ALBERT

Description: The multidrug resistance transporter, ABCG2, is expressed in primitive hematopoietic stem cells from a variety of sources. These cells are detected in dual wave-length fluorescent FACS profiles as a “side population” (SP cells) on the basis of their ability to efflux the fluorescent dye, Hoechst 33342. We have previously shown that 2 types of human short term repopulating cells (STRC) can be enumerated by limiting dilution analysis of their efficient ability to regenerate exclusively myeloid cells after 3 weeks (STRC-Ms), or both myeloid and lymphoid cells after 6–12 weeks (STRC-MLs) in NOD/SCID-b2microglobulin-/- (b2m-/-) mice. Previous findings also implicated these STRCs as determinants of the rapidity of early hematologic recovery in patients transplanted with cultured mobilized peripheral blood (mPB) cells. Here we asked whether any human STRCs have an SP phenotype and hence whether the isolation of SP cells would retain the rapid repopulating activity of a clinical transplant. CD3- SP and non-SP cells were isolated by FACS from low-density (LD) mPB cells after Hoechst staining and transplanted at limiting dilutions into 117 sublethally irradiated b2m-/- mice. The numbers and types of human hematopoietic cells present in the bone marrow of these mice were subsequently monitored by FACS analysis of bone marrow cells aspirated serially, 3, 8 and 12 wks post-transplant. A verapamil-sensitive SP population was reproducibly detected in all 5 patients’ samples studied (0.039 ± 0.012% of the CD3- LD cells). The in vivo assays failed to detect either STRC-Ms or STRC-MLs in the SP fraction and all these activities were obtained from the non-SP cells. If even a single recipient of the largest dose of SP cells transplanted had been positive, this would have detected 10% of the STRCs present. Thus, 〉90% of all STRC-M and STRC-ML in mPB are non-SP cells. However, 4 of 40 mice transplanted with SP mPB cells produced some B-lymphoid cells only starting 12 wks post-transplant. However, this result is difficult to interpret since subjecting the STRC-Ms to the Hoechst 33342 staining and FACS isolation procedure alone eliminated their ability to generate megakaryocytic progeny in vivo, although this did not occur when these cells were just stained for CD34 and then isolated by FACS. In addition, the differentiation behaviour of STRC-MLs was not affected by the Hoechst staining and subsequent FACS isolation procedure. In summary, we demonstrate that purification of SP cells depletes human mPB transplants of STRCs, thereby raising serious concerns about the safety of any clinical use of SP cell-enriched transplants as stem cell support after myeloablation. Our results also suggest that the staining and enrichment procedure for isolating SP human cells may differentially affect the lineage potential of some types of STRCs, including those whose activity may be indispensable for rapid and multi-lineage hematologic recovery.

Description: Successful selection of gene modified hematopoietic stem cells against the majority of non-modified cells may increase the efficiency and safety of clinical gene therapy. Especially if a reduction of the intensity of the myelotoxic pre-transplant conditioning treatment is sought, selection of the initially low percentage of retrovirally gene modified stem cells is required. In addition to the decreased drug related toxicity, such a procedure reduces the risk of genotoxicity caused by insertional mutagenesis simply by diminishing the likelihood of transplanting stem cells that carry unwanted insertion sides compared to strategies that attempt to increase efficiency by increasing vector dose and/or the number of engrafted gene-modified cells. To date, the mutant O6-methylguanine-DNA methyltransferase (MGMT) enzyme that confers resistance to nitrosoureas such as BCNU is the drug-resistance gene that allows most efficient selection at the stem cell level. In the murine model, MGMT selection by BCNU and O6-BG has mostly been performed at LD 50 dose levels of the selecting agents, that is associated with very considerable toxicity. We now established minimal dosage requirements of transplanted cells as well as of BCNU and O6-BG that allow an efficient selection of murine long-term hematopoiesis. Bone marrow cells from 5-FU treated C57 BL/6J mice were transduced with an MGMT/IRES/eGFP encoding retroviral vector and transplanted at two dilutions (1x105 and 4x105) into 84 lethally irradiated syngeneic recipient mice. Starting 4 weeks post-transplant, the mice were treated monthly with two reduced dosages of O6-BG and BCNU (either 10 mg/kg O6-BG and 2,5 mg/kg BCNU or 20 mg/kg O6-BG and 5 mg/kg BCNU) as compared to the commonly used regimens. Kinetics, differentiation and clonality of the transduced hematopoiesis were monitored by FACS and LAM-PCR analysis of serial peripheral blood samples over a total of 5 rounds of selection. 3 of 48 mice died during the selection procedure. Although in 97% of all mice GFP+ blood cells were detectable 4 weeks after transplantation, an efficient selection (〉twofold increase in transduced cells) was only seen in mice with 〉3% GFP+ initial blood cells. The percentage of transduced hematopoiesis in 9 of 24 mice treated with the higher BCNU and O6-BG dosage increased at least twofold as compared to 5 of 24 treated with the lower dosage and 1 of 24 mice without selection. Using Poisson statistics, the frequency of selectable hematopoietic units was calculated to be as low as 1 per 5x105 cells transplanted. Nevertheless, using LAM-PCR we detected stable oligoclonal hematopoiesis after transplantation of 4x105 cells for more than 6 months, suggesting that a subfraction of repopulating cell clones present were not selectable, or that the contribution of individual clones was small. No differences in lineage differentiation of the GFP+ hematopoiesis as determined in lineage marker expression by FACS could be observed after selection as compared to unselected control mice. In summary, our results demonstrate that a reduced dosage of BCNU and O6-BG with lower toxicity allows selection of MGMT expressing murine hematopoietic stem cells with unperturbed multilineage differentiation potential. Moreover, under these conditions, a subfraction of transduced repopulating cell clones was selectable.

Description: We have previously reported that ex vivo retroviraly-mediated gc gene transfer into CD34 (+) bone marrow precursor cells led to the correction of the immunodeficiency in 9 out of 10 patients with X-linked severe combined immunodeficiency. Follow-up now reaches more that 6 years for the first 2 treated patients. Patients’immune function has been restored. The distribution of both TCR Vb family usage and TCR Vb CDR3 length still reveals a broadly diversified T cell repertoire. Moreover 6 years after treatment the thymus is still seeded by transduced progenitor cells as attested by the presence of TRECS in peripheral blood RTE. Among these patients, three (P4, P5 and P10) developed at 30 to 34 months after gene therapy a monoclonal T cell proliferation requiring a chemotherapy. P4 received also an allogenic HSCT from a MUD but died 26 months after the occurence of the lymphoproliferation. For P5 and P10, chemotherapy has led to an overall control of the clonal proliferation. These two patients are doing well and P5 is off treatment with a good immunological recovery. Genetic analysis of the blastic cells showed that in the two first cases the vector had integrated within or upstream of the LMO2 locus causing an insertional activation of LMO2 transcription. The last case revealed the involvement of several targeted sites, but their exact contribution to the lymphoproliferation is still under investigation. The repeated involvment of LMO2 as a site of vector integration in the proliferating T-cells points to an insertional activation of this gene as at least one of the causes of the oncogenic process. However, the long latency observed in all cases (〉 30 months) suggests that additional “hits” have been required for overt desease. Synergy with gc expression and thereby induced proliferative signals (explaining occurrence in SCID-X1 patients only) is the most obvious hypothesis which we are trying to analyse in a mouse model. A deep analysis of retroviral integration patterns has been performed on patients’PBMCS by LAM-PCR to estimate the frequency of potentially harmful integration events and to assess the risk factors associated with the LTR’s strong enhancer effect of the MLV-based retroviral vector. 708 unique integration sites (IS) have been obtained from all analysed patients post-gene therapy and among them, 577 could be mapped unequivocally to the human genome. *Most of these insertions (63%) are located in the vicinity of 10kb or within the coding sequence of a known gene*. A significant peak of insertion frequency is related closely to the transcription strart site *among the 577 IS, 43 are common integration sites. Among the latter, we found out a high selection of genes involved in human oncogenic process.