fluorescence in situ hybridization: a sensitive method for trisomy

Fluorescence In Situ Hybridization: A Sensitive Method for Trisomy 8 Detection in Bone Marrow Specimens

By Robert B. Jenkins, Michelle M. Le Beau, William J. Kraker, Thomas J. Borell, Paul G. Stalboerger, Elizabeth M. Davis, Lolita Penland, Anthony Fernald, Rafael Espinosa 111, Daniel J. Schaid, Pierre Noel, and Gordon W. Dewald

Trisomy 8 is a common anomaly in bone marrow (BM) cells of patients with myeloproliferative disorders (MPD), myelodysplastic syndromes (MDS), or acute nonlymphocytic leukemia (ANLL). We studied the efficacy of fluorescence in situ hybridization (FISH) detection of trisomy 8 in patients with MPD, MDS, or ANLL using directly labeled fluorescent a-satellite and whole chromosome paint (WCP) DNA probes specific for chromosome 8. Using FISH, we analyzed interphase nuclei and metaphase spreads from randomized series of BM specimens from normal individuals and patients with varying proportions of trisomy 8 as determined by conventional cytogenetic analysis. The BM of all normal donors contained ~ 2 . 0 % nuclei with 3 interphase FISH signals and s1 metaphase with 3 WCP FISH signals. Ninety-five percent and 98% of BM specimens with at least two metaphase cells with trisomy 8 by cytogenetic analysis contained greater than 2.0% nuclei with 3 interphase FISH and greater than 2 metaphases with 3 WCP FISH signals, respectively. Thirteen

RISOMY 8 is a common cytogenetic abnormality T found in the bone marrow (BM) of patients with myeloproliferative disorders (MPD), myelodysplastic syndromes (MDS), or acute nonlymphocytic leukemia (ANLL).1-5 Using conventional cytogenetic methods in which 20 to 30 metaphase cells are typically examined, it is possible to find 2 or more metaphases with trisomy 8 in many of these patients. According to the ISCN definition,6 these specimens would have an abnormal clone. However, in some patients, a single metaphase cell with trisomy 8 is found by conventional studies. These uncertain results are not clonal by ISCN criteria and may reflect either artifact or the potential emergence of an abnormal clone. For at least some of these patients, subsequent BM specimens are found to have 2 or more metaphase cells with trisomy 8, demonstrating the presence and proliferation of an abnormal clone.’ In other patients, subsequent BM specimens may have only normal metaphase cells, raising the possibil- ity that the single metaphase with trisomy 8 in the initial study was artifactual. At the time of the initial study, it would be important to know, using a simple and independent method, which patients with uncertain cytogenetic results might have an emerging abnormal clone.

In addition, the BM of many patients with MPD, MDS, or ANLL often contain no apparent chromosome abnormality by conventional cytogenetic studies. The observation of only normal metaphase cells may mean that the neoplastic cells were not dividing at the time of cytogenetic analysis or were missed because of sampling error. Alternatively, the malignant cells may contain submicroscopic molecular genetic abnormalities or subtle cytogenetic abnormalities. A simple and sensitive method to detect chromosomal aneuploidy (eg, trisomy 8) in such patients might be diagnostically and prognostically relevant.

The technique of fluorescent in situ chromosomal hybridization (FISH) is a technique that is very useful for gene mapping purposes and for the identification of marker

patients had 1 in 20 or 1 in 30 metaphase cells with trisomy 8 by conventional cytogenetic studies. Of these patients, four had greater than 2.0% nuclei with 3 interphase FISH signals. The BM of all four patients contained positive metaphase FISH results. We then studied the usefulness of FISH analysis to detect occult trisomy 8 by analyzing BM nuclei from 144 patients who had MPD, MDS, or ANLL and either 20 normal metaphase cells or an abnormal karyotype without trisomy 8. Seven patients had greater than 2.0% nuclei with 3 interphase FISH signals (range, 2.10% to 3.40%) and six patients had 2 or more cells with trisomy 8 upon metaphase FISH or extensive conventional cytogenetic analysis. Our results show that interphase and metaphase FISH analyses are useful methods to detect trisomy 8 cells in BM specimens, especially for specimens with normal or uncertain conventional cytogenetic results. 0 1992 by The American Society of Hematology.

chromosomes.8 In addition, the use of a-satellite DNA and whole chromosome paint (WCP) probes specific for individual chromosomes in conjunction with FISH has been proposed as a method to enumerate chromosomes within interphase nuclei and metaphase s p r e a d ~ . ~ - l ~ To maximize the probe signal, most enumeration analyses use an indirect FISH technique. Biotin-labeled DNA probes are hybridized to metaphase and/or interphase preparations, and fluorescently labeled avidin or antiavidin molecules are used to detect the p r o b e ~ . ~ - l ~ This method can be time- consuming. Recently, directly labeled DNA probes have been developed that emit enough light to be easily detected even with relatively inexpensive fluorescent light micro- s c o p e ~ . ~ ~ Such probes simplify the technique by eliminating the secondary detection step and, thus, may be ideally suited for the detection of trisomies.

The first purpose of this study was to investigate the usefulness in clinical practice of directly labeled fluorescent a-satellite DNA probes for chromosome 8 to detect trisomy 8 in interphase BM nuclei. We attempted to determine the sensitivity and specificity of the interphase FISH method by

From the Section of Laboratory Genetics, the Division of Biostatis- tics, and the Division of Hematology and Internal Medicine, Mayo Clinic and Foundation, Rochester, MN; and the Section of Hematol- ogylOncology, The University of Chicago Pritzker School of Medicine, Chicago, IL.

Submitted September 20, 1991; accepted February 21, 1992. R.B.J., M.M.L., and G. W.D. aresupported by grants fiom Imagenet-

ics, Inc, Naperville, IL. Address reprint requests to Robert B. Jenkins, MD, PhD, Cytogenet-

ics Laboratoy, Mayo Clinic, 200 First St, SW; Rochester, MN 55905. The publication costs of ihis article were defrayed in part by page

charge payment. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. section 1734 solely to indicate this fact.

0 1992 by The American Society of Hematology. 0006-4971 1921 7912-0030$3.00/0

Blood, Vol79, No 12 (June 151, 1992: pp 3307-3315 3307

For personal use only.on April 12, 2019. by guest www.bloodjournal.orgFrom

http://www.bloodjournal.org/

http://www.bloodjournal.org/site/subscriptions/ToS.xhtml

3308 JENKINS ET AL

analysis of two randomized series of normal BM donors and patients with varying proportions of metaphase cells with trisomy 8. Each specimen was studied by two different technologists who were not aware of each other's results, the patient's clinical status, or the cytogenetic findings until all studies were complete. A similar randomized series was performed to assess ability of directly labeled WCP probes to detect trisomy 8 within metaphase spreads. We then studied patients with clinically proven MPD, MDS, or ANLL whose BM had either no apparent chromosome abnormality or an abnormal karyotype without trisomy 8 to determine if any of these patients had occult trisomy 8 that was not detected by routine cytogenetic analysis.

MATERIALS AND METHODS

Patients and normal donors studied. For many years both the Mayo Clinic (MC) and The University of Chicago (UC) Hemato- logic Cytogenetics Laboratories have stored residual BM cell pellets from specimens that were processed for clinical purposes. The hematologic disorders and cytogenetic results associated with these specimens are maintained in computerized databases. For the first set of MC interphase FISH experiments, we retrieved all stored specimens from normal BM transplant donors collected between July 1988 and December 1990. These BM donors were hematologically normal by history, general physical examination, peripheral blood counts, and by morphologic/cytogenetic analysis. The conventional cytogenetic results on these and other normal BM specimens studied at MC have been reported e1se~here. l~ Briefly, metaphase cells with trisomy 8 were not observed in any specimens from the normal donors. For the initial study, the MC also identified specimens with trisomy 8 collected during the same time period and retrieved 57 specimens with different proportions of metaphase cells with trisomy 8 for interphase FISH analysis. By conventional cytogenetic analysis, 12 of the specimens had 1 of 30 metaphase cells with trisomy 8, one had 1 of 20 metaphases with trisomy 8, and the remaining 44 had from 10% (2 of 20) to 100% (20 of 20) trisomy 8 metaphase cells.

The MC series of 57 patients with trisomy 8 and 25 normal donors were then randomly ordered. The cellular suspensions were processed and the resultant slides were given to the two technologists performing interphase FISH analysis. Neither technologist was aware of the original cytogenetic findings, the clinical history, nor the results of the other individual until all specimens were analyzed. A metaphase FISH study of the same group of patients was performed by a single technologist. This set of experiments was separately randomized.

For the second MC study, we searched the MC database for BM specimens that had either no apparent chromosome abnormality or a clonal abnormality other than trisomy 8 by conventional cytogenetic analysis. The clinical history for each of these patients was reviewed, and those patients who did not have morphologically proven MPD, MDS, or ANLL were excluded. Patients with chronic granulocytic leukemia (CGL) or CGL in accelerated phase or blast crisis were also excluded. We selected residual cell pellets from 93 specimens for interphase and metaphase FISH analysis.

For the studies at UC, BM specimens were obtained from 67 patients with ANLL ascertained between April 1989 and July 1990, and from 10 healthy volunteer individuals. These normal BM donors were hematologically normal by morphologic and cytogenetic analysis. The ANLL patients included patients who had apparently normal karyotypes (25 patients), abnormalities other than trisomy 8 (26 patients), or trisomy 8 (16 patients). The cell suspensions were processed and hybridized with the directly

labeled CEP-8 probe (normal individuals and 35 ANLL patients) or with a biotin-labeled CEP-8 probe (32 ANLL patients). The hybridized slides were coded and randomized for analysis by two individuals. The patients studied at UC had between 9% (2 of 22) and 91% (20 of 22) metaphase cells with trisomy 8.

For each specimen, metaphase and interphase nuclei preparations were made from direct BM pellets stored at -70°C in fixative (MC) or from BM cells derived from short-term cultures (24, 48, or 72 hours) stored at -20°C in fixative (UC). Pellets were resuspended in fresh fixative (3:l methano1:glacial acetic acid) and dropped on clean slides. Slides were cleared in 100% ethanol and air dried before hybridization.

Interphase FISH was performed at MC using a directly labeled a-satellite DNA probe specific for chromosome 8 (SpectrumOr- ange CEP-8; Imagenetics, Naperville, IL) as follows: 2 pL Spec- trumorange-labeled CEP-8 and 1 pL sterile water in 7 pL of hybridization mix (MM2.1; Imagenetics) was applied to chromosome preparations, coverslipped, and sealed with rubber cement. Chromosomal and probe DNA were denatured simultaneously for 3 minutes in a 90°C oven and hybridization was allowed to take place at 42°C overnight in a humidified chamber. Posthybridization washes were as follows: three rinses in 50% formamide/2X SSC at 45°C for 15 minutes each, two rinses in 2X SSC at 45°C for 15 minutes each, one rinse in 2X SSC/O.l% NP-40 at 45°C for 15 minutes, followed by one rinse in 2X SSC/O.l% NP-40 at room temperature. Interphase nuclei were counterstained with 1 pg/mL DAPI (Sigma, St Louis, MO) containing the antifade compound 1,4 diazabicyclo[2.2.2]octane (DABCO; Sigma). Hybridization sites were analyzed using microscopes equipped with appropriate filter sets for visualizing rhodamine and DAPI counterstain. Two individuals each scored 1,000 nuclei from each BM specimen for the number of hybridization signals. Individual nuclei were first visual- ized using the DAPI filter set; then the Spectrumorange filter set was placed in the optical axis and the number of hybridization signals per nucleus was determined. Broken or overlapping nuclei were excluded.

The procedure used for interphase FISH at UC was as follows: slide preparations were treated with RNAase (100 pg/mL in 2X SSC for 1 hour), rinsed extensively in 2X SSC, and dehydrated in a graded ethanol series. DNA was denatured by immersion of the slides in 70% formamide-4X SSC (pH 7.0,70"C for 2 minutes). The hybridization mixture (described above) containing the directly labeled probe was heated at 75°C for 5 minutes and then incubated at 37°C for 5 to 10 minutes to promote partial reannealing. Ten microliters of the probe solution was applied to each slide, and the slides were coverslipped and incubated at 37°C for 16 hours. After hybridization, the slides were washed in 50% formamide-4X SSC (three washes, 5 minutes each at 40"C), followed by three washes in 4X SSC at 40°C (10 minutes each). Interphase nuclei were counterstained with DAPI (200 ng/mL in 2X SSC for 5 minutes) and the slides were mounted in 20 mmol/L tris-HCl,90% glycerol containing 2.3% DABCO antifade. The procedure for analysis was identical to that used at MC with the exception that each technologist analyzed 500 nuclei per specimen. Specimens from 32 individuals were analyzed using a biotinylated CEP-8 probe (Biotin- CEP-8; Imagenetics) using previously published indirect methods."

Metaphase FISH using a directly labeled WCP probe specific for chromosome 8 (Spectrumorange WCP-8; Imagenetics) was performed as follows: 1.7 pL Spectra orange-labeled WCP-8, 1.3 pL Cot 1 DNA in 7 pL of hybridization mix (MM1.l; Imagenetics) was applied to chromosome preparations, coverslipped, and sealed with rubber cement. Chromosomal and probe DNA were denatured simultaneously for 3 minutes in a 90°C oven and hybridization was allowed to take place at 37°C overnight in a humidified

FISH methods.




TRISOMY 8 DETECTION BY FISH 3309

chamber. Posthybridization washes were performed as previously described for the CEP-8 probe. Up to 500 metaphase cells were scored for the number of hybridization signals. A few slides to be hybridized with WCP-8 were first stained with Wright’s stain and suitable metaphase cells were analyzed and photographed with Technical Pan 2415 film (Kodak, Rochester, NY). These slides were destained in 100% ethanol and air-dried for WCP-8 analysis. In all instances, only the cytologically identifiable chromosome 8 was stained with the WCP-8 probe.

These studies were performed at MC and involved three BM specimens. The first BM sample contained 100% (20 of 20) metaphase cells with trisomy 8 by standard cytogenetic analysis, the second sample contained 7% (2 of 30) trisomy 8 metaphase cells, and the third contained no apparent chromosome abnormality. To study the intrarun variation, 10 slides prepared from each BM specimen were hybridized overnight with aliquots of the same probe solution. Two technologists each analyzed five slides from each specimen. To study the interrun variation, two slides were prepared from each of the three BM specimens on each of 4 consecutive days. A newly diluted Spectrumorange-labeled CEP-8 probe solution was prepared and hybridized to the three slide pairs each day. Each technologist analyzed one slide from each specimen on the day hybridization was complete.

Interrun and intrarun interphase FISH variation studies.

RESULTS

We found no statistically significant interrun and intrarun interphase FISH variation between slides from the same patient (data not shown). Thus, we developed two randomized studies at each of two institutions to evaluate the use of a directly labeled fluorescent a-satellite probe for chromosome 8 to detect trisomy 8 within interphase nuclei. Figure 1 shows representative microscopic views using FISH and the chromosome 8 a-satellite probe for two BM specimens.

Interobserver and interinstitutional interphase FISH variation studies. We next attempted to establish the interobserver variation for the number of chromosome 8 signals in interphase nuclei for the 35 normal BM specimens in this series. We first compared the findings between observers for each patient with respect to the number of nuclei with 0, 1,2 ,3 , or 2 4 hybridization signals. We found that, for any single normal BM specimen, the difference between observers was never more than 1.2%, 2.2%, 3.0%, 2S%, or 0.8%, for nuclei with 0, 1,2,3, or 2 4 signals, respectively.

We then compared the distribution of hybridization signals per nucleus for the two observers at each institution. We calculated the mean and standard deviation for nuclei with 0, 1,2,3, and 4 signals in the normal samples for each observer at each institution. No statistically significant difference between the technologists at MC was found for the mean number of nuclei with 2, 3, or 4 signals using a paired-difference t-test. A statistically significant difference was found between the MC technologists for the mean number of nuclei with 0 or 1 signals, but this difference was less than 0.6%. At UC no statistically significant difference between technologists was found for the mean number of nuclei with 0, 1,2,3, or 4 signals.

Importantly, although for the normal specimens there were statistically significant interobserver and interinstitutional differences in some interphase FISH signal catego-

ries, there was no statistically significant difference between observers or between institutions with respect to the percent of nuclei with 3 signals.

Figure 2 illustrates the interobserver variation in the number of nuclei with 3 signals for the 60 BM specimens that had trisomy 8 in at least 2 of 20 metaphases by conventional cytogenetic analysis. The findings of the technologists correlated closely at both UC and MC, especially among those cases with a higher proportion of nuclei with 3 signals. None of the slopes of the calculated regression lines were significantly different from 1.

Detection of trisomy 8: A comparison of interphase FISH with routine cytogenetic analysis. Because the results for the chromosome 8 interphase FISH results correlated strongly between technologists at each institution, we used the average of the technologists’ values for each specimen in the remaining analyses. Figure 3 illustrates the correlation between the percentage of trisomy 8 cells as determined by conventional cytogenetic studies and the percentage of nuclei with 3 signals for the 60 patients with at least 2 trisomy 8 metaphase cells. The estimates of the percentage of interphase cells with 3 signals generally correlated with the percentage of metaphase cells with trisomy 8 (correlation coefficient = .90). However, the interphase FISH estimate of trisomy 8 cells was usually less than the cytogenetic estimate (slope = 0.83; 95% confidence limits 0.73 to 0.93).

Sensitivity and speciJicity of interphase FISH detection of trisomy 8. The cumulative distribution of BM specimens whose percent of nuclei with 3 signals was less than or equal to a given percent are presented in Fig 4 for three groups of patients: the 35 normal donors, the 13 patients with only 1 trisomy 8 metaphase cell, and the 60 patients with 2 or more trisomy 8 metaphase cells. Figure 4 clearly illustrates that there is minimal overlap between the distribution of the normal specimens and that of the specimens with 2 or more trisomy 8 metaphase cells. Figure 4 can also be used to judge specificity (the cumulative percent of normal patients) and sensitivity (one minus the cumulative percent for patients with 2 or more trisomy 8 metaphase cells) of a particular cutoff percentage value of 3 nuclear signals. Figure 4 also shows that the distribution of the specimens with only 1 trisomy 8 metaphase cell overlaps the other two distributions. These patients were excluded from the following sensitivity and specificity analysis because we could not be certain about their results.

Thirty-three of the 35 normal BM specimens (a specificity of 94%) contained s 1.75% nuclei with 3 signals. If this cutoff value for the upper limit of normal is arbitrarily selected, then 58 of 60 patients with 2 or more trisomy 8 metaphases would have been determined to be abnormal (a sensitivity of 97%). Each of the normal BM specimens contained 12.0% nuclei with 3 signals (a specificity of 100%). If this cut-off value is arbitrarily selected as the upper limit of normal, then 57 of 60 patients with 2 or more metaphases would have been determined to be abnormal (a sensitivity of 95%). The three specimens that would have been classified as normal by this criteria each had 2 of 20 metaphases with trisomy 8 by conventional cytogenetic analysis (Table 1). Seven other BM specimens with 2 of 20




3310 JENKINS ET AL

Fig 1. Representation microscopic fields (original magnification x500) from BM specimens analyzed using FISH and the CEP-8 probe that by conventional cytogenetic analysis contained (A) 100% trisomy 8 metaphase cells (patient no. 25671) and (B) approximately 50% trisomy 8 (patient no. 46594). A dual pass rhodamine/DAPI filter was used to obtain these photo- graphic images. For details of experimental procedures see Ma- terials and Methods.

or 2 of 30 metaphase cells with trisomy 8 by cytogenetic analysis had greater than 2.0% of nuclei with 3 signals (Table 1). No BM specimen with greater than 2 of 20 metaphases (50 samples) with trisomy 8 metaphase cells contained I 2.0% of nuclei with 3 signals.

Metaphase FISH detection of trisomy 8. In an attempt to more accurately estimate the percent of metaphases with trisomy 8, at MC we performed an independent randomized study of the same group of normal BM specimen donors and specimens from patients with trisomy 8 using a directly labeled WCP probe for chromosome 8 (WCP-8). When the code was broken and the data analyzed, none of the normal BM specimens contained more than 1 metaphase with three WCP-8 hybridization signals, even when 500 metaphase cells are analyzed. Based on this result,

specimens with 2 or more than 3 metaphases with 3 WCP-8 signals were arbitrarily classified as abnormal. Using these criteria, the metaphase FISH method had a sensitivity of 98% (42 of 43 patients with cytogenetic trisomy 8 were abnormal by metaphase FISH) and a specificity of 100% (24 of 24 of normal BM specimens were normal by metaphase FISH). The percent of metaphases with trisomy 8 by WCP-8 analysis correlated better with the percent of nuclei with 3 interphase FISH signals than with the percent of metaphases with trisomy 8 by routine cytogenetic analysis (data not shown).

FISH analysis of specimens with equivocal cytogenetic results. Nine of the 10 specimens with 2 of 20 or 2 of 30 metaphase cells with trisomy 8 by conventional cytogenetic analysis provided a suitable number of additional met-




TRISOMY 8 DETECTION BY FISH 331 1

IlXl I

0 10 20 30 40 50 60 70 80 90 100

Observer 1 (% of Nuclei with 3 interphase FISH Signals)

Fig 2. The interobserver variation in the percent of nuclei with 3 interphase FISH signals on 60 BM specimens with at least 2 of 20 metaphase cells with trisomy 8 by conventional cytogenetic analysis. (0) 44 MC patients; (- - -) fitted regression line for these 44 patients [fl = ,996. slope = 0.995); (0) 7 UC patients (fl = .930, slope = 1.20); [---) fitted regression line for these seven patients; (+) nine UC patients analyzed using indirect interphase FISH; [---) fitted regression line for these nine patients (fl = .91, slope = 0.81).

aphase cells for WCP-8 analysis. Eight of the specimens had at least 4 metaphases with 3 specific WCP-8 signals (Table 1 and Fig 5). These findings support the conclusion that the BM of each of these 8 patients most likely contained clonal trisomy 8. The BM of one patient (52956) with 1.75% nuclei with 3 interphase FISH signals did not contain any metaphase cells with 3 WCP-8 signals (500 metaphases were

0 10 20 30 40 50 60 70 80 9 0 1 0 0

% Metaphases with +8

Fig 3. Correlation of the percent nuclei with 3 interphase FISH signals with the percent of trisomy 8 metaphase cells by conventional cytogenetic analysis. Each BM specimen contained at least 2 of 20 trisomy 8 metaphase cells by conventional cytogenetic analysis. (0) 44 MC patients; (0) 16 UC patients; (- - -1 fitted regression line for all patients (fl = .90, slope 0.83); (-) predicted unit slope line.

examined). This patient had a 7-year history of the 5q- syndrome, had a de1(5)(q13q33) in 18 of 20 metaphase cells and trisomy 8 in 2 of 20 metaphase cells.

Four of 13 patients with only one trisomy 8 metaphase cells by conventional cytogenetic studies had greater than 2.0% nuclei with 3 interphase FISH signals (Table 2). Each of these four patients proved to have at least 4 metaphase cells with three WCP-8 signals when additional metaphases were studied. The medical records for 2 of the latter 4 patients were available for review and both had hematologic and/or morphologic evidence of MDS or ANLL. For example, patient 50516 presented with a possible MDS and 2.95% nuclei with 3 signals. He returned 4 months later and a second BM specimen (no. 53879) demonstrated overt MDS and 2 of 20 metaphase cells with trisomy 8. In this specimen we observed 8.70% of nuclei with 3 signals. In these two patients, the positive interphase FISH and metaphase FISH results and the positive clinical histories suggest that the single metaphase with trisomy 8 found by routine cytogenetics may truly be a result of a clonal neoplastic process. Of the remaining 9 BM specimens with I 2.0% nuclei with 3 signals, two had at least 4 metaphases with 3 WCP-8 signals (patients 39974 and 45009 in Table 2), also suggesting that they may truly have a clonal neoplastic process, including trisomy 8. The remaining 7 BM specimens had no apparent trisomy 8 metaphases by WCP-8 analysis. The medical records of 4 of these 7 patients were available for review. Three patients had equivocal hematologic and morphologic workups, suggestive but not diagnostic of a neoplastic process. One other patient had ANLL in progression from agnogenic myeloid metaplasia. The absence of trisomy 8 cells by interphase FISH and metaphase FISH analysis, and the equivocal clinical histories in 3 of the 4 patients, suggest that the single metaphase cell with trisomy 8 found by conventional cytogenetic analysis in these seven patients may not represent an abnormal clone.

FISH detection of trisomy 8 in patients with MPD, MDS, or ANLL with normal cytogenetics or an abnormal karyotype without trisomy 8. Using the directly labeled chromosome 8 probes, we next analyzed BM specimens from 144 patients with clinically proven MDS (43 patients), MPD (14 patients), or ANLL (87 patients) that by conventional cytogenetic analysis contained no apparent chromosome abnormality or a clonal abnormality other than trisomy 8. Using the a-satellite probe two technologists at MC each independently analyzed 1,000 interphase nuclei from 93 of these patients. Using the WCP-8 probe, a single technologist from MC also analyzed up to 500 metaphases from the same patients. Two technologists at UC used the a-satellite probe to analyze 500 interphase nuclei from 51 patients. The BM specimens of seven patients contained greater than 2.0% nuclei with 3 interphase FISH signals (Table 3). Three of these patients had MDS, two MPD, and two ANLL. Two of the seven patients had another acquired chromosome abnormality. Three of the seven patients had at least 3 metaphase cells with 3 WCP-8 signals and a fourth had 2 metaphases with trisomy 8 when 90 additional metaphase cells were cytogenetically analyzed. A fifth patient was discovered to have a large number of tetrasomy




JENKINS ET AL

Fig 4. The cumulative distribution of BM specimens categorized bv the Dercent of nuclei with 3 /

/ - interphase FISH signals. (-----) 35 normal BM donors; (-.-.-) 13 patients with only 1 trisomy 8 metaphase cell; (-) 60 patients with 2 or more trisomy 8 metaphase cells. The vertical lines indicate 1.75%

I 1 I I I I 1 3 1 0 30 10 0 30 0 l o b o

Percent of nuclei with 3 Interphase FISH Signals

and octosomy 8 metaphases upon WCP-8 analysis. Two additional patients were found to have normal interphase FISH results and 2 metaphases with three WCP-8 signals. Both had other cytogenetic anomalies and one had an increased percentage of tetraploid cells.

DISCUSSION

Many patients present to hematologists and other physi- cians with overt signs and symptoms of MDS, MPD, or acute leukemia. Other patients present with vague clinical findings. Both sets of patients often have BM examinations and, if MDS, MPD, or acute leukemia is in the differential diagnosis, then cytogenetic analysis is usually performed. By conventional cytogenetic studies many of these patients are found to have an abnormal clone [eg, trisomy 8, monosomy 7, del(20q), etc] that may help explain their clinical signs and symptoms. However, the cytogenetic findings in other patients can be difficult to interpret. For example, a single metaphase cell may be found in the BM of these patients with a cytogenetic anomaly commonly observed in hematologic disorders [eg, trisomy 8, monosomy

and 2.0% nuclei with 3 signals.

7, del(20q), etc]. On one hand, this finding may be the sole evidence that the patient has a neoplastic hematologic disorder because morphologic examination of the BM may be nondiagnostic. On the other hand, that single abnormal metaphase cell could also be due to an artifact of the cytogenetic methodology. Thus, an independent method of determining the presence or absence of specific chromosome anomalies would be extremely useful in clinical practice. Consequently, we undertook an evaluation of new methods to detect trisomy 8: interphase FISH using a directly labeled DNA probe specific for the centromeric region of chromosome 8 and metaphase FISH using a directly labeled WCP probe.

We believe that it is important to establish the normal range for chromosome-specific interphase FISH analysis before this method is used in clinical practice. Known problems and artifacts of the FISH procedure (partial hybridization, variations in slide and sample quality, auto- fluorescence and fluorescent debris, how one defines a “specific” fluorescent signal, superimposition of fluorescent signals, and polyploidy) make an appropriate normal value

Table 1. Chromosome 8 FISH Results on BM Specimens With 2 of 20 Metaphases With Trisomy 8 by Routine Cytogenetics

Metaphases Cytogenetic Results CEP-8 Hybridization Signals

(average per 1,000 nuclei) File Cells With Other With 3 No. Trisomy8 Abnormalities 0 1 2 3 4 5 26 WCP-8 Signals BM Diagnosis Clinical Diagnosis

MDS-RAEB 49932 2/20 None 0 10.5 974 14.5 1 0 0 4/500 MDS, 10% blasts 52956 2/20 de1(5)(q13q33) 0.5 13 965.5 17.5 3.5 0 0 0/500 MDS, 12% blasts 5q- syndrome 24651 2/20 None 0 11.5 970 18 0.5 0 0 4/200 MDS, 5% blasts MDS-RAEB

26388 2/20 None 1.5 10 954.5 33.5 0.5 0 0 15/500 NA NA 22955 2/30 None 13 31 911 42.5 2.5 0 0 ND NA NA 49759 2/20 None 1.5 6.5 941 49 1.5 0 0.5 4/353 Plasmacytosis MGUS, possible ther-

44979 2/20 None 0 6.5 931 61.5 1 0 0 25/155 NA NA 48208 2/20 None 1.5 8 908 77.5 4.5 0.5 0 25/151 MDS, <5% blasts MDS-RARS 23927 2/20 None 0 3.5 874 118 3.5 0.5 0.5 39/100 ANLL, 12% peripheral ANLL

24521* 2/20 None 3 9 946 40.5 1 0 0.5 8/200 Remission marrow ANLLin remission

apy-related MDS

blasts

Abbreviations: NA, not available; ND. not done; RARS. refractory anemia with ringed sideroblasts; RAEB, refractory anemia with excess blasts;

*A second specimen from patient no. 23927; 2 months later after the initiation of chemotherapy. MGUS, monoclonal gammopathy of uncertain significance.




TRISOMY 8 DETECTION BY FISH 331 3

Fig 5. (A a d B) Two mpmsontative metaphases with 3 WCPB *nab found in a BM specimen from a patient (no. 50516) with 1 of 30 metaphose cells wtth trisomy 8 by conventional cytogenetic analysis (original magnification ~ 5 0 0 ) .

study critical. To accomplish this, we set up blinded protocols at two institutions to analyze a combined series of normal BM donors and patients who were found to have trisomy 8 by conventional cytogenetic analysis. This blinded analysis provided an accurate assessment of the FISH methods' sensitivityand specificity and thusestimatesof the upper limit of normal. Furthermore, the random inclusion of patients with a single metaphase with trisomy 8 by

conventional techniques provided a means by which we could assess the diagnostic usefulness of FISH analysis in such patients.

After analysis of the data, we decided to select an upper limit of normal for interphase FISH analysis that would be 95% sensitive and 100% specific. This value proved to be greater than 2.0% nuclei with 3 signals. We also decided that greater than 2 metaphases with 3 metaphase WCP-8

Table 2 C h " o 8 FISH Rsrultr on BM Specimens From Patients With Uncertlin Knyotyp. Results

Wogemtic Rerub CEP-8 Hybridization Signals MelsPhsSSS

With 3 (average per 1.OOO nuclei) File Calla WRh Other

No. Triaomv8 AbnormalRh 0 1 2 3 4 5 r6 WCP-8Signnla BM Diagnosis Clinicel Diagnoaia

43792 1/30 None 1 16 975.5 4 3 0.5 0 01200 NA NA' 23766 1/30 None 1.5 27 965 5 1.5 0 0 01500 Suggests MDS MDS-RA or anemia of

39953 1/30 None 1.5 20.5 971.5 6 0.5 0 0 01500 NA NA 51314 1/30 del(ll)(q?21q23)* 1 16.5 973 6.5 3 0 0 01500 Fibrotic, 6% blasts AMM progressing to

39974 1/30 None 0 9.5 981.5 8 1 0 0 61300 NA NA 47129 1/30 None 3 14 971 10.5 1.5 0 0 01500 NA NA 52599 1/30 None 1.5 7.5 969.5 17.5 4 0 0 01200 Relative erythroid hy- Microangiopathic ane-

perplasia mia, possible MDS 49485 1/30 None 0.5 11.5 966.5 19 2 0 0.5 01200 Erythroid hyperplasia, MDS-RAoranemiaof

Dimorphic periph- chronic disease era1 blood smear

chronlc disease

ANLL

45009 1/30 None 1 10 968.5 19 1.5 0 0 41200 NA NA

25360 1/30 Nonet 0.5 11 963.5 24 1 0 0 51200 ANLL-M4 ANLL-M4 50516 1/30 None 1 12 951.5 29.5 4.5 0.5 1 41500 Erythroid hyperplasia Possible MDS 53879* 2/20 None b 6 9 0 7 8 7 0 0 0 ND MDS MDS-RA or MDS-RAEB 42274 1/30 None 2.5 11.5 942.5 39.5 4 0 0 221200 NA NA 42193 1/20 None 0.5 8 889.5 100 2 0 0 241200 NA NA

Abbreviations: RA, refractory anemia; AMM, agnogenic myeloid metaplasia. *One cell only. tPatient presented 9 months later with 14 of 20 metaphases with +21. *Not part of randomized study; second presentation of patient no. 50516 4 months later.




3314

Table 3. Chromosome 8 FISH Results on BM Specimens From Patients With ANLL, MDS, or MPD With No Apparent Cytogenetic Abnormality of Chromosome 8

JENKINS ET AL

File No. Karyotype

CEP-8 Hybridization Signals (average per 1,000 nuclei)

Clinical With 3 Metaphases

Diagnosis 1 2 3 4 5 26 WCP-8Signals

41 344 24889 24613 46730 39245 26092 26343 41228 14070

20=45,XY,-7,inv(3)(q21 q25) 29=46,XY/1=47,XY,+9 20=46,XY 20=46,XX,inv(2)(pl1.2q13)c 20=46,XY 20=46,XY 20=46,XY 4=46,XX/ 16=46,XX,t(l; 15)(pU.?1 ;ql3) 28=46,XY/16=46,XY,t(17;?)(pl l;?)

ANLL CMD MDS-RAEB CMD CMD MDS-RA MDS-RA ANLL-M2 ANLL

5.5 986 8 981.5

14 962 8 966

12.5 962.5 8.5 965 8 945.5

17 946 26 884

7.5 10 21 22 23 26 31.5 34 79

1 0 0 0.5 0 0 2 0 1 3.5 0 0.5 1.5 0 0.5 0.5 0 0

13 1 1 3 0 0 5 0 0

2/500* 21246 01563 01500 4/ 100 31500 01500t 31100 2190*

Abbreviations: CMD, chronic myeloproliferative disorder; c, constitutional anomaly. *This specimen contained 13 metaphases with 4 WCP-8 signals. tThis specimen contained 23 and 6 metaphases with 4 and z 6 WCP-8 signals, respectively. *Seventy metaphases were reanalyzed by routine cytogenetics.

signals was also abnormal. This value proved to be 98% sensitive and 100% specific.

We identified 4 BM specimens with uncertain conventional cytogenetic results that interphase FISH analysis suggested might have clonal trisomy 8. The WCP-8 analysis of metaphases from these BM specimens and the clinical and other laboratory data for these patients were consistent with this conclusion. In fact, one of these patients clearly progressed to overt clonal trisomy 8 with MDS 4 months later. Similar findings have recently been reported in studies using only conventional cytogenetic' and interphase ana1y~is.l~ In our series, 9 BM specimens with uncertain conventional cytogenetic results had I 2.0% nuclei with 3 interphase FISH signals. Based on the WCP-8 analysis, 2 of the 9 specimens actually had clonal trisomy 8, and so might be considered false-negative results of the interphase FISH method.

Both interphase and metaphase FISH analysis clearly detected all of the BM specimens with greater than 2 of 20 metaphase cells with trisomy 8 by conventional cytogenetic studies. In addition, interphase FISH analysis detected 7 of 10 BM specimens with 2 of 20 or 2 of 30 metaphase cells with cytogenetic trisomy 8. Two of the three interphase FISH negative specimens were found to have additional trisomy 8 metaphases upon metaphase FISH. These could be considered false-negative results of the interphase FISH method. One of these three patients had negative interphase FISH and metaphase FISH studies. This patient had a 7-year history of 5q- syndrome and had chromosome studies performed on several previous BM specimens. Trisomy 8 metaphase cells were observed only in one specimen by conventional cytogenetic analysis and neyer in cells with a del(5q). Thus, it is quite possible that the trisomy 8 found in this specimen by conventional cytogenetic analysis was a false-positive result. The results of studies on future specimens from this patient may help clarify this issue.

Our study also shows that a small but significant fraction of patients with MDS, MPD, and ANLL with either no apparent chromosome abnormality or a clonal karyotype without trisomy 8 have occult trisomy 8 demonstrable by

interphase and metaphase FISH analysis. Although increased polyploidy must be excluded as a cause of positive results, we believe that FISH methodology is an excellent way to detect occult trisomy 8. Importantly, the positive patients by metaphase FISH analysis might be considered false-negative results of conventional cytogenetics. The true clinical relevance of low levels of trisomy 8 awaits a prospective analysis of a large patient series.

If the metaphase FISH method (used at MC) and extensive conventional metaphase analysis (UC) is used as the standard reference method of trisomy 8 detection, then interphase FISH correctly identified the presence of a trisomy 8 clone in 96.0% (242 of 252) of the BM specimens we studied (Table 4). Conventional cytogenetic analysis correctly identified a trisomy 8 clone in 91.6% (231 of 252) of the specimens.

Thus, we believe that FISH detection of trisomy 8 is a useful follow-up screening test in cases with uncertain or normal conventional cytogenetic results. Once at least 20

Table 4. Predictive Value of Interphase FISH and Conventional Cytogenetic Analysis of Trisomy 8

Interphase FISH Results

Total Not Correct Cytogenetics Correct

Correct 227 4' 231 Equivocal 11 2 13 Not Correct 4 t 4* 8 Totals 242 10 252

In cases of uncertainty, extensive metaphase analysis served as the deciding factor for the classification as correct or incorrect. The interphase FISH results were classified by using a criteria of greater than 2.0% nuclei with 3 signals as abnormal.

*Includes two patients that had negative cytogenetic results and positive interphase FISH results (Table 3) and two patients that had 2 metaphases with trisomy 8 by cytogenetic analysis and negative interphase FISH results (Table 1).

tlncludes the four patients that had negative cytogenetic results and positive interphase and metaphase FISH results (Table 3).

Slncludes one patient with 5q- syndrome, one patient with increased tetrasomy 8, and the two patients with positive metaphase FISH results (Tables 1 and 3).




TRISOMY 8 DETECTION BY FISH 3315

nuclei with 3 interphase FISH signals have been identified by one observer in 500 randomly sampled nuclei (eg, 2 4.0% nuclei with 3 signals), further analysis of more cells by the same or another observer may not be necessary. Although in our series the interobserver variation was quite small, when the percentage of nuclei with 3 signals found by a single observer is low there may be a significant number of false-positive and false-negative results. Thus, we recom- mend that when less than 20 nuclei with 3 signals are observed per 500 cells then a second individual should independently analyze another 500 nuclei (if that many cells are available). In addition, we are currently performing extensive metaphase analysis with the WCP-8 probe when 1 in 30 metaphases is positive by conventional cytogenetics or when the number of nuclei with 3 interphase FISH signals is greater than 15 or less than 40 per 1,000 nuclei (500 nuclei analyzed by two observers).

The interphase and metaphase FISH methods consis- tently identified a smaller percentage of trisomy 8 cells than was found by conventional cytogenetic analysis. FISH methods might underestimate the percentage of trisomy 8 cells for several reasons. First, chromosomally abnormal cells may spread more efficiently than normal cells under the conditions of cytogenetic harvest and thus be selected for conventional chromosome analysis. Second, the probes may not efficiently penetrate all nuclei or metaphases, producing inefficient hybridization, or the hybridization signals may overlap. It is possible to correct for these artifacts,16 but we did not use them in this study. Finally, because conventional cytogenetic analysis usually involved the examination of only a small number of metaphase cells

there is ample opportunity for sampling error. In fact, the error bars surrounding each of our cytogenetic estimates of metaphase cells with trisomy 8 would in many cases encompass the fitted regression line illustrated in Fig 2, but not the predicted unit slope line. Although sampling error can explain some of the difference between the two methods, there are too many cases in which the percentage of trisomy 8 cells estimated by interphase FISH was less than the estimate by cytogenetic analysis for this to be the sole explanation. The data suggests that the differences are due more to real biologic processes rather than to methodologic differences.

Our study shows that FISH analysis using directly labeled DNA probes specific for chromosome 8 is an excellent way to detect trisomy 8 cells in BM specimens. FISH could be used to detect trisomy 8 in patients with myeloid neoplasms examined prospectively, or it could be used to supplement conventional cytogenetic studies in patients with normal or uncertain results. This technique could also be used to monitor the effects of treatment, such as hematopoietic growth factor therapy, in patients found to have trisomy 8 at the time of diagnosis, or to detect early relapse. We believe this study predicts that such analyses could be extended to other chromosomal trisomies and monosomies; although the latter will require even stricter attention to normal value studies.

ACKNOWLEDGMENT

We thank Karen Erwin and Robin Hythecker for processing the manuscript and we thank Sherry Cooke for preparation of the figures.

REFERENCES

1. Second International Workshop on Chromosomes in Leuke- mia, 1979: Chromosomes in preleukemia. Cancer Genet Cytogenet 2:108,1980

2. Rowley JD, Golomb H, Vardiman J: Nonrandom chromosome abnormalities in acute leukemic and dysmyelopoietic syndromes in patients with previously treated malignant disease. Blood 58:759, 1981 3. Fourth International Workshop bn Chromosomes in Leuke-

mia, 1982. Cancer Genet Cytogenet 11:251,1984 4. Yunis J J Recurrent chromosomal defects are found in most

patients with acute nonlymphocytic leukemia. Cancer Genet Cyto- genet 11:125, 1984

5. Knapp RH, Dewald GW, Pierre R V Cytogenetic studies in 174 consecutive patients with preleukemic or myelodysplastic syndromes. Mayo Clin Proc 60507,1985

6. ISCN (1985): An international system for human cytogenetic nomenclature. Birth Defects 21:1,1985 7. McConnell TS, Duncan MH, Foncar K, Southwestern Oncol-

ogy Group Leukemia Cytogenetics Subcommittee: Do random [non-clonal] chromosome abnormalities in bone marrow predict a clone to come? Cancer Genet Cytogenet 53:257,1991

8. Trask BJ: Fluorescence in situ hybridization: Applications in cytogenetics and gene mapping. Trends Genet 7:149,1991 9. Pinkel D, Straume T, Gray JW: Cytogenetic analysis using

quantitative, high-sensitivity, fluorescence hybridization. Proc Natl Acad Sci USA 83:2934,1986 10. Cremer T, Landegent J, Bruckner A, Scholl HP, Schardin

M, Hager HD, Devilee P, Pearson P, van de Ploeg M: Detection of chromosome aberrations in the human interphase nucleus by

visualization of specific target DNAs with radioactive and non- radioactive in situ hybridization techniques: Diagnosis of trisomy 18 with probe L1.84. Hum Genet 74:346,1986

11. Rowley JD, Diaz MO, Espinosa R, Patel YD, van Melle E, Ziemin S, Taillon-Miller P, Lichter P, Evans GA, Kersey JH, Ward DC, Domer PH, Le Beau MM: Mapping chromosome band llq23 in human acute leukemia with biotinylated probes: Identification of llq23 breakpoints with a yeast artificial chromosome. Proc Natl Acad Sci USA 87:9358,1990 12. Poddighe PJ, Moesker 0, Smeets D, Awwad BH, Ramaek-

ers FCS, Hopman AHN: Interphase cytogenetics of hematologic cancer: Comparison of classical karyotyping and m situ hybridization using a panel of eleven chromosome specific DNA probes. Cancer Res 51:1959,1991 13. Kibbelaar RE, van Kamp H, Dreef EJ, Wessels JW, Bever-

stock GC, Raap AK, Fibbe WE, den Ottolander GJ, Kluin PM: Detection of trisomy 8 in hematologic disorders by in situ hybridization. Cytogenet Cell Genet 56:132, 1991 14. Morrison L, Legator M: Multi-color in situ hybridization

using direct-labeled fluorescent probes. Cytometry Suppl 5:131, 1991

15. Kuffel DG, Schultz CG, Ash RC, Dewald GW: Normal cytogenetic values for bone marrow based on studies of bone marrow transplant donors. Cancer Genet Cytogenet 55:39,1991

16. Eastmond DA, Pinkel D: Detection of aneuploidy and aneuploidy-inducing agents in human lymphocytes using fluorescence in situ hybridization with chromosome-specific DNA probes. Mutat Res 234303,1990




1992 79: 3307-3315

R 3d Espinosa and DJ SchaidRB Jenkins, MM Le Beau, WJ Kraker, TJ Borell, PG Stalboerger, EM Davis, L Penland, A Fernald, detection in bone marrow specimensFluorescence in situ hybridization: a sensitive method for trisomy 8

http://www.bloodjournal.org/content/79/12/3307.full.htmlUpdated information and services can be found at:

Articles on similar topics can be found in the following Blood collections

http://www.bloodjournal.org/site/misc/rights.xhtml#repub_requestsInformation about reproducing this article in parts or in its entirety may be found online at:

http://www.bloodjournal.org/site/misc/rights.xhtml#reprintsInformation about ordering reprints may be found online at:

http://www.bloodjournal.org/site/subscriptions/index.xhtmlInformation about subscriptions and ASH membership may be found online at:

Copyright 2011 by The American Society of Hematology; all rights reserved.Society of Hematology, 2021 L St, NW, Suite 900, Washington DC 20036.Blood (print ISSN 0006-4971, online ISSN 1528-0020), is published weekly by the American


http://www.bloodjournal.org/content/79/12/3307.full.html

http://www.bloodjournal.org/site/misc/rights.xhtml#repub_requests

http://www.bloodjournal.org/site/misc/rights.xhtml#reprints

http://www.bloodjournal.org/site/subscriptions/index.xhtml





fluorescence in situ hybridization: a sensitive method for trisomy

Documents