AN 03

Automated spot distance measurements using Metafer: quantitative detection of interphase nuclei with balanced chromosomal translocations

Gábor Méhes

Laboratory of Tumor Cytogenetics, Department of Pathology, University of Pécs, Szigeti út 12, H-7643 Pécs, Hungary
eMail: mehesg@pathology.pote.hu
WWW: www.pte.hu

received: August 2002

Introduction

Molecular cytogenetic analysis in the interphase relies on the statistical evaluation of count and relative topography of fluorescence in situ hybridisation (FISH) signals. Thus, the sensitivity of the method is highly depending on the number of cells analyzed. As the evaluation is rather time consuming, just a few hundred cells are usually examined on a routine basis. The analysis of FISH results is also complicated due to inadequate probe specificity. Random signal colocalization is especially limiting the reliable detection of low frequency cytogenetic events.

Tumor cells can be specifically identified by targeting typical (characteristic) molecular cytogenetic aberrations. To improve their diagnostic value, numerous DNA-probe combinations have been introduced to interphase cytogenetics. Their evaluation is based on combinations of spot counts, split signals and single or multiple signal colocalizations (1, 2, 3). However, statistical evaluations become much more difficult with increasing complexity of probe sets. The aim of our present work is the elaboration of an automated image processing search protocol for the quantitative detection of cells with balanced chromosomal translocations by accurate multicolor spot distance measurements.

As a model system, follicular lymphomas were analyzed for the occurrance of the balanced translocation t(14;18)(q32;q21) which was reported to be present in approx. 70% of these lymphomas (4, 5). Quite often it is difficult to differentiate this entity from reactive follicular hyperplasias or other indolent lymphomas. A quantitative detection of cells with the characteristic translocation would be of significant practical advantage for diagnosis and clinical follow up.

Figure 1.
FISH spot distribution following the hybridization with the two-color IgH/BCL2 probe in normal and translocated lymphatic cells. Isolated two red and two green signals according to intact chromosomal loci (upper part). Double red and green signal   colocalisation due to balanced t(14;18) (lower part). Red and green spot distance ranking is displayed on the corresponding drawing. As a result of a balanced chromosomal translocation, spot distances for the fused signals get significantly shorter.

Materials and Methods

Tissue imprints from follicular lymphomas and lymph nodes with follicular hyperplasia were analysed for the presence of t(14;18) by using a dual color IgH/BCL2 FISH probe mix (Vysis Inc.). The major criteria for the presence of the translocation were colocalizations of the red and green hybridization signals on both involved alleles (Figure 1). The range of spot distances in case of dual fusions was demonstrated by a break apart probe on a separate area of the same slide (C-MYC, Vysis Inc.) (Figure 2).

For the analysis of a large cell number the automatic image analysis system Metafer-MetaCyte (MetaSystems, Germany) was used. The system was mounted to an Axioplan 2 motorized microscope (Carl Zeiss, Germany), and included a high resolution CCD camera, an 8 slide motorized scanning stage (Maerzhaeuser, Germany) and a microcomputer system (Pentium IV, 1.8 GHz, Windows 2000 operating system). A predefined area on a slide was scanned without gaps, isolated cells (nuclei) were identified and images captured in three colors. As signals are randomly distributed across the 3-dimensional nuclei, extended focus images calculated from focus stacks were used for FISH spot analysis. For at least 1000 cells per sample Metafer-MetaCyte automatically detected the FISH signals and measured 3-dimensional distances between spots of different colors. The two shortest distances between spots of different colors are indicative for the presence or absence of a double fusion. As a short second smallest distance implies an even shorter smallest distance, it is sufficient to check for the second smallest distance as a criterion for a double fusion (Figure 1+2).

Figure 2.
Graphical display of the distribution  of the break-apart probe for C-MYC hybridized to lymphoma tissue imprints. As C-MYC is not involved in translocations the signals are not  expected to be separated. The distance is displayed in pixel. The   shortest distance ranged from 1-12 pixels, the 2nd shortest from 1-18  pixels.

Results

The values of spot distances measured in non-neoplastic lymph node samples hybridized with the IgH/BCL2 probe (Figure 3A) were comparable to values from a control slide hybridized with probes of similar size (data not shown). In both cases purely random colocalisations give rise to a sub-population of cells showing a low value of the smallest red-green spot distance. In follicular lymphomas with t(14;18) translocation the two shortest spot distances were significantly smaller as a result of the translocation (Figure 3B, 3C). Whereas the shortest spot distance reflects random signal colocalizations as well as true fusions, a low value of the second shortest spot distance was clearly specific for translocated cells. Differences between histograms of control and translocated cell populations suggested a distinctive distance value for translocated cells. The break apart c-myc probe showed the distribution range for fused signals (1-12 pixels for the first and 1-18 pixels for the second shortest spot distance, Figure 2). Considering a safety zone between 18 and 12 pixel we concluded, that cells with less than 12 pixel second shortest spot distance can be all interpreted as translocated cells (Figure 2+3).

Figure 3.
Demonstration of t(14;18) in tissue imprints by automatic image analysis of IgH/BCL2 FISH spots.  Graphical  display of the 2nd shortest spot distances in non-malignant follicular hyperplasia (A), and follicular
lymphoma (B,C). The quantity of lymphoma cells with the IgH/BCL2 fusion was determined as 2.2% in B and as 67.6% in C. The spot distance is displayed in pixel.

Conclusion

FISH signal count can be influenced by technical factors (proteolytic and stringency variables) as well as by biological factors (e.g. chromosomal breaks, losses and gains). The detection of balanced translocations in interphase nuclei can be specifically carried out by reciprocal signal fusions on both alleles involved (Figure 1). The probability of a random signal fusion in the interphase nucleus is generally at 3% (unpublished data). In conclusion, the theoretical probability of random double signal fusions in the absence of a translocation is at 0.09% (3%x3%). This would reflect a higher sensitivity for a given translocation than many currently used PCR-based molecular approaches. However, to reach this level the evaluation of high cell numbers is necessitated. High throughput automatic measurement of the spot distance distribution using MetaSystems’ Metafer-MetaCyte significantly improves the power of FISH. The described approach enables a sensitive statistical evaluation of large cell numbers and is clinically applicable for the quantitative demonstration of cytogenetically aberrant small cell populations.

References

König M, Reichel M, Marschalek R, Haas OA, Strehl S (2002)
A highly specific and sensitive fluorescence in situ hybridization assay for the detection of t(4;11)(q21;q23) and concurrent submicroscopic deletions in acute leukaemias.

Br J Haematol 116, 758-764.

Frater JL, Tsiftsakis EK, Hsi ED, Pettay J, Tubbs RR (2001)
Use of novel t(11;14) and t(14;18) dual-fusion fluorescence in situ hybridization probes in the differential diagnosis of lymphomas of small lymphocytes.

Diagn Mol Pathol 10, 214-222.

Mancini M, Nanni M, Sirleto P et al. (2001)
Detection of BCR/ABL rearrangements in adult acute lymphoblastic leukemia using a highly sensitive interphase fluorescence in situ hybridization method (D-FISH).

Hematol J 2,54-60.

Tilly H, Rossi A, Stamatoullas A et al. (1994)
Prognostic value of chromosomal abnormalities in follicular lymphoma.

Blood 84, 1043-1049.

Knutsen T (1997)
Cytogenetic mechanisms in the pathogenesis and progression of follicular lymphoma.

Cancer Surv 30, 163-192.