| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Advances in Brief |
Laboratory of Head and Neck Cancer Biology, The University of Michigan, Ann Arbor, Michigan 48109-0506 [S. T., T. O., K. Y. J., A. M., T. E. C.]; Department of Otorhinolaryngology, Hamamatsu University School of Medicine, Hamamatsu 431-3192, Japan [H. M.]; Department of Obstetrics and Gynecology, Loyola University, Chicago, Illinois 60153 [S. G. F.]; and Department of Otorhinolaryngology-Head and Neck Surgery, Department of Medical Biochemistry, Turku University, FIN-20521 Turku, Finland [R. G.]
 |
ABSTRACT |
|---|
 Top ABSTRACT IntroductionMaterials and MethodsResultsDiscussionREFERENCES |
|---|
|
Introduction |
|---|
|
TopABSTRACTIntroductionMaterials and MethodsResultsDiscussionREFERENCES |
|---|
60%), and 21 (52%; Refs. 1
and
2
). The high frequency of loss affecting 18q caused us to
take special interest in this region. In addition to HNSCC, LOH on 18q
is implicated in several other tumor types, notably colon carcinoma
(3)
, but also in non-small cell lung carcinoma, prostate
carcinoma, esophageal squamous cell carcinoma, and transitional cell
carcinoma of the bladder. This frequent loss suggests that one or more
genes on 18q may be important in the genesis or progression of these
tumors. Evidence that 18q LOH is associated with poor survival in HNSCC
(4
, 5) and other tumor types (6)
supports
this concept. We reported previously that 18q LOH is more common in cell lines and tumors from recurrent or metastatic HNSCCs than in primary tumors or in cell lines derived from the primary tumor, suggesting a role in tumor progression (7) . That study also showed that tumor cell lines accurately represent the status of 18q in the tumor tissue from which they were derived. This strengthens the value of using tumor cell lines to characterize the minimal regions of loss. Furthermore, the absence of normal tissue or cells contaminating the tumor cell population in cultured lines reduces the ambiguity in evaluating LOH. In addition, tumor cell lines that have been characterized for chromosome deletion provide useful models for assessing gene inactivation and functional evaluation of candidate TSGs. An example of how effective cell lines can be for this type of study is the use of human pancreatic carcinoma cell lines and pancreatic tumor xenografts in nude mice to define the minimal region of loss on 18q that affects DPC4/Smad4 in carcinoma of the pancreas (8) .
Although chromosome aberrations have been landmarks to identify cancer genes in many tumor types, the mechanisms of altered gene expression in tumors cannot be deduced solely from the type of chromosome rearrangement. Translocations and breakpoints are often associated with activation of growth-promoting genes, whereas chromosome losses are associated with inactivation of growth regulatory or TSGs. Thus, the breakpoint in a deleted chromosome could be important if a nearby gene is consistently affected by the rearrangement or alternatively, the most commonly lost region may indicate the locus of a gene for which both copies are inactivated during tumor progression. The tumor suppressor hypothesis proposes that if one allele has a mutation and the other is lost or inactivated, the tumor suppressor function of the mutated gene will also be lost. Depending on the role of the affected gene, this may lead to cell cycle deregulation and increased growth capacity or perhaps to increased propensity for metastasis. Whatever the regulatory function of a TSG may be, its inactivation is likely to increase the aggressive behavior of the new clone. However, until the important gene affected by the deleted region is known, it will not be possible to understand the mechanism by which this genetic event contributes to the development and progression of the tumor. The definition of common breakpoints and MLRs is a necessary first step to localize and identify the 18q gene(s) involved.
Therefore, in this study a large number of microsatellite markers spanning 18q were systematically analyzed to assess breakpoints and MLRs in a large series of well-characterized HNSCC cell lines for which normal cells were available as a source of genomic DNA.
|
Materials and Methods |
|---|
|
TopABSTRACTIntroductionMaterials and MethodsResultsDiscussionREFERENCES |
|---|
DNA was extracted by using the Nucleon II extraction kit (Scotlab, Shelton, CT), according to the manufacturer’s instructions, except that silica beads were not used. Instead, proteins were removed by chloroform extraction, after which DNA was precipitated from the aqueous phase by ethanol and ammonium acetate solution. Cultured cells were trypsinized and collected by centrifugation, and then 2 ml of reagent B (400 mM Tris-HCl, 60 mM EDTA, 150 mM NaCl, and 1% SDS) were added and mixed for 2 h for cell lysis. Five hundred µl of 5 M sodium perchlorate were added and mixed for 1 h for deproteinization. Two ml of chloroform stored at -20°C were added and mixed for 10 min and then centrifuged at 2300 rpm for 5 min at 4°C. Only the upper aqueous phase was carefully transferred into new tube, to which four equal volumes of ice-cold ammonium acetate/ethanol solution were added. The tubes were inverted gently and kept at -20°C overnight to precipitate DNA. The samples were centrifuged at 4000 rpm for 5 min at 4°C to pellet the DNA, and the supernatant was discarded. The pellets were resuspended and washed twice with 1 ml of cold 70% ethanol (4°C), after which they were air dried for 15 min and dissolved in 100 µl of Tris-EDTA buffer.
Microsatellite Markers and PCR Amplification.
Forty-two highly polymorphic microsatellite markers on 18q spaced less
than 5 cM apart (mean distance, 1.87 cM) from 18q11.1
(D18S44) to 18q23 (D18S1141) were used to detect
LOH on 18q. D18S71 (18p11.21) was used to detect loss on the
p arm. All of the readily available polymorphic microsatellite markers
mapped to 18q that have heterozygosity scores >50% were used for this
study. All primer sets were purchased from Research Genetics
(Huntsville,
AL).6
The locations of genes mapped to 18q were also obtained from the
Genetic Location Database, except for MADR2/Smad2
(10)
, SCCA1, and SCCA2
(11)
.
PCR amplification of 50–200 ng of genomic DNA was carried out in a reaction mixture consisting of 1.5 pmol of sense primer 32P-end labeled by T4 polynucleotide kinase (Life Technologies, Inc., Gaithersburg, MD), 1.5 pmol of unlabeled antisense primer, 150 µM each deoxynucleotide triphosphate, 10x PCR buffer, 1.5 mM MgCl2, 1% DMSO, and 1 unit of Taq polymerase (Life Technologies, Inc.) with Taq start antibody (Clontech, Palo Alto, CA). The 10x PCR buffer contained 200 mM Tris-HCl (pH 8.4) and 50 mM KCl. Fourteen µl of the PCR mixture for each sample and 2 µl of genomic DNA were overlaid with mineral oil. Using the optimal annealing temperature for each primer set, the PCR mixture and genomic DNA were denatured at 94°C for 3 min and amplified in a Perkin-Elmer Thermal Cycler (Norwalk, CT) for 30 cycles as follows: 94°C for 20 s, 49–58°C for 30 s, and 72°C for 40 s, followed by a final 5-min extension at 72°C. Ten µl of stopping solution (95% formamide, 10 mM EDTA, 0.1% bromphenol blue, and 0.1% xylene cyanol) were added to each PCR product. The PCR products were denatured at 94°C for 3 min and loaded onto prewarmed 6% bis-acrylamide denaturing gels containing 7 M urea and electrophoresed at 115 W and 47°C for 1.5–5 h, depending on the expected PCR product size. The gels were transferred to Whatman paper, dried at 80°C for 1 h, and then autoradiographed with intensifying screens.
|
Results |
|---|
|
TopABSTRACTIntroductionMaterials and MethodsResultsDiscussionREFERENCES |
|---|
shows all of the markers used for the LOH analysis and a summary of the
results. On the left side of the table, each of the polymorphic markers
are listed with their relative positions on 18q (from 18p at the top to
18qter at the bottom) and the heterozygosity score for each. The
locations of several known genes are indicated in the center portion of
the table. The relative location of each marker and gene and the most
likely cytogenetic band for each was determined from the Genetic
Location Database composite map constructed under a priority rule in
which physical > radiation hybrid > genetic linkage map > cytogenetic assignment was used.
In subsequent columns, for each marker the expected and observed rate
of informative cases among the 57 cases we tested are listed followed
by the number of cases with NL, LOH, and AI. In our sample a
significantly lower or higher rate of heterozygosity/informatives was
demonstrated in 5 of the 41 markers (P < 0.001). Two markers, D18S44 and D18S877 (marked
with two asterisks), exhibited a higher observed than expected rate of
heterozygosity in our sample. Those with lower than expected rate of
heterozygosity, D18S470, D18S364, and
D18S977, are marked with a single asterisk in Table 1
. All
others did not differ significantly from the reported rates. Such
differences most likely reflect differences in the genetic
characteristics of the study population and the populations used to
determine the relative frequency of these alleles. In the last two
columns, the number of cases with combined LOH + AI and the
LOH frequency (which includes the cell lines with AI for individual
markers) are given. Data rows for the three loci with LOH + AI >70% of cases are shaded. Although tumor cell lines were used in
this analysis, we observed some cases in which the products from the
two alleles were not equal. AI in the tumor cell lines was interpreted
as LOH in most cases. The rationale for this interpretation is that
with balanced heterozygosity (NL) at some loci and AI at others, it is
likely that these tumor cell lines contained one population of cells
with NL and another population that has lost the allele on one copy of
chromosome 18. To score a locus as AI = LOH required
retention of both alleles at least two other loci to help rule out the
likely alternative explanation that the observed AI was attributable to
duplication of one homologue. However, it is not possible to rule out
local amplification of small regions with the methods we used. Examples
of LOH and AI at the most commonly deleted markers are illustrated in
Fig. 1
. This figure shows representative examples of the PCR results for
UM-SCC-19, UT-SCC-4, UM-SCC-91, and UM-SCC67 at the D18S39,
D18S61, D18S70, and D18S1122
loci. Results from these cell lines were selected for
illustration because they have very limited LOH and have in common LOH
at D18S70 (Fig. 2)
. D18S1122 is also shown because these tumor lines do not
have loss affecting this locus, which is mapped immediately distal from
D18S70. AI was noted at D18S70 in UM-SCC-91 and
at both D18S39 and D18S70 in UT-SCC-4 (open
arrows). LOH was present in UM-SCC-19 and UM-SCC-67 at
D18S70 (solid arrows) but not at the other loci
shown, although D18S39 and D18S61 were NI in
UM-SCC-19. MI was observed in UT-SCC-4 at D18S1122
(shaded arrow).
|
|
|
shows the detailed LOH map on 18q. The 57 tumor cell lines are
arrayed across the page from left to right, with those having the
greatest degree of 18q loss on the left to least amount of loss on the
right (except for UM-SCC-83A, which will be discussed later). The
markers are arranged from the p arm at the top of the page to the
telomere of the q arm at the bottom. LOH frequency for individual
markers on 18q varied from 13% (D18S44) at 18q11.1 to 75%
(D18S70) at 18q23 (Table 1)
. Forty-three of 57 HNSCCs (75%)
have LOH or AI for at least one locus on 18q. Eight cell lines have
whole-arm loss (five of these probably lost the entire chromosome
because there is also loss on the p arm). Thirty-five cell lines have
partial loss of 18q. Eleven cell lines, UM-SCC-14A, UM-SCC-17A,
UM-SCC-19, UM-SCC-54, UM-SCC-73A, UM-SCC-77, UM-SCC-83A, UM-SCC-99,
UT-SCC-4, UT-SCC-16A, and UT-SCC-20A, show interstitial losses or AIs
with retention of two alleles at some loci. Of these lines with
interstitial losses, UM-SCC-54 and UM-SCC-99 are heterozygous for only
a single marker, and UM-SCC-77 retains heterozygosity at only four
markers. Fourteen cell lines (those on the right side of the figure)
have no detectable loss. Nine cell lines exhibit AI (UM-SCC-15,
UM-SCC-42, UM-SCC-73A, UM-SCC-77, UM-SCC-91, UT-SCC-4, UT-SCC-16A,
UT-SCC-20A, and UT-SCC-43A). MSI (indicated in Fig. 2
by an underscore)
was noted in eight cell lines, at least at one locus (UM-SCC-87,
UM-SCC-16, UM-SCC-23, UM-SCC-12, UM-SCC-73A, UM-SCC-83A, UT-SCC-4, and
UT-SCC-46A). Instability was observed at 12 loci in UM-SCC-83A, at 10
loci in UM-SCC-16, at two loci in UM-SCC-12 and UM-SCC-73A, and at only
one locus in each of the other six lines.
For tumor lines with contiguous deleted regions, the breakpoint is the
most proximal LOH marker, with an adjacent more proximal marker
exhibiting NL. The majority of the partial loss cell lines, 22 of 35
(63%) have a breakpoint located between D18S480 (18q11.1)
and D18S57 (18q12.3; Fig. 2
). Two have a breakpoint at the
centromere. Seven have a breakpoint at D18S480, four at
D18S877, two at D18S463, and three at
D18S57.
We identified three MLRs (Fig. 2
, shaded markers), which contain at
least one marker with LOH frequencies of 70% or more. The most
proximal is defined by D18S39 in band 18q21.1 with LOH or AI
in 28 of 38 (74%) informative cases. The second and largest MLR is
bounded by D18S61 in band 18q22.2 (loss in 28 of 40; 70%)
and D18S50 in band 18q23 (loss in 27 of 42; 64%). This MLR
is defined by three cell lines, UT-SCC-20A, UT-SCC-16A, and UM-SCC-12,
which have retained markers at either D18S554 (distal;
UT-SCC-20A and UT-SCC-16A) or D18S392 (proximal; UT-SCC-16A
and UM-SCC-12). The third (and most distal) region is centered on
D18S70 (30 of 40; 75%) in band 18q23. The LOH frequency of
75% at D18S70 is the highest of all 43 alleles tested. In
comparison, at the D18S461 locus, 0.34 cM centromeric from
D18S70, LOH frequency is only 53% (18 of 34), and at the
D18S871 locus, 0.57cM centromeric from D18S70,
LOH frequency is only 52% (14 of 27). Similarly, the frequency of LOH
at D18S1122 and D18S1141 telomeric to
D18S70 is 63 and 66%, respectively. Six tumor cell lines
with minimal losses at other markers (UM-SCC-12, UM-SCC-73A, UM-SCC-19,
UM-SCC-67, UM-SCC-91, and UT-SCC-4; Fig. 2
, bold type) have LOH or AI
at D18S70. UM-SCC-67 has LOH only at D18S70, and
UM-SCC-91 shows AI only at this locus. UM-SCC-19 and UM-SCC-67 each
exhibit LOH at D18S70 with retention of alleles in both the
immediately proximal and distal markers. Likewise, UM-SCC-91 has AI at
D18S70 with retention of heterozygosity at both flanking
alleles. UT-SCC-4 has AI only at the D18S70 locus and at one
of the other MLRs, D18S39. Similarly, UM-SCC-73A has distal
LOH affecting the region surrounding D18S70 and isolated AI
affecting D18S39. UM-SCC-12 has loss that probably
encompasses the entire region from D18S61 to 18qter and
includes the MLRs at both D18S61 and D18S70.
The distance between the retained markers centromeric and telomeric
from the most frequently lost marker determined the relative size of
the MLR in each case. In the case of the MLR centered on
D18S39, the distance between adjacent markers is 1.56 cM.
For the MLR beginning at D18S61, the distance between
adjacent markers is 15.79 cM. This is defined by UT-SCC-16A and
UM-SCC-12, which each have a retained heterozygous allele at
D18S392 on the proximal side, and by UT-SCC-20A and
UT-SCC-16A, which have retained both alleles at D18S554 on
the telomeric side. The third MLR centered on D18S70 spans
3.67 cM (Table 1)
. The pattern of loss between sequential markers (from
centromere to telomere) was examined using a McNemar test for matched
samples. Rates of loss increased significantly between the incremental
markers from D18S44 to D18S57. Although rates of
loss tended to increase along the chromosomal arm, the only other
significant increase in loss was detected between D18S461
(53% loss) and D18S70 (75% loss; P = 0.08). This further supports the concept that LOH at
D18S70 is important.
Of the cell lines with the lowest frequency of allelic loss, only two
cell lines, UM-SCC-83A and UM-SCC-17A shown in Fig. 2
(to the right of
the bold column), exhibit a sporadic interstitial pattern of 18q loss
that does not include any of the three MLRs. UM-SCC-17A has LOH at only
a single marker, D18S977. It is interesting that the tumor
cell lines established from secondary tumors in these patients each
have extensive 18q LOH (7)
, and in both cases the losses
in the secondary tumor lines include the three
MLRs.7
|
Discussion |
|---|
|
TopABSTRACTIntroductionMaterials and MethodsResultsDiscussionREFERENCES |
|---|
For the analysis reported here, tumor cell line DNA was used instead of tumor tissue DNA. DNA isolated from tumor tissue is contaminated with normal cells, which can make it difficult to generate unambiguous data for defining MLRs. The advantage of the cell lines is that there is no contamination by normal cells, which results in very clean estimates of loss in most cases. Furthermore, we showed previously that the tumor cell lines accurately represent the in vivo situation with respect to loss of heterozygosity on 18q (7) . In every case we analyzed, the allele that gave the weaker signal in tumor DNA was the allele that was lost in the cell line DNA (7) . Thus, it is possible to be confident in using well-characterized cell lines established in our own laboratories for this analysis. Although between our combined laboratories HNSCC cell lines have been developed from more than 150 different patients, the availability of normal tissue DNA for LOH analysis limited our study to tumor lines from 57 patients.
It is of interest that we observed AI for some 18q alleles in some tumor cell lines. Because the cell lines had all been carried for >20 passages at the time of DNA extraction, the likelihood that residual normal cells remained in the culture population is small. Therefore, AI indicates that there are two populations of tumor cells that differ with respect to 18q copy number. One alternative is that there are different numbers of each homologue in all of the cells in the tumor lines showing this pattern. If that were the best explanation, then we would expect to see AI at every allele on the portion of the duplicated chromosome. Instead, balanced retention of both alleles was observed at some loci with imbalance at others. From this, it was concluded that a subpopulation of tumor cells with interstitial loss on one chromosome was a more likely explanation. In fact, the latter is consistent with our earlier observation that 18q LOH occurs with progression within the same patient (7) and suggests that in the lines with two populations, the 18q LOH progression step has already begun in the primary tumor population. Furthermore, additional studies of paired sets of A and B cell lines from same donor support this concept. As one example, we observed that regions affected by AI in the primary tumor line UT-SCC 20A appear as LOH in the secondary tumor cell line UT-SCC-20B from the same donor.7
In this study, three MLRs were identified. The most proximal MLR (D18S39) in this study corresponds to the third and most distal of the three regions that Papadimitrakopoulou et al. (12) defined. Our second MLR is distal to this and is defined by D18S61 to D18S50 in 18q22.2–18q23, and the third and most distal MLR we defined is centered on D18S70 in band 18q23. D18S70 has the highest LOH frequency of all 43 alleles tested. Furthermore, LOH or AI affects this locus in 6 of 10 cell lines that have loss at 10 or fewer markers. Interestingly, two of three tumor cell lines with very minimal losses (i.e., loss at five or fewer loci on 18q) exhibit LOH only at D18S70, and four of the five tumors with loss or AI at only one or two loci have loss or AI at D18S70. One cell line exhibited AI only at this locus, and another cell line showed AI at this locus and D18S39. Together, these data suggest that this marker is very close to the important gene locus affected most commonly by loss on 18q. Furthermore, one cell line with minimal interstitial losses, UM-SCC-83A, showed retention of D18S70 but exhibited loss at D18S1122, the next marker telomeric from D18S70. This is consistent with the locus of an important gene falling between these two markers. With respect to the significance of a distal gene locus being the important region, it is of interest that Pearlstein et al. (5) found that LOH at MBP, which is midway between D18S61 and D18S70, was significantly associated with poor survival.
Papadimitrakopoulou et al. (12)
also showed
three proximal MLRs on 18q using 19 polymorphic markers and tumor DNA
from 50 patients. The first MLR they defined had an LOH frequency of
42% and corresponded to D18S470 and D18S474, the
second one (48% LOH) corresponded to D18S1099 and
D18S487, and the last one (42% LOH) corresponded to
D18S39 and D18S41, which we also identified in
our study (Table 1
and Fig. 2
). In the first and second MLRs they
defined, our LOH frequencies are from 53 to 60%, which are not as
frequently lost as are the markers in the three MLRs (74, 70, and 75%)
defined in this report. In their study, twenty-eight of 42 (67%) tumor
DNA samples showed interstitial losses, as opposed to 10 of 57 (18%)
cell line DNA samples in our study. It may be that because our study
used only cell lines that it is easier for us to demonstrate that
losses are contiguous in many cases. An alternative possibility is that
in some cases LOH begins with chromosome replication errors and
multiple regions of interstitial loss that later (i.e., with
progression) converts to whole-arm or additional regions of loss,
resulting in larger contiguous missing regions. Perhaps our cell lines
represent more abundantly than tumor tissue those clones from the tumor
population that have the greatest growth potential and survival
advantage, whereas many less aggressive or even nonviable clones in the
tumor may retain large segments of 18, giving the impression of
segmental loss. If tumors progress from small regions of interstitial
loss to stable populations of cells with whole-arm loss or large
regions of loss, then our cell lines might reflect this progression. In
fact, our previous (7)
and ongoing
analysis7
of primary and secondary tumor cell
lines support this concept. For example, UM-SCC-83A from a primary
tumor shows multiple discontinuous small 18q losses (Fig. 2)
. In
comparison, the UM-SCC-83B cell line from a metastatic tumor in the
same patient exhibits large regions of loss that encompass the small
regions present in the primary tumor line.7
Another factor that may contribute to the discrepancy between this
study and that of Papadimitrakopoulou et al.
(12)
is that it is difficult to unequivocally identify LOH
in tumor tissue DNA because there is a variable degree of normal
cell contamination in each tumor sample. As a result, many more cases
of LOH become judgment calls. Furthermore, the length of exposure of
each autoradiograph becomes a factor in identifying differences between
AI and NL. These factors can lead to a higher estimate of interstitial
loss than what actually occurs. Nevertheless, we did find examples of
small interstitial losses in our tumor cell lines, and some of these
cases helped to define our minimal regions of loss. That both this and
the previous studies found an MLR surrounding D18S39 is
consistent with this region containing a higher frequency of loss than
in other regions, and fine mapping is indicated for this region.
Papadimitrakopoulou et al. (12) , who tested only from D18S34 in 18q12.2 to D18S61 in 18q22.2, and El-Naggar et al. (13) did not find the regions we found centered on D18S61 or D18S70, although each study used one of these markers. We suspect that the difficulty in detecting LOH on clinical specimens is one reason. A second reason could be that their tumor panels contained a greater proportion of early stage tumors than does our set of cell lines, which generally represent more advanced tumors. However, the stage distribution in EL-Naggar et al. (13) is similar to our report by Jones et al. (2) , and this distribution is also representative of the stage distribution for our cell line donors (9) . Presumably, there is a lower frequency of 18q loss in early-stage tumors, as our previous results suggest (7) , although our previous studies also found the same rate of 18q LOH in the tumor samples sent to our laboratory as we have observed in the established cell lines (2) .
Several putative TSGs and other genes implicated in cancer have been
mapped to 18q. These include MADR2/Smad2 (10)
,
DPC4 (MADR4/Smad4; Ref. 8
), DCC
(3)
, SCCA1 and SCCA2
(11)
, and BCL2 (Ref. 14
; which
inhibits programmed cell death), listed in order from most centromeric
to most telomeric. However, none of the putative TSGs (MADR2,
DPC4/MADR4, and DCC) map to the MLRs we identified, and
some of these genes have already been studied in SCC.
Papadimitrakopoulou et al. (12)
used reverse
transcription-PCR and sequencing to examine the expression of
MADR2 in nine HNSCC cell lines. No mutations or
polymorphisms were found, suggesting that this gene is not altered in
most HNSCCs. Similarly, Kim et al. (15)
found
only one example of alteration of DPC4/MADR4 among 16 HNSCC
cell lines, representing primary and secondary tumors from 11 patients
and 20 primary tumors. All cell lines expressed full-length transcripts
by reverse transcription-PCR. Only UM-SCC-22A and UM-SCC-22B contained
a nonsense mutation, and there was one polymorphism in the normal and
tumor tissue of one patient (15)
. Similarly, we
(7)
found homozygous deletion of exon one of
DPC4 in UM-SCC-81B, although this defect was not present in
UM-SCC-81A, a cell line from a primary tumor in the same patient. In
fact, no LOH was found in UM-SCC-81A (Table 1)
. The cell line FaDu also
has homozygous deletion that affects DPC4 and extends into
DCC (16)
. Furthermore, transfer of a wild-type
copy of chromosome 18 reduced the tumorigenic and invasive behavior of
FaDu clones (16)
. Reiss et al.
(16)
concluded that because the parental cells contained
one full-length copy of DCC, the altered behavior was most likely the
result of restoring DPC4 expression. Taken together, these data suggest
that DPC4 may have a role in a small subset of HNSCC. The role of
DCC remains unclear. LOH affecting the DCC locus
is common in HNSCCs (4)
. In esophageal squamous cancers,
point mutations and allelic deletions affecting DCC have
been reported (17)
. Similarly, we have preliminary
evidence indicating that DCC is not expressed in most HNSCC
cell
lines,8
but it is not yet known whether this gene is normally expressed in
normal mucosa or if alterations of DCC are involved in the
development or progression of this tumor type. It was a little
surprising that we did not find any areas of consistent homozygous loss
among 57 tumor cell lines. If the mechanism behind the LOH is loss of
both copies of a particular TSG, then homozygous deletion might occur
in a subset of cases, if markers close enough to the important gene are
used. In pancreatic carcinomas, Hahn et al. (8)
found frequent examples of homozygous loss affecting DPC4
exon 1 in a high frequency of pancreatic carcinoma xenografts, once the
MLR was narrowed sufficiently. Thus, we might see common homozygous
deletion when the critical MLRs in HNSCCs are more narrowly defined.
Alternatively, in HNSCC it may be that there will be a consistent
activating mutation of a TSG that requires only the loss of the
wild-type allele to become an activated tumor growth-promoting gene.
SCCA1 and SCCA2 (11)
encode related
serine protease inhibitors and were originally identified as a squamous
cell antigens that are found circulating in the serum of patients with
advanced squamous cancers. These genes map to the region between
D18S39 and D18S61. At this time, the role of
SCCA1 and SCCA2 in tumor progression, if any, is
unknown. BCL2 also maps to this same region. BCL2
expression is activated in B-cell lymphomas by a t(14;18) chromosome
translocation, which brings it into proximity with the actively
transcribed and expressed immunoglobulin heavy chain gene.
BCL2 encodes a mitochondrial membrane protein that blocks
apoptosis. Although BCL2 protein is strongly expressed in some
HNSCC tumors, it is not known if BCL2 expression is affected by 18q
LOH. Furthermore, it appears that the product of a different
BCL2 gene family member, BCL-xL, is more
frequently overexpressed in HNSCCs, suggesting that BCL2 is
also not a primary target of 18q rearrangement in this tumor type.
The absence of a strong and consistent link to any of the known tumor-related genes on 18q supports the possibility that the most important gene(s) in SCC development and progression may be located at one of the most frequently lost regions defined in this study. The MLRs we identified range in size from 1.56 cM at D18S39 to 15.79 cM at D18S61. The region of minimal loss flanking the most commonly lost marker, D18S70, is 3.67 cM. These are large regions, and until chromosome 18 is completely sequenced, further study will require building of contigs over each of the MLRs to identify new markers and to find the possible candidate genes within the consistently lost regions. However, even without physical mapping, there are some known genes mapped within the MLRs. For example, within the D18S70 region there are three known genes as well as several expressed sequence tags that may be candidates. The known genes are PEPA, a cytoplasmic peptidase; NAFTC1, a nuclear transcription factor in activated T cells; and GALNR1, the galanin receptor 1. The latter is a seven transmembrane G-protein coupled receptor that is activated by the ubiquitous 37-amino acid neuropeptide galanin (18) . Alterations of this class of G-protein coupled receptor could affect growth and differentiation signals, as is the case with the bombesin or gastrin-releasing peptide receptor in small cell lung cancer (19) . However, it is unknown at this time whether the galanin receptor is expressed or active in squamous mucosa.
An alternative way of considering the data that 18q is frequently lost in many tumors is that the breakpoint region may be the important event in tumor development. The common proximal break point of 18q LOH located between D18S480 and D18S56 encompasses the DSC (desmocollin) and DSG (desmoglein) family of genes. The DSC and DSG family genes are components of desmosomes, which are critical in maintaining cell-to-cell adhesion in the squamous epithelium. Disruption or loss of function of these genes may play an important role in the metastatic spread of malignant tumors (20) . It will be of interest to determine whether these genes are functional in the tumor cell lines with breakpoints at these regions.
Three MLRs are identified: D18S39 in 18q21.1; D18S61 in 18q22.2; and D18S70 in 18q23. Thus far, no known candidate TSGs are mapped to these MLRs. Further efforts are needed to confirm these MLRs in more tumors and to identify candidate genes in these regions that may be inactivated by LOH and mutation.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
1 This work was supported by the Munn Fund of the
University of Michigan Comprehensive Cancer Center, a Fellowship from
the Elsa U. Pardee Foundation (to K- Y. J.), Hamamatsu University
School of Medicine (to S. T.), and Grant R01 DE12477 from the
NIH-National Institute of Dental and Craniofacial Research. 
2 Present address: Department of
Otorhinolaryngology, Hamamatsu University School of Medicine, 3600
Handa-cho, Hamamatsu 431-3192, Japan.
3 Present address: Department of
Otorhinolaryngology-Head and Neck Surgery, Korea University College of
Medicine, 26-1, 5Ka, Anam-Dong, Sungbuk-ku, Seoul 136-705, Korea.
4 To whom requests for reprints should be
addressed, at Laboratory of Head and Neck Cancer Biology, 6020 KHRI,
University of Michigan, 1301 East Ann Street, Ann Arbor, MI 48109-0506.
Phone: (734) 764-4371; Fax (734) 764-0014; E-mail: careyte{at}umich.edu
5 The abbreviations used are: HNSCC, squamous cell
carcinoma of the head and neck region; LOH, loss of heterozygosity; AI,
allelic imbalance; NL, no loss; NI, not informative; MSI,
microsatellite instability; MLR, minimally lost region; TSG, tumor
suppressor gene.
6 Data regarding the primer sets were obtained
from the Genome Database (http://www.gdbwww.gdb.org/), the Genetic
Location Database
(http://www.cedar.genetics.sonton.ac.uk/public_html/), and the
vendor’s catalogue.
7 S. Takebayashi, A. Hickson, T. Ogawa, K-Y. Jung,
H. Mineta, R. Grenman, and T. E. Carey. Tumor progression associated
with loss of heterozygosity on 18q in head and neck squamous cell
carcinoma, manuscript in preparation.
8 K. Y. Jung, B. P. Phan, A. Hickson, S.
Takebayashi, J. O. Choi, and T. E. Carey. Expression of candidate
TSGs, DCC and DPC4, in primary and
metastatic squamous cell carcinomas of head and neck, manuscript in
preparation.
Received 1/27/00. Accepted 5/17/00.
|
REFERENCES |
|---|
|
TopABSTRACTIntroductionMaterials and MethodsResultsDiscussionREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  M. Takkunen, R. Grenman, M. Hukkanen, M. Korhonen, A. Garcia de Herreros, and I. Virtanen Snail-dependent and -independent Epithelial-Mesenchymal Transition in Oral Squamous Carcinoma Cells J. Histochem. Cytochem., November 1, 2006; 54(11): 1263 - 1275. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. S. Henson, R. R. Neubig, I. Jang, T. Ogawa, Z. Zhang, T. E. Carey, and N. J. D'Silva Galanin Receptor 1 Has Anti-proliferative Effects in Oral Squamous Cell Carcinoma J. Biol. Chem., June 17, 2005; 280(24): 22564 - 22571. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Baez, A. Cantor, S. Fonseca, M. Marcos-Martinez, L. A. Mathews, C. A. Muro-Cacho, and T. Munoz-Antonia Differences in Smad4 Expression in Human Papillomavirus Type 16-Positive and Human Papillomavirus Type 16-Negative Head and Neck Squamous Cell Carcinoma Clin. Cancer Res., May 1, 2005; 11(9): 3191 - 3197. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Akervall, X. Guo, C.-N. Qian, J. Schoumans, B. Leeser, E. Kort, A. Cole, J. Resau, C. Bradford, T. Carey, et al. Genetic and Expression Profiles of Squamous Cell Carcinoma of the Head and Neck Correlate with Cisplatin Sensitivity and Resistance in Cell Lines and Patients Clin. Cancer Res., December 15, 2004; 10(24): 8204 - 8213. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. A. Ginos, G. P. Page, B. S. Michalowicz, K. J. Patel, S. E. Volker, S. E. Pambuccian, F. G. Ondrey, G. L. Adams, and P. M. Gaffney Identification of a Gene Expression Signature Associated with Recurrent Disease in Squamous Cell Carcinoma of the Head and Neck Cancer Res., January 1, 2004; 64(1): 55 - 63. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. R. Bradford Genetic Markers of Head and Neck Cancer: Identifying New Molecular Targets Arch Otolaryngol Head Neck Surg, March 1, 2003; 129(3): 366 - 367. [Full Text] [PDF]
|
|
|
|
 |
|
 D J Smiraglia, L T Smith, J C Lang, L J Rush, Z Dai, D E Schuller, and C Plass Differential targets of CpG island hypermethylation in primary and metastatic head and neck squamous cell carcinoma (HNSCC) J. Med. Genet., January 1, 2003; 40(1): 25 - 33. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. P. Tabor, R. H. Brakenhoff, V. M. M. van Houten, J. A. Kummer, M. H. J. Snel, P. J. F. Snijders, G. B. Snow, C. R. Leemans, and B. J. M. Braakhuis Persistence of Genetically Altered Fields in Head and Neck Cancer Patients: Biological and Clinical Implications Clin. Cancer Res., June 1, 2001; 7(6): 1523 - 1532. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. A. Muro-Cacho, K. Rosario-Ortiz, S. Livingston, and T. Munoz-Antonia Defective Transforming Growth Factor {beta} Signaling Pathway in Head and Neck Squamous Cell Carcinoma as Evidenced by the Lack of Expression of Activated Smad2 Clin. Cancer Res., June 1, 2001; 7(6): 1618 - 1626. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Cancer Research | Clinical Cancer Research |
| Cancer Epidemiology Biomarkers & Prevention | Molecular Cancer Therapeutics |
| Molecular Cancer Research | Cancer Prevention Research |
| Cancer Prevention Journals Portal | Cancer Reviews Online |
| Annual Meeting Education Book | Meeting Abstracts Online |
