In the prechemotherapy era arsenic derivatives were used for treatment of chronic myelogenous leukemia, a myeloproliferative disorder characterized by the t(9;22) translocation, the Philadelphia chromosome(Ph+). In acute promyelocytic leukemia response to arsenic trioxide(As2O3) has been shown to be genetically determined by the acute promyelocytic leukemia-specific t(15;17)translocation product PML/RARα. Hence, we reasoned that As2O3 might have a selective inhibitory effect on proliferation of BCR-ABL-expressing cells.
Here, we report that: (a)As2O3 induced apoptosis in Ph+ but not in Ph−lymphoblasts; (b) enforced expression of BCR-ABL in U937 cells dramatically increased the sensitivity to As2O3; (c) the effect of As2O3 was independent of BCR-ABL kinase activity; and (d)As2O3 reduced proliferation of chronic myelogenous leukemia blasts but not of peripheral CD34+ progenitors. In summary, these data establish As2O3 as a tumor cell-specific agent, making its clinical application in Ph+ leukemia feasible.
Arsenic derivatives represent one of the oldest treatments for leukemia. In the last century, Fowler’s solution (potassium arsenite)was administered to patients suffering from leukemia. In the 1930s,treatment with Fowler’s solution achieved remarkable clinical response in CML3(Ref. 1 and references therein). Until the 1950s, arsenic has been used in combination with radiotherapy to treat CML(2). More than 95% of CMLs are Ph+ and, thus, express the p210(BCR-ABL). The other possible t(9;22)translocation product, p185(BCR-ABL), is present predominantly in adult Ph+ ALLs (20–25% of adult ALLs; Ref.3). In both cases, the t(9;22) translocation leads to mutated forms of the genes encoding the tyrosine kinases BCR and ABL,which become constitutively activated, thereby inducing aberrant proliferation and neoplastic transformation (3).
Recently, it has been reported that As2O3 is capable of inducing complete remissions in patients with t(15;17) APL(4, 5, 6). The response to As2O3 in APL patients is genetically determined by expression of the PML/RARαfusion protein specific for t(15;17). Furthermore, transfection of PML/RARα into naturally As2O3-resistant U937 cells renders these cells sensitive to As2O3-induced apoptosis(7).
Therefore, we investigated whether there is also a functional relationship between the expression of the translocation product BCR-ABL and As2O3-induced apoptosis in Ph+ leukemia.
Here, we show that As2O3induced apoptosis in Ph+, but not in Ph− lymphoblastic cell lines. BCR-ABL mediated As2O3-induced apoptosis,analogous to PML/RARα. This activity was independent of the aberrant kinase activity of BCR-ABL. As2O3 was effective on Ph+peripheral blasts of patients with CML in blast crisis but did not influence colony formation activity of primary CD34+ hematopoietic precursors.
The presented data establish the basis for the application of As2O3 as a tumor cell-specific agent in the treatment of Ph+ CML, as well as Ph+ ALL.
Materials and Methods
Cell Lines, Cell Culture, and Western Blotting.
Nalm-6, MOLT-3, SEM, Jurkat, Daudi, BV173, SD-1, and U937 cells were maintained in RPMI 1640 supplemented with 10% FCS (Life Technologies,Inc., Karlsruhe, Germany). TOM-1 cells were maintained in Iscove Medium with 10% FCS and Sup B15 in RPMI 1640 with 15% FCS. The U937 MT B45 and PML/RARα-positive P/R9 cells were obtained as described previously (8, 9). The p185 (BCR-ABL) and p210 (BCR-ABL) encoding cDNA was cloned into the ZnSO4 (Zn2+)-inducible pGMTSVneo expression vector (9). The BCR-ABL-positive U937 MTp185 and MTp210 bulk populations were obtained by electroporation of these p185 (BCR-ABL)- and p210 (BCR-ABL)-carrying expression vectors and G418 selection. The expression of the exogenous protein by Zn2+ treatment was induced as described(9) and evaluated by Western blotting. Anti-ABLantibody (α-ABL), anti-bcl-X(α-bcl-X), and anti-PARP (α-PARP)were purchased from Santa Cruz Biotechnology (Santa Cruz, CA);anti-bcl-2 (α-bcl-2) from DAKO (Hamburg,Germany); and anti-phosphotyrosine (α-PY) from Upstate Biotechnology,Inc. (Lake Placid, NY). All were applied according to widely used protocols.
Zn2+-treated U937 cells were washed twice with PBS to eliminate Zn2+, thus avoiding interference with apoptosis. All cell types were diluted to 1 × 105 cell/ml and exposed to a final concentration of 0.1–2 μmAs2O3 (Sigma Chemical Co.,St. Louis, MO). For analysis of the rate of apoptosis, the FACS-based 7-AAD method was used as described elsewhere (7).
Fresh CD34+ progenitor cells were purified from leukapheresis samples after 4–5 days of mobilization with granulocyte colony-stimulating factor (10 μg/kg body weight) of two patients with Ph+ ALLs in CR. The CML blasts derived from the peripheral blood of two patients with newly diagnosed myeloid or lymphatic blast crisis. CD34+ cells were isolated by Ficoll-Hypaque density-gradient centrifugation, followed by separation with the VarioMACS system, with the appropriate columns using the Direct CD34+ Progenitor Isolation Kit according to manufacturer’s instructions (Myltenyi Biotec, Bergisch-Gladbach,Germany).
Fresh CD34+ cells and CML PMNCs were plated at 400 cells/ml on day 2 of treatment with As2O3, in 0.9% semisolid methyl-cellulose Methocult complete medium (StemCell Technologies Inc., Vancouver, Canada) and incubated at 37°C in a humidified atmosphere of 5% CO2. Each assay was plated as triplicate. After 9–14 days of incubation colonies (>50 cells) were counted.
As2O3 Induces Apoptosis in BCR-ABL-positive Lymphoblastic Cell Lines.
To determine whether Ph+ cell lines recapitulate the response of CMLs to As2O3, analogous to the PML/RARα-expressing NB4 and U937 cells (4, 7), we exposed different ALL- and CML-derived lymphoblastic cell lines to 2 μmAs2O3. SD-1, Tom-1, and Sup-B15 cells are p185(BCR-ABL)-positive ALL cell lines, and BV173 is a p210(BCR-ABL)-positive CML cell line (10). As control, we used Ph− lymphoblastic cell lines such as Nalm-6 (B-lineage ALL), MOLT-3 (T-lineage ALL), SEM[t(4;11)-positive B-lineage-ALL], Jurkat (leukemic T-cell lymphoma),and Daudi (Burkitt lymphoma). Fig. 1,A depicts the BCR-ABL protein expression levels of the above cited cell lines. The rate of apoptosis was determined by FACS analysis, measuring the percentage of 7-AAD-positive cells after 72–96 h of As2O3 exposure(11). In Fig. 1,B, we show one representative experiment of three that gave similar results. In contrast to untreated cells (BV173, 16%; SD-1, 16%; Sup-B15, 26%; Tom-1, 35%) all Ph+lymphoblastic cell lines showed a high rate of apoptosis on treatment with As2O3 (BV173, 74%;SD-1, 49%; Sup-B15, 93%; Tom-1, 79%). None of the Ph− cell lines exhibited a significant response to the As2O3 treatment (Fig. 1 B). No average was calculated because the kinetics of apoptosis in the Ph+ cell lines differed between separate experiments,resulting in considerable variability, in particular, at the early time points. Moreover, in contrast to Ph− cells, none of the Ph+ cell lines recovered from treatment with As2O3 in prolonged culture(data not shown).
Fig. 1 C demonstrates a typical 7-AAD FACS profile of BV173 cells exposed to increasing concentrations of As2O3. Already, 0.1μ mAs2O3 induced a significant rate of apoptosis with respect to untreated cells. Furthermore,activity of As2O3 was dose dependent, as demonstrated by an increased percentage of apoptotic cells with rising dosages of As2O3.
Taken together, these data indicate a specific activity of As2O3 in BCR-ABL-expressing lymphoblastic cell lines comparable with the known activity of As2O3on PML/RARα-positive cells. Moreover,As2O3 exerts its effect on Ph+ cells, regardless of the type of product of t(9;22) present[i.e., the ALL-specific p185(BCR-ABL)is able to mediate sensitivity to As2O3 to the same extent as the CML specific p210(BCR-ABL) ].
As2O3-induced Apoptosis in Ph+ Cells Is Genetically Determined by the Presence of t(9;22).
To examine whether As2O3-induced apoptosis in Ph+ cell lines is specifically mediated by BCR-ABL, and to exclude the possibility that the effect of As2O3 is due to a yet unknown common feature of the different Ph+ cell lines, we transfected U937 cells with different expression vectors. The expression vectors contained cDNA encoding either p185(BCR-ABL) or p210(BCR-ABL) under the control of the Zn2+-inducible metallothionein (MT-1)promoter (9).
U937 cells are myeloid precursors blocked at the promonocytic stage,which do not undergo As2O3-induced apoptosis(4, 7). In our experiments, we analyzed the effect of As2O3 on p185(BCR-ABL)- and p210(BCR-ABL)-expressing U937 cells. To avoid the bias of clonal variability, we used highly expressing bulk populations selected after transfection only for G418 resistance without further subcloning (MTp185 and MTp210). On Zn2+ induction the transfected cells expressed the BCR-ABL fusion proteins to a similar level as BV173, SD-1, Sup-B15, and Tom-1 cells (Fig. 1,A). As negative control for As2O3-induced apoptosis we used the MT B45 cells, transfected with the “empty” expression vector, and as positive control we used the PML/RARα-expressing P/R9 cells, as described previously(7). Twelve h of Zn2+ treatment was by itself not able to induce apoptosis in any cell lines (Fig. 2). On exposure to 1 μmAs2O3 and in the absence of Zn2+, MT B45 control cells showed a nearly identical apoptosis rate than with Zn2+ treatment alone (20% and 19%, respectively). Even without Zn2+-induced protein expression there was a pronounced increase in apoptosis in BCR-ABL p210 and P/R9 cells (29% and 40%, respectively) when compared with MT B45 control cells (19%) and BCR-ABL p185 cells (12%; Fig. 2). This effect was most likely due to “leakage” protein expression from the transgenes, as demonstrated by Western blot (Fig. 1 A). When cells were treated for 12 h with Zn2+ to induce protein expression prior to As2O3 exposure, apoptosis increased dramatically in the MTp185 and MTp210 populations (77% and 56%, respectively). Taking into account that the BCR-ABL-positive U937 cells are bulk populations, where only 50–70% of cells express the transgenes (determined by immunofluorescence studies and further subcloning of the bulk populations; data not shown), sensitivity of the p185(BCR-ABL)- and p210(BCR-ABL)-expressing cells to As2O3 reached nearly the same extent than the P/R9 clone (96%).
Taken together, these data indicate that the response to As2O3 in Ph+ leukemia is genetically determined by the presence of the t(9;22) translocation products p185(BCR-ABL) or p210(BCR-ABL).
Response to As2O3 Is Independent of the Constitutive ABL Kinase Activity of BCR-ABL.
BCR-ABL transforms cells via its aberrant constitutive kinase activity (3). To investigate a possible role of BCR-ABL kinase activity in As2O3-induced apoptosis, we exposed BV173 and SD-1 cells to the specific ABL-kinase inhibitor STI 571 [kindly provided by E. Buchdunger (Novartis, Basel,Switzerland)]. Cells were treated with As2O3 after a 6-h exposure to 0.5 μm STI571, to guarantee that ABL kinase activity was switched off. This is the lowest possible concentration of STI571, ensuring inhibition of BCR-ABL kinase activity. In BV173 cells as well as in SD-1 cells STI571 exhibited its known proapototic effects on BCR-ABL-transformed cells, but had no considerable influence on response to As2O3. Fig. 3 A shows one of three representative experiments that gave similar results.
To answer the question whether As2O3 induces apoptosis by interfering with the BCR-ABL kinase activity, we probed immunoblots of cellular lysates of BV173 and SD-1 cells after 8, 24,48, and 72 h of As2O3treatment with an antiphospho-tyrosine monoclonal antibody (see“Material and Methods”). There was no decrease in the expression of phosphorylated BCR-ABL at any time point (Fig. 3 B).
In summary, the response to As2O3 in Ph+ lymphoblasts seems to be independent of the constitutive ABL kinase activity of BCR-ABL.
In BCR-ABL-positive Cells As2O3 Activates Apoptosis without Caspase-3 Activation or bcl-2 Regulation.
In APL, the role of bcl-2- and caspase-3-like activity in As2O3-induced apoptosis is controversially discussed. Reportedly, one of the mechanisms of decreased susceptibility to apoptosis in BCR-ABL-positive cells is due to up-regulation of bcl-2 (12). Therefore, we assessed bcl-2 expression by immunoblotting in BV173 cells on As2O3 treatment. In comparison with untreated BV173 cells, no modification of bcl-2 expression was noted at 8, 24, 48, or 72 h of As2O3 treatment (Fig. 3 B).
As shown previously, bcl-X expression plays an important role in protection from various apoptotic stimuli in BCR-ABL-transfected HL-60 cells (13). Two bcl-X gene products are known: bcl-XL, an inhibitor of apoptosis, and bcl-XS, a promoter of apoptosis (reviewed in Ref. 14). To answer the question whether variations of bcl-X isoform expression explains the mechanism of As2O3-induced apoptosis in BV173, we probed the above-mentioned immunoblots with an antibody recognizing bcl-XL as well as bcl-XS (Santa Cruz Biotechnology). Only bcl-XL was detected, and no difference in its expression level between untreated and As2O3-treated BV173 cells was seen (Fig. 3 B).
Another key player discussed in the context of As2O3-induced apoptosis is caspase-3. (7, 15). Caspases constitute a family of cysteine proteases with aspartic acid substrate specificity,thought to be crucial for apoptosis in multicellular organisms(reviewed in Ref. 14). To address whether As2O3-induced apoptosis is mediated by caspase-3 activity, we probed the same samples with an antibody specific for PARP (Santa Cruz Biotechnology), a known substrate for several caspases, including caspase-3. In presence of activated caspase-3, PARP is cleaved and the 113 kDa species is replaced by a 81 kDa species, which is also recognized by theα-PARP antibody. In our experiments, no cleavage of endogenous PARP was observed after 8, 24, 48, or 72 h of As2O3 exposure (Fig. 3 B), indicating that caspase-3 is not involved in As2O3-induced apoptosis of BCR-ABL-positive BV173 cells.
Taken together, these data give further proof that As2O3 induces apoptosis independent of bcl-2 expression level and caspase-3-like activity.
As2O3 Has No Effect on the CFUs of CD34+Primary Hematopoietic Precursors.
Clinical studies indicate that APL patients treated with As2O3 alone do not experience aplasia of the bone marrow, commonly seen in cytotoxic chemotherapy regimen (6). To assess a possible effect of As2O3 on normal hematopoietic progenitor cells, we exposed CD34+ cells isolated from two patients with Ph+ ALL in CR to 2 μmAs2O3 and tested the CFU in a methyl-cellulose colony formation assay. The experiments were performed in triplicates. CD34+ cells were seeded in methyl-cellulose after 2 days of exposure to As2O3. There was no considerable difference regarding number, morphology, composition, or relationship between GM-CFU, BFU-E, and CFU mix between the treated and untreated population. A representative depiction of the growth pattern of the colonies of one of two patients’ samples is given in Fig. 4.
These data confirm the hypothesis that the toxic effect of As2O3 on normal hematopoietic precursors is minimal, as long as As2O3 is applied in clinically relevant concentrations.
As2O3 Significantly Reduces the CFU of Ph+Blasts of CML Patients in Blast Crisis.
In BV173 cells a 6-h exposure to As2O3 is sufficient to irreversibly induce apoptosis (data not shown). To examine the effect of As2O3 on primary blasts of CML patients, we treated PMNCs of five CML patients in myeloid or lymphatic blast crisis. On As2O3 exposure there was a very high variability in viability and total cell number between different experiments and patient samples, most likely due to the different sensitivity to culture conditions (data not shown). For that reason, PMNC samples of two patients, which showed no effect on As2O3 exposure regarding total cell number or viability, were seeded in a methyl-cellulose colony assay (see “Material and Methods”). At the 9th day, the number of CFUs of the As2O3-treated samples was significantly lower than in untreated control cells (Fig. 4). Interestingly, on As2O3-exposure most of the CFUs exhibited the characteristics of differentiated granulocytic colonies rather than BFU-E or CFUs (Fig. 4).
In summary, these data indicate that BCR-ABL increases sensitivity to As2O3-induced apoptosis in Ph+ CML blasts, but not in CD34+ progenitors.
Here, we show that As2O3, an agent known to induce apoptosis in PML/RARα-positive APL, also exhibits potent and specific activity against BCR-ABL-expressing cells. Without exception, all Ph+ lymphoblastic cell lines (SD-1, Tom-1, Sup-B15, and BV173) examined were highly sensitive to As2O3-induced apoptosis. In contrast, Ph− cell lines, including the t(4;11)-positive SEM cells,responded to As2O3-induced apoptosis. Furthermore, there was no notable difference regarding response to As2O3 between the ALL-derived (p185(BCR-ABL) positive) SupB15,TOM-1, SD-1, and the CMLderived(p210(BCR-ABL) positive) BV173 cell lines or between p185(BCR-ABL)- and p210(BCR-ABL)-transfected U937 cells. Thus, we conclude that p185(BCR-ABL), as well as p210(BCR-ABL), is able to mediate response to As2O3.
Moreover, sensitivity of BCR-ABL-transfected U937 cells to As2O3-induced apoptosis also excludes the possibility that the effect of As2O3 is due to a common feature of bone marrow cells arrested at the B-cell precursor stage of differentiation. Instead, it demonstrates that As2O3-induced apoptosis is genetically determined by the presence of the t(9;22)-specific chimeric gene products p185(BCR-ABL) and p210(BCR-ABL). This effect of BCR-ABLis analogous to PML/RARα, which determines As2O3 sensitivity of APL blasts (6, 7). Other translocation products, such as HRX-AF4, the product of t(4;11), present in the SEM cells,did not mediate As2O3-induced apoptosis. Initial support for the hypothesis of a genetic determination of the As2O3 response in Ph+leukemia was given by the fact that, to the best of our knowledge, only CML patients were reported to respond to treatment with arsenic derivatives (1, 2).
Nevertheless, the mechanism by which BCR-ABL mediates As2O3-induced apoptosis remains unclear. As2O3 does not interfere with the constitutive kinase activity of BCR-ABL, and response to As2O3 is not influenced by the abrogation of BCR-ABL kinase activity. Our data indicate that As2O3-induced apoptosis does not interfere with signaling pathways used by BCR-ABL to transform cells. This is, in particular,supported by the fact that the overall tyrosin-phosphorylation pattern of BV173 and SD-1 cells was unaltered in response to As2O3 (data not shown).
All originally BCR-ABL-positive cell lines, as well as BCR-ABL-transfected U937 cells, have shown clear evidence of apoptosis after 48–72 h on As2O3-treatment. As assessed by immunoblotting, bcl-2 expression was unaffected by As2O3-treatment up to 72 h. These data confirm recent data on PML/RARα-transfected U937 and PML/RARα-positive NB4 cells, which underwent apoptosis without down-regulation of bcl-2 (7, 16). In our study, we extended the investigation to bcl-X, another regulator of apoptosis that seems to be influenced by BCR-ABL (13). BV173 expressed high levels of bcl-XL, but As2O3-treatment neither led to down-regulation of bcl-XL nor to up-regulation of bcl-XS, the proapoptotic form of bcl-X. PARP cleavage is an important indicator of caspase activation during apoptosis. Activated caspase-3 is one of the PARP-cleaving caspases (17). We excluded an involvement of PARP-cleaving caspase activity in As2O3-treated BV173 cells. This confirms recent data that showed that PML/RARα is degraded by PARP-cleaving activity on retinoic acid treatment, but not by As2O3(7, 18). Thus, we conclude that in BCR-ABL-transformed lymphoblasts As2O3-induced apoptosis is not mediated by any of the formerly discussed key players, such as bcl-2, bcl-X or caspase-3-like activity.
Because we show that the described effects of As2O3 are specific for Ph+CML blasts, as well as for p185(BCR-ABL)-expressing ALL-derived- and p210(BCR-ABL)-expressing CML-derived cell lines,our data establish As2O3 as a potential agent for the treatment of patients with Ph+ leukemia.
Supported by grants from Deutsche Forschungsgemeinschaft and “Dr. Mildred Scheel–Stiftung” der Deutschen Krebshilfe e.V. E. P. is a recipient of a fellowship from the “Deutsche José Carreras Leukämiestiftung e.V.”(DJCLS-99/NAT-1).
The abbreviations used are: CML, chronic myelogenous leukemia; ALL, acute lymphocytic leukemia; APL, acute promyelocytic leukemia; FACS, fluorescence-activated cell-sorting;PARP, poly-(ADP ribose) polymerase; 7-AAD, 7-amino-actinomycin D; CFU,colony-forming unit; As2O3, arsenic trioxide;Ph+, Philadelphia chromosome; CR, complete remission; PMNC, peripheral mononuclear cells; PML/RARα, promyelocytic leukemia gene/retinoic acid receptor α; GM-CFU, granulocyte-macrophage CFU; BFU-E,blast-forming unit(s) erythroid.