- Email: [email protected]
Expression of type 2 nitric oxide synthase and vascular endothelial growth factor in oral dysplasia
J Oral Maxillofac Surg 60:1455-1460, 2002
Expression of Type 2 Nitric Oxide Synthase and Vascular Endothelial Growth Factor in Oral Dysplasia Peter A. Brennan, MD, FDSRCS, FRCS, FRCSI,* Tijjani Umar, MB BS, MSc, MRCPath, FMCPath,† Alan W. Wilson, FDSRCS, FRCS,‡ and Timothy K. Mellor, FDSRCS, FRCS§ Purpose:
The small molecule nitric oxide (NO), produced by a family of enzymes called NO synthase (NOS), has a diverse array of functions in both physiologic and pathologic states. Prolonged production of NO by the isoform NOS2 has been implicated in human cancer progression. NO has an important role in angiogenesis, being both an upstream signal and a downstream effector molecule to vascular endothelial growth factor (VEGF). The latter group of proteins are crucial for vascular endothelial cell proliferation and permeability. The expression of VEGF increases with cancer progression. Because angiogenesis is a prerequisite for the development of invasive cancer, this immunohistochemical study investigated the expression of NOS2 and VEGF in oral epithelial dysplasia. Materials and Methods: An immunohistochemical study was performed using monoclonal antibodies to NOS2 and VEGF on archival formalin-fixed, paraffin-embedded tissue of 33 cases of oral dysplasia. Results: A significant correlation was found between NOS2 and VEGF expression in oral dysplasia (P ⬍ .001). Expression of both NOS2 and VEGF also correlated with the severity of dysplasia (P ⬍ .001, P ⬍ .002). Conclusions: These findings may provide further understanding to the complex transformation of oral epithelial dysplasia to invasive carcinoma and the role of angiogenesis in this process. © 2002 American Association of Oral and Maxillofacial Surgeons J Oral Maxillofac Surg 60:1455-1460, 2002 The biologic actions of nitric oxide (NO) are diverse and include effects ranging from the regulation of vascular tone and blood pressure to pathologic roles in septic shock and cancer. Although much has been written about the effects of NO in cancer, only a handful of reports relate to the head and neck region. *Specialist Registrar, Department of Oral and Maxillofacial Surgery, Poole Hospital, Poole, Dorset, England. †Specialist Registrar in Histopathology, Bournemouth Hospital, Bournemouth, Dorset, England. ‡Specialist Registrar, Department of Oral and Maxillofacial Surgery, Queen Alexandra Hospital, Portsmouth, Hants, England. §Consultant Oral and Maxillofacial Surgeon, Queen Alexandra Hospital, Portsmouth, Hants, England. The British Association of Oral and Maxillofacial Surgeons provided financial assistance via endowment grants. Address correspondence and reprint requests to Dr Brennan: 11 Oxlease Close, Romsey, Hants, SO51 7HA, United Kingdom; e-mail: [email protected] © 2002 American Association of Oral and Maxillofacial Surgeons
0278-2391/02/6012-0013$35.00/0 doi:10.1053/joms.2002.36122
In vivo, NO is synthesized from the terminal guanidino group of the amino acid L-arginine by the enzyme nitric oxide synthase (NOS). This is a complex family of enzymes that share significant homology to cytochrome P450. There are at least 3 NOS isoenzymes, each being the product of a distinct gene. The expression of type 2 NOS (NOS2) is induced by various stimuli, including various cytokines (interleukin [IL]-1b, IL-2, IL-6), interferon-␥ (IFN-␥), tumor necrosis factor-␣ (TNF-␣), and hypoxia.1 NOS2 is therefore also termed inducible NOS (iNOS). NOS2 has been found in a variety of cells, including macrophages and natural killer cells of the immune system, and some tumor cells, including oral squamous carcinoma cells, also express it. Because its activity is independent of local calcium levels, NOS2 continues to produce NO for many hours or even days after its induction. This is in contrast to the constitutively expressed isoenzymes, NOS1 and NOS3, also called neural and endothelial NOS (nNOS and eNOS, respectively), which produce NO in a pulsed fashion. Because the function of these 2 isoenzymes is dependent on elevated local calcium levels,
1455
1456 probably to beyond 500 nmol/L (from a resting level of 70 to 100 nmol/L), NO production is transient. Working on the assumption that sustained production on NO is more potentially significant, the majority of studies of NO in cancers have been directed at NOS2. The understanding of NO in tumor biology is still in its infancy, because NO appears to have both tumorpromoting as well as inhibitory actions. The specific actions of NO are now known to be concentration dependent, and the levels found in human tumors are thought to be responsible for enhanced angiogenesis and tumour invasiveness.2,3 NO appears to play a pivotal role in angiogenesis because L-NAME (NG-nitro-L-arginine-methyl ester), a stereospecific competitive inhibitor of NOS, blocks the angiogenic process.4 The existence of a link between NO and vascular endothelial growth factor (VEGF) is increasingly recognized.5,6 VEGF and its receptors, the Flt family, are mainly responsible for the mechanism of tumor angiogenesis.7,8 VEGF, a heparin-binding growth factor, acts directly on vascular endothelial cells promoting cell proliferation and permeability, inducing angiogenesis in both physiologic and pathologic situations. Expression of VEGF and the onset of mutations in cancer-related genes such as p53 increase with cancer progression.9,10 Like NOS2, expression of VEGF is also induced by hypoxia, although this is mediated by the molecule hypoxiainduced factor-1 (HIF).11 The progression of dysplasia to invasive carcinoma is a complicated and poorly understood process. Animal model studies have underlined the importance of angiogenesis during the progressive transformation of keratinocytes to invasive cancer. Dysplastic lesions displayed abundant capillaries leading to intense vascularization in invasive squamous cell carcinoma.12 Because tumors cannot exceed 1 to 2 mm3 of volume without developing new blood vessels, they must produce angiogenic factors at an early point of development. Surprisingly, there are few reports in the literature of VEGF expression in both oral cancer13,14 and dysplasia,15-17 a known premalignant condition with 3 commonly recognized degrees of severity. One study found that VEGF mRNA was significantly increased in severe dysplasia and invasive oral cancer and absent in normal mucosa and mild dysplasia.15 However, an immunohistochemical study showed that VEGF was downregulated in oral dysplasia compared with normal oral mucosa.16 A more recent immunohistochemical study found varying VEGF expression depending on the type of VEGF antibody used.17 Clearly, further research is required to establish the role of VEGF and its isoforms (the protein can exist in
NOS2 AND VEGF IN DYSPLASIA
5 isoforms with molecular weights between 121 and 206 kDa), in both oral dysplasia and invasive cancer. We recently found that the immunohistochemical expression of NOS2 correlated with severity of oral dysplasia, with the most intense and widespread NOS2 staining being found in cases of severe dysplasia.18,19 Because the relationship between NO, NOS2, and VEGF is increasingly recognized, the purpose of the current immunohistochemical study was to determine whether an association could be found in oral epithelial dysplasia.
Materials and Methods Histologic slides were prepared from formalin-fixed paraffin-embedded archival specimens, initially stained with hematoxylin and eosin and examined by a consultant histopathologist both to confirm the diagnosis and to grade the severity of dysplasia into mild, moderate, or severe. Cases were excluded if the specimen included any other associated pathology (eg, chronic candidiasis or frank malignancy). In addition, 6 biopsy samples of macroscopically and histologically normal buccal mucosa were studied, obtained after approval by the local ethics committee and informed consent. Multiple sections of 4-m thickness were obtained from each paraffin block and stained as follows: Sections were placed onto positively charged slides, heated for 40 minutes at 50°C, dewaxed in xylene for 10 minutes, rehydrated in alcohol solutions (70%, 50%), and placed in 5% hydrogen peroxide in absolute alcohol for 10 minutes to block endogenous peroxidase activity. The sections were allowed to drain for 5 minutes by placing them upright in a slide carrier. Rehydration was completed by placement in absolute alcohol and finally water. The slides were treated with a boiling solution of freshly prepared 0.05 mol/L citrate buffer, pH 6.0, for 2 minutes in a pressure cooker. The sections were treated with normal goat serum with 1 drop of avidin block/mL for 20 minutes and incubated at room temperature for 1 hour with either anti-NOS2 monoclonal antibody (Transduction Laboratories, Lexington, KY) or VEGF antibody (Santa Cruz Biotechnology, Santa Cruz, CA) at dilutions of 1:100 and 1:200, respectively. The VEGF monoclonal antibody used in this study is marketed by the manufacturer for use on foramalin-fixed, paraffin-embedded tissue, and we have recently confirmed the specificity of the NOS2 antibody with Western blotting.19 The slides were rinsed in Trisbuffered saline (TBS) before treatment with biotinylated anti-mouse immunoglobulin (Dako, Glostrup, Denmark) at a dilution of 1:200 for 30 minutes. After further rinsing, the slides were treated with avidin-biotin-peroxidase complex (Dako) for an additional 30 minutes before being washed again with
1457
BRENNAN ET AL
TBS. The immunostaining was visualized by development in diaminobenzidine (DAB), and the slides were then rinsed in distilled water and counterstained using Mayer’s hematoxylin. Negative controls were treated in the same manner but with the omission of the primary antibodies. Additionally, an irrelevant antibody, smooth muscle actin (SMA) (Dako), with the same isotype as NOS2 and VEGF (IgG2a) was used at a dilution of 1:100 to assess nonspecific primary antibody binding. As positive controls, we used parotid tissue, the ducts of which stain intensely for NOS2,20 together with normal kidney tissue (obtained at post mortem) for confirmation of VEGF expression (both positive controls being formalin-fixed paraffin-embedded tissue). Staining was scored by a histopathologist (T.U.) by evaluating both the percentage of staining within representative regions of each specimen and the intensity of staining. Slides were randomly reviewed so as minimize possible bias. At least 3 representative regions were scored for each section. In sections displaying staining heterogeneity, the mean average score was used for analysis. The region of staining (viewed at a magnification of ⫻40) was scored as follows: 0, no staining of cells in any microscopic field; 1⫹, less than 25% of tissue stained positive; 2⫹, between 25% and 50% staining positive; 3⫹, between 50% and 75% staining positive; and 4⫹, more than 75% staining positive. For the purposes of statistical analysis, expression was graded as positive if more than 25% of epithelial cells showed immunoreactivity. The intensity of staining (viewed at a magnification of ⫻200) was recorded on the following scale: 0, no staining seen; 1⫹, mild staining; 2⫹, moderate staining; and 3⫹, intense staining. Staining intensity was objectively recorded in relation to the positive control sections, which stained intensely (score, 3) for NOS2 and VEGF in all cases. The scores of region and intensity of staining were combined to give a value between 0 and 7. The results were analyzed using Arcus Quickstat Biomedical Software (Longman, Cambridge, United Kingdom). Possible statistical significance was assessed using Fisher’s exact test and Spearman’s rank correlation. The level of significance for the tests was chosen to be P ⬍ .01.
Results Thirty-three cases were studied and graded as mild (n ⫽ 9), moderate (n ⫽ 10), and severe (n ⫽ 14) dysplasia. Specimens were taken from patients with an age range of 33 to 82 years (mean average age, 64 years). The specimens were taken from the floor of the mouth (n ⫽ 10), lateral border of the tongue (n ⫽ 18), and base of the tongue (n ⫽ 5). The immunostaining was of high quality and easy to interpret (Figs
1 to 3). Positive controls stained intensely in all cases. The use of the irrelevant antibody SMA confirmed that nonspecific binding had not occurred because staining was confined to blood vessel walls. NOS2 and VEGF staining were not shown in normal oral mucosa epithelium. In mild dysplasia, NOS2 staining was seen in 2 of 9 cases (22%), in the basal layers of epithelium only. NOS2 staining was seen in the majority of moderate and severe dysplasia cases (Fig 1), with a significant correlation being found between its expression and the severity of dysplasia (r ⫽ 0.55, P ⬍ .001) (Table 1). It was also found in the submucosal connective tissue in cells presumed to be macrophages, by virtue of their morphology. In mild dysplasia, immunolocalization of VEGF was found in the basal and suprabasal epithelial layer in 3 cases (33%) (Fig 2), with one case showing both VEGF and NOS2 expression. VEGF expression was also found to correlate with the severity of dysplasia (r ⫽ 0.50, P ⬍ .002) (Table 1). Expression of both VEGF and NOS2 was seen in 18 of 24 cases (75%) of moderate and severe dysplasia (Fig 3), and a significant correlation was found between them (P ⬍ .001). Like NOS2, VEGF was also expressed by inflammatory cells in the underlying connective tissue (Fig 3), and it was also seen with variable distribution in blood vessel endothelium.
Discussion The VEGF results found in this study are in agreement with an earlier study of VEGF mRNA expression.15 Although a recent immunohistochemical study did not find a correlation of VEGF expression in oral dysplasia or carcinoma,17 this is likely to be a result of differences in the monoclonal antibodies used in recognizing the various isoforms of VEGF.17 The NOS2 results are also in agreement with recent immunohistochemical studies of oral dysplasia.18,19 VEGF is the only angiogenic peptide known to act specifically on endothelial cells, and it is therefore considered to be the leading candidate of the factors that promote angiogenesis.21 The relationship between NO and VEGF is increasingly recognised. It would appear that NO may have actions both upstream and downstream of VEGF. Parenti et al22 showed that NO was an upstream signal for VEGF-related kinases, whereas Ziche et al6 found that breast carcinoma cells that overexpressed VEGF require the NO pathway to induce angiogenesis in vivo. Our study has shown a significant association between the immunohistochemical expression of NOS2 and VEGF. However, we also found that expression of NOS2 and VEGF occurred independently in 6 cases, suggesting that other factors may be involved in the expression of these proteins. Because this would ap-
1458
NOS2 AND VEGF IN DYSPLASIA
FIGURE 3. Intense expression of VEGF can be seen throughout the epithelium and in submucosal inflammatory cells in severe dysplasia (DAB/hematoxylin stain, original magnification ⫻40).
FIGURE 1. Widespread NOS2 expression by epithelial cells, with submucosal inflammatory cell staining in moderate and severe dysplasia (DAB/hematoxylin stain, original magnification ⫻100). (Reprinted with permission.18)
pear to be the first study to assess the expression of both of these proteins in oral dysplasia, it is not possible to comment on whether NO produced by NOS2 was responsible for VEGF expression or vice versa. Further research is required to establish the relationship between these proteins more fully. Because there are 5 VEGF isoforms, it would be interesting to determine whether one or more of these proteins are involved in this interaction with NOS2 in dysplasia. It
would also be interesting to compare the results obtained with dysplastic changes that are often found at the margins of invasive cancer. Although these regions are termed dysplastic, they are likely to represent field change close to the invading cancer margin and therefore may have different expressions of these proteins.17 No attempt was made in this study to evaluate vessel microdensity. It is particularly difficult to interpret the results of vessel microdensity (usually assessed with an endothelial marker such as CD31) in the oral cavity because of the large degree of heterogeneity found in vessel density in oral mucosa and because it is not possible to readily distinguish neovascularization from preexisting vessels. This may explain why, in a recent immunohistochemical study, Carlile et al17 found that VEGF did not correlate with vessel microdensity in dysplasia or invasive cancer. The roles of NO in tumour biology are complicated and multifaceted because NO appears to have both cytotoxic and tumor-promoting effects. This apparent controversy is probably explained by the NO concentrations that are found in different experimental and clinical models.2 It is likely that the levels of NO produced in human cancers are insufficient to cause apoptosis and cell death but instead facilitate the processes of angiogenesis and tumor dissemination.2,4
Table 1. THE MEAN AND RANGE OF STAINING SCORES FOR NOS2 AND VEGF
Mild dysplasia (n ⫽ 9) Moderate dysplasia (n ⫽ 10) Severe dysplasia (n ⫽ 14) FIGURE 2. Expression of VEGF confined to the basal and suprabasal layer of epithelium in mild dysplasia (DAB/hematoxylin stain, original magnification ⫻40).
NOS2
VEGF
0.6 (0-3) 3.8 (3-4) 4.6 (0-6)
1.2 (0-4) 3.2 (0-5) 3.64 (0-7)
NOTE. Significant correlations were found between NOS2 (P ⬍ .001) and VEGF (P ⬍ .002) expression and the degree of severity of dysplasia.
1459
BRENNAN ET AL
Recently, the effect on NOS2 expression of the chemical carcinogen 7,12-dimethylbenz[a]anthracene (DMBA) was studied in the hamster buccal pouch.24 Immunohistochemical expression of NOS2 expression was increased in DMBA-induced buccal mucosa dysplasia and carcinoma compared with untreated animals. In the latter group, no NOS2 staining was found. It was postulated that NOS2 was implicated in the transformation of normal mucosa through dysplasia to invasive malignancy. However, the explanation for NOS2 activation in dysplastic epithelium is difficult to explain. It may occur as a result of cytokine release from macrophages and other cytokine-releasing inflammatory cells. We found both NOS2 and VEGF staining in cells of the underlying connective tissue. Although these cells are likely to be macrophages (by virtue of their morphology), a double staining technique using the myelomonocytic marker CD68 would be required to confirm the exact cell of origin. Even though there are a considerable number of reports that address this subject (none in head and neck cancer or dysplasia), the exact pathway remains unclear. Hajri et al25 investigated the role of NO in pancreatic adenocarcinoma and suggested that the interaction of tumor cell and macrophage might be due to direct contact or contingent paracrine activating factors. Whatever the activation pathways, the expressions of both NOS2 and VEGF increase with the severity of oral dysplasia, and it is likely that they are implicated in the complex transformation to malignancy. Certainly, the development of an invasive tumor requires the process of angiogenesis to occur at an early stage. This has been shown to occur in such premalignant conditions as breast and cervical dysplasia.12 The interesting question therefore arises about clinically inhibiting NOS2 or VEGF in dysplasia and oral cancer as part of the overall management of these diseases. The use of experimental drugs such as N-(3(aminomethyl)benzyl)acetamidine (or 1400W), a specific antagonist of NOS2, results in reduced tumour growth, thereby confirming the role of NOS2 and NO in this process.26 Because the configuration of the arginine binding site for each NOS isoenzyme is slightly different, the specificity for this agent to NOS2 is 2,000 times that of either NOS1 or NOS3. Therefore the side effects of these specific drugs (eg, potential modulation of blood pressure via NOS3 inhibition) are minimal. Monoclonal antibodies against VEGF protein are currently undergoing early clinical trials in advanced ovarian cancer. It is possible that these or similar agents may also become part of the overall management of cancer and precancerous conditions such as dysplasia in the future.
A significant relationship was found between NOS2 and VEGF expression in oral dysplasia. Further research is required to more fully establish the roles of VEGF and NOS2 in the development and progression of oral cancer and to assess the possible therapeutic applications of inhibiting the expression of these proteins. Acknowledgments We thank Dr Anne Spedding, Consultant Histopathologist, and Sue Tant for technical assistance. We are also grateful to Bernard Higgins, Medical Statistician, University of Portsmouth, and the Photographic Department, Poole Hospital, for their help in this study.
References 1. Melillo G, Musso T, Sica A, et al: A hypoxia-responsive element mediates a novel pathway of activation of the inducible nitric oxide promoter. J Exp Med 182:1683, 1995 2. Thomsen LL, Miles DW: Role of nitric oxide in tumour progression: Lessons from human tumours. Cancer Metastasis Rev 17:107, 1998 3. Fukumara D, Jain RK: Role of nitric oxide in angiogenesis and microcirculation in tumors. Cancer Metastasis Rev 17:77, 1998 4. Gallo O, Masini E, Morbidelli L, et al: Role of nitric oxide in angiogenesis and tumour progression in head and neck cancer. J Natl Cancer Inst 90:587, 1998 5. Chiarugi V, Magnelli L, Gallo O: Cox-2, iNOS and p53 as play-markers of tumor angiogenesis. Int J Mol Med 2:715, 1998 6. Ziche M, Morbidelli L, Choudhuri R, et al: Nitric oxide synthase lies downstream from vascular endothelial growth factor-induced but not basic fibroblast growth factor-induced angiogenesis. J Clin Invest 99:2625, 1997 7. Kim KJ, Li B, Winer J, et al: Inhibition of vascular endothelial growth factor-induced angiogenesis suppresses tumour growth in vivo. Nature 362:841, 1993 8. Millauer B, Shawver LK, Plate KH, et al: Glioblastoma growth inhibited in vivo by a dominant-negative Flk-1 mutant. Nature 843:576, 1994 9. Keiser A, Weich HA, Brandner G, et al: p53 potentiates protein kinase C induction of vascular endothelial growth factor. Oncogene 9:963, 1994 10. Mazure NM, Chen EY, Yeh P, et al: Oncogenic transformation and hypoxia synergistically act to modulate vascular endothelial growth factor expression. Cancer Res 56:3436, 1996 11. Forsythe JA, Jiang BH, Iyer NV, et al: Activation of vascular endothelial growth factor gene transcription gene by hypoxiainducible factor 1. Mol Cell Biol 16:4604, 1996 12. Hanahan D, Folkman J: Patterns and emerging mechanisms of the angiogenic switch during tumorigenesis. Cell 86:353, 1996 13. Maeda T, Matsumura S, Hiranuma H, et al: Expression of vascular endothelial growth factor in human oral squamous cell carcinoma: Its association with tumour progression and p53 gene status. J Clin Pathol 51:771, 1998 14. Moriyama M, Kumagai S, Kawashiri S, et al: Immunohistochemical study of tumour angiogenesis in oral squamous cell carcinoma. Oral Oncol 33:369, 1997 15. Denhart BC, Guidi AJ, Tognazzi K, et al: Vascular permeability factor/vascular endothelial growth factor and its receptors in oral and laryngeal squamous cell carcinoma and dysplasia. Lab Invest 77:659, 1997 16. Feng L, Lee JJ, Hong WK, et al: Expression of vascular endothelial growth factor and microvessel density in head and neck tumorigenesis. Clin Cancer Res 6:2821, 2000 17. Carlile J, Harada K, Baillie R, et al: Vascular endothelial growth factor (VEGF) expression in oral tissues: Possible relevance to angiogenesis, tumour progression and field cancerisation. J Oral Pathol Med 30:449, 2001 18. Brennan PA, Palacios-Callender M, Sinclair D, et al: Does expression of inducible nitric oxide synthase correlate with se-
1460
19.
20.
21. 22.
verity of oral epithelial dysplasia? J Craniomaxillofac Surg 28: 44, 2000 Brennan PA, Palacios-Callender M, Spedding AV, et al: Does type II nitric oxide synthase expression correlate with cellular proliferation in oral squamous cell carcinoma and dysplasia? Head Neck 27:217, 2001 Brennan PA, Umar T, Zaki GA, et al: Are myoepithelial cells responsible for the widespread expression of inducible nitric oxide synthase in pleomorphic adenoma? An immunohistochemical study. J Oral Pathol Med 29:279, 2000 Dvorak HF, Brown LF, Detmar M, et al: Vascular permeability factor/vascular endothelial growth factor, microvascular hyperpermeability and angiogenesis. Am J Pathol 146:1029, 1995 Parenti A, Morbidelli L, Cui XL, et al: Nitric oxide is an upstream signal of vascular endothelial growth factor-induced
NOS2 AND VEGF IN DYSPLASIA
23. 24.
25.
26.
extracellular signal-regulated kinase1/2 activation in postcapillary endothelium. J Biol Chem 273:4220, 1998 Albina JE: On the expression of nitric oxide synthase by human macrophages. Why no NO? J Leukoc Biol 58:643, 1995 Chen YK, Lin LM: Immunohistochemical expression of inducible nitric oxide synthase in DMBA-induced hamster buccal pouch carcinogenesis. Oral Oncol 36:221, 2000 Hajri A, Metzger E, Vallat F, et al: Role of nitric oxide in pancreatic tumour growth: In vivo and in vitro studies. Br J Cancer 78:841, 1998 Thomsen L, Scott JMJ, Topley P, et al: Selective inhibition of inducible nitric oxide synthase inhibits tumour growth in vivo: Studies with 1400W, a novel inhibitor. Cancer Res 57:3300, 1997
Expression of Type 2 Nitric Oxide Synthase and Vascular Endothelial Growth Factor in Oral Dysplasia Peter A. Brennan, MD, FDSRCS, FRCS, FRCSI,* Tijjani Umar, MB BS, MSc, MRCPath, FMCPath,† Alan W. Wilson, FDSRCS, FRCS,‡ and Timothy K. Mellor, FDSRCS, FRCS§ Purpose:
The small molecule nitric oxide (NO), produced by a family of enzymes called NO synthase (NOS), has a diverse array of functions in both physiologic and pathologic states. Prolonged production of NO by the isoform NOS2 has been implicated in human cancer progression. NO has an important role in angiogenesis, being both an upstream signal and a downstream effector molecule to vascular endothelial growth factor (VEGF). The latter group of proteins are crucial for vascular endothelial cell proliferation and permeability. The expression of VEGF increases with cancer progression. Because angiogenesis is a prerequisite for the development of invasive cancer, this immunohistochemical study investigated the expression of NOS2 and VEGF in oral epithelial dysplasia. Materials and Methods: An immunohistochemical study was performed using monoclonal antibodies to NOS2 and VEGF on archival formalin-fixed, paraffin-embedded tissue of 33 cases of oral dysplasia. Results: A significant correlation was found between NOS2 and VEGF expression in oral dysplasia (P ⬍ .001). Expression of both NOS2 and VEGF also correlated with the severity of dysplasia (P ⬍ .001, P ⬍ .002). Conclusions: These findings may provide further understanding to the complex transformation of oral epithelial dysplasia to invasive carcinoma and the role of angiogenesis in this process. © 2002 American Association of Oral and Maxillofacial Surgeons J Oral Maxillofac Surg 60:1455-1460, 2002 The biologic actions of nitric oxide (NO) are diverse and include effects ranging from the regulation of vascular tone and blood pressure to pathologic roles in septic shock and cancer. Although much has been written about the effects of NO in cancer, only a handful of reports relate to the head and neck region. *Specialist Registrar, Department of Oral and Maxillofacial Surgery, Poole Hospital, Poole, Dorset, England. †Specialist Registrar in Histopathology, Bournemouth Hospital, Bournemouth, Dorset, England. ‡Specialist Registrar, Department of Oral and Maxillofacial Surgery, Queen Alexandra Hospital, Portsmouth, Hants, England. §Consultant Oral and Maxillofacial Surgeon, Queen Alexandra Hospital, Portsmouth, Hants, England. The British Association of Oral and Maxillofacial Surgeons provided financial assistance via endowment grants. Address correspondence and reprint requests to Dr Brennan: 11 Oxlease Close, Romsey, Hants, SO51 7HA, United Kingdom; e-mail: [email protected] © 2002 American Association of Oral and Maxillofacial Surgeons
0278-2391/02/6012-0013$35.00/0 doi:10.1053/joms.2002.36122
In vivo, NO is synthesized from the terminal guanidino group of the amino acid L-arginine by the enzyme nitric oxide synthase (NOS). This is a complex family of enzymes that share significant homology to cytochrome P450. There are at least 3 NOS isoenzymes, each being the product of a distinct gene. The expression of type 2 NOS (NOS2) is induced by various stimuli, including various cytokines (interleukin [IL]-1b, IL-2, IL-6), interferon-␥ (IFN-␥), tumor necrosis factor-␣ (TNF-␣), and hypoxia.1 NOS2 is therefore also termed inducible NOS (iNOS). NOS2 has been found in a variety of cells, including macrophages and natural killer cells of the immune system, and some tumor cells, including oral squamous carcinoma cells, also express it. Because its activity is independent of local calcium levels, NOS2 continues to produce NO for many hours or even days after its induction. This is in contrast to the constitutively expressed isoenzymes, NOS1 and NOS3, also called neural and endothelial NOS (nNOS and eNOS, respectively), which produce NO in a pulsed fashion. Because the function of these 2 isoenzymes is dependent on elevated local calcium levels,
1455
1456 probably to beyond 500 nmol/L (from a resting level of 70 to 100 nmol/L), NO production is transient. Working on the assumption that sustained production on NO is more potentially significant, the majority of studies of NO in cancers have been directed at NOS2. The understanding of NO in tumor biology is still in its infancy, because NO appears to have both tumorpromoting as well as inhibitory actions. The specific actions of NO are now known to be concentration dependent, and the levels found in human tumors are thought to be responsible for enhanced angiogenesis and tumour invasiveness.2,3 NO appears to play a pivotal role in angiogenesis because L-NAME (NG-nitro-L-arginine-methyl ester), a stereospecific competitive inhibitor of NOS, blocks the angiogenic process.4 The existence of a link between NO and vascular endothelial growth factor (VEGF) is increasingly recognized.5,6 VEGF and its receptors, the Flt family, are mainly responsible for the mechanism of tumor angiogenesis.7,8 VEGF, a heparin-binding growth factor, acts directly on vascular endothelial cells promoting cell proliferation and permeability, inducing angiogenesis in both physiologic and pathologic situations. Expression of VEGF and the onset of mutations in cancer-related genes such as p53 increase with cancer progression.9,10 Like NOS2, expression of VEGF is also induced by hypoxia, although this is mediated by the molecule hypoxiainduced factor-1 (HIF).11 The progression of dysplasia to invasive carcinoma is a complicated and poorly understood process. Animal model studies have underlined the importance of angiogenesis during the progressive transformation of keratinocytes to invasive cancer. Dysplastic lesions displayed abundant capillaries leading to intense vascularization in invasive squamous cell carcinoma.12 Because tumors cannot exceed 1 to 2 mm3 of volume without developing new blood vessels, they must produce angiogenic factors at an early point of development. Surprisingly, there are few reports in the literature of VEGF expression in both oral cancer13,14 and dysplasia,15-17 a known premalignant condition with 3 commonly recognized degrees of severity. One study found that VEGF mRNA was significantly increased in severe dysplasia and invasive oral cancer and absent in normal mucosa and mild dysplasia.15 However, an immunohistochemical study showed that VEGF was downregulated in oral dysplasia compared with normal oral mucosa.16 A more recent immunohistochemical study found varying VEGF expression depending on the type of VEGF antibody used.17 Clearly, further research is required to establish the role of VEGF and its isoforms (the protein can exist in
NOS2 AND VEGF IN DYSPLASIA
5 isoforms with molecular weights between 121 and 206 kDa), in both oral dysplasia and invasive cancer. We recently found that the immunohistochemical expression of NOS2 correlated with severity of oral dysplasia, with the most intense and widespread NOS2 staining being found in cases of severe dysplasia.18,19 Because the relationship between NO, NOS2, and VEGF is increasingly recognized, the purpose of the current immunohistochemical study was to determine whether an association could be found in oral epithelial dysplasia.
Materials and Methods Histologic slides were prepared from formalin-fixed paraffin-embedded archival specimens, initially stained with hematoxylin and eosin and examined by a consultant histopathologist both to confirm the diagnosis and to grade the severity of dysplasia into mild, moderate, or severe. Cases were excluded if the specimen included any other associated pathology (eg, chronic candidiasis or frank malignancy). In addition, 6 biopsy samples of macroscopically and histologically normal buccal mucosa were studied, obtained after approval by the local ethics committee and informed consent. Multiple sections of 4-m thickness were obtained from each paraffin block and stained as follows: Sections were placed onto positively charged slides, heated for 40 minutes at 50°C, dewaxed in xylene for 10 minutes, rehydrated in alcohol solutions (70%, 50%), and placed in 5% hydrogen peroxide in absolute alcohol for 10 minutes to block endogenous peroxidase activity. The sections were allowed to drain for 5 minutes by placing them upright in a slide carrier. Rehydration was completed by placement in absolute alcohol and finally water. The slides were treated with a boiling solution of freshly prepared 0.05 mol/L citrate buffer, pH 6.0, for 2 minutes in a pressure cooker. The sections were treated with normal goat serum with 1 drop of avidin block/mL for 20 minutes and incubated at room temperature for 1 hour with either anti-NOS2 monoclonal antibody (Transduction Laboratories, Lexington, KY) or VEGF antibody (Santa Cruz Biotechnology, Santa Cruz, CA) at dilutions of 1:100 and 1:200, respectively. The VEGF monoclonal antibody used in this study is marketed by the manufacturer for use on foramalin-fixed, paraffin-embedded tissue, and we have recently confirmed the specificity of the NOS2 antibody with Western blotting.19 The slides were rinsed in Trisbuffered saline (TBS) before treatment with biotinylated anti-mouse immunoglobulin (Dako, Glostrup, Denmark) at a dilution of 1:200 for 30 minutes. After further rinsing, the slides were treated with avidin-biotin-peroxidase complex (Dako) for an additional 30 minutes before being washed again with
1457
BRENNAN ET AL
TBS. The immunostaining was visualized by development in diaminobenzidine (DAB), and the slides were then rinsed in distilled water and counterstained using Mayer’s hematoxylin. Negative controls were treated in the same manner but with the omission of the primary antibodies. Additionally, an irrelevant antibody, smooth muscle actin (SMA) (Dako), with the same isotype as NOS2 and VEGF (IgG2a) was used at a dilution of 1:100 to assess nonspecific primary antibody binding. As positive controls, we used parotid tissue, the ducts of which stain intensely for NOS2,20 together with normal kidney tissue (obtained at post mortem) for confirmation of VEGF expression (both positive controls being formalin-fixed paraffin-embedded tissue). Staining was scored by a histopathologist (T.U.) by evaluating both the percentage of staining within representative regions of each specimen and the intensity of staining. Slides were randomly reviewed so as minimize possible bias. At least 3 representative regions were scored for each section. In sections displaying staining heterogeneity, the mean average score was used for analysis. The region of staining (viewed at a magnification of ⫻40) was scored as follows: 0, no staining of cells in any microscopic field; 1⫹, less than 25% of tissue stained positive; 2⫹, between 25% and 50% staining positive; 3⫹, between 50% and 75% staining positive; and 4⫹, more than 75% staining positive. For the purposes of statistical analysis, expression was graded as positive if more than 25% of epithelial cells showed immunoreactivity. The intensity of staining (viewed at a magnification of ⫻200) was recorded on the following scale: 0, no staining seen; 1⫹, mild staining; 2⫹, moderate staining; and 3⫹, intense staining. Staining intensity was objectively recorded in relation to the positive control sections, which stained intensely (score, 3) for NOS2 and VEGF in all cases. The scores of region and intensity of staining were combined to give a value between 0 and 7. The results were analyzed using Arcus Quickstat Biomedical Software (Longman, Cambridge, United Kingdom). Possible statistical significance was assessed using Fisher’s exact test and Spearman’s rank correlation. The level of significance for the tests was chosen to be P ⬍ .01.
Results Thirty-three cases were studied and graded as mild (n ⫽ 9), moderate (n ⫽ 10), and severe (n ⫽ 14) dysplasia. Specimens were taken from patients with an age range of 33 to 82 years (mean average age, 64 years). The specimens were taken from the floor of the mouth (n ⫽ 10), lateral border of the tongue (n ⫽ 18), and base of the tongue (n ⫽ 5). The immunostaining was of high quality and easy to interpret (Figs
1 to 3). Positive controls stained intensely in all cases. The use of the irrelevant antibody SMA confirmed that nonspecific binding had not occurred because staining was confined to blood vessel walls. NOS2 and VEGF staining were not shown in normal oral mucosa epithelium. In mild dysplasia, NOS2 staining was seen in 2 of 9 cases (22%), in the basal layers of epithelium only. NOS2 staining was seen in the majority of moderate and severe dysplasia cases (Fig 1), with a significant correlation being found between its expression and the severity of dysplasia (r ⫽ 0.55, P ⬍ .001) (Table 1). It was also found in the submucosal connective tissue in cells presumed to be macrophages, by virtue of their morphology. In mild dysplasia, immunolocalization of VEGF was found in the basal and suprabasal epithelial layer in 3 cases (33%) (Fig 2), with one case showing both VEGF and NOS2 expression. VEGF expression was also found to correlate with the severity of dysplasia (r ⫽ 0.50, P ⬍ .002) (Table 1). Expression of both VEGF and NOS2 was seen in 18 of 24 cases (75%) of moderate and severe dysplasia (Fig 3), and a significant correlation was found between them (P ⬍ .001). Like NOS2, VEGF was also expressed by inflammatory cells in the underlying connective tissue (Fig 3), and it was also seen with variable distribution in blood vessel endothelium.
Discussion The VEGF results found in this study are in agreement with an earlier study of VEGF mRNA expression.15 Although a recent immunohistochemical study did not find a correlation of VEGF expression in oral dysplasia or carcinoma,17 this is likely to be a result of differences in the monoclonal antibodies used in recognizing the various isoforms of VEGF.17 The NOS2 results are also in agreement with recent immunohistochemical studies of oral dysplasia.18,19 VEGF is the only angiogenic peptide known to act specifically on endothelial cells, and it is therefore considered to be the leading candidate of the factors that promote angiogenesis.21 The relationship between NO and VEGF is increasingly recognised. It would appear that NO may have actions both upstream and downstream of VEGF. Parenti et al22 showed that NO was an upstream signal for VEGF-related kinases, whereas Ziche et al6 found that breast carcinoma cells that overexpressed VEGF require the NO pathway to induce angiogenesis in vivo. Our study has shown a significant association between the immunohistochemical expression of NOS2 and VEGF. However, we also found that expression of NOS2 and VEGF occurred independently in 6 cases, suggesting that other factors may be involved in the expression of these proteins. Because this would ap-
1458
NOS2 AND VEGF IN DYSPLASIA
FIGURE 3. Intense expression of VEGF can be seen throughout the epithelium and in submucosal inflammatory cells in severe dysplasia (DAB/hematoxylin stain, original magnification ⫻40).
FIGURE 1. Widespread NOS2 expression by epithelial cells, with submucosal inflammatory cell staining in moderate and severe dysplasia (DAB/hematoxylin stain, original magnification ⫻100). (Reprinted with permission.18)
pear to be the first study to assess the expression of both of these proteins in oral dysplasia, it is not possible to comment on whether NO produced by NOS2 was responsible for VEGF expression or vice versa. Further research is required to establish the relationship between these proteins more fully. Because there are 5 VEGF isoforms, it would be interesting to determine whether one or more of these proteins are involved in this interaction with NOS2 in dysplasia. It
would also be interesting to compare the results obtained with dysplastic changes that are often found at the margins of invasive cancer. Although these regions are termed dysplastic, they are likely to represent field change close to the invading cancer margin and therefore may have different expressions of these proteins.17 No attempt was made in this study to evaluate vessel microdensity. It is particularly difficult to interpret the results of vessel microdensity (usually assessed with an endothelial marker such as CD31) in the oral cavity because of the large degree of heterogeneity found in vessel density in oral mucosa and because it is not possible to readily distinguish neovascularization from preexisting vessels. This may explain why, in a recent immunohistochemical study, Carlile et al17 found that VEGF did not correlate with vessel microdensity in dysplasia or invasive cancer. The roles of NO in tumour biology are complicated and multifaceted because NO appears to have both cytotoxic and tumor-promoting effects. This apparent controversy is probably explained by the NO concentrations that are found in different experimental and clinical models.2 It is likely that the levels of NO produced in human cancers are insufficient to cause apoptosis and cell death but instead facilitate the processes of angiogenesis and tumor dissemination.2,4
Table 1. THE MEAN AND RANGE OF STAINING SCORES FOR NOS2 AND VEGF
Mild dysplasia (n ⫽ 9) Moderate dysplasia (n ⫽ 10) Severe dysplasia (n ⫽ 14) FIGURE 2. Expression of VEGF confined to the basal and suprabasal layer of epithelium in mild dysplasia (DAB/hematoxylin stain, original magnification ⫻40).
NOS2
VEGF
0.6 (0-3) 3.8 (3-4) 4.6 (0-6)
1.2 (0-4) 3.2 (0-5) 3.64 (0-7)
NOTE. Significant correlations were found between NOS2 (P ⬍ .001) and VEGF (P ⬍ .002) expression and the degree of severity of dysplasia.
1459
BRENNAN ET AL
Recently, the effect on NOS2 expression of the chemical carcinogen 7,12-dimethylbenz[a]anthracene (DMBA) was studied in the hamster buccal pouch.24 Immunohistochemical expression of NOS2 expression was increased in DMBA-induced buccal mucosa dysplasia and carcinoma compared with untreated animals. In the latter group, no NOS2 staining was found. It was postulated that NOS2 was implicated in the transformation of normal mucosa through dysplasia to invasive malignancy. However, the explanation for NOS2 activation in dysplastic epithelium is difficult to explain. It may occur as a result of cytokine release from macrophages and other cytokine-releasing inflammatory cells. We found both NOS2 and VEGF staining in cells of the underlying connective tissue. Although these cells are likely to be macrophages (by virtue of their morphology), a double staining technique using the myelomonocytic marker CD68 would be required to confirm the exact cell of origin. Even though there are a considerable number of reports that address this subject (none in head and neck cancer or dysplasia), the exact pathway remains unclear. Hajri et al25 investigated the role of NO in pancreatic adenocarcinoma and suggested that the interaction of tumor cell and macrophage might be due to direct contact or contingent paracrine activating factors. Whatever the activation pathways, the expressions of both NOS2 and VEGF increase with the severity of oral dysplasia, and it is likely that they are implicated in the complex transformation to malignancy. Certainly, the development of an invasive tumor requires the process of angiogenesis to occur at an early stage. This has been shown to occur in such premalignant conditions as breast and cervical dysplasia.12 The interesting question therefore arises about clinically inhibiting NOS2 or VEGF in dysplasia and oral cancer as part of the overall management of these diseases. The use of experimental drugs such as N-(3(aminomethyl)benzyl)acetamidine (or 1400W), a specific antagonist of NOS2, results in reduced tumour growth, thereby confirming the role of NOS2 and NO in this process.26 Because the configuration of the arginine binding site for each NOS isoenzyme is slightly different, the specificity for this agent to NOS2 is 2,000 times that of either NOS1 or NOS3. Therefore the side effects of these specific drugs (eg, potential modulation of blood pressure via NOS3 inhibition) are minimal. Monoclonal antibodies against VEGF protein are currently undergoing early clinical trials in advanced ovarian cancer. It is possible that these or similar agents may also become part of the overall management of cancer and precancerous conditions such as dysplasia in the future.
A significant relationship was found between NOS2 and VEGF expression in oral dysplasia. Further research is required to more fully establish the roles of VEGF and NOS2 in the development and progression of oral cancer and to assess the possible therapeutic applications of inhibiting the expression of these proteins. Acknowledgments We thank Dr Anne Spedding, Consultant Histopathologist, and Sue Tant for technical assistance. We are also grateful to Bernard Higgins, Medical Statistician, University of Portsmouth, and the Photographic Department, Poole Hospital, for their help in this study.
References 1. Melillo G, Musso T, Sica A, et al: A hypoxia-responsive element mediates a novel pathway of activation of the inducible nitric oxide promoter. J Exp Med 182:1683, 1995 2. Thomsen LL, Miles DW: Role of nitric oxide in tumour progression: Lessons from human tumours. Cancer Metastasis Rev 17:107, 1998 3. Fukumara D, Jain RK: Role of nitric oxide in angiogenesis and microcirculation in tumors. Cancer Metastasis Rev 17:77, 1998 4. Gallo O, Masini E, Morbidelli L, et al: Role of nitric oxide in angiogenesis and tumour progression in head and neck cancer. J Natl Cancer Inst 90:587, 1998 5. Chiarugi V, Magnelli L, Gallo O: Cox-2, iNOS and p53 as play-markers of tumor angiogenesis. Int J Mol Med 2:715, 1998 6. Ziche M, Morbidelli L, Choudhuri R, et al: Nitric oxide synthase lies downstream from vascular endothelial growth factor-induced but not basic fibroblast growth factor-induced angiogenesis. J Clin Invest 99:2625, 1997 7. Kim KJ, Li B, Winer J, et al: Inhibition of vascular endothelial growth factor-induced angiogenesis suppresses tumour growth in vivo. Nature 362:841, 1993 8. Millauer B, Shawver LK, Plate KH, et al: Glioblastoma growth inhibited in vivo by a dominant-negative Flk-1 mutant. Nature 843:576, 1994 9. Keiser A, Weich HA, Brandner G, et al: p53 potentiates protein kinase C induction of vascular endothelial growth factor. Oncogene 9:963, 1994 10. Mazure NM, Chen EY, Yeh P, et al: Oncogenic transformation and hypoxia synergistically act to modulate vascular endothelial growth factor expression. Cancer Res 56:3436, 1996 11. Forsythe JA, Jiang BH, Iyer NV, et al: Activation of vascular endothelial growth factor gene transcription gene by hypoxiainducible factor 1. Mol Cell Biol 16:4604, 1996 12. Hanahan D, Folkman J: Patterns and emerging mechanisms of the angiogenic switch during tumorigenesis. Cell 86:353, 1996 13. Maeda T, Matsumura S, Hiranuma H, et al: Expression of vascular endothelial growth factor in human oral squamous cell carcinoma: Its association with tumour progression and p53 gene status. J Clin Pathol 51:771, 1998 14. Moriyama M, Kumagai S, Kawashiri S, et al: Immunohistochemical study of tumour angiogenesis in oral squamous cell carcinoma. Oral Oncol 33:369, 1997 15. Denhart BC, Guidi AJ, Tognazzi K, et al: Vascular permeability factor/vascular endothelial growth factor and its receptors in oral and laryngeal squamous cell carcinoma and dysplasia. Lab Invest 77:659, 1997 16. Feng L, Lee JJ, Hong WK, et al: Expression of vascular endothelial growth factor and microvessel density in head and neck tumorigenesis. Clin Cancer Res 6:2821, 2000 17. Carlile J, Harada K, Baillie R, et al: Vascular endothelial growth factor (VEGF) expression in oral tissues: Possible relevance to angiogenesis, tumour progression and field cancerisation. J Oral Pathol Med 30:449, 2001 18. Brennan PA, Palacios-Callender M, Sinclair D, et al: Does expression of inducible nitric oxide synthase correlate with se-
1460
19.
20.
21. 22.
verity of oral epithelial dysplasia? J Craniomaxillofac Surg 28: 44, 2000 Brennan PA, Palacios-Callender M, Spedding AV, et al: Does type II nitric oxide synthase expression correlate with cellular proliferation in oral squamous cell carcinoma and dysplasia? Head Neck 27:217, 2001 Brennan PA, Umar T, Zaki GA, et al: Are myoepithelial cells responsible for the widespread expression of inducible nitric oxide synthase in pleomorphic adenoma? An immunohistochemical study. J Oral Pathol Med 29:279, 2000 Dvorak HF, Brown LF, Detmar M, et al: Vascular permeability factor/vascular endothelial growth factor, microvascular hyperpermeability and angiogenesis. Am J Pathol 146:1029, 1995 Parenti A, Morbidelli L, Cui XL, et al: Nitric oxide is an upstream signal of vascular endothelial growth factor-induced
NOS2 AND VEGF IN DYSPLASIA
23. 24.
25.
26.
extracellular signal-regulated kinase1/2 activation in postcapillary endothelium. J Biol Chem 273:4220, 1998 Albina JE: On the expression of nitric oxide synthase by human macrophages. Why no NO? J Leukoc Biol 58:643, 1995 Chen YK, Lin LM: Immunohistochemical expression of inducible nitric oxide synthase in DMBA-induced hamster buccal pouch carcinogenesis. Oral Oncol 36:221, 2000 Hajri A, Metzger E, Vallat F, et al: Role of nitric oxide in pancreatic tumour growth: In vivo and in vitro studies. Br J Cancer 78:841, 1998 Thomsen L, Scott JMJ, Topley P, et al: Selective inhibition of inducible nitric oxide synthase inhibits tumour growth in vivo: Studies with 1400W, a novel inhibitor. Cancer Res 57:3300, 1997