ALBERT

Description: Stem cells can be identified by a side population (SP) phenotype in a variety of adult and embryonic tissues. We have previously shown that expression of the Abcg2 serves as a prospective marker for isolating HSCs suggesting that Abcg2 expression may also serve as a marker for stem cell activity in other non-hematopoetic tissues. In particular, skeletal muscle SP cells have been shown to have stem cell activity in muscle reconstitution experiments and the SP population in skeletal muscle is significantly reduced in Abcg2 null mice. To investigate the possibility that Abcg2 can serve as a muscle stem cell marker, we used our mouse strain in which a GFP reporter gene was inserted into the Abcg2 locus. Skeletal muscle cells from adult Abcg2/GFP knock-in mice were isolated based on GFP expression and tested for stem cell activity. To exclude contamination by hematopoetic cells, all experiments were performed on cells gated for the CD45 −/Ter119− phenotype. Flow cytometric analysis showed that 11.6 ± 4.2 % of these muscle cells expressed the Abcg2/GFP allele. Since myogenic progenitor cells have the CD34+/ Sca-1−phenotype, GFP positive and negative cell populations were further analyzed for CD34 and Sca-1 expression. This analysis showed that 15.6 ± 5.3 % of Abcg2/GFP+ cells were CD34+/ Sca-1−. In contrast, 51.3 ±18.3 % of Abcg2/GFP− cells were CD34+/ Sca-1−. These results indicated that Abcg2/GFP− cell population may have a higher frequency of myogenic progenitor cells when compared to Abcg2/GFP+ cells. Analysis of skeletal muscle SP cells for GFP expression showed that 57.5±12 % of the SP and 10.8±0.9 % of non-SP or main population (MP) cells expressed the Abcg2/GFP allele. When SP and MP cell populations were analyzed for CD34 and Sca-1 expression, the highest percentage of CD34+/Sca-1− cells were found in MP/GFP− cell population (33.8±5.3%). Since 61.7 % of total cells were MP/GFP− cells, the greatest absolute number of cells with the myogenic phenotype were found to be located in MP/GFP− population. The growth characteristics and differentiation potential of Abcg2/GFP+ and Abcg2/GFP− cells were then assessed in a myogenic clonal culture assay. Sorted Abcg2/GFP+ and Abcg2/GFP− cells were plated in collagen-coated plates in proliferation medium. Both cell populations increased in number and formed large colonies after 7 days in culture. When these cells were then cultured in myogenic differentiation medium for 4 days, only GFP− cells differentiated into contracting myofibers. In contrast, GFP+ cells differentiated mostly into adherent fibroblast like cells. This data was further validated by DNA micro-arrays analysis of GFP+ and GFP− cell populations. We found that GFP− cells expressed skeletal muscle-specific genes such as MyoD, myf-5, myogenin and troponin whereas GFP+ population did not express any of these genes. Based on these data, we conclude that myogenic progenitor cells did not express the Abcg2/GFP allele. We are currently characterizing the Abcg2/GFP+ population for potential mesenchymal stem cell activity. Transplantation assays to determine myogenic activity of GFP+ and GFP− populations in vivo are in progress.

Description: Neural stem cells have been identified in both the cerebellum and forebrain of fetal and adult mice. These cells can form neurospheres in culture and differentiate into both glia and neurons either in vitro or in vivo. In embryonic day 14 forebrain, neural stem cells are found to exist exclusively in a subpopulation with the Side Population (SP) phenotype, and express Abcg2, a member of the ABC transporter family that is responsible for the SP phenotype in hematopoietic stem cells (HSCs). The expression of Abcg2 in stem cells in the cerebellum has not been characterized. We have generated an Abcg2/GFP knock-in mouse model in which expression of GFP is under control of the endogenous Abcg2 locus and used this model to demonstrate that Abcg2 expression can be used for HSC enrichment. Here we report the use of this mouse model to explore the relationship between Abcg2 expression and neural stem cell function in neonatal cerebellum. Single cells were prepared from cerebellum of 4–9 day old mice by digesting with papain. We then stained the cells with anti-CD45 and anti-Ter119 antibody to exclude the resident hematopoietic cells in subsequent flow cytometry analysis and cell sorting. We found that a small but consistent subpopulation of cells, comprising 0.7±0.12% of total CD45−Ter119- single cell preparations, expressed the Abcg2/GFP allele. To determine whether these GFP+ cells were enriched for neural stem cells, we sorted the CD45−Ter119- cells into GFP+ and GFP− subpopulations and analyzed for their neurosphere forming activity in the presence of epidermal growth factor and basic fibroblast growth factor. We found that the GFP+ subpopulation formed 21 fold more neurospheres compared with the GFP− subpopulation. These neurosphere forming cells can self-renew as evidenced by their capacity to form secondary neurospheres when replated. These results demonstrate that similar to what is seen with HSCs and with embryonic forebrain cells, Abcg2 is expressed in the neural stem cells in neonatal cerebellum, and Abcg2/GFP expression in this mouse model could also be used as a marker to prospectively purify neural stem cells from cerebellum. Ongoing studies are focused on defining the in vivo multilineage differentiation potential of the Abcg2/GFP+ cells and determining whether Abcg2 expression could be used as a marker for purification of medulloblastoma stem cells.

Description: Lentiviral vectors derived from the Simian Immunodeficiency Virus (SIV) mediate relatively efficient transduction of hematopoietic stem cells (HSCs) from rhesus macaques. While integration sites associated with onco-retroviral vectors have been extensively studied in primate transplantation experiments, much less in known about lentiviral vector integration site patterns. The existing literature is limited to one report showing that SIV vectors have a distinctive genomic integration pattern compared with onco-retroviral vectors (Hematti et al 2004). Here we report our results mapping 263 integration sites for SIV vectors in an autologous rhesus macaque transplantation model. Two SIV vectors were used that expressed either MGMT-P140K alone or MGMT-P140K together with HOXB4 from an internal MSCV promoter. Two rhesus macaques were transplanted with autologous CD34+ cells, half of which were transduced with the MGMT vector and half were transduced with MGMT-HOXB4 vector. The first animal was treated with 7 courses of temozolomide and 6-BG which has resulted in selection of transduced cells in vivo, both at the level of myeloid progenitors, and to a lesser degree, in HSCs. A total of 152 integration sites were identified from this animal based on LAM-PCR. Sequence analysis showed a favored preference for integration into transcription units, which comprised 70% of all integrations, with 64% integrations occurring within introns and 6% within exons. The highest density of SIV integration sites per Mbp were on chromosomes 17 and 19 (0.17 and 0.2 respectively). At different time points during drug treatment, multiple clones contributed to hematopoiesis and 24 clones were identified repetitively. The second animal was treated with two courses of TMZ/BG and two courses of BCNU/BG resulting in selection of transduced cells in all lineages. So far, a total of 111 integration sites have been identified in this animal and a similar general integration pattern was observed as seen in the first animal. Integration into transcription units was favored (71%) with 65% occurring within introns and 6% within exons. The three most gene-dense chromosomes 17, 19 and 22 had the highest density of SIV integration sites (0.11, 0.16 and 0.18 respectively). In this animal, 10 out 111 integration sites were identified repetitively during the drug treatments. Vector integrations near previously described oncogenes were identified in both animals (19 out 152 and 11 out of 111 integration sites for each animal respectively). However, no common integration sites (CIS) into a single oncogene were observed and no abnormal hematopoietic proliferation developed in either animal. Moreover, there were no integrations seen within the MDS/Evi locus that has been previously shown to be a CIS for onco-retroviral vectors. Our study shows that the SIV integration pattern is distinctly different from that obtained with murine oncoretroviral vectors and is consistent with the previous study. The lack of integrations within the MDS1/Evi locus represents a potential safety advantage, however further study will be necessary to determine whether the overall propensity for insertional mutagenesis and transformation is decreased. We also show that multiple clones contributed to hematopoiesis before and after MGMT-mediated selection suggesting that this approach is not necessarily associated with restrictions in clonal numbers contributing to hematopoiesis.

Description: Three patients in the French X-SCID gene therapy trial have developed T-cell lymphomas associated with integration of a γ c-expressing oncoretroviral vector into the LMO2 locus. These occurrences have raised important questions regarding the safety of gene therapy for hematopoietic diseases: 1) are there unique risk factors for XSCID gene therapy that increase the risk of insertional mutagenesis; 2) does deregulated expression of the vector-encoded γ c gene contribute to transformation; 3) what other genetic lesions may contribute to transformation; 4) can safer vectors be designed that result in lower levels of cellular gene activation? To address these questions, we generated a mouse model of XSCID gene therapy in which both the Arf tumor suppressor gene and the γ c gene were ablated. Bone marrow cells from these Arf −/−, γ c−/− mice were transduced with a MSCV-γ c-ires-GFP vector or with a control MSCV-GFP vector. These transduced cells were then transplanted into lethally-irradiated, CD45.1+ wild-type mice. After 1 year of follow-up, 13 out of 15 mice in γ c-transplanted group developed lymphomas in which 12 were T-cell lymphomas and 1 was a B-cell lymphoma. All of these lymphomas except one were highly positive for GFP expression and were derived from transplanted donor cells. In contrast, there were only 3 lymphomas in the MSCV-GFP control group, all of which lacked GFP expression and two of which were derived from recipient hematopoietic cells. Southern blot analysis of lymphoma cells in the γ c-group demonstrated that the lymphomas were clonally-derived. Ligation-mediated PCR analysis showed integrations near or within established proto-oncogenes in 8 cases demonstrating that T-cell transformation was associated with potential insertional oncogene activation. To examine whether the X-SCID background was essential to the increased transformation rate, a second transplant experiment was performed in which bone marrow cells from ARF−/−, γ c+/+ mice were transduced with a MFG-γ c vector. Only 4 out of 18 mice developed lymphomas at 54 weeks indicating that the XSCID background was required for accelerated transformation and that there was no direct effect from deregulated γ c expression per se. We hypothesized that γ c gene deletion could lead to an increase in the number of early progenitors with a resultant increase in target cells susceptible to insertional mutagenesis. Flow cytometry analysis indeed revealed that XSCID mice had a 5-fold increase in the number of Lin−, c-kit+, Sca1+ progenitor cells in the bone marrow relative to wild-type mice, suggesting potential expansion of common lymphocyte progenitors was present. Our results show that unique risk factors exist for gene therapy-induced transformation in XSCID suggesting that the risk for gene therapy in other hematopoietic disorders may be significantly less. One unique risk factor is likely to be an expansion of early progenitors resulting from loss of γ c gene function. Loss of tumor suppressor gene function is likely to be a required secondary event; a hypothesis that is consistent with the relatively long latency for tumor development in patients. Lastly, our animal model should now allow us to test vector safety modifications such as the use of self-inactivation vectors and chromatin insulators and could define a new γ c vector suitable for use in XSCID gene therapy.

Description: Retroviral vectors containing internal promoters, chromatin insulators, and self-inactivating (SIN) long terminal repeats (LTRs) may have significantly reduced genotoxicity relative to the conventional retroviral vectors used in recent, otherwise successful clinical trials. Large-scale production of such vectors is problematic, however, as the introduction of SIN vectors into packaging cells cannot be accomplished with the traditional method of viral transduction. We have derived a set of packaging cell lines for HIV-based lentiviral vectors and developed a novel concatemeric array transfection technique for the introduction of SIN vector genomes devoid of enhancer and promoter sequences in the LTR. We used this method to derive a producer cell clone for a SIN lentiviral vector expressing green fluorescent protein, which when grown in a bioreactor generated more than 20 L of supernatant with titers above 107 transducing units (TU) per milliliter. Further refinement of our technique enabled the rapid generation of whole populations of stably transformed cells that produced similar titers. Finally, we describe the construction of an insulated, SIN lentiviral vector encoding the human interleukin 2 receptor common γ chain (IL2RG) gene and the efficient derivation of cloned producer cells that generate supernatants with titers greater than 5 × 107 TU/mL and that are suitable for use in a clinical trial for X-linked severe combined immunodeficiency (SCID-X1).

Description: Stem cells from a variety of tissues could be identified by a side population (SP) phenotype. We have identified an association between expression of the ABCG2 transporter and the SP phenotype, although the exact relationship between these two cellular characteristics is still not well understood. To further explore the relationship between Abcg2 expression and SP phenotype, we have generated a Abcg2/GFP reporter mouse model in which an IRES-GFP expression cassette was inserted down-stream of the stop codon of Abcg2 gene by homologous recombination, so that the expression of GFP is under the control of Abcg2 locus but endogenous Abcg2 expression was unperturbed. Immunohistochemistry of tissue sections using an anti-GFP antibody confirmed expression of GFP in tissues that normally express Abcg2, such as renal proximal tubules and intestinal epithelium. Flow cytometry analysis showed that about 10% of total bone marrow mononuclear cells expressed the Abcg2/GFP allele. Within this total GFP+ population, greater than 99% of the cells fall outside the SP region. Staining with lineage specific antibodies showed that 77% of the non-SP/GFP+ cells are Ter119+ erythroid cells. In contrast, only 0.5% of the total GFP+ population falls within the SP region. Approximately 90% of these bone marrow SP cells expressed the Abcg2/GFP allele. Transplantation studies with bone marrow cells that were depleted of lineage positive cells were then performed. All 5 mice transplanted with 100 Lin−, SP+, GFP+ cells were reconstituted in both myeloid and lymphoid lineages. In contrast, no repopulating activity was detected with up to 10,000 Lin−, non-SP, GFP− cells. When Lin−, non-SP, GFP+ cells were transplanted, only 2 of 5 animals were reconstituted with 5000 cells, and no repopulation was seen with lower cell doses. These results show a complex relationship between Abcg2 expression and the SP phenotype. Virtually all SP cells express the Abcg2/GFP allele, and these cells are highly enriched for repopulating activity in transplant assays. However, the majority of bone marrow cells that expressed the Abcg2/GFP allele were Ter119+ erythroid cells that do not bear the SP phenotype. In addition, there are non-SP, GFP+ cells that lack expression of Ter119 and other mature lineage markers. These Lin−, non-SP, GFP+ cells have relatively lower repopulating activity compared to SP cells.

Description: There is now accumulating evidence for the existence of rare cancer stem cells that resemble adult stem cells in their ability to replicate and produce more specialized cells constituting the bulk of the tumor. Neuroblastoma is the most common childhood cancer, developing extracranially from neuroblasts of the body. Approximately 70–80% of patients have metastatic disease at the time of diagnosis, and fewer than half of these patients are cured. Human neuroblastoma cells have been shown to contain a subpopulation of cells with a high capacity to efflux Hoechst 33342 nuclear dye, resulting in a distinct side population (SP) phenotype (Hirschmann-Jax et al., PNAS, 2004). These SP cells also express high levels of ABCG2 and ABCA3 transporter genes. We have used a mouse model to further investigate the relationship between the SP phenotype and Abcg2 expression in neuroblastoma stem cells. Mice, expressing N-myc in neural-crest cells, develop neuroblastomas at early age (Weiss et al., EMBO J, 1997). We have found that these neuroblastomas can be divided into three groups according to their SP phenotype; no SP cells present, low SP cells (0.6–2% of total cell number) and high SP cells (20–40% SP cells in total neuroblastoma cell population). When present, the SP fraction was significantly decreased after treatment of the cells with gleevec and fumitremorgin C, inhibitors of Abcg2 function ( 4.7% with treatment vs. 30.5% untreated in one case). This result indicates that Abcg2 is a major determinant of SP phenotype in these tumors. Quantitative PCR, performed on sorted SP and non-SP cells confirmed about 6 fold higher level of Abcg2 expression in the SP cell fraction in comparison with non-SP. In order to determine the clonogenic capacity of different tumor cell populations, varying numbers of tumor cells were injected in the flanks of NOD/SCID/gamma null mice. Transplantability of the tumors was found to correlate with SP phenotype. At a dose of 106 cells per recipient, neuroblastomas with no SP cells did not form tumors (0 tumors developed in 6 recipients). Neuroblastoma cells with low SP cell numbers (0.6% of total cells) formed tumors in 2 out of 4 transplants at this cell dose. Neuroblastomas with a high SP cell population (30% of the cells) had the highest clonogenic activity, forming tumors in 6 out of 6 transplants at 106 cells per injection. These results indicate that tumor stem cells are more abundant in high SP tumors in comparison with tumors with lower SP cell fractions. Next, sorting experiments based on the SP phenotype indicated that SP cells are enriched for neuroblastoma stem cells. In one experiment using a neuroblastoma sample with 22% SP cells, recipients were inoculated with a dose of 105 sorted cells. Three out of 4 mice formed tumors after transplantation with sorted SP cells while only 1 of 4 mice transplanted with non-SP cells developed tumors. Secondary tumors, developed from sorted SP cells, had themselves higher proportion of SP cells in comparison with tumors, developed from non-SP cells (35–40% and 8–10%, respectively). We are now using this genetic mouse model to further study the use of Abcg2 expression to isolate neuroblastoma stem cells.

Description: ABCG2 encodes a transmembrane transporter associated with multi-drug resistance in various cancer cells. ABCG2 is also highly expressed in hematopoietic stem cells (HSCs) and is down-regulated in most committed progenitors, while expression is sharply up-regulated during erythroid differentiation and significant levels of protein expression are detected on mature red blood cells. The expression of ABCG2 in these hematopoietic cells likely reflects a protective role against both endogenous and exogenous toxins. The mechanisms for regulation of ABCG2 expression in hematopoietic cells are poorly understood. We have recently identified three novel leader exons (termed E1a, E1b and E1c) located in the 5′ untranslated region (5′-UTR) of mouse Abcg2 mRNA by EST database searches and reverse-transcription PCR (RT-PCR). In genomic DNA, the distance of these noncoding exons from exon 2, which contains the translation start codon, is 58.5 kb for E1a, 15.0 kb for E1b and 5.1 kb for E1c. In a mouse erythroid cell line (MEL), RT-PCR analysis showed that the transcript containing E1b exon was the only isoform detected. Furthermore, in primary Ter119+ erythroid cells from mouse bone marrow (BM), real time PCR showed that expression levels of the E1b-containing transcript were at least 100 fold greater than those of the E1a transcript. For comparison, cDNA was isolated from c-Kit+, Sca-1+, Lin- (KSL) BM cells, a population devoid of erythroid cells but highly enriched for repopulating HSCs. Real-time PCR showed that the E1a-containing transcript was the major expressed isoform in KSL cells, while the E1b transcript was present at significantly lower levels. The differential expression pattern of Abcg2 mRNA isoforms in HSCs and erythroid cells indicates that at least 2 different promoters control Abcg2 transcription during hematopoiesis. We also analyzed the human EST database and found 3 different ESTs all containing exon 2, similar to mouse Abcg2 gene. RT-PCR analysis of cDNA derived from human placenta and BM mononuclear cells showed expression of all 3 isoforms, which was confirmed by plasmid cloning and sequencing. These results indicate that the human ABCG2 locus also has three leader exons which are alternatively used and give rise to three isoforms of ABCG2 mRNA that differ in their 5′-UTR. We are currently defining the expression pattern of these isoforms in different human hematopoietic cell subpopulations. In summary, our data show the expression of Abcg2 during hematopoiesis is regulated at the level of transcription by alternative use of multiple noncoding leader exons in lineage-specific manner. Functional analysis of different putative promoter regions is underway and should provide a better understanding of the developmental regulation of Abcg2 expression in hematopoiesis.