Re:Correlation of hypoxia-inducible factor-1 alpha (HIF-1α) and vascular endothelial growth factor (VEGF) expressions with clinico-pathological featu

[문형철]

BEYOND REASON

Correlation of hypoxia-inducible factor-1 alpha (HIF-1α) and vascular endothelial growth factor (VEGF) expressions with clinico-pathological features of oral squamous cell carcinoma (OSCC)

Author links open overlay panelHebaEl-Sayed Mohammed YoussefNahed EmadEldeen Abo-AzmaEman MohieEldeen Megahed

https://doi.org/10.1016/j.tdj.2015.05.010

Get rights and content

Under a Creative Commons license

open access

Abstract

Oral squamous cell carcinoma (OSCC) is a characteristic aggressiv e tumor representing a significant public health threat all over the world. Up to 50–60% of solid tumors may exhibit hypoxic tissue areas that are heterogeneously distributed within the tumor stroma. Hypoxia inducible factor-1α (HIF-1α) is a central regulator in the adaptive cellular response to hypoxia; however its role is still uncovered. Vascular endothelial growth factor (VEGF) is an important molecule playing crucial role not only in inflammation, but also in angiogenesis and thus tumor growth and progression. So, the aim of this work is to investigate the expression‎ of HIF-1α and VEGF in OSCC and correlate their expressions with clinical and histopathological features of OSCC. In this work, tissue specimens from a total of 45 cases with OSCC were stained immunohistochemically with HIF-1α and VEGF antibodies and examined microscopically.

The results of this study revealed that HIF-1α and VEGF expressions appeared to be significantly positive and directly related to histopathological grades, lymph node (LN) status as well as clinical stages of OSCC. Moreover, HIF-1α expression‎ was significantly correlated to angiogenic activity measured by VEGF immunostaining. Thus, it was concluded that HIF-1α and VEGF expressions were up-regulated with increased malignancy and can be used as predictive markers of tumor behavior.

• Previous article in issue
• Next article in issue

Keywords

OSCC

HIF-1α

VEGF

1. Introduction

Worldwide, oral cancer (OC) is a public health problem, representing the sixth most common malignant neoplasm. Oral squamous cell carcinoma (OSCC) is considered the most frequent of all oral neoplasms representing about 95% of these neoplasms [1].

The process of tumorigenesis involves numerous changes in tumor microenvironment (TME) that lead to the production of new components and/or upregulation of its molecules that play significant roles in tumor progression. These changes that occur in OSCC progressively increase the ability of transformed cells to proliferate and invade. The heterogeneity of these changes explains why tumors at the same clinical stage and localization often show significant differences in their clinical outcomes and treatment responses [2], [3].

Hypoxia is one of the most common changes that occur in the TME and it is a common characteristic of human solid tumors, and it can drive malignant cells to undergo adaptive changes that enable them to survive in oxygen-depleted regions. Hypoxia inducible factor-1(HIF-1) is a transcription factor composed of two subunits namely HIF-1α and HIF-1β. HIF-1β is a constitutively nuclear-located subunit that is found under all oxygen conditions. However, HIF-1α is a highly regulated subunit that is extremely labile in normoxia with a half-life of less than 5 min. The rapid and continuous degradation of HIF-1α is effectively blocked if oxygen availability is reduced. This permits dimerization with HIF-1β to form the active HIF-1 complex [4], [5].

HIF-1α contains a unique oxygen-dependent degradation domain (ODDD) that is central to the oxygen-regulated stability of this protein. Specific degradation of HIF-1α in normoxia is triggered through this domain. Therefore, hypoxia increases HIF-1 levels in cells by strongly stabilizing HIF-1α through the inhibition of its rapid degradation by the proteasome [5]. These striking criteria in oxygen sensitivity & trans-activation ability of HIF-1α make it the main functional protein of HIF-1 complex [4]. More than 60 genes have been found to be induced by HIF-1α that are implicated in many different cellular functions such as cell survival, cell proliferation, apoptosis, glucose metabolism and angiogenesis [6].

Many studies have shown that the immunohistochemical (IHC) expression‎ of HIF-1α is associated with a poor prognosis, however others have only shown a trend for poor prognosis and some have been not been statistically significant. In three studies, in head and neck (HN) cancer, OC and non-small cell lung cancer, HIF-1α expression‎ predicted a good prognosis [7], [8], [9].

One of the studies where HIF-1α expression‎ was a good prognostic factor showed a strong correlation between HIF-1α and the expression‎ of pro-apoptotic factors such as caspase-3, Fas, and Fas ligand [9].

In addition to prognosis, HIF-1α expression‎ was also investigated in some studies on esophageal cancer patients to predict response to chemo-radiotherapy. HIF-1α expression‎ predicted a poor treatment response [5], [10].

Liang et al. [11], has reported that HIF-1α overexpression‎ was correlated significantly with the clinical stage, histopathological grade, lymph node (LN) metastasis and worse prognosis in patients with OSCC of the tongue. Furthermore, it was found that HIF-1α over-expression‎ in combination with deficiency or mutation of tumor suppressor genes such as von Hipple Lindu (vHL), p53, and amplification of oncogenes as Akt and Ras was frequently seen in human cancer and these genetic alterations have been associated with tumor growth, invasion and metastasis. Most cancers over-expressing HIF-1α are associated with increased mortality [12].

Angiogenesis is also essential for the development of solid tumors and the survival of their cells. Tumor cells that are located more than 100 mm away from blood vessels become hypoxic. If new blood vessels do not form, tumor clones will be confined within 1–1.5 mm3 diameter. Such clones remain dormant from months to years before they switch to an angiogenic phenotype [13]. In the beginning, angiogenesis develops by incorporating existing host blood vessels; no solid tumors can probably grow more than 2 mm3 unless they synthesize their own network of new microvessels. This angiogenic switch is dependent mainly on hypoxia and oncogenic transformation which occurs in tumor cells, or in an autocrine growth factor loop [14].

Vascular endothelial growth factor (VEGF) is considered a potent angiogenic activator and its associate receptor (VEGFR-2). VEGF is a member of platelet-derived growth factor (PDGF) superfamily of growth factors that triggers mobility and maturation of endothelial cells (ECs) towards the hypoxic environment that allows blood and oxygen to reach hypoxic cells [15]. It has long been established that ECs of tumor-associated neovasculature proliferate 20–2000 times more rapidly than ECs of normal tissue [16].

Macluskey et al. [17] and Carlile et al. [18], observed a significant increase in vascularity during the transition from normal oral mucosa, through different degrees of dysplasia, to invasive carcinoma. Several studies have also reported a close association between tumor angiogenesis and tumor progression from early to late OSCC [19], [20]. However, conflicting results were reported relating to HIF-1α and VEGF expressions in HN cancers.

So, the aim of the present work is to investigate the expression‎ and the distribution patterns of HIF-1α and VEGF immunostainings in OSCC and to correlate their expressions with the clinico-pathological features of OSCC to clarify their role in the biological behavior of this tumor.

2. Material and methods

-

The present study included 45 cases with OSCC who were diagnosed and subjected to surgery at Tanta Cancer Center and Oral and Maxillofacial Surgery Department, Faculty of Dentistry, Tanta University. Neither pre-operative chemotherapy nor radiotherapy was performed. Informed consent was obtained from all the patients examined in this study. Detailed case history and thorough clinical examination with clinical staging was performed for each case.

-

The fresh tissue specimens of OSCC were fixed in 10% neutral buffered formalin and routinely processed for hematoxylin and eosin (H&E), examined under light microscope and graded histologically.

-

For IHC study, the Avidin–Biotin Complex (ABC) technique was used to detect the following reagents:

-

Primary antibodies against HIF-1α (Clone: 54/HIF-1α, Mouse monoclonal, Becton Dickinson (BD) Transduction Laboratories, Franklin Lakes, USA) and VEGF (Clone: SP6, Rabbit monoclonal, Neomarkers, Fremont CA, USA). The Avidin Biotin Complex (ABC) Universal Kit (Neomarkers, Fremont CA, USA) was used.

-

The immunostaining procedure was performed according to the manufacturer' instructions. The deparaffinized sections associated with immunostaining kit were processed acting as positive controls whereas negative controls consisting of tissue sections on which primary antibody was replaced with non-immune serum was also performed.

-

The sections were eval‎uated for HIF-1α and VEGF expressions by assessing the site of staining (nuclear and/or cytoplasmic). Scoring the IHC positivity was done using two parameters; staining intensity and percentage of stained positive cells within 4 different microscopic fields of the same section at X200 magnification as follows: Percentage of positive cells: (0) if the tumor cells showed no reactivity of antigen (1), (Focal) if <30% of the tumor cells are positive (2), (Partial) if 30–60% of the tumor cells are positive (3), (Diffuse) if >60% of the tumor cells are positive. The intensity: 1 = [weak], 2 = [moderate], 3 = [strong].

-

Final score for each case was the sum of both intensity and percentage and was represented as follows: Negative = (0), Weak = (2–3), Moderate = (4), Strong = (5–6).

-

All IHC eval‎uation was performed by two independent observers blind to clinical information of the cases and the average of both observers' scoring was used.

-

Statistical analysis was done using Monte Carlo exact test by SPSS program V.19 and Spearman's rank correlation coefficient (rho). The p value for significance was adopted at P < 0.05.

3. Results

3.1. Clinical findings

-

All the observed clinical findings of OSCC cases are showed in Table 1. It was found that the mean age of the studied cases was 54 years old and the tongue was the most commonly affected site (17 cases) followed by buccal mucosa (10 cases). Twenty cases were classified clinically as advanced stages (Stage III and IV) and 22 cases of OSCC appeared as exophytic fungating masses with or without surface ulcerations (Fig. 1A), while 17 cases appeared as endophytic ulcerative lesions (Fig. 1B). The remaining 6 cases appeared clinically as white and/or red lesions (Fig. 1C). Thirteen cases showed LN metastasis and 2 cases recurred locally.

Table 1. Clinical findings of OSCC.

Clinical data	Range	Mean ± SD
Age (years)	28–76	54.5 ± 11.18
	n	(%)
Sex
Male	24	(53.3)
Female	21	(46.7)
Size of primary tumor
<2 cm	13	(28.8)
2–4 cm	16	(35.6)
>4 cm	16	(35.6)
Site of primary tumor
Tongue	17	(37.8)
Buccal mucosa	10	(22.2)
Gingiva	5	(11.1)
Lip	4	(8.9)
Palate	4	(8.9)
Floor of the mouth	3	(6.7)
Alveolar mucosa	2	(4.4)
Lymph node involvement	13	(28.9)
Clinical stage (UICC)
Stage I	13	(28.9)
Stage II	12	(26.7)
Stage III	12	(26.7)
Stage IV	8	(17.7)
Clinical presentation
Exophytic	22	(48.9)
Endophytic	17	(37.8)
Red and/or white lesion	6	(13.3)
Duration
<12 months	20	(44.4)
>12 months	25	(55.6)
Local recurrence	2	(4.4)
Total	45	(100)

1. Download : Download full-size image

Fig. 1. Intra-oral photographs are showing:(A) OSCC appears as exophytic mass at the right buccal mucosa with shallow ulceration. (B) Deep malignant ulcer of OSCC at the left postero-lateral side of the tongue. (C) A case of OSCC appears as mixed red and white lesion at the floor of the mouth.

1. Download : Download full-size image

Fig. 2. Analysis of the relation between histopathological grades and different clinical stages of OSCC.

3.2. Conventional hematoxylin and eosin staining (H&E)

-

By light microscopic examination, 19 cases out of the 45 cases of OSCC were diagnosed as well-differentiated (Grade I) (Fig. 5a), 17 cases were diagnosed as moderately differentiated (Grade II) (Fig. 6a) and the remaining 9 cases were poorly differentiated tumors (Grade III) (Fig. 7a).

1. Download : Download full-size image

Fig. 3. Analysis of HIF-1α expression‎ in different clinical stages of OSCC.

1. Download : Download full-size image

Fig. 4. Analysis of VEGF expression‎ in different clinical stages of OSCC.

1. Download : Download full-size image

Fig. 5. Well-differentiated OSCC (a) Epithelial and keratin pearls invading the connective tissue (H&E, ×200), (b) Localized weak cytoplasmic HIF-1α expression‎ around keratinized area (ABC, ×200), (c) Localized weak HIF-1α reactivity at the peripheral cell layer of epithelial pearls (ABC, ×200), (d) Negative HIF-1α immunostaining (ABC, ×200), (e) Weak endothelial HIF-1α at the tumor stroma (Red arrows) (ABC, ×400), (f) Localized weak cytoplasmic VEGF immunostaining of tumor nests as well as focal VEGF staining of stromal cells (Blue arrow) (ABC, ×200).

1. Download : Download full-size image

Fig. 6. Moderately-differentiated OSCC: (a) Malignant epithelial nests invading the connective tissue (H&E, ×200). (b) Localized cytoplasmic HIF-1α expression‎ (ABC, ×200). (c) Diffuse strong nuclear HIF-1α expression‎ of tumor cells of OSCC (ABC, ×200). (d) The same previous field of OSCC shows strong reactivity of VEGF of both tumor and stromal cells (ABC, ×200). (e) Weak nuclear and cytoplasmic HIF-1α immunostaining of OSCC cells (ABC, ×200). (f) Weak VEGF expression‎ of both tumor and stromal cells in the same previous case (ABC, ×200). (g) HIF-1α immunoreactivity of endothelial cells (ECs) of the tumor stroma (ABC, ×400). (h) Strong VEGF staining of the stromal inflammatory and vascular ECs (ABC, ×400).

1. Download : Download full-size image

Fig. 7. Poorly-differentiated OSCC, (a) Individual malignant epithelial cells invading the connective tissue with prominent hyperchromatism and pleomorphism (H&E, ×200), (b) Diffuse strong nuclear and cytoplasmic HIF-1α expression‎ of tumor cells (ABC, ×400), (c) Diffuse VEGF reactivity of malignant cells and stromal cells (ABC, ×200), (d) Weak HIF-1α immunostaining in a case classified clinically as stage-I (ABC, ×200), (e) Weak VEGF staining in the previous stage-I case (ABC, ×400), (f) Strong VEGF immunostaining of stromal fibroblasts at the deep invasive front where muscles are invaded by malignant cells (Blue arrow) ((ABC, ×200).

-

It was observed that 14 out of 25 cases with early stages of OSCC (Stage I and II) were well-differentiated histologically and only 3 cases were poorly differentiated tumors. Five cases of OSCC out of 20 with advanced clinical stages (Stage III and IV) were diagnosed as well-differentiated tumors, while 6 cases were poorly differentiated (Table 2) (Fig. 2). The clinical stages showed non-significant correlation to the histopathological grades (p = 0.067, rho = 0.276) (i.e. the lesser degree of differentiation of the tumor cells is not essentially associated with more advanced clinical stage of OSCC).

Table 2. Histopathological grades in relation to the different clinical stages of OSCC.

Clinical stages	Histopathological grades						Total
	Well differentiated		Moderately differentiated		Poorly differentiated
	n	(%)	n	(%)	n	(%)	n	(%)
Stage I	6	(46)	5	(38.7)	2	(15.3)	13	(28.9)
Stage II	8	(66.7)	3	(25)	1	(8.3)	12	(26.7)
Stage III	4	(33.3)	5	(41.7)	3	(25)	12	(26.7)
Stage IV	1	(12.5)	4	(50)	3	(37.5)	8	(17.7)
Total	19	(42.2)	17	(37.8)	9	(20)	45	(100)

3.3. Immunohistochemical staining

3.3.1. Correlation of antigen reactivity to the clinical findings

-

Regarding the age of the patients, most cases of OSCC at the 7th and 8th decades of life showed high (moderates and strong) HIF-1α and VEGF reactivities representing 68.7% (11 out of 16 cases) and 81% (13 cases) for each antigen respectively. However, all patients at the 3rd and 4th decades of life (4 cases) showed weak expressions of HIF-1α and VEGF except one case of them revealed no positivity of HIF-1α. OSCC cases from the tongue showed the highest levels of HIF-1α and VEGF stainings representing 67% and 35% respectively. However, the correlation was not significant between the site of OSCC and expressions of the studied antigens (Table 3, Table 4).

Table 3. Expression‎ of HIF-1α in relation to the age and site of OSCC cases.

Age range (years)	HIF-1α expression‎								Total
	Negative		Weak		Moderate		Strong
	n	(%)	n	(%)	n	(%)	n	(%)	n
20–39	1	(25)	3	(75)	–	–	–	–	4
40–59	–	–	12	(48)	8	(32)	5	(20)	25
60–69	1	(8.3)	4	(33.3)	4	(33.3)	3	(25)	12
70–76	–	–	–	–	–	–	4	(100)	4
Site of OSCC	n	(%)	n	%)	n	(%)	n	(%)	Total
Tongue	1	(5.9)	6	(35.3)	3	(17.6)	7	(41.2)	17
Buccal mucosa	–	–	6	(60)	3	(30)	1	(10)	10
Gingiva and Alveolar mucosa	1	(14.3)	2	(28.6)	4	(57.1)	–	–	7
Lip	–	–	2	(50)	1	(25)	1	(25)	4
Palate	–	–	2	(50)	1	(25)	1	(25)	4
Floor of the mouth	–	–	1	(33.3)	–	–	2	(66.7)	3

Table 4. VEGF expression‎ in relation to the age and site of OSCC cases.

Age range (years)	VEGF expression‎						Total
	Weak		Moderate		Strong
	n	%	n	%	n	%	n
30–39	4	(100)	–	–	–	–	4
40–59	12	(48)	7	(28)	6	(24)	25
60–69	3	(23)	7	(53.8)	2	(15.2)	13
70–74	–	–	2	(50)	2	(50)	4
Site of OSCC	n	%	n	%	n	%	Total
Tongue	6	(35.3)	6	(35.3)	5	(29.4)	17
Buccal mucosa	4	(40)	4	(40)	2	(20)	10
Gingiva and Alveolar mucosa	3	(42.9)	3	(42.9)	1	(14.3)	7
Lip	2	(50)	2	(50)	–	–	4
Palate	2	(50)	1	(25)	1	(25)	4
Floor of the mouth	1	(33.3)	1	(33.3)	1	(33.3)	3

-

HIF-1α immunostaining increased significantly with advanced clinical stages of OSCC (p = 0.002), hence 11 out of 20 cases with stage III and IV showed strong HIF-1α reactivity, while only 1 case out of 25 with stages I and II showed marked upregulation of HIF-1α expression‎ (Table 5) (Fig. 3). Similarly, VEGF reactivity was significantly correlated to clinical stages of OSCC (p = 0.002), since VEGF was weak in 11 cases of stage I cases. However, 6 cases of stage IV-OSCC showed strong VEGF reactivity and 2 cases were moderately-stained with VEGF, while no cases showed weak VEGF (Table 5) (Fig. 4).

Table 5. HIF-1α and VEGF expressions in clinical stages and histo-pathological grades of OSCC.

Antigen expression‎		Clinical stages								Histopathological grades
		Stage I		Stage II		Stage III		Stage IV		Grade I		Grade II		Grade III
		n	%	n	%	n	%	n	%	n	%	n	%	n	%
HIF-1α	Negative	1	7	1	8	–	–	–	–	1	5	1	5	–	–
	Weak	10	8	6	50	4	33	–	–	12	63	6	35	2	22
	Moderate	2	15	4	34	4	33	1	12	4	21	5	30	2	22
	Strong	–	–	1	8	4	34	7	88	2	11	5	30	5	56
VEGF	Negative	–	–	–	–	–	–	–	–	–	–	–	–	–	–
	Weak	11	85	5	42	3	25	–	–	12	63	5	30	2	22
	Moderate	2	15	7	58	4	33	2	25	5	26	6	35	4	44
	Strong	–	–	–	–	5	42	6	75	2	11	6	35	3	34

-

HIF-1α and VEGF expressions were significantly correlated to LN involvement (p = 0.001 for each) as 9 cases out of 13 with LN metastasis (69%) showed strong staining of both antigens, while 65% of cases (21 out of 32 cases) without any nodal involvement showed negative or weak HIF-1α and VEGF immunostainings. Strong HIF-1α expression‎ was also observed in cases that showed local recurrence.

3.3.2. Correlation of antigen reactivity to the histo-pathological findings

3.3.2.1. Well-differentiated OSCC (grade I, 19 cases)

-

HIF-1α immunoreactivity appeared as a localized or focal granular cytoplasmic and/or nuclear brown staining of mainly the peripheral malignant cell layers in the keratin and epithelial pearls. It was more pronounced around areas of necrosis and keratinization (Fig. 5b and c). In most cases, HIF-1α expression‎ appeared more intense in areas away from the closest blood vessels.

-

Cytoplasmic HIF-1α was predominant than nuclear staining or may appear the only expression‎ pattern in 12 cases of well-differentiated OSCC (63%) (Fig. 6b). Almost all of these cases (11 cases) were clinically classified as stage I and II. HIF-1α staining was not positive in only one case (Fig. 5d). HIF-1α expression‎ was also seen in vascular ECs of the tumor stroma in 9 cases (47%) (Fig. 5e).

-

VEGF expression‎ appeared mostly as localized cytoplasmic brown staining of the tumor cells with higher intensity at the outer cell layers and less intense toward the center. In addition, stromal cells including inflammatory cells, ECs, fibroblasts also demonstrated cytoplasmic VEGF reactivity especially those cells at the tumor–stroma interface (Fig. 5f). Generally, it was observed that 12 cases of well-differentiated tumors showed weak immunostaining of HIF-1α and VEGF. Only 2 cases revealed strong reactivity of the studied antigens and these cases were classified clinically as stage III and IV with LN metastasis (Fig. 8, Fig. 9) (Table 5).

1. Download : Download full-size image

Fig. 8. Analysis of HIF-1α expression‎ in different histopathological grades of OSCC.

1. Download : Download full-size image

Fig. 9. Analysis of VEGF expression‎ in different histopathological grades of OSCC.

3.3.2.2. Moderately-differentiated OSCC (grade II, 17 cases)

-

Few cases showed localized HIF-1α immunoreactivity either nuclear or cytoplasmic (Fig. 6b), while most cases of moderately-differentiated OSCC demonstrated diffuse HIF-1α expression‎ throughout the entire tumor which was independent of vessel proximity. Nuclear HIF-1α was predominant than cytoplasmic staining in 14 cases (82.3%) (Fig. 6c). Almost all of these cases (16 cases) were clinically classified as stage III and IV. Endothelial HIF-1α immunostaining was also seen in 12 cases (70%) (Fig. 6g). VEGF reactivity appeared mostly as diffuse moderate brown cytoplasmic staining in the malignant cells with higher intensity at the outer cell layers (Fig. 6d and f).

-

By analyzing the expression‎ level of HIF-1α in moderately-differentiated OSCC, it was found that 10 cases demonstrated high scores of immunoreactivity (5 cases were moderate and 5 cases were strong) (Fig. 6, Fig. 8), while 6 cases showed weak HIF-1α immunoreactivity (Fig. 6, Fig. 8). Only one case was not positive for HIF-1α expression‎. On the other hand, VEGF expression‎ showed high scores of staining in 12 cases (6 were moderate and 6 cases were strong) (Fig. 6d). The remaining 5 cases revealed weak reactivity for VEGF (Fig. 6, Fig. 9) (Table 5).

-

VEGF immunostaining was also observed markedly in the stromal fibroblasts, inflammatory cells and vascular ECs (Fig. 6h).

3.3.2.3. Poorly-differentiated OSCC (grade III, 9 cases)

-

HIF-1α expression‎ in most cases of poorly-differentiated OSCC (7 out of 9 cases) appeared mainly as intense diffuse nuclear reaction of the invading malignant cells, with or without cytoplasmic localization (Fig. 7b). Only one case showed cytoplasmic HIF-1α reactivity without nuclear pattern of expression‎.

-

As shown in Table 5, the majority of poorly-differentiated specimens (7 cases) showed high HIF-1α and VEGF expressions (moderate and strong stainings) (Fig. 7b and c). Whereas, only few cases showed weak expression‎ levels of these antigens (2 cases out of 9 for both HIF-1α and VEGF) (Fig. 8, Fig. 9). Interestingly, the cases that demonstrated weak expression‎ of the studied antigen were classified clinically as stage I and II (Fig. 7d–e). Strong VEGF reactivity was also observed in the stromal cells especially fibroblasts at the deep invasive front (Fig. 7f). Generally, in all the studied 45 cases of OSCC HIF-1α and VEGF expressions showed significant correlation with the degree of differentiation of OSCC (p = 0.001 for each antigen) (Table 6). Moreover, HIF-1α expression‎ was significantly correlated to tumor cell angiogenesis measured by VEGF reactivity (p = 0.001) (Table 7).

Table 6. Correlation between histopathological grade, HIF-1α and VEGF expressions.

Variables	Histopathological grade
	rho	p
HIF-1α expression‎	0.572	0.001*
VEGF expression‎	0.528	0.001*

Monte Carlo Exact Test: p = 0.001 (*Significant).

Table 7. Correlation between HIF-1α and VEGF expressions in OSCC.

Variables	HIF-1α expression‎
	rho	p
VEGF expression‎	0.859	0.001*

Monte Carlo Exact Test: p = 0.001 (*Significant).

4. Discussion

OSCC remains one of the most difficult malignancies to control and there has been little improvement in the survival rate over the past 20 years. Despite many efforts, precise assessment of progression and behavior of OSCC with a view to predict prognosis remains an elusive goal.

This study showed that men were more affected than women; however, the male: female ratio was less pronounced (1.2:1). Similar trends have been reported in literature regarding the gender with increasing incidence in women that has been attributed to increased participation of women in risk factor behaviors such as smoking and alcohol consumption in recent decades [21], [22]. In this study, Tongue was the most affected site (37.8%) and mostly associated with metastasis (53.8%). This is supported by previous studies that showed the same results [23], [24], [25] and may be explained by the rich blood supply and lymphatic drainage of the tongue.

The present study revealed a non-significant direct correlation between clinical stages and the degree of differentiation of OSCC. These results are consistent with the findings of Dantas et al. [26], who reported the same results in OSCC of the tongue with the tumor size fails to signify any information related to the actual histological condition of the tumor.

Hypoxia is a common characteristic of human solid tumors, and it can drive malignant cells to undergo adaptive changes that enable them to survive in oxygen-depleted regions. It has been suggested that tumor hypoxia plays an essential role in promoting chromosome instability, cancer cell invasiveness, and metastasis. It is associated with aggressive tumor growth and treatment failure in several human solid tumors [2], [3].

When correlating HIF-1α reactivity with clinical features of the studied cases regarding the age of the patients, the correlation was positive. This may be due to mutations that are more common to occur in older people. Consequently, these mutations might activate oncogenes that encourage the expression‎ of many transcriptional factors such as HIF-1α and their activation [27].

Angiogenesis is another adaptive response of tumor cells to hypoxia. VEGF is a multifunctional cytokine expressed by both tumor and stromal cells. Its angiogenic effect is based on its ability to induce various responses by ECs during vascular development, including cell proliferation, migration, specialization and survival [28].

In the current study, enhanced VEGF expression‎ also appeared in old-aged patients than in young patients. This could be due to the epigenetic changes that are increasingly recognized as a part of aging as hypermethylation of p53 genes. This may lead to gene silencing and suppress its function, thus facilitate VEGF expression‎ to escape the control of p53 suppressor gene [27], [29].

Furthermore, the observed significant correlation between HIF-1α, clinical stages and LN status of OSCC in this study was consistent with that reported by Eckert et al. [30] and Tilakaratne and Nissanka-Jayasuriya. [31], who found a positive correlation between HIF-1α, tumor size, recurrences, nodal and distant metastasis. This may be a result of upregulation of matrix metalloproteinases (MMPs) in hypoxic OSCC that are responsible for degradation of the extracellular matrix (ECM). These findings are indicative that HIF-1α overexpression‎ in OSCC might be responsible for enhancement of tumor invasion as well as metastasis, resulting in poor prognosis of OSCC [32]. Furthermore, Ryu et al. [33], postulated that HIF-1α accumulation in tumor cells modulates α2-integrin and fibronectin (invasion-related factors) favoring invasion by enhancement of new cell-ECM attachments by fibronectin-rich ECM deposition.

Imai et al. [34], also observed downregulation of E-cadherin expression‎ in OSCC cells under hypoxic conditions that leads to loss of cellular cohesiveness, consequently resulting in more invasive and metastatic potential of tumor cells.

In addition, VEGF expression‎ was also found in this study to be positively correlated with clinical stage of OSCC and LN metastasis. Similar results were reported in previous studies in HN cancer and OSCC [35], [36], [37]. This correlation may be attributed to the role of VEGF as a selective mitogen for vascular ECs that induce tumor angiogenesis by promoting the differentiation and proliferation of ECs. Then, VEGF guides the angiogenic sprouts to expand the primary plexus and vascularize the growing tumor that facilitate tumor expansion and invasion [38]. Additionally, VEGF was also reported to induce vasodilatation and increases vascular permeability. This makes the newly-formed capillaries more penetrable by tumor cells than normal vessels, consequently, increasing the risk of metastasis [39].

The direct correlation between HIF-1α and local recurrence is supported by the results of Katsuta et al. [40], Schrijvers et al. [41], in esophageal and glottic laryngeal carcinoma respectively. These results may reveal the role of HIF-1α in attaining an aggressive phenotype of the tumor cells that have the ability for invasion, metastasis and local recurrence.

Histologically, the localized HIF-1α expression‎ around keratinized and necrotic areas of the tumor or in regions distal to the nearest blood vessel appeared to corroborate with the observations of previous studies [9], [42]. This supports the hypothesis that HIF-1α expression‎ is induced mainly by hypoxia and the presence of blood vessels in the tumor stroma does not necessarily ensure that a region of tumor is receiving appropriate perfusion and oxygenation. This could be due to the tumor blood flow that may be transiently compromised in regions where microvessels are immature, deformed or occluded. Moreover, areas where cells beyond the diffusion distance of oxygen from microvessel may be subjected to chronic hypoxia that strongly induce cells to produce HIF-1α as an adaptive mechanism to this hypoxia [42], [43].

These previous findings are also supported by the observations of a more recent study by Zhu et al. [44], who observed that HIF-1α protein and mRNA levels were hardly detected under normoxic condition, but were remarkably increased at hypoxic condition analyzed by using Western blot and quantitative transcriptase-polymerase chain reaction (RT-PCR).

Furthermore, the diffuse HIF-1α expression‎ observed in some cases of OSCC irrespective with the distance from the closest blood vessel, may be explained by the hypothesis that HIF-1α production is not only stimulated by hypoxia, but also with other signals as growth factors as well as the activation of other oncogenic signal transduction pathways that may play a role in upregulation of HIF-1α expression‎ in addition to hypoxia [45].

Additionally, the nuclear expression‎ of HIF-1α was more predominant and intense in advanced stages and increased degree of malignancy of OSCC in this study. These observations were in accordance with previous studies made by Ogane et al. [46] and Zhu et al. [44], who eval‎uated HIF-1α expression‎ in SCC focusing on the nuclear pattern in comparison to the cytoplasmic one. They found that nuclear pattern predominance is considered to be strongly related to the activated status of HIF-1α as it becomes trans-located to the nucleus to dimerize with HIF-1β subunit to form HIF-1 molecule that may control many pathways involved in aggressive behavior of the tumors [47].

The significant correlation of HIF-1α expression‎ with the histopathological grade of OSCC in the current study is supported by the results of previous studies that revealed similar findings in OC [7], [48]. This may indicate that HIF-1α upregualtion occurs well in early carcinogenesis and an essential step before acquiring invasive phenotype. Emerging experimental evidence indicates that hypoxic HIF-1α expression‎ plays important role in the regulation of phenotype and function of cancer stem cells (CSCs) by inducing the activation of signature genes of CSCs that enhance self-renewal capacity and maintenance of undifferentiated state of these cells [49].

In this study, only two cases of OSCC revealed no positivity of HIF-1α immunoreactivity. At present, there is no certain explanation for this finding. However, the prolonged fixation of tissue specimens is known to substantially compromise antigen detection, so that failure to stain these cases might be artificial. Otherwise, this negativity could be also explained that HIF-1α protein expression‎ in these cases was at levels below the limits of detection by IHC methodology [50].

In the present study, VEGF expression‎ was found to be significantly upregulated with increased malignant criteria of OSCC, as it was markedly higher in poorly-differentiated OSCC compared to well-differentiated cases. However, VEGF reactivity was not positive or very weak in the normal epithelium. Similar findings were reported by Macluskey et al. [11], and Shang et al. [51], who observed a significant increase in VEGF expression‎ during the transition from normal oral mucosa through different degrees of dysplasia. These findings suggest the hypothesis that VEGF is strongly involved in angiogenesis and it is essential in the development of OSCC.

In the current study, HIF-1α and VEGF immunoreactivities were also observed in the vascular ECs of the tumor stroma and HIF-1α showed significant correlation with tumor angiogenesis measured by VEGF staining. These findings were in a line with a previous study that found that the endothelial HIF-1α mediates pathway for VEGF expression‎ and deletion of HIF-1α in ECs disrupts an autocrine loop necessary for hypoxic induction of VEGFR-1 and VEGFR-2 signaling by preventing the binding of HIF-1 to an enhancer element located in VEGFR-1 and the post-trancriptional regulation of VEGFR-2 by HIF-1. Thus, this will result in decreased endothelial proliferation and tube formation in vitro leading to severely impaired tumor angiogenesis and reduced tumor growth in vivo [52].

Accordingly, all the previous observations of this study shed a light on the importance of TME changes, specifically hypoxia and angiogenesis, in controlling the behavior of OSCC. This confirm‎s the hypothesis that the adaptive pathways to hypoxia as HIF-1α and VEGF expressions and their upregulation may play a crucial role in tumor angiogenesis, progression and consequently, controlling the behavior of OSCC.

Acknowledgment

This work was supported in part by Oral Pathology Department, Faculty of Dentistry, Tanta University and Tanta Cancer Center.

References

[1]

S. Choi, J.N. MyersMolecular pathogenesis of oral squamous cell carcinoma: implications for therapy

J Dent Res, 87 (2008), pp. 14-32

CrossRefView Record in ScopusGoogle Scholar

[2]

I.J. Hoogsteen, H.A. Marres, J. Bussink, A.J. van der Kogel, J.H. KaandersTumor microenvironment in head and neck squamous cell carcinomas: predictive value and clinical relevance of hypoxic markers: a review

Head Neck, 29 (2007), pp. 591-604

CrossRefView Record in ScopusGoogle Scholar

[3]

S.M. van de Pol, P.A. Doornaert, R. de Bree, C.R. Leemans, B.J. Slotman, J.A.LangendijkThe significance of anemia in squamous cell head and neck cancer treated with surgery and postoperative radiotherapy

Oral Oncol, 42 (2006), pp. 131-138

ArticleDownload PDFView Record in ScopusGoogle Scholar

[4]

M.A.C. Dery, M.D. Michaud, D.E. RichardHypoxia-inducible factor 1: regulation by hypoxic and non-hypoxic activators

Int J Biochem Cell Biol, 37 (2005), pp. 535-540

ArticleDownload PDFView Record in ScopusGoogle Scholar

[5]

E. Berra, E. Benizri, A. Ginouves, V. Volmat, D. Roux, J. PouysségurHIF prolyl-hydroxylase 2 is the key oxygen sensor setting low steady-state levels of HIF-1alpha in normoxia

EMBO J, 22 (2003), pp. 4082-4090

View Record in ScopusGoogle Scholar

[6]

G.L. SemenzaTargeting HIF-1 for cancer therapy

Nat Rev Cancer, 3 (10) (2003), pp. 721-732

CrossRefView Record in ScopusGoogle Scholar

[7]

M. Volm, R. KoomageHypoxia-inducible factor-1 (HIF-1) and its relationship to apoptosis and proliferation in lung cancer

Anticancer Res, 20 (2000), pp. 1523-1527

Google Scholar

[8]

N.J. Beasley, R. Leek, M. Alam, H. Turley, G.J. Cox, K. Gatter, et al.Hypoxia-inducible factors HIF-1 alpha and HIF-2 alpha in head and neck cancer: relationship to tumor biology and treatment outcome in surgically resected patients

Cancer Res, 62 (2002), pp. 2493-2497

View Record in ScopusGoogle Scholar

[9]

L. Kopfstein, G. ChristoforiMetastasis: cell-autonomous mechanisms versus contributions by the tumor microenvironment

Cell Mol Life Sci, 63 (2006), pp. 449-468

CrossRefView Record in ScopusGoogle Scholar

[10]

M. Sohda, H. Ishikawa, N. Masuda, H. Kato, T. Miyazaki, M. Nakajima, et al.Pre-treatment eval‎uation of combined HIF-1alpha, p53, and p21 expression‎ is a useful and sensitive indicator of response to radiation and chemotherapy in oesophageal cancer

Int J Cancer, 110 (6) (2004), pp. 838-844

View Record in ScopusGoogle Scholar

[11]

X. Liang, M. Zheng, J. Jiang, G. Zhu, J. Yang, Y. TangHypoxia-inducible factor -1 alpha in association with TWIST2 and SNIP1 is a critical prognostic factor in patients with tongue squamous cell carcinoma

Oral Oncol J, 47 (2011), pp. 92-97

ArticleDownload PDFView Record in ScopusGoogle Scholar

[12]

H.E. Ryan, J. Lo, R.S. JohnsonHIF-1 alpha is required for solid tumor formation and embryonic vascularization

EMBO J, 17 (1998), pp. 3005-3015

View Record in ScopusGoogle Scholar

[13]

T. Tonini, F. Rossi, P.P. ClaudioMolecular basis of angiogenesis and cancer

Oncogene, 22 (2003), pp. 6549-6556

CrossRefView Record in ScopusGoogle Scholar

[14]

A. BikfalviAngiogenesis

Encycl Endocrinol Dis, 1 (2004), pp. 227-233

ArticleDownload PDFGoogle Scholar

[15]

M. Tajima, Y. Kurashima, K. Sugiyama, T. Ogura, H. SakagamiThe redox state of glutathione regulates the hypoxic induction of HIF-1

Euro J Pharmacol, 606 (2009), pp. 45-49

ArticleDownload PDFView Record in ScopusGoogle Scholar

[16]

M. HerlynEndothelial cells as targets for tumor therapy

J Immunother, 22 (1999), pp. 185-186

CrossRefView Record in ScopusGoogle Scholar

[17]

M. Macluskey, L.M. Chandrachud, S. Pazouki, M. Green, D.M. Chisholm, G.R.Ogden, et al.Apoptosis, proliferation, and angiogenesis in oral tissues possible relevance to tumor progression

J Pathol, 191 (4) (2000), pp. 368-375

View Record in ScopusGoogle Scholar

[18]

J. Carlile, K. Harada, R. Baillie, M. Macluskey, D.M. Chisholm, G.R. Ogden, et al.Possible relevance to angiogenesis, tumor progression and field cancerization

J Oral Pathol Med, 30 (8) (2001), pp. 449-457

View Record in ScopusGoogle Scholar

[19]

C.N. Penfold, M. Partridge, R. Rojas, J.D. LangdonThe role of angiogenesis in the spread of oral squamous cell carcinoma

Br J Oral Maxillofac Surg, 34 (1) (1996), pp. 37-41

ArticleDownload PDFView Record in ScopusGoogle Scholar

[20]

S. Pazouki, D.M. Chisholm, M.M. Adi, G. Carmichael, M. Farquharson, G.R.Ogden, et al.The association between tumor progression and vascularity in the oral mucosa

J Pathol, 183 (1) (1997), pp. 39-43

View Record in ScopusGoogle Scholar

[21]

S. WarnalulasuriyaGlobal epidemiology of oral and oropharyngeal cancer

Oral Oncol, 45 (2009), pp. 309-316

Google Scholar

[22]

D.M. Parkin, L. Boyd, L.C. WalkerThe fraction of cancer attributable to lifestyle and environmental factors in the UK in 2010

BJC, 105 (2) (2011), pp. 77-81

Google Scholar

[23]

E. Halboub, M. Al-Mohaya, M. Abdulhuq, A. Al-Mandili, Y. Al-AnaziOral squamous cell carcinoma among Yemenis: onset in young age and presentation at advanced stage

J Clin Exp Dent, 4 (4) (2012), pp. 221-225

Google Scholar

[24]

R. Shenoi, V. Devrukhar, R.K. Chaudhauri, B.K. Sharma, S.B. Sapre, A. ChikhaleDemographic and clinical profile of oral squamous cell carcinoma patients: a retrospective study

Ind J Cancer, 49 (2013), pp. 21-26

Google Scholar

[25]

V. Mehta, G.P. Yu, S.P. SchantzPopulation-based analysis of oral and oropharyngeal carcinoma: changing trends of histopathologic differentiation, survival and patient demographics

Laryngoscope, 120 (2010), pp. 2203-2212

CrossRefView Record in ScopusGoogle Scholar

[26]

D.D. Dantas, C.C. Ramos, A.L. Costa, L.B. Souza, L.P. PintoClinicopathological parameters in squamous cell carcinoma of the tongue

Braz Dent J, 14 (1) (2003), pp. 22-25

CrossRefView Record in ScopusGoogle Scholar

[27]

A.Y. Maslov, J. VijgGenome instability, cancer and aging

Biochim Biophys Acta J, 209 (1790), pp. 963-969

View Record in ScopusGoogle Scholar

[28]

P.A. Brennan, T. Umar, G.I. Smith, P. McCauley, W.J. Peters, J.D. LangdonExpression‎ of type 2 nitric oxide synthase and p53 in Warthin's tumor of the parotid

J Oral Pathol Med, 31 (2002), pp. 458-462

View Record in ScopusGoogle Scholar

[29]

J. Rak, R.S. KerbelRas regulation of vascular endothelial growth factor and angiogenesis

Meth Enzymol, 333 (2001), pp. 267-283

ArticleDownload PDFView Record in ScopusGoogle Scholar

[30]

A.W. Eckert, A. Launter, K. Schutze, K. Bolte, M. Bache, M. Kappler, et al.Co-expression‎ of HIF-1 alpha and CAIX is associated with poor prognosis in oral squamous cell carcinoma patients

J Oral Pathol Med, 39 (2010), pp. 313-317

View Record in ScopusGoogle Scholar

[31]

W.M. Tilakaratne, E.H. Nissanka-JayasuriyaValue of HIF-1α as an independent prognostic indicator in oral squamous cell carcinoma

Expert Rev Mol Diagn, 11 (2) (2011), pp. 145-147

CrossRefView Record in ScopusGoogle Scholar

[32]

Y. Miayazaki, A. Itara, K. Kato, T. Oyama, Y. Yamada, H. Mori, et al.The effect of hypoxic microenvironment on matrix metalloproteinase expression‎ in xenografts of human oral squamous cell carcinoma

Int J Oncol, 32 (2008), pp. 135-151

Google Scholar

[33]

M.H. Ryu, H.M. Park, J. Chung, C.H. Lee, H.R. ParkHypoxia inducible factor-1 alpha mediates oral squamous cell carcinoma invasion via upregulation of α5-integrin and fibronectin

Biochem Biophys Res Communic, 393 (2010), pp. 11-15

ArticleDownload PDFView Record in ScopusGoogle Scholar

[34]

T. Imai, A. Horiuchi, C. Wang, K. Oka, S. Ohira, T. Nikaido, et al.Hypoxia attenuates the expression‎ of E-cadherin via up-regulation of SNAIL in ovarian carcinoma cells

Am J Pathol, 163 (2003), pp. 1437-1447

ArticleDownload PDFView Record in ScopusGoogle Scholar

[35]

O. Gallo, G. Masini, L. Morbidelli, A. Franchi, I. Fini-Storchi, W.A. Vergari, et al.Role of nitric oxide in angiogenesis and tumor progression in head and neck cancer

J Natl Cancer Inst, 90 (1998), pp. 587-596