Stem cell marking with promotor-deprived self-inactivating retroviral vectors does not lead to induced clonal imbalance

Abstract

Stable genetic modification of stem cells holds great promise for gene therapy and marking, but commonly used gamma-retroviral vectors were found to influence growth/survival characteristics of hematopoietic stem cells (HSCs) by insertional mutagenesis. In this article, we show that promoter-deprived gamma-retroviral self-inactivating (pd-SIN) vectors allow stable genetic marking of serially reconstituting murine HSC. In contrast to findings with gamma-retroviral long terminal repeat (LTR) vectors, serial transplantation of pd-SIN-marked HSC in a sensitive mouse model was apparently not associated with induced clonal imbalance of gene-marked HSC. Furthermore, insertions of pd-SIN into protooncogenes, growth-promoting and signaling genes occurred significantly less frequent than in control experiments with LTR vectors. Also, transcriptional dysregulation of neighboring genes potentially caused by the pd-SIN insertion was rarely seen and comparatively weak. The integration pattern of promotor-deprived SIN vectors in reconstituting HSC seems to depend on the transcriptional activity of the respective gene loci reflecting the picture described for LTR vectors. In conclusion, our data strongly support the use of SIN vectors for gene-marking studies and suggest an increased therapeutic index for vectors lacking enhancers active in HSC.

Figure 1

Promoter-deprived γ-retroviral self-inactivating (pd-SIN) vector design and transcriptional activity. (a) Construction of the vector SF91-eGFP was described earlier.²² pd-SIN was derived from SINSFeGFP (refs. 19,23) by deleting the SF-promotor/enhancer sequences. It thus does not contain any promotor/enhancer element. (b) No enhanced green fluorescent protein (eGFP) transcription was detectable in murine SC-1 fibroblasts after transduction with pd-SIN (performed in duplicates) by northern blot analysis using an eGFP probe (lanes 2,3). In contrast, SC-1 cells transduced in parallel with SINSFeGFP (in duplicates) revealed strong eGFP transcription (arrow, lanes 4,5). Lane 1 shows a mock control. RNA load was controlled by hybridization with an 18S probe (lower blot, arrow). (c) Since pd-SIN-transduced cells were weakly eGFP-positive in fluorescence-activated cell sorting analysis, quantitative reverse transcription–PCR analyses were performed. As shown, normalized transcription levels were reduced by a factor of >25.

Figure 2

Experimental setup. The design of the mouse study follows the well established, sex-mismatched diallelic (CD45.2/Ly5.2 into CD45.1/Ly5.1) serial bone marrow transplantation model.^7,8,21 This model had enabled us to discover serious side effects of γ-retroviral gene transfer before their occurrence in clinical gene therapy studies, in particular the phenomenon of induced clonal dominance. Shortly, Lin– bone marrow cells from male donors were transduced with the promoter-deprived γ-retroviral self-inactivating vector and transplanted into female mice after TBI (10 Gy). Seven months later, hematopoietic organs of these mice were thoroughly investigated before secondary transplantation into irradiated mice took place. After a further observation period of 5 months, final analysis of hematopoietic organs, including molecular analysis of vector insertion sites, was performed.

Figure 3

Analysis of promoter-deprived γ-retroviral self-inactivating insertion sites by different PCR techniques. (a) A modified ligation-mediated PCR (LM-PCR) method²⁶ was used to assess genomic locations of integrated vector copies in reconstituting hematopoietic cells (~12 months after initial transduction and 5 months after serial bone marrow transplantation). A representative agarose gel analysis showing the five secondary recipients from Exp. 1 is depicted. Twelve different insertions representing dominant clones could be identified. One insertion (I-2) was dominant in four out of five secondary animals (arrow). (b) Insertion site specific (is-PCR) revealed the distribution of hematopoietic clones bearing specific insertion in individual animals of the first and second cohorts of transplanted mice. As a representative example, is-PCR to detect the hematopoietic clone marked by insertion I-2 is shown. As illustrated, this clone took part in the reconstitution of hematopoiesis in all secondary animals. Insertion I-2 had occurred into the Efcab2 gene not known to be involved in hematopoietic stem cell (HSC) growth regulation. (c) Shot-gun cloning of LM-PCR products followed by sequencing of individual clones was performed to assess the relative contribution of specifically marked HSC clones to hematopoiesis in individual animals. Again, insertion I-2 has been selected as a representative example for this type of analysis. Quantitative data are in very good agreement with data from (a) LM-PCR and (b) is-PCR.

Figure 4

Specific distribution of promoter-deprived γ-retroviral self-inactivating (pd-SIN) insertion sites in reconstituting hematopoietic stem cell. Insertion sites were categorized based on the genes present in a 150-kb window at both the 5′ and the 3′ sites of each individual insertion. Genes were classified into (putative) protooncogenes (class 1), signaling genes (class 2), other genes (class 3), and genes of unknown function as previously described.²⁵ For control, respective data obtained in the two parallel control experiments with the γ-retroviral long terminal repeat (LTR) vector SF91-eGFP are shown. As evident, pd-SIN insertions show a strongly different insertion pattern as compared to γ-retroviral LTR vectors. Based on the frequency of hits into putative growth–promoting genes (classes 1 and 2), this difference is highly significant (P < 0.002; Fisher's exact test). In contrast, no significant differences were found between insertion patterns for SF91-eGFP (control group) and historical data from the IDDb.²⁵

Figure 5

Minor transcriptional dysregulation in gene loci targeted by promoter-deprived γ-retroviral self-inactivating (pd-SIN). Real-time reverse transcription–PCR using QuantiTect Primer Assays were used to assess transcription levels of genes adjacent to pd-SIN vector insertions. RNA levels as compared to healthy control mice splenocytes were established using the ΔΔC_t method.^8,49 Only comparatively slight changes in transcription levels were detected. In particular, genes adjacent to insertion sites found to mark clones dominant in more than half of the secondary recipients did not show relevant dysregulation of transcription (I-2: Efcab2; II-3: B3bp). Relative error bars indicate SDs from at least three independent assays.

Figure 6

Promoter-deprived γ-retroviral self-inactivating (pd-SIN) preferentially integrates into transcriptionally active regions. (a) Insertion sites of pd-SIN were correlated with transcription levels of the targeted gene loci. To do so we made use of an expression database for murine hematopoietic stem cells established by Goodell and colleagues.³³ All genes contained in this database were plotted against their respective expression values. The shown line consists of >20,000 individual data points representing the highest expression values for each gene on the array. As shown, most genes hit by pd-SIN (gray triangles) were transcriptionally highly active in hematopoietic stem cell (HSC). (b) The relative transcriptional activity of genes targeted by pd-SIN was compared with that of all genes contained in the HSC expression database.²⁹ Clearly, pd-SIN preferred actively transcribed gene loci for integration.