numerical abnormalities of chromosomes 17 and 18 in sporadic colorectal cancer: incidence and...

Numerical Abnormalities of Chromosomes 17 and 18in Sporadic Colorectal Cancer: Incidence and

Correlation With Clinical and Biological Findings andthe Prognosis of the Disease

Jacinto Garcia,1 Angel Duran,2 Maria Dolores Tabernero,3 Asuncion Garcia Plaza,1

Teresa Flores Corral,4 Maria Luisa Najera,2 Alberto Gomez-Alonso,1 and Alberto Orfao2*1Servicio de Cirugia, Hospital Universitario and Departamento de Cirugia, Centro de Investigaciones del Cancer, Universidad

de Salamanca, Salamanca, Spain2Servicio General de Citometria, Departamento de Medicina and Centro de Investigaciones del Cancer, Salamanca, Spain

3Unidad de Investigacion, Hospital, Universitario de Salamanca, Salamanca, Spain4Servicio de Anatomia Patologica, Universidad de Salamanca, Salamanca, Spain

Background: In recent years important information has accumulated on the genetic alterations present incolorectal tumors. However, thus far few studies have analyzed the impact of numerical abnormalities ofchromosomes 17 and 18, which carry the p53 and DCC plus SHAD4/DPC4 genes involved in colorectalcancer, on the clinical and biological behaviors of the disease.

Methods: With the use of interphase fluorescence in situ hybridization (FISH), we analyzed the incidenceof numerical abnormalities of chromosomes 17 and 18 in a series of malignant colorectal tumors andexplored its potential association with clinicobiological behavior and the prognosis of the disease. For thispurpose, 94 consecutive patients newly diagnosed with colorectal cancer were analyzed. In all cases, FISHanalyses of the number of copies and nuclei of chromosomes 17 and 18 were performed in interphase nucleiwith the use of double stainings. For all patients, information on age, sex, tumor size, Dukes’ stage, tumorlocalization, DNA ploidy status, and the proportion of S-phase tumor cells was recorded. Median follow-upwas 38 months.

Results: Numerical abnormalities of chromosomes 17 and 18 were present in most patients withcolorectal cancer (57% and 52%, respectively). Gains of chromosome 17 and monosomy 18 were found in51% and 29% of cases, respectively, and they were the most frequent individual abnormalities for eachchromosome. The simultaneous analysis of the number of copies of both chromosomes in the same cellshowed that, in most cases displaying numerical abnormalities for these chromosomes, two or more differenttumor cell clones were present. From a clinical point of view, numerical abnormalities of chromosome 17,especially monosomy 17, were associated with a significantly higher incidence of rectal tumors (P � 0.001)and Dukes’ stage D (P � 0.02) and a lower median of disease-free survival among patients who underwentcurative surgery (P � 0.05), as compared with diploid cases. In addition, cases with an altered number ofcopies of chromosome 17 showed a higher incidence of DNA aneuploidy (P � 0.0001) and a greaterproportion of S-phase cells (P � 0.001) by flow cytometry. In contrast, no clear association was foundbetween the presence of numerical abnormalities of chromosome 18 and clinicobiological disease char-acteristics, except for a higher incidence of DNA aneuploidy by flow cytometry (P � 0.001) and a lowermedian of disease-free survival (P � 0.06). Multivariate analysis showed that numerical abnormalities ofchromosome 17, but not of chromosome 18, are an independent prognostic factor for predicting disease-freesurvival in patients with colorectal cancer.

*Correspondence to: Professor Alberto Orfao, MD, PhD, Centro deInvestigaciones del Cancer (Laboratorio 11), Universidad de Salamanca,Paseo de la Universidad de Coimbra s/n, 37007 Salamanca, Spain.

E-mail: [email protected] 19 February 2002; Accepted 30 August 2002Published online in Wiley InterScience (www.interscience.wiley.com).

DOI: 10.1002/cyto.b.10006

Cytometry Part B (Clinical Cytometry) 51B:14–20 (2003)

© 2002 Wiley-Liss, Inc.

Conclusions: Numerical abnormalities of chromosomes 17 and 18 were relatively common findings inpatients with colorectal cancer, with chromosome 17 being associated with a higher incidence of tumorslocalized to the rectum and a worse clinical outcome. Cytometry Part B (Clin. Cytometry) 51B:14–20, 2003.© 2002 Wiley-Liss, Inc.

Key terms: abnormalities of chromosomes 17 and 18; colorectal cancer; incidence; clinicobiologicalfeatures; prognosis

INTRODUCTIONColorectal cancer is the second leading cause of malig-

nant mortality in humans in developed countries, account-ing for more than 10% of all cancer-related deaths (1). Inpast decades, many relevant genetic events occurring dur-ing the malignant transformation of the intestinal epithe-lium have been identified and well characterized (2–5).According to these studies, p53 mutations in chromosome17q13.1 and deletion of the DCC gene together withdeletion and/or mutation of the SMAD4/DPC4 gene, bothlocated on the long arm of chromosome 18 (18q21) rep-resent two of the most significant events (4–15). P53 hasbeen identified as a tumor-suppressor gene that regulatesthe transition between the G1 and S phases of the cellcycle (16–18) whose product also acts as a pro-apoptoticprotein (19). In many human tumors including sporadiccolorectal cancer, one the of the p53 alleles is mutated inthe presence or absence of the other allele (20). Allelicdeletions of the DCC gene encoding for a protein display-ing high homology with neural cell adhesion molecules(7,9) also have been reported frequently in colorectaltumors, where they have been associated with a metastaticphenotype (4,6–9,21,22). Abrogation of the function ofSMAD4/DPC4 may cause a breakdown of the transformedgrowth factor-beta signaling pathway and loss of transcrip-tion of genes critical to cell cycle control (23,24). Additionalchromosomal or genetic aberrations relevant in the onco-genic process include molecular abnormalities of the RASand APC genes (4,5) and chromosome aneuploidization ortetraploidization (25,26). However, the genetic aberra-tions present in patients with colorectal cancer at diagno-sis are extremely complex (7); to the best of our knowl-edge, no individual genetic marker has been shown tocontribute to the prediction of any patient’s survival(4,6,9,22,27) except for the numerical abnormalities ofchromosome 18 (28). However, it should be noted thatmost information available on the cytogenetic abnormali-ties in sporadic colorectal tumors derives from studiesbased on conventional cytogenetic techniques (29–32).Although these techniques have provided a general over-view of all the genetic abnormalities present in singletumor cells, they have demonstrated some limitations re-lated, on many occasions and among other factors, to thefailure in obtaining tumoral mitosis. In addition, the needto culture the tumor cells may induce the selection ofspecific clones within a tumor and, hence, only a smallpercentage of all tumor cells is analyzed (33–37). In recentyears, fluorescence in situ hybridization (FISH) has be-

come a solid, objective, and reliable tool for the assess-ment of numerical and structural chromosomal aberra-tions present in interphase nuclei from tumor cells indifferent hematologic malignancies and solid tumors(36,38,39). In contrast to the more traditional cytogenetictechniques, interphase FISH allows the analysis of all cellnuclei in a tumor sample from all patients to whom it isapplied (36,39). Moreover, it provides highly specific andsensitive information on the genetic abnormalities ex-plored (33–35). Despite the advantages of interphase FISHmethods, few studies (6,28,33,37,40,41) have reported onthe potential prognostic value of the abnormalitiespresent on chromosomes that contain relevant genes inthe pathogenesis of colorectal cancer, such as chromo-somes 17 and 18.

The aim of the present study was to analyze by inter-phase FISH the incidence of numerical aberrations of twochromosomes (chromosomes 17 and 18) that carry genes(p53 and DCC) plus SMAD4/DPC4 relevant to the malig-nant transformation of colorectal epithelia and to exploreits potential association with the clinicobiological behav-ior and prognosis of the disease. For this purpose, arelatively large series of colorectal cancer patients under-going curative surgery was studied.

MATERIAL AND METHODSPatients

Ninety-four consecutive patients (47 men, 47 women;mean age � 66 years, age range � 36–89 years) newlydiagnosed with sporadic colorectal cancer who under-went curative surgery at the Department of Surgery of theUniversity Hospital of Salamanca were included in thepresent study. From the histopathologic point of view, 61cases corresponded to Dukes’ stages A and B, 23 to stageC, and 10 to stage D. In all cases, information was re-corded with regard to tumor size, tumor localization, andcytologic grade. Levels of carcinogenic embryonic antigenin sera also were assessed at diagnosis, in most cases withthe use of radioimmune analysis. At the moment of closingthis study, 56 patients were alive and free of disease,whereas 20 had relapsed (12 of whom had died). Medianfollow-up was 38 months. Fifteen patients died after diag-nosis due to non–tumoral-related complications and threeindividuals were lost during follow-up.

FISH Studies

In all cases, FISH analysis was performed on single-cellsuspensions obtained by mechanical disintegration of pri-

15CHROMOSOMES 17 AND 18 IN COLORECTAL CANCER

mary tumor specimens with the use of Medimachine(Dako, Glosturp, Denmark). Briefly, cells were fixed inCarnoy’s medium and dropped onto slides cleaned withethanol:ether (1:1, vol/vol) according to conventional cy-togenetic techniques. Slides then were sequentially incu-bated with solutions containing 0.1 mg/ml of RNAse A (1h at 37°C) and 0.1 mg/ml pepsin (10 min at 37°C); theywere fixed in 1% acid-free paraformaldehyde (10 min atroom temperature) and dehydrated in progressively stron-ger concentrations of ethanol (70%, 95%, and 100%) aspreviously reported (42). Afterward, slides containingDNA from tumor cells and 10 ng of each of the two DNAprobes assayed, the biotinylated pZ17-1.6A probe and thedigoxigenin-labeled p2X ba probe (Boehringer Mannheim,Mannheim, Germany), specifically directed for the identi-fication of repetitive DNA sequences of the centromericregions of chromosomes 17 and 18, respectively, weredenatured in an 80°C oven for 100 s.

After denaturation, the slides were hybridized overnightin a humid chamber at 37°C. The immunologic detectionof the biotin-labeled pZ17-1.6A and the digoxigenin-la-beled pZX ba hybridized probes was performed with animmunologic blocking incubation with 4 mol/l of buffer(30 min at 37°C) followed by another incubation (30 minat 37°C) with a solution containing avidin-fluorescein iso-thiocyanate (Vector Laboratories, Burlingame, CA) and amouse anti-digoxigenin monoclonal antibody conjugatedwith rhodamine (Boehringer Mannheim).

Nuclei were counterstained with 4,6-diamidino-2-phe-nylindole (Sigma, St. Louis, MO) in the presence ofVectashield (Vector Laboratories) as an anti-fading agent,as previously described (42). The number of hybridizationspots was evaluated with a DMRB fluorescence micro-scope (Leica, Wezler, Germany) equipped with a 100� oilobjective, with a count of at least 200 nuclei/sample. In allslides analyzed, the number of nonhybridized cells in theareas assessed was lower than 1% and only those spotswith a similar size, intensity, and shape were counted.

Mononuclear cells from 10 age- and sex-matched healthyindividuals displaying normal karyotypes were used as con-trols. In these control samples, the mean percentages oftrisomic and monosomic cells were 0.5 � 0.8 and 5.6 � 2.8for chromosome 17 and 0.8 � 0.4 and 1.3 � 0.7% forchromosome 18, respectively. A patient was considered tocarry a numerical chromosomal abnormality when the per-centage of cells displaying an abnormal number of spots forchromosomes 17 and 18 was higher than the mean valueplus two standard deviations obtained for the same chromo-some in the normal control samples.

Flow Cytometric Analysis of Cell DNA Contents

Flow cytometric analysis of cell DNA contents was per-formed in all cases on single-cell suspensions obtained fromthe diagnostic tumor sample after its mechanical disaggrega-tion with the procedure described above. Between 0.5 and1 � 106 cells were washed once in a citrate buffer (85.5 g ofsucrose [Sigma], 11.76 g of trisodium citrate H2O2 [Merck,Darmstadt, Germany], and 50 ml of dimethyl sulfoxide[Merck] in 1 liter of distilled water, pH 7.6) and sequentially

incubated with 1.8 ml of a solution containing trypsin in astock solution –of citrate buffer plus 1.628 g/l of sperminetetrahydrochloride (Sigma) and 0.1% (vol/vol) Nonidet P40(Sigma), 1.5 ml of a solution containing RNAse (0.5 mg/ml)and a trypsin inhibitor (0.1 mg/ml; Sigma) in stock solutionand 1.5 ml of a solution containing propidium iodide (PI;0.42 mg/ml; Sigma) in stock solution for 10, 10, and 15 min(at room temperature), respectively. Immediately after theseincubations, measurements of propidium iodide–associatedfluorescence were performed in a FACSort flow cytometer(Becton Dickinson Biosciences, San Jose, CA) equipped withan argon-ion laser set at 488 nm. Information on a minimumof 3 � 105 events corresponding to the cellularity of theentire sample was collected for each tumor sample with theCellFit software program (Becton Dickinson Biosciences).The instrument was calibrated before data acquisition byusing chicken erythrocyte nuclei so that normal humanG0/G1 diploid cells, stained with identical sample prepara-tion conditions, would show a PI-associated fluorescenceintensity of 200 fluorescence channels (arbitrary units scaledfrom 0 to 1,023). A mixture of tumor cells and peripheralblood mononuclear cells from sex-matched, healthy individ-uals was stained in parallel to position the G0/G1 peak cor-responding to normal human diploid cells.

Data analysis was performed with the CellFit softwareprogram. DNA aneuploidy was defined on the basis ofmore than one G0/G1 peak different from that of thenormal human diploid cells measured simultaneously.DNA index was calculated as the ratio between the modalfluorescence channel of G0/G1 tumor cells and that of thenormal G0/G1 diploid cells. To analyze the cell cycle dis-tribution of tumor cells, the Poly mathematical modelincluded in the CellFit software program was used in allcases after excluding cell debris and cell doublets in for-ward light scatter versus PI � FL2 area and a PI � FL2 areaversus PI � FL2 width bivariate dot plots, respectively.

Statistical Methods

Median, mean, standard deviation, and range valueswere calculated for all variables under study (SPSS, SanDiego, CA). To ascertain the statistical significance of thedifferences between groups, analysis of variance or theKruskal-Wallis test was used (SPSS). Survival curves wereplotted according to the method of Kaplan and Meier andthe existence of significant differences in survival be-tween groups was explored with the log-rank test (SPSS)for those patients undergoing curative surgery. Multivari-ate analysis of prognostic factors was performed with theCox regression model (SPSS).

RESULTSOf the 94 patients included in the present study, most

showed numerical abnormalities of 57% and 52% for chro-mosomes 17 and 18, respectively. Gains of chromosome17 were much more frequent than monosomy 17 (51% vs.5%), whereas monosomy was the most frequent numeri-cal abnormality detected for chromosome 18 (29%). Ofthe 94 cases analyzed, only 13 (14%) showed a diploidkaryotype with regard to both chromosomes; 22 showed

16 GARCIA ET AL.

only trisomy 17 (n � 18) or only trisomy 18 (n � 4), and20 individuals gained both chromosomes simultaneously;losses of chromosome 18 alone or in combination withgain or loss of chromosome 17 were detected in 30%, 4%,and 8% of cases, respectively.

The use of a double-staining FISH technique allowed thedetection of up to four different clones per tumor sampleaccording to the numerical abnormalities detected percell within a tumor specimen for chromosomes 17 and 18.As might be expected, the incidence of cases with morethan one tumor cell clone was higher among patients whodisplayed numerical abnormalities for chromosomes 17(80% of cases with monosomy 17 and 78% of tumors withgains of chromosome 17) and 18 (75% of cases withmonosomy 18 and 88% of cases with gains for chromo-some 18) than among those who displayed two copies ofchromosome 17 (45%) and 18 (51%).

Tables 1 and 2 show the clinical and biological charac-teristics of colorectal cancer patients according to thepresence of numerical abnormalities of chromosome 17.As can be seen in both tables, tumors showing a diploidcontent for chromosome 17 were more frequently in thecolon (P � 0.01) and showed a higher frequency of

Dukes’ stages A and B (P � 0.02) and a lower relapse rate(P � 0.1) as compared with cases displaying numericalabnormalities, especially losses, of chromosome 17. Asexpected, patients with numerical abnormalities of chro-mosome 17 showed a significantly higher incidence ofDNA aneuploidy (P � 0.0001) and a greater proportion ofS-phase tumors cells (P � 0.0001). In contrast, no signif-icant differences were found between these groups ofindividuals with regard to age, sex, tumor size, and carci-nogenic embryonic antigen serum levels.

Interestingly, numerical abnormalities for chromosome18 did not show an association with most of the clinical(Table 3) and biological (Table 4) disease characteristicsanalyzed, except for a significantly higher incidence ofDNA aneuploidy (P � 0.001).

From the prognostic point of view, numerical abnor-malities for chromosomes 17 and 18 were associated witha significantly lower (P � 0.05 and 0.06, respectively)disease-free survival (Fig. 1), and no major differenceswere observed between cases displaying gains and mono-somy for either chromosome.

Multivariate analysis of prognostic factors with respectto disease-free survival showed that the best combination

Table 1Clinical and Histopathologic Characteristics of Patients With Colorectal Cancer According to the

Presence and Type of Numerical Abnormalities Detected for Chromosome 17

N. of Copies of chromosome 17P1 (n � 5) 2 (n � 40) �2 (n � 48)

Age (years)a 68 � 14 64 � 11 67 � 11 0.37Sex (male/female) 1/4 20/20 25/23 0.39Tumor size (cm)a 4.6 � 1.0 (4–6) 4.8 � 1.7 (2–10) 4.8 � 1.8 (1.8–9) 0.95Localization

Colon (%) 1 (20) 31 (78) 26 (53)Rectum (%) 4 (80) 9 (22) 22 (47) 0.01

Dukes’ stageA � B (%) 0 (0) 30 (75) 30 (64)C (%) 3 (60) 7 (17) 13 (25) 0.02D (%) 2 (40) 3 (8) 5 (11)

Relapseb 1/1 (50) 28/5 (15) 26/13 (50) 0.15aResults expressed as mean � standard deviation (range).bOnly for Dukes’ stages A, B, and C.

Table 2Colorectal Cancer: Biological and Serologic Disease Characteristics According to the Presence and Type

of Numerical Aberrations Detected for Chromosome 17


DNA ploidy statusHypodiploid (%) 2 (40) 2 (5) 0 (0)Diploid (%) 2 (40) 26 (65) 6 (12) 0.001Hyperdiploid (%) 1 (20) 12 (30) 42 (88)

Mean DNA index of aneuploid casesa 1.06b 1.38 � 0.29 (1.07–1.83) 1.65 � 0.30 (1.06–2.71) 0.01%S-phase tumor cellsa 15.5 � 7.4 (5.8–23) 10.9 � 5.3 (0.5–24.6) 19 � 9.4 (3.7–47) 0.001Carcinogenic embryonic antigen serum

levels (UI/ml)a 2.8 � 1 (2.1–3.5) 8.9 � 12.2 (1–58) 10.4 � 16.8 (0.3–63) 0.86aResults expressed as mean � standard deviation (range).bCould be calculated in one case.


of variables for predicting survival was age, sex, and thepresence or absence of numerical abnormalities of chro-mosome 17.

DISCUSSIONIn the present study we confirmed and extended pre-

vious observations which show that numerical abnormal-ities of chromosomes 17 and 18 are a common findingamong patients with sporadic colorectal cancer (3,5,28).Our results showed that gains of chromosome 17 arepresent in most patients with colorectal cancer at diagno-sis, whereas monosomy 17 is a rather infrequent finding.In contrast, numerical abnormalities of chromosome 18were more frequently associated with monosomy 18 thanwith an increased number of chromosomal copies. Previ-ous studies have shown that monosomy 18 is one of themost frequent genetic abnormalities in patients with spo-radic colorectal cancer, with an incidence of approxi-mately one-fourth to one-third of cases (4–8,25,30,32,37).In contrast, monosomy 17 has been reported to be a farrarer abnormality (33) as confirmed by our observations.Abnormalities of chromosomes 17 and 18 are believed toplay an important role in the pathogenesis of sporadiccolorectal cancer because these two chromosomes carry

relevant genes in the malignant transformation of the gutepithelium: the p53 and DCC plus SMAD4/DPC4 genes,respectively. Allelic loss of one DCC gene in chromosome18 has been associated with loss of the DCC gene function(7) and altered cell cycle regulation (23,24) respectively.Interestingly, for the p53� dominant type of mutations,the malignant phenotype would develop even without anallelic loss of another copy of the gene (20), which mightin part explain the low frequency of monosomy 17 incontrast to the high frequency of p53 alterations reported(32). In cases with gains of chromosome 17, duplicationof the chromosome carrying a mutated p53 gene couldconfer an even more dominant malignant phenotype tothe tumor cell (43,44). Interestingly, in the present study,we found that the combined analysis of chromosomes 17and 18 in the same cells allowed the identification of morethan one tumor cell clone in a significant proportion ofcases. These observations are in line with findings re-ported by Di Vinci et al. (25) who used a similar approach,thus supporting the existence of great levels of intratumorgenetic heterogeneity in colorectal cancer.

Although an increasingly large amount of informationon the molecular genetic alterations in colorectal tumors

Table 3Clinical and Histopathologic Characteristics of Patients with Colorectal Cancer According to the Presence

and Type of Numerical Chromosome Abnormalities Detected for Chromosome 18


Age (years)a 64.8 � 11.5 (43–86) 66.6 � 11.5 (36–89) 65 � 10 (42–81) 0.84Sex (male/female) 12/16 23/22 12/9 0.6Tumor size (cm)a 4.9 � 1.5 (4–9) 4.8 � 1.6 (2–10) 4.8 � 1.4 (1.8–8) 0.92Localization

Colon (%) 16 (59) 30 (67) 13 (62)Rectum (%) 12 (41) 15 (33) 8 (38) 0.71

Dukes’ stageA � B (%) 16 (57) 33 (75) 12 (57)C (%) 9 (32) 8 (16) 6 (29) 0.45D (%) 3 (11) 3 (9) 3 (14)

Relapseb 8/22 (36) 6/32 (16) 5/15 (33) 0.15aResults expressed as mean � standard deviation (range).bOnly for Dukes’ stages A, B, and C.

Table 4Colorectal Cancer: Biological and Serologic Disease Characteristics According to the Presence and Type

of Numerical Aberrations Detected for Chromosome 18


DNA ploidy statusHypodiploid (%) 3 (11) 1 (2) 0 (0)Diploid (%) 15 (53) 18 (41) 2 (10) 0.001Hyperdiploid (%) 10 (20) 26 (58) 19 (90)

Mean DNA index aneuploid casesa 1.43 � 0.3 (1.06–1.79) 1.51 � 0.23 (1.06–1.83) 1.76 � 0.38 (1.15–2.71) 0.01%S-phase of tumor cellsa 13.6 � 6.1 (3.9–30.8) 14.4 � 7.4 (0.5–36.7) 19.4 � 12.6 (3.7–46.8) 0.14Carcinogenic embryonic antigen serum

levels (Ul/ml)a 10.9 � 13.5 (0.3–55.4) 8.7 � 16 (0.5–63.2) 8.7 � 4.9 (0.9–13.7) 0.87aResults expressed as mean � standard deviation (ranges).

18 GARCIA ET AL.

is available, to date few studies have been devoted specif-ically to the analysis of the potential clinical and prognos-tic effects of numerical abnormalities of chromosomes 17and 18 by interphase FISH (32,33,45,46). In the presentstudy we found that numerical abnormalities of chromo-some 17 are significantly more frequent in tumors in therectum and also display a higher frequency of tumors withadvanced Dukes’ stages. In addition, these patientsshowed a significantly higher incidence of DNA aneu-ploidy and more fractions of proliferating cells as evalu-ated by the proportion of S-phase tumor cells. Interest-ingly, once cases with gains of chromosome 17 werecompared with those with monosomy 17, those withmonosomy 17 showed a significantly higher incidence ofrectal, Dukes’ stage D, and DNA hypodiploid tumors.Previous studies have associated abnormalities of chromo-some 17 with more advanced stages of the disease(33,47,48), as in the present series. However, to the best

of our knowledge, this is the first study in which a corre-lation between numerical abnormalities of chromosome17 and tumor localization in the rectum has been re-ported. These findings suggest that development of colo-rectal cancer, at least at the stage of aneuploidization, mayfollow different genetic pathways depending on the local-ization of the tumor.

As expected, losses and gains of chromosome 17 weresignificantly associated with a greater incidence of DNAhypodiploidy and DNA hyperdiploidy, respectively. Theseobservations concurred with those of previous studies(32,33) and may reflect that the overall tumor cell DNAcontent depends on the DNA ploidy state as a measure ofthe numerical chromosomal abnormalities carried by thetumor cells; moreover, our observations suggested thatchromosome 17 is usually involved in the aneuploidiza-tion process.

In contrast to chromosome 17, no significant associa-tion was found between abnormalities in the number ofcopies of chromosome 18 in tumor cells and the clinicaland biological characteristics of the disease, except for ahigher incidence of DNA aneuploidy by flow cytometryamong cases displaying an abnormal number of copies ofchromosome 18. Previous reports, based on relativelysmall series of colorectal cancer patients, have associatedabnormalities of chromosome 18, especially monosomy18, with a higher incidence of metastasis and a moreaggressive disease course (4,6,21,22). In the presentstudy, no clear association was observed between mono-somy 18 (or gains of chromosome 18) and Dukes’ stage,tumor size, or the proportion of S-phase tumor cells.Nevertheless, it should be noted that patients with numer-ical abnormalities of chromosome 18 who underwentcurative surgery displayed a higher incidence of relapsesand a significantly lower disease-free survival rate(4,6,22,28).

Identification of prognostic factors among patients un-dergoing curative surgery represents a challenge for themanagement of colorectal cancer (49). Although manyindividual genetic markers have been associated with dis-ease outcome (50), none of them is currently used inroutine methods for predicting patient outcome from thetime of diagnostic surgery (33,51). Interestingly, in thepresent study, numerical abnormalities of chromosome 17appeared to have an independent prognostic value forpredicting disease-free survival in patients with colorectalcancer undergoing curative surgery. However, becausethese patients showed a higher incidence of tumors in therectum, differences in the timing of tumor diagnosis andthe type of treatment received also might influence pa-tient outcome. Further studies based on larger series ofpatients with long follow-up periods are necessary toconfirm these observations.

LITERATURE CITED1. Olandis SH, Murray T, Bolden S, Wingo PA. Cancer statistics 1999. CA

Cancer J Clin 1999;49:8–31.2. Powell SM; Zilz N, Beazer-Barclay Y, Bryan TM, Hamilton SR, Thi-

bodeau SN, Vogelstein B, Kinzler KW. APC mutations occur earlyduring colorectal tumorigenesis. Nature 1992;359:235–237.

FIG. 1. Disease-free survival of patients with colorectal cancer whounderwent curative surgery, according to the presence or absence ofnumerical abnormalities of chromosomes 17 (A) and 18 (B) by inter-phase fluorescence in situ hybridization.


3. Vogelstein B, Kinzler KW. The multistep nature of cancer. TrendsGenet 1993;9:138–141.

4. Jen J, Kim H, Piantodosi S, Liu ZF, Lewitt RC, Sistonen P, Kinzler KW,Vogelstein B, Hamilton SR. Allelic loss of chromosome 18q andprognosis in colorectal cancer. N Engl J Med 1994;331:213–221.

5. Bodmer W, Bishop T, Karran P. Genetic steps in colorectal cancer.Nat Genet 1994;6:217–219.

6. Frank CJ, McClatchey KD, Devaney KO, Carey TE. Evidence that lossof chromosome 18q associated with tumor progression. Cancer Res1997;57:824–827.

7. Oci A, Huang CHD, Masayoshi M, Nakaminshi I. Numerical chromo-some alterations in colorectal carcinoma detected by fluorescence insitu hybridisation. Relationship to 17p and 18q allelic losses. Vir-chows Arch 1996;428:243–251.

8. Fearon FR, Cho KK, Nigro JM, Kern SE, Simonds JV, Ruppert JM,Hamilton SR, Preisinger AC, Thomas G, Kinzler KW. Identification ofchromosome 18q gene that is colorectal cancers. Science 1990;247:49–56.

9. Halmiton SR. Molecular genetic alterations as potential prognosticindicator in colorectal carcinoma. Cancer 1992;69:1589–1591.

10. Vogestein B, Fearon ER, Hamilton SR. Genetic alterations duringcolorectal tumor development. N Engl J Med 1988;319:525–532.

11. Nanashina A, Tagawa Y, Yasutake T, Sawai T, Tuji T, Sasano O,Nakagoe T, Ayabe H. Aneusomy of chromosome 18 is associated withthe development of colorectal carcinoma. J Gastroenterol 1997;32:487–491.

12. Thiagalingam S, Lengauer C, Leach ES, Schutte M, Halm SA, Over-hause J, Willson JKV, Markowitz S, Hamilton SR, Kern SE, KinzlerKW, Volgelstein B. Evaluation of candidate tumor suppressor geneson chromosome 18 in colorectal cancer. Nat Genet 1996;13:343–346.

13. Salovaara R, Roth S, Loukola V, Sistonen P, Avizienyte E, et al.Frequent loss of SMAD4/DPC4 protein in colorectal cancers. Gut2002;51:59.

14. Takagi Y, Kohmura H, Futamura M, Kida H, Tanemura H, ShimokawaK, et al. Somatic alterations of the DPC4 gene in human colorectalcancers in vivo. Gastroenterology 1996;111:1387–1389.

15. Miyaki M, Iijima T, Konishi M, Sakai K, Ishii A, Yasuno M, et al. Higherfrequency of SMAD4 gene mutation in human colorectal cancer withdistant metastsis. Oncogene 1999;18:3098–3103.

16. Agarwal UL, Agarwal A, Taylor WR, Stark GR. P53 controls both G2/Mand G1 cell cycle checkpoints and mediates reversible growth arrestin human fibroblast. Proc Soc Natl Acad Sci USA 1995;92:8493–8497.

17. Steward N, Nicks GG, Paraskevas F, Mowat M. Evidence for a secondcells cycle block at G2/M by p 53. Oncogene 1995;10:109–115.

18. Shackney SE, Shankey V. Common patterns of genetic evolution inhuman solid tumors. Cytometry 1997;29:1–27.

19. Kostic C, Shaw PH. Isolation and characterization of sixteen novelp53 response genes. Oncogene 2000;19:3978–3987.

20. Delattre O, Olschwang S, Law DJ, Melot T, Remvikos Y, Salmon RJ,Sastre X, Validire P, Feinberg AP, Thomas G. Multiple genetic alterationsin distal and proximal colorectal cancer. Lancet 1989;ii:353–356.

21. Ookawa K, Sakamoto M, Mirohashi S, Yoshida Y, Suginura Terada M,Yokota J. Concordant p53 and DCC alterations and allelic losses onchromosomes 13q and 14q associated liver metastases of colorectalcarcinoma. Int J Cancer 1993;53:382–387.

22. Lino H, Fukuyana M, Maeda Y, Koike M, Mori T, Takahashi T, Kikuchi-Yanoshitz R, Miyaki M, Mizuno S, Watanabe S. Molecular genetic forchemical management of colorectal carcinoma 17p, 18q and 22q lossof heterozygosity and decreased DCC expression are correlated withthe metastatic potential. Cancer 199;73:1324–1331.

23. De Caestecker MP, Piek E, Roberts AB. Role of transforming growthfactor-beta signaling in cancer. J Natl Cancer Inst 2001;93:555–557.

24. Grady WM, Markowitz SD. Genetic and epigenetic alterations incolon cancer. Annu Rev Genomics Hum Genet 2002;3:101–128.

25. Di Vinci A, Infusini E, Peveri C, Sciutto A, Orecchia R, Geido E,Monaco R, Giaretti W. Intratumor heterogeneity of chromosome1,7,17 and 18 aneusomies obtained by FISH and association with flowcytometric DNA index in human colorectal adenocarcinomas. Cytom-etry 1999;35:369–375.

26. Giaretti W. A model of DNA aneuploidization and evolution in colo-rectal cancer. Lab Invest 1994;71:904–910.

27. Jernwall P, Makinen MJ, Karttunen TJ, Makela J, Vihko P. Loss ofheterozygosity at 18q21 is indicative of recurrence and thereforepoor prognosis in a subset of colorectal cancers. Br J Cancer 1999;79:903–908.

28. Zhou W, Goodman SN, Galizia G, Lieto E, Ferraraccio F, Pignatelli C,Purdie CA, Piris J, Morris R, Harrison DJ. Counting alleles to predictrecurrence of early-stage colorectal cancers. Lancet 2002;359:219–225.

29. Bomme L, Bardi G, Pandis N, Fenger G, Kronborg O, Hein S. Clonalkaryotypic abnormalities in colorectal adenomas. Clues to the earlygenetic events in the adenoma-carcinoma sequence. Genes Chromo-somes Cancer 1994;10:190–196.

30. Muleris M, Zafrani B, Validice P, Girodet J, Salman RJ, Duitrillaux B.Cytogenetic study of 30 colorectal adenomas. Cancer Genet Cyto-genet 1994;74:104–108.

31. Bardi G, Sukhikh T, Panlis N, Fenger C, Kronborg O, Helm S. Karyo-typic characterization of colorectal adenocarcinoma. Genes Chromo-somes Cancer 1995;12:97–109.

32. Steiner MG, Harlow SP, Colombo E, Bauer KD. Chromosomes 8,12 and17 copy number in Astler-Coller stage C colon cancer in relation toproliferative activity and DNA ploidy. Cancer Res 1993;53:681–686.

33. Tagawa Y, Yantake T, Sawai T, Nanashima A, Jibiki M, Morinaga M,Akama F, Nakagoe T, Ayabe H. Clinical and pathological significanceof numerical aberrations of chromosomes 11 and 17 in colorectalneoplasm. Clin Cancer Res 1997;3:1587–1592.

34. Hopman AHN, Moesker O, Smeets AWGB, Pauwels RPE, Vooijs GP,Ramekes FCS. Numerical chromosome 1, 7, 9 and 11 aberrations inbladder cancer detected by in situ hybridization. Cancer Res 1991;51:644–651.

35. Van Dekken M, Pizzolo JG, Kelsen DP, Melaned MR. Targeted cyto-genetic analysis of gastric tumor by in situ hybridization with a set ofchromosome specific DNA probes. Cancer 1990;66:491–497.

36. Wolman SR. Fluorescence in situ hybridization; a new tool for thepathologist. Hum Pathol 1994;25:586–590.

37. Sasaki K, Sato T, Kurose A, Uesugi N, Ikeda E. Monosomy of chro-mosome 18 detected by fluorescence in situ hybridization in colorec-tal tumors. Cancer 1995;76:1132–1138.

38. Criado B, Barros A, SuijKerbuijk RF, Weghuis MO, Seruca R, FonsecaE, Castedo S. Detection of numerical alterations for chromosome 7and 12 in benign thyroid lesion by in situ hybridization. Am J Pathol1995;145:136–144.

39. Trask BJ. Fluorescence in situ hybridization. Trends Genet 1991;7:149–154.

40. Paredes-Zaghul A, Kang JJ, Essig YP, Mao W, Irby R, Wloch M,Yeatman TJ. Analysis of colorectal cancer by comparative genomichybridization: Evidence for induction of the metastatic phenotype byloss of tumor suppressor genes. Clin Cancer Res 1998;4:879–886.

41. Chen TR, Nierman WC. Chromosome painting and quantitativekaryotyping of colon adenocarcinoma cell lines DLD-1 and HTC-15.Anticancer Res 1994;14:109–112.

42. Tabernero MD, San Miguel JF, Garcia Sanz R, Najera ML, GarciaIsidoro M, Perez Simon JA, Gonzalez M, Wiegant J, Raap AK, Orfao A.Incidence of chromosome numerical changes in multiple myeloma.Fluorescence in situ hybridization analysis using chromosome-spe-cific probes. Am J Pathol 1996;149:153–161.

43. Takahashi DM, Angriman T, Yaremko ML, Montag A, Westrook CA,Michelassi F. Chromosome 17p allelic loss in colorectal carcinoma.Clinical significance. Arch Surg 1995;130:585–589.

44. Gerdes M, Chen Q, Elahi AH, Sircar A, Goldberg E, Winawer D, Urma-cher C, Winawer SJ, Jhannar SC. Recurrent deletions involving chromo-somes 1,5,17 and 18 in colorectal carcinomas: possible role in biologicaland clinical behavior of tumors. Anticancer Res 1995;15:13–24.

45. Laurent-Puig P, Olschwang S, Delattre O, Remvikos Y, Asselain B,Melot T, Validire P, Muleris M, Girodet J, Salmon RJ. Survival andacquired genetic alterations in colorectal cancer. Gastroenterology1992;102:1136–1141.

46. Herbergs J, Hopman AH, De Bruine AP, Raemakers FC, Arends JW. Insitu hybridization and flow cytometric analysis of colorectal tumorssuggests two routes of tumorigenesis characterized by gain of chro-mosome 7 and loss of chromosomes 17 and 18. J Pathol 1996;179:243–247.

47. Kern SE, Fearon ER, Tersmette KWF, Enterline JP, Leppert M, Naka-mura Y, White R, Vogelstein B, Hamilton SR. Allelic loss in colorectalcarcinoma. JAMA 1989;261:3099–3103.

48. Baker SJ, Preisinger AC, Jessup JM, Paraskeva C, Markovitz S, WillsonJK, Hamilton SR, Vogelstein B. P53 gene mutations occur in combi-nation with 17p allelic delections as late events in colorectal tumor-igenesis. Cancer Res 1990;50:7717–7722.

49. Font A, Abad A, Monzo M, Sanchez JJ, Guillot M, Manzano JL, Pinol M,Ojanguren Y, Rosell R. Prognostic value of K-ras mutations and allelicimbalance on chromosome 18q in patients with resected colorectalcancer. Dis Colon Rectum 2001;44:549–557.

50. O’Connell JM, Shaid DJ, Gangu V, Cunningham J, Kovach JS, Thi-bodeau SN. Current status of adjuvant chemotherapy for colorectalcancer. Can molecular markers play a role in predicting prognosis?Cancer 1992;70:1732–1739.

51. Offit K. Genetic prognostic markers for colorectal cancer. N EnglJ Med 2000;342:124–125.

20 GARCIA ET AL.

numerical abnormalities of chromosomes 17 and 18 in sporadic colorectal cancer: incidence and...

Documents