validation of a targeted dna microarray for the clinical evaluation of recurrent abnormalities in...

Validation of a targeted DNA microarray for the clinical evaluationof recurrent abnormalities in chronic lymphocytic leukemia

Ankita Patel,1* Sung-Hae Kang,1y Patrick Alan Lennon,2y Yin Feng Li,1 P. Nagesh Rao,3 Lynne Abruzzo,4

Chad Shaw,1 Alan Craig Chinault,1 and Sau W Cheung1

Recurrent genomic alterations, mainly losses and gains of specific chromosomes and/or regions, in chroniclymphocytic leukemia (CLL) are recognized as important independent predictors of prognosis and diseaseprogression. The current standard clinical practice for identifying these alterations is chromosome analysisand in situ hybridization with probes targeting 4–5 chromosome regions. We sought to apply array compar-ative genomic hybridization (array-CGH) technology for the simultaneous detection of genomic imbalancesof all loci implicated in CLL. DNA from enriched B-cells from CLL patients were analyzed by array-CGH ona customized CLL BAC array. Copy number changes were detected in 87% of samples with a sensitivity of100% in samples with clonal abnormalities present in at least 23% of the cells. Furthermore, in nine casesgenomic alterations were observed that were undetectable by standard cytogenetic and/or FISH analyses.One of these patients had a 13q14 deletion that was missed by the clinical CLL FISH panel probe set. Ourresults suggest that a subset of potentially significant genomic alterations in CLL is being missed by thecurrent available techniques. Furthermore, this pilot study clearly shows the robustness, high sensitivity,and high specificity for the targeted CLL microarray analysis as well as the potential for use in routinescreening in CLL. Am. J. Hematol. 83:540–546, 2008. VVC 2007 Wiley-Liss, Inc.

IntroductionChronic lymphocytic leukemia (CLL) is the most common

form of leukemia in the Western Hemisphere and accountsfor about 25% of all leukemia in the United States [1]. Theclinical course of CLL is heterogeneous, ranging from nor-mal life expectancy to rapid progression leading to death. Amajor advance in predicting the prognosis of patients withCLL was the development of clinical staging systems byRai et al. and Binet et al. [2,3]. Although these clinical stag-ing systems are useful for classifying patients into broadrisk categories, they cannot predict clinical course for earlystage CLL to identify patients with poor prognosis. There-fore, over the past decade there has been active interest inidentifying molecular markers that would be useful for pre-dicting the clinical course in these patients. Among the old-est and most promising markers are cytogenetic aberra-tions, first published in the late 1970s after the discovery ofB-cell mitogens to induce the proliferation of leukemic cellsfrom B-cell tumors [4,5]. It was found that CLL is quiteunique among leukemias in that genomic imbalances suchas deletions and gains are more frequently seen thantranslocations. However, CLL B-cells are difficult to grow inculture using routine cytogenetic methods. Karyotype anal-ysis from cultured cells stimulated with CD40 ligand or withimmunostimulatory CpG-oligonucleotide and interleukin 2,identified complex chromosome rearrangements in 24–40%of the cases with the majority being in the form of unbal-anced translocations [6,7]. It is important to note howeverthat the clinical significance of these rearrangementsremain unclear. The advent of a panel of FISH probes to di-agnosis CLL on unstimulated interphase cells overcomesthis obstacle and is currently the standard of care for CLLpatients, but additional potentially significant abnormalitiesmay be missed using interphase FISH alone.A subset of genomic aberrations have been identified as

important independent predictors of disease progressionand survival [8]. Trisomy 12 is the most common clonalabnormality seen in CLL by karyotype analysis and it hasbeen correlated to advanced stage disease and shorter

survival time [9]. Deletion of 13q14 is the most commoncytogenetic abnormality seen by FISH analysis in CLL thatis correlated with favorable prognosis in the absence ofother abnormalities [10], whereas deletions of 11q and 17pare associated with disease progression and reduced sur-vival [11–13]. Therefore, routine diagnostic work up of CLLpatients include conventional cytogenetics and interphaseFISH (iFISH) analysis with a commercially available panelof five probes representing four chromosome regions(11q23, 12, 13q14, 13q34, and 17p13) and some laborato-ries include 6q21 to detect deletion in 6q and the IgH probeto detect translocation and deletion of 14q32 [14–17].Recently, array-CGH analysis has shown great promise

as a high throughput tool for the analysis of genetic altera-tions in complex cancer genomes. Therefore, the develop-ment of a clinical chip that targets all the genomic regionsimplicated in CLL can potentially improve diagnostics, andwith accumulation of more array data, the risk stratificationof CLL patients. Schwaenen et al. [18] have publishedresults using a customized array for diagnosis of geneticaberrations in CLL. They showed that in testing 106patients with CLL there was high specificity and 100% sen-

1Department of Molecular and Human Genetics, Baylor College of Medicine,Houston, Texas; 2School of Health Sciences, MD Anderson Cancer Center,Houston, Texas; 3Department of Pathology and Laboratory Medicine, DavidGeffen School of Medicine at UCLA, Los Angeles, California; 4Departmentof Hematopathology, MD Anderson Cancer Center, Houston, Texas

*Correspondence to: Ankita Patel, Ph.D., Department of Molecular andHuman Genetics, Baylor College of Medicine, One Baylor Plaza, NAB 2015,Houston, TX 77024. E-mail: [email protected]

yThese authors contributed equally to this study.

Received for publication 16 August 2007; Revised 14 December 2007;Accepted 19 December 2007

Am. J. Hematol. 83:540–546, 2008.

Published online 27 December 2007 in Wiley InterScience (www.interscience.wiley.com).DOI: 10.1002/ajh.21145

VVC 2007 Wiley-Liss, Inc.

American Journal of Hematology 540 http://www3.interscience.wiley.com/cgi-bin/jhome/35105

sitivity for cases with clonal abnormalities in >53% of thecell population as confirmed by FISH analysis. However,their array included over 270 control regions spanning thewhole genome thus complicating the analysis by potentiallydetecting normal copy number variation (CNV).The aim of this pilot study was to demonstrate potential

clinical utility of array CGH by (1) detecting clonal changesin CLL with greater sensitivity using extracted DNA from aB-cell enriched population and (2) designing a chip focusingon genomic regions implicated in CLL (Fig. 1). This tar-geted small chip should also simplify data analyses by min-imizing the number of changes detected due to normalCNV. Such variation would present a significant problem inthe CLL population because a majority of the patientswould be over the age of 50 and parental studies to deter-

mine whether the changes are de novo or inherited maynot be feasible.

Results and Discussion

SensitivityTo evaluate whether the sensitivity of array-CGH for

detecting clonal changes in CLL could be improved, CLLpatient DNA extracted from enriched B-cells instead offrom whole blood was used to interrogate our custom CLLchip. Thirty-one CLL samples were analyzed, in a blindedfashion. Interphase FISH and/or cytogenetic analysesdetected 34 abnormalities in 25 samples whereas six werenegative. Array-CGH analyses detected 37 abnormalities in27 samples and four samples showed no change (Table I).

Figure 1. Ideogram showing chromosome regions represented on the CLL microarray chip. Solid lines to the right of the ideogram represent chromosome regionsassayed by the CLL chip. The asterisks mark the chromosomes implicated in CLL.

American Journal of Hematology 541

Overall, the sensitivity of array-CGH analysis in this studywas 82% with specificity of 100% for cases with clonalaberrations in 6–96% of the cell population by iFISH analy-sis. For constitutional abnormalities, it has been establishedthat array-CGH can detect genomic copy number changewhen present in at least 25% of the cell population [unpub-lished data], and sometimes as low as 10–20% when itinvolves the whole chromosome [19–21]. Using 25% as thecutoff value for detecting clonal changes by array-CGH forthe CLL chip pilot study, the sensitivity is 100%. Schwae-nen et al. reported in their series of CLL patients studiedby array-CGH analysis a sensitivity of 100% using 53% asthe cutoff for detection [18]. Therefore, our clonal detectioncutoff value of 25% is two times more sensitive than whathas been previously reported. This increase in sensitivity ismost likely due to using DNA from an enriched populationof cells in our study. Therefore, these results suggest thatin CLL patients, enriching for B-cells allows for higher sen-sitivity for detecting clonal abnormalities using array-CGHanalysis.

Genomic copy number loss and gain detected byarray-CGH versus iFISHArray-CGH analyses detected 31 aberrations in 26 sam-

ples that were concordant with the iFISH and/or cytoge-netic analysis while it failed to detect at least one of theiFISH detected abnormalities in seven cases (Table II). Infour of these cases (CLL 14, 16, 21, 30), array-CGH analy-sis failed to detect one of the clonal changes. Three of thefour cases (CLL 14, 21, 30) involved the clonal loss of the13q34 or the 13q14 region in 6–12% of the cells examinedby iFISH analysis (Table II) which is very close to or at thenormal cut-off value for a positive iFISH analysis [22,23]. Incase CLL16, array-CGH detected the clonal loss in 13q14that was present in 11% of the cells by FISH analysis. Itdid not, however, identify the clonal loss of 11q23 (ATM) asobserved by iFISH analysis in 23% of the cells. The dis-crepancy between the iFISH and array results in this casemay be because of the population of cells used for eachassay. During the timeframe these samples were analyzedin the clinical cytogenetics laboratory at MD Anderson Can-cer Center, FISH using the CLL panel was routinely per-formed on B-cell stimulated cultured cells. In contrast, thearray was performed on DNA extracted directly fromenriched B-cells from whole blood without culturing.Unfortunately, additional samples were not available to testthis hypothesis. It is also possible that the observed dis-crepancy may be due to differences in sensitivity betweenthe two assays. Smaller rearrangements are more difficultto detect using BAC array-CGH when present in lower lev-els. In all four cases described above the clonal abnormal-ities detected by iFISH were less than 25% which in gen-eral, are beyond the resolution of array-CGH analysis.Although array-CGH did not identify the clone that was

detected by iFISH or cytogenetic analysis in the remainingthree of the seven cases (CLL7, 27, 28), it did however,identify other additional genomic rearrangements (Table II).In CLL27 array-CGH showed a gain of chromosome 18q

but did not identify a trisomy 12 clone present in 6% of thecells as determined by FISH analysis (Fig. 2). The Bcl-2gene maps to 18q21 has been shown to be overexpressedin CLL and has been associated with poor response tochemotherapy and advanced stages of CLL.[24–26] InCLL28, deletion of 11q23 was observed in 8% of the cellsexamined by iFISH analysis. Although array-CGH failed todetect this loss, which was present at a level below thethreshold of detection by this method, it did identify a lossof the 6q14-q21 region that has been previously reportedto be found in 7% of CLL cases [27]. Use of a FISH probethat hybridizes to 6q21 is now standard in many clinical lab-oratories performing CLL iFISH analysis. In the last case(CLL7), cytogenetic analysis showed two cells with at(14;18) which is characteristic of follicular cell lymphomaand has also been described CLL [14–17]. However, it hasbeen found that the translocation seen in CLL is molecu-larly distinct from that observed in follicular B-cell lym-phoma and the currently available FISH assays are notable to distinguish between the two. In general, characteris-tic translocations are often seen in acute leukemia and lym-phomas but are rare in CLL. Array-CGH analysis is notable to detect balanced rearrangements, however in thiscase, an additional deletion of the 9p21 region wasdetected and subsequently confirmed by FISH analysiswith probes from the region (Fig. 3). The 9p21 region har-bors the p14 and p16 tumor suppressor genes and hasbeen shown to be deleted in a wide variety of cancersincluding progressing follicular cell lymphoma [28]. Althougharray-CGH cannot detect balanced translocations, inclusionof loci on the array where rearrangements occur in B-cellmalignancies may be helpful in differentiating CLL fromother more aggressive forms of cancer. These results high-light the need to examine a broader selection of loci thanwhat is currently available using the CLL FISH panel tomore accurately diagnose CLL patients and potentiallyidentify new prognostically significant aberrations in CLL.

Additional findings by array-CGHThere were three cases (CLL 4, 24, 29) with the sole ab-

normality of trisomy 12 by FISH or chromosome analysiswhere the array identified an additional genomic rearrange-ment (Table II). CLL4 showed two cells with trisomy 12 bykaryotype analysis. Array-CGH not only identified the tris-omy 12 clone but it also identified an additional gain of the3q region which has been described in CLL. AlthoughCLL24 harbors a trisomy 12 clone detected by both FISHanalysis (61%) and array-CGH, array-CGH identified addi-tional gains of chromosome 19 and the 22q13 region (Fig.2). There are now several reports of trisomy 19 in CLLpatients with trisomy 12 [18,29,30]. Deletions in the 22q13region are seen in 50% of MCL cases and gains inthis region have also been reported [31]. Additional studiesare required to determine if rearrangements in the 22q13region will have any prognostic significance in CLL or MCL.For now, the translocation (11;14)(q13;q32), which is patho-gnomonic for MCL should remain a part of the CLL FISHpanel test to differentiate between CLL and MCL as bal-anced translocations cannot be detected by array-CGH.[14] In CLL29, in addition to trisomy 12, array-CGHdetected a gain in copy number in the N-myc gene on chro-mosome 2p24. Gain of N-myc and gain of 2p has nowbeen described by at least two groups in CLL [7,18].

Detection of 13q14 deletions by array-CGHArray-CGH in this series also identified two cases of

13q14 deletions (CLL13 and CLL26) that were initiallymissed by the clinical iFISH analysis (Table II). There were16 cases with a 13q14 deletion in the present study. Based

TABLE I. Comparison of Abnormalities Detected by

FISH/Chromosome versus Array-CGH Analyses

Cases

Abnormalities detected

FISH Array-CGH

25 34 35

6 0 2

31 34 37

542 American Journal of Hematology

on the clone coverage of the 13q14 region on our chip (10BAC clones covering 16 Mb), it is apparent that the size ofthis deletion is heterogeneous in CLL (Table III). CLL26had a 1.5 Mb 13q14 deletion and is located proximal to thelocus assayed by the commercially available 13q14 FISHprobe (D13S319) currently used by most clinical laborato-ries. In ten out of sixteen cases with the 13q14 deletion,the deletion size appeared to be at least 10 Mb or less.Deletions of this size are usually cryptic at the chromosomeresolution seen in conventional cancer cytogenetics. Theseresults confirm recent studies using FISH to show that the13q deletions in B-CLL metaphases are cryptic [32].Theother 13q14 deletion case (CLL13) which was missed bythe CLL iFISH panel, should have been detected by theD13S319 FISH probe. Additional FISH analyses on an

enriched B-cell preparation using the probes on the arrayand the commercial 13q14 probe (D13S319) showed a de-letion in 93% of the cells (Fig. 4). Again, it is possible thatdifferences in the population of cells used in each assaycontribute to the observed discrepancy but a lab error can-not be ruled out. This study demonstrates that array-CGHhas the potential to identify atypical deletions in criticalregions in CLL that could be missed by iFISH analysisusing a single locus probe.

Conclusions and future directionsIn summary, we show that a highly specific and sensitive

analytical tool that offers a robust approach for detectingchromosomal aberrations can be achieved using enrichedB-cell extracted DNA from CLL patients on a targeted

TABLE II. Comparison of Known FISH/Chromosome Abnormalities to the Copy Number Changes Detected by Array CGH

Case FISH (% cells) or karyotype results Array interpretation Comparison

CLL2 13q14 deletion (94%) Loss of 13q14 Consistent



CLL6 13q14 biallelic deletion (39%) Loss of 13q14 (biallelic) Consistent

CLL8 11q23 deletion (51.2%) Loss of 11q23 Consistent

CLL9 13q14 deletion (85.5%), biallelic 13q14 deletion (21.5%) Loss of 13q14 Consistent

CLL10 13q14 deletion (96.8%) Loss of 13q14 Consistent


CLL12 Negative FISH No losses or gains Consistent



13q14 deletion (84%) Loss of 13q14 Consistent

CLL18 13q14 biallelic deletion (79%) Loss of 13q14 (biallelic) Consistent




CLL22 Trisomy 12 (62%) Gain of chromosome12 Consistent

13q14 deletion (43.5%) Loss of 13q14 Consistent

CLL23 13q14 biallelic deletion (87%), monosomy 13 (10%) Loss of 13q14 (biallelic) Consistent

17p13 deletion (93%) Loss of 17p13 Consistent


CLL31 Trisomy 12 (77%) Gain of chromosome 12 Consistent

CLL32 17p13 deletion (81%) Loss of 17p13 Consistent

CLL4 47,XX112[cp3]/46,XX[17] Gain of chromosome 12 Consistent

Gain of 3q Additional finding


Gain of 2p Additional


Gain of chromosome 19 Additional finding

Gain of 22q13 Additional finding


13q34 deletion (8%) Not detected

CLL16 11q23 deletion (23%)

Loss of 13q14

Not detected

13q14 deletion (11%) Consistent

CLL21 13q14 deletion (12%) Not detected





Loss 10q24 Additional

CLL7 47,XY,t(14;18)(q32;q21),1mar[2]/46,XY[18] Not detected

Loss of 9p21 Additional finding

CLL27 Trisomy 12(7%) Not detected

Gain of chromosome 18 Additional finding


Loss of 6q21 Additional finding

CLL13 Negative FISH Loss of 13q14 Additional finding

CLL26 Negative FISH Loss of 13q14 Additional finding


chip that includes regions implicated in CLL. This array-based method gives results that are very comparable withthose from iFISH analysis and is superior to standard cyto-genetics.Both array-CGH and iFISH have advantages and disad-

vantages. The current iFISH test only assays a handful ofgenomic regions in one experiment while a single array-CGH experiment is limited only by the number probes thatare on the chip. Balanced translocations may be missed byboth techniques. An advantage of array-CGH is that addi-tional probes can be easily added to assay to identity otherpotentially significant rearrangements. Although the clinical

significance of additional genomic regions is not known atthis time, increased use of array-CGH on patients diag-nosed with CLL facilitates identification of new regions thatmay prove to be significant in CLL. Currently the level ofsensitivity of array-CGH is a disadvantage. There were sixprognostically significant abnormalities that were present inless than 25% and was detected by iFISH analysis thatwere missed by the array analysis. It has been shown thatthe level of sensitivity can be increased by increasing thenumber of probes interrogating the region of interest [21].This is clearly illustrated in the ability of array-CGH todetect whole chromosome gains and losses in as low as10-20% of the cell population [19,20]. Furthermore, oligonu-cleotide-based arrays are becoming more common for bothresearch and clinical applications. In our experience withconstitutional abnormalities, we find that targeted oligonu-cleotide arrays designed with oligonucleotides that areselected with stringent Tm criteria are more sensitive thanBAC arrays [unpublished data]. We predict that theincreased dynamic range observed with oligo arrays couldfacilitate detection of abnormalities present in smaller cellpopulations. Regardless, array-CGH is clearly not a usefulassay to measure minimal residual disease (MRD). Moni-toring patients for MRD require much more sensitiveassays, and is probably best performed using FISH orquantitative PCR.We propose that the robustness of the assay and small

design of this CLL chip make routine clinical use of thistechnology an attractive approach for screening CLLpatients. The compactness of the chip would also affordcost-effectiveness so that it can be easily incorporated intoa clinical diagnostic laboratory. Also, the results from thispilot study are promising and likely warrant a larger studyusing higher resolution oligonucleotide arrays with carefullyselected probes to further demonstrate the clinical utility ofthis technology as well as the potential for discovering newgenomic alterations in CLL.

Materials and Methods

Patient samplesA total of 31 blood samples from untreated CLL patients that were

received for routine cytogenetic and/or CLL panel iFISH analysis to rule

Figure 2. Array-CGH ratio plots of cases CLL27 and CLL24. Averaged normal-ized log2 ratios from two hybridizations are plotted on the x-axis as dots with errorbars. The clones are ordered linearly by chromosome. (A) Array-CGH profile ofCLL 27 showing a gain in copy number of all chromosome 18q clones (circled).(B) Array-CGH profile of CLL24 showing a gain in copy number of all chromo-somes 12, 19 and 22q13 clones (circled).

Figure 3. Array-CGH analysis and FISH confirmation of deletion 9p21 in CLL7.(A) Array-CGH profile of CLL7 showing a deletion of chromosome region 9p21(circled). (B). FISH analysis with RP11-149I2 (green signal) and control probe forcentromere 9 (red signals) showing a deletion on one chromosome, confirming thearray-CGH analysis.

Figure 4. Array-CGH and FISH analyses of CLL13. (A) Array-CGH profile ofCLL13 showing a deletion of chromosome 13q14 region (circled). (B) FISH analy-sis with clones RP11-259O18 and RP11-277G5 (represented on the CLL arraychip) on B-cells showing only one copy. (C) FISH analysis on enriched B-cellsfrom sample CLL13 with the Vysis CLL panel probe D13S310 showing only onecopy.


out CLL at MD Anderson Cancer Center and UCLA cytogenetic labora-tories were de-identified and forwarded to Baylor College of Medicinefor microarray analysis.

B-cell enrichmentB-cells were enriched from the CLL blood samples with the Rosette-

Sep Human B-Cell Enrichment kit (cat# 15705; StemCell Technologies,Vancouver, British Columbia) using the manufacturer’s protocol. Theexpected purity of CD191 B cells is 90% ± 3.5% according to themanufacturer (http://www.stemcell.com/technical/8_RS%20DML%20v1.1.pdf). DNA was either extracted immediately for use in the array-CGH,or cells were stored overnight in PBS 1 2% FBS.

DNA extractionGenomic DNA was extracted from enriched B-cells using the Pure-

gene1 Genomic DNA Purification Kit (Gentra Systems, Minneapolis,MN) according to the manufacturer’s protocol.

Array productionThe CLL array consists of 220 BAC/PAC clones, selected from the

publication Schwaenen et al., which were selected from NCBI/UCSCpublic databases. These clones covered 12 chromosomes (2p, 3q, 6,8q,10, 11, 12, 13, 14, 17, 18, 19) implicated in CLL and four chromo-some regions (4q, 9p, 20, 22) implicated in B-cell disorders such aslarge B-Cell and mantle cell lymphoma (Fig. 1). Chromosomal locationand specificity of each clone were verified by FISH analysis prior toarray printing. DNA from each clone was prepared by a standard alka-line lysis method and chemically modified for array printing as pre-viously described. Briefly, the DNA was chemically cross-linked using(3-glycidoxypropyl) trimethoxysilane (Sigma, St. Louis, MO) and printedon alkaline/acid cleaned Bis(trichlorosilyl) octane (Gelest, Morrisville,PA)-treated glass slides (VWR Scientific micro slides cat. no. 48300-047) using a GeneMachine Omnigrid Accent Microarrayer (GenomicSolutions, Ann Arbor, MI). All clones were printed in duplicate from384-well plates in spatially distinct subarrays.

FISH analysisFISH analyses for mapping verification of all BACs/PACs spotted on

the array and for confirmation of the array-CGH analysis results wereperformed using standard procedures. Briefly, the BAC and PAC clonesof interest were grown in TB media with either 20 lg/ml chlorampheni-col or 25 lg/ml kanamycin, respectively. DNA was isolated using astandard alkaline lysis kit (Eppendorf Plasmid Mini Prep; Hamburg,Germany). DNA was labeled using digoxigenin-11-UTP or biotin-UTPand detected with anti-digoxigenin-rhodamine or avidin-FITC fragments,respectively. Digital FISH images were captured using a Power Macin-tosh G3 System and MacProbe version 4.4 (Applied Imaging; SanJose, CA).

CLL FISH panel analysisInterphase cells from CLL cultures were analyzed by FISH using the

commercially available CLL FISH panel probe set from Vysis (Abbott

Molecular, Des Plaines, IL) according to the manufacturer’s instruc-tions. A total of 200 nuclei were scored for each probe. Results wereconsidered abnormal based on the following cut-off values as deter-mined by probe validation procedures established at MD AndersonCancer Center [33]: 5% for trisomy 12, 5.7% for ATM deletion, 5.6% for13q14 deletion, 6.5% for 13q34 deletion and 6.6% for p53 deletion.

Array-CGHArray-CGH analysis was performed as previously described [34].

Briefly, 500 ng each of fragmented DNA from patient and control sam-ples were differentially labeled with Cyanine 3-dCTP (Cy3) and Cyanine5-dCTP (Cy5) (Perkin–Elmer, Boston, MA) using the BioPrime randompriming kit (Invitrogen, Carlsbad, CA). Both probes were simultaneouslyhybridized onto the microarray slides for 20–24 hr at 378C. The slideswere washed and then scanned using an Axon microarray scanner(GenePix 4000B, Molecular Devices, Sunnyvale, CA) to visualize thefluorescent signals. Microarray image files (TIFF) were quantified usingGenePix Pro 6 software or BlueFuse software (BlueGnome, Cam-bridge, UK). For each patient sample, parallel experiments were per-formed with reversal of the dye. The quantitation data from both hybrid-izations were combined and subjected to normalization to determinefold-change detected by each clone. All analyses were performed onlog2 ratios using code for the normalization and inference that wasimplemented in the R statistical programming language.

AcknowledgmentsThe authors thank George M. Weissenberger and Ellen

K. Brundage in the Microarray Printing Laboratory for theirexpert technical assistance.

References1. American Cancer Society. Cancer Facts and Figures 2006. Atlanta; 2006.2. Binet JL, Auquier A, Dighiero G, et al. A new prognostic classification of

chronic lymphocytic leukemia derived from a multivariate survival analysis.Cancer 1981;48:198–206.

3. Rai KR, Sawitsky A, Cronkite EP, et al. Clinical staging of chronic lymphocyticleukemia. Blood 1975;46:219–234.

4. Robert KH, Moller E, Gahrton G, Eriksson H, et al. B-cell activation of periph-eral blood lymphocytes from patients with chronic lymphatic leukaemia. ClinExp Immunol 1978;33:302–308.

5. Schroder J, Vuopio P, Autio K. Chromosome changes in human chronic lym-phocytic leukemia. Cancer Genet Cytogenet 1981;4:11–21.

6. Buhmann R, Kurzeder C, Rehklau J, et al. CD40L stimulation enhances theability of conventional metaphase cytogenetics to detect chromosome aberra-tions in B-cell chronic lymphocytic leukaemia cells. Br J Haematol 2002;118:968–975.

7. Dicker F, Schnittger S, Haferlach T, et al. Immunostimulatory oligonucleotide-induced metaphase cytogenetics detect chromosomal aberrations in 80% ofCLL patients: A study of 132 CLL cases with correlation to FISH, IgVH status,and CD38 expression. Blood 2006;108:3152–3160.

8. Dohner H, Stilgenbauer S, Benner A, et al. Genomic aberrations and survivalin chronic lymphocytic leukemia. N Engl J Med 2000;343:1910–1916.

TABLE III. Size of 13q14 Deletions by Array-CGH in CLL Samples

Clones Position

Cases

2 3 5 6 9 10 11 13 14 16 17 18 20 22 23 26

RP11-9F13 40.6 Mb 2 1 1 1 1 1 1 + 1 1 1 1 1 2 1 +RP11-196P14 47.6 Mb 2 2 1 1 1 1 1 + 2 1 2 1 2/1 2 2 2

RP1-278N3 49.1 Mb 2 2 1 2 1 2 1 − 2 2 2 1 2 2 2 2

RP1-173A12 49.6 Mb 2 2 2 2 2 2 2 − 2 2 2 2 2 2 2 +RP1-105L4 49.6 Mb 2 2 2 2 2 2 2 − 2 2 2 2 2 2 2 +D13S319 CLL FISH panel probe 49.8 Mb

RP1-295O18 49.9 Mb 2 2 2 2 2 2 2 − 2 2 2 2 2 2 2 +RP1-269F22 50 Mb 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 +RP1-277G5 50 Mb 2 2 2 2 2 2 2 − 2 2 2 2 2 2 2 +RP11-205J24 56.8 Mb 1 1 1 1 1 1 1 + 1 1 1 1 1 1 2/1 +

Boldface: Cases that were negative for 13q14 deletion by FISH analysis.

1, Normal array-CGH analysis; 2, loss of copy number by array-CGH.


9. Matutes E, Oscier D, Garcia-Marco J, et al. Trisomy 12 defines a group ofCLL with atypical morphology: Correlation between cytogenetic, clinical andlaboratory features in 544 patients. Br J Haematol 1996;92:382–388.

10. Juliusson G, Oscier DG, Fitchett M, et al. Prognostic subgroups in B-cellchronic lymphocytic leukemia defined by specific chromosomal abnormalities.N Engl J Med 1990;323:720–724.

11. Fegan C, Robinson H, Thompson P, et al. Karyotypic evolution in CLL: Identi-fication of a new sub-group of patients with deletions of 11q and advanced orprogressive disease. Leukemia 1995;9:2003–2008.

12. Neilson JR, Auer R, White D, et al. Deletions at 11q identify a subset ofpatients with typical CLL who show consistent disease progression andreduced survival. Leukemia 1997;11:1929–1932.

13. Geisler CH, Philip P, Christensen BE, et al. In B-cell chronic lymphocytic leu-kaemia chromosome 17 abnormalities and not trisomy 12 are the single mostimportant cytogenetic abnormalities for the prognosis: A cytogenetic andimmunophenotypic study of 480 unselected newly diagnosed patients. LeukRes 1997;21:1011–1023.

14. Nowakowski GS, Dewald GW, Hoyer JD, et al. Interphase fluorescence in situhybridization with an IGH probe is important in the evaluation of patients witha clinical diagnosis of chronic lymphocytic leukaemia. Br J Haematol 2005;130:36–42.

15. Aoun P, Blair HE, Smith LM, et al. Fluorescence in situ hybridization detectionof cytogenetic abnormalities in B-cell chronic lymphocytic leukemia/small lym-phocytic lymphoma. Leuk Lymphoma 2004;45:1595–1603.

16. Dyer MJ, Oscier DG. The configuration of the immunoglobulin genes in B cellchronic lymphocytic leukemia. Leukemia 2002;16:973–984.

17. Yabumoto K, Akasaka T, Muramatsu M, et al. Rearrangement of the 50 clusterregion of the BCL2 gene in lymphoid neoplasm: A summary of nine cases.Leukemia 1996;10:970–977.

18. Schwaenen C, Nessling M, Wessendorf S, et al. Automated array-basedgenomic profiling in chronic lymphocytic leukemia: Development of a clinicaltool and discovery of recurrent genomic alterations. Proc Natl Acad Sci USA2004;101:1039–1044.

19. Ballif BC, Rorem EA, Sundin K, et al. Detection of low-level mosaicism byarray CGH in routine diagnostic specimens. Am J Med Genet A 2006;140:2757–2767.

20. Cheung SW, Shaw CA, Scott DA, et al. Microarray-based CGH detects chro-mosomal mosaicism not revealed by conventional cytogenetics. Am J MedGenet A 2007;143:1679–1686.

21. Garnis C, Coe BP, Lam SL, et al. High-resolution array CGH increases heteroge-neity tolerance in the analysis of clinical samples. Genomics 2005;85: 790–793.

22. Dewald GW, Brockman SR, Paternoster SF, et al. Chromosome anomaliesdetected by interphase fluorescence in situ hybridization: Correlation with sig-nificant biological features of B-cell chronic lymphocytic leukaemia. Br J Hae-matol 2003;121:287–295.

23. Goorha S, Glenn MJ, Drozd-Borysiuk E, Chen Z. A set of commercially avail-able fluorescent in-situ hybridization probes efficiently detects cytogeneticabnormalities in patients with chronic lymphocytic leukemia. Genet Med2004;6:48–53.

24. Robertson LE, Plunkett W, McConnell K, et al. Bcl-2 expression in chroniclymphocytic leukemia and its correlation with the induction of apoptosis andclinical outcome. Leukemia 1996;10:456–459.

25. Aviram A, Rabizadeh E, Zimra Y, et al. Expression of bcl-2 and bax in cellsisolated from B-chronic lymphocytic leukemia patients at different stages ofthe disease. Eur J Haematol 2000;64:80–84.

26. Faderl S, Keating MJ, Do KA, et al. Expression profile of 11 proteins and theirprognostic significance in patients with chronic lymphocytic leukemia (CLL).Leukemia 2002;16:1045–1052.

27. Stilgenbauer S, Bullinger L, Benner A, et al. Incidence and clinical signifi-cance of 6q deletions in B cell chronic lymphocytic leukemia. Leukemia1999;13:1331–1334.

28. Elenitoba-Johnson KS, Gascoyne RD, Lim MS, et al. Homozygous deletionsat chromosome 9p21 involving p16 and p15 are associated with histologicprogression in follicle center lymphoma. Blood 1998;91:4677–4685.

29. Schlette E, Rassidakis GZ, Canoz O, Medeiros LJ. Expression of bcl-3 inchronic lymphocytic leukemia correlates with trisomy 12 and abnormalities ofchromosome 19. Am J Clin Pathol 2005;123:465–471.

30. Sellmann L, Gesk S, Walter C, et al. Trisomy 19 is associated with trisomy 12and mutated IGHV genes in B-chronic lymphocytic leukaemia. Br J Haematol2007;138:217–220.

31. Schraders M, Pfundt R, Straatman HM, et al. Novel chromosomal imbalancesin mantle cell lymphoma detected by genome-wide array-based comparativegenomic hybridization. Blood 2005;105:1686–1693.

32. Stockero KJ, Fink SR, Smoley SA, et al. Metaphase cells with normal G-bands have cryptic interstitial deletions in 13q14 detectable by fluorescencein situ hybridization in B-cell chronic lymphocytic leukemia. Cancer GenetCytogenet 2006;166:152–156.

33. Glassman AB, Hayes KJ. The value of fluorescence in situ hybridization inthe diagnosis and prognosis of chronic lymphocytic leukemia. Cancer GenetCytogenet 2005;158:88–91.

34. Cheung SW, Shaw CA, Yu W, et al. Development and validation of a CGHmicroarray for clinical cytogenetic diagnosis. Genet Med 2005;7:422–432.


validation of a targeted dna microarray for the clinical evaluation of recurrent abnormalities in...

Documents