Human chromosome telomeres are important for maintaining chromosome structure and function1. The adjacent subtelomeric chromosome regions are transcriptionally active, GC-rich and contain a high density of CpG islands and genes2,
3. In addition, the frequency of genetic recombination increases towards the telomeres4. These features are reflected in the finding that telomeric chromosome rearrangements are common in a number of genetic diseases5. The detection of subtle chromosome rearrangements involving telomeric regions poses a particular problem for diagnostic cytogenetics. Conventional cytogenetic analysis relies on the interpretation of a monochrome G-banding pattern; however, the telomeric chromosome regions are usually G-band negative and small deletions or translocations with other chromosomal telomeric regions may be almost impossible to discriminate. Nevertheless, there is mounting evidence that many apparently normal karyotypes may contain subtle (cryptic) deletions or translocations involving telomeres, particularly in clinical referrals featuring mental retardation6,
7. Similarly, an increasing number of studies report cryptic telomeric chromosome rearrangements in hematological malignancies8,
9,
10.
One approach to the accurate detection of chromosome abnormalities is the use of fluorescence in situ hybridization (FISH) with specific gene probes11. FISH is now a successful adjunct to conventional cytogenetic diagnosis, but requires a pre-existing knowledge of the region likely to be affected. Whole-genome screening is provided by the 24-color karyotyping methods multiplex FISH (M-FISH)12 and spectral karyotyping (SKY)13. However, the resolution of these whole-chromosome−painting techniques for subtle chromosome rearrangements is limited14,
15. Similarly, comparative genomic hybridization (CGH) to metaphase chromosomes can only detect deletions of 10 Mb, based on approximately 80% of cells carrying the deletion16.
The development of an accurate and sensitive method to screen all telomeric regions has important ramifications for the clinical diagnosis of both constitutional and acquired telomeric abnormalities. The first step in the development of such an assay was provided by the isolation and characterization of a set of 41 cosmid, P1 and PAC clones specific for all human subtelomeric chromosome regions17. We describe here the development of a 12-color FISH assay for telomeric rearrangements (termed M-TEL) using an updated set of telomeric probes18, which permits the screening of all telomeres in only two hybridizations. An integral part of the assay was the application of an innovative software program to analyze the telomeric signals. We have validated the M-TEL assay in a blind positive control study using constitutional karyotypes with known telomeric rearrangements, the majority of which were not visible by G-banding. We have also demonstrated the applicability of the M-TEL assay to leukemic karyotypes.
Development of the M-TEL assay To develop the M-TEL assay, we used a definitive `second-generation' set of subtelomeric probes cloned in PACs or BACs, all except 5p confirmed as within 500 kb of the telomere18 (Table 1). We used the `combinatorial' labeling approach19 and four spectrally distinct fluorochromes to generate 12 different colors (Fig. 1). This strategy included detecting both the short (p) and long (q) arm probes for each chromosome in the same color. Although this has the potential limitation of being unable to detect some pericentric inversions, this consideration was far outweighed by the benefits in probe labeling and image analysis. One of the most important factors in the successful development of a robust set of labeled probes for the M-TEL assay was the construction of DNA `pools' for labeling. For this, we combined all probes requiring labeling with a particular fluorochrome into a single nick-translation reaction. This not only reduced the number of nick-translation reactions (from more than 70 to 8), but also provided a standardized source of labeled probes, ensuring reproducible hybridization experiments.
Figure 1. Combinatorial labeling scheme for the M-TEL assay.
For each chromosome the p and q arm subtelomeric probes were labeled with a unique combination of fluorochromes. Only Cy3 was incorporated directly in the labeling reaction (as Cy3-dUTP). Cy3.5, FITC and Cy5.5 were used as avidin, anti-digoxigenin and anti-estradiol conjugates for secondary detection of biotin (bio), digoxigenin (dig) and estradiol (est), respectively. The two sets of probes required for a full survey of all telomeric regions are termed M-TEL1 and M-TEL2, respectively. `Pseudocolor' refers to the false color assigned to the telomeric signals using the goldFISH analysis approach.
Another important parameter for successful M-TEL hybridizations was a requirement for well spread metaphases (with no visible cytoplasm), in order to ensure high hybridization efficiencies. We found the criteria established for assessing metaphases for CGH (ref. 20) were also useful in this respect. The hybridization efficiency (defined as the percentage of cells with fluorescent signal on both sister chromatids of both chromosome homologs) for the majority of probes when hybridized individually was 98%. In practice, we found that the majority of metaphases ( 8/10 analyzed) had all telomeric signals correctly hybridized, assessed by viewing the individual fluorochrome channels in karyotype format.
M-TEL image analysis We used two software programs to analyze the telomeric probe signals. First, we used a new analysis system, termed goldFISH, which performs classification of telomeric signals by anisotropic, non-linear diffusion filtering21 (Fig. 2). Second, we used a modified version of a commercially available software package, MacProbe v4.1 (Applied Imaging, Newcastle, UK), to view the individual fluorochrome channels in a karyotype format (Fig. 3). This was important to resolve any anomalies in the color classification of individual probes (Fig. 2).
Figure 2. Color classification (goldFISH) analysis of the telomeric signals in metaphases from patient 3.
a, The M-TEL1 probes identified a deletion of 7q on one chromosome 7 homolog (arrow). b, The M-TEL2 probes identified an additional chromosome 2 signal (orange) on the derivative chromosome 7 (arrow). One further FISH experiment with 2p and 2q subtelomeric probes was sufficient to confirm an unbalanced translocation resulting in monosomy for 7q and trisomy for 2q in this patient. The respective pseudocolors are given on the left. In both a and b, there are some apparent misclassifications, for example 5p and q, 11p, 16q, 19q. However, in all cases the misclassification involved only one out of a possible two sister chromatids on one chromosome arm. Note also that the 9q and XqYq PACs were not included in this early version of the M-TEL assay, because of cross-hybridization with several other telomeres18. Instead, the first-generation cosmids were applied separately in a dual-color hybridization. However, the newer BAC probes for 9q and XqYq do not cross-hybridize and were applied in the more recent M-TEL 1 hybridizations (see Table 1 and Fig. 4).
Figure 3. MacProbe v4.1 analysis of telomeric signals in metaphases from patient 3.
Metaphases were karyotyped from the inverted DAPI images and the individual fluorochrome channels were assessed for the presence or absence of telomeric signals. a−c Individual fluorochrome images of chromosomes 7 hybridized with M-TEL1 (a and b) and M-TEL2 probes (c). The M-TEL1 probes showed a missing signal from 7q in both the Cy3 (a) and Cy3.5 channels (b) from the same metaphase, confirming a deletion of chromosome 7. The M-TEL2 probes showed an additional signal on 7q in the FITC channel (c), confirming trisomy for chromosome 2.
Validation of the M-TEL assay First, we applied the M-TEL assay to metaphases from at least five normal individuals and obtained no false-positive results. Second, to test the efficacy of the M-TEL assay for clinical diagnostic cases, we performed a positive-control study using eight patients with known constitutional rearrangements (Table 2). The karyotypes we chose included two telomeric deletions, five unbalanced translocations and one balanced translocation. Slides were coded and the analysis was done by an investigator who had no knowledge of the karyotype details. In each case, the M-TEL assay correctly identified the abnormality. For the unbalanced translocations, however, it was necessary to perform an additional dual-color hybridization to determine whether the p or the q arm was involved. This is similar to M-FISH or SKY, which both require additional experiments to give regional information. Finally, to validate the assay for leukemia and other malignancies, we applied the M-TEL assay to a series of leukemic karyotypes (Table 2, patients 9−11). In all cases, the abnormalities were accurately identified by the M-TEL assay (Fig. 4).
Figure 4. Color classification (goldFISH) analysis of bone-marrow metaphases with trisomy 8 from an AML patient.
a, In the M-TEL1 hybridization, all chromosome telomeres were correctly classified. One chromosome 21 homolog is missing, however analysis of 10 further cells showed that this was not a clonal abnormality. The XpYp probe has failed to hybridize to Xp, although the Yp signal is present. However, this finding was not confirmed in the other metaphases analyzed. b, In the M-TEL2 hybridization, 3 copies of chromosome 8 were identified in all cells analyzed (arrows). In this metaphase, all telomeres except chromosome 22 were correctly classified. However, one homolog of chromosome 22 was entirely misclassified as chromosome 8. On reviewing the individual fluorochromes, this was found to be due to a weak signal in the Cy3.5 channel.
Utility of the M-TEL assay There are currently only two methods available for telomere screening. The first uses a set of DNA markers close to the telomere to investigate the inheritance of polymorphic alleles22. This has the advantage of being able to detect isodisomy (the inheritance of two chromosome homologs from the same parent). However, the disadvantages are that it requires the analysis of both parents and relies on a high level of marker informativeness. The second is a commercially available FISH-based technique, the Multiprobe T assay, which uses 23 hybridizations on a single microscope slide23. However, the Multiprobe T assay requires a high mitotic index and is therefore not applicable in cases where the clinical material is limited. In addition, because only a small number of cells can be analyzed for each telomere, the Multiprobe T assay is not suitable for the analysis of the majority of leukemic karyotypes, where there is likely to be a mixture of normal and abnormal cells. Overall, the M-TEL assay is more widely applicable than either of the aforementioned assays. The ability to survey all chromosome telomeres in only two hybridizations is an important advantage for many clinical diagnostic situations; for example, in situations where the mitotic index is low or there are clonal populations of abnormal cells (such as leukemic bone marrow) or where the sample is small (such as blood samples from infants). We have recently demonstrated the usefulness of the M-TEL assay in one such case of an infant with multiple congenital abnormalities15.
The application of CGH to an array of mapped DNA or cDNA sequences immobilized on a glass slide (microarray) holds great promise for the sensitive screening of whole genomes for regions of imbalance24,
25. The reported sensitivity of CGH to DNA microarrays is currently approximately 40 kb for the detection of deletions24. However, DNA microarrays cannot detect balanced translocations, therefore the M-TEL assay provides a significant advantage in this respect. The identification of balanced translocations is important for a number of applications, including the investigation of couples experiencing multiple miscarriage and the identification of new non-random translocations in hematological malignancies.
Cytogenetic analysis provides important diagnostic and prognostic information in many types of leukemia. In acute myeloid leukemia, the incidence of apparently normal karyotypes is as high as 42% (ref. 26). The identification of hidden translocations in this group may have important implications for the stratification of patients into appropriate treatment protocols. There is compelling evidence for the existence of cryptic translocations in apparently normal leukemic karyotypes, notably the t(12;21)(p13;q22) in childhood B-cell acute lymphoblastic leukaemia8, as well as the t(5;11)(q35;p13)9 and t(7;12)(q36;p13)10 in childhood leukemia. We have verified that the M-TEL assay works equally well for leukemic as for constitutional metaphases. Indeed, in contrast to M-FISH, the use of contracted chromosomes (such as are often obtained from leukemic bone marrow) does not compromise the sensitivity of the M-TEL assay. The assay can also be successfully applied to archival metaphase preparations on slides stored at -20 °C for up to 18 months, and on archival fixed chromosome suspensions that are greater than 5 years old.
Subtelomeric probe screening is expected to have a profound impact on the diagnosis and genetic counseling of families with several members affected by mental retardation. In addition, the detailed analysis of these rearrangements is expected to identify dosage-sensitive genes involved in genetic disease, as well as providing insight into the structure and function of telomeric regions. In hematological malignancies, the identification of new, non-random chromosome rearrangements is expected to identify new regions and genes involved in the leukemogenic process. It is expected that the M-TEL assay will contribute significantly to all of these areas.
Methods Nick translation. All probes requiring labeling with a particular fluorochrome or hapten were combined in the same nick translation reaction. Two sets of digoxigenin, biotin, estradiol, and Cy3-DNA pools were prepared (Fig. 1). The amount of each probe in the respective pools was adjusted to ensure approximately the same fluorescent intensity after FISH. In most cases, 1 g of each probe DNA was used. The exceptions were: Bio Pool 1: 13q (3 g), 21q (0.5 g), XpYp (4 g); Dig pool 1: 13q (3 g), 17p (2 g), 17q (2 g), 21q (0.5 g); Dig pool 2: 2q (2 g) 4q (1.5 g), 8p (1.5 g), 8q (1.5 g); Est pool 1: 9p (2 g), 15q (3 g); Est pool 2: 8p (1.5 g), 8q (1.5 g), 18p (4 g), 22q (2 g); Cy3 pool 1: 1p (0.5 g), 7q (2 g), 13q (3 g); Cy3 pool 2: 12q (2 g). Before nick translation, the DNA was treated with RNase (0.02 U per 1 g of DNA) (Sigma), for 30 min at 37 °C. Nick translation was carried out at 16 °C for 2 h using the following reaction mixture (for each 1 g of DNA): 50 mM Tris-HCl, 5 mM MgCl2, 2.5 g BSA, 10 mM -mercaptoethanol, 50 mM dAGC, 20 M hapten/fluor [estradiol 15-dUTP, biotin-16-dUTP, digoxigenin-11-dUTP (all from Roche); Cy3 dUTP (Amersham)], 0.04 U DNase and 10 U DNA polymerase I, made up to a final volume of 50 l with H20.
M-TEL hybridization. Two separate hybridization mixtures were prepared, M-TEL1 and M-TEL2 (Fig. 1). The appropriate volume of each labeled pool (to contain 100 ng of each probe) was ethanol precipitated with 3 g placental DNA (per 100 ng of probe) and 2.2 g salmon sperm DNA (per 100 ng of probe). After precipitation, the pellet was resuspended in hybridization mix containing 50% formamide, 15% dextran sulphate and 2 SSC. Before hybridization, metaphase chromosomes on slides were treated with 100 g/ml RNase for 1 h at 37 °C and 30 g/ml pepsin (in 0.01 N HCl) for 2−5 min at 37 °C, depending on the amount of cytoplasm present. Postfixation was carried out using 1% formaldehyde (in 50 mM PBS/MgCl2 solution) for 10 min at room temperature. The M-TEL probe mixtures were denatured at 75 °C for 8 min, then transferred to 37 °C for at least 1 h before hybridization. Metaphase chromosomes on slides were denatured in 70% formamide in 2 SSC, pH 7.0 at 70 °C for 5 min. Hybridization was carried out in a humid chamber for 1−2 days at 37 °C. After hybridization, slides were washed 3 times (5 min each) in 50% formamide, 2 SSC (pH 7.0), followed by one 5-min wash each in 2 SSC and 4 SSC with 0.05% Tween (SSCT) (all at 50 °C). After a blocking step in 3% BSA in SSCT at 37 °C for 1 h, the following detection steps were applied: 1) Layer 1: avidin Cy3.5 (3.3 g/ml, Amersham Pharmacia Biotech), rabbit anti-estradiol (1 g/ml, Roche), sheep anti-dig FITC (2 g/ml, Roche); 2) Layer 2: donkey anti rabbit Cy5.5 (2 g/ml, Amersham Pharmacia Biotech), donkey anti-sheep FITC (11 g/ml, Sigma) (both at 37 °C for 20 min). All antibody solutions were made up in 1% BSA in SSCT. Finally, slides were washed twice in PBS (5 min each) at room temperature, dehydrated through an ethanol series, then mounted in Vectashield (Vector Laboratories, Peterborough, UK) containing 1 g/ml DAPI.
Image capture and analysis. Images were captured using a Sensys CCD camera with a Kodak KAF 1400 chip (Photometrics, Tucson, Arizona) mounted on an Olympus AX70 epifluorescence microscope, equipped with an 8-position filter turret with individual filter blocks specificf or DAPI, FITC, Cy3, Cy3.5, Cy5, and Cy5.5 (Chroma Technology, Brattleboro, Vermont). A 75-W Xenon lamp was used to capture all of the fluorochrome channels.
goldFISH. The individual fluorochrome channels from the same metaphase images used for step 2 (above) were subjected to color classification analysis. After identification of the telomeric signals by anisotropic diffusion filtering and segmentation, each signal was classified according to its average color composition. We set the minimum requirement for accurate classification21 as one out of two sister chromatids with the correct classification. Misclassifications, such as those in Fig. 2, would therefore not cause any diagnostic errors.
MacProbe v 4.1. For each patient, 10−20 metaphases were captured, and at least 10 of these were fully analyzed. Each metaphase was karyotyped using the inverted DAPI image. After karyotyping, each fluorochrome channel was viewed individually and the signal for each chromosome recorded manually. This was important to resolve any anomalies in color classification. In the majority of cases, the misclassified signal was shown to be due to a weak signal for one of the constituent fluorochromes, often due to the signals from different fluorochromes being in a different focal plane. Weak signals such as this would be eliminated by the thresholding step that precedes the color classification21.
Zakian, V.A. Telomeres: Beginning to understand the end. Science270, 1601–1607 (1995). | PubMed | ISI | ChemPort |
Craig, J.M. & Bickmore, W.A. The distribution of CpG islands in mammalian chromosomes. Nature Genet.7, 376–382 (1994). | Article | PubMed | ISI | ChemPort |
Saccone, S., De Sario, A., Valle, D.G. & Bernardi, G. The highest gene concentration in the human genome are in telomeric bands of metaphase chromosome. Proc. Natl. Acad. Sci. USA89, 4913–4917 (1992). | PubMed | ChemPort |
Donis-Keller, H. et al. A genetic linkage of the human genome. Cell51, 319–337 (1987). | Article | PubMed | ChemPort |
Raynham, H., Gibbons, R., Flint, J. & Higgs, D. The genetic basis for mental retardation. Quart. J. Med.89, 169–175 (1996). | ChemPort |
Knight, S.J.L. et al. Subtle chromosomal rearrangements in children with unexplained mental retardation. Lancet354, 1676–1681 (1999). | Article | PubMed | ISI | ChemPort |
Holinski-Feder, E. et al. Familial mental retardation syndrome ATR-16 due to an inherited cryptic subtelomeric translocation, t(3;16)(q29;p13.3). Am. J. Hum. Genet.66, 16–25 (2000). | Article | PubMed | ISI | ChemPort |
Romana, S.P., Le Coniat, M. & Berger, R. t(12;21): A new recurrent translocation in acute lymphoblastic leukemia. Genes Chromosom. Cancer9, 186–191 (1994). | PubMed | ISI | ChemPort |
Jaju, R.J. et al. A new recurrent translocation, t(5;11)(q35;p15.5), associated with del(5q) in childhood AML. Blood94, 773–780 (1999). | PubMed | ISI | ChemPort |
Tosi, S. et al. t(7;12)(q36;p13), a new recurrent translocation involving ETV6 in infant leukaemia. Genes Chromosom. Cancer29, 325–332 (2000). | Article | PubMed | ISI | ChemPort |
Buckle, V.J. & Kearney, L. New methods in cytogenetics. Curr. Opin. Genet. Dev.4, 374–382 (1994). | Article | PubMed | ChemPort |
Schröck, E. et al. Multicolor spectral karyotyping of human chromosomes. Science273, 494–497 (1996). | PubMed |
Uhrig, S. et al. Multiplex-FISH for pre- and postnatal diagnostic applications. Am. J. Hum. Genet.65, 448–462 (1999). | Article | PubMed | ISI | ChemPort |
Brown, J. et al. The identification of a subtle t(16;19)(p13.3;p13.3) in an infant with multiple congenital abnormalities using a 12-colour multiplex FISH assay, M-TEL. Eur. J. Hum. Genet.8, 903–910 (2000). | Article | PubMed | ISI | ChemPort |
Bentz, M., Plesch, A. Stilgenbauer, S. Döbner H., Lichter, P. Minimal sizes of deletions detected by comparative genomic hybridization. Genes Chromosom. Cancer21, 172–175 (1998). | Article | PubMed | ISI | ChemPort |
National Institutes of Health and Institute of Molecular Medicine Collaboration. A complete set of human telomeric probes and their clinical application. Nature Genet.13, 86–89 (1996).
Knight, S.J.L. et al. An optimized set of human telomere clones for studying telomere integrity and architecture. Am. J. Hum. Genet.67, 320–332 (2000). | Article | PubMed | ISI | ChemPort |
Nederlof, P.M. et al. Multiple fluorescence in situ hybridization. Cytometry11, 126–131 (1990). | PubMed | ISI | ChemPort |
Karhu, R. et al. Quality control of CGH: Impact of metaphase chromosomes and the dynamic range of hybridization. Cytometry28, 198–205 (1997). | Article | PubMed | ISI | ChemPort |
Saracoglu, K. et al. New concepts to improve resolution and sensitivity of molecular cytogenetic diagnostics by multicolor FISH. Cytometry (in the press).
Knight, S.J.L. et al. Development and clinical application of an innovative fluorescence in situ hybridization technique which detects submicroscopic rearrangements involving telomeres. Eur. J. Hum. Genet.5, 1–8 (1997). | PubMed | ISI | ChemPort |
Pinkel, D. et al. High resolution analysis of DNA copy number variation using comparative genomic hybridization to microarrays. Nature Genet.20, 207–211 (1998). | Article | PubMed | ISI | ChemPort |
Pollack, J.R. et al. Genome-wide analysis of DNA copy-number changes using cDNA microarrays. Nature Genet.23, 41–46 (1999). | Article | PubMed | ISI | ChemPort |
.Grimwade, D. et al. The importance of diagnostic cytogenetics on outcome in AML: Analysis of 1,612 patients entered into the MRC AML 10 trial. Blood92, 2322–2333 (1998). | PubMed | ISI | ChemPort |
Macina, R.A. et al. Sequence organization of the human 2q telomere. Hum. Mol. Genet.3, 1847–1853 (1994). | PubMed | ISI | ChemPort |
Nesslinger, N.J. et al. Clinical, cytogenetic, and molecular characterization of seven patients with deletions of chromosome 22q13.3. Am. J. Hum. Genet.54, 464–472 (1994). | PubMed | ISI | ChemPort |
Horsley, S.W. et al. Del(18p) shown to be a cryptic translocation using a multiprobe FISH assay for subtelomeric chromosome rearrangements. J. Med. Genet.35, 722–726 (1998). | PubMed | ISI | ChemPort |
Acknowledgments We thank J. Flint and S. Knight for supplying the second-generation subtelomeric probes; V. Buckle for critical reading of the manuscript; and M. Schulze for modifications to the MacProbe v4.1 software. This work was supported in part by the Leukaemia Research Fund, UK (to J.B.) and the Medical Research Council (to L.K). R.E. was supported by the German Minister for Education and Research (BMBF, BioFuture grant AZ 11880), the Deutsche Forschungsgemeinschaft (Ja 395/6-2; Ei-358/1-1) and the German-Israeli foundation for research and technology (GIF G-112-207.04/97).
8/10 analyzed) had all telomeric signals correctly hybridized, assessed by viewing the individual fluorochrome channels in karyotype format.
g of each probe DNA was used. The exceptions were: Bio Pool 1: 13q (3 
-mercaptoethanol, 50 mM dAGC, 20 
SSC. Before hybridization, metaphase chromosomes on slides were treated with 100 
