S/MAR sequence confers long-term mitotic stability on non-integrating lentiviral vector episomes without selection

Abstract

Insertional oncogene activation and aberrant splicing have proved to be major setbacks for retroviral stem cell gene therapy. Integrase-deficient human immunodeficiency virus-1-derived vectors provide a potentially safer approach, but their circular genomes are rapidly lost during cell division. Here we describe a novel lentiviral vector (LV) that incorporates human ß-interferon scaffold/matrix-associated region sequences to provide an origin of replication for long-term mitotic maintenance of the episomal LTR circles. The resulting 'anchoring' non-integrating lentiviral vector (aniLV) achieved initial transduction rates comparable with integrating vector followed by progressive establishment of long-term episomal expression in a subset of cells. Analysis of aniLV-transduced single cell-derived clones maintained without selective pressure for >100 rounds of cell division showed sustained transgene expression from episomes and provided molecular evidence for long-term episome maintenance. To evaluate aniLV performance in primary cells, we transduced lineage-depleted murine hematopoietic progenitor cells, observing GFP expression in clonogenic progenitor colonies and peripheral blood leukocyte chimerism following transplantation into conditioned hosts. In aggregate, our studies suggest that scaffold/matrix-associated region elements can serve as molecular anchors for non-integrating lentivector episomes, providing sustained gene expression through successive rounds of cell division and progenitor differentiation in vitro and in vivo.

Figure 1.

Effect of wPRE replacement with S/MAR sequence on transgene expression and production of non-integrating vector episomes. (A) Diagram of HIV-1-based self-inactivating LV pLVCG with wPRE cloned within the transgene expression cassette pLVCG harboring human S/MAR from the ß-interferon gene cluster. The GFP-wPRE fragment was released by digestion with XbaI and EcoRI to replace with the S/MAR fragment. The dotted line corresponds to the DIG-labeled probe binding site. S/MAR is cloned at the transcriptionally active region between GFP reporter ORF and 3′LTR after replacing the wPRE fragment in LVCG plasmid. (B) Four LVs classified on integration status and S/MAR presence were used in the study. iLV: integrating lentiviral vector harboring wPRE sequence, niLV: non-integrating lentiviral vector harboring wPRE, iLV-S/MAR: integrating lentiviral vector harboring S/MAR sequence, aniLV: anchoring non-integrating LV harboring S/MAR sequence. (C) Microscopy images of GFP expression from iLV and iLV-S/MAR-transduced 293T cells at the day-2 time point. Integrating vector stocks were produced by using the helper plasmids pLP-1, pLP-2 and pLP/VSVG. Non-integrating vector stocks were produced by using the helper plasmids pCD/NL2 ΔInt, pLP-2 and pLP/VSVG. (D) Diagram representing the LTR circle formation due to abortive integration of niLVs. The class I integrase mutant used to produce niLVs is a D64V mutant that limits IN function regarding cleavage and integration but maintains normal levels of viral DNA. The niLVs are produced by using the helper plasmids pCD/NL2 ΔInt, pLP-2 and pLP/VSVG. After transduction, the abortive integration event of the vector genome results the formation of reverse-transcribed circular DNA vector genomes with one or two LTRs. All LTR sequences used in this study are promoterless self-inactivating (SIN) vector sequences.

Figure 2.

Transgene expression from LVs and long-term GFP expression from aniLV-transduced 293T cells. (A) Time-course evaluation of GFP expression from the LV-transduced 293T cells at post-transduction time points. The population of GFP-positive cells was analyzed at intervals to measure the effect of wPRE replacement by S/MAR and functional integrase replacement by defective integrase. (B) Long-term stability of GFP expression from the aniLV-transduced and enriched cells after sorting. Mitotic stability and rate of gene expression from integration defective S/MAR containing LV in relation to control vectors. (C) Microscopy images of GFP florescence from the LV-transduced 293T cells. GFP signal was consistently lower in integrase-defective niLV vectors and S/MAR harboring vectors. Scale bars: 400 µm. (D) Comparison of GFP expression levels (MFI) from LV-transduced 293T cells two days post-transduction. GFP expression was significantly lower in integrase-defective wPRE-harboring vector than the integrating wPRE vector.

Figure 3.

Episomal persistence and gene expression in 293T LV clones. (A) GFP-positive cells in LV-transduced single cell clones of long-term cultured 293T cells. All clones were maintained in culture without selection pressure. Clonal variability of GFP-positive cells in aniLV clones (green) was observed compared with iLV-transduced 293T clones (blue). (B) MFI in LV-transduced single cell clones 8 weeks after transduction with aniLV or iLV vector at matched MOI. Decreased MFI from aniLV clones compared with iLV clones. (C). Agarose gel electrophoresis of PCR and Southern blot on LTR junctions (bottom) in aniLV clones using PPT F and PBS R primers. M1: DIG-labeled Lambda HindIII DNA ladder (Roche), M2: 50 bp DNA ladder. Numbers represent DNA template from the respective clone used for PCR and Southern blot analysis. (D) Sequence of LTR circles of aniLV clones and the primer orientation on the vector episome. LTR junctions of the episomes are sequenced and analyzed. Complete 2-LTR episomes are not observed in the screened clones. Eighteen clones have 1-LTR junctions and one clone showed one full-length LTR and a partial LTR (ΔLTR) spanning by PPT and PBS sequence on both sites.

Figure 4.

Mitotic stability and episomal status of anchoring niLV. (A) Southern blot analysis to identify the episomal and integrated vector copies. An EcoRV site located upstream to GFP ORF was used to linearize the vector. EcoRV-digested DNA samples were used, and the signal was detected by DIG-labeled probe as shown in Figure 1A. A single ≈6-kb DNA signal was observed in the analyzed aniLV clones. No signal was observed to indicate integrated copies in the analyzed aniLV clones. M1: DIG-labeled Lambda HindIII DNA ladder (Roche). Numbers designate DNA from each clone used for Southern blot analysis. Asterisk corresponds to the average position of the target DNA. (B) Alu-PCR analysis to validate the background integration status of aniLV episomes. Primer targets from each vector backbone (PPT-F) and Alu sequence (Alu-1) are used in combination to amplify the region between the integrated vector copy and nearby Alu fragment. M1: 100 bp DNA ladder, C1: 293T gDNA used as template (C) Agarose gel electrophoresis of episome-specific PCR on aniLV clones using PPT F and GFP R and the corresponding Southern blot. M1: DIG-labeled Lambda HindIII DNA ladder, M2: 1-kb DNA ladder, C1: pEpi plasmid DNA used as template, C2: 293T gDNA used as template. Number represents DNA template of each clone used for Southern blot analysis, asterisk corresponds to the average position of the target DNA. (D) FISH analysis on multiple metaphases of one aniLV clone (12) shows nuclear vector genomes.

Figure 5.

LV transduction into murine HPC in vitro and in vivo. (A) Colony-forming assay of LV-transduced mHPCs in cytokine-supplemented methylcellulose. GFP expression from the colonies was analyzed at 10–12 days after transduction. Total number and GFP-positive colonies were counted. (B) Microscopic images of GFP expression from iLV and aniLV-transduced mHPC-derived colonies. Lower GFP expression was observed in aniLV compared with iLV-transduced colonies. (C) GFP expression profile in aniLV-transduced mHPC-transplanted animals. Each cohort represents five animals that received aniLV-transduced mHPCs. Representative FACS analysis of whole-blood cell sample of post-transduction and transplantation of mHPCs into the mice at 3 and 10 weeks, respectively. Lower MOI was consistent with CFU assay and 293T aniLV clones compared with cells transduced with iLV. (D) PCR analysis to detect the presence of episomal vector DNA in the vector-transduced CFUs. LTR junction amplification using PPT-F and PBS-R primers from aniLV CFUs suggests that colonies still harbor vector episomes at detectable levels. GFP PCR as vector control and murine-specific GapDH PCR to confirm the presence of genomic DNA in positive control sample and in untransduced control CFU-c.

Figure 6.

Model of aniLV delivery and nuclear retention. The vector genome undergoes reverse transcription and forms pre-integration complex after cellular entry and uncoating. The nuclear transport of the preintegration complex across the nuclear membrane delivers the vector DNA in to the host nucleus. Owing to the abortive integration, vector DNA circularizes at LTR junctions and forms S/MAR harboring episomes. Owing to S/MAR-mediated nuclear retention, episomes persist in the nucleus and attach to the nuclear matrix followed by replication and segregation during cell division.