Clinical Laboratory Correlates and Prognosis in AML

High Bad and Bax mRNA expression correlate with negative outcome in acute myeloid leukemia (AML)


The search for molecular markers in AML that allow prediction of outcome has recently focused on genes involved in the regulation of programmed cell death (PCD). The aim of our study was to determine whether mRNA levels of Mdm-2, Bcl-2, Bcl-xL, Bad, and Bax are independent prognostic parameters for outcome. Transcript levels were analyzed by real-time quantitative RT-PCR in 232 samples collected either at diagnosis or following induction chemotherapy (ICT). Multivariate COX regression analysis adjusted for chemotherapy protocol, de novo vs secondary AML, and de novo vsrelapsed AML indicated: (1) At diagnosis, high expression of Bad (P = 0.015) and even more so high Bax and Bad levels (P = 0.018) predicted adverse outcome, regardless of the response to ICT. In patients who subsequently failed to enter complete remission (CR), high levels of Bad, Bax and Bax high/Bad high were associated with an increased relative risk (RR) to die from tumor (RR = 5.0 for Bad, 3.4 for Bax and 6.14 for Bax high/Bad high). (2) Following ICT, high expression of Bax (P = 0.005) and high Bcl-2/Bax ratios (P = 0.004) were independent predictors of unfavorable outcome, regardless of response to ICT. We conclude that high levels of Bax and Bad correlate with poor outcome, particularly in patients who do not enter CR and may serve as prognostic markers in AML.


Inspite of considerable advances in combination chemotherapy, a significant number of malignant tumors become resistant to treatment. One major cause of either intrinsic or acquired resistance is the selective activation of a number of highly effective cellular defense mechanisms which have evolutionarily developed to protect the cell from exogenous toxic compounds. The first defense barrier to a great variety of cytotoxic drugs is formed by prominent members of the ATP-binding cassette membrane transporter family, such as P-glycoprotein and the multidrug resistance-associated protein (MRP)1 which may lower intracellular drug concentrations to sublethal levels by active transport mechanisms.2 Agents which escape the primary efflux mechanisms are generally believed to effect tumor cell killing by inducing apoptosis.2 A second mechanism associated with primary resistance to chemotherapy-induced cell death may be the failure to activate the apoptotic machinery. Therefore, the search for new prognostic markers which may reflect tumor prognosis has recently focused on genes involved in the regulation of PCD. Experimental evidence has accumulated that PCD genes may be dysregulated in several types of human malignancies.3,4,5,6,7,8,9,10,11,12,13,14,15,16 Particularly, members of the expanding Bcl-2 superfamily have been extensively studied. Current opinion holds that, if proapoptotic family members (eg Bax, Bad, Bak, Bid, Bcl-xS) are in excess, cells execute a death command, but if antiapoptotic members (eg Bcl-2, Bcl-xL, Mcl-1) dominate, this program is inhibited.17,18 Death ligands such as Bad, represent an evolving set of proapoptotic molecules which appear to reside more proximal to Bcl-2 in the apoptotic pathway and connect upstream signal transduction paths with the Bcl-2 family thus modulating apoptosis.18 There is evidence that Bax is transcriptionally activated by the tumor suppressor p53 and may be a key regulator of apoptosis.18 p53 itself is controlled by murine double minute-2 (Mdm-2) which promotes inactivation of p53 by rapid degradation or by masking the p53 transactivation domain.19,20 We recently showed that low levels of Mdm-2 mRNA are associated with poor prognosis in soft tissue sarcoma.21

Several attempts have been made to correlate expression levels of resistance genes with clinical outcome of leukemia.14,22,23,24,25 Only a few studies have assessed the association between candidate gene expression and survival in acute myelogenous leukemia (AML), most of them with rather small patient numbers, and sometimes with conflicting results.9,22,26,27 Therefore, the potential prognostic value of apoptosis-related genes in AML remains to be elucidated. In this retrospective analysis, we used a semi-automated high-throughput methodology to study the expression of apoptosis-related genes (Mdm-2, Bcl-2, Bcl-xL, Bad, Bax) in 232 AML specimens and correlated expression with response to ICT and survival.

Materials and methods

Patients, samples, and treatment

Two-hundred and thirty-two peripheral blood or bone marrow (BM) aspirates were collected from adult AML patients at diagnosis or after treatment with the first or second ICT cycle (71% paired samples) according to either the low-dose (AML’93) or the intermediate-dose (AML’96) protocols of the East-German Hematology and Oncology Study Group (OSHO) (Table 1). The AML’93 ICT protocol comprised two cycles of cytarabine (ARA-C) 120 mg/m2 i.v. every 12 h (days 1–7), idarubicin (IDA) 12 mg/m2 i.v. (days 1–3), followed by a first and second consolidation therapy. The AML’96 ICT protocol consisted either of ARA-C 2 g/m2 as a 8 h infusion (days 1, 3, 5, 7) and IDA 12 mg/m2 (days 1–3) or ARA-C 1 g/m2 every 12 h as a 3 h infusion (days 1, 3, 5, 7) and IDA 12 mg/m2 (days 1–3). Complete remission was defined as 5% marrow blasts by morphology. All other patients were classified as non-responders (NR). At diagnosis, the mean percentage of blasts in samples were 65 ± 22% (median: 70%). After ICT, the NR group showed 46 ± 31% blasts (median: 48%) and the CR group 4 (1% blasts (median: 4%). Patients with acute promyelocytic leukemia (FAB-M3) and patients who died from toxicity were excluded from the analysis.

Table 1 AML study group, patient characteristics and composition of samples either obtained at the time of diagnosis or following ICT treatment

FACS analysis

CD34+ cells were counted with a FACscan flow cytometer following incubation with the phycoerythrin-labeled HPCA-2 monoclonal antibody (Becton Dickinson, San Jose, CA, USA).

Clinical sample preparation

Mononuclear cells from 5–10 ml blood (9% of samples) or BM samples (91% of samples) were collected after separation on sterile Histopaque-1077 lymphocyte separation medium (Sigma, St Louis, MO, USA), washed three times with phosphate-buffered saline, and homogenized with 2 ml of RNAzol ‘B’ (Tel-Test, Friendswood, TX, USA) or 1 ml of Reagent 14 (Integrated Separation Systems, Natick, MA, USA). Homogenates were stored at −20°C until RNA preparation was performed.

RNA isolation and reverse transcription (RT)

Total RNA was isolated manually or in an Autogen 540 nucleic acid extraction robot (Integrated Separation Systems). cDNA was synthesized from 1 μg aliquots in a 20-μl standard reaction mixture containing AMV reverse transcriptase buffer (250 mM Tris/HCl; pH 8.3, 250 mM KCl, 50 mM MgCl2, 50 mM dithiothreitol, 2.5 mM spermidine), 5 U AMV reverse transcriptase, 0.5 mM of each dNTP (Promega, Madison, WI, USA), 10 U recombinant RNase inhibitor (AGS, Heidelberg, Germany), and 200 ng oligo (dT) (Amersham Pharmacia Biotech, Uppsala, Sweden) at 42°C for 1 h. RNA and cDNA samples were stored at −80°C until use.

High-throughput transcript analysis by quantitative real-time fluorescence PCR

Mdm-2 (splice variants A, B, C, D; not E), Bcl-2, Bcl-xL, Bax (splice variants alpha, beta, delta, sigma, zeta; not gamma and epsilon), Bad, and glyceraldehyde-3-phosphate dehydrogenase GAPDH) transcripts were measured from cDNA in duplicate experiments by ready-to-use PCR testkits (Roboscreen Gesellschaft für molekulare Biotechnologie, Leipzig, Germany). Briefly, conventional 96-well bases were loaded with one eight-well ready-to-use reference-DNA strip which was storage-stable coated with eight different amounts of respective reference-DNA for quantitation of the desired transcripts, forward and reverse primer, and the TaqMan probe. The probes were either 5′-labeled with the fluorescent reporter dye 6-carboxyfluorescein (FAM) or 2,7-dimethoxy-4,5-dichloro-6-carboxyfluorescein (JOE) for GAPDH, and the common 3′-fluorescent quencher dye 6-carboxy- tetramethylrhodamine (TAMRA) in order to generate the respective standard reference curves for each run. The remaining free positions were loaded with the required number of sample tubes containing only the respective TaqMan oligonucleotide sets. Reaction premixes containing PCR buffer, the passive fluorescence dye 6-carboxy-tetramethyl-rhodamin (ROX), dNTPs, and 0.5–1 U of AmpliTaq ‘GOLD’ (Applied Biosystems, Weiterstadt, Germany) were mixed according to the manufacturer's guidelines and supplemented with H2O to a final volume of 25 μl. Aliquots of the mixes were added to each reaction tube using a BIOMEK 2000 laboratory automation workstation (Beckman Instruments, Fullerton, CA, USA). Sample reactions were completed by the addition of 2-μl aliquots of cDNA. PCR amplification and detection was performed on an ABI PRISM 7700 Sequence Detection System (Applied Biosystems). Amounts were calculated from a reference curve obtained from the simultaneously processed reference DNA strips. Mdm-2, Bcl-2, Bcl-xL, Bax, and Bad data were expressed as zeptomoles [zmol, 10−21 mole] cDNA per attomole [amol, 10−18 mole] of GAPDH cDNA which was calculated from the same cDNA sample.

Statistical analyses

Statistics were performed using the SPSS 9.0 software program (SPSS, Chicago, IL, USA). Univariate survival analysis was performed using the Kaplan–Meier method and log-rank test. A multivariate Cox regression model was used to estimate the influence of individual gene expression parameters on prognosis, adjusted for the covariates AML sub-type and chemotherapy protocol. Changes of likelihood interactions between Bax and Bad gene expression were analyzed using an interaction term.28 High Bax and Bad were classified as ‘1’, low Bax or Bad values as ‘0’. P values of less than 0.05 were considered statistically significant.


Validation of quantitative RT-PCR

The expression of the Bcl-2 family genes Bcl-2, Bcl-xL, Bad, Bax, and the p53 antagonist Mdm-2 were assessed by recently developed ready-to-use fluorescence RT-PCR assays based on the TaqMan technology. The repeatability of the Roboscreen assays was good: the coefficients of variation of the mean reference curves slopes were usually below 5% which indicates reliable standard stability.21,29 The precision of the assays was tested by analysis of cDNAs prepared from the K562 and CCRF ADR5000 cell lines for Bcl-2 and GAPDH expression, respectively. Input analyte amounts at the attomole detection level, ie typical for GAPDH reference transcripts, were detected within a 10% precision rate, whereas Bcl-2 transcripts which are usually at a low zeptomole detection level were detected within a precision of ±28% (eight repeated analyses performed for each cell line, data not shown). All values were adjusted for the number of GAPDH transcripts, which have been reported to be constant in various tumors, tissues and cell lines.30,31,32

Univariate and multivariate correlation between transcript levels and outcome at the time of diagnosis

The most important conventional prognostic parameters are shown in Table 2. Univariate Kaplan-Meier regression analysis and log-rank test showed that poor prognosis was strongly associated with the low-dose therapy protocol compared to the intermediate-dose protocol (P = 0.014) and with later NR compared to CR (P = 0.004). A trend towards unfavorable prognosis was found for relapsed and secondary AML compared to de novo AML (P = 0.097) and for unfavorable karyotypes compared with standard or favorable karyotypes (P = 0.186) confirming earlier data from the literature.33,34 There was no significant effect of age, sex, or CD34+ expression on survival.

Table 2 Univariate analysis on prognostic factors in AML at diagnosis and following ICT

To assess the influence of individual transcript amounts on prognosis, medians were calculated for each transcript separately at diagnosis and following ICT in order to group patients into high and low expressing individuals (Table 3). For this analysis, AMLs at presentation and newly diagnosed relapsed leukemias were analyzed together. Patients were divided according to their expression level of the respective transcript into high (expression above median) and low (expression below or equal to median value) expressing groups, and survival was compared between the two groups. In univariate Kaplan–Meier analysis, the most important predictors of survival at diagnosis were Bad (P = 0.026) and the combination of high Bad and high Bax levels (P = 0.014), whereas Bax alone showed only a marginal impact on survival (P = 0.095) (Table 2, Figure 1). These univariate findings remained stable when Bad and Bax expression were analyzed by a multivariate COX model adjusted for (1) chemotherapy protocol (AML 93 vs AML 96), (2) de novo vs secondary AML, and (3) de novo vs relapsed AML (Table 4). Cox analysis further revealed that Bad and combined Bax/Bad levels above the median values are independent negative prognostic parameters for survival (P = 0.015 and P = 0.018, respectively) at the time of initial diagnosis, whereas Bax alone remained borderline significant (P = 0.064) confirming the univariate analysis. Furthermore, the data indicate that AML patients characterized by Bad, Bax or Bax as well as Bad levels above median showed a 1.78-fold, 1.52-fold and 1.83-fold higher relative risk of tumor-related death compared to patients with levels equal to or below the median (Table 4, Figure 1). All other transcripts were without predictive value at diagnosis.

Table 3 Medians and ranges of apoptosis-related transcripts
Figure 1

Kaplan–Meier plots of Bax, Bad, and interacting Bax as well as Bad levels measured at diagnosis. High expression is defined as greater than the median, and low expression is defined as less than or equal to the median value. Statistical analysis performed by the log-rank test, corresponding P values indicated in the figure. Symbols: ‘+’: censored data, P and RR values labeled by asterisk are obtained from multivariate Cox regression analysis, N, number of samples.

Table 4 Gene expression and relative risk (RR) of tumor-related death

Univariate and multivariate correlation between transcript levels and outcome after ICT

After ICT (regardless of response), the initially observed correlation between survival and expression of Bad disappeared, whereas a poor prognosis was now associated with high Bax levels (P = 0.040) and high Bcl-2/Bax ratios (P = 0.044, Table 2, Figure 2). Analysis by the COX model showed that Bax expression above the median, high levels of Bax and Bad, and high Bcl-2/Bax ratios were independent prognostic factors of adverse outcome and correspond to a 2.40-fold, 2.80-fold and 2.62-fold RR, respectively, of tumor-related death (Table 4, Figure 2). Mdm-2 mRNA levels had only borderline impact on survival after treatment (P = 0.098, Figure 2).

Figure 2

Kaplan–Meier plots of Mdm-2, mdr-1, and Bax expression, and Bcl-2/Bax ratios after ICT treatment. High and low expression is defined as described previously. Statistical analysis is performed by the log-rank test (P), *data confirmed by multivariate Cox regression analysis.

Since the response to ICT is an important prognostic factor in itself, the groups of complete responders and non-responders were analyzed separately. Bax and Bad expression influenced survival only within the NR group. NR expressing high levels of Bax, Bad or high levels of Bax and Bad had a much worse prognosis than patients with low expression, both at diagnosis (P = 0.001, P = 0.01 and P = 0.002, respectively) and after ICT (P = 0.001, P = 0.045 and P = 0.004). The corresponding RR values are 4.97, 3.42 and 6.14 respectively, at the time of diagnosis and 9.62, 3.46 and 10.82, respectively, after ICT (Table 5, Figure 3). Except in one patient, failure to achieve CR together with high expression of Bax at diagnosis predicted survival of less than 12 months, whereas none of the group characterized by high Bax and Bad levels survived the first 9 months after ICT start (Figure 3). In contrast, a similar tendency but no statistical significance was observed in patients who achieved CR (Figure 3). Consequently, the strongest parameter for predicting poor prognosis in NR is the combination of increased Bax and Bad levels.

Table 5 Transcript amounts and ratios correlated with survival of patients divided into non-responders (NR) and complete responders (CR)
Figure 3

Bax, Bad and interacting Bax and Bad expression at diagnosis and response to treatment. Kaplan–Meier survival analysis and log-rank test separately performed for NR and CR AML patients expressing high or low mRNA amounts of Bax, Bad, or high Bax and Bad compared with low Bax or Bad levels, respectively. *Data confirmed by multivariate Cox regression analysis.

In patients achieving CR Bad and Bax levels were not correlated with outcome. However, patients with high Bcl-2/Bax and Bcl-xL/Bad ratios after treatment were characterized by a 2.3-fold and 2.5-fold increased RR to die from their leukemia (Table 5).


The aim of our study was to determine whether mRNA levels of the apoptosis-regulators Mdm-2, Bcl-2, Bcl-xL, Bad, and Bax could serve as independent parameters to predict clinical outcome in AML. The semi-automated high-throughput methodology used in our study allows for a much more rapid quantitation (up to 200 samples per day and method) of expression of apoptosis-related genes than would be possible by traditional techniques such as quantification at the protein level by Western blotting,9,35,36 immunohistochemistry,6,8,13,15,37 or flow cytometry.36 We show that mRNA expression of several genes has a distinct prognostic value at the time of initial diagnosis (eg Bad, Bax) as well as after completing ICT (eg Bax, Bad, Bcl-2/Bax ratio, Mdm-2).

The clearest correlation exists between high expression of the proapoptotic genes Bad and Bax and poor prognosis. This is surprising since one might expect an inverse correlation, at least at the protein level.25,28,37 Remarkably, the combination of high Bad and Bax values appear to be as strong predictive parameters as the conventional risk factors, relapse and treatment protocol, but with the considerable advantage of being prospectively available. Regarding Bax, similar observations were made in non-Hodgkin's lymphoma,15 acute lymphoblastic leukemia (ALL),38 breast,13 pancreatic4 and ovarian carcinomas.8 Our findings are also consistent with the observation that patients with small-cell lung cancer and Bax-positive/Bcl-2-negative tumors have a poor prognosis,7 and that higher ratios of Bcl-2/Bax correlate with failure to achieve CR in AML39,40,41 or chronic lymphocytic leukemia (CLL).24 In contrast, increased Bax protein expression was found to be correlated with improved in vitro survival of CLL cells,24 cervical,12 pancreatic,4 low-grade urinary bladder cancer,11 esophageal squamous cell carcinoma25 and hepatic metastases of colorectal cancer.6 It should be stated here that in a significant number of studies only the Bax-alpha isoform was characterized,3,25,42 whereas the assay used in this study detects the majority of human alternative spliced Bax mRNA isoforms (ie alpha, beta, delta, sigma, zeta) except gamma and epsilon. A recent study found no correlation to prognosis in AML at diagnosis.41

Furthermore, an important observation is that the uniform NR group may be subdivided into a high and low risk subgroup based on Bax or Bad expression at diagnosis as well as after ICT. Surprisingly, this correlation was restricted to refractory patients. It is not clear why high Bax mRNA levels at diagnosis are not predictive of poor outcome if patients achieve a CR. One possible explanation is that if ICT fails to eliminate the clone that expresses high levels of Bax and Bad, all subsequent attempts will also fail to achieve this. Alternatively, one could speculate that the physiological function of Bax in an AML cell prior to chemotherapy is different from its function after the cell has evolved under the selective pressure of cytotoxic agents. An explanation for the unexpected correlation between high Bax and Bad mRNA levels and poor prognosis may be that transcripts may not necessarily reflect protein levels. This is, however, somewhat unlikely because a significant number of studies that investigated protein levels yielded results very similar to ours.7,8,9,13,15 On the other hand, overexpression of Bax may be the result of mutations in the DNA-binding motif of p53; such mutations have recently been demonstrated in non-small-cell lung cancer.43 p53 mutations have also been reported in a significant proportion of AML44 and we are currently investigating whether such mutations are correlated with high levels of Bax. Another possible explanation is loss-of-function mutations of Bax, which have been found in 21% of human hematopoetic cell lines, most commonly in ALL subset45 and some gastrointestinal cancers.46 These mutations may affect the capability of Bax to dimerize,45 and may result in a compensatory increase of Bax expression. To our knowledge, no data with regard to Bax mutations in AML are available. Since Bax plays a role in the control of proliferation as well as apoptosis,18,47 a third explanation for the relationship between Bax overexpression and poor prognosis could be that its major function has shifted towards enhancement of cell proliferation.9,18 Because Bad is able to displace Bax from Bcl-xL48 its simultaneously high-level expression may further enhance the putative proliferative effect of Bax.

Little is known from the literature that could explain the potential physiological role of increased Bad mRNA levels and their strong correlation with poor survival in AML. Interestingly, in CLL, higher levels of the anti-apoptotic proteins Mcl-1 and Bag-1, which are also death ligands, were strongly correlated with failure to achieve CR after chemotherapy.14 The Bad protein has been shown to be rapidly inactivated by serine-phosphorylation on residues 112 and 136 followed by its sequestration in the cytosol by 14–3-3 proteins.18 This inactivation can be induced by survival-promoting cytokines such as IL-3 which lead to the activation of kinases like Akt and Raf-1 that phosphorylate Bad.18,49 Interestingly, up to 90% of AML tumor cells appear to secrete IL-1β, which can induce cells in the microenvironment to produce IL-3.50 Therefore, up-regulation of proapoptotic Bad may be an attempt to compensate for increased post-translational inactivation of the protein. In line with this concept, a recent publication demonstrated that in 40 out of 41 AML samples the Bad protein was phosphorylated on both possible residues.51 Whether this is the case in our samples is currently under investigation.

In summary, our data provide a rationale to use Bad and Bax transcript levels as independent predictive factors to assess prognosis in AML patients at diagnosis as well as after ICT. If confirmed in a prospective study, measurement of certain apoptosis-related genes could guide risk-adapted therapy to aid treatment stratification in the future.


  1. 1

    Licht T, Pastan I, Gottesman MM, Herrmann F . The multidrug-resistance gene in gene therapy of cancer and hematopoietic disorders Ann Hematol 1996 72: 184–193

  2. 2

    McKenna SL, Padua RA . Multidrug resistance in leukaemia Br J Haematol 1997 96: 659–674

  3. 3

    Bargou RC, Wagener C, Bommert K, Mapara MY, Daniel PT, Arnold W, Dietel M, Guski H, Feller A, Royer HD, Dorken B . Overexpression of the death-promoting gene bax-alpha which is downregulated in breast cancer restores sensitivity to different apoptotic stimuli and reduces tumor growth in SCID mice J Clin Invest 1996 97: 2651–2659

  4. 4

    Friess H, Lu Z, Graber HU, Zimmermann A, Adler G, Korc M, Schmid RM, Buchler MW . bax, but not bcl-2, influences the prognosis of human pancreatic cancer Gut 1998 43: 414–421

  5. 5

    Feinmesser M, Halpern M, Fenig E, Tsabari C, Hodak E, Sulkes J, Brenner B, Okon E . Expression of the apoptosis-related oncogenes bcl-2, bax, and p53 in Merkel cell carcinoma: can they predict treatment response and clinical outcome? Hum Pathol 1999 30: 1367–1372

  6. 6

    Sturm I, Kohne CH, Wolff G, Petrowsky H, Hillebrand T, Hauptmann S, Lorenz M, Dorken B, Daniel PT . Analysis of the p53/BAX pathway in colorectal cancer: low BAX is a negative prognostic factor in patients with resected liver metastases J Clin Oncol 1999 17: 1364–1374

  7. 7

    Apolinario RM, van der Valk P, de Jong JS, Deville W, van Ark-Otte J, Dingemans AM, van Mourik JC, Postmus PE, Pinedo HM, Giaccone G . Prognostic value of the expression of p53, bcl-2, and bax oncoproteins, and neovascularization in patients with radically resected non-small- cell lung cancer J Clin Oncol 1997 15: 2456–2466

  8. 8

    Marx D, Binder C, Meden H, Lenthe T, Ziemek T, Hiddemann T, Kuhn W, Schauer A . Differential expression of apoptosis associated genes bax and bcl-2 in ovarian cancer Anticancer Res 1997 17: 2233–2240

  9. 9

    Hogarth LA, Hall AG . Increased BAX expression is associated with an increased risk of relapse in childhood acute lymphocytic leukemia Blood 1999 93: 2671–2678

  10. 10

    Deng G, Lane C, Kornblau S, Goodacre A, Snell V, Andreeff M, Deisseroth AB . Ratio of bcl-xshort to bcl-xlong is different in good- and poor-prognosis subsets of acute myeloid leukemia Mol Med 1998 4: 158–164

  11. 11

    Gazzaniga P, Gradilone A, Vercillo R, Gandini O, Silvestri I, Napolitano M, Albonici L, Vincenzoni A, Gallucci M, Frati L, Agliano AM . Bcl-2/bax mRNA expression ratio as prognostic factor in low-grade urinary bladder cancer Int J Cancer 1996 69: 100–104

  12. 12

    Harima Y, Harima K, Shikata N, Oka A, Ohnishi T, Tanaka Y . Bax and Bcl-2 expressions predict response to radiotherapy in human cervical cancer J Cancer Res Clin Oncol 1998 124: 503–510

  13. 13

    Binder C, Marx D, Binder L, Schauer A, Hiddemann W . Expression of Bax in relation to Bcl-2 and other predictive parameters in breast cancer Ann Oncol 1996 7: 129–133

  14. 14

    Kitada S, Andersen J, Akar S, Zapata JM, Takayama S, Krajewski S, Wang HG, Zhang X, Bullrich F, Croce CM, Rai K, Hines J, Reed JC . Expression of apoptosis-regulating proteins in chronic lymphocytic leukemia: correlations with in vitro and in vivo chemoresponses Blood 1998 91: 3379–3389

  15. 15

    Bairey O, Zimra Y, Shaklai M, Okon E, Rabizadeh E . Bcl-2, Bcl-X, Bax, and Bak expression in short- and long-lived patients with diffuse large B-cell lymphomas Clin Cancer Res 1999 5: 2860–2866

  16. 16

    Daniel PT . Dissecting the pathways to death Leukemia 2000 14: 2035–2044

  17. 17

    Kroemer G . The proto-oncogene Bcl-2 and its role in regulating apoptosis Nat Med 1997 3: 614–620

  18. 18

    Korsmeyer SJ . BCL-2 gene family and the regulation of programmed cell death Cancer Res 1999 59: 1693s–1700s

  19. 19

    Haupt Y, Maya R, Kazaz A, Oren M . Mdm2 promotes the rapid degradation of p53 Nature 1997 387: 296–299

  20. 20

    Sigalas I, Calvert AH, Anderson JJ, Neal DE, Lunec J . Alternatively spliced mdm2 transcripts with loss of p53 binding domain sequences: transforming ability and frequent detection in human cancer Nat Med 1996 2: 912–917

  21. 21

    Taubert H, Koehler T, Meye A, Bartel F, Lautenschlager C, Borchert S, Bache M, Schmidt H, Wurl P . mdm2 mRNA level is a prognostic factor in soft tissue sarcoma Mol Med 2000 6: 50–59

  22. 22

    Tsurusawa M, Saeki K, Katano N, Fujimoto T . Bcl-2 expression and prognosis in childhood acute leukemia. Children's Cancer and Leukemia Study Group Pediatr Hematol Oncol 1998 15: 143–155

  23. 23

    Hart SM, Ganeshaguru K, Hoffbrand AV, Prentice HG, Mehta AB . Expression of the multidrug resistance-associated protein (MRP) in acute leukaemia Leukemia 1994 8: 2163–2168

  24. 24

    Aguilar-Santelises M, Rottenberg ME, Lewin N, Mellstedt H, Jondal M . Bcl-2, Bax and p53 expression in B-CLL in relation to in vitro survival and clinical progression Int J Cancer 1996 69: 114–119

  25. 25

    Sturm I, Petrowsky H, Volz R, Lorenz M, Radetzki S, Hillebrand T, Wolff G, Hauptmann S, Dorken B, Daniel PT . Analysis of p53/BAX/p16(ink4a/CDKN2) in esophageal squamous cell carcinoma: high BAX and p16(ink4a/CDKN2) identifies patients with good prognosis J Clin Oncol 2001 19: 2272–2281

  26. 26

    Campos L, Sabido O, Viallet A, Vasselon C, Guyotat D . Expression of apoptosis-controlling proteins in acute leukemia cells Leuk Lymphoma 1999 33: 499–509

  27. 27

    Campos L, Rouault JP, Sabido O, Oriol P, Roubi N, Vasselon C, Archimbaud E, Magaud JP, Guyotat D . High expression of bcl-2 protein in acute myeloid leukemia cells is associated with poor response to chemotherapy Blood 1993 81: 3091–3096

  28. 28

    Prokop A, Wieder T, Sturm I, Essmann F, Seeger K, Wuchter C, Ludwig WD, Henze G, Dorken B, Daniel PT . Relapse in childhood acute lymphoblastic leukemia is associated with a decrease of the Bax/Bcl-2 ratio and loss of spontaneous caspase-3 processing in vivo Leukemia 2000 14: 1606–1613

  29. 29

    Köhler T, Lerche D, Meye A, Weisbrich C, Wagner O . Automated analysis of nucleic acids by quantitative PCR using DNA coated ready-to-use reaction tubes J Lab Med 1999 23: 408–414

  30. 30

    Dukas K, Sarfati P, Vaysse N, Pradayrol L . Quantitation of changes in the expression of multiple genes by simultaneous polymerase chain reaction Anal Biochem 1993 215: 66–72

  31. 31

    Finnegan MC, Goepel JR, Hancock BW, Goyns MH . Investigation of the expression of housekeeping genes in non-Hodgkin's lymphoma Leuk Lymphoma 1993 10: 387–393

  32. 32

    Spanakis E . Problems related to the interpretation of autoradiographic data on gene expression using common constitutive transcripts as controls Nucleic Acids Res 1993 21: 3809–3819

  33. 33

    Kern W, Schoch C, Haferlach T, Braess J, Unterhalt M, Wormann B, Buchner T, Hiddemann W . Multivariate analysis of prognostic factors in patients with refractory and relapsed acute myeloid leukemia undergoing sequential high-dose cytosine arabinoside and mitoxantrone (S-HAM) salvage therapy: relevance of cytogenetic abnormalities Leukemia 2000 14: 226–231

  34. 34

    Chang M, Raimondi SC, Ravindranath Y, Carroll AJ, Camitta B, Gresik MV, Steuber CP, Weinstein H . Prognostic factors in children and adolescents with acute myeloid leukemia (excluding children with Down syndrome and acute promyelocytic leukemia): univariate and recursive partitioning analysis of patients treated on Pediatric Oncology Group (POG) Study 8821 Leukemia 2000 14: 1201–1207

  35. 35

    Salomons GS, Smets LA, Verwijs-Janssen M, Hart AA, Haarman EG, Kaspers GJ, Wering EV, Den Berg DD-V, Kamps WA . Bcl-2 family members in childhood acute lymphoblastic leukemia: relationships with features at presentation, in vitro and in vivo drug response and long-term clinical outcome Leukemia 1999 13: 1574–1580

  36. 36

    Nuessler V, Stotzer O, Gullis E, Pelka-Fleischer R, Pogrebniak A, Gieseler F, Wilmanns W . Bcl-2, bax and bcl-xL expression in human sensitive and resistant leukemia cell lines Leukemia 1999 13: 1864–1872

  37. 37

    Ong YL, McMullin MF, Bailie KE, Lappin TR, Jones FG, Irvine AE . High bax expression is a good prognostic indicator in acute myeloid leukaemia Br J Haematol 2000 111: 182–189

  38. 38

    Srinivas G, Kusumakumary P, Nair MK, Panicker KR, Pillai MR . Mutant p53 protein, Bcl-2/Bax ratios and apoptosis in paediatric acute lymphoblastic leukaemia J Cancer Res Clin Oncol 2000 126: 62–67

  39. 39

    Karakas T, Maurer U, Weidmann E, Miething CC, Hoelzer D, Bergmann L . High expression of bcl-2 mRNA as a determinant of poor prognosis in acute myeloid leukemia Ann Oncol 1998 9: 159–165

  40. 40

    Stoetzer OJ, Nussler V, Darsow M, Gullis E, Pelka-Fleischer R, Scheel U, Wilmanns W . Association of bcl-2, bax, bcl-xL and interleukin-1 beta-converting enzyme expression with initial response to chemotherapy in acute myeloid leukemia Leukemia 1996 10 (Suppl. 3): S18–S22

  41. 41

    Kornblau SM, Vu HT, Ruvolo P, Estrov Z, O'Brien S, Cortes J, Kantarjian H, Andreeff M, May WS . BAX and PKCalpha modulate the prognostic impact of BCL2 expression in acute myelogenous leukemia Clin Cancer Res 2000 6: 1401–1409

  42. 42

    Bargou RC, Daniel PT, Mapara MY, Bommert K, Wagener C, Kallinich B, Royer HD, Dorken B . Expression of the bcl-2 gene family in normal and malignant breast tissue: low bax-alpha expression in tumor cells correlates with resistance towards apoptosis Int J Cancer 1995 60: 854–859

  43. 43

    Huang C, Kohno N, Inufusa H, Kodama K, Taki T, Miyake M . Overexpression of bax associated with mutations in the loop-sheet-helix motif of p53 Am J Pathol 1999 155: 955–965

  44. 44

    Nakano Y, Naoe T, Kiyoi H, Kunishima S, Minami S, Miyawaki S, Asou N, Kuriyama K, Saito H, Ohno R . Poor clinical significance of p53 gene polymorphism in acute myeloid leukemia Leuk Res 2000 24: 349–352

  45. 45

    Meijerink JP, Mensink EJ, Wang K, Sedlak TW, Sloetjes AW, de Witte T, Waksman G, Korsmeyer SJ . Hematopoietic malignancies demonstrate loss-of-function mutations of BAX Blood 1998 91: 2991–2997

  46. 46

    Gil J, Yamamoto H, Zapata JM, Reed JC, Perucho M . Impairment of the proapoptotic activity of Bax by missense mutations found in gastrointestinal cancers Cancer Res 1999 59: 2034–2037

  47. 47

    Manfredini R, Capobianco ML, Trevisan F, Rauzi F, Barbieri D, Citro G, Tagliafico E, Ferrari S . Antisense inhibition of Bax mRNA increases survival of terminally differentiated HL60 cells Antisense Nucleic Acid Drug Dev 1998 8: 341–350

  48. 48

    Yang E, Zha J, Jockel J, Boise LH, Thompson CB, Korsmeyer SJ . Bad, a heterodimeric partner for Bcl-XL and Bcl-2, displaces Bax and promotes cell death Cell 1995 80: 285–291

  49. 49

    Majewski M, Nieborowska-Skorska M, Salomoni P, Slupianek A, Reiss K, Trotta R, Calabretta B, Skorski T . Activation of mitochondrial Raf-1 is involved in the antiapoptotic effects of Akt Cancer Res 1999 59: 2815–2819

  50. 50

    Preisler HD, Bi S, Venugopal P, Raza A . Cytokines, molecular biological abnormalities, and acute myelogenous leukemia Leuk Res 1997 21: 299–312

  51. 51

    Andreeff M, Jiang S, Zhang X, Konopleva M, Estrov Z, Snell VE, Xie Z, Okcu MF, Sanchez-Williams G, Dong J, Estey EH, Champlin RC, Kornblau SM, Reed JC, Zhao S . Expression of Bcl-2-related genes in normal and AML progenitors: changes induced by chemotherapy and retinoic acid Leukemia 1999 13: 1881–1892

  52. 52

    Hoelzer D, Seipelt G . Acute leukemias and myelodysplastic syndromes. The role of haemopoietic growth factors. Consultant Series No. 6. Gardiner-Caldwell Communications Ltd 1994

Download references


The authors gratefully acknowledge the dedicated assistance of all involved doctors, nurses, medical and research assistants, data managers, haematopathologists and cytogeneticists of all hospitals involved in the study. This work was supported by the University of Leipzig and the OSHO study group.

Author information

Correspondence to T Köhler.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Köhler, T., Schill, C., Deininger, M. et al. High Bad and Bax mRNA expression correlate with negative outcome in acute myeloid leukemia (AML). Leukemia 16, 22–29 (2002).

Download citation


  • acute myeloid leukemia
  • prognosis
  • gene expression
  • quantitative
  • Bax
  • Bad

Further reading