Leukemia (2004) 18, 375–384. doi:10.1038/sj.leu.2403234 Published online 22 January 2004

Acute promyelocytic leukemia: where does it stem from?

D Grimwade1 and T Enver2

  1. 1Department of Medical and Molecular Genetics, Guy's, King's and St Thomas' School of Medicine and Department of Haematology, University College London Hospitals, London, UK
  2. 2MRC Molecular Haematology Unit, Weatherall Institute of Molecular Medicine, John Radcliffe Hospital, Oxford, UK

Correspondence: D Grimwade, Department of Medical and Molecular Genetics, Guy's, King's and St Thomas' School of Medicine and Department of Haematology, University College London Hospitals, London, UK. Fax: +44 207 955 8762; E-mail: ; T Enver, MRC Molecular Haematology Unit, Weatherall Institute of Molecular Medicine, John Radcliffe Hospital, Oxford, UK. Fax: +44 1865 222449; E-mail:

Received 8 September 2003; Accepted 24 October 2003; Published online 22 January 2004.



A fundamental issue in cancer biology is the identification of the target cell in which the causative molecular lesion arises. Acute myeloid leukemia (AML) is thought to reflect the transformation of a primitive stem cell compartment. The resultant 'cancer stem cells' comprise only a minor portion of the leukemic clone but give rise through differentiation to more committed progenitors as well as differentiated blasts that constitute the bulk of the tumor. The maintenance of the leukemic clone is dependent on the self-renewal capacity of the cancer stem cell compartment, which is revealed by its ability to re-initiate leukemia in a transplant setting. The cellular basis of acute promyelocytic leukemia (APL) is however less clear. APL has traditionally been considered to be the most differentiated form of AML and to arise from a committed myeloid progenitor. Here we review apparently conflicting evidence pertaining to the cellular origins of APL and propose that this leukemia may originate in more than one cellular compartment. This view could account for many apparent inconsistencies in the literature to date. An understanding of the nature of the target cell involved in transformation of APL has important implications for biological mechanism and for clinical treatment.


acute promyelocytic leukemia (APL), acute myeloid leukemia (AML), hematopoietic stem cells, lineage commitment



Acute promyelocytic leukemia (APL) is one of the commonest subtypes of acute myeloid leukemia (AML) and has been of particular interest given that it is the first disease for which molecularly targeted therapies in the form of all-trans-retinoic acid (ATRA) and arsenic trioxide have been successfully deployed, yielding substantial improvements in clinical outcome (reviewed in Mistry et al1 and Ohno et al2). Two major subtypes of the disease can be distinguished on morphological grounds, namely hypergranular classical APL and the hypogranular (microgranular) variant form, which are designated as M3 and M3v respectively in the French–American–British (FAB) classification of AML.3,4 In the former, the bone marrow is replaced by hypergranular abnormal promyelocytes, including cells containing bundles of Auer rods (faggot cells), which are diagnostic of the disease. Hypergranular APL commonly presents with leukopenia with few abnormal cells in the peripheral blood. This contrasts with the hypogranular form, which is defined by a predominance of less granular blasts with bilobed nuclei, is typified by an elevated presenting leukocyte count and has a higher frequency of activating mutations of FLT3 (reviewed in Mistry et al1) However, despite their morphological differences, both are characterized by chromosomal translocations involving 17q21, leading to rearrangements of the gene encoding RARalpha, which is a member of the steroid hormone family of transcription factors and which has been implicated in modulating normal myelopoiesis (reviewed in Collins5).

To date, five fusion partners of RARalpha have been identified, namely PML, PLZF, NPM, NuMA and STAT5b (designated X-proteins), associated with t(15;17)(q22;q12–21), t(11;17)(q23;q12–21), t(5;17)(q35;q12–21), t(11;17)(q13;q12–21) and interstitial deletion of 17q respectively (reviewed in Redner6). The vast majority of APL cases are associated with the PML-RARalpha fusion (>95%); PLZF-RARalpha and NPM-RARalpha are detected in approximately 1 and 0.5% APL respectively, while NuMA-RARalpha and STAT5b-RARalpha have been identified in only a single patient each.1 In all cases, the RARalpha moiety of the fusion protein retains the DNA binding domain and carboxy-terminal E-domain which mediates interaction with ligand (retinoic acid (RA)/ATRA) and retinoid-X-receptor (RXR) which is necessary for high-affinity binding at retinoid response elements (RARE). The nature of the fusion partner has an important bearing on disease characteristics, determining the sensitivity to ATRA and arsenic trioxide (reviewed in Mistry et al1). Work from a number of laboratories has revealed that at physiological levels of RA, X-RARalpha proteins bind corepressor/histone deacetylase (HDAC) complexes at significantly higher affinity than wild-type RARalpha; which leads to repression of retinoid target genes which is likely to play a key role in mediating the differentiation block that characterizes APL (reviewed in Pitha-Rowe et al7). The transcriptional repression of target genes has recently been shown to be further compounded by recruitment of DNA methyltransferases leading to methylation of key promoters.8 In the presence of pharmacological levels of RA, ligand binding induces a conformational change in the PML-RARalpha fusion protein accompanied by the release of corepressor/HDAC complex in favor of coactivator binding leading to transcriptional activation of downstream targets.7 In addition, RA leads to degradation of PML-RARalpha and release of sequestered RXR, which all together lead to differentiation of the leukemic clone culminating in apoptosis, associated with upregulation of TRAIL.9,10 This scenario contrasts with the rarer subtype of APL with the PLZF-RARalpha fusion in which corepressor/HDAC complexes additionally bind to the amino-terminal PLZF moiety in a retinoid-insensitive fashion, such that this subgroup of APL responds poorly to ATRA as single agent therapy.11,12,13,14,15

Thus the molecular basis of APL is relatively well understood. However, the mechanisms by which the t(15;17) arises have not been so extensively explored. Characterization of genomic breakpoints in patients with APL has revealed a complex pattern with microhomologies, deletions and inversions suggesting that the t(15;17) is likely to arise as a result of double-strand DNA breakage followed by nonhomologous end-joining.16,17 However, an important issue that remains to be resolved is the nature of the hematopoietic progenitor that is the normal target of the PML-RARalpha fusion in this disease. In this review, we consider the pertinent experimental evidence to date, particularly with respect to different neoplastic conditions and its implications for the origin of APL.


The case for cancer as a stem cell disease

Evidence suggests that cancer is a clonal disease that initiates in a single cell whose progeny make up the tumor. In many instances however, two other aspects relating to the cellular origins of cancer are less well understood. First, the nature of the cell in which the initiating mutation occurred, often referred to as the target cell, is unknown. Second, while it has been appreciated that tumors are functionally heterogeneous at a cellular level, in that only a subpopulation of cells have the capacity to sustain the tumor or re-initiate it in the experimental setting of transplantation, it is not always clear which cells within the population have this capacity.

Two conceptually very distinct models have been proposed to account for this kind of tumor heterogeneity. The stochastic model proposes that all cells within the tumor are equivalent in respect of their potential to re-initiate the clone. In this scheme, heterogeneity is only an apparent property reflecting the fact that the probability of re-initiation by any given cell is significantly less than 1. The alternative view is that re-initiation potential is a fixed property of a distinct subset of cells within the tumor.

This latter model is consistent with a stem cell-based framework for tumors. Thus the target cell may be viewed as a cancer stem cell which clonally gives rise to progeny. The extent to which the progeny, which will most likely constitute the bulk of the tumor, retain the functional properties and morphological characteristics of the target cell depends on the extent to which the target cell retains the capacity to differentiate. This will also determine how easy or not it is to identify the initiating cell itself. An example is provided by chronic myelogenous leukemia (CML), which presents as a predominantly granulocytic disease. However, the BCR-ABL mutation characteristic of this disease does not arise in a granulocyte or exclusively granulocytic-committed progenitor, but rather in a cell that retains both lymphoid and myeloid differentiation capacity and resides at the top of the hematopoietic differentiation hierarchy.18 The true origins of CML in this cell type are revealed in the observation that the disease may progress to myeloid or lymphoid blast crisis.19

If tumors follow the rules of a stem cell-based differentiation hierarchy, albeit a somewhat corrupted one, then it follows that self-renewal capacity, of which transplantability is one measure, will be preferentially associated with cells at the top rather than the bottom of the hierarchy. It also follows that these 'cancer stem cells', while likely to numerically represent only a minority of cells within the tumor, will constitute the key therapeutic target; any therapy that fails to eliminate these cells, irrespective of its efficiency against the bulk of the tumor, will leave the door open to relapse (see below).

The notion of stem or progenitor cells constituting preferential cellular targets for initiating mutations in cancer is attractive for two reasons. Firstly, these cells have intrinsically high self-renewal capacity. Thus the molecular machinery that sustains self-renewal is presumably already active within these cells and could be easily hijacked or dysregulated to provide the unrestrained self-renewal that is an essential feature of cancer cells (Figure 1). Achieving this in more differentiated cells that lack intrinsic self-renewal capacity would presumably be more difficult. Secondly, stem and progenitor cells are, in general, considerably more long-lived than their differentiated progeny. It stands to reason therefore that it is harder to accumulate the number of mutations required to develop a tumor in a short-lived cell than in a long-lived one. An exception to this argument would be if the initial mutation immortalized a cell which would then by definition have an extended lifetime during which it could accumulate cancer-associated or predisposing mutations.

Figure 1.
Figure 1 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact or the author

The stem and progenitor cell basis of cancer. (a) Normal differentiation hierarchy with stem cells (purple) giving rise to progenitors (lilac) and more mature differentiated cells (blue). (b) Modelling a 'first hit' in cancer. Mutations that arise in stem cells (dark green) are maintained indefinitely due to the high intrinsic self-renewal capacity of HSC. Mutations that occur in more downstream cells will persist as these cells have more limited self-renewal potential and are destined to die. The particular event shown on the right-hand side of the diagram occurs in a committed progenitor but could equally occur in a terminally differentiated cell. Note that the extent to which more differentiated forms carrying the first hit mutation are observed will depend on the extent to which the first hit mutation blocks further differentiation of the stem cell in question. (c) Stem cell-based differentiation hierarchy in cancer. A second hit mutation within a stem cell gives rise to a cancer stem cell with tumor-initiating potential. Although more differentiated cells within the cancer clone harbor the mutation, they lack the ability to re-initiate the tumor and as such cannot be considered to be cancer stem cells. While elimination of these more differentiated cells will ease the tumor burden, the tumor will reappear unless the cancer stem cell itself is eliminated.

Full figure and legend (217K)

In many cases of adult AML, myelodysplastic syndromes and CML, one can make a reasonably cogent argument along the lines articulated above that a multipotent stem cell constitutes the initial target cell. Can this notion be extrapolated to all hematological malignancies? If one accepts that key features of a stem cell that make it a productive transformation target are its longevity and its self-renewal capacity, then in principle at least one could consider most blood cancers as being 'stem cell' diseases. For example, myeloma cells are end-stage cells of the B-lymphoid differentiation pathway. However, it is likely that the disease originates in and is sustained by clonogenic or memory B cells that (i) are long-lived, (ii) retain the potential to undergo extensive proliferation, or self-renewal, in response to antigen, and (iii) can differentiate into plasma cells. In principle, similar considerations may apply to malignant diseases involving cells from the T lineage.

Does the stem cell credo have broader currency beyond blood cell cancers? This line of argument may explain why neuroblastoma and rhabdomyosarcoma are diseases of childhood, since neural and muscle stem cells respectively are likely to be both more frequent and proliferating more in early development. The pattern of onset of diseases like Wilms tumor, which could represent the transformation of a developmentally restricted multipotent mesodermal progenitor, might also be explained in this way. Finally, it follows that cancers of adulthood would be most frequent in stem cell-based organ systems like the gut, liver, skin and other epithelial linings. The high frequency of carcinomas in adulthood would tend to favor this interpretation.


Cancer stem cells

Unambiguous identification of a cancer stem cell is provided by the demonstration of tumor-initiating potential in an appropriate fraction of tumor cells purified to homogeneity. Transplantation of human leukemic cells into immunocompromised murine models such as the NOD/SCID mouse has proved invaluable in identifying leukemia-initiating cells.20,21 Perhaps the best worked example is that of AML where Dick and co-workers20 showed that only the CD34+CD38- fraction, which comprised from a 100th to a 1000th of the leukemia cells, had the capability of initiating an AML-like disease upon transplantation in the NOD/SCID model. More detailed analyses of the surface phenotype of these cells demonstrated further similarities with normal human stem cells. CD34+CD38+ cells, as well as the CD34- fraction, which together constituted the bulk of the leukemic population, could not initiate tumors, although these populations did contain some progenitor cell activity, exhibiting in vitro clonogenic potential. Interestingly, analysis of AML cases of different FAB types revealed that despite the morphological differences, the NOD/SCID repopulating cells, which have subsequently been shown to be largely quiescent,22 consistently resided within the CD34+CD38- fraction.20 Furthermore, morphological differences were maintained in serial transplantation studies. This led to the conclusion that noncommitted progenitors are targeted in AML, with the intrinsic molecular lesion (eg expression of leukemia-associated fusion gene product) influencing lineage commitment, degree of differentiation block and hence the morphological characteristics. Thus an AML clone may be viewed as a differentiation hierarchy organized in a similar fashion to normal hematopoiesis and containing a stem, progenitor and more mature nonclonogenic population. This model squares with the cancer stem cell model of tumor heterogeneity – cancer stem cells represent the target cell in transformation and retain tumor-initiating function, with heterogeneity within the malignant population arising as a consequence of their differentiation.

Recent developments in understanding the stem cell biology of the breast coupled with the identification of breast cancer stem cells23 encourage the view that the intimate relationship between tumor cells and stem cells seen in the blood system will also emerge as a key feature of solid tumors. Thus, a population of cells has been identified in both normal murine24,25 and human breast tissue25 which has the characteristics of side population or SP cells. In other tissues such as blood, SP cells are enriched for stem cells. Similarly, using expression of the blood stem cell-associated marker gene Sca1, Rosen and co-workers24 have enriched a population of transplantable murine mammary stem cells. Finally, Al Hajj et al23 have examined the transplantability of human breast tumor samples using the NOD/SCID model. In these experiments, they identified a breast cancer stem cell population that has tumor re-initiating potential and that may be prospectively isolated using a combination of lineage depletion and FACS sorting for CD44+, CD24 lo/negative cells.

Understanding the nature of progenitors subject to malignant change is not only important as a means of gaining information regarding the normal mechanisms of lineage commitment and differentiation and the processes leading to neoplastic transformation, but also, as already alluded to above, carries therapeutic implications. As far as leukemias are concerned, it is important to establish whether the transforming event occurs in primitive multipotent progenitors and whether a residual population of primitive normal stem cells remains that has the potential to repopulate the marrow after myeloablative therapy. For hematological malignancies in which multipotent progenitors are involved and in which there is no residual stem cell pool, conventional chemotherapy or transplantation approaches involving use of autologous stem cells are unlikely to be successful, and allogeneic transplantation which generates a 'graft-versus-leukemia' effect currently remains the sole treatment option available that is likely to be curative. Identification of the leukemia-initiating cell also has important implications for targeted therapeutic approaches, since failure to eliminate residual disease in the leukemia repopulating fraction could lead to subsequent relapse. Moreover, it is worth bearing in mind that leukemia-initiating cells may differ from their more differentiated progeny, which form the bulk of the disease burden, in terms of susceptibility to various treatment modalities. This could reflect differences in gene expression profile and/or functional characteristics including cell cycle status that render them relatively resistant to therapy. For example, immune-directed therapeutic approaches could be hampered by lack of expression of appropriate target molecules, for example CD33. Furthermore, investigation of the characteristics of normal human multipotent hematopoietic progenitors and leukemia-initiating cells has revealed a number of similarities. This suggests that the latter could be inherently resistant to induction of apoptosis by various agents due to expression of proteins functioning as multidrug efflux pumps, or promoting cell survival such as BCL2.26,27 Indeed, recent studies have revealed that cases of CML harbor a population of BCR-ABL-positive quiescent stem cells that are resistant to treatment including targeted therapies such as imatinib.28 This could very well account for the prolonged evidence of residual disease at the molecular level in the majority of patients treated with this agent (reviewed in Hochhaus29).

While lack of sustained response or relapse could reflect relative insensitivity of cancer stem cells to conventional therapy, it nevertheless seems likely that improved characterization of cancer stem cells could yield substantial advantages potentially affording the opportunity for more precise targeting of therapy, eliminating tumors at source. Success of this strategy could also entail substantial reductions in treatment-related toxicity.

Thus the cancer stem cell concept provides a useful framework for thinking about the biology of many types of cancer and has implications for how they might best be treated. How then does APL fit within this framework? Of particular interest in this respect is a transplantation study performed by Dick and co-workers20 where the behavior of primary APL cells appeared to be distinct from other forms of AML. These workers observed failure of engraftment in NOD/SCID mice in each of the three patient samples that they examined. This led the investigators to draw the conclusion that APL may be a highly distinct subtype of AML that arises in committed progenitors as opposed to multipotent stem cells.


The case for APL arising in committed myeloid progenitors

The proposal that APL arises in committed progenitors has been supported by a number of lines of evidence. Most persuasively, investigation of clonality and PML-RARA expression status of purified bone marrow subsets in a series of three cases of APL suggested that the leukemic clone arose in CD34+CD38+ progenitors, with no evidence of involvement of the more primitive CD34+CD38- cell population.30 Furthermore, a number of mouse models of APL have been established using promoters that are switched on relatively late in myeloid development, that is, MRP8 and cathepsin G.31,32,33,34,35 Whereas high-level ubiquitous expression of PML-RARalpha using the beta-actin promoter was found to induce embryonic lethality,33 transgenic approaches suggested that the range of progenitors subject to transformation by the PML-RARalpha fusion is somewhat restricted (reviewed in Westervelt and Ley35). Expression of PML-RARA under the control of CD11b regulatory elements, leading to expression in myelocytes and mature neutrophils, did not give rise to APL.36 Similarly, no leukemias resulted from targeting fusion gene expression to very early progenitors using c-fes (reviewed in Westervelt and Ley35). However in the latter study, expression of PML-RARalpha in the hematopoietic compartment was not clearly detected at the protein level; hence, it remains a possibility that expression levels were not comparable to those associated with disease in man, thereby accounting for the disappointing result. An alternative experimental approach, using mathematical modelling based on population-based leukemia registry data to address the number of oncogenic steps to the development of APL, also predicted that a committed progenitor was the target of the rate-limiting step for leukemogenesis, with a translocation involving RARA being considered the most likely candidate.37


The case against APL arising in committed myeloid progenitors

While there are a number of lines of evidence supporting a transformation event at the level of committed myeloid progenitors, some experimental evidence is very difficult to reconcile with such a hypothesis. In particular, immunophenotypic characterization of primary APL blasts has revealed some highly atypical expression profiles. A proportion express CD34 (approximately 25%), and intriguingly for what is considered the most differentiated form of AML, a significant number express the natural killer (NK) cell-affiliated marker CD56 (approximately 15%), the B-cell marker CD19 (approximately 15%) and T-lineage-affiliated glycoprotein CD2 (approximately 25%).38,39,40 Furthermore, it has been confirmed that surface positivity for CD2 is a reflection of the respective mRNA expression level, rather than modulation of surface receptor, with levels in CD2-positive APL being comparable to those detected on normal resting T cells.41,42 The expression of CD34 and lymphoid markers in APL is associated with the hypogranular variant form of the disease, bcr3 PML breakpoint and higher presenting leukocyte count.38,41,43,44

The mechanism accounting for aberrant expression of lymphoid markers in AML still remains obscure and has been a matter of debate for over two decades. According to the 'lineage infidelity' model,45 this phenomenon is a consequence of deregulation of lineage-affiliated genes occurring during the process of leukemic transformation, whereas, according to the 'lineage promiscuity' model,46 coexpression of myeloid and lymphoid markers reflects the immunophenotype of the progenitor population subject to leukemic transformation which is perpetuated in the leukemic progeny. In order to gain further insights into mechanisms underlying coexpression of the T-lineage-affiliated CD2 gene in AML, a recent study investigated long-range chromatin structure surrounding the CD2 locus by DNase I hypersensitivity assay in primary APL blasts.42 The majority of cases examined were the hypogranular variant form of APL and in each case the CD2 locus was found to lie within an open chromatin environment, irrespective of the CD2 status at the cell surface. Given that previous studies have suggested that in early hematopoietic progenitors, multiple lineage-affiliated genes lie in an open chromatin configuration poised for transcription,47 it is possible that the results obtained in primary APL blasts could be indicative of the long-range chromatin configuration surrounding the CD2 locus in the progenitors targeted by the PML/RARA rearrangement. Recent studies that have considered the gene expression profiles of highly selected murine bone marrow progenitor populations suggest that myeloid and lymphoid genes are coexpressed in multipotent progenitors prior to lineage commitment.48,49 The process of myeloid commitment is correlated with progressive silencing of lymphoid- and NK-affiliated genes such that only myeloid and erythroid genes are expressed in common myeloid progenitors (CMPs). Conversely, lymphoid commitment sees a progressive silencing of myeloid genes, such that only B-, T- and NK-lineage-affiliated genes are expressed in common lymphoid progenitors (CLPs). If this widely accepted model of hematopoietic lineage commitment and differentiation is indeed correct, the 'lineage promiscuity' model would imply that some cases of APL arise in progenitors that have not undergone lineage restriction (see Figure 2a).

Figure 2.
Figure 2 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact or the author

Potential target cell populations for the PML-RARA fusion in APL, in the context of normal hematopoietic differentiation pathways. (a) Hematopoietic differentiation pathway according to the model developed by Weissman and co-workers.48 In the context of this scheme, evidence to date suggests that AML generally arises in the CD34+CD38-CD71-Thy1- cell fraction, which includes early hematopoietic progenitors prior to lineage commitment.20,63,64,65 Potential points at which PML-RARA could arise leading to induction of the APL phenotype are indicated by red arrows. According to this widely accepted model for the major hematopoietic differentiation pathways, one would predict that APL cases expressing lymphoid markers arise in multipotent progenitors, prior to lineage commitment, whereas in cases of APL lacking such markers, the t(15;17) could arise in myeloid committed cells equivalent to CMP, which expresses CD34. (b) Hematopoietic differentiation pathway according to the model recently proposed by Katsura and co-workers,74,75 which is based on analysis of hematopoiesis in murine fetal liver and postulates the existence of a range of distinct progenitor populations including those with myeloid and lymphoid differentiation potential. If these populations are shown to exist in human marrow and are permissive to transformation by the t(15;17), it could account for conflicting data regarding lineage involvement in APL as well as the phenomenon of lymphoid antigen expression, which occurs in over a quarter of cases. Potential points at which PML-RARA could arise leading to induction of the APL phenotype are indicated by red arrows. Panel a: HSCs, hematopoietic stem cells; MPPs, multipotent progenitors; CLP, common lymphoid progenitor; CMP, common myeloid progenitor; GMP, granulocyte/monocyte restricted progenitors; MEP, megakaryocyte/erythrocyte restricted progenitors; NK, natural killer cells; T, T cells/lineage; B, B cells/lineage; Monos, monocyte lineage; PMNs, granulocyte series; Megas, megakaryocytes; Erythro, erythroid lineage. In panel b, differentiation potentials of progenitor (p-) populations are denoted: B, B lymphoid; T, T lymphoid; M, myeloid; E, erythroid.

Full figure and legend (182K)

Interestingly, a recent study sought to address the nature of the progenitors subject to leukemic transformation by fusion genes involving MLL which is disrupted by translocations involving 11q23 and which are variably associated with AML, acute lymphoblastic leukemia (ALL) or mixed lineage leukemia. Using MLL-GAS7 which is generated by the t(11;17)(q23;p13) as a model, So et al50 convincingly showed that the leukemic phenotype generated by a single fusion gene is highly dependent on the nature of the progenitor targeted. Transformation of hematopoietic stem cells and multipotent progenitors induced AML, ALL or biphenotypic leukemia in in vivo models, whereas transformation of CMPs or CLPs gave rise solely to myeloid or lymphoid colonies, respectively.50 These data provide further support for the 'lineage promiscuity' hypothesis, whereby the immunophenotype of leukemic blasts is largely a reflection of their cell of origin. Indeed, Edwards et al51 have demonstrated the presence of the PML-RARA fusion by fluorescence in situ hybridization (FISH) within the majority of cells residing within the CD34+CD38- fraction in two cases of CD2-positive hypogranular variant APL. Taken together, these data suggest that a significant proportion of APL cases arise in more primitive progenitors than previously considered to be the case.


Clues to the origins of APL gained from understanding the activity of the PML-RARalpha oncoprotein

The functional activity of the APL-associated fusion proteins themselves also provides clues as to the likely nature of the hematopoietic progenitors that are subject to leukemic transformation. If APL is actually to arise in late myeloid progenitors that under normal circumstances have limited self-renewal capacity, it is clear that the fusion proteins generated by 17q21 rearrangements themselves or other associated 'second hits' must impart such characteristics. Experimental approaches to targeting expression of X-RARalpha proteins to human cellular fractions enriched for CMPs or more mature myeloid progenitors and determining their impact on self-renewal, proliferation and differentiation are fraught with difficulties and hence there are no published data. However, in the murine system, transgenic animals have been generated in which APL-associated fusion genes are expressed under the control of promoters that are expressed at the promyelocyte stage and even later in myelopoiesis. A number of groups have used a human cathepsin G (hCG) minigene to drive myeloid-specific expression of APL-associated fusion genes, including PML-RARA, PLZF-RARA and NPM-RARA. In each of these models, myelopoiesis was deregulated leading to an expansion in the myeloid compartment and increased colony formation, which in some lines was correlated with a myeloproliferative disorder with an incomplete block in myeloid differentiation.13,32,33,52 Expression of PML-RARA under the MRP8 promoter induced a shift from mature neutrophilic cells to more immature cells with a small increase in promyelocytes and enhanced cell survival following growth factor deprivation.31 A common feature of these models was that the frequency of APL development was relatively low (less than 30%) and invariably followed a long latency period. This led to the proposal that X-RARalpha proteins are insufficient in their own right to mediate the pathogenesis of APL, but nevertheless play a critical role in determining disease characteristics, particularly with regard to the response to RA and arsenic trioxide. Evidence supporting a requirement for 'second hits' in the pathogenesis of APL has been provided by the frequent observation of chromosomal aberrations at the point of leukemic transformation in PML-RARalpha mouse models53,54 and the increased penetrance of the leukemic phenotype in PML-RARalpha transgenic mice in the presence of the reciprocal fusion product RARalpha-PML, FLT3 ITD or BCL2 overexpression.55,56,57,58 Moreover, coexpression of the reciprocal RARalpha-PLZF fusion in PLZF-RARalpha transgenic mice transformed the phenotype from a myeloproliferative disorder into a disease more reminiscent of APL.59 Despite these data, the promyelocyte, which is a cell with limited proliferative potential and no self-renewal capacity, seems an unlikely target for transformation by APL-associated chromosomal translocations in man. Although it has been proposed that expression of the PML-RARA fusion under hCG and MRP8 promoters enhanced the survival of hematopoietic progenitors and may induce genomic instability, which together predispose to the acquisition of additional oncogenic events triggering leukemic transformation, it could be argued that the lesions found to cooperate with PML-RARalpha expressed at the level of promyelocytes are merely bestowing properties upon the targeted cells such as a proliferative response (FLT3 ITD) or favoring 'self-renewal' (BCL2, FLT3 ITD) that are characteristics already held to varying degrees by more primitive progenitors. A significant contributing factor to the prolonged latency period and low penetrance of current transgenic mouse models of APL may reflect the targeting of progenitors that may be much later than those in which APL-associated translocations are transforming in man. Indeed, it remains a possibility that the development of APL in these transgenic mouse models could be the result of leaky expression of the transgene in earlier progenitors, but the level and stage specificity of expression do not precisely match the pattern mediating the human disease.

Two recent studies support the notion that the APL-associated translocations arise prior to the promyelocyte stage. Minucci et al60 transduced purified murine lin- hematopoietic progenitors with retroviral vectors carrying the PML-RARA fusion, which led to increased proliferative potential and impaired the ability to undergo terminal myeloid differentiation in vitro. Moreover, inoculation of transduced lin- cells into syngeneic irradiated mice gave rise to a disease reminiscent of hypogranular variant APL in terms of clinical features and retinoid sensitivity in the majority of animals (>80%) associated with a short latency period (median 4 months). The leukemic blasts expressed CD34, were intensely positive for Sudan black, but were relatively agranular and lacked Auer rods by May–Grünwald–Giemsa stain. Although this mouse model was characterized by high penetrance and short latency in comparison to pre-existing transgenic models (see above), the lack of gross abnormalities in preleukemic marrow together with monoclonality or oligoclonality as determined by characterization of viral integration sites in some tumors still points to a requirement for additional 'oncogenic hits' for the pathogenesis of APL in the murine system.60 Nevertheless, an earlier study revealed that the PML-RARA fusion when expressed under the control of retroviral vectors in early human hematopoietic progenitors (CD34+lin-CD71-) exhibits a number of properties that are pertinent to the development of the APL phenotype.61 Expression of PML-RARalpha led to a rapid induction of progenitor differentiation to the promyelocyte stage, followed by an RA-responsive maturation block. Furthermore, irrespective of the hematopoietic growth factor stimulus, there was a differentiation bias toward the granulocytic lineage, implying a PML-RARalpha-mediated switch in lineage programming. Expression of PML-RARalpha was found to render hematopoietic progenitors resistant to apoptosis induced by hematopoietic growth factor withdrawal. The central role of transcriptional repression induced by the PML-RARalpha fusion protein in mediating the biological characteristics of APL was highlighted by the lack of impact of PML-RARalpha harboring a mutation in the N-CoR binding site on hematopoietic lineage commitment or myeloid differentiation. Overall, these experiments support the notion that the leukemogenic event in APL could occur in early progenitors. However, Grignani et al found that PML-RARalpha did not alter the proliferation/differentiation program of Dexter type 12-week extended long-term culture initiating cells (LTC-IC),61 representing stem cells62 suggesting that the progenitors targeted in APL lie at a distal stage in the differentiation pathway.61


Summing up

Many different experimental approaches have been adopted in an attempt to determine the nature of hematopoietic progenitors subject to leukemic transformation. As far as most forms of AML are concerned, all data available point to transformation at the level of primitive CD34+CD38-CD71-Thy1- hematopoietic progenitors prior to lineage commitment.20,63,64,65 However, the nature of the normal hematopoietic counterparts that are the target of chromosomal translocations in APL has been a matter of some debate. A widely held view is that APL arises at the level of myeloid-committed progenitors. This is based on experimental findings undertaken on primary leukemic cells taken from a very limited number of patients (less than 10) and the development of APL (albeit at low penetrance and long latency) in transgenic mouse models utilizing regulatory elements leading to fusion gene expression at the promyelocyte stage. However, a number of other studies using similar techniques have yielded contradictory results, which point toward APL arising in more primitive progenitors.

With respect to studies that have used primary patient material to investigate the origins of APL, evidence supporting the case that it arises from committed progenitors is based to a considerable degree on the study by Turhan et al.30 These investigators analyzed clinical material derived from three patients, finding the leukemic clone to be restricted to the myeloid lineage, with expression of PML-RARA being detected in the CD34+CD38+ cell population, but not in the more primitive CD34+CD38- subset.30 A subsequent study, which employed morphological analysis in conjunction with the FICTION technique, combining FISH and immunophenotype analysis, was in accordance.66 In each of the six cases of hypergranular APL examined, which all happened to lack expression of lymphoid-affiliated antigens, the PML-RARA fusion was restricted to the myeloid lineage, with no evidence of involvement of erythroblasts, plasma cells, or B or T lymphocytes.66 Similarly, the t(15;17) was found to be restricted to the granulocytic series in an earlier study that utilized the MAC method (morphology–antibody–chromosomes).67 However, two further studies yielded results with quite different implications regarding the origins of APL. In particular, Takatsuki et al68 detected PML-RARA mRNA in BFU-E in 2/5 cases examined. Unfortunately, precise characteristics of the patients studied in terms of morphological and immunological features were not provided; however, these findings could place the transformation event upstream of CMP in some cases of APL, according to current models of hematopoietic differentiation (see Figure 2a). Similar conclusions were reached by Edwards et al,51 who detected the PML-RARA fusion by FISH not only in the CD34+CD38+ population in accordance with Turhan et al,30 but also in CD34+CD38- cell progenitors. The differences in the experimental findings between the studies reported by Takatsuki et al and Turhan et al could potentially be ascribed to any number of reasons including possible differences in the sensitivity of RT-PCR assays, or lack of purity of the cell populations sampled for PCR analysis. However, the demonstration of the PML-RARA fusion by FISH in the majority of the CD34+CD38- cell population (>90%) in two cases with M3v51 provides strong evidence that APL arises in early progenitors in at least a proportion of cases. Moreover, since promyelocytes do not express CD34, it seems highly improbable that APL arises in myeloid progenitors that have differentiated beyond the granulocyte–monocyte precursor (GMP) stage, in which CD34 is still expressed (see Figure 2a). Indeed cases of hypergranular M3 that may be considered to represent the most differentiated form of the disease, which are CD34 negative according to conventional immunophenotype criteria, have nevertheless been shown to harbor the PML-RARA fusion in the majority of CD34+ cells.66 This is consistent with the existence of a hierarchy in APL as well as other subsets of AML.20 Furthermore, it would suggest that the current transgenic models of APL may not be successfully targeting expression of PML-RARalpha at an appropriate level to the progenitors that are subject to leukemic transformation in man.

The other key line of evidence supporting the hypothesis that APL arises from committed myeloid progenitors was the lack of engraftment of primary APL blasts derived from three patients in NOD/SCID mice, as distinct from other subsets of AML that consistently engrafted.20 However, this has not proved to be a consistent finding; indeed using a similar model, Ailles et al21detected engraftment in 3/5 APL cases examined, although levels of human cells at 8-weeks post-inoculation were significantly lower (mean 1.8%) in comparison to other subsets of AML, particularly those with adverse karyotypic features (mean 20.5%). Again, these data support the occurrence of the leukemia-initiating event in an early stem cell, at least in a subset of APL cases. The low levels of engraftment observed at 8 weeks suggest that NOD/SCID leukemia-initiating cells, that is, the leukemic stem cell pool, may be reduced in APL and/or that rates of proliferation and differentiation into more mature progeny are lower than in other subsets of AML. This does not seem unreasonable given the frequency with which APL cases present with leukopenia with few leukemic cells detectable in the peripheral blood.


Reconciliation and speculation

While at first glance it would appear that the reported studies are providing conflicting information regarding the stem cell targeted in APL, it is important to take into account the fact that sample sizes in all studies carried out to date have been small. As such, they cannot be expected to have encompassed the biological heterogeneity within the disease entity of APL, particularly with regard to morphological and immunophenotypic features. Moreover, the disparate experimental findings are readily reconciled if one considers that a range of hematopoietic progenitors may be permissive for the PML-RARA fusion, which include multipotent as well as myeloid committed progenitors (see Figure 2a). It is possible that targeting the latter population may give rise to classical hypergranular APL, which is a clinically less aggressive disease characterized by low presenting leukocyte count, bcr1 PML breakpoint and lower rates of FLT3 mutation (Figure 3). Whereas, data derived from a variety of experimental approaches suggest that the hypogranular variant form of APL may arise in much earlier progenitors; such cases have been shown to harbor the PML-RARA fusion in the CD34+CD38- cell fraction and coexpress lymphoid-affiliated antigens (see above). Interestingly, expression of PML-RARA in CD34+lin- hematopoietic progenitors in mice yields a leukemia reminiscent of hypogranular variant APL.60

Figure 3.
Figure 3 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact or the author

Hypothesis: clinical and biological characteristics of APL may vary according to the stem cell targeted. APL cases with hypogranular variant morphology (M3v) are typified by hyperleukocytosis and are associated with bcr3 PML breakpoint, activating mutations of FLT3 and coexpression of lymphoid antigens. Cases with hypergranular APL (M3) typically present with leukopenia, have a higher frequency of bcr1 PML breakpoint, more commonly have wild-type FLT3 and typically lack lymphoid markers. This raises the possibility that these distinct subsets of disease might be determined by the nature of the progenitor subject to leukemic transformation. LT-HSC, long-term hematopoietic stem cells; MPP, multipotent progenitors; CMP, common myeloid progenitor; GMP, granulocyte/monocyte restricted progenitors.

Full figure and legend (60K)

Nevertheless, there remain a number of areas of uncertainty. Of particular interest are the origins of APL cases expressing lymphoid-affiliated markers. According to the commonly accepted model of hematopoietic lineage commitment and differentiation, these features are consistent with transformation prior to myeloid specification (Figure 2a). However, recent studies have revealed some degree of plasticity within primary committed progenitors.69,70 These and related data have been reviewed by Graf,71 and in the case of APL it may be particularly pertinent to consider the possibility that leukemogenic, transcriptional deregulation of the myeloid-affiliated regulator CEBPalpha might result in lymphoid to myeloid switching. Along these lines, the documentation of T-lineage-affiliated chromatin features in APL42 and the observation that PML-RARalpha exerts lineage-altering properties61 raises the interesting possibility that even T-lineage progenitors could be potential targets in APL. Indeed, the U937 cell line, which harbors the CALM-AF10 fusion which is associated with acute undifferentiated leukemia and T-ALL,72 is one of the few cell lines to tolerate the PML-RARA fusion.73 Furthermore, application of a novel clonal assay that enables determination of myeloid, erythroid, B- and T-lymphoid potential of progenitor populations to investigate pathways involved in fetal hematopoiesis74 has called into question the commonly accepted hematopoietic differentiation pathway (reviewed in Katsura75). This assay supports the existence of common myeloid/lymphoid progenitors, which give rise to myeloid/B and myeloid/T populations (see Figure 2b). If such progenitor populations are indeed represented in human bone marrow and are permissive for the PML-RARA fusion, it could explain the disparate immunophenotypic features observed in APL, thereby lending further support to the 'lineage promiscuity' model46 to account for aberrant expression of lymphoid-affiliated antigens in AML. Furthermore, it would provide an explanation for the somewhat conflicting data regarding the lineages involved in the disease.

While it seems highly plausible that a variety of hematopoietic populations are permissive for the PML-RARA fusion on the basis of published data, the development of full-blown APL appears to be dependent upon the acquisition of a variable number of additional mutations. It will be interesting to determine to what extent the nature of the progenitors targeted dictate the nature and number of additional hits necessary to mediate the leukemic phenotype, or indeed influence the biological characteristics including response to therapy. There is an intriguing correlation between PML-RARalpha isoform, FLT3 ITD and morphological features,76,77,78 (Figure 3) and it will be interesting to dissect out to what extent the morphological features and leukocytosis associated with M3v reflect the targeting of a more primitive progenitor and/or the expression of the short isoform of PML-RARalpha or presence of activating mutations of FLT3. This raises even more fundamental questions regarding the mechanisms by which APL-associated translocations arise in the first place and whether distinct subsets of progenitors have differing propensities to the development of particular chromosomal breakpoints. It will be interesting to try to determine whether the nature of the stem cell influences the likelihood of breakpoints occurring in particular genomic hotspots or whether only certain subtypes of fusion protein are transforming in particular subsets of cells. Although seemingly esoteric, addressing these questions is likely to improve substantially our understanding of basic hematopoiesis as well as the processes involved in leukemic transformation.


Note added in proof

A further twist to the tale is provided by a very recent study from Lane and Ley,79 in which they provide evidence that proteolytic cleavage of the PML-RARalpha fusion protein by neutrophil elastase is important for the development of APL in mice expressing PML-RARA under the cathepsin G promoter. These findings are consistent with the putative target progenitors for the t(15;17) proposed in Figure 2, with PML-RARalpha conferring a survival advantage and promoting differentiation along the myeloid lineage. In addition, Lane and Ley's study suggests that as cells progress towards the promyelocyte stage, neutrophil elastase is upregulated leading to cleavage of PML-RARalpha contributing to leukemic transformation. It will be interesting to determine the implications of these findings for the pathogenesis of APL in man given that the preferential cleavage sites lie within regions encoded by PML exon 5 which is commonly subject to alternative splicing in primary patient samples. The nature of proteolytic cleavage appears to be distinct between bcr1 and bcr3 PML-RARalpha isoforms;79 as to whether this influences morphological and biological features encompassed within the spectrum of acute promyelocytic leukemia (Figure 3) remains to be established.



  1. Mistry AR, Wessel Pedersen E, Solomon E, Grimwade D. The molecular pathogenesis of acute promyelocytic leukaemia: implications for the clinical management of the disease. Blood Rev 2003; 17: 71–97. | Article | PubMed | ISI |
  2. Ohno R, Asou N, Ohnishi K. Treatment of acute promyelocytic leukemia. Strategy toward further increase of cure rate. Leukemia 2003; 17: 1454–1463. | Article | PubMed | ISI | ChemPort |
  3. Bennett JM, Catovsky D, Daniel MT, Flandrin G, Galton DA, Gralnick HR et al. Proposals for the classification of acute leukaemias. French–American–British (FAB) co-operative group. Br J Haematol 1976; 33: 451–458. | PubMed | ISI | ChemPort |
  4. Bennett JM, Catovsky D, Daniel MT, Flandrin G, Galton DA, Gralnick HR et al. Proposed revised criteria for the classification of acute myeloid leukaemia. A report of the French–American–British cooperative group. Ann Intern Med 1985; 103: 620–625. | PubMed | ISI | ChemPort |
  5. Collins SJ. The role of retinoids and retinoic acid receptors in normal hematopoiesis. Leukemia 2002; 16: 1896–1905. | Article | PubMed | ISI | ChemPort |
  6. Redner RL. Variations on a theme: the alternate translocations in APL. Leukemia 2002; 16: 1927–1932. | Article | PubMed | ISI | ChemPort |
  7. Pitha-Rowe I, Petty WJ, Kitareewan S, Dmitrovsky E. Retinoid target genes in acute promyelocytic leukemia. Leukemia 2003; 17: 1723–1730. | Article | PubMed | ISI | ChemPort |
  8. Di Croce L, Raker VA, Corsaro M, Fazi F, Fanelli M, Faretta M et al. Methyltransferase recruitment and DNA hypermethylation of target promoters by an oncogenic transcription factor. Science 2002; 295: 1079–1082. | Article | PubMed | ISI | ChemPort |
  9. Altucci L, Rossin A, Raffelsberger W, Reitmair A, Chomienne C, Gronemeyer H. Retinoic acid-induced apoptosis in leukemia cells is mediated by paracrine action of tumor-selective death ligand TRAIL. Nat Med 2001; 7: 680–686. | Article | PubMed | ISI | ChemPort |
  10. Zhu J, Lallemand-Breitenbach V, de Thé H. Pathways of retinoic acid- or arsenic trioxide-induced PML/RARalpha catabolism, role of oncogene degradation in disease remission. Oncogene 2001; 20: 7257–7265. | Article | PubMed | ISI | ChemPort |
  11. Hong SH, David G, Wong CW, Dejean A, Privalsky ML. SMRT corepressor interacts with PLZF and with the PML-retinoic acid receptor alpha (RARalpha) and PLZF-RARalpha oncoproteins associated with acute promyelocytic leukemia. Proc Natl Acad Sci USA 1997; 94: 9028–9033. | Article | PubMed | ChemPort |
  12. Guidez F, Ivins S, Zhu J, Söderström M, Waxman S, Zelent A. Reduced retinoic acid sensitivities of nuclear receptor corepressor binding to PML- and RARA underlie molecular pathogenesis and treatment of acute promyelocytic leukemia. Blood 1998; 91: 2634–2642. | PubMed | ISI | ChemPort |
  13. He LZ, Guidez F, Tribioli C, Peruzzi D, Ruthardt M, Zelent A et al. Distinct interactions of PML-RARalpha and PLZF-RARalpha with co-repressors determine differential responses to RA in APL. Nat Genet 1998; 18: 126–134. | Article | PubMed | ISI | ChemPort |
  14. Grignani F, De Matteis S, Nervi C, Tomassoni L, Gelmetti V, Cioce M et al. Fusion proteins of the retinoic acid receptor-alpha recruit histone deacetylase in acute promyelocytic leukaemia. Nature 1998; 391: 815–818. | Article | PubMed | ISI | ChemPort |
  15. Lin RJ, Nagy L, Inoue S, Shao W, Miller WH, Evans RM. Role of the histone deacetylase complex in acute promyelocytic leukaemia. Nature 1998; 391: 811–814. | Article | PubMed | ISI | ChemPort |
  16. Yoshida H, Naoe T, Fukutani H, Kiyoi H, Kubo K, Ohno R. Analysis of the joining sequences of the t(15;17) translocation in acute promyelocytic leukemia: sequence non-specific recombination between the PML and RARA genes within identical short stretches. Genes Chromosomes Cancer 1995; 12: 37–44. | Article | PubMed | ISI | ChemPort |
  17. Reiter A, Saubetaele S, Grimwade D, Wiemels JL, Segal MR, Lafage-Pochitaloff M et al. Genomic anatomy of the specific reciprocal translocation t(15;17) in acute promyelocytic leukemia. Genes Chromosomes Cancer 2003; 36: 175–188. | Article | PubMed |
  18. Shet AS, Jahagirdar BN, Verfaillie CM. Chronic myelogenous leukemia: mechanisms underlying disease progression. Leukemia 2002; 16: 1402–1411. | Article | PubMed | ISI | ChemPort |
  19. Kantarjian HM, Keating MJ, Talpaz M, Walters RS, Smith TL, Cork A et al. Chronic myelogenous leukemia in blast crisis. Analysis of 242 patients. Am J Med 1987; 83: 445–454. | Article | PubMed | ISI | ChemPort |
  20. Bonnet D, Dick JE. Human acute myeloid leukemia is organized as a hierarchy that originates from a primitive hematopoietic cell. Nat Med 1997; 3: 730–737. | Article | PubMed | ISI | ChemPort |
  21. Ailles LE, Gerhard B, Kawagoe H, Hogge DE. Growth characteristics of acute myelogenous leukemia progenitors that initiate malignant hematopoiesis in nonobese diabetic/severe combined immunodeficient mice. Blood 1999; 94: 1761–1772. | PubMed | ISI | ChemPort |
  22. Guan Y, Gerhard B, Hogge DE. Detection, isolation, and stimulation of quiescent primitive leukemic progenitor cells from patients with acute myeloid leukemia (AML). Blood 2003; 101: 3142–3149. | Article | PubMed | ISI | ChemPort |
  23. Al-Hajj M, Wicha MS, Benito-Hernandez A, Morrison SJ, Clarke MF. Prospective identification of tumorigenic breast cancer cells. Proc Natl Acad Sci USA 2003; 100: 3983–3988. | Article | PubMed | ChemPort |
  24. Welm BE, Tepera SB, Venezia T, Graubert TA, Rosen JM, Goodell MA. Sca-1(pos) cells in the mouse mammary gland represent an enriched progenitor cell population. Dev Biol 2002; 245: 42–56. | Article | PubMed | ISI | ChemPort |
  25. Alvi AJ, Clayton H, Joshi C, Enver T, Ashworth A, Vivanco MM et al. Functional and molecular characterisation of mammary side population cells. Breast Cancer Res 2003; 5: R1–8. | Article | PubMed |
  26. Scharenberg CW, Harkey MA, Torok-Storb B. The ABCG2 transporter is an efficient Hoechst 33342 efflux pump and is preferentially expressed by immature human hematopoietic progenitors. Blood 2002; 99: 507–512. | Article | PubMed | ISI | ChemPort |
  27. Williams GT. Programmed cell death: apoptosis and oncogenesis. Cell 1991; 65: 1097–1098. | Article | PubMed | ISI | ChemPort |
  28. Graham SM, Jorgensen HG, Allan E, Pearson C, Alcorn MJ, Richmond L et al. Primitive, quiescent, Philadelphia-positive stem cells from patients with chronic myeloid leukemia are insensitive to STI571 in vitro. Blood 2002; 99: 319–325. | Article | PubMed | ISI | ChemPort |
  29. Hochhaus A. Minimal residual disease in chronic myeloid leukaemia patients. Best Pract Res Clin Haematol 2002; 15: 159–178. | Article | PubMed |
  30. Turhan AG, Lemoine FM, Debert C, Bonnet ML, Baillou C, Picard F et al. Highly purified primitive hematopoietic stem cells are PML-RARA negative and generate nonclonal progenitors in acute promyelocytic leukemia. Blood 1995; 85: 2154–2161. | PubMed | ISI | ChemPort |
  31. Brown D, Kogan S, Lagasse E, Weissman I, Alcalay M, Pelicci PG et al. A PML-RARalpha transgene initiates murine acute promyelocytic leukemia. Proc Natl Acad Sci USA 1997; 94: 2551–2556. | Article | PubMed | ChemPort |
  32. Grisolano JL, Wesselschmidt RL, Pelicci PG, Ley TJ. Altered myeloid development and acute leukemia in transgenic mice expressing PML-RARalpha under control of cathepsin G regulatory sequences. Blood 1997; 89: 376–387. | PubMed | ISI | ChemPort |
  33. He LZ, Tribioli C, Rivi R, Peruzzi D, Pelicci PG, Soares V et al. Acute-leukemia with promyelocytic features in PML/RARalpha transgenic mice. Proc Natl Acad Sci USA 1997; 94: 5302–5307. | Article | PubMed | ChemPort |
  34. Westervelt P, Lane AA, Pollock JL, Oldfather K, Holt MS, Zimonjic DB et al. A high penetrance mouse model of acute promyelocytic leukemia with very low levels of PML-RARalpha expression. Blood 2003; 102: 1857–1865. | Article | PubMed | ISI | ChemPort |
  35. Westervelt P, Ley TJ. Seed vs soil: The importance of the target cell for transgenic models of human leukemias. Blood 1999; 93: 2143–2148. | PubMed | ISI | ChemPort |
  36. Early E, Moore MA, Kakizuka A, Nason-Burchenal K, Martin P, Evans RM et al. Transgenic expression of PML/RARalpha impairs myelopoiesis. Proc Natl Acad Sci USA 1996; 93: 7900–7904. | Article | PubMed | ChemPort |
  37. Vickers M, Jackson G, Taylor P. The incidence of acute promyelocytic leukemia appears constant over most of a human lifespan, implying only one rate limiting mutation. Leukemia 2000; 14: 722–726. | Article | PubMed | ISI | ChemPort |
  38. Guglielmi C, Martelli MP, Diverio D, Fenu S, Vegna ML, Cantù-Rajnoldi A et al. Immunophenotype of adult and childhood acute promyelocytic leukaemia: correlation with morphology, type of PML gene breakpoint and clinical outcome. A cooperative Italian study on 196 cases. Br J Haematol 1998; 102: 1035–1041. | Article | PubMed | ISI | ChemPort |
  39. Murray CK, Estey E, Paietta E, Howard RS, Edenfield WJ, Pierce S et al. CD56 expression in acute promyelocytic leukemia: a possible indicator of poor treatment outcome? J Clin Oncol 1999; 17: 293–297. | PubMed | ChemPort |
  40. Ferrara F, Morabito F, Martino B, Specchia G, Liso V, Nobile F et al. CD56 expression is an indicator of poor clinical outcome in patients with acute promyelocytic leukemia treated with simultaneous all-trans-retinoic acid and chemotherapy. J Clin Oncol 2000; 18: 1295–1300. | PubMed | ISI | ChemPort |
  41. Biondi A, Luciano A, Bassan R, Mininni D, Specchia G, Lanzi E et al. CD2 expression in acute promyelocytic leukemia is associated with microgranular morphology (FAB M3v) but not with any PML gene breakpoint. Leukemia 1995; 9: 1461–1466. | PubMed | ISI | ChemPort |
  42. Grimwade D, Outram SV, Flora R, Ings S, Pizzey AR, Morilla R et al. The T-lineage-affiliated CD2 gene lies within an open chromatin environment in acute promyelocytic leukemia cells. Cancer Res 2002; 62: 4730–4735. | PubMed | ChemPort |
  43. Claxton DF, Reading CL, Nagarajan L, Tsujimoto Y, Andersson BS, Estey E et al. Correlation of CD2 expression with PML gene breakpoints in patients with acute promyelocytic leukemia. Blood 1992; 80: 582–586. | PubMed | ISI | ChemPort |
  44. Foley R, Soamboonsrup P, Carter RF, Benger A, Meyer R, Walker I et al. CD34-positive acute promyelocytic leukemia is associated with leukocytosis, microgranular/hypogranular morphology, expression of CD2 and bcr3 isoform. Am J Hematol 2001; 67: 34–41. | Article | PubMed |
  45. Smith LJ, Curtis JE, Messner HA, Senn JS, Furthmayr H, McCulloch EA. Lineage infidelity in acute leukemia. Blood 1983; 61: 1138–1145. | PubMed |
  46. Greaves MF, Chan LC, Furley AJW, Watt SM, Molgaard HV. Lineage promiscuity in hemopoietic differentiation and leukemia. Blood 1986; 67: 1–11. | PubMed | ISI | ChemPort |
  47. Hu M, Krause D, Greaves M, Sharkis S, Dexter M, Heyworth C et al. Multilineage gene expression precedes commitment in the hemopoietic system. Genes Dev 1997; 11: 774–785. | Article | PubMed | ISI | ChemPort |
  48. Miyamoto T, Iwasaki H, Reizis B, Ye M, Graf T, Weissman IL et al. Myeloid or lymphoid promiscuity as a critical step in hematopoietic lineage commitment. Dev Cell 2002; 3: 137–147. | Article | PubMed | ISI | ChemPort |
  49. Akashi K, He X, Chen J, Iwasaki H, Niu C, Steenhard B et al. Transcriptional accessibility for genes of multiple tissues and hematopoietic lineages is hierarchically controlled during early hematopoiesis. Blood 2003; 101: 383–389. | Article | PubMed | ISI | ChemPort |
  50. So CW, Karsunky H, Passegue E, Cozzio A, Weissman IL, Cleary ML. MLL-GAS7 transforms multipotent hematopoietic progenitors and induces mixed lineage leukemias in mice. Cancer Cell 2003; 3: 161–171. | Article | PubMed | ISI | ChemPort |
  51. Edwards RH, Wasik MA, Finan J, Rodriguez R, Moore J, Kamoun M et al. Evidence for early hematopoietic progenitor cell involvement in acute promyelocytic leukemia. Am J Clin Pathol 1999; 112: 819–827. | PubMed | ChemPort |
  52. Cheng GX, Zhu XH, Men XQ, Wang L, Huang QH, Jin XL et al. Distinct leukemia phenotypes in transgenic mice and different corepressor interactions generated by promyelocytic leukemia variant fusion genes PLZF-RARalpha and NPM-RARalpha. Proc Natl Acad Sci USA 1999; 96: 6318–6323. | Article | PubMed | ChemPort |
  53. Le Beau MM, Davis EM, Patel B, Phan VT, Sohal J, Kogan SC. Recurring chromosomal abnormalities in leukemia in PML-RARA transgenic mice identify cooperating events and genetic pathways to acute promyelocytic leukemia. Blood 2003; 102: 1072–1074. | Article | PubMed | ChemPort |
  54. Zimonjic DB, Pollock JL, Westervelt P, Popescu NC, Ley TJ. Acquired, non-random chromosomal abnormalities associated with the development of acute promyelocytic leukemia in transgenic mice. Proc Natl Acad Sci USA 2000; 97: 13306–13311. | Article | PubMed | ChemPort |
  55. Pollock JL, Westervelt P, Kurichety AK, Pelicci PG, Grisolano JL, Ley TJ. A bcr-3 isoform of RARalpha-PML potentiates the development of PML-RARalpha-driven acute promyelocytic leukemia. Proc Natl Acad Sci USA 1999; 96: 15103–15108. | Article | PubMed | ChemPort |
  56. Kelly LM, Kutok JL, Williams IR, Boulton CL, Amaral SM, Curley DP et al. PML/RARalpha and FLT-ITD induce an APL-like disease in a mouse model. Proc Natl Acad Sci USA 2002; 99: 8283–8288. | Article | PubMed | ChemPort |
  57. Sohal J, Phan VT, Chan PV, Davis EM, Patel B, Kelly LM et al. A model of APL with FLT3 mutation is responsive to retinoic acid and a receptor tyrosine kinase inhibitor, SU11657. Blood 2003; 101: 3188–3197. | Article | PubMed | ISI | ChemPort |
  58. Kogan SC, Brown DE, Schultz DB, Truong B-TH, Lallemand-Breitenbach V, Guillemin MC et al. BCL-2 cooperates with promyelocytic retinoic acid receptor alpha chimeric protein (PML-RARalpha) to block neutrophil differentiation and initiate leukemia. J Exp Med 2001; 193: 531–543. | Article | PubMed | ISI | ChemPort |
  59. He LZ, Bhaumik M, Tribioli C, Rego EM, Ivins S, Zelent A et al. Two critical hits for promyelocytic leukemia. Mol Cell 2000; 6: 1131–1141. | Article | PubMed | ISI | ChemPort |
  60. Minucci S, Monestiroli S, Giavara S, Ronzoni S, Marchesi F, Insinga A et al. PML-RAR induces promyelocytic leukemias with high efficiency following retroviral gene transfer into purified murine hematopoietic progenitors. Blood 2002; 100: 2989–2995. | Article | PubMed | ISI | ChemPort |
  61. Grignani F, Valtieri M, Gabbianelli M, Gelmetti V, Botta R, Luchetti L et al. PML/RARalpha fusion protein expression in normal human hematopoietic progenitors dictates myeloid commitment and the promyelocytic phenotype. Blood 2000; 96: 1531–1537. | PubMed | ISI | ChemPort |
  62. Hao QL, Thiemann FT, Petersen D, Smogorzewska EM, Crooks GM. Extended long-term culture reveals a highly quiescent and primitive human hematopoietic progenitor population. Blood 1996; 88: 3306–3313. | PubMed | ISI | ChemPort |
  63. Lapidot T, Sirard C, Vormoor J, Murdoch B, Hoang T, Caceres-Cortes J et al. A cell initiating human acute myeloid leukaemia after transplantation into SCID mice. Nature 1994; 367: 645–648. | Article | PubMed | ISI | ChemPort |
  64. Blair A, Hogge DE, Ailles LE, Lansdorp PM, Sutherland HJ. Lack of expression of Thy-1 (CD90) on acute myeloid leukemia cells with long-term proliferative ability in vitro and in vivo. Blood 1997; 89: 3104–3112. | PubMed | ISI | ChemPort |
  65. Blair A, Hogge DE, Sutherland HJ. Most acute myeloid leukemia progenitor cells with long-term proliferative ability in vitro and in vivo have the phenotype CD34(+)/CD71(-)/HLA-DR-. Blood 1998; 92: 4325–4335. | PubMed | ISI | ChemPort |
  66. Haferlach T, Loffler H, Nickenig C, Ramm-Petersen L, Meeder M, Schoch R et al. Cell lineage specific involvement in acute promyelocytic leukaemia (APL) using a combination of May–Grunwald–Giemsa staining and fluorescence in situ hybridization techniques for the detection of the translocation t(15;17)(q22;q12). Br J Haematol 1998; 103: 93–99. | Article | PubMed |
  67. Knuutila S, Teerenhovi L, Larramendy ML, Elonen E, Franssila KO, Nylund SJ et al. Cell lineage involvement of recurrent chromosomal abnormalities in hematologic neoplasms. Genes Chromosomes Cancer 1994; 10: 95–102. | PubMed |
  68. Takatsuki H, Sadamura S, Umemura T, Abe Y, Goto T, Yufu Y et al. PML/RARalpha fusion gene is expressed in both granuloid/macrophage and erythroid colonies in acute promyelocytic leukaemia. Br J Haematol 1993; 85: 477–482. | PubMed | ChemPort |
  69. Kondo M, Scherer DC, Miyamoto T, King AG, Akashi K, Sugamura K et al. Cell-fate conversion of lymphoid-committed progenitors by instructive actions of cytokines. Nature 2000; 407: 383–386. | Article | PubMed | ISI | ChemPort |
  70. Heyworth C, Pearson S, May G, Enver T. Transcription factor-mediated lineage switching reveals plasticity in primary committed progenitor cells. EMBO J 2002; 21: 3770–3781. | Article | PubMed | ISI | ChemPort |
  71. Graf T. Differentiation plasticity of hematopoietic cells. Blood 2002; 99: 3089–3101. | Article | PubMed | ISI | ChemPort |
  72. Asnafi V, Radford-Weiss I, Dastugue N, Bayle C, Leboeuf D, Charrin C et al. CALM-AF10 is a common fusion transcript in T-ALL and is specific to the TCRitalic gammadelta lineage. Blood 2003; 102: 1000–1006. | Article | PubMed | ISI | ChemPort |
  73. Ferrucci PF, Grignani F, Pearson M, Fagioli M, Nicoletti I, Pelicci PG. Cell death induction by the promyelocytic leukemia-specific PML/RARalpha fusion protein. Proc Natl Acad Sci USA 1997; 94: 10901–10906. | Article | PubMed | ChemPort |
  74. Lu M, Kawamoto H, Katsube Y, Ikawa T, Katsura Y. The common myelolymphoid progenitor: a key intermediate stage in hemopoiesis generating T and B cells. J Immunol 2002; 169: 3519–3525. | PubMed | ISI | ChemPort |
  75. Katsura Y. Redefinition of lymphoid progenitors. Nat Rev Immunol 2002; 2: 127–132. | Article | PubMed | ISI | ChemPort |
  76. Kiyoi H, Naoe T, Yokota S, Nakao M, Minami S, Kuriyama K et al. Internal tandem duplication of FLT3 associated with leukocytosis in acute promyelocytic leukemia. Leukemia Study Group of the Ministry of Health and Welfare (Kohseisho). Leukemia 1997; 11: 1447–1452. | Article | PubMed | ISI | ChemPort |
  77. Noguera NI, Breccia M, Divona M, Diverio D, Costa V, De Santis S et al. Alterations of the FLT3 gene in acute promyelocytic leukemia: association with diagnostic characteristics and analysis of clinical outcome in patients treated with the Italian AIDA protocol. Leukemia 2002; 16: 2185–2189. | Article | PubMed | ISI | ChemPort |
  78. Shih LY, Kuo MC, Liang DC, Huang CF, Lin TL, Wu JH et al. Internal tandem duplication and Asp835 mutations of the FMS-like tyrosine kinase 3 (FLT3) gene in acute promyelocytic leukemia. Correlation with molecular subtypes of PML/RARalpha and clinical features. Cancer 2003; 98: 1206–1216. | Article | PubMed | ISI | ChemPort |
  79. Lane AA, Ley TJ. Neutrophil elastase cleaves PML-RARalpha and is important for the development of acute promyelocytic leukemia in mice. Cell 2003; 115: 305–318. | Article | PubMed | ISI | ChemPort |


DG is supported by the Leukaemia Research Fund of Great Britain. TE is supported by the MRC and a specialist programme from the Leukaemia Research Fund of Great Britain.