354 the integration profile of lentiviral vector in CD34+ hematopoietic stem cells is dependent on the cell cycle status of the target cell

354 The Integration Profile of Lentiviral Vector in CD34+ Hematopoietic Stem Cells Is Dependent on the Cell Cycle Status of the Target Cell Molecular Therapy Volume 20, Supplement 1, May 2012 Copyrigh[.]

Molecular Therapy Volume 20, Supplement 1, May 2012 Copyright © The American Society of Gene & Cell Therapy

S138

RNA VIRAL VECTORS

are partially recapitulated with the ML6 chimera and entirely with

the ML14, which includes an additional mutation translating in an

enhanced targeting specifi city by one order of magnitude Preliminary

high throughput integration sites analysis reveals a shift in the histone

H3 targeting chimeras profi le Our approach could signifi cantly

reduce integration into open chromatin sensitive sites in stem cells at

the time of transduction (Biasco et al, 2011), a feature which might

signifi cantly decrease subsequent risk for insertional mutagenesis

through undesired enhancer-activation, in combination with a potent

and stable genetic insulation, regardless of the integration site inside

heterochromatin

352 Effi cient and Non-Toxic Targeting of a

Human Safe Harbor Locus with a Lentiviral Vector

Associated Meganuclease

Chenxia He,1 Agnès Gouble,2 Alix Bourdel,1 Aleksander Edelman,1

Laurent Poirot,2 Frédéric Paques,2 Olivier Danos.1,3

1 INSERM U845, Université Paris Descartes, Faculté de Médecine

Necker, Paris, France; 2 Cellectis SA, Paris, France; 3 Cancer

Institute, University College London, London, United Kingdom.

Meganucleases (MN) are site specifi c endonucleases, with 12

to 45 bp DNA recognition sites They generate double-strand

breaks that can be repaired either by homologous recombination

or by non-homologous end joining (NHEJ) MN can be engineered

for custom recognition of any genetic locus and used for gene

targeting Our interest is to develop effi cient and non-toxic means

of targeting transgenes at specifi c loci for therapeutic purpose We

have previously shown that an active MN could be delivered to

cells as a protein associated to a lentiviral vector particle Here, we

further explore the possibilities of associating an I-CreI-derived

single chain meganuclease to lentiviral vectors We have designed

a MN, CLS4076, with a unique cutting site on human chromosome

14, at a locus selected as a potential safe harbor for the insertion

of therapeutic transgenes The initial characterization of CLS4076

indicates a high level of activity, associated with signifi cant cell

toxicity when introduced by transfection at high doses Protein

fusions were constructed to drive the incorporation of the MN into

lentiviral particles Viral proteins (Vpr and Vpx) and other candidates,

all of which interact with HIV-1 Gag, were tested as fusion partners

Different fusions were found to efficiently associate CLS4076

to virions A quantitative extra-chromosomal assay based on the

recombination-mediated rescue of a luciferase reporter gene was used

for the initial characterization of MN carrying lentiviral particles Cell

toxicity was measured over 7 days following transduction and found

to be absent or minimal at high doses of particles In comparision,

transduction of a regular CLS4076 coding lentiviral vector was

associated with up to 45% of cell death MN transducing virions are

being evaluated for targeting the same safe harbor locus on human

chromosome 14 Data will be presented on the amount of NHEJ

induced at the locus, and on the analysis of homologous recombination

in single cell clones

353 bInSiGHT: Bioinformatics Integration Sites

Tool for Gene Therapy with High-Throughput

Platforms

Andrea Calabria,1 Fabrizio Benedicenti,1 Davide Cittaro,2 Elia

Stupka,2 Christof von Kalle,3 Manfred Schmidt,4 Luigi Naldini,1

Eugenio Montini.1

1 San Raffaele Telethon Institute for Gene Therapy, Milano, Italy;

2 Center for Translational Genomics and Bioinformatics, San

Raffaele, Milano, Italy; 3 National Center for Tumor Diseases,

Heidelberg, Germany.

Insertional mutagenesis is one of the major hurdles of gene therapy

with integrating vectors Analysis of chromosomal vector integration

sites in vector marked cells from gene therapy patients and preclinical models has enabled to detect in vivo selection of gain-of-function insertional mutants even before they progress to overt malignancy For this reason, over the last years there has been a constant increase

in the amount of sequencing and mapping of vector/genomic DNA junctions and statistical tools to improve biological investigations and their interpretation Recently, Next Generation Sequencing (NGS) approaches have been exploited in Gene Therapy to greatly enhance the analytical power of integration site analysis However, NGS analytical advances are balanced by computational drawbacks such as the fl ooding of data that needs to be carefully managed and processed with smart pipelines on distributed high-performances infrastructures Thus NGS-structured and standardized pipelines for biologists are strongly required We developed a NGS bioinformatics pipeline, both for Illumina and Roche platforms, for integration sites analysis with intuitive web-based interfaces, enabling quality controls and high-performance properties: bInSiGHT We designed bInSiGHT data processing activities grouped in: (a) sequence reads quality control, avoiding unreliable data analysis, (b) raw data cleaning and vector sequences trimming, (c) sequence reads grouping and clustering, to estimate the number of integration sites before mapping sequences, (d) reads mapping to reference genome (e) post-processing data

fi ltering and refi nement of integration sites for subsequent analyses (i.e integration sites association to annotated genomic features, common integration site identifi cation and complexity estimation as surrogate of clonal diversity of vector marked cells) We developed and tested the system on cluster environment We successfully validated bInSiGHT characterizing the integration profi le on 3 patients enrolled in a lentiviral vector-based hematopoietic stem cell gene therapy trial for metachromatic leukodystrophy (MLD) performed in our institute Exploiting both optimized alignment methods and parallelization algorithms, we obtained a speedup of 20x when compared to our previously developed version of data analysis pipeline Moreover, the identifi cation and quantifi cation of genuine integration sites based on clustering of vector integrations without the need of genomic mapping has increased the numbers of integrations that can be tracked during time and between different hematopoietic lineages By the use of bInSiGHT we found that MLD treated patients have a large repertoire of vector integrations suggesting high levels of polyclonal reconstitution Moreover, no signifi cant clonal expansions were detected and no skewing towards CIS or cancer genes during time suggesting lack of genotoxic events

at one year after transplantation

354 The Integration Profi le of Lentiviral Vector

in CD34+ Hematopoietic Stem Cells Is Dependent

on the Cell Cycle Status of the Target Cell

Eleni Papanikolaou,1,2 Ekati Drakopoulou,1,2 Evangelos Stamateris,1,2 Alexandros Polyzos,3 Anna Paruzynski,4 Cynthia Bartholomae,4 Manfred Schmidt,4 Christof von Kalle,4 Nicholas P Anagnou.1,2

1 Laboratory of Biology, University of Athens School of Medicine, Athens, Greece; 2 Laboratory of Cell and Gene Therapy, Biomedical Research Foundation of the Academy of Athens, Athens, Greece; 3 Deparment of Molecular Biology, Biomedical Research Foundation of the Academy of Athens, Athens, Greece;

4 National Center for Tumor Diseases (NCT), German Cancer Research Center (DKFZ), Heidelberg, Germany.

Gene therapy utilizing lentiviral-vectors (LVs) is postulated as a dynamic therapeutic alternative for monogenic diseases However, retroviral gene transfer may activate proto-oncogenes via viral integration, a phenomenon called insertional mutagenesis Although, such risks had been originally estimated as extremely low, the report of leukemia due to insertional activation of the LMO2 gene following gene therapy for X-SCID in a minority of patients, led to a

Molecular Therapy Volume 20, Supplement 1, May 2012

RNA VIRAL VECTORS

re-evaluation of the mechanisms operating in insertional mutagenesis

To this end, we sought to determine the viral integration pattern based

on the cycle stage of the target cell For that reason, cord blood CD34+

cells were transduced with a GFP-LV after synchronization at the G1

phase (G1 population) by addition of thymidine, or after arrest at

the G1/S phase (G1/S population) by addition of aphidicoline or at

mitosis by addition of nocodazole or taxol (MN or MT populations

respectively) Overall, we mapped 197 integration sites (IS), i.e 49 in

the untreated cells, 73 in the G1 population, 12 in G1/S population and

53 and 10 in the MN or MT populations, respectively These results

show that taxol and aphidicoline severely inhibit viral integration

Relative to the integration pattern, the percentage of the “targeted”

cancer-correlated genes was signifi cantly lower in the thymine

population (4% vs 34% in the untreated cells), while we observed

an increased probability (1%) of viral integration in the SPATS2 and

OR13H1 genes These novel data provide the impetus for preclinical

in vitro models for the engineered cells, leading eventually to patients’

safety

355 Generation of a Barcoded Lentiviral

Plasmid Library for Analyzing Clonal Complexity

and Evaluation of Library Complexity Using Next

Generation Sequencing

Claire T Deakin,1 Samantha L Ginn,1 Claus V Hallwirth,1 Ian E

Alexander.1,2

1 Gene Therapy Research Unit, Children’s Medical Research

Institute and Children’s Hospital at Westmead, Australia;

2 Discipline of Paediatrics and Child Health, University of Sydney,

Australia.

Successful clinical trials of gene therapy targeting the hematopoietic

compartment have demonstrated retroviral gene delivery to relatively

small numbers of hematopoietic progenitors can cure a disease

phenotype However, little is known about how many progenitor

clones contribute to differentiation of different hematopoietic lineages

nor the long-term dynamics of differentiation Such knowledge may

help defi ne optimal cell doses in clinical applications Furthermore,

the development of leukemia in fi ve infants treated in gene therapy

trials for X-linked severe combined immunodefi ciency highlighted

the need for sensitive methods to detect malignancies in clinical

applications using integrating vectors Existing methods based on

integration site analysis are associated with biases introduced by the

use of restriction endonucleases leading to preferential amplifi cation

of smaller amplicons The development of barcoded lentiviral vectors

may enable individual cells to be uniquely tagged during transduction

and subsequently monitored within a mixed population in an unbiased

fashion Analyzing such samples requires accurate identifi cation of

large numbers of unique barcodes present at low frequency and in

the order of thousands to millions of combinations Developments

in next generation sequencing (NGS) have opened the potential for

such analysis A barcoded lentiviral plasmid library was generated

by inserting a short sequence containing degenerate nucleotides into

a lentiviral construct containing the elongation factor 1α (EF1α)

promoter and the interleukin-2 receptor common gamma chain (γc)

Preliminary analysis of library complexity was conducted using

an Illumina HiSeq2000, however, analysis was confounded by the

instrument’s high error rate under the conditions used and the absence

of a reference sequence for a complex library where the sequence of

each barcode was unknown Analysis of error rates revealed the most

accurate region of the Illumina read was within the fi rst 35 bases,

such that sequencing of the barcode region, located after position 40,

was impugned by instrument error Furthermore, a 33-base stretch of

homogeneous sequence at the front-end of the read did not favor the

platform’s algorithms for registering individual nucleotide clusters

and calculating phasing and dephasing parameters associated with

error The barcode insert was redesigned in a platform-specifi c manner

to locate the barcode at the front-end of a sequence read and maximize use of the highest quality Illumina sequence data A second barcoded EF1α.γc lentiviral plasmid library specifi c for the Illumina platform was produced and is currently undergoing analysis with the inclusion

of an internal standard curve of barcode complexity Collectively, these measures should allow us to defi ne the utility of the Illumina platform for this type of analysis This study has shown careful design based on an understanding of sequencing technology is required when NGS is used to analyze the complexity of barcoded lentiviral vectors

Glycoprotein Leads to Lack of Correct Processing and Therefore Correct Function of the Envelope

Eleonora Zucchelli,1 Anna Stornaiuolo,1 Sergio Bossi,1 Claudio Bordignon,1 Gian Paolo Rizzardi,1 Chiara Bovolenta.1

1 MolMed S.p.A., Milano, Italy.

MolMed S.p.A has leading expertise in cell and gene therapy product development, with an in-house cGMP facility authorized for the production and release of clinical-grade medicinal products for human use Several years ago we initiated a long-term development plan to generate packaging cell lines for semi-stable and stable production of HIV based lentiviral vector (LV) for clinical application Our strategy is based on the sequential integration of the constitutively

expressed viral genes (gag, pol, rev and the feline endogenous retroviral RD114-TR env) through viral vector transduction

Packaging genes have been inserted by means of an AAV Rep78-mediated baculo-AAV hybrid vector obtaining an intermediate

cell clone, named PK-7, whereas the env gene has been integrated

through SIN-LV-specifi c delivery Although the titer of the

RD114-TR pseudotyped LV resulting upon transient transfection of SIN-LV

is in the range of those reported in literature (1 × 105 TU/ml), and that of LTR-LV upon stable integration is even higher (1 × 106 TU/ ml), we sought to increase LV effi ciency by codon optimization of RD114-TR gene To this aim, we initially codon optimized (co) the entire RD114-TR ORF, RD114-TRco (GENEART, Germany) and fully characterized the new product by assessing its expression and function by Western blot analysis and RD114-TRco pseudotyped LV titer determination, respectively We used two specifi c

anti-RD114-TR Abs, the anti-SU and anti-TM, both recognizing the precursor (PR) besides the surface (SU) and trans-membrane (TM) specifi c subunits In fact, RD114-TR is normally translated as a 564-aa PR, that is processed by the membrane-associated endoprotease furin in 361-aa SU and 203-aa TM subunits The migration of RD114-TR molecules in a SDS-PAGE depend on their glycosylation status, that is, the glycosylated forms of both PR and SU migrate at 75 kDa, whereas, after deglycosylation, the PR at 60 kDa and the SU at 40 kDa, and the glycosylated and deglycosylated TM at ca 18 kDa and

ca 15 kDa, respectively Remarkably, we found that RD114-TRco was transcribed and its PR abundantly translated, but not processed Western blot analysis in fact revealed that in the glycosylated and deglycosylated conditions, respectively, the 75-kDa and 60-kDa bands were normally present, but not the 18-kDa and 15-kDa bands As a consequence the titer of the corresponding RD114-TRco pseudotyped

LV was negative Next, we generated two half-recoded proteins, the RD114-TR-5’co and RD114-TR-3’co, by codon optimizing only the 5’ half and the 3’ half of the ORF, hypothesizing that partial recoding could eliminate the negative effect likely generated by silent mutation(s) introduced during recoding causing either abnormal speed

of the transcription/translation process and/or abnormal folding of the protein, which could eventually prevent furin cleavage Unfortunately, also the two half-optimized envelopes were not functional and their overall maturation was severely impaired In conclusion, our studies suggest that the RD114-TR envelope glycoprotein is not suitable for codon optimization and this strategy cannot be applied to improve its performance