Despite numerous studies showing that mutagen sensitivity is a cancer predisposition factor, the heritability of mutagen sensitivity has not been clearly established. In this report, we used a classic twin study design to examine the role of genetic and environmental factors on the mutagen sensitivity phenotype. Mutagen sensitivity was measured in peripheral blood lymphocytes from 460 individuals [148 pairs of monozygotic (MZ) twins, 57 pairs of dizygotic (DZ) twins, and 50 siblings]. The intraclass correlation coefficients were all significantly higher in MZ twins than in dizygotes (DZ pairs and MZ-sibling pairs combined) for sensitivity to four different mutagen challenges. Applying biometric genetic modeling, we calculated a genetic heritability of 40.7%, 48.0%, 62.5%, and 58.8% for bleomycin, benzo[a]pyrene diol epoxide, γ-radiation, and 4-nitroquinoline-1-oxide sensitivity, respectively. This study provides the strongest and most direct evidence that mutagen sensitivity is highly heritable, thereby validating the use of mutagen sensitivity as a cancer susceptibility factor. (Cancer Res 2006; 66(12): 5993-6)

It is widely recognized that both genetic and environmental factors influence cancer in a complex fashion. Aberration in many cellular functions may be involved in the etiology of cancer. DNA repair capacity is a cellular defense system designed to protect genomic integrity. There is considerable interindividual variation in DNA repair capacity. Numerous epidemiologic studies have consistently yielded significant associations between reduced DNA repair capacity and increased cancer occurrence (1, 2), suggesting that DNA repair capacity is a cancer susceptibility factor. Mutagen sensitivity, measured by quantifying the chromatid breaks induced by mutagen in short-term cultures of peripheral blood lymphocytes (PBL), has been used as an indirect measure of DNA repair capacity. In response to mutagen exposure, genetic damages would accumulate more quickly in people with suboptimal DNA repair capacity than in normal people. Depending on the level of chromatid breaks induced by mutagen treatment, individuals can be classified as nonsensitive or sensitive, which is believed to reflect their ability to repair DNA damages induced by selected mutagen. The original assay used bleomycin as the mutagen challenge (3). The assay has since been expanded to employ other mutagens to reflect different cancer etiologies, including benzo[a]pyrene diol epoxide (BPDE), a tobacco smoke carcinogen, for smoking-related carcinogenesis (4, 5); 4-nitroquinoline-1-oxide (4-NQO), a UV-mimetic mutagen, for melanoma initiation (6); and γ-radiation exposure for breast and brain cancer etiology (7, 8). Bleomycin, a radiomimetic drug, induces oxidative base damage and single-strand and double-strand breaks in DNA that necessitates base excision repair and double-strand break repair. BPDE covalently binds to DNA to form BPDE-DNA adducts, which require nucleotide excision repair. 4-NQO is a UV-mimetic agent and causes bulky lesions mainly repaired by nucleotide excision repair. γ-Radiation induces DNA strand breaks that require double-strand break repair and base excision repair.

Mutagen sensitivity has been increasingly recognized as a cancer-predisposing factor (18). However, almost all of the evidence supporting mutagen sensitivity as a cancer susceptibility factor is derived from case-control studies that compared mutagen sensitivity in unrelated individuals. The extent to which mutagen sensitivity is influenced by genetic factors remains unclear. Without strong evidence supporting high genetic heritability, the merit of mutagen sensitivity—an intermediate phenotype—as a cancer risk factor will continue to be scrutinized for the possibility of “reverse causation.” In this report, we applied the classic twin study design to examine the relative contribution of genetic and environmental factors to the variability in mutagen sensitivity. The twin model, by comparing similarities between monozygotic (MZ) twins (who are genetically identical) and dizygotic (DZ) twins (who share, on average, half their genes), is an ideal model to partition interindividual variability into components reflecting genetic and environmental factors. Twin studies have been widely used to validate the heritability of recognized phenotypic risk factors for cancer and other complex diseases (9, 10). In this study, we determined the heritability of mutagen sensitivity by challenging PBL with the above-mentioned four different mutagens.

Twin recruitment. The study was carried out within the framework of a study of pharmacokinetics of nicotine in twins in California (11, 12). The procedures for twin recruitment and exclusion criteria have been described in full detail elsewhere (11, 12). Informed consent was obtained from each subject. All methods for recruitment, informed consent, screening, and data collection were reviewed and approved by the Institutional Review Boards of SRI International, California.

Blood collection and mutagen sensitivity assays. For each participant, two 10-mL tubes of blood samples were collected at the Clinical Research Center at University of California-San Francisco at the baseline and delivered to the Department of Epidemiology, the University of Texas M.D. Anderson Cancer Center, by Federal Express mail within 24 hours of blood drawing for lab assays. Mutagen sensitivity assays in blood cultures were done as previously described (5, 6).

Statistical analysis. Descriptive statistics (mean, median, range, SD) were computed for all the participants and for subgroups. Differences between groups with respect to mutagen sensitivity were assessed by Student's t test or ANOVA while adjusting for nonindependence of data in families. Intraclass correlations were calculated between the set of MZ twins and the set of DZ twins (as well as the twin-sibling pairs). The intraclass correlation is given by r = (MSA − MSW) / (MSA + MSW), where MSA is the mean square among twin pair and MSW is the mean square within pair. Genetic heritability was quantified by a biometric genetic modeling as previously described (13, 14).

A total of 460 individuals participated in the study, including 148 pairs of MZ twins, 57 pairs of DZ twins, and 50 siblings of twins (either MZ or DZ). There were 353 (76.7%) Caucasians, 12 (2.6%) African Americans, 54 (11.7%) Hispanics, 17 Asians (3.7%), and 24 (5.2%) other ethnicities. The numbers of never, former, and current smokers were 270 (60.8%), 104 (22.7%), and 76 (16.6%), respectively.

Whereas the mutagen sensitivity assay has been shown to be reproducible and reliable, no previous study has been done to systematically evaluate the measurement error in duplicate cultures from the same subjects. In our study, we did the four mutagen sensitivity assays in duplicate samples to assess measurement error. We found that the correlations between duplicate samples were all highly significant in all four assays. The correlation coefficients for bleomycin, BPDE, γ-radiation, and 4-NQO assays were 0.70 (P < 0.001; n = 25 pairs), 0.75 (P < 0.001; n = 63 pairs), 0.53 (P < 0.001; n = 55 pairs), and 0.82 (P < 0.001; n = 68 pairs), respectively. The repeatability of the mutagen sensitivity assay of multiple samples from the same person at various time intervals has previously been shown (15, 16).

The sensitivity to the four mutagens was generally not associated with epidemiologic variables, including ethnicity, gender, age, or smoking status (data not shown). Correlations between the four mutagen sensitivity scores were evaluated and significant correlations were only seen between BPDE and bleomycin (r = 0.30, P < 0.001), between bleomycin and γ-radiation (r = 0.22, P < 0.001), and between 4-NQO and γ-radiation (r = 0.11, P = 0.03). Intraclass correlations were calculated between the set of MZ twins and the set of dizygotes (DZ pairs and MZ-sibling pairs combined). The intraclass correlation coefficients for MZ twins were 0.67, 0.72, 0.55, and 0.61 for bleomycin, BPDE, γ-radiation, and 4-NQO sensitivity (all P < 0.001; Fig. 1, left), respectively, whereas the intraclass correlation coefficients for dizygotes were 0.49, 0.27, 0.26, and 0.39 for these four markers, respectively (Fig. 1, right). The correlations were significantly higher within MZ pairs than within dizygotes for all four mutagen sensitivities, suggesting a strong genetic influence.

Figure 1.

Intraclass correlation of mutagen sensitivity (mean breaks per cell) between twin 1 and twin 2 in MZ twins (left) and between person 1 and person 2 in dizygotes (right), as calculated by use of ANOVA. Due to the low number of DZ twins, we combined DZ twin pairs with MZ-sibling pairs. In families with MZ pairs and siblings, we randomly picked one co-twin from the twin pair and one sibling to form a MZ-sibling pair. Similarity for MZ twins is significantly greater than for dizygotes in all four mutagen sensitivity assays.

Figure 1.

Intraclass correlation of mutagen sensitivity (mean breaks per cell) between twin 1 and twin 2 in MZ twins (left) and between person 1 and person 2 in dizygotes (right), as calculated by use of ANOVA. Due to the low number of DZ twins, we combined DZ twin pairs with MZ-sibling pairs. In families with MZ pairs and siblings, we randomly picked one co-twin from the twin pair and one sibling to form a MZ-sibling pair. Similarity for MZ twins is significantly greater than for dizygotes in all four mutagen sensitivity assays.

Close modal

We then applied biometric genetic modeling by partitioning the total variance into variance due to genetic, shared environment, and nonshared environment components and included age as a covariate (refs. 13, 14; Table 1). The null hypothesis, which assumes no genetic effects, was rejected because the restricted model, assuming no genetic influence, fitted the data significantly worse than the full model, suggesting the genetic component contributed significantly to all four markers (P < 0.01). The model generated a heritability estimate of 40.7%, 48.0%, 62.5%, and 58.8% for bleomycin, BPDE, γ-radiation, and 4-NQO sensitivity. In addition, the contribution of nonshared environment was also substantial, accounting for 33.3%, 31.8%, 37.5%, and 39.7% for the four markers, respectively, whereas the contribution of shared environment was relatively small (26%, 20.2%, 0%, and 1.5% for the four markers, respectively).

Table 1.

Contribution of genetic and environmental variance components to mutagen sensitivity

Bleomycin
BPDE
γ-Radiation
4-NQO
Full model*No genetic influenceFull modelNo genetic influenceFull modelNo genetic influenceFull modelNo genetic influence
Log-likelihood 199.967 194.664 206.199 201.101 491.806 487.430 237.290 230.99 
χ2  10.606  10.195  8.750  12.60 
P (two sided)  <0.01  <0.01  <0.01  <0.01 
Degrees of freedom     
Age coefficient 0.006 0.006 −0.004 −0.004 0.001 0.001 0.003 0.002 
Variances         
    Genetic 0.077 Fixed 0.083 Fixed 0.0316 Fixed 0.077 Fixed 
    Shared 0.049 0.109 0.035 0.112 0.000 0.015 0.002 0.015 
    Nonshared 0.063 0.081 0.055 0.069 0.019 0.026 0.052 0.026 
Heritability (%) 40.7  48.0  62.7  58.8  
Bleomycin
BPDE
γ-Radiation
4-NQO
Full model*No genetic influenceFull modelNo genetic influenceFull modelNo genetic influenceFull modelNo genetic influence
Log-likelihood 199.967 194.664 206.199 201.101 491.806 487.430 237.290 230.99 
χ2  10.606  10.195  8.750  12.60 
P (two sided)  <0.01  <0.01  <0.01  <0.01 
Degrees of freedom     
Age coefficient 0.006 0.006 −0.004 −0.004 0.001 0.001 0.003 0.002 
Variances         
    Genetic 0.077 Fixed 0.083 Fixed 0.0316 Fixed 0.077 Fixed 
    Shared 0.049 0.109 0.035 0.112 0.000 0.015 0.002 0.015 
    Nonshared 0.063 0.081 0.055 0.069 0.019 0.026 0.052 0.026 
Heritability (%) 40.7  48.0  62.7  58.8  

NOTE: The pedigree-based maximum likelihood method (13, 14) was used to quantify the influences of genetic and environmental contributions to mutagen sensitivity.

*

The full model includes variances due to genetic, nonshared, and shared environmental factors. The pedigree-based maximum likelihood method fitted to the data by specifying age on the mean phenotype (breaks per cell) as a fixed effect and partitioning the total variance by the contribution of genetic influence, shared environment (common environmental factors shared by family members), and nonshared environment (environmental factors unique to each individual subject) to the full model.

The major concern about the validity of mutagen sensitivity, an intermediate phenotype, as a genetic marker for cancer susceptibility is the problem of “reverse causation” (i.e., whether disease status and/or prior environmental exposure cause variation in mutagen sensitivity phenotype). The heritability of mutagen sensitivity was initially implicated in early studies showing that first-degree relatives of mutagen-sensitive individuals were also mutagen sensitive (1719), suggesting genetic predisposition to mutagen sensitivity. One strong support for the heritability of mutagen sensitivity came from a study of measuring chromosomal radiosensitivity in families of patients with breast cancer. Twenty-three of the 37 (62%) first-degree relatives of sensitive patients were also sensitive compared with 1 of 24 (4.2%) first-degree relatives patients with normal responses. Segregation analysis of 95 family members showed clear evidence of heritability of radiosensitivity, with a single major gene accounting for 82% of the variance between family members (18, 19). However, family studies cannot distinguish between shared environment and genetic factors. Previous case-control studies indicated that mutagen sensitivity remained a significant predictor of cancer risk independent of demographic and environmental risk factors (5, 20). However, as limited by study design, the cause-effect relationship between mutagen sensitivity and cancer cannot be answered in case-control studies. A twin study is the most powerful tool for direct assessment of a genetic component and offers the capability of differentiating genetic effects from shared environment within families (21). Two previous pilot studies estimated the heritability of bleomycin sensitivity using the twin design: one involved 25 pairs of MZ twins and 14 pairs of dizygotes (twin pairs and siblings; ref. 22) and the other had only 9 MZ twins and 10 DZ twins (23). Cloos et al. (22) estimated a high heritability of 75% for bleomycin sensitivity in a study involving 25 pairs of MZ twins and 14 pairs of dizygotes (twin pairs and siblings), suggesting a clear genetic basis for bleomycin sensitivity. Tedeschi et al. (23) reported higher correlations of bleomycin sensitivity score in MZ twins as compared with DZ twins and yielded a heritability of 0.58, although this estimate was based on intraclass correlations only and the biometric genetic modeling was not applied in their study. Our results corroborate these pilot studies with a substantially larger sample size (148 pairs of MZ twins and 82 pairs of dizygotes). Moreover, we provided the first systematic evaluation of the genetic heritability of mutagen sensitivity by extending to other more commonly used and biologically relevant mutagens. We also estimated, for the first time, the substantial contribution of nonshared environmental factors to mutagen sensitivity, highlighting the presence of gene-environment interaction in modulating cancer risk. It is apparent that the genetic heritability varies among the cellular sensitivity to different agents and the nonshared and shared environment contributes to mutagen sensitivity with differing degrees among the four agents. The mechanisms behind these variations are unclear. Mutagen sensitivity is an indirect measurement of DNA repair capacity and different mutagens elicit different DNA repair pathways. It is understandable that base excision repair, nucleotide excision repair, and double-strand break repair involve different genes, follow distinct repair kinetics, and therefore exhibit variations in heritability. Even for the similar damages (e.g., in the case of BPDE and 4-NQO), the systems involved may differ significantly. Theoretically, BPDE and 4-NQO induce similar sensitivity profiles and the removal of both 4-NQO and BPDE adducts mainly requires nucleotide excision repair. However, there are other mechanisms involved besides nucleotide excision repair. BPDE is already a highly reactive intermediate, which does not require bioactivation. 4-NQO, on the other hand, must be bioactivated through nitroreduction to produce a reactive intermediate. The in vitro 4-NQO assay may therefore reflect the cumulative bioactivation ability and DNA repair capacity (6). Similarly, radiomimetic bleomycin and γ-radiation both mainly induce DNA strand breaks; however, these DNA damages are structurally complex and follow different repair kinetics (24). Our results suggest that γ-radiation sensitivity has the highest heritability (62.5%), confirming the high heritability of cellular radiosensitivity from family studies of breast cancer (18, 19). The biological mechanism for the variation in sensitivity to different agents remains to be investigated.

In conclusion, our data provide the strongest evidence that genetic influence contributes significantly to mutagen sensitivity phenotypes, validating the use of mutagen sensitivity as a genetic susceptibility factor. Identifying genes involved in this susceptibility and determining the segregation pattern of these genes could provide further insight into the etiology of cancer.

Grant support: National Cancer Institute grants CA085576 (X. Wu) and DA11170 (G.E. Swan).

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.

We thank Mary R. Mcelroy, Ruth E. Krasnow, and Jill Rubin for subject recruitment.

1
Berwick M, Vineis P. Markers of DNA repair and susceptibility to cancer in humans: an epidemiologic review.
J Natl Cancer Inst
2000
;
92
:
874
–97.
2
Spitz MR, Wei Q, Dong Q, Amos CI, Wu X. Genetic susceptibility to lung cancer: the role of DNA damage and repair.
Cancer Epidemiol Biomarkers Prev
2003
;
12
:
689
–98.
3
Hsu TC, Johnston DA, Cherry LM, et al. Sensitivity to genotoxic effects of bleomycin in humans: possible relationship to environmental carcinogenesis.
Int J Cancer
1989
;
43
:
403
–9.
4
Wu X, Hsu TC, Spitz MR. Mutagen sensitivity exhibits a dose-response relationship in case-control studies.
Cancer Epidemiol Biomarkers Prev
1996
;
5
:
577
–8.
5
Wu X, Gu J, Amos CI, Jiang H, Hong WK, Spitz MR. A parallel study of in vitro sensitivity to benzo[a]pyrene diol epoxide and bleomycin in lung cancer cases and controls.
Cancer
1998
;
83
:
1118
–27.
6
Gu J, Bondy ML, Sigurdson A, Spitz MR, Hsu TC, Wu X. Three measures of mutagen sensitivity in a cancer-free population.
Cancer Genet Cytogenet
1999
;
110
:
65
–9.
7
Buchholz TA, Wu X. Radiation-induced chromatid breaks as a predictor of breast cancer risk.
Int J Radiat Oncol Biol Phys
2001
;
49
:
533
–7.
8
Bondy ML, Kyritsis A, Gu J, et al. Mutagen sensitivity and risk of gliomas: a case-control analysis.
Cancer Res
1996
;
56
:
1484
–6.
9
Bataille V, Snieder H, MacGregor AJ, Sasieni P, Spector TD. Genetics of risk factors for melanoma: an adult twin study of nevi and freckles.
J Natl Cancer Inst
2000
;
92
:
457
–63.
10
Boyd NF, Dite GS, Stone J, et al. Heritability of mammographic density, a risk factor for breast cancer.
N Engl J Med
2002
;
347
:
886
–94.
11
Swan GE, Benowitz NL, Jacob P III, et al. Pharmacogenetics of nicotine metabolism in twins: methods and procedures.
Twin Res
2004
;
7
:
435
–48.
12
Swan GE, Benowitz NL, Lessov CN, Jacob P, Tyndale RF, Wilhelmsen K. Nicotine metabolism: the impact of CYP2A6 on estimates of additive genetic influence.
Pharmacogenomics
2005
;
15
:
115
–25.
13
Amos CI, Zhu DK, Boerwinkle E. Assessing genetic linkage and association with robust components of variance approaches.
Ann Hum Genet
1996
;
60
:
143
–60.
14
de Andrade M, Amos CI, Thiel TJ. Methods to estimate genetic components of variance for quantitative traits in family studies.
Genet Epidemiol
1999
;
17
:
64
–76.
15
Cloos J, Steen I, Joenje H, et al. Association between bleomycin genotoxicity and non-constitutional risk factors for head and neck cancer.
Cancer Lett
1993
;
74
:
161
–5.
16
King TM, Trizna Z, Wu X, et al. A clinical trial to evaluate the effect of vitamin C supplementation on in vitro mutagen sensitivity.
Cancer Epidemiol Biomarkers Prev
1997
;
6
:
537
–42.
17
Knight RB, Parshad R, Price FM, Tarone RE, Sanford KK. X-ray-induced chromatid damage in relation to DNA repair and cancer incidence in family members.
Int J Cancer
1993
;
54
:
589
–93.
18
Roberts SA, Spreadborough AR, Bulman B, Barber JBP, Evans DGR, Scott D. Heritability of cellular radiosensitivity: a marker of low-penetrance predisposition genes in breast cancer.
Am J Hum Genet
1999
;
65
:
784
–94.
19
Scott D. Chromosomal radiosensitivity and low penetrance predisposition to cancer.
Cytogenet Genome Res
2004
;
104
:
365
–70.
20
Cloos J, Spitz MR, Schantz SP, et al. Genetic susceptibility to head and neck squamous cell carcinoma.
J Natl Cancer Inst
1996
;
88
:
530
–4.
21
Boomsma D, Busjahn A, Peltonen L. Classical twin studies and beyond.
Nat Rev Genet
2002
;
3
:
872
–82.
22
Cloos J, Nieuwenhuis EJ, Boomsma DI, et al. Inherited susceptibility to bleomycin-induced chromatid breaks in cultured peripheral blood lymphocytes.
J Natl Cancer Inst
1999
;
91
:
1125
–30.
23
Tedeschi B, Cicchetti R, Argentin G, et al. Aphidicolin and bleomycin induced chromosome damage as biomarker of mutagen sensitivity: a twin study.
Mutat Res
2004
;
546
:
55
–64.
24
Pastwa E, Neumann RD, Mezhevaya K, Winters TA. Repair of radiation-induced DNA double-strand breaks is dependent upon radiation quality and the structural complexity of double-strand breaks.
Radiat Res
2003
;
159
:
251
–61.