patients with acute myeloblastic leukaemia (AML).1â3 However, as reported in a recent paper in Leukemia, pgp expression in AML blast cells has been shown ...
Correspondence
1468
P-glycoprotein expression is associated with resistance to spontaneous in vitro apoptosis in AML TO THE EDITOR Expression of p-glycoprotein (pgp), the product of the MDR1 gene, is associated with impaired uptake of daunorubicin by blasts from patients with acute myeloblastic leukaemia (AML).1–3 However, as reported in a recent paper in Leukemia, pgp expression in AML blast cells has been shown to be unrelated to in vitro daunorubicin chemoresistance.4 We would like to offer a possible explanation for this apparently puzzling finding. P-glycoprotein has hitherto been thought of as primarily a drug efflux molecule. However, the ability of this versatile protein to inhibit the apoptosis induced by serum withdrawal, first demonstrated in MDR1-transfected Chinese hamster ovary fibroblasts,5 has now been found in blasts from AML patients: we have shown an active role for pgp in enhancing cell viability by showing that use of the pgp blocking antibody UIC2, as well as the agent PSC 833, increases the in vitro apoptosis of pgp-positive AML blast samples following short-term culture in the absence of drugs (Ref. 6 and manuscript submitted for publication). Others have also demonstrated a role for pgp in resisting fas-dependent, drug-independent apoptosis.7 Because of the drug-efflux independent role for pgp in promoting leukaemic cell survival, we hypothesised that in vitro chemoresistance in pgp+ leukaemic blasts may have two components: resistance to apoptosis induced by growth and survival factor withdrawal under standard suspension culture conditions, as well as resistance due to drug efflux. We have taken the first step towards validating this hypothesis by comparing the viability of pgp− and pgp+ blasts from patients with untreated AML after 18 h culture in a medium consisting of RPMI 1640 with 1% FCS, 2 mm l-glutamine, 100 U/ml penicillin and 100 g/ml streptomycin. Fifteen cryopreserved samples from patients with AML were thawed and processed in batches of one pgp-positive sample to one or two pgp-negative samples to control for inter-assay variation. Pgp expression had previously been determined using flow cytometry according to consensus recommendations8 with Kolmogoroff Smirnoff analysis of cells stained with the MRK-16 antibody as described.3 Thawed cells were rested for 90 min in medium with 20% FCS, without heparin, to allow dead cells to clump. Following this procedure all samples were greater than 90% viable. The cells were then pelleted, rinsed in RPMI and cultured at 5 × 105/ml in sterile falcon tubes in medium with 1% FCS at 37°C overnight in a humidified 5% CO2 incubator. At 18 h 12.5 g/ml 7-aminoactinomycin D was added and the cells were assayed flow cytometrically for loss of viability, measured as the proportion of 7AAD dim/high, low forward scatter events.9 We found that the loss of viability in pgp-positive samples (median 26.7%) was significantly lower than that of their pgpnegative counterparts (median 51.9%) (P ⬍ 0.03) (see Figure 1). These findings are clearly of relevance to in vitro chemosensitivity assays in AML. One would expect that a pgp-positive blast treated with a pgp substrate drug would be relatively chemoresistant. However, one might now also expect that following in vitro culture, an untreated pgp-positive blast is more likely to survive than a pgp-negative blast. Therefore, it may be the absolute number of drug treated cells surviving at the end of the in vitro culture period, rather than the ratio of drug treated to untreated cells, which provides a sensitive measure of pgp-mediated resistance. The finding of Norgaard and colleagues10 that lack of spontaneous apoptosis in vitro was an indicator of poor prognosis emphasises the clinical relevance of this component of blast cell resistance to apoptosis. In conclusion, we have found evidence for a mechanistic, pgprelated, interpretation of spontaneous in vitro blast cell apoptosis in
Correspondence: M Pallis, David Evans Medical Research Centre, Nottingham City Hospital, Nottingham NG5 1PB, UK; Fax: 0115 985 8864 Received 13 April 1999; accepted 20 May 1999
Figure 1 Percentage loss of viability in AML samples after 18 h suspension culture in medium containing 1% FCS.
AML on the basis of which we would suggest to other groups engaged in chemoresistance research that they incorporate a standardised outcome measure into their assays in order to be able to compare treated blast cell viability between patients, as well as measuring the relative (treated cell:untreated cell) viability of individual samples.
M Pallis N Russell
Department of Haematology Nottingham City Hospital and University of Nottingham Nottingham, UK
References 1 Guerci A, Merlin JL, Missoum N, Feldmann L, Marchal S, Witz F, Rose C, Guerci O. Predictive value for treatment outcome in acute myeloid leukemia of cellular daunorubicin accumulation and Pglycoprotein expression simultaneously determined by flow cytometry. Blood 1995; 85: 2147–2153. 2 Marie JP, Faussatsuberville AM, Zhou DC, Zittoun R. Daunorubicin uptake by leukemic cells – correlations with treatment outcome and mdr1 expression. Leukemia 1993; 7: 825–831. 3 Pallis M, Turzanski T, Harrison G, Wheatley K, Langabeer S, Burnett A, Russell N. Use of standardised flow cytometric determinants of multidrug resistance to analyse response to remission induction chemotherapy in patients with acute myeloblastic leukaemia. Br J Haematol 1999; 104: 307–312. 4 Broxterman H, Sonneveld P, Pieters R, Lankelma J, Eekman C, Loonen A, Schoester M, Ossenkooppele G, Lowenberg B, Pinedo H, Schuurhuis G. Do P-glycoprotein and major vault protein (MVP/LRP) expression correlate with in vitro daunorubicin resistance in acute myeloid leukemia? Leukemia 1999; 13: 258–265. 5 Robinson L, Roberts W, Ling T, Lamming D, Sternberg S, Roepe P. Human MDR1 protein expression delays the apoptotic cascade in Chinese hamster ovary fibroblasts. Biochemistry 1997; 36: 11169–11178. 6 Pallis M, Russell N. A drug efflux independent role for p-glycoprotein in augmenting the apoptosis induced by growth factor withdrawal in acute myeloid leukemia. Br J Haematol 1999; 105: 77. 7 Smyth M, Krasovskis E, Sutton V, Johstone R. The drug efflux protein, P-glycoprotein, additionally protects drug-resistant tumor cells from multiple forms of caspase-dependent apoptosis. Proc Natl Acad Sci USA 1998; 95: 7024–7029.
Correspondence
8 Marie JP, Huet S, Faussat AM, Perrot JY, Chevillard S, Barbu V, Bayle C, Boutonnat J, Calvo F, Campos Guyotat L, Colosetti P, Cazin JL, de Cremoux P, Delvincourt C, Demur C, Drenou B, Fenneteau O, Feuillard J, Garnier Suillerot A, Genne P, Gorisse MC, Gosselin P, Jouault H, Lacave R, Le Calvez G, Leglise MC, Leonce S, Manfait M, Maynadie M, Merle Beral H, Merlin JL, Mousseau M, Morjani H, Picard F, Pinguet F, Poncelet P, Racadot E, Raphael M, Richard B, Rossi JF, Schlegel N, Vielh P, Zhou DC, Robert J.
Multicentric evaluation of the MDR phenotype in leukemia. Leukemia 1997; 11: 1086–1094. 9 Schmid I, Uittenbogaart CH, Keld B, Giorgi JV. A rapid method for measuring apoptosis and dual color immunofluorescence by single laser flow cytometry. J Immunol Meth 1994; 170: 145–157. 10 Norgaard J, Langkjer S, Palshof T, Clausen N, Pedersen B, Hokland P. Relation of blast cell survival and proliferation to chemotherapy resistance in AML. Br J Haematol 1996; 93: 888–897.
TEL-AML1 positivity in relapsed B cell precursor acute lymphoblastic leukemia in childhood TO THE EDITOR TEL-AML1 fusion resulting from the cryptic translocation t(12;21)(p13;q22) constitutes the most frequent genetic rearrangement in initial childhood B cell precursor (BCP) acute lymphoblastic leukemia (ALL) (20–25%), and has been associated with a favorable prognosis in life-table analyses assessing outcome at 3 or 4 years.1–4 In contrast, the communicated incidences of TEL-AML1 positivity in relapsed ALL either supporting or questioning the feature of TEL-AML1 fusion as positive prognostic indicator vary strongly. In a recent issue of Leukemia, Zuna et al5 and Rubnitz et al6 reported on the incidence of TEL-AML1 fusion in relapsed childhood ALL treated initially on ALL BFM (Berlin–Frankfurt–Mu¨nster) protocols 83 to 95 in the Czech Republic, and on three consecutive protocols at St Jude Children’s Hospital, respectively. Four out of 45 relapsed ALL were TEL-AML1 positive in the Czech study (8.9%) and 5/49 (10.2%) in the analysis of St Jude. This incidence is lower than in previous Japanese and German reports (19–28%).7–9 In yet another recent retrospective study, the DFCI study group found an exceedingly low incidence of relapsed TEL-AML1-positive ALL (1/32, 3.1%).10 These variable frequencies (3.1–28%) require comment at this point and clarification in prospective clinical trials. General characteristics of TEL-AML1-positive childhood ALL both at diagnosis and at relapse are the restriction to B cell precursor immunophenotype, good response to current combination chemotherapy, and mostly WBC ⬍50 × 103 l.2–6,8–10 In spite of different relapse rates, studies on relapsed TEL-AML1-positive BCP-ALL are concordant with respect to the long median duration of first complete remission (CR).5,6,8–10 Major points of criticism regarding these retrospective studies on frequency of TEL-AML1 in first BCP-ALL relapses concern selection biases due to retrospective analysis of cryopreserved bone marrow samples, and analysis of only a small cohort of patients treated furthermore on consecutive frontline regimens. Methodological aspects and – although very unlikely – therapy-induced secondary leukemia possibly influencing the relapse rates cannot be ruled out completely at present. In addition to our previous report on 142 patients with relapsed BCP-ALL,9 bone marrow samples from a total of 268 children with first BCP-ALL relapse enrolled in multicentric BFM study group relapse trials, ALL-REZ BFM 90/96, including 178 patients screened prospectively during the last 2. years, were analyzed for the presence of TELAML1 fusion transcripts. Chimeric TEL-AML1 transcripts were detected by nested RT-PCR, and integrity of cDNA was controled by internal co-amplification of c-ABL cDNA, as described previously. To confirm results, one additional diagnostic sample was analyzed separately. Methodology of nested RT-PCR was verified by interphase FISH analysis of 25 TEL-AML1-positive and 58 TEL-AML1-negative ALL samples. Results were in complete concordance in all comparative analyses. The frequency of TEL-AML1 positivity in first BCP-ALL relapses was 17.5% (50/268) in the total study population, and 17.4% (31/178) in the prospective analysis. The predominant majority of patients ana-
Correspondence: K Seeger; Fax: 49–30–450 66906 Reveived 22 April 1999; accepted 25 May 1999
lyzed prospectively had initial chemotherapy for ALL according to three consecutive multicenter BFM (144/178; ALL-BFM 86–95) or cooperative CoALL (24/178) frontline protocols. Among second BCPALL relapses reported to the study center, the incidence of TEL-AML1 fusion was 27.8% (15/54) and 33.3% (12/36), respectively. Statistical analysis of 178 patients with first BCP-ALL relapse analyzed prospectively corroborates our previous analysis on relapsed TEL-AML1-positive BCP-ALL. TEL-AML1-positive and -negative patients differed significantly with respect to duration of first CR (median 46.1 vs 26 months; P ⬍ 0.0001), and age at initial diagnosis (median 54 vs 76 months; P = 0.02). TEL-AML1-positive relapses occurred preponderantly off-therapy (28/31; 90%), in contrast to 67/147 (46%; P ⬍ 0.0001) TEL-AML1-negative relapses. Similar results are obtained if only patients with first BCP-ALL relapse (26 TEL-AML1-positives among 123 patients, 21.1%) treated uniformly on frontline trial ALL BFM 90 are considered. The median duration of first CR and median age at diagnosis for TEL-AML1-positive and -negative patients were 45.4 vs 28.9 months (P ⬍ 0.001) and 53 vs 77 months (P = 0.004), respectively. TEL-AML1-positive patients had standard or intermediate risk arm therapy of trial ALL BFM 90, 15 TEL-AML1negatives were assigned to the high risk group. No significant differences were detected in all analyses between TEL-AML1 defined groups regarding WBC and peripheral blast cell count, sex, and age at relapse. The rather high incidence of TEL-AML1 fusion in second BCP-ALL relapses may be biased in the way that not all second relapses are reported to the study center, and that TEL-AML1 are overrepresented due to association of late-relapse and better outcome. Considering all published reports, irrespective of frontline treatment and size of study cohort TEL-AML1-positive BCP-ALL are characterized by a long duration of first CR (median 46 months; range 13– 125 months).5,6,8–10 Half of the patients suffer a relapse beyond 46 months of diagnosis. Therefore, analyses of frontline treatment assessing outcome at 3 to 4 years omit half of all relapse events. This accounts also to survival analyses only including patients with ALL at risk beyond 4 years.5 Moreover, in order to prevent biases based on dissimilar distribution of ALL immunophenotypes and/or low number of patients prone to marked variations, the incidence of TEL-AML1 fusion should be related to B cell precursor phenotype. An additional relapse in the study of Zuna et al5 would increase the relapse rate of TEL-AML1-positive ALL among BCP-ALL from 13.3% (4/30) to 16.1% (5/31), an incidence similar to the presented results based on a representative number of patients with BCP-ALL. Concerning the St Jude study, the immunophenotypes of the analyzed ALL were not specified.6 Prospective data are lacking from other single or multicentric centers. However, the discrepancies on frequency of TEL-AML1 fusion at ALL relapse clearly demonstrates that many issues remain to be resolved in prospective clinical trials on sufficiently large cohorts. Evaluation of both frontline and relapse studies have to take into account the long relapse-free intervals of TEL-AML1-positive patients. The international comparison of different frontline trials will permit evaluation of the effect and efficacy of therapeutic regimens in the treatment of TEL-AML1-positive ALL.
1469
