Smoking prevention will decrease lung cancer incidence in time. However,early detection would improve lung cancer prognosis in subjects at risk provided that specific markers could be identified. We previously reported that retinoic acid receptor (RAR) and retinoid X receptor(RXR) expression was altered in lung tumors. RAR-βgene status could be derived from corresponding allelotyping and immunohistochemistry data. We now report the continued study on lung cancer precursor lesions. Fluorescence PCR-based assays were used for allelotyping at the RAR/RXR loci of: (a) 66 lung precursor lesions found at the free resection margins of 41 patients undergoing surgery for lung cancer (+ 31 paired tumors); and(b) bronchial cells also found at the free resection margins from 16 current and 8 never smokers operated on for noncancerous diseases. Three microsatellites located at 3p14–21 and 9p21 were also used for interwork comparison. Immunohistochemistry was additionally performed to evaluate P53 and RAR-β expression in precursor lesions. χ2 tests showed significant differences(P < 0.05) when comparing the results obtained from never smokers, smokers, squamous metaplasia,dysplasia + in situ carcinoma, and tumors. Microsatellite changes occurred frequently in all samples, but without specificity for any group (P < 0.08–0.52). They were globally correlated with tobacco exposure(P < 0.04), for which the RAR-γ marker appeared as a preferential target(P < 0.004). Few reparation error phenotypes were observed, mostly at the RXR-α and RXR-γ markers for which combined changes were also linearly increasing from never smokers to dysplasia + in situ carcinoma(P < 0.05 and P < 0.03). RAR-β marker losses also increased from the first to the last group studied (P < 0.01), with a concomitant decrease in RAR-β protein expression and correlated p53 increased immunoreactivity (P < 0.02). Losses at 3p14, 3p21, and P16 were frequent, but no significant differences between groups could be found. Unexpectedly, high constitutive homozygosity was observed near the RAR-α locus in squamous cell lung cancer cases. RARs/RXRs form homodimers or heterodimers involved in ligand binding. Their added alterations could result in a state of functional vitamin A deficiency in the affected bronchial cells. Further deletion events drawn from a limited repertoire of specific regions such as 3p14–21 and 9p21 could subsequently drive the deficient cells to invasive carcinoma.

Stepwise modifications affecting the microscopic organization of the bronchial epithelium precede its malignant transformation. They are described as squamous metaplasia, dysplasia, and ISC3and are widely accepted as the natural history of squamous cell lung carcinoma. Metaplasia is the physiological repair process for the injured bronchial epithelium and is reversible, whereas dysplasia and ISC have been shown to be specific preneoplastic lesions(1, 2, 3). They are not believed to precede small cell lung cancer and proximal adenocarcinoma, although they are commonly associated with these tumors. Lung cancer precursor lesions demonstrate molecular alterations affecting different chromosome loci (for review,see Ref. 4). These alterations have also been observed in lung tumors and even in the normal bronchial mucosa of smokers(5, 6), where morphological changes caused by cigarette smoking were first noted years ago (7). It results in a“field cancerization” in which tobacco smoke carcinogens injure the entire lung. In the past, studies of premalignant lesions were difficult to perform due to technical limitations. However, new methods have been designed to visually improve identification of these lesions using fluorescence-based fibroscopy. Tissue microdissection has also been optimized to prepare specific DNAs (8) for genome screening for microsatellite changes such as LOH which are being used to locate specific lung cancer tumor supressor genes(9, 10, 11).

Vitamin A and related retinoids are known to regulate normal lung development, maturation, and maintenance of the bronchial epithelium. Chronic vitamin A deficiency in hamsters results in the replacement of normal tracheal epithelium by a pseudostratified squamous epithelium. Reversal occurs when vitamin A is restored to the diet. A similar phenomenon occurs in vitro in retinoid-deprived bronchial cells (12). The control of gene expression by retinoids is complex and depends on the nature of the ligand, the type of ligand-binding proteins, and the interacting nuclear retinoid receptor genes. They include two different families: the RAR and RXR, with three subtypes for each (α, β, and γ) and several isoforms arising from promoter usage and alternate splicing. In addition, the RARs/RXRs form homodimers and/or heterodimers that bind to cis-acting response elements of retinoid target genes and interact also with varied coactivators or corepressors. RXRs are unusual because they bind to their response elements as homodimers. They also associate with many other hydrophobic ligand receptors such as the peroxisome proliferator-activated receptors. Thus, the retinoids extend their function to cross-modulate specific cell surface receptor signaling pathways (for review, see Ref. 13).

Retinoids have been tested in cancer prevention, but with somewhat puzzling results (14, 15). RAR-β is the best-studied member of the RAR family in the lung cancer process. It is believed to function as a tumor supressor gene (16, 17). By studying nucleic acids as well as the encoded proteins(18), we reported that RAR/RXR had modified combined expression in lung tumors. The extension of the study to lung precursor lesions was difficult because tissue fixation is necessary for reliable identification and results in nucleic acid degradation. Moreover, the small size of the lesion precludes extensive immunohistochemistry screening. However, a microsatellite located near the RAR-β gene (D3S1283) was found to be very useful in reflecting the gene status in lung tumors. Therefore, RARs/RXRs were targeted for more allelotyping studies testing RAR-β protein expression in premalignant lesions in an effort to further define the early roles of RAR/RXR in lung cancer.

Patient Population.

The patient population (lung cancer patients, n = 41) included patients who underwent lung surgical resection with a curative intent between 1989 and 1998 at the Fédération de Pneumologie, Center Hospitalier Universitaire (Nancy, France). pTNM(tumor-node-metastasis) tumor staging and histological grading were performed following the WHO guidelines. Examination of H&E-stained sections of the free resection margins detected a lung cancer precursor lesion that was graded by consensus as squamous metaplasia, dysplasia,or ISC as described previously (6). These samples were added to specimens left over from our previous study(2) of six subjects for whom more than one lung cancer precursor lesion was found by serial sections of the bronchial tree.

A control population without lung cancer was recruited from the archived samples of consecutive patients operated on at our institution between 1997 and 1999. It included patients who were either current smokers (n = 16) or lifelong never smokers(n = 8). As described above, an examination of the free resection margin was conducted.

Other sets of patients recruited at our institution were tested only for their constitutional status at the RAR-α and RAR-γ markers as described in the DNA section. Set 1 consisted of 50 current smokers who were enrolled successively in 1999 for an ongoing smoking cessation program. They were paired in sex,age (± 2 years), and tobacco consumption (±2 pack/year) with the smoker population described above. Set 2 included the 40 patients with squamous cell lung carcinoma who were described previously (2, 18). Set 3 included 62 cases of pleural mesothelioma diagnosed between 1988 and 1998 at our institution that have been confirmed by the French Mesopath Panel. The patient and control populations are described in Table 1.

All graded resection free margins were serially cut (4-μm sections)and laid on slides for tissue microdissection and immunohistochemistry testing.

DNA Purification.

Five to ten consecutive sections of each lung cancer precursor lesion(n = 66), tumor (n = 4), and free resection margins from never smokers (n= 8) and smokers (n = 16) were used to separately collect all types of bronchial cells and paired normal control cells whenever necessary (n = 44)under the basal membrane by microdissection as described previously(8). Normal healthy tissue and paired tumor(macrodissected) obtained during surgery and kept frozen at −80°C were also used for DNA preparation (n = 31)with tissue proteinase K digestion for 24–72 h, phenol/chloroform extraction, and further ethanol precipitations.

For the set 1 subjects, blood DNA was prepared with the Nucleon BACC3 kit from Pharmacia (Orsay, France). For sets 2 and 3, frozen normal lung tissue from patients with squamous cell lung carcinoma or microdissected (mesothelioma) normal tissue provided constitutional DNA.

DNA Amplification.

All of the reagents and the apparatus were from Pharmacia unless otherwise specified. All sense primers were labeled in 5′ with CY5. The microsatellites were CA repeats chosen on the Genethon site,4where the sequences of the flanking primers were also found. They were as follows: (a) at 17q12, D1751804(RAR-α); (b) at 3p24.2, DS1283(RAR-β); (c) at 12q13.13, D12S368(RAR-γ); (d) at 9q34.3, D9S158(RXR-α); (e) at 6p21, D6273(RXR-β); (f) at 1q23, D1S2635(RXR-γ); (g) at 3p14.2, D3S1300;(h) at 3p21, D3S1582; and (i) at 9p21, D9S171 (P16). Microsatellite amplifications were performed in a minithermocycler from MJ Research (Watertown, MA). Duplex PCRs were accomplished in a 10-μl final volume including 0.20μl of Taq polymerase, 0.8 μl of PCR mix, 0.8 μl of the four deoxynucleotide triphosphate mixture (2 mm), 0.8 μl of each primer pair (10 pm), and 1 μl of DNA (10–50 ng). The PCR cycles were as follows: (a) a 10-min hot start at 95°C;(b) 38 cycles of 94°C for 35 s, 65°C for 1 min, and 70°C for 1 min; and (c) a final extension step at 70°C for 10 min. At the end of the PCR, 4 μl of loading dye (Dextran 2000;5 mg/ml in deionized formamide) were mixed with each reaction, and 9μl of the mixture were heat-denatured and further electrophoresed on 7 m urea-acrylamide (6%) gels using an Alf express sequencing machine fitted with short plates. CY5-labeled molecular weight markers (100, 150, and 200 bp) were deposited on each side of each gel. Duplex PCRs were always performed using the following primer sets together: (a) 3p14/3p21;(b) RAR-α/RXR-γ; (c) RAR-γ/P16; and/or (d) RAR-β/RAR-γ. PCRs for only one microsatellite at a time were also performed with the same protocol. The results were analyzed online with the Allele Links software. LOH was defined as a complete disappearance of either microsatellite alleles in the bronchial cells or in the tumors when compared with the heterozygous paired normal sample. All patients were first screened for their constitutional status for a given microsatellite before proceeding to allelotyping of the sample. RER was present when, in a constitutional heterozygous sample, the size of the paired lung cancer precursor lesion or tumor alleles shifted. Whenever LOH or RER was observed, the experiment was repeated for confirmation. Representative data are shown on Fig. 1.

Immunohistochemistry.

p53 antibodies (DO7; Novacastra, Newcastle-upon-Tyne, United Kingdom;1:200) were used on lung cancer precursor lesions as described previously (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18). Using an optic grid fitted onto the ocular (×480), the results were quantified by counting all of the premalignant cells and the number of those cells that were positively stained for p53. Positive specimens contained 30% stained cells. The RAR-β antibodies [Rpβ(F)] were generous gifts from Dr. C. Rochette-Egly and P. Chambon (Institut de Génétique et Biologie Moleculaire de Strasbourg,Strasbourg, France). They were used only on the larger lung cancer precursor lesions (12a, 12e, 14, 24, 25, and 35, as described in Ref.18).

Microsatellite Changes.

The allelotyping results were broken down into five sample“progression” groups: (a) never smokers; (b)smokers; (c) squamous metaplasia; (d)dysplasia + ISC; and (e) tumors. Because DNA was not available for all experiments, the number of observed abnormalities or microsatellite changes in each group and for a given microsatellite is expressed as a percentage and as a ratio: LOH + RER number in the samples:number of heterozygous paired samples. All of the data are shown in Table 2, in addition to the immunohistochemistry results of our previous studies (2, 18).

Microsatellite Informativity.

Because one patient can have several precursor lesions, the overall microsatellite informativity status in each group was expressed as a percentage and as a ratio of the total number of heterozygous patients:total number of homozygous + heterozygous patients in the group (Table 3).

Statistical Analysis.

All of the computations were performed using BMDP software. All results were statistically analyzed by two-sided tests. Significance of difference for contingency tables (2 × 2) was assessed by Fisher’s exact test; as for contingency tables (nxm) by Pearson’s χ2 test. ANOVA was realized with the Kruskal-Wallis test, and mean comparisons were performed using the Mann-Whitney test. Microsatellite changes from never smokers to tumors and from never smokers to dysplasia + ISC were also tested with a χ2 test for trended linear progression. The association of any microsatellite change with smoking, age, p53 immunohistochemistry results, and differences in the repartition of constitutional heterozygous samples was considered for each microsatellite and between groups.

Tobacco exposure and age were quite similar in all groups(40 ± 20 pack/years with P < 0.74 and 62 ± 10 years and P < 0.08). As seen in Fig. 1, microdissection was very efficient to prepare separate specific DNAs with consequently complete rather than near complete LOH in the specimens. The amplicon peak was often larger for the remaining allele. Although tumors were carefully macrodissected(70% of tumor cells), it is possible that LOHs have been hidden by the stromal cells. RER by allele expansion (Fig. 1, panel 6) was observed once at the RAR-α marker in sample 23P, three times at the RXR-γ marker in samples 4T, 39T, and 18(g)P, once at 3p14 in 57 but 7 times at the RXR-α marker in samples 7T, 10T, 20P, 24P, 30T,37P, and 61 with DNA prepared from either fixed or frozen tissue(underlined sample numbers). There were neither global(P < 0.63) nor individual(P < 0.08–0.52) microsatellite change differences between the five sample groups (Table 3). The overall number of microsatellite changes was correlated to the level of tobacco consumption (P < 0.04) but not to the age of the patients (P < 0.82). Microsatellite changes at the RAR-α marker were rare in most specimens, but the microsatellite informativity was low in precursor lesions (22%) and tumors (23%). LOHs at the RAR-β marker were present in all types of precursor lesions (35%) and tumors (60%) and were even seen in smokers (22%). There was an increasing number of LOHs at this marker from group 1 to group 5 (P < 0.01), with no significant difference in the repartition of heterozygous samples in all groups(P = 0.3). LOH at the RAR-γmarker was common, but the microsatellite informativity was low in premalignant specimens (28%) and tumors (29%). LOH at this locus was strongly correlated with tobacco consumption expressed in pack/years(P < 0.004). LOH or, more often, RER for D9S158 (RXR-α) was rarely observed in precursor specimens(11%) and in smokers (29%) but was more frequently seen in tumors(35%). However, the repartition of heterozygous samples in the five groups was nonhomogenous (P = 0.2). LOH at the RXR-β marker was frequent, but with a low overall microsatellite informativity in all groups. There was a significant increase in the microsatellite changes for the RXR-γmarker from groups 1 through 4 (P < 0.03). Microsatellite changes for D3S1300 (3p14) were frequent in all specimens, but with a linear decrease in the number of heterozygous samples from groups 1 to 5. LOH at the 3p21locus was also common, with a higher but not significant prevalence seen in the dysplasia + ISC group. LOH at the P16microsatellite was present in 40% of smokers, 44% of precursor lesions, and 71% of tumors. Six patients had several premalignant lesions for which the allelic losses were not similar. Moreover,different microsatellite alleles were also lost in the different specimens (patients 1, 6, and 12).

Lung cancer precursor lesions analyzed in this study (Table 2) were paired with nine tumors that were screened previously by semiquantitative immunohistochemistry for the expression of RAR-α, RAR-β, RXR-α, and RXR-β and by quantitative reverse transcription-PCR for RAR-γ expression in specimens 27T and 29T. The allelotyping data are fully consistent with the immunohistochemistry results: a decrease in RAR-β is accompanied by LOH in heterozygous samples; whereas increased or normal RAR-α and RXR-α expression did not coincide with any microsatellite change. With regard to the RXR-β and RAR-γtesting, the data are inconsistent. In premalignant specimens 12a, 14,and 35, LOH at the RAR-β marker and decreased RAR-βimmunostaining were concomitant (Fig. 2) whereas no decrease in RAR-β expression and no change for D3S1283 were seen in premalignant samples 12e, 23, and 24. P53-positive staining was observed for only 35% of dysplasias but in all ISCs and in all paired tumors (except for one case) and therefore strongly correlated with the higher grade of lung cancer precursor lesion (P < 0.001). RAR-β LOH was also correlated with P53-positive staining (P < 0.02).

For D17S1804 and D12S368, an unexpectedly high rate of constitutional homozygosity was found in premalignant and tumor samples. Based on these results, we proceeded with genotyping for the RAR-αand RAR-γ markers: 40 more subjects (test 1 population), for a total of 66 squamous cell lung carcinoma cases; 48 more smokers (test 2 population) without lung cancer, for a total of 64 subjects and for comparison: 62 mesothelioma (test 3 population). Patients with squamous cell lung carcinoma were less heterozygous than expected for both the RAR-α (36%) and RAR-γ(29%) markers. Smokers were more heterozygous for the RAR-α microsatellite (58%) but were equally as homozygous for the RAR-γ microsatellite (27%). The low rate of heterozygosity at both loci persisted when the 18 additional nonsquamous cell lung tumors were included (36% and 29%,respectively). By contrast, in mesothelioma, 70% and 44% of subjects,respectively, were found to be heterozygous.

We observed combined RAR/RXR expression changes in lung tumors(18) and undertook a new study to further define the status of RAR/RXR in lung precursor lesions. Within the technical restrictions detailed above, allelotyping for microsatellites located near the RAR/RXR gene loci was performed. The microsatellites chosen were the closest possible microsatellites to the concerned genes, except for the more distant RAR-γ marker,which is, in fact, close to the RAS gene, an oncogene known to play a role in lung cancer, and to the fragile site FRA 12A. Interestingly, losses of this marker were correlated with tobacco exposure. However genetic losses were not consistent with our previous results in tumors 27 and 29. D6S273 has just been moved on the Genethon map from 6p11 to 6p22, near to the keratin gene cluster. High levels of homozygosity were found in all studied samples, and hemiallelic losses did not match the previous immunohistochemistry results. However, Virmani et al.(19) reported that 6p21 is a region of frequent allelic losses in non-small cell lung cancer. Unfortunately, there were no reliable RXR-γ antibodies to match the RXR-γ marker allelotyping results. As shown in Table 2, the allelotyping data of the other microsatellites were mostly consistent with the previous tumor immunohistochemistry screening. For the RAR-α marker,there was neither tumoral LOH nor decreased gene expression, but there were too many homozygous samples. Immunohistochemistry performed on five lung cancer precursor lesions with or without LOH for D3S1283 demonstrated decreased expression of the RAR-β gene when there was also a LOH. A similar concordance was found previously in 75–86% of 76 lung tumors (18). Allelic losses were already present in smokers, but not in lifelong never smokers. However,from so few cases, it is difficult to conclude any tobacco carcinogen-specific action. Decreased RAR-β quantity would impair heterodimerization with RXR partners. In combination with other RAR/RXR disruptions (promoter methylation), this could result in a functional cellular retinoid deficiency. Similar events have been described recently for RAR-γ and RXR-α in the skin as a consequence of UV irradiation(20). Lung tumors were found to overexpress the RXR-αprotein in combination with RAR-β underexpression (18),possibly as an adaptation to retinoid deficiency. Normal expression or overexpression of RXR-α has also been observed in breast tumors and mouse skin tumors (21). Conflicting results placing microsatellite instability in lung tumors between 0% and 60% have been published previously (22, 23). We recently participated in extensive lung tumor allelotyping (18, 24), and although different methodologies were used, the results were consistent: RERs were rare. For the present study, fixed or unfixed tissues were used, and, as found previously, RERs were rare. The discrepancies might depend on: (a) the type of polymorphic marker used (CA repeats versus trinucleotides or tetranucleotides, for which the spontaneous rate of mutation can be very high); (b) the criteria used to define RER as a shifting of one or both alleles; and (c) microsatellite instability as RER at one or several markers. Low polymorphism of microsatellites at certain loci may be explained by the rather conservative structure of the genes lying at such loci. In squamous cell lung carcinoma, the frequency of homozygosity at D17S1804(RAR-α, 17q12) was unusual and was not found in smokers or in patients with mesothelioma. However, higher rates of homozygosity at D12S368 (RAR-γ, 12q13.13) were found in both squamous cell lung carcinoma and smokers by comparison with mesothelioma cases. The clinical and biological significance of such homozygosities needs to be investigated carefully, even in comparable ethnic populations, to eliminate a founder effect. Indeed,homozygosities could favor altered DNA recombinations.

In general, it is thought that molecular damage incidence increases as histopathological lung cancer precursor lesions progress from hyperplasia to ISC. Identifying the genes targeted at each step would reconstitute the natural history of squamous cell lung carcinoma. However, comparisons between investigations are difficult because of differences in methodology and in criteria for studying and scoring the molecular disruptions. The systematic screening of lung tumors for allelic losses led to the identification of multiple distinct regions of recurrent deletions at 3p, 17p, 9q, 5q, 13q, 8p, and 11p(approximately ranked by decreasing frequency), suggesting that these regions contain unidentified genes involved in lung cancer(24). Based on these findings, allelic losses have been reported in lung cancer precursor lesions mainly on 3p and 9p regions (4, 9, 10, 11, 25). Four hot spots have been identified in lung tumors at 3p12, 3p14, 3p21, and 3p25, and a candidate tumor supressor gene at 3p14, fragile histidine triad (FHIT), has been cloned (26). Whereas LOH at 3p14 has been associated with the presence of aberrant transcripts that involve partial deletions of this gene, other studies have suggested that LOH can occur without abnormalities and is influenced by the proximity of the FRA3B region on which tobacco exposure may be causal (27). This might also be the case for the RAR-γ marker used here. Transfer of DNA fragments from 3p21.3 into tumor cell lines suggested that the region has tumor supressor gene activity (28). The protein tyrosine phosphatase gene and a mitogen-activated protein kinase are potential candidates located on 3p21, whereas others have been eliminated (25). P16/CDKN2 is located on 9p21; a high percentage of alterations of this gene has been observed in many tumors types, but the frequency of LOH found in lung tumors is higher than the frequency of mutations, suggesting that other tumor supressor genes reside on 9p(25). The size of the lung cancer precursor lesion prevents fine mapping, but the 3p21 region remains a good candidate for tumor supressor gene localization.

Allelotyping data have been obtained in current smokers, former smokers, and nonsmokers for normal and abnormal bronchial epithelium. Wistuba et al.(6) studied several microsatellites at the 3p14–21-24 loci and the retinoblastoma, P53, and P16 (D9S171) regions. No molecular changes were found in nonsmokers, but interpretation of the findings was potentially limited by the different age distribution of the nonsmokers, who were significantly younger than the smokers. Among the smokers, there was a modest correlation between the number of molecular changes/subject and smoking exposure, but the variation between current and former smokers was not significant. LOH occurred mainly at 3p (38%) and P16 (23%). At 3p21, LOH was detected in histologically normal epithelium,whereas 3p14 losses were detected only in dysplasia. RER was detected in smokers (64%), even in normal histological specimens. In addition, there was a loss of the same allele of a polymorphic marker in the same patient. By contrast, Mao et al.(5) found few 3p14 LOHs in nonsmokers (20%)but found more in current smokers (85%) than in former smokers (45%);premalignant lesions were also more frequent in current smokers. P16 losses were found in 23% of the specimens. P53 losses at 17p13 reached 18%, but no RER was reported. The alleles lost in the different biopsies for the studied polymorphic markers were different in the same patient. It is difficult to reach a consensus from these studies, but our results concerning the smokers are more in accordance with the latter work.

Several studies have repeatedly shown p53 increased immunoreactivity usually reflective of P53 mutations in premalignant lung lesions (2, 29, 30). Other abnormalities, including an increased cellular proliferation rate (2) and c-myc(2) and bcl-2 overexpression with changes in the keratin pattern toward squamous epithelia, have also been described previously(31, 32).

Smoking prevention is the first tool against lung cancer prevention. For former smokers, early detection is necessary because lung cancer prognosis remains poor, despite several therapeutic improvements. Many efforts have been directed toward the identification of biomarkers for early detection of lung cancer (33) and chemoprevention(14, 25). This study and our previous study(18) suggest that retinoid deficiency may be among the first events contributing to lung tumorigenesis and imply that retinoids could be used in lung cancer chemoprevention. Aerosolized early on site, they could reverse the deficiency in stabilizing RAR/RXR expression for increased ligand binding to restore normal cellular differentiation. Early lung cancer detection could associate yeast functional assays of RAR-β and P53 in shed bronchial cells and complementary testing of microsatellites, such as D3S1283, whose loss is correlated with decreased RAR-β expression.

Fig. 1.

Allelotyping results including specific gel aspects. MW, molecular weight standards (100, 150, and 200 bases). Constitutive DNA (1) and DNA purified from the normal bronchial epithelium (2) of patient 44; duplex PCR for the P16 and RAR-γ markers with LOH for RAR-γ. Constitutive DNA (3) and DNA from the premalignant lesion(4) of patient 8; duplex PCR for the RAR-β and RAR-γmarkers with concomitant LOH for both markers. Constitutive DNA(5) and tumor DNA (6) from patient 4 with a RER for the RAR-α marker.

Fig. 1.

Allelotyping results including specific gel aspects. MW, molecular weight standards (100, 150, and 200 bases). Constitutive DNA (1) and DNA purified from the normal bronchial epithelium (2) of patient 44; duplex PCR for the P16 and RAR-γ markers with LOH for RAR-γ. Constitutive DNA (3) and DNA from the premalignant lesion(4) of patient 8; duplex PCR for the RAR-β and RAR-γmarkers with concomitant LOH for both markers. Constitutive DNA(5) and tumor DNA (6) from patient 4 with a RER for the RAR-α marker.

Close modal
Fig. 2.

Immunohistochemical investigation of RAR-β protein in normal bronchial epithelium and preinvasive lesions. Normal bronchial epithelium (a) and squamous metaplasia lesion(c) show a normal and similar expression of RAR-βprotein in nuclei from both epithelial cells and submucosal cells;corresponding controls, b and d. Squamous metaplasia (e), squamous metaplasia with dysplasia(f), and ISC (g) lesions show a strong decrease in expression of RAR-β, as judged in comparison to the degree of expression in submucosal cells.

Fig. 2.

Immunohistochemical investigation of RAR-β protein in normal bronchial epithelium and preinvasive lesions. Normal bronchial epithelium (a) and squamous metaplasia lesion(c) show a normal and similar expression of RAR-βprotein in nuclei from both epithelial cells and submucosal cells;corresponding controls, b and d. Squamous metaplasia (e), squamous metaplasia with dysplasia(f), and ISC (g) lesions show a strong decrease in expression of RAR-β, as judged in comparison to the degree of expression in submucosal cells.

Close modal

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Supported by Action Integrée Franco/Belge INSERM; Ligue contre le Cancer: Meuse, Moselle, and Vosges; and Comité National Contre les Maladies Respiratoires et la Tuberculose.

The abbreviations used are: ISC, in situ carcinoma; RAR, retinoic acid receptor; RXR, retinoid X receptor; LOH, loss of heterozygosity; RER, replication error phenotype.

ftp://ftp.genethon.fr/pub/Gmap.

Table 1

Patient population

Tumor histologyNo. of patientsNo. of squamous metaplasiasNo. of dysplasiasNo. of ISCsTotal no. of lung precursor lesions
SQCLCa 26 20 16 42 
ADC 12 10 19 
SCLC 
      
Total 41 32 26 66 
      
Controlsb 24c 
Tumor histologyNo. of patientsNo. of squamous metaplasiasNo. of dysplasiasNo. of ISCsTotal no. of lung precursor lesions
SQCLCa 26 20 16 42 
ADC 12 10 19 
SCLC 
      
Total 41 32 26 66 
      
Controlsb 24c 

SQCLC, squamous cell lung cancer; ADC, adenocarcinoma;SCLC, small cell lung carcinoma.

No cancer, 16 smokers and 8 never smokers.

This number includes 21 histologically normal epithelia.

Table 2

Study results; panel A for squamous metaplasia; panel B for dysplasia + ISC with paired lung tumors (T); panel C for smokers and never smokers (bronchial epithelium)a

Study results; panel A for squamous metaplasia; panel B for dysplasia + ISC with paired lung tumors (T); panel C for smokers and never smokers (bronchial epithelium)a
Study results; panel A for squamous metaplasia; panel B for dysplasia + ISC with paired lung tumors (T); panel C for smokers and never smokers (bronchial epithelium)a
Study results; panel A for squamous metaplasia; panel B for dysplasia + ISC with paired lung tumors (T); panel C for smokers and never smokers (bronchial epithelium)a
Study results; panel A for squamous metaplasia; panel B for dysplasia + ISC with paired lung tumors (T); panel C for smokers and never smokers (bronchial epithelium)a
 
 

MS, microsatellite name; N, patient number (fixed or frozen tissue (bold); P, precursor lesion; T, tumor; |B*, homozygous;○, heterozygous; •, LOH; ⧫, RER; nd, not done; ↑, RAR/RXR expression increased; ↓, RAR/RXR expression decreased; =, RAR/RXR expression normal.

Table 3

Allelotyping results

Microsatellite names and lociD17S1804 RAR-α 17q12D3S1283 RAR-β 3p24.2D12S368 RAR-γ 12q13.13D9S158 RXR-α 9q34.3D6S273 RXR-β 6p21-3D1S2635 RXR-γ 1q23D3S1300 3p14.2 3p14D3S1582 3p21 3p21D9S171 p16 9p21
Heterozygous vs. homozygousa          
NSb  6 vs. 2   2 vs. 6   0 vs. 8   4 vs. 4   0 vs. 8   3 vs. 5   5 vs. 3   3 vs. 5   2 vs. 6  
10 vs. 6   9 vs. 7  10 vs. 6   7 vs. 9   7 vs. 9  12 vs. 4  12 vs. 4   7 vs. 9   5 vs. 11 
12 vs. 48 34 vs. 29 17 vs. 44 35 vs. 28 13 vs. 50 36 vs. 18 27 vs. 36 42 vs. 21 24 vs. 38 
 7 vs. 20 20 vs. 11  9 vs. 22 22 vs. 7   8 vs. 24 19 vs. 12  8 vs. 23 14 vs. 17 14 vs. 17 
Specimen microsatellite changes (%)c          
NS 16 50 0 40 100 
20 22 70 29 14 33 58 29 40 
31 35 88 11 45 57 30 40 44 
60 56 35 56 44 38 43 71 
Microsatellite informativity (%)d          
NS 75 33 50 37 62 37 25 
63 56 63 44 44 75 75 44 31 
20 54 28 56 20 67 43 67 63 
25 65 29 75 25 61 25 45 45 
Expected informatives (%) 82 68 86 70 85 86 72 82 72 
Microsatellite names and lociD17S1804 RAR-α 17q12D3S1283 RAR-β 3p24.2D12S368 RAR-γ 12q13.13D9S158 RXR-α 9q34.3D6S273 RXR-β 6p21-3D1S2635 RXR-γ 1q23D3S1300 3p14.2 3p14D3S1582 3p21 3p21D9S171 p16 9p21
Heterozygous vs. homozygousa          
NSb  6 vs. 2   2 vs. 6   0 vs. 8   4 vs. 4   0 vs. 8   3 vs. 5   5 vs. 3   3 vs. 5   2 vs. 6  
10 vs. 6   9 vs. 7  10 vs. 6   7 vs. 9   7 vs. 9  12 vs. 4  12 vs. 4   7 vs. 9   5 vs. 11 
12 vs. 48 34 vs. 29 17 vs. 44 35 vs. 28 13 vs. 50 36 vs. 18 27 vs. 36 42 vs. 21 24 vs. 38 
 7 vs. 20 20 vs. 11  9 vs. 22 22 vs. 7   8 vs. 24 19 vs. 12  8 vs. 23 14 vs. 17 14 vs. 17 
Specimen microsatellite changes (%)c          
NS 16 50 0 40 100 
20 22 70 29 14 33 58 29 40 
31 35 88 11 45 57 30 40 44 
60 56 35 56 44 38 43 71 
Microsatellite informativity (%)d          
NS 75 33 50 37 62 37 25 
63 56 63 44 44 75 75 44 31 
20 54 28 56 20 67 43 67 63 
25 65 29 75 25 61 25 45 45 
Expected informatives (%) 82 68 86 70 85 86 72 82 72 

In number of tested specimens.

NS, nonsmokers; S, smokers; P, precursor lesions; T, tumors.

LOH + RER in the specimens/number of heterozygous paired specimens.

Number of heterozygous subjects/total number of subjects. Results are underlined when they concern at least 50% of heterozygous subject.

1
Sundaresan V., Ganly P., Hasleton P., Rudd R., Sinha G., Bleehen N. M. P53 and chromosome 3 abnormalities, characteristic of malignant lung tumors, are detectable in preinvasive lesions of the bronchus.
Oncogene
,
7
:
1989
-1997,  
1992
.
2
Klein N., Vignaud J. M., Sadmi M., Plenat F., Borelly J., Duprez A., Martinet Y., Martinet N. Squamous metaplasia expression of proto-oncogenes and P53 in lung cancer patients.
Lab. Invest.
,
68
:
26
-32,  
1993
.
3
Hammond W. G., Teplitz R. L., Benfield J. R. Variable regression of experimental bronchial preneoplasia during carcinogenesis.
J. Thorac. Cardiovasc. Surg.
,
101
:
800
-806,  
1991
.
4
Martinet Y., Brambilla E., Martin J. P., Martinet N., Vignaud J. M. Les bases biologiques de la prévention du cancer du poumon.
Médecine Sciences
,
13
:
1465
-1471,  
1997
.
5
Mao L., Lee J. S., Kurie J. M., Fan Y. H., Lippman S. M., Lee J. J., Ro J. Y., Broxson A., Yu R., Morice R. C., Kemp B. L., Khuri F. R., Walsh G. L., Hittelmann W. N., Hong W. F. Clonal genetic alterations in the lungs of current and former smokers.
J. Natl. Cancer Inst.
,
89
:
857
-862,  
1997
.
6
Wistuba I. I., Lam S., Behrens C., Virmani A. K., Fong K. M., LeRiche J., Samet J. M., Srivastava S., Minna J. D., Gazdar A. F. Molecular damage in the bronchial epithelium of current and former smokers.
J. Natl. Cancer Inst.
,
89
:
1366
-1373,  
1997
.
7
Auerbach O., Hammond E. C., Garfinkel L. Changes in bronchial epithelium in relation to smoking and cancer of the lung.
N. Engl. J. Med.
,
265
:
253
-267,  
1961
.
8
Going J. J., Laub R. F. Practical histological microdissection for PCR analysis.
J. Pathol.
,
179
:
121
-124,  
1979
.
9
Kishimoto Y., Sugio K., Hung J. Y., Virmani A. K., McIntre D. D., Minna J. D., Gazdar F. Allele-specific loss in chromosome 9p loci in preneoplastic lesions accompanying non-small-cell lung cancers.
J. Natl. Cancer Inst.
,
87
:
1224
-1291,  
1995
.
10
Gazdar A. F., Bader S., Hung J., Kishimoto Y., Sekido Y., Sugio K. Molecular genetic changes found in human lung cancer and its precursor lesions.
Cold Spring Harbor Symp. Quant. Biol.
,
59
:
565
-572,  
1994
.
11
Thiberville L., Payne P., Vielkinds J., Leriche J., Horsman D., Nouvet G., Palcic B., Lam S. Evidence of cumulative gene losses with progression of premalignant epithelial lesions to carcinoma of the bronchus.
Cancer Res.
,
55
:
5133
-5139,  
1995
.
12
Jetten A. M. Retinoids, growth factors and the tracheobronchial epithelium.
Lab. Invest.
,
59
:
1
-4,  
1988
.
13
Chambon P. A decade of molecular biology of retinoic acid receptors.
FASEB J.
,
10
:
940
-954,  
1996
.
14
Lee J. S., Lippman S. M., Benner S. E., Lee J. J., Ro J. Y., Lukeman J. M., Morice R. C., Peters E. J., Pang A. C., Fritsche H. A., Jr. Randomized placebo controlled trial of isotretinoin in chemoprevention of bronchial squamous metaplasia.
J. Clin. Oncol.
,
12
:
937
-945,  
1994
.
15
The α-Tocopherol, β-Carotene Cancer Prevention Study Group. The effect of vitamin E and β carotene on the incidence of lung cancer and other cancers in male smokers.
N. Engl. J. Med.
,
330
:
1029
-1035,  
1996
.
16
Houle B., Leduc F., Bradlevy W. E. Implication of RARβ in epidermoid squamous lung cancer.
Genes Chromosomes Cancer
,
3
:
358
-366,  
1991
.
17
Xc X. U., Sozzi G., Lee J. S., Lee J. J., Pastorino U., Pilotti S., Kurie J. M., Hong W. K., Lotan R. Suppression of RARβ in NSCLC in vivo: implications for lung cancer development.
J. Natl. Cancer Inst.
,
89
:
624
-629,  
1997
.
18
Picard E., Seguin C., Monhoven N., Rochette-Egly C., Siat J., Borelly J., Martinet Y., Martinet N., Vignaud J. M. Expression of retinoid receptors genes and proteins in non-small-cell lung cancer.
J. Natl. Cancer Inst.
,
91
:
59
-66,  
1999
.
19
Virmani A. K., Fong K. M., Kodagoda D., McIntire D., Hung J., Tonk V., Minna J. D., Gazdar A. F. Allelotyping demonstrates common and distinct patterns of chromosomal loss in human lung cancer types.
Genes Chromosomes Cancer
,
21
:
308
-319,  
1998
.
20
Lawrence J. A., Merino M. J., Simpson J. F., Manrow R. E., Page D. L., Steeg P. S. A high-risk lesion for invasive breast cancer, ductal carcinoma in situ, exhibits frequent overexpression of retinoid X receptor.
Cancer Epidemiol. Biomark. Prev.
,
7
:
29
-35,  
1998
.
21
Wang Z., Boudjelal M., Kang S., Voorhees J. J., Fisher G. J. Ultraviolet irradiation of human skin causes functional vitamin A deficiency, preventable by all-trans retinoic acid pre-treatment.
Nat. Med.
,
5
:
418
-422,  
1999
.
22
Miozzo M., Sozzi G., Musso K., Pilotti S., Incarbone M., Pastorino U., Pierotti M. A. Microsatellite alterations in bronchial and sputum specimens of lung cancer patients.
Cancer Res.
,
56
:
2285
-2288,  
1996
.
23
Pylkkänen L., Karjalainen A., Antilla S., Vainio H., Husgafvel-Pursiainen K. No evidence of microsatellite instability but frequent loss of heterozygosity in primary resected lung cancer.
Environ. Mol. Mutagen.
,
30
:
217
-223,  
1997
.
24
Lerebours F., Olschwang S., Thuille B., Schmitz A., Fouchet P., Buecher B., Martinet N., Galateau F., Thomas G. Fine deletion mapping of chromosome 8p in non-small-cell lung carcinoma.
Int. J. Cancer
,
81
:
854
-858,  
1999
.
25
Wiest J. S., Franklin W. A., Drabkin H., Gemmill R., Sidransky D., Anderson M. W. Genetic markers for early detection of lung cancer and outcome measures for response to chemoprevention.
J. Cell Biochem.
,
28/29
:
64
-73,  
1997
.
26
Fong K. M., Biesterveld E. J., Virmani A., Wistuba I., Sekido Y., Bader S. A., Ahmadian M., Ong S. T., Rassool F. V., Zimmerman P. V., Giaccone G., Gazdar A. F., Minna J. D. FHIT and FRA3B 3p14.2 allele losses are common in lung cancer and preneoplastic bronchial lesions and are associated with cancer-related FHIT cDNA splicing aberrations.
Cancer Res.
,
57
:
2256
-2267,  
1997
.
27
Todd S., Franklin W. A., Varella-Garcia M., Kennedy T., Hilliker C. E., Jr., Hahner L., Anderson M., Wiest J. S., Drabkin H. A., Gemmill R. M. Homozygous deletions of human chromosome 3p in lung tumors.
Cancer Res.
,
57
:
1344
-1352,  
1997
.
28
Killary A. M., Wolf M. E., Giambermardi T. A., Naylor S. L. Definition of a tumor suppressor locus within human chromosome 3p21–p22.
Proc. Natl. Acad. Sci. USA
,
89
:
1721
-1729,  
1992
.
29
Bennett W. P., Colby T. V., Travis W. D., Borkowski A., Jones R. T., Lane D. P., Metcalf R. A., Samet J. M., Takeshima Y., Gu J. R., Vähäkangas K. H., Soini Y., Pääkkš P., Welsh J. A., Trump B. F., Harris C. C. P53 protein accumulates frequently in early bronchial neoplasia.
Cancer Res.
,
53
:
4817
-4822,  
1993
.
30
Sozzi G., Miozzo M., Donghi R., Pilotti S., Cariani C. T., Pastorino U., Della Porta G., Pierotti M. A. Deletions of 17p and p53 mutations in preneoplastic lesions of the lung.
Cancer Res.
,
52
:
6079
-6082,  
1992
.
31
Colby V. Precursor lesions to pulmonary neoplasia Brambilla C. Brambilla E. eds. .
Lung Tumor
,
:
61
-87, M. Dekker New York  
1999
.
32
Fisseler-Eckhoff A., Erfkamp S., Muller K. M. Cytokeratin expression in preneoplastic lesions and early squamous cell carcinoma of the bronchi.
Pathol. Res. Pract.
,
192
:
552
-559,  
1996
.
33
Zhou J., Mulshine J. L., Unsworth E. J., Scott F. M., Avis I. M., Vos M. D., Treston A. M. Purification and characterization of a protein that permits early detection of lung cancer.
J. Biol. Chem.
,
271
:
10760
-10766,  
1996
.