FLT3-ITD–, but not BCR/ABL-transformed cells require concurrent Akt/mTor blockage to undergo apoptosis after histone deacetylase inhibitor treatment

The BCR/ABL chromosomal translocation (Ph⁺) is detected in up to 30% of adult acute lymphoblastic leukemia (ALL) patients and associated with a detrimental prognosis.^1,2  The FMS-like tyrosine kinase 3 (FLT3)–internal tandem duplication (ITD) oncogene has been identified as a frequent aberration in acute myeloid leukemia (AML).³  Both the FLT3-ITD and the BCR/ABL oncogenes cause malignant transformation and confer factor-independent growth (transformation) when expressed in the interleukin-3 (IL-3)–dependent cell line 32D.^4,5  Specific BCR/ABL inhibition with imatinib mesylate (IM, Gleevec; Novartis, Basel, Switzerland) antagonizes transformation and causes impressive clinical remissions;^6,7  however, IM resistance inevitably occurs in Ph⁺ ALL,⁸  and the overall prognosis in this disease remains poor even after allogenic transplantation.⁹  Histone deacetylase inhibitors (HDIs) block tumor growth via pleiotropic mechanisms that are only partially understood.^10,11  Valproic acid (VPA) is a well-known anticonvulsant drug with HDI activity. VPA alone or in combination with all-trans retinoic acid (ATRA) has been reported to inhibit tumor growth and trigger differentiation^12,13  and apoptosis^11,14  of AML cells. However, in vivo, we and others recently demonstrated that the efficacy of VPA/ATRA in inducing remissions in myelodysplastic syndromes/AML is rather poor, whereas disease stabilization may be obtained more frequently.^15-17  In BCR/ABL-positive cell lines, HDI have been shown to overcome IM resistance, particularly when combined with other inhibitors classes.¹⁸  Here we investigated determinants for the heterogenous efficacy of HDI in inducing apoptosis in acute leukemias.

ATRA, VPA, z-VAD-fmk, and rapamycin were purchased from Sigma (Steinheim, Germany); SH6 was obtained from Alexis (Grünberg, Germany). The final concentrations of the different compounds used in this study were as follows: ATRA at 1 μM, VPA at 1 mM, SH6 at 10 μM, rapamycin at 10 nM, and caspase inhibitor z-VAD-fmk at 100 μMor200 μM.

Peripheral blood and bone marrow aspirates were obtained during routine punctures after obtaining written informed consent. Sample acquisition was approved by the local ethics committee of the University Marburg. In total, specimens of 9 patients with Ph⁺ ALL and 9 consecutive patients with AML were analyzed. For patient details, see Table S1, available on the Blood website (see the Supplemental Table link at the top of the online article). Mononuclear cells were isolated using Ficoll-Hypaque density gradient. The percentage of blasts was more than 80%. Fresh leukemic cells were cultured in RPMI 1640 supplemented with 30% fetal calf serum (FCS), 5 ng/mL IL-3 (Peprotech, Rocky Hill, NJ), 10 ng/mL IL-6 (Peprotech), and 20 ng/mL stem cell factor (SCF; Peprotech).

For assessment of apoptosis, the Annexin V–FITC Apoptosis detection kit (Sigma-Aldrich, Steinheim, Germany) was used essentially as recommended by the manufacturer and as previously described.¹⁹  Cell cycle was analyzed by flow cytometry as previously described.²⁰ 

Western blotting was performed as previously described¹⁹  using the following antibodies for detection of p21 (sc-6246), p27 (sc-528), and p45^skp2 (sc-7164) (all from Santa Cruz Biotechnology, Santa Cruz, CA), and total Akt, pAKT (Ser 473), total p70S6 kinase, pp70S6 kinase (Thr 389), acetyl-histone H3 (lys9), and acetyl-histone H4 (lys8) from Cell Signaling Technology (Danvers, MA).

Analysis of the statistical significance of a difference between the apoptosis induction frequency in different samples was done with the Mann Whitney U test. Differences were considered significant with P levels less than .05.

32D cells were transformed by BCR/ABL and FLT3-ITD, resulting in IL-3–independent growth, respectively (Figure 1A). VPA/ATRA synergistically inhibited factor-independent growth in transformed lines and mediated a G₀/G₁ arrest (Figure 1A-B). VPA or VPA/ATRA blocked deacetylation of acetyl-histone H3 and H4 and up-regulated 2 known targets of HDI, p21^waf1 and p27^kip1 (Figure 1C).¹⁰  The level of accumulation of p27^kip1 correlated with the degree of down-regulation of its specific ubiquitin ligase p45^{SKP2 21} and with growth arrest (Figure 1B-C). This suggested that VPA/ATRA synergistically inhibited proliferation via induction of p21^waf1 and p27^kip1. Notably, VPA/ATRA-induced growth arrest was comparable in BCR/ABL- and FLT3-ITD–transformed cell lines (Figure 1A-B), but apoptosis occurred exclusively in BCR/ABL-transformed 32D cells (Figure 1B, D). Likewise, BCR/ABL-transformed BaF3 pre-B cells, and BCR/ABL-positive K562 cells were susceptible to undergo VPA/ATRA-stimulated apoptosis (data not shown). This implied that BCR/ABL, but not the AML oncogene FLT3-ITD, confers a particular responsiveness of transformed cells to VPA/ATRA-induced apoptosis. Since Akt/mTor signaling provides important survival impulses for AML blasts,²²  we investigated the Akt activation status before and after HDI treatment. Unexpectedly, VPA/ATRAdid not cause inhibition, but rather increased activation (phosphorylation) of Akt and p70S6 kinase in FLT3-ITD–and BCR/ABL-transformed cells (Figure 1E). Consequently, VPA/ATRA-induced cell death in BCR/ABL-positive cells occurred independently from Akt activation levels. In contrast, Akt could have contributed to apoptosis resistance against HDI in FLT3-ITD cells. Indeed, when Akt was specifically blocked in FLT3-ITD cells by SH6,²³  as demonstrated by means of reduction of the phospho-Akt and phospho-p70S6K levels (Figure 1E), apoptosis resistance to VPA/ATRA was overcome (Figure 1D-E). Therefore, Akt blockage was required for apoptosis induction by VPA/ATRA in FLT3-ITD–positive cells, but not in BCR/ABL-expressing cells (Figure 1D-E). Furthermore, we found that HDI-mediated apoptosis in BCR/ABL-positive 32D cells and Ph⁺ primary cells was strictly caspase dependent because it could be blocked almost entirely by pan-caspase inhibition using z-VAD-fmk (Figure 2A). In contrast, VPA/ATRA/SH6-induced apoptosis in FLT3-ITD cells was not influenced by caspase inhibition (not shown).

In order to validate our findings on primary cells, BCR/ABL-positive ALL blast samples (n = 9) and BCR/ABL-negative AML (n = 9) (Table S1) with greater 80% blasts in the peripheral blood were compared for HDI-stimulated apoptosis. In line with our hypothesis, VPA/ATRA elicited considerably more apoptosis in BCR/ABL-positive acute leukemias than in AML (P < .01; Figure 2B). In addition, as seen in the AML model cell line 32D-FLT3-ITD, resistance of primary AML blasts to undergo apoptosis after VPA/ATRA treatment was overcome by blocking Akt signaling (P < .03; Figure 2C). To inhibit the Akt pathway, the clinically available inhibitor of mTor, rapamycin (Rap), was used. Neither Rap nor VPA/ATRA alone was capable of eliciting apoptosis (Figure 2C). As shown in 2 patients with AML (no. 15 and no. 11), only the combination of VPA/ATRA and Rap, but neither drug alone, antagonizedAkt and p70S6K activity, which resulted in apoptosis (Figure 2D). This underscored the importance of Akt signal inhibition as a prerequisite for HDI-induced apoptosis in primary AML.

An in vivo proof for the concept that Ph⁺ ALLs are particularly susceptible to HDI-induced apoptosis was observed in a patient who had failed all previous chemotherapeutic therapies, including IM. In this patient a combination treatment of oral VPA (increased to 300 mg twice daily) and ATRA (40 mg daily) was commenced. After 1 month of therapy, platelet and neutrophil blood counts had improved to almost normal values, and he recovered from a febrile pneumonia (Figure 2Ei). Whereas a bone marrow aspiration documented a hypercellular marrow with nearly 100% blastic infiltration before treatment (Figure 2Eii, left panel), the infiltration decreased to 20% to 30% of CD34⁺ and TdT⁺ blasts under VPA/ATRA therapy (Figure 2Eii, middle and right panels).

Together, Akt differentially dictates sensitivity to VPA/ATRA-mediated apoptosis in different types of acute leukemia. Whereas BCR/ABL-positive leukemias die after VPA/ATRA stimulation, FLT3-ITD–transformed cells and primary AML blasts require an independent Akt pathway inhibition to undergo HDI-stimulated apoptosis. These data warrant the clinical application of HDI in poor prognosis, Ph⁺ ALL. In AML, Akt/mTor signaling blocker may contribute to overcome HDI resistance.

The authors wish to thank Christine Barett for excellent technical assistance and Dr Hubert Serve for providing the 32D-FLT3-ITD cell line.