SEARCH

SEARCH BY CITATION

Keywords:

  • Acute lymphocytic leukemia;
  • B lymphocytes;
  • Cancer stem cells;
  • Self-renewal

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES
  11. Supporting Information

The initial steps involved in the pathogenesis of acute leukemia are poorly understood. The TEL-AML1 fusion gene usually arises before birth, producing a persistent and covert preleukemic clone that may convert to precursor B cell leukemia following the accumulation of secondary genetic “hits.” Here, we show that TEL-AML1 can induce persistent self-renewing pro-B cells in mice. TEL-AML1+ cells nevertheless differentiate terminally in the long term, providing a “window” period that may allow secondary genetic hits to accumulate and lead to leukemia. TEL-AML1-mediated self-renewal is associated with a transcriptional program shared with embryonic stem cells (ESCs), within which Mybl2, Tgif2, Pim2, and Hmgb3 are critical and sufficient components to establish self-renewing pro-B cells. We further show that TEL-AML1 increases the number of leukemia-initiating cells that are generated in collaboration with additional genetic hits, thus providing an overall basis for the development of novel therapeutic and preventive measures targeting the TEL-AML1-associated transcriptional program. STEM CELLS2013;31:236–247


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES
  11. Supporting Information

Evidence is accumulating that shows the presence of multistep molecular mechanisms for leukemogenesis. Although the cascade of genetic “hits” and their contribution to leukemogenesis remains poorly understood, cells with the initial hit that initiates the process are expected to provide a reservoir for subsequent oncogenic hits to accumulate during the covert stage of acute leukemia.

Hematopoietic stem cells (HSCs) are unique among blood-forming cells in their ability to self-renew, thus providing a reservoir for multistep molecular events to accumulate and culminate in the development of leukemia. However, the involvement of HSCs in clinically frank acute leukemia remains a matter of debate [1–3]. In line with this, experimental expression of at least some potent oncoproteins associated with leukemia such as MLL-ENL, MLL-AF9, and MOZ-TIF2 fusions confers progenitor cells downstream of HSCs with sustained self-renewal [4–6]. However, these potent oncoproteins rapidly induce leukemia in mice models, and therefore do not provide many insights into how the clinically covert stage preceding frank leukemia (“pre”-leukemic stage) is initiated and maintained as a reservoir.

TEL (ETV6)-AML1 (RUNX1) is the most prevalent fusion gene associated with pediatric acute lymphoblastic leukemia (ALL) and is exclusively associated with precursor B cell leukemia. This chimeric gene arises predominantly during fetal hemopoiesis and TEL-AML1-expressing fetal clones can expand and persist in a clinically covert pre-leukemic state in the long term [7]. Although TEL-AML1 fusion is supposed to impede AML1 functions through the recruitment of corepressor molecules, examination of the manner by which transcriptional deregulation might lead to a sustained preleukemic clone is difficult with clinical samples, as an analysis of TEL-AML1 function will be confounded by secondary genetic hits the cells acquired to become frankly leukemic.

Several animal models have been developed in an effort to delineate the manner by which TEL-AML1 initiates and maintains preleukemic clones. However, the findings of these investigations have often been inconsistent, probably reflecting different aspects of TEL-AML1 functions. Given that analysis of the rearrangement of immunoglobulin genes in clinical leukemia samples suggests that TEL-AML1 translocation occurs somewhere in or between stem and early stages of B cells [8–11], TEL-AML1 has been accordingly programmed for expression in such cells. Investigations using a knock-in mouse model where TEL-AML1 is expressed lymphoid-specifically in common lymphoid progenitors and its downstream cells reveal no apparent changes in either differentiation or growth [12]. In another knock-in mouse model where TEL-AML1 is expressed in the early embryo, fetal liver-derived cells exhibit increased self-renewal activity assayed in vitro, which is lost in newborns [12]. In a third knock-in mouse model where TEL-AML1 is expressed in and after the HSC stage, the stem cell fraction expands and myeloid cells emerge in abundance, but B cells are virtually undetectable [12]. A similar observation to that seen in the third knock-in mouse model was made in mouse bone marrow (BM) transplantation models using adult hematopoietic cells as a target of TEL-AML1 expression, although in this case a small number of B cells emerge and the B cell differentiation program was inhibited in the transition of the pro-B to pre-B cell stages [13, 14]. Transgenic zebrafish models where TEL-AML1 was expressed under control of ubiquitously active promoters exhibit fatal hyperplasia of lymphoid blasts (6% occurrence), however, lineage of the blasts is unclear, and whether the hyperplasia is cell autonomous is not tested [15]. A mouse BM transplantation model where TEL-AML1 was retrovirally expressed in fetal hematopoietic cells exhibited increased repopulation of B cells. However, the differentiation of TEL-AML1+ B cells remained unperturbed (therefore, with no implication of pro-B cell expansion), and an increase in repopulation of myeloid cells resulted [16], a phenomenon that has not been detected in humans [17].

These animal models provide implications for the functional roles of TEL-AML1 in altered hematopoiesis and leukemia development, and TEL-AML1+ cells in some of these animal models can develop leukemia in collaboration with additional genetic hits. However, the lineages of leukemia are not restricted to B lymphoid [18, 19] or mainly T lymphoid [12]. By contrast, TEL-AML1 in humans is exclusively associated with leukemia of the precursor B cell type [7]. In addition, examination of human cord blood with the TEL-AML1 translocation through FISH analysis revealed that the fusion gene was detected in a considerable proportion of the B, but not myeloid or T, cell fraction [17], indicating the ability of TEL-AML1 to expand B cells in a covert stage. Examination of immunoglobulin rearrangement of a nonleukemic clone harboring the TEL-AML1 fusion contained in a healthy twin of a leukemia patient revealed genetic evidence of pro-B cells, and the nonleukemic clone is identical, on a clonal basis, to the leukemic clone in the patient [9], thus likely representing the preleukemic stage. These findings in humans raise the possibility that TEL-AML1 can induce pro-B cells that persist and self-renew in vivo. However, such a possibility has not been explored in animal models.

Here, we therefore investigated the effects of TEL-AML1 on self-renewal activity of pro-B cells in vitro and in vivo in mice. We subsequently investigated the manner by which, if at all, TEL-AML1 collaborates with an additional hit for leukemogenesis. Finally, we focused on the genetic program influenced by TEL-AML1 to exert its functions.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES
  11. Supporting Information

Isolation of Cells and Cell Culture

B220+c-kit+ pro-B cells were Fluorescence-activated cell sorting (FACS)-purified from fetal liver or BM of 10-week-old BALB/c mice, or induced from fetal liver cells that were cultured on OP9 cells in Iscove's modified Dulbecco's medium (Invitrogen, Carlsbad, CA, http://www.invitrogen.com) supplemented with 15% fetal calf serum, stem cell factor (SCF), flt3-ligand (FL), interleukin-7 (IL7), and 2-mercaptoethanol. TEL-AML1 cell lines, Reh and KOPN41 [20] were obtained by the American Type Culture Collection, and generously provided by Dr. Kanji Sugita (Yamanashi University, Japan).

Retroviral Infection and Transplantation of Cells

cDNA for TEL-AML1 tagged with the myc-epitope at the carboxyl terminus [13] was (a) cloned upstream of the internal ribosome entry site (ires) element of the murine stem cell virus (MSCV)-ires- green fluorescent protein (GFP) vector, or (b) cloned downstream of the GFP-ires element of the MSCV-GFP-ires vector. Other retrovirus plasmids used, and target sequences for shRNAs with references are described in Supporting Information Methods. Frequency of leukemia-initiating cells was calculated using L-Calc software (STEMCELL Technologies, Vancouver, Canada, http://www.stemcell.com). All animal experiments were performed in accordance with the protocols approved by the Institutional Animal Care and Use Committee at the Aichi Cancer Center.

Flow Cytometry

Anti-B220 (RA3-6B2), anti-CD19 (1D3), anti-CD3 (145-2C11), anti-TER-119 (TER-119), anti-Gr-1(RB6-8C5), anti-c-kit (2B8), anti-CD43 (S7), anti-CD25 (PC61.5), and anti-IgM (II/41) antibodies were used.

Western Blotting

Anti-myc tag (9E10; Sigma, St. Louis, MI, www.sigma-aldrich.com), anti-tubulin (DM1A; Sigma), anti-actin (AC40; Sigma), and anti-RUNX1 (Ab-2, Lot D11394-3; Oncogene) antibodies were used, in combination with ECL (GE Healthcare, Uppsala, Sweden, www.gelifesciences.com) or SuperSignal West Femto (Thermo, Waltham, MA, http://www.thermofisher.com).

In Vitro Colony-Forming Assays

Cells were plated in triplicate in methylcellulose under multimyeloid (SCF, IL-3, IL-6) or B cell (SCF, FL, IL-7) conditions. Colonies were counted and replated every 7 days.

Microarray and Bioinformatics Analyses

Agilent Mouse Whole Genome 4 × 44K microarrays, and Agilent Feature Extraction were used. Microarray data are available in the ArrayExpress database under accession no. E-MEXP-3297 and were processed as described elsewhere [21]. Gene set enrichment analyses (GSEAs) were performed with the use of GSEA ver. 2.0 software (http://www.broad.mit.edu/gsea) with the S2N metric for gene ranking and 1,000 data permutations.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES
  11. Supporting Information

Expression Level-Dependent Impact of Leukemia Associated-TEL-AML1 Protein on Differentiation and Growth of B Cells

Functions of a given transcription factor are known to be expression level-dependent [22–24]. To determine whether TEL-AML1 protein expression can induce self-renewing pro-B cells, we therefore used two types of retroviral vectors to achieve different levels of protein expression and compared their effects on pro-B cells; one expressing TEL-AML1 cloned upstream of an ires-GFP element (designated TEL-AML1“high”), and the other expressing TEL-AML1 cloned downstream from an ires element (designated TEL-AML1“low”) (Fig. 1A). Protein expression level obtained downstream of the ires element is known to be lower than that obtained upstream of the ires element [25, 26]. The expression level of TEL-AML1 protein obtained with the TEL-AML1“low” virus in fetal pro-B cells was ∼ 1/6 of that obtained with the TEL-AML1“high” virus and was comparable to those of human TEL-AML1 cell lines Reh and KOPN41 (Fig. 1B and Supporting Information Fig. 1). It is known that although TEL-AML1 transcript levels could vary from patient to patient, they are not dissimilar to that of Reh [27].

thumbnail image

Figure 1. Analyses of colony-forming activities of TEL-AML1-expressing cells. (A): Schematic drawings of retrovirus vectors (not to scale). (B): Western blot analysis of cell lysates using anti-myc and anti-tubulin antibodies (left), and anti-RUNX1 and anti-actin antibodies (right). Three times as many TEL-AML1“low”-infected cells compared to TEL-AML1“high”-infected cells were used (left). Expression levels of TEL-AML1 in TEL-AML1“low”-infected pro-B cells, and human TEL-AML1 cell lines Reh and KOPN41. GFP-only virus-infected pro-B cells served as negative controls (right). (C): Lineage-negative cells were isolated from e14 fetal liver, infected with the indicated retrovirus and cultured on OP9 cells without sorting. Cells were analyzed for expression of GFP and B220 on the indicated days after the beginning of the culture. (D): Colonies formed per 104 c-kit+ e12 fetal liver cells infected with the indicated retrovirus, sorted for GFP+ and plated in semisolid medium in the presence of stem cell factor (SCF), flt3-ligand, and interleukin-7 (IL7) (B cell condition). Colonies were counted and replated every 7 days (upper panel). Photomicrographs show the relative numbers of colonies formed (middle). Expressions for GFP and B220 were analyzed upon replatings (lower panel). (E): Colonies formed per 104 fetal liver-derived pro-B cells infected with the indicated retrovirus, sorted for GFP+, and assayed under the B cell condition. Colonies were counted and replated every 7 days (upper left). Photomicrographs show relative numbers and sizes of the colonies generated (upper right). Cells recovered from the second and third colonies were analyzed for expressions of GFP, B220, CD19, c-kit, and CD43 (lower panel). Photomicrpgraphs of colonies in (E) were taken with original magnification ×40 (UPlan FLN4x/0.13 PhP objective). (F): Colonies formed per 104 adult bone marrow-derived pro-B cells infected with the indicated retrovirus, sorted for GFP+, and assayed under the B cell condition. Colonies formed in the B cell condition were counted and replated every 7 days. (G): Colonies formed per 104 e14 lineage-negative fetal liver cells infected with the indicated retrovirus, sorted for GFP+ and plated in semisolid medium in the presence of SCF, IL3 and IL6 (multimyeloid condition). Colonies were counted and replated every 7 days. Typical data of at least three experiments are shown (C–G). Colony-forming assays were conducted in triplicate (D–G). Abbreviations: GFP, green fluorescent protein; IRES, internal ribosome entry site; NC, negative control.

Download figure to PowerPoint

We first chose immature hematopoietic cells as a target of TEL-AML1 expression. E14 fetal liver-derived lineage-negative cells infected with TEL-AML1“high,” TEL-AML1“low,” or GFP-only control virus were cultured on OP9 in the presence of SCF, FL, and IL7 (B cell condition), without cell sorting (Fig. 1C). B220+CD19+ B cells were easily identified in both the GFP-positive (+) and negative (−) fraction in the culture of GFP-only-infected cells. In contrast, B cells were undetectable from the GFP+ fraction in the culture of TEL-AML1“high”-infected cells, even after 3 weeks of culture, but readily detected in the GFP− (therefore, TEL-AML1-negative) fraction, suggesting that a high level of TEL-AML1 expression inhibits immature hematopoietic cells from giving rise to B cells. In sharp contrast, the low level of TEL-AML1 expression allowed immature cells to give rise to B cells. Significantly, TEL-AML1“low” cells appeared to dominate the culture after, but not before, B220+ B cells emerged (Fig. 1C). This finding was further confirmed by a colony-formation assay using c-kit+ cells from e12 fetal liver: B cells are normally undetectable in e12 fetal liver (Fig. 1D). In the first-round plating where B220+ B cells represented a minor population in the culture, the colony-forming abilities were comparable among TEL-AML1“low,” TEL-AML1“high,” and GFP-only-infected cells. Upon replatings, where B cells normally became the major population, TEL-AML1“low” cells yielded more colonies than the GFP-only control. In contrast, TEL-AML1“high” cells did not appreciably yield colonies, and cells grown were non-B cells (Fig. 1D). These overall findings suggest that low-level expression of TEL-AML1 confers upon cells an enhanced replating efficiency after, but not before, B cells emerge. Additionally, high-level expression of TEL-AML1 inhibits the emergence of B cells.

We next directly transduced pro-B cells isolated from fetal liver with TEL-AML1 virus. CD19+c-kit+CD43+ pro-B cells were induced from e14 lineage-negative cells, infected with TEL-AML1“low,” TEL-AML1“high,” or GFP-only control virus, FACS-sorted for GFP+, and assayed for colony-forming ability in the B cell condition (Fig. 1E). Whereas GFP-only control cells showed a gradual decline in colony formation as replatings proceeded, TEL-AML1“low” cells retained their colony-forming ability. The TEL-AML1“low” cells in this culture remained largely B220+CD19+c-kit+CD43+ pro-B cells. In contrast, TEL-AML1“high” cells showed no appreciable replating activity. Experiments using pro-B cells directly isolated from e18 fetal liver yielded similar results (Supporting Information Fig. 2). Interestingly, the effect of TEL-AML1 expression on self-renewal of pro-B cells appears to be specific to cells of fetal origin, as pro-B cells isolated from adult BM did not show increased replating abilities upon infection with TEL-AML1“low” virus (Fig. 1F). Additionally, low-level TEL-AML1 expression had no appreciable effects on replating ability of e14 fetal liver-derived lineage-negative cells assayed under myeloid-supporting condition, although interestingly high-level TEL-AML1 expression resulted in increased replating abilities of cells (Fig. 1G).

thumbnail image

Figure 2. Competitive transplantation of pro-B cells into mice. (A): Equal numbers of GFP- and KO-, or TEL-AML1- and KO-expressing pro-B cells were mixed and transplanted into the bone marrow (BM) cavity of NOD-SCID mice. One month later, BM cells were analyzed for expression of GFP and KO. Fluorescence-activated cell sorting analyses for expression of GFP and KO in pro-B cells before transplantation (left), and GFP and KO of the B220+ fraction in BM after transplantation (right). Typical data from three experiments using eight recipients each for GFP/KO and TEL-AML1/KO are presented. (B): A graph presenting ratios of GFP to KO in BM. (C): Analysis of expression of GFP and KO in the B220+ fraction of BM following serial transplantation. Experiments are conducted as in (A) and B220+ B cells were isolated from the BM of recipient mice and serially transplanted into mice. Typical data from three experiments using a total of 40 mice are presented. Abbreviations: GFP, green fluorescent protein; KO, Kusabira Orange.

Download figure to PowerPoint

To conclude, TEL-AML1 protein has different effects on cells depending on the expression level of the protein and the cell type; a “low” level of expression of TEL-AML1 exhibits no obvious effects on the growth of hematopoietic cells in the myeloid-supporting condition but produced profound effects in promoting growth and self-renewal of pro-B cells. We detected no apparent difference in the effects on pro-B cells, whether they were descended from TEL-AML1-infected immature hematopoietic cells or pro-B cells directly infected with TEL-AML1 virus (Fig. 1D, 1E). In contrast, a “high” level of expression of TEL-AML1 resulted in increased self-renewal activities in hematopoietic cells cultured in the myeloid-supporting condition but negatively impacted growth and self-renewal of pro-B cells. Our findings suggest that protein expression level is important for TEL-AML1 to induce self-renewing pro-B cells.

TEL-AML1 Expression Confers Fetal Pro-B Cells with Enhanced Self-Renewal In Vivo

We next explored the effects of TEL-AML1 expression in vivo. Fetal pro-B cells were infected with TEL-AML1“low,” GFP-only control, or Kusabira Orange (KO)-only control virus, sorted for GFP or KO, and then mixed so that TEL-AML1-expressing (therefore, GFP+) cells and KO-expressing cells were equal in number. Similarly, GFP-only-expressing and KO-expressing cells were mixed in a 1:1 ratio (Fig. 2A). Mixed cells were then transplanted into the BM cavity of NOD-SCID mice; fetal liver-derived pro-B cells are known to easily repopulate to immune-deficient mice [28]. One month later, mice were analyzed for the expression of GFP and KO in the B cell compartment in BM (Fig. 2A). Although the GFP/KO ratio was ∼ 1.0 in mice that received a transplant with the mixture of GFP-control and KO-expressing cells, the ratio was ∼ 3-6 in mice that received a transplant with the mixture of TEL-AML1-expressing and KO-expressing cells (Fig. 2B). Therefore, TEL-AML1 conferred a competitive ability on pro-B cells over the control in vivo.

Importantly, TEL-AML1+ B cells were serially transplantable (Fig. 2C). B cells were purified from the BM of primary recipient mice and secondarily transplanted into new mice. In secondary recipient of B cells purified from the primary recipient of a mixture of GFP-only and KO-only cells, both GFP+ and KO+ cells were barely detectable. In contrast, while KO+ cells were barely detectable, GFP+ (therefore, TEL-AML1+) cells were easily detectable in secondary recipient of B cells purified from the primary recipient of a mixture of TEL-AML1 (GFP) and KO-only cells. In a similar way, TEL-AML1+ cells were found to be thirdly transplantable (Fig. 2C), suggestive of enhanced self-renewal activity.

Flow-cytometric analysis showed the inhibitory effects of TEL-AML1 on B cell differentiation (Fig. 3). Given that the transplanted cells were pro-B cells, they were positive for CD43 and c-kit, but negative for CD25 and IgM before transplantation, in both vector-only control- and TEL-AML1-transduced cells (Fig. 3A). The vector-only infected cells partly became CD43-negative and IgM-positive in BM, and largely did so in spleen after the transplantation (Fig. 3B). In contrast, TEL-AML1+ cells remained largely CD43-positive both in BM and spleen; the negativity for CD25 was lost to a variable degree from mouse to mouse, but only a small fraction of the cells became IgM+ (Fig. 3C). In secondary recipient mice, GFP-only+ cells were virtually undetectable in BM, but a small number of GFP+ cells detected in spleen were largely IgM+ (Fig. 3B). In contrast, TEL-AML1+ cells still largely remained CD43+ with a small fraction being IgM+ (Fig. 3C). These findings suggest that while control cells differentiated from pro-B to IgM+ matured B cells, TEL-AML1+ cells were inhibited, albeit incompletely, in the differentiation program. Consistent with the incompleteness of the inhibition in B cell differentiation, TEL-AML1+ pro-B cells eventually differentiated to IgM+ matured B cells in the long term (5 months after the third transplantation) (Fig. 3D). TEL-AML1+ pro-B cells are thus able to persist in vivo, but eventually differentiate terminally in our mouse model. This incompleteness may at least partly account for the inability of TEL-AML1 to induce leukemia in our mouse model.

thumbnail image

Figure 3. Analysis of B cell differentiation of pro-B cells following transplantation into mice. (A): Cells were infected with virus for GFP-only and TEL-AML1, sorted for GFP+ and analyzed for the indicated cell surface molecules just before transplantation. (B, C): Analysis of differentiation of control GFP+ (B) and TEL-AML1+ (C) cells in primary and secondary recipient mice. A low proportion of control GFP+ cells in the BM of secondary recipients precluded the analysis. GFP+ cells were analyzed for expression of the indicated molecules associated with B cell differentiation. Typical data from three experiments using a total of 40 mice are presented. (D): Analysis of B cell differentiation in tertiary recipient mice. TEL-AML1+ cells were analyzed for the indicated molecules for differentiation, 5 months following the tertiary transplantation. Typical data of three experiments using three mice each are presented. Abbreviations: BM, bone marrow; GFP, green fluorescent protein; Spl, spleen.

Download figure to PowerPoint

The enhanced self-renewal and inhibited differentiation was additionally confirmed with the use of a lentivirus vector driving TEL-AML1 expression under the control of the CD19 promoter (CD19-TEL-AML1-iresGFP) (Fig. 4). In mice that received a transplant with lineage-negative fetal liver cells infected with CD19-TEL-AML1-iresGFP virus (Fig. 4A), GFP expression was restricted to the B cell compartment as expected (Fig. 4B). The differentiation of CD43+ pro-B cells to CD43- pre-B cells was not apparently perturbed in the GFP- fraction where CD43+ cells accounted for ∼ 15% of the B cell fraction of BM, which is comparable to the percentage of CD43 in normal BM cells. In contrast, the differentiation of pro-B cells to pre-B cells was inhibited in the GFP+ fraction; CD43+ cells accounted for ∼ 60% of the B cell fraction of BM (Fig. 4B). On plating in vitro, TEL-AML1+ pro-B cells isolated from the mice yielded more colonies than their GFP- counterparts, and only TEL-AML1+ cells exhibited replating abilities.

thumbnail image

Figure 4. Analysis of differentiation and colony-forming activity of TEL-AML1-expressing pro-B cells generated in vivo. (A): A schematic drawing of the lentivirus vector CS-EμMarCD19-TEL-AML1myc-ires GFP-WPRE (not to scale). (B): Fluorescence-activated cell sorting analysis of bone marrow cells of mice that received a transplant with e14 fetal liver cells infected with the lentivirus. GFP+ and GFP− cells (upper panel) were analyzed for B220+CD43+ pro-B cells. Percentages of CD43+ cells among B220+ B cells are also presented (lower panel). (C): In vitro colony-forming activity of sorted pro-B cells in coculture with OP9 under the B cell condition. Colonies were counted and replated every 7 days. Typical data from three experiments conducted in triplicate using three mice per experiment are presented. Abbreviations: GFP, green fluorescent protein; WPRE, woodchuck hepatitis virus posttranscriptional regulatory element.

Download figure to PowerPoint

TEL-AML1 Plays Crucial Roles for the Development of Leukemia in Collaboration with Additional Genetic Hits

The ability of TEL-AML1 to enhance self-renewal could cooperate with additional genetic hits for the development of leukemia. To test this possibility, we manipulated TEL-AML1+ fetal pro-B cells to coexpress mutated RAS (N-RASG12D), transplanted the cells into mice, and then monitored the incidence of leukemia. Mutated N-RAS or K-RAS is found in some 10% of TEL-AML1 leukemia clinical samples [29], and enforced expression of a mutant N-RAS (N-RASG12D) in mouse pro-B cells is capable of eliciting leukemia on its own [30]. Here, we used N-RASG12D retrovirus that coexpresses the extracellular domain of human CD8 (hCD8) as a surrogate marker. Although % GFP and % hCD8 were comparable between GFP/N-RASG12D(hCD8)- and TEL-AML1(GFP)/N-RASG12D(hCD8)-infected cells before transplantation (Fig. 5A), mice that received a transplant with TEL-AML1/N-RASG12D cells succumbed to leukemia significantly earlier than their GFP/N-RASG12D counterparts (Fig. 5B).

thumbnail image

Figure 5. Roles of TEL-AML1 in leukemia development in the context of an additional oncogene being coexpressed. Fetal pro-B cells were infected with GFP- or TEL-AML1-virus in combination with N-RASG12D that coexpress an extracellular domain of human CD8 as a surrogate marker, and transplanted into irradiated syngeneic hosts. (A): Fluorescence-activated cell sorting profile of pro-B cells after the infection. (B): Kaplan-Meier survival curves of mice that received a transplant with cells indicated in (A). The difference was statistically significant (p = .014), as assessed by the log-rank test. (C): Cells as shown in (A) were sorted for GFP+ hCD8+ double-positivity, and transplanted into mice in a limiting-dilution fashion. Mice were then monitored for survival for one year. Frequencies of leukemia-initiating cells are also presented. Two experiments yielded a similar result. (D): Replating potentials of pro-B cells coexpressing TEL-AML1 and N-RASG12D. Pro-B cells coexpressing GFP and N-RASG12D, or TEL-AML1 and N-RASG12D were prepared and assayed for replating potential in colony-formation in vitro under the B cell condition. Photomicrographs of colonies in the fifth replating are also presented: original magnification ×40 (UPlan FLN4x/0.13 PhP objective). Typical data obtained from four experiments conducted in triplicate are presented. Abbreviations: CI, confidence interval; GFP, green fluorescent protein.

Download figure to PowerPoint

Next, fetal pro-B cells doubly infected with TEL-AML1/N-RASG12D or GFP-only control/N-RASG12D were flow-sorted and transplanted in a limiting-dilution fashion. Results (Fig. 5C) showed that TEL-AML1+N-RASG12D+ cells caused leukemic death at lower cell dosages than GFP-only+N-RASG12D+ cells. The frequency of leukemia-initiating cells was estimated to be 1/5,398 (95% CI = 1/2,153-1/13,533) for TEL-AML1+N-RASG12D+ cells, whereas it was only 1/98,704 (95% CI = 1/42,283-1/218,700) for GFP-only+N-RASG12D+ cells. The difference in frequency is unlikely to be accounted for by the difference in differentiation (Supporting Information Fig. 3).

To corroborate the increase in the frequency of leukemia-initiating cells conferred by TEL-AML1 expression, a serial replating assay of fetal pro-B cells conducted in vitro under the B cell condition showed that TEL-AML1 expression enhanced self-renewal activity even in the presence of coexpressed N-RASG12D (Fig. 5D). Although the possible occurrence of additional mutations in cells to elicit leukemia is not excluded, overall, these findings suggest that TEL-AML1 and mutant N-RAS cooperatively induce leukemia.

TEL-AML1 Uses ESC-like Gene Modules for Persistence of Pro-B Cells

We next searched for genes linking TEL-AML1 expression to the enhanced self-renewal ability of pro-B cells (Fig. 6). GSEA with molecular signature database c2.v.2 (http://www.broadinstitute.org/gsea/index.jsp) comparing TEL-AML1 “low”- and GFP-only-infected fetal pro-B cells revealed enrichment of ESC-like program genes (STEMCELL_EMBRYONIC_UP) in TEL-AML1+ cells (Fig. 6A). Expression of genes commonly associated with embryonic, neural, and hematopoietic stem cells (STEMCELL_ COMMON_UP) was also enriched. Subsequent application of additional gene sets with an ESC-like program [31–34] confirmed this association. In contrast, the reverse association was found with an adult stem cell core program [34] and a gene program associated with differentiated cells compared with stem cells. Enrichment for an ESC-like program was not driven by proliferation genes, since the significance of the correlation was maintained when genes with cell-cycle functional annotations were excluded from the analysis [34]. Given the fact that both TEL-AML1+ and GFP-only control cells were pro-B cells (Fig. 3A) when subjected to the analysis, the difference in gene expression is not likely to represent a difference of the cell population. Reverse enrichment of adult tissue stem-cell modules in TEL-AML1+ pro-B cells suggests that the mechanisms TEL-AML1 uses to convert pro-B cells to preleukemia cells are distinct from those involved in the self-renewal of adult stem cells.

thumbnail image

Figure 6. Analysis of genes associated with TEL-AML1-mediated self-renewal activity. (A): Gene set enrichment analysis (GSEA) of genes influenced by TEL-AML1 expression in pro-B cells. Fetal pro-B cells were infected with virus for GFP-only control or TEL-AML1, sorted for GFP, and subjected to microarray gene expression analysis. Data obtained from four experiments were analyzed by GSEA. Summary of the analysis and enrichment plots for the indicated gene sets are presented. 1) Results obtained using molecular signature database c2.v2. as provided by the Broad Institute (http://www.broadinstitute.org/gsea/index.jsp). 2) Results obtained using gene sets associated with ESC, as described by Somervaille et al. [34]. They are ADULT_TISSUE_ STEM_CELL_MODULE, CORE_ESC_LIKE_GENE_MODULE, CORE_ESC_LIKE_GENE_ WITHOUT_CELL_CYCLE, CELL_CYCLE, ES_EXPRESSED_1, ES_EXPRESSED_2, NOS_TFS (transcription factors regulated by Nanog, Pou5f1, and Sox2.), GOOD_PROGNOSIS_AML, and POOR_PROGNOSIS_AML. (B): Heat-map presentation of genes listed as the top 30 in the leading edge of the indicated gene sets. Genes marked with asterisks were analyzed for their functionality in (C–E). (C): Effects of shRNA-mediated silencing of genes on TEL-AML1-mediated self-renewal. Pro-B cells were infected with virus for TEL-AML1 and shRNA for the indicated gene, and doubly infected cells were assayed for replating potential in colony-formation under the B cell condition. Mean and S.D. of assays conducted in triplicate are presented. Typical results of four independent experiments are presented. (D): Effects of expression of the indicated genes on self-renewal assayed in vitro. Pro-B cells were infected with a mixture of viruses for the indicated genes (denoted as “+”) and GFP-only control virus (denoted as “−”), and assayed for replating potential in colony-formation under the B cell condition. Mean and S.D. of assays conducted in triplicate are presented. Typical results of four experiments are presented. (E): Effects of expression of Mybl2, Tgif2, Pim2, Hmgb3 and Sall4 on self-renewal of pro-B cells as assayed in vivo. Pro-B cells infected with a mixture of viruses for Mybl2, Tgif2, Pim2 and Hmgb3 that coexpress GFP were mixed with an equal number of KO-infected cells (left), and transplanted into NOD-SCID mice. One month later, B220+ cells in BM and spleen were analyzed for expression of GFP and KO (middle). B220+ cells isolated from BM were secondarily transplanted into new mice, and analyzed for expression of GFP and KO in secondary recipient mice one month later (right). Pro-B cells infected with a mixture of viruses for Mybl2, Tgif2, Pim2, Hmgb3 and Sall-4 were similarly tested. Expression of CD43, a marker of pro-B cells, on GFP+ B cells in the secondary recipient mice, is also presented. Typical data of three experiments using a total of nine mice each are presented. Abbreviations: BM, bone marrow; ES, enrichment score; FDR q-val, false discovery rate q-value; GFP, green fluorescent protein; KO, Kusabira Orange; NES, net enrichment score; NOM p-val, nominal p-value.

Download figure to PowerPoint

In an effort to determine genes responsible for the TEL-AML1 activity, among those found in gene sets associated with the ESC-like program, we focused on genes for regulators of gene expression and kinases. This strategy is similar to that taken by Somervaille et al. when elucidating essential genes for MLL fusion-mediated leukemia stem cell activity [34]. We thus chose Mybl2, TCF7L1(TCF3), Smad1, Pim2, Tgif2, Sall4, Ryk, Ctbp2, Cbx5, and Hmgb3 because these genes are recurrently included in top 30 of the leading edges of gene sets for the ESC-like program that are associated with TEL-AML1 expression (Fig. 6B), and ESC-associated genes with transcription or chromatin annotation [34] (Supporting Information Fig. 4). We then used shRNA retrovirus and examined their effects on TEL-AML1-mediated self-renewal activities (Fig. 6C). shRNA against luciferase had no apparent effects on the replating ability of TEL-AML1+ pro-B cells in vitro, and therefore served as a control. shRNA-mediated silencing of one of the genes among Mybl2, Cbx5, TCF7L1, Smad1, Pim2, Hmgb3, Tgif2, Sall4, and Ctbp2 consistently, although to a variable degree, inhibited the replating; numbers of colonies formed in the third plating were less than 1/4 of those in the first plating. Although two shRNAs for Ryk did not show consistent results, they did not show such inhibitory effects (Fig. 6C).

We therefore next conducted reciprocal experiments, using a library consisting of retroviruses to express Mybl2, Cbx5, TCF7L1, Smad1 (constitutively active), Pim2, Hmgb3, Tgif2, Sall4, and Ctbp2. Fetal pro-B cells infected with the retrovirus library (a mixture of retroviruses individually produced) yielded colonies at the third replating, whereas those infected with GFP-only control virus did not show such a response in three independent experiments. Polymerase chain reaction analysis of genomic DNA extracted from cells remaining at the third replating consistently detected Mybl2, Pim2, Hmgb3, Tgif2, and Sall4, but not Smad1, TCF7L1, or Ctbp2. Inclusion of Smad4, a binding partner of Smad1, and constitutively active β-catenin, a mediator of Wnt signaling, in the library did not change the results. We then focused on the five genes (Mybl2, Pim2, Hmgb3, Tgif2, and Sall4) for further experiments (Fig. 6D). Cells infected with either one of the viruses exhibited no enhanced replating ability compared with GFP-alone infected cells; infection efficiencies of these viruses were 30%-40% when individually used. Then, retroviruses for the five genes were mixed so that cells were infected with viruses of every possible combination containing at least two viruses. Although combinations of any two or three genes did not consistently produce replatable cells, a mixture of five viruses reproducibly produced replatable cells. Among mixtures of four viruses, a mixture of viruses for Mybl2, Pim2, Hmgb3, and Tgif2 appeared sufficient, although addition of Sall4 virus (i.e., a total of five viruses) consistently yielded more colonies (Fig. 6D); here, we chose combinations of genes which generated more than twice as many colonies in the third plating as in the first plating.

We finally tested the transplantability of pro-B cells infected with viruses for the four genes. While pro-B cells infected with GFP-control virus were not transplantable in secondary recipients (Fig. 2C), those infected with a mixture of viruses for the four genes were serially transplantable, although addition of Sall4 again appeared to enhance the ability (Fig. 6E). All these findings suggest that Mybl2, Pim2, Hmgb3, and Tgif2 are required for TEL-AML1-mediated self-renewal and are themselves sufficient to cooperatively induce self-renewal in pro-B cells.

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES
  11. Supporting Information

Although several lines of evidence indicate that the TEL-AML1 fusion gene is usually generated by a prenatal initiating mutation in the common B cell precursor ALL, the manner by which the TEL-AML1 fusion protein establishes a persistent preleukemic clone remains unclear [7]. Although attempts have been made to establish a preleukemic stage, findings using animal models have often been inconsistent. Studies using human clinical leukemia samples suggest that the TEL-AML1 fusion arises somewhere in or between the stem and early stages of B cells. However, definitive information with respect to the cell of origin for the TEL-AML1 fusion is lacking. Nevertheless, the finding that a persistent B cell clone harboring the TEL-AML1 fusion identified in a healthy twin of a leukemia patient shows genetic evidence of pro-B cells raises the possibility that TEL-AML1 confers self-renewal activity upon pro-B cells. Such a possibility has not been investigated in animal models. In the study presented here, we have shown for the first time that TEL-AML1 can inhibit B cell differentiation at the pro-B stage, albeit incompletely, and at the same time enhance self-renewal of pro-B cells in vivo.

The results we obtained when comparing the effect of different levels of TEL-AML1 expression on murine cells revealed an expression level- and cell lineage-dependent impact of TEL-AML1 expression on self-renewal. Although our modeling uses normal cells as a target of expression of the fusion gene, as opposed to human TEL-AML1+ cells harboring only one allele for both TEL and AML1, a similar strategy to ours is widely used and has provided much information with respect to the function of a given fusion gene [35]. In addition, although protein levels obtained by retrovirus may vary depending on cell types, the TEL-AML1 protein level obtained in mouse fetal pro-B cells using our low virus was comparable to that of two TEL-AML1 leukemia cell lines. However, the TEL-AML1 protein level in preleukemic clones in humans is not known and might be different from that in leukemia cells. These caveats aside, our results nevertheless show that low-level expression of TEL-AML1 promoted self-renewal of B cells but not of myeloid cells, while high-level expression of TEL-AML1 promoted self-renewal of myeloid cells, but not of B cells.

Results obtained using low-level expression of TEL-AML1 also showed that pro-B cells may expand, regardless of whether they are descended from uncommitted immature cells expressing TEL-AML1, or pro-B cells themselves directly targeted for TEL-AML1 expression (Fig. 1). In many clinical leukemia samples, analyses of immunoglobulin receptor rearrangements have demonstrated that preleukemic clones can be propagated after B-lineage commitment [8, 9, 36], while other studies have suggested that the initial hit may be propagated in cells not yet committed to the B cell lineage [10, 11]. These findings and the results of our modeling suggest that a TEL-AML1 translocation may occur in both uncommitted stem/progenitor cells and committed B cells.

Detailed analysis of self-renewing TEL-AML1+ B cells in vivo revealed that cells are capable of differentiating to IgM+ cells that are normally accompanied by a rearranged immunoglobulin light chain, consistent with the findings of Mori et al. in human cord blood samples [17]. Interestingly, a predominance of IgM+ cells takes some time to achieve and is concomitant with a decrease in the proportion of more immature B cells. This observation may indicate the presence of a time “window” for secondary genetic hits to accumulate to elicit leukemia. This finding might be consistent with the reported findings that the TEL-AML1 fusion is detectable in adults at a rate comparable to that in cord blood, but in low cell dosage [7, 37, 38]. Differentiated TEL-AML1+ B cells could, in part, live for a long period in vivo following proper selection in the periphery and antigen stimulation; TEL-AML1+IgM+ B cells retain the ability to undergo class-switching and differentiation to plasma cells (Supporting Information Fig. 5). The ability of TEL-AML1+ cells to differentiate terminally over the long-term might account for the rare occurrence of TEL-AML1 leukemia in adults [39, 40]. Additionally, our results suggest that TEL-AML1 has a reduced impact on adult compared to fetal pro-B cells (Fig. 1F), which is in line with reported findings [12, 16].

We then sought to identify the genes involved in TEL-AML1-associated functions. A comparison of gene expression signatures of TEL-AML1+ versus control pro-B cells facilitated identification of an enrichment of genes overrepresented in ESC compared to differentiated cells. It was recently suggested that in many instances cancer stem cells use ESC-like gene modules to maintain or provoke their stemness [31, 33, 41], although the genes involved in the ESC-like module that are crucial for the generation and maintenance of cancer stem cells remain largely unknown. The roles of ESC-like genes in leukemia have only just begun to be investigated in terms of the leukemia stem cell activity of modeled acute myelogenous leukemia [34]. Here, we addressed the roles of ESC-like genes in the pre-eukemic stage of ALL. Although differences between humans and mice may make direct extrapolation difficult, and our analytical measures of gene expression data may have limitations (including the fact that there is no evidence for the direct regulation of genes by TEL-AML1), our analysis nevertheless suggests that Mybl2, Pim2, Tgif2 and Hmgb3 are crucial components for TEL-AML1 to establish a modeled preleukemic state. Given the nature of overexpression, retroviral transduction of pro-B cells with the four genes might drive a genetic program which differs from the one caused by TEL-AML1, although our findings nevertheless showed that the four genes subverted by TEL-AML1 are themselves sufficient to induce self-renewing pro-B cells. However, the possibility remains that alternative transcriptional programs other than that driven by the four genes also operate in TEL-AML1+ cells.

The functions of a given transcriptional regulator such as TEL-AML1 are largely dependent on the cell context. Therefore, the functional roles of TEL-AML1 might be largely or completely different in initiating preleukemic clones and in cells that acquired additional genetic hits. However, our mouse models showed that TEL-AML1 enhances self-renewal of mutated N-RAS-expressing pro-B cells, and thus contribute to the development of leukemia. Moreover, our use of human TEL-AML1 cell lines suggests that the genes TEL-AML1 influences (such as MYBL2, TGIF2, PIM2, and HMGB3) in initiating self-renewing pro-B cells in our mouse model still play roles in established human TEL-AML1 leukemia, since shRNA-mediated silencing of any one of the four genes attenuated the growth of TEL-AML1 cell lines (Supporting Information Fig. 6). In addition, gene expression analysis comparing TEL-AML1, E2A-PBX1, MLL-rearranged, hyperdiploid, hyperdiploid with <50 chromosomes, hypodiploid and pseudodiploid subgroups of childhood B cell leukemia (GSE12995) [42] shows that TEL-AML1 subgroup is among the highest ones with respect to gene expression of the four gene (Supporting Information Fig. 7). Overall, these findings raise the alternative possibility that gene networks provided or initiated by TEL-AML1 play a role both in the establishment of preleukemia and in cooperation with additional genetic hits that lead to leukemia, although the extent to which TEL-AML1 contributes toward the development of leukemia may vary depending on the nature of the additional hits.

It is known that Mybl2 is involved in regulation of the cell cycle and promotion of survival [43], Pim2 is involved in the regulation of growth and inhibition of apoptosis [44], Tgif2 antagonizes TGFβ signaling [45], and Hmgb3 is involved in regulation of the B cell differentiation program and the balance between self-renewal and differentiation in hematopoietic stem cells [46, 47]. However, the mechanisms by which these four genes collaborate and mediate self-renewal of pro-B cells are currently unclear. Mybl2 appears to be essential for cell proliferation [43]. Pim2-deficient mice are healthy and exhibit no apparent abnormality [48]. Additionally, Tgif2-deficient mice appear uncompromised [49], and Hmgb3 is dispensable for HSCs to renew and reconstitute hematopoiesis [47]. The observation that children in lengthy remission can suffer late relapse with the appearance of a novel leukemic clone, which actually appears to be derived from a preleukemic clone or a minor clone present at the time of initial disease manifestation [50–55], suggests that the TEL-AML1+ clone may persist even after the cells propagating the overt leukemia have been eradicated. Overall, our findings therefore may provide a basis for cancer eradication and prevention, where specific targeting of precancerous cells might be advantageous, while mitigating their adverse effects on HSCs.

CONCLUSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES
  11. Supporting Information

The TEL-AML1 fusion gene is the most prevalent of the genetic abnormalities in childhood leukemia. In the study presented here, we highlight the ability of TEL-AML1 to confer self-renewing ability upon mouse fetal pro-B cells. Such cells can persist in vivo in mice, but differentiate terminally in the long-term, suggesting that additional hits for leukemogenesis need to be acquired before the differentiation. The self-renewing ability also functions to increase the number of leukemia-initiating cells. Gene expression analysis suggests the link between TEL-AML1-mediated self-renewal of fetal pro-B cells and the ECS-associated transcriptional program, in which Mybl2, Tgif2, Pim2, and Hmgb3 appear to be critical components. Exploiting the TEL-AML1-mediated transcriptional program may provide a basis for eradication and prevention of the leukemia.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES
  11. Supporting Information

We thank Seiko Sato for technical assistance and help with animal husbandry. This work was supported by a grant-in-aid for the Second Term Comprehensive 10-Year Strategy for Cancer Control from the Ministry of Health and Welfare (M.S.); a grant-in-aid for scientific research from The Ministry of Education, Culture, Sports, Science and Technology (M.S.); a grant-in-aid for scientific research from the Japan Society for the Promotion of Science (S.T.); and a Research Grant of the Princess Takamatsu Cancer Research Fund (11-24307) (S.T.).

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES
  11. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES
  11. Supporting Information

Additional Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
sc-12-0743_sm_SupplFigure1.TIF60KSupplemental Figure 1. Western blot analysis of TEL-AML1 expression of fetal pro-B cells infected with TEL-AML1“low” and TEL-AML1“high” viruses. Anti-myc and anti-tubulin blots are presented, along with relative cell numbers used for the analysis.
sc-12-0743_sm_SupplFigure2.TIF49KSupplemental Figure 2. Analyses of colony-forming activities of TEL-AML1-expressing pro-B cells. Fetal pro-B cells were isolated from e18 mouse fetal liver, infected with the indicated virus, sorted for GFP, and assayed for replating potential in semi-solid media under the Bcell condition. Note that only TEL-AML1“low”-infected cells efficiently yielded colonies at the 3rd plating. Typical data of three experiments conducted in triplicate are presented.
sc-12-0743_sm_SupplFigure3.TIF130KSupplemental Figure 3. Analysis for differentiation of GFP+/N-RASG12D+ and TEL-AML1+/ N-RASG12D+ cells in transplanted mice. Mice described in the legend to Figure 5C were analyzed for expression of the indicated molecules. Typical data of 4 mice are presented.
sc-12-0743_sm_SupplFigure4.TIF209KSupplemental Figure 4. GSEA for genes with transcription or chromatin annotations. Gene expression data presented in Figure 6 are analyzed for the enrichment of genes with transcription or chromatin annotation in ES_EXPRESSED_1 and ES_EXPRESSED_2 [5]. Enrichment plot and heat map of the leading edge are presented along with NES and nominal p-value. Genes marked with asterisks were included in the analysis conducted in Figure 6.
sc-12-0743_sm_SupplFigure5.TIF94KSupplemental Figure 5. Differentiation of spleen B cells in vitro. TEL-AML1+ B220+ B cells were flow-sorted from spleen of mice that received a transplant with pro-B cells infected with TEL-AML1“low” virus, and cultured in the condition that allows differentiation in vitro. Flow-cytometric analyses of B cells before and after the differentiation induction are presented. Note that IgG1+ B cells and B220-CD138+ plasma cells emerged after the induction.
sc-12-0743_sm_SupplFigure6.TIF66KSupplemental Figure 6. Effects of shRNA-mediated silencing of genes on growth of TEL-AML1cell lines, Reh and KOPN41. (A) Cell number versus time elapsed. Reh and KOPN41 cells were infected with lentivirus vector encoding shRNA for the indicated gene. Infected cells were then selected with puromycin-resistance, and monitored for cell growth. (B) Efficiencies of respective shRNAs on gene expression, as assessed by quantitative PCR, in Reh. Relative levels of transcripts of cells infected with shRNA for the indicated gene are presented, as compared with those in cells infected with shRNA for luciferase.
sc-12-0743_sm_SupplFigure7.TIF66KSupplemental Figure 7. Expression of MYBL2, TGIF2, HMGB3, PIM2 and its related PIM1 genes in childhood ALL cases. (A) Expression of the indicated genes according to subgroups of childhood ALL: TEL-AML1 (n=34), E2A-PBX1 (n=16), MLL-rearranged (n=13), BCR-ABL (n=20), hyperdiploid (n=27), hyperdiploid with <50 chromosomes (n=16), hypodiploid (n=8), and pseudodiploid (n=16). Box plots show the range of gene expression. Median expression is shown by the horizontal line, the size of the box represents the borders that contain 50% of the values, and the error bars represent the upper and the lower quartile of the values. Data were obtained from Affymetrix gene arrays (GSE12995) [27] with the use of a BioConductor program with RMA correction, and normalized values are presented. The statistical significance comparing values of TEL-AML1 samples and those of any one group was calculated using Mann-Whitney U-test. * <0.05, ** <0.01, NS: not significant.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.