Transgenic Mice Expressing the p75 CCAAT-Displacement Protein/Cut Homeobox Isoform Develop a Myeloproliferative Disease–Like Myeloid Leukemia

The CCAAT-displacement protein/cut homeobox (CDP/Cux) belongs to a family of transcription factors present in all metazoans and is involved in the control of proliferation and differentiation (reviewed in ref. 1). The gene was first identified in Drosophila melanogaster and was named after the “cut wing” phenotype (2). In higher vertebrates, there are two Cut-like genes called Cut-like 1 (CUTL1) and Cut-like 2 (CUTL2) in human, and Cut-homeobox 1 (Cux1) and Cut-homeobox 2 (Cux2) in mouse and chicken (3–5). Hereafter in the text, the term CDP/Cux will be used to describe the protein encoded by the human CUTL1 gene. Although Cux1 is expressed in most tissues, Cux2 expression is restricted primarily to nervous tissues (3). Cux1 knockout mice generally die during the perinatal period, but 20% survive into adulthood. Such surviving mice displayed phenotypes in various organs including curly whiskers, growth retardation, delayed differentiation of lung epithelia, altered hair follicle morphogenesis, male infertility, and a deficit in T and B cells (6–8). In contrast to the small size of the Cux1 knockout mice, transgenic mice expressing Cux1 displayed multiorgan hyperplasia and organomegaly (9). Thus, from genetic studies both in Drosophila and the mouse, it is clear that the CDP/Cux/Cut gene plays an important role in the development and homeostasis of several tissues.

At least three CDP/Cux protein isoforms can be expressed as the result of proteolytic processing or transcription initiation at an alternative start site: p200, p110, and p75. The full-length protein, p200 CDP/Cux, is a complex protein with four evolutionarily conserved DNA-binding domains: three Cut repeats (CR1, CR2, and CR3) and a Cut homeodomain (Fig. 1; refs. 4, 10–12). The NH₂-terminal end of the full-length protein harbors an autoinhibitory domain that inhibits DNA binding (13). Two active transcriptional repression domains are present within the carboxyl-terminal domain (R1 and R2; refs. 14–16). The full-length protein was found to be proteolytically processed at the G₁-S transition of the cell cycle, thereby generating the p110 CDP/Cux isoform which contains three DNA-binding domains, CR2, CR3, and HD (17). In addition, a tissue-specific mRNA species was found to code for the p75 CDP/Cux isoform which contains only two DNA-binding domains: CR3 and HD (3, 18). Molecular studies showed that the full-length protein, p200, binds rapidly but transiently to DNA and carries the CCAAT displacement activity (17, 19). In contrast, the p110 and p75 isoforms behave like classical transcription factors that engage in slow but stable interactions with DNA (17, 18). CDP/Cux was originally shown to function in precursor cells of various lineages as a transcriptional repressor that down-modulates lineage-specific genes that later become expressed in terminally differentiated cells (1, 20–24). In addition, recent evidence suggests that the processed isoform can also participate in transcriptional activation and can stimulate cell proliferation by accelerating entry into S phase (25, 26).

The I20-mRNA and p75 proteins were found to be expressed primarily in the placenta and thymus, but aberrant expression was observed in breast cancer cells (18). In invasive tumors, a significant association was established between higher I20-mRNA expression and a diffuse infiltrative growth pattern (18). In the present study, we set out to test the hypothesis that p75 CDP/Cux may play a causal role in cancer. We used a knock-in strategy into the hypoxanthine phosphoribosyltransferase (hprt) locus to generate transgenic mice containing a p75 CDP/Cux transgene under the control of the mouse mammary tumor virus-long terminal repeat (MMTV-LTR) (27). To our surprise, transgenic mice from the first generations of backcrosses to both the FVB and C57BL/6 strains of mouse succumbed to a disease involving the hematopoietic system. In this report, we characterize the cellular abnormalities encountered in these transgenic mice.

Plasmids. Plasmid sequences and maps will be provided upon request. Expression vectors for CDP/Cux included the complete MMTV-LTR as well as coding sequences for the indicated amino acids of the human CDP/Cux protein: 1-1505, 817-1505, and 1062-1505 (GenBank accession no., M74099). For introduction into embryonic stem cells, all of the above constructs were directionally inserted into a slightly modified version of the hprt targeting vector, pMP8SKB (a gift from Sarah Bronson, Pennsylvania State University, Hershey, PA; ref. 27).

Cell culture and electroporation. BK4 embryonic stem cells were originally derived from the 129/Ola strain and contain a deletion within the hprt locus that prevents expression of the hprt gene (28). Cell culture and electroporation were done as previously described (29). PCR analysis on genomic DNA was done using a forward primer from the 5′-flanking hprt genomic sequences (5′-ggcagaagtagaattaggcttttcagg-3′) and a reverse primer from the MMTV-LTR sequence (5′-caaccccttggctgcttctcc-3′).

Generation of transgenic mice. All experiments involving animals were conducted in accordance with the McGill University Animal Care Guidelines. Targeted embryonic stem cells were injected into C57BL/6-derived blastocysts that were then transplanted into the uteri of recipient females. Resulting chimeric males were bred with C57BL/6 females, and the F1 agouti female offspring were backcrossed with C57BL/6 males. Genotyping was done by PCR analysis of genomic DNA prepared from mouse tail biopsy using a forward primer from the MMTV-LTR sequence and a reverse primer from the CDP/Cux sequence. Two lines of p75 CDP/Cux transgenic mice were generated, p75-48 and p75-50, each from an independent blastocyst microinjection with different embryonic stem cell clones. Because the transgene was integrated into the hprt locus on chromosome X, the transgene would be expected to be expressed in ∼50% of the cells in females and in 100% of cells in males. Statistics on penetrance were obtained with female mice exclusively.

Monitoring mice. Mice were palpated every week for the development of ascites. Premoribund mice were anesthetized with isofluorane and euthanized by cervical dislocation. Hematopoietic organs (spleen, lymph nodes, thymus, and bone marrow), liver and lungs were harvested, weighed, and analyzed by histology, flow cytometry, reverse transcription-PCR (RT-PCR), and Western blot. The measurement of hematologic variables was done upon sacrifice (Diagnostic Laboratory, Animal Resource Center, McGill University) and blood smears were stained with a Wright stain.

Histology. Tissues were fixed in 4% paraformaldehyde for 24 hours. Paraffin-embedded sections were then stained with H&E. CD11b+ splenocytes were stained with Diff-Quik Stain (Dade Behring, Düdingen, Switzerland).

RT-PCR. RNA was prepared using TRIzol (Invitrogen, Carlsbad, CA) and cDNA was prepared using the Superscript II RNase H-reverse transcriptase kit (Invitrogen). Real-time PCR was done on a LightCycler (Roche, Basel, Switzerland) using the FastStart DNA Master SYBR Green kit (Roche).

Preparation of protein extracts. Total protein extracts were prepared by homogenizing tissue in NP40 buffer [150 mmol/L of NaCl, 50 mmol/L of Tris (pH 8.0), 1% NP40, 10% glycerol, 0.5 mmol/L of DTT, protease inhibitor mix tablet (from Roche)] and mixing for 30 minutes at 4°C. The extracts were then centrifuged for 15 minutes at 4°C and the supernatants collected. Nuclear extracts were prepared as described previously (17).

Flow cytometry analysis. Single cell suspensions were prepared from bone marrow, spleen, lymph nodes, thymus, blood, and infiltrated organs (liver and lung), and RBC were lysed with ACK buffer (0.15 mol/L of NH₄Cl, 1 mmol/L of KHCO₃, 0.1 mmol/L of Na₂-EDTA; adjusted to pH 7.2-7.4). Livers were incubated for 20 minutes at 37°C with collagenase prior to preparation of single cell suspensions. Cells were then resuspended in 37% Percoll and centrifuged to isolate hematopoietic cells. From the single cell suspensions obtained, 10⁶ cells were incubated for 15 minutes on ice in blocking solution (2.4G2) and were then stained with monoclonal antibodies conjugated with phycoerythrin, FITC, or biotin, to detect either myeloid cells, B lymphocytes, or T lymphocytes. The following antibodies, obtained from BD PharMingen (Palo Alto, CA) and Cedarlane (Hornby, ON, Canada), were used: CD11b (Mac-1), Gr-1, 7/4, CD4, CD8, B220, IgM, F4/80, and CD11c. Cells were submitted to a FACScan flow cytometer (Becton Dickinson, Franklin Lakes, NJ) and data were analyzed with FLOWJO software developed by Tree Star (San Carlos, CA).

Immunoblotting. Western blot analyses with actin (Santa Cruz Biotechnology, Santa Cruz, CA; 1/2,000) and 1300 (1/1,000) were done as previously described (17).

Isolation of CD11b-positive splenocytes. Splenocytes were stained with CD11b-biotin antibody after RBC lysis with ACK. The stained cells were applied on streptavidin magnetic beads (MACS) for 15 minutes and then passed through a magnetic column (MACS). Greater than 97% of the cells recovered were CD11b-positive. TRIzol was added to these cells to extract mRNA and for RT-PCR analyses.

Transplantation. Ten million splenocytes were transplanted into sublethally irradiated (650 rad) immunocompromised nude mice. Recipient mice were followed by observation of their health status and also by blood test analyses.

The p75 CDP/Cux isoform can be expressed from the MMTV-LTR. The p110 and p75 CDP/Cux isoforms were previously reported to be aberrantly expressed in uterine leiomyomas and breast tumors, respectively (18, 30). To assess the oncogenic potential of p110 and p75 CDP/Cux, we set out to generate transgenic mice that would express these isoforms specifically in mammary epithelial cells. We generated an expression vector carrying the coding sequences for p110 or p75 CDP/Cux downstream of the MMTV-LTR. These regulatory sequences were previously shown to drive expression in mammary epithelial cells, but transgene expression has been reported in other cell types, particularly in the hematopoietic system (31–33). To investigate expression of the three CDP/Cux isoforms under the control of the MMTV-LTR, the corresponding expression plasmids were introduced by transfection into 293 cells, nuclear protein extracts were prepared and Western blot analysis was done using the CDP/Cux antibody 1300 (Fig. 1A). Both the p75 and p110 isoforms were expressed in 293 cells, but little, if any, p200 expression was detected (Fig. 1A , lanes 1-3). We consider it likely that the role of p200 as a transcriptional repressor of MMTV-LTR precludes its expression from the same regulatory sequences (34, 35). However, our results clearly showed that both the p110 and p75 CDP/Cux isoforms can be expressed using the MMTV-LTR regulatory sequences.

Generation of p75 CDP/Cux transgenic mice by specific transgenesis. To avoid complications resulting from variations in copy number and integration site effects, we used the method of targeted transgenesis to insert the construct into the mouse hprt locus (27). To confirm integration by homologous recombination, PCR analysis was done using a forward primer from the 5′-flanking hprt genomic sequences and a reverse primer from the MMTV-LTR sequences (Fig. 1B). Two independent lines of p75 CDP/Cux transgenic mice were generated, p75-48 and p75-50. In previous studies, development and cancer of the mammary gland has generally been analyzed in the FVB strain of mouse (36, 37). Therefore, we initiated a series of backcrosses to the FVB strain and, as a control, the C57BL/6 strain. As the backcrosses were under way, a high proportion of mice from the first generations, both from the FVB and C57BL/6 backcrosses, succumbed to what seemed to be a similar disease characterized by splenomegaly and infiltration of WBC to other organs (Table 1; Supplemental Table S1). In contrast, the p110 transgenic mice did not display a higher incidence of this disease (Table 1). Here, we describe the phenotype of the p75-transgenic mice.

Expression of the p75 CDP/Cux transgene in several tissues. A 10-month-old healthy virgin female transgenic mouse from the backcross 1 (BC1) generation to C57BL/6 was sacrificed and transgene expression was investigated by RT-PCR (Fig. 1C). We observed elevated expression of the transgene in the spleen, kidney, and brain, and relatively lower expression in the liver (Fig. 1C , lanes 2, 4, 5, and 7). Because the mouse had not been subjected to perfusion, we cannot exclude that the expression detected in some tissues, in particular in the liver and brain, was not due to blood contamination. Considering that the transgene was integrated into the hprt locus and that one chromosome X is inactivated in each cell, in females, ∼50% of the cells would be expected to express the transgene in those tissues showing expression.

A fraction of p75 CDP/Cux transgenic mice developed hepatosplenomegaly. A cohort of p75 CDP/Cux transgenic (n = 60) and wild-type mice (n = 35) were kept until moribund. Mice were palpated every week for the development of ascites and enlarged abdomen. A fraction of transgenic mice (20 of 60) developed what seemed to be a hematopoietic disorder characterized by hepatosplenomegaly (Fig. 2; Supplemental Table S1). The enlarged spleens of diseased mice weighed, on average, 18.8 times that of normal spleens (Table 2). The latency was fairly short in two cases (9 months) but was generally much longer with an average of 20 months. Overall, mice from backcrosses 1 to 3 in C57BL/6, and from backcross 1 in FVB, developed the disease with a penetrance of 33% (20 of 60) and an average latency of 20.3 months (Table 1). Note that statistics on penetrance were obtained with female mice exclusively. In comparison, nontransgenic littermates developed the disease with a penetrance of 8.6% (3 of 35 mice) and an average latency of 20.5 months (Table 1). Therefore, the susceptibility to this disease was increased approximately four times in p75 transgenic mice.

Splenomegaly was caused by myeloid hyperplasia. Histologic analysis revealed that the enlarged spleens of transgenic mice did not maintain a normal architecture. Regions of white pulp were compressed, whereas there was a large number of cells resembling polymorphonuclear leukocytes (Fig. 2B). Indeed, ring-shaped nuclei characteristic of neutrophils were observed following the purification of CD11b-positive cells and staining with Diff-Quik (Fig. 2D). In three cases (mice p75-48-54, p75-50-116, and p75-50-129), an increased number of both granulocytes and megakaryocytes was observed in the spleen (Fig. 2D).

Nonhematopoietic organs were infiltrated in many sick mice. When observed at autopsy, the lungs and livers of sick mice were of unusual color and texture, suggesting infiltration by another cell type (Fig. 2A; Supplemental Table S1). Histologic examination confirmed the presence of an overrepresented population of cells in the liver, lungs, kidneys, and blood (Fig. 2B and C). In contrast, in most cases, the thymus and lymph nodes of sick mice seemed normal upon morphologic and histologic examinations, and exhibited normal staining patterns in flow cytometry analysis (data not shown).

Anemia and thrombocytopenia in sick transgenic mice. Hematologic variables were measured in the peripheral blood of affected transgenic and wild-type littermate mice (Table 3; Fig. 2C). Although the hemoglobin concentration, hematocrit, and number of RBC and platelets were significantly decreased in sick transgenic mice, the number of WBC was significantly increased. In most sick transgenic mice, the increase in WBC resulted from an increase in the number of neutrophils. Morphologic analysis of blood smears did not reveal the presence of progranulocytes, myelocytes, metamyelocytes, or band cells. Myeloblasts were detected in only one sick transgenic mouse (Table 3, mouse #116). Lymphopoiesis did not seem to be impaired in affected transgenic mice because the number of lymphocytes did not vary significantly. However, because the number of neutrophils was increased, the percentage of neutrophils in the blood was significantly increased at the expense of the percentage of lymphocytes. In summary, the analysis of blood smears revealed that affected transgenic mice suffered from anemia and thrombocytopenia, whereas their blood showed an increase in the number of neutrophils.

The expanded cell population expressed neutrophil surface markers. A large proportion of cells extracted from the spleen, bone marrow, liver, and blood of transgenic mice displayed increased size and granulosity as shown from side and forward scatter analyses (Supplemental Fig. S1). Flow cytometry analysis was done on 15 affected transgenic mice and 16 wild-type littermates using antibodies specific for cell surface markers of myeloid cells (CD11b, 7/4, and Gr-1), B lymphocytes (B220, IgM), and T lymphocytes (CD4, CD8; Table 2; Supplemental Fig. S1; Supplemental Table S2). With the exception of two mice that presented with a hematopoietic disorder involving the lymphoid compartment (p75-48-129 and p75-48-136, see below), most sick transgenic mice displayed an excess of CD11b-positive cells in the spleen, bone marrow, liver, and blood. In contrast, B220/IgM-positive cells were underrepresented in the spleen, bone marrow, and liver of diseased transgenic mice. CD11b/Gr-1/7/4 triple staining revealed a large overrepresentation of neutrophils in the spleen and blood of diseased transgenic mice (Table 2; Supplemental Fig. S1D). Moreover, in the spleen, we observed both low and high Gr-1 expression levels among these cells. These findings suggest that the overrepresented cells are neutrophils and that complete maturation can occur in this cell lineage. Altogether, data from histologic and flow cytometry analyses indicated that most of the affected p75 CDP/Cux transgenic mice displayed an overrepresentation of neutrophils in the blood, spleen, liver, and bone marrow, and in many cases, in nonhematopoietic organs like the kidneys and lungs.

The disease is not transplantable. Splenocytes from seven sick transgenic mice were transplanted into sublethally irradiated immunocompromised nude mice. After a period of 10 to 13 months, no recipient has yet developed a hematopoietic disorder. We conclude that the disease is not transplantable.

Other less frequent hematopoietic disorders in p75 CDP/Cux transgenic mice. A few transgenic mice succumbed to different hematopoietic disorders. The presence of peripheral myeloblasts in the blood smear of a transgenic mouse suggested a blast transformation (Table 3, p75-50-116). Two transgenic mice seemed to suffer from a disorder involving the lymphoid compartment: flow cytometry analysis of the spleen revealed an overrepresentation of CD4-positive cells in one mouse (Table 2, p75-48-129), whereas examination of spleen sections revealed the presence of a large number of atypical lymphocytes with numerous mitotic figures in the other (p75-48-136, data not shown). Finally, one mouse showed both an excess of myeloid cells in the spleen and a grossly enlarged lymph node containing mostly B cells (B220⁺IgM⁻; Table 2, p75-48-73; data not shown).

The p75 CDP/Cux transgene is expressed in the spleen, liver, and CD11b⁺ cells of leukemic transgenic mice. The transgene mRNA was expressed in the spleen and liver of transgenic mouse and in CD11b⁺ cells purified from the spleen (Fig. 3A,, lanes 2, 4, and 5). Transgene protein expressions were investigated in four groups of healthy or leukemic transgenic mice and their respective wild-type littermates. The p75 protein was observed in three leukemic transgenic mice, but very little or no expression was detected in the wild-type littermates or in the healthy transgenic mice (Fig. 3B, leukemic transgenic mice in lanes 2, 5, and 7; wild-type littermates in lanes 1, 3, 6, and 8; healthy transgenic mouse in lanes 4 and 9). These results suggest that the development of the disease is associated with increased expression of the p75 CDP/Cux protein.

Expression of CDP/Cux target genes in the spleen and CD11b⁺ cells of leukemic mice. Previous studies have identified a number of CDP/Cux putative targets, some with relevance to cancer (9, 25, 38, 39). Total RNA was prepared from the enlarged spleen of four leukemic mice and from CD11b⁺ cells purified from the enlarged spleens of two leukemic mice. As a control, total RNA was prepared from the spleen of a wild-type littermate. However, because the spleen of leukemic mice contained a higher proportion of CD11b cells than that from normal mice, as a second control, we prepared RNA from CD11b⁺ cells isolated from a pool of spleens from five wild-type C57BL/6 mice. Changes in the expression levels of some genes were detected, however, apart for the p27 CDK inhibitor, no change in expression was consistently observed in all leukemic mice. Interestingly, the levels of p27 mRNA seemed to be decreased in the spleen and CD11b+ cells of all leukemic mice (Fig. 3C). Semiquantitative real-time PCR analysis was done to verify mRNA levels for both p27 and p21 (Fig. 3D). The results confirmed that p27 mRNA levels were reduced in the spleen and CD11b+ cells of all leukemic mice.

Our study revealed that p75 CDP/Cux transgenic mice display increased susceptibility to a myeloproliferative disease. In investigating the etiology of the disease, it was particularly important to determine whether the disease was the result of reactive conditions as opposed to a neoplastic process. Several criteria enabled us to exclude that an inflammatory response was responsible for the expansion of the CD11b/Gr-1 compartment in various organs. First, cells of affected organs were predominantly of granulocytic lineage, whereas reactive myeloid hyperplasia is characterized by a heterogeneous population of myeloid, erythroid, and megakaryocytic cells. Second, in contrast to reactive processes, in many transgenic mice, CD11b/Gr-1 cells spread to the liver and caused its enlargement, and sometimes spread to some nonhematopoietic organs like the lungs and kidneys. Third, the disease developed without apparent contribution from any environmental factor. Nonhematopoietic tumors were absent from the affected mice. The mice were maintained in a specific pathogen–free environment and were not exposed to drugs or toxins. Transgenic and nontransgenic littermates were exposed to identical environmental conditions because they were maintained in the same cage yet, only 3 of 35 nontransgenic mice developed disease. Moreover, the latency was significantly longer in mice obtained from the FVB versus the C57BL/6 backcross populations, further supporting the genetic basis of the disease.

The Bethesda proposals for the classification of nonlymphoid hematopoietic neoplasms in mice provides precise criteria for accurate diagnosis (40). The disease that developed most often in p75 CDP/Cux transgenic mice meets several of the criteria that define a nonlymphoid leukemia. The disease diffusely involved hematopoietic tissues with an increase in myeloid cells in both the spleen and bone marrow, and was accompanied by anemia and thrombocytopenia (Fig. 2; Tables 2 and 3; Supplemental Tables S1 and S2; Supplemental Fig. S1). Myeloid cells were often disseminated in nonhematopoietic organs including the liver, lungs, and kidneys (Fig. 2). Leukocytosis was present except for one exception (p75-48-48, 32% CD34⁺ cells in spleen), nonlymphoid immature forms/blasts did not make up 20% of leukocytes in the peripheral blood, spleen, or bone marrow (Tables 2 and 3; Supplemental Table S2; data not shown). Indeed, myeloid cells did not lose their capacity to differentiate as shown by the presence of CD11b/7/4 cells with high Gr-1 expression levels (Table 2; Supplemental Fig. S1; Supplemental Table S2). The absence of a block to differentiation was also in agreement with the slow progression of the disease and, except for one case, the absence of an acute phase (data not shown). In addition, we note that the disease had not developed 10 months after the transplantation of neoplastic myeloid cells into sublethally irradiated histocompatible recipients. Altogether, these criteria would define the disease as a myeloproliferative disease–like myeloid leukemia.

Although the diseased p75 mice did not exhibit the typical cytogenetic rearrangements involving the BCR/ABL oncogene, the myeloproliferative disease in p75 CDP/Cux transgenic mice shares several features with chronic myelogenous leukemia (CML) in humans. In particular, not only did the disease evolve slowly, but except for two mice that became moribund at 9 months of age, it affected old individuals without appreciably shortening their life spans. In this respect, the p75 CDP/Cux transgenics might prove to be a useful mouse model to study some aspects of CML. In addition to a BCR-ABL transgenic line in which p210 was driven by the tec promoter, three knockouts were also found to exhibit a CML-like phenotype. Inactivation of either the IFN consensus sequence binding protein (ICSBP), JunB or the estrogen receptor β (ERβ) invariably resulted in a CML-like disease that is characterized by an elevation of neutrophils in hematopoietic tissues (refs. 41–44, reviewed in ref. 45). The latency was short for ICSBP and JunB knockouts, but latency was long for the ERβ knockout. One important difference between these mouse models and the transgenics described here is that the inactivation of these genes occurred early, affected all cells, and had an effect on the behavior of hematopoietic stem cells, whereas the p75 CDP/Cux transgene was expressed in only half of the cells and later in ontogeny, most likely within a committed myeloid progenitor. Future studies should determine whether there are common transcriptional targets or signal transduction pathways operating in p75 CDP/Cux or BCR-ABL-expressing cells, and in ICSBP, JunB, or ERβ knockout mice. In light of the phenotypic similarities in such mice, it is possible that CDP/Cux, ICSBP, JunB, and ERβ are involved in a common transcriptional regulatory network or even function as direct regulators of one another. Also, it remains to be determined if up-regulation of p75 CDP/Cux transcription or activity is an etiologic factor in human CML, in particular, among the 5% of patients who carry a diagnosis of CML but do not harbor the Philadelphia (Ph) chromosome.

Our finding that MMTV-p75 CDP/Cux transgenic mice developed a myeloproliferative disease was unexpected. We believe two reasons explain the low incidence of tumors in mammary glands: the weak expression of the transgene in mammary glands and the apparent cell type–specific effect of p75 CDP/Cux. The genetic background of transgenic mice, and the fact that almost all (56 of 60) transgenic mice were virgin, contributed to the weak expression in mammary glands. As MMTV-directed transgene experiments have been done in the FVB strain, we expect that in future backcrosses, transgene expression in mammary glands will augment in parallel with the FVB genetic component and will further increase in multiparous females (32, 33). This said, MMTV-dependent transgene expression alone cannot explain the observed phenotypes. We consider that the heightened susceptibility to a myeloproliferative disease in p75-transgenic mice likely reveals a particular tropism of the p75 CDP/Cux isoform towards certain myeloid precursor cells. This notion is reinforced by the contrasting small incidence of myeloproliferative diseases that developed in MMTV-p110 CDP/Cux mice (Table 1). We note that the p75 CDP/Cux transgene was expressed weakly, if at all, in the spleen of healthy transgenic mice and that disease was associated with increased expression of the transgene. This indicates that increased p75 expression was selected for, and probably played a causative role, in the neoplastic process. The increase in p75 expression observed in affected transgenic mice also suggests that another event was required for the activation of transgene expression. It will be important to determine the cell type in which this event takes place. The failure to transplant the disease suggests that the event leading to the activation of p75 transgene expression did not occur in a hematopoietic stem cell but, most likely, in a committed myeloid progenitor. This could be the reason for the long latency period and would further suggest that the ability to transplant and a shorter latency could be obtained by combining p75 CDP/Cux with a promoter that enables expression in the hematopoietic stem cell or an early myeloid progenitor. Indeed, several studies have shown that the phenotype of murine transgenic models of human leukemia is critically dependent on the cellular compartment that is targeted (reviewed in refs. 46–48). For example, the type of myeloproliferative diseases that were generated depended on whether the inactivation of JunB or transduction of BCR-ABL occurred in hematopoietic stem cells or in granulocyte-macrophage progenitor cells (41, 48). Most of the hematopoietic disorders that developed in the cohort of MMTV-p75 transgenic mice involved myeloid cells; however, in a few mice, the lymphoid compartment was affected. It is not clear at this point whether this spectrum of diseases reflects the tropism of the p75 protein or the cell type specificity of transgene expression. Future studies should investigate the phenotype resulting from the expression of CDP/Cux isoforms in early hematopoietic precursor or stem cells.

Grant support: Canadian Breast Cancer Research Alliance grant no. 014348 (A. Nepveu) and the Regulatory Genetics Program of Genome Quebec/Genome Canada (A.C. Peterson).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

A. Nepveu and S. Fournier are the recipients of scholarships from the Fonds de la Recherche en Santé du Québec. C. Cadieux is the recipient of studentships from the Royal Victoria Hospital Research Institute (2005) and from the Department of Defense Breast Cancer Research Program (2006).

We acknowledge the expertise of Ms. Jo-Ann Bader for the histological work.