SEARCH     advanced search my account e-alerts subscribe register
Journal home
Advance online publication
Current issue
Press releases
For authors
For referees
Contact editorial office
About the journal
For librarians
Contact NPG
Customer services
Site features
NPG Subject areas
Access material from all our publications in your subject area:
Biotechnology Biotechnology
Cancer Cancer
Chemistry Chemistry
Dentistry Dentistry
Development Development
Drug Discovery Drug Discovery
Earth Sciences Earth Sciences
Evolution & Ecology Evolution & Ecology
Genetics Genetics
Immunology Immunology
Materials Materials Science
Medical Research Medical Research
Microbiology Microbiology
Molecular Cell Biology Molecular Cell Biology
Neuroscience Neuroscience
Pharmacology Pharmacology
Physics Physics
Browse all publications
20 September 1999, Volume 18, Number 38, Pages 5278-5292
Table of contents    Previous  Article  Next   [PDF]
In vivo analysis of the molecular pathogenesis of acute promyelocytic leukemia in the mouse and its therapeutic implications
Li-Zhen He, Taha Merghoub and Pier Paolo Pandolfi

Department of Human Genetics and Molecular Biology Program, Memorial Sloan-Kettering Cancer Center, Sloan-Kettering Division, Graduate School of Medical Sciences, Cornell University, 1275 York Avenue, New York, NY 10021, USA

Correspondence to: Pier Paolo Pandolfi, Department of Human Genetics and Molecular Biology Program, Memorial Sloan-Kettering Cancer Center, Sloan-Kettering Division, Graduate School of Medical Sciences, Cornell University, 1275 York Avenue, New York, NY 10021, USA


Acute promyelocytic leukemia (APL) is characterized by the expansion of malignant myeloid cells blocked at the promyelocytic stage of hemopoietic development, and is associated with reciprocal chromosomal translocations always involving the retinoic acid receptor alpha (RARalpha) gene on chromosome 17. As a consequence of the translocation RARalpha variably fuses to the PML, PLZF, NPM and NUMA genes (X genes), leading to the generation of RARalpha-X and X-RARalpha fusion genes. The aberrant chimeric proteins encoded by these genes may exert a crucial role in leukemogenesis. Retinoic acid (RA), a metabolite of vitamin A, can overcome the block of maturation at the promyelocytic stage and induce the malignant cells to terminally mature into granulocytes resulting in complete albeit transient disease remission. APL has become, for this reason, the paradigm for `cancer differentiation therapy'. Furthermore, APL associated with translocation between the RARalpha and the PLZF genes (PLZF-RARalpha) shows a distinctly worse prognosis with poor response to chemotherapy and little or no response to treatment with RA, thus defining a new APL syndrome. Here we will focus our attention on the recent progresses made in defining the molecular mechanisms underlying the pathogenesis of this paradigmatic disease in vivo in the mouse. We will review the critical contribution of mouse modeling in unraveling the transcriptional basis for the differential response to RA in APL. We will also discuss how this new understanding has allowed to propose, develop and test in these murine leukemia models as well as in human APL patients novel therapeutic strategies.


APL; transgenic mice; cell differentiation; retinoic acid; HDACI


Nonrandom, somatically acquired chromosomal translocations or inversions are the most common genetic lesions in acute leukemia (Rabbitts, 1994). Specific chromosomal rearrangements have been linked to distinct subtypes of leukemia or associated with particular stages of disease progression or prognostic outcomes. These structural rearrangements frequently involve genes encoding transcription factors whose altered functions can interfere with regulatory cascades that have been demonstrated to be critical for controlling the growth, differentiation and survival of normal blood cell precursors (Look, 1997). Acute promyelocytic leukemia (APL), has been well characterized at the molecular level to become one of the most compelling examples of aberrant transcriptional regulation in cancer pathogenesis, as well as a unique example of successful osmosis between molecular biology research and clinical research at the bedside.

APL, M3 subtype of acute myeloid leukemia (AML), accounts for more than 10% of all AMLs and is characterized by three distinctive and unique features (reviewed in Warrell et al., 1993b; Grignani et al., 1994; Pandolfi, 1996; Kalantry et al., 1997; Melnick and Licht, 1999): (1) the accumulation in the bone marrow of tumor cells with promyelocytic features; (2) the invariable association with specific translocations which always involve chromosome 17; and (3) the exquisite sensitivity of APL blasts to the differentiating action of retinoic acid (RA). From this point of view APL has become the paradigm for therapeutic approaches utilizing differentiating agents. This therapeutic approach is conceptually new in that it does not involve chemical or physical agents to eradicate the tumor by `killing' the neoplastic cells, but rather reprograms these cells to differentiate normally (Figure 1). However, although effective, treatment with RA alone in APL-1 patients induces disease remission transiently and relapse is inevitable if remission is not consolidated with chemotherapy. In addition, in the majority of cases, relapse is accompanied by RA resistance (Warrell et al., 1993b; Grignani et al., 1994; Pandolfi, 1996; Kalantry et al., 1997; Melnick and Licht, 1999).

At the molecular level APL is associated with reciprocal translocation always involving the retinoic receptor alpha (RARalpha) gene on chromosome 17 that translocates and fuses to four distinct partner genes (for brevity, hereafter referred as X genes; Figure 2). In the vast majority of cases RARalpha fuses to the PML gene (promyelocytic leukemia gene, originally named myl) on chromosome 15 (de The et al., 1991; Goddard et al., 1991; Kakizuka et al., 1991; Pandolfi et al., 1991). In a few cases RARalpha fuses to the promyelocytic leukemia zinc finger (PLZF) gene, to the nucleophosmin (NPM) gene, or to the nuclear mitotic apparatus (NuMA) gene located on chromosome 11, 5 or 11, respectively (Chen et al., 1993a,b; Redner et al., 1996; Wells et al., 1997).

It is of note that from a clinical stand point, APL associated with the t(11;17)/PLZF-RARalpha shows a distinctly worse prognosis with poor response to chemotherapy and little or no response to treatment with ATRA (Licht et al., 1995). The observation that this subset of APL patients respond differentially to RA prompted a comparative analysis of the phenotype observed in PML-RARalpha and PLZF-RARalpha transgenic mice, and of the transcriptional properties of these two fusion proteins, that was ultimately decisive in redefining our understanding of the molecular mechanisms underlying APL pathogenesis (see below).

PML is a member of the RING finger family of proteins (de The et al., 1991; Goddard et al., 1991; Kakizuka et al., 1991; Pandolfi et al., 1991; Reddy et al., 1992; Lovering et al., 1993). PLZF gene encodes a transcription factor with transcriptional repressive activity, and is a member of the POK (POZ and Krüppel) family of proteins which shares an N-terminal POZ motif and a C terminal DNA binding domain made by Krüppel-like C2-H2 zinc-fingers (Chen et al., 1993b; Bardwell and Treisman, 1994; Li et al., 1997a). NPM is a major nonribosomal nucleolar phosphoprotein (Feuerstein et al., 1988a,b). NuMA is a nuclear matrix associated protein which is critical for coordination of mitosis (Gaglio et al., 1995; Wells et al., 1996). Evidence recently has been accumulating in support of a role for the various X proteins in the control of the transduction of the mitogenic, differentiating and survival signals (Pandolfi, 1996; Kalantry et al., 1997; Wang et al., 1998a,b; Melnick and Licht, 1999).

RARs are members of the superfamily of nuclear hormone receptors which act as RA inducible transcriptional activators, in their heterodimeric form with retinoid-X-receptors (RXRs), a second class of nuclear retinoid receptors (Chambon, 1996). The invariable involvement of RARalpha made APL the first convincing example of aberrant transcriptional regulation in cancer pathogenesis (Warrell et al., 1993b; Grignani et al., 1994; Pandolfi, 1996; Melnick and Licht, 1999). In the absence of RA, RAR-RXR heterodimers can repress transcription through histone deacetylation by recruiting nuclear receptor co-repressors (N-CoR or SMRT), Sin3A or Sin3B, which in turn form complexes with histone deacetylases (HDAC1 or 2), thereby resulting in nucleosome assembly and transcriptional repression (Grunstein, 1997). RA causes the dissociation of the co-repressors complex, and the recruitment of transcriptional co-activators to the RAR-RXR complex, thus resulting in the activation of gene expression which, in turn, can induce terminal differentiation and growth arrest of cells of various histological origin including normal myeloid hemopoietic cells (Smith et al., 1992; Gudas et al., 1994).

The various translocations of APL are always balanced and reciprocal resulting in the formation of X-RARalpha and RARalpha-X fusion genes and the co-expression of their products in the leukemic blasts (Figure 2). These X-proteins have no structural similarities (Figure 2). At a first glance, it therefore appears that the contribution of the various RARalpha partners to the two chimeric products is, at a structural level, dramatically different and that the only common feature among the fusion proteins is the presence of the RARalpha domains, suggesting that the disruption of the RARalpha function is the major and only cause of APL. Despite this diversity, the various X-RARalpha fusion molecules have the capacity of heterodimerizing with either PML, PLZF, NPM or NuMA, since the regions which can normally mediate X proteins homodimerization are retained in the fusion moiety (Liu and Chan, 1991; Kalantry et al., 1997; Perez et al., 1993; Bardwell and Treisman, 1994; Wells et al., 1997). Similarly, the RARalpha portion is able to mediate heterodimerization with RXR, as well as DNA and ligand binding through its RA and DNA binding domains (Perez et al., 1993). Therefore these fusion products have invariably the potential ability to interfere with both X and RAR/RXR pathways (Figure 3a).

However, the initial information accumulated on the function of the fusion proteins of APL was solely based on studies performed in various leukemic cell lines, which, obviously, do not constitute an ideal system for the study of oncogenic transformation. Key questions remain to be answered such as whether the X-RARalpha fusion proteins are indeed the major oncogenic players in APL, whether they are necessary and/or sufficient for leukemogenesis, and whether they can be regarded as identical RARalpha mutants in view of their identity in their RARalpha moiety, or if, on the contrary, they are biologically distinct. This is particularly relevant in view of the differential response to RA in t(11;17) APL, which raises the corollary question if and how the X-RARalpha protein would directly mediate differential response to RA. Finally, the role of RARalpha-X in APL pathogenesis, if any, remains to be explored. The recent use of mouse models to study APL is proving instrumental in addressing these questions.

Modeling APL in the mouse

X-RARalpha proteins are necessary but not sufficient to cause leukemia

In order to elucidate, in vivo, the possible leukemogenic role of the fusion proteins present in APL (X-RARalpha and RARalpha-X) several mouse models have been generated. We and others have generated a human cathepsin-G (hCG) minigene expression vectors to direct the expression of the X-RARalpha and RARalpha-X transgenes to the promyelocytic cellular compartment in mice (Grisolano et al., 1994, 1997; He et al., 1997; Cheng et al., 1999), or took advantage of the hMRP8 expression cassette which also drives myeloid specific expression in transgenic mice (Lagasse and Weissman, 1994; Brown et al., 1997). The cathepsin G gene expression, whose product accumulates in the granules found in the cytoplasm of myeloid progenitors and granulocytes, peaks at the promyelocytic stage of myeloid differentiation while the MRP8 gene is expressed at high levels in early myeloid progenitors but also in more mature myeloid cells (Figure 4).

hCG-PML-RARalpha transgenic mice: Approximately 25% of our hCG-PML-RARalpha transgenic mice developed acute leukemia and died within 4 weeks after full-blown leukemia displayed in peripheral blood (PB) (He et al., 1997, 1998a). In the vast majority of cases leukemia occurred after 12 months of age. The leukemia was characterized by a profound leukocytosis, anemia, thrombocytopenia and extensive organ infiltration by leukemic cells, including bone marrow, spleen, liver, lymph nodes, kidneys, reproductive organs and lungs. The leukemic cellular population in bone marrow and spleen consisted of myeloid blasts, promyelocytes and myelocytes that partially retained the ability to terminally differentiate toward mature granulocytes. Flow cytometric analysis with Gr-1 and Mac-1 markers, whose expression increase while myeloid cellular differentiation progresses, confirmed the morphological observations (He et al., 1997, 1998a). It is interesting to note that in the leukemic phase, hCG-PML-RARalpha mice showed a marked accumulation of C-Kit (a cell surface marker for hematopoietic precursors) positive cells in the spleen. Leukemia was preceded by a myeloproliferative disorder of variable duration characterized by the accumulation of immature myeloid cells in the spleen and bone marrow (He et al., 1997, 1998a). Transplantation of bone marrow or spleen cells collected from hCG-PML-RARalpha transgenic mice with leukemia or myeloproliferative disorder caused leukemia in the recipient nude mice (He et al., 1997, 1998a).

Grisolano et al. reported that 100% of their hCG-PML-RARalpha transgenic mice displayed altered myeloid development, manifested by myeloid expansion in bone marrow and spleen as well as by splenic extramedullary hematopoiesis, but normal peripheral blood counts and myeloid differentiation (Grisolano et al., 1997). Over the course of 6 - 13 months, 30% of the transgenic mice from three different founder lines developed AML. These leukemia showed similar hematological features to our hCG-PML-RARalpha leukemia, i.e., profound leukocytosis, extensive organ infiltration by leukemic cells with disruption of normal histologic architecture, markedly increased numbers of promyelocytes, but partial myeloid maturation in PB and bone marrow (Grisolano et al., 1997).

hMRP8-PML-RARalpha transgenic mice: The hMRP8-PML-RARalpha mice generated by Brown and colleagues developed APL at 3 - 9 months of age (Brown et al., 1997). The features of leukemia in these mice were similar to those in hCG-PML-RARalpha mice on some points, especially the extensive infiltration of leukemic cells. However, the hCG-PML-RARalpha and hMRP8-PML-RARalpha mice differed in several aspects. First of all, the leukocytosis in the PB was not observed. Secondly the leukemic cells were namely promyelocytes with low expression levels of Gr-1 and Mac-1 antigens. Once again, prior to overt leukemia, the hMRP8-PML-RARalpha mice showed a `preleukemic state', characterized by lower levels of Gr-1 antigen expression and lower granularity (lower side-scatter detected by flow cytometry) of the peripheral blood and bone marrow cells (Brown et al., 1997). Transplantation of leukemic cells into unirradiated FVB/N mice, same strain as the leukemic donor, and the preleukemic cells into lethally irradiated FVB/N mice, resulted in leukemia with identical characteristics to the donor (Brown et al., 1997). In addition, all the hMRP8-PML-RARalpha mice developed epidermal papillomas before or simultaneously to the leukemia onset. This further demonstrates the neoplastic activity of PML-RARalpha, but complicated the analysis of the natural history of the disease rendering even more difficult the maintenance and propagation of these transgenic lines.

Collectively, these data demonstrate that PML-RARalpha plays indeed a critical role in APL pathogenesis since the mice eventually develop leukemia, but it is not sufficient to trigger full blown leukemia since in all these transgenic models leukemia does not occur at birth and is preceded by a preleukemic period of variable length where the cells do not invade the PB nor other organs. Thus, other events have to occur in order to transform the aberrant hemopoiesis observed in the various transgenic lines in full blown leukemia.

The X-RARa fusion proteins are biologically distinct RARalpha mutants

We have generated hCG-transgenic mice for all the four X-RARalpha fusion genes of APL (He et al., 1997, 1998a, and unpublished data). This gave us the opportunity to study, on a comparative basis, the biological activity of these molecules and to determine if they cause identical or different phenotypes (He et al., 1997, 1998a). hCG-PLZF-RARalpha transgenic mice developed leukemia and died between 6 and 18 months of age. This leukemia also displayed a dramatic leukocytosis accompanied with modest anemia and thrombocytopenia in the PB, but were characterized by the infiltration in all organs examined of leukemic cells that fully retained the capacity to terminally mature. Flow cytometric analysis showed that the majority of the leukemic cells in bone marrow and spleen were Gr-1 and Mac-1 double positive. In addition, the expression level (fluorescence intensity) of both Gr-1 and Mac-1 were higher than those in hCG-PML-RARalpha leukemia. Furthermore, unlike in hCG-PML-RARalpha leukemias, the percentage of C-Kit positive cells was the same as in wild-type mice. Thus, the leukemia in hCG-PLZF-RARalpha mice resembled human chronic myeloid leukemia (CML) more than classical APL. In agreement with our observation, Cheng et al. (1999) recently confirmed that the leukemia in their hCG-PLZF-RARalpha transgenic mice were characterized by CML like morphological features. The leukemia were once again preceded by a myeloproliferative disorder of various duration (He et al., 1997, 1998a). In this phase, proliferation of myeloid precursors increased, but the cells retained the ability to terminally differentiate into granulocytes. Various mature stages of myeloid cells progressively accumulated in the bone marrow and spleen, and in some cases the peripheral blood. As for the hCG-PML-RARalpha mice, transplantation of bone marrow or spleen cells from hCG-PLZF-RARalpha mice with leukemia or myeloproliferative disorder caused leukemia in the recipient nude mice.

More recently, Cheng et al. reported that hCG-NPM-RARalpha mice develop leukemia after a period of latency exceeding 1 year. The leukemia in hCG-NPM-RARalpha mice were characterized by heterogeneity in cytology/pathology, with a spectrum of manifestations from typical APL to CML-like syndrome (Cheng et al., 1999).

Thus, the various X-RARalpha mutants in spite of their identity in the RARalpha portion do not represent identical RARalpha mutants. This difference is particularly prominent in comparing hemopoiesis of hCG-PLZF-RARalpha and hCG-PML-RARalpha transgenic mice. In particular the hCG-PLZF-RARalpha fusion protein fails to cause the block at the promyelocytic stage which is distinctive of APL. In human t(11;17) blasts, the differentiation block at the promyelocytic stage might be conferred by some additional events such as the coexisting RARalpha-PLZF fusion protein.

The X-RARalpha fusion proteins directly mediate differential response to RA

Administration of RA to hCG-PML-RARalpha leukemia mice at a dose equivalent to the one that would be utilized for the treatment of human APL patients induced, as in human APL, complete albeit transient disease remission (Brown et al., 1997; Grisolano et al., 1997; He et al., 1997). The cells from hCG-PML-RARalpha leukemia differentiated upon RA treatment in vitro and in vivo (Brown et al., 1997; Grisolano et al., 1997; He et al., 1997). The bone marrow cells of hCG-NPM-RARalpha were also sensitive to RA treatment in vitro, as reported for human patients harboring t(5;17) (Cheng et al., 1999).

On the contrary, while RA could prolong the survival of the hCG-PLZF-RARalpha leukemic mice, complete remissions were never achieved, precisely as observed in human t(11;17) APL upon RA treatment (He et al., 1998a).

These data provide conclusive evidence demonstrating that, in vivo, X-RARalpha directly mediate differential response to RA (Brown et al., 1997; Grisolano et al., 1997; He et al., 1997, 1998a).

Differences and similarities between human and murine APL

Taken together, the various mouse models of APL generated by various laboratories develop leukemias with features which closely mimic human APL (Brown et al., 1997; Grisolano et al., 1997; He et al., 1997, 1998a; Cheng et al., 1999). For instance, PML-RARalpha or NPM-RARalpha mice are responsive to RA treatment, while PLZF-RARalpha mice develop RA-resistant leukemia (Brown et al., 1997; Grisolano et al., 1997; He et al., 1997, 1998a; Cheng et al., 1999). However the disease observed in some of these APL models exhibited several discrepancies when compared to the human APL. For example, while human APL is characterized by leukopenia, leukemias in the various hCG-transgenic mice are characterized by high white cell counts (Grisolano et al., 1997; He et al., 1997, 1998a; Cheng et al., 1999). In addition, the infiltrating cellular population partially retains the ability to terminally differentiate towards mature granulocytes, which is not observed in hMRP8-PML-RARalpha mice nor in human APL (Grisolano et al., 1997; He et al., 1997, 1998a; Cheng et al., 1999).

Implications for the molecular pathogenesis of APL

As aforementioned, the leukemic phenotype in the various X-RARalpha transgenic mice was preceded by a variable latency or long pre-leukemic period and no mice were leukemic at birth. In addition, in all the transgenic lines leukemia has an incomplete penetrance with the exception of the PLZF-RARalpha mice. Several interpretations which are not mutually exclusive have been proposed in order to explain these findings: (1) the reciprocal translocation (RARalpha-X) is also required for full-blown leukemogenesis; (2) other mutations (or secondary hits) are required; and (3) the relative level of expression of the X-RARalpha transgene with respect to the wild-type X and RARalpha genes is not appropriate. From this point of view it is important to remember that in the blasts of human APL, in view of the translocation the normal gene dosage of X and RARalpha is reduced to heterozygosity, whereas in the X-RARalpha transgenic mice the two normal copies of the gene X and RARalpha are still intact. This aspect is of particular relevance if the X-RARalpha was to act as dominant negative inhibitors of RARalpha and X functions, since, in this case, the products of the two normal RARalpha and X genes could antagonize X-RARalpha leukemogenetic potential in the transgenic system (Figure 3a).

In order to test whether the increase in the expression level of the X-RARalpha transgene or the decrease in the dose of the gene X and RARalpha, i.e., changes of the relative expression level of wild-type genes and fusion genes, would potentiate the leukemogenic capacity of the APL specific fusion molecules, two experiments can be performed taking advantage of the availability of both the transgenic models and X and RARalpha knock out mice (Lufkin et al., 1993; Barna et al., 1998; Wang et al., 1998b): firstly, to test the relation between transgene expression level and the leukemia development, it is possible to analyse the latency and leukemia features between homozygous and heterozygous transgenic mice within each transgenic line; secondly, to test the relation between the reduction of normal X and RARalpha dosages and leukemogenesis, it is possible to analyse the latency and leukemia features in X-RARalpha/X+/-/RARalpha+/- and X-RARalpha/X-/-/RARalpha-/- mutants. Having already generated PML-/-, PLZF-/- mice (Barna et al., 1998; Wang et al., 1998b) and having obtained RARalpha mice (Lufkin et al., 1993) we are currently performing these experiments in a systematic manner. Preliminary results obtained interbreeding hCG-PML-RARalpha transgenic mice with PML-/- reveal a significant increase in the frequency of leukemia in both PML+/- and PML-/- background cohorts, as well as an earlier onset of the disease (Wang et al., 1998a). Furthermore, mice homozygous for the hCG-PML-RARalpha transgenes also develop leukemia at higher incidence and earlier presentation (Wang et al., 1998a). The intercrossing between PLZF-/- with hCG-X-RARalpha is underway. The acceleration of leukemia observed reducing the dose of PML could be due on one hand to the change of the ratio between the fusion gene and the wild-type PML gene since, as above mentioned, X-RARalpha can act as a dominant negative on X function, or on the other hand due to haploinsufficiency or loss of function of the PML protein per se. X proteins, at least PML and PLZF, have been in fact shown to function as growth/tumor suppressors and to promote cell death (Mu et al., 1994; Koken et al., 1995; Quignon et al., 1998; Shaknovich et al., 1998; Wang et al., 1998d; Yeyati et al., 1999). PML-/- mice are highly susceptible to carcinogen-induced tumor development though no spontaneous leukemia or tumors observed (Wang et al., 1998b). Moreover, PML has been demonstrated to be a critical component of the RA pathway required for the ability of RA to induce growth inhibition and or myeloid differentiation (Wang et al., 1998b). However, even in PML-RARalpha/PML-/- mice leukemia does not occur at birth but after 4 - 6 months. Thus, the reduction of RARalpha to heterozygosity and/or the presence of the reciprocal RARalpha-PML fusion protein, or the presence of other mutational events may further contribute to leukemogenesis. We are presently testing these various possibilities.

RARalpha-X proteins play a role in leukemogenesis

The reciprocal RARalpha-PML fusion protein consists of RARalpha transactivation domain A (Chambon, 1996), fused to the COOH-terminus of PML to which no discrete biochemical functions have been attributed to date (Goddard et al., 1991; Kakizuka et al., 1991; Alcalay et al. 1992; Chang et al., 1992) (Figure 3b). RARalpha-PML transcript are present in 70 - 80% of APL patients with t(15;17) (Alcalay et al., 1992; Chang et al., 1992; Grimwade et al., 1996). No apparent difference has been observed in RA sensitivity or clinical outcomes in APL patients who do or do not harbor the RARalpha-PML transcript (Grimwade et al., 1996; Li et al., 1997b). APL patients with prolonged remissions do not appear to express the PML-RARalpha transcript, as evaluated by RT - PCR, but are in some cases still positive for the expression of the RARalpha-PML fusion transcript which can be nevertheless attributed to a differential sensitivity of the RT - PCR assay (Tobal et al., 1995). Interestingly, however, an APL patient whose blasts harbored the RARalpha-PML but not the PML-RARalpha fusion gene was also identified (Lafage-Pochitaloff et al., 1995). Therefore, if RARalpha-PML is necessary and/or even sufficient for leukemogenesis is still an open question. We and others generated RARalpha-X transgenic mice (Pollock et al., 1997; He et al., 1998b). Pollock et al. generated hCG-RARalpha-PML transgenic mice which they further interbred with their hCG-PML-RARalpha transgenic mice to determine whether RARalpha-PML play a role in leukemogenesis and to assess whether expression of both fusion genes might affect the latency and/or penetrance of leukemia. They concluded that the expression of RARalpha-PML alone did not lead to APL, but that co-expression of RARalpha-PML with PML-RARalpha increased the penetrance and the onset of leukemia development (Pollock et al., 1997).

In all analysed cases of t(11;17) APL the reciprocal RARalpha-PLZF transcript was consistently found expressed (Licht et al., 1995; Grimwade et al., 1997). The RARalpha-PLZF fusion protein contains the RARalpha transactivation domain A fused to the last seven zinc fingers of PLZF (Licht et al., 1995; Grimwade et al., 1997) (Figure 3b). As previously mentioned, PLZF is a transcription factor expressed in early hematopoietic cells and down-regulated during hemopoietic differentiation (Reid et al., 1995). Its DNA binding domain has been mapped to the last five zinc fingers (Sitterlin et al., 1997), while its N-terminal POZ/BTB domain interacts with N-CoR/SMRT/HDAC and mediate transcriptional repression (Grignani et al., 1998; He et al., 1998a; Lin et al., 1998). Unlike the wild-type PLZF, RARalpha-PLZF can bind DNA, but can no longer repress transcription, thus, in principle, interfering with the transcriptional repressive activity of PLZF by competitive binding to its responsive elements (Li et al., 1997a; Sitterlin et al., 1997; Ivins and Zelent, 1998; Yeyati et al., 1999). To date, PLZF responsive elements have been identified in regulatory elements of several genes such as Cyclin A, Hox B2 and HOXD11 genes (Barna et al., 1998; Ivins and Zelent, 1998; Yeyati et al., 1999). hCG-RARalpha-PLZF transgenic mice have been generated and preliminarily characterized (He et al., 1998b). In a 2 year follow-up, no hCG-RARalpha-PLZF mice developed full-blown leukemia. Automatic and differential counting of peripheral blood were also found normal. However, hCG-RARalpha-PLZF mice showed splenomegaly with marked extramedullary hematopoiesis. Differential counts and flow cytometric analysis of bone marrow and spleen revealed abnormal hematopoiesis with increased percentages of myeloid cells, macrophages and eosinophils. This observation demonstrates that RARalpha-PLZF, albeit not sufficient for leukemogenesis, can impair myelopoiesis. Intercrosses between PLZF-RARalpha transgenic mice and RARalpha-PLZF transgenic mice are ongoing and will tell to which extent RARalpha-PLZF can modify and/or accelerate leukemogenesis in PLZF-RARalpha mice. It must be remembered, in fact, that leukemia in PLZF-RARalpha mice lacks the classical block of differentiation at the promyelocytic stage which characterizes APL. It will be also interesting to test whether RARalpha-PLZF could further contribute to the unresponsiveness to RA observed in t(11;17) APL.

The various methodological approaches in modeling APL in the mouse and their biological implications

Classical transgenic approach

In principle, the generation of a transgenic animal which carries in the germ line a desired mutation exogenously introduced (transgene) by physically microinjecting the `unicellular' mouse embryo (fertilized egg) is still to date the most straightforward approach available in order to test, in vivo, the pathogenic potential of any aberrant gene product (Boyd and Samid, 1993). Since the genetic alterations associated with cancer, such as chromosome translocations in the case of leukemia, are acquired and somatic lesions which occur in a specific cell type, the key point for this approach is to identify and utilize appropriate expression vectors in order to direct the expression of the aberrant gene to a specific target tissue or to a restricted cell compartment, thus mimicking the human malignancy of interest in the experimental animal.

The first successful attempt of modeling leukemia in the experimental mouse was carried out testing the in vivo leukemogenic potential of the BCR-ABL fusion gene (Heisterkamp et al., 1990). Two different isoforms of BCR-ABL, P210 and P190, are found in the blasts of human chronic myeloid leukemia (CML) and of acute lymphoblastic leukemia (ALL) patients respectively, due to various breakpoints in BCR locus (Shtivelman et al., 1985; Mes-Masson et al., 1986; Griffiths et al., 1992). Initial attempts to express a BCR-ABLP210 transgene under the control of human BCR promoter failed because of deleterious effects of the construct on embryogenesis (Heisterkamp et al., 1991; Voncken et al., 1995). However, when the expression of BCR-ABL was driven by the metallothionein (MTT) promoter, which supposedly drives in vivo unrestricted and Zn inducible transgene expression, MTT-BCR-ABL transgenic mice were viable and developed T- or B-lymphoid leukemia (Heisterkamp et al., 1991). Nevertheless, none of these transgenic lines developed classical CML even with the BCR-ABLP210 isoform, which in human is strictly associated with CML (Voncken et al., 1995).

While `ubiquitous' expression of BCR-ABLP210 or BCR-ABLP190 is compatible with adult life and eventually results in leukemia, the same is not true in the case of the PML-RARalpha fusion protein associated with APL, as its unrestricted or untimely expression in transgenic mice is either incompatible with life or ineffective. In fact, our initial attempts to generate PML-RARalpha transgenic mice utilizing ubiquitous promoters such as the one from the beta-actin gene or CMV was unsuccessful precisely because it invariably resulted in embryonic lethality (He et al., 1997, and unpublished observations). The generation of transgenic mice harboring PLZF-RARalpha under the control of the CMV or the SV40 promoter also resulted in neither expression of the fusion gene, nor in leukemogenesis (Cheng et al., 1999). Using the MTT promoter to drive the expression of PML-RARalpha, the transgenic mice did not develop leukemia either, but did show liver pathology including the development of hepatocellular carcinomas after 5 days zinc stimulation (David et al., 1997). The PML-RARalpha protein was found expressed in the altered tissues at much higher levels than in the surrounding normal liver tissues. In addition, overexpression of PML-RARalpha was accompanied by a strong proliferative response in hepatocytes (David et al., 1997). These data on one hand confirmed the oncogenic role of PML-RARalpha, on the other hand, further corroborated the importance of using more appropriate and tissues specific promoters in order to reproduce the APL phenotype in mouse models.

Several options are available to direct in vivo the expression of an aberrant gene product to the lymphoid compartment. The CD2 locus control region (Badiani et al., 1996) and the Lck proximal and distal promoters, as examples, retain regulatory elements which have been tested to be sufficient to drive in vivo high levels as well as T-cell specific expression of the transgene (Voncken et al., 1995; Wildin et al., 1995), thus providing an invaluable instrument for testing in the mouse the activity of aberrant gene products associated to ALL and/or lymphoma pathogenesis. Furthermore, utilizing the aforementioned regulatory elements it is also possible to restrict the expression of the transgene either to immature progenitors (e.g. utilizing the Lck proximal promoter which is active in thymocytes (Wildin et al., 1995)), or to more mature lymphoid cells (e.g. utilizing the Lck distal promoter which is active in mature thymocytes and peripheral T lymphocytes (Wildin et al., 1995)).

However, the search for the appropriate myeloid specific promoter in order to model APL in the mouse proved not to be an easy task, which prompted our own group and others to test and develop alternative expression vectors for in vivo transgene expression (Figure 4). For example, transgenic mice for the PML-RARalpha fusion gene under the control of the myeloid specific CD11b promoter which drives transgene expression relatively late in myeloid maturation (Figure 4), did not develop leukemia (Early et al., 1996). However, these mice displayed a defect in myeloid precursor response to cytokines and a profound neutropenia after sublethal irradiation (Early et al., 1996). Similarly, the restricted expression of PML-RARalpha to the CD34+/hemopoietic stem cell/early myeloid compartment by means of c-Fes expression vector was not leukemogenic either (P Greer, personal communications) (Figure 4). The comparative meta-analysis of these results along with the results obtained in X-RARalpha mice which do develop leukemia, lead to the conclusion that the lineage- and stage-specific expression of the fusion genes of APL is critical for their leukemogenic potential.

In conclusion, the conventional transgenic approach constitutes a very powerful tool to study leukemogenesis in vivo, but it has several limitations: (1) frequently an appropriate expression vector to target the expression of aberrant gene in the `right' cell compartment is not available; (2) the expression level of the transgene may be influenced along with the promoter activity by the copy number and the integration site of the transgene; and (3) the endogenous wild-type counterparts of the genes involved in the mutation are still present in normal dosage. This aspect is of particular relevance when the aberrant gene product is thought to interfere with the normal function of its parental genes.

Modeling APL by a `knock in' approach

To circumvent the disadvantages encountered utilizing a conventional transgenic approach a number of investigators have used an alternative strategy referred to as `knock in' approach to introduce an aberrant gene in the mouse germline. In this approach the mutated gene is targeted directly to a predefined locus in the mouse genome by homologous recombination in mouse embryonic stem (ES) cells. These cells are subsequently utilized to generate a mutant mouse/embryo and/or chimeras (Castilla et al., 1996; Yergeau et al., 1997; Okuda et al., 1998; Pandolfi, 1998). This approach has been used to replace a normal gene with an aberrant version of the same gene, thus closely mimicking what occurs in the altered cells in human disease. For example, the `knock in' approach has been successfully utilized to target the BCR-ABLP190 fusion gene into the BCR locus and to generate chimeric mice which developed B-cell ALL, exactly reproducing in the mouse the phenotype observed in human patients (Castellanos et al., 1997). Similarly, AML1-ETO fusion cDNA resulting from the t(8;21) observed in the AML-M2 subtype of myeloid leukemia was used to replace the normal AML1 gene in order to model this AML-M2 subtype in the mouse. However, this replacement resulted in the embryonic lethality of the heterozygous mutants due to the block of fetal hemopoiesis, with a phenotype similar to the one observed in the AML-1 `knock out' homozygous mice (Yergeau et al., 1997; Okuda et al., 1998). Despite the impossibility of characterizing the leukemogenic role of AML1-ETO, the `knock in' experiment was extremely informative since it supported the notion that the fusion protein could act in a dominant negative manner thus blocking AML1 function (Okuda et al., 1998). We have utilized the `knock in' approach to generate mice harboring the fusion PLZF-RARalpha targeted to the PLZF locus (Merghoub, unpublished observation). This experiment will be crucial to determine if the fusion protein is indeed dominant over X and RARalpha functions. If this is the case, the PLZF-RARalpha `knocked-in' heterozygous mutants should show the phenotype observed in PLZF and RARalpha null mutants. We are presently characterizing the resulting mice and chimeras.

The `knock in' approach has also been utilized to target the aberrant gene into a desired locus other than the involved gene itself. In this case, the gene to be targeted is selected on the basis of its expression pattern which should be restricted to the cell type where the aberrant gene is found to be expressed in the human disease. Ideally, the locus to be targeted should be expressed exclusively in that cell type. In this respect, a successful example of targeting by `knock in' the PML-RARalpha to the murine cathepsin G locus has recently been reported (Westervelt et al., 1997). Two types of ES cells were generated, cells with the `PGK neo' selectable marker left within the targeted locus and cells where the `PGK neo' gene was excised using the Cre-loxP system (see next section). Recombinant ES cells with either an intact (+) or excised (-) `PGKneo' cassette were used to generate mice. No leukemia was observed in PGKneo (+) animals over a period of 57 - 80 weeks. However, `PGKneo' (-) `knock in' mice had a 77% cumulative probability of leukemia by 71 weeks of age, compared with 15% leukemia incidence by the same age observed in a concurrent cohort of hCG-PML-RARalpha transgenic mice (P<0.05) (Westervelt et al., 1997). Thus, the expression of the knocked-in PML-RARalpha under the control of murine Cathepsin G promoter caused leukemia with higher incidence than previously reported in standard transgenic experiments. Phenotypic characteristics of leukemia in the `knock in' animals were similar to those of hCG-PML-RARalpha transgenic animals, including leukocytosis, anemia, thrombocytopenia, massive hepatosplenomegaly with extensive leukemic cell infiltration, and accumulation of promyelocytes in the marrow and spleen. The difference between the leukemia onset in transgenic and `knock in' mice might be explained by the relatively higher level of expression driven by the promoter of the murine Cathepsin G gene. The failure of PML-RARalpha to cause leukemia when the `PGKneo' cassette was retained in the locus can be explained by a silencing effect due to the transcriptional interference of the `PGKneo' gene with the expression of nearby genes (the Cathepsin G gene in this case). This `neighborhood effect' which was previously described (Pham et al., 1996), must be taken into account in the design of animal models through this strategy.

Obviously, the classical `knock in' approach offers several advantages: (1) the appropriate level of expression of the aberrant gene. The aberrant gene which replaces the normal gene is in fact under the control of its natural promoter thus closely mimicking what is occurring, at the somatic level, in the leukemic cell, and avoiding the tedious search for the `appropriate' promoter; (2) the relative dosage of the wild-type counterpart (at least for one of them) and the aberrant gene is, in principle, comparable and their expression should be concordant; (3) the elimination of the variegation effect due to the random insertion of the aberrant gene at different location of the genome in transgenic mice thus eliminating variability of expression among founders.

Nonetheless, the `knock in' approach presents several fundamental limitations, too. The translocations associated with leukemia occur at the somatic level, and, presumably, leukemia has a clonal origin arising from a single hemopoietic progenitor, while the fusion gene introduced by `knock in' is carried in the germ cell as an inheritable character. Therefore, unlike somatic translocation, which results in the expression of an aberrant gene in a specific subset of cells, the `knock in' results in the expression of the altered gene in all tissues and throughout the developmental stages of the embryo where the targeted gene is normally found expressed. Many of the genes involved in cancer associated translocations play an important role not only in hemopoiesis but also in general development. Thus, the disruption and the expression of the fusion genes in other tissues may result in more broad phenotypes or even cause early embryonic lethality. Hence, the major disadvantage of `knock in' approach is the lack of tissue specificity.

Cre-LoxP-mediated, tissue specific, transgene expression and chromosomal translocations

The Cre-loxP system represents an ideal instrument to manipulate chromosomal organization and to control gene expression in vivo, in a tissue specific manner, and at a specific differentiation moment. Cre, the recombinase of bacteriophage P1, is able to mediate the specific excision of DNA, which is flanked by loxP recognition sites, from the genome of mammalian cells in culture and transgenic mice (Kilby et al., 1993; Zou et al., 1994). The experimental procedure to generate mouse models by using this system consists of three distinct phases: (1) generate a transgenic animal in which the gene of interest is silent because of an intervening sequence flanked by two loxP sites which is placed between the promoter and the gene; (2) generate transgenic animals in which the Cre gene is expressed in a tissue specific manner; and (3) intercross the transgenic mice generated in step 1 with those generated in step 2. Animals generated from these breedings will be transgenic for both constructs. As a result, Cre will excise the loxP-separated intervening sequence, moving the promoter close to the coding region of the gene and resulting in the expression of the gene only in the tissue or cells in which the Cre gene is expressed. The main advantage of this three-step approach is that it allows the dissociation of the `tissue specificity' of expression from the `level' of expression since the level of expression is controlled by the promoter that drives the expression of the altered gene while the tissue specificity is conferred by the promoters that drives the expression of the Cre gene. It is thus possible to achieve high levels of expression in a tissue specific manner and overcome the embryonic lethality caused by an unrestricted expression of the faulty gene. We have generated transgenic mice where the PML-RARalpha fusion gene would fall under the control of either the hCMV immediate early gene enhancer/promoter or the human beta-actin promoter when a loxP-flanked Lac-Z reporter/spacer gene is removed by the tissue specific activity of the Cre recombinase. We are interbreeding these mice with several Cre transgenic mice to further study the oncogenic potential of PML-RARalpha.

The Cre-loxP system has also been used to generate tissue specific (conditional) gene knock-out mutants where a loxP flanked critical region of a gene is excised after expression of Cre in the tissue of interest (Gu et al., 1993; Sauer, 1998). A similar approach has been utilized, in the case of the APL `knocked in' mice, to remove the `PGKneo' gene cassette from the cathepsin G locus in ES cells, thus resulting in the expression of PML-RARalpha and the consequent leukemia phenotype (Westervelt et al., 1997).

Recently, several reports have explored the possibility of using Cre-loxP system for chromosome manipulation in mammalian cells in vivo, breaking ground for chromosome engineering in the mouse (Ramirez-Solis et al., 1995; Smith et al., 1995; Van Deursen et al., 1995; Lewandoski and Martin, 1997; Herault et al., 1998). The strategy involves the generation of a mouse where loxP recombination sequences will be introduced by `knock in' approach in trans, into two distinct pre-defined chromosomal loci and the interbreeding of this mouse with a transgenic mouse where the expression of Cre is under the control of a tissue specific promoter. This will result in a site-specific recombination between the loxP sites, leading to the desired chromosomal rearrangement in the Cre expressing cells. Depending on the position and relative orientation of the two loxP sites, different recombination products can be generated, such as deletions, duplications, inversions and translocations (Ramirez-Solis et al., 1995; Smith et al., 1995). Thus, the Cre-loxP system represents a powerful instrument to generate a chromosomal translocation in vivo, in a tissue specific manner. For instance, to generate APL mouse models by this approach, mice carrying X and RARalpha genes flanked by loxP sites introduced by a `knock in' approach, will be crossed with transgenic mice expressing the Cre recombinase under the control of a specific promoter such as hCG promoter. The resulting mice will harbor a reciprocal translocation (X-RARalpha and RARalpha-X) that occurs only in promyelocytes. The advantage of this system is that the translocation and the simultaneous expression of the resulting fusion proteins (X-RARalpha and RARalpha-X) occur at the somatic level as in human APL patients. Thus, the generation of a reciprocal and somatic translocation in the mouse might lead to generation of the `perfect' mouse model of the human disease which may constitute the ideal tool for drug testing in vivo. However, a very important limitation of this approach might be constituted by the low frequency of Cre mediated chromosomal translocation, although, in principle, a single event should be enough to lead to neoplastic transformation.

Transgenic models of APL for the development of new therapeutic strategies

APL has become the paradigm for `cancer differentiation therapy' being the first convincing example of a successful application of a therapeutic strategy which targets the activities of a specific oncoprotein with a cell differentiation agent, RA (Warrell et al., 1991, 1993b; Grignani et al., 1994; Pandolfi, 1996). RA induces complete remission in 70 - 95% of APL patients (Fenaux and Degos, 1997; Tallman et al., 1997). The reasons for the spontaneous RA resistance in the remaining APL patients have not been elucidated. It has been demonstrated that APL characterized by translocations between PLZF and RARalpha genes do not respond or respond poorly to RA (Licht et al., 1995). Moreover, upon prolonged RA treatment, RA responsive APL eventually becomes RA resistant (Warrell et al., 1991). As aforementioned, X-RARalpha transgenic mice have been used to test response to RA and the results are consistent with human APL patients (Brown et al., 1997; Grisolano et al., 1997; He et al., 1998a). PML-RARalpha mice attained transient complete remission upon RA treatment, while PLZF-RARalpha mice, at the same dose of RA, never achieved remission though RA treatment prolonged survival (He et al., 1998a). However, in both PML-RARalpha and PLZF-RARalpha mice, leukemia relapsed becoming RA insensitive (He et al., 1998a). Thus, X-RARalpha mice represent an invaluable model system to study the mechanisms underlying resistance to RA in APL as well as to develop therapeutic approaches to overcome RA resistance and to potentiate the therapeutic effects of RA.

RA and cytokine associations for APL treatment

The potential synergism between RA and cytokines is a controversial issue. RA and cytokine combinations, including interferon (IFN), granulocyte colony-stimulating factor (G-CSF), interleukin (IL)-1alpha, IL-4, tumor necrosis factor (TNF), were found to increase differentiation of HL60 and U937 cells though these cytokines had no differentiating effect by themselves (Peck and Bollag, 1991; Nason-Burchenal et al., 1996). Lack of TNFalpha expression or the expression of IL-3, G- or GM-CSF by blasts from APL patients significantly reduced the differentiating potential of RA in APL cells (Dubois et al., 1994). It was reported that IFN could restore RA sensitivity in a relapsed APL patient and complete remission was achieved after combination therapy (Koller et al., 1991), but this result could not be confirmed in other patients (Warrell, 1993a). It is known that RA can lead to the degradation of the PML-RARalpha fusion protein (Raelson et al, 1996; Koken et al., 1994, 1999; Weis et al., 1994), while IFN stimulates PML, as well as PML-RARalpha expression (Nason-Burchenal et al., 1996, Lavau, et al., 1995). The potential efficacy of RA and IFN in combinations can be precisely and reliably tested in the various X-RARalpha mice.

More promising data have been obtained utilizing RA and G-CSF combinations. Cassinat et al. reported that these drugs in combination enhanced the terminal differentiation of fresh t(15;17) positive APL blasts (Cassinat et al., 1998). Striking granulocytic maturation was also observed in vitro when APL blasts from a t(11; 17) patient were cultured with RA and G-CSF though treatment with a single agent did not induce differentiation of the leukemic cells. When the patient was administered RA and G-CSF at leukemia relapse, in vivo, granulocytic maturation and consequent complete remission at hematological, cytogenetic and RT - PCR levels was achieved (Jansen et al., 1998). In addition, bone marrow cells isolated from hCG-PLZF-RARalpha leukemic mice responded much better to RA when cultured in a 2% murine spleen cell conditioned medium (MSCCM) containing IL-3 and GM-CSF (unpublished observation). These observations encourage investigations aiming at evaluating the efficacy of combination of RA and hematopoietic growth factors in RA resistant leukemia. Therapeutic trials with RA and G-CSF combinations in hCG-PLZF-RARalpha leukemic mice are currently ongoing.

Arsenic trioxide (As2O3) and RA/As2O3 combinations

Arsenic trioxide (As2O3), a chemical used in traditional Chinese medicine, has recently been proven to be extremely effective in the treatment of APL. About 90% of APL patients treated with As2O3 alone achieved complete remission, even though they were resistant to RA and/or conventional chemotherapy (Shen et al., 1997; Soignet et al., 1998). In vitro studies on APL cell lines or fresh APL cells showed that As2O3 could act as a growth inhibitory and pro-apoptotic agent (Chen et al., 1997; Shao et al., 1998; Wang et al., 1998c). Low dose of As2O3 may trigger partial differentiation of the leukemic blasts (Chen et al., 1997). As2O3-induced apoptosis might be mediated by down-regulation of BCL-2 gene expression, upregulation of the expression of the proenzymes of caspase 2 and 3 and activation of both caspase 1 and 3 (Chen et al., 1996; Soignet et al., 1998; Wang et al., 1998c). However, the information obtained from in vitro studies is controversial in respect to two main questions: (1) whether As2O3 works specifically and only on APL cells harboring PML-RARalpha or if it can be utilized also for the treatment of APL harboring variant translocations; and (2) whether combinations of As2O3 and RA act in a synergistic or antagonistic manner in the treatment of APL.

It was proposed that the degradation of PML-RARalpha would be critical in mediating the biological effects of As2O3 in APL (Chen et al., 1996). As2O3 may induce the degradation of the PML-RARalpha protein through ubiquitination of the PML moiety (Chen et al., 1996; Kamitani et al., 1998; Puccetti et al., 1998; Wang et al., 1998c). However, NB4-306 cells, a RA-resistant cell line derived from NB4 that no longer expresses the intact PML-RARalpha fusion protein, responded to As2O3 as the parental NB4 cells (Shao et al., 1998; Wang et al., 1998c). These findings suggest that As2O3 inhibit growth and induce apoptosis in PML-RARalpha independent manner and that this drug could also be effective in t(11;17) APL and other malignancies. However, so far no data are available in t(11;17) patients. It has been reported that As2O3 failed to induce apoptosis and differentiation in vitro in PLZF-RARalpha positive cells isolated from a relapsed t(11;17) APL patient (Koken et al., 1999). Therapeutic trials in PLZF-RARalpha transgenic mice using As2O3 will provide conclusive information on whether As2O3 would be effective in APL which do not harbor PML-RARalpha.

RA and As2O3 are both very effective anti APL drugs. RA triggers differentiation while As2O3 induces both apoptosis and a partial differentiation (Wang et al., 1998c; Gianni et al., 1998). They would therefore be expected to synergize. In fact, APL cell lines resistant to one agent are sensitive to the other (Wang et al., 1998c; Gianni et al., 1998). However, it has been proposed that As2O3 and RA might act through different and opposing mechanisms. On one hand, RA would induce differentiation which requires the transcriptional activity of PML-RARalpha. On the other hand, As2O3 would induce the degradation of the PML-RARalpha protein in turn inducing apoptosis. As2O3 and RA could therefore antagonize each other, limiting the possibility to utilize combinations of these compounds in clinical trials (Shao et al., 1998). Studies using fresh leukemia cells from patients with APL showed that As2O3 interfered with RA-induced differentiation. Co-treatment with both RA and As2O3 reduced the percentage of differentiated cells, in agreement with the decrease in type II transglutaminase activity, which had been shown to be increased during RA-induced differentiation in NB4 cells (Shao et al., 1998). To test the benefit of combining RA and As2O3 in vivo, PML-RARalpha leukemic mice and syngenic grafts of leukemia blast from PML-RARalpha transgenic mice have been used to evaluate and compare the efficacy of RA or As2O3 alone and their combination (Rego et al., 1998; Lallemand-Breitenbach et al., 1999). The results demonstrated that As2O3 alone induced apoptosis and modest differentiation of leukemic cells and prolonged mouse survival. Combining As2O3 with RA accelerated tumor regression and dramatically prolonged survival through enhanced differentiation and apoptosis (Lallemand-Breitenbach et al., 1999).

RA and histone deacetylase inhibitors (HDACIs)

Recent studies have proposed a model for the mechanisms of transcriptional silencing by the X-RARalpha fusion proteins, which can explain both the molecular pathogenesis of APL and the differential response to RA (He et al., 1998a; Lin et al., 1998; Grignani et al., 1998; Guidez et al., 1998). The X-RARalpha fusion proteins homodimerize or replace RARalpha in heterodimerizing with RXRalpha, to form co-repressor complexes with NcoR/SMRT-Sin3A-HDAC, which are less sensitive to RA (Figure 5) (He et al., 1998a, Lin et al., 1998; Grignani et al., 1998).

PLZF-RARalpha can form co-repressor complexes via its PLZF moiety, which are insensitive to RA (He et al., 1998a; Cheng et al., 1999; Lin et al., 1998; Grignani et al., 1998; Grunstein, 1997; Chambon, 1996). Therefore, the X-RARalpha fusion proteins of APL can exert their leukemogenic potential through an aberrant HDAC-dependent chromatin remodeling and transcriptional repressive ability. Histone deacetylase inhibitors (HDACIs) such as trichostatin A (TSA), in combination with RA, can overcome the transcriptional repressor activity of PML-RARalpha and PLZF-RARalpha. Furthermore, TSA can overcome the RA resistance of leukemic cells from PLZF-RARalpha transgenic mice (He et al., 1998a). These observations suggest that HDACIs alone or in combination with RA, which we refer to as `transcription therapy', might be useful in the treatment of APL (He et al., 1998c).

Several classes of HDACIs have been identified including short chain fatty acids (e.g., butyrates) and organic hydroxamic acids [e.g., TSA and hybrid polar compounds (HPCs)] (Richon et al., 1998). Sodium phenylbutyrate (SPB) has previously been used as a single agent in the treatment of beta-thalassemia, due to its ability to induce the expression of the gamma-globin gene (Dover et al., 1992). Recently, we have successfully utilized SPB in combination with RA for the treatment of one case of APL refractory to multiple chemotherapeutic regimens as well as to RA (Warrell et al., 1998). Preliminary results indicate that HDACIs such as SPB, TSA and suberanilohydroxamic acid (SAHA), a newly synthesized hybrid polar compound (Richon et al., 1998), invariably exerted a dramatic growth inhibitory and pro-apoptotic activity in leukemia cell lines such as NB4 (APL cell line), or in HL60 and U937 cell lines which do not harbor the APL fusion genes (He et al., 1998c). Inhibition of proliferation and induction of apoptosis by HDACIs were potentiated by RA. RA-induced differentiation was potentiated by these HDACIs, though no differentiation could be induced by HDACIs themselves. Despite their marked antitumoral activity these compounds display negligible toxicity in mice at doses that were able to induce accumulation of acetylated histones in peripheral blood and bone marrow cells (He et al., 1998c). These promising preliminary results prompted us to initiate in our X-RARalpha leukemia models the first in vivo `transcription therapy' trials with HDACIs or HDACIs/RA combinations (Figure 6).

Since HDACIs work as growth inhibitors and apoptosis inducers irrespectively of the presence of X-RARalpha proteins, these drugs may be useful in combination with RA or alone in the therapy of other leukemias and cancers. Indeed, differentiating, growth inhibitory and/or pro-apoptotic activity by HDACIs has recently been reported in other tumor cell lines (Richon et al., 1998, and unpublished observations). The molecular mechanisms underlying the broad antitumoral actions of HDACIs have not been clarified. It has been proposed however that HDACIs may specifically derepress a pre-programmed set of genes whose transcriptional activation induces cell-cycle arrest, apoptosis and/or cell differentiation (He et al., 1998c).


It is more and more apparent that murine models of human diseases go far beyond the simple mimicry of the human pathological condition, providing invaluable insights in the definition of the mechanisms underlying any etiopathogenetic process, while allowing the validation, in vivo, under physiological conditions, of any biochemical hypothesis. In particular, modeling APL in vivo in the mouse has been, and still is, a most rewarding an exciting adventure, not only because the mouse models generated are so accurately reproducing the human condition, but, more importantly, because these models have been continuously prompting the elaboration of new mechanistic hypotheses, also revealing new potential therapeutic strategies which can subsequently be validated in vivo.


We would like to thank all the past and present members of the laboratory of Molecular and Developmental Biology (MADB) lab at Memorial Sloan-Kettering Cancer Center, working on APL and related subjects: Maria Barna, Mantu Bhaumik, Laurent Delva, Mirella Gaboli, Domenica Gandini, Marco Giorgio, Ailan Guo, Nicola Hawe, Sundeep Kalantry, Daniela Peruzzi, Eduardo Rego, Roberta Rivi, Simona Ronchetti, Davide Ruggero, Paolo Salomoni, Carla Tribioli, Zhu-Gang Wang, Hui Zhang, Sue Zhong as well as Letizia Longo in our Department. L-Z He was partially supported by the Charles H Revson Foundation. PP Pandolfi is a Scholar of the Leukemia Society of America. Our work is supported, by the NCI, the De Witt Wallace Fund for Memorial Sloan-Kettering Cancer Center, and NIH Grants to PP Pandolfi.


Alcalay M, Zangrilli D, Fagioli M, Pandolfi PP, Mencarelli A, Lo Coco F, Biondi A, Grignani F and Pelicci PG. (1992). Proc. Natl. Acad. Sci. USA 89, 4840-4844. MEDLINE

Badiani PA, Kioussis D, Swirsky DM, Lampert IA and Weston K. (1996). Oncogene 13, 2205-2212. MEDLINE

Bardwell VJ and Treisman R. (1994). Genes Dev. 8, 1664-1677. MEDLINE

Barna M, Merghoub T, Cattoretti G, Wolgemuth D, Niswander L and Pandolfi PP. (1998). Blood 92, (Suppl 1), S310a.

Boyd AL and Samid D. (1993). J. Anim. Sci. 71, (Suppl 3) S1-S9. MEDLINE

Brown D, Kogan S, Lagasse E, Weissman I, Alcalay M, Pelicci PG, Atwater S and Bishop JM. (1997). Proc. Natl. Acad. Sci. USA 94, 2551-2556. Article MEDLINE

Cassinat B, Balitrand N, Zassadowski F, Barbey C, Delva L, Bastie JN, Rain JD, Degos L, Fenaux P and Chomienne C. (1998). Blood 92, (Suppl 1), S596a.

Castellanos A, Pintado B, Weruaga E, Arevalo R, Lopez A, Orfao A and Sanchez-Garcia I. (1997). Blood 90, 2168-2174. MEDLINE

Castilla LH, Wijmenga C, Wang Q, Stacy T, Speck NA, Eckhaus M, Marin-Padilla M, Collins FS, Wynshaw-Boris A and Liu PP. (1996). Cell 87, 687-696. MEDLINE

Chambon P. (1996). FASEB J. 10, 940-954. MEDLINE

Chang KS, Stass SA, Chu DT, Deaven LL, Trujillo JM and Freireich EJ. (1992). Mol. Cell. Biol. 12, 800-810. MEDLINE

Chen GQ, Shi XG, Tang W, Xiong SM, Zhu J, Cai X, Han ZG, Ni JH, Shi GY, Jia PM, Liu MM, He KL, Niu C, Ma J, Zhang P, Zhang TD, Paul P, Naoe T, Kitamura K, Miller W, Waxman S, Wang ZY, de The H, Chen SJ and Chen Z. (1997). Blood 89, 3345-3353. MEDLINE

Chen GQ, Zhu J, Shi XG, Ni JH, Zhong HJ, Si GY, Jin XL, Tang W, Li XS, Xong SM, Shen ZX, Sun GL, Ma J, Zhang P, Zhang TD, Gazin C, Naoe T, Chen SJ, Wang ZY and Chen Z. (1996). Blood 88, 1052-1061. MEDLINE

Chen SJ, Zelent A, Tong JH, Yu HQ, Wang ZY, Derre J, Berger R, Waxman S and Chen Z. (1993a). J. Clin. Invest. 91, 2260-2267. MEDLINE

Chen Z, Brand NJ, Chen A, Chen SJ, Tong JH, Wang ZY, Waxman S and Zelent A. (1993b). EMBO J. 12, 1161-1167. MEDLINE

Cheng G, Zhu X, Men X, Wang L, Huang Q, Jin XL, Xiong S, Zhu J, Guo W, Chen J, Xu S, So E, Chan L, Waxman S, Zelent A, Chen G, Dong S, Liu J and Chen S. (1999). Proc. Natl. Acad. Sci. USA 96, 6318-6323. Article MEDLINE

David G, Terris B, Marchio A, Lavau C and Dejean A. (1997). Oncogene 14, 1547-1554. MEDLINE

de The H, Lavau C, Marchio A, Chomienne C, Degos L and Dejean A. (1991). Cell 66, 675-684. MEDLINE

Dover GJ, Brusilow S and Samid D. (1992). N. Engl. J. Med. 327, 569-570. MEDLINE

Dubois C, Schlageter MH, de Gentile A, Guidez F, Balitrand N, Toubert ME, Krawice I, Fenaux P, Castaigne S, Najean Y, Degos L and Chomienne C. (1994). Blood 83, 3264-3270. MEDLINE

Early E, Moore MA, Kakizuka A, Nason-Burchenal K, Martin P, Evans RM and Dmitrovsky E. (1996). Proc. Natl. Acad. Sci. USA 93, 7900-7904. Article MEDLINE

Fenaux P and Degos L. (1997). N. Engl. J. Med. 337, 1076-1077. MEDLINE

Feuerstein N, Chan P-K and Mond J. (1988a). J. Biol. Chem. 263, 10608-10612.

Feuerstein N, Spiegel S and Mond J. (1988b). J. Cell. Biol. 107, 1629-1642.

Gaglio T, Saredi A and Compton DA. (1995). J. Cell. Biol. 131, 693-708. MEDLINE

Gianni M, Koken MH, Chelbi-Alix MK, Benoit G, Lanotte M, Chen Z and de The H. (1998). Blood 91, 4300-4310. MEDLINE

Goddard AD, Borrow J, Freemont PS and Solomon E. (1991). Science 254, 1371-1374. MEDLINE

Griffiths SD, Healy LE, Ford AM, Bennett CA, Voncken JW, Heisterkamp N, Groffen J and Greaves MF. (1992). Oncogene 7, 1391-1399. MEDLINE

Grignani F, De Matteis S, Nervi C, Tomassoni L, Gelmetti V, Cioce M, Fanelli M, Ruthardt M, Ferrara FF, Zamir I, Seiser C, Lazar MA, Minucci S and Pelicci PG. (1998). Nature 391, 815-818. Article MEDLINE

Grignani F, Fagioli M, Alcalay M, Longo L, Pandolfi PP, Donti E, Biondi A, Lo Coco, F and Pelicci PG. (1994). Blood 83, 10-25. MEDLINE

Grimwade D, Gorman P, Duprez E, Howe K, Langabeer S, Oliver F, Walker H, Culligan D, Waters J, Pomfret M, Goldstone A, Burnett A, Freemont P, Sheer D and Solomon E. (1997). Blood 90, 4876-4885. MEDLINE

Grimwade D, Howe K, Langabeer S, Davies L, Oliver F, Walker H, Swirsky D, Wheatley K, Goldstone A, Burnett A and Solomon E. (1996). Br. J. Haematol. 94, 557-573. MEDLINE

Grisolano JL, Sclar GM and Ley TJ. (1994). Proc. Natl. Acad. Sci. USA 91, 8989-8993. MEDLINE

Grisolano JL, Wesselschmidt RL, Pelicci PG and Ley TJ. (1997). Blood 89, 376-387. MEDLINE

Grunstein M. (1997). Nature 389, 349-352. Article MEDLINE

Gu H, Zou YR and Rajewsky K. (1993). Cell 73, 1155-1164. MEDLINE

Gudas L, Sporn M and Roberts A. (1994). In: Sporn MB, Roberts AB, Goodman DS. (eds). The retinoids: biology, chemistry and medicine. Raven Press: New York, pp. 443-520.

Guidez F, Ivins S, Zhu J, Soderstrom M, Waxman S and Zelent A. (1998). Blood 91, 2634-2642. MEDLINE

He LZ, Guidez F, Tribioli C, Peruzzi D, Ruthardt M, Zelent A and Pandolfi PP. (1998a). Nat. Genet. 18, 126-135. MEDLINE

He LZ, Ivins S, Zelent A and Pandolfi PP. (1998b). Blood 92, (Suppl 1) S480a.

He LZ, Tolentino T, Warrell Jr RP, Zelent A, Rifkind RA, Marks PA, Richon VM and Pandolfi PP. (1998c). Blood 92, (Suppl 1) S510a.

He LZ, Tribioli C, Rivi R, Peruzzi D, Pelicci PG, Soares V, Cattoretti G and Pandolfi PP. (1997). Proc. Natl. Acad. Sci. USA 94, 5302-5307. Article MEDLINE

Heisterkamp N, Jenster G, Kioussis D, Pattengale PK and Groffen J. (1991). Transgenic Res. 1, 45-53. MEDLINE

Heisterkamp N, Jenster G, ten Hoeve J, Zovich D, Pattengale PK and Groffen J. (1990). Nature 344, 251-253. MEDLINE

Herault Y, Rassoulzadegan M, Cuzin F and Duboule D. (1998). Nat. Genet. 20, 381-384. Article MEDLINE

Ivins S and Zelent A. (1998). Blood 92, (Suppl 1) S308a.

Jansen JH, de Ridder MC, Geertsma WMC, Erpelinck CAJ, Smit B, Slater R, wd Reijden BA, de Greef GE, Sonneveld P and Löwenberg B. (1998). Blood 92, (Suppl 1) S402a.

Kakizuka A, Miller Jr WH, Umesono K, Warrell Jr RP, Frankel SR, Murty VV, Dmitrovsky E and Evans RM. (1991). Cell 66, 663-674. MEDLINE

Kalantry S, Delva L, Gaboli M, Gandini D, Giorgio M, Hawe N, He LZ, Peruzzi D, Rivi R, Tribioli C, Wang ZG, Zhang H and Pandolfi PP. (1997). J. Cell. Physiol. 173, 288-296. MEDLINE

Kamitani T, Kito K, Nguyen HP, Wada H, Fukunda-Kamitani T and Yeh EH. (1998). J. Biol. Chem. 273, 26675-26682. Article MEDLINE

Kilby NJ, Snaith MR and Murray JA. (1993). Trends Genet. 9, 413-421. MEDLINE

Koken MH, Daniel MT, Gianni M, Zelent A, Licht J, Buzyn A, Minard P, Degos L, Varet B and de The H. (1999). Oncogene 18, 1113-1118. MEDLINE

Koken MH, Linares-Cruz G, Quignon F, Viron A, Chelbi-Alix MK, Sobczak-Thepot J, Juhlin L, Degos L, Calvo F and de The H. (1995). Oncogene 10, 1315-1324. MEDLINE

Koken MH, Puvion-Dutilleul F, Guillemin MC, Viron A, Linares-Cruz G, Stuurman N, de Jong L, Szostecki C, Calvo F and Chomienne C. (1994). EMBO J. 13, 1073-1083. MEDLINE

Koller E, Krieger O, Kasparu H and Lutz D. (1991). Lancet 338, 1154-1155. MEDLINE

Lafage-Pochitaloff M, Alcalay M, Brunel V, Longo L, Sainty D, Simonetti J, Birg F and Pelicci PG. (1995). Blood 85, 1169-1174. MEDLINE

Lagasse E and Weissman IL. (1994). J. Exp. Med. 179, 1047-1052. MEDLINE

Lallemand-Breitenbach V, Guillemin MC, Janin A, Daniel MT, Degos L, Kogan SC, Bishop JM and de The H. (1999). J. Exp. Med. 189, 1043-1052. MEDLINE

Lavau C, Marchio A, Fagioli M, Jansen J, Falini B, Lebon P, Grosveld F, Pandolfi PP, Pelicci PG and Dejean A. (1995). Oncogene 11, 871-876. MEDLINE

Lewandoski M and Martin GR. (1997). Nat. Genet. 17, 223-225. MEDLINE

Li JY, English MA, Ball HJ, Yeyati PL, Waxman S and Licht JD. (1997a). J. Biol. Chem. 272, 22447-22455. Article MEDLINE

Li YP, Andersen J, Zelent A, Rao S, Paietta E, Tallman MS, Wiernik PH and Gallagher RE. (1997b). Blood 90, 306-312. MEDLINE

Licht JD, Chomienne C, Goy A, Chen A, Scott AA, Head DR, Michaux JL, Wu Y, DeBlasio A, Miller Jr WH et al. (1995). Blood 85, 1083-1094. MEDLINE

Lin RJ, Nagy L, Inoue S, Shao W, Miller Jr WH and Evans RM. (1998). Nature 391, 811-814. Article MEDLINE

Liu QR and Chan PK. (1991). Eur. J. Biochem. 200, 715-721. MEDLINE

Look AT. (1997). Science 278, 1059-1064. MEDLINE

Lovering R, Hanson IM, Borden KL, Martin S, O'Reilly NJ, Evan GI, Rahman D, Pappin DJ, Trowsdale J and Freemont PS. (1993). Proc. Natl. Acad. Sci. USA 90, 2112-2116. MEDLINE

Lufkin T, Lohnes D, Mark M, Dierich A, Gorry P, Gaub MP, LeMeur M and Chambon P. (1993). Proc. Natl. Acad. Sci. USA 90, 7225-7229. MEDLINE

Melnick A and Licht JD. (1999). Blood 93, 3167-3215. MEDLINE

Mes-Masson AM, McLaughlin J, Daley GQ, Paskind M and Witte ON. (1986). Proc. Natl. Acad. Sci. USA 83, 9768-9772. MEDLINE

Mu ZM, Chin KV, Liu JH, Lozano G and Chang KS. (1994). Mol. Cell. Biol. 14, 6858-6867. MEDLINE

Nason-Burchenal K, Gandini D, Botto M, Allopenna J, Seale JR, Cross NC, Goldman JM, Dmitrovsky E and Pandolfi PP. (1996). Blood 88, 3926-3936. MEDLINE

Okuda T, Cai Z, Yang S, Lenny N, Lyu CJ, van Deursen JM, Harada H and Downing JR. (1998). Blood 91, 3134-3143. MEDLINE

Pandolfi PP. (1996). Haematologica 81, 472-482. MEDLINE

Pandolfi PP. (1998). Semin Hematol. 35, 136-148. MEDLINE

Pandolfi PP, Grignani F, Alcalay M, Mencarelli A, Biondi A, LoCoco F and Pelicci PG. (1991). Oncogene 6, 1285-1292. MEDLINE

Peck R and Bollag W. (1991). Eur. J. Cancer 27, 53-57. MEDLINE

Perez A, Kastner P, Sethi S, Lutz Y, Reibel C and Chambon P. (1993). EMBO J. 12, 3171-3182. MEDLINE

Pham CT, MacIvor DM, Hug BA, Heusel JW and Ley TJ. (1996). Proc. Natl. Acad. Sci. USA 93, 13090-13095. Article MEDLINE

Pollock J, Grisolano J, Goda P, Westervelt P and Ley T. (1997). Blood 90, (Suppl 1) S320a.

Puccetti E, Sternsdorf T, Hoelzer D, Hans Will H, Ottmann OG, Ruthardt M. (1998). Blood 92, (Suppl) S212a.

Quignon F, De Bels F, Koken M, Feunteun J, Ameisen JC and de The H. (1998). Nat. Genet. 20, 259-265. Article MEDLINE

Rabbitts TH. (1994). Nature 372, 143-149. MEDLINE

Raelson JV, Nervi C, Rosenauer A, Benedetti L, Monczak Y, Pearson M, Pelicci PG and Miller Jr WH. (1996). Blood 88, 2826-2832. MEDLINE

Ramirez-Solis R, Liu P and Bradley A. (1995). Nature 378, 720-724. MEDLINE

Reddy BA, Etkin LD and Freemont PS. (1992). Trends Biochem. Sci. 17, 344-345. MEDLINE

Redner RL, Rush EA, Faas S, Rudert WA and Corey SJ. (1996). Blood 87, 882-886. MEDLINE

Rego E, He LZ, Wang ZG, Peruzzi D, Warrell Jr RP and Pandolfi PP. (1998). Blood 92, (Suppl 1) S403a.

Reid A, Gould A, Brand N, Cook M, Strutt P, Li J, Licht J, Waxman S, Krumlauf R and Zelent A. (1995). Blood 86, 4544-4552. MEDLINE

Richon VM, Emiliani S, Verdin E, Webb Y, Breslow R, Rifkind RA and Marks PA. (1998). Proc. Natl. Acad. Sci. USA 95, 3003-3007. Article MEDLINE

Sauer B. (1998). Methods 14, 381-392. Article MEDLINE

Shaknovich R, Yeyati PL, Ivins S, Melnick A, Lempert C, Waxman S, Zelent A and Licht JD. (1998). Mol. Cell. Biol. 18, 5533-5545. MEDLINE

Shao W, Fanelli M, Ferrara FF, Riccioni R, Rosenauer A, Davison K, Lamph WW, Waxman S, Pelicci PG, Lo Coco F, Avvisati G, Testa U, Peschle C, Gambacorti-Passerini C, Nervi C and Miller Jr WH. (1998). J. Natl. Cancer Inst. 90, 124-133. MEDLINE

Shen ZX, Chen GQ, Ni JH, Li XS, Xiong SM, Qiu QY, Zhu J, Tang W, Sun GL, Yang KQ, Chen Y, Zhou L, Fang ZW, Wang YT, Ma J, Zhang P, Zhang TD, Chen SJ, Chen Z and Wang ZY. (1997). Blood 89, 3354-3360. MEDLINE

Shtivelman E, Lifshitz B, Gale RP and Canaani E. (1985). Nature 315, 550-554. MEDLINE

Sitterlin D, Tiollais P and Transy C. (1997). Oncogene 14, 1067-1074. MEDLINE

Smith AJ, De Sousa MA, Kwabi-Addo B, Heppell-Parton A, Impey H and Rabbitts P. (1995). Nat. Genet. 9, 376-385. MEDLINE

Smith MA, Parkinson DR, Cheson BD and Friedman MA. (1992). J. Clin. Oncol. 10, 839-864. MEDLINE

Soignet SL, Maslak P, Wang ZG, Jhanwar S, Calleja E, Dardashti LJ, Corso D, DeBlasio A, Gabrilove J, Scheinberg DA, Pandolfi PP and Warrell Jr RP. (1998). N. Engl. J. Med. 339, 1341-1348. MEDLINE

Tallman MS, Andersen JW, Schiffer CA, Appelbaum FR, Feusner JH, Ogden A, Shepherd L, Willman C, Bloomfield CD, Rowe JM and Wiernik PH. (1997). N. Engl. J. Med. 337, 1021-1028. MEDLINE

Tobal K, Saunders MJ, Grey MR and Yin JA. (1995). Br. J. Haematol. 90, 615-618. MEDLINE

Van Deursen J, Fornerod M, Van Rees B and Grosveld G. (1995). Proc. Natl. Acad. Sci. USA 92, 7376-7380. MEDLINE

Voncken JW, Kaartinen V, Pattengale PK, Germeraad WT, Groffen J and Heisterkamp N. (1995). Blood 86, 4603-4611. MEDLINE

Wang Z-G, Rego E, Peruzzi D, He LZ and Pandolfi PP. (1998a). Blood 92, (Suppl 1) S479a.

Wang ZG, Delva L, Gaboli M, Rivi R, Giorgio M, Cordon-Cardo C, Grosveld F and Pandolfi PP. (1998b). Science 279, 1547-1551. Article MEDLINE

Wang ZG, Rivi R, Delva L, Konig A, Scheinberg DA, Gambacorti-Passerini C, Gabrilove JL, Warrell Jr RP and Pandolfi PP. (1998c). Blood 92, 1497-1504. MEDLINE

Wang ZG, Ruggero D, Ronchetti S, Zhong S, Gaboli M, Rivi R and Pandolfi PP. (1998d). Nat. Genet. 20, 266-272. Article MEDLINE

Warrell Jr RP. (1993a). Lancet 341, 126.

Warrell Jr RP, de The H, Wang ZY and Degos L. (1993b). N. Engl. J. Med. 329, 177-189. MEDLINE

Warrell Jr RP, Frankel SR, Miller Jr WH, Scheinberg DA, Itri LM, Hittelman WN, Vyas R, Andreeff M, Tafuri A and Jakubowski A. (1991). N. Engl. J. Med. 324, 1385-1393. MEDLINE

Warrell Jr RP, He LZ, Richon V, Calleja E and Pandolfi PP. (1998). J. Natl. Cancer Inst. 90, 1621-1625. Article MEDLINE

Westervelt P, Link D, Graubert T, Grass T, Corbin T, Goforth R and Ley TJ. (1997). Blood 92:, (Suppl 1) 322a.

Weis K, Rambaud S, Lavau C, Jansen J, Carvalho T, Carmo-Fonseca M, Lamond A and Dejean A. (1994). Cell 76, 345-356. MEDLINE

Wells RA, Catzavelos C and Kamel-Reid S. (1997). Nat. Genet. 17, 109-113. MEDLINE

Wells RA, Hummel JL, De Koven A, Zipursky A, Kirby M, Dube I and Kamel-Reid S. (1996). Leukemia 10, 735-740. MEDLINE

Wildin RS, Wang HU, Forbush KA and Perlmutter RM. (1995). J. Immunol. 155, 1286-1295. MEDLINE

Yergeau DA, Hetherington CJ, Wang Q, Zhang P, Sharpe AH, Binder M, Marin-Padilla M, Tenen DG, Speck NA and Zhang DE. (1997). Nat. Genet. 15, 303-306. MEDLINE

Yeyati PL, Shaknovich R, Boterashvili S, Li J, Ball HJ, Waxman S, Nason-Burchenal K, Dmitrovsky E, Zelent A and Licht JD. (1999). Oncogene 18, 925-934. MEDLINE

Zou YR, Muller W, Gu H and Rajewsky K. (1994). Curr. Biol. 4, 1099-1103. MEDLINE


Figure 1 Differentiation therapy in APL. APL is characterized by the clonal expansion of malignant myeloid cells blocked at the promyelocytic stage of hemopoietic differentiation. All-trans retinoic acid (RA) induces the malignant promyelocytes to terminally differentiate towards mature granulocytes. Occasionally, Auer rods (indicated by the yellow arrows), pathognomonic structures found in the cytoplasm of APL malignant promyelocytes, are still observed in the mature granulocytes upon RA treatment demonstrating that these mature cells originated from the neoplastic clone

Figure 2 Modular organization of the various wild-type proteins involved in APL and of the resulting aberrant fusion products. (a) Schematic representation of RARalpha, PML, PLZF, NPM and NuMA proteins. RARalpha is subdivided into its conserved A-F functional domains (Chambon, 1996). C and E designate the DNA and ligand binding domain respectively. The various functional regions of PML, PLZF, NPM and NuMA are designated as follows: P=proline rich region; R=Ring-finger domain; B1 and B2=B-boxes 1 and 2; S/P=serine-proline rich region, POZ=BTB/POZ domain; GD=globular domains. The breakpoints where the genes are fused with RARalpha are indicated by the black arrowheads and phosphorylation sites in PML and NPM are designated by asterisks. The nine zinc fingers of PLZF are represented by checked circles. MBD in the NPM scheme denotes the potential metal binding domain and the two boxes numbered 1 and 2 indicate acidic amino acid clusters. The dimerization interfaces are underlined. (b) Schematic representation of the various fusion proteins generated by the four variant translocations. Symbols are as before. (c) Schematic representation of the reciprocal fusion proteins. The existence of the RARalpha-NuMA fusion molecules has not been demonstrated

Figure 3 Hypothetical model for the role of the various X-RARalpha and RARalpha-X fusion proteins in APL pathogenesis. (a) X-RARalpha fusion proteins have the ability to heterodimerize with PML, PLZF, NPM and NuMA (X proteins) as well as with RXR. X-RARalpha proteins can also interfere with RARalpha transcriptional function on DNA, since they invariably retain the RARalpha DNA binding domain, as well as through their ability to bind the ligand (RA). The various X-RARalpha proteins can therefore interfere with both X or RARalpha-RXR pathways at multiple levels. The functional inactivation/interference with these pathways would result in growth advantage, tumor susceptibility and the differentiation block at the promyelocytic stage. (b) The biochemical functions of the various RARalpha-X proteins are presently unclear although preliminary data obtained in transgenic mice reveal that their contribution to APL leukemogenesis could be critical. Unlike the other RARalpha-X fusion protein, RARalpha-PLZF exerts a discrete biochemical function since it can bind DNA through the zinc finger domains which constitute the PLZF DNA binding domain but loses the repressive transcriptional ability which is mediated, at least in part, by the PLZF POZ domain. Thus, RARalpha-PLZF can interfere with PLZF biological functions through deregulation of PLZF target genes

Figure 4 In vivo expression of the PML-RARalpha fusion gene under the control of different promoters and resulting leukemia phenotypes. The upper portion of the figure represents schematically the various stages of myeloid maturation: HSC (hemopoietic stem cell), MB (myeloblast), PMC (promyelocyte), MC (myelocyte), MMC (metamyelocyte), BC (band cells), SG (segmented granulocyte). The patterns of transgene expression theoretically obtained in vivo in the transgenic animal utilizing various promoter/minigene expression vectors are shown underneath. The phenotype observed in PML-RARalpha transgenic mice utilizing these vectors is summarized on the right column, (-) indicates absence of leukemia, (+) development of leukemia, (*) these mice developed hepatocellular carcinoma. Red color highlights expression vectors whose utilization resulted in leukemogenesis

Figure 5 Mechanisms of transcriptional repression in APL pathogenesis. In the absence of the ligand, RARalpha heterodimerizes with RXRalpha and acts as a transcriptional repressor by recruiting nuclear receptor co-repressors, NcoR, SMRT, Sin3 and in turn HDACs. The presence of RA, at physiological concentrations, induces an allosteric change in the receptor leading to the release of the co-repressor complex and the recruitment of the co-activator complex thus leading to the activation of transcription of genes required for cellular differentiation and growth inhibition. At this concentration of the ligand, both PML-RARalpha and PLZF-RARalpha are potent transcriptional repressors in view of an increased and aberrant affinity for nuclear co-repressors and HDACs. At pharmacological dose of RA, while PML-RARalpha can be freed from co-repressors interactions thus directly mediating trans-activation of RARalpha target genes, PLZF-RARalpha, via the PLZF moiety, renders the leukemic cells RA unresponsive through co-repressor interactions which are insensitive to RA

Figure 6 Combinations of HDAC inhibitors (HDACIs) and RA in the treatment of APL. HDACIs might synergize with RA by blocking HDAC activity within the aberrant repressive transcription complex, thus antagonizing transcriptional repression by the APL fusion proteins as well as their leukemogenic potential

20 September 1999, Volume 18, Number 38, Pages 5278-5292
Table of contents    Previous  Article  Next    [PDF]