Joint Effect of Mutagen Sensitivity and Insulin-Like Growth Factors in Predicting the Risk of Developing Secondary Primary Tumors and Tumor Recurrence in Patients with Head and Neck Cancer

Head and neck cancer (including malignancies in the oral cavity, pharynx, and larynx) will occur with an estimated incidence of 40,500 new cases (22,040 in oral cavity, 8,950 in pharynx, and 9,510 in larynx) and cause ∼11,170 deaths in the United States in 2006 (1). Cigarette smoking and alcohol consumption are the predominant risk factors for head and neck cancer, estimated to be responsible for three fourths of all the cases (2). However, only a small fraction of smokers and drinkers develop these cancers, suggesting that variations in genetic susceptibility may play important roles in cancer etiology. There have been numerous studies supporting this concept, including evidence of increased risk of head and neck cancer in first-degree relatives of cancer probands (3–6), significant associations of genetic polymorphisms in genes involved in critical cellular functions (carcinogen metabolism, DNA repair, cell cycle control, and apoptosis) with head and neck cancer risk (7, 8), and increased head and neck cancer risk conferred by reduced DNA repair capacity and increased genetic instability (9, 10).

Early-stage head and neck cancer patients are often cured with surgery or radiation. However, local-regional recurrence and secondary primary tumors (SPT) pose major threats to the long-term survival of early-stage head and neck patients. Approximately 10% of early-stage head and neck cancer patients develop local recurrences (11–13). SPTs occur in 15% to 25% of patients during the first 5 years after initial diagnosis with a relatively constant rate of about 3% to 6% yearly (13–19) and have become the leading cause of mortality in early-stage head and neck cancer patients (20–22). Although the term SPT indicates that these tumors and the index (primary) tumors developed independently, recent genetic and biochemical studies have suggested that, in a significant portion of patients, the SPTs and index tumors have a common clonal origin (23–26). The risk factors for SPT have been extensively investigated. SPT is not treatment related but is related to the same and/or distinct environmental or genetic factors that cause the index cancers. For example, analogous to the risk of index cancer, smoking and alcohol are unquestionable risk factors for SPT (15, 27, 28). Index tumor site and disease stage also contribute significantly to the risk of SPT and recurrence (13).

Diminished DNA repair capacity has been implicated as a risk factor for a variety of cancers (29). The mutagen sensitivity assay was developed to indirectly measure an individual's DNA repair capacity after in vitro mutagen challenge and may reveal latent genetic instability (30). An earlier pilot study suggested that mutagen sensitivity was a significant predictor of risk for developing SPT in previously untreated early-stage head and neck cancers (31). We have also shown previously from a subset of patients from the current study population that in vitro sensitivity of peripheral blood lymphocytes to bleomycin was associated with risk (10) and recurrence (12) of head and neck cancer. Insulin-like growth factor (IGF)-I is a potent stimulator of normal and tumor cell growth and also has antiapoptotic activities. The major form of circulating IGF-I is complexed with IGF-binding protein-3 (IGFBP-3), which inhibits the mitogenic activity of IGF-I by preventing IGF-I from binding to its receptor. There is considerable interindividual variability in circulating levels of IGF-I and IGFBP-3. Many studies have shown that elevated serum IGF-I levels are associated with increased risks for a variety of cancers (32–35). A recent study on a subset of patients with head and neck cancer from the current study population showed that elevated IGF-I levels were associated with a significantly increased risk of subsequent development of SPTs (36).

We therefore hypothesized that an individual's increased intrinsic carcinogen sensitivity (measured by the mutagen sensitivity assay) and proliferation potential (using serum IGF level as an indicator) contribute to elevated risk of SPT/recurrence and, furthermore, that there is a synergistic effect of these two adverse phenotypes. To test these hypotheses, we designed this large nested case-control study and determined the joint effects of mutagen sensitivity and IGF levels on the risk of SPT/recurrence.

Study population. The subjects included in this study were derived from the Retinoid Head and Neck Second Primary Trial that began in November 1991 and closed to new patient accrual in June 30, 1999. The primary end point of this trial was to evaluate whether a daily low dose of 13-cis-retinoic acid (13 cRA) for 3 years prevents SPTs in early-stage head and neck cancer patients who had been successfully treated with surgery or radiation or both. The detailed study design has been published previously (13, 15). Briefly, patients with histologically confirmed stage I or II squamous cell carcinoma of head and neck (including the larynx, oral cavity, or pharynx), who were cancer-free for at least 16 weeks at the time of recruitment, were eligible for randomization. Patients were enrolled from the Radiation Therapy Oncology Group, The University of Texas M. D. Anderson Cancer Center, the Clinical Community Oncology Group, and the Southwest Oncology Group. Patients were assigned to either low-dose (30 mg/d) 13 cRA treatment or the placebo arm with equal probability. The stage (I or II), primary tumor site (larynx, oral cavity, or pharynx), and smoking status (never/former, recent/current smoker) were used as stratifying factors to allocate treatment before randomization. Patients took 13 cRA or placebo for a total of 3 years followed by 4 years of follow-up. Follow-up information on SPTs and tumor recurrence was collected at 3, 6, 9, 12, 16, 20, 24, 28, and 36 months after randomization with additional evaluations at 6-month intervals for 4 years. A SPT was defined using the criteria established by Warren and Gates (37). These were new cancer of a different histologic type, one of identical histologic type occurring >3 years after therapy of the primary tumor or one separated from the initial primary tumor by >2 cm of clinically normal epithelium. The major sites of SPT in this population were lung (29.8%), head and neck (28.0%), prostate (14.2%), and bladder (5.1%). A local recurrence was defined as any tumor of similar histology appearing within 2 cm or 3 years of the primary tumor.

Data collection. Before randomization, participants were given a structured questionnaire that elicited information on sociodemographic factors, clinical information, tobacco exposure, and alcohol consumption. After the interview was completed, blood was drawn into heparinized tubes and delivered by overnight mail to M. D. Anderson Cancer Center to be used in cytogenetic and molecular analyses. The study was approved by all of the relevant review boards and in accordance with an assurance filed with, and approved by, the U.S. Department of Health and Human Services. Written informed consent was obtained for each participant.

Measurement of IGFs. Blood was centrifuged at 3,000 × g for 10 minutes at room temperature to separate the plasma. Collected plasma was stored at −80°C. The levels of IGF-I and IGFBP-3 were determined by an ELISA (Diagnostic Systems Laboratories, Webster, TX) according to the manufacturer's protocol. Cross-reaction of the antibodies with other members of the IGF family is not detected at physiologic concentrations according to the manufacturer. We separated IGF-I from its binding proteins by mixing plasma specimens with acid-ethanol extraction buffer and measured the serum levels of total IGF-I. For IGFBP-3, the specimens were diluted 100-fold in an assay buffer before conducting the ELISA test. All assays for IGF-I and IGFBP-3 were done in duplicate, and the average of the two measurements was used in the data analysis.

Mutagen sensitivity assays. Mutagen sensitivity was measured in vitro in lymphocytes by counting chromatid breaks induced by bleomycin as described previously (10). Briefly, blood cultures were incubated for 3 days and then exposed to bleomycin (0.03 units/mL) for 5 hours. Cells were harvested, and chromatid breaks were scored in 50 metaphases per sample and recorded as the mean number of breaks per cell. Laboratory personnel who read the slides were blinded to the case and control status.

Statistical analysis. Plasma levels of IGF-I and IGFBP-3 as well as bleomycin sensitivity were analyzed as continuous and categorical variables. As continuous variables, the differences between cases (patients with an event) and controls (patients without an event) were calculated with a two-sample Student's t test for demographic variables and mutagen sensitivity and a Wilcoxon rank sum test for IGFs (which are not normally distributed). As categorical variables, high levels of IGFs and mutagen sensitivity have been associated with increased SPT risk. The median value of plasma IGF-I and IGFBP-3 in controls was used as a cutoff point for high or low IGFs. A cutoff point determined by classification and regression tree analysis was applied to categorize each subject as either in the high or in the low bleomycin sensitivity stratum. The recursive partitioning process in classification and regression tree analysis does not depend on any underlying distributional assumptions and allows for nonlinear relations between predictive factors and outcomes. The cutoff points obtained correspond to the lowest cross-validation error rates. Univariate and multivariate Cox proportional hazards models were applied to calculate hazard ratios (HR) and corresponding 95% confidence intervals (95% CI) and analyze the effect of mutagen sensitivity and other variables on the development of SPT/recurrence. The joint effects between IGFs and bleomycin sensitivity were estimated by stratified analysis in which we formed joint strata of IGFs and bleomycin sensitivity values. All Ps were two sided. Associations were considered statistically significant at P < 0.05.

Table 1 summarizes the characteristics of 991 early-stage head and neck cancer patients available for analyses. An event (recurrence or SPT) occurred in 298 patients (referred to as cases) and 693 patients were event-free (referred to as controls). The median duration between blood draw and event (recurrence or SPT) development was ∼26 months (range, 0-92 months). The cases were significantly older than the controls (mean age, 62.6 ± 10.3 versus 60.3 ± 10.9 years; P = 0.003). Gender was not significantly associated with event development (P = 0.709). Compared with controls, cases included more patients with oral (32.55% versus 28.14%) and pharyngeal cancers (13.09% versus 9.09%) but fewer patients with laryngeal cancers (54.36% versus 62.77%). The cases had a significantly higher percentage of stage II patients (40.60%) than the controls (32.32%; P = 0.012). Compared with controls, cases were more likely to be current smokers (41.95% versus 35.79%) and current drinkers (52.35% versus 46.03%) and less likely to be never smokers (11.07% versus 14.14%) and never drinkers (15.44% versus 21.65%). Among smokers, the cases smoked more cigarettes daily (28.6 ± 16.6 versus 26.3 ± 15.8; P = 0.048) for a longer duration (37.4 ± 13.9 versus 34.8 ± 13.8 years; P = 0.005) than the controls, and consequently, the mean pack-years was significantly greater for the cases than the controls (55.2 ± 39.4 versus 47.6 ± 35.2; P = 0.010). The mean number of chromatid breaks per cell was higher in the cases (0.86 ± 0.44) than in the controls (0.82 ± 0.46; P = 0.083) but only of borderline statistical significance. Radiation and 13 cRA chemoprevention did not have an obvious effect on SPT/recurrence, whereas surgery provided some benefit in reducing the occurrence of SPT/recurrence (P = 0.076). Because the majority of the participants were Caucasians, our subsequent analyses were limited to this group.

We next calculated the HRs using the Cox proportional hazards model (Table 2). A cutoff of 0.5 chromatid breaks per cell was chosen by the recursive partitioning procedure to dichotomize the subjects into two groups of high and low bleomycin sensitivity. Overall, high bleomycin sensitivity was associated with a statistically significantly elevated risk for developing SPT/recurrence (adjusted HR, 1.38; 95% CI, 1.02-1.86). This risk was more evident in younger (HR, 1.84; 95% CI, 1.10-3.08) than in older patients (HR, 1.09; 95% CI, 0.75-1.58), in women (HR, 2.17; 95% CI, 1.05-4.46) than in men (HR, 1.24; 95% CI, 0.89-1.72), in former smokers (HR, 1.66; 95% CI, 1.03-2.68) than in current smokers (HR, 1.18; 95% CI, 0.76-1.81), and in lighter smokers (HR, 1.63; 95% CI, 0.98-2.69) than in heavy smokers (HR, 1.20; 95% CI, 0.80-1.80). Interestingly, the elevated risk was more evident in patients who had surgery (HR, 1.89; 95% CI, 1.15-3.12) than those who did not (HR, 1.12; 95% CI, 0.76-1.63) and in patients who did not receive radiotherapy (HR, 2.07; 95% CI, 1.21-3.55) than those who did (HR, 1.19; 95% CI, 0.82-1.72).

To evaluate the joint effects of bleomycin sensitivity and IGFs on the risk of SPT/recurrence, we measured the plasma levels of IGF-I and IGFBP-3 in a matched subset of patients (107 cases and 94 controls; Table 3). Controls were randomly chosen from the entire pool of patients who did not develop SPT/recurrence and were frequency matched to the cases on age (±5 years), sex, ethnicity, year of randomization, and length of follow-up. The cases exhibited significantly higher levels of IGF-I (mean, 160 ng/mL) and IGFBP-3 (mean, 3,164 ng/mL) than the controls (mean, 142 and 2,852 ng/mL; P = 0.022 and 0.042, respectively). The molar ratios of IGF-I/IGFBP-3 were not significantly different between the cases and controls (P = 0.957). In this subset of patients, the mean number of chromatid breaks per cell was significantly higher in the cases (0.76 ± 0.42) than in the controls (0.66 ± 0.48; P = 0.009). The joint effects of bleomycin sensitivity and IGF-I or IGFBP-3 on the risk of developing SPT/recurrence were next assessed (Table 4). The mean values of IGF-I and IGFBP-3 levels in controls were used as cutoff points to dichotomize the subjects into high and low IGF-I or IGFBP-3 groups. Using patients with low IGF-I and low bleomycin sensitivity as the reference group, the adjusted odds ratios (OR) of developing SPT/recurrence for patients with high IGF-I level alone, high mutagen sensitivity alone, and both high IGF-I level and high mutagen sensitivity were 2.85 (95% CI, 0.92-8.82), 3.92 (95% CI, 1.28-11.97), and 6.16 (95% CI, 2.03-18.71), respectively. A similar joint effect was observed for bleomycin sensitivity and IGFBP-3 level. Compared with patients with low IGFBP-3 and low bleomycin sensitivity, patients with high IGFBP-3 level alone (OR, 3.49; 95% CI, 1.02-11.86), high mutagen sensitivity alone (OR, 5.39; 95% CI, 1.56-18.63), and both elevated IGFBP-3 level and mutagen sensitivity phenotypes (OR, 7.07; 95% CI, 2.06-24.19) exhibited significantly elevated risks of SPT/recurrence.

We have previously reported that bleomycin-induced mutagen sensitivity was a significant predictor of SPT or recurrence in a subset of the current study population (12, 31). In this study, we confirmed this finding with substantially larger sample size and using SPT and recurrence as the combined outcome event. To our knowledge, this is the largest prospective study to evaluate mutagen sensitivity as a prognosis marker in head and neck cancer, and the data confirmed that high mutagen sensitivity was associated with significantly elevated SPT/recurrence risk in early-stage head and neck cancer patients. In addition, we showed for the first time the joint effects between elevated IGF levels (high proliferative potential) and the mutagen sensitivity phenotype (latent genetic instability) in the risk of developing SPT/recurrence.

Genetic instability may contribute to a high genetic heterogeneity in a tumor and hence contribute to poor clinical outcome (38). Charuruks et al. (39) showed a statistically significant increment of genetic instability in terms of normalized chromosome index and polysomy index in recurrent primary tumors and SPTs of head and neck cancer patients compared with tumors without recurrence and SPT, suggesting that genetic instability might be a potential molecular marker for risk assessment of recurrence and SPT in head and neck tumorigenesis. Likewise, in vitro mutagen sensitivity, reflecting latent genetic instability, may also be a biomarker for tumor recurrence and SPT in head and neck cancer patients. As shown in our previous small-scale studies (12, 31) and confirmed in this large-scale study, high mutagen sensitivity indeed confers an increased risk of developing recurrence and SPT in head and neck cancer patients. The large sample size allowed us to do stratified analyses based on age, gender, and smoking status. The elevated risk was more evident in women than in men, in younger cases than in older cases, in former smokers than in current smokers, and in lighter smokers than in heavy smokers. These observations are consistent with previous studies. We have shown that the association between benzo(a)pyrene diol epoxide (a tobacco carcinogen) sensitivity and head and neck cancer risk was higher in former than in current smokers and in younger than in older patients. In a case-control lung cancer study, younger cases, women, and lighter smokers exhibited the lowest DNA repair capacity and the highest cancer risk among their subgroups (40). The theory is that exposed individuals with susceptible phenotypes might develop primary or secondary tumors earlier and with less carcinogen exposure than those with more resistant phenotype. Genetic differences in risks for primary or secondary tumors tend to be smaller at high doses of carcinogens. For example, the current smokers in this population were generally heavy smokers; therefore, the risk of smoking might overpower a weaker genetic risk factor, which explains why former smokers and light smokers with bleomycin-sensitive phenotypes exhibited higher risks than current and heavy smokers with similar sensitive phenotypes. Therefore, our data, obtained from the largest phenotypic assays to date, strengthened the notion that mutagen sensitivity constitutes a susceptible phenotype not only for cancer risk but also for SPT/recurrence.

Several studies have shown that IGF-I is a stimulator of cell proliferation and inhibitor of apoptosis and that elevated serum IGF-I levels are associated with increased risks for a variety of cancers (32–35). We recently reported that elevated IGF-I levels were associated with a significantly increased risk of SPT in a subset of patients from the current study population (36). The data in this study were consistent with previous observations. More significantly, for the first time, we found that there is a joint effect between mutagen sensitivity and serum IGF-I level in modulating SPT/recurrence risk. Biologically, IGF level and mutagen sensitivity reflect two distinct yet interacting cellular functions (i.e., cell proliferation and DNA repair). Cells with deficient DNA repair plus uncontrolled proliferation will lead to a disproportionate increase in cells with damaged DNA. With a higher rate of cell turnover, there is increased likelihood for propagation of genetic errors. A previous study has shown that IGF-I and mutagen sensitivity jointly increased lung cancer risk (41). In this study, we found that IGF-I and mutagen sensitivity jointly increased risks of secondary tumors. These observations are consistent and biologically plausible.

Although the positive association between IGF-I level and various cancer risks has been consistently reported in the literature (32–35), the relationship between IGFBP-3 level and cancer risk remains controversial. Some studies showed that elevated serum levels of IGFBP-3 were associated with either increased or decreased cancer risk (34, 35, 42–47), whereas other studies did not detect an association between serum levels of IGFBP-3 and cancer risk (35, 48). The dual function of IGFBP-3 may explain these contradictory results. Circulating IGFBP-3 modulates the amount of bioavailable free IGF, thereby preventing their binding to IGF-I receptor and suppressing cell proliferation (49, 50). IGFBP-3, in this respect, inhibits IGF-I activity and may play a protective role against cancer risk. On the other hand, IGFBP-3 can also enhance IGF activity by presenting and slowly releasing IGF-I for receptor interactions while protecting the receptor from down-regulation by high IGF-I exposure (50). In this case, IGFBP-3 may serve as a risk factor for cancer. In this current study, the high risk of developing an SPT/recurrence conferred by higher level of IGFBP-3 may be explained by the dual function of IGFBP-3 in regulating IGF-I.

In our analyses, we combined SPT and recurrence as a single end point to increase sample size and power. In addition, it is difficult to clearly distinguish SPT and recurrent tumor biologically because recent genetic and biochemical studies have suggested that, in a significant portion of patients, the SPTs and index tumors have a common clonal origin (23–26). We also did the same analyses separating these two events. The results were very similar to what we presented here with combined events. Higher bleomycin sensitivity conferred borderline significantly increased risks for SPT or recurrence (data not shown). We did not elaborate the effect of tumor characteristics and smoking status on SPT and recurrence, which have been presented in detail previously (13, 15).

There are several strengths to this study. This is the largest prospective study of mutagen sensitivity to date. There is no selection bias because the population is a well-defined population of early-stage head and neck cancer patients and biological samples were obtained before an event developed, which suggests that the observations in this study were more likely to be the contributor rather than the effect of the event. A major limitation of the study is the limited number of subjects with serum IGF values and a one-time point measurement of IGF levels, which may be subjected to temporal confounders and laboratory variability. However, Kaaks et al. (46) reported that blood samples drawn 0.75 to 4.75 years apart produced highly correlated values for both IGF-I (r = 0.87) and IGFBP-3 (r = 0.73). Additional studies are warranted to validate the joint effects between IGF and mutagen sensitivity.

In conclusion, this prospective study confirmed that high mutagen sensitivity was associated with significantly elevated SPT/recurrence risk in early-stage head and neck cancer patients. In addition, this is the first study to find a joint effect between elevated IGF levels and mutagen sensitive phenotype in elevating SPT/recurrence risk. By combining these two pathways, we may be able to improve our ability to define high-risk populations of SPT/recurrence in early-stage head and neck cancer patients, which may have clinical implications for patient surveillance and chemoprevention.