573 high throughput screening identifies effective enhancers of lentiviral transduction

573 High Throughput Screening Identifies Effective Enhancers of Lentiviral Transduction Molecular Therapy Volume 20, Supplement 1, May 2012 Copyright © The American Society of Gene & Cell Therapy S222[.]

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

S222

CELL PROCESSING AND VECTOR MANUFACTURE

Cell Processing and Vector Manufacture

571 Manufacture of Clinical-Grade Lentiviral

Vectors for Ex Vivo Use

Anne Galy,1 Emmanuel Galene,1 Than-Hoa Le,1 Otto Merten,1

Frederic Barnay-Toutain,1 Sabine Charrier,1 Mehdi Gasmi.1

1 Genethon, Evry, France.

Hematopoietic gene therapy has now been tested for over 10

years in pilot phase I/II trials in several inherited diseases such

as primary immune defi ciencies (PID) As vector technology is

evolving, new bioprocesses are needed to manufacture clinical-grade

products Lentiviral vectors (LV) present signifi cant advantages for

gene transfer into hematopoietic stem/progenitor cells, in terms of

biological effi cacy, biosafety and medicinal product characterization

We have developed an advanced generation LV derived from

HIV-1 and pseudotyped with VSVg to treat Wiskott Aldrich syndrome

(WAS), a rare PID characterized by a combined platelet and immune

defect This LV leads to the physiological expression of the WAS

gene in hematopoietic cells To produce the clinical lots we have

developed a scaleable lentiviral manufacturing process that has been

approved for phase I/II clinical trials in the EU (UK, France) and in

the US The LV upstream process involves a 4-plasmid transfection

system in HEK293T cells in Cell Factory stacks Downstream process

involves ion exchange chromatography coupled to ultrafi ltration

and gel fi ltration steps to purify, concentrate and formulate the bulk

vector, which in turn is fi lter-sterilized to generate the product fi lled

for clinical use At present six clinical grade preparations of the

WAS LV have been manufactured and released from our facility

Although the process has not yet been validated, to a large extent

analytical data generated show good reproducibility and robustness

The vector is stable for at least 36 months The process is now adapted

to manufacture another VSVg-pseudotyped LV vector for a different

PID, X-linked chronic granulomatous disease In an effort to improve

residual contaminant profi les of the vector, we have introduced a

modifi cation of the DNA removal step in the process The fi rst GMP

lot of the X-CGD vector has been manufactured Process yield and

analytical data will be discussed in the perspective of regulatory

requirements

572 Effi cient High Titer Production of

Recombinant Lentiviral Vectors for Cancer

Immunotherapy

Bernd Hauck,1 Guang Qu,1 Tiffany Kincaid,1 William Righter,1

John Scholler,2 Olga Zelenaia,1 Anne Chew,2 Junwei Sun,1 Carl

June,2 Katherine A High,1,3 J Fraser Wright.1,2

1 Center for Cellular and Molecular Therapeutics, Children’s

Hospital of Philadelphia, Philadelphia, PA; 2 Pathology and

Laboratory Medicine, University of Pennsylvania Perelman School

of Medicine, Philadelphia, PA; 3 Howard Hughes Medical Research

Institute, Children’s Hospital of Philadelphia, Philadelphia, PA.

Recombinant lentiviral vectors (rLenti) expressing a chimeric

antigen receptor with specifi city for the B-cell antigen CD19 used

to modify autologous T lymphocytes have recently shown great

promise for effective immunotherapy in human clinical trials for

chronic lymphocytic leukemia (Porter et al, NEJM 365:725) One

challenge facing development and large scale implementation of this

therapeutic approach is the ability to manufacture a suffi cient quantity

of clinical grade rLenti We have developed and optimized a clinical

Phase I/II GMP manufacturing process that has provided consistent

high level production and effi cient purifi cation Vector generation is

performed by four plasmid transient transfection of HEK293T cells

in CF5 cell culture units using an optimized scalable methodology

to achieve effi cient DNA transfection A two harvest strategy was

developed following matrix optimization experiments to maximize

rLenti TU recovery from the transfected cells using a strategy that balances vector generation during culture and loss of infectivity of the labile rLenti particles that occurs in a time dependent manner after budding from production cells The harvest strategy involved

<10L harvest volumes, consistent with handling of materials in a BSL2 environment with appropriate safety controls Weekly crude harvest titers ranged from 5.4 - 8.7 E6 TU/mL as measured using SupT1 cells, a human T cell line Each weekly cell culture harvest was subject to tangential fl ow fi ltration for harvest concentration, treatment with nuclease (Benzonase) to digest non vector nucleic acid impurities, and two centrifugation steps performed in series to further concentrate and purify the vector particles rLenti vectors were accumulated in weekly batches, with each batch assessed for p24 titer, TU titer, sterility and endotoxin prior to pooling to provide the Lot of purifi ed vector A representative fi nal Lot of rLenti expressing CD19 chimeric antigen receptor was determined to have a fi nal titer

of 8E8 TU/mL, corresponding to a total of 1.3E11 total TU The weekly process vector yield and quality after batch purifi cation was consistent, ranging from 6.6 - 9.4 E8 TU/mL in weekly batches This rLenti manufacturing process provides suffi cient capacity to effectively support early phase clinical development for large cohort cancer immunotherapy investigational product development

573 High-Throughput Screening Identifi es Effective Enhancers of Lentiviral Transduction

Jennifer M Johnston,1 Dresden Whitehead,1 Christopher B Doering,1 H Trent Spencer.1

1 Pediatrics, Emory University, Atlanta, GA.

Lentiviral vectors containing Factor VIII (FVIII) transgenes consistently have lower titers and transduction effi ciencies compared

to similar vectors containing other transgenes GFP encoding lentiviral vectors are produced with titers ten-fold higher than our lentiviral vectors containing a FVIII transgene In addition, murine hematopoietic stem and progenitor cells, sca-1+ cells, are transduced three times more effectively with a lentiviral vector encoding GFP than a lentiviral vector encoding FVIII To overcome the transduction barrier, a high-throughput screen was performed to identify compounds that enhance lentiviral transduction A library of

1280 pharmacologically active compounds (Sigma’s LOPAC1280 library) initially was screened using K562 cells in the presence of a self-inactivating (SIN) lentiviral vector encoding a GFP transgene The screening protocol was optimized for cell number, culturing conditions, and multiplicity of infection (MOI) in a 384 well format Of the 1280 compounds screened, 20 were identifi ed as possible enhancers of viral transduction, as determined by a dose response increase in fl uorescent intensity The 20 compounds were then individually analyzed utilizing a 10 point dose ranging from 0.1μM to 100μM Among the positive hits with a recognizable dose response were camptothecin and etoposide, which have been shown previously to be enhancers of lentiviral transduction However, as topoisomerase inhibitors they also are known to cause DNA strand breaks A potentially safer compound identifi ed from the screen was phorbol 12-myristate 13-acetate (PMA) PMA, a diester phorbol, resembles diacylglycerol and is able to bypass the signal transduction pathway leading to PKC activation Using K562 cells, an EC50 for PMA was found to be between 0.4nM and 1nM, and GFP expression was increased more than 3-fold at an MOI of 0.5 in the presence

of PMA compared to transduction without PMA In addition, the transduction of a FVIII encoding SIV lentiviral vector was 5-times greater in the presence of PMA The enhancement in viral transduction appeared to correlate with an inhibition of cell division When PMA was added to sca-1+ cells isolated from a hemophilia A mouse, cell division was not impaired and, viral transduction was not enhanced

In contrast, PMA did enhance viral transduction of CD34+ cells isolated from human bone marrow Notably, GFP expression was

Molecular Therapy Volume 20, Supplement 1, May 2012

CELL PROCESSING AND VECTOR MANUFACTURE

increased from 7% in the absence of PMA to greater than 20% in

the presence of 1nM PMA, and cell division was hindered The lack

of enhanced viral transduction with sca-1+ cells may be due to the

different PKC isoform found in murine sca-1+ cells compared to

those in human cells Delineation of the mechanism whereby PMA is

acting, specifi cally focusing on which PKC isoform is most affected

by the presence of PMA, is currently underway

574 Magnetic Colocalization of Viral Vectors

and Target Cells Improves Transduction Effi ciency

in Human Hematopoietic Cells

Constanze Lehmann,1 Sigrid Espenlaub,2 Florian Kreppel,2 Ian C

D Johnston.1

1 Miltenyi Biotec GmbH, Bergisch Gladbach, Germany;

2 Department of Gene Therapy, University of Ulm, Ulm, Germany.

A major factor limiting viral transduction of target cells in

cell culture is the diffusion of virus particles to the cell surface

Polycationic reagents, recombinant fi bronectin, spinoculation or the

use of polycationic magnetic particles and a magnetic fi eld can be

used to increase vector concentration at the cell surface In a further

refi nement of this latter method, we have labeled the target cells with

standard magnetic cell separation reagents (MACS®MicroBeads)

in addition to magnetic labeling of the viral vector particles before

placing these on a magnetic cell separation column Magnetic

labeling of both virus particles and the target cells ensures an

effi cient colocalization within the high-gradient magnetic fi eld of a

MACS Separator leading to reduced viral reagent requirements and

specifi c transduction of only the magnetically-labeled target cells

(Blood: 117: e171–181) In this study, both polycationic magnetic

particles and antibody-conjugated superparamagnetic nanoparticles

(MicroBeads) were assessed for their ability to bind to viral vector

particles and transduce human cell lines and primary hematopoietic

cells Polycationic magnetic particles complexed GFP-encoding

adenoviral (AdV) and lentiviral (LV) vector particles effectively

After AdV transduction (pMOI, physical particles per cell=200-500)

of human cells of low permissivity (K562, M-07e, HuT78), up to

15-fold more cells were transduced after magnetic colocalization with

increases in target gene expression of over 40-fold Similarly, CD34+

cells isolated from peripheral blood with CD34 MicroBeads could

be transduced effi ciently (pMOI=2000, 15% GFP+ cells (3 fold over

control) and 8-fold higher MFI) Transduction of human T cells and

CD34+ cells with LV vectors was also supported using this protocol

Both numbers of transduced cells and the expression intensity of the

transgene were increased (T cells, MOI=0.5, 30% GFP+; CD34+ cells,

MOI=50, 75% GFP+) During LV budding from the cell membrane,

host membrane proteins are also incorporated into the viral membrane

MicroBeads that bind these molecules can be used to isolate

wild-type HIV from patient samples, but can also be used to magnetically

label recombinant LV vectors (effi ciency >90%) and direct them to

a target cell Using this approach, vector particles could be directed

to target cells expressing the same surface molecule An enhanced

transduction was observed (3-fold) when both magnetically labeled

vector and cells were combined in MACS columns as described

above (T cells, MOI=0.5, 30% GFP+; CD34+ cells, MOI=25, 69%

GFP+) These novel transduction reagents and protocols enable a fast,

fl exible and reproducible transduction of target cells to be performed

that is independent of vector titer As current cell therapy and gene

therapy approaches require many manual handling steps between

collecting the patient or donor cell sample and returning the modifi ed

cell product to the patient, these protocols are currently being assessed

for their suitability for incorporation into a functionally closed and

fully automated cell processing device for the manufacture of gene

therapeutic cellular products

for Clinical Applications

Roopa Mucharla,1 Usanarat Anurathapan,2 Natalia Lapteva,3 Ann

M Leen,4 Cliona Rooney,5 Juan F Vera.6

1 Center for Cell and Gene Therapy, Baylor College of Medicine, Houston, TX.

Chimeric antigen receptors (CARs) are artifi cial molecules which can be use to redirect T cell immune response against antigens expressed on the surface of tumor cells Recent encouraging clinical data from our group and others has shown that T cells engineered with these molecules can produce complete clinical response Although promising, most current protocols expand engineered T cells non-specifi cally (using IL2 and OKT3), which often results in a decrease

in transgenic populations over time Additionally, cell expansion using conventional cultureware is complicated and labor intensive, limiting the broader application of this therapy With the purpose of optimizing and streamlining CAR-T cell manufacture we assessed whether cell expansion could be improved by; (i) substituting non-specifi c stimuli with an artifi cial antigen presenting cell (a-APC) expressing cognate antigen, and (ii) culturing cells in a simple and scalable gas permeable culture device (G-Rex) To expand T cells engineered with a CAR targeting the prostate cancer antigen PSCA we fi rst generated an a-APC cell line using K562 cells engineered to express PSCA antigen and different co-stimulatory molecules including CD80, CD86 and

41BB When co-cultured with CAR-PSCA T cells in vitro we found

that a-APCs co-expressing PSCA, CD80 and 41BB in combination were most effective in inducing T cell expansion, with a 1.9 fold increase in cell numbers when compared with CAR T cells cultured in the presence of IL2 alone In addition this culture condition enriched for engineered T cells as illustrated by a 2.4±1.2 fold increase in the frequency of transgenic cells after only 7 days of culture To next assess whether we could simply CAR T cell manufacture we transferred our engineered a-APCs and transgenic T cells to a static G-Rex with a surface area of 600cm2, developed by Wilson Wolf Manufacturing as a closed and GMP-compliant culture system In this fl ask O2 and CO2 are exchanged across a silicone membrane at the base, allowing an increased depth of medium above the cells, which provides more nutrients while waste products are diluted These culture conditions resulted in an increase in the cell output when compared with conventional commercial products such as bags, fl asks, and 24-well tissue culture plates, without increasing the number of cell doublings From an initial seeding density of 1.5E+08

T cells (0.25E+06 cells per cm2) we achieved on average a 110 fold cell expansion in 7 days of culture resulting in the production of 1.5-1.8E+10 CAR-T cells (25-30E+06 T cells per cm2) using only 6L of culture media (without media change) We found that 10ml media/cm2

(volume:surface area) supported maximal cell output and cell viability was maintained at >95% for 3 days However, while cell output could not be improved by adding additional media the use of an additional 5ml/cm2 sustained cell viability at >95% for 5 days This optimized bioprocess is GMP compatible and could be adapted for clinical manufacture of CAR-T cells, decreasing the cost and complexity of this technology and making this therapy more accessible

576 Exploiting the Mechanism of Intron-Splicing in Insect Cells To Produce Viral Vectors Harboring Toxin Genes for Cancer Gene Therapy

Haifeng Chen.1

1 VIROVEK Incorporation, Hayward, CA.

Suicide gene therapy has been widely investigated for the treatment of human immunodefi ciency virus (HIV) infection, for controlling graft-versus-host disease, and also for the treatment of cancer While the production of viral vectors carrying suicide genes such as the herpes simplex virus thymidine kinase (HSV-TK) or