Fax: (713) 792-0807
BPDE-induced lymphocytic 3p21.3 aberrations may predict head and neck carcinoma risk
Version of Record online: 23 JUL 2002
Copyright © 2002 American Cancer Society
Volume 95, Issue 3, pages 563–568, 1 August 2002
How to Cite
Zhu, Y., Spitz, M. R., Zheng, Y.-L., Hong, W. K. and Wu, X. (2002), BPDE-induced lymphocytic 3p21.3 aberrations may predict head and neck carcinoma risk. Cancer, 95: 563–568. doi: 10.1002/cncr.10689
- Issue online: 23 JUL 2002
- Version of Record online: 23 JUL 2002
- Manuscript Accepted: 11 MAR 2002
- Manuscript Revised: 25 FEB 2002
- Manuscript Received: 26 NOV 2001
- NCI. Grant Numbers: PO1 CA 52051, CA 86390, RO1 74880
- chromosomal aberration;
Tobacco exposure is an established risk factor for head and neck squamous cell carcinoma (HNSCC). Benzo[α]pyrene diol expoxide (BPDE), a main metabolic product of the tobacco smoke constituent benzo[α]pyrene, induces chromosomal aberrations at specific loci. Chromosomal aberrations in peripheral blood lymphocytes (PBLs) induced by BPDE may reflect individuals' genetic susceptibility to tobacco carcinogens.
This study was designed to detect BPDE-induced aberrations in PBLs at locus 3p21.3 in cultured lymphocytic cells. Our hypothesis is that the presence of BPDE-induced 3p21.3 aberrations is a biomarker of an individual's genetic susceptibility and that individuals with these aberrations are at an increased risk for HNSCC. PBL cultures from 52 cases and 54 controls were treated with 2 μM BPDE for 24 hours before the 3p21.3 aberrations were assessed by flourescence in situ hybridization. One thousand lymphocyte interphases were scored for each sample.
We found that BPDE-induced chromosome 3p21.3 aberrations occurred more frequently in cases (mean: 31.4 per 1000 cells) than in controls (mean: 22.1 per 1000 cells; P < 0.001). However, when 6q27 was selected as a control locus, no such difference was observed (P = 0.545). When the 75th percentile value of induced aberrations in the controls was used as a cutoff point to classify 3p21.3 BPDE-induced sensitivity, 30 of the 52 cases (57.69%) and only 14 of the 54 controls (25.93%) were sensitive to BPDE exposure. This approach resulted in an odds ratio of 4.8 (95% confidence interval: 1.87–12.28) for HNSCC risk associated with BPDE-induced 3p21.3 aberrations. There was also a dose-response relationship between the number of BPDE-induced aberrations at 3p21.3 and risk for HNSCC.
The results from this study demonstrated that 3p21.3 may be a specific molecular target of tobacco carcinogens and that BPDE sensitivity at this locus may reflect an individual's genetic susceptibility to HNSCC. Cancer 2002;95:563–8. © 2002 American Cancer Society.
Although genetic alterations underlying malignant initiation and progression of head and neck squamous cell carcinoma (HNSCC) are not fully understood, it is clear that HNSCC is an environmentally induced disease. More than 90% of patients have a history of previous tobacco and alcohol exposure. The interaction between carcinogen exposures and host susceptibility may be responsible for chromosomal aberrations conferring genetic instability that may consequently initiate occurrence of HNSCC. Therefore, identification of specific molecular targets associated with tobacco carcinogen exposure may provide insight into HNSCC development as well as potential inherited susceptibility.
The short arm of chromosome 3 (3p) has attracted much attention in cancer biology recently because aberrations on 3p are one of the most frequently identified genetic aberrations in head and neck carcinomas and in other tobacco-induced malignancies.1–3 In addition to cytogenetically visible deletions, excess loss of heterozygosity (LOH) at loci on 3p has been reported in HNSCC cell lines and tumor tissue samples.4, 5 Three discrete regions of deletion on 3p (3p24-ter, 3p21.3, and 3p14-cen) have been identified in head and neck carcinomas4 and support a common oncogenic pathway in lung carcinogenesis, for which a similar pattern of 3p deletion has been described.3 Allelic loss on chromosome 3p has been further identified as an early event1, 4, 6 and an unfavorable prognostic factor in patients with primary HNSCC.5, 7 Tumor suppressor activity has been detected in tumor cell lines with transferred subchromosomal fragments of 3p21.2-3p21.3.8, 9 These data suggest that genetic instability at 3p may play a role in the pathogenesis of HNSCC by altering one or more tumor suppressor genes (TSG).
The association between these alterations and HNSCC, a tobacco-related carcinoma, also implies that there might be some specific molecular targets of tobacco carcinogens on the short arm of chromosome 3. Several lines of evidence have suggested that chromosome fragility and instability are not distributed randomly but are located at specific sites in the human genome.10–12 The onset of head and neck tumors may be due to clonal expansion from a common clone with 3p alterations damaged by carcinogens. In support of this notion, recent mutagen sensitivity studies have shown that 3p is a site of nonrandom aberrations induced by bleomycin and 4-nitroquinoline N-oxide (4NQO) in peripheral blood lymphocytes (PBLs) from HNSCC patients.13 In addition, 3p losses in HNSCC tumor tissue samples were significantly associated with bleomycin sensitivity.14 In this molecular cytogenetic study, we extended previous findings by focusing on 3p21.3 aberrations induced by benzo[α]pyrene diol expoxide (BPDE), a major constituent of tobacco smoke. We hypothesized that BPDE-induced chromosome 3p aberrations reflect host susceptibility to carcinogens in tobacco smoke and individuals with a higher number of these aberrations are at an elevated risk for HNSCC. A locus-specific fluorescent in situ hybridization (FISH) probe for 3p21.3 was utilized to examine BPDE-induced chromosomal alterations in PBLs from both HNSCC patients and matched healthy controls.
MATERIALS AND METHODS
We identified 52 HNSCC cases from The University of Texas M. D. Anderson Cancer Center. All cases were newly diagnosed, previously untreated (chemotherapy or radiotherapy), and histologically confirmed. There were no age, gender, ethnicity, or stage restrictions. The 54 controls with no history of cancer with the exception of nonmelanoma skin cancer were identified from the rosters of the largest multispecialty managed maintenance organization in Houston, Texas. The majority of the controls visited the clinic for their annual checkups, not for treatment of chronic illnesses. The controls were matched to the cases by age (+ 5 years), ethnicity, and smoking status.
Epidemiologic Data Collection
Epidemiologic data were collected from questionnaires that comprehensively elicited recent and not so recent histories of cigarette smoking and alcohol consumption. Blood was drawn into sodium heparinized tubes for cytogenetic and molecular genetic analyses. The personnel who performed the laboratory analyses were blinded with regard to case–control status.
Lymphocyte Cultures and BPDE Treatment
Lymphocyte cultures were set up following the routine protocol of adding 1 mL whole blood to 9 mL of RPMI 1640 tissue culture medium (JRM Biosciences, Lenexa, KS) with 10% fetal calf serum and 1% phytohemagglutinin (Wellcome Research Laboratories, Research Triangle Park, NC) at 37 °C for 72 hours. (+/−)-anti-BPDE was purchased from Midwest Research Institute (Kansas City, MO). Aliquots of the 1-mM stock were placed in microcentrifuge tubes and stored at −20 °C in the dark. Each culture was treated with BPDE with a final concentration of 2 μM for 24 hours. After routine blocking with colcemid, hypotonic treatment, and fixation, cell suspensions were stored at −20 °C until they were processed for FISH analysis.
Chromosomal aberrations at 3p21.3 were detected by FISH with a specific DNA probe covering the human semaphorin IV gene (approximately 75 kb; Oncor, Gaithersberg, MD). Another probe for chromosome 6q27 (Oncor) was used as an internal control locus in the analysis because 6q27 has not been reported to be deleted in HNSCC and is located in a Giemsa-light region, as is 3p21.3. The 6q27 probe also served as a control for differences in chromatin structure (which may affect hybridization) between 3p21.3 and the rest of the genome.
FISH with the locus-specific 3p21.3 probe was performed according to the supplier's instructions. Briefly, 10 μl of the probe was added to each interphase smear slide. The slides had been pretreated with 2 × SSC (0.3 M sodium chloride 0.03 M sodium citrate, pH 7.0) at 37 °C for 30 minutes, dehydrated in 70%, 85%, and 100% ethanol for 1 minute each at room temperature, and denatured with 70% formamide and 2 × SSC (pH 7.0) at 73 °C for 5 minutes. Hybridizations were performed at 37 °C for at least 16 hours. Posthybridization, the slides were washed in 2 × SSC (pH 7.0) for 5 minutes at 72 °C without agitation. Signals were detected by incubating the slides for 10 minutes at 37 °C in fluorescein isothiocyante (FITC)-conjugated avidin (Oncor) and then washing them in 1 × BST buffer (0.5 M NaHCO3, 1.5 M NaCl, and 0.25% Tween 20 [pH 8.0]) three times for 3 minutes each time. If a signal was weak, it was amplified by incubation with antiavidin antibody for 10 minutes at 37 °C. Then the slides were washed in 1 × BST, incubated with FITC-antiavidin antibody for 10 minutes at 37 °C, and then washed again in 1 × BST buffer. Finally, the slides were counterstained with propodium iodide/antifade and the cells were viewed through a fluorescent microscope (Leeds, Irving, TX) with individual FITC filters.
The criteria for scoring FISH signals were the same as previously described.15 Only nonoverlapped nuclei with clearly separated FISH signals were scored. Minor hybridization spots, which are smaller and less intense than real signals, were excluded. Paired or close signals were counted as one signal. The laboratory personnel who did the scoring were blinded to the sample's case–control status.
All statistical analyses were performed using STATA statistical software (Stata Corp., College Station, TX). Pearson's chi-square was used to test the differences in the distribution between cases and controls. We examined the association between 3p BPDE sensitivity, assessed as the number of 3p21.3 aberrations in 1000 interphases, and risk of HNSCC in this case–control study. Odds ratios (ORs) with 95% confidence intervals (CI) were calculated as estimates of the relative risk. Our analytic approach was to dichotomize 3p BPDE sensitivity at the 75th percentile value for controls and to study it as a continuous variable. The ORs were calculated by logistic regression adjusted for multiple covariates.
The characteristics of the study subjects are summarized in Table 1. This case–control study comprised 52 HNSCC case subjects and 54 control subjects. The majority of our subjects were Caucasians (92.31% in cases and 79.63% in controls). There were significantly more males in the cases than in the control population. The mean age of the cases (63.8 years) and the controls (60.5 years) did not differ significantly. There were also no significant differences between the cases and the controls in terms of smoking status. The mean number of cigarettes per day was 24.71 for the cases and 23.44 for the controls, but this difference was not statistically significant.
|Variable||Cases (N = 52)||Controls (N = 54)||P value|
|White||48 (92.31)||43 (79.63)|
|Hispanic||2 (3.85)||6 (11.11)|
|Black||2 (3.85)||5 (9.26)||0.172|
|Gender||N (%)||N (%)|
|Male||41 (78.85)||28 (51.85)|
|Female||11 (21.15)||26 (48.15)||0.004|
|M (SD)||63.8 (11.15)||60.5 (9.64)||0.106|
|Never||3 (5.77)||3 (5.66)|
|Former||22 (42.31)||24 (45.28)|
|Current||27 (51.92)||26 (49.06)||0.953|
|No. cigarettes per dayb|
|M (SD)||24.71 (19.40)||23.44 (13.97)||0.494|
Two FISH signals represented two copies of the chromosome region in a normal cell. An aberration was defined as one or more than two signals in an interphase cell. The most common abnormal event was one FISH signal, representing a deletion of the locus-specific probes (Fig. 1). Overall, we observed more BPDE-induced chromosomal aberrations at locus 3p21.3 than at control locus 6q27 in both cases and controls. The mean number of BPDE-induced chromosome 3p21.3 aberrations was higher in patients with HNSCC (31.4 ± 8.87 per 1000 cells) than in healthy controls (22.1 ± 10.7 per 1000 cells; P < 0.001). The mean number of BPDE-induced chromosome 6q27 aberrations detected was not statistically significant between cases (13.4 ± 1.86 per 1000 cells) and controls (11.9 ± 1.38 per 1000 cells; P = 0.545). Using the 75th percentile value in the controls (29 aberrations per 1000 cells) as a cutoff point to classify 3p21.3 BPDE-induced sensitivity, 30 of the 52 cases (57.69%) but only 14 of the 54 controls (25.93%) were sensitive to BPDE (Table 2). This locus-specific mutagen sensitivity was associated with a significantly increased risk of HNSCC. After adjustment by age, ethnicity, and smoking status, the relationship was still evident with an OR of 4.80 (95% CI: 1.87–12.28). When we further categorized the subjects by the quartile distribution of 3p21.3 aberrations in the controls, we found a significant gradient of elevated risk of HNSCC with increasing numbers of BPDE-induced 3p21.3 aberrations. The adjusted ORs for subjects in the second, third, and highest quartiles of BPDE-induced 3p21.3 aberrations relative to the first quartile were 2.14 (95% CI: 0.18–25.66), 13.46 (95% CI: 1.36–133.09), and 27.65 (95% CI: 2.84–269.5), respectively. The trend was statistically significant (P < 0.05).
|3p21.3 BPDE sensitivity||No. (%)||Adjusteda OR (95% CI)|
|Nonsensitive||22 (42.31)||40 (74.07)||1|
|Sensitiveb||30 (57.69)||14 (25.93)||4.80 (1.87–12.28)|
|Number of aberrations|
|0–15||1 (1.92)||13 (24.07)||1|
|16–19||4 (7.69)||13 (24.07)||2.14 (0.18–25.66)|
|20–28||17 (32.69)||14 (25.93)||13.46 (1.36–133.09)|
|≥ 29||30 (57.69)||14 (25.93)||27.65 (2.84–269.5)|
The results of this study demonstrate a significant association between BPDE-induced 3p21.3 aberrations detected in PBLs and the risk for HNSCC. First, our data show that the average number of BPDE-induced chromosomal alterations is significantly higher at locus 3p21.3 than at the control locus 6q27 in both our case and control populations. This result suggests that the short arm of chromosome 3 might be one of the specific molecular targets of BPDE, supporting the notion that expression of locus-specific mutagen hot spots is not random and that these specific genetic aberrations may be etiologically related to specific cancers.16, 17 Second, we detect more BPDE-induced 3p21.3 aberrations in HNSCC cases than in healthy controls. In contrast, there is no significant difference in the number of aberrations observed at the control locus 6q27 between cases and controls. This finding illustrates a higher risk of HNSCC for individuals with greater BPDE-sensitivity at 3p21.3 and implies a role of the 3p21.3 region in HNSCC tumorigenesis. Finally, a dose-response relationship between the number of BPDE-induced aberrations and HNSCC risk provides further evidence for 3p21.3 as a hot spot associated with HNSCC risk by indicating that individuals with more 3p21.3 aberrations are at higher risk to develop HNSCC. These findings are consistent with results from our previous study on another smoking-related cancer, lung carcinoma.15
Ideally, normal target tissue (e.g., epithelial cells) should be used as the experimental material in studies of early events in carcinogenesis. However, because it is difficult to obtain large samples of cultured epithelial cells, lymphocytes may be used initially as a surrogate tissue sample. Fragile sites may be inherited throughout the whole human genome.18 For example, similar patterns of BPDE-induced adducts have been found in three cell types: HeLa cells, bronchial cells, and normal human fibroblasts. This demonstrates that the same molecular targets of a mutagen are probably present in different target cells during tumor transformation. Given the observation that mutagen hot spots in lymphocytes may be similar to those in target tissues, screening nonrandom cytogenetic defects in PBLs may be used as a tool to identify people who are predisposed to a specific cancer.10, 11, 19, 20
The findings of our molecular cytogenetic analysis in PBLs are also consistent with reported tumor tissue observations. A high frequency of LOH has been mapped to several distinct regions on the short arm of chromosome 3 in HNSCC patients by many studies utilizing microsatellite markers.14, 21, 22 Among these regions, 3p21 has the highest rate of allelic deletion,23 which supports our observation that deletion is the most common chromosomal aberration detected in the FISH analysis (data not shown). A cluster of chromosomal breaks was also detected on chromosome 3 of HNSCC patients in a cytogenetic study that applied Giemsa banding techniques to PBLs treated with various mutagens such as bleomycin and 4NQO.13 More importantly, 3p21.3 was the most often affected region among these hot spots. Thus, the congruence from different studies reaches the consensus that 3p21.3 may contribute to tumor development and 3p21.3 sensitivity to carcinogens may reflect an individual's susceptibility to HNSCC.
Although numerous studies have indicated a role of the chromosome region 3p21.3 in cancer development, the mechanism of these chromosomal fragilities in tumorigenesis is far from clear. The break sites may play a role in carcinogenesis through various pathways such as protooncogene activation, TSG inactivation, or DNA methylation.24 Inactivation of TSG could account for our findings, because a convergence of evidence from allele loss mapping, mutagen sensitivity, and FISH assays strongly suggests the presence of one or more putative TSGs at 3p21.3.25, 26 It is noteworthy that dysfunction of the mismatch repair gene hMLH1, located at 3p21.3, may not only generate a tendency for mutations in the genome but also increase the mutation incidence in oncogenes and TSGs.27 Other candidate TSGs including D8 (CHCM, ubiquitin-activating enzyme), ACY 1 (aminoacylase), APEH (D3S48E), and PTP gamma (protein tyrosine phosphatase) may also play a role in HNSCC development. In addition, there are two other genes located at 3p21.3 that are mutated in head and neck carcinomas. One is the human homolog of the ribosomal protein L14 (RPL14) gene,28 and the other is an arginine-rich protein gene.29 Damage to these genes, which are involved in growth factor control, DNA repair, and cell cycle regulation, would contribute to the initiation or promotion of the tumorigenic process. Our results support this contention by showing that subjects with greater sensitivity to BPDE at 3p21.3 have higher HNSCC risks.
In summary, our molecular cytogenetic study suggests that BPDE-induced aberrations are not random and that the 3p21.3 region is a specific molecular target of tobacco carcinogens. The degree of BPDE sensitivity at 3p21.3 may be a suitable marker for assessing predisposition to HNSCC.
- 13Nonrandom distribution of mutagen-induced chromosome breaks in lymphocytes of patients with different malignancies. Int J Oncol. 1994; 5: 733–740., , , .
- 22Three discrete regions of deletion at 3p in head and neck cancers. Cancer Res. 1993; 53: 5779–5885., , , , , .