| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
From the Cytogénétique Moléculaire et Oncologie,
*
Unité Mixte de Recherche 147, Centre National de Recherche Scientifique, Institut Curie, Paris; the Département de Radiobiologie et Radiopathologie-Direction de Sciences du Vivant,
Commissariat à l’Energie Atomique, Fontenay-Aux-Roses; the Cytopathologie et Cytométrie Clinique,
Laboratoire de Recherche Correspondant, Commissariat à l’Energie Atomique
#
14, Institut Curie, Paris; the Département d’Oncologie Médicale,
§
Institut Curie, Paris; and the Hôpital Marie Lannelongue,
¶
Le Plessis-Robinson, France
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Several significant recurrent chromosome rearrangements, however, have been identified, such as loss of the short arm of chromosomes 3, 9, and 17.3 The loss of the short arm of chromosome 3 (3p) is the most frequently observed,1, 2, 3, 4 appearing in 70 to 80% of SCLCs and 50% of NSCLCs. This deletion is confirmed by data obtained from molecular biology, ie, allelic losses. The target gene remains unknown, but several possible tumor suppressor genes could be involved.5, 6, 7, 8 Moreover, the 3p loss is an early event in the tumor progression of lung cancer, in that it is also observed in preneoplastic lesions.9, 10 Thus, the loss of 3p leads to an imbalance between the long and the short arms of chromosome 3 within lung tumors.
Recently, we have developed a method to investigate chromosome rearrangements within interphase nuclei using fluorescence in situ hybridization (FISH) and image cytometry.11 Briefly, our method consisted of detecting the imbalance between the long and short chromosome arms. After two-color FISH of chromosome arm paintings for both arms, the imbalance for each nucleus was determined by calculating the fluorescence ratio of the two arms. This method has the advantage of being free of any technical and biological bias encountered during spot counting.12
In the present study, we applied this approach to lung cancer to develop a new marker for the detection of cancer cells. Using image cytometry and chromosome arm paintings, we first analyzed xenografted and short-time cultured lung tumors for which data from classical cytogenetics were available. Pleural effusions were then analyzed to investigate whether rare tumors cells were detectable among heterogeneous cell populations.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Case Description
Fourteen cases of lung carcinomas were analyzed. Eight and six
cases were histologically proven SCLCs and NLSCLCs, respectively. Of
these cases, eight were xenografted, cultured, and analyzed; the
remaining cases were cultured directly. Three samples of pleural
effusions from lung cancer were treated as mentioned above. Control
experiments were performed on lymphocytes from healthy donors.
DNA Probes and Labeling
One bladder cancer cell line OBR containing
del3(p10-pter), kindly provided by Dominique
Bellet (Institut Gustave Roussy, Villejuif, France) was used to prepare
chromosome 3 arm-specific probes. Chromosome suspensions and chromosome
sorting by flow cytometry were performed as described.17
Chromosome DNA was amplified and labeled according to Milan et
al18
using digoxigenin-11-dUTP (Roche Diagnostics, Meylan,
France) for 3q. Commercially available chromosome 3p arm painting
(biotin-labeled from AL Technologies, Arlington, VA) was used.
FISH
In situ hybridization for analysis by image cytometry
was carried out as described,11
with minor modifications.
Briefly, slides were pretreated with RNase (100 µg/ml, Sigma, France,
Saint Quentin Fallavier, France) and pepsin (4 µg/ml, Sigma). Thermal
probe denaturation was performed at 70°C for 10 minutes in classical
hybridization buffer. Amplified probe DNA was about 30 ng per slide.
Target DNA was denaturated at 70°C for 3 minutes in 70% formamide
(Fluka, Neu-Ulm, Germany) /2x SSC. Hybridization was carried
out overnight at 37°C. Rapid posthybridization washes were performed
in 2x SSC at 72°C for 5 minutes.
For fluorescent detection, slides were incubated with anti-digoxigenin FITC antibodies (Roche Diagnostics, Meylan, France) for the 3q arm, whereas mouse anti-biotin antibodies followed by anti-mouse IgG antibodies labeled with Texas Red (Molecular Probes, Leiden, The Netherlands) were applied for 3p arm detection. Slides were counterstained with DAPI (1 µg/ml) and mounted in p-phenylene-diamine.
Systematically, reference slides with normal lymphocytes were hybridized together with tumor slides.
Quantitative FISH Measurements
Quantitation of the hybridization signals was performed using a
Discovery automated image analyzer (Becton Dickinson Image Cytometry
Systems, Leiden, The Netherlands). Detection procedure was same as
described,11
with slight modifications. Nuclei were first
detected by DAPI excitation at 330 nm. Next, the integrated
fluorescence of the hybridization signal was measured separately for
fluorescein isothiocyanate (FITC; 3q arm, green fluorescence,
excitation at 485 nm) and for Texas Red (3p arm, red fluorescence,
excitation at 577 nm). For each experiment, optimal integration time
was determined with the reference slide and kept constant for all tumor
slides. Generally, 500 to 700 nuclei were quantified and one
measurement was done per slide. Artifacts, incorrectly segmented
nuclei, and nuclei containing either no green or no red
fluorescence were excluded interactively. The percentage of
excluded cells could reach up to 20%, depending on the hybridization
efficiency. At least 500 cells with analyzable data were obtained. The
resulting histograms for each fluorochrome were generated by the
Discovery software.
Visual Observations
Visual observations and photography were performed with an
epifluorescence microscope (Leitz, Welmar, Germany) equipped
with a mercury arc lamp and a 100x oil immersion objective.
Determination of the Imbalance between Chromosome 3 Arms
Interphase Nuclei
The signal ratio between 3q and 3p arms was determined first. Two
different methods were used. First, mean integrated fluorescence of
each color was determined by setting the gates at half the height of
the histogram peak, and the Discovery software calculated the mean
values. The level of green fluorescence (3q) divided by that of
red fluorescence (3p) was the 3q/3p ratio. For cases containing two
different cell populations, integrated fluorescence of each color and
for each nucleus was plotted against each other in a scatter plot.
After setting of a region, mean values and the corresponding standard
deviations (SDs) of each color were automatically calculated by the
Discovery software. The SD of each mean fluorescence was taken as an
indicator of the dispersion of the values. Different regions were used,
and mean and SD were determined for each region separately. The mean
green fluorescence (3q) divided by red fluorescence (3p) was the 3q/3p
ratio. Imbalances for interphase nuclei (Ii) were obtained by dividing
the 3q/3p ratio of tumor samples by the 3q/3p ratio of the reference
slide (normal lymphocytes).
Metaphases
When available after short-term cultures, metaphases on the same
slides were also used to determine 3q/3p imbalance as a control for
nuclei measurements.
R-banding was performed together with FISH. For complex karyotypes, the number of arm copies was determined by summing all labeled arm fragments in the genome and by taking the normal arm as a reference. For xenografts, about 10 metaphases could have been analyzed, whereas the amount of analyzable metaphases for tumor samples was lower (about 5). Imbalance detected on metaphases (Im) was calculated by dividing the amount of 3q (green fluorescence) by the amount of 3p copies (red fluorescence).
Background Measurements
To estimate nonspecific background signal, slides without probe
were analyzed. Comparison with normally hybridized slides showed that
the mean background signal was <20% of the specific signal. Thus,
background noise did not significantly interfere with the measurements.
Control Experiments
Metaphases from normal lymphocytes gave two specifically marked 3q
and 3p arms (data not shown). The corresponding nuclei showed two large
spots, each representing the long or short arm. Determination of
imbalances within lymphocyte samples showed differences of up to 10%.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
), indicating an excess gain of 3q over 3p. No excess loss of 3q over
that of 3p was observed in any case. For two cases of pleural
effusions, the samples were split to obtain data after cell culturing
(Figure 1C)
and after direct spreading on the slides. Results on interphase
nuclei were concordant in all cases for both cultured and directly
spread samples. Cytogenetic control data from R-banding and metaphase
FISH were available in 60% of the cases representing the actual
percentage of successfully karyotyped cases. Chromosome rearrangements
were so complex in some cases that no precise imbalance value could
have been determined by R-banding data (cases NSCLC1, NSCLC3, and
SCLC110). The analysis of metaphase FISH showed precise imbalance
values in some cases (SCLC41 and NSCLC3) but was not sufficient in
other cases (SCLC110 and NSCLC1), due to the high complexity of the
chromosome rearrangements. Overall, only complete or partial loss of 3p
and no gain of 3q were detected, leading especially to formation of
derivative chromosomes der(1;3)(q10;p10) and
der(3;17)(q10;q10) (Figure 1A
and 1B)
, except for case SCLC108,
where the formation of an isochromosome i(3q) was observed (Figure 1D)
.
For cases where data on nuclei and metaphases were available, measured
imbalances were similar except for two cases.
|
|
. Compared to the standard method where one single population was
considered (Figure 2A)
, Ii were calculated as if at least two different
subpopulations were present (Figure 2B)
. To separate these populations,
different criteria based on the scatter plot were applied. The
existence of two regions with significantly different imbalance values
based on the visual distinction was observed (Figure 2B)
, and the
corresponding regions were set. To confirm the first rough analysis,
another refinement was necessary. One criterion was applied: the change
of the SDs between the two populations. The limit of each region was
defined when a significant change in the SD of green and red
fluorescence occurred. Hence, three regions were obtained, regions 1
and 2 as shown, representing about 33 to 50% of the whole population,
and region 3, containing less than 10% of the cells of the whole
population. Iis determined for each region showed that region 1, with
the normal cells, gave an Ii = 1 and region 2, with tumor cells,
gave an Ii = 1.5, corresponding to what was obtained by Ims. The
cells having an Ii = 0.8 seemed to represent a population with
loss of 3q. Because this loss has never been observed in lung cancer
and only a few cells are in this region, the Ii can be
considered as representing artifacts generated during the hybridization
procedure.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Different cell populations could be distinguished based on the detection of the imbalance between chromosome arms. However, the distinction must be made in an objective and rigorous manner using computer programs with mathematical models. In our study, because no appropriate program was available, SDs were used as the criterion. This separation of the populations is very useful because many cytological samples are contaminated with mesothelial cells as well as normal blood cells, which may be difficult to distinguish morphologically. Because the actual amount of tumor cells within the samples remains low, our approach could provide data for better diagnoses with cytological samples of lung cancer. More rigorous analysis of the data would allow the detection of small populations of tumor cells in a reliable way. Indeed, the lowest percentage of tumor cells with chromosome 3 imbalances would determine the sensitivity of the method. The essential requirement for improved diagnosis would be the specificity of this chromosome rearrangement for lung cancer. Hence, malignant metastatic cells from lung cancer in pleural effusions could be distinguished from reactive mesothelial hyperplasia and mesothelioma,19 and atypical cells or secondary carcinoma from other sites such as the breast,20 because these cells do not show imbalance of the chromosome 3 as a recurrent event.
Recent findings21 about tumor extension in lung carcinomas by analysis of cell proliferation showed evidence for higher proliferative activity of cells located at the periphery of the tumor. It would therefore be possible, using our approach, to confirm whether these cells have already undergone chromosome rearrangements and have become tumoral. This would give more insight into tumor progression.
Another potential avenue of diagnostic or perhaps prognostic improvement would be the application of our method to preneoplastic lesions such as bronchial epithelial dysplasia. Many studies using molecular biology show loss of heterozygosity at loci on the short arm of chromosome 3 occurring in all histological types of lung cancer,7, 8 so that our marker could be of prognostic significance in such lesions. Because they tend to be noninvasive but premalignant, a detailed analysis could show at what stage they become invasive carcinomas. Using an in situ approach, it would also be possible to obtain additional data on routinely stained slides such as Papanicolaou-stained bronchial cytological samples. For this purpose, a series of bronchial brush and aspiration specimens is currently analyzed by FISH after destaining of the slides. Additionally, we intend to compare these data with clinical data obtained from conventional cytology and corresponding histological biopsies.
In conclusion, quantitative FISH of lung cancer samples allows detection of the presence of malignant cells of lung cancer in a reliable manner, as previously shown for breast cancer cells, by analyzing the imbalance of chromosome 1.11 This new tool could improve the detection of lung cancer cells and help in the characterization of premalignant lesions.
|
Footnotes |
|---|
Supported by grants from the Ligue Nationale contre le Cancer (Comité National, Paris et Hauts-de-Seine) and the Association pour la Recherche sur le Cancer (ARC).
Accepted for publication August 31, 1999.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
