Mutations in ras proto-oncogenes are associated with lower mdr1 gene expression in adult acute myeloid leukaemia

Authors


Markus Schaich, M.D., Department of Medicine I, University Hospital, Fetscherstr. 74, 01307 Dresden, Germany. E-mail: schaich@oncocenter.de

Abstract

Mutations in ras genes have been found to be the most frequent genetic aberrations in adult myeloid leukaemia (AML). Some reports have shown an improved outcome of ras-mutated AML. In order to understand the biology of ras mutation in AML, we studied a cohort of patients treated in a prospective multicentre trial for ras mutational status and resistance gene expression. Blast samples from 162 adult patients with de novo or secondary AML were examined for resistance gene expression (mdr1, mrp1 and lrp) and ras mutations using reverse transcription-polymerase chain reaction and protein nucleic acid-competitive polymerase chain reaction strategies respectively. Ras mutations were confirmed using DNA sequencing. Ras mutations leading to an exchange of amino acids were found in 40 (25%) patients. Thirty AML patients had N-ras mutations and nine patients had K-ras mutations. One patient showed both N-ras and K-ras mutations. Resistance gene expression was positive for mdr1 in 30%, for mrp1 in 43% and for lrp in 62% of patients. There was a strong inverse correlation between the presence of ras mutation and mdr1 expression (P = 0·005). However, no significant difference was seen between patients with or without ras mutations and mrp1 or lrp expression. Whereas mdr1 expression was associated with a lower complete remission rate (P < 0·04), ras mutations had no significant influence on remission status. Neither ras mutation nor mdr1 expression had a significant impact on overall or disease-free survival to date. For the first time, there is evidence that activated ras genes are associated with lower mdr1 expression in AML.

The expression of resistance genes, like mdr1, mrp, or lrp has been associated with the occurrence and overexpression of a multidrug resistance phenotype (Bradshaw & Arceci, 1998). Both the mdr1 gene product, P-glycoprotein, and multidrug resistance-associated protein (MRP) function as transmembrane pumps that actively extrude cytotoxic drugs out of the cells (Endicott & Ling, 1989; Grant et al, 1994; Schaich et al, 1997). The function of lung resistance-related protein (LRP) is not completely elucidated, but it is thought to be involved in the transport of cytotoxic drugs between the nucleus and the cytoplasm of the cells (Schuurhuis et al, 1991). Overexpression of mdr1 is frequently seen in de novo and relapsed acute myeloid leukaemia (AML) patients (Beck et al, 1996) and has been shown to correlate with significantly reduced remission rates in AML patients (Leith et al, 1999). Reports evaluating the prognostic significance of mrp1 or lrp revealed conflicting data (List et al, 1996; Leith et al, 1999).

Recently, Smeets et al (1999) reported a downregulation of mdr1 expression and function if they induced proliferation in leukaemic blasts of AML patients using a cytokine cocktail.

One potential transmitter of proliferation signals is the ras pathway. In the GTP-bound state, ras proteins activate signal transduction cascades including Raf and mitogen activated protein (MAP) kinase (Prendergast & Gibbs, 1993; Denhardt, 1996). This pathway leads to the activation of downstream transcription factors and proto-oncogenes, such as NF-IL6, Elk-1, c-Jun and c-Myc, that regulate the expression of different effector genes and play a major role in cell cycle and cellular proliferation (Prendergast & Gibbs, 1993; Matsuda et al, 1994). Furthermore, some autocrine growth factors in haematological malignancies, such as interleukin-1 beta (IL-1β) or IL-6, can be induced by transfection of activated ras (Demetri et al, 1990). In the presence of a mutation, however, the intrinsic ability of ras to hydrolyse GTP is impaired, which leads to a constitutive signalling (Lowy & Willumsen, 1993). Moreover, Darley & Burnett (1999) showed that mutant ras maintained proliferation and blocked differentiation in the transfected multipotent haematopoetic cell line FDCP-mix in the presence of granulocyte-macrophage colony-stimulating factor (GM-CSF).

Hot spots for ras mutations are usually found in the codons 12, 13 and 61 (Lowy & Willumsen, 1993).

Ras mutations are detected relatively often in AML, affecting the subfamily members K-ras and N-ras. The influence of ras mutations on treatment outcome in AML, however, is controversial (Radich et al, 1990; Coghlan et al, 1994; Neubauer et al, 1994; De Melo et al, 1997; Kiyoi et al, 1999).

Therefore, in this study we examined a series of AML patients for resistance gene expression (mdr1, mrp, lrp) and ras mutations to evaluate the influence of mutational activated ras on resistance gene expression. This led to a possible explanation of the conflicting results regarding ras mutations and treatment outcome in AML.

PATIENTS and METHODS

Patients One hundred and sixty-two patients with de novo or secondary AML with a mean age of 52 years (range, 18–77 years) were studied. Patients diagnosed with the subtype FAB-M3 were not included in the study. The patients were uniformly treated according to the German multicentre protocol of the Süddeutsche Hämoblastosegruppe (SHG). Patients over 60 years of age received two induction therapies containing daunorubicin 45 mg/m2 (d 3–5) and cytosine arabinoside (ara-C) 100 mg/m2 (d 1–7). Individuals ≤ 60 years were treated with double induction therapy and priority-based post-remission therapy in different cytogenetic risk groups, as shown in Fig 1.

Figure 1.

Therapy protocol of the German SHG AML 96 study for AML patients younger than 60 years of age. Risk stratification was carried out by cytogenetics. Allo, allogeneic; auto, autologous; I-MAC, post-remission therapy with intermediate dose Ara-C; H-MAC, post-remission therapy with high-dose Ara-C; VP16, etoposide; m-Amsa, amsacrine; KI, short-time infusion.

Complete remission (CR) was defined as the presence of < 5% of blast cells in a standardized bone marrow puncture after the second course of induction therapy. Only patients with fully regenerated blood counts were considered to be in CR.

The control group for drug resistance gene expression consisted of 20 healthy bone marrow donors. The study was approved by the ethics committees of the Universities of Dresden and Marburg. Each patient gave written informed consent.

Sample handling Bone marrow or peripheral blood samples were taken at diagnosis. Samples were divided for routine analysis, cytogenetics and immunocytochemical analysis. One part of the sample was frozen in liquid nitrogen. At the time of RNA or DNA extraction, samples were thawed according to routine protocols.

mRNA and DNA extraction/c-DNA synthesis Probes containing less than 80% of blasts were referred to CD-3 depletion, as T cells may perturb resistance gene expression analysis. We performed depletion using CD-3-coated dynabeads (Dynal, Hamburg, Germany) according to the manufacturer's recommendations. CD-3-positive cells could be eliminated with a sensitivity of 98% (data not shown). RNA extraction and cDNA synthesis was carried out as described previously (Schmidt et al, 1998).

DNA-extraction was performed using standard phenol/chloroform extraction or using the Qiagen DNA-extraction kit (Qiagen, Hilden, Germany).

Reverse transcription-polymerase chain reaction (RT-PCR) for resistance gene expression PCR was performed in a final volume of 50 μl containing 1× reaction buffer (Perkin Elmer-Applied Biosystems, Weiterstadt, Germany), 2·0 mmol/l MgCl2, 15 pmol of each primer, 200 μmol/l of each dNTP (Pharmacia, Freiburg, Germany) and 1·5 U of AmpliTaq Gold-Polymerase (Perkin Elmer-Applied Biosystems). GAPDH primers were from Clontech (Clontech, Heidelberg, Germany), mrp primers and mdr1 oligonucleotides were used as previously described (Beck et al, 1995). Lrp primers were deduced from a previously published sequence (Scheffer et al, 1995) (Gen Bank accession number X79882). Lrp primers spanning position 479–962 of the published sequence were as follows: upper primer, 5′-CGC TGC TTG ATT TTG AGG AT; lower primer, 5′-CGA GAA TCA CGC AGT AGT TG. Lrp PCR conditions included 30 cycles of 96°C denaturation (30 s), 60°C annealing (45 s) and 72°C extension (60 s). All primer pairs were tested in cycle kinetic analysis to ensure amplification in the exponential range of PCR (data not shown). All PCR reactions were run at least twice.

GAPDH, mdr-1, lrp and mrp1 PCR reaction products were ethanol-precipitated and subsequently loaded onto a 5% polyacrylamide gel. After electrophoresis, the gel was ethidium bromide-stained and evaluated using the BioDoc II video documentation system (Biometra, Göttingen, Germany). Densitometrical analysis was carried out using the ScanPack 3·0 software (Biometra). Relative amounts of resistance gene expression were determined by division with the observed GAPDH value. Accuracy of PCR amplification was controlled using published positive reference cell lines: CCRF-VCR100 for mdr1 and mrp (Beck et al, 1995), and HT 29 for lrp (Izquierdo et al, 1996).

To assess overexpression of each investigated resistance gene, we compared the relative mdr1, lrp, and mrp1 patient values with the mean values of a control group of healthy bone marrow donors. The probes of the control group were T cell-depleted, handled exactly as the probes of the study population and consisted of previously taken bone marrow aspirates.

Protein nucleic acid (PNA)-competitive PCR and DNA sequencing for analysis of ras mutations PNA-mediated PCR clamping was carried out according to our previously published method (Thiede et al, 1996), in which the conditions and primer sequences of the PNA-PCR for detection of K-ras mutations at codon 12,13 were described. As a control gene we used the human growth hormone (HGH) gene.

Conditions of PNA-PCR for detection of N-ras mutations at codon 12,13 were as follows: primer and PNA sequences: NRPNA-1 5′-gTA CTg TAg ATg Tgg CTC gCC A, NRPNA-2 5′-ATT gTC AgT gCg CTT TTC CCA AC, PNA-N1 5′-AAC ACC ACC TgC TCC, HGHs 5′-gCC TTC CCA ACC ATT CCC TTA and HGHas 5′-TCA Cgg ATT TCT gTT gTg TTT C. Primer and PNA concentrations: NRPNA-1 and NRPNA-2 0·2 μmol/l, HGHs and HGHas 0·1 μmol/l, PNA-N1 2·5 μmol/l, Taq Polymerase 0·625 Unit/reaction (Perkin Elmer), 200 μmol/l dNTPs, 10 mmol/l Tris-HCl, pH 8·3, 50 mmol/l KCl, 1·5 mmol/l MgCl2, 0·001% (w/v) gelatine and 5% glycerol. PCR was performed using a Perkin Elmer 9600 (Norwalk, CT, USA) thermal cycler under the following conditions: soak file, 94°C 2 min 30 s; cycles 1–27, 94°C 30 s, 70°C 50 s, 58°C 50 s, 72°C 60 s; cycle 28, 94°C 30 s, 60°C 5 min.

Primer and PNA sequences for N-ras gene Codon 61 PNA-PCR: NRPNA-11 5′-CCA gAT Agg CAg AAA Tgg gCT Tg, NRPNA-12 5′-ggT CTC TCA Tgg CAC TgT ACT CT and PNA-N2 5′-CTT CTT gTC CAg CTg. Primer concentrations were 0·5 μmol/l for NRPNA-11 and NRPNA-12, 0·0625 μmol/l for HGHs and HGHas, PNA concentration was 6·0 μmol/l and Taq polymerase 1·0 unit/reaction. The annealing temperature was 56°C and 32 cycles were performed.

The PCR products were separated on a 3% agarose gel stained with ethidium bromide.

In order to confirm and differentiate ras mutations, all PNA-PCR-positive samples were analysed by cycle-sequencing with an ABI Prism DNA-sequencer 377. The ABI Prism BigDye Terminator Cycle Sequencing Ready Reaction kit was used. The PNA-PCR product was reamplified with a semi-nested PCR procedure using a Perkin Elmer 9600 (Norwalk) thermal cycler under the following conditions: soak file, 94°C 2 min 30 s; cycles 1–29, 94°C 30 s, 58°C 30 s, 72°C 60 s; cycle 30, 94°C 30 s, 60°C 5 min. Primer sequences for the semi-nested PCR were as follows: N-ras codon 12,13, 5′-ggC TCg CCA ATT AAC CCT gA-3′, N-ras codon 61, 5′-gCA ATA gCA TTg CAT TCC CTg Tg-3′ and Ki-ras codon 12,13, 5′-gTg TgA CAT gTT CTA ATA TAg TC-3′. The primer concentration was 0·2 μmol/l for each primer, Taq Polymerase 0·6125 Units/reaction (Perkin Elmer), 200 μmol/l dNTPs, 10 mmol/l Tris-HCl, pH 8·3, 50 mmol/l KCl, 1·5 mmol/l MgCl2 and 0·001% (w/v) gelatine. The PCR products were purified by Sephadex G50 (Pharmacia) columns, added to a prepared sequencing mix, loaded on a 5% Long Ranger gel (6 mol/l urea) and sequenced with an ABI Prism 377 DNA Sequencer.

Flow cytometry For discrimination of CD34-positive cells, the monoclonal antibody QBEnd1D (Coulter-Immunotech Diagnostics, Hamburg, Germany) was used according to previously published protocols (Gramatzki et al, 1998).

Cytogenetics Chromosome analysis was performed on metaphases from direct preparations, as well as from 24 h and 48 h cultures of bone marrow and/or peripheral blood samples. The cytogenetic preparation and G-banding were carried out according to routine laboratory procedures. Cytogenetic risk groups were defined as follows: high risk: −5/del(5q), −7/del(7q), hypodiploid karyotypes (besides 45,X,–Y or –X), inv(3q), abnl12p, abnl11q, +11, +13, +21, +22, t(6;9); t(9;22); t(9;11); t(3;3), multiple aberrations; intermediate risk: patients without low-risk or high-risk constellation; low risk: t(8;21) and t(8;21) combined with other aberrations. Cytogenetic stratification was performed only in patients ≤ 60 years of age.

Statistical analysis Statistical analysis was carried out using a SPSS software package. Univariate P-values were calculated using a double-sided Fisher's exact test or using the Mann–Whitney U-test for non-parametric results. Multivariate analysis was performed by stepwise logistic regression analysis.

Results

Frequencies of ras mutations

One hundred and sixty-two adult patients with de novo (n = 133) or secondary (n = 29) AML were examined for resistance gene expression (mdr1, mrp and lrp) and N-ras or K-ras mutations. Fifty-six patients were older than 60 years and 106 patients were ≤ 60 years of age respectively.

Ras mutations leading to an exchange of amino acids, confirmed by DNA sequencing, were found in 40 (25%) of the examined patients. The proportion of ras mutations were 34% in secondary compared with 23% in de novo AML patients (see Table I).

Table I.  Correlations of ras mutations in the 162 examined AML patients with disease status, age, FAB subtypes, cytogenetics, CD34 expression and resistance gene expression on mRNA level obtained by RT-PCR.
  Number of patients with ras mutationsNumber of patients without ras mutations
  1. Statistical analysis was carried out using the two-tailed Fishers exact test. mdr1, multi drug resistance gene; mrp1, multi drug resistance-related protein; lrp, lung resistance protein.

StatusDe novo30103
Secondary10 19 
Age≤ 60 years27 79
> 60 years13 43 
FAB subtypeM214 46
Other than M226 76 
Cytogeneticshigh risk 9 30
Other than high risk31 92 
CD34Positive22 64
Negative16 44 
Mdr1Positive 5 44
Negative35 78 
Mrp1Positive14 56
Negative26 66 
LrpPositive26 74
Negative14 48 

N-ras was mutated in 30 patients with codon 12 mutations in 16 patients, codon 13 mutations in eight patients and codon 61 mutations in five patients. A simultaneous N-ras mutation in codon 12 and 61 was found in one patient (see Table II).

Table II.  Clinical characteristics, WBC count, CD34 expression, cytogenetics, resistance gene expression and treatment response of the 40 adult AML patients with ras mutations leading to an exchange of amino acids.
PatientAge at diagnosis
(years)/Sex
DiagnosisWBCCD34KaryotypeRas mut.Mdr1Mrp1LrpTreatment
response
  1. s, secondary AML; WBC, white blood cell count (× 109/l) at diagnosis; CD34, CD34-positive cells in percentage; ras mut., ras mutational status; N, N-ras; K, K-ras; 12,13,61, mutated codons; –, negative;+, positive, CR, complete remission according to CALGB criteria; PR, partial response; NR, no response; n.d., not done.

 124/Females M2  419NormalK 13NR
 254/MaleM42242NormalN 12CR
 364/FemaleM2  579t(8;21), del (9q)–XN 13++CR
 456/Males M4 2058− 7N 12+++NR
 564/MaleM41988+ 8N 61+NR
 658/Females M1 2391NormalN 61+NR
 758/Females M4 4472NormalN 12NR
 874/MaleM5b 531NormalK 12+CR
 938/FemaleM4Eo 1360inv(16),abnl11q,
t(1;11), + 8,+19
N12 + 61CR
1030/FemaleM2  565−7K12++CR
1130/MaleM112067del(16)N12+NR
1248/FemaleM5b 2313NormalN12+CR
1360/Females M5b 987NormalN13+CR
1465/Males M2 5466NormalN61++CR
1557/FemaleM2  420NormalN13CR
1650/Males M5a2142NormalN12++NR
1757/FemaleM2  2n.d.NormalN13++CR
1833/MaleM5a 4753NormalN12++CR
1946/FemaleM213092del(9q)N13+CR
2061/MaleM4 112NormalN12,K12+NR
2147/FemaleM4Eo 8235inv(16)N61CR
2239/Females M2  73NormalN12+PR
2343/Females M5a 2043t(8;11),del(9q)N12++NR
2430/FemaleM4 7060t(6;11),abnl11qN12++CR
2567/MaleM21016NormalK12+CR
2660/MaleM1 6334NormalN12CR
2769/FemaleM4 722t(11;17), +19K12+NR
2857/FemaleM5b 837NormalK13++CR
2928/FemaleM4 447NormalN12CR
3071/FemaleM5a  811NormalK13+NR
3163/FemaleM2  1n.d.+ 8N12PR
3228/FemaleM0 4226t(9;11),abnl11q,del(1)N12+CR
3360/MaleM2 667NormalK13+PR
3454/MaleM5b  34NormalN13CR
3562/FemaleM2 1574NormalN13+NR
3668/FemaleM4 1148abnl12pN12+PR
3777/MaleM7 1635+8K12++NR
3863/Males M4 3043t(12;17),–7N12+NR
3932/FemaleM2 3132t(8;21),del(9q),add(2q),–XN61+CR
4039/FemaleM2 1345t(8;21),–XN13+CR

K-ras mutations were found in nine AML patients. Five patients had a mutation in codon 12, four patients in codon 13. Finally, one patient was found to be positive for both N-ras (codon 12) and K-ras (codon 12) mutations (see Table II).

Mutational activated ras did not correlate with age, FAB subtype, cytogenetics and CD34 expression (see Table I). Furthermore, white blood cell counts or bone marrow blast percentages were not significantly different in ras-mutated patients compared with their non-mutated counterparts (data not shown).

Resistance gene expression

Mdr1 mRNA was expressed in 49 (30%) out of 162 AML patients. Age correlated with mdr1 expression. Forty-two percent of patients older than 60 years were mdr1 positive compared with 24% of younger patients (P = 0·03). Almost all mdr1-positive patients were CD34 positive (P < 0·0001) (data not shown).

Mrp1 or lrp gene expression was found in 70 (43%) or 100 (62%) of the examined 162 AML patients respectively. A representative gel electrophoresis of resistance gene RT-PCR products of AML patients is shown in Fig 2.

Figure 2.

Polyacrylamide gel electrophoresis of gapdh, lrp, mrp1 and mdr1 RT-PCR reaction products. Lane 1 shows a 100-bp molecular weight marker, lane 2 a negative control, lane 3 the cell line CCRF VCR100 serving as a positive control for mrp1 and mdr1, and lane 4 the cell line HT29 serving as a positive control for lrp. Lanes 5–9 display RT-PCR products of five AML patients at diagnosis.

Activated ras and resistance gene expression

We found a strong inverse correlation between mutated ras genes and mdr1 mRNA expression in the examined AML patients. Five out of the 40 AML patients with ras mutations were positive for mdr1 expression (P = 0·005) compared with 44 out of 122 patients without a ras mutation (see Table I). Three of those mdr1-positive patients had a N-ras mutation, two in codon 12 and one in codon 13. Two showed K-ras mutations in codon 12. Mdr1 expression was completely absent in patients with N-ras mutations of codon 61 or K-ras mutations of codon 13 (see Table II).

There was no significant difference between patients with or without ras mutations and mrp1 or lrp expression (see Table I).

Prognostic impact

Fifty-eight percent of all AML patients reached CR criteria after induction therapy. The CR rate was not significantly different between patients with ras mutations and patients without (55% vs. 59%). This was also true for N-ras or K-ras mutations analysed separately (see Table III).

Table III.  Remission rates of all AML patients (n = 162) dependent on ras mutation or resistance gene mRNA expression status.
Variable% in CR 
  1. Statistical analysis was carried out using the two-tailed Fishers exact test. The % in CR reflects the complete remission rate in positive vs. negative cases.

All58 
Ras mutation (N-or K-ras)55 vs. 59P = 0·71
K-ras mutation40 vs. 59P = 0·32
N-ras mutation58 vs. 58P = 1·00
Mdr145 vs. 64P < 0·04
Mrp155 vs. 60P = 0·63
Lrp57 vs. 59P = 0·87

However, mdr1 expression was correlated with a lower CR rate of 45% in all AML patients compared with 64% in their mdr1-negative counterparts (P < 0·04) (see Table III).

Three of the five mdr1-positive AML patients with ras mutations were non-responders, whereas the other two patients achieved CR (see Table II).

In the multivariate analysis, an age of over 60 years and high-risk karyotype were the strongest predictors for treatment failure (data not shown).

To date, neither ras mutations nor mdr1 had shown a significant impact on overall survival (OS) or disease-free survival (DFS) in the examined AML patients, with a median observation time of 12·6 months (0·1–38·3 months) (data not shown).

Discussion

In this study, we analysed the relationship between mutational activated ras, clinical aspects, morphology, cytogenetics and resistance gene expression in adult AML. To determine overexpression of resistance genes, we used an RT-PCR assay because we found that mRNA resistance gene expression is more sensitive than protein- or efflux-based methods (Illmer et al, 1999).

We demonstrated an inverse correlation between mdr1 mRNA expression and ras mutations. Only five patients with ras mutations had detectable mdr1 mRNA levels. Moreover, these mdr1 expression levels were quantitatively estimated using a previously published PCR MIMIC strategy (Illmer et al, 1999), and mdr1 levels of the patients with ras mutations and mdr1 expression were lower than the levels found in patients without ras mutations (data not shown).

Although there are alternative mechanisms of ras activation, mutations in the ras genes are the most common activating events in AML patients.

Mutational activated p21ras causes a more proliferative state of leukaemic cells by itself or by altering genes that associate with ras signalling pathways (Shannon, 1995). This enhanced proliferation capacity might explain the decreased mdr1 expression in patients with ras activating mutations. Smeets et al (1999) showed that normal non-cycling haematopoetic progenitors of volunteers and the leukaemic blasts of 10 AML patients exhibited an increased anthracyline retention and toxicity after triggering proliferation by cytokines in those cells. This was as a result of downregulation of the mdr1 phenotype measured using the rhodamine efflux assay. Transfection of the multipotent haematological cell line FDCP-mix with mutated ras resulted in the inhibition of terminal neutrophil differentiation and the promotion of continued proliferation of metamyelocytes in the presence of GM-CSF (Darley & Burnett, 1999). As a high percentage of the clonogenic leukaemic myeloblasts are non- or slowly proliferating cells (Raza et al, 1987; Butturini et al, 1990), there is a reservoir of blast cells in mdr1-positive and mdr1-negative patients that might be susceptible to proliferation induction.

However, there are conflicting results about mdr1 expression during different cell cycle phases. Tarasiuk et al (1993) found no differences in mdr-dependent efflux during different cell cycle phases and, finally, Ramachandran et al, (1995) revealed a higher mdr1 expression in S-phase cells than G1 and G2+ M-phase cells. Notably, these results were achieved with multidrug-resistant cell lines.

Moreover, the promoter of the human mdr1 gene was shown to be influenced by the H-ras oncogene, which implies that the mdr1 gene could be activated during tumour progression associated with mutations in H-Ras (Chin et al, 1992). However, H-ras mutations do not play a major role in AML (Neubauer et al, 1994).

Very little is known about the dependency of mrp1 or lrp expression on the proliferative state or cell cycle. No correlation between proliferation and lrp expression was detectable in non-small cell lung carcinomas (Volm et al, 1997), whereas Bader et al (1999) found a higher proliferation rate associated with enhanced mrp1 expression in neuroblastoma cells. No data about mrp1 or lrp expression and proliferation in leukaemia currently exists. Mrp and lrp did not correlate with ras mutational status in AML patients in our analysis.

Mdr1 expression revealed AML patients with a lower remission rate than the mdr1-negative group. This is in accordance to previously published data (Leith et al, 1999).

The implication of ras mutations on treatment outcome in AML is more controversial. In previously published analyses of Neubauer et al (1994) with 99 patients and Coghlan et al (1994) with 219 patients, a better OS in young AML patients with mutations of K-ras and N-ras was found, whereas De Melo et al (1997) postulated a shortened OS in eight ras-positive Brazilian AML patients compared with 32 ras-negative patients. Radich et al (1990) found no influence of ras mutations on CR, OS and DFS in a series of 55 AML patients. More recently, Kiyoi et al (1999) screened AML patients for FLT3 ligand mutations and N-ras mutations. There was a trend for a lower CR rate in patients with N-ras mutations but, in contrast to FLT3 mutations, N-ras mutation as a single factor had no influence on survival.

In the present study, a large number of 162 AML patients was analysed and no impact of ras mutations on CR rate, OS or DFS was detected. This might be because of different factors affecting chemotherapy response in AML patients. A recent report (Koo et al, 1999) found an enhanced sensitivity to anthracylines in human leukaemic cell lines with activated ras. This finding supports a better treatment outcome in ras-positive patients, as shown in an earlier survey (Neubauer et al, 1994) in which AML patients received a daunorubicin-containing induction therapy compared with a mitoxantrone-containing induction therapy for the majority of patients, i.e. patients younger than 60 years, in the present study. As mdr1-positive cells are more resistant to daunorubicin than mitoxantrone (Koo et al, 1999), the positive effect of ras mutations on outcome in the study of Neubauer et al (1994) could have been partly as a result of a lower mdr1 expression and, therefore, an increased sensitivity to daunorubicin in the ras-mutated cases. So, owing to the use of mitoxantrone in the present paper, the positive effect of activated ras on cell killing may have been diminished.

In conclusion, we demonstrated an inverse correlation between ras mutations and mdr1 expression in AML, whereas no correlation was found between lrp or mrp1 expression and ras mutations. Therefore, some of the effects of ras mutations on outcome may have been affected by lower mdr1 expression in ras-mutated patients and may depend on the drugs used within the induction regimens.

Moreover, in future prospective analyses, proliferation markers should be correlated to ras mutational status and resistance gene expression in AML patients. This would give more rationales for the use of haematopoetic growth factors together with mdr1 modulators within induction regimens, as proposed by Smeets et al (1999), to improve the treatment outcome in mdr1-positive AML patients.

Acknowledgments

The expert technical assistance of E. Harbich-Brutscher, C. Rabolt and B. Ziegs is highly appreciated. Furthermore, we thank S. Soucek and S. Freund for statistical support. The CD34 data were kindly provided by M. Gramatzki (Erlangen). Parts of the cytogenetic analyses were carried out by U. Pascheberg (Dortmund). The work was partly supported by grants from the Deutsche Krebshilfe to G.E., and the Deutsche Forschungsgemeinschaft (Ne 310/6–3) and the Wilhelm Sander Stiftung to A.N.

Finally, our thanks go to all the members and clinicians of the German SHG AML 96 study group who entered their patients into the trial.

Ancillary