Interphase cytogenetic studies of human hepatocellular carcinomas by fluorescent in situ hybridization

Interphase Cytogenetic Studies of Human Hepatocellular Carcinomas by Fluorescent In Situ Hybridization CHANTAL HAMON-BENAIS,1 OLIVIER INGSTER,1 BENOIT TERRIS,2 MARIE-HE´LE`NE COUTURIER-TURPIN,1 ALAIN BERNHEIM,3 AND GE´RARD FELDMANN1

Although numerous allelic chromosome losses have been reported in hepatocellular carcinomas (HCC), chromosome analysis by cytogenetic methods has rarely been performed in these tumors, unlike in other solid malignant tumors. The purpose of the current study was to analyze primary liver tumors by conventional cytogenetic methods and by a new molecular cytogenetic technique, called fluorescent in situ hybridization (FISH), a technique that has been recently proposed to count the number of chromosome copies in interphase nuclei with chromosome centromeric probes. Primary cultures of tumoral cells were prepared to obtain metaphases. Specific chromosomes probes 7, 17, and 20 were used to perform in situ hybridization on isolated intact tumoral cells. Seven cases of primary liver tumors (six cases of HCC and one case of benign focal hepatic nodular hyperplasia) were investigated. A few metaphases were obtained in five of the seven tumors, and in most cases numerical abnormalities were difficult to interpret. In contrast with in situ hybridization, all cases of HCC showed losses and/or gains of chromosomes. Loss of one to three chromosomes occurred in five tumors. A gain of two chromosomes was observed in two of these five tumors. In only one case, a gain of only three chromosomes occurred. In addition, a loss of chromosome 17 was recorded for the benign tumor. These results demonstrate that FISH with specific probes can provide information on chromosome number in the tumoral cells of primary liver tumors even in the absence of analyzable metaphases. This technique opens new possibilities for the investigation of chromosome abnormalities in HCC. (HEPATOLOGY 1996;23:429-435.)

Abbreviations: HCC, hepatocellular carcinoma; FISH, fluorescent in situ hybridization; SSC, sodium saline citrate. From the 1Laboratoire de Biologie Cellulaire, Inserm U327, Faculte´ de me´decine Xavier Bichat, Universite´ Paris; 2Service d’Anatomie Pathologique, Hopital Beaujon, Clichy, 3and Laboratoire de Cytoge´ne´tique, Institut Gustave Roussy, Villejuif, France. Received April 26, 1995; accepted September 25, 1995. Dr. Hamon-Benais was supported by a grant from the ‘‘Fondation SingerPolignac’’ Paris. Drs. Hamon-Benais and Ingster contributed equally to this study. Address reprint requests to Ge´rard Feldmann, M.D., Inserm U. 327, Faculte´ de me´decine Xavier Bichat, Universite´ Paris 7 Denis Diderot, BP 416, 75870 Paris Cedex 18, France. Copyright q 1996 by the American Association for the Study of Liver Diseases. 0270-9139/96/2303-0006$3.00/0

Cancer is thought to be caused by a series of alterations of a limited number of specific genes, the socalled oncogenes and tumor suppressor genes.1,2 Some of these alterations have been shown to be related to chromosomal changes because specific chromosomal abnormalities have been demonstrated by cytogenetic studies of cancer cells.3 For instance, more than 100 recurrent chromosomal translocations have been described in approximately 14,000 cases of different malignant neoplasms.4 The cytogenetic study of chromosomal changes and the possible consequences on cellular function are therefore important for a more complete understanding of carcinogenesis. Hepatocellular carcinoma (HCC) is one of the most common cancers worldwide. Epidemiological studies have shown frequent associations between HCC and chronic hepatitis B or C virus infections, alcohol abuse, or hemochromatosis.5-7 The oncogenic effects of these factors remain unclear despite numerous hypotheses and extensive investigations.6,8,9 Although numerous chromosome allelic losses have been reported in HCC,10,11 specific regions have not been determined. Moreover, very few primary HCCs have been studied by conventional cytogenetic techniques,12-14 and, as far as we know, none have been studied by molecular cytogenetic techniques. A major advance in cytogenetics was the introduction of fluorescent in situ hybridization (FISH) with cloned DNA probes that can recognize chromosome-specific repeat sequences, such as a-satellite centromeric DNAs. An important advantage of this method is that the chromosome copy number can be determined in the interphase nuclei of solid tumors. This study analyzed six cases of primary HCC and one case of focal hepatic nodular hyperplasia, a benign liver tumor, by conventional cytogenetic methods and by FISH performed on nuclei of isolated intact tumoral cells. MATERIALS AND METHODS Tumor Material Tumor tissue from six cases of HCC and one case of focal hepatic nodular hyperplasia were collected during surgery. A part of the tissue was fixed in formol for histopathological studies. The other part was left unfixed and immediately placed in a modified Eagle medium. The size of the specimens varied from 0.3 to 4.0 cm3. The age and sex of the patients

429

5p0b$$0003

02-16-96 18:02:46

hepa

WBS: Hepatology

430 HAMON-BENAIS ET AL.

HEPATOLOGY March 1996

TABLE 1. Clinical and Laboratory Data of Seven Patients With Benign or Malignant Primary Tumors of Liver

Patient

Age (yr)

Sex

Size* (cm)

Cause

Histology and Edmondson Grade

1 2

31 34

M F

2.5 18

Unknown Viral chronic hepatitis B

Nodular hyperplasia‡ Grade III

3 4 5 6 7

59 54 51 67 38

M M F M M

4 4.5 4 7 6

Alcohol-induced cirrhosis Viral chronic hepatitis C Viral chronic hepatitis C Unknown Viral chronic hepatitis B

Grade Grade Grade Grade Grade

III II II II II

No. of Metaphases Obtained

1 9 0 5 0 1 22

No. of Chromosomes†

46 42, 45, 47, 49, 70, 71, 75, 99, 127 40, 43 40 46 for 10 mitoses, 45, 39, 43, 37

* Maximum size of tumor nodule. † Chromosome number is not indicated for every metaphase obtained in cases 4 and 7, some metaphases being innumerable. ‡ This case is a benign liver tumor.

and the causes of the tumors are given in Table 1. Results of the histopathological examination for each malignant tumor classified according to Edmondson15 and the number of metaphases obtained in each case are also provided in Table 1. Cytogenetic Methods Fresh tumor tissue was cut in 1-mm3 pieces by mechanical mincing; the pieces were exposed to a 1% collagenase A solution (Sigma, St. Louis, MO) for 10 to 60 minutes and filtered on 70-mm millipores. Cell viability was tested by trypan blue exclusion and varied from 18% to 95%. Cells were cultured in plastic flasks in the modified Eagle medium supplemented with 10% fetal bovine serum (Gibco, Paisley, Scotland), insulin (5 mg/mL), albumin (1 mg/mL), hydrocortisone (7.1005 mol/L), penicillin (100 U/mL), streptomycin (100 mg/mL), and gentamicin (10 mg/mL). Primary cultures were incubated in a humidified atmosphere with 5% CO2 in air, at 377C, in a continuous CO2 flow incubator. Twenty-four hours later, the medium was changed. In most cases, the cultured cells were harvested at various times between 2 hours and 10 days. Metaphase spreads from HCC were obtained by blocking cell mitosis with 0.2 mg/mL colcemid overnight. The cells were harvested, treated with 0.075 mol/L KCl at 377C for 12 minutes, and then fixed in a mixture of methanol:glacial acetic acid (3:1). Cell suspensions containing metaphases and intact nuclei were dropped onto glass slides and air dried. Metaphases were analyzed after Giemsa staining only because G banding was unsuccessful. Some slides were stored at 0207C for FISH. Normal human lymphocytes and hepatocytes were used for controls. Lymphocytes were obtained according to conventional cytogenetic techniques. Normal human adult hepatocytes obtained by collagenase perfusion of a hepatic lobe that had not been used for liver transplantation were provided by Dr. Y. Calmus (Hopital Cochin, Paris, France). Hepatocytes were harvested under the same conditions used for the tumor cells. DNA Probes and Fluorescent In Situ Hybridization The a-satellite DNA probes specific for the centromeres of chromosomes 7, 17, and 20 were used in this study. All probes, D7Z1,16,17 D17Z1,17 and D20Z118 were purchased in the digoxigenin-labeled form from Oncor (Gaithersburg, MD). The hybridization protocol followed was described by Pinkel et al19 with minor modifications. Before hybridization, slides

5p0b$$0003

02-16-96 18:02:46

hepa

were treated with enzymes to remove cell proteins and improve DNA probe penetration. The slides were treated with RNase A (1.0 mg/mL) (Boehringer, Mannheim, Germany) for 1 hour at 377C in a humidified chamber, washed twice in 21 SSC (sodium saline citrate), pH 7.0, for 5 minutes each at room temperature, followed by dehydration in an ethanol series (70%, 95%, 100%), and air dried. The slides were then incubated for 10 minutes at 377C in 0.01 mol/L HCl containing 5% pepsin and washed in phosphate-buffered saline for 5 minutes and air dried. The nuclei were postfixed in 4% formaldehyde in phosphate-buffered saline for 7 minutes at room temperature. Then the slides were washed in phosphate-buffered saline, dehydrated in an ethanol series, and air dried. Chromosomal DNA was denatured by immersion in 70% formamide/21 SSC, pH 7.0, at 707C for 5 minutes, followed by dehydration in an ethanol series at 0207C. According to the supplier, probes and hybridization buffer (hybrisol VI, Oncor) were prewarmed at 377C for 5 minutes; denaturation was made at 707C for 5 minutes. Ten to 30 mL probe mixture was added to each slide under coverslip and sealed with rubber cement. The hybridization was performed at 377C, for 16 to 18 hours in a humidified chamber. Posthybridization washings were carried out at 727C in 0.251 SSC, pH 7.0, for 5 minutes. The slides were immersed in 1 1 phosphatebuffered detergent (0.2% Tween/41 SSC) at room temperature for 2 minutes. Digoxigenin-labeled probes were detected with rhodamine-labeled anti-digoxigenin antibodies (Oncor), and, if necessary, signals were amplified as described by the supplier. Cells were examined with an epifluorescence Zeiss Axiophot (Zeiss, Oberkochen, Germany) equipped with a filter combination for 4*,6*-diamidino-2-phenylindole and tetramethylrhodamine isothiocyanate. Photomicrographs were taken with an objective 63 with Kodak Ektachrome 400 ASA films (Eastman Kodak, Rochester, NY). One hundred to 376 interphase nuclei were counted for each analyzed chromosome (Table 2), except for tumors 3 and 5, where fewer nuclei were counted (between 70 and 84). In most cases, one observer examined one half of the hybridization area, and the other examined nuclei in the other half. The number of nuclei containing 0, 1, 2, 3, 4, 5, and more than 5 signals was counted, and the percentage of nuclei was determined for each chromosome. The specificity of each probe was tested on normal lymphocytes. In these cells, the percentage of nuclei with two spots for chromosomes 7, 17, and 20 varied from 80% to 86%, whereas the percentage of nuclei with one spot

WBS: Hepatology

HEPATOLOGY Vol. 23, No. 3, 1996

HAMON-BENAIS ET AL.

431

TABLE 2. FISH Results in Normal Hepatocytes and Tumor Cells Percentage of Centromere Signals per Nucleus (mean 6 2 SEM) Cells

Hepatocytes Chromosome

Tumor 1 Chromosome

Tumor 2 Chromosome

Tumor 3 Chromosome

Tumor 4 Chromosome

Tumor 5 Chromosome

Tumor 6 Chromosome

Tumor 7 Chromosome

3

4

5

5

No. of Examined Nuclei

0

1

2

7 17 20

060 0.3 6 0.4 060

5.0 6 1.9 9.5 6 2.2 7.3 6 3.7

66.4 6 4.0 64.7 6 3.7 62.3 6 7.0

7.1 6 2.2 8.6 6 2.1 6.8 6 3.6

20.6 6 3.4 14.2 6 2.7 22.5 6 6.0

0.9 6 2.4* 2.7 6 1.2* 1.0 6 1.4*

— — —

553 663 191

7 17 20

060 0.4 6 0.7 3.0 6 3.4

8.2 6 5.2 18.8 6 4.8 12.0 6 6.5

81.8 6 7.4 71.8 6 5.5 83.0 6 7.5

8.2 6 5.2 8.3 6 3.4 2.0 6 2.8

1.8 6 2.5 0.7 6 1.0 060

060 060 060

060 060 060

109 266 100

7 17 20

060 0.4 6 0.7 060

0.4 6 0.8 2.6 6 1.9 8.3 6 5.2

25.5 6 5.6 11.3 6 3.8 27.5 6 8.5

16.3 6 4.8 20.8 6 5.0 43.1 6 9.5

31.0 6 6.0 25.0 6 5.3 16.5 6 7.1

10.1 6 3.9 31.8 6 5.7 3.7 6 3.6

16.7 6 4.8 8.2 6 4.7 0.9 6 1.8

239 268 109

7 17 20

060 1.4 6 2.2 3.6 6 4.1

28.7 6 8.7 28.6 6 10.8 23.8 6 9.3

55.6 6 9.5 51.4 6 12.0 60.7 6 10.6

13.0 6 6.5 12.9 6 8.0 8.3 6 6.0

2.7 6 3.1 4.3 6 4.8 3.6 6 4.1

060 060 060

060 1.4 6 4.8 060

108 70 84

7 17 20

060 0.5 6 0.9 9.5 6 3.8

8.8 6 3.8 10.2 6 4.1 33.5 6 6.1

43.1 6 6.7 60.2 6 6.6 43.4 6 6.4

31.5 6 6.3 19.4 6 5.3 8.3 6 3.5

4.6 6 2.8 8.8 6 3.8 3.3 6 2.3

12.0 6 4.4 0.9 6 1.3 2.0 6 1.8

060 060 060

216 216 242

7 17 20

0.8 6 1.6 0.8 6 1.6 1.3 6 2.6

26.0 6 8.0 30.6 6 8.3 18.2 6 8.8

68.1 6 8.5 64.5 6 8.7 67.5 6 10.7

3.4 6 3.3 3.3 6 3.2 7.8 6 6.1

1.7 6 2.4 0.8 6 1.6 3.9 6 4.4

060 060 1.3 6 2.6

060 060 060

119 121 77

7 17 20

0.5 6 0.9 0.4 6 0.8 5.0 6 4.3

46.2 6 6.8 4.2 6 2.6 24.0 6 5.5

50.0 6 6.8 42.7 6 6.4 50.0 6 10.0

2.8 6 2.3 47.3 6 6.5 18.0 6 7.7

0.5 6 0.9 3.3 6 2.3 3.0 6 3.4

060 1.7 6 1.6 060

060 0.4 6 0.8 060

212 239 100

7 17 20

1.8 6 1.3 4.0 6 2.4 0.9 6 1.8

37.2 6 4.9 15.3 6 4.5 18.2 6 7.4

58.3 6 5.0 73.4 6 5.6 69.1 6 8.8

2.7 6 1.7 5.3 6 2.8 1.8 6 2.5

060 2.0 6 1.7 1.8 6 2.5

060 060 060

060 060 060

376 248 110

NOTE. Boxes correspond to significant results (see text). *Percentage corresponding to hepatocytes with $5 spots.

varied from 9% to 11%. These results support already-published data20-22 in which 82% to 98% of nuclei from normal lymphocytes presented two spots and 2% to 14% of nuclei showed one spot. Statistical Analysis of FISH Results Statistical analysis of hybridization results was performed with the Statgraphic system (CTSC, Rockville, MD) using the x2 test to determine if the signal distribution in tumor cells was significantly different from that of normal hepatocytes. Moreover, the results (Table 2) were considered significant for an aneusomy according to the following criteria: For chromosome 7, tumor cells showing one spot were considered to present a chromosome loss (usually a monosomy) when their percentage exceeded 6.9%. This percentage was determined by adding the percentage of normal hepatocytes showing one spot for chromosome probe 7 with 2 standard errors of mean. Tumor cells showing 3, 4, 5, and more spots were considered to present a chromosome gain for chromosome 7 when the percentage exceeded 36.6%. This percentage was determined by adding the percentage of normal hepatocytes

5p0b$$0003

02-16-96 18:02:46

hepa

showing 3, 4, 5, or more spots with 2 corresponding standard errors of mean. The same criteria were used for chromosome 17, where a loss was considered to occur when the percentage of tumor cells showing one spot exceeded 11.7%. For this chromosome, there was a chromosome gain when the percentage of tumor cells showing 3, 4, 5, or more spots exceeded 31.5%. For chromosome 20, a chromosome loss was defined when the percentage of tumor cells showing one spot exceeded 11%, and a chromosome gain was considered to be present when the percentage of tumor cells showing 3, 4, 5, or more spots exceeded 41.3%. When the global test was significant for the chromosome gain, the x2 test was also performed separately for the cells with 3, 4, or 5 spots to confirm if a trisomy, a tetrasomy, or a pentasomy had occurred. A value of P õ .05 was considered statistically significant. RESULTS Conventional Cytogenetic Results

The main difficulty in the conventional cytogenetic study of our tumor cells was obtaining metaphases and

WBS: Hepatology

432 HAMON-BENAIS ET AL.

HEPATOLOGY March 1996

FIG. 1. Representative interphase nuclei processed by fluorescence in situ hybridization. (A, B) Normal hepatocytes showing two spots with the chromosome 7 (A) and 17 (B) centromeric probes. (C, D) Interphase nuclei from tumors showing one or two spots for tumor 1 (focal nodular hyperplasia) (C) and the varying distribution of signals for tumor 2 (D) with a chromosome 17 centromeric probe.

analyzable chromosome preparations in each case. For two tumors (tumors 3 and 5), the metaphases could not be obtained, and for the other tumors, metaphasis analysis was particularly difficult because of unsuccessful G banding. For example, if nine metaphases were obtained in tumor 2, with a broad range of chromosome numbers from 42 to 127 chromosomes per cell (Table 1), only two, with 70 and 99 chromosomes, respectively, were classified according to the usual criteria of chromosome morphology. In these two metaphases, numerous polysomies were demonstrated, particularly for chromosomes 7, 17, and 20. In the five other cases, chromosome numbers varied from 37 to 46 (Table 1), but two metaphases could be classified in tumors 1 and 7, and no abnormalities were observed in these metaphases. FISH Results

For normal interphase hepatocyte nuclei, the results of FISH with specific probes for chromosomes 7, 17, and 20 (Table 2) showed that between 62% and 66% of nuclei had two copies (Fig. 1A and B), and 14% to 22% of nuclei had four copies. Only a few hepatocyte interphase nuclei (0.9% to 2.7%) presented five spots. These results are comparable with those obtained by flow cytometry of the normal human adult liver, in which approximately 55% of hepatocytes are diploid, and the remainder are mostly tetraploid.23,24 Because

5p0b$$0003

02-16-96 18:02:46

hepa

of the difficulty of interpreting the FISH results on a limited number of metaphases, only the results of this technique on the interphase nuclei of tumor cells (Fig. 1C and D and Table 2) are reported here. According to statistical analysis, the signal distribution in the cells from the six cases of HCC and from the case of focal nodular hyperplasia was statistically different (P õ .001) from that of normal hepatocytes for chromosomes 7, 17, and 20. Using criteria described in Materials and Methods for the demonstration of an aneusomy, we attempted to determine if the differences were caused by a loss or a gain of chromosomes by comparing the results in tumor cells with those in normal hepatocytes. Losses in Chromosomes 7, 17, and 20. The results of tumors 3, 5, and 7 for the three chromosome were statistically different (P õ .001) from those of normal hepatocytes (Table 2, Fig. 2); in tumor 5, the difference was also significant for chromosome 20 with, however, a P õ .05. For tumors 4 and 6, the results were statistically different (P õ .001) for chromosome 20 (tumor 4) and for chromosomes 7 and 20 (tumor 6). In the case of focal nodular hyperplasia (tumor 1), a loss of chromosome 17 was observed (P õ .001), whereas no difference was found (P Å .2) for chromosomes 7 and 20. In tumor 2, no loss was observed for the three chromosomes. Gains in Chromosomes 7, 17, and 20. Chromosome gains were observed in tumors 2, 4, and 6 (Table 2, Fig. 2). The results were statistically different (P õ .001) from normal hepatocytes when the global test for 3, 4, 5, and more spots was performed, except for tumor 6, which did not fulfill the conditions defined in Materials and Methods for chromosome 20. When the x2 test was performed separately on cells with 3, 4, or 5 spots, a polysomy was observed in tumor 2, and the results were statistically different (P õ .001) between chromosomes 7 and 17 and normal hepatocytes. For chromosome 20, this tumor showed a gain in cells exhibiting three, five, or more spots (P õ .001), but the test was not statistically different between the cells showing four spots and normal hepatocytes (P Å .2). For tumor 4, trisomies of chromosomes 7 and 17 and a pentasomy of chromosome 7 were observed (P õ .001). For tumor 6, trisomies of chromosomes 17 and 20 were noted with P õ .001 and P õ .01, respectively. The other tumors did not show a chromosome gain. In all cases, the tumors analyzed with our centromeric probes showed numerical chromosome abnormalities. In the six cases of HCC, the loss of one to three chromosomes was observed in five of six tumors, whereas a chromosome gain of two to three chromosomes was observed in only three of six tumors. For chromosome 17, a loss or a gain was present in the six tumors. For tumors 2 and 7, where it was also possible to analyze some metaphases by conventional cytogenetics techniques (Table 1), there was agreement between the chromosome numbers counted in metaphases and FISH results in nuclei (Table 2). Finally, there was no correlation of loss (or gain) with the histopathological Edmondson grades or with the cause of the HCC. However, this series is too small to be conclusive.

WBS: Hepatology

HEPATOLOGY Vol. 23, No. 3, 1996

HAMON-BENAIS ET AL.

433

FIG. 2. Centromeric probe distribution of the number of hybridization signals per nucleus for HCC specimens after FISH in black columns compared with normal hepatocyte nuclei in open columns. Tumor 2 presented gains of chromosomes 7 (A), 17 (B), and 20 (C). Tumor 3 showed losses of chromosome 7 (D), 17 (E), and 20 (F). Tumor 6 showed a loss of chromosome 7 (G), a gain of chromosome 17 (H), and a loss and gain for chromosome 20 (I).

DISCUSSION

This study provides an analysis of seven cases of primary liver tumors (six cases of HCC and one case of benign tumor) by conventional cytogenetic techniques and by FISH using specific centromeric probes for chromosomes 7, 17, and 20. Conventional cytogenetic techniques were unable to analyze numerical chromosomal abnormalities. In contrast, an aneuploidy pattern was detected by FISH and confirmed statistically in all the tumors. In five of the six HCC, chromosomal losses were observed, and chromosomal gains were observed in three cases. A particular case was represented by the benign tumor, in which a chromosomal loss was noted. Numerous studies have been published with FISH using specific centromeric probes to enumerate chromosomes within interphase nuclei and to detect aneusomy in tumoral cells.25-28 As in other solid tumors, the cytogenetic analysis of liver tumors is limited by the

5p0b$$0003

02-16-96 18:02:46

hepa

technical problems29 of obtaining primary cell cultures and mitoses, although several hepatoma cell lines derived from primary tumors have been reported.12-14 Another difficulty of cytogenetic analysis is obtaining a sufficient number of nuclei to perform FISH under optimal conditions on interphase nuclei. This was illustrated by the relatively few nuclei recovered in tumors 3 and 5. However, despite these technical problems, a number of cells, more than usually examined by conventional cytogenetic studies, were analyzed in this work with FISH. Interpretation of FISH with specific probes depends on the correspondence between the number of apparent hybridization signals and the specific target sites of the probes.30 The number of signals can be smaller than the number of target sites if DNA is lost during sample preparations or remains inaccessible to the probe or if signals overlap.30 Although we did not control these points in the current study, our results with normal

WBS: Hepatology

434 HAMON-BENAIS ET AL.

HEPATOLOGY March 1996

lymphocytes were similar to those already published20-22 and confirm that our probes functioned correctly. In the case of liver tumors, where mitoses were very difficult to obtain, the FISH procedure represents an essential technique for counting chromosome number. Although the pericentromeric probes used do not necessarily represent the whole intact chromosome, this technique is considered reliable for this purpose.20,31 In addition, FISH can demonstrate tumor monosomy or trisomy by specifically identifying each chromosome. In contrast, flow cytometry used for tumoral DNA analysis does not distinguish specifically involved chromosomes. Aneuploidy has been reported between in 42% to 92% of cases of HCC with flow cytometry,32-35 and our results for chromosomes 7, 17, and 20 confirm the results of this technique because we observed aneuploidy in all cases of HCC. In the normal human adult liver, hepatocytes are diploid or tetraploid,23,24,36 and our studies by FISH confirmed that a mixture of diploid and tetraploid hepatocytes exist in the normal liver such as we observed with specific probes for chromosomes 7, 17, and 20. Our results with tumor cells showed significant differences when the criteria of statistical analysis described in Materials and Methods were used. In all cases, the tumors analyzed with our centromeric probes showed numerical chromosome abnormalities. Monosomic cells were more frequently observed than trisomic cells in our HCC cases; they were seen in five of six tumors. Tumors 3, 5, and 7 were almost exclusively disomic or monosomic for the three studied chromosomes. We can suppose that the origin of the cancerous process is monoclonal in these cases, from a diploid cell, and that monosomic cells originated from diploid and not from tetraploid cells. Monosomy is known to be one of the mechanisms of tumor suppressor gene loss in malignant tumors.1 Allelic losses of chromosomes 7, 17, and 20 have been found in HCC.37 No monosomy has been reported in the few HCCs cytogenetically analyzed,12,13 except for one case in which a monosomy was observed for chromosome 17.13 The tumor suppressor gene p53 is located in this chromosome.37 Monosomy 7 has been described in hematologic malignancies such as myelodysplastic disorders26,38 as well as in the liver, whereas monosomy 17 has been observed in colorectal tumors39 and monosomy 20 in acute lymphoblastic leukemia.40,41 Trisomic cells were observed in only three of our cases: trisomy 7 and trisomy 20 twice; trisomy 17, three times. A polysomy for the three chromosomes was observed in tumor 2, with a significant increase in cells with 3, 4, 5, or more spots, whereas tumor 4 presented a trisomy and a pentasomy for chromosome 7. In tumor 6, trisomy and monosomy for chromosome 20 coexist: one hypothesis is that some successive mitotic non-disjunctions occur in the cancer, leading in this case to the presence of a monosomic 20 cell near a trisomic 20 cell. The biological significance of the trisomies is different if they occur in a DNA diploid hepatic cell or in a DNA tetraploid hepatic cell. In tumors 4 and 6,

5p0b$$0003

02-16-96 18:02:46

hepa

there were more trisomic cells and fewer cells with two and four spots compared with normal hepatocytes. This could be attributable either to a chromosome loss from original tetraploid cells or to a chromosome gain from diploid cells, the latter possibly by a mechanism of a tetraploidization followed by losses of selected chromosomes. It should be noted that polysomies of chromosome 20 have been observed in some cases of hepatoblastoma42-44 and polysomies of chromosome 7 in other extrahepatic tumors.39 An advantage of interphase cytogenetics lies in the possibility of defining the part of tumor cell heterogeneity and of quantifying the aberrant clones. In most of our cases (cases 3, 4, 5, 6, 7), there is a high percentage of disomic cells for the three studied chromosomes near the aneusomic population. It could be a normal diploid cell population, which reflects contamination by nonneoplastic elements (mononuclear cells, fibroblasts, endothelial cells, and perhaps also normal peritumoral hepatocytes). A loss of chromosome 17 was observed in our case of focal nodular hyperplasia (tumor 1). It is interesting because several specific chromosomal abnormalities have been reported in various types of benign tumors.45 Particularly a chromosome 17 loss has been reported in other benign epithelial tumors, such as in one case of breast hyperplasia,27 suggesting that some hyperplasias may be part of a sequence of progression to malignancy in breast cancer, and also in a tubulovillous adenoma of the colon.46 However, focal nodular hyperplasia is a benign liver tumor that generally does not become malignant.47 Nevertheless, this observation needs confirmation. The current study shows the feasibility of studying numerical chromosome aberrations by interphase cytogenetics in HCC in the absence of karyotype results. Abnormalities of chromosomes 7, 17, and 20 were present in all tumors. Other studies would be necessary, for example, using protooncogenes probes for detecting some gene amplifications. Examination of other specimens of HCC is obviously necessary, and we expect that the FISH procedure will provide important information for the understanding of hepatocarcinogenesis. Acknowledgment: We thank Dr. Yvon Calmus for providing human hepatocytes. We are grateful to Alain Loiseau and Myriam Rosenberg-Bourgin for their contribution to statistical analysis. REFERENCES 1. Bishop JM. Molecular themes in oncogenesis. Cell 1991;64:235248. 2. Weinberg RA. The integration of molecular genetics into cancer management. Cancer 1992;70:1653-1658. 3. Solomon E, Borrow J, Goddard AD. Chromosome aberrations and cancer. Science 1991;254:1153-1160. 4. Mitelman F, Johansson B, Mertens F. Catalogue of chromosomal aberrations in cancer. Ed 5. New-York: Wiley-Liss, 1994. 5. Tiollais P, Pourcel C, Dejean A. The hepatitis B virus. Nature 1985;317:489-495. 6. Kew MC. Hepatitis C virus and hepatocellular carcinoma. FEMS Microbiol Rev 1994;14:211-220.

WBS: Hepatology

HEPATOLOGY Vol. 23, No. 3, 1996

HAMON-BENAIS ET AL.

7. Sherlock S, Dooley J. Diseases of the liver and biliary system. Ed. Cambridge, MA: Blackwell Scientific, 1993:503-531. 8. Harris CC. Chemical and physical carcinogenesis: advances and perspectives for the 1990s. Cancer Res 1991;51:5023-5044. 9. Nishida N, Fukuda Y, Ishizaki K. Molecular aspects of hepatocarcinogenesis and their clinical implications. Int J Oncol 1994;4:615-622. 10. Fujimori M, Tokino T, Hino O, Kitagawa T, Imamura T, Okamoto E, Mitsunobu M, et al. Allelotype study of primary hepatocellular carcinoma. Cancer Res 1991;51:89-93. 11. Nishida N, Fukuda Y, Kokuryu H, Sadamoto T, Isowa G, Honda K, Yamaoka Y, et al. Accumulation of allelic loss on arms of chromosomes 13q, 16q and 17p in the advanced stages of human hepatocellular carcinoma. Int J Cancer 1992;51:862-868. 12. Simon D, Munoz SJ, Maddrey WC, Knowles BB. Chromosomal rearrangements in a primary hepatocellular carcinoma. Cancer Genet Cytogenet 1990;45:255-260. 13. Bardi G, Johansson B, Pandis N, Heim S, Mandahl N, Andre´nSandberg A, Ha¨gerstrand I, et al. Cytogenetic finding in three primary hepatocellular carcinomas. Cancer Genet Cytogenet 1992;58:191-195. 14. Werner M, Nolte M, Georgii A, Klempnauer J. Chromosome 1 abnormalities in hepatocellular carcinoma. Cancer Genet Cytogenet 1993;66:130. 15. Edmondson HA. Tumors of the liver and intrahepatic bile ducts. Section VII, fascicle 25. Washington, DC: Armed Forced Institute of Pathology, 1958. 16. Waye JS, England SB, Willard HF. Genomic organization of alpha satellite DNA on human chromosome 7: evidence for two distinct alphoid domains on a single chromosome. Mol Cell Biol 1987;7:349-356. 17. Willard HF, Waye JS. Hierarchical order in chromosome-specific human alpha satellite DNA. Trends Genet 1987;3:192-198. 18. Wevrick R, Willard HF. Long-range organization of tandem arrays of alpha satellite DNA at the centromeres of human chromosomes: high-frequency array-length polymorphism and meiotic stability. Proc Natl Acad Sci U S A 1989;86:9394-9398. 19. Pinkel D, Straume T, Gray JW. Cytogenetic analysis using quantitative, high-sensitivity, fluorescence hybridization. Proc Natl Acad Sci U S A 1986;83:2934-2938. 20. Kiechle-Schwarz M, Decker H-JH, Berger CS, Fiebig HH, Sandberg AA. Detection of monosomy in interphase nuclei and identification of marker chromosomes using biotinylated alpha-satellite DNA probes. Cancer Genet Cytogenet 1991;51:23-33. 21. Poddighe P, Moesker O, Smeets D, Awwad BH, Ramaekers FCS, Hopman AHN. Interphase cytogenetics of hematological cancer: comparison of classical karyotyping and in situ hybridization using a panel of eleven chromosome specific DNA probes. Cancer Res 1991;51:1959-1967. 22. Balazs M, Mayall BH, Waldman FM. Interphase cytogenetics of a male breast cancer. Cancer Genet Cytogenet 1991;55:243-247. 23. Saeter G, Lee C-Z, Schwarze PE, Ous S, Chen DS, Sung JL, Seglen PO, et al. Changes in ploidy distributions in human liver carcinogenesis. J Natl Cancer Inst 1988;80:1480-1485. 24. Feldmann G. Liver ploidy. J Hepatol 1992;16:7-10. 25. Brown JA, Alcaraz A, Takahashi S, Persons DL, Lieber MM, Jenkins RB. Chromosomal aneusomies detected by fluorescent in situ hybridization analysis in clinically localized prostate carcinoma. J Urol 1994;152:1157-1162. 26. Brizard F, Brizard A, Guilhot F, Tanzer J, Berger R. Detection of monosomy 7 and trisomies 8 and 11 in myelodysplastic disorders by interphase fluorescent in situ hybridization: comparison with acute non lymphocytic leukemia. Leukemia 1994;8:10051011. 27. Micale MA, Visscher DW, Gulino S, Wolman SR. Chromosomal aneuploidy in proliferative breast disease. Hum Pathol 1994; 25:29-35.

5p0b$$0003

02-16-96 18:02:46

hepa

435

28. Hopman HAN, Moesker O, Smeets AWGB, Pauwels RPE, Vooijs GP, Ramaekers FCS. Numerical chromosome 1, 17, 9 and 11 aberrations in bladder cancer detected by in situ hybridization. Cancer Res 1991;51:644-651. 29. Teissier JR. Analysis of human solid tumor. Cancer Genet Cytogenet 1989;37:103-125. 30. Matsumura K, Kallioniemi A, Kallioniemi O, Chen L, Smith HS, Pinkel D, Gray J, et al. Deletion of chromosome 17p loci in breast cancer detected by fluorescence in situ hybridization. Cancer Res 1992;53:3474-3477. 31. Chen Z, Morgan R, Berger CS, Sandberg AA. Application of fluorescence in situ hybridization in hematological disorders. Cancer Genet Cytogenet 1992;63:62-69. 32. Fujimoto J, Okamoto E, Yamanaka N, Toyosaka A, Mitsunobu M. Flow cytometric DNA analysis of hepatocellular carcinoma. Cancer 1991;67:939-944. 33. Chen MF, Hwang TL, Tsano KC, Sun CF, Chen TJ. Flow cytometric DNA analysis of hepatocellular carcinoma: preliminary report. Surgery 1991;109:455-458. 34. Hamazaki K, Kato T, Yunoki Y, Mori M, Gochi A, Mimura H, Orita K. An analysis of DNA ploidy pattern of hepatocellular carcinoma. Acta Med Okayama 1993;47:413-416. 35. Cottier M, Jouffre C, Maubon I, Sabido O, Barthelemy C, Cuilleron M, Laurent JL, et al. Prospective flow cytometric DNA analysis of hepatocellular carcinoma specimens collected by ultrasound-guided fine needle aspiration. Cancer 1994;74:599605. 36. Kudryavtsev BN, Kudryavtseva MV, Sakuta GA, Stein GI. Human hepatocyte polyploidization kinetics in the course of life cycle. Virchows Arch B Cell Pathol 1993;64:387-393. 37. Nishida N, Fukuda Y, Ishizaki K. Molecular aspects of hepatocarcinogenesis and their clinical implications. Int J Oncol 1994;4:615-622. 38. Flactif M, Lai JL, Preudhomme C, Fenaux P. Fluorescence in situ hybridization improves the detection of monosomy 7 in myelodysplastic syndromes. Leukemia 1994;8:1012-1018. 39. Couturier-Turpin MH, Couturier D, Nepveux P, Louvel A, Chapuis Y, Guerre J. Human chromosome analysis in 24 cases of primary carcinoma of the large intestine: contribution of the Gbanding technique. Br J Cancer 1982;46:856-869. 40. Silengo M, Vassallo E, Barisone E, Miniero R, Madon E. Monosomy 20 in childhood acute lymphoblastic leukemia. Cancer Genet Cytogenet 1992;59:177-179. 41. Bonet C, Sole F, Woessner S, Florensa L, Besses C, Sans-Sabrafen J. A case of monosomy 20 in an adult acute lymphoblastic leukemia. Cancer Genet Cytogenet 1993;6:165. 42. Mascarello JT, Jones MC, Kadota RP, Krous HF. Hepatoblastoma characterized by trisomy 20 and double minutes. Cancer Genet Cytogenet 1990;47:243-247. 43. Soukup SW, Lampkin BL. Trisomy 2 and 20 in two hepatoblastomas. Genes Chromosom Cancer 1991;3:231-234. 44. Vijay ST, Wilson KS, Timmons CF, Schneider NR. Trisomy 2, trisomy 20, and del(17p) as sole chromosomal abnormalities in three cases of hepatoblastoma. Genes Chromosom Cancer 1994;11:199-202. 45. Mitelman F, Kaneko Y, Trent J. Report of the committe on chromosomes changes in neoplasia. Cytogenet Cell Genet 1991;58: 1053-1079. 46. Couturier-Turpin MH, Louvel A, Couturier D, Esnous C, Poirier Y, Nepveux P. Tubulovillous adenoma of the colon with hyperdiploidy, double-minute chromosome, and inversion of chromosome 1. Cancer Genet Cytogenet 1988;32:253-262. 47. Thung SN, Gerber MA. Development of malignancy in benign tumors of the liver. In: Tabor E, Di Bisceglie AM, Purcell RH, eds. Etiology, pathology, and treatment of hepatocellular carcinoma in North America. Ed 1. Houston: Gulf Publishing Company, 1991:209-215.

WBS: Hepatology