MULTICENTERS. Uncontrolled and open phase II study. Evaluation of the effectiveness of a treatment associating 5 Azacytidine,Valproic acid ,Retinoic Acid at subjects-reached of syndromes myelodysplasia and acute MYELOID leukaemia Hematological response at 6 months Uncontrolled prospective cohort.

Chromatin Demethylation Apart from histone acetylation deacetylation, promoter hypermethylation is another important and relevant mechanism involved in gene transcription regulation (reviewed in Herman, 2003). Chromatin remodeling might thus be also targeted using nucleoside analogues, such as 5 azacytidine or decitabine, which reactivate gene transcription through DNA demethylation (Silverman, 2001). Recently, in in VITRO studies, induction of gene expression by 5 AzaC has been obtained in primary AML and MDS cells by DNA methylation dependent and independent mechanisms (SCHMELZ, 2005)

Again, AzaC has been demonstrated as capable to induce clinical hematological responses in patients with MDS. A controlled study conducted by the US Cancer and Leukemia Group B (CALBG)has reported a higher response rate, a lower incidence of leukemic transformation and a prolonged survival as compared with supportive care alone in these patients Silverman, 2002. Another confirmatory Phase 3 study is ongoing.

AzaC, in combination with valproic acid, in leukemic cell line (HL60 and MOLT4, has demonstrated a synergistic activity to induce gene reexpression (reactivation of p21 CIP1) and a synergistic effect in terms of growth inhibition, induction of apoptosis (Yang H, 2005).

Histone Acetylation Numerous investigator groups have tried to elucidate the molecular mechanisms underlying the ATRA induced differentiation in NB4 cells, fresh APL cells, APL mice, or APL patients (Melnick 1999). One of the main issue was to understand the crucial role of the PML RARα fusion protein in the differentiation response to RA. It was observed first that therapeutic concentrations of ATRA resulted in the reformation of PML nuclear bodies associated with a cleavage of the PML RARα fusion protein. Disappearance of this fusion product which acts as a dominant negative regulator of RA target genes transcription gave an explanation for the rerun of the differentiation process. The dominant negative role of PML RARα was secondly explained by the association of the fusion protein to the N-CoR-SMRT-Sin3 corepressor complex, leading to histone deacetylase (HDAC) activities recruitment and to the lack of target genes transcription (REDNER, 1999). Of interest, a similar recruitment of corepressor HDAC activities has been reported in other fusion gene leukemia, including PLZF RARα ,AML1 ETO, CBFß MYH11, and TEL AML1 acute leukemia. In PML RARα APL cells, therapeutic concentrations of ATRA allow the release of corepressor HDAC activities, histone acetylation, chromatin remodeling, and transcription of target genes potentially responsible for terminal granulocytic differentiation (REDNER, 1999; DILWORK, 2001). From this point of view, ATRA therapy of APL is the first example of a gene targeted therapy which specifically targets pathogenic genetic abnormalities in a human leukemia.

In VITRO and in vivo resistance to ATRA-induced differentiation observed in patients with PLZF RARα leukemia has been related to a more potent recruitment of corepressor HDAC activities in this APL subset, as compared to classical PML- RARα APL (two corepressor binding sites on PLZF instead of one on PML). Very interestingly, it has been recently demonstrated that PLZF RARα leukemic cells are not actually completely resistant to differentiation induction, especially if appropriate COSTIMULI are given. First, these cells can differentiate in the presence of higher concentration of ATRA (3 microM instead of 1 microM).

Secondly the addition of a HDAC inhibitor (trichostatin A) restores the ATRA sensitivity at 1 microM (KITAMURY 2000).

Thirdly G CSF signaling may force these cells to undergo terminal differentiation (JANSEN 2001).

HDAC inhibitors have also been shown as able to induce remission in transgenic models of therapy resistant acute promyelocytic leukemia (He 2001).

The sensitivity of HL60 cells, which do not display any chromosomal rearrangement involving the RARα locus, to RA induced differentiation may be related to these observations. One may hypothesize that some unknown COSTIMULI including chromatin remodeling events contribute to the RA sensitivity of HL60 cells.

From a therapeutic point of view, these observations lead to evaluate non targeted transcriptional therapies combining ATRA with non targeted HDAC inhibitors and/or cAMP inducers in non APL leukemias. Among the known HDAC inhibitors (trichostatin A, trapoxin A, butyrate, oxamflatin, depsipeptide, and MS 275), sodium phenylbutyrate has been successfully administered in combination with ATRA to a patient with clinically ATRA resistant APL (WARRELL, 1998). This case report represents the first example of a targeted transcriptional therapy in a human leukemia. It has recently been demonstrated that valproic acid (VPA) belongs to the HDAC inhibitor family (PHIEL, 2001). Valproic acid is a short chained fatty acid widely used as an anticonvulsant and mood stabilizer.

The characteristic delay in response to VPA and its teratogenic potential had led for a long time to the proposal that it acts through modulation of gene expression. It has also been reported that VPA can activate AP1-dependent transcription (ASGHARI, 1998; Chen, 1999 1; Yuan, 2001) and upregulate bcl 2 (Chen, 1999 2). VPA was also recently demonstrated as capable to induce differentiation of F9 teratocarcinoma cells, which are known to be also capable to differentiate in the presence of RA or cAMP (WERLING, 2001). Finally, VPA might sensitize neoplastic cells to pro apoptotic stimuli through an inhibition of glutathione (GSH) reductase, an enzyme required for maintaining high cellular levels of reduced GSH (Moog, 1996), or an inhibition of the NF-κB pathway (ICHIYAMA, 2000). Interestingly, it was also shown that lithium chloride (another mood stabilizer) acts synergistically with ATRA to induce terminal differentiation of WEHI-3B leukemia cells. As observed with the combination of ATRA and G CSF, this observed synergism appeared to be related to the prevention of RAR protein loss usually observed under ATRA exposure (Finch, 2000).

Of interest, VPA as single agent or administered in combination with ATRA has been recently demonstrated as capable to induce clinical hematological responses in patients treated for myelodysplastic syndromes (KUENDGEN, 2004). In this study, a pretreatment with VA seemed to be required for further positive effects of ATRA.

Retinoic Acid :

Retinoids represent a large group of compounds structurally related to vitamin A (retinol). They act through binding to and activating specific nuclear receptors, which bind the DNA. Retinoic acid (RA), the natural acidic derivative of retinol, is a key differentiating factor involved in specific phases of the embryonic development, differentiation of the visual system, and (of interest here) hemopoietic granulocytic maturation (CORNIC, 1994). In VITRO, RA was demonstrated as capable to induce granulocytic differentiation of the HL60 cell line. This cell line was established in 1977 from a patient with AML. The cells largely resemble promyelocytes but can be induced to differentiate terminally. Some reagents, including RA, cause HL60 cells to differentiate to granulocyte-like cells, others to monocyte/macrophage-like cells. The HL60 cell genome contains an amplified c-myc proto-oncogene and c-myc mRNA levels decline rapidly following induction of differentiation (BIRNIE, 1988).

Recurrent alterations of the gene coding for the RA alpha receptor (RARα, located on the chromosome 17q12) are associated with some acute myeloid leukemia (AML) subsets. The most common RARα gene alteration is the reciprocal t(15;17) chromosomal translocation observed in the vast majority of acute promyelocytic leukemia (APL) corresponding to the AML-M3 subset of the French-American-British (FAB) classification. This translocation fuses the RARα gene to the PML gene (located on the chromosome 15q21), resulting in PML- RARα fusion products. Variant translocations fusing the RARα gene with other partners including PLZF on chromosome 11, NPM on chromosome 5, NuMA on chromosome 11, and STAT5 on chromosome 17, have been rarely or occasionally reported. The causal role of the RARα fusion proteins (at least PML- RARα and PLZF- RARα) in APL has been demonstrated in murine models (KOGAN, 1999). It has been shown that these fusion proteins may act as a dominant transcriptional repressor in APL cells (Melnick, 1999).

The NB4 cell line was established in 1991 from a patient with APL (LANOTTE, 1991). This cell line has been widely used to study the biology of this disease. Conversely to HL60, NB4 is carrying the t(15;17) translocation. As in HL60 cells, RA is capable to induce granulocytic differentiation of NB4 cells leading to cell death through terminal induction of apoptosis. These differentiating effects have been confirmed in vivo using different murine models of APL (KOGAN, 1999; He, 1999).

Based on these pre-clinical observations, oral all-trans RA (ATRA) has been successfully administered to APL patients. In patients with relapsing APL, front-line therapy with ATRA (45 mg/m2/day) induces in vivo differentiation of leukemic promyelocytes into abnormal granulocytes still carrying the PML- RARα fusion, resulting in approximately 90% hematological remission and 20% molecular remission rates as assessed by specific PML- RARα RT-PCR negativation (Huang, 1988; CASTAIGNE, 1990; DEGOS,1995). This represented the first example of a differentiation therapy in a human leukemia. Unfortunately, such beneficial effects have not been observed in patients with other AML subtypes when treated with ATRA using similar dosage and schedule. Following these results obtained in relapsing APL patients, ATRA was then evaluated in combination with chemotherapy during front-line treatment of newly-diagnosed APL patients. Several controlled trials have established the ATRA-chemotherapy combination as the current standard therapy for newly-diagnosed APL. Apparently, the best results are obtained when ATRA and chemotherapeutic agents are administered simultaneously.

Other therapeutic interventions might be considered to increase the RA sensitivity in RA-resistant cells. Actually, explanations for the RA resistance include RA-induced increased expression of cytochrome P450 isoforms (CYPs), RA-induced increased expression of cytoplasmic RA-binding proteins type II (CRABP-II), overexpression of P-glycoprotein (P-gp), and acquired mutations of the ligand-binding region of the RARα gene. Agents interacting with ATRA metabolism, such as HIV-1 protease inhibitors (indinavir, ritonavir, saquinavir) which inhibit CYPs and P-gp and compete with ATRA for CRABP-I binding, could enhance the induction of differentiation in RA-resistant cells (IKEZOE, 2000).

Inclusion Criteria:

• Patients ≥ 18 years

• high Risk Acute Myelogenous Leukemia (FAB-M3 excluded), including :

• AML in first relapse in patients with secondary AML(after MDS and CMML)

• AML in first relapse in patients with a CR duration < 12 months

• Second Relapse or > 2

• de novo AML without previous treatment in elderly patients (FAB-M3 excluded) , if :

  • 70 years

• with de novo AML or secondary AML (Transformation of myelodysplasia)

• Unfit for Intensive chemotherapy

• High risk myelodysplasia, including :

• RAEB or t-RAEB (FAB)

• With IPSS score Intermediate-2 or High risk (Greenberg, 1997)

• non eligible for allogeneic HSC transplantation

• Women of childbearing potential (WOBP) must be using an adequate method of contraception

• Men with WOBP have to use an acceptable method to avoid pregnancy

• Signed Written informed consent

Exclusion Criteria:

• APL(FAB)
• Clinical CNS involvement
• Uncontrolled infectious disease
• Adequate hepatic function defined as total bilirubin < 3 times ULN ALAT and ASAT < 2.5 times ULN
• Adequate renal function (serum creatinine < 1.5x ULN anc Creatinine clearance < 25ml/min)
• Included in an other clinical trial
• Previous treatment with 5-aza &/or Valproic acid &/or retinoic acid
• Positive pregnancy test
• Women who are breastfeeding