Abnormal Chromosome 8 Copy Number in Stage I to Stage IV Breast Cancer Studied by Fluorescence In Situ Hybridization

Abnormal Chromosome 8 Copy Number in Stage I to Stage IV Breast Cancer Studied by Fluorescence In Situ Hybridization Hon Fong L. Mark, William Taylor, Stephen Brown, Mangala Samy, Ci-Lin Sun, Kathleen Santoro, and Kirby I. Bland

ABSTRACT: To test the hypothesis that the frequency of abnormal chromosome 8 copy number increases with the severity of the disease as defined by an increase in clinical stage, we conducted a fluorescence in situ hybridization (FISH) study of a sample of 42 breast cancer specimens utilizing a protocol that was optimized by our laboratory. Cytogenetic results, obtained from blinded analyses of archival specimens, demonstrated that the higher clinical stages (i.e., stages III and IV) yield higher frequencies of abnormal chromosome 8 copy number. Specifically, 45.45% and 50% of the stage I and stage II cases, respectively, were abnormal, whereas 63.64% and 60% of the stage III and stage IV cases, respectively, were abnormal for chromosome 8 copy number. The overall frequency of abnormal chromosome 8 copy number was 54.76% (23 of 42 tumors studied). When the results of a control probe were taken into account, 34.78% (8 of 23) of the abnormal cases were trisomic, whereas the remaining cases were likely triploid. Thus, the present data not only established that chromosome 8 trisomy is a recurrent finding in breast cancer, but also confirmed a higher frequency of occurrence of abnormal chromosome 8 copy number with the higher clinical stages. Future experiments utilizing additional specimens in this laboratory and from other laboratories are necessary to confirm and extend the findings of the present study. © Elsevier Science Inc., 1998

INTRODUCTION Recent developments of molecular cytogenetic techniques such as fluorescence in situ hybridization (FISH) offer an unprecedented opportunity for the study of abnormal chromosome copy number in cancer cells. The major advantage of FISH is that it is not restricted to fresh tumor tissues, but can also be used on more plentiful archival materials. Trisomy 8 is a frequent acquired chromosomal abnormality encountered most notably in the myelodysplastic syndromes as well as other hematopoietic disorders. It has also been reported in breast and other cancers. These include lymphoma, melanoma, sarcoma, fibrosarcoma, Wilms tumors, desmoid tumors, mesoblastic nephroma, leiomyosarcoma, mediastinal germ cell tumor, prostate carcinoma,

From the Lifespan Academic Medical Center Cytogenetics Laboratory (H. F. L. M., S. B., M. S., C.-L. S., K. S.), the Department of Pathology (H. F. L. M., W. T.), and the Department of Surgery (K. I. B.), Rhode Island Hospital, and Brown University School of Medicine (H. F. L. M., K. I. B.), Providence, Rhode Island, USA. Address reprint requests to: Dr. Hon Fong Louie Mark, Director, Laboratory of Cytogenetics, FISH and Genotoxicology, Rhode Island Hospital, 593 Eddy Street, Providence, RI 02903. Received February 11, 1998; accepted May 13, 1998. Cancer Genet Cytogenet 108:1–5 (1999)  Elsevier Science Inc., 1998 655 Avenue of the Americas, New York, NY 10010

ovarian carcinoma, salivary gland tumors, squamous cell carcinoma of the head and neck, benign nevi, colon adenoma, solitary fibrous tumor, endometrial tumors, Ewing’s sarcoma, and gestational trophoblastic disease. The long-term goal of the present research project is to identify a subset of breast cancer characterized by chromosome 8 trisomy using FISH on formalin-fixed, paraffinembedded tissues. Pilot studies reported by our laboratory [1, 2] indicated that trisomy 8 is a recurrent, numerical chromosomal abnormality found in infiltrating ductal carcinoma and is associated more often with stage II than with stage I disease. Subsequent experiments [3] confirmed and extended the finding of a higher frequency of trisomy 8 in the higher clinical stage tumors, as reported in our first study. The present study was carried out to test the hypothesis that this trend continues for stages I and II breast cancer. In addition, the analysis in stages I and II breast cancers was extended to stages III and IV breast disease in this study. MATERIALS AND METHODS Study Cases Cases of stages I–IV breast cancer were identified via computer searches for this retrospective study. Formalin-fixed,

0165-4608/99/$–see front matter PII S0165-4608(98)00112-5

2 paraffin-embedded tissue blocks were retrieved from the routine surgical pathology files of the Department of Pathology at Rhode Island Hospital. Identified sections were cut and applied directly to silanized glass slides. At least one slide from the consecutive sections was stained by hematoxylin & eosin and evaluated for histological grade according to the Bloom and Richardson [4] criteria. The other adjacent sections were then processed for FISH. FISH Protocol and Photography For FISH modifications of the procedures of Pinkel et al. [5], Mark [6], Miranda et al. [7], Afify et al. [1], Afify and Mark [2] and Mark et al. [3, 8] as well as manufacturer’s instructions (Oncor, Gaithersburg, MD) were followed. Briefly, formalin-fixed, paraffin-embedded, 5–7-micron-thick tissue sections were applied to silanized slides. Slides were then baked at 608C for 60 minutes. Specimens were deparaffinized in xylene and fixed in 100% ethanol for 10 minutes each. Pretreatment with 30% bisulfide sodium in 2 3 SSC (pH 7.0) for 15–18 minutes was followed by 1-minute washes of 2 3 SSC (pH 7.0) and 2 minutes of dehydration in an increasing series of 70%, 80%, 90%, and 100% ethanol solutions. Treatment with proteinase (400 ml of a stock solution, 25 mg/mL), in 40 mL 2 3 SSC (pH 7.0) for 18 minutes was followed by three washes in 2 3 SSC (pH 7.0) for 1 minute each and by dehydration with the same series of ethanol solutions mentioned above, plus 100% acetone for 2 minutes. Hybridization was performed with a mixture of 2 ml of a chromosome 8, a-satellite digoxigenin-FITC probe and 30 ml of Hybrisol VI (Oncor, Gaithersburg, MD) per specimen, which was then covered with a glass coverslip and sealed. The specimens were denatured at 908C for 12 minutes and left in a humidified chamber at 378C overnight to hybridize. The following day the slides were post-washed twice in a 50% formamide/2 3 SSC solution (pH 7.0) at 438C for 15 minutes each, in 0.1 3 SSC (pH 7.0) for 30 minutes at 438C and in a final wash in 1 3 PBD (phosphate buffered detergent). For detection, slides were treated with 50 ml of fluorescein-labeled anti-digoxigenin at 378C for 20 minutes. The interphase nuclei were counterstained with 20 ml of a 1:1 dilution of propidium iodide/antifade for a final concentration of 1.25 mg/mL per slide. The slides were covered with glass coverslips and examined under a Zeiss epifluorescence photomicroscope using an FITC exciter filter set. Photography was performed using Ektachrome ISO 400 color slide film. Signal Scoring Scoring was performed independently by cytogenetic technicians without any clinical information or knowledge of other histological and pathological results at the time of scoring. Non-overlapping nuclei with well-defined nuclear outlines were chosen for scoring fluorescent signals, following the guidelines adopted by Mark et al. [9– 13], Kim et al. [14], and Hopman et al. [15]. Only bright, punctate signals were scored as positive. Small, pale, irregular, and ambiguous fluorescent deposits in the nuclear area were discarded. Because the nuclei were usually

H. F. L. Mark et al. present in slightly different planes of the section, it was necessary to refocus the microscope constantly so that all positive signals could be included in the scoring. In general, attempts were made to score 400 nuclei per specimen, although scoring of large numbers of nuclei for each specimen was not always possible.

Table 1 Results of scoring for chromosome 8 copy number using FISH with a chromosome 8-specific α-satellite probe on formalin-fixed paraffin-embedded tumor tissue Case no. Stage I 96-913-A8 96-1310-A3 96-6773-A3 97-5058-B1 96-7158-A2 96-7053-A1 95-16855-A4 96-4826-B29 97-2942-A5 97-151-A4 97-3934-A1 Stage II 97-4826-A4 97-4033-A1 97-5821-A8 93-9493-B1 96-6510-A1 97-163-A5 97-4222-A8 97-1025-A2 97-2788-A2 97-6040-B10 Stage III 97-5007-A4 97-4909-A3 97-66-B1 97-15757-A5 96-6727-A5 96-4520-A2 96-5613-A3 97-4409-A9 96-5028-A1 96-5187-A1 97-6896-A1 Stage IV 96-7276-A2 96-15015-A1 96-4979-A1 95-280-A1 96-12148-A2 95-2580-A6 95-11771-A1 96-187-A5 96-8448-B5 96-13089-A4 a

91.2% monosomy.

Total # cells scored

Cells with 3 signals (%)

470 432 237 434 166 269 428 415 484 97 170

0 0.7 4.2 4.4 7.2 8.2 10.5 16.6 21.5 26.8 38.2

308 450 226 200 420 764 206 424 223 294

0 0.4 4.4 7.0 8.3 11.8 16.5 22.0 27.4 27.9

465 202 218 200 209 400 400 415 341 229 444

0.9 6.9 6.9 9.0 10.5 12.3 13.5 16.1 19.4 29.0 29.5

400 244 477 200 243 345 330 437 270 200

0a 1.2 2.5 4.0 14.0 14.5 15.3 15.3 18.5 33.0

3

FISH Study of Stage I–IV Breast Cancer Table 2 Summary of the distributions of chromosome copy number in stages I–IV breast cancer Chromosome 8 copy number Normal (2 signals) Abnormal (3 signals) Total

Clinical Stage I

II

III

IV

Total

6

5

4

3

18

5 11

5 10

7 11

6 10a

23 42a

a One case of a stage IV tumor was found to be monosomic with 91.25% of the cells showing only one fluorescent signal.

RESULTS A total of 42 specimens of formalin-fixed, paraffin-embedded breast cancer tissues were analyzed cytogenetically in a blinded fashion for the frequency of abnormal chromosome 8 copy number using FISH and the previously described protocol optimized by our laboratory. Of the 42 cases studied, 11 were stage I, 10 were stage II, 11 were stage III, and 10 were stage IV. The results of FISH scoring are given in Table 1. Cases with 10% or greater cells with three signals were counted as abnormal, presumably trisomic, whereas those

with less than 10% cells with three signals were counted as normal, or disomic, consistent with the criterion used in our previous studies of chromosome 8 copy number. Although this criterion is not as stringent as the one used previously for tumors that we studied for the first time, it is a conservative approach and is more stringent than most of those adopted in the literature, as discussed in Mark et al. [3]. Using the 10% cutoff value, the overall frequency of abnormal chromosome 8 copy number among stage I tumors was calculated to be 45.45% (5 of 11 tumors). The frequency of abnormal chromosome 8 copy number among stage II tumors was 50% (5 of 10 tumors). The frequency of abnormal chromosome 8 copy number among stage III tumors was 63.64% (7 of 11 tumors). The frequency of abnormal chromosome 8 copy number among stage IV tumors was 60% (6 of 10 tumors). In addition, in one case of a stage IV tumor, 91.25% of the cells showed chromosome 8 monosomy, which leads to a frequency of combined aneusomy (combined trisomy and/or triploidy and monosomy) of 70% (7 of 10 tumors). The overall frequency of abnormal chromosome 8 copy number not including the monosomy case was 54.76% (23 of 42 tumors studied). The chromosome 8 copy numbers of the tumors in the different stages are summarized in Table 2. Representative

Figure 1 Fluorescence in situ hybridization using a chromosome 8-specific a-satellite probe on formalin-fixed, paraffin-embedded breast tumor tissues, demonstrating the presence of abnormal chromosome 8 copy number.

4 interphase cells exhibiting abnormal chromosome 8 copy number are shown in Figure 1. In previous studies, when a case was scored as “trisomic,” control probes were used whenever additional slides were available in order to determine whether the case was indeed trisomic or more likely triploid. If the case were truly trisomic, then one would expect three signals using the chromosome 8 probe and two signals when the control probes were used. If the case were triploid (or in the rare instance of the case being doubly or triply trisomic for the other control chromosomes) one would expect three signals on all probes used. However, as noted previously [3], the presence of triploidy or tetraploidy would in no way invalidate the conclusions drawn, but may only require their modification. An abnormal copy of chromosome 8, irrespective of ploidy background, may be associated with a subset of breast cancer associated with a poorer prognosis. When the data on a control (chromosome 17 centromere enumeration) probe for the cases were taken into account, it was found that of the 23 cases with abnormal copies of chromosome 8 (3 signals), 8 (34.78%) were truly trisomic, yielding three signals for chromosome 8 but only two signals for chromosome 17. The rest were most likely triploid. Thus, the overall frequency of chromosome 8 trisomy was found to be 19.04% (8 of 42), even when a conservative standard (10% cutoff value) was used.

DISCUSSION Our initial study of breast cancer [1] involved 30 informative cases of archival patient materials. In that initial study, we found that 36% of the infiltrating ductal carcinomas (NOS) were trisomic, whereas 30% of the ductal carcinoma in situ were trisomic. Furthermore, all cases of lobular carcinoma, papillary carcinoma, and the benign lesions were found to be normal (disomic) for the number of copies of chromosome 8. In a second paper [2] Afify and Mark studied 34 cases of stages I and II infiltrating ductal carcinoma of the breast, again using FISH and a chromosome 8 specific, a-satellite probe. In that second study, we found that 24% of the stage I tumors were trisomic, whereas 82% of the stage II tumors were trisomic. A third study was performed to confirm and extend these results in stages I and II disease. Thirty-six additional specimens of formalin-fixed, paraffin-embedded breast cancer tissues were analyzed cytogenetically in a blinded fashion for the frequency of abnormal chromosome 8 copy number using FISH and the previously described protocol optimized for our laboratory. The overall frequency of interphase cells with three chromosome 8 centromere signals was 28% for stage I and 61% for stage II tumors. Thus, the observed trend was consistent with those reported in our initial study of stages I and II disease, indicating that a higher frequency of aberrant chromosome 8 copy number occurs with a higher clinical stage as compared with a lower stage. In addition, that study confirmed and firmly established that chromosome 8 trisomy is a recurrent finding in a subset of breast cancer.

H. F. L. Mark et al. The present study was conducted to extend the findings in stages I and II breast cancer to higher stages of disease. The results of the present study are summarized in Tables 1 and 2. Based on a 10% cutoff value, the overall frequency of abnormal chromosome 8 copy number among stage I tumors was calculated to be 45.45%, as can be seen from the tables. The frequency of abnormal chromosome 8 copy number among stage II tumors was 50%. The frequency of abnormal chromosome 8 copy number among stage III tumors was 63.64%. The frequency of abnormal chromosome 8 copy number among stage IV tumors was 60%, while the frequency of combined aneusomy for chromosome 8 was 70%. The overall frequency of abnormal chromosome 8 copy number was 54.76%. Although a strictly linear relationship cannot be unequivocally established from this limited sample, the trend of increasing frequency of abnormal chromosome 8 copy number with higher clinical stage was clearly evident in this sample. In addition, it is of interest to note that when stages I and II cases were combined, the frequency of abnormal chromosome 8 copy number was 47.62% (10 of 21 tumors). Analogously, when stages III and IV were combined, the frequency of abnormal chromosome 8 copy number was 61.90% (13 of 21 tumors). The finding of a higher frequency of abnormal chromosome 8 copy number with the higher clinical stages was consistent with the results of previous breast cancer studies. In conclusion, the present data not only establish that abnormal chromosome 8 copy number is a recurrent finding in a subset of breast cancer, but also confirm that higher frequencies of abnormal chromosome 8 copy number occur with higher clinical stages. Future experiments utilizing additional specimens in this laboratory, as well as in other laboratories are necessary to confirm and extend the findings of the present study. We would like to thank Dr. Roger Mark for reading the manuscript and Mr. William Campbell for occasional technical assistance while he was training in this laboratory. The continued support of Dr. Roger Mark and the dedicated staff of the Lifespan Academic Medical Center Cytogenetics Laboratory, which celebrates its third decade of service, is also acknowledged.

REFERENCES 1. Afify A, Bland KI, Mark HFL (1996): Fluorescent in situ hybridization assessment of chromosome 8 copy number in breast cancer. Breast Cancer Res Treat 38:201–208. 2. Afify A, Mark HFL (in press): FISH assessment of chromosome 8 copy number in stage I and stage II infiltrating ductal carcinoma of the breast. Cancer Genet Cytogenet. 3. Mark HFL, Taylor W, Afify A, Riera D, Rausch M, Huth A (in press): Stage I and stage II infiltrating ductal carcinoma of the breast analyzed for chromosome 8 copy number using FISH. Pathobiology. 4. Bloom HJG, Richardson WW (1957): Histological grading and prognosis in breast cancer. A study of 1409 cases of which 359 have been followed for 15 years. Br J Cancer 11:359–377. 5. Pinkel D, Strawne R, Gray J (1986): Cytogenetic analysis using quantitative high sensitivity, fluorescence hybridization. Proc Nat Acad Sci USA 83:2934–2938.

FISH Study of Stage I–IV Breast Cancer 6. Mark HFL (1994): Fluorescent in situ hybridization as an adjunct to conventional cytogenetics. Ann Clinic Lab Sci 24:153–163. 7. Miranda RN, Mark HFL, Medeiros LJ (1994): Fluorescent in situ hybridization in routinely processed bone marrow aspirate clot and core biopsy sections. Am J Pathol 145:1–6. 8. Mark HFL, Grollino MG, Sulaiman RA, Lathrop JC (1995): Fluorescent in situ hybridization (FISH) assessment of chromosome copy number in gestational trophoblastic disease. Ann Clin Lab Sci 25:291–296. 9. Mark HFL, Mills D, Kim E, Santoro K, Quddus M, Lathrop JC (1996): Fluorescent in situ hybridization assessment of chromosome copy number in buccal mucosal cells. Cytobios 87:117–126. 10. Mark HFL, Mills D, Santoro K, Quddus M, Lathrop JC (1997): Fluorescent in situ hybridization (FISH) analysis of cervical smears: a pilot study of 20 cases. Ann Clin Lab Sci 27:224– 229. 11. Mark HFL, Afify A, Taylor W, Santoro K, Lathrop JC (1997):

5 A subset of gestational trophoblastic disease (GTD) characterized by abnormal chromosome 8 copy number detected by FISH? Cancer Genet Cytogenet 96:1–6. 12. Mark HFL, Huth A, Santoro K, Ferreira K, Barker B (1997): Sequential flow cytometry and FISH for the study of formalin-fixed paraffin-embedded breast cancer tissues. Cancer Genet Cytogenet 98:1–5. 13. Mark HFL, Rehan J, Mark S, Santoro K, Zolnierz K (1997): FISH analysis of single-cell trisomies for determination of clonality. Cancer Genet Cytogenet 100:1–5. 14. Kim SY, Lee JS, Ro JY, Gay ML, Hong WK, Hittleman WN (1993): Interphase cytogenetics in paraffin sections of lung tumors by non-isotopic in situ hybridization: mapping/genotype heterogeneity. Am J Pathol 142:307–317. 15. Hopman AHN, Van Hooren E, Van de Kaa, Vooijs CA, Ramaekers FCS (1991): Detection of numerical chromosome aberrations using in situ hybridization in paraffin sections of routinely processed bladder cancer. Mod Pathol 4:503–513.