Interphase FISH assays for the detection of translocations with breakpoints in immunoglobulin light chain loci

Abstract

Many B-cell malignancies bear chromosomal translocations juxtaposing immunoglobulin (IG) genes with oncogenes, resulting in deregulated expression of the latter. Translocations affecting the IG heavy chain (IGH) locus in chromosomal region 14q32 are most prevalent. However, variant translocations involving the IG kappa (IGK) locus in 2p12 or the IG lambda (IGL) locus in 22q11 occur recurrently in B-cell neoplasias. No routine methods for the detection of all breakpoints involving IG light chain loci independently of the translocation partner have been described. For this reason, we have designed 2 novel interphase fluorescence in situ hybridization (FISH) assays using differentially labeled probes flanking the IGK and IGL locus, respectively. Based on extensive control studies, the diagnostic thresholds for the detection of breakpoints were set at 0.3% for IGK and 1.4% for IGL. Fifteen cases of B-cell malignancies with cytogenetically detectable chromosomal abnormalities in 2p11-14 were investigated with the FISH assay for IGK. Breakpoints affecting the IGK locus were detected in 7 cases including all 4 variant Burkitt's translocations t(2;8)(p12;q24) and a variant BCL2-associated translocation t(2;18)(p12;q21). Other translocation partners were chromosome bands 7q21 and 16q24. Ten cases with abnormalities in 22q11-12 were investigated with the FISH assay for IGL. Breakpoints in the IGL locus were diagnosed in 7 cases including both variant Burkitt's translocations t(8;22)(q24;q11) and a t(3;22)(q27;q11) involving the BCL6 locus. Other translocation partners were 2p13-14, 4q13 and 16p12. Our results show that these FISH assays provide flexible, simple and reliable tools in the diagnosis and characterization of genetic changes in B-cell malignancies. © 2002 Wiley-Liss, Inc.

The most common cytogenetic feature of B-lymphoid malignancies is the presence of translocations between the immunoglobulin (IG) genes and oncogenes that are subsequently deregulated. Frequently, IG rearrangements involve the immunoglobulin heavy chain (IGH) locus at 14q32, which is juxtaposed to over 30 different partner genes. The most recurrent of these translocations are the t(14;18)(q32;q21) in follicular lymphoma, the t(8;14)(q24;q32) in Burkitt′s lymphoma, the t(11;14)(q13;q32) in mantle cell lymphoma and myeloma and the t(3;14)(q27;q32) in diffuse large B-cell lymphomas involving the oncogenes BCL2, MYC, BCL1 (PRAD1/CCND1) and BCL6, respectively.1

In addition to translocations involving the IGH locus, variant translocations have been described in 5–10% of B-cell neoplasms involving either the immunoglobulin kappa (IGK) light chain locus at 2p12 or the immunoglobulin lambda (IGL) light chain locus at 22q11.2 The best known of these translocations involving IG light chain loci are the variant Burkitt's translocations t(2;8)(p12;q24) and t(8;22)(q24;q11), present in up to 25% of all Burkitt's lymphomas.3 Other well-characterized translocations with breakpoints in the IGK or the IGL locus are the t(2;3)(p12;q27) and t(3;22)(q27;q11) involving the BCL6 oncogene4 and the t(2;18)(p12;q21) and t(18;22)(q21;q11) affecting the BCL2 locus.5

The different translocations involving immunoglobulin genes have a diagnostic value, can be used to monitor the clinical course of the disease and might help to identify chromosomal regions containing new potential oncogenes.6, 7 Fluorescence in situ hybridization (FISH) with 2 differentially labeled probes flanking the breakpoint has been shown to be the most reliable method to detect all translocations affecting a promiscuous locus.8, 9 If the locus flanked by the probes is intact, the differently colored signals appear colocalized. In case of a translocation, the signals are separated independently of the partner chromosome.10 Such FISH assays for the IGH locus have been successfully applied not only to diagnose well-known recurrent translocations but also to detect cytogenetically invisible, so-called cryptic aberrations in up to half of all multiple myelomas.11–13

Comparable FISH assays have not been published so far for the IG light chain loci, so that detection of breakpoints in these loci has relied mainly on chromosome banding analysis. Nevertheless, conventional chromosome analysis in B-cell malignancies is sometimes hampered by a low mitotic index of the tumor cells, poor chromosome spreading and the cytogenetic complexity of the tumor clones.14, 15 To overcome these limitations, we designed novel dual-color FISH assays using probes flanking the IGK and IGL loci. Moreover, we show that these novel assays are applicable to interphase and metaphase cells, and we establish their diagnostic efficiency.

MATERIAL AND METHODS

Samples and controls

One cell line (BL104) and 23 primary B-cell neoplasms with a total of 25 cytogenetically proven 2p11-14 (n = 15) or 22q11-12 (n = 10) rearrangements, excluding the t(9;22)(q34;q11) Philadelphia chromosome, were studied. One patient (case number 15) with a t(2;22)(p13-14;q11) translocation was included in both groups. Cytogenetic and molecular studies in patient 2 were published previously.16 Patient features are summarized in Table I. Karyotyping of tumor cells from affected tissues was performed according to standard methods.17 All cases studied were extracted from the files of the Institute of Human Genetics of the University Hospital Kiel and the Department of Genetics of the University of Navarra. Lymphocytes from 5 donors with normal karyotypes were used as negative controls. FISH analyses were performed on cells fixed in methanol/acetic acid left from cytogenetic analysis.

Clone selection

The IGK flanking probes were selected by searching the genome sequence database using the BLAST algorithm at NCBI (http://www. ncbi.nlm.nih.gov/blast) with sequences from the IGK variable (accession number X72820) and constant (accession number U72063) regions. Moreover, the physical maps deposited in the HUMANace database at Washington University (http://genome.wustl.edu/gsc) were analyzed for appropriate clones. Clones flanking the IGL locus at 22q11 were obtained from the chromosome 22 sequence deposited in the Genebank at NCBI (accession number NT_001454).

Probe preparation and FISH

BAC clones were kindly provided by the Sanger Centre (Hinxton, UK), the University of Oklahoma (Tulsa, OK) and the German Resource Center (Berlin, Germany). BAC clones were grown overnight in LB media supplemented with chloramphenicol, and DNA was isolated with the Perfectprep Plasmid Maxi Kit (Eppendorf, Köln, Germany). For the IGK and IGL assays, probes flanking the breakpoint were labeled differentially with dUTP-SpectrumOrange (Vysis, Downers Grove, IL) or dUTP-SpectrumGreen (Vysis) by use of the Bioprime labeling kit (Gibco/Life Technologies, Eggenstein, Germany) exchanging dNTP-Bio with the appropriate modified dNTP mixture. Labeled products were purified with Sephadex G-50 columns. Two hundred nanograms of each probe were coprecipitated with 5 μl of Cot1-DNA (Gibco/Life Technologies) and resuspended in 10 μl of master mix containing 2× standard saline citrate (SSC), 50% formamide and 10% dextran sulfate. Additional FISH probes used to confirm the involvement of partner genes were LSI MYC and LSI IGH/MYC (Vysis; kind prerelease by J. Proffit), LSI IGH/BCL2 (Vysis), BCL11A (2p13) and REL (2p13)18 and BCL6 (3q27).19 FISH was performed as previously described.17 Slides were analyzed by use of a Zeiss Axioskop2 fluorescence microscope (Göttingen, Germany) equipped with the appropriate filter sets (AHF, Tübingen, Germany) and documented using the ISIS imaging system (MetaSystems, Altlussheim, Germany).

Determination of cutoff levels for false-positive findings

For the determination of the diagnostic thresholds of the new IGK and IGL probe sets, about 200 nuclei were evaluated in each negative control, and the largest distance between the differently colored signals of a signal pair within the nucleus was scored. Six categories were created: (i) overlapping signals; (ii) nonoverlapping signals; (iii) distance between the signals more than once the signal diameter; (iv) distance between the signals more than twice the signal diameter; (v) distance between the signals more than 3 times the signal diameter; and (vi) distance between the signals more than 4 times the signal diameter. In analogy to previous studies,20 the diagnostic cutoffs were calculated as the mean of false-positive nuclei in controls plus 3 SD.

RESULTS

Probe identification and design of the FISH assays

Translocations involving the IG light chain loci usually lead to breakage within the joining region of the respective locus, which is flanked by constant and variable sequences. Therefore, our FISH assays applied BAC clones publicly available from the Roswell Park Cancer Institute (RPCI) or the Caltech A (CTA) libraries, which mapped to or bordered the constant and variable regions of IGK and IGL.

Based on sequence comparisons and physical maps of the chromosome region 2p11-13 deposited in the HUMANace database at Washington University, BAC RPCI-11 316G9 (labeled in red) as well as the pooled overlapping BACs RPCI-11 1021F11 and RPCI-11 525L16 (labeled in green) were selected as telomeric and centromeric probes for IGK, respectively (Fig. 1a).

Figure 1. (a–c) IG kappa. (a) Schematic representation of the IGK FISH assay showing the expected signal patterns in metaphase and interphase cells lacking and containing a breakpoint in the IGK locus. In cells with an intact IGK locus, 2 colocalizations of the signals for the centromeric and telomeric probes would be expected. A break event in the IGK locus leads to the dissociation of 1 colocalization, as shown for the example of a t(2;8)(p12;q24). (b) Metaphase of case number 4 with a der(8)t(2;8)(p12;q24) showing a fusion of the dual-color probe on the chromosome 2 with an intact IGK locus, an isolated red signal from the telomeric probe on derivative chromosome 8 and a single green signal from the centromeric probe on a marker chromosome. (c) Interphase FISH in case number 3. All 3 nuclei show a dissociation of 1 red and 1 green signal (arrows), indicating a breakpoint in the IGK locus and a fusion signal from the intact IGK locus. (d–f) IG lambda. (d) Schematic representation of the IGL FISH assay displaying the expected signal patterns in metaphase and interphase cells lacking and containing a breakpoint in the IGL locus. In cells with an intact IGL locus, 2 colocalizations of the signals for the centromeric and telomeric probes would be expected. A break event in the IGL locus leads to the dissociation of 1 colocalization, as shown for the example of a t(8;22)(q24;q11). (e) Metaphase of case number 16 with a t(8;22)(q24;q11) showing a fusion of the dual-color probe on chromosome 22 with an intact IGL locus, an isolated red signal from the centromeric probe on the derivative chromosome 22 and a single green signal from the telomeric probe on derivative chromosome 8. (f) Interphase FISH in the same case. The 2 nuclei show a dissociation of 1 red and 1 green signal (arrows), indicating a breakpoint in the IGL locus and a fusion signal from the intact IGL locus.

Clones flanking the IGL locus at 22q11 were obtained from the complete chromosome 22 sequence. As the centromeric probe, BAC CTA 526G4 (labeled in red), located immediately proximal to the IGL variable region, was selected. As the telomeric probe, the pooled overlapping BACs CTA 60B5 and CTA 865E9 (labeled in green), which are located distal to the IGL constant region, were used. (Fig. 1d).

Cutoff level

At least 200 nuclei from each of 5 controls were evaluated for the distance between signals derived from the differentially labeled flanking probes. For the IGK “break apart” probe, mean and SD (in parentheses) for nonoverlapping signals and for the signal distances of more than once and more than twice the signal diameter in these controls were 5.5% (1.3%), 1.3% (0.4%) and 0.04% (0.09%), respectively. Accordingly, the cutoff levels (mean + 3SD) were 9.4%, 2.5% and 0.3%, respectively. No split was observed if a signal distance of more than 3 signal diameters was considered. Based on these controls, a signal split was defined for the IGK FISH assay as a distance between the signals derived from the centromeric and telomeric probe of more than twice the estimated signal diameter. Based on this restrictive criterion, the cut-off level was 0.3%.

For the IGL “break apart” probe, means and SDs for nonoverlaping signals and for the signal distances of more than once, twice, 3 times and 4 times the signal diameter were 40.2% (3.8%), 10.5% (1.4%), 2.4% (0.8%), 0.5% (0.3%) and 0.2% (0.2%), respectively. Therefore, the cutoff levels calculated were 51.6%, 14.8%, 4.9%, 1.4% and 0.8%, respectively. A split was thus defined for the IGL FISH assay as a distance between the signals of the flanking probes of more than 3 times the estimated signal diameter, with the cutoff level being 1.4%.

Interphase FISH in B-cell neoplasias with breakpoints at 2p11-14 and 22q11-12

To investigate the diagnostic validity of our new IGK and IGL FISH assays, we studied 23 primary B-cell neoplasms with a total of 24 breakpoints affecting chromosomal bands 2p11-14 (n = 15) or 22q11-12 (n = 9). Patient number 15 had a t(2;22)(p13-14;q11) and therefore was studied with both assays. Additionally, the Burkitt′s lymphoma cell line BL104 with a t(8;22)(q24;q11) was investigated. The evaluation criteria defined in the control studies described above were applied. It has to be emphasized that, in addition to typical splits in part of the tumors as well as in tumor subclones, variant signal constellations can occur in interphase cells that include the absence or extra copies of red or green signals due to loss of derivative chromosomes or presence of extra derivative chromosomes, respectively. Table I summarizes the partial karyotypes and results of our FISH experiments with regard to the total percentage of interphase nuclei with aberrant signal constellations.

Table I. Lymphoid Malignancies with Cytogenetically Detectable Breaks at 2p11-14 or 22q11-12 Studied by Interphase FISH for IGK and IGL Loci1

Case no.	Diagnosis	Age (yr)	Gender	Sample	Partial karyotypes with 2p11-14 and 22q11-12 karyotypic abnormalities	Interphase FISH results
						% Aberrant	Oncogene
1 B-PLL, B-cell prolymphocytic leukemia; MCL, mantle cell lymphoma; FL, follicular lymphoma; MALT, mucosa-associated lymphoid tissue lymphoma; DLBCL, diffuse large B-cell lymphoma; AILD, angioimmunoblastic lymphadenopathy; Pre-B ALL, precursor B-cell acute lymphoblastic leukemia; B-CLL, B-cell chronic lymphocytic leukemia; BL, Burkitt's lymphoma/leukemia; MM, multiple myeloma; BM, bone marrow; LN, lymph node; PB, peripheral blood. 2 Confirmed by FISH. 3 No evidence for cryptic MYC involvement.
IG kappa (2p12)
1	B-PLL	63	F	BM	t(2;8)(p12;q24)	89	MYC2
2	MCL	67	M	LN	t(2;8)(p12;q24)	61	MYC2
3	FL	54	F	PB	t(2;8)(p12;q24)	89	MYC2
4	DLBCL	29	F	BM	der(8)t(2;8)(p12;q24)	75	MYC2
5	FL	59	M	LN	t(2;18)(p12;q21)	75	BCL22
6	MALT	80	M	LN	t(2;7)(p12;q21)	69	CDK6?
7	FL	50	M	LN	t(2;16)(p11;q24)	79	MAF?
8	AILD	62	F	LN	t(2;4)(p14;q21)	0	—
9	Pre-B ALL	18	M	BM	t(2;5)(p14;q12)	0	—
10	B-CLL	64	M	BM	t(2;5)(p13;q25)	0	—
11	Pre-B ALL	7	F	BM	t(2;6)(p14;q24)	0	—
12	DLBCL	74	F	LN	t(2;11)(p14;q23)	0	—
13	FL	64	M	LN	t(2;18)(p13;q22)	0	—
14	Pre-B ALL	23	M	BM	t(2;19)(p13;p13)	0	—
15	FL	56	M	LN	t(2;22)(p13-14;q11)	0	—
IG lambda (22q11)
16	BL	BL104	Cell line	t(8;22)(q24;q11)	78	MYC2
17	BL	51	M	BM	t(8;22)(q24;q11)	7	MYC2
18	FL	54	F	LN	t(3;22)(q27;q11)	30	BCL62
15	FL	56	M	LN	t(2;22)(p13-14;q11)	90	REL/BCL11A2
19	DLBCL	55	M	LN	t(4;22)(q13;q11)	41	Unknown
20	BL	73	F	LN	t(16;22)(p12;q11)	85	Unknown3
21	MM	90	F	BM	der(22)(q11)	42	Unknown
22	B-CLL	85	F	BM	der(22)(q11)	1	—
23	FL	79	F	LN	t(6;22)(q12;q12)	1	—
24	DLBCL	49	F	BM	t(19;22)(q13;q11)	1	—

Translocations affecting 2p11-14.

A total of 15 cases with rearrangements at 2p11-14 were studied with our IGK break-apart assay. In 7 cases (Nos. 1–7), IGK translocations were detected (Fig. 1a–c). These included all 4 cases (Nos. 1–4) with a variant Burkitt's translocation t(2;8) as well as a case (No. 5) with a variant BCL2-associated translocation t(2;18). Further FISH studies confirmed involvement of the MYC gene in 8q24 and of the BCL2 gene at 18q21 in these cases, respectively. In cases No. 6 and 7, bearing t(2;7)(p12;q21) and t(2;16)(p11;q24), respectively, the oncogenes CDK6 (7q21) and MAF (16q23) might be the translocation partners. The percentage of interphase nuclei with a signal constellation indicating a break in the IGK locus in the positive cases ranged from 61 to 89% and, thus, exceeded by far the cutoff level of 0.3%. A breakpoint in the IGK locus was ruled out by FISH in the remaining 8 cases (Nos. 8–15). Remarkably, according to the results of the cytogenetic analysis, none of these cases contained a translocation previously reported to affect the IGK locus recurrently. Moreover, the 2p breaks in these cases were located cytogenetically to 2p13-14 rather than to 2p11-12.

Translocations affecting 22q11-12.

A total of 10 samples, including the cell line BL104, were studied with our break-apart IGL probe. Seven cases (Nos. 15–21) displayed translocations involving IGL (Fig. 1d–f). Among those cases, there were both variant Burkitt′s translocations t(8;22) (Nos. 16 and 17) and a t(3;22) (No. 18), in which MYC and BCL6 involvement, respectively, was confirmed by further FISH studies. In case No. 15 with IGL-involvement (confirmed by FISH), additional studies revealed that the REL/BCL11A locus in 2p13 was affected, rendering these genes candidate oncogenes deregulated by the IGL translocation. Two cases had translocations affecting chromosomal regions 4q13 (No. 19) and 16p12 (No. 20) from which no B-cell associated oncogenes have been cloned so far. As patient No. 20 had a Burkitt′s lymphoma, a cryptic MYC translocation was ruled out by FISH. Patient No. 21 also displayed IGL involvement, but the partner chromosome could not be identified due to the poor spreading of the metaphases. The percentage of interphase nuclei containing aberrant signal patterns ranged from 7 to 90% compared with the cutoff level of 1.4%. Breaks affecting the IGL locus were ruled out in the other 3 cases (Nos. 22–24).

DISCUSSION

The accurate detection of IG illegitimate rearrangements is important for the diagnostic process and the clinical management of patients with B-cell malignancies; it is also a useful tool for discovering new genes associated with these neoplasms.1 The most simple, reliable and flexible method to detect all translocations affecting IG loci is a FISH assay using probes flanking the recurrent breakpoints. The establishment of such FISH assays takes advantage of the fact that most translocations, regardless of the translocation partner, involve clustered breakpoints within the IG loci; most translocations involve the joining or the switch regions of the IGH or the joining region of the IG light chain loci.21

So far, no reliable method has been described for the detection of IG light chain translocations. Consequently, we designed 2 novel FISH assays to detect translocations of IGK and IGL located at 2p12 and 22q11, respectively. Both regions bear characteristics making the establishment of a classical break-apart FISH assay difficult. Due to its genomic size of around 1 Mb (from 5.9 Mb to 6.9 Mb from the centromere of chromosome 22 according to http://www.sanger.ac.uk/HGP/Chr22), probes flanking the IGL locus have to be considerably far away from each other. Therefore, based on extensive controls, a split was not considered unless the distance between the differentially labeled signals exceeded 3 times the signal diameter. Applying this strict criterion resulted in an acceptable cutoff level of 1.4%.

Because of this criterion, the false-negative rate might be relatively high, as shown by the results obtained from the cell line BL104. In this cell line, which should contain 100% of aberrant cells, only 78% of the nuclei met the strict criterion, yielding a false-negative rate in this cell line of 22%. Remarkably, cases 7 and 9 displayed 85 and 90% of aberrant cells, respectively and thus a lower false-negative rate. This might be explained by the different size of the nuclei in these cases compared with the cell line.

For the IGK locus, designing a FISH assay was difficult because the IGK variable region is duplicated with some 800 Kb in between.22 The probe covering this region appeared frequently as paired dots in the hybridization experiments. This fact does not prevent the interpretation of hybridization results, but it has to be taken into account to score the experiments correctly. The cutoff level for this IGK break-apart assay was set to 0.3% when the distance between the differentially labeled signals was at least 2 times the signal diameter.

Comparable break-apart FISH assays like those for MLL and BCL2 genes present cutoff levels of around 3%.23, 24 Nevertheless, for most of these assays the normal signal pattern has not been systematically evaluated scoring the distance between dual-color signals in interphase nuclei. As highlighted in our study, to calculate the threshold for signal disruption while maintaining an acceptable false-positive rate, it is important to calculate cutoffs by analyzing the distance between signals. This holds particularly true if both flanking probes are separated by a considerable genomic distance, as might be necessary for the detection of widely scattered breakpoints.

We applied our assay to a total of 24 cases with cytogenetically confirmed breakpoints at 22q11-12 or 2p11-14. To ensure the reliability of the novel probes, variants of known translocations affecting IGH were studied. Eight cases displayed variants of the classical t(8;14)(q24;q32), t(14;18)(q32;q21) and t(3;14)(q27;q32) translocations, which were further confirmed by FISH using probes for MYC, BCL2 and BCL6, respectively. In all these cases, IG light chain breakpoints were detected by our assays confirming their reliability. Two additional cases probably involve the genes CDK6 and MAF, respectively.25, 26

Our study also points to new candidate oncogenes targeted by IG translocations that are located in 2p13, 4q13 and 16p13. A breakpoint in the region where BCL11A and REL are located was identified by FISH in a t(2;22)(p13-14;q11) translocation. Remarkably, both these genes in 2p13 are supposed to play a role in B-cell lymphomagenesis.18, 27 All cases lacking IGK translocations in our study contained breakpoints in 2p13-14 rather than 2p11-12, which might indicate that the BCL11A/REL or other loci distal to IGK might be prone to chromosomal breakage in B-cell neoplasms.

In conclusion, the assays designed in our study were shown to be reliable tools for the detection of both known and unknown IGK- and IGL-related translocations, constituting a flexible and accurate method for the diagnosis and follow-up of patients with B-cell disorders. Furthermore, these assays might help to discover new oncogenes involved in the pathogenesis of B-cell malignancies.

Acknowledgements

Dr. J.I. Martín-Subero is a scholar of the Gobierno de Navarra. The skilled technical assistance of Mrs. C. Becher and Mrs. R. Zühlke-Jenisch is gratefully acknowledged.