• oligonucleotide microarrays;
  • immunophenotyping;
  • leukemia;
  • gene expression;
  • protein expression


  1. Top of page
  2. Abstract
  5. Acknowledgements


Microarray analysis is considered a future diagnostic tool in leukemias. Whereas data accumulate on specific gene expression patterns in biologically defined leukemia entities, data on the correlation between flow cytometrically determined protein expression, which are essential in the diagnostic setting today, and microarray results are limited.


The results obtained by microarray analysis were compared using the Affymetrix GeneChip HG-U133 system in parallel with flow cytometric findings of 36 relevant targets in 814 patients with newly diagnosed acute and chronic leukemias as well as in normal bone marrow samples.


In a total of 21,581 individual comparisons between signal intensities obtained by microarray analysis and percentages of positive cell as determined by flow cytometry, coefficients of correlation in the range of 0.171 to 0.807 were obtained. In particular, the degree of correlation was high in the following genes critical in the diagnostic setting: CD4, CD8, CD13 (ANPEP), CD33, CD23 (FCER2), CD64 (FCGR1A), CD117 (KIT), CD34, MPO, CD20 (MS4A1), CD7 (range of r, 0.589–0.807).


The present data prove the high degree of correlation between findings obtained by microarray analysis and flow cytometry. They are in favor of a future application of the microarray technology as a robust diagnostic tool in leukemias. Cancer 2006. © 2006 American Cancer Society.

Microarray analysis provides an innovative and powerful tool capable of determining the expression status of the whole genome in 1 approach.1 Among a variety of present and future applications, its use in the diagnostic setting of acute and chronic leukemias is anticipated within the next years.2–5 Thereby, current standard diagnostic methods will be supplemented and perhaps even replaced. Among these, multiparameter flow cytometry (MFC) represents an essential application needed to diagnose and subclassify acute lymphoblastic and chronic lymphocytic leukemias (ALL and CLL) and some acute myeloid leukemias (AML).6 Furthermore, MFC is used to monitor minimal residual disease taking advantage of leukemia-associated aberrant immunophenotypes based on which leukemic bone marrow cells can be distinguished from normal bone marrow cells with high sensitivity.7–9

A large variety of microarray studies have demonstrated specific gene expression signatures being present in well-defined subtypes of leukemia.10–22 Whereas these analyses have built the basis for the use of microarrays as a diagnostic tool in leukemias, the correlation of gene expression data with data obtained by standard methods for specific genes, e.g., immunophenotyping, may in addition lead to an increase in acceptance of this innovative method. Along this line, it has been shown that there is a good correlation between positivity for peroxidase and nonspecific esterase as determined by cytochemistry and the expression level of these genes as determined by microarrays in AML.2 Similarly, the fluorescence intensity in positive cells as obtained by flow cytometry has been shown to highly correlate with the gene expression levels obtained by microarrays in a large number of genes used in diagnosing acute leukemias.23 However, in today's routine diagnostic setting the standard immunophenotyping report gives the percentages of cells positive for the expression of the respective proteins as determined using isotype controls. In the present study, we therefore compared these percentages of positive cells with the gene expression intensities of the respective genes obtained by microarray analysis for 36 targets in 814 individual cases of acute and chronic leukemias.


  1. Top of page
  2. Abstract
  5. Acknowledgements


Fresh bone marrow samples from thoroughly characterized patients with newly diagnosed and untreated AML, ALL, CLL, biphenotypic acute leukemias (BAL), and myelodysplastic syndromes (MDS) as defined by the WHO classification24 and EGIL25 as well as from healthy volunteers were used and processed immediately. Samples were analyzed by cytomorphology, cytochemistry, immunophenotyping, cytogenetics, and molecular genetics in all cases and were characterized by specific chromosomal aberrations or molecular genetic alterations (Table 1).26 The studies complied with the rules of the local Internal Review Board and the tenets of the revised Helsinki protocol.

Table 1. Sample Characteristics (Total: N = 814)
DiseaseSubclassSample number
  1. ALL indicates acute lymphoblastic leukemia; AML, acute myeloid leukemia; BAL, biphenotypic acute leukemia; CLL, chronic lymphocytic leukemia; MDS, myelodysplastic syndrome.

Mature B-ALL/t(8;14)6
Cortical T-ALL9
Mature T-ALL2
AMLNormal karyotype201
Complex aberrant karyotype60
Other karyotype aberrations203
BAL 27
CLL 29
MDS5q- syndrome1
RAEB1 and RAEB216
Normal bone marrow 6

Flow Cytometry

The studies were performed on cells isolated from bone marrow (peripheral blood in CLL cases) by Ficoll-Hypaque density gradient centrifugation as described previously.27 Applying triple- and 5-fold-stainings and isotype controls, monoclonal antibodies against antigens were used in the combinations given in Tables 2 and 3. The respective combinations of antibodies were added to 1 × 106 cells (volume, 100 μL) and incubated for 19 minutes at room temperature. The samples were then washed twice in phosphate-buffered saline (PBS) and resuspended in 0.5 mL PBS. For the analysis of cytoplasmic antigens, cells were fixed and permeabilized before staining using Fix & Perm (Caltag, Burlingame, CA). The flow cytometric analysis was performed using a FACSCalibur flow cytometer (Becton Dickinson, San Jose, CA) and an FC500 flow cytometer (Beckman Coulter, Miami, FL). Analysis of list-mode files was performed by means of CellQuest Pro software (Becton Dickinson) and RXP software (Beckman Coulter). In order to obtain analyses on the same cells for both flow cytometry and microarray analysis, i.e., all nucleated cells of each sample, the analysis gate was set in a forward-scatter/side-scatter plot and included lymphocyte, blast, monocyte, and granulocyte populations. The percentage of positive cells was determined in comparison to isotype controls, which were set to 1%.

Table 2. Antibodies Used in Triple Combinations
  • All antibodies were purchased from Immunotech (Marseilles, France), except for:

  • *

    Medarex (Annandale, NJ)

  • Milteny Biotech (Bergisch Gladbach, Germany)

  • Caltag (Burlingame, CA)

  • §

    Dako (Glostrup, Denmark)

  • Pharmingen (San Diego, CA).

Table 3. Antibodies Used in 5-Fold Combinations
  • All antibodies were purchased from Immunotech (Marseilles, France), except for:

  • a

    * Medarex (Annandale, NJ)

  • b

    † Milteny Biotech (Bergisch Gladbach, Germany)

  • Caltag (Burlingame, CA)

  • §

    Dako (Glostrup, Denmark)

  • Pharmingen (San Diego, CA).


Microarray Analyses

The GeneChip System (Affymetrix, Santa Clara, CA) and the HG-U133 microarray set were used for microarray analyses.11, 23 This 2-array set provides comprehensive coverage of the human genome. The U133A and U133B arrays represent 44,000 probe sets and 1,000,000 distinct oligonucleotide features. For gene expression profiling, cell lysates of the leukemia samples were thawed, homogenized (QIAshredder, Qiagen, Hilden, Germany), and total RNA was extracted (RNeasy Mini Kit, Qiagen). The subsequent target preparation steps as well as hybridization, washing, and staining of the probe arrays were performed according to recommended protocols (Affymetrix Technical Manual). The Affymetrix software package (Microarray Suite 5.1) extracted fluorescence intensities from each element on the microarrays as detected by confocal laser scanning.28 Signal intensity values were calculated by scaling the raw data intensities to a common target intensity (U133 mask file; TGT value: 5000). Each human GeneChip expression array features 100 human maintenance genes that serve as a tool to normalize and scale the data before performing data comparisons. As recommended by the manufacturer, these 100 probe sets were used for normalization ( The minimal quality control parameters for inclusion of an expression profile in our study took into account more than 30% cells present (U133A microarray) and a low 3′/5′ ratio of represented glyceraldehyde-3′-phosphate dehydrogenase gene (GAPDH) probe sets.


Microarray results and flow cytometry from a total of 36 genes were compared in 814 cases with leukemias applying Spearman rank correlation using SPSS v. 12.0.1 (Chicago, IL).


  1. Top of page
  2. Abstract
  5. Acknowledgements

A total of 36 genes were analyzed in parallel by microarray analysis and flow cytometry in 814 cases, resulting in a total of 21,581 individual comparisons. Table 1 gives the classification characteristics of these cases.

The comparisons between signal intensities obtained by microarray analysis and percentages of positive cells as determined by flow cytometry using isotype controls resulted in highly significant correlations for the vast majority of the genes analyzed (Table 4), (Fig. 1). In these cases the coefficient of correlation ranged from 0.171 to 0.807. Specifically, with regard to the genes that are critical in immunophenotyping as used for diagnostic purposes in leukemias, i.e., CD4, CD8, CD13 (ANPEP), CD33, CD23 (FCER2), CD64 (FCGR1A), CD117 (KIT), CD34, MPO, CD20 (MS4A1), and CD7, high coefficients of correlation in the range of 0.589 to 0.807 were found.

thumbnail image

Figure 1. Correlation of flow cytometric and microarray analysis results. Flow cytometric results are given as percentages of positive cells as rated using isotype controls. Microarray analysis results are given as mean fluorescence intensities.

Download figure to PowerPoint

Table 4. Correlation between Microarray Signal Intensity and Percentage of Positive Cells by Flow Cytometry
Antigen (gene symbol*)Coefficient of correlationP
  • *

    Only given if different from commonly used antigen name.

NG2 (CSPG4)0.301<.001
CD116 (CSF2RA)0.648<.001
CD15 (FUT4)0.492<.001
CD13 (ANPEP)0.599<.001
CD23 (FCER2)0.723<.001
CD64 (FCGR1A)0.589<.001
CD135 (FLT3)0.552<.001
CD235a (GYPA)0.399<.001
CD11b (ITGAM)0.632<.001
TdT (DNTT)0.449<.001
CD117 (KIT)0.733<.001
CD20 (MS4A1)0.803<.001
CD56 (NCAM1)0.395<.001
CD90 (THY1)0.291<.001
CD87 (PLAUR)0.443<.001
CD133 (PROM1)0.683<.001
CD45 (PTPRC)0.244<.001

In general, several aspects of these analyses have to be considered. Thus, for some genes, i.e., CD13 and CD33 (Fig. 1), in most cases high levels were recorded by both flow cytometry and microarray analysis. Clearly, this is related to the respective numbers of different entities analyzed, with an emphasis on AML. Nonetheless, the resulting degree of correlation is very high, demonstrating the comparability of both methodologically different approaches. Another aspect is that, given the good correlations in most genes, the relation between flow cytometric and microarray findings differing from gene to gene resulted in more or less steep correlation curves.

For some comparisons, quite high ranges of microarray signal intensities in flow cytometrically negative cases were observed. These cases in fact were anticipated because large amounts of mRNA of the respective genes may be present in the cells without the resulting protein being readily expressed, at least on the cell surface.

The opposite was observed for CD235a (glycophorin A) in quite a number of cases. Thus, flow cytometry gives an intensive signal, whereas microarray signal intensities are very low (Fig. 1). In fact, this was anticipated to occur because erythroid cells are known to harbor significant amounts of RNAse,29 leading to a degradation of glycophorin A mRNA, whereas the protein is still strongly expressed on erythroid cells in the respective samples.

For 2 genes, however, the degree of correlation was lower than for those discussed above, i.e., CD3 and CD19, although the correlation is still significant. With regard to CD3 (r = 0.257, P < .001), it is obvious that the majority of cases reveal a flow cytometric positivity in the range of 10%. Thus, the levels are relatively low and interfering signals due to both nonspecific binding of antibodies and normal T-cells present in the bone marrow samples therefore must be considered. Accordingly, the correlation between microarray and flow cytometry clearly is better in cases with a flow cytometric positivity higher than 10% (Fig. 1). This is also true when considering only T-ALL cases in which the coefficient of correlation amounts to r = 0.672 (P = .00009).

Similarly, as for CD3, cases with a low level of positivity for CD19 show various signal intensities by microarray analysis and the same reason as discussed above must be considered. However, for CD19 even some cases with a strong flow cytometric positivity yield weak microarray signal intensities only, calling into question the limited sensitivity of the respective probe set rather than an unspecific positivity of the flow cytometric results. Importantly, differences of cells analyzed cannot account for these discrepancies because both methods were applied to the same samples of cells obtained by Ficoll-Hypaque density gradient centrifugation.

There are limited data on the comparison between mRNA and protein levels of distinct genes expressed in leukemias. Specifically, results obtained using the microarray technology were shown to correlate with the quantification of protein levels for a number of genes.16, 30 Previous data from our group demonstrated good correlations between gene expression profiling data and immunophenotyping data23; however, that analysis had focused on the signal intensities of flow cytometric measurements rather than on the percentages of positive cells, which is the clinically more convenient result. Further analyses were performed focusing on the quantification of minimal residual disease in acute leukemias.31, 32 Whereas good correlations between quantitative reverse-transcriptase polymerase chain reaction (RT-PCR) and flow cytometry were reported, this applies to the amount of leukemic cells rather than to the expression level of distinct genes, at least with regard to flow cytometry.

The results of the present report thus are in agreement with the limited data published so far in the field and give a comprehensive and solid set of data, assuring an overall high degree of correlation between measurements performed by microarray analysis and flow cytometry. Most importantly, this adds to previously published data indicating a potential role for gene expression profiling as a diagnostic tool in leukemias.2 Because the diagnosis and the subclassification of many leukemia entities is based, in combination with cytomorphology,33 cytogenetics and fluorescence in situ hybridization,34–39 and molecular genetics,40–43 on immunophenotyping today6, 44 microarray technology must be capable of accurately reproducing flow cytometric findings before its general acceptance in the diagnostic setting. The present data clearly are in favor of this concept and thereby are strongly in line with further analyses demonstrating the reproducibility of gene expression profiling with regard to the detection of biologically defined leukemia entities.10, 11, 16, 45, 46 Recently, gene expression patterns have been shown to be robust under various conditions.47 Further studies should therefore be undertaken in prospective and multicenter clinical trials to prove the capability of the microarray technology as a diagnostic tool in leukemia.48


  1. Top of page
  2. Abstract
  5. Acknowledgements

Supported by grants from the “Else Kroner-Fresenius-Stiftung” and from “Wilhelm Sander-Stiftung.” The gene expression studies were supported in part by Roche Diagnostics GmbH, Penzberg, Germany.


  1. Top of page
  2. Abstract
  5. Acknowledgements
  • 1
    Lipshutz RJ, Fodor SP, Gingeras TR, Lockhart DJ. High density synthetic oligonucleotide arrays. Nat Genet. 1999; 21: 2024.
  • 2
    Haferlach T, Kohlmann A, Kern W, Hiddemann W, Schnittger S, Schoch C. Gene expression profiling as a tool for the diagnosis of acute leukemias. Semin Hematol. 2003; 40: 281295.
  • 3
    Grimwade D, Haferlach T. Gene-expression profiling in acute myeloid leukemia. N Engl J Med. 2004; 350: 16761678.
  • 4
    Staudt LM. Molecular diagnosis of the hematologic cancers. N Engl J Med. 2003; 348: 17771785.
  • 5
    Ebert BL, Golub TR. Genomic approaches to hematologic malignancies. Blood. 2004; 104: 923932.
  • 6
    Campana D, Behm FG. Immunophenotyping of leukemia. J Immunol Methods 2000; 243: 5975.
  • 7
    Kern W, Voskova D, Schoch C, Hiddemann W, Schnittger S, Haferlach T. Determination of relapse risk based on assessment of minimal residual disease during complete remission by multiparameter flow cytometry in unselected patients with acute myeloid leukemia. Blood. 2004; 104: 30783085.
  • 8
    Rawstron AC, Kennedy B, Evans PA, et al. Quantitation of minimal disease levels in chronic lymphocytic leukemia using a sensitive flow cytometric assay improves the prediction of outcome and can be used to optimize therapy. Blood. 2001; 98: 2935.
  • 9
    Coustan-Smith E, Behm FG, Sanchez J, et al. Immunological detection of minimal residual disease in children with acute lymphoblastic leukaemia. Lancet. 1998; 351: 550554.
  • 10
    Schoch C, Kohlmann A, Schnittger S, et al. Acute myeloid leukemias with reciprocal rearrangements can be distinguished by specific gene expression profiles. Proc Natl Acad Sci U S A. 2002; 99: 1000810013.
  • 11
    Kohlmann A, Schoch C, Schnittger S, et al. Pediatric acute lymphoblastic leukemia (ALL) gene expression signatures classify an independent cohort of adult ALL patients. Leukemia. 2004; 18: 6371.
  • 12
    Kern W, Kohlmann A, Schoch C, et al. Gene expression profiling in adult acute lymphoblastic leukemia, biphenotypic acute leukemia, and acute myeloid leukemia without differentiation: confirmation of immunophenotypic and cytogenetic diagnostic findings. Blood. 2004; 104.
  • 13
    Haferlach T, Schnittger S, Kern W, Hiddemann W, Schoch C. Genetic classification of acute myeloid leukemia (AML). Ann Hematol. 2004; 83( Suppl 1): S97100.
  • 14
    Alizadeh AA, Eisen MB, Davis RE, et al. Distinct types of diffuse large B-cell lymphoma identified by gene expression profiling. Nature. 2000; 403: 503511.
  • 15
    Rosenwald A, Alizadeh AA, Widhopf G, et al. Relation of gene expression phenotype to immunoglobulin mutation genotype in B cell chronic lymphocytic leukemia. J Exp Med. 2001; 194: 16391647.
  • 16
    Yeoh EJ, Ross ME, Shurtleff SA, et al. Classification, subtype discovery, and prediction of outcome in pediatric acute lymphoblastic leukemia by gene expression profiling. Cancer Cell. 2002; 1: 133143.
  • 17
    Ross ME, Zhou X, Song G, et al. Classification of pediatric acute lymphoblastic leukemia by gene expression profiling. Blood. 2003; 102: 29512959.
  • 18
    Kohlmann A, Schoch C, Schnittger S, et al. Molecular characterization of acute leukemias by use of microarray technology. Genes Chromosomes Cancer. 2003; 37: 396405.
  • 19
    Ross ME, Mahfouz R, Onciu M, et al. Gene expression profiling of pediatric acute myelogenous leukemia. Blood. 2004; 104: 36793687.
  • 20
    Haferlach T, Kohlmann A, Schnittger S, et al. AML M3 and AML M3 variant each have a distinct gene expression signature but also share patterns different from other genetically defined AML subtypes. Genes Chromosomes Cancer. 2005; 43: 113127.
  • 21
    Haferlach T, Kohlmann A, Schnittger S, et al. A global approach to the diagnosis of leukemia using gene expression profiling. Blood. 2005; 106: 11891198.
  • 22
    Kohlmann A, Schoch C, Dugas M, et al. New insights into MLL gene rearranged acute leukemias using gene expression profiling: shared pathways, lineage commitment, and partner genes. Leukemia. 2005; 19: 953964.
  • 23
    Kern W, Kohlmann A, Wuchter C, et al. Correlation of protein expression and gene expression in acute leukemia. Cytometry. 2003; 55B: 2936.
  • 24
    Jaffe ES, Harris NL, Stein H, Vardiman JW. World Health Organization classification of tumours. Pathology and genetics of tumours of haematopoietic and lymphoid tissues. Lyon, France: IARC Press, 2001.
  • 25
    Bene MC, Castoldi G, Knapp W, et al. Proposals for the immunological classification of acute leukemias. European Group for the Immunological Characterization of Leukemias (EGIL). Leukemia. 1995; 9: 17831786.
  • 26
    Haferlach T, Schoch C. Modern techniques in leukemia diagnosis. Internist. 2002; 43: 11901202.
  • 27
    Kern W, Voskova D, Schoch C, Schnittger S, Hiddemann W, Haferlach T. Prognostic impact of early response to induction therapy as assessed by multiparameter flow cytometry in acute myeloid leukemia. Haematologica. 2004; 89: 528540.
  • 28
    Hubbell E, Liu WM, Mei R. Robust estimators for expression analysis. Bioinformatics. 2002; 18: 15851592.
  • 29
    Moenner M, Vosoghi M, Ryazantsev S, Glitz DG. Ribonuclease inhibitor protein of human erythrocytes: characterization, loss of activity in response to oxidative stress, and association with Heinz bodies. Blood Cells Mol Dis. 1998; 24: 149164.
  • 30
    Chen G, Zeng W, Miyazato A, et al. Distinctive gene expression profiles of CD34 cells from patients with myelodysplastic syndrome characterized by specific chromosomal abnormalities. Blood. 2004; 104: 42104218.
  • 31
    Malec M, van der Velden VH, Bjorklund E, et al. Analysis of minimal residual disease in childhood acute lymphoblastic leukemia: comparison between RQ-PCR analysis of Ig/TcR gene rearrangements and multicolor flow cytometric immunophenotyping. Leukemia. 2004; 18: 16301636.
  • 32
    Kern W, Schoch C, Haferlach T, Voskova D, Hiddemann W, Schnittger S. Complemental roles for multiparameter flow cytometry and quantitative RT-PCR for the quantification of minimal residual disease in patients with acute myeloid leukemia. Blood. 2003; 102.
  • 33
    Haferlach T, Schoch C, Loffler H, et al. Morphologic dysplasia in de novo acute myeloid leukemia (AML) is related to unfavorable cytogenetics but has no independent prognostic relevance under the conditions of intensive induction therapy: results of a multiparameter analysis from the German AML Cooperative Group studies. J Clin Oncol. 2003; 21: 256265.
  • 34
    Schoch C, Schnittger S, Klaus M, Kern W, Hiddemann W, Haferlach T. AML with 11q23/MLL abnormalities as defined by the WHO classification: incidence, partner chromosomes, FAB subtype, age distribution, and prognostic impact in an unselected series of 1897 cytogenetically analyzed AML cases. Blood. 2003; 102: 23952402.
  • 35
    Schoch C, Haferlach T, Kern W, et al. Occurrence of additional chromosome aberrations in chronic myeloid leukemia patients treated with imatinib mesylate. Leukemia. 2003; 17: 461463.
  • 36
    Schoch C, Kern W, Schnittger S, Hiddemann W, Haferlach T. Karyotype is an independent prognostic parameter in therapy-related acute myeloid leukemia (t-AML): an analysis of 93 patients with t-AML in comparison to 1091 patients with de novo AML. Leukemia. 2004; 18: 120125.
  • 37
    Hoelzer D. Therapy and prognostic factors in adult acute lymphoblastic leukaemia. Baillieres Clin Haematol. 1994; 7: 299320.
  • 38
    Pui CH, Relling MV, Downing JR. Acute lymphoblastic leukemia. N Engl J Med. 2004; 350: 15351548.
  • 39
    Dohner H, Stilgenbauer S, Benner A, et al. Genomic aberrations and survival in chronic lymphocytic leukemia. N Engl J Med. 2000; 343: 19101916.
  • 40
    Schnittger S, Schoch C, Dugas M, et al. Analysis of FLT3 length mutations in 1003 patients with acute myeloid leukemia: correlation to cytogenetics, FAB subtype, and prognosis in the AMLCG study and usefulness as a marker for the detection of minimal residual disease. Blood. 2002; 100: 5966.
  • 41
    Schnittger S, Weisser M, Schoch C, Hiddemann W, Haferlach T, Kern W. New score predicting for prognosis in PML-RARA+, AML1-ETO+, or CBFBMYH11+ acute myeloid leukemia based on quantification of fusion transcripts. Blood. 2003; 102: 27462755.
  • 42
    Crespo M, Bosch F, Villamor N, et al. ZAP-70 expression as a surrogate for immunoglobulin-variable-region mutations in chronic lymphocytic leukemia. N Engl J Med. 2003; 348: 17641775.
  • 43
    Frohling S, Schlenk RF, Stolze I, et al. CEBPA mutations in younger adults with acute myeloid leukemia and normal cytogenetics: prognostic relevance and analysis of cooperating mutations. J Clin Oncol. 2004; 22: 624633.
  • 44
    Kern W, Diem H, Haferlach T. Immunophenotyping in modern leukemia diagnostics. Dtsch Med Wochenschr. 2005; 130: 21529.
  • 45
    Bullinger L, Dohner K, Bair E, et al. Use of gene-expression profiling to identify prognostic subclasses in adult acute myeloid leukemia. N Engl J Med. 2004; 350: 16051616.
  • 46
    Valk PJ, Verhaak RG, Beijen MA, et al. Prognostically useful gene-expression profiles in acute myeloid leukemia. N Engl J Med. 2004; 350: 16171628.
  • 47
    Kohlmann A, Schoch C, Dugas M, et al. Pattern robustness of diagnostic gene expression signatures in leukemia. Genes Chromosomes Cancer. 2005; 42: 299307.
  • 48
    Kern W, Kohlmann A, Schnittger S, Hiddemann W, Schoch C, Haferlach T. Gene expression profiling as a diagnostic tool in acute myeloid leukemia. Am J Pharmacogenom. 2004; 4: 225237.