- Email: [email protected]
DNA in situ hybridization (interphase cytogenetics) versus comparative genomic hybridization (CGH) in human cancer: detection of numerical and structural chromosome aberrations
acta histochem. 102, 85±94 (2000) Ó Urban & Fischer Verlag
DNA in situ hybridization (interphase cytogenetics) versus comparative genomic hybridization (CGH) in human cancer: detection of numerical and structural chromosome aberrations Herman Van Dekken, Pieter J. Krijtenburg, and Janneke C. Alers Department of Pathology, Josephine Nefkens Institute, Erasmus University Rotterdam, P. O. Box 17 83, 3000 DR Rotterdam, The Netherlands Received 22 September 1999 and in revised form 15 November 1999; accepted 20 November 1999
Summary DNA in situ hybridization techniques for cytogenetic analyses of human solid cancers are nowadays widely used for diagnostic and research purposes. The advantage of this methodology is that it can be applied to cells in the interphase state, thereby circumventing the need for high-quality metaphase preparations for karyotypic evaluation. In situ hybridization (ISH) with chromosome specific (peri)centromeric DNA probes, also termed ªinterphase cytogeneticsº, can be used to detect numerical changes, whereas comparative genomic hybridization (CGH) discloses chromosomal gains and losses, i.e. amplifications and deletions. We wanted to compare both methods in human solid tumors, and for this goal we evaluated ISH and CGH within a set of 20 selected prostatic adenocarcinomas. Chromosomes 7 and 8 were chosen for this analysis, since these chromosomes are frequently altered in prostate cancer. ISH with chromosome 7 and 8 specific centromeric DNA probes was applied to standard, formalinfixed and paraffin-embedded, histological sections for numerical chromosome analysis. CGH with DNA's, extracted from the same histologic area of the archival specimens, was used for screening of gains and losses of 7 and 8. ISH with centromeric probes distinguished a total of 26 numerical aberrations of chromosome 7 and/or 8 in the set of 20 neoplasms. In the same set CGH reCorrespondence to: Dr. H. Van Dekken, phone: +31-10/4 08 79 01, fax: +31-10/4 08 94 87, e-mail: [email protected]. http://www.urbanfischer.de/journals/actahist
0065±1281/00/102/1±085 $ 12.00/0
86
H. Van Dekken et al.
vealed a total of 35 losses and gains. CGH alterations of 7 and 8 were seen in twenty-two of the 26 chromosomes (85%) that showed aberrations in the ISH analysis. Concordance between ISH and CGH was seen in 11 (of 26; 42%) chromosomes. Eight chromosomes were involved in gains (5 ´ #7, 3 ´ #8), three in losses (3 ´ #8). This included both complete (3/11) and partial (8/11) CGH confirmation of the numerical alteration. Partial CGH confirmation was defined as loss or gain of a chromosome arm with involvement of the centromeric region. In the majority of these cases it concerned a whole chromosome arm, mostly the long arm. We conclude that generally a fair correlation was found between ISH and CGH in interphase preparations of a series of prostate cancers. However, when specified in detail, most of the numerical ISH aberrations were only partly represented in the CGH analysis. On the one hand, it suggests that CGH does not adequately discriminate numerical abnormalities. On the other hand, it likely implies that not all numerical changes, as detected by interphase cytogenetics, are truly involving the whole chromosome. A part of these discrepancies might be caused by structural mechanisms, most notably isochromosome formation. Key words: in situ hybridization ± comparative genomic hybridization ± chromosome aberration ± cancer
Introduction Despite the frequent incidence of human solid tumors, relatively little is known about the cytogenetic aberrations that characterize them. Mostly, technical problems have for a long time hampered the acquisition of cytogenetic data on solid tumors (Sandberg, 1990). Solid tumor biopsies yield low numbers of viable cells, have a low mitotic index and frequently contain non-neoplastic cells. Furthermore, the quality of metaphases is relatively poor. Therefore, interphase approaches, both at molecular and cytological levels, are better equipped to disclose genetic changes occurring in human epithelial cancers. Over the past decade, interphase hybridization technology, i.e. DNA in situ hybridization (ISH; reviewed in e. g. Van Dekken et al., 1997), and comparative genomic hybridization (CGH; first introduced by Kallioniemi et al. in 1992), have emerged as important instruments for the delineation of genetic changes in human solid cancers. ISH can be used for rapid detection of numerical chromosome changes, whereas CGH is optimally suited for evaluation of chromosomal imbalances, i.e. deletions and amplifications. Results from molecular and karyotyping studies can be confirmed in situ and, in addition, new genetic aberrations are being disclosed. For example, in lung cancer ISH with region-specific probes showed loss of the p15/p16 region at 9p, which is consistent with allelotyping studies (Xiao et al., 1995). In head and neck tumors, CGH studies repeatedly identified a new amplified chromosomal region, 3q26-qter (Brzoska et al., 1995; Speicher et al., 1995). This region has
ISH versus CGH in cancer
87
not been described before and may harbor a new oncogene that is important in these neoplasms. ISH on colon cancer specimens have confirmed the presence of numerical aberrations that are frequently encountered in colorectal tumors, such as trisomy 7 and monosomy 18 (Herbergs et al., 1994). It has also added some new numerical aberrations, e. g. gain of chromosome 12 (Steiner et al., 1993). In prostatic adenocarcinoma, ISH analysis rendered possible markers for biological aggressiveness, such as gain of chromosomes 7 and 8 (Alcaraz et al., 1994; Alers et al., 1997). Additionally, CGH provided some new regions of interest, most notably the amplification of the 8 q arm (Visakorpi et al., 1995). Thus, it can be stated that ISH and CGH have established their importance in solid cancer genetics. In this report, we wanted to compare ISH and CGH in human solid cancers. Sofar, such a study is lacking in the literature. A well-defined set of prostatic adenocarcinomas was used for this purpose (Alers et al., 1997). The following issues were specifically addressed: 1] What is the sensitivity of ISH versus CGH in detecting numerical chromosomal aberrations? and 2] Are alterations detected by ISH indeed caused by changes in the copy number of a specific chromosome?
Material and methods We collected representative paraffin blocks from 20 prostatic adenocarcinoma specimens (6 primary tumors, 6 lymph node metastases, 8 distant metastases). The 20 tumors were taken from a previously reported series, since they were found to be altered for chromosome 7 and/or 8 by DNA in situ hybridization (Alers et al., 1997). In 5 cases a neighbouring block was used for CGH. DNA ISH: ISH was performed on consecutive tissue sections (4 lm thick) that were adhered to standard coated microscope glass slides. The (peri)centromeric DNA probes for chromosomes 1 (reference probe), 7 and 8 (test probes) were labeled with biotin-14-dATP (BioNick Kit, Gibco BRL, Gaithersburg MD, USA). The ISH procedure was carried out as described before (Alers et al., 1997). Briefly, after appropriate pepsin digestion the sections were heat-denatured for 2 min in 70% formamide in 2 ´ SSC, and hybridized overnight at 37 °C with the denatured probes in a hybridization mixture containing 2 ng/ll DNA probe, 500 ng/ll herring sperm DNA (Sigma, St. Louis MO, USA), 0.1% Tween-20, 10% dextran sulphate, and 60% formamide in 2 ´ SSC at pH 7.0. Then a series of stringent washes followed to remove the unbound probe. Histochemical detection was performed by immunoperoxidase staining. Slides were subsequently incubated with mouse anti-biotin, biotin-labeled goat anti-mouse, and peroxidase conjugated avidin-biotin complex (Dako, Glostrup, Denmark). The probe-related ISH signals were developed with diaminobenzidine (DAB), and the slides were counterstained with hematoxylin. The DNA probes were evaluated in predefined tissue areas. For each of the test probes (7 and 8), as well as the reference chromosome 1-specific DNA probe, 100 non-overlapping intact nuclei were counted by 2 independent investigators. The DNA spot distribution patterns were then compared and averaged. In these series, discrepancies between the 2 investigators were not encountered using this procedure. A DNA spot distribution was considered aberrant, when the associated P value was < 0.01 (Kolmogorov-Smirnov test; Alers et al., 1997). In this series, aberrations of chromosome 1 were not observed.
88
H. Van Dekken et al.
CGH: Isolation of DNA from the formalin-fixed paraffin-embedded tumor material was performed as described before (Alers et al., 1999). Shortly, tissue blocks were counterstained in 4,6-diamidino-2-phenyl indole (DAPI; 0.1 lg/ml in distilled water) for 5 min and placed under a fluorescence microscope enabling precise selection of the tumor areas. Microdissection was performed using a hollow bore coupled to the microscope. Lower boundaries were checked for the presence of tumor on hematoxylin-eosin-stained tissue sections. Samples were digested in extraction buffer and incubated at 55 °C for 3 to 4 days. Concentration, purity and molecular weight of the DNA were estimated using both a fluorometer and ethidium bromide-stained agarose gels with control DNA series. Test DNA was labeled with biotin by nick translation (Nick Translation System, Gibco BRL, Gaithersburg MD, USA). Low molecular test DNA (< 1 000 bp) was chemically labelled with biotin-ULS (Universal Linkage System; Kreatech Diagnostics, Amsterdam, The Netherlands) in order to maintain an acceptable DNA fragment size for CGH. Male reference DNA (Promega, Madison WI, USA) was labeled with digoxigenin (Boehringer Mannheim, Indianapolis IN, USA) by nick translation. The reaction time and the amount of DNAse were adjusted to obtain a matching probe size for reference DNA and test DNA. Molecular weight of both test and reference DNA was checked by gel electrophoresis. Probe sizes were between 300 bp and 1.5 kb. CGH was essentially performed according to the procedure described by Kallioniemi et al. (1992). In brief, normal male metaphase chromosomes were denatured, dehydrated and air dried. Four hundred ng of each labeled tumor DNA, 200 ng of reference DNA, and 15 mg of unlabeled Cot-1 DNA were precipitated in ethanol and dissolved in 10 ml of hybridization mixture (50% formamide, 0.1% Tween-20, and 10% dextran sulphate in 2 ´ SSC at pH 7.0). The probe mixture was denatured and hybridized to the normal metaphase chromosomes. After washing of the slides, fluorescent detection of the biotin- and digoxigenin-labeled DNA probes was accomplished with avidin-fluorescein isothiocyanate and anti-digoxigenin rhodamine, respectively, for 1 h at 37 °C. Samples were counterstained with DAPI in anti-fade solution. CGH analysis was accomplished with QUIPS XL software (Vysis, Downers Grove IL, USA). For the profiles, losses of DNA sequences were defined as chromosomal regions where the mean green to red fluorescent ratio was below 0.85, while gains were defined as chromosomal regions where this ratio was above 1.15. The threshold values were based on measurements from a series of normal controls.
Results We have compared DNA ISH and CGH within a human solid cancer system, i.e. a defined set of 20 prostatic adenocarcinomas. Chromosomes 7 and 8 were selected, since they show frequent aberrations in prostate cancer. ISH with chromosome 7 and 8 specific centromeric DNA probes was applied to standard formalin-fixed and paraffin-embedded histological sections for numerical chromosome analysis. CGH with DNA's, extracted from the same histologic area of the archival specimens, was used for screening of gains and losses of 7 and 8. ISH with centromeric probes distinguished a total of 26 numerical aberrations of chromosome 7 and/or 8 in the set of 20 neoplasms (Table 1). In the same set CGH revealed a total of 35 losses and gains on 26 chromosomes 7 and/or 8 (Fig. 1). CGH alterations were observed in 22 of the 26 chromosomes (85%) that showed aberrations in the ISH analysis. Examples of CGH and ISH are shown in Fig. 2. Concordance between ISH and
ISH versus CGH in cancer
89
Table 1. ISH and CGH of chromosomes 7 and 8 in 20 prostatic cancers (corresponding ISH/ CGH results in bold) Tumor
ISH #71
ISH #81
CGH #72
CGH #82
±8
+7q11.2-32
+7
+8
+7
+8 +8 ±8
+7p15 +7q21
±8p11.2-ter +8q21.1-ter ±8p12-22 +8q21-ter
13 2 3 43 53 6 7 8 9 10 11 123 13
+7
+7
+7pter-q31
+7q11.2
+8 +8
+7q11.2 +7q22-ter
14
+8
15
+8
+7p15-ter +7q31-34
163 17 18 19 20
+7
+8 +8 +8 ±8 +8
+7p21-ter
+7 +7 +7 +7
±8 +8 +8
+7pter-qter +7pter-qter +7q31-ter
±8p11.2-22 ±8p11.2-ter +8q24.1-ter +8q13-ter ±8p11.2-ter ±8p11.2-22 ±8p21-ter +8q11.2-ter ±8p11.2-21 ±8p21-ter +8q11.2-ter ±8p11.2-ter +8q11.2-ter ±8p21-ter +8q21.2-ter ±8p11.2-ter +8q22-ter +8q12-ter
1
No aberrations were detected for the chromosome 1-specific (reference) DNA probe. CGH gain or loss of chromosome bands p11.2 or q11.2 strongly suggest involvement of the centromeric region of the chromosome. 3 In these cases, an adjacent tissue block was taken for CGH. 2
CGH was observed in 11 (of 26; 42%) chromosomes (Table 1). Eight chromosomes were involved in gains (5 ´ #7, 3 ´ #8), 3 in losses (3 ´ #8). This included both complete (3/11) and partial (8/11) CGH confirmation of the numerical alterations. Partial CGH confirmation was defined as loss or gain of a chromosome arm with involvement of the centromeric region. In the majority of these cases, it concerned a whole chromosome arm, mostly the long arm. Combination of the latter CGH pattern and ISH aberration suggests the presence of isochromosome formation, most notably an isochromosome of the long arm of chromosome 8.
90
H. Van Dekken et al.
Fig. 1. Chromosome 7 and 8 ideograms showing the summary of DNA copy number changes, detected by CGH, in 20 prostatic cancers. Losses are displayed on the left of the ideogram, gains are depicted on the right. Chromosome 7 only shows gains, whereas chromosome 8 displays loss on the short arm and/or gain on the long arm. DNA ISH with centromeric probes to the same set of tumors revealed 9 ´ +7, 4 ´ ±8 and 13 ´ +8.
" Fig. 2. Three cases showing both ISH (A±C, E±G, I±K) and CGH (D, H, L) of archival formalin-fixed and paraffin-embedded prostate tumor specimens. The ISH related spots on the tissue sections were visualized with immuno-peroxidase/DAB (black); hematoxylin was used as a counterstain (gray; 40 ´ objective). The bar histograms (C, G, K) illustrate ISH alterations that are depicted in the microphotographs. Note the truncation phenomenon in the distribution pattern of the ISH spots, which is caused by ISH to thin histological sections. The CGH ideograms (D, H, L) are shown along with the ratio profiles (bar on the left of the chromosome represents loss, bar on the right represents gain). A±D] Case 19: ISH with centromeric DNA probes for chromosome 1 and 7 (panel A and B, respectively). ISH shows +7, when compared with the chromosome 1 reference DNA probe. Note more DNA probe spots per nucleus for the chromosome 7 specific probe (a few are arrowed), than for chromosome 1. This is also reflected in the bar histogram as a shift to the right of the chromosome 7 DNA probe spot distribution pattern (C). CGH reveals gain of the complete chromosome (D). Thus, there is a complete confirmation of ISH by CGH. E±H] Case 13: ISH with DNA probes for chromosome 1 and 8 (panel E and F, respectively), displaying +8 (E±G). CGH shows gain of the complete long arm with concomitant loss of (most of) the short arm (H), suggestive of i(8q). In this tumor, CGH partly confirmed the ISH findings. I±L] Case 8: ISH with DNA probes for chromosome 1 and 8 (panel I and J, respectively) reveals +8 (I±K), whereas CGH distinguishes loss of 8p (L). No gain of 8 is discriminated by CGH. There is no concordance between ISH and CGH.
ISH versus CGH in cancer
91
Discussion We have performed an evaluation of the application of ISH and CGH for detection of copy number changes in human solid tumors. In a large number of cases, it appeared that numerical chromosome aberrations, as revealed by ISH with centromeric DNA probes, were not observed concordantly in the CGH profiles. We suggest that this may be due to isochromosome formations, i.e. a structural chromosome alteration containing identical (parts of) chromosome arms, bridged by centromeric DNA. This abnormality is encountered in approx. 10% of human malignant neoplasms (Mertens et al., 1994). The 9 most common isochromosomes (in decreasing order of frequency) are i(17q), i(8q), i(1q), i(12p), i(6p), i(7q), i(9q), i(5p), and i(21q). Isochromosomes can be of clinical significance. For example, i(12p) is a frequently found diagnostic alteration in germ cell tumors, and the chromosome 12-specific centromeric
92
H. Van Dekken et al.
DNA probe can be used for detection (Looijenga and Oosterhuis, 1999). Thus, if the isochromosome represents a well known aberration, interphase cytogenetics with the appropriate centromeric DNA probe can be adequately used for evaluation. Moreover, in many neoplasms specific and frequent numerical changes have been identified, e. g. trisomy 7 in prostate cancer (Lundgren et al., 1992). In these cases, ISH yields an excellent tool for detection and enumeration (reviewed in Van Dekken et al., 1997). However, when information is not available, one should be cautious with the interpretation, as this study indicates. We detected 9 cases with extra copies of chromosome 7 by ISH with a centromeric DNA probe. However, only 3 could be fully distinguished in the CGH profiles (Table 1; cases 6, 18, 19). This might point at a reduced potential of CGH in the delineation of numerical abnormalities. The chromosome 8 numerical aberrations, as discriminated by ISH, could not be completely confirmed by CGH. This may indeed be caused by isochromosome formation, i.e. i(8q). In 5 cases, a combination of 8p loss with simultaneous 8q gain, often the complete long arm, was found by CGH, whereas ISH showed +8 (Table 1; cases 2, 11, 13, 14, 15). This combinatorial pattern of 8p loss with 8q gain is very suggestive of i(8q) formation. In those cases, in which only a part of the 8q arm was gained, one could speculate that this was the result of deletion after isochromosome formation. We observed i(8q) formation by CGH and classical karyotyping in 2 prostate cancer cell lines, PC 133 and PC 346 (Van Weerden et al., 1996). In addition, another CGH study of prostate cancer also suggested i(8q) formation (Cher et al., 1994). Importantly, it should be noted that in CGH the centromeres are excluded from analysis to avoid artefacts caused by repetitive (peri)centromeric DNA sequences (Kallioniemi et al., 1992). One could, however, speculate on its involvement in certain aberrations to explain the combined ISH-CGH patterns found in this and other studies. Another karyotypic possibility for 8q gain by CGH is duplication of part of the q arm, but in this scenario ISH changes are not expected. In 4 tumors numerical loss of chromosome 8 was discriminated by ISH. In 3 of those cases, CGH revealed loss of the complete short arm of chromosome 8 likely also involving the centromeric region. Several review studies have clearly illustrated that the same genetic aberrations can be detected by karyotyping, in situ and molecular techniques (see e. g. Van Dekken et al., 1997). However, a detailed CGH-ISH comparison, as described above, has sofar not been published. A CGH study on prostatic cancer was reported in which ISH with centromere probes was used only in a few selected cases (Cher et al., 1994). In these cases, results were obtained that correlate well with the results of our investigation. These authors further found good correlation between CGH-loss on 8p and loss of heterozygosity (LOH). Iwabuchi et al. (1995) compared LOH and CGH in ovarian tumors. There was a straightforward concordance between allelotyping and reduced copy number detected by CGH. Recently, CGH results in ovarian borderline tumors were confirmed by ISH (Wolf et al., 1999). The detection of gene amplifications in breast cancer specimens was studied by CGH and Southern
ISH versus CGH in cancer
93
blotting (Courjal and Theillet, 1997). A good agreement was found, although CGH revealed gains that remained undetected by Southern blotting. In conclusion, ISH can be used adequately in those circumstances in which (numerical) chromosomal aberration has been fully characterized. CGH is the method of choice for genome-wide evaluations of chromosomal gains and losses. However, this CGH-ISH study demonstrates that when the chromosomal constitution of an abnormality is unknown, ISH results should be interpreted with caution. The underlying alteration might be structural, and the possibility of an isochromosome should be considered. Acknowledgment: This work was supported by the Dutch Cancer Society, Grant no's EUR 97-1404 and EUR 92-35.
References Alers JC, Krijtenburg PJ, Rosenberg C, Hop WCJ, Verkerk AM, SchroÈder FH, Van der Kwast ThH, Bosman FT, and Van Dekken H (1997) Interphase cytogenetics of prostatic tumor progression: specific chromosomal abnormalities are involved in metastasis to the bone. Lab Invest 77: 437±448 Alers JC, Rochat J, Krijtenburg PJ, van Dekken H, Raap AK, and Rosenberg C (1999) Universal Linkage System: an improved method for labeling archival DNA for comparative genomic hybridization. Genes Chromosom Cancer 25: 301±305 Brzoska PM, Levin NA, Fu KK, Kaplan MJ, Singer MI, Gray JW, and Christman MF (1995) Frequent novel DNA copy number increase in squamous cell head and neck tumors. Cancer Res 55: 3055±3059 Cher ML, MacGrogan D, Bookstein R, Brown JA, Jenkins RB, and Jensen RH (1994) Comparative genomic hybridization, allelic imbalance, and fluorescence in situ hybridization on chromosome 8 in prostate cancer. Genes Chromosom Cancer 11: 153±162 Courjal F, and Theillet C (1997) Comparative genomic hybridization analysis of breast tumors with predetermined profiles of DNA amplification. Cancer Res 57: 4368±4377 Herbergs J, De BruõÈne AP, Marx PTJ, Vallinga MIJ, Stockbrugger RW, Ramaekers FCS, Arends JW, and Hopman AHN (1994) Chromosome aberrations in adenomas of the colon, proof of trisomy 7 in tumor cells by combined interphase cytogenetics and immunohistochemistry. Int J Cancer 57: 781±785 Iwabuchi H, Sakamoto M, Sakunaga H, Ma YY, Carcangiu ML, Pinkel D, Yang-Feng TL, and Gray JW (1995) Genetic analysis of benign, low-grade, and high-grade ovarian tumors. Cancer Res 55: 6172±6180 Kallioniemi A, Kallioniemi OP, Sudar D, Rutovitz D, Gray JW, Waldman F, and Pinkel D (1992) Comparative genomic hybridization for molecular cytogenetic analysis of solid tumors. Science 258: 818±821 Looijenga LH, and Oosterhuis JW (1999) Pathogenesis of testicular germ cell tumours. Rev Reprod 4: 90±100 Lundgren R, Mandahl N, Heim S, Limon J, Henrikson H, and Mitelman F (1992) Cytogenetic analysis of 57 primary prostatic adenocarcinomas. Genes Chromosom Cancer 4: 16±24 Mertens F, Johansson B, and Mitelman F (1994) Isochromosomes in neoplasia. Genes Chromosom Cancer 10: 221±230 Sandberg AA (1990) The Chromosomes in Human Cancer and Leukemia. 2nd ed. Elsevier Science Publishing, New York
94
H. Van Dekken et al.
Speicher MR, Howe C, Crotty P, Du Manoir S, Costa J, and Ward DC (1995) Comparative genomic hybridization detects novel deletions and amplifications in head and neck squamous cell carcinomas. Cancer Res 55: 1010±1013 Steiner MG, Harlow SP, Colombo E, and Bauer KD (1993) Chromosomes 8, 12, and 17 copy number in Astler-Coller stage C colon Cancer in relation to proliferative activity and DNA ploidy. Cancer Res 53: 681±686 Takahashi S, Qian J, Brown JA, Alcaraz A, Bostwick DG, Lieber MM, and Jenkins RB (1994) Potential markers of prostate cancer aggressiveness detected by fluorescence in situ hybridization in needle biopsies. Cancer Res 54: 3574±3579 Van Dekken H, Rosenberg C, Krijtenburg PJ, and Alers JC (1997) Interphase cytogenetics and comparative genomic hybridization of human epithelial cancers and precursor lesions. Histochem Cell Biol 108: 419±430 Van Weerden WM, de Ridder CMA, Verdaasdonk CL, Romijn JC, Van der Kwast ThH, SchroÈder FH, and Van Steenbrugge GJ (1996) Development of seven new human prostate tumor xenograft models and their histopathological characterization. Am J Pathol 149: 1055±1062 Visakorpi T, Kallioniemi A, SyvaÈnen A-C, Hyytinen ER, Karhu R, Tammela T, Isola JJ, and Kallioniemi O-P (1995) Genetic changes in primary and recurrent prostate cancer by comparative genomic hybridization. Cancer Res 55: 342±347 Wolf NG, Abdul-Karim FW, Farver C, SchroÈck E, Du Manoir S, and Schwartz S (1999) Analysis of ovarian borderline tumors using comparative genomic hybridization and fluorescence in situ hybridization. Genes Chromosom Cancer 25: 307±315 Xiao S, Li D, Corson JM, Vijg J, and Fletcher JA (1995) Codeletion of p15 and p16 genes in primary non-small cell lung carcinoma. Cancer Res 55: 2968±2971
DNA in situ hybridization (interphase cytogenetics) versus comparative genomic hybridization (CGH) in human cancer: detection of numerical and structural chromosome aberrations Herman Van Dekken, Pieter J. Krijtenburg, and Janneke C. Alers Department of Pathology, Josephine Nefkens Institute, Erasmus University Rotterdam, P. O. Box 17 83, 3000 DR Rotterdam, The Netherlands Received 22 September 1999 and in revised form 15 November 1999; accepted 20 November 1999
Summary DNA in situ hybridization techniques for cytogenetic analyses of human solid cancers are nowadays widely used for diagnostic and research purposes. The advantage of this methodology is that it can be applied to cells in the interphase state, thereby circumventing the need for high-quality metaphase preparations for karyotypic evaluation. In situ hybridization (ISH) with chromosome specific (peri)centromeric DNA probes, also termed ªinterphase cytogeneticsº, can be used to detect numerical changes, whereas comparative genomic hybridization (CGH) discloses chromosomal gains and losses, i.e. amplifications and deletions. We wanted to compare both methods in human solid tumors, and for this goal we evaluated ISH and CGH within a set of 20 selected prostatic adenocarcinomas. Chromosomes 7 and 8 were chosen for this analysis, since these chromosomes are frequently altered in prostate cancer. ISH with chromosome 7 and 8 specific centromeric DNA probes was applied to standard, formalinfixed and paraffin-embedded, histological sections for numerical chromosome analysis. CGH with DNA's, extracted from the same histologic area of the archival specimens, was used for screening of gains and losses of 7 and 8. ISH with centromeric probes distinguished a total of 26 numerical aberrations of chromosome 7 and/or 8 in the set of 20 neoplasms. In the same set CGH reCorrespondence to: Dr. H. Van Dekken, phone: +31-10/4 08 79 01, fax: +31-10/4 08 94 87, e-mail: [email protected]. http://www.urbanfischer.de/journals/actahist
0065±1281/00/102/1±085 $ 12.00/0
86
H. Van Dekken et al.
vealed a total of 35 losses and gains. CGH alterations of 7 and 8 were seen in twenty-two of the 26 chromosomes (85%) that showed aberrations in the ISH analysis. Concordance between ISH and CGH was seen in 11 (of 26; 42%) chromosomes. Eight chromosomes were involved in gains (5 ´ #7, 3 ´ #8), three in losses (3 ´ #8). This included both complete (3/11) and partial (8/11) CGH confirmation of the numerical alteration. Partial CGH confirmation was defined as loss or gain of a chromosome arm with involvement of the centromeric region. In the majority of these cases it concerned a whole chromosome arm, mostly the long arm. We conclude that generally a fair correlation was found between ISH and CGH in interphase preparations of a series of prostate cancers. However, when specified in detail, most of the numerical ISH aberrations were only partly represented in the CGH analysis. On the one hand, it suggests that CGH does not adequately discriminate numerical abnormalities. On the other hand, it likely implies that not all numerical changes, as detected by interphase cytogenetics, are truly involving the whole chromosome. A part of these discrepancies might be caused by structural mechanisms, most notably isochromosome formation. Key words: in situ hybridization ± comparative genomic hybridization ± chromosome aberration ± cancer
Introduction Despite the frequent incidence of human solid tumors, relatively little is known about the cytogenetic aberrations that characterize them. Mostly, technical problems have for a long time hampered the acquisition of cytogenetic data on solid tumors (Sandberg, 1990). Solid tumor biopsies yield low numbers of viable cells, have a low mitotic index and frequently contain non-neoplastic cells. Furthermore, the quality of metaphases is relatively poor. Therefore, interphase approaches, both at molecular and cytological levels, are better equipped to disclose genetic changes occurring in human epithelial cancers. Over the past decade, interphase hybridization technology, i.e. DNA in situ hybridization (ISH; reviewed in e. g. Van Dekken et al., 1997), and comparative genomic hybridization (CGH; first introduced by Kallioniemi et al. in 1992), have emerged as important instruments for the delineation of genetic changes in human solid cancers. ISH can be used for rapid detection of numerical chromosome changes, whereas CGH is optimally suited for evaluation of chromosomal imbalances, i.e. deletions and amplifications. Results from molecular and karyotyping studies can be confirmed in situ and, in addition, new genetic aberrations are being disclosed. For example, in lung cancer ISH with region-specific probes showed loss of the p15/p16 region at 9p, which is consistent with allelotyping studies (Xiao et al., 1995). In head and neck tumors, CGH studies repeatedly identified a new amplified chromosomal region, 3q26-qter (Brzoska et al., 1995; Speicher et al., 1995). This region has
ISH versus CGH in cancer
87
not been described before and may harbor a new oncogene that is important in these neoplasms. ISH on colon cancer specimens have confirmed the presence of numerical aberrations that are frequently encountered in colorectal tumors, such as trisomy 7 and monosomy 18 (Herbergs et al., 1994). It has also added some new numerical aberrations, e. g. gain of chromosome 12 (Steiner et al., 1993). In prostatic adenocarcinoma, ISH analysis rendered possible markers for biological aggressiveness, such as gain of chromosomes 7 and 8 (Alcaraz et al., 1994; Alers et al., 1997). Additionally, CGH provided some new regions of interest, most notably the amplification of the 8 q arm (Visakorpi et al., 1995). Thus, it can be stated that ISH and CGH have established their importance in solid cancer genetics. In this report, we wanted to compare ISH and CGH in human solid cancers. Sofar, such a study is lacking in the literature. A well-defined set of prostatic adenocarcinomas was used for this purpose (Alers et al., 1997). The following issues were specifically addressed: 1] What is the sensitivity of ISH versus CGH in detecting numerical chromosomal aberrations? and 2] Are alterations detected by ISH indeed caused by changes in the copy number of a specific chromosome?
Material and methods We collected representative paraffin blocks from 20 prostatic adenocarcinoma specimens (6 primary tumors, 6 lymph node metastases, 8 distant metastases). The 20 tumors were taken from a previously reported series, since they were found to be altered for chromosome 7 and/or 8 by DNA in situ hybridization (Alers et al., 1997). In 5 cases a neighbouring block was used for CGH. DNA ISH: ISH was performed on consecutive tissue sections (4 lm thick) that were adhered to standard coated microscope glass slides. The (peri)centromeric DNA probes for chromosomes 1 (reference probe), 7 and 8 (test probes) were labeled with biotin-14-dATP (BioNick Kit, Gibco BRL, Gaithersburg MD, USA). The ISH procedure was carried out as described before (Alers et al., 1997). Briefly, after appropriate pepsin digestion the sections were heat-denatured for 2 min in 70% formamide in 2 ´ SSC, and hybridized overnight at 37 °C with the denatured probes in a hybridization mixture containing 2 ng/ll DNA probe, 500 ng/ll herring sperm DNA (Sigma, St. Louis MO, USA), 0.1% Tween-20, 10% dextran sulphate, and 60% formamide in 2 ´ SSC at pH 7.0. Then a series of stringent washes followed to remove the unbound probe. Histochemical detection was performed by immunoperoxidase staining. Slides were subsequently incubated with mouse anti-biotin, biotin-labeled goat anti-mouse, and peroxidase conjugated avidin-biotin complex (Dako, Glostrup, Denmark). The probe-related ISH signals were developed with diaminobenzidine (DAB), and the slides were counterstained with hematoxylin. The DNA probes were evaluated in predefined tissue areas. For each of the test probes (7 and 8), as well as the reference chromosome 1-specific DNA probe, 100 non-overlapping intact nuclei were counted by 2 independent investigators. The DNA spot distribution patterns were then compared and averaged. In these series, discrepancies between the 2 investigators were not encountered using this procedure. A DNA spot distribution was considered aberrant, when the associated P value was < 0.01 (Kolmogorov-Smirnov test; Alers et al., 1997). In this series, aberrations of chromosome 1 were not observed.
88
H. Van Dekken et al.
CGH: Isolation of DNA from the formalin-fixed paraffin-embedded tumor material was performed as described before (Alers et al., 1999). Shortly, tissue blocks were counterstained in 4,6-diamidino-2-phenyl indole (DAPI; 0.1 lg/ml in distilled water) for 5 min and placed under a fluorescence microscope enabling precise selection of the tumor areas. Microdissection was performed using a hollow bore coupled to the microscope. Lower boundaries were checked for the presence of tumor on hematoxylin-eosin-stained tissue sections. Samples were digested in extraction buffer and incubated at 55 °C for 3 to 4 days. Concentration, purity and molecular weight of the DNA were estimated using both a fluorometer and ethidium bromide-stained agarose gels with control DNA series. Test DNA was labeled with biotin by nick translation (Nick Translation System, Gibco BRL, Gaithersburg MD, USA). Low molecular test DNA (< 1 000 bp) was chemically labelled with biotin-ULS (Universal Linkage System; Kreatech Diagnostics, Amsterdam, The Netherlands) in order to maintain an acceptable DNA fragment size for CGH. Male reference DNA (Promega, Madison WI, USA) was labeled with digoxigenin (Boehringer Mannheim, Indianapolis IN, USA) by nick translation. The reaction time and the amount of DNAse were adjusted to obtain a matching probe size for reference DNA and test DNA. Molecular weight of both test and reference DNA was checked by gel electrophoresis. Probe sizes were between 300 bp and 1.5 kb. CGH was essentially performed according to the procedure described by Kallioniemi et al. (1992). In brief, normal male metaphase chromosomes were denatured, dehydrated and air dried. Four hundred ng of each labeled tumor DNA, 200 ng of reference DNA, and 15 mg of unlabeled Cot-1 DNA were precipitated in ethanol and dissolved in 10 ml of hybridization mixture (50% formamide, 0.1% Tween-20, and 10% dextran sulphate in 2 ´ SSC at pH 7.0). The probe mixture was denatured and hybridized to the normal metaphase chromosomes. After washing of the slides, fluorescent detection of the biotin- and digoxigenin-labeled DNA probes was accomplished with avidin-fluorescein isothiocyanate and anti-digoxigenin rhodamine, respectively, for 1 h at 37 °C. Samples were counterstained with DAPI in anti-fade solution. CGH analysis was accomplished with QUIPS XL software (Vysis, Downers Grove IL, USA). For the profiles, losses of DNA sequences were defined as chromosomal regions where the mean green to red fluorescent ratio was below 0.85, while gains were defined as chromosomal regions where this ratio was above 1.15. The threshold values were based on measurements from a series of normal controls.
Results We have compared DNA ISH and CGH within a human solid cancer system, i.e. a defined set of 20 prostatic adenocarcinomas. Chromosomes 7 and 8 were selected, since they show frequent aberrations in prostate cancer. ISH with chromosome 7 and 8 specific centromeric DNA probes was applied to standard formalin-fixed and paraffin-embedded histological sections for numerical chromosome analysis. CGH with DNA's, extracted from the same histologic area of the archival specimens, was used for screening of gains and losses of 7 and 8. ISH with centromeric probes distinguished a total of 26 numerical aberrations of chromosome 7 and/or 8 in the set of 20 neoplasms (Table 1). In the same set CGH revealed a total of 35 losses and gains on 26 chromosomes 7 and/or 8 (Fig. 1). CGH alterations were observed in 22 of the 26 chromosomes (85%) that showed aberrations in the ISH analysis. Examples of CGH and ISH are shown in Fig. 2. Concordance between ISH and
ISH versus CGH in cancer
89
Table 1. ISH and CGH of chromosomes 7 and 8 in 20 prostatic cancers (corresponding ISH/ CGH results in bold) Tumor
ISH #71
ISH #81
CGH #72
CGH #82
±8
+7q11.2-32
+7
+8
+7
+8 +8 ±8
+7p15 +7q21
±8p11.2-ter +8q21.1-ter ±8p12-22 +8q21-ter
13 2 3 43 53 6 7 8 9 10 11 123 13
+7
+7
+7pter-q31
+7q11.2
+8 +8
+7q11.2 +7q22-ter
14
+8
15
+8
+7p15-ter +7q31-34
163 17 18 19 20
+7
+8 +8 +8 ±8 +8
+7p21-ter
+7 +7 +7 +7
±8 +8 +8
+7pter-qter +7pter-qter +7q31-ter
±8p11.2-22 ±8p11.2-ter +8q24.1-ter +8q13-ter ±8p11.2-ter ±8p11.2-22 ±8p21-ter +8q11.2-ter ±8p11.2-21 ±8p21-ter +8q11.2-ter ±8p11.2-ter +8q11.2-ter ±8p21-ter +8q21.2-ter ±8p11.2-ter +8q22-ter +8q12-ter
1
No aberrations were detected for the chromosome 1-specific (reference) DNA probe. CGH gain or loss of chromosome bands p11.2 or q11.2 strongly suggest involvement of the centromeric region of the chromosome. 3 In these cases, an adjacent tissue block was taken for CGH. 2
CGH was observed in 11 (of 26; 42%) chromosomes (Table 1). Eight chromosomes were involved in gains (5 ´ #7, 3 ´ #8), 3 in losses (3 ´ #8). This included both complete (3/11) and partial (8/11) CGH confirmation of the numerical alterations. Partial CGH confirmation was defined as loss or gain of a chromosome arm with involvement of the centromeric region. In the majority of these cases, it concerned a whole chromosome arm, mostly the long arm. Combination of the latter CGH pattern and ISH aberration suggests the presence of isochromosome formation, most notably an isochromosome of the long arm of chromosome 8.
90
H. Van Dekken et al.
Fig. 1. Chromosome 7 and 8 ideograms showing the summary of DNA copy number changes, detected by CGH, in 20 prostatic cancers. Losses are displayed on the left of the ideogram, gains are depicted on the right. Chromosome 7 only shows gains, whereas chromosome 8 displays loss on the short arm and/or gain on the long arm. DNA ISH with centromeric probes to the same set of tumors revealed 9 ´ +7, 4 ´ ±8 and 13 ´ +8.
" Fig. 2. Three cases showing both ISH (A±C, E±G, I±K) and CGH (D, H, L) of archival formalin-fixed and paraffin-embedded prostate tumor specimens. The ISH related spots on the tissue sections were visualized with immuno-peroxidase/DAB (black); hematoxylin was used as a counterstain (gray; 40 ´ objective). The bar histograms (C, G, K) illustrate ISH alterations that are depicted in the microphotographs. Note the truncation phenomenon in the distribution pattern of the ISH spots, which is caused by ISH to thin histological sections. The CGH ideograms (D, H, L) are shown along with the ratio profiles (bar on the left of the chromosome represents loss, bar on the right represents gain). A±D] Case 19: ISH with centromeric DNA probes for chromosome 1 and 7 (panel A and B, respectively). ISH shows +7, when compared with the chromosome 1 reference DNA probe. Note more DNA probe spots per nucleus for the chromosome 7 specific probe (a few are arrowed), than for chromosome 1. This is also reflected in the bar histogram as a shift to the right of the chromosome 7 DNA probe spot distribution pattern (C). CGH reveals gain of the complete chromosome (D). Thus, there is a complete confirmation of ISH by CGH. E±H] Case 13: ISH with DNA probes for chromosome 1 and 8 (panel E and F, respectively), displaying +8 (E±G). CGH shows gain of the complete long arm with concomitant loss of (most of) the short arm (H), suggestive of i(8q). In this tumor, CGH partly confirmed the ISH findings. I±L] Case 8: ISH with DNA probes for chromosome 1 and 8 (panel I and J, respectively) reveals +8 (I±K), whereas CGH distinguishes loss of 8p (L). No gain of 8 is discriminated by CGH. There is no concordance between ISH and CGH.
ISH versus CGH in cancer
91
Discussion We have performed an evaluation of the application of ISH and CGH for detection of copy number changes in human solid tumors. In a large number of cases, it appeared that numerical chromosome aberrations, as revealed by ISH with centromeric DNA probes, were not observed concordantly in the CGH profiles. We suggest that this may be due to isochromosome formations, i.e. a structural chromosome alteration containing identical (parts of) chromosome arms, bridged by centromeric DNA. This abnormality is encountered in approx. 10% of human malignant neoplasms (Mertens et al., 1994). The 9 most common isochromosomes (in decreasing order of frequency) are i(17q), i(8q), i(1q), i(12p), i(6p), i(7q), i(9q), i(5p), and i(21q). Isochromosomes can be of clinical significance. For example, i(12p) is a frequently found diagnostic alteration in germ cell tumors, and the chromosome 12-specific centromeric
92
H. Van Dekken et al.
DNA probe can be used for detection (Looijenga and Oosterhuis, 1999). Thus, if the isochromosome represents a well known aberration, interphase cytogenetics with the appropriate centromeric DNA probe can be adequately used for evaluation. Moreover, in many neoplasms specific and frequent numerical changes have been identified, e. g. trisomy 7 in prostate cancer (Lundgren et al., 1992). In these cases, ISH yields an excellent tool for detection and enumeration (reviewed in Van Dekken et al., 1997). However, when information is not available, one should be cautious with the interpretation, as this study indicates. We detected 9 cases with extra copies of chromosome 7 by ISH with a centromeric DNA probe. However, only 3 could be fully distinguished in the CGH profiles (Table 1; cases 6, 18, 19). This might point at a reduced potential of CGH in the delineation of numerical abnormalities. The chromosome 8 numerical aberrations, as discriminated by ISH, could not be completely confirmed by CGH. This may indeed be caused by isochromosome formation, i.e. i(8q). In 5 cases, a combination of 8p loss with simultaneous 8q gain, often the complete long arm, was found by CGH, whereas ISH showed +8 (Table 1; cases 2, 11, 13, 14, 15). This combinatorial pattern of 8p loss with 8q gain is very suggestive of i(8q) formation. In those cases, in which only a part of the 8q arm was gained, one could speculate that this was the result of deletion after isochromosome formation. We observed i(8q) formation by CGH and classical karyotyping in 2 prostate cancer cell lines, PC 133 and PC 346 (Van Weerden et al., 1996). In addition, another CGH study of prostate cancer also suggested i(8q) formation (Cher et al., 1994). Importantly, it should be noted that in CGH the centromeres are excluded from analysis to avoid artefacts caused by repetitive (peri)centromeric DNA sequences (Kallioniemi et al., 1992). One could, however, speculate on its involvement in certain aberrations to explain the combined ISH-CGH patterns found in this and other studies. Another karyotypic possibility for 8q gain by CGH is duplication of part of the q arm, but in this scenario ISH changes are not expected. In 4 tumors numerical loss of chromosome 8 was discriminated by ISH. In 3 of those cases, CGH revealed loss of the complete short arm of chromosome 8 likely also involving the centromeric region. Several review studies have clearly illustrated that the same genetic aberrations can be detected by karyotyping, in situ and molecular techniques (see e. g. Van Dekken et al., 1997). However, a detailed CGH-ISH comparison, as described above, has sofar not been published. A CGH study on prostatic cancer was reported in which ISH with centromere probes was used only in a few selected cases (Cher et al., 1994). In these cases, results were obtained that correlate well with the results of our investigation. These authors further found good correlation between CGH-loss on 8p and loss of heterozygosity (LOH). Iwabuchi et al. (1995) compared LOH and CGH in ovarian tumors. There was a straightforward concordance between allelotyping and reduced copy number detected by CGH. Recently, CGH results in ovarian borderline tumors were confirmed by ISH (Wolf et al., 1999). The detection of gene amplifications in breast cancer specimens was studied by CGH and Southern
ISH versus CGH in cancer
93
blotting (Courjal and Theillet, 1997). A good agreement was found, although CGH revealed gains that remained undetected by Southern blotting. In conclusion, ISH can be used adequately in those circumstances in which (numerical) chromosomal aberration has been fully characterized. CGH is the method of choice for genome-wide evaluations of chromosomal gains and losses. However, this CGH-ISH study demonstrates that when the chromosomal constitution of an abnormality is unknown, ISH results should be interpreted with caution. The underlying alteration might be structural, and the possibility of an isochromosome should be considered. Acknowledgment: This work was supported by the Dutch Cancer Society, Grant no's EUR 97-1404 and EUR 92-35.
References Alers JC, Krijtenburg PJ, Rosenberg C, Hop WCJ, Verkerk AM, SchroÈder FH, Van der Kwast ThH, Bosman FT, and Van Dekken H (1997) Interphase cytogenetics of prostatic tumor progression: specific chromosomal abnormalities are involved in metastasis to the bone. Lab Invest 77: 437±448 Alers JC, Rochat J, Krijtenburg PJ, van Dekken H, Raap AK, and Rosenberg C (1999) Universal Linkage System: an improved method for labeling archival DNA for comparative genomic hybridization. Genes Chromosom Cancer 25: 301±305 Brzoska PM, Levin NA, Fu KK, Kaplan MJ, Singer MI, Gray JW, and Christman MF (1995) Frequent novel DNA copy number increase in squamous cell head and neck tumors. Cancer Res 55: 3055±3059 Cher ML, MacGrogan D, Bookstein R, Brown JA, Jenkins RB, and Jensen RH (1994) Comparative genomic hybridization, allelic imbalance, and fluorescence in situ hybridization on chromosome 8 in prostate cancer. Genes Chromosom Cancer 11: 153±162 Courjal F, and Theillet C (1997) Comparative genomic hybridization analysis of breast tumors with predetermined profiles of DNA amplification. Cancer Res 57: 4368±4377 Herbergs J, De BruõÈne AP, Marx PTJ, Vallinga MIJ, Stockbrugger RW, Ramaekers FCS, Arends JW, and Hopman AHN (1994) Chromosome aberrations in adenomas of the colon, proof of trisomy 7 in tumor cells by combined interphase cytogenetics and immunohistochemistry. Int J Cancer 57: 781±785 Iwabuchi H, Sakamoto M, Sakunaga H, Ma YY, Carcangiu ML, Pinkel D, Yang-Feng TL, and Gray JW (1995) Genetic analysis of benign, low-grade, and high-grade ovarian tumors. Cancer Res 55: 6172±6180 Kallioniemi A, Kallioniemi OP, Sudar D, Rutovitz D, Gray JW, Waldman F, and Pinkel D (1992) Comparative genomic hybridization for molecular cytogenetic analysis of solid tumors. Science 258: 818±821 Looijenga LH, and Oosterhuis JW (1999) Pathogenesis of testicular germ cell tumours. Rev Reprod 4: 90±100 Lundgren R, Mandahl N, Heim S, Limon J, Henrikson H, and Mitelman F (1992) Cytogenetic analysis of 57 primary prostatic adenocarcinomas. Genes Chromosom Cancer 4: 16±24 Mertens F, Johansson B, and Mitelman F (1994) Isochromosomes in neoplasia. Genes Chromosom Cancer 10: 221±230 Sandberg AA (1990) The Chromosomes in Human Cancer and Leukemia. 2nd ed. Elsevier Science Publishing, New York
94
H. Van Dekken et al.
Speicher MR, Howe C, Crotty P, Du Manoir S, Costa J, and Ward DC (1995) Comparative genomic hybridization detects novel deletions and amplifications in head and neck squamous cell carcinomas. Cancer Res 55: 1010±1013 Steiner MG, Harlow SP, Colombo E, and Bauer KD (1993) Chromosomes 8, 12, and 17 copy number in Astler-Coller stage C colon Cancer in relation to proliferative activity and DNA ploidy. Cancer Res 53: 681±686 Takahashi S, Qian J, Brown JA, Alcaraz A, Bostwick DG, Lieber MM, and Jenkins RB (1994) Potential markers of prostate cancer aggressiveness detected by fluorescence in situ hybridization in needle biopsies. Cancer Res 54: 3574±3579 Van Dekken H, Rosenberg C, Krijtenburg PJ, and Alers JC (1997) Interphase cytogenetics and comparative genomic hybridization of human epithelial cancers and precursor lesions. Histochem Cell Biol 108: 419±430 Van Weerden WM, de Ridder CMA, Verdaasdonk CL, Romijn JC, Van der Kwast ThH, SchroÈder FH, and Van Steenbrugge GJ (1996) Development of seven new human prostate tumor xenograft models and their histopathological characterization. Am J Pathol 149: 1055±1062 Visakorpi T, Kallioniemi A, SyvaÈnen A-C, Hyytinen ER, Karhu R, Tammela T, Isola JJ, and Kallioniemi O-P (1995) Genetic changes in primary and recurrent prostate cancer by comparative genomic hybridization. Cancer Res 55: 342±347 Wolf NG, Abdul-Karim FW, Farver C, SchroÈck E, Du Manoir S, and Schwartz S (1999) Analysis of ovarian borderline tumors using comparative genomic hybridization and fluorescence in situ hybridization. Genes Chromosom Cancer 25: 307±315 Xiao S, Li D, Corson JM, Vijg J, and Fletcher JA (1995) Codeletion of p15 and p16 genes in primary non-small cell lung carcinoma. Cancer Res 55: 2968±2971