to the subcellular patterns of BNIP3 expression (low expression versus nuclear expression versus strong cytoplasmic expression; B), and after double stratification for BNIP3 and bcl-2 expression (C). A, solid line, BNIP3low (55pts); dashed line, BNIP3high (41pts); P = 0.004. B, solid line A, negative (53pts); thick solid line B, nuclear (14pts); dotted line C, cytoplasmic (29pts); A vs. B, P = 0.0007; A vs. C, P = 0.01. C, dashed line A, BNIP3neg/bcl-2neg (41pts); solid line B, BNIP3neg/bcl-2pos (14pts); dotted line C, BNIP3pos/bcl-2neg (41pts); A vs. B, P = 0.07; A vs. C, P = 0.002; B vs. C, P = 0.05.

Close modal
Table 1

Details of the antibodies, dilutions, and antigen retrieval methods used in this study

Primary antibodyDilution/incubation timeAntigen retrievalSpecificitySourceRef.
ESEE 122 1:20 (90 min*MW HIF-1α Oxford University  15  
EP 190b Neat (90 min*MW HIF-2α Oxford University  15  
M75 1:20 MW CA9 Oxford University  16  
CW27 (90 min) MW DEC1 Oxford University  17  
Ab9002 1:200 (overnight) MW LDH-5 Abcam, UK  18  
VG1 1:4 (90 min*MW VEGF Oxford University  12, 13  
P-GF.44C 1:4 (60 min*No TP Oxford University  11  
FGF-2(147)-6 1:20 (90 min*MW bFGF Santa Cruiz Biothech.  14  
JC70 (CD31) 1:50 (30 min*Protease XXIV Endothelium Dako, Denmark  10  
124 1:80 overnight at 4°C MW bcl-2 Dako  10  
DO-7 1:30 overnight at 4°C MW p53 Dako  10  
Ki67 1:50 (90 min) MW Proliferation Dako  
Primary antibodyDilution/incubation timeAntigen retrievalSpecificitySourceRef.
ESEE 122 1:20 (90 min*MW HIF-1α Oxford University  15  
EP 190b Neat (90 min*MW HIF-2α Oxford University  15  
M75 1:20 MW CA9 Oxford University  16  
CW27 (90 min) MW DEC1 Oxford University  17  
Ab9002 1:200 (overnight) MW LDH-5 Abcam, UK  18  
VG1 1:4 (90 min*MW VEGF Oxford University  12, 13  
P-GF.44C 1:4 (60 min*No TP Oxford University  11  
FGF-2(147)-6 1:20 (90 min*MW bFGF Santa Cruiz Biothech.  14  
JC70 (CD31) 1:50 (30 min*Protease XXIV Endothelium Dako, Denmark  10  
124 1:80 overnight at 4°C MW bcl-2 Dako  10  
DO-7 1:30 overnight at 4°C MW p53 Dako  10  
Ki67 1:50 (90 min) MW Proliferation Dako  

Abbreviations: MW, microwave heating; HIF, hypoxia-inducible factor; VEGF, vascular endothelial growth factor; bFGF, basic fibroblast growth factor.

*

At room temperature.

Table 2

BNIP3 association with histology and molecular parameters

BNIP3
LowHighP value
Histology parameters     
 T-stage     
 28 16 0.42 
 33 28  
 N-stage     
 44 27 0.29 
 17 17  
 Histology     
 Adenocarcinoma 18 16 0.52 
 Squamous cell 43 28  
 Grade     
 1–2 28 18 0.69 
 33 26  
 Necrosis     
 No 16  
 Focal 20 16 0.62 
 Extensive 25 20  
Hypoxia-related parameters     
 HIF-1α     
 Low 28 0.003 
 High 33 36  
 HIF-2α     
 Low 32 17 0.17 
 High 29 27  
 HIF-1α/2α     
 Low 21 0.002 
 High 40 40  
 LDH-5     
 Low 38 17 0.01 
 High 23 27  
 CA9     
 Low 40 20 0.04 
 High 21 24  
 DEC1     
 Low 41 21 0.06 
 High 20 23  
Angiogenesis parameters     
 VEGF     
 Low 27 18 0.84 
 High 34 26  
 TP     
 Low 45 26 0.14 
 High 16 18  
 bFGF     
 Low 33 20 0.43 
 High 28 24  
 MVD     
 Low/Medium 37 30 0.66 
 High 18 12  
Proliferation and bcl-2     
 Ki67     
 Low 42 40 0.007 
 High 19  
 Bcl-2     
 Low 47 42 0.009 
 High 14  
BNIP3
LowHighP value
Histology parameters     
 T-stage     
 28 16 0.42 
 33 28  
 N-stage     
 44 27 0.29 
 17 17  
 Histology     
 Adenocarcinoma 18 16 0.52 
 Squamous cell 43 28  
 Grade     
 1–2 28 18 0.69 
 33 26  
 Necrosis     
 No 16  
 Focal 20 16 0.62 
 Extensive 25 20  
Hypoxia-related parameters     
 HIF-1α     
 Low 28 0.003 
 High 33 36  
 HIF-2α     
 Low 32 17 0.17 
 High 29 27  
 HIF-1α/2α     
 Low 21 0.002 
 High 40 40  
 LDH-5     
 Low 38 17 0.01 
 High 23 27  
 CA9     
 Low 40 20 0.04 
 High 21 24  
 DEC1     
 Low 41 21 0.06 
 High 20 23  
Angiogenesis parameters     
 VEGF     
 Low 27 18 0.84 
 High 34 26  
 TP     
 Low 45 26 0.14 
 High 16 18  
 bFGF     
 Low 33 20 0.43 
 High 28 24  
 MVD     
 Low/Medium 37 30 0.66 
 High 18 12  
Proliferation and bcl-2     
 Ki67     
 Low 42 40 0.007 
 High 19  
 Bcl-2     
 Low 47 42 0.009 
 High 14  

Abbreviations: HIF, hypoxia-inducible factor; VEGF, vascular endothelial growth factor; bFGF, basic fibroblast growth factor; MVD, microvessel density.

Table 3

Multivariate analysis of BNIP3 in comparison with histological features and markers of hypoxia/angiogenesis

Model 1Model 2Model 3
t-ratioP valuet-ratioP valuet-ratioP value
BNIP3 2.13 0.035 — — 2.92 0.004 
BNIP3/bcl-2 — — 2.40 0.018 — — 
2.02 0.047 1.79 0.077 2.69 0.008 
1.90 0.060 1.95 0.054 3.41 0.017 
Grade 0.53 0.594 0.40 0.683 — — 
Histology 0.32 0.744 0.34 0.731 — — 
BCL2 0.73 0.462 — — — — 
HIF-2α 0.92 0.355 0.97 0.334 — — 
HIF-1α 0.12 0.901 0.08 0.934 — — 
CA9 0.12 0.904 0.10 0.913 — — 
LDH 0.52 0.600 0.90 0.370 — — 
DEK 1.19 0.237 1.48 0.142 — — 
CD31 1.61 0.110 1.67 0.098 2.02 0.041 
VEGF 0.71 0.478 0.88 0.377 — — 
PD-ECGF 0.03 0.972 0.12 0.902 — — 
bFGF 0.14 0.883 0.13 0.891 — — 
KI67 0.03 0.972 0.07 0.938 — — 
Model 1Model 2Model 3
t-ratioP valuet-ratioP valuet-ratioP value
BNIP3 2.13 0.035 — — 2.92 0.004 
BNIP3/bcl-2 — — 2.40 0.018 — — 
2.02 0.047 1.79 0.077 2.69 0.008 
1.90 0.060 1.95 0.054 3.41 0.017 
Grade 0.53 0.594 0.40 0.683 — — 
Histology 0.32 0.744 0.34 0.731 — — 
BCL2 0.73 0.462 — — — — 
HIF-2α 0.92 0.355 0.97 0.334 — — 
HIF-1α 0.12 0.901 0.08 0.934 — — 
CA9 0.12 0.904 0.10 0.913 — — 
LDH 0.52 0.600 0.90 0.370 — — 
DEK 1.19 0.237 1.48 0.142 — — 
CD31 1.61 0.110 1.67 0.098 2.02 0.041 
VEGF 0.71 0.478 0.88 0.377 — — 
PD-ECGF 0.03 0.972 0.12 0.902 — — 
bFGF 0.14 0.883 0.13 0.891 — — 
KI67 0.03 0.972 0.07 0.938 — — 

NOTE. Model 1, including all variables examined; Model 2, considering the BNIP3 and bcl-2 expression as one combined factor; Model 3, considering BNIP3 expression together with the major T,N-pathology parameter and with angiogenesis.

Abbreviations: HIF, hypoxia-inducible factor; VEGF, vascular endothelial growth factor; bFGF, basic fibroblast growth factor.

1
Boyd JM, Malstrom S, Subramanian T, et al Adenovirus E1B 19 kDa and Bcl-2 proteins interact with a common set of cellular proteins.
Cell
,
79
:
341
-51,  
1994
.
2
Matsushima M, Fujiwara T, Takahashi E, et al Isolation, mapping, and functional analysis of a novel human cDNA (BNIP3L) encoding a protein homologous to human NIP3.
Genes Chromosomes Cancer
,
21
:
230
-5,  
1998
.
3
Chen G, Cizeau J, Vande Velde C, et al Nix and Nip3 form a subfamily of pro-apoptotic mitochondrial proteins.
J Biol Chem
,
274
:
7
-10,  
1999
.
4
Imazu T, Shimizu S, Tagami S, et al Bcl-2/E1B 19 kDa-interacting protein 3-like protein (Bnip3L) interacts with bcl-2/Bcl-xL and induces apoptosis by altering mitochondrial membrane permeability.
Oncogene
,
18
:
4523
-9,  
1999
.
5
Bruik RK. Expression of the gene encoding the proapoptotic BNIP3 protein is induced by hypoxia.
Proc Natl Acad Sci USA
,
97
:
9082
-7,  
2000
.
6
Chong TW, Horwitz LD, Moore JW, Sowter HM, Harris AL. A mycobacterial iron chelator, desferri-exochelin, induces hypoxiainducible factors 1 and 2, NIP3, and vascular endothelial growth factor in cancer cell lines.
Cancer Res
,
62
:
6924
-7,  
2002
.
7
Sowter HM, Ratcliffe PJ, Watson P, Greenberg AH, Harris AL. HIF-1-dependent regulation of hypoxic induction of the cell death factors BNIP3 and NIX in human tumors.
Cancer Res
,
61
:
6669
-73,  
2001
.
8
Kothari S, Cizeau J, McMillan-Ward E, Israels SJ, et al BNIP3 plays a role in hypoxic cell death in human epithelial cells that is inhibited by growth factors EGF and IGF.
Oncogene
,
22
:
4734
-44,  
2003
.
9
Vande Velde C, Cizeau J, Dubik D, et al BNIP3 and genetic control of necrosis-like cell death through the mitochondrial permeability transition pore.
Mol Cell Biol
,
20
:
5454
-68,  
2000
.
10
Giatromanolaki A, Koukourakis MI, Theodossiou D, Barbatis C, Harris AL, Gatter KC. Comparative evaluation of angiogenesis assessment with anti-Factor VIII and anti-CD31 immunostaining in non small cell lung cancer.
Clin Cancer Res
,
3
:
2485
-93,  
1997
.
11
Koukourakis M, Giatromanolaki A, O’Byrne K, et al Platelet-derived endothelial cell growth factor expression correlates with tumour angiogenesis and prognosis in non small cell lung cancer.
Br J Cancer
,
75
:
477
-81,  
1997
.
12
Giatromanolaki A, Koukourakis MI, Kakolyris S, et al Vascular endothelial growth factor, wild-type p53 and angiogenesis in early operable non-small cell lung cancer.
Clin Cancer Res
,
4
:
3017
-24,  
1998
.
13
Koukourakis MI, Giatromanolaki A, Thorpe PE, et al Vascular endothelial growth factor/KDR activated micorvessel density versus CD31 standard microvessel density in non-small cell lung cancer.
Cancer Res
,
60
:
3088
-95,  
2000
.
14
Giatromanolaki A, Koukourakis MI, Sivridis E, et al Co-expression of MUC1 glycoprotein with multiple angiogenic factors in non-small cell lung cancer suggets co-activation of angiogenic and migratory pathways.
Clin Cancer Res
,
6
:
1917
-21,  
2000
.
15
Giatromanolaki A, Koukourakis MI, Sivridis E, et al Relation of hypoxia inducible factor 1 alpha and 2 alpha in operable non-small cell lung cancer to angiogenic/molecular profile of tumours and survival.
Br J Cancer
,
85
:
881
-90,  
2001
.
16
Giatromanolaki A, Koukourakis MI, Sivridis E, et al Expression of hypoxia-inducible carbonic anhydrase-9 relates to angiogenic pathways and independently to poor outcome in non-small cell lung cancer.
Cancer Res
,
61
:
7992
-8,  
2001
.
17
Giatromanolaki A, Koukourakis MI, Sivridis E, et al DEC1 (STRA13) protein expression relates to hypoxia- inducible factor 1-alpha and carbonic anhydrase-9 overexpression in non-small cell lung cancer.
J Pathol
,
200
:
222
-8,  
2003
.
18
Koukourakis MI, Giatromanolaki A, Sivridis E, et al Tumour and Angiogenesis Research Group. Lactate dehydrogenase-5 (LDH-5) overexpression in non-small-cell lung cancer tissues is linked to tumour hypoxia, angiogenic factor production and poor prognosis.
Br J Cancer
,
89
:
877
-85,  
2003
.
19
Sowter HM, Ferguson M, Pym C, et al Expression of the cell death genes BNip3 and NIX in ductal carcinoma in situ of the breast; correlation of BNip3 levels with necrosis and grade.
J Pathol
,
201
:
573
-80,  
2003
.
20
Kubasiak LA, Hernandez OM, Bishopric NH, Webster KA. Hypoxia and acidosis activate cardiac myocyte death through the Bcl-2 family protein BNIP3.
Proc Natl Acad Sci USA
,
99
:
12825
-30,  
2002
.
21
Graeber TG, Osmanian C, Jacks T, et al Hypoxia-mediated selection of cells with diminished apoptotic potential insolid tumours.
Nature
,
379
:
88
-91,  
1996
.
22
Blancher C, Moore JW, Talks KL, Houlbrook S, Harris AL. Relationship of hypoxia-inducible factor (HIF)-1alpha and HIF-2alpha expression to vascular endothelial growth factor induction and hypoxia survival in human breast cancer cell lines.
Cancer Res
,
60
:
7106
-13,  
2000
.
23
Krishnamachary B, Berg-Dixon S, Kelly B, et al Regulation of colon carcinoma cell invasion by hypoxia-inducible factor 1.
Cancer Res
,
63
:
1138
-43,  
2003
.
24
Akakura N, Kobayashi M, Horiuchi I, et al Constitutive expression of hypoxia-inducible factor-1alpha renders pancreatic cancer cells resistant to apoptosis induced by hypoxia and nutrient deprivation.
Cancer Res
,
61
:
6548
-54,  
2001
.
25
Harris AL. Hypoxia–a key regulatory factor in tumour growth.
Nat Rev Cancer
,
2
:
38
-47,  
2002
.