Carcinogenesis vol.21 no.6 pp.1239–1246, 2000
Mutagen sensitivity and environmental exposures as contributing
causes of chromosome 3p losses in head and neck cancers
Stimson P.Schantz5, Qiang Huang, Kinner Shah,
V.V.V.S.Murty2, T.C.Hsu3, Guopei Yu, Peter E.Andersen4,
Andrew G.Huvos1 and Raju S.K.Chaganti2
Department of Surgery, 1Department of Pathology and 2Department of
Human Genetics, Memorial Sloan-Kettering Cancer Center, New York,
NY 10021, 3Department of Cell Biology, The University of Texas M.D.
Anderson Cancer Center, Houston, TX 77030 and 4Department of
Otolaryngology–Head and Neck Surgery, Oregon Health Sciences
University, Portland, OR 97201, USA
5To
whom correspondence should be addressed at Department of
Otolaryngology, New York Eye and Ear Infirmary, New York Medical
College, 310 East 14th Street, New York, NY 10003, USA
Email: [email protected]
The interaction between environmental exposures and
host susceptibility may lead to specific mutational events
within head and neck squamous cell carcinoma (HNSCC).
Furthermore, this interplay may determine not only
the probability of cancer development but also the
biologic characteristics of the tumor once it occurs. To
better understand the relationship of mutagen sensitivity
and tobacco and/or alcohol consumption on HNSCC
carcinogenesis, we examined loss of heterozygosity on
chromosome 3p in 58 HNSCCs using 10 microsatellite
markers. Mutagen sensitivity was determined in vitro by
quantitating bleomycin-induced chromatid breaks utilizing
peripheral blood lympocytes from respective patients.
Forty-six of the 58 invasive cancers showed allelic loss at
one or more loci. Consistent with previous investigations,
three discrete regions of deletions were identified: 3p13–
14.2, 3p21.1–21.2, and 3p25.1–26.1. The frequency and
types of deletions were dependent upon tobacco and alcohol
exposures. The distal region of 3p but not the remaining
two regions was most frequently influenced by tobacco
exposure. In contrast, heavy alcohol use when combined
with tobacco use was associated with whole-arm loss of 3p
rather than identifiable site-specific damage. Furthermore,
this combined influence of alcohol and tobacco exposures
on whole-arm loss was most apparent in those patients
who expressed mutagen-sensitivity; the odds ratio of wholearm loss increasing from 2.67 (95% CI 0.21–33.49) in those
individuals who were mutagen resistant to 13.5 (95% CI
1.3–136.0; P ⍧ 0.02 by Fisher’s exact test) in those
who were mutagen sensitive. An assessment of clinical
parameters in this population demonstrated that patients
with whole-arm loss were more likely to present with
cervical lymph node metastases and advanced stage disease
than patients with partial losses. Results indicate that
various environmental exposures as well as the expression
of mutagen sensitivity will influence the types of chromosome 3p allelic losses in head and neck cancers as well as
the behavior of disease once it develops.
Abbreviations: HNSCC, head and neck squamous cell carcinoma; LOH, loss
of heterozygosity.
© Oxford University Press
Introduction
The development of head and neck cancer may depend
upon the interaction between host susceptibility factors and
environmental carcinogens. Tobacco and alcohol have been
implicated as the etiologic agents in ~80% of individuals (1).
Host susceptibility factors are only now being elucidated and
may include abnormalities in carcinogen metabolism, factors
linked to blood group antigen expression, as well as deficiencies
in DNA repair. Furthermore, exposure to tobacco carcinogens
may lead to specific mutational events commonly identified in
head and neck cancers, the most common of which involves
the p53 gene. This gene is observed to be mutated in 42–69%
of head and neck cancers, dependent upon the extent of a
patient’s tobacco use (2–5). Mutations commonly involve a
G→A transversion which can be duplicated in the laboratory
with benzo[a]pyrene (6).
One of the more frequently identified genetic alterations in
head and neck cancers involves the short arm of chromosome
3 in which chromosomal rearrangements and deletions
predominate. Indeed, deletion mapping of chromosome 3p
shows losses involving at least one locus in ⬎50% of the
patients. Using allelotype analysis, Maestro et al. and others
found three regions of chromosomal loss, 3p14-cen, 3p21.3
and 3p24-ter (reviewed in ref. 7). Though not universally
confirmed, Roz et al. identified similar losses and demonstrated
that such mutations commonly occurred early in the progression
from dysplasia to invasive disease (reviewed in ref. 7). The
high frequency of similar losses in other tobacco-related
diseases such as the lung and esophagus would suggest a
causal relationship; an observation reinforced in the study by
Sozzii et al. (8) in which the probability of mutations involving
the FHIT gene at 3p14.2 is significantly increased within lung
cancers from smokers as compared with non-smokers, but no
point mutations or deletions of this gene were found in head
and neck squamous cell carcinoma (HNSCC) (9).
A premise in this study is that damage to chromosome 3p
in head and neck cancers may be influenced by the interaction
between carcinogen exposures and host susceptibility factors.
One such factor within head and neck cancer patients may be
reflected in the mutagen-sensitivity assay (10–12). The assay
makes use of peripheral blood lymphocytes in order to test for
bleomycin-induced chromosomal breakage in vitro. Bleomycin
induces chromosomal damage through the generation of free
radical oxygen and, thus, is reflective of one measure of
tobacco-induced damage. Previous studies have demonstrated
that head and neck cancer patients may be abnormally sensitive
to bleomycin-induced chromosomal damage as compared with
age- and sex-matched healthy controls (11,13–20). Furthermore, risk assessments in these studies suggest an interaction
between carcinogen exposure and mutagen-sensitivity measures, risk estimates being highest in those individuals who
both consume tobacco and express sensitivity to free radical
oxygen damage in vitro (14,16). The basis of this mutagen
sensitivity may reflect either an underlying DNA repair
1239
S.P.Schantz et al.
deficiency or factors which control susceptibility to initial
clastogenic influences (19).
Relevant to our presented study, Dave et al. (21) has also
suggested that chromatid break sites induced by bleomycin
in vitro were not random but rather were predetermined by
both host-susceptibility factors as well as specific mechanisms
related to free radical oxygen–chromatid interactions. In
support of this hypothesis, lymphocytes obtained from head
and neck cancer patients more frequently demonstrated breaks
on the short arm of chromosome 3 following bleomycin
exposure in vitro than similarly treated lymphocytes from
either healthy controls or patients with melanoma. In light of
these observations, we present here additional evidence of
deletional events involving chromosome 3p with three discrete
regions of deletion. Furthermore, we were able to extend
previous 3p allelotypic studies by assessing for the relationship
of specific deletional events to the duration and intensity of
tobacco and alcohol exposures; the latter information obtained
as part of a prospective case–control epidemiologic study
(16). The above results were then examined with respect to
quantitative measures of mutagen sensitivity expressed in vitro
by peripheral blood lymphocytes from the respective patients.
The data presented here will demonstrate that 3p deletional
events will vary, dependent upon the level of tobacco and
alcohol exposure and their overall frequency will be influenced
by the patient’s expression of mutagen sensitivity.
Materials and methods
Patient population
Fifty-eight patients with previously untreated squamous cell carcinoma of the
upper aerodigestive tract formed the basis of this study. Information regarding
carcinogen exposures was obtained using a self-administered questionnaire
prior to treatment, mutagen-sensitivity assessment and analysis of 3p deletions
in the patient’s respective cancer. This questionnaire included information
regarding the duration and intensity of tobacco and alcohol exposure. History
regarding tobacco exposure included type of tobacco consumed, years of
smoking and the age at initiation. Alcohol consumption patterns included the
types of alcohol, frequency and quantities of alcohol consumption, and years
of alcohol ingestion. The validation of this self-administered, comprehensive
questionnaire has been reported previously (22). Of note, in three patients no
questionnaire information was provided and information regarding tobacco
and alcohol exposures were obtained from the medical chart without knowledge
of mutagen-sensitivity and allelotype assessments. Informed consent was
obtained from all patients prior to questionnaire administration using a protocol
approved by the institutional review board of Memorial Sloan-Kettering
Cancer Center.
Mutagen-sensitivity assay
The assay utilized in this study has been described in detail previously (16,23).
A peripheral blood sample was collected from each donor in a heparinized
tube prior to the initiation of lymphocyte culture. The standard lymphocyte
culture procedure utilized RPMI-1640 medium, supplemented with 15% fetal
calf serum and phytohemagglutinin in a ratio of blood to medium of 1:9. At
67 h of incubation, cultures are treated with bleomycin (0.03 U/ml) for 5 h.
Colcemid (0.04 mg/ml) is added in the last hour to induce mitotic arrest prior
to harvesting. Conventional cell harvesting procedure follows: the cells are
treated with hypotonic KCl (0.06 M) solution for 15–20 min, fixed, washed
with a freshly prepared mixture of methanol and acetic acid (3:1) and air-dried
on wet slides. The slides are stained with Giemsa solution without banding.
Fifty well-spread metaphases are examined from coded slides. Chromatid
aberrations are recorded as frank chromatid breaks or exchanges. Bleomycin
tends to induce few chromatid exchanges (which if present, are considered as
two breaks). Chromatid gaps or attenuated regions are disregarded. The
frequency of breakage was expressed as breaks per cell (b/c) for purposes of
comparison.
The reliability of cytogenetic scoring has previously been evaluated by
comparing four separate blood samples from a respective donor with a
minimum interval between samples of 1 week (23). Using a random effect
one-way analysis of variance model, significance within group variation was
noted, suggesting that sensitivity appeared to be stable and representative.
1240
Table I. Frequency of LOH on chromosome 3pa
Locus (D3S-)
Map positionb
Infc
LOH (%)d
1038
1110
656
1007
1611
966
1076
1067
1228
659
p25.2–26.1
p25.1–25.3
p25.1
p25
p22–24.1
p21.31–21.32
p21.1–21.2
p14.3–21.1
p14.1–14.2
p13
40
40
28
34
33
36
29
45
38
49
31
25
10
19
15
19
20
19
23
30
(78)
(63)
(36)
(56)
(45)
(53)
(69)
(42)
(61)
(61)
aResults were generated from
bThe map positions are based
analysis of cancers from 58 patients.
on Genome Data Base (GDB).
number of tumors which were informative for that particular loci.
dThe number of tumors which showed LOH and as a percentage of
informative cases.
cThe
Microdissection and DNA isolation
Normal tissue, dysplasia and frankly invasive cancer were identified by a
pathologist (A.G.H.) using a 5 µm H&E stained section. Four to eight
additional 8 µm serial sections were obtained from the same specimens above
for microdissection. Using the 5 µm stained sections as a guide, suitable areas
for tumor, dysplasia and normal non-epithelial tissue contained within the
corresponding 8 µm sections were microdissected using a microcapillary
pipette that had been reconfigured into a fine pipette tip. The microdissected
cells, deparaffinized in xylene, were collected in a 1.5 ml Eppendorf tube
containing 100 µl of digestion buffer (50 mM Tris–HCl pH 8.5, 1 mM EDTA
and 0.5% Tween-20) and then subjected to proteinase K (at a final concentration
of 400 µg/ml) digestion at 50°C for 72 h. The digested samples, which were
incubated at 95°C for 10 min to inactivate the proteinase K, were centrifuged
and the supernatant was extracted with phenol–chloroform and ethanol
precipitation before being used for PCR analysis.
Loss of heterozygosity (LOH) analysis
Ten microsatellite markers mapping to chromosome 3p were used (Table I).
All primer pairs were obtained from Research Genetics (Huntsville, AL). A
standard PCR reaction was carried out in a toal volume of 15 µl containing 25–50 ng of DNA, 10 pmol of each primer, 0.3 µCi of [α-32P]dCTP
(3000 Ci/mmol; DuPont New England Nuclear, Boston, MA). The PCR
products were denatured in sequencing stop solution and subjected to electrophoresis in 6% denaturing polyacrylamide gels and the dried gels were
autoradiographed for 4–72 h. Either the absence or the decrease in signal
intensity by ⬎50% of one allele in tumor or dysplasia DNA compared with
the same allele in normal DNA by visual examination was considered as
LOH. Alterations were judged as microsatellite instability when additional
bands appeared in tumor or dysplasia DNA. All samples showing LOH or
microsatellite instability were subject to repeat analysis and in all cases, the
replicate test gave identical results.
Results
Allelic deletions on 3p
The frequency of LOH in tumor DNA among the tested loci
ranged from 36% (10 of 28 cases) at D3S656 to 78% (31 of
40 cases) at D3S1038 (Table I). Forty-six of the 58 invasive
carcinoma specimens (79%) showed allelic loss at one or
more loci. Of these, 17 (37%) tumors showed LOH at all
informative loci suggestive of entire 3p deletions. The
remaining 29 tumors exhibited interstitial deletion. The
boundaries for minimal deletion were defined by retention of
heterozygosity of markers on either side or on one side of a
loss. The patterns of LOH in this group identified three
common regions of deletions at 3p25.1–26.1, 3p21.1–21.2,
and 3p13–14.2. The distal 3p25.1–26.1 region spanning the
markers D3S1038 and D3S1110 was deleted in 22 (76%)
tumors with interstitial deletions. The tumors p-19, p-6, o-36,
y-14, p-9, y-10, p-3, y-16 and p-22 define the boundaries. The
Chromosome 3p losses in head and neck cancers
Fig. 1. Patterns of LOH and replication errors on chromosome 3p in 46 head and neck cancers.Vertical lines on right of the idiogram indicate sites of three
discrete regions of deletion. Solid, hatched and blank squares represent LOH, retention of heterozygosity and homozygous/non-informative cancers,
respectively. Solid circles within squares indicate replication error-type microsatellite instability. ND, not done.
Table II. The relationship between tobacco use and LOH within three discrete regions of chromosome 3pa
Map position
p25.1–26.1
p21.1–21.2
p13–14.2
Locus (D3S-)
1038
1110
1076
1228
659
Non-users
⬎20 cigarettes/day
1–20 cigarettes/day
P-valueb
Inf.
LOH
(%)
Inf.
LOH
(%)
Inf.
LOH
(%)
7
3
(43)
16
8
(50)
11
10
(91)
0.045
3
7
1
3
(33)
(43)
10
17
7
8
(70)
(47)
7
14
4
7
(57)
(50)
0.594
1.000
Inf, the number of tumors which were informative for that particular loci.
aLOH is defined as a loss of informative loci within the indicated region. Results were generated from the analysis of cancers from 55 patients and each
informative case represents an individual patient.
bP-value was determined by Fisher’s exact test (two-tail).
proximal 3p13–14.2 region containing the markers D3S659
and D3S1228 was deleted in 20 (69%) tumors with interstitial
deletions. The boundaries of the 3p13–14.2 were defined by
the tumors y-20, y-6, y-5, y-9, p-33, p-6, y-2, y-17, p-10, o-36,
y-10, p-3 and y-11. The third region, 3p21.1–21.2, spanned
by the marker D3S1076 was deleted in 13 (45%) tumors
with the boundaries identified by y-17, y-14, p-26 and y-8
(Figure 1). Of the 29 tumors identified with common regions
of deletion, 10 showed deletions at all three regions, 11 at two
regions, and eight tumors exhibited deletions at one region.
Thus, these data have identified three specific regions of
deletions on 3p in head and neck cancer.
To determine the relative timing of chromosome 3p alterations in head and neck cancer progression, we analyzed LOH
on this chromosome in 18 paired dysplasia and tumor DNA
samples. In all instances, the dysplasia from each patient was
identified as adjacent to the respective individual’s cancer. All
samples were graded as mild to moderate dysplasia with no
evidence of severe dysplasia or carcinoma in situ. As a
comparison, we similarly assessed for losses involving the
9p21 locus using the D9S171 primer. This region has
characteristically been identified as deleted in dysplastic oral
lesions and is thereby considered an early event (24,25).
Although five of 18 (28%) dysplastic samples in which DNA
had been obtained showed LOH in at least one locus, the
overall frequency of loci showing LOH was very low (five of
115 informative loci; 4%). In contrast, the corresponding tumor
samples adjacent to the dysplastic lesions showed a high
incidence of LOH. Eleven (61%) of the 18 tumors showed
evidence of loss in at least one informative locus. Overall, 55
(45%) of the 121 informative loci within the cancers displayed
LOH. Assessment of mutational events at chromosome 9p21
showed LOH in 15% (two of 13 informative cases) and 54%
(seven of 13 informative cases) in the same dysplasias and
cancers, respectively. Of note, the regions of LOH within the
dysplastic lesions were similar to those within tumors and
included three at 3p25.1–26.1, one at 3p21.1–21.2 and one at
3p13–14.2.
Tobacco, alcohol and LOH
Fifty-six patients had information available regarding tobacco
and alcohol consumption as well as allelotype assessment. The
relationship between the intensity of tobacco and alcohol
exposure and deletions at any of the three discrete regions on
chromosome 3p is detailed in Tables II and III. Patients were
categorized as either non-users or categorized by increasing
use of each of the two substances. Of the 56 patients, 55 had
information regarding tobacco use and 47 (86%) patients gave
a history of tobacco consumption. Among cigarette smokers,
the probability of LOH within any of the three discrete
1241
S.P.Schantz et al.
Table III. The relationship between alcohol use and LOH within three discrete regions of chromosome 3pa
Map position
p25.1–26.1
Locus (D3S-)
1038
1110
1076
1228
659
p21.1–21.2
p13–14.2
Non-users
Inf.
LOH
15
8
9
17
9
7
1–11 drinks/week
(%)
12–28 drinks/week
艌29 drinks/week
P-valueb
Inf.
LOH
(%)
Inf.
LOH
(%)
Inf.
LOH
(%)
(53)
5
3
(60)
3
3
(100)
12
8
(67)
0.521
(100)
(41)
4
5
1
3
(25)
(60)
2
3
1
2
(50)
(67)
6
13
2
6
(33)
(46)
0.004
0.851
Inf, the number of tumors which were informative for that particular loci.
aLOH is defined as a loss of informative loci within the indicated region. Results were generated from the analysis of cancers from 54 patients and each
informative case represents an individual patient.
bP-value was determined by Fisher’s exact test (two-tail).
Table IV. The relationship of whole-arm loss of chromosome 3p in head and neck cancers to tobacco and alcohol usea
Tobacco and alcohol use
ORb
Whole-arm loss
No
Yes
Tobacco
Non-users
1–20 cigarettes/day
⬎20 cigarettes/day
7
17
15
1
9
6
1.00
3.71
2.80
Alcohol
Non-users
4–11 drinks/week
12–28 drinks/week
艌29 drinks/week
17
5
4
13
1
2
3
9
1.00
6.80
12.75
11.77
Interaction
Non-users
Smokers only
Alcohol only
Both
2
14
5
17
0
1
1
13
(95% CI)
(0.4–35.0)
(0.3–27.9)
(0.5–91.5)
(1.0–157.1)d
(1.3–105.0)d
–
1.00
3.20
12.24
ORc
1.00
6.47
5.24
1.00
7.58
29.61
19.00
(95% CI)
(0.6–75.5)
(0.4–65.3)
(0.5–124.1)
(1.8–498.2)d
(1.7–219.5)d
–
(0.2–61.0)
(1.4–104.6)d
1.00
2.83
28.76
(0.1–68.2)
(2.0–417.6)d
aWhole-arm
bORs were
cORs were
d P ⬍ 0.05
loss is defined as the loss (yes) or incomplete loss (no) of all informative loci on the short arm of chromosome 3 within an analyzed cancer.
not adjusted.
adjusted for age.
by Wald χ2 statistics based on logistic regression model.
regions increased with increasing tobacco use (Table II). The
probability of LOH at a particular region among smokers was
most evident in the most distal region of chromosome 3p, i.e.
3p25.1–26.1. Within this latter group, the probability of a
deletion approximated 91% as compared with 30% probability
in those who were non-users (χ2 for trend ⫽ 4.96, P ⫽ 0.03).
Alcohol history could be obtained in 54 of the 56 patients
and its use was identified in 36 (67%) of these individuals. It
is of note that 83% of these individuals who used alcohol
also gave a history of tobacco use. In contrast to tobacco
consumption, increasing alcohol exposure was not associated
with an increase in the probability of an interstitial deletion at
a particular site (Table III). Indeed, the highest probability of
loss was noted in non-users of alcohol at the 3p21.1–21.2
locus. Losses were identified in all informative patients within
this latter group.
We then assessed for the relationship between whole-arm
loss, i.e. loss of all informative loci on 3p in a respective
individual’s tumor, and alcohol and tobacco exposure.
As seen in Table IV, the most significant relationship between
whole-arm loss and exposure was seen in those individuals who
consumed alcohol. This was most evident in individuals who
consumed more than 12 drinks per week either unadjusted or
after adjusting for age (χ2 ⫽ 6.33; P-value for trend ⫽ 0.01).
1242
When assessing for interactions, the greatest probability of
whole-arm loss was seen in those individuals who used both
tobacco and alcohol, after adjusting for age [odds ratio
(OR) ⫽ 28.76, 95% confidence interval (95% CI) 2.0–417.6;
P ⫽ 0.01] with the greatest probability occurring in the
younger populations. Of note, just as we examined for the
presence of LOH at any of the loci within dysplastic lesions,
we also examined for the presence of whole-arm loss. None
of the dysplastic lesions demonstrated evidence of a wholearm loss of chromosome 3p.
Mutagen sensitivity and LOH
Mutagen-sensitivity values were obtained from 43 HNSCC
patients in whom respective tumors were allelotyped for
LOH. Mutagen-sensitivity results of these 43 patients
had been published previously as part of a larger casecontrol study (16). The use of one break per cell as a cut-off
value to classify individuals as mutagen sensitive or nonsensitive has previously been reported by us (16). Overall, 26
(60%) of these 43 patients demonstrated mutagen sensitivity,
i.e. ⬎1.0 break/cell. The relationship between the expression
of mutagen sensitivity and both the interstitial deletions as
well as whole-arm loss was examined. In individuals whose
tumors demonstrated interstitial deletions, the probability of a
Chromosome 3p losses in head and neck cancers
Table V. The relationship between mutagen sensitivity and LOH within three discrete regions of chromosome 3pa
Map position
Locus (D3S-)
Mutagen-sensitivityb
P-valuec
艋1 b/c
Inf.
p25.1–26.1
p21.1–21.2
p13–14.2
1038
1110
1076
1228
659
⬎1 b/c
LOH
(%)
Inf.
LOH
(%)
9
6
(67)
15
11
(73)
1.000
4
11
2
5
(50)
(46)
11
15
7
8
(64)
(53)
1.000
1.000
Inf, the number of tumors which were informative for that particular loci.
aLOH is defined as a loss of informative loci within the indicated region. Results were generated from the analysis of cancers from 43 patients.
bPatients were categorized as mutagen sensitive by the number of breaks per cell (b/c) induced in peripheral blood lymphocytes by bleomycin (9,11,15).
cP-value was determined by Fisher’s exact test (two-tailed).
Table VI. The relationship of whole-arm loss of chromosome 3p in head and neck cancers to tobacco and alcohol use and mutagen sensitivity
Tobacco and alcohol usea
Whole-arm loss
Mutagen sensitivityb
Total
No
Others
Both
13
12
OR ⫽ 7.04 (1.3–37.9)
P ⫽ 0.020c
Yes
2
13
⬍1.0 b/c
艌1.0 b/c
No
Yes
No
Yes
4
6
OR ⫽ 2.67 (0.2–33.5)
P ⫽ 0.600c
1
4
9
6
OR ⫽ 13.50 (1.3–136.0)
P ⫽ 0.018c
1
9
aBoth
is defined by a combined habit of smoking and drinking, and others represent individuals who used either substance alone or did not use either
substance.
were categorized as mutagen sensitivity by the number of breaks per cell (b/c) induced in peripheral blood lymphocytes by bleomycin (9,11,15).
cFisher’s exact test (two-tailed).
bPatients
deletion at a particular region was increased slightly but not
significantly if the patient expressed mutagen sensitivity. This
was true for all three regions of interstitial deletion (Table V).
Table VI shows the relationship between whole-arm loss
in head and neck cancers and both tobacco and alcohol
exposures and mutagen sensitivity. The greatest probability of
demonstrating whole-arm loss was in those individuals who
used both substances and were mutagen sensitive [OR ⫽ 13.5,
P ⫽ 0.02 by Fisher’s exact test (two-tailed)].
Taken together, the patterns of deletional events on the short
arm of chromosome 3 were dependent on the type of exposure
and the presence of mutagen sensitivity. Increasing alcohol
exposure when combined with tobacco use was more likely
to be associated with whole-arm loss while increasing tobacco
usage alone induced changes related to specific loci. The
presence of mutagen sensitivity within the respective head and
neck cancer patient was more likely to enhance certain genetic
events; specifically, whole-arm loss in those who were heavy
alcohol users. As noted above, the vast majority of this latter
group also smoked cigarettes, suggesting that it was the
combined presence of these three factors which contributed to
the whole-arm loss (Table VI).
Clinical characteristics and allelic losses
Results suggest that two types of mutational events are likely
to occur in head and neck cancer patients after accounting for
tobacco and alcohol use and mutagen sensitivity: (i) allelic
losses limited to specific regions on 3p in patients who used
tobacco and (ii) whole-arm loss on chromosome 3p, i.e. the
loss of all informative loci within an individual patient’s tumor
in those who consumed heavy alcohol (the majority of whom
used tobacco) and who were also mutagen sensitive. It should
be noted again that though losses at particular regions could
be identified in dysplastic mucosa adjacent to a particular
cancer, no dysplastic lesion demonstated whole-arm loss.
The clinical characteristics of those who demonstrated
whole-arm loss versus those with either partial or no loss was
assessed. As noted above, patients with whole-arm loss tended
to be younger, i.e. more likely ⬍50 years of age [six of 16
individuals (38%)] than those with partial or no losses [five
of 40 individuals (12%)] (χ2 ⫽ 5.38, P ⬍ 0.05). No other
significant relationship could be identified when assessing for
either of these mutational events with patient sex and race.
When assessing for significant tumor characteristics, patients
with whole-arm loss were more likely to present with advanced
stage disease than the remaining population (Table VII).
Indeed, no cases of stage one disease could be identified in
the group with whole-arm loss. Patients with whole-arm loss,
in contrast to those with partial or no losses, were also more
likely to present with clinical evidence of cervical lymph node
metastases (P ⬍ 0.05).
Discussion
The identification of three discrete regions of loss on the short
arm of chromosome 3, the frequency of these lesions, as well
as the timing, i.e. their presence within dysplastic mucosa
adjacent to invasive head and neck cancer, are mostly similar
1243
S.P.Schantz et al.
Table VII. Head and neck cancer disease stage and its relationship with
whole-arm loss on chromosome 3p
Stage
I
II
III
IV
Whole-arm lossa
Yes
(%)
No
(%)
0
4
7
5
(0)
(25)
(44)
(31)
13
8
8
11
(33)
(20)
(20)
(27)
aThe
number of individuals whose respective cancer did (yes) or did not
(no) show the indicated loss as a function of disease stage. Results are
expressed as the percentage individuals within the categorical grouping. The
patients with whole-arm loss presented less frequently with early stage
disease; P ⬍0.05 by Fisher’s exact test (two-tailed).
in this study with that of previous reports (26–30). Some
variability in frequency can be identified, however, when
examining particular loci. For instance, though Maestro et al.
(26) found a similarly high rate of loss at 3p25 using the
microsatellite marker D3S1038 (53 versus 78% in our analysis,
respectively), Wu et al. found somewhat lower rates (19%)
(29). The results of our study benefited from a larger patient
sample in order to determine areas of loss (58 cases, over
twice that used in previous reports). Furthermore, as we will
describe below, our larger patient sample size was important
when considering the influence of identifiable host and
environmental factors such as mutagen sensitivity, tobacco
and alcohol use. These variables had not been accounted for
in previous studies and will influence results.
Also similar to other reports, we identified the presence of
3p allelic losses within dysplastic lesions adjacent to the
invasive cancer suggesting an early event in the carcinogenic
process (29–31). The frequency of loss in our dysplastic lesions
were lower, however, than those reported previously (29–31).
This may reflect the degree of dysplasia as none of our samples
demonstrated severe epithelial dysplasia or carcinoma-in-situ.
The report of Roz et al. (30), for instance, demonstrated LOH
within 3p regions to be as high as 67% in carcinoma in situ
and as low as 33% in less-severe intraepithelial lesions. The
probability of allelic loss anywhere on 3p was also lower
within intraepithelial tumors in our study as compared with
the matched invasive tumors (28 versus 68%, respectively).
The probability of invasive disease is unlikely to be influenced
by any single allelic loss. Rather, accumulated damage to
multiple genes on the chromosome 3 short arm may be a
prerequisite. Only eight of the 58 cancer samples had loss
confined to a single region and these eight cases had losses
equally distributed throughout the three identifiable regions
(Figure 1). Indeed, in support of this concept as we will later
discuss, the most aggressive disease was found in patients
whose cancers demonstrated losses of all informative loci.
As mentioned above, one critical part of this investigation
was to expand past analyses to account for the influence of
tobacco and alcohol use to account for chromosome 3p allelic
losses. Using data gained from questionnaires in a prospective
case–control study (16), we have shown that the probability
of loss within any of the three discrete regions increased with
increasing tobacco use. The region most significantly affected
was 3p25.1–26.1 in which LOH increased ⬎2-fold as compared
with non-smokers. Given that most tumors which show allelic
losses at at these three regions are tobacco-related (32–36), it
is not surprising that a quantitative relationship should be
1244
identified here. An unexpected finding in this study regarding
chromosomal 3p damage, however, related to alcohol exposure.
Interestingly, damage associated with alcohol was not site
specific but related to 3p losses in cancers which involved all
informative loci, a process we termed whole-arm loss. Alcohol
has been considered to contribute to the carcinogenic process
by principally promotional means and not through a direct
carcinogenic effect (37). Of note, the vast majority of alcohol
users in this study, likewise, used tobacco. Thus, any process
other than promotional cannot be inferred. A limitation in this
study, however, is the number of patients analyzed. Categorical
groupings comprised of small populations increase the risk of
random findings.
A particular focus of this study was the influence of mutagen
sensitivity on chromosomal 3p deletional events. Significantly,
the probability of the combined tobacco- and alcohol-related
deletional losses was further increased in those patients who
expressed this sensitivity phenotype. The ORs of whole-arm
loss in those individuals who used both tobacco and alcohol
increased nearly 6-fold in those individuals who were mutagen
sensitive as compared with non-sensitive individuals. It is
relevant that Dave et al. (21) examined for specificity of
bleomycin-induced chromosomal damage within peripheral
blood lymphocytes from various populations including patients
with head and neck cancer. Bleomycin was shown to induce
breaks within all chromosomes. However, the short arm of
chromosome 3 was most commonly effected. Thus, our results
provide additional evidence, when viewed in concert with the
study by Dave et al. (21), to suggest that 3p may have a
constituitively expressed predisposition site for mutations
within head and neck cancer patients. Distinct from the study
of Dave et al. in which 3p21 was most often affected, we
found no allele that was more commonly involved in mutagensensitive individuals. It should be emphasized that limited
markers were utilized at the 3p21 region; only one at 3p21.3.
Caution should thus be taken in interpreting these results.
Furthermore, more direct evidence to support the presence of
constitutively expressed fragility sites would have to come by
assessing site-specific damage induced within chromosomes
of peripheral blood lymphocytes from the respective patients.
Are the patients with losses at this 3p chromosomal site within
their tumor also the same patients who show chromatid breaks
following bleomycin exposure? Likewise, further evidence that
this represents a host site predisposition to free radical oxygen
damage would come from analyzing for site-specific damage
within normal mucosal cells from the upper aerodigestive tract
in a method similar to Dave et al. (21). These studies remain
to be performed. Whether or not sensitivity to bleomycin is a
heritable host susceptibity factor to tobacco-induced cancer
also needs to be established. In support of that contention, Li
et al. (38) have reported that first degree relatives of mutagensensitive head and neck cancer patients also show sensitivity
to bleomycin. Bondy et al. (36) have related that the risk of
cancer among first degree relatives is also increased in those
head and neck cancer patients who are mutagen sensitive.
The influence of mutagen-sensitivity assessments on the
frequency of alcohol-related 3p deletions provides a potential
clue as to underlying mechanisms. For instance, in vitro studies
by Hsu et al. (39) have shown that the repair of free radical
chromosomal damage is impaired in the presence of alcohol,
which could account for our findings. The end result would
be increased clastogenic processes in individual patients,
dependent upon level of tobacco exposure, a process which
Chromosome 3p losses in head and neck cancers
would be enhanced by mechanisms which relate to decreased
DNA repair, such as potentially reflected by alcohol use and
the mutagen-sensitivity assay. The greater the clastogenic
effect, the more likely the end result would be the loss of
an entire chromosomal arm. Our two previous case–control
epidemiologic studies would support this interaction as a
contributing influence to head and neck cancer development
(13,16). These two separate studies demonstrated that the OR
of head and neck cancer was most influenced by the presence
of both alcohol exposure and mutagen sensitivity. The OR of
disease increased ⬎40-fold in those who both consumed
alcohol and were mutagen sensitive, and indicated a multiplicative effect. Either variable alone was associated with an
OR of disease which did not exceed 7-fold.
Numerous genes relevant to carcinogenesis have been
identified within the three identified regions of loss on 3p,
including genes involved in growth factor control, DNA repair
and cell-cycle regulation (40–45). It would stand to reason
that the loss of all these genetic regulatory elements would
contribute to a more progressive tumorigenic process. Our
results would support that contention. It is of note that the von
Hippel–Lindau tumor suppressor gene, localized at chromosome 3p25–26 and responsible for tumorigenesis in clear
cell renal cell carcinoma (46), is not likely involved in the
pathogenesis of HNSCC (47). Mutagen-sensitive patients with
heavy environmental exposures were more likely to lose the
entire short arm of chromosome 3. The loss of the entire short
arm was associated with a greater probability of presenting
with advanced stage disease and clinically evident lymph node
metastases as compared with those whose losses were limited
to a specific locus, such as 3p21. Also in contrast to both our
results here as well as previous findings involving allelic losses
at 3p21, none of the dysplastic mucosa adjacent to invasive
cancers showed evidence of whole-arm loss (30). It would
suggest that intensive levels of environmental exposures,
potentially influenced by host susceptibility factors, would
lead to a more rapidly progressive process, one in which early
disease is unlikely to be identified. The clinical relevance of
this hypothesis may relate to the problem of head and neck
cancer among certain ethnic populations. Numerous studies
have suggested that head and neck cancer is a worse disease
among black males (48–52). Previous epidemiologic studies
have suggested that this population can also be characterized
by both more intensive alcohol exposure as well as a greater
probability of expressing mutagen sensitivity (48,52). Further
research should confirm more direct evidence of whole-arm
losses within head and neck cancers, such as through the use
of comparative genomic hybridization analysis, and relate
findings to environmental exposures and other ethnic and
socioeconomic factors. The identification of a more virulent
disease influenced by these factors may impact upon screening
strategies as well as the use of multimodalities therapies
designed to improve quality of life as well as survival.
Acknowledgement
This work was partially supported by National Cancer Institute Grant RO1
CA-57155
References
1. Blot,W.J., McLaughlin,J.K., Winn,D.M., Austin,D.F., Greenberg,R.S.,
Preston-Martin,S., Bernstein,L., Schoenberg,J.B., Stemhagen,A. and
Fraumeni,J.F.Jr (1988) Smoking and drinking in relation to oral and
pharyngeal cancer. Cancer Res., 48, 3282–3287.
2. Field,J.K., Spandidos,D.A., Malliri,A., Gosney,J.R., Yiagnisis,M. and
Stell,P.M. (1991) Elevated P53 expression correlates with a history of
heavy smoking in squamous cell carcinoma of the head and neck. Br. J.
Cancer, 64, 573–577.
3. Boyle,J.O., Hakim,J., Koch,W., van der Riet,P., Hruban,R.H., Roa,R.A.,
Correo,R., Eby,Y.J., Ruppert,J.M. and Sidransky,D. (1993) The incidence
of p53 mutations increases with progression of head and neck cancer.
Cancer Res., 53, 4477–4480.
4. Somers,K.D., Merrick,M.A., Lopez,M.E., Incognito,L.S., Schechter,G.L.
and Casey,G. (1992) Frequent p53 mutations in head and neck cancer.
Cancer Res., 52, 5997–6000.
5. Koch,W.M., Lango,M., Sewell,D., Zahurak,M. and Sidransky,D. (1999)
Head and neck cancer in nonsmokers: a distinct clinical and molecular
entity. Laryngoscope, 109, 1544–1551.
6. Mazur,M. and Glickman,B.W. (1988) Sequence specificity of mutations
induced by benzo[a]pyrene-7,8-diol-9,10-epoxide at endogenous aprt gene
in CHO cells. Somat. Cell Mol. Genet., 14, 393–400.
7. Kok,K., Naylor,S.L. and Buys,C.H. (1997) Deletions of the short arm of
chromosome 3 in solid tumors and the search for suppressor genes. Adv.
Cancer Res., 71, 27–92.
8. Sozzi,G., Sard,L., De Gregorio,L., Marchetti,A., Musso,K., Buttitta,F.,
Tornielli,S., Pellegrini,S., Veronese,M.L., Manenti,G., Incarbone,M.,
Chella,A., Angeletti,C.A., Pastorino,U., Huebner,K., Bevilaqua,G.,
Pilotti,S., Croce,C.M. and Pierotti,M.A. (1997) Association between
cigarette smoking and FHIT gene alterations in lung cancer. Cancer Res,
57, 2121–2123.
9. Gonzalez,M.V., Pello,M.F., Ablanedo,P., Suarez,C., Alvarez,V. and Coto,E.
(1998) Chromosome 3p loss of heterozygosity and mutation analysis of
the FHIT and beta-cat genes in squamous cell carcinoma of the head and
neck. J. Clin. Pathol., 51, 520–524.
10. Hsu,T.C. (1983) Genetic instability in the human population: a working
hypothesis. Hereditas, 98, 1–9.
11. Schantz,S.P. and Hsu,T.C. (1989) Mutagen-induced chromosome fragility
within peripheral blood lymphocytes of head and neck cancer patients.
Head Neck, 11, 337–342.
12. Cloos,J., Nieuwenhuis,E.J., Boomsma,D.I., Kuik,D.J., van der Sterre,M.L.,
Arwert,F., Snow,G.B. and Braakhuis,B.J. (1999) Inherited susceptibility
to bleomycin-induced chromatid breaks in cultured peripheral blood
lymphocytes. J. Natl Cancer Inst., 91, 1125–1130.
13. Spitz,M.R., Fueger,J.J., Beddingfield,N.A., Annegers,J.F., Hsu,T.C.,
Newell,G.R. and Schantz,S.P. (1989) Chromosome sensitivity to
bleomycin-induced mutagenesis, an independent risk factor for upper
aerodigestive tract cancers. Cancer Res., 49, 4626–4628.
14. Spitz,M.R., Fueger,J.J., Halabi,S., Schantz,S.P., Sample,D. and Hsu,T.C.
(1993) Mutagen sensitivity in upper aerodigestive tract cancer: a casecontrol analysis. Cancer Epidemiol. Biomarkers Prev., 2, 329–333.
15. Cloos,J., Braakhuis,B.J., Steen,I., Copper,M.P., de Vries,N., Nauta,J.J.
and Snow,G.B. (1994) Increased mutagen sensitivity in head-and-neck
squamous-cell carcinoma patients, particularly those with multiple primary
tumors. Int. J. Cancer, 56, 816–819.
16. Schantz,S.P., Zhang,Z.F., Spitz,M.S., Sun,M. and Hsu,T.C. (1997) Genetic
susceptibility to head and neck cancer: interaction between nutrition and
mutagen sensitivity. Laryngoscope, 107, 765–781.
17. Cloos,J., Spitz,M.R., Schantz,S.P., Hsu,T.C., Zhang,Z.F., Tobi,H.,
Braakhuis,B.J. and Snow,G.B. (1996) Genetic susceptibility to head and
neck squamous cell carcinoma. J. Natl Cancer Inst., 88, 530–535.
18. Li,A.T., Wang,T.D. and Yang,R.T. (1994) Pingyangomycin-induced
chromosome damage in lymphocyes of laryngeal cancer patients and
healthy control subjects. Head Neck, 16, 510.
19. Pandita,T.K. and Hittelman,W.N. (1995) Evidence of a chromatin basis
for increased mutagen sensitivity associated with multiple primary
malignancies of the head and neck. Int. J. Cancer, 61, 738–743.
20. Schantz,S.P., Spitz,M.R. and Hsu,T.C. (1990) Mutagen sensitivity in
patients with head and neck cancers: a biologic marker for risk of multiple
primary malignancies. J. Natl Cancer Inst., 82, 1773–1775.
21. Dave,B.J., Hsu,T.C., Hong,W. and Pathak,S. (1994) Nonrandom distribution
of mutagen-induced chromosome breaks in lymphocytes of patients with
different malignancies. Int. J. Oncol., 5, 733–740.
22. Spitz,M.R., Fueger,J.J., Borrud,L.G. and Newell,G.R. (1988) The
development of a comprehensive, institution-based patient risk evaluation
program: I. Development content and data management. Am. J. Prev.
Med., 4, 183–187.
23. Hsu,T.C., Johnston,D.A., Cherry,L.M., Ramkissoon,D., Schantz,S.P.,
Jessup,J.M., Winn,R.J., Shirley,L. and Furlong,C. (1989) Sensitivity to
genotoxic effects of bleomycin in humans: possible relationship to
environmental carcinogenesis. Int. J. Cancer, 43, 403–409.
1245
S.P.Schantz et al.
24. van der Riet,P., Nawroz,H., Hruban,R.H., Corio,R., Tokino,K., Koch,W.
and Sidransky,D. (1994) Frequent loss of chromosome 9p21–22 early in
head and neck cancer progression. Cancer Res., 54, 1156–1158.
25. Wistuba,I.I., Montellano,F.D., Milchgrub,S., Virmani,A.K., Behrens,C.,
Chen,H., Ahmadian,M., Nowak,J.A., Muller,C., Minna,J.D. and
Gazdar,A.F. (1997) Deletions of chromosome 3p are frequent and early
events in the pathogenesis of uterine cervical carcinoma. Cancer Res., 57,
3154–3158.
26. Maestro,R., Gasparotto,D., Vukosavljevic,T., Barzan,L., Sulfaro,S. and
Boiocchi,M. (1993) Three discrete regions of deletion at 3p in head and
neck cancers. Cancer Res., 53, 5775–5779.
27. el-Naggar,A.K., Lee,M.S., Wang,G., Luna,M.A., Goepfert,H. and
Batsakis,J.G. (1993) Polymerase chain reaction-based restriction fragment
length polymorphism analysis of the short arm of chromosome 3 in
primary head and neck squamous carcinoma. Cancer, 72, 881–886.
28. Ah-See,K.W., Cooke,T.G., Pickford,I.R., Soutar,D. and Balmain,A. (1994)
An allelotype of squamous carcinoma of the head and neck using
microsatellite markers. Cancer Res., 54, 1617–1621.
29. Wu,C.L., Sloan,P., Read,A.P., Harris,R. and Thakker,N. (1994) Deletion
mapping on the short arm of chromosome 3 in squamous cell carcinoma
of the oral cavity. Cancer Res., 54, 6484–6488.
30. Roz,L., Wu,C.L., Porter,S., Scully,C., Speight,P., Read,A., Sloan,P. and
Thakker,N. (1996) Allelic imbalance on chromosome 3p in oral dysplastic
lesions: an early event in oral carcinogenesis. Cancer Res., 56, 1228–1231.
31. el-Naggar,A.K., Hurr,K., Batsakis,J.G., Luna,M.A., Goepfert,H. and
Huff,V. (1995) Sequential loss of heterozygosity at microsatellite motifs
in preinvasive and invasive head and neck squamous carcinoma. Cancer
Res., 55, 2656–2659.
32. Latif,F., Fivash,M., Glenn,G., Tory,K., Orcutt,M.L., Hampsch,K., Delisio,J.,
Lerman,M., Cowan,J., Beckett,M. et al. (1992) Chromosome 3p deletions
in head and neck carcinomas: statistical ascertainment of allelic loss.
Cancer Res., 52, 1451–1456.
33. Meltzer,S.J. (1996) The molecular biology of esophageal carcinoma. Rec.
Res. Cancer Res., 142, 1–8.
34. Brauch,H., Johnson,B., Hovis,J., Yano,T., Gazdar,A., Pettengill,O.S.,
Graziano,S., Sorenson,G.D., Poiesz,B.J., Minna,J. et al. (1987) Molecular
analysis of the short arm of chromosome 3 in small-cell and non-smallcell carcinoma of the lung. N. Engl. J. Med., 317, 1109–1113.
35. Kok,K., Osinga,J., Carritt,B., Davis,M.B., van der Hout,A.H., van der
Veen,A.Y., Landsvater,R.M., de Leij,L.F., Berendsen,H.H., Postmus,P.E.
et al. (1987) Deletion of a DNA sequence at the chromosomal region
3p21 in all major types of lung cancer. Nature, 330, 578–581.
36. Bondy,M.L., Spitz,M.R., Halabi,S., Fueger,J.J., Schantz,S.P., Sample,D.
and Hsu,T.C. (1993) Association between family history of cancer and
mutagen sensitivity in upper aerodigestive tract cancer patients. Cancer
Epidemiol. Biomarkers. Prev., 2, 103–106.
37. Blot,W.J. (1992) Alcohol and cancer. Cancer Res., 52 (suppl.), 2119s–
2123s.
38. Li,A.T., Wang,T.D. and Yang,R.F. (1996) High sensitivity to mutagens in
healthy blood relatives of laryngeal and hypopharyngeal cancer patients.
In Werner,J.A. and Lippert,B.M. (eds) Head and Neck Cancer—Advances
in Basic Research. Elsevier Science, Amsterdam, The Netherlands, pp. 3–7.
1246
39. Hsu,T.C., Furlong,C. and Spitz,M.R. (1991) Ethyl alcohol as a cocarcinogen
with special reference to the aerodigestive tract: a cytogenetic study.
Anticancer Res., 11, 1097–1101.
40. Killary,A.M., Wolf,M.E., Giambernardi,T.A. and Naylor,S.L. (1992)
Definition of a tumor suppressor locus within human chromosome 3p21p22. Proc. Natl Acad. Sci. USA, 89, 10877–10881.
41. LaForgia,S., Morse,B., Levy,J., Barnea,G., Cannizzaro,L.A., Li,F.,
Nowell,P.C., Boghosian-Sell,L., Glick,J., Weston,A. et al. (1991) Receptor
protein-tyrosine phosphatase gamma is a candidate tumor suppressor gene
at human chromosome region 3p21. Proc. Natl Acad. Sci. USA, 88,
5036–5040.
42. Gebert,J.F., Moghal,N., Frangioni,J.V., Sugarbaker,D.J. and Neel,B.G.
(1991) High frequency of retinoic acid receptor beta abnormalities in
human lung cancer [published erratum appears in Oncogene (1992), 7,
821]. Oncogene, 6, 1859–1868.
43. Mathew,S., Murty,V.V., Cheifetz,S., George,D., Massague,J. and
Chaganti,R.S. (1994) Transforming growth factor receptor gene TGFBR2
maps to human chromosome band 3p22. Genomics, 20, 114–115.
44. Papadopoulos,N., Nicolaides,N.C., Wei,Y.F., Ruben,S.M., Carter,K.C.,
Rosen,C.A., Haseltine,W.A., Fleischmann,R.D., Fraser,C.M., Adams,M.D.
et al. (1994) Mutation of a mutL homolog in hereditary colon cancer.
Science, 263, 1625–1629.
45. Virgilio,L., Shuster,M., Gollin,S.M., Veronese,M.L., Ohta,M., Huebner,K.
and Croce,C.M. (1996) FHIT gene alterations in head and neck squamous
cell carcinomas. Proc. Natl Acad. Sci. USA, 93, 9770–9775.
46. Decker,H.J., Weidt,E.J. and Brieger,J. (1997) The von Hippel–Lindau
tumor suppressor gene. A rare and intriguing disease opening new insight
into basic mechanisms of carcinogenesis. Cancer Genet. Cytogenet., 93,
74–83.
47. Waber,P.G., Lee,N.K. and Nisen,P.D. (1996) Frequent allelic loss at
chromosome arm 3p is distinct from genetic alterations of the Von-Hippel
Lindau tumor suppressor gene in head and neck cancer. Oncogene, 12,
365–369.
48. Day,G.L., Blot,W.J., Austin,D.F., Bernstein,L., Greenberg,R.S., PrestonMartin,S.,
Schoenberg,J.B.,
Winn,D.M.,
McLaughlin,J.K.
and
Fraumeni,J.F.Jr (1993) Racial differences in risk of oral and pharyngeal
cancer: alcohol, tobacco and other determinants. J. Natl Cancer Inst., 85,
465–473.
49. Devesa,S.S., Blot,W.J. and Fraumeni,J.F.Jr (1990) Cohort trends in
mortality from oral, esophageal and laryngeal cancers in the United States.
Epidemiology, 1, 116–121.
50. Slotman,G.J., Swaminathan,A.P. and Rush,B.F.Jr (1983) Head and neck
cancer in a young age group: high incidence in black patients. Head Neck
Surg., 5, 293–298.
51. Boring,C.C., Squires,T.S. and Health,C.W.Jr (1992) Cancer statistics for
African Americans [published erratum appears in CA Cancer J. Clin.
(1992), 42, 125]. CA Cancer J. Clin., 42, 7–17.
52. Spitz,M.R., Hsu,T.C., Wu,X., Fueger,J.J., Amos,C.I. and Roth,J.A. (1995)
Mutagen sensitivity as a biological marker of lung cancer risk in African
Americans. Cancer Epidemiol. Biomarkers Prev., 4, 99–103.
Received December 9, 1999; revised February 17, 2000;
accepted February 29, 2000