Original Article

Oncogene (2008) 27, 6356–6364; doi:10.1038/onc.2008.233; published online 4 August 2008

Evidence for Hox and E2A–PBX1 collaboration in mouse T-cell leukemia

J Bijl1,2,3, J Krosl1, C-E Lebert-Ghali2, J Vacher4, N Mayotte1 and G Sauvageau1,2,3,5

  1. 1Laboratory of Molecular Genetics of Hematopoietic Stem Cells, Institute for Research in Immunology and Cancer (IRIC), Université de Montréal, Montréal, Quebec, Canada
  2. 2Division of Hematology, Research Center of the Maisonneuve-Rosemont Hospital, Montréal, Quebec, Canada
  3. 3Department of Medicine, Université de Montréal, Montréal, Quebec, Canada
  4. 4Section of Cellular Interactions and Development, Clinical Research Institute of Montreal, Montréal, Quebec, Canada
  5. 5Leukemia Cell Bank of Québec, Maisonneuve-Rosemont Hospital, Montréal, Quebec, Canada

Correspondence: Dr J Bijl, Centre de Recherche Hôpital Maisonneuve-Rosemont, 5415 Boul. De l'Assomption, Montréal, Quebec, Canada H1T 2M4. E-mail: jbijl.hmr@ssss.gouv.qc.ca

Received 12 September 2007; Revised 3 June 2008; Accepted 10 June 2008; Published online 4 August 2008.



Using murine Moloney leukemia virus (MMLV)-based proviral insertional mutagenesis, we previously showed a preferential targeting of a small region in the Hoxa gene locus in E2A–PBX1-induced lymphoid leukemia resulting in the overexpression of several Hoxa genes including Hoxa10, Hoxa9 and Hoxa7. This observation suggested a functional interaction between Hox gene overexpression and E2A–PBX1 in lymphoid tumors. To further explore this possibility, we generated a series of compound E2A–PBX1 × Hox transgenic mice and tested the genetic interaction between these genes in the generation of lymphoid leukemia in vivo. Results presented in this report show that the onset of T-cell leukemia is significantly accelerated in E2A–PBX1 × Hoxb4 compound transgenic animals when compared with control E2A–PBX1 or Hoxb4 littermates. Hoxa9 appears less potent than Hoxb4 to accelerate E2A–PBX1-induced T-cell leukemia in mice. E2A–PBX1-induced T-cell leukemias express much higher levels of Hoxa genes than MMLV-induced counterparts, possibly suggesting a contribution of these genes to T-cell transformation by E2A–PBX1. Collectively, these data provide the first genetic evidence showing oncogenic collaboration between E2A–PBX1 and a Hox gene in lymphoid malignancies in vivo and document the specific deregulation of a subgroup of Hoxa genes in these leukemias.


Hoxb4, E2A–PBX1, collaborator oncogenes, T-ALL



Chromosome translocations and their corresponding gene fusions play a major role in the initiation of malignant transformation. The translocation t(1;19)(q23;p13) encoding the E2A–PBX1 fusion protein is present in about 6% of all B-cell acute lymphocytic leukemias (ALLs), 25% of pediatric pre-B-ALL, and in rare cases of myeloid and T-cell leukemias (Troussard et al., 1995; Look, 1997; Armstrong and Look, 2005). The translocation fuses the transactivation domains of the basic helix–loop–helix transcription factor E2A with the C-terminal region of the homeobox gene PBX1 (Kamps et al., 1990; Nourse et al., 1990). The E2A gene products, E12 and E47, are essential for early B- and T-cell development (Bain et al., 1994, 1997), and loss of their activity results in the development of T-cell lymphomas (Bain et al., 1997). Studies of Pbx1/ mutant mice revealed that Pbx1 is also required for the generation of common lymphoid precursors of B, NK (natural killer) and T cells (Sanyal et al., 2007). Results of in vitro studies suggest that fusion with the E2A domain converts the PBX1 homeodomain protein into a constitutive transcriptional transactivator (van Dijk et al., 1993; LeBrun and Cleary, 1994; Lu et al., 1994), indicating that the transforming potential of the E2A–PBX1 fusion likely results from the combined effects of haploinsufficiency at E2A and PBX1 loci, and the aberrant transcriptional activity of the E2A–PBX1 fusion protein.

PBX1 is a member of the TALE (for three amino-acid loop extension) family of homeodomain proteins, which also includes several MEIS and PREP members. PBX1, but not E2A–PBX1, forms DNA-binding heterodimers with MEIS1/PREP1 (Chang et al., 1997). All members of the TALE family, including E2A–PBX1, also interact with HOX proteins to form multiprotein complexes that participate in the regulation of gene expression. The cooperative interaction between PBX1 and HOX proteins was proposed to enhance the DNA-binding affinity and specificity of HOX proteins (Mann, 1995; van Dijk et al., 1995), and is required for the execution of some HOX-mediated functions (Azpiazu and Morata, 1998; Kroon et al., 1998; Medina-Martinez and Ramirez-Solis, 2003). Moreover, structure/function analyses showed that one of the E2A–PBX1 domains required for cellular transformation is also critical for cooperative DNA binding with HOX proteins and maps to a short region of PBX1 termed Hox Cooperativity Motif (HCM; Chang et al., 1997), indicating that E2A–PBX1/Hox interactions may play a role in E2A–PBX1-induced leukemias.

In a mouse bone marrow transplantation model, overexpression of E2A–PBX1 resulted in the development of myeloid leukemias with long latency (Kamps and Baltimore, 1993). Supporting the role for E2A–PBX1/Hox interactions in leukemogenesis, our studies have shown that co-overexpression of E2A–PBX1 and Hoxa9 or Hoxb3 significantly accelerates the onset of the disease (Thorsteinsdottir et al., 1999). Lymphoid-restricted overexpression of E2A–PBX1 induces T-cell lymphomas after approximately 5 months, clearly indicating a requirement for additional genetic events (Dedera et al., 1993; Bijl et al., 2005). Using murine Moloney leukemia virus (MMLV) insertional mutagenesis in a transgenic mouse model of E2A–PBX1-induced pre-B-cell leukemia, we recently reported that all leukemias analysed harbored proviral integrations in a small region of the Hoxa locus between Hoxa7 and Hoxa9 (Bijl et al., 2005). These leukemias expressed high levels of several Hoxa genes, suggesting that Hox gene products potentially collaborate with E2A–PBX1 in this disease.

In this study, we directly address whether E2A–PBX1 genetically interacts with Hox genes when co-overexpressed in lymphoid cells.



Description of transgenic mice used in these studies

Genetic interactions between E2A–PBX1 and Hoxa9 or Hoxb4 were studied using transgenic mice that express these transgenes in lymphoid cells. Hoxa9 and Hoxb4 were chosen as members of the Hox gene family showing potent and weak oncogenic potential, respectively (Kroon et al., 1998, 2001; Sauvageau et al., 2001; Thorsteinsdottir et al., 2002; Pineault et al., 2004). A brief summary of the different transgenics lines used in this study is provided in Table 1. Hoxa9 and E2A–PBX1 (isoform 1a) transgenic mice were previously described (Thorsteinsdottir et al., 2002; Bijl et al., 2005). Hoxb4 transgenic mice were never reported before and they are further described below. All mice used in these experiments were backcrossed over 10 generations and, except for E2A–PBX1 transgenics, they do not develop spontaneous tumors unless exposed to mutagens such as irradiation (data not shown) or MMLV (see Table 1 and below).

Hoxb4 transgenic mice were generated using the same regulatory elements as described previously for the Hoxa9 and E2A–PBX1 lines (see also Figure 1a). Western blot analyses revealed high levels of HOXB4 in the thymus, moderate levels in the spleen, but no detectable expression in the bone marrow (Figure 1b). This expression pattern is reminiscent to that observed with HOXA9 in the lines previously generated in our laboratory (Thorsteinsdottir et al., 2002). The impact of Hoxb4 overexpression on B- and T-cell development was first analysed in adult (>3-month-old) transgenic mice. Fluorescence-activated cell sorting (FACS) analyses revealed that the size of the lymphoid populations were comparable between Hoxb4 transgenic mice and littermate controls (Figure 1c and data not shown). In addition, no differences were observed in the proportions and absolute numbers of pre-B (B220+CD43+) and maturing B-cell populations (B220+IgM+), or in subpopulations of thymocytes (CD8CD4, CD8+CD4+, CD8+CD4 and CD8CD4+). The double-negative (DN) populations (CD8CD4Lin) were slightly more abundant in the Hoxb4 transgenic thymus, but the proportions in DN1–4 (from more primitive_1 to mature_4) subpopulations were comparable between Hoxb4 and control littermates (Supplementary Table 1). In accordance with these observations, similar frequencies of bone marrow cells capable of initiating long-term B-cell cultures (WW-IC) and long-term myeloid cultures (LTC-IC), as well as myeloid clonogenic progenitors were determined for Hoxb4 and control mice (Figure 1d and Supplementary Table 2). Despite normal mature B-cell populations, a tendency toward reduced B-cell progenitor numbers was observed in Hoxb4 transgenic mice (P=0.064, t-test, Figure 1d). Together and within the power of this study, these results suggest that lymphoid and myeloid cell development appears rather normal in our Hoxb4 transgenic mice.

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 help@nature.com or the author

Hoxb4 transgenic mice. (a) Scheme of the transgenic construct with lymphoid promoter and enhancer elements from the TCR Vβ, lck and immunoglobin-μ genes. (b) Western blot analysis of Hoxb4 protein levels in bone marrow (BM), spleen (S) and thymus (T) of Hoxb4 transgenics and control littermates. Tubulin levels are shown as a loading control. (c) Fluorescence-activated cell sorting (FACS) profiles of B cells (B220/CD43/IgM) and T cells (CD4/CD8) in the BM, spleen and thymus of Hoxb4 animals. (d) Left panel: determination of early B (Withlock–Witte culture-initiating cells, WW-ICs, n=3) and myeloid long-term culture-initiating cells (LTC-ICs, n=6) in Hoxb4 animals and littermate controls; center panel: B-lymphoid clonogenic progenitor numbers in Hoxb4 animals and littermate controls; right panel: myeloid clonogenic progenitors (colony-forming cells, CFCs, n=8) in the BM and spleen of Hoxb4 animals and littermate controls. *Note nonsignificant reduction in B-cell clonogenic progenitors (two-tail Student's t-test, P=0.064).

Full figure and legend (177K)

Proviral insertional mutagenesis in Hoxb4 and Hoxa9 transgenic mice

Using proviral insertional mutagenesis, we previously reported a high frequency of MMLV integrations in the Hoxa locus of lymphoid tumors that developed in our E2A–PBX1 transgenic mice. Using a similar approach, we investigated whether MMLV infection could accelerate the occurrence of lymphoid leukemias in cohorts of Hoxa9 and Hoxb4 transgenic mice when compared with control animals. To achieve this, newborn Hoxa9 (n=15) or Hoxb4 (n=18) transgenic mice and their control littermates (n=43) were injected intraperitoneally with MMLV, and monitored for disease development. Although no spontaneous lymphoid malignancies were ever observed in Hoxb4 or Hoxa9 transgenic animals that had been kept under observation for over 2 years (n=15 for Hoxb4 and n=13 for Hoxa9), we detected acute leukemia in 100% of the mice injected with MMLV whether they were transgenic (Hoxa9 and Hoxb4) or not (Figure 2a). Leukemia occurrence was significantly shorter in Hoxa9 mice than in Hoxb4 transgenics with mean survivals of 125±40 versus 195±65 days, respectively (P<0.01, Figures 2a and b). The shortest latency for disease development in Hoxa9 mice was 72 days, whereas the first Hoxb4 and control mice succumbed after 86 and 97 days, respectively (Figure 2a). Interestingly, the average onset of MMLV-induced leukemia was similar between the Hoxb4 transgenic mice and the control group (P=0.76, Figures 2a and b). Diseased animals presented with enlarged thymi and spleens compared with healthy controls that were not injected with MMLV, and showed infiltration of leukemic blasts in their bone marrow, livers, lungs and kidneys (Figure 2b and data not shown). FACS analyses indicated that both MMLV-infected transgenic and control mice succumbed to T-cell leukemias of CD4+, CD8+ or CD4+/CD8+ phenotypes. Representative FACS profiles from a CD4+ (Hoxb4) and a CD8+ (Hoxa9) leukemia are presented in Figure 2c.

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 help@nature.com or the author

(a) Proviral insertional mutagenesis. Survival curves of Hoxb4 and Hoxa9 transgenic and their control littermates injected intraperitoneally with murine Moloney leukemia virus (MMLV) 1 day after birth. Hoxa9 significantly accelerates the onset of the MMLV-induced T-cell leukemia (P<0.01), whereas MMLV-induced leukemia is not accelerated in the Hoxb4 background. (b) Tumor characteristics of MMLV-induced leukemia's in transgenic and normal context. T, thymus; LN, lymph nodes. (c). Fluorescence-activated cell sorting (FACS) profiles for bone marrow (BM), spleen and thymus of leukemic mice. Examples of a CD4- and CD8-positive leukemia induced by MMLV in either Hoxa9 or Hoxb4 transgenic background are given. Note that the leukemia infiltrated all three organs.

Full figure and legend (142K)

From this analysis, it appears that MMLV-induced T-cell tumors were accelerated in Hoxa9 transgenics and not in the Hoxb4 animals. Southern blot analyses of DNA isolated from MMLV-induced tumors showed that up to 32% of them had rearrangement in the c-myc gene. Few rearrangements were also detected in Pim1, Bmi1 and Notch1 (Supplementary Table 3).

T-cell leukemia development in compound Hox/E2A–PBX1 transgenic mice

To examine if Hoxa9 or Hoxb4 can accelerate the development of E2A–PBX1-induced lymphoid leukemia, Hox(b4 or a9) transgenic mice (1 line of each) were crossed with E2A–PBX1 transgenic animals (2 lines: 19 and 23) and their progeny were monitored for disease development. Spontaneous T-cell leukemias developed in the two lines of compound Hoxb4/E2A–PBX1 mice with a much shorter latency than observed in both lines of E2A–PBX1 mice (Figures 3a and b). As a result, compound Hoxb4/E2A–PBX1.19 and Hoxb4/E2A–PBX1.23 transgenic mice survived 103±20 and 90±15 days compared with 187±66 and 189±89 days determined for the cohorts of E2APBX1.19 and E2APBX1.23 animals, respectively (P<0.02, Figure 3b). The T-cell leukemias that were observed in compound Hoxb4/E2A–PBX1 mice expressed both the E2A–PBX1 (detected by the E47 antibody, Figure 3c) and the HOXB4 proteins (Figure 3c), and they shared similar phenotypes to those isolated from the two lines of E2A–PBX1 transgenic animals. The majority of these leukemias were CD4+CD8, CD4CD8+ or CD4+CD8+ (data not shown, see Figure 3d for two representative profiles and Supplementary Table 4).

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 help@nature.com or the author

(a) Survival curves of compound E2A–PBX1/Hoxb4 and E2A–PBX1/Hoxa9 transgenic mice and single E2A–PBX1 transgenic mice (lines19 and 23). For single E2A–PBX1 transgenic mice, only mice that succumbed to T-cell leukemia were included. (b) Survival and tumor characteristics of compound and single E2A–PBX1 transgenic mice with T-cell leukemia. (c) Western blot analysis of E2APBX1, and Hoxb4 protein levels in total cell lysates derived from thymoctes from a healthy Hoxb4 transgenic mouse (lane 1), leukemia cells from compound (lanes 2–5) and single transgenic mice (lane 6). Lanes 2 and 3 represent different organs from the same mouse. T, thymus; M, mediastinal, including thymus and lymph nodes; LNmes, mesenterial lymph node. (d) Fluorescence-activated cell sorting (FACS) profiles of the T-cell tumors derived from compound transgenic mice. Cells were stained with antibodies directed to CD4 and CD8 T-cell markers. Examples of CD4+/CD8+ and CD4low/CD8+ leukemias are given for E2APBX1-induced leukemias in either Hoxb4 or Hoxa9 background, respectively.

Full figure and legend (210K)

T-cell leukemias were also observed in Hoxa9/E2A–PBX1.23 compound transgenic mice, but the acceleration of disease latency was not as impressive as that determined for Hoxb4/E2A–PBX1.23 (Figure 3a). Although the collaboration with Hoxa9 did not reach statistical significance, there was a trend for disease acceleration as indicated by a reduction of 40 days in mean survival time between Hoxa9/E2A–PBX1.23 (149 days) and E2A–PBX1.23 (189 days, Figure 3c) and a noticeable acceleration in lethality for the longest survivors of both cohorts (compare the right portion of the survival curves in Figure 3a for Hoxa9/E2A–PBX1.23 (gray line) versus E2A–PBX1.23). Together, these results indicated that Hoxb4, and possibly Hoxa9, genetically interact with E2A–PBX1 in T-cell leukemia in mice.

Hoxa gene expression levels in E2A–PBX1 and MMLV-induced T-cell leukemias

We previously observed Hoxa genes deregulation in a mouse model of E2A–PBX1-induced leukemia (Bijl et al., 2005). Recent evidence further supports a role for Hoxa genes deregulation in several subtypes of human and mouse T-cell leukemias (Soulier et al., 2005; Speleman et al., 2005; Su et al., 2006; Van Vlierberghe et al., 2008), (Dik et al., 2005; Bergeron et al., 2006; Caudell et al., 2007; Cauwelier et al., 2007). To gain further insight into the expression levels of all 11 Hoxa genes in our T-cell leukemia specimens, we used quantitative reverse transcription–PCR analyses as previously described (Thompson et al., 2003). For this analysis, absolute Hoxa gene expression levels were first determined in normal unpurified thymocytes (n=4 specimens: two wild-type and two Hoxb4 transgenic mice). This analysis showed that normal thymocytes express high levels of certain Hox genes, namely Hoxa7, Hoxa4 and Hoxa5 all with copy numbers above 500. Other 3′ Hox genes, such as Hoxa1, a2, a3 and a6, are moderately expressed. 5′ Hox genes including Hoxa9, a10, a11 and a13 are expressed at very low levels (from 5 to 65 copies, Figure 4a). Although there was a tendency to decreased expression of 3′ Hoxa genes (Hoxa2 to Hoxa7) in the presence of the Hoxb4 transgene, Hoxb4 overexpression did not significantly change the profile of individual Hoxa gene expression in thymocytes except for Hoxb4 itself as copy numbers differ less than two-fold (thus <1 cycle time; Figures 4a and c).

Figure 4.
Figure 4 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

The expression of Hoxa cluster genes and Hoxb4 (in copy numbers) (a) in normal thymocytes derived from wild-type and Hoxb4 transgenic mice and (b) in five groups of T-cell acute lymphocytic leukemias (T-ALLs) induced by murine Moloney leukemia virus (MMLV) (columns 1 and 2) or E2A–PBX (columns 3–5) in a transgenic or wild-type background. The expression was determined in 50ng RNA, normalized for GAPDH (CT=18). Absolute copy numbers for each group are calculated from the average CT value in each group according to the following formula 2(38−ct). (c) Fold difference in Hoxa and Hoxb4 gene expression of Hoxb4 transgenic thymocytes and control or transgenic T-cell leukemias induced by murine Moloney leukemia virus (MMLV) or E2A–PBX1 normalized to the expression in wild-type thymocytes. Note that eight-fold difference (23), which corresponds to a ΔCT of 3, is considered significant.

Full figure and legend (223K)

We then compared Hoxa genes and Hoxb4 expression profiles between T-ALL from two subgroups: MMLV versus E2A–PBX1 (see Figure 4b). These results showed an important difference between the expression levels of several 3′ Hox genes (that is, Hoxa1 to Hoxa9) that are expressed at much higher levels in E2A–PBX1 tumors when compared with MMLV-induced tumors. This result cannot be attributed to phenotypal differences between the specimens as the expression levels of Hoxa genes in individual leukemia samples with different T-cell phenotype induced by MMLV (CD4+/CD8+ or CD4+, n=2 each) or by E2A–PBX1 (CD4+/CD8+ or CD8, n=2 each) were comparable in each group (data not shown). Although the number of tumors that could be analysed remains low, it is important to note that neither Hoxb4 nor Hoxa9 overexpression changed this trend except for a noticeable reduction in Hoxa6 and Hoxa7 expression levels in E2A–PBX1 × Hoxb4 leukemias (see Hoxb4 genotype in Figure 4b). 5′ Hoxa genes, such as Hoxa9, a10 and a11, are also more expressed in E2A–PBX1 leukemias than in MMLV-induced T-ALL. This difference was also blunted by the overexpression of Hoxb4 in the E2A–PBX1 tumors (compare fourth to third column in Figure 4b). Thus, MMLV-induced diseases, either from Hoxb4 transgenic mice or from littermate controls, showed the lowest values for Hoxa gene expression with the exception of Hoxa7 at 128 copies (Figure 4b). Similar to the expression in normal thymocytes, the tendency toward lower expression of Hoxa3 to Hoxa7 persisted in Hoxb4 transgenic compared with control MMLV-induced leukemias. These leukemias expressed 10–100-fold lower levels of Hoxa1 to Hoxa7 genes than detected in normal thymocytes (Figure 4c). Remarkably, Hoxb4 is expressed at moderate levels in all leukemias from both subgroups (Figure 4b). Interestingly, Hoxa7 and a6 expression levels are reduced in T-ALL that occurred in Hoxb4 transgenics (genotype in Figure 4b), possibly indicating some level of cross-regulation or that Hoxb4 expression compensates for these two genes. Finally, we observed that Hoxa10 and Hoxa11 expression is mostly restricted to E2A–PBX1 or to E2A–PBX1/Hoxa9 leukemias (Figures 4a–c).



A role for HOXA genes in human acute T-cell leukemia was first suggested by Ferrando et al. (2003), who observed increased expression of HOXA9 and HOXA10 in MLL rearranged T- and B-ALLs. Since then transcriptional activation of specific HOXA genes has been reported for several subsets of T-ALL with different cytogenetic abnormalities (Dik et al., 2005; Soulier et al., 2005; Speleman et al., 2005; Van Vlierberghe et al., 2008), identifying the activation of the HOXA cluster genes as an important event in T-cell leukemogenesis. Despite these observations, no mouse model has yet demonstrated a role for Hox genes in T-cell leukemia. The genetic interaction between Hoxb4 and E2A–PBX1 in the pathogenesis of mouse T-ALL as reported herein partly fills this gap. Moreover, this paper shows that Hoxa genes are expressed at higher levels in murine E2A–PBX1 T-cell leukemias than in similar diseases induced by proviral insertional mutagenesis (MMLV), thus supporting a role for the Hox network in E2A–PBX1-induced T-cell transformation.

The profound perturbations in T-cell development in bone marrow chimeras generated from the retroviral overexpression of Hoxb3 or Hoxa9 or Hoxa10 (Sauvageau et al., 1997; Thorsteinsdottir et al., 1997; Kroon et al., 1998), together with the reduction in lymphocyte numbers in Hoxa9/ mice, suggest that some of these genes might play an important role in T-cell development. Indeed, the expression of HOXA7, HOXA9, HOXA10 and HOXA11 was demonstrated in early thymic T progenitors, which was progressively lost with the differentiation collinear with their 3′ to 5′ localization in the cluster (Taghon et al., 2003). The required downregulation of HOXA genes for T-cell differentiation was illustrated by enforced expression of HOXA10 in human cord blood progenitors in fetal thymic organ cultures (Taghon et al., 2002). Thus, the expression of Hoxa10 and Hoxa11, and the relative high expression of the 3′ located genes, including Hoxa5 and Hoxa7, might contribute to disease development in our E2A–PBX1 transgenic mice possibly by preventing T-cell maturation.

The observed difference in collaboration of Hoxa9 versus Hoxb4 with E2A–PBX1 remains intriguing. Although our small screen could not detect a significant contribution of Hoxa9 to the E2a-PBX1-dependent leukemogenesis, we cannot exclude that Hoxa9 and E2A–PBX1 do collaborate. It is possible that T-cell progenitors at stages before transgene expression are sensitive to high levels of Hoxa9 (Taghon et al., 2003). The collaboration between Hoxa9 and MMLV-induced leukemogenesis dampens—but does not eliminate—this possibility.

This screen also highlighted the low probability for the creation of conditions required to uncover the Hoxb4 transformation ability. In contrast to Hoxa9, overexpression of Hoxb4 alone has never resulted in hematological malignancies. However, some recipients of bone marrow cells co-infected with Hoxb4 and antisense Pbx1 cDNA carrying recombinant retroviruses presented with AML (Cellot et al., 2007). Leukemic clones in all cases harbored high numbers of the integrated proviruses, suggesting that retroviral insertions perturbed the activity of oncogenes that promoted transformation in the Hoxb4highPbx1low cellular context. Recently, a study in large animals showed that overexpression of Hoxb4 alone can eventually lead to myeloid leukemia. The leukemic cells contained a high number of retroviral integrations near oncogenes such as LMO2 and prdm16 after extended (~2 years) latency (Zhang et al., 2008). Finally, co-overexpression of Hoxb4 and Meis1 was reported to result in the development of AML (Pineault et al., 2004 and N Mayotte and GS, data not shown). Hoxb4 thus appears to posses a weak leukemogenic potential that can be activated with a limited subset of oncogenes (Meis1, LMO2 and prdm16), which now includes E2A–PBX1.

Our data suggest that the cellular context restricts the oncogenic activity of Hox genes to specific oncogenes that are activated by insertional mutagenesis (MMLV) for Hoxa9 and by E2A–PBX for Hoxb4. Interestingly, our expression data presented in Figures 4b and c suggest that Hoxb4 expression (provided as a transgene) may overcome the need for certain Hoxa gene activation in E2A–PBX T-ALL, implying that interactions between E2A–PBX and Hox genes may be critical and deterministic.


Materials and methods

Transgenic animals

Transgenic mice for Hoxb4 and Hoxa9 were generated using the same vector (Thorsteinsdottir et al., 2002) and methodology as that described for the recently reported E2A–PBX1 lines (Bijl et al., 2005). The backbone of the pLIT3 vector has been described elsewhere (Hough et al., 1994). In short, cDNA was cloned downstream of a T-cell receptor Vβ promoter, immunoglobulin (Ig) enhancer elements and sequences from the proximal promoter of the lck gene. The human growth hormone (hGH) gene containing a frameshift mutation in the coding region was inserted 3′ of the transgene to provide introns for enhanced expression. Transgenic animals were generated using standard techniques (Hogan, 1983). Basically, the transgene was injected into pronuclei of (C57Bl/6J × C3H) F2 hybrid zygotes. A total of 2–4h after injection, surviving eggs were transplanted into oviducts of pseudopregnant CD-1 host females. Transgenic progeny were identified by dot blot analysis of tail genomic DNA using an hGH probe.

Statistical analysis

Evaluation of significant acceleration of disease in the Hox transgenic background for MMLV-induced or E2A–PBX1-induced lymphomagenesis was obtained by performing a two-tailed Student's t-test. The rate of acceleration is considered significant when P-value <0.05.

Flow cytometry

Bone marrow, spleen and thymus cells were isolated from leukemic mice and incubated with the following antibody sets to detect T cell: CD4-FITC and CD8-PE (BD Bioscience Pharmingen, San Diego, CA, USA); B cell: B220-PE, B220-PE-cy5, CD19-APC, BP1-PE, CD43-FITC, IgM-bio and IgD-FITC (all from BD Bioscience Pharmingen), myeloid cell populations: Mac1-bio and Gr1-FITC (both from BD Bioscience Pharmingen). Biotinylated antibodies were detected with phycoerythrin (PE), allophycoerythrin (APC) (BD Bioscience Pharmingen) or PE-cy7 (e-Bioscience, San Diego, CA, USA)-conjugated streptavidin. Fluorescence was analysed using the FACS Calibur (BD Bioscience, San Jose, CA, USA) or the LSR II (BD Bioscience), using DIVA software. Multicolor FACS data were analysed with WinMDI or FCS Express (De Novo Software, Los Angeles, CA, USA) software.

Quantitative RT–PCR

Total RNA was isolated from fresh tumor cells (BM, spleen, thymus and enlarged lymph nodes,) by Trizol (Invitrogen, Life Technologies, Carlsbad, CA, ISA), and Dnase I treated. cDNA was generated from 5μg of RNA using MMLV reverse transcriptase (Invitrogen) according to the manufacturer's instructions. Quantitative PCR was performed in triplicate for each sample as described before (Bijl et al., 2005), using SYBR Green (Applied Biosystems, Toronto, Canada) and run on a thermal cycler ABI 7500 (Applied Biosystems). Primers for the detection of Hox gene expression were used as designed by Thompson et al., and were validated for use with SYBR green. Triplicates were accepted in a 0.5 CT range.

Protein analysis

Preparation of cellular extracts derived from control, single or compound transgenic mice and western blot analyses were performed as described (Krosl and Sauvageau, 2000). Briefly, proteins were separated by gel electrophoresis using 10% polyacrylamide–SDS and transferred to Immobilon P membranes (Millipore, Bedford, MA, USA). Membranes were blocked with 5% non-fat milk in TBST (20mM Tris-Cl, pH 7.6, 140mM NaCl and 0.05% Tween 20) and incubated with E47 antibody (Santa Cruz Biotechnology Inc., Santa Cruz, CA, USA) to detect E2APBX1 or an antibody detecting Hoxb4 (Krosl et al., 2003). Bound antibodies were detected using horseradish peroxidase-conjugated anti-rabbit antibodies (Sigma, St Louis, MO, USA) followed by enhanced chemiluminescence (ECL; Amersham, Buckinghamshire, UK).

In vitro clonogenic progenitor assays

For myeloid clonogenic progenitor assays, cells were plated in 35mm dishes in 1% methylcellulose in DMEM (Dulbecco's modified Eagle media) supplemented with 10% fetal calf serum (FCS), 5.7% bovine serum albumin, 10−5 β-mercaptoethanol (β-ME), 5U of erythropoietin (Epo) per milliliter, IL-3 10ng/ml, IL-6 10ng/ml, steel 50ng/ml, 2mM glutamine and 200mg transferrin per milliliter. Fetal liver (FL) cells of mutant and control embryos were plated at concentrations of 0.5 × 105 cells per milliliter. BM and spleen from Hoxb4 mice were plated at 3 × 104 and 1 × 106 cells per milliliter, respectively. Colonies were scored on days 12–14 of incubation and identified according to standard criteria. Pre-B-cell clonogenic progenitor assays were only performed using cells isolated from Hoxb4 and control mice. For this assay, 3 × 105 bone marrow cells or 1 × 106 spleen cells were plated in 1% methylcellulose in DMEM supplemented with 30% FCS (selected for B cells), 10−4 β-ME, 2mM glutamine and 0.2ng/ml IL-7. Pre-B-cell colonies were scored on day 8.



  1. Armstrong SA, Look AT. (2005). Molecular genetics of acute lymphoblastic leukemia. J Clin Oncol 23: 6306–6315. | Article | PubMed | ISI | ChemPort |
  2. Azpiazu N, Morata G. (1998). Functional and regulatory interactions between Hox and extradenticle genes. Genes Dev 12: 261–273. | Article | PubMed | ISI | ChemPort |
  3. Bain G, Engel I, Robanus Maandag EC, Te Riele HP, Voland JR, Sharp LL et al. (1997). E2A deficiency leads to abnormalities in alphabeta T-cell development and to rapid development of T-cell lymphomas. Mol Cell Biol 17: 4782–4791. | PubMed | ISI | ChemPort |
  4. Bain G, Maandag EC, Izon DJ, Amsen D, Kruisbeek AM, Weintraub BC et al. (1994). E2A proteins are required for proper B cell development and initiation of immunoglobulin gene rearrangements. Cell 79: 885–892. | Article | PubMed | ISI | ChemPort |
  5. Bergeron J, Clappier E, Cauwelier B, Dastugue N, Millien C, Delabesse E et al. (2006). HOXA cluster deregulation in T-ALL associated with both a TCRD-HOXA and a CALM-AF10 chromosomal translocation. Leukemia 20: 1184–1187. | Article | PubMed | ISI | ChemPort |
  6. Bijl J, Sauvageau M, Thompson A, Sauvageau G. (2005). High incidence of proviral integrations in the Hoxa locus in a new model of E2a-PBX1-induced B-cell leukemia. Genes Dev 19: 224–233. | Article | PubMed | ChemPort |
  7. Caudell D, Zhang Z, Chung YJ, Aplan PD. (2007). Expression of a CALM-AF10 fusion gene leads to Hoxa cluster overexpression and acute leukemia in transgenic mice. Cancer Res 67: 8022–8031. | Article | PubMed | ChemPort |
  8. Cauwelier B, Cave H, Gervais C, Lessard M, Barin C, Perot C et al. (2007). Clinical, cytogenetic and molecular characteristics of 14 T-ALL patients carrying the TCRbeta-HOXA rearrangement: a study of the Groupe Francophone de Cytogenetique Hematologique. Leukemia 21: 121–128. | Article | PubMed | ISI | ChemPort |
  9. Cellot S, Krosl J, Chagraoui J, Meloche S, Humphries RK, Sauvageau G. (2007). Sustained in vitro trigger of self-renewal divisions in Hoxb4hiPbx1(10) hematopoietic stem cells. Exp Hematol 35: 802–816. | Article | PubMed | ISI | ChemPort |
  10. Chang CP, De Vivo I, Cleary ML. (1997). The Hox cooperativity motif of the chimeric oncoprotein E2a-Pbx1 is necessary and sufficient for oncogenesis. Mol Cell Biol 17: 81–88. | PubMed | ISI | ChemPort |
  11. Dedera DA, Waller EK, LeBrun DP, Sen-Majumdar A, Stevens ME, Barsh GS et al. (1993). Chimeric homeobox gene E2A–PBX1 induces proliferation, apoptosis, and malignant lymphomas in transgenic mice. Cell 74: 833–843. | Article | PubMed | ISI | ChemPort |
  12. Dik WA, Brahim W, Braun C, Asnafi V, Dastugue N, Bernard OA et al. (2005). CALM-AF10+ T-ALL expression profiles are characterized by overexpression of HOXA and BMI1 oncogenes. Leukemia 19: 1948–1957. | Article | PubMed | ISI | ChemPort |
  13. Ferrando AA, Armstrong SA, Neuberg DS, Sallan SE, Silverman LB, Korsmeyer SJ et al. (2003). Gene expression signatures in MLL-rearranged T-lineage and B-precursor acute leukemias: dominance of HOX dysregulation. Blood 102: 262–268. | Article | PubMed | ISI | ChemPort |
  14. Hogan B. (1983). Molecular biology. Enhancers, chromosome position effects, and transgenic mice. Nature 306: 313–314. | Article | PubMed | ISI | ChemPort |
  15. Hough MR, Takei F, Humphries RK, Kay R. (1994). Defective development of thymocytes overexpressing the costimulatory molecule, heat-stable antigen. J Exp Med 179: 177–184. | Article | PubMed | ChemPort |
  16. Kamps MP, Baltimore D. (1993). E2A–Pbx1, the t(1;19) translocation protein of human pre-B-cell acute lymphocytic leukemia, causes acute myeloid leukemia in mice. Mol Cell Biol 13: 351–357. | PubMed | ISI | ChemPort |
  17. Kamps MP, Murre C, Sun XH, Baltimore D. (1990). A new homeobox gene contributes the DNA binding domain of the t(1;19) translocation protein in pre-B ALL. Cell 60: 547–555. | Article | PubMed | ISI | ChemPort |
  18. Kroon E, Krosl J, Thorsteinsdottir U, Baban S, Buchberg AM, Sauvageau G. (1998). Hoxa9 transforms primary bone marrow cells through specific collaboration with Meis1a but not Pbx1b. EMBO J 17: 3714–3725. | Article | PubMed | ISI | ChemPort |
  19. Kroon E, Thorsteinsdottir U, Mayotte N, Nakamura T, Sauvageau G. (2001). NUP98-HOXA9 expression in hemopoietic stem cells induces chronic and acute myeloid leukemias in mice. EMBO J 20: 350–361. | Article | PubMed | ISI | ChemPort |
  20. Krosl J, Sauvageau G. (2000). AP-1 complex is effector of Hox-induced cellular proliferation and transformation. Oncogene 19: 5134–5141. | Article | PubMed | ISI | ChemPort |
  21. Krosl J, Austin P, Beslu N, Kroon E, Humphries RK, Sauvageau G. (2003). In vitro expansion of hematopoietic stem cells by recombinant TAT-HOXB4 protein. Nat Med 9: 1428–1432. | Article | PubMed | ISI | ChemPort |
  22. LeBrun DP, Cleary ML. (1994). Fusion with E2A alters the transcriptional properties of the homeodomain protein PBX1 in t(1;19) leukemias. Oncogene 9: 1641–1647. | PubMed | ISI | ChemPort |
  23. Look AT. (1997). Oncogenic transcription factors in the human acute leukemias. Science 278: 1059–1064. | Article | PubMed | ISI | ChemPort |
  24. Lu Q, Wright DD, Kamps MP. (1994). Fusion with E2A converts the Pbx1 homeodomain protein into a constitutive transcriptional activator in human leukemias carrying the t(1;19) translocation. Mol Cell Biol 14: 3938–3948. | PubMed | ISI | ChemPort |
  25. Mann RS. (1995). The specificity of homeotic gene function. Bioessays 17: 855–863. | Article | PubMed | ISI | ChemPort |
  26. Medina-Martinez O, Ramirez-Solis R. (2003). In vivo mutagenesis of the Hoxb8 hexapeptide domain leads to dominant homeotic transformations that mimic the loss-of-function mutations in genes of the Hoxb cluster. Dev Biol 264: 77–90. | Article | PubMed | ChemPort |
  27. Nourse J, Mellentin JD, Galili N, Wilkinson J, Stanbridge E, Smith SD et al. (1990). Chromosomal translocation t(1;19) results in synthesis of a homeobox fusion mRNA that codes for a potential chimeric transcription factor. Cell 60: 535–545. | Article | PubMed | ISI | ChemPort |
  28. Pineault N, Abramovich C, Ohta H, Humphries RK. (2004). Differential and common leukemogenic potentials of multiple NUP98-Hox fusion proteins alone or with Meis1. Mol Cell Biol 24: 1907–1917. | Article | PubMed | ISI | ChemPort |
  29. Sanyal M, Tung JW, Karsunky H, Zeng H, Selleri L, Weissman IL et al. (2007). B-cell development fails in the absence of the Pbx1 proto-oncogene. Blood 109: 4191–4199. | Article | PubMed | ChemPort |
  30. Sauvageau G, Thorsteinsdottir U, Hough MR, Hugo P, Lawrence HJ, Largman C et al. (1997). Overexpression of HOXB3 in hematopoietic cells causes defective lymphoid development and progressive myeloproliferation. Immunity 6: 13–22. | Article | PubMed | ISI | ChemPort |
  31. Sauvageau G, Thorsteinsdottir U, Mayotte N, Jerome LJ. (2001). Evidence that chromosome instability and stem cell expansion cooperate to induce leukemia in mice. Blood 98(Part: I): 836a.
  32. Soulier J, Clappier E, Cayuela JM, Regnault A, Garcia-Peydro M, Dombret H et al. (2005). HOXA genes are included in genetic and biologic networks defining human acute T-cell leukemia (T-ALL). Blood 106: 274–286. | Article | PubMed | ISI | ChemPort |
  33. Speleman F, Cauwelier B, Dastugue N, Cools J, Verhasselt B, Poppe B et al. (2005). A new recurrent inversion, inv(7)(p15q34), leads to transcriptional activation of HOXA10 and HOXA11 in a subset of T-cell acute lymphoblastic leukemias. Leukemia 19: 358–366. | Article | PubMed | ISI | ChemPort |
  34. Su X, Drabkin H, Clappier E, Morgado E, Busson M, Romana S et al. (2006). Transforming potential of the T-cell acute lymphoblastic leukemia-associated homeobox genes HOXA13, TLX1, and TLX3. Genes Chromosomes Cancer 45: 846–855. | Article | PubMed | ChemPort |
  35. Taghon T, Stolz F, De Smedt M, Cnockaert M, Verhasselt B, Plum J et al. (2002). HOX-A10 regulates hematopoietic lineage commitment: evidence for a monocyte-specific transcription factor. Blood 99: 1197–1204. | Article | PubMed | ISI | ChemPort |
  36. Taghon T, Thys K, De Smedt M, Weerkamp F, Staal FJ, Plum J et al. (2003). Homeobox gene expression profile in human hematopoietic multipotent stem cells and T-cell progenitors: implications for human T-cell development. Leukemia 17: 1157–1163. | Article | PubMed | ISI | ChemPort |
  37. Thompson A, Quinn MF, Grimwade D, O'Neill CM, Ahmed MR, Grimes S et al. (2003). Global down-regulation of HOX gene expression in PML-RARalpha+acute promyelocytic leukemia identified by small-array real-time PCR. Blood 101: 1558–1565. | Article | PubMed | ChemPort |
  38. Thorsteinsdottir U, Krosl J, Kroon E, Haman A, Hoang T, Sauvageau G. (1999). The oncoprotein E2A–Pbx1a collaborates with Hoxa9 to acutely transform primary bone marrow cells. Mol Cell Biol 19: 6355–6366. | PubMed | ISI | ChemPort |
  39. Thorsteinsdottir U, Mamo A, Kroon E, Jerome L, Bijl J, Lawrence HJ et al. (2002). Overexpression of the myeloid leukemia-associated Hoxa9 gene in bone marrow cells induces stem cell expansion. Blood 99: 121–129. | Article | PubMed | ISI | ChemPort |
  40. Thorsteinsdottir U, Sauvageau G, Hough MR, Dragowska W, Lansdorp PM, Lawrence HJ et al. (1997). Overexpression of HOXA10 in murine hematopoietic cells perturbs both myeloid and lymphoid differentiation and leads to acute myeloid leukemia. Mol Cell Biol 17: 495–505. | PubMed | ISI | ChemPort |
  41. Troussard X, Rimokh R, Valensi F, Leboeuf D, Fenneteau O, Guitard AM et al. (1995). Heterogeneity of t(1;19)(q23;p13) acute leukaemias. French Haematological Cytology Group. Br J Haematol 89: 516–526. | Article | PubMed | ChemPort |
  42. van Dijk MA, Peltenburg LT, Murre C. (1995). Hox gene products modulate the DNA binding activity of Pbx1 and Pbx2. Mech Dev 52: 99–108. | Article | PubMed | ChemPort |
  43. van Dijk MA, Voorhoeve PM, Murre C. (1993). Pbx1 is converted into a transcriptional activator upon acquiring the N-terminal region of E2A in pre-B-cell acute lymphoblastoid leukemia. Proc Natl Acad Sci USA 90: 6061–6065. | Article | PubMed | ChemPort |
  44. Van Vlierberghe P, van Grotel M, Tchinda J, Lee C, Beverloo HB, van der Spek PJ et al. (2008). The recurrent SET-NUP214 fusion as a new HOXA activation mechanism in pediatric T-cell acute lymphoblastic leukemia. Blood 111: 4668–4680. | Article | PubMed | ChemPort |
  45. Zhang XB, Beard BC, Trobridge GD, Wood BL, Sale GE, Sud R et al. (2008). High incidence of leukemia in large animals after stem cell gene therapy with a HOXB4-expressing retroviral vector. J Clin Invest 118: 1502–1510. | Article | PubMed | ChemPort |


We thank Melanie Frechette for assistance with the animal care and transplantation experiments, and to the personnel from the transgenic animal facility at the Institute for Clinical Research of Montreal (IRCM) for the generation of Hoxb4 and Hoxa9 transgenic lines. We also thank Eric Massicotte and Martine Dupuis from the flow cytometry service of IRCM, and Danielle Gagné from the flow cytometry service of IRIC for assistance in FACS analysis. Dr Alex Thompson (Queen's University, Belfast, UK) is thanked for critically reading the manuscript. This work was supported by funds from the Canadian Cancer Society through a NCIC grant to GS (no. 018478) and the Canada Research Chair program also to GS.

Supplementary Information accompanies the paper on the Oncogene website (http://www.nature.com/onc)