Published ahead of print on February 8, 2007, doi:10.1164/rccm.200612-1770OC
© 2007 American Thoracic Society doi: 10.1164/rccm.200612-1770OC
The Influence of Fluticasone Inhalation on Markers of Carcinogenesis in Bronchial Epithelium1 Department of Pulmonology, VU University Medical Center (VUMC), Amsterdam, The Netherlands; 2 Departments of Biometrics and 3 Thoracic Oncology, The Netherlands Cancer Institute/Antoni van Leeuwenhoek Hospital (NKI/AvL), Amsterdam, The Netherlands; and 4 Department of Pathology, VU University Medical Center (VUMC), Amsterdam, The Netherlands Correspondence and requests for reprints should be addressed to E. F. Smit, M.D., Ph.D., Department of Pulmonology, VU Medical Centre, PO Box 7057, 1007 MB Amsterdam, The Netherlands. E-mail: ef.smit{at}vumc.nl
Rationale: Bronchial epithelium exposed to cigarette smoke undergoes a series of histologic changes that may ultimately lead to invasive cancer. Inhaled corticosteroids reduce the number of lung tumors developing in rats exposed to cigarette smoke. Objectives: We studied the effect of inhaled fluticasone on premalignant lesions in smokers and patients curatively treated for head and neck cancer or lung cancer. Methods: Participants were screened for premalignant lesions by bronchoscopy. Biopsies were taken from three to five locations and classified using WHO criteria. In case of a metaplasia index of > 15%, participants were randomized to receive a powder inhalation device containing either fluticasone 500 µg or a placebo, to be used twice a day. After 6 months, biopsies were obtained from the same locations as previously sampled. Efficacy of treatment was assessed by reversal of metaplasia/dysplasia; secondary endpoints were reversal of increased p53 and KI-67 immunoreactivity and expression of human telomerase reverse transcriptase. Measurements and Main Results: Of the 201 subjects that were screened, 108 were included. Mean age was 53 yr (35–71), mean number of pack-years 48 (18–99), mean metaplasia index 48%, and 32% had some degree of dysplasia at baseline. The two treatment arms did not differ with respect to response or change in either metaplasia index or the expression of the markers p53, KI-67, or human telomerase reverse transcriptase. Conclusions: Inhaled fluticasone in a dose of 500 µg twice a day does not affect the natural course of premalignant lesions in the central airways.
Key Words: cancer chemoprevention • hTERT • KI-67 • p53 • inhalational corticosteroid
Despite advances in treatment, lung cancer kills more patients than any other malignancy (1). Its main cause is tobacco smoking (2), which is widely recognized as a danger to public health. In spite of anti-smoking campaigns and legislation intended to reduce the number of smokers, lung cancer will remain an important cause of cancer mortality in the foreseeable future. The process of carcinogenesis takes many years, and although lung cancer risk declines in the years after quitting, many former smokers develop lung cancer. This is illustrated by a study in which 42% of 1,039 patients with lung cancer were former smokers (3). Of the male patients, 15% had stopped smoking over 15 years before diagnosis. Successes in the preventive treatment of risk factors in cardiovascular medicine suggest that a similar course of action may be helpful in the fight against lung cancer. In theory, cancer chemoprevention is feasible. Certain drugs or (dietary) supplements may counteract the effects of carcinogenic substances on healthy subjects at risk of developing cancer. Many mechanisms of carcinogenesis have been elucidated over the past decades, and drugs affecting this process have proven their benefit in in vitro and animal studies. In large clinical trials tamoxifen reduced the risk of recurrence of breast cancer (4), which is a proof of principle for cancer chemoprevention. Inhalational corticosteroids (ICS) are promising for lung cancer chemoprevention. Corticosteroids inhibit tumor growth in vitro (5), and both ingested and inhaled, this class of drugs reduces the number of lung tumors developing in rats exposed to cigarette smoke (6) and in a transgenic (p53/p16 null) mouse model treated with benzo[a]pyrene (7). ICS have been used in the treatment of asthma for years and were proven safe in long-term use, which is important if a substance is to be used for large-scale chemoprevention. Bronchial epithelium exposed to cigarette smoke will often undergo a sequence of histologic changes, from hyperplasia through squamous metaplasia and different grades of dysplasia. These changes may eventually lead to invasive squamous cell carcinoma. This sequential morphologic progression is thought to be driven by an accumulation of genetic and epigenetic changes. In this trial we examined the effect of fluticasone on premalignant epithelial lesions of the bronchus: reversal of histologic abnormalities of the bronchial epithelium was the primary endpoint. As a secondary endpoint we chose the expression of three biomarkers: p53, KI-67, and hTERT (human telomerase reverse transcriptase, a subunit of the enzyme telomerase, important for tumor cell immortalization). These markers are indicative of changes on a molecular level and may add to the predictive value of histology (8–11). Some of the results of this study have been previously reported in the form of an abstract (12).
Subjects and Screening Volunteers with a smoking history of over 20 pack years (one pack of cigarettes a day for over 20 yr or equivalent) were recruited by advertisements in local and national newspapers. Patients who had undergone treatment with curative intent of head and neck or lung cancer of any stage and did not show signs of recurrence on computed tomography (CT) were also asked to participate. Subjects with serious comorbid disease, FEV1 values below 1,000 ml, or use of systemic or inhaled corticosteroids in the year before enrollment were excluded. Participants were screened with autofluorescence- (VUMC) or white light bronchoscopy (NKI). Biopsies were taken from three to five locations. Standard locations (main carina, and carinas of right and left upper lobes) were biopsied when no suspicious lesions could be detected. From each location, one biopsy was snap-frozen in liquid nitrogen, and another was formalin fixed and paraffin embedded for immunohistochemistry and histologic assessment on a series of hematoxylin and eosin–stained sections by a staff pathologist according to WHO guidelines. At least one biopsy was to contain squamous metaplasia or dysplasia for the subject to be included in the study. Protocols were approved by the institutional review boards, and written informed consent was obtained from all subjects.
Randomization and Treatment
Evaluation of Efficacy
Markers. Twenty-five 10-µm sections were cut from the snap-frozen biopsies and lysed in RNAzol Bee (Campro Scientific, Veenendaal, The Netherlands). RNA was isolated according to the manufacturer's instructions. Levels of hTERT mRNA expression were determined with a quantitative one-step real-time reverse transcriptase PCR, using the light cycler instrument and the light cycler TeloTAGGG hTERT quantification kit (Roche, Mannheim, Germany) according to the manufacturer's instructions. The housekeeping gene porphobilinogen deaminase was amplified and detected in the same reaction, serving as a reference as well as internal control for RNA quality. Experiments were performed in duplicate. Normalized mean hTERT level was used for analysis. p53 and KI-67 immunostaining was performed as previously described (9). Antibodies (DO-7 and MIB-1) were obtained from Dako (Glostrup, Denmark).
Statistics
Characteristics Between February 2002 and October 2004, 201 subjects were screened. Two were excluded because of lung cancer detected during initial screening. Of the volunteers who were screened, 108 were included (Figure 1). Mean metaplasia index at baseline was 47% (median 33%); 35% had some degree of dysplasia (Table 2). Baseline characteristics were well balanced among the intervention and the placebo group, except for one: seven subjects with a history of lung cancer or head and neck cancer (of whom three were treated more than 7 yr before) were allocated to the placebo group, whereas two subjects with a history of head and neck cancer were allocated to the fluticasone group. Disease stage ranged from T1N0 to T2N1 in lung and T1N0 to T4N2 in head and neck cancer; median time since treatment was 33 months (range, 9 mo to 10 yr; see Table E1 in the online supplement). In one subject a carcinoid tumor was discovered at the time of the second bronchoscopy; another developed bronchioloalveolar carcinoma. Both were assigned to the placebo group and are included in the analysis. Seventeen subjects failed to complete the trial according to protocol. One subject, allocated to the fluticasone arm of the study, underwent the second bronchoscopy 1 month earlier than planned because of side effects (recurrent stomatitis). This person had a complete response and was included in the analysis. Of the 16 subjects who were not included in the analysis, 9 were in the placebo and 7 in the fluticasone arm (Figure 1). Subjects failed to complete the trial because of side effects, because they refused to undergo the second bronchoscopy, or were lost to follow up. One subject was excluded because an acute exacerbation of chronic obstructive pulmonary disease necessitated treatment with systemic corticosteroids. Data on five subjects were incomplete. The average number of doses taken per month was 56 (20–60; median 57, SD 5.0), without any difference between subjects assigned to the fluticasone or placebo arm. Data from 92 subjects were available for analysis. Six subjects out of 92 quit smoking during the study, of which 3 (6.8%) were allocated to the placebo group and 3 (6.4%) to the fluticasone group. Of the subjects allocated to the placebo arm, 21 were subsequently included in the crossover part of the trial. Adverse events related to toxicity of study medication were reported on 68 occasions and were limited to thrush, cough, and hoarseness of the voice. All toxicity-related adverse events were below CTC grade 3.
Histology Lesion-specific analysis. At baseline a total of 437 locations were biopsied. Of these locations 320 yielded biopsies of sufficient quality from subjects who finished the trial according to protocol. After 6 months, 291 locations that were sampled at baseline were biopsied a second time. Fifty-three locations were sampled that were not biopsied at baseline. The complete response rate on a lesion level was 27% in the fluticasone group and 22% in the placebo group (p > 0.30; see Table 3). The progressive disease rate was 13% and 18% for fluticasone and placebo, respectively. For analysis of the crossover part of the trial, the response of 39 locations (21 subjects) after using open-label fluticasone for 6 months was compared with the response of all locations in the placebo arm. Again, fluticasone did not influence the response (data not shown).
Subject-specific analysis. Ninety-two subjects had results suitable for analysis. Complete response rate was 34% in the fluticasone group and 38% in the placebo group; progressive disease was found in 43% in the fluticasone and 38% in the placebo group (p = 0.75; see Table 3). Stratified analysis showed that institution, number of pack-years, sex, and history of cancer did not confound the main comparison. Before intervention mean metaplasia index was 50% and 51%, and after treatment 35% and 37% for fluticasone and placebo groups, respectively (p = 0.78; Wilcoxon 2-sided).
Markers
KI-67.
hTERT.
We have performed a randomized, controlled trial of fluticasone, followed by a one-way crossover from placebo to active treatment, in a high-risk group of smokers. Lung cancer in smokers is a relatively rare occurrence, and when used as the endpoint in a chemoprevention trial, it would take a large cohort and a long follow-up time to obtain a number of cases large enough for sufficient statistical power. Therefore, we investigated the chemopreventive potential of fluticasone by examining its effect on squamous precursor lesions of the bronchus. No difference in change of histology between the fluticasone or placebo arms could be detected after a 6-month treatment period. The same result was found in the crossover part of the trial. However, because of the absence of a proper control group, the value of the crossover data is limited. A similar study has been performed by Lam and colleagues (13). Their study showed that inhalation of budesonide (like fluticasone and ICS) has no effect on squamous dysplasia in high-risk smokers, but does cause a reduction of the percentage of p53-positive cells as well as the number of CT-detected nodules. Metaplasia rather than dysplasia was used as inclusion criterion because dysplastic lesions are relatively rare in this population and we were unable to screen enough volunteers to obtain an equal number of subjects with dysplastic lesions. The metaplasia index has been used before as endpoint for chemoprevention studies (14), and although it has not been validated it is clearly related to smoking (15). Metaplastic changes are manifestations of the repair of epithelial damage (16), which may be caused by smoking-induced inflammation (17, 18). The effect of ICS on carcinogenesis could stem from an influence on this inflammation. If we found a reaction of the metaplasia index to treatment in this study, this would have been indicative of an influence of ICS early on in the process of carcinogenesis. In a study by Lee and coworkers, the number of subjects who quit smoking was of greater influence on the metaplasia index than the intervention that was tested (19). The number of subjects that quit in the current study was small (just over 6% in both arms) and unlikely to affect the comparison between the treatment arms. Could results be different in former smokers? Retinoic acid, a substance that has a chemopreventive effect in oral squamous cell carcinoma, was found to have a lung cancer–promoting effect in subjects who continued to smoke (20). There were indications of a beneficial effect in former smokers. However, no effect was found on metaplasia in a trial of retinoic acid exclusively in former smokers (21). The number of former smokers in our study was too small to make any statement about a possible effect of fluticasone on this subgroup. To investigate a possible effect later on in the process of carcinogenesis, we also included (change of) dysplasia in our analysis as well as three biomarkers: KI-67, p53, and hTERT. KI-67 protein is a proliferation marker, which cells in any stage of proliferation will express. The ratio of KI-67–positive and –negative cells, the proliferation index, is used as a measure of the speed of turnover, which is higher in (pre)malignant lesions. Smoking causes an elevated proliferative index regardless of the presence of premalignant lesions (14). The change in KI-67 percentage did not show a significant difference between the two treatment arms. We have shown that suprabasal p53 staining has additional value as a marker in premalignant lesions of the lung, in a study of patients with a history of lung cancer or high-grade premalignancy, in which lung cancer was the primary outcome (9). In the current study there was convincing suprabasal staining in lesions of only one subject. Possibly p53 mutation does not occur as often in this population consisting mainly of healthy smokers. In addition to scoring the lesions with suprabasal staining, the samples were categorized according to the percentage of p53-positive cells (no staining; < 10% staining; between 10 and 50%; and > 50%). As opposed to the results of the study by Lam and colleagues (13), we found no significant difference between the placebo and fluticasone arms with respect to the percentage of p53-positive cells. The reason for the difference in findings with regard to p53 percentage may be that all study subjects in the budesonide study had bronchial dysplasia, whereas our main inclusion criterion was squamous metaplasia. Telomerase activation is a crucial step in development of most tumors, since it allows tumor cells to bypass senescence and divide indefinitely. hTERT, one of its components, is the rate-limiting enzyme (22). Increased hTERT expression was found in 33 out of 38 squamous cell carcinomas in a recent study (23). Elevated expression in bronchial epithelium was found to be predictive of malignant transformation in several recent retrospective series including our own (10, 24, 25). We found no significant difference in the change of mean hTERT level between the placebo and fluticasone groups. Corticosteroids were previously proven effective in reducing the number of lung tumors in A/J mice and rats exposed to cigarette smoke. In a study by Yao and coworkers (26), the expression profile of mouse lung tumors treated with budesonide was analyzed by microarray. The authors concluded that budesonide affected mouse lung tumorigenesis through modulation of the Bcl-2– and caspase-2–regulated apoptosis pathways as well as the Mad2/3-regulated mitotic checkpoint. As Lam and colleagues (13) speculated, the explanation for the discrepancy between the promising results of trials of ICS in rodents and the lack of effect in humans may lie in the fact that in the rodent lung carcinogenesis model the tumors that develop are all classified as adenocarcinomas and their precursor lesions as adenomas. Since the tumors that originate in epithelium of the large bronchi are squamous carcinomas, the mouse model may not be appropriate for investigating central bronchial carcinogenesis. The reduction in CT detected nodules in the budesonide study (13) could be indicative of an effect on adenomatous precursor lesions, similar to the lesions that develop in the mouse model. We conclude that ICS do not influence squamous bronchial lesions. A possible effect on peripheral adenomatous lesions found in a previous study warrants further investigation.
Supported by a grant from GlaxoSmithKline, Zeist, The Netherlands. This article has an online supplement, which is accessible from this issue's table of contents at www.atsjournals.org Originally Published in Press as DOI: 10.1164/rccm.200612-1770OC on February 8, 2007
Conflict of Interest Statement: R.M.v.d.B. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. H.v.T. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. N.v.Z. has received £2,500 for serving on an advisory board for Eli Lilly and Company and has received £14.000 during the last three years for speaking at conferences organized by Eli Lilly and Company and AstraZeneca. C.V. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. A.P. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. C.K. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. T.G.S. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. P.B. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. K.G. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. W.J.M. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. P.J.F.S. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. P.E.P. received Received in original form December 6, 2006; accepted in final form February 8, 2007
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
AT A GLANCE COMMENTARY
2= 0.05). The SAS version 9 program (SAS Institute, Inc., Cary, NC) was used for statistical tests. Categorical variables were tested using Fisher exact test (location specific analysis) or Jonkheere Terpstra test when appropriate. Continuous variables were tested by t tests. Logistic regression analysis was performed to adjust for sex, institution where screening and follow up took place, number of pack-years, and history of cancer. For the lesion-specific analysis, patients were considered a random factor in a mixed-effect model to account for multiple observations per patient. 
57,176 for the studies discussed in this manuscript in 2004/05; 
