AML Progenitor Cells Research Articles

SFK Inhibition with Dasatinib Results In Selective Targeting of Primitive Human Acute Myeloid Leukemia Stem and Progenitor Cells.

AbstractAbstract 1053The Src family tyrosine kinases (SFKs) are abnormally activated in AML compared to normal CD34+ hematopoietic progenitors. Studies using pharmacological and siRNA approaches indicate an important role for the SFK Lyn in AML progenitor cell growth and survival (Dos Santos et al., 2008). However the role of SFKs in AML leukemic stem cell (LSC) growth and survival is not clear. SFK activity in Lin- CD34+ CD38- primitive stem/progenitor cells, Lin- CD34+ CD38+ committed progenitors, and Lin-CD34- cells from primary human AML (n=14) and cord blood (CB) (n=6) samples was measured by analyzing SFK phosphorylation using flow cytometry. We observed significant increase of SFK phosphorylation in AML compared to CB Lin- CD34+CD38- cells (3.8±1.9 versus 1.9±0.7, p=0.006), Lin- CD34+ CD38+ (3.9±1.7 versus 1.9±0.6, p=0.013) and Lin- CD34- cells (3.4±1.8 versus 1.1±0.5, p=0.0005). Dasatinib, a potent SFK and ABL kinase inhibitor, is approved for clinical use in chronic myeloid leukemia. We evaluated the effect of SFK inhibition using Dasatinib on the growth and viability of AML stem and progenitor cells. Dasatinib exposure resulted in dose-dependent inhibition of SFK phosphorylation in each subpopulation (in CD34+ CD38-, 3.0 for the control versus 1.9 with 100nM and 1.6 with 500nM, in CD34+ CD38+, 2.8 for the control versus 1.9 with 100nM and 1.6 with 500nM) after 30 minutes of drug treatment). The addition of Dasatinib (10-500nM) to methylcellulose progenitor assays resulted in dose-dependent inhibition of AML colony forming cell (CFC) growth (83.9±16.1% inhibition with 500nM, and 48.1±6.8% inhibition with 10nM Dasatinib, n=8), to a greater extent than CB CFC to (CFU-GM inhibition 62.6±1.5% with 500nM, and 1.1±16.3% with 10nM Dasatinib, n=4). Short-term exposure to Dasatinib (10-500nM) for 48 hours also resulted in significantly greater inhibition of AML CFC (73.6±13% with 500nM, and 41.7±10.8% with 10nM Dasatinib, n=8) compared to CB CFC (CFU-GM inhibition 23±9.1% with 500nM, and 1.3±11.3% with 10nM Dasatinib, n=4). Importantly Dasatinib treatment (200nM) also resulted in reduction of AML stem/primitive progenitor growth in long term culture-initiating cells (LTC-IC) assays (56±23, 8 % inhibition, p=0.003, n=4), suggesting that SFK inhibition may inhibit AML stem cell maintenance. The effect of Dasatinib on apoptosis was evaluated by labeling cells with Annexin V and DAPI. Treatment with Dasatinib resulted in significant increase in apoptosis of Lin- AML cells (41.5% ±10.7 of apoptosis with 200nM Dasatinib versus 25%±10.8 for the control, p=0,004, n=5) We studied the effects of Dasatinib on differentiation of Lin- CB (n=3) and AML cells (N=5) cultured with SCF, IL-3, GM-CSF, G-CSF and EPO. In normal CD34+, Dasatinib (100 nM) treatment resulted in increased CD33+ and CD14+ cells and reduced CD34+, CD11b+, CD15+, GPA+ and CD71+ cell numbers, indicating that SFK increased monocytic but reduced granulocytic and erythroid differentiation. Treatment of AML cells with Dasatinib resulted in markedly reduced numbers of CD34+, CD33+ and CD71+ cells, but increased numbers of CD11b cells, in 3 of 5 samples, indicating a trend towards increased granulocytic differentiation in contrast to normal progenitors. Our results indicate that SFK activity is increased in primary human AML stem and progenitor cells and suggest that SFK blockade with Dasatinib may reduce maintenance of AML LSC/ primitive progenitors, through inhibition of progenitor proliferation, induction of apoptosis and enhancement of differentiation. These results support further evaluation of SFK blockade with Dasatinib for targeting of AML stem and progenitor cells in preclinical and clinical studies. Disclosures:Bhatia:Novartis: Consultancy, Honoraria.

The deacetylase inhibitor LAQ824 induces notch signalling in haematopoietic progenitor cells

AML progenitor cells (AML-PC) undergo significant apoptosis in response to the deacetylase inhibitor (DACi) LAQ824 and lose the replating capacity which was not observed with the DACi valproic acid. Treatment of normal hematopoietic progenitor cells (HPC) with LAQ824 resulted in (i) inhibition of differentiation, (ii) an G2/M cell cycle arrest exclusively in multipotent CD34 + HPC and (iii) induction of apoptosis predominantly in committed CD34 − HPC. Gene expression analysis showed induction of coactivator and target genes of the notch pathway as well as cell cycle arrest-inducing genes in the most primitive CD34 + CD38 − HPC population which may in part be responsible for the considerable, but reversible haematotoxicity of this drug.

Targeting Levels, Aberrant Localization or Oligomerization of Mutant Nucleophosmin Induces Differentiation and Loss of Survival of Human AML Cells with Mutant NPM1.

Abstract 2656Poster Board II-632NPM1 is a nucleolar phosphoprotein that forms oligomers and functions as a molecular chaperone for both proteins and nucleic acids. NPM1 normally shuttles between the nucleus and cytoplasm but is mutated and aberrantly localized to the cytoplasm in 35% of patients with AML, Mutant (m) NPM1 contains a 4-base insert that results in extra C-terminal residues encoding a nuclear export signal (NES), which causes mNPM1 to remain in the cytoplasm. In AML, the presence of mNPM1 is often co-detected with internal tandem duplications of FLT-3, which confers a poor prognosis. The effects of the cytosolic mNPM1 on the biology and drug-sensitivity of AML are unclear. In the present studies we determined the expression and biologic effects of mNPM1 in cultured and primary AML cells. Utilizing RT-PCR, immunoblot analysis and immunoflourescence microscopy (with a polyclonal anti-mNPM1 antibody, we confirmed that in the cultured AML OCI-AML3 cells (hetrozygous for NPM1 mutation) mNPM1 protein was localized in the cytosol, while the un-mutated NPM1 was nuclear. In AML HL-60 cells NPM1 was un-mutated and entirely nuclear. As compared to control siRNA, treatment with siRNA to NPM1 for 72 hours significantly induced p21 expression, decreased % of S phase cells in the cell cycle, as well as induced morphologic differentiation of OCI-AML3 but not of HL-60 cells. This was associated with marked induction of CEBPα in OCI-AML3 cells. NPM1 siRNA also attenuated the expression of HOXA9 and MEIS1 (which are known to be leukemogenic), associated with a marked loss of clonogenic survival of OCI-AML3 cells. NPM1 siRNA mediated depletion of Meis1, which is known to transactivate FLT3 tyrosine kinase in AML progenitor cells, was associated with down regulation of FLT-3 expression in OCI-AML3 cells. Importantly, treatment with NPM1 siRNA significantly sensitized OCI-AML3 more than HL-60 cells to all-trans retinoic acid (ATRA, 0.25 to 2.0 uM) and Ara-C (0.5 to 5.0 uM). Treatment with leptomycin B (5 nM for 24 hours), which is an inhibitor of the NES receptor CRM1/exportin-1, shifted the localization of mNPM1 from the cytosol to the nucleus, abrogated the levels of Meis1 and HoxA9 and induced apoptosis of OCI-AML3 but not HL-60 cells. Recently, treatment with NSC348884, a small molecule inhibitor of NPM1 oligomerzation, was shown to induce apoptosis of colon and prostate cancer cells. Our studies determined that NSC348884 (3.0 uM) was highly active in disrupting the oligomerization of mNPM1 and induce apoptosis of 90 % of OCI-AML but only 20% of HL-60 cells. Additionally, treatment with 5.0 uM of NSC34884 induced apoptosis of approximately 75% of primary AML cells with mNPM1, without inducing apoptosis of normal CD34+ progenitor cells. Collectively, these findings suggest that strategies targeting mNPM1 levels or oligomerization, or reversing the aberrant cytosolic localization of mNPM1, would induce relatively selective differentiation and apoptosis, as well as sensitize AML cells with mNPM1 to Ara-C and ATRA. Disclosures:No relevant conflicts of interest to declare.

Chemical Derivatives of the Anti-Leukemia Stem Cell Compound 4-Benzyl-2-Methyl-1,2,4-Thiadiazolidine-3,5-Dione (TDZD-8) with Improved Activity.

Abstract 3764Poster Board III-700We have recently described the novel and unique anti-leukemic properties of 4-benzyl-2-methyl-1,2,4-thiadiazolidine-3,5-dione (TDZD-8). Indeed, TDZD-8 was shown to eradicate leukemia at the bulk, stem, and progenitor level with rapid kinetics (typically < 2 h) with minimal toxicity to normal hematopoietic cells. Moreover, the cytotoxic activity of this drug is observed only in hematologic malignancies. The precise mechanism of TDZD-8 is not fully appreciated, but the compound has been shown to inhibit NFkappaB, GSK3beta, protein kinase C, FMS-like tyrosine kinase (Flt3), AKT and KDR. In addition, TDZD-8 rapidly depletes free thiols and appears to disrupt membrane integrity. Interestingly, TDZD-8 results in rapid mitochondrial swelling, followed by vacuole formation and accumulation. Despite possessing potent and specific anti-leukemia activity, the clinical utility of TDZD-8 is limited by the need for high concentrations (20 microM) and poor solubility. Therefore, to improve the pharmacological properties of TDZD-8, we initiated efforts to generate derivatives with greater anti-leukemia activity at lower concentrations and with greater water-solubility. To this end, an extensive structure-activity relationship study was carried out to examine the effects on anti-leukemic activity resulting from introduction of substituents at the C-2 and C-4- positions of the thiadiazolidine ring of TDZD-8, and of replacement of the TDZD ring with isosteric scaffolds. In total, we have screened more than 60 new analogs. Preliminary screens were performed using two different AML cell lines, evaluating the LD50 for each of the analogs relative to the parental compound. We found that introducing substitutuents into the main TDZD ring resulted in loss of anti-leukemic activity of the compounds. Moreover, substitution in the benzyl/phenyl ring does not affect anti-leukemia activity of the TDZD analogs. Importantly, we found that N-2 halogenoethyl analogs, exhibit exceptional activity against leukemic cells. Of the halogenoethyl analogs evaluated, the iodoethyl analog TD-361 was the most active compound with an LD50 of 0.49 microM in MV-411 cells. Compounds exhibiting increased anti-leukemia activity were subjected to further testing in phenotypically-defined AML stem/progenitor cells. All analogs demonstrated efficacy in primary AML cells at lower concentrations than TDZD-8. Moreover, these analogs still maintained the rapid kinetics observed with TDZD-8. Finally, we performed colony assays to determine the effect of new analogs on progenitor cells from normal and leukemic cells. These studies demonstrated that the more active TDZD analogs retained the selective ability of TDZD-8 to abate AML progenitor cells without harming normal hematopoietic cells. In summary, we have identified the critical chemical moieties for the observed activity of TDZD-8, and have also discovered analogs with improved anti-leukemia activity. Going forward, the most active derivatives are being optimized for water-solubility and will then be evaluated using primary human AML specimens engrafted into immune deficient mice. Based on evidence to date, we propose that the TDZD family of compounds may represent a new class of drugs for the treatment of leukemia and related hematologic malignancies. Disclosures:No relevant conflicts of interest to declare.

Inhibition of KSP by ARRY-520 induces cell cycle block and cell death via the mitochondrial pathway in AML cells

Kinesin spindle protein (KSP), a microtubule-associated motor protein essential for cell cycle progression, is overexpressed in many cancers and is a potential anti-tumor target. We found that inhibition of KSP by a selective inhibitor, ARRY-520, blocked cell cycle progression, leading to apoptosis in acute myeloid leukemia cell lines that express high levels of KSP. Knockdown of p53, overexpression of XIAP and mutation in caspase-8 did not significantly affect sensitivity to ARRY-520, suggesting that the response is independent of p53, XIAP and the extrinsic apoptotic pathway. Although ARRY-520 induced mitotic arrest in both HL-60 and Bcl-2-overexpressing HL-60Bcl-2 cells, cell death was blunted in HL-60Bcl-2 cells, suggesting that the apoptotic program is executed through the mitochondrial pathway. Accordingly, inhibition of Bcl-2 by ABT-737 was synergistic with ARRY-520 in HL-60Bcl-2 cells. Furthermore, ARRY-520 increased Bim protein levels prior to caspase activation in HL-60 cells. ARRY-520 significantly inhibited tumor growth of xenografts in SCID mice and inhibited AML blast but not normal colony formation, supporting a critical role for KSP in proliferation of leukemic progenitor cells. These results demonstrate that ARRY-520 potently induces cell cycle block and subsequent death in leukemic cells via the mitochondrial pathway and has the potential to eradicate AML progenitor cells.

The effect of cantharidins on leukemic stem cells

To identify an agent with specific activity against leukemic stem cells (LSCs), we evaluated compounds that targeted hepatic leukemia factor (HLF), a gene implicated in hematopoietic stem cell (HSCs) regulation, that we found overexpressed in LSCs. Cantharidin, a natural toxin from blister beetles, used as medicinal agent since antiquity, has been described to modulate the HLF competitor NFIL3 and is under clinical evaluation as an antitumor and antimetastatic agent. The molecule is not a substrate for multidrug resistant pumps and does not cause myelosuppression, and therefore it represents a promising compound for selective ablation of LSCs. Cantharidin and norcantharidin, a derivative with reduced toxicity, decreased HLF protein levels and induced apoptosis in the AML cell line MV4-11 by modulating the expression of several molecules that govern survival pathway, including HLF, SLUG, NFIL3 and c-myc, thereby inducing p53 and the mitochondrial caspase cascade. In vitro, cantharidin readily targeted primary AML stem and progenitor cells in contrast to conventional chemotherapeutic agents, such as Ara-C and daunorubicin, that mainly targeted more differentiated leukemic cells. In vitro the compound did not exhibit a therapeutic window, being equally toxic to normal HSCs and LSCs. In vivo cantharidin did not produce myelosuppression. Because of dose-limiting toxicity in vivo, neither cantharidin nor norcantharidin proved therapeutical benefit in AML xenograft models as a single agent. However, its potent in vitro LSC activity and pathway targeting may still be exploited clinically with a new generation of cantharidin derivatives or formulations and with appropriate drug combinations.

Leukemia-Derived DC for T-Cell Therapy against AML Progenitor Cells

Presentation of leukemic antigens (LAA) can be improved by conversion of leukemic cells to leukemia derived DC (DCleu), thereby enabling the generation of leukemia specific CTL. DC/DCleu can be generated and quantified from every AML case with at least one of 3 different DC generating methods (Schmetzer 2007/2008). We want to enlight the role of the composition and quality of DC and (DC or blast trained) T cells to mediate leukemia cytotoxic reactions or to predict the clinical response to therapy. Autologous patients', allogeneic donor T cells or T cells at relapse after SCT were trained with DC or blasts from 25 AML-cases in a ‘Mixed lymphocyte culture’ (MLC) and DC/T cell profiles and antileukemic Tcell cytotoxicity evaluated. We generated DC/mature DC/DCleu from every patient (Ø27/45/83%). DC training of T cells increased proliferating, CD4+ and memory T cells and decreased CD8+ T cells; blast training did not increase memory T cells. An antileukemic, very efficient T cell cytotoxicity was achieved in 47% of cases after DC/DCleu training but only in 24% after blast training of T cells. A comparison of cases with a gain of antileukemic T cell cytotoxicity to those without a lytic activity showed higher proportions of mature DC/DCleu and CD4/memory T cells and higher amounts of secreted IFNgamma and IL 6 in the lytically active, DC trained group. The differences were most distinct in the group with DC trained T cells prepared at relapse after SCT. Cases with a response to therapy showed higher proportions of DCleu, proliferating, memory or CD4+ T cells. We showed that >67% of all cases gained an antileukemic T cell cytotoxicity after DC training if >45% proliferating/>65% CD4+/>42% memory T cells or >40% mature DC/>65% DCleu were in the DC training setting. Moreover, 90% of DC trained T cells gained a lytic activity if >65% DCleu were in the MLC. AML patients presenting with a relapse after SCT showed better ex vivo convertibility of blasts to DCleu if they had responded to a GM CSF/DLI based therapy of their relapse after SCT compared to cases with no response (72 vs 36% blasts convertible to DCleu; 44 vs 29% generable DC). By spectratyping of the Vβ TCR region in an AML case we demonstrated a more extended clonal restriction of donor T cells after DC training of T cells compared to blast trained T cells. Moreover, the restricted pattern was also found in T cells from the patient after SCT. In summary, DC/DCleu can be generated in any given case independent from karyotype. A DC training of T cells improves the antileukaemic CTL, but can also mediate a T cell anergy. The composition of DC and T cells is predictive for the lytic efficiency of the trained T cells: A successful DC training of T cells is associated with high mature DC/DCleu counts and high rates of proliferating, CD4+ and memory T cells. Patients responding to a DLI/GM CSF based therapy are characterized by a better convertibility of blasts to DCleu and more mature DC. Identical clonal restrictions of T cells were found in blast trained and even more in DC trained T cells. Identical clonal patterns were found in ex vivo trained and in vivo selected T cells. We can contribute to understand biological mechanisms behind cytotoxic reactions and escape mechanisms and to develop adoptive immunotherapies with specific, antileukemia directed LAA specific T cells, e.g. selected by multimers from SCT donors or with specifically trained and selected T cells after DC training without side effects.

HDAC Inhibition by SNDX-275 (Entinostat) Restores Expression of Silenced Leukemia-Associated Transcription Factors Nur77 and Nor1 and of Key Pro-Apoptotic Proteins in AML

Orphan nuclear receptors Nur77 & Nor1 are highly conserved members of the nuclear receptor family 4. They are involved in various physiological and cellular functions. Both proteins are products of immediate early genes triggered by growth factors or chemokines. Induction and activation of Nor-1 and/or Nur77 have been implicated in apoptosis for negative selection of T-lymphocytes and activation induced cell death of macrophages. In a previous study (Nat Med. 2007 Jun; 13(6):730–5) we reported that double knockout of Nur77 and Nor1 in mice resulted in the rapid development of AML. The disease was characterized with quick death of mice (2–4 weeks), abnormal expansion of hematopoietic stem cells (HSCs) and myeloid progenitors, hepatosplenomegaly, and lymphadenopathy. Molecular analysis revealed that the absence of both Nur77 and Nor1 leads to the downregulation of AP1 transcription factors c-Jun and JunB, and extrinsic apoptosis mediators TRAIL and FasL. Real- time RT-PCR analysis of AML patient samples (n=16) revealed the universal down-regulation of Nur77 and Nor1 transcripts in AML progenitor and stem cells (CD34+38−). These findings indicate that Nur77 and Nor1 play roles in maintaining hematopoiesis and silencing of Nur77 and Nor1 is a critical step in AML development. To investigate the role of Nur77/Nor1 in AML, we hypothesized that silenced gene expression of Nur7 and Nor1 is mediated by histone deacetylases (HDACs). HDACs are enzymes involved in chromatin modifications that play an important role in the epigenetic regulation of gene expression and mediate abnormal epigenetic alterations associated with cancer, which make HDAC inhibitors promising anti-cancer drugs. SNDX-275, a benzamide HDAC inhibitor that preferentially inhibits activities of nuclear HDACs, including HDAC1, HDAC2, and HDAC3, has been shown to induce extensive apoptosis in various cancer types and has progressed to clinical phase I and II trials. We demonstrated that HDAC inhibition with SNDX-275 restored the expression of Nur77 and Nor1 in AML cell lines and primary AML samples (n=5) to levels comparable to those found in normal bone marrows. In the knockout mouse model, TRAIL and AP-1 transcription factors (c-Jun and JunB) were identified as downstream target genes of Nur77 and Nor1. In our present study, we observed strong upregulation by SNDX-275 of TRAIL, c-Jun and JunB in leukemia cell lines and primary AML blasts. As is common with HDAC inihibitors, apoptosis induced by SNDX-275 was associated with dramatic H3 and H4 acetylation. We further exploited the lack of Nur77 and Nor1 in leukemic stem cells (CD34+/38−) and demonstrated that SNDX-275 restored their expression resulting in apoptotic cell death of leukemia stem cells. Besides, pro-apoptotic Bim and Noxa were transcriptionally upregulated by SNDX-275 in AML cell lines, primary AML samples and also leukemia stem cells. Altogether, our results suggest a novel mechanism of HDAC inhibitor-induced apoptosis in AML which involves the restoration of the silenced nuclear receptors Nur77 and Nor1 and the activation of their downstream target genes, in addition to affecting apoptotic machinery. These findings also characterize HDAC-mediated transcriptional silencing of nuclear receptors as one of the key mechanisms driving the pathogenesis of AML.

Synergistic Induction of Apoptosis by Simultaneous Disruption of the Bcl-2 and mTOR/Akt Pathways in Acute Myeloid Leukemia.

Activation of the phosphatidylinositol 3-kinase (PI3K)/protein kinase B (Akt)/mammalian target of rapamycin (mTOR) signaling pathway provides survival signals for leukemic cells and blockade of this pathway may facilitate cell death. Akt/mTOR activation is a frequent event in AML that translates into significantly shorter survival of AML patients (S.M. Kornblau, ASH 2007). Bcl-2 family proteins are key regulators of apoptosis that are known to promote tumorigenesis and chemoresistance. We have recently reported that ABT-737, a small molecule BH3 mimetic, effectively kills AML progenitor cells via the disruption of Bcl-2:Bax, but its activity is largely diminished in cells overexpressing Mcl-1 (M. Konopleva, Cancer Cell 2006). We have now investigated anti-leukemia efficacy of concomitant blockade of mTOR/Akt, known to regulate Mcl-1 levels, and of Bcl-2 signaling. Combined use of rapamycin derivative temsirolimus and ABT-737 induced striking apoptotic responses in AML cell lines OCI-AML3 and MOLM13 leukemic cells. Mechanistically, temsirolimus downregulated expression of the anti-apoptotic protein Mcl-1 and induced expression levels of the proapoptotic BH3-only protein Noxa which specifically neutralizes Mcl-1. These effects were seen upon prolonged temsirolimus treatment that inhibited mTORC2 complex (Zeng, Blood 2007) and downregulated Akt phosphorylation. In primary AML samples co-cultured with bone marrow-derived stromal cells, ABT-737 at 25nM induced apoptotic cell death in CD34+ AML progenitor cells from 13 of the 16 samples tested. Temsirolimus/ABT-737 combination markedly enhanced apoptosis induction in 8/9 primary samples that were sensitive to temsirolimus alone (specific apoptosis, temsirolimus 15±5%; ABT-737, 28±5.9%; temsirolimus+ABT, 45±5.9%, p<0.001) and in 3/7 samples resistant to temsirolimus (p=0.7). Further, combined blockade of mTOR/AKT and Bcl-2 pathway effectively induced apoptosis in 3 AML samples resistant to ABT-737 alone. Taken together, these results indicate that temsirolimus can directly alter the Bcl-2 apoptotic rheostat via induction of Noxa and decreased Mcl-1 expression sensitizing leukemia cells to apoptosis induced by BH3 mimetics. The combined use of agents interfering with AKT/mTOR and Bcl-2 pathways may be a highly effective anti-leukemia strategy that induces apoptosis in AML progenitor cells in the context of bone marrow microenvironment.

5-Azacytidine Augments the Cytotoxicity of Mylotarg toward AML Blasts In Vitro and In Vivo.

Binding of CD33 on AML cells by monoclonal antibodies (mAb) mediates cytotoxicicity on AML cells modulated by the protein tyrosine kinase Syk and the phosphatase SHP-1. We have demonstrated that Syk- negative AML cells are relatively resistant to the effects CD33 ligation, but after exposure to the hypomethylating agent 5-azacytidine (5-aza) they become sensitive to the effects of both unconjugated and chemically-conjugated (gemtuzumab ozogamycin, GO) anti-CD33 mAb. Here we tested a panel of 40 primary AML samples for the expression of SHP-1. In 13% of the cases SHP-1 was undetectable. Anti-CD33 mAb and GO induced growth inhibition more effectively in AML cells that expressed SHP-1. Among SHP-1-positive samples, 69% demonstrated significant growth inhibition in response to CD33 ligation. In contrast, none of the SHP-1-negative samples responded to anti-CD33 mAb. These results show a correlation between SHP-1 expression and responsiveness of AML cells to CD33 ligation. However, 5-aza treatment restored SHP-1 expression and, therefore, increased the anti-proliferative effects of anti-CD33 mAb and GO. In 40% of SHP-1-negative samples, the AML cells were only marginally inhibited by 5-aza or anti-CD33 mAb alone, whereas the combination produced a more than additive effect in AML cells where 5-aza induced re-expression of SHP-1. The effect of GO was more than doubled by 5-aza in these cells. Total inhibition of DNA synthesis in the presence of 5-aza plus anti-CD33 mAb or GO reached 60–70% and was similar in SHP-1-positive and SHP-1-negative cells. Moreover, 5-aza significantly enhanced of GO-mediated cytotoxicity in AML progenitor cells. In a NOD/scid mouse model, which permits growth of human AML cells and allows measurement of in vivo therapeutic effects of therapeutic strategies for AML, we tested whether combined treatment of the mice with 5-aza would enhance the cytotoxicity of anti-CD33 mAb and GO. Suboptimal doses of 5-aza by itself, as well as treatment with murine anti-CD33 mAb alone did not cause significant cytoreduction. However, combined treatment of mice with anti-CD33 mAb and 5-aza, resulted in a significant response. Treatment with GO mediated up to 60% inhibition of AML cell proliferation. Combined treatment of mice with GO and 5-aza resulted in reduction of leukemia cells by >80%. These data show an interaction of 5-aza and anti-CD33/GO in an in vivo AML model. Based on these data, we hypothesize that the combination of 5-aza and GO may be a potent therapy for patients with AML. Moreover, Syk and SHP-1 may serve as biomarkers of leukemia cell response. Therefore, we initiated a clinical trial of 5-aza and GO combined therapy. Six patients with relapsed AML have been treated in a dose escalation of 5-aza preceeding GO (6 mg/m2 times two). All 6 were Syk-positive, while SHP-1 expression was detected in 4 samples and absent in two. 2 days of 5-aza treatment in vivo induced re-expression of SHP-1 in both previously SHP-1 negative patient cells. Moreover, significant increases were observed in the levels of Syk protein in one baseline positive sample and 1 SHP-baseline positive sample. Study of the effects of 5-aza alone ex vivo on the baseline patient cells showed no significant effect on leukemia cell proliferation. However, importantly, in all 6 samples 5-aza more than doubled the AML cells' response to cytotoxic effects of GO and naked anti-CD33 mAb. These results suggest that 5-aza may augment the effects of anti-CD33 mAb therapy through demethylation of SYK, SHP-1, and possibly other genes. The clinical efficacy of the combined therapy requires further study.

Defining the Leukemia Epigenome: Distinct Genome Wide Histone H3 Modification Patterns Exist in AML, ALL and Healthy Hematopoietic Progenitor Cells.

Aberrant transcriptional regulation plays a crucial role in the pathogenesis of acute leukemias. Chromosomal translocations and other mutations frequently affect transcription and chromatin regulation-associated genes in these diseases. Based on these findings we hypothesized that distinct global chromatin modification patterns exist that can distinguish between progenitor cells (CD34+ HSC) and acute leukemia as well as its subtypes. We used high density oligonucleotide ChIP-Chip assays querying more than 31,000 genomic loci to analyze global Histone H3 acetylation (H3Ac) and Lysine 9 (H3K9me3) trimethylation patterns each in a large number of AML (n=115), ALL (n=30), CD34+ HSC (n= 21) and peripheral blood cell (N=18) specimens. Class comparisons and predictions as well as unsupervised analyses were performed. Bioinformatic analyses led to histone modification maps across the genome with Histone H3 acetylation levels peaking around the predicted transcriptional start sites. More than 1000 loci differed in H3 acetylation and H3K9me3 between AML and CD34+ specimens (5% FDR). Among the regulatory classes over-represented among the altered genes were those involved in oncogenesis and cellular proliferation/differentiation. Histone H3 acetylation and H3K9me3 patterns also differed at hundreds of loci between ALL and AML samples (5% FDR). Genes involved in transcriptional regulation were significantly altered between the two leukemia subtypes. Specific differences in global chromatin modifications allowed support vector machine-based classification of CD34+ progenitor cells, AML, and ALL, with about 90% sensitivity and specificity based on Histone H3 acetylation patterns. A similar classification power was observed for H3K9me3. Within the AML patients, several groups of patients were identified that clustered together in unsupervised hierarchical cluster analysis due to similar histone modification patterns. These patterns did not primarily depend on patients′ karyotypes, indicating that it may be possible to define some, as yet unknown, different types of AML based on chromatin modification. On a global scale, acute leukemias are associated with specific histone modification patterns that distinguish AML from ALL and CD34+ hematopoietic progenitors. Taken together, these first data on the leukemia epigenome provide the basis for improved understanding of genes involved in leukemia pathogenesis.

Inhibition of KSP by ARRY-520 Induces Cell Cycle Block and Cell Death Via the Mitochondrial Pathway in AML Cells.

Kinesin spindle protein (KSP or Eg5), a microtubule-associated motor protein, plays an important role in establishing a bipolar spindle during mitosis and is essential for cell cycle progression. It has been demonstrated that inhibition of KSP prevents bipolar spindle formation leading to mitotic arrest and cell death. As such, a KSP inhibitor may have anti-tumor potential without toxicities associated with anti-microtubule agents such as taxanes. We found that KSP is highly expressed in various acute leukemic cells. Inhibition of KSP by a specific inhibitor, ARRY-520, at low nanomolar concentrations, blocked cell cycle progression and led to subsequent cell death in OCI-AML3, Molm13, HL-60, U937, and Jurkat cells. Knocking down p53 by p53shRNA in OCI-AML3 cells did not alter the effectiveness of ARRY-520. U937 cells overexpressing XIAP (U937XIAP) and Jurkat cells lacking caspase 8 (JurkatI2.1) showed unchanged sensitivities to ARRY-520 compared with their respective control cells, suggesting that cell cycle block and cell death induced by KSP inhibition is independent of p53 status, XIAP levels, and the activation of the extrinsic apoptotic pathway. However, although ARRY-520 blocked cell growth and induced mitotic arrest in both HL-60 cells and HL-60 cells overexpressing Bcl-2 (HL-60Bcl-2), cell death (determined by annexin V staining and changes in mitochondrial potential) was significantly abolished in HL-60Bcl-2 cells. These results suggest that cell death following cell cycle blockade by KSP inhibition is mediated through the intrinsic mitochondrial pathway. Furthermore, ARRY-520 induced the protein level of Bim, a proapoptotic BH3-only Bcl-2 family protein, prior to the activation of caspases in HL-60 cells. Although KSP inhibition by ARRY-520 had no effect on the survival of non-dividing AML blasts in vitro, ARRY-520 significantly inhibited the colony formation capacities of AML blasts, further supporting the critical role of KSP in cell proliferation. Our studies demonstrated that inhibition of KSP by ARRY-520 potently induces blockage of cell cycle progression which leads to cell death of various leukemic cells via the mitochondrial pathway and has the potential to eradicate AML progenitor cells.

The Endosteum Region Keeps Human Leukemic Stem Cells Alive

Only a subset of hierarchically arranged, primitive human AML progenitor cells can initiate the disease in transplanted immune-deficient mice. A publication in <i>Nature Biotechnology</i> by Ishikawa et al. (2007) describes a new animal model to study chemotherapy-resistant, quiescent human AML CD34⁺CD38⁻ stem cells, which are retained in the endosteum region.

Effect of histone deacetylase inhibitor valproic acid on progenitor cells of acute myeloid leukemia

Histone deacetylase inhibitor valproic acid (VPA) was recently shown to enhance proliferation and self-renewal of normal hematopoietic stem cells, raising the possibility that VPA may also support growth of leukemic progenitor cells (LPC). Here, VPA maintains a significantly higher proportion of CD34+ LPC and colony forming units compared to control cultures in six AML samples, but selectively reduces leukemic cell numbers in another AML sample with expression of AML1/ETO. Our data suggest a differential effect of VPA on the small population of AML progenitor cells and the bulk of aberrantly differentiated blasts in the majority of AML samples tested.

Primitive AML progenitors from most CD34+ patients lack CD33 expression but progenitors from many CD34− AML patients express CD33

AML blast populations are heterogeneous in their phenotype and functional properties, and contain a small subset of cells that regenerate leukemia in immunocompromised mice or produce clonogenic progeny in long-term cultures. This suggests the existence of a hierarchy of AML progenitor cells. CD33 is a myeloid marker absent on normal hematopoietic stem cells but expressed in about 75% of AML patients, and has been used for BM purging strategies and Ab-targeted therapies. These CD33 Ab therapies benefit only a minority of AML patients, suggesting that AML stem cells are heterogeneous in their CD33 expression. In order to evaluate this question, we determined expression levels of CD34 and CD33 on AML progenitors with long-term in vitro proliferative ability and NOD/SCID engrafting ability. The CD34(+) CD33(-) subfraction contained the majority of progenitors detected in vitro and most often engrafted the mice. This proliferation was leukemic from the CD34(+) AML patients, however from the CD34(-) AML patients only normal progenitors were detected in this fraction in some cases. These data suggest that most leukemic progenitors of CD34(+) patients do not express CD33. In contrast, CD34(-) AML primitive leukemic progenitors may be CD33(+). CD34(-) AML patients could potentially benefit most from CD33-targeted therapies or purging.

MMRI-23: A Potential Role in AML Antibody-Mediated Therapy.

Novel targets expressed on early AML precursor cell populations, with restricted expression on normal tissues, provide attractive options for antibody based therapeutics. We identified the CD300 family as Ig superfamily members with the capacity, through signaling motifs, to induce triggering and inhibitory signals to the leukocytes on which they are expressed. Here we report on an inhibitory member of this family, 35-L5, as a potential target for antibody-mediated therapy in AML. To assess its potential, we analyzed 35-L5 transcript and cell surface expression on normal blood cells, cell lines, and AML cells using quantitative RT-PCR and a monoclonal antibody (MMRI-23) made against a recombinant 35-L5 Ig fusion protein. RT-PCR showed its expression in normal PBMC was restricted to CD14+ monocytes, CD11c+ dendritic cells (myeloid DCs), monocyte derived dendritic cells (MoDC), with limited expression by B cells. MMRI-23 bound CD14+ monocytes, myeloid DC and MoDC, but not CD19+ B, CD4+ T, CD56+ NK cells and CD11c− (plasmacytoid) DC. Interestingly, MMRI-23 binding varied between donors and 35-L5 expression was upregulated on mature MoDC, when compared to immature MoDC from the same donor (3 from 6 donors). However, MMRI-23 binding to PBMC remained stable when cultured with activation agents PMA and ionomycin, LPS, or PHA for 24 and 48 hours. 35-L5 transcript and surface protein expression was evident in the monocytic derived U937, leukemic derived HL60, erythroleukemic derived HEL, CML derived K562, and B lymphoblast derived Daudi cell lines, but was limited in the KMH2 and HDLM2 cell lines and absent in the Jurkat, Raji, L428 and Mann cell lines. Preliminary studies in AML using quantitative RT-PCR showed that 35-L5 transcript was increased compared to control normal CD14 cell expression (n=9). Furthermore, MMRI-23 bound to 6 of 8 AML samples of the FAB M1, M2 and M5 subtypes. MMRI-23 crosslinking induced neither cell apoptosis, nor cell differentiation and change of CD14, CD33, CD64, CD16, CD11b expression. Defining the role of 35-L5 on normal and AML progenitor cells is necessary to evaluate its potential as a possible target for antibody-mediated therapy. MMRI-23 may have potential for drug conjugated antibody-mediated therapy, but 35-L5's inhibitory function may also be exploited to develop a novel mechanism for controlling myeloid leukemias.

Targeting Mutant Nucleophosmin (NPM)1: A Promising Approach To Induce Growth Inhibition and Differentiation in Human AML.

NPM1 is a nucleolar phosphoprotein that functions as a molecular chaperone for both proteins and nucleic acids and is capable of histone and nucleosome assembly. NPM1 shuttles between nucleus and cytoplasm, but is mutated and aberrantly localized to the cytoplasm in 35% of patients with AML. NPM1 is involved in regulating the activity of p53 and p14ARF, such that cells lacking NPM1 show decreased p14ARF. Mutant (m) NPM1 contains a 4-base insert that results in extra C-terminal residues encoding a nuclear export signal (NES). This causes mNPM1 to remain in the cytoplasm. In AML, mNPM1 has been correlated with the presence of FLT3-ITD and normal karyotype. NPM1 maps to 5q35 locus, which is often deleted in myelodysplastic syndrome. The role of the cytosolic mNPM1 in the pathogenesis of AML is not clear. In the present studies we determined the expression and biologic effects of mNPM1 in cultured and primary AML cells. Mutant nucleophosmin can be readily identified by reverse transcription followed by amplification with PCR primers specific for the mutant sequence. Sub-cellular fractionation and immunoblot analysis and/or immunoflourescent microscopy, as well as RT-PCR to detect the mNPM1 transcripts, confirmed that mNPM1 is localized in the cytosol in the cultured AML OCI-AML3 but is unmutated and nuclear in HL-60 cells. As compared to control siRNA, treatment with siRNA to NPM1 for 72 hours significantly induced p21 expression, inhibited cell proliferation and induced morphologic differentiation of OCI-AML3 cells. NPM1 siRNA did not induce significant apoptosis or loss of cell viability as determined by trypan blue uptake assay. OCI-AML3 cells co-express HoxA9 and Meis1 proteins, which are known to be leukemogenic in mouse myeloid progenitor cells. Notably, treatment with NPM1 siRNA depleted Meis1 and HoxA9 expressions in OCI-AML3 cells. Treatment with leptomycin B (5 nM for 48 hours), an inhibitor of the NES receptor CRM1/exportin-1, shifted the localization of mNPM1 from the cytosol to nucleus and also depleted the levels of Meis1 and HoxA9 in OCI-AML3 cells. Meis1 is known to transactivate FLT3 tyrosine kinase in AML progenitor cells, and a significant association of FLT-3-ITD with the presence mNPM1 has been documented. In the present studies, of the 16 primary AML samples, four that expressed FLT3-ITD also demonstrated expression of mNPM1. Taken together, these findings suggest that abrogation of the levels or reversing the aberrant cytosolic localization of mNPM1 may be a promising anti-leukemia treatment strategy, especially for targeting AML that co-express FLT3-ITD and mNPM1.

A Novel Orally Available Parthenolide Analog Selectively Eradicates AML Stem and Progenitor Cells.

We have previously demonstrated that Parthenolide (PTL), the principal component of the medicinal plant Feverfew, selectively induces apoptosis in AML stem cells while sparing normal counterparts. Recent reports show that PTL is also effective against CLL cells. Additionally, we observed PTL efficacy for ALL cells in vitro. These findings suggest that PTL-derived drugs may provide a unique means of leukemia therapy. A clinical study using Feverfew showed PTL plasma levels were not sufficient to achieve the concentration needed for AML targeting as established by our in vitro studies. Therefore, we synthesized a family of PTL derivatives designed to be more water-soluble. An amino analog, LC-1, retained the anti-leukemia properties of PTL with 1000× improved water solubility. In vitro studies indicate that LC-1 induces irreversible apoptosis of primary human AML cells within 8 h of treatment. Analysis of phenotypically primitive cells (CD34+CD38−, n=27) treated with 5.0–7.5 uM LC-1 for 18–24 h demonstrated an average cell kill of 85–90%. In contrast, normal CD34+CD38− cells showed less than 10% death under the same conditions. Similarly, colony-forming potential of AML specimens (n=5) was inhibited by >90% but by <20% for normal specimens (n=5). Further, engraftment of NOD/SCID mice was reduced by 91% for primary AML specimens (n=3), but only 28% for normal controls. Taken together, these data indicate LC-1 is a potent and selective agent for ablation of primitive AML cells. Moreover, we observed that similar concentrations of LC-1 were effective in ablating blast crisis CML (n=13) and ALL (n=7) specimens, for both total and phenotypically primitive cells. Molecular studies demonstrate that LC-1 induces a strong oxidative stress response characterized by rapid up-regulation of Nrf2 and its downstream target HO-1, followed by p53 activation and NF-kB inhibition. Thus, activation of stress responses and concomitant inhibition of survival mechanisms are evident for LC-1 treated AML cells, and represent potential biomarkers for drug activity. Pharmacodynamic studies of NOD/SCID mice 8 wks after transplantation with primary human AML cells indicate that oral dosing of LC-1 (50–100mg/kg) mediates a clear biological response. Microscopy studies show induction of Nrf2 and HO-1 in transplanted human AML cells within 1 h of treatment, while strong NF-kB inhibition is evident within 2 h. Finally, we also examined the activity of LC-1 for spontaneous acute canine leukemia. In vitro treatment of primary canine leukemia cells (n=4) with 7.5 uM LC-1 resulted in 80% apoptosis. Further, in a case study, a dog with advanced disease was treated for 3 consecutive days with increasing oral doses of LC-1 (25, 50, and 100 mg/kg). Within 5 days of the last treatment, CD34+ tumor cells were reduced from 45% to 3% of circulating WBCs. Thus, the animal demonstrated a clear hematologic response to LC-1 treatment. Based on these findings, we propose that LC-1 may be appropriate for future therapeutic regimens and has the potential to selectively target leukemic stem and progenitor cells in vivo.

Raf Inhibitor BAY 43-9006 Induces Bim Dephosphorylation and Activates the Intracellular Apoptotic Pathway in AML.

We have previously demonstrated constitutive activation of MAPK signaling in 70% of primary AML samples (Millela et al, JCI 108:851–859, 2001), suggesting that upstream kinases (Raf and MEK) may play a role in the leukemic transformation of myeloid cells. BAY 43-9006 is a small molecule Raf kinase inhibitor that has demonstrated potent anti-tumor activity against solid human tumors in xenograft models. In this study, we tested the hypothesis that BAY 43-9006 inhibits leukemia cell growth and/or induces apoptosis by suppressing the activity of the MAPK pathway. In the in vitro kinase assay, BAY 43-9006 inhibited both Raf-1 and B-Raf-mediated MEK1 phosphorylation in a dose-dependent manner, with Raf-1 kinase being more sensitive to the inhibitory effects of BAY 43-9006 (IC50Raf-1, 1.37 μM vs. IC50B-Raf, 4.64 μM). BAY 43-9006 suppressed MEK1/2 and ERK phosphorylation in the AML cell lines OCI-AML3, HL-60, U937 and KG-1 in a dose-dependent manner after 24 hr treatment. Unexpectedly, BAY 43-9006 also inhibited AKT phosphorylation on Ser473 (after 4.5 hrs). BAY 43-9006 inhibited growth of AML cells in a dose- and time-dependent manner. The 50% inhibitory concentration (IC50) of BAY 43-9006 was 0.39, 1.14, 2.86 and 2.80 μM, respectively in OCI-AML3, HL-60, U937 and KG-1 cells after 72 hrs. This growth-inhibitory effect was mediated by a dose-dependent induction of cell cycle arrest in G1 mediated by the down-regulation of the cell cycle-related proteins cyclin E, cdk2 and cdc2, followed by induction of apoptosis after 72 hrs. In primary AML patient samples, BAY 43-9006 not only inhibited cell growth and induced apoptosis after 48–72 hrs in vitro, but also preferentially inhibited colony formation of AML progenitor cells compared to normal bone marrow cells [IC50: 2.33 μM vs. 9.34μM (CFU-GM), 5.69 μM (Erythroid) and 3.75 μM (Mixed), respectively]. Time-course analyses demonstrated that BAY 43-9006 suppressed phosphorylation of the pro-apoptotic protein Bim (at 4.5 hrs), caused loss of the mitochondrial membrane potential and cytochrome c release (at 6 hrs) followed by cleavage of caspases-3 and -9 but not of caspase-8, suggesting primary involvement of the intrinsic mitochondrial pathway. Furthermore, the pro-apoptotic proteins Bim and Bax were up-regulated after 48 hrs of BAY 43-9006 treatment, and the level of the inhibitor-of-apoptosis protein Survivin was down-regulated after 48 hrs. In summary, our data demonstrates that BAY 43-9006 inhibits Raf-MEK-ERK signaling and induces apoptosis in AML via Bim de-phosphorylation and activation of the intrinsic apoptotic pathway. The potential of BAY 43-9006 in the therapy of AML patients will be tested in a Phase I clinical trial.

Triterpenoid Methyl-CDDO Is a Potent Inducer of Apoptosis in CD34+ AML Progenitor Cells Via Activation of SAPK Pathways and Inhibition of MAPK Cascades.

Outcome results in AML are a continued challenge for the development of novel therapeutic strategies. C-28 methyl ester of 2-cyano-3,12-dioxoolen-1,9-dien-28-oic acid, CDDO-Me, a novel triterpenoid, induces apoptosis in myeloid and lymphoid leukemic cell lines and in primary AML samples in sub-micromolar concentrations. We reported previously that CDDO-Me inhibits the activation of ERK1/2, blocks Bcl-2 phosphorylation, and promoted apoptosis in AML-derived U937 cells (Blood 2002, 99(1):326–35). Here, we examined the effects of CDDO-Me on CD34+ AML progenitor cells in vitro. Exposure to 1μM CDDO-Me induced apoptosis in all but one AML sample (46±4% annexin(+) CD34+ cells, n=20). To assess the anti-leukemia activity of CDDO-Me in vivo, scid mice intravenously injected with U937 cells were treated with liposomally-delivered CDDO-Me (20mg/kg/day IV every other day, starting at day 7, for a total of 9 injections). While CDDO-Me was not toxic to the mice, pathological and flow cytometry analysis revealed reduced (55–86%) leukemia burden in the bone marrow, liver, and spleen of mice. Since we had shown that CDDO-Me inhibits phosphorylation of pERK in U937 cells, a further goal of this study was to assess the role of ERK in CDDO-Me-induced cell death in primary AML samples. ERK was expressed and phosphorylated in all (n=15) patients' samples studied. CDDO-Me inhibited ERK phosphorylation in 9 of 15 patient samples and promoted apoptosis. However, CDDO-Me still induced apoptosis in 5/6 samples that displayed no significant changes in pERK levels in response to the drug. This finding suggests that ERK is not the sole target of the compound. The stress-activated protein kinases JNK and p38 are related to ERK but in general activate pathways that are opposed to ERK. By Western blot analysis, CDDO-Me induced early (30 min) phosphorylation of JNK and p38, which preceded induction of cell death. Pre-treatment of U937 cells with JNK and p38 inhibitors SP600125 and SB203580 partially abrogated induction of apoptosis, while MEK inhibitor PD-98059 significantly enhanced cytotoxicity. CDDO-Me induced p38 phosphorylation in 5 of 6 primary AML samples tested. Collectively, these finding indicate that CDDO-Me markedly shifts signaling toward the JNK and p38 MAPK stress-related pathways and away from the cytoprotective MAPK cascade. In summary, the triterpenoid CDDO-Me is a potent inducer of apoptosis in primary AML including CD34+ AML progenitor cells. Induction of apoptosis is in part mediated via inhibition of ERK signaling combined with JNK/p38 activation. These studies in primary AML samples and the anti-leukemia activity of the compound in vivo suggest potential utility of CDDO-Me for the treatment of AML patients.