Evidence of neoangiogenesis and an increase in the number of proliferating cells within the bronchial epithelium of smokers
Article first published online: 17 SEP 2002
Copyright © 2002 American Cancer Society
Volume 95, Issue 7, pages 1539–1545, 1 October 2002
How to Cite
Hiroshima, K., Iyoda, A., Shibuya, K., Hoshino, H., Haga, Y., Toyozaki, T., Shiba, M., Baba, M., Fujisawa, T. and Ohwada, H. (2002), Evidence of neoangiogenesis and an increase in the number of proliferating cells within the bronchial epithelium of smokers. Cancer, 95: 1539–1545. doi: 10.1002/cncr.10850
- Issue published online: 17 SEP 2002
- Article first published online: 17 SEP 2002
- Manuscript Accepted: 17 APR 2002
- Manuscript Revised: 8 APR 2002
- Manuscript Received: 22 OCT 2001
- Smoking Research Foundation
- lung carcinoma;
- vascular endothelial growth factor;
Normal bronchial epithelium gradually acquires cellular and genetic changes that result in the formation of invasive tumors. The objective of this study was to evaluate the degree of proliferative change and the amount of neovascularization in both normal and preneoplastic lesions in smokers who were at high risk for developing lung carcinoma.
The authors studied bronchial biopsy specimens from 7 nonsmokers and 52 smokers. Immunohistochemical staining of the specimens with antibodies for the presence of p53 protein, Ki-67 and CD34 antigens, and vascular endothelial growth factor was performed. The proliferation index (PI) was assessed by immunohistochemical staining for Ki-67 antigen.
Overexpression of p53 protein was observed frequently in regions of squamous dysplasia and in squamous cell carcinoma tissue. The PI of normal epithelium from smokers was increased compared with nonsmokers, and the difference was statistically significant (P < 0.05). The microvessel count (MC) in normal mucosa obtained from smokers was higher compared with the MC in normal mucosa obtained from nonsmokers (P < 0.05). A significant difference in MC also was observed between regions of squamous metaplasia or dysplasia with projections of capillary loops into the bronchial mucosa and similar lesions without capillary loops (P < 0.005); however, there was no difference in either the PI or the incidence of p53 overexpression between these groups.
These results show that smoking appears to induce both a proliferative response and neovascularization in bronchial mucosa. The projection of capillary loops into the bronchial mucosa also may be a result of neovascularization occurring within the lamina propria of the bronchial wall. Cancer 2002;95:1539–45. © 2002 American Cancer Society.
Lung carcinoma is the leading cause of cancer deaths throughout the industrialized world, and tobacco smoking is recognized as the major cause of lung carcinoma.1 Recent studies indicate that lung carcinoma is not the result of a sudden transformational event occurring within the bronchial epithelium, but, in fact, is due to a multistep process.2 Normal bronchial epithelium gradually acquires cellular and genetic changes that ultimately result in the formation of invasive tumors. The third edition of the World Health Organization classification of lung tumors recognizes the importance of this process by including preinvasive lesions within a separate section. Mucosal changes in the large airways that may precede invasive squamous cell carcinoma (SCC) include hyperplasia, squamous metaplasia, squamous dysplasia, and carcinoma in situ. Many of these changes are not observed during routine white-light bronchoscopy, and their detection requires the use of fluorescence bronchoscopy.3
Areas consisting of capillary blood vessels closely juxtaposed to and projecting into metaplastic and dysplastic squamous bronchial epithelium often are observed in bronchial preneoplastic lesions.4 This type of preneoplastic change was called angiogenic squamous dysplasia (ASD) by Keith et al.5 It also has been reported that ASD occurs in cervical intraepithelial neoplasia.6
It has been suggested that histologically normal bronchial epithelium in smokers harbors genetic changes and that these changes do not disappear with cessation of smoking.7, 8 We previously reported that irreversible genetic alterations, similar to those seen in adjacent bronchial carcinoma, are detectable within regions of squamous dysplasia.9 These observations are consistent with a multistep model of carcinogenesis and with the field carcinogenesis model of lung carcinoma.
Microsatellite alterations, hypermethylation of the p16 gene, and high telomerase activity have been observed in preneoplastic lesions of bronchial epithelium.10–13 However, whether these changes are premalignant and/or can be used to predict lung carcinoma is unknown.
It is believed that abnormal epithelial proliferation is one of the hallmarks of tumorigenesis. The objectives of this study were to evaluate the amount of proliferative activity and the degree of neoangiogenesis in bronchial mucosa obtained from smokers and to compare these values with the values observed in nonsmokers.
MATERIALS AND METHODS
Sputum examinations were performed as part of mass screening of patients who were at high risk for lung carcinoma. Participants with sputum cytology that was suspicious or positive for malignancy were referred to our institute and investigated using both white-light and fluorescence bronchoscopy.3 The nonsmokers included in this study were patients who were referred to our institute because of chronic cough, bloody sputum, or an abnormal shadow on chest X-ray but who did not have evidence of a major airway tumor. Biopsy samples from smokers were obtained between 1995 and 1999, and samples from nonsmokers were obtained between 1991 and 1998. Eighty-three lesions were analyzed retrospectively from 59 patients (52 smokers and 7 nonsmokers). All smokers and one nonsmoker were male; the remaining nonsmokers were female. Biopsies from all of the nonsmokers and from 12 smokers consisted of normal mucosa. The remaining samples from smokers included hyperplasia (n = 13 samples), squamous metaplasia without dysplasia (n = 7 samples), squamous dysplasia (n = 23 samples), and SCC (n = 21 samples).
After fixation in formalin, the sections were embedded in paraffin and stained with hematoxylin and eosin. The sections were then immunostained with the monoclonal antibody DO-7 (1:1600 dilution; Dako, Glostrup, Denmark), which reacts with wild type and mutant type p53 protein; a monoclonal antibody against Ki-67 antigen (MIB-1) (1:100 dilution; Immunotech, Marseille, France); a monoclonal antibody against CD34 (1:10 dilution; Immunotech); or a polyclonal antibody against vascular endothelial growth factor (VEGF; 1:400 dilution; Santa Cruz Biotechnology, Santa Cruz, CA). The streptavidin-biotin method (Histofine SAB-PO Ki;, Nichirei, Tokyo, Japan) was used for immunohistochemical staining. To improve the staining pattern, the tissues were pretreated by microwave treatment for 15 minutes in 10 mM citrate buffer, pH 6.0, before immunostaining with DO-7, MIB-1, and anti-VEGF antibody. DO-7 immunoreactivity was classified as positive if tumor cell nuclei were stained. The proliferation index (PI) was calculated as the percentage of Ki-67 positive cells in bronchial epithelium or carcinoma after counting at least 1000 nuclei of epithelial cells from these specimens. The evaluation of microvessel count (MC) and the determination of VEGF expression were not performed on all biopsy samples, because appropriate sections were not obtained from the tissue blocks.
We used the method described by Weidner et al. to calculate the highest microvessel density of each biopsy sample examined.14 Sections immunostained with endothelial marker CD34 were scanned at ×100 magnification for areas with the highest vascular density. Individual microvessels were then counted on a ×400 field (i.e., ×40 objective lens and ×10 ocular lens; 0.1885 mm2 per field). The presence of a vessel lumen was not required for a structure to be defined as a microvessel. The MC for each biopsy sample was calculated from the mean counts from two or three areas with the greatest microvessel density.
The immunohistochemical results for PI and MC were analyzed using a Student t test for unpaired data. The p53 protein and VEGF expression data were analyzed using a chi-square test. P values < 0.05 were considered statistically significant.
Frequency of ASD
ASD was observed in 17 of 30 biopsy samples (56.7%) that had evidence of squamous metaplasia or dysplasia.
Overexpression of p53 Protein
None of the biopsies that contained normal epithelium from nonsmokers (n = 7 samples) or from smokers (n = 12 samples) showed overexpression of p53 protein. One of 13 hyperplastic lesions (7.7%) from smokers showed overexpression of p53 protein. Overexpression of p53 protein was observed in 2 of 7 squamous metaplastic lesions (28.6%) (Fig. 1a), in 9 of 23 dysplastic lesions (39.1%), and in 9 of 21 SCC lesions (42.9%) (see Fig. 1b). The incidence of p53 overexpression in biopsy samples with squamous dysplasia or with SCC was significantly greater compared with the incidence observed in biopsies of normal epithelia from nonsmokers (P < 0.05 and P < 0.01, respectively), in biopsies of normal epithelia from smokers (P < 0.05 and P < 0.05, respectively), and in biopsies from hyperplastic lesions (P < 0.05 and P < 0.05, respectively) (see Table 1). It is interesting to note that, in one biopsy, the hyperplastic epithelium showed immunoreactivity for p53 protein, but a synchronous metaplastic lesion did not. There was no difference in the frequency of overexpression of p53 protein between squamous metaplasia or dysplasia with or without projections of capillary loops into the bronchial mucosa.
|Tissue source||No. of patients||Negative||Positive||Incidence (%)|
The nuclei of several basal cells and parabasal cells that were stained with MIB-1 in bronchial epithelium (Fig. 2a) and of Ki-67 positive tumor cells were distributed randomly throughout the tumors in biopsy sample with SCC. The PI of histologically normal epithelium from smokers (n = 12 samples; 1.26% ± 1.17%) was greater compared with the PI observed in biopsy samples from nonsmokers (n = 7 samples; 0.16% ± 0.15%), and this difference was statistically significant (P < 0.05). The PI of biopsies from hyperplastic areas (n = 13 samples; 2.55% ± 1.44%) was significantly greater compared with the PI observed in biopsies of normal epithelium from smokers (P < 0.05) and from nonsmokers (P < 0.001). The PI of biopsies from squamous metaplasia (n = 7 samples; 4.94% ± 3.42%) was significantly higher compared with the PI of biopsies from areas of hyperplasia (P < 0.05) or biopsies of normal mucosa from smokers (P < 0.005) or nonsmokers (P < 0.005). The PI of biopsies from squamous dysplasia (Fig. 2b) (n = 23 samples; 8.45% ± 5.79%) also was significantly greater compared with the PI of biopsies from hyperplasia (P < 0.002) and biopsies of normal mucosa from smokers (P < 0.001) or nonsmokers (P < 0.001). The PI of SCC (n = 21 samples; 14.53% ± 13.78%) was significantly greater compared with the PI of biopsies from hyperplasia (P < 0.005) and biopsies of normal mucosa from smokers (P < 0.005) or nonsmokers (P < 0.001). There was no significant difference in the PI between biopsies of squamous metaplasia, squamous dysplasia, and SCC (Fig. 3). The PI of squamous metaplasia or dysplasia with or without projections of capillary loops into the bronchial mucosa was 8.54% ± 6.18% (n = 17 samples) and 6.43% ± 4.39% (n = 13 samples), respectively, and there was no statistically significant difference between them.
Figure 4a,b shows examples of immunohistochemical staining of normal epithelium (from nonsmokers) and areas of squamous dysplasia (from smokers) for the CD34 antigen. The mean MC of normal mucosa from nonsmokers was 17.00 ± 5.97, whereas the mean MC of normal mucosa from smokers was 26.50 ± 5.66. This difference was statistically significant (P < 0.005). The mean MC of biopsies from hyperplastic lesions was 32.00 ± 6.54, which was significantly greater compared with the mean MC of normal mucosa from nonsmokers or smokers (P < 0.001 and P < 0.05, respectively). The mean MC of biopsies of squamous metaplasia was 27.71 ± 9.60, which was significantly greater compared with the mean MC of normal mucosa from nonsmokers (P < 0.05). The mean MC of biopsies of squamous dysplasia was 34.00 ± 8.38, which, again, was significantly greater compared with the mean MC of normal mucosa from nonsmokers or smokers (P < 0.005 and P < 0.01, respectively) (see Fig. 5). The mean MC of biopsies of SCC was 25.79 ± 10.46. However, because the stromal component of SCC biopsy samples was scant, the mean MC value for SCC could not be compared with the mean MC of other lesions. The mean MC of biopsies of squamous metaplasia or dysplasia that demonstrated projections of capillary loops into the bronchial mucosa (37.58 ± 6.36) was significantly greater compared with the mean MC observed in biopsies of squamous dysplasia or metaplasia without capillary loop projections (27.31 ± 8.77; P < 0.005).
VEGF staining was seen in the bronchial epithelium, in smooth muscle, and in bronchial glands. VEGF expression was observed in 3 of 7 biopsies of normal epithelia from nonsmokers, in 3 of 7 biopsies of normal epithelia from smokers, in 4 of 6 biopsies of hyperplastic lesions, in 10 of 20 squamous dysplastic lesions (Fig. 6), and in 14 of 19 biopsies of SCC. It was not observed in any biopsies of squamous metaplasia. Differences in the frequency of VEGF expression among these lesions were not statistically significant. The degree of VEGF expression was strong in biopsies of SCC compared with the degree of VEGF expression observed in normal epithelium.
The p53 protein inhibits cell proliferation by arresting cells in the G1 phase of the cell cycle. Loss of this activity can lead to neoplastic transformation.15 Mutation of the p53 gene is observed not only in carcinoma but also in preneoplastic lesions.9, 16 We previously reported the occurrence of identical p53 gene mutations in preneoplastic lesions and synchronous lung carcinomas.9 Franklin et al. reported that dysplastic bronchial epithelium from a smoker without overt lung carcinoma had an identical mutation of the p53 gene at several endobronchial sites.16 Positive p53 immunoreactivity within a tumor usually implies the existence of a mutation of the p53 gene,17 because mutated p53 proteins have longer half-lives than the wild type protein.18 However, not all tumors with immunoreactivity for p53 protein harbor mutations of the p53 gene.19, 20 After DNA damage from a variety of agents, transient elevation of wild type p53 protein levels can occur.21, 22 In this study, p53 immunoreactivity was observed in approximately 50% of biopsies that had areas of squamous dysplasia or SCC. In one smoker, the hyperplastic epithelium showed overexpression of p53 protein.
Previous studies have noted a positive correlation between the degree of proliferative activity, as measured by proliferative cell nuclear antigen levels, and the grade of dysplasia in squamous metaplastic lesions in bronchial biopsies.23, 24 Keith et al. reported that the proliferative activity of ASD lesions was significantly greater compared with the proliferative activity of normal epithelium.5 Although asymptomatic volunteers were included in their study, those authors did not specify the smoking status of patients with histologically normal epithelium. In the current study, the PI of normal epithelium from nonsmokers was 0.16%, whereas the PI of normal epithelium from smokers was 1.26%. The PI of normal epithelium from nonsmokers in the current report was similar to that found in a previous report.25 Thus, smoking appears to elicit a proliferative response in bronchial epithelium. This may be due to airway injury and chronic inflammation of bronchial epithelium or to proliferative stimuli resulting from early genetic mutations. We also observed that biopsy samples that consisted of areas of squamous metaplasia, dysplasia, or SCC from smokers had even greater proliferative activity. Because this increase of PI occurs prior to the development of preneoplastic lesions, we suggest that an increase in proliferating cells may be the first step in the multistep process of lung carcinogenesis. Long-term follow-up of these patients with increased proliferative activity within their bronchial epithelia is needed to evaluate the prognostic significance of these observations.
Angiogenesis is essential for solid tumor growth26 and metastasis.27 The significance of neoangiogenesis in the development of cervical intraepithelial neoplasia,6, 28, 29 bronchial tree carcinoma,30 and carcinoma of the oral mucosa31 has been demonstrated previously. Weidner et al. reported that the MC is a predictor of breast carcinoma metastasis.14 However, Tae et al.32 reported that the MC both in premalignant lesions and in carcinoma of head and neck was significantly lower compared with the MC measured in normal control tissue. Fontanini et al., in a study of patients with nonsmall cell lung carcinoma, also noted that metaplastic lesions had a lower MC compared with dysplastic lesions or regions of carcinoma in situ and, thus, hypothesized that they did not represent precursor malignant change in the bronchial epithelium.4 In the current study, the MC of areas of normal epithelium, hyperplasia, and squamous metaplasia in smokers was greater compared with the MC of normal epithelium from nonsmokers. It is suggested that, in smokers, neoangiogenesis occurs in histologically normal bronchi and that this change persists as the tissue evolves to hyperplasia, squamous metaplasia, and dysplasia. To our knowledge, a direct correlation between bronchoscopic biopsy sample neovascularity and a patient's smoking status has not been reported previously. Our results suggest that tobacco smoke stimulates angiogenesis within bronchial epithelium, and it may be an early step in the progression of lung carcinoma.
VEGF is a key angiogenic factor and a regulator of endothelial function.33 Obermair et al. reported that the MC increased with increasing cervical dysplasia and observed a significant correlation between high VEGF expression levels and high MC.29 In our study, the expression of VEGF in SCC lesions was greater compared with the expression of VEGF observed in other lesions, although the difference was not statistically significant. Although neovascularization occurs in preneoplastic lesions, the correlation between VEGF expression and neovascularization has not been clarified. The antibody against VEGF used in this study reacts with the 121 amino acid, 165 amino acid, and 189 amino acid splice variants of VEGF. A previous report suggested that expression of the VEGF 189 amino acid mRNA isoform (VEGF 189) is correlated with tumor angiogenesis.34 The use of a specific antibody against VEGF 189 may demonstrate a correlation between VEGF expression and the observed neovascularization of bronchial epithelium from smokers.
ASD was observed in > 50% of biopsy samples that consisted of areas of squamous metaplasia or dysplasia in this study. From our analysis of serial sections of these specimens, we hypothesize that the capillaries observed in ASD are stromal vascular papillae extending into the dysplastic epithelium.
The MC of lesions with ASD was higher compared with the MC of lesions without ASD, and the difference was statistically significant. This result suggests that the growth of stromal vascular papillae into the epithelium is an important part of the process of neovascularization. There was no statistically significant difference in either the PI or the incidence of overexpression of p53 protein between bronchial lesions with and without evidence of ASD. The presence of ASD, therefore, does not appear to be related to the proliferative activity of the bronchial epithelium. Keith et al. reported that 53% of lesions with ASD had demonstrable LOH at the chromosomal region 3p but that lesions without ASD had a similar rate of LOH.5 This may indicate that ASD is due to a mechanism similar to the mechanism that initiates neovascularization in the lamina propria of the bronchial wall. Thus, ASD may serve as an important histologic marker of preneoplastic lesions; however, currently, its biologic significance is unclear.
The proliferative activity of normal epithelium from smokers was greater compared with the activity of normal epithelium from nonsmokers. An increased number of microvessels also was observed in normal mucosa from smokers compared with normal mucosa from nonsmokers. ASD was observed in approximately 50% of metaplastic or dysplastic lesions. The same mechanism that results in neovascularization of the lamina propria of the bronchial wall may be responsible for the development of ASD.
The authors thank Ms. Ayaka Sato, Ms. Michiko Hanazono, Ms. Kazuko Abe, and Ms. Tamiyo Taniguchi for their technical assistance.
- 1World Health Organization. Histological typing of lung and pleural tumours, 3rd ed. Berlin: Springer-Verlag, 1999: 28–30.