Note : Les descriptions sont présentées dans la langue officielle dans laquelle elles ont été soumises.
CA 02238434 1998-06-08
W O 97/21824 PCT~US96/19904
METE~OD FOR OBTAINING RETROVIRAL VE{~TOR SUPERNATANT
HAVING HIGH TRANSDU~TION EFFICIENCY
This application is a con~inll~tion-in-part of United States patent application Serial
No. 08/572,943, filed December 15, 1995.
s
TFCHNICAT FIEI,D OF ~F D!~VENTION
This invention generally relates to the derivation and use of packaging cell lines for
the production of retroviral transducing supern~ nt
BACKGROUND OF THF INVF,NTION
Human gene transfer involves the transfer of one or more therapeutic genes and the
sequences controlling their expression to ~lol,liate target cells. A number of vector
systems have been developed for the transfer of the therapeutic genes for various clinical
indications. In vivo gene transfer involves the direct ~flmini~tration of vector to the target
cells within a patient. E~c vivo gene transfer entails removing target cells from an
individual, modifying them ex vivo and l~ul~ g the modified cells to the patient.
The majority of gene therapy protocols approved for clinical trials by the NIH
Recombinant DNA Advisory Committee (RAC) have used amphotropic retroviral vectors
(ORDA Reports Recombinant DNA Advisory Committee (RA~ Data Management
Report, June 1994, (1994) Human Gene The~ap~ 5:1295-1302). Retroviral vectors are the
vehicle of choice primarily due to the generally high rate of gene transfer obtained in
experiments with cell lines and the ability to obtain stable integration of the genetic
material, ensuring that the progeny of the modified cell will contain the transferred genetic
material. For a review of retroviral vectors and their use in the kansfer and ~pl~S~,iOn of
foreigngenes,seeGilboa(1988)Adv. Exp. Med. Biol. 241:29;Luskeyetal. (1990)Ann
l~.Y. Acad. Sci. 612:398, and Smith (1992)J. Hematother. 1:155-166.
Many retroviral vectors currently in use are derived from the Moloney murine
leukemia virus (MMLV). In most cases, the viral gag, pol and env sequences are removed
from the virus, allowing for insertion of foreign DNA sequences. Genes encoded by the
foreign DNA are often expressed under the control of the strong viral promoter in the LTR.
Such a construct can be packaged into vector particles efficiently if the gag, pol and env
CA 02238434 1998-06-08
W O 97/21824 PCTAUS96/19904
functions are provided in tra~s by a pachaging cell line. Thus, when the vector construct is
introduced into the p~t~.k~gin~ cell, the Gag-Pol and Env proteins produced by the cell
assemble with the vector RN~ to produce replication-defective or transducing virions that
are secreted into the culture medium. The vector particles thus produced can i~fect and
integrate into the DNA of the target cell, but generally will not produce infectious virus
since it is lacking es.~nti~l viral sequences.
Most of the pRc.k~ging cell lines ~ lly in use have been transfected with
separate plasmids encoding Gag-Pol and Env, so that multiple recombination events are
necessary before a replication-competent retrovirus (RCR) can be produced. Cornmonly
used retroviral vector p~c.k~gin~: cell lines are based on the murine NIH/3T3 cell line and
include PA317 (Miller & Buttimore (1986) Mol. Cell Biol. 6:2895; Miller & Rosman(1989) BioTech~iques 7:980), CRIP (Danos & Mulligan (1988) Proc. Natl ~cad Sci USA
85:6460), and gp + aml2 (Markowitz et al. (1988) Virology 167:400). Although splitting
the gag-pol and env genes within the packaging cell genome decreases the incidence of
R(~R, RCR is occasionally observed in clinical-scale productions of retroviral vector
prc~aldLions and is a major safety concern. This is likely due, at least in part, to the fact
thatNIH/3T3 cells contain endogenous MLV sequences (Irving et al. (1993) Bio/Technol.
11:1042-1046) which could participate in recombination to form RCR (Cosset et al. (1993)
Yirology 193:385-395 and Vanin et al. (1994) J. I~irology 68:4241 -4250), particularly in
mass culture during large-scale clinical vector production.
The range of host cells that may be infected by a retrovirus or transduced by a
retroviral vector is determined by the viral Env protein. The recombinant virus can be used
to infect virtually any cell type recognized by the Env protein provided by the packaging
cell, resulting in the integration of the viral genome in the transduced cell and the stable
production of the foreign gene product. The efficiency of infection is also related to the
level of expression of the receptor on the target cell. In gcneral, murine ecotropic Env of
MMLV allows infection of rodent cells, whereas amphotropic Env allows infection of
rodent, avian and some primate cells, including human cells. Xenotropic vector systems
ntili7inp murine xenotropic Env would also allow transduction of human cells.
The host range of retroviral vectors has been altered by substituting the Env protein
of the base virus with that of a second virus. The resulting, "pseudotyped" vector particle
has the host range of the virus donating the envelope protein and expressed by the
-
CA 02238434 l998-06-08
W O 97/21824 PCT~US96/19904
p~.k~in~ cell line. For example, the G-glycoprotein from vesicular stomatitis virus
(VSV-G) has been substituted for the MMLV Env protein, thereby bro;~ nin~ the host
range. ~ee, e.g., Burns et al. (1993) Proc. Natl. Acad. Sci USA 90.8033-8037 andInternational PCT patent application Publication No. WO 92/14829.
Incon~i~t(?nt results and inefficient gene transfer to some target cell types are two
additional problems associated with current retroviral vector systems. For example,
hematopoietic stem cells are an attractive target cell type for gene therapy because of their
self-renewal capacity and their ability to differentiate into all hematopoietic lineages,
thereby repopulating a patient with the modified cells. Yet retroviral gene transfer into
hematopoietic stem cells has been inconsistent and disappointingly inefficient. Kantoff et
al. (1987) J. E7cp. Med. 166:219-234; Miller, A.D. (1990) Blood 76:271-278, and Xu et al.
(1994) J. Virol. 68:7634. Efforts to increase gene transfer efficiency include producing
higher end-point-titer retroviral vector supernatants. End-point titer is a measure of the
number of functional vector particles in a preparation which, when increased, should
theoretically increase transduction efficiency by increasing the ratio of functional vector to
target cells, i.e. increasing multiplicity of infection (m.o.i.). Despite increased end-point
titers, however, retroviral gene transfer efficiency (transduction efficiency) has not
increased correspondingly (Xu et al. (1994), supra, Paul (1993) Hum. ~ene
Therapy 4:609-615; Fraes-Lutz et al. (1994) 22:857-865).
Efforts to increase end-point titer have included improving production of retroviral
vector supernatants (see Kotani et al. (1994) Human Gene Therapy, 5:19-28) and physical
concentration of vector particles by ultrafilkation (Paul, et al. (1993), supra. and Kotani, et
al. (1994) supra). It was shown that incubation of producer cells at 32~C rather than at
37~C yielded sup~rn~t~nt~ with higher end-point titers, but transduction efficiencies were
not compared (See Kotani, et al. (1994) supra). The authors of Kotani et al. (1994) supra,
postulated that the higher titers were due to a lower rate of inactivation combined with a
faster rate of virion production at 32~C. In another study, transduction efficiency was
measured before and after concentration of three supernatants with similar end-point titers
Paul et al. (1993) supra. In each case, concentration increased end-point titer and modestly
improved the transduction efficiency. However, the transduction efficiency achieved with
one of the unconcentrated supernatants was significantly higher than that achieved with the
other concentrates (Paul et al. (1993) supra).
CA 02238434 1998-06-08
W O 97~1824 PCT~US96/19904
For in vivo gene therapy applications, it is important that the retroviral vector not be
inactivated by human serum before transducing the target cells. Reports show that human
serum inactivates a number of recombinant retroviruses, apparently via a complement
pathway. ~3oth viral envelope and producer cell components have been reported to be
5 responsible for viral sensitivity to human complement (Takeuchi et al. (1994) J. Virol.
68(12):8001).
Thus, a need exists for methods of reproducibly increasing transduction
efficiency and for providing stable, safe pa~k~ging cell lines for producing high
transduction efficiency retroviral preparations. This invention satisfies these needs and
10 provides related advantages as well.
SUMMARY OF THF ~NVFNTION
This invention provides, in~er alia, a method for obtaining a recombinant retroviral
packaging cell capable of producing retroviral vectors. In one embodiment, the method
comprises the steps of selecting a retrovirus and obtaining a cell free of endogenous
retroviral nucleic acid. These steps are interchangeably performed. However, after
selection of the retrovirus, a min;m~l gag-pol open reading frame (ORF) insert is isolated
from the retrovirus. Alternatively, a nucleic acid molecule coding for functionally
equivalent retroviral minim~l ORF can be isolated and used in the methods disclosed
herein. In the same manner, a minim~l env ORF is isolated from wild type retrovirus or an
equivalent nucleic acid molecule is obtained. The minim~l ORF nucleic acid molecules
are then amplified, either by insertion into a suitable replication vector or plasmid and
replication of a host cell co~ ; l l;l lp the vector and/or plasmid or by other non-biological
methods (PCR). After amplification, and consistent with the method of amplification, the
minim~l ORF nucleic acid molecules are inserted into a cell pre-selected to be devoid o~
endogenous retroviral nucleic acid. The transformed cells are then propagated under
conditions favorable for expression of the min;m~l retroviral gag-pol and em~ ORF.
Suitable candidate packz~ginP cell lines include, but are not limited to mzlmm~ n
cells such as COS, Vero, HT-1080, Dl7 MRC-5, FS-4, TE671, human embryonic kidney(293), and HeLa.
CA 02238434 1998-06-08
W O 97/21824 PCT~US96/19904
The invention also provides the use of an ELISA method to screen for production
of retroviral structural proteins to identify a retroviral packaging cell capable of producing
recombinant, tr~n~ cing retroviral vector particles. In further embodiment of this
method, the cells that produce high levels of the retroviral Gag-Pol protein and the
retroviral Env protein are identified or selected for by assaying for gag-pol and/or env ORF
translation products in the supernatant.
This invention also provides the recombinant p~ck~gin~ cell obtained by the
methods in all various embodiments described herein, including the amphotropic cell line
de~igns-ted ProPak-A and the xenotropic p~ck~ging cell line designated ProPak-X. Other
10 embodiments of the packaging cell lines produced b~v the methods disclosed herein are
packaging cell lines characterized by having the ability to produce transducing supern~t~nt
that is: resistant to human complement; has a transducing efficiency of greater than or
equal to 5Q% when assayed on NIH/3T3 cells, or greater than that achieved with
supernatant from PA3 1 7-based cells, or has a transducing efficiency of greater than or
equal to 20% when assayed on 293 cells; and substantially free of RCR after interaction of
retroviral vector sequences and continuous culture of more than 2 weeks with an indicator
cell line.
Further provided by this invention is a method of producing recombinant retroviral
particles obtained by introducing into the p~ck~ging cells obtained according to the
20 methods disclosed herein, a recombinant retroviral vector and pro~g~ting the resulting
producer cells under conditions favorable for the production and secretion of retrovira~
vector supernatant.
With respect to the propagation of the producer cells, methods are provided for
obtaining a retroviral vector supernatant having high transduction efficiency by culturing
the producer cells in a packed-bed bioreactor and harvesting the resultant supernatant. In
one method, the packed-bed bioreactor has a surface to volume ratio of 5 to 50 cm2/ml, and
the producer cells are cultured at a temperature of 30~C to about 37~C
The retroviral supern~t~nt~ produced by these methods are also claimed herein.
This invention also provides a method for screening retroviral vector supernatant
30 for high transduction efficiency and methods for producing retroviral vector supernatant
for transducing cells with high efficiency in gene therapy applications.
CA 02238434 1998-06-08
WO 97/21824 PCT~US96/19904
Yet another embodiment of the invention is a method of increasing the transduction
efficiency of a cell by tr~n~(lu-~inp a cell with a retroviral vector supernatant obtained from
the culture of one or more recombinant p~( k:lging cells produced by the methods of the
invention. Vector supernatant cont~inin~ particles of more than one tropism can be
5 produced by the co-culture of two or more producer cells of complementary kopisms.
The present invention provides a method of producing viral vector sup~rn~t~nt ofhigh k~n~ ction efficiency. The prevailing view in the art is that increasing the viral titer
will result in an increase in the kansduction efficiency. As shown herein, this is not the
case. Rather, higher transduction efficiency is correlated with increased numbers of
10 functional vector particles. Vector particles are not all produced identically and high titer
supernatants can contain many non-transducing particles.
nF~CRIPTION OF THF FIGURFS
Figure 1 schematically shows the plasmid constructs for expression of MLV structural
genes, that were used as insertion plasmids for construction of the p~ck~in~ cell lines
ProPak -A and Pro-Pak-X. The top three constructs are for expression of Gag-Pol and the
bottom three are Env expression constructs. Minimal MLV sequences are a common
20 feature among these plasmids. CMV-IE denotes cytomegalovirus immediate early
promoter. SD/SA denotes splice donor/splice acceptor site. MLV is the murine leukemia
viral promoter present in the viral LTR. RSV LTR is the LTR promoter of Rous Sarcoma
Virus. SV40 denotes the simian virus 40 early promoter. pA is the poly-adenylation site.
Ea refers to the amphotropic envelope gene, Ex refers to the xenotropic envelope gene, and
25 Eax denotes the chimeric ampho/xenotropic envelope gene.
Figure 2 shows the structure of several of the retroviral vectors used in this study. The
name of each vector is indicated on the left. Open boxes represent the MLV LTR
elements; RevM 10, trans-dominant mutant of the HIV Rev protein, TK, Herpes Simplex
30 virus-1 thymidine kinase promoter; Neo, neomycin phosphokansferase; SV, simian virus
40 early promoter; nls, nuclear localization signal; IRES, internal ribosomal entry site
element; NGFR, nerve growth factor receptor, puro, puromycin resistancc gene; CMV,
CA 02238434 1998-06-08
W O 97/21824 PCT~US96/19904
CMV-IE promoter. LMiLy and LLySN encode the Lyt2 surface antigen expressed either
directly from the retroviral I,TR promoter in LLySN, or via an IRES element ~ires) in
LMiLy. LLySN contains the neomycin phosphotransferase gene (Neo) expressed from an
intern~l SV40 promoter (SV).
Figure 3 shows the result of exposing to human serum, supernatants cont~inin~ lacZ-
encoding vector prepared from stable producer cells (PA317; PES01), by transienttransfection of vector into p~kiqgin~ cells (ProPak-A) or by co-transfection of p~ .k~gin~
and vector constructs into 293 cells (293). Supenn~ts~nt~ were mixed with an equal volume
10 of a pool of human serum from 4 healthy donors, incubated for 1 hour at 37~C, and the
residual titer determined on NIHt3T3. The serum was either untreated (HS) or had been
heat-inactivated for 30 min at 56~C (HI-HS). The human serum pool had a hemolytic titer
(CH50; E~Z Complement Assay, Diamedix, Miami, Florida) of 117 to 244 before, and < 8
after heat-inactivation. End-point titers (cfu x 10-5/ml) of supernz~t~nts treated with heat-
inactivated serum (100 %) were: PA317, 5.0; ProP~-A, 1.0; PE501,1.~, and 293, 1.1.
The bars indicate the range for duplicate samples. HI-HS denotes heat inactivated hurnan
serum. HS denotes human serum.
~igures 4A, 4B and 4C show transduction efficiencies of vector supernatants. Figure 4A
20 is a comparison of the transduction efficiency of Lyt-2-encoding (LMiLy vector) viral
supern:~t~nt~ produced from ProPak-A and PA317 cells cultured in a packed-bed
bioreactor, as assayed on NIH/3T3 cells. Figure 4B is a comparison of the transduction
efficiency of vector supernat~nt~ ~LMiLy vector) produced from ProPak-A and PA 317
cells cultured in an aerated packed-bed bioreactor, as assayed on 293 cells. Figure 4C is a
25 comparison of the transduction efficiencies achieved with Lyt2-encoding (LLySN) vector
supernzt~nf~ produced from PA317 or ProPak-A-based producer cells, as the proportion
(%) of NIH/3T3 cells that stained with anti-Lyt2 alltibody (Pharrningen, San Diego, CA) 2
days after inoculation of the NIH/3T3 cells with the dilutions of vector supernatant shown.
Supern:~f~nf~ were prepared from confluent producer cell cultures after culturing for 12
30 hours at 32~C.
CA 02238434 1998-06-08
w o 97nl824 PCT~US96/19904
Figures 5A and 5B show the results of comparison of transduction mediated by various
viral vector supernatants. 5A is a comparison of end-point titer and transduction efficiency
for PA.SVNLZ supernatants measured on NIH/3T3 cells. Figure 5B is a comparison of
transduction efficiencies measured on NIH/3T3, HeLa or Jurkat cells for three PA.LMTNL
S supen1~t~nt~ with the end-point titer measured on NIH/3T3 cells. 1 E + 05 means an end-
point titer of IxlOs cfu/ml on NIH/3T3 cells. Error bars show the standard error for three
det~ nnin~tions.
Figure 6 sho~,vs the transduction efficiency of PA.SVNLZ sUp~rn~t~nt~ before or after
10 ~ concentration. The original supen~:~t~nt and the concentrate (see Example 2 and Table 4)
were diluted with medium and inoculated onto NIH/3T3 cells.
Figures 7A and 7L show measurement of tr~n~cluçtion efficiency (Figure 7A) or end-
point titer of PA.SVNLZ rekoviral vector supern~t~nt on NIH/3T3 cells (Figure 7B),
showing the inactivation of vector incubated at 37~C, 32~C or 0~C.
Figures 8A and 81B show the time-course of production of PA.SVNLZ retroviral vector in
a 75 cm2 tissue culture flask at 32~C or 37~C measured as either transduction efficiency
(Figure 8A) or end-point titer on NIH/3T3 cells (Figure 8B) or.
Figure 9 is a time-course of P~.LMTNL vector production from a connuent roller bottle
culture at 32~C as measured by transduction efficiency or end-point titer.
Figures lOA and lOB show a time-course of vector production for PA.SVNLZ producer
cultures in a roller bottle (900 cm2) or packed bed-bioreactor (12,000 cm2) shown by
transduction efficiency (Figure lOA) or end-point titer (Figure lOB). lE + 05 means an
end-point titer of lx105 cfi~/ml on NIH/3T3 cells.
Figure 11 illustrates a bioreactor that is suitable for the production of retroviral vector
30 = supern~t~nt The numbers indicate the following: 1: medium inlet from 10 L vessel; 2: air
inlet medipure air/S% CO2; 3: Fibercell discs lOg; 4: agitation rate 80 rpm see-~in~;, 250
rpm production: 5: medium outlet: to 2 L vessels on ice; and 6: supernatant samples.
CA 02238434 1998-06-08
WO 97/21824 PCT~US96/19904
Figure 12 illustrates the set up of a continuous perfusion packed-bed bioreactor that is
suitable for the production of retroviral vector supernatant. The parts of the set-up are
indicate by nurnbers as follows: 1, peristaltic pump; 2, medium feed tank (10 L); 3, packed
bed bioreactor (500 rnl), 4, magnetic stirrer; 5, air/CO2 mix; 6, 2 L vessels for daily
supernatant collection m~in~in~-l at 0~C. Incubator temperatures are at 37~C for growth
of the producer cells and at 32~-37~C for vector production.
Figure 13 show the results of the comparison of vector production from ProPak-A.855
producer cells in T -flasks versus aerated packed-bed bioreactor, and as measured on 293
cells. Time (h) is in hours. Transduction efficiency is presented as the percentage of cells
expressing the vector-encoded surface marker protein. See Figure 2 for structure of vector
PG855.
Figure 14 shows transduction of 293 cells with ProPak-X-derived supernatants harvested
at different times post seeding of cells in either a T-flask or the packed-bed bioreactor.
Figure 15 is a comparison of transduction efficiency achieved with supern~t~nts harvested
from ProPak-A.LMiLy cells cultured in different vessels, as shown.
Figures 16A and 161~ shows transduction of bone marrow (CD34 ) with LMiLy packed-
bed supernatants. Figure 16A shows transduction after spinoculation with single
supernatants. Figure 16B shows transduction after spinoculation with single or mixed
(ampho + xeno) supernatant.
Figures 17Aand 17B show transduction of cell lines with Lyt2-encoding vector
plepalalions with various tropisms. Average values for duplicate samples are plotted. In
Figure 17A, cell lines were inoculated at unit gravity with LLySN vector from ProPak-X
(PP-X), ProPak-A (PP-A), or PG13 (PG)-based producer populations. MLV(V-G). LMiLy
is an MLV(V~V-G) pseudotype supernatant prepared by transient transfection. In Figure
17B, Quail cells (Qcl. 3; Cullen et al., 1983) were inoculated with LMiLy vector particles
CA 02238434 1998-06-08
W O 97/21824 PCT~US96/19904
bearing the chimeric envelope (Eax; prepared by transient transfection), the xenotropic
(ProPak-X) or the amphotropic (ProPak-A) envelope.
Figure 18. Complement resistance of vector packaged in ProPak-X, ProPak-A, or PA317
S packaging cell lines. Supern~t~nt~ were incubated for 30 min at 37~C with an equal
volume of human serum. The residual transducing activity relative to samples incubated in
medium was determined on 293 (ProPak-X) or NIH/3T3 cells (ProPak-A, PA317). The
human serum, a pool from 4 healthy donors, had a hemolyt;c titer (CH50) of 150 to 313
(EZ Complement Assay, Diamedix, Miami, FL). Bars represent the range for duplicate
10 samples.
Figure 19 shows a comparison of vector production methods. ProPak-A.52.LMiLy
supernzl~nt~ were harvested one day after confluence (visual monolayer in T-flask and
roller bottle cultures), diluted 1 in 8 and inoculated onto NIH/3T3 cells. Lyt2 expression
15 was analyzed 3 days later by FACS.
Figure 20 shows phenotypic analysis 3 days after spinoculation of CD34-positive cells.
Cells were stained with anti-Lyt2-phyco~ in antibody, and either anti-CD34-sulphur
rhodamine (panels A and B) or a panel of fluorescein isothiocyanate-labeled, lineage-
20 specific antibodies (panels C and D). The panel of lineage-specific antibodies contained
antibodies to the following hematopoietic cell lineages: thymocytes (CD2); granulocytcs,
monocytes and macrophages (CD14, CD15); natural killer cells (CD16); B lymphocytes
(CDl9); and erythrocytes (glycophorin). Samples were analyzed on a Becton-Dickinson
Vantage flow cytometer. Panels A and C are from uninoculated, and panels B and D from
25 inoculated cells. The values represent the proportion of the cell population in the
respective quadrants.
CA 02238434 1998-06-08
W O 97/21824 11 PCT~US96/19904
MOPF(S) FOR CARRYING OUT THE I~VENTION
Definition~
Unless otherwise specified herein, common definitions are intended by the words
and terrns used herein. For example, "retrovirus" denotes a class of viruses which use
5 RNA-directed DNA polymerase, or "reverse transcriptase" to copy a viral RNA genome
into a double-stranded DNA intermediate which can be incorporated into chromosomal
DNA (a "provirus") of an avian or m~mmz~ n host cell. Retrovirus also exist as free
virions, that contain the structural and enzymatic proteins of the retrovirus (including
reverse transcriptase), two copies of the viral genome, and portions of the host cell's
10 plasma membrane in which is embedded the viral envelope glycoprotein. Many such
retroviruses are known to those skilled in the art and are described, for example, in Weiss
et al., eds, ~NATIlmor Viruses~ 2nd ed., Cold Spring Harbor Press, Cold Spring ~arbor,
New York (1984 and l9g5). Plasmids cont:~ining retroviral genomes also are widely
available from the American Type Culture Collection (ATCC3, and other sources as15 described in Gacesa and Ramji~ Vectors: F~ntial 7 )ata. John Wiley & Sons, New York
~1994). The nucleic acid sequences of a large number of these viruses are kno~,vn and are
general3y available from databases such as GENBANK, for example. The complete
nucleic acid sequence of the MoMLV and other MLVs is known in the art.
"P~ckz~ging cell line" is a recombinant cell line cont~ining nucleic acid expressing
~0 retroviral Gag, Pol and Env structural proteins. Because the p7~ck:~ging cell line lacks the
retroviral nucleic acid coding for packaging signal and other cis-acting elements, infectious
virions cannot be produced.
A "producer cell" is a packaging cell as defined above which also contains a
replication-defective retroviral vector which is packaged into the vector particle. The
25 producer cell produces transducing retroviral-based particles cont~;ning "foreign" (i.e.,
non-retroviral) genes, such as therapeutic or marker genes.
A "target cell" is a cell to be transduced with a recombinant retroviral vector.Thus, target cells may be, for example, a cell line used for assessing the quality of a
retroviral vector ~l~p~ ion. a primary cell for genetic modification ex vivo, or a cell
30 within a patient that will be modified by in vivo introduction of a retroviral vector.
CA 02238434 1998-06-08
W O 97/21824 12 PCT~US96/19904
The terms "polynucleotide", "oligonucleotide", "nucleic acids" and "nucleic acidmolecules" are used interchangeably, and refer to a polymeric form of nucleotides of any
length, either deoxyribonucleotides or ribonucleotides, or analogs thereof. Polynucleotides
may have any three-dimensional structure, and may perforrn any function, known or
S unknown. The following are non-limiting examples of polynucleotides: a gene or gene
fr~nent, exons, introns, messenger RNA (mRNA), transfer RNA, ribosomal RNA,
ribozymes, cDNA, recombinant polynucleotides, branched polynucleotides, pl~mi~ls7
vectors, isolated DNA of any sequence, isolated RNA of any sequence, nucleic acid
probes, and primers. A polynucleotide may comprise modified nucleotides, such asmethylated nucleotides and nucleotide analogs. If present, modifications to the nucleotide
structure may be imparted before or after assembly of the polymer. The sequence of
nucleotides may be interrupted by non-nucleotide components. A polynucleotide may be
further modified after polymerization, such as by conjugation with a labeling component.
The term polynucleotide, as used herein, refers interchangeably to double- and
single-stranded molecules. Unless otherwise specified or required, any embodiment of the
invention described herein that is a polynucleotide encompasses both the double-stranded
form and each of two complementary single-stranded forrns known or predicted to make
up the double-stranded form.
A "gene" can refer to a polynucleotide or a portion of a polynucleotide comprising
a sequence that encodes a protein. It is often desirable for the gene also to comprise a
promoter operatively linked to the coding sequence in order to effectively promote
transcription. Enhancers, repressors and other regulatory sequences also can be included in
order to modulate activity of the gene, as is well known in the art (see, e.g., the references
cited herein).
A "detectable marker" gene is a gene that allows cells carrying the gene to be
specifically detected (i.e., to be distinguished from cells which do not carry the marker
gene). A large variety of such marker genes are known in the art. Preferred examples of
such marker genes encode proteins appearing on cellular surfaces, thereby facilit~ting
simplified and rapid detection and/or cellular sorting.
A "selectable marker" gene is a gene that allows cells carrying the gene to be
specifically selected for or against, in the presence of a corresponding selective agent. By
way of illustration, an antibiotic resistance gene can be used as a positive selectable marker
-
CA 02238434 1998-06-08
W O 97/21824 13 PCT~US96/19904
gene that allows a host cell to be positively selected for in the presence of the
corresponding antibiotic. A variety of positive and negative selectable markers are known
in the art, some of which are described herein.
In the context of polynucleotides, a "linear sequence" or a "sequence" is an order
S of nucleotides in a polynucleotide in a 5' to 3' direction in which residues that neighbor
each other in the sequence are contiguous in the primary structure of the polynucleotide. A
"partial sequence" is a linear sequence of part of a polynucleotide which is known to
comprise additional residues in one or both directions.
"Hybridization" refers to a reaction in which one or more polynucleotides react to
form a complex that is stabilized via hydrogen bonding between the bases of the nucleotide
residues. The hydrogen bonding may occur by Watson-Crick base pairing, Hoogsteenbinding, or in any other sequence-specific manner. The complex may comprise two
strands forming a duplex structure, three or more strands forrning a multi-stranded
complex, a single self-hybridizing strand, or any combination of these. A hybridization
reaction may constitute a step in a more extensive process, such as the initiation of a PCR,
or the enzymatic cleavage of a polynucleotide by a ribozyme.
Hybridization reactions can be performed under conditions of dirrerellt
"stringency". Conditions that increase the stringency of a hybridization reaction are widely
known and published in the art: See, for example. Sambrook et al. (1989) infia
"Tm" is the temperature in degrees Centigrade at which 50% of a polynucleotide
duplex made of complementary strands hydrogen bonded in an antiparallel direction by
Watson-Crick base paring dissociates into single strands under the conditions of the
experiment. Tm may be predicted according to standard formula; for example:
Tm= 81.5 + 16.6 log [Na ~ + 0.41 (%G/C) - 0.61 (%F3 - 600/L
where Na+ is the cation concentration (usually sodium ion) in mol/L; (%G/C) is the
number of G and C residues as a percentage of total residues in the duplex; (%F) is the
percent form~ in solution (wt/vol), and L is the number of nucleotides in each strand
of the duplex.
A linear sequence of nucleotides is "identical" to another linear sequence, if the
order of nucleotides in each sequence is the same. and occurs without substitution,
deletion, or material substitution. It is understood that purine and pyrimidine nitrogenous
bases with similar structures can be functionally ec~uivalent in terms of Watson-Cricl~
CA 02238434 1998-06-08
WO 97/21824 PCT~US96/19904 14
base-pairing; and the inter-substitution of like nitrogenous bases, particularly uracil and
thymine, or the modification of nitrogenous bases, such as by methylation, does not
constitute a m~t~ri~l substitution. An RNA and a DNA polynucleotide have identical
sequences when the sequence for the RNA reflects the order of nitrogenous bases in the
polyribonucleotide, the sequence for the DNA reflects the order of nitrogenous bases in the
polydeoxyribonucleotide, and the two sequences satisfy the other requirements of this
definition. Where at least one of the sequences is a degenerate oligonucleotide comprising
an ambiguous residue, the two sequences are identical if at least one of the alternative
forms of the degenerate oligonucleotide is identical to the sequence with which it is being
compared. For example, AYAAA is identical to ATAAA, if AYAAA is a mixture of
ATAAA and ACAAA.
When comparison is made between polynucleotides, it is implicitly understood that
complementzlry strands are easily generated, and the sense or antisense strand is selected or
predicted that m~xi~ S the degree of identity between the polynucleotides being
compared. For example, where one or both of the polynucleotides being compared is
double-stranded, the sequences are identical if one strand of the first polynucleotide is
identical with one strand of the second polynucleotide. Similarly, when a polynucleotide
probe is described as identical to its target, it is understood that it is the compl~ment~ry
strand of the target that participates in the hybridization reaction between the probe and the
target.
A linear sequence of nucleotides is "ec~enti~lly identical" to another linear
sequence, if both sequences are capable of hybridizing to form duplexes with the same
complementary polynucleotide. Sequences that hybridize under conditions of greater
stringency are more preferred. It is understood that hybridization reactions canaccommodate insertions, deletions, and substitutions in the nucleotide sequence. Thus,
linear sequences of nucleotides can be essentially identical even if some of the nucleotide
residues do not precisely correspond or align. Sequences that correspond or align more
closely to the invention disclosed herein are comparably more preferred. Generally, a
polynucleotide region of about 25 residues is essentially identical to another region, if the
sequences are at least about 80% identical; more preferably, they are at least about 90%
identical; more preferably~ they are at least about 95% identical, still more preferably, the
sequences are 100% identical. ~ polynucleotide region of 40 residues or more will be
CA 02238434 1998-06-08
W O 97/21824 PCT~US96/19904
essenti~lly identical to another region, after alignrnent of homologous portions if the
sequences are at least about 75% identical; more preferably, they are at least about 80%
identical; more preferably, they are at least about 85% identical; even more preferably,
they are at least about 90% identical; still more preferably, the sequences are 100%
identical.
In dett~.rrninin~ whether polynucleotide sequences are essentially identical, a
sequence that preserves the functionality of the polynucleotide with which it is being
compared is particularly ~ r~ d. Functionality can be ~let~rmine~l by ~ nt
parameters. For example, if the polynucleotide is to be used in reactions that involve
hybridizing with another polynucleotide, then plcrt~ d sequences are those whichhybridize to the same target under similar conditions. In general, the Tm of a DNA duplex
decreases by about 1 0~C for every 1% decrease in sequence identity for duplexes of 200 or
more residues; or by about 50~C for duplexes of less than 40 residues, depending on the
position of the mi~ms~tched residues. Essentially identical sequences of about 100 residues
will generally form a stable duplex with each other"s respective complementary sequence
at about 20~C less than Tm; preferably, they will form a stable duplex at about 15~C less;
more preferably, they will form a stable duplex at about 1 0~C less; even more preferably,
they will form a stable duplex at about 50~C less; still more preferably, they will form a
stable duplex at about Tm. In another example, if the polypeptide encoded by thepolynucleotide is an important part of its functionality, then preferred sequences are those
which encode identical or essenti~lly identical polypeptides. Thus, nucleotide differences
which cause a conservative amino acid substitution are preferred over those which cause a
non-conservative substitution, nucleotide differences which do not alter the amino acid
sequence are more ~l~f~ d, while identical nucleotides are even more preferred.
Insertions or deletions in the polynucleotide that result in insertions or deletions in the
polypeptide are preferred over those that result in the down-stream coding region being
rendered out of phase; polynucleotide sequences comprising no insertions or deletions are
even more preferred. The relative importance of hybridization properties and the encoded
polypeptide sequence of a polynucleotide depends on the application of the invention.
A polynucleotide has the same "characteristics" of another polynucleotide if both
are capable of forming a stable duplex with a particular third polynucleotide under similar
CA 02238434 1998-06-08
W O 97/21824 16 PCT~US96/19904
conditions of m~xim~l stringency. Preferably, in addition to similar hybridization
properties, the polynucleotides also encode çcs~ntizllly identical polypeptides."Conserved" residues of a polynucleotide sequence are those residues which occurunaltered in the same position of two or more related sequences being compared. p~e~idues
S that are relatively conserved are those that are conserved zlmonp~t more related sequences
than residues ~pe~il~g elsewhere in the sequences.
"Related" polynucleotides are polynucleotides that share a significant proportion of
identical residues.
A "probe" when used in the context of polynucleotide manipulation refers to an
10 oligonucleotide which is provided as a reagent to detect a target potentially present in a
sample of interest by hybridizing with the target. Usually, a probe will comprise a label or
a means by which a label can be attached, either before or subsequent to the hybridization
reaction. Suitable labels include, but are not limited to radioisotopes, fluorochromes,
chemilnminescent compounds, dyes, and proteins, including enzymes.
A "primer" is an oligonucleotide, generally with a free 3' -OH group, that binds to
a target potentially present in a sample of interest by hybridizing with the target, and
thereafter promotes polymerization of a polynucleotide complementary to the target.
Processes of producing replicate copies of the same polynucleotide, such as PCR or
gene cloning, are collectively referred to herein as "arnplification" or "replication". For
20 example, single or double-stranded DNA may be replicated to form another DNA with the
same sequence. RNA may be replicated, for example, by an RNA-directed RNA
polymerase, or by reverse-transcribing the DNA and then perforrning a PCR. In the latter
case, the amplified copy of the RNA is a DNA with the identical sequence.
A "polymerase chain reaction" ("PCR") is a reaction in which replicate copies are
25 made of a target polynucleotide using one or more primers. and a catalyst of
polymerization, such as a reverse transcriptase or a DNA polymerase, and particularly a
thermally stable polymerase enzyme. Generally, a PCR involves reiteratively forming
three steps: "~nne~ling", in which the temperature is adjusted such that oligonucleotide
primers are permitted to form a duplex with the polynucleotide to be amplified,
30 "elongating", in which the temperature is adiusted such that oligonucleotides that have
formed a duplex are elongated with a DNA polymerase, using the polynucleotide to which
they have formed the duplex as a template, and "melting", in which the temperature is
CA 02238434 1998-06-08
W O 97/21824 17 PCTAJS96/19904
adjusted such that the polynucleotide and elongated oligonucleotides dissociate. The cycle
is then repeated until the desired amount of amplified polynucleotide is obtained. Methods
for PCR are taught in U.S. Patent Nos. 4,683,195 (Mullis~ and 4,683,202 (Mullis et al.).
A "control element" or "control sequence" is a nucleotide sequence involved in an
interaction of molecules that contributes to the functional regulation of a polynucleotide,
including replication, duplication, transcription, splicing, translation, or degradation of the
polynucleotide. The regulation may affect the frequency, speed, or specificity of the
process, and may be enhancing or inhibitory in nature. Control elements are known in the
art. For example, a "promoter" is an example of a control element. A promoter is a DNA
region capable under certain conditions of binding RNA polymerase and initiatingtranscription of a coding region located downstream (in the 3' direction) from the
promoter. Retroviral long terminal repeat sequences (LTR) contain strong promoters that
are suitably used in the inventions described herein.
"Operatively linked" refers to a juxtaposition of genetic elements, wherein the
elements are in a relationship permi~ing them to operate in the expected manner. For
instance, a promoter is operatively linked to a coding region if the promoter helps initiate
transcription of the coding sequence. There may be intervening residues between the
promoter and coding region so long as this functional relationship is m~intz-;ned.
The "gag" gene of a retrovirus refers to the 5' gene on retrovirus genomes and is an
abbreviation for group-specific antigens. It is translated to give a precursor polyprotein
which is subsequently cleaved to yield three to five capsid proteins.
The "po~' gene refers to a gene encoding a polymerase. Thus, the pol gene encodes
for a retrovirus reverse transcriptase and also encodes the IN protein needed for viral
integration into cell DNA.
The "env" or envelope region of a retrovirus genome codes for the envelope
proteins. For the purpose of this invention, the "env" gene is to include not only the
naturally occurring env sequence from a virus, but also modifications to the env gene, such
as env genes that are modified to alter target specificity of retrovirus or alternative env
genes that are used to generate "pseudotyped" retrovirus. Preferred env genes for use in
this invention include, but are not limited to amphotropic enV, murine xenotropic env,
Gibbon ~pe Leukemia virus (GaLV) env and the VSV-G protein-encoding gene.
CA 02238434 1998-06-08
WO 97/21824 18 PCT~US96/19904
The terms "polypeptide", "peptide" and "protein" are used interchangeably hereinto refer to polymers of amino acids of any length. The polymer may be linear or branched,
it may comprise modified amino acids, and it may be interrupted by non-amino acids. The
terms also encompass an amino acid polymer that has been modified, for example,
5 fli~ fide bond formation, glycosylation, lipidation, acetylation, phosphorylation, or any
other manipulation, such as conjugation with a labeling component.
The "biochemical function" or "biological activity" of a polypeptide includes any
feature of the polypeptide detectable by suitable experimental investigation. "Altered"
biochemical function can refer to a change in the primary, secondary, tertiary, or
10 quaternary structure of the polypeptide; detectable, for example, by molecular weight
det~rmin~tion, circular dichroism, antibody binding, difference spectroscopy, or nuclear
magnetic resonance. It can also refer to a change in reactivity, such as the ability to
catalyze a certain reaction, or the ability to bind a cofactor, substrate, inhibitor, drug,
hapten, or other polypeptide. A substance may be said to "interfere" with the biochemical
15 function of a polypeptide if it alters the biochemical function of the polypeptide in any of
these ways.
A "fusion polypeptide" is a polypeptide comprising regions in a dirr~lellL position
in the sequence than occurs in nature. The regions may normally exist in separate proteins
and are brought together in the fusion polypeptide; or they may normally exist in the same
20 protein but are placed in a new arrangement in the fusion polypeptide. A fusion
polypeptide may be created, for example, by chemical synthesis, or by creating and
tr~n~l~tin~ a polynucleotide in which the peptide regions are encoded in the desired
relationship.
An "isolated" polynucleotide, polypeptide, protein, antibody, nucleic acid, oracid,or
25 other substance refers to a pl~dlion of the substance devoid of at least some of the other
components that may also be present where the substance or a similar substance naturally
occurs or is a purification technique to enrich it from a source mixture. Enrichment can be
measured on an absolute basis, such as weight per volume of solution, or it can be
measured in relation to a second, potentially interfering substance present in the source
30 = mixture. A substance can also be provided in an isolated state by a process of artificial
assembly, such as by chemical synthesis or recombinant expression.
-
CA 02238434 1998-06-08
W O 97/218~4 PCTrUS96/19904
19
A polynucleotide used in a reaction, such as a probe used in a hybridizat;on
reaction, a primer used in a P~R, or a polynucleotide present in a ph~rm~euticald~ion7 is referred to as "specific" or "selective" if it hybridizes or reacts with the
int~nc~ed target more frequently, more rapidly, or with greatcr duration than it does with
alternative substances. Similarly, a polypeptide is referred to as "specific" or "selective" if
it binds an in~n~led target, such as a ligand, hapten, substrate, antibody, or other
polypeptide more frequently, more rapidly, or with greater duration than it does to
~ltern~tive substances. An antibody is referred to as "specific" or "selective" if it binds via
at least one antigen recognition site to the intended target more frequently, more rapidly, or
with greater duration than it does to alternative substances. A polynucleotide, polypeptide,
or antibody is said to "selectively inhibit" or "selectively interfere with" a reaction if it
inhibits or interferes with the react;on between particular substrates to a greater degree or
for a greater duration than it does with the reaction between alternative substrates.
A "cell line" or "cell culture" denotes higher eukaryotic cells grown or m~int~ined
in vitro. It is understood that the descendants of a cell may not be completely identical
(either morphologically, genotypically, or phenotypically) to the parent cell.
The term "primate" as used herein refers to any member of the highest order of
m~mm~ n species. This includes (but is not limited to) proc imi~n.~, such as lemurs and
lorises; tarsioids, such as tarsiers; new-world monkeys, such as squirrel monkeys (Saimiri
sciureus) and tamarins; old-world monkeys such as macaques (including Macaca
nemestrina, Macaca fascicularis, and Macaca fuscata); hylobatids, such as gibbons and
si~m~ngC; pongids, such as orangutans, gorillas, and ~himr~n7~?~?c~; and hominids, including
hlTm~n~
"Mean residence time" is the average amount of time the culture medium remains
in contact with the producer cells during the production phase. The optimal meanresic1~n~e time is determined by the following considerations: 1) cell specific vector
production rate; 2) rate of vector inactivation; 3) cell specific nutrient uptake rates; 4) cell
specific metabolite production rates; 5) temperature; 6) volumetric cell density; and 7) the
target cells for which the vector supernatant is intended. Given a volumetric producer cell
density of> lx106 celLs/ml, the optimal mean residence time ranges from 3 to 6 hours
with PA317-based cultures; from 6 to 12 hours with ProPak-A-based cultures; and 12 to 24
CA 02238434 1998-06-08
W O 97/21824 PCTnUS96/19904
hours with ProPak-X or PG13-based cultures. These mean residence times are based upon
maximal tr~n~d~lction of cell lines.
nF~C~TPTlO~ C)F F~BODJMENTS
This invention provides a method for obtaining a safe recombinant retroviral
packaging cell capable of producing retroviral-based vectors and retroviral vector
supern~t~nt The method comprises the steps of selecting a retrovirus that will provide the
retroviral env and gag-pol oligonucleotide sequences for the recombinant production of
retroviral env and gag-pol gene products and obtaining a eukaryotic cell free ofendogenous related retroviral nucleic acids or oligonucleotides of the same retroviral type.
Although any retrovirus can be suitably used in the method of this invention, the use of the
murine leukemia virus (MLV) is presently described. Thus, if any of gag, pol or env genes
are to be derived from a MLV, the candidate packaging cell should be screened for absence
of endogenous MLV retroviral nucleic acid and those sequences closely related to MLV
which, by recombination, would produce replication competent retrovirus.
The env gene product determines the target cell specificity of the recombinant
retrovirus particle and will, therefore, be selected to provide optimal transduction of the
target cells of interest. Env may be the MLV amphotropic Env or any envelope capable of
combining to form infectious "pseudotyped" retrovirus particles. For example, MLV
amphotropic and murine xenotropic retroviral vectors are known to transduce human cells.
Other env genes of interest include those from Gibbon Ape Leukemia Virus (GaLV),RD114, FeLV-C, FeLV-B, BLV, and HIV-I . See PCT Publication No. WO 92/14829
(page 25, line 1 through I through page 34, line 1). In addition, the em~ gene can be
modified to more specifically target the recombinant retrovirus to the target cell of interest.
For, example, the Env protein may be modified by combination with an antibody binding
site specific for a cell surface antigen on the target cells of interest, e.g. an anti-CD34
antibody for targeting to hematopoietic stem and progenitor cells (Cossett et al. (1995) J.
Yirol. 69:6314-632'2 and Kasahara, et al. (1994) Science 266:1373-1376). The e~v gene
can also be mod;fied so as to broaden the cell tropism of the virus, for e.g. by constructing
a chimeric env that can bind to both amphotropic and xenotropic receptors or to a unique
receptor on cells. In one embodiment, a chimeric amphotropic/xenotropic envelope gene,
CA 02238434 1998-06-08
WO 97/21824 PCT/lJS96/19904
21
Eax, was constructed as described in Examples 6 and 7. Vector particles with a Eax
envelope displayed dual cell tropism and will therefore transduce a broader range of target
cells.
For clinical gene therapy applications it is important that the retroviral vector
sequences and structural gene sequences be tlesi~ne~l to minimi7~ recombination to form
replication competent retrovirus (RCR)~ By introducing the gag-pol and env gene sequences
into the p~ck~gin~ cell separately so that they integrate in different areas of the p~ gin~
cell genome, the rate of RCR formation is decreased since multiple recombination events are
required to generate RCR. Nevertheless, RCR are sometimes found in recombinant retroviral
vector ~L~dlions. One possible reason is the presence of endogenous retroviral sequences
in the murine cells (NIH/3T3) that the most commonly used pa(~kzlging cell lines are based
on, which may recombine with the introduced retroviral sequences. Therefore, the safe
p:~k~gin~ cells of the present invention are generated from a eukaryotic cell line lacking
endogenous retroviral sequences which would be capable of producing RCR by
recombination with introduced retroviral sequences.
Thus, in one embodiment, the p:~ kzlging cells of the present invention are derived
from a cell line having no detectable endogenous retroviral sequences related to MLV. In
addition, as described in the Examples herein, the cell line is preferably screened for the
ability to stably secrete Gag-Pol and Env proteins and to ef~lciently tr~n~ ce target cells
rather than to produce high end-point titers alone. Preferably the eukaryotic cell line will be a
non-murine cell line, more preferably a primate cell line, and most preferably a hurnan cell
line. The inventors have found that human 293 cells are free of retroviral sequences related
to MLV, and when used as the basis for stable p~k~ginE cells, are able to produce high
transduction efficiency retroviral vectors.
A cell free of related retroviral nucleic acid is obtained by screening a candidate cell
for endogenous retroviral nucleic acid using methods well known to those of skill in the art
and exemplified below. For exarnple, several available cell lines such as the Mus dunni tail
fibroblasts (see Lander and Chattopa&yay (1984) J. Virol. 52:695-698) are reportedly free of
endogenous retroviral nucleic acid and thus, are suitably used in the methods disclosed
herein. Alternatively, one of skill in the art can determine if the ccll line contains
endogenous retroviral nucleic acid by isolating a nucleic acid sarnple from the c~nc~ te cell
line and probing for the endogenous DNA or RNA using methods such as traditional
CA 02238434 1998-06-08
WO 97/21824 22 PCT~US96/19904
Southern and Northern hybridization analysis or the polymerase chain reaction ("PCR"),
using retroviral specific probes, and when available, commercially available PCR kits
(Invitrogen, San Diego, CA). Southern and Northern hybridization analyses are described,
for example in Sarnbrook et al. (1989) inJ;a. PCR methods are described in Gene Fxpression
5 Technolo~y, Goeddel, et al. eds., ~c~-lemic Press, Inc. New York (1991). A cell is free of
endogenous related retroviral nucleic acid if no hybridization is detected, even at low
stringency conditions of 500mM sodium ions, or if the primer used for the PCR analysis does
not provide amplified nucleic acid. Preferably highly conserved sequences sp:~nninp viral
LTR, p~k~ging sequence and gag-pol gene regions are used as probes.
The candidate eukaryotic cell is of any suitable type, i.e., murine, non-murine,m~mm~ n, primate, canine and human, provided that the cell line lacks endogenousretroviral sequences, grows well in culture, can be transfected or transduced with the
aL)~ iate gag-pol and env ~ t;s~ion constructs and can express the viral proteins. The
c~ntlic1zlte cell line is preferably primate, and most preferably human. It has been found that
15 primate and preferably human-based p~ck~gin~ cells can be used to produce retroviral
vectors resistant to human complement.
In addition, the method further comprises using minim~l gag-pol and env sequences
to further decrease the chances of recombination to produce RCR. A minimz~l gag-pol open
reading frame (ORF) and minim~l env 0~ are obtained from the selected retroviruses. The
20 minimz~l ORF of the retroviral sequences are defined to include only those retroviral
sequences from the ATG through the stop codon of the gene with no flzlnking sequences.
Fr~gment~ of the gene as well as biological equivalents thereof also can be used provided that
functional protein is produced when introduced into the c~n~ te p~k~ging cell line. In one
embodiment, isolated retroviral nucleic acid coding for the minim~l gag-pol and env OR~ is
25 selected for use. In a preferred embodiment, the nucleic acid is selected from MLV and the
minim~l sequences are ~lettormined to consist of nucleotides from about 621 to 5837 (gag-
pol) (numbering from Shinnick et al. (1981)) and about nucleotides 37 to 2000 (env)
(numbering from Ott et al. ~1990). These nucleotide positions will vary with different
MLVs. It should be understood, although not always explicitly stated, that nucleic acid
30 sequences or molecules that are "equivalent" are determined to produce the same phenotypic
effect as the isolated minim~l ORF described herein, can be utilized as the minim~l ORF
CA 02238434 l998-06-08
W O 97/21824 23 PCTnJS96/19904
sequences in the methods described herein. For exarnple, altered, but phenotypically
equivalent nucleic acid molecules are referred to as "equivalent nucleic acids".The minimz~l gag-pol and env ORF nucleic acid molecules can be isolated using the
technique described in the experimental section described below or replicated using PCR
5 (Perkin-Elmer) and published sequence inforrnation. For example, the sequence can be
replicated by PCR (Perkin-Elmer) which, in combination with the synthesis of
oligonucleotides, allows easy reproduction of DNA sequences. The PCR technology is the
subject matter of United States Patent Nos. 4,683,195, 4,800,1 5g, 4,754,065, and 4,683,202
and described in PCR: The Polymerase Ch~in Reaction Mullis et al. eds., Birkh~ er Press,
10 Boston (1994) and references cited therein. As is a~t;l~l to those of skill in the art,
modifications and/or additions to the viral sequences are made to facilitate isolation and
ion of the amplified DNA.
It is conceived that arnphotropic and xenotropic cell lines are produced by thismethod and is determin~d by the selection of the env gene. Thus, the selection of the gag-pol
15 and env gene are not restricted to isolation from the same virus, or the same virus type. An
example of an amphotropic-producing packaging cell line produced by this method is
ProPak-A and an example of a xenotropic p~ek~ging cell line, also produced by this method,
is ProPak-X.
Thus, the invention further provides the isolated genes operatively linked to a
promoter of RNA transcription, as well as other regulatory sequences for replication and/or
transient or stable t;x~ ion of the DNA. To minimi7P the chance of RCR it is preferred to
avoid using extraneous viral sequences. The term "operatively linl~ed" is defined above.
The promoter may be a minim:~l MLV-LTR, and is preferably "heterologous" to the
retroviral gene. Suitable promoters are those that drive stable, high-level expression of gag-
pol and env. Examples of suitable promoters include, but are not limited to, thecytomegalovirus imme~ te early (CMV) promoter, lRous Sarcoma Virus ~RSV) LTR,
murine moloney lenkemi~ virus (MMLV) LTR, or other viral LTR sequences. Vectors and
plasmids which contain a promoter or a promoter/enhancer, with tPrmin~tion codons and
selectable marker sequences, as well as a cloning site into which an inserted piece of DNA
30 can be operatively linked to that promoter, are well known in the art and commercially
available. To minimi7.~ the chance of RC~R formation, gag-p~ol and env genes are preferably
incorporated into ~;e~ expression plasmids and the plasmids introduced sequentially into
CA 02238434 1998-06-08
WO 97/218Z4 PCTnJS96/19904
24
the eukaryotic host cells. In plasmid amplification, the bacterial host cells are propagated at
temperatures in the range from about 28~C to about 32~C and more preferably at about 30~C,
to prevent recombination of viral sequences carried on the plasmids and to m~int~in the
integrity of the plasmids.
The separate ~ es~ion plasmids cont:~inin~ the retroviral gag-pol and env genes are
then introduced in separate, sequential steps into the candidate p~r~ginp cells by techniques
well known to those of skill in the art, such as calcium phosphate precipitation,
electroporation and lipofection (Sambrook et al. (1989) supra). The insertion technique also
can involve the use of a modified integrase enzyme that will recogni_e a specific site on the
target cell genome. Such-site specific insertion allows the genes to be inserted at sites on
host cells' DNA that will minimi7r the chances of insertional mutagenesis, minimi7~
interference from other host cellular sequences, and allow insertion of sequences at specific
target sites so as to reduce or elimin~te the expression of undesirable genes.
Introduction of gag-pol and env genes can be done in either order. After insertion
and integration of the gag, pol and env genes, the cells are screened for retroviral gene
ion. If a selectable marker gene, such as antibiotic resistance, was used in
combination with the retroviral sequences, the population can be enriched for cells for
ex~les~il.g the selectable marker (and therefore the retroviral genes) by growing the
candidate cells in the presence of the antibiotic. Cells which survive and propagate will
contain both the antibiotic resistance gene and the retroviral sequences. Still further, an
ELISA with the ~plo~-iate antibody to a product of the expressed sequences will be a
simple and quick assay to determine whether and to what extent the cells contain and express
the retroviral genes.
Thus, the invention also provides an ELISA method to screen for expression of
retroviral genes in c~n~ te packaging cells to identify a retroviral p~rk~gin~ cell capable
of producing recombinant, transducing retroviral vector particles. In particular, the ELISA
will be used to screen for production of retroviral structural proteins such as Gag, Pol and
Env. The ELISA method of the invention can also have diagnostic use in patients who
have undergone gene therapy.
In one specific embodiment, Env protein is detected in a sandwich ELISA assay
using an antibody from hybridoma 83A25 as primary antibody, to capture the protein. The
captured protein is then detected using as secondary antibody, antisera 79S-83~, followed
CA 02238434 1998-06-08
W O 97/21824 25 PCTrUS96/19904
by an enzyme-conjugated antispecies antibody and the substrate for the enzyme. In a
similar manner, Gag is detected separately using as primary antibody, an antibody from
hybridoma R187 followed by antisera 77S-227, enzyme-conjugated antispecies antibody
and enzyme substrate. The primary antibody can be provided in the form of hybridoma
culture supernatant, ascites or purified antibody. The anti-species antibody is preferably
conjugated to a label such as an enzyme. Suitable enzymes include horeseradish
peroxidase and alkaline phosphatase. The a~ pliate substrate to the particular enzyme is
used. This particular embodiment of the ELISA assay is described in detail in the Examples
below.
Preferably, the candidate packaging cell lines are further screened for the ability to
produce vector supernatant having high transduction efficiency. Transduction efficiency is
measured by the ability of the vector supernatant to transduce a target cell population. The
target cell population carl be human 293 cells, NIH/3T3 cell or primary cells.
This invention also provides recombinant p~rk~ging cells obt~ined by the method
15 described above. The recombinant p~k~ginp cell lines are an improvement over prior art
p~- k~ging cell lines because, when tr~n~dll~e-1 with a suitable retroviral vector and
propagated, the cell lines of this invention produce a retroviral titer having high transduction
efficiency. Thus, the cell line of this invention is characterized by producing viral
supern~n~ that is resistant to human complement; has a high tr~n~ ction efflciency; and is
20 substantially free of RCR after continuous culture of more than 2 weeks and up to at least 12
weeks. As used herein, "transduction efficiency" refers to the percentage of the inoculated
target cell population which has been marked with the vector. Gene m~rking can be
measured by determining the fraction of cells which have integrated proviral vector (eg. by
PCR) or by detPrminin~ the fraction of cells expressing the k~n~gPnP (eg. by ~ACS
25 analysis). A recombinant retrovirus ~l~d~ion (e.g., supernatant or supernatant concentrate)
that has high trzm~clt~ction efficiency will produce an increased percentage of inoculated
target cells that have been marked with the viral vector. P~ck~ging cells of the present
invention are able to produce recombinant retroviral supernatants capable of higher
transduction efficiency than standard murine, PA3 17 cells as assayed on NIH/3T3 cells, and
30 particularly as assayed on human 293 or primary cells. Suitable reagents and methods for
pel~llllillg this analysis are provided in the expPrim~?nt~l section below.
CA 02238434 1998-06-08
WO 97/21824 PCT~US96/~9904
26
Also provided herein is a method of producing rctroviral vector sup~rn~t~nt
comprising tr~nc d~lring the novel pa~k~ging cells produced according to the method
described above with a suitable retroviral vector to generate producer cells. This invention
further provides the retroviral supernatant so produced. In addition to the gene of interest to
be delivered to the host cell, the retroviral vector will contain the "p~rk~ging signal" that
allows the retroviral vector nucleic acid to be packaged in the vector particle in the
recombinant park~gin~ cell line, and the long t~nin~l repeat that allow the vector nucleic
acid to become effectively integrated into the target cell genome. The LTRs are positioned at
either end of the proviral nucleic acid and also generally contain regulatory sequences such as
promoter/enhancers that affect e~,ession of the therapeutic or marker gene of interest. The
gene or genes of interest also can be operably linked to a suitable promoter which can be
constitutive, cell-type specific, stage-specific and /or modulable. Enhancers, such as those
from other virus, can also be included.
In a separate embodirnent, the retroviral vectors can contain genes coding for
selectable and/or detectable markers that facilitate isolation of tran~hlced cells. In other
embollim~nt~, it may be desirable for the vector to include a "suicide gene" that allows
recipient cells to be selectively elimin~t~d at will. Isolation, insertion and use of such
markers and suicide genes are well known to those of skill in the art as exemplified in PCT
PublicationNos. WO 92/08796 and WO 94/28143.
Prior art methods to improve gene transfer (tran~ rtion) efficiency with retroviral
vectors often focused on increasing the end-point titer. The evidence presented herein does
not support a correlation between end-point titer and transduction efficiency. The evidence
p~ L~d herein also does not support the conclusion that increasing end-point titer
nec~s~rily increases tr:~n.~dl~ction efficiency because non-transducing particles i~ lrt;l~ with
tr~n~ducing virions and reduce transduction efficiency without reducing-end point titer.
Indeed, the evidence shows that pac~ ging cell line clones should be screened for ability to
generate supernz-~nt based on transduction efficiency rather than on end-point titers.
Thus, the invention further provides a method of producing retroviral vector
supernatant having higher transduction efficiency which is more suitable for gene therapy,
comprising culturing producer cells in a packed-bed bioreactor having a surface to vo}ume
ratio of about 5 to 50 cm2/ml, culturing the producer cells population at a temperature of
around 30~C to about 37~C and harvesting sup~rnzlt~nt produced by said producer cells at
CA 02238434 1998-06-08
W O 97/Z1824 PCTnJS96/19904
27
such time as the transduction efficiency of the supernatant is optimal for a target cell, thereby
obtaining a high tr:~n~dllrtion efficiency supernatant.
As used herein, a packed-bed bioreactor refers to a cell culture vessel comprising a
bed matrix in which cells grow in a confined three-dimensional space. Reactor volume size,
5 and volume of packed-bed matrix can vary depending on the quantity of retroviral
supernatant required.
The matrix which comprises the packed-bed has the following properties:
1) The matrix is made of a material which allows for the ~ltt~hment and growth of
anchorage-dependent cells. The material can also be surface-coated or surface-treated to
10 achieve this ~uality; 2) The matrix can entrap cells normally cultured in suspension so as to
confine them to a three-~limen~ional space; 3) The matrix has a large surface area to volurne
ratio (S to ~0 cm2/ml), enabling cells to grow to high volumetric ~1en~ities (10~ to 1 o8
cells/ml). This property ensures a high volumekic vector production. In a specific
embodiment, the surface to volume ratio is from about 20 to about 30 cm2/ml and preferably
is about 24 cm2/ml; 4) The matrix has a large void volume when packed to minimi7e the
pressure-drop across the bed during operation. A negligible pressure drop ensures an even
supply of nutrients to cells throughout the bed.
The bioreactor is available commercially and the bed matrix can be packed
separately. Preferably, the bed matrix is constructed such that the pressure drop across the
bed matrix is in the range of 0-0.25 mbar/cm. Ideally, the plCS~;UlC drop is negligible.
It is int~n~1ed that any retroviral producer cell line can be cultured in the packed-bed
bioreactor. However, the producer cells derived from the p~k~ging cell lines produced
according to the above method are particularly well suited for use in this method to produce
retroviral vector supern~t~nt~ Such cells include, but are not limited to, producer cells
derived from 3T3 or 293 cells. ~Examples of 3T3 and 293-based producer cell lines are those
based on the PA3 17, GP-Aml2, CRIP. PG13, ProPak-A and ProPak-X packaging cell lines.
After seeding of producer cells into the packed-bed reactor, cells can be cultured at
temperatures ranging from 30~ to 37~C. Preferably, the cells will be initially cultured at
37~C to allow for maximal cell growth. Once the culture achieves a high cell density (106
cells/rnl or greater), the temperature can be lowered to 32~ during the production phase to
P the loss of transducing vector due to thermal inactivation. The tc~ cldLur~
selected for the production phase will depend on the following variables: LelllpcldL~re
CA 02238434 1998-06-08
W O 97~1824 PCT~US96/19904
28
dirre~ ces in the cell specific vector production rate; the rate of vector inactivation at a given
t~ ,.dl lre, cell density; and the rate at which the vessel is perfused with fresh medium.
The time to harvest the retroviral ~u~c;lll~L~lL to obtain optimal tr~n~duction
efficiency will vary with the parental cell type of the producer cells. "Mean residence time"
5 and "tr~n~luction efficiency" are defined above. Optimal tr;m~duction efficiency for a
retroviral sup~ lf refers to the highest percentage of the inoculated target cell population
that is marked with the viral vector, obtained under the culture conditions tested. If the
producer cells are derived from PA3 17 cells, the ~u~ can be harvested after a mean
residence time of from 2 to 12 hours starting from the time the producer cells reach a cell
density of at least 1 o6 cells per ml. For human 293-derived producer cells, the supernatant is
harvested after a mean residence time of from 1 to 24 hours, usually from 6 to 24 hours.
Mean residence time can be decreased as cell density increases.
The packed-bed bioreactor can be operated in a fed-batch or perfusion mode. Fed-batch operation refers to the ch~nging of a part or all of the culture medium at discrete time
intervals. Perfusion mode refers to the continual supply of fresh medium at a specified
perfusion rate which is equivalent to the rate at which vector suF~rn~t~nt is contin~ lly
h~ ~ L~d. The perfusion mode of operation is pl~rt;llc~d as it I l l~ l l l i ;l; l l.C a more stable culture
environm~nt, yielding less stress on the producer cells. Depending on the producer cell line
and the volumetric cell density, the optimal perfusion rate can vary from 1 reactor volume per
day to 24 reactor volumes per day. Preferably, the perfusion rate is from 2-12 reactor
volurnes per day, even more preferably at 2-6 reactor volumes per day, and most preferably,
at a constant rate of about 4 reactor volumes per day. In the most ~l~r~ d embodiment, the
perfusion rate increases proportionally to the cell density. The cell density can be monitored
indirectly by nutrient consurnption, metabolite production, or preferably, by oxygen uptake.
The optimal perfusion rate~ for a given producer cell line at a given volumetric cell density, is
established by ~letermining the mean residence time ofthe culture medium required to
produce a vector supern~t~nt which yields the maximum transduction efficiency of the target
cell population.
The cell culture conditions and parameters described above are also applied to amethod of producing retroviral supern~nt having high transduction efficiency, from non-
murine derived producer cells. After seeding in the packed-bed bioreactor, the cells can be
continuously cultured at 37~C or 32~C. Alternatively, the cells are initially cultured at 37~C
CA 02238434 1998-06-08
W O 97/21824 PCTAUS96/19904
29
until the cells produce a confluent monolayer, and are thereafter cultured at 32~C. The
producer cells derived from the non-murine paek~ging cell lines produced according to the
above method are particularly suited for use in this method for obtaining vector supernatant.
Such cells include, but are not limited to producer cells derived from 293 cells, HTl 080 cells
S and D17cells. In one particular embo-liment, the non-murine producer cells are derived from
293 cells and the retroviral sup~" ~ " is harvested after a mean residence time of from 6 to
24 hours, starting from the time the producer cells reach a cell density of at least 1 o6 cells per
ml.
To achieve m~im~l producer cell densities in the packed-bed bioreactor, proper
10 aeration of the vessel is important. It has been notcd that with sufficient oxygen transfer,
supern~t~nt~ with higher trzm~ tion efficiencies are obtained. Aeration conditions
providing air saturation greater than 20%, preferably between 20% and about 60% are
necessary to achieve efficient vector production. Most preferred are aeration conditions
providing greater than 40% of air saturation.
One of skill in the art can easily determine the most suitable culture medium for the
cells using commercially available basal medium and supplemçnting the medium as provided
by the art. Cells can be cultured in the medium supplemented with fetal bovine serum (FBS),
and vector supernatants produced in the presence of FBS. However, for clinical use, serum-
free medium is preferred. To achieve this, cells can be initially e~pSln-le-l and grown in the
20 packed-bed bioreactor in medium supplemented with FBS, and switched to a serum-free
medium formulation for the production of serum-free vector supernz~t~nt.s Most preferably,
cells will be cultured and vector supern~t~nt~ produced in serum-free media starting with
cells derived from a serum-free master cell bank.
The disclosed method also can be utilized to produce retroviral vector supernatant
25 having high transduction efficiency by seeding primate producer cells in a packed-bed
bioreactor. Suitable primate producer cells include those derived from 293 cells, ProPak-A
and ProPak-X cells. When the producer cell is deri~ed from a 293 cell, it is preferable to
harvest the retroviral supernatant when the culture medium's mean residence time in the
reactor is about 6-24 hours, starting from the time the producer cells reached a cell density of
30 at least 1 o6 cell per ml.
The invention further provides a method of obtaining high transduction efficiency
MLV-based retroviral supernatant suitable for gene therapy, comprising culturing human-
CA 02238434 1998-06-08
W O 97/21824 PCTrUS96/19904
derived producer cells under the conditions described above. Human-derived producer cells
are seeded in a packed-bed bioreactor having a surface area to volume ratio of about S to 50
cm2/ml, at a concentration starting from about 2 to 3 x 104 cells/cm2, under constant
perfusion, with aeration sufficient to ~ ill the culture m~ m at about 20% to 60% air
S saturation, ~l~rt;l~ly at or greater than 40% air saturation. The cells are cultured at a
te~ d~ure of about 37~C until they grow to a cell density of at least 1 o6 cells per ml
whereupon the t~"~ dlure is then lowered to 32~C. Retroviral sUp~rn~t~nt is then collected
after an additional 6 to about 24 hours of incubation time. Human cells useful in this method
include, but are not limited to those derived from 293 cells, HT 1080 cells, ProPak-A cells
10 and ProPak-X cells.
The present method of producing retroviral vector supern~t~ntc also encompasses co-
culturing of two or more producer cell lines under the conditions described above. In this
circumstance, the packed-bed bioreactor is seeded with two or more different producer cell
lines, preferably cont~ining the same vector and having complementary tropisms, e.g.
15 producer cells derived from human amphotropic and xenotropic p~ek~ging cell lines. This
co-culture technique results in an amplification of vector copy number within the producer
cells, and yields vector supern~t~ntc which contain a mixture of retroviral vector particles
targeted to distinct receptors expressed on most human cells.
The culture sup~ produced by the methods described herein is also within the20 scope of this invention. The supernatant has been shown to transduce cells, and in particular,
stem cells very eff1ciently.
Prior to tr:~ncduction~ the stem cells are isolated and selected. Methods of isolating
and selecting cells are well known in the art. For example, sorted CD34+Thyl+Lin~ cells
from either adult bone marrow (ABM) or mobilized peripheral blood (MPB) are used.
25 CD34+ Thyl+Lin~ are highly enriched in human hematopoietic stem cells. See. U.S. Patent
No. 5,061,620.
Various methods can be used to isolate stem cells. For the purpose of illustrating one
method of preparing stem cells from ABM, 20 mL of fresh bone marrow can be isolated by
aspiration of the iliac crest from normal human volunteers. Marrow is separated by taking
30 the mononuclear cell fraction following a Ficoll-Perque separation, positive-selected for
CD34+ cells according to the method described by Sutherland et al. (1992) E~p. Hematol.
20:590. Briefly, cells are resuspended in .ct:~ining buffer (SB) (HBSS contz~inin~ 10 mM
CA 02238434 1998-06-08
W O 97/21824 31 PCT~US96/19904
~IEPES, 2% heat-inactivated FCS) at 5 x107 cells/mL. QBEND10 (anti-CD34) (Amac,
Westbrook, ME) is added at 1/100 dilution, and then the cells are incubated on ice for 30
min. Cells are then washed in SB with a FCS underlay, and resuspended at 4 x 1 07/mL in
SB. An e~ual volume of washed Dynal sheep anti-mouse IgG,Fc m~gnetic beads (Dynal,
Oslo, Norway), is added at a 1:1 bead to cell ratio, to give a final cell concentration of 2 x 107
cells/mL. After 30 min incubation on ice, with gentle inversion, the tube was placed against
a Dynal magnet (Dynal) for 2 minutes, and CD34- cells removed. Following two washes, 20
mL of 'glycoprotease' (O-sialoglyco~ tein endopeptidase, Accurate Ch~micz~l, Westbury,
New York) plus 180 mL of RPMI (JRH Biosciences)/20% FCS is added and the beads
incubated at 37~C for 30 min to cleave the QBEND10 epitope, and release CD34~ cells from
the beads. Beads are then washed three times to m~imi7.o cell recovery. The glycoprotease
used for the release step in the positive selection procedure has been shown not to effect
subsequent ex vivo expansion of progenitors (Marsh et al. (1992) l,eukemia 6:926).
Mobilized peripheral blood (MPB) samples can be obtained with informed consent
from multiple myeloma patients. Patients are treated on day 1 with cyclophosphamide at 6
g/m2 (1.5 g/m2 every 3 hours x 4 doses). From day 1 until the start of leukopheresis (usually
10-28 days), granulocyte macrophage colony ~tim~ ting factor (GM-CSF) is given at 0.25
mglm2/day. Apheresis for total white cells is started when the peripheral blood white cell
count is greater than 500 cells/ml and the platelet count is greater than 50,000 cells/ml.
Patients are apheresed daily until from 6 x 1 o8 mononuclear cells (MNC) are collected.
Fresh MPB samples are then elutriated with a JES.0 Beckrnan counterflow elutriator
equipped with a Sanderson chamber (Beckman, Palo Alto, CA). Cells are resuspended in
elukiation medium (Biowhittaker, Walkersville, MD) at pH 7.2, supplemented with 0.5%
human serum albumin (HSA). The rotor speed is set at 2000 RPM, the cells are inkoduced7
and the first fraction collected at a fiow rate of 9.6 ml/min. Fractions 2 and 3 are collected at
the respective flow rates of 14 and 16 ml/min. The larger cells ~ g in the chamber are
collected after stopping the rotor. Cells are resuspended in RPMI supplem~nt. d with 5%
HSA, 10 mg/ml DNAse I and penicillin/streptomycin at 50 U/ml and 50 mg/ml, respectively.
Fractions 2 and 3 are pooled and incubated with 1 mg/ml heat-inactivated human gamma-
globulin to block non-specific Fc binding. Granulocytes are further depleted by incubation
with CD15 conjugated to m~gne~ic beads (Dynal M450, Oslo, Norway) followed by
magnetic selection.
CA 02238434 1998-06-08
W O 97~1824 32 PCT~US96/19904
CD34~Thyl+Lin~ cells are isolated from ABM and MP~3 by flow cytometry as
follows. Antibodies to CD14 and CD15 were obtained as FITC conjugates from ~ecton-
Dickinson. Antibodies to Thy-1 (GM201) can be obtained from a commercial source or
prepared and ~i~tected with anti-IgG1-PE conjugate from Caltag. Antibody to CD34 (Tuk 3)
5 also can be commercially obtained and detected with an anti-IGg3-Texas Red conjugate
(Southern Biotechnologies).
Anti-CD34 antibody or an IgG3 isotype m~t~hPd conkol is added to cells in staining
buffer (HBSS, 2% FCS, 10 mM HEPES) for 20 mtnllt~c on ice, together with anti-Thy-l
antibody at 5 ~g/ml. Cells are washed with a FCS underlay, and then incubated with Texas
10 Red conjugated goat anti-mouse IgG3 antibody and phycoerythrin-conjugated goat anti-
mouse IgG1 antibody for 20 minut~s on ice. Blocking IgGl is then added for 10 minlltf~s
Afterblocking, the FITC-conjugated lineage antibody panel (CD14 and CD15) is added, and
incubated for another 20 minutes on ice. After a final washing, cells are resuspended in
staining buffer c~ i . , i, .g propidium iodide (PI).
Cells are sorted on a Vantage cell sorter (Becton Dickinson) equipped with dual
argon ion lasers, the primary laser ~mitting at 488 nm and a dye laser (Rho-l~mine 6G)
emitting at 600 nm (Coherent Innova 90, Santa Cruz, CA). Residual erythrocytes, debris and
dead cells are excluded by light scatter gating plus an FL3 (PI) low gate. The sorted cell
population is diluted 1:1 in HBSS, pelleted, and resuspended in HBSS for hemocytometer
20 counting.
For transduction, the sorted cells are suspended in frcsh or freshly thawed retroviral
supern~t~nt diluted in a~~ iate media c~ cytokines. Cells and vector are then
centrifuged and resuspended and cultured in cytokine-enriched media for about three days.
For culturing CD34+ hematopoietic stem or progenitor cells~ the cytokines are preferably a
25 combination including, but not exclusively, IL-3, IL-6, LIF and SCF. After three days, the
- cells are harvested and used to cletf rmine bulk transduction frequency as described below.
Alternatively, the cells can be introduced into a patient by methods well known to those of
skill in the art for gene therapy.
Further provided by the invention is a method of increasing the gene tr~n~dllction
30 efficiency of a cell by transducing target cells with a retroviral vector sup~rn~t~nt derived
from one or more recombinant packaging cells produced by the methods of the present
invention. Preferably, the target cell is a primary hematopoietic cell such as a stem cell. In a
CA 02238434 1998-06-08
W O 97/21824 PCT/US96/19904
33
specific embodiment, the retroviral supernatant is derived from the culture of ProPak-A.6 or
ProPak-A.52 and ProPak-X. The tr~n~dl-rtion efficiency of a cell population can also be
increased by inoculating or tr~n~illring a target cell population with vector particles of more
than one tropism. This is achieved by tr~n~clnr.ing the cells with a vector particle having a
modif1ed envelope protein as described above, wherein the envelope protein can bind to more
than one receptor. Examples 6 and 7 below provide tr~n.~-luçtion of target cells with a vector
particle having a chimeric amphotropic/xenotropic envelope encoded by the Eax env gene.
Alternatively, the cells are simultaneously tr~n~ lr.erl with vector supernatants c~-nt~inin~
particles of complementary kopisms, eg. amphotropic and xenotropic vector sup~
mixed into a single inoculum. The inoculum Co~ irlg vector supern~t~nt~ of different
tropisms can be prepared by mixing the two supern~t~nt~ obtained from their respective,
separately cultured producer cells, or the two complementary producer cell lines can be co-
cultured to produce a single supernatant cont~ining the two types of vector particles. In a co-
culture, a production vessel, preferably a packed-bed bioreactor, is seeded with a mixture of
producer cells with complementary tropisms that produce vector particles which are capable
oftr~n~d~lcing human cells.
In one embodiment of the present invention, cells are trs1n~cluce~1 with a retroviral
~U~ produced from the co-culture of a ProPak-A and a ProPak-X producer cell line,
specifically, ProPak-A.52.LMiLy and ProPak-X.LMiLy. The co-culture of ProPak-A and
ProPak-X producer cells have the advantage that they do not produce RCR. Close to 100%
gene transfer was achieved using such co-culture supernatants on MPB-derived C~34+ cells.
Yet another aspect of the invention is a method of producing a p~ck~gin~ cell line
capable of ~ e~hlg more than one envelope protein with tropism for human cells, e.g.
amphotropic and xenotropic envelope proteins. This p~kz~ging cell line is derived by
introduction of ~ es~ion plasmids for other envelope genes into an existing pack~ginp cell
line. For example, ProPak-~.6 cells can be transfected with the expression plasmid pCI-Ex
and cell clones isolated which secrete vector particles that tr~n~iuce quail cells or other cells
which are tr~n.~d~lced by xenotropic particles but not by arnphotropic particles. Such multi-
tropism particles would be expected to tr~n~duce target cells with higher efficiency in the
situation where the target cells express the individual receptors but at concentrations too low
to allow stable vector-cell complex formation by the pure amphotropic or xenotropic
particles.
CA 02238434 1998-06-08
W O 97/21824 PCTnUS96/19904
34
The following examples are int~?nc1ec~ to illustrate, not limit the scope of the invention
disclosed herein.
Experiment~l
As noted above, successful retroviral-mediated gene therapy requires safe and
efficient packaging cell lines for vector particle production. Existing p~ ging lines for
murine leukemia virus-based vectors are predomin~ntly derived from NIH/3T3 cells which
carry endogenous murine viral sequences which could participate in recombination to form
replication competent retrovirus (RCR), thereby rendering them unsafe. Provided herein
are methods for constructing safe and efficient retroviral packaging cell lines for vector
particle production as well as the cell lines so obtained.
M~terizll~ and Methods
Unless otherwise stated, materials and reagents are prepared and methods are
con~ terl, as set forth below. Producer cell lines are named by noting the p~eksl~in~ cell
line and inserted vector. For example, PA.SVNLZ denotes a producer cell line derived
from pa~k~ging cell line PA317 carrying the vector SVNLZ. PP-A and PP-X denote
ProPak-A and ProPak-X, respectively.
P~17-P~ ed Cell T ines and Retroviral Vectors
The retroviral vector producer cell lines disclosed herein were generated from the
amphotropic PA317 p~eL-~ging cell line produced according to the methods of Miller and
Buttimore, (1986) Mol. Cell Biol. 6:2895-2902. Producer cell lines, the GP+E86
packaging cell line produced according to the method of Markowitz, et al. (19883 J: Virol.
62: 1120-1124, and NIH/3T3 cells (ATCC CRL 1658) were cultured in Dulbecco"s
modified Eagle mediurn (DMEM) supplemented with 4.5 g/L of glucose, 4 mM L-
glnt~min~ and 5% Cosmic Calf Serum (CCS, Hyclone, U17. The human epithelial
carcinoma line HeLa (ATCC CCL2) was cultured in DMEM supplemented with 4.5 g/L of
glucose, 4 mM L-glutamine and 10% fetal bovine serum (FBS, Hyclone, UT). The human
lymphoid cell line Jurkat was cultured in RPMI supplemented with 4.5 g/L of glucose, 2
mM L-glutamine, and 10% FBS. All cell lines were m~intzl;ned in an incubator at 37~C
under 5% C02 unless otherwise stated.
CA 02238434 lsss-06-08
W o 97/21824 pcTrus96/l99o4
The SVNLZ (see Bonnerot, et al. (1987) Proc. Nafl. ~4cad. Sci. U.S.~. 84:6795-
6799) and LMTNL vectors are known in the art (See for example, Escaich et al. (1995)
Human Ge~le Therapy 6:625-634). Figure 2 sets forth a map of these vectors. The
SVNLZ vector encodes the LacZ gene with a nuclear localization signal expressed from
S the simian virus 40 early promoter (Figure 2). The LMTNL vector encodes a trans-
dominant mutant of the HIV Rev protein (RevM10) expressed from the MMLV-LTR, andthe neomycin phosphotransferase (neo) gene expressed from the thymidine kinase
promoter (Figure 2).
10 P~17 Culture Con~litions
For vector production, PA317-based producer cells were seeded into culture vessels
at a density of 3xl 04cells/cm2 unless otherwise noted. Once the cells had forrned a
confluent monolayer (approximately 3 days), the medium was changed and the
temperature lowered from 37~C to 32~C. Three types of culture vessels were used: (1) 75-
cm2 tissue culture flasks with filter caps (Costar, Cambridge, MA); (2) 900-cm2 roller
bottles with filter caps (Costar, Carnbridge, MA); or (3) a 500 ml bench scale packed-bed
bioreactor (New Brunswick Scientific, Edison, NJ) with 10 g of Fibra Cell discs (12,000
cm2). Sa7nples from the bioreactor (I ml) were removed with a sterile syringe through a
sampling port. Supernatant samples were cleared of cellular debris by filtration through
20 0.45 ~Lm nitrocellulose filters (Nalgene, Rochester, NY), snap-frozen in methanol/dry-ice,
and stored at -70~C. Frozen supern~t~nf~ were thawed at 37~~ and kept on ice until
assayed. All supernatants used in this study were free of replication competent retrovirus
as determined by the S~L- assay on PG4 cells as described in Cornetta, et al. (1993)
Human Gene Therapy 4:579-588 and Forestell et al. (1996) J. Virological Methods
60:171-178.
A~ys for Retroviral Vectors
T 5~7. en~l-point ~itration
NIEI/3T3 cells were plated into 24 well plates at 5x104 cells per well, and
inoculated the next day with 0.2 ml of diluted vector supernatant (10~! to 10-6 dilution
30 series) in the presence of polybrene (8 ~g/ml). After three hours, the inoculum was
CA 02238434 1998-06-08
W O 97~1824 PCT~US96/19904
36
aspirated and 0.5 ml of fresh medium added. Three days post inoculation, the cells were
f1xed with 0.5% glutaraldehyde and stained with X-gal using the protocol of Bagnis et al.
(1993) Oncogene 8:737-743. The end-point titer (cfu/ml) was calculated by counting a
st~ti~tif~zllly representative number of blue colonies in a well, multiplied by the dilution
5 factor, and divided by the volume of the inoculum. All end-point titrations were
performed in duplicate.
For the xenotropic vectors prepared using ProPak-X, titrations were performed
using the same procedure except that human 293 cells were used as the target cells.
T z-~7 trsin~duction efficienry
Cells were plated into 24 well plates at 5x104 cells per well, and inoculated the next
day with 0.2 ml of a 1:1 dilution of vector supernatant with medium in the presence of
polybrene (8 ~Lg/ml). Three hours later, the inoculum was aspirated and 0.5 ml of fresh
medium added. Transduction efficiencies were determined three days post-inoc~ tion by
flow cytometry using the FluoReporter lacZ detection kit (Molecular Probes, Inc., Eugene,
OR). For NIH/3T3 or 293 cells, the reaction with fluorescein di-~-D-galactopyranoside
(FDG) was quenched after one minute with lmM phenethylthio-~-D-galactopyranoside(PETG). For Jurkat and HeLa cells, PETG was added after a one hour incubation on ice.
Acquisition and analysis were performed using a Becton Dickinson FACScan and theLYSYS software package. Transduction efficiency was rletermined as the percentage of
20 cells expressing the lacZ gene (green fluorescence intensity) above the basal fluorescence
levels defined as that of non-transduced control cells.
G4 18 resistance end-point titer
Titrations were performed in 6 well plates seeded with 2.5x104 NIH/3T3 cells perwell, and inoculated the next day with 1.0 ml of diluted vector supernatant ( 10-~ to I o-6
25 dilution series). Three hours later, the inoculum was aspirated and 2.0 ml of fresh medium
was added. One day after inoculation, G4 18 was added to each well at a final
concentration of 0.7 mg/ml, and the medium was changed as required until individual
colonies could be seen (typically 10 to 14 days). The end-point titer (cfu/ml), was
calculated by counting a statistically representative number of G418 resistant colonies in a
30 well and multiplied by the dilution factor.
-
CA 02238434 1998-06-08
W O 97/21824 PCTnUS96/19904
For the xenotropic vectors prepared using ProPak-X, titrations were performed
using the same procedure except that human 293 cells were used as the target cells.
~ 418 r~ t~nc ~ tr~nsduction errficiency
Cells were seeded into 6 well plates at lxlOs cells per well. The next day, cells
5 were inoculated with 1.0 ml of a 1: 1 dilution of vector supernatant with medium in the
presence of polybrene (8 ,ug/ml). Three hours later, the inoculum was aspirated and 2.0 ml
of fresh medium was added. Three days post-inoculation, adherent cells were trypsinized
and seeded into 6 well plates at a 10-fold dilution series from 105 cells/well to 1 cell/well.
Jurkat cells were plated in methylcellulose medium to allow individual colonies to be
10 scored. Duplicate plates were seeded, and incubated in the presence or absence of G418.
The percentage transduction efficiency was calculated as the number of G418-resistant
colonies divided by the number of colonies that formed in the absence of G418, multiplied
by 100.
The following methods relate to Examples 6-12.
Retrovir~l pack~ing cell cul~lre
P~ ginp and producer cell lines were grown in Dulbecco's modified Eagle's
medium (DMEM: JR~ Biosciences, Lenexa, KS) supplem~.nte~l with fetal bovine serum
(FBS: ~IyClone Laboratories Inc., Logan, UT) at either 5% (PA 317 (Miller and Buttimore,
(1986) Mol. Cell. Biol. 6: 2895-2902), PG13 (Miller et al., (1991) J. Virol. 65: 2220-2224))
or 10% (ProPak).
FxI ression Co~fructs and Retroviral Vectors
Gag--Pol and Env proteins were expressed from separate plasmids which, along
with the retroviral vectors used in this study, are described in Figures 1 and 2. Methods are
described herein for derivation and analysis of cells e~pressing MLV proteins, analysis of
Gag-Pol and Env proteins in tissue culture supernatants, and titration of retroviral vectors
encoding the Lyt2 antigen (Rigg et al. (1995) J. Imm1mol. Meth. 188: 501-509). Aproducer cell line and the respective retroviral vector supernatant are indicated by the cell
name followed by a period and the name of the vector, c.g., ProPak-A.LLySN
(Amphotropic), ProPak-X.LLySN (Xenotropic), PA317.LMiLy or P(~13.LMiLy.
CA 02238434 1998-06-08
W O 97~1~24 PCTAUS96/19904
38
Co-cultures of ProPak-X and ProPak-A-based producer cell lines are indicated with a
slash, i.e. ProPak-AtX.LMiLy.
Vector Supe~ t~nt Production
Producer cells were seeded into culture vessels at a density of 3xl 04cells/cm2 or
S greater. Once the cells had forrned a confluent monolayer (approximately 3 days), the
medium was changed and the temperature lowered from 37~C to 32~C. Thereafter,
supernatant was harvested at 12 hour intervals. For supernatant production in T-flasks
from pools of LLySN-carrying producer cells, medium was repeatedly harvested for up to
10 days at 32~C. Supernatants were then compared for their ability to tr~ncd~lce NIH/3T3
10 or 293 cell lines, and those supernatants that yielded the highest transduction were used in
the studies reported here. Three types of culture vessels were used: (1) 75--, 162--or
225-cm2 tissue culture flasks with filter caps (Costar, Carnbridge, MA), (2) 900-cm2 roller
bottles with filter caps (Costar, Cambridge, MA); and (3) a 500 ml packed-bed bioreactor
(New Brunswick Scientific, Edison, NJ) with 10 g of Fibra Cell discs (12,000 cm2). The
15 bioreactor was seeded with 5x105 cells/ml and the mediurn was circulated by a marine-type
impeller. Supplemental aeration was achieved by direct rnicro-sparging of a medical grade
mix of air with 5% CO2, and 0.01% Pluronic F-68 (Sigma, St. Louis, MO) was added to
prevent cell damage from shear. Vector supernatants were cleared of cellular debris by
filtration through 0.45 ~Lm nitrocellulose filters (Nalgene, Rochester, NY), snap-frozen in
20 methanol/dry-ice, and stored at -70~C. Aliquots were thawed at 37~C and kept on ice until
use. All supernatants used for gene transfer experiments were free of replication-
competent retrovirus (RCR) as determined by the ext~n~l~cl S+L- assay on PG4 cells
(Forestell et al., (1996) J. Virol. Meth. supra).
Tno~ tiorl Proce~lllres
Unless otherwise stated, cells were inoculated with vector once at unit gravity for
3 h at 37~C. Inoculation under centrifugation, termed spinoculation, (Kotani et al., (1994)
Hum. Gene Ther.5: 19-28; Bahnson et al., (1995) J. Virol. Mefh. 54: 131-143; Forestell et
al., ~1996) supra) was at 2550g for 3 or 4 h. Polybrene (8 ~ug/rnl) or protarninc sulfate (4
g/ml) was added to inocula for cells lines or primary cells, respectively.
CA 02238434 1998-06-08
W O 97/21824 PCT~US96/19904
39
Prim~ry Hen~t~poietic Cell Purification
PBL were isolated from peripheral blood mononuclear cells as described (Rigg et
al., ~1995) supra) and depleted of CD8-positive cells. The CD34-positive (CD34+)fraction was isolated from mobilized peripheral blood (MPB) or adult bone marrow- 5 (ABM) with the Baxter Isolex affinity column (Baxter Healthcare, Deerfield, IL). Further
purification to obtain the fraction carrying the Thy-1 antigen (CD34+/Thy+) was achieved
by high-speed flow cytometric sorting, as described by Sasaki et al, (1995) .~: Hemat. 4:
503-514). Hematopoietic stem and progenitor cells (HSPC) were cultured in a 1:1 mixture
of IMDM and R~MI media cont~ininP: 10% ~BS, IL-3 and IL-6 (20 ng/ml each, SandozPhar~na, Basel, Swit~rl~nc~), and stem cell factor (SCF; Amgen Inc., Thousand Oaks, CA)
or leukemia inhibitory factor (LIF; Sandoz) (100 ng/ml).
Tr~n~duction F.fficiency Assays
Transduction efficiencies achieved with Lyt2-encoding vector supern~t~nt~ were
qll~ntit~ted as the proportion of target cells that expressed Lyt2 antigen 2 or 3 days
following inoculation (Rigg et al., (1995) supra). Integration ofthe RevM10 gene into
~SPC clonogenic progeny was determined by PCR. Methylcellulose (Stem Cell
Technologies, Vancouver, Canada) was supplemented with IL-3, IL-6, SCF, erythropoietin
(Arngen, Thousand Oaks, CA) and granulocyte-macrophage colony stimulating factor(Immunex, Seattle, WA). Individual colonies of all lineages (CFU-C) were picked after 12
to 14 days culture in methylcellulose culture, and assayed for the revM10 gene in cellular
DNA using a PCR assay as described in Plavec et al. (1996) Gene Therapy 3:723.
Colonies were scored positive if both revM 10 and ,~-globin sequences were detected.
Fx~n~le 1: (~onstruction of Packaging Cells
To minimi7~ the chance of recombination with endogenous viral sequences to form
RCR, ç~n~ tf cell lines were screened for endogenous retroviral, and in particular,
endogenous MMLV nucleic acids. Genomic DNA from a variety of cell lines was
analyzed by Southern blot hybridization by the method generally disclosed in Sambrook et
al., (1989) Molecular Cloning: A Laboratory Manual 2nd ed., Cold Spring ~arbor
Laboratory, New York, using probes specific to retrovirus MMLV long terminal repeat
CA 02238434 1998-06-08
W O 97~1824 PCTnJS96/19904
(LTR) or MMLV gag-pol sequences. The following cell lines were screened: CHO cells
which are used to produce recombinant proteins; Vero and MRC-5 cells in which vaccines
are produced (WHO (1989) "Technical Report Series No. 786" WHO, Geneva), and human
embryonic kidney 293 cells (ATCC C~L 1573) that are used to produce adenovirus
S vectors for clinical gene therapy applications and Lt;l~o~ dl vectors (Pear et al. (1993)
Proc. Natl. Acad Sci. U.S.~. 90:8392-8396; Finer et al. (1994) Blood 83:43-50). Mus
dunni tail fibroblasts (NIH) also were included since these cells are reportedly free of
endogenous MMLV sequences (Lander and Chattopadhyay (1984) J. ~irol. ~3:695-698).
Genomic DNA from NIH/3T3 cells, the basis for the majority of existing p~k~g;ng cell
10 lines, hybridized very strongly with both MMLV-specific probes at low or high stringency.
Hybridization conditions are provided in the key below Table 1. In contrast, neither probe
cross-hybridized with genomic DNA from 293 or MRC-5 cells, even at low stringency
(Table 1). In addition, no cross-hybridization was seen with genomic DNA from Mus
dunni, MDCK, Vero or Fox Lung cells at high-stringency (see Table 1, below).
CA 02238434 1998-06-08
WO 97121824 41 PCT/US96/19904
Table 1
Screening of cell lines for cross-hybridization to MMLV LTR or gag-pol sequences.
Fn~1~en~us MMLV Sequences ff~vbridization~
~Iybridization Probe: LTR gag/pol
W:~h Strir~ency: Low E~i~ Low
Cell Lines Tested
293 (ATCC CRL 1573) - - - -
MDCK (ATCC CCL 34) ~ - +
Mus dunni tail fibroblasts ~ - ~t
Vero (ATCC CCL 81) - - ~ -
Fox Lung (ATCC CCL 168
MRC-5 (ATCC CCL 171)
NIH/3T3 (ATCC CR~ 1658) ++ ++ ++ ++
CHO-Kl (ATCC 61j ++ ++ ++ ++
Kev:
Hybridization signal strength: -, none; +, weak; +, moderate; ++, strong.
Probes: LTR (positions relative to cap site of genomic RNA): nucleotides -232 (Eco RV) to
563 (Pst I of 5' leader sequence); gag-pol: nucleotides 739 (Pst I in gag) to 3705 (Sal I in
pol).
Stringency (65~C in all instances): Low, 500 mM Na+, High, 50 mM Na+.
ATCC: American Type Culture Collection, 12301 Parklawn Drive, Rockville, Maryland,
20852, U.S.A.
CA 02238434 1998-06-08
W O 97/21824 42 PCT~US96/19904
Fox Lung and MRC-5 were discounted due to poor growth or limited cell division
potential respectively, which would preclude sub-cloning of stably-transfected cells. Thus?
293, MDCK, MUS dunni and Vero cells were identified as candidate cell lines from which
to derive pa(~k~ging cell lines.
To decrease the probability of RCR formation, separate expression plasmids were
constructed forgag-pol and env, as disclosed in Danos (1991) Proc. Natl. ~cad. Sci. US.~.
85:6460-6464, Danos and Mulligan, (1988) Proc Nafl~ Acad. Sci. U.S.A. 85:6460-6464;
Markovitz et al. (1988) ~ Virol 62:1120-1124, Miller (1990) Human Gene Therapy 1:5-
14; Morgenstern and Land (1990) Nucl. Acids Res. 18:3587-3596. In contrast to existing
packaging cells, however, only the Illhlilllulll genetic information required to encode Gag-
Pol and Env proteins was included. The structural gene sequences were ampli~led by
polymerase chain reaction (PCR) using the primers shown below to obtain the openreading frames (ORFs) from the initiation to the termin~t;on codons of the gag-pol or env
genes flanked by Not I reskiction sites, and the fr~gment~ were subcloned into pBluescript
SK~ (Stratagene, La Jolla, California). Oligonucleotide primers (Genosys
Biotechnologies, Woodlands, Texas) corresponding to the N-terminus of the genes also
placed the AUG in the ideal context for translation (Kozak (1987)) J: Mol. Biol. 196:947-
950), and those corresponding to the C-l~".~ le encoded a second in-frame stop codon.
The integrity of PCR products was verified by DNA sequencing. The gag-pol ORF was
amplified from the plasmid pVH-2, which carries the infectious Moloney MLV se~uence
(Miller and Verma (1984) J. Virol. 49:214-222), using the primer pair:
5'-AAAAAAAAGCGGCCGCGCCGCCACCATGGGCCAGACTGTTACCAC-3' (SEQ
ID NO: 1),
and 5'-~GCGGCCGCTCAttaGGGGGCCTCGCGGG-3 ' (SEQ ID NO: 2).
The underlined ATG is that of the pl 5Gag (bases 621 to 623) and the codon in
lower case corresponds to the pol stop codon (bases 5g35 to 5837). Thc expression
plasmid pCMV-gp with the human cytomegalovirus immediate early (CMV) promoter was
constructed by inserting the gag-pol fragment into the pcDNA3 plasmid (Invitrogen, San
Diego, CA) from which the neomycin resistance ~ ssion cassette (DraIII to BsmI) had
been deleted. Figure 1 schem~tically shows the plasmid constructs used for ~plession of
MMLV structural genes. Plasmids carrying the gag-pol ORF were propagated at 30~C to
CA 02238434 1998-06-08
W O 97t21824 PCTrUS96/19904
43
prevent recombination under the conditions set forth in Joshi and Jeang (1993)
BioTechniques 14:883-886.
To amplify the env gene, a contiguous amphotropic envelope sequence was
constructed from p4070A constructed according to the method of Ott et al. (1990) J. Virol.
64:757-766, and amplified with the primer pair:
5'-TAATCTACGCGGCCGCCACCATGGCGCGTTCAACGCTC-3' (S~Q ID NO: 3)
and 5'-AATGTGATGCGGCCGCtcaTGGCTCGTACTCTATGG-3' (SEQ ID NO: 4).
The underlined ATG corresponds to bases 37 to 39, and the stop codon (lower
case), bases 1998 to 2000 (Ott et al. (1990) supra. The CMV promoter-env expression
plasmid pCMV*Ea was created by insertion of the env 0~ in place of the beta-
galactosidase gene of pCMV13 (~lontech, Palo Alto, CA) modified by mutation of an
extraneous ATG in the SV40 intron to ACG (SD/SA~). The integrity of the PCR products
was verified by DNA sequencing.
To further identify the optimal cell line, sandwich ELISA assays were developed to
detect Gag and Env proteins in transfected cells and particularly in sUpern~tslntc. Plates
were coated with hybridoma culture supernatants from either 83A25 (available from the
NIH) as disclosed in Evans et al. (1990) J. Virol. 64:6176-6183, for Env, orR187 (ATCC
CRL 1912) for Gag. Captured proteins were detected with 79S-834 and 77S-227 anti-sera
(Quality Biotech, Camden, New Jersey), respcctively, and horseradish peroxidase-conjugated anti-species antibodies and substrate 2,2-Azinobis (3-ethylbenzothiazoline-6-
sulfonic acid) (Pierce, Rockford, Illinois).
The ability to produce vector particles was assessed by transfection of a gc~g-pol
expression plasmid into candidate cell lines and selection of drug-resistant pools. Only the
2~ supern~t~ntc from gag-pol -transfected 293 or Mus dzmni cells contained Gag protein. No
Gag was secreted by transfected Vero or MDCK cells, although Gag was present in the cell
lysates. Vero and MDCK cells were discounted as candidate cell lines, and human 293
cells were elected to derive packaging cells.
_
CA 02238434 l998-06-08
W O 97/21824 PCT~US96/19904
44
Con~tructiorl of ProP~k-A
To derive arnphotropic packaging cells, the pCMV*Ea plasmid was introduced into
293 cells (ATCC CRL 1573) by co-transfection (Profection Kit, Promega, Madison,
Wisconsin) at a ratio of 15: 1 with the pHA58 plasmid ~as disclosed in Riele et al. (1990)
Nature 348:649-651) conferring resistance to hygromycin B (250 mg/ml; Boehringer,
Tnt1t~3n:~polis, Indiana). Stably-selected populations were stained with anti-Env antibody
(83A25, available from the NIH) and individual Env-positive cells were isolated by
automatic cell deposition on a FACStar Plus (Becton Dickinson, San Jose, CA). Three
clones with the highest fiuorescence intensity were further characterized. All three yielded
equivalent titers upon transient co-transfection with gag-pol and vector plasmids.
Next, the pCMV-gp construct was stably transfected into one of the three 293-Envclones by cotransfection with the plasmid pSV2pac (as disclosed in Vara et al. (1986)
Nucl. Acids Res. 14:4617-4624). Puromycin-resistant (I mg/ml; Sigma, St. Louis,
Missouri) clones were grown to confluence, medium was exchanged and supernatants were
collected 16 hours later, filtered, and analyzed for Gag and Env production by using the
sandwich ELISA as described in the previous page. Sixteen clones were identified (16/37)
that secreted high levels of Gag and Env antigens. Of these, six clones produced virus in
transient transfections at titers within 2 to 3-fold of Oz 2 cells ~Table 2A), the arnphotropic
equivalent of BOSC: 23 cells ~see Pear et al. (1993) Proc. ~atl. Acad. Sci. USA. 90:8392-
8396).
CA 02238434 1998-06-08
W O 97/21824 PCTrUS96/19904
Table 2A
Comparison of end-point titers from transiently-transfected ProPak-A cell clones
Transient Titers - MFG lac Z Vector
Cell Line or End-Point Titer
Clone # (cfu x 10~5/ml)
Oz2 13.8+0.3
ProPak.12 8.8 i 0.8
ProPak.31 8.0 + 0.5
ProPak.6 7.5 i 0.0
ProPak.27 6.8 i 1.2
ProPak.21 6.0 + 1.0
ProPak.S 6.0 i 2.0
Sup~m~t~nt~ were collected after 16 hours at 32~C, and end-point titers deterrninecl on
NIH/3T3 cells using the materials and methods disclosed above.
cfu: colony forming units.
Cells were seeded at 2 xl Os cells/cm2 in 6-well plates, and transfected 16 hours later with 2.5
10 mg/well MFG-lacZ DNA (Dranoff et al. (1993) supra and described herein) was in the
presence of 25 mM chloroquine (Pear et al. (1993), supra). Titers are the average and range
for duplicate transfections. Oz 2 cells also are called Bing cells and are the amphotropic
equivalent of BOSC23 cells (Pear et al., 1993).
CA 02238434 1998-06-08
WO 97/21824 PCT~US96/19904
46
Table 2B
Comparison of end-point titers from
stable producer cell clones with the LMTNL Vector
End-Point Titer
Producer Clone(G4 1 8r cfu x 1 0~6/ml
PA3 1 7.LMFNL 1.7 + 0.7
ProPak-A.6.LMTNL.6 2.1 + 0.3
ProPak-A.6.LMTNL.7 2.2 ~ 0.8
Supern~t~nt.~ were collected after 16 hours at 32~C, and end-point titers cletPrnnintocl on
NIH/3T3 cells
cfu: colony forming units.
Supern~t~n~ were harvested from confluent cultures of producer cell clones in T-75 flasks.
The average and range for triplicate sarnples is given.
One clone with high transient transfection titers was selected and designated
ProPak-A.6. ProPak-A.6 was deposited with the Arnerican Type Culture Collection
(ATCC), 12301 Parklawn Drive, Rockville, MD 20852 U.S.A. on December 15, 1995,
under the provisions for the Budapest Treaty for the Deposit of Microor~ni~m~ for the
Purposes of Patent Procedure, and was accorded ATCC Accession No.CRL 12006.
Transient titers reflect the efficiency of transient transfection, and the titers obtained with
ProPak-A cells are lower than those achieved with Oz 2 cells, possibly because Oz 2 is
based on a 293T cell clone selected for high transient transfection efficiency (Pear et al.
(1993) supra).
CA 02238434 1998-06-08
W O 97~1824 PCT~US96/19904
47
Co~truction of ProP~k-X
A xenotropic p~ ging cell line, de~ te-l ProPak-X, was constructed as follows.
The ATG in the splice donor/splice acceptor of pCMV plasmid was mllt~tP~l to ACG as
described above. The CMV promoter was excised (EcoRl/XhoI, blunt-ended), and
S replaced with the MoMLV LTR (Asp 718/HindIII, blunt-ended) from plasmid pVH2. The
~3-galactosidase gene was replaced by the gag-pol 0~ (NotI fragment) to generatepMoMLVgp. pMoMLVgp was co-transfected with pHA58 into 293 cells (ATCC CRL
1573) by calcium phosphate co-precipitation and hygromycin B-resistant cells were
selected. Clones were screened for the level of Gag secretion and one clone secreting high
levels of Gag was selected ~cle~1gn~fçd ProGag); this clone yielded high viral titers in
transient transfection.
The plasmid pNZBxeno, contzlinin~ the murine xenotropic env gene, was obtained
from Christine Kozak (NIH). The non-contiguous env sequences in pNZBxeno were made
contiguous by digestion with SalI and EcoRI and ligation into the SalI site of pBluescript
(Stratagene). The xeno e~v ORF was amplified by PCR and cloned into XhoI/XbaI
digested pCI (Promega) to generate the expression plasmid pCI*Ex. pCI*Ex was co-transfected with pSV2pac into the cell line selected as above (ProGag) by calcium
phosphate precipitation and pulolllycin-resistant cells were selected. The resulting cells
were screened for Env e~,e~ion by flow cytometry and clones designated ProPak-X,expressing high levels of Env were screened for ability to produce transducing vector.
Samples of one clone, clone 36 designated ProPak-X.36, have been deposited with the
American Type Culture Collection (ATCC), 12301 Parklawn Drive, Rockville, MD 20852
U.S.A. on December 15, 1995, under the provisions for the Budapest Treaty for the
Deposit of Microorg~ni~m~ for the Purposes of Patent Procedure, and the deposit was
accorded ATCC Accession No.CRL 12007.
Supernatant produced by ProPak-X based producer cells were tested for end-point
titer arld transduction efficiency as described above, except that human 293 cells were used
as the target cells.
In a similar manner to ProPak-X, amphotropic cell lines were also derived by
transfection of the amphotropic envelope-encoding plasmid, pCMV*Ea, into ProGag cells
CA 02238434 1998-06-08
WO 97/21824 PCT~US96/19904
48
and isolation of clones yielding high transduction efficiencies. One of these, clone number
52 (ProPak-A.52), is used extensively in experiments described herein.
T~rkofRCR
The safety of the ProPak-A cells was ~let~rmined by stringent testing for the ability
5 to recombine to generate RCR. In previous work it was found that the vector
BC140revM10 (Bevec et al. (1992) Proc. Natl. Acad. Sci. U.S.A. 89:9870-9874)
reproducibly gave rise to RCR in PA317 cells. BC 140revM 10 carries the extendedpackaging sequence, including the ATG of the gag ORF. The LMINL vector (constructed
as described in Escaich et al. (1995) supra3, in contrast, lacks part of the 5' untr~n~l~t~A
10 region and contains no gag sequences and is therefore less likely to recombine and form
RCR. (See Figure 2). The BC140revM10 or LMTNL vectors were each introduced into
PA317 or ProPak-A cells and culture supernat nts were tested for RCR (using the methods
disclosed in Haapala et al. (1985) J. T~irol. 53:827-833 and Printz et al. (1995) Gene
Therapy 2:143-150) at weekly intervals. The PA317/BC140revM10 combination
15 (transfected or tr~ncdllced) gave rise to RCR detectable by direct inoculation of culture
supernatant onto PG4 cells at 4 weeks (Table 3, below). Cultures were m~int~ined for 4
more weeks, and also tested by co-culture of producer cells with A~: dunni cells to amplify
any RCR in the culture, followed by S+L- assay on PG4 cells. Even by this stringent assay
for RCR, the ProPak-A-based producer pools were all free of RCR (Table 3).
CA 02238434 1998-06-08
WO 97/21824 PCTrUS96/19904
49
Table 3
Assay for presence of RCR in cultures canying the BC 140revM10 or LMTNL vectors
s
p~9rk~ging Transfected Supernatant Co-culture
Cell Line with: Transduced with: RCR (wk) RCR (wk 8)
ProPak-A pBC140revM10 N/A (>8) Negative
ProPak-A pLMTNL N/A (>8) Negative
ProPak-A N/A M(G).BC140revM10 (>8) Negative
ProPak-A N/A M(G).LMTNL (>8) Negative
PA317 pBC 140revM 10 N/A 4 Not tested
PA317 pLMTNL N/A (>8) Negative
PA317 N/A M(G).BC140revM10 4 Positive
PA317 N/A M(G).LMTNL (>8) Negative
RCR detected by S+L- assay on PG4 cells (ATCC CRL 2032) by inoculation with
supernatant from producer cell cultures, or after 3 passages of co-culture with Mus dunni
cells.
10 N/A: not applicable.
(>8): no RCR ~letecte-l 8 weeks after &418-resistant pools established.
M(G).: transient MMLV(VSV-G) pseudotype (Yee et al.(1994) Proc. Natl. Acad. Sci.91:9564-9568) used as inoculurn.
CA 02238434 1998-06-08
W O 97/21824 PCT~JS96/19904
Resi~f~nce to ;nz~ ;v~tion by hllm~n serllm
Recently, interest has arisen in the in vivo application of retroviral gene transfer by
direct ~t1minictration of vector particles into human beings. In addition, targeting of
particles bearing hybrid ligand-ecotropic env genes to specific receptors has been reported
(Kasahara et al. (1994); Science 266:1373-1376 Somia et al. (1995); Proc. Natl. Acad. Sci.
U.S.A. 92:7570-7574; Cosset et al. (1995) J: I~irol. 69:6314-6322). A pre-requisite is that
the particles are not inactivated by human serum. Therefore, ProPak-A.6 or A.52, ProPak-
X or PA317-packaged vector particles were analyzed for susceptibility to inactivation by
human serum. In addition ecokopic supernatants packaged in either PE501 cells
(NIH/3T3-based; see Miller and Rosman, (1989) BioTechniques 7:980-990) or in 293 cells
were analyzed. Vector particles with the various envelopes produced from 293 cells were
resi~t~nt, while supernzlt~nt~ packaged in NIH/3T3 cells were inactivated by incubation
with human serum (Figure 3). Takeuchi et al. (1994) supra, concluded that resi~t~nce of
vector particles to human serum was ~let~-nnined by both the host cell typc and the viral
envelope. The data submitted herein shows that packaging of amphotropic? xenotropic and
ecotropic vectors in 293-based cells is sufficient to confer reei~t~nce to humancomplement.
For the data presented in Examples 2 and 3, herein, the parental packaging cell line
was PA317, unless otherwise noted.
Fx~m~le 2- ~nn~zlri~on of Fnd-Point Titer and Tr~n~ fion Efficiency
For gene therapy applications, it is necessary to generate large volumes of
characterized sup~llld~ll~, which cannot be easily prepared by transient kansfection~
It was therefore necessary to determine the stable end-point titers and the
transduction efficiencies. End-point titers were determined for supernatants from producer
cell clones which had been tr~n~ lced with the PA.LMTNL vector, in which an in~rnsll
thymidine kinase promoter (T in vector name) drives the neomycin phosphotransferase
gene (N). As shown in Table 2B, end-point titers from ProPak-A-based producer cells
were marginally higher than those for the best PA317-based producer clone. In addition,
CA 02238434 1998-06-08
W O 97/21824 PCT~US96/~9904
the titers from ProPak-A.LMTNL producer pools were stable when passaged for 3 months
in the absence of drug selection.
As shown herein, while end-point titers are broadly used, transduction efficiency is
a better measure of gene transfer potency. However, the assay can be laborious with
vectors encoding drug resistance genes. Therefore PA3 17- or ProPak-A-based producer
cell populations carrying a vector (LLySN, Figure 2) derived from the LXSN vector
(Miller and Rosman (1989) supra) were prepared by insertion of the Lyt2 surface marker
gene (Tagawa et al. (1986) supra). Surface expression of the Lyt2 antigen allows simple,
q~ e determination of transduction efficiency by FACS. Higher transduction
e~1~1ciencies were achieved with supernatants from two independently-derived ProPak-A-
LLySN populations than with sUpern~t~n~ from three PA3 1 7.LLySN pools assayed on
NIH/3T3 cells (see Figure 4C). Surprisingly, when the same supernatants were assayed for
transduction efficiency on human 293 cells, the superiority of ProPak-A is even greater
(Figure 4B).
To determine the relationship of end-point titer to transduction efficiency,
supern~t~nts were harvested from producer cells cultured under a variety of conditions to
optimize production of retroviral vector supern~t~nt~. In addition to assaying end-point
titers, the proportion of cells tr~n~ ed after a single inoculation (i.e. transduction
efficiency) also was determined. Figure 5A shows the end-point titers and transduction
efficiencies obtained with 70 different PA317-derived -,B-galactosidase-encoding SVNLZ
(PA.SVNLZ, Figure 2) vector supern~t~nt~ No direct correlation between the end-point
titer and transduction efficiency of the supern~1~n~ was found (correlation factor, r =
0.07). To confirrn that the lack of correlation between tr~n~ çtion efficiency and end-
point titer was not specific to the NIH/3T3 cells or the flow cytometry method used to
quantitate the plO~Ol ~ion of cells transduced with the PA3 1 7-derived SVNL7 vector,
supern~t:~nt~ cont~ining an amphotropic retrovirus encoding the neomycin
phosphotransferase gene (LMTNL, Figure 2) were tested on NIH/3T3 as well as Jurkat and
HeLa cells (Figure 5B). The results demonstrate that the supernatant which yielded the
highest transduction efficiency on NIH/3T3 cells also gave the highest transduction
efficiency on both Jurlcat and HeLa cells (Figure 5~3). Furthermore, the supernatant which
yielded the highest transduction efficiency did not have the highest end-point titer, again
distinguishing between these two functional measurements. Similar results also were
CA 02238434 1998-06-08
WO 97nl824 52 PCT~US96/19904
obtained with the PA.SVNLZ vector on several different cell lines suggesting that the lack
of correlation between end-point titer and transduction efficiency is neither specific to the
indicator gene nor to the target cell species.
Con~ntration of Vector Supern~t:~nt~
The lack of correlation between end-point titer and transduction efficiency also was
cllL when the effect of physically concentrating retroviral vector supernatants by
ultrafiltration was examined. Vector sUpern~tz'nt~ were concentrated using three dirf~ ;nt
ultrafiltration systems. The Sartocon Mini cross flow ultrafiltration system (Sartorius,
Bohemia, NY) was used with a 77.4 cmZ, 100,000 kDa molecular weight cut-off (MWCO)
polysulfone module at a feed pressure of 3 pounds per square inch (psi). Concentration
was achieved within one hour at room temperature. An Amicon Stirred Cell model 8050
(Amicon, Beverly, MA) was used with a YM100 ultrafilter (3.4 cm2, 100,000 kDa
MWCO). Positive ~l'eS:iUlC was m~int~in~d with sterile, filtered regulated air, and
concentration achieved in 30 minlltPs at room temperature. Small volumes (5 to 20 ml),
were concentrated using Filtron 300,000 kDa MWCO centrifugal concentrators (Filkon
Tech. Corp., Northborough, MA). Concentration was achieved within 45 minlltPs bycentrifugation at 3,000 g in a Beckrnan GS-6KR centrifuge ~Beckman, Palo Alto, CA) at
4~C.
Table 4 sllmm~ri7Ps data from five independent experiments using two different
retroviral vectors and the three different ultrafiltration systems. Lnd-point titers increased
in proportion to the volurne reduction (up to 1 9-fold), indicating that the vector was not
inactivated by this procedure. However, higher transduction efficiencies were not
achieved. Equivalent transduction efficiencies were achieved with supern~t~nt~ produced
at 32~C before and after concentration (Table 4). In contrast, retroviral vector supernatants
produced at 37~C had lower transduction efficiencies following ultrafiltration (Table 4).
Interestingly, the lower transduction efficiencies could be restored to the level of the
original supernatant by diluting the concentrates (see Figure 6), suggesting that an
inhibiting agent had been co-concentrated with the transducing particlcs. In principle, the
inhibitor could be non-transducing retroviral particles. non-vector-associated envelope
protein, or a non-viral component of the tissue culture supernatant.
~0296~0~3740
Table 4
CONCENTRATION OF RETROVIRAL VECTORS BY DIFFERENT ULTRAFILTR~TION SYSTEMS
End Point Titer Transduction
(cfu/ml x 10 6) Efficiency (%)
Expt. Temp VectorUlll ~rlllr~lion Volume Before After Before After D
(~C) System Reduction Conc. Conc. Conc. Conc. ~
37 SVNLZSartocon 100kD 8.4x 0.34 2.4 ND 45 r
2 37 SVNLZAmicon 100kD 6.7x 1.6 7.0 25.9 14.3
3 37 SVNLZFiltron 300kD 5.9x 8.0 50.0 17.6 5.7 o
4 32 SVNLZFiltron 300kD 14.3x 0.3 5.7 22.3 21.1
32 LMINLFiltron 300kD 7.5x 2.9 44.0 28.5 27.0
CA 02238434 1998-06-08
W O 97~1824 PCTAJS96/19904
54
Fnvelol;~e-Specific Inhibition of ~etroviral Vector Tr~n~ tion
To address the nature of the inhibiting agent, tissue culture supernatant from
different packaging cell lines or NIH/3T3 cells was tested to determine whether it could
inhibit kansduction. Sup~rn~f~nt~ from packaging cell lines contain all the necessary
5 vector particle pl~L~;illS, but lack vector genomes and hence are unable to transduce cells.
Ecokopic vectors use different receptors to enter cells than arnphotropic vectors and
competition for binding to receptors can occur only between viral particles with the same
kopism. PA.SVNLZ supernatant was mixed with supernatant from the following:
parental arnphotropic PA3 17 packaging cells; sup~ t~nt from the ecokopic GP+E8610 p~qc.k~gin~ cell line; PA.SVNLZ concentrate from above or supernatant from NIH/3T3
cells, and the tr~n~ ction efficiency and end-point titers were measured.
Addition of supernatant from NIH/3T3 cells or thc GPE+86 packaging cell line hadno effect on transduction efficiency or end-point titer (Table 5). In conkast, addition of
either parental PA3 17 or concenkated vector sup~:rn~t~nt derived from producer cell line
15 PA.SVNLZ reduced transduction efficiency, even though in the latter case? the end-point
titer was increased by addition of the concenkated PA.SVNLZ supernatant (Table 5).
These data indicate that inhibition of amphotropic vector transduction by amphotropic
envelope protein in particulate or free for~n.
,
CA 02238434 1998-06-08
W O 97/21824 PCT~US96/19904
Table 5
Comparison of the transductioll of NIH/3T3 cells
achieved with PA.SVNLZ supernatant diluted 1:1 with other supernatants.
Supernatant l~ransduction ~ End-Point ~
F.ff ~ ~ (%)Titer (cfulml x 10~)
~IH/3T3 22.9 ~ 0.5 1.4 + 0.5
GP+E86 (no vector)24.0 ~ 0.5 2.0 + 0.3
PA317 (no vector)12.1 + 0.4 2.5 + 0.3
Conc. PA.SVNLZ*12.7 + 0.4 3.7 + 0.1
* 8.4-fold concentrate with Sartocon 100 kD ultrafilter (Experiment 1, Table 4)
t values given are the average of two samples, and the range.
10 Stabil i$r of Retroviral Vector Particles at Different T~ l dL~res
Concentration of supernatant produced at 37~C reduced trAn~ ction efficiency
relative to the original supernatant (Table 4). It is possible that these supernatants contain
a higher proportion of inactivated vector than supernatants produced at 32~C. To test
vector particle stability at different temperatures, SVNI,Z supe~llaL~llL produced at 32~C
was incubated at either 37~C, 32~C, or 0~C for various lengths of time. The original
supernatant had a transduction efficiency of 25% (see Figure 7A), and after incubation for
24 hours at 37~C, 32~C or 0~C the transduction efficiencies were reduced to 2, 14 and 22%
respectively. Incubation at 0~C for 4 days further reduced the transduction efficiency to
12%, while the end-point titer remained relatively stable (see Figure 6B). At all
20 tenllJe~dLules ex~rnined, transduction efficiency declined more rapidly than end-point titer.
The half-life of vector particles at 37~C was calculated as 4.1 hour based on the end-point
titer, similar to previous reports. The half-life of vector at 32~C and 0~C was 12.0 hours
and 123.4 hours, respectively. It is interesting to note that at all temperatures examined,
end-point titer remained relatively stable (2.5 ~ ().4 x 106 cfu/ml) for at Icast the first eight
CA 02238434 1998-06-08
WO 97/21824 PCTnJS96/19904
56
hours. This initial stability in end-point titer, which was not observed for transduction
efficiency, has been seen in three separate experiments using either the PA.SVNLZ or
PA.LMTNL vector.
Re~roviral Productiorl Kin~tics
Given the greater stability of retroviral vector particles at 32~C than at 37~C,experiments were performed to study the kinetics of vector production at these
tel,lpeldlllres. SVNLZ producer cells were grown to approximately 80% confluency at
37~C, at which time the medium was changed and the cells placed at either 37~C or 32~C.
Supernatant samples were collected at different time points, snap-frozen, and assayed for
both end-point titer (Figure 8A~ and tr~n~dll~tion efficiency (Figure 8B). Over the first 6
hours, similar amounts of transducing vector accumulated in cultures at 3 7~C or 32~C
(Figures 7A and 7B). Calculations incorporating the inactivation rate of the vector at 37~C
or 32~C show that the rate of vector production is slightly higher at 37~C than at 32~C
(0.031 cfu/cell/hour co~ d to 0.027 cfu/cell/hour). Consistent with this, the amount of
p30 capsid or gp70 envelope protein present in the samples measured by ELISA
demonstrated that vector secretion is also marginally higher at 37~C than at 32~C. Similar
results were also found with the LMTNL vector.
While the rate of virion inactivation is lower at 32~C than 37~C, the inactivation of
vector particles at 32~C is still significant (Figure 7A), and the time that supernzlt~nt~
remain at this temperature should therefore be minimi7rd. Experiments were performed to
determine the mi,~ --., time required to produce supern~t~nt~ with maximal transduction
efficiency. Figure 9 gives the time-course of vector production from a confluentPA.LMTNL producer cell culture in a roller bottle at 32~C, and shows that the transduction
efficiency reaches a plateau 3 hours after medium exchange.
l~x~mple 3: Comparison of Production Methods
The kinetics of vector production from PA3 1 7-based producer cell lines suggestthat once producer cells reach confluency, supernatants should be collected every 3 to 5
hours to obtain supernatant with a high transduction efficiency. To achieve these
conditions, a bench scale packed-bed bioreactor was operated under either fed-batch
CA 02238434 1998-06-08
W O 97/21824 PCT~US96/19904
57
(periodic medium exchange) or perfusion mode (continual medium perfusion). In the
latter, supernatant leaving the bioreactor was collected into a container at 0~C to minimi
inactivation. Initial experiments with the PA.SVNLZ producer cell line cultured in the
packed-bed bioreactor operated in fed-batch mode showed comparable vector production
S by end-point titer and transduction efficiency to that in tissue culture flasks. Next, the
packed-bed bioreactor was operated in perfusion mode and compared to a roller bottle
operated in fed-batch mode (Table 6). Sarnpling of supernzltzlnt~ from the bioreactor or
roller bottles began one day after cells were seeded. In both cultures, incubation was
initially at 37~C and lowered to 32~C after 50 hours. The cells in both cultivation systems
10 reached confluency after approxim~tely 100 hours of incubation. Tr~n~hlction efficiencies
achieved with supern~t~nt~ produced in the packed-bed bioreactor were greater than those
from roller bottle cultures (Figure 1 OA). However, supernatants produced in the packed-
bed bioreactor had lower end-point titers than supern~t~n1~ from the roller bottle (Figure
1 OB) possibly due to the shorter residence time of the medium in the reactor (5.75 hours
compared to 24 hours for the roller bottle). Over a 5 day period a total of 1 OL of
supern~t~nt was collected using the packed-bed bioreactor, compared to lL from the roller
bottle.
Table 6
Operating Pa~ameters of Dirr~ .t Production System~
Parameter Packed Bed B~oreactor* Roller Bottle+
Surface Area (cm2) 12,000 1,700
Volume (ml) 500 200
Mode of Operation perfusion fed-batch
Dilution Rate (Vol./day) 4.2 1.0
Total Production Vol. ~ml) 10,000 1,000
Seeding Density (cells/ml) 6.5 x 105 2.3 x 105
Final Density (cells/ml) 5.91 x 106 9.06 x 106
End-Point Titer (cfi~/ml) 1.45 x 1 o6 2.40 x107
Transduction Ef~ciency (%) 36.0 29.5
~ 1 0g of Fibercell discs in bench-top New Brunswick Scientific bioreactor
+ ~orning exT1~n~ 1 surface area roller bottle
CA 02238434 1998-06-08
WO 97~182~ 58 PCT~US96/19904
F~mple 4: Corr~rison of VectorProduction from ProPak-A and P~17 Cells Super~t~ntProduction from Bioreactor
1 Q g of the New Brunswick Scientific Fiber Cell disks (catalog #M 1176-9984)
were placed into the spirmer basket of the packed-bed bioreactor (catalog #M 1222-9990)
S and washed several times with PBS before autoclaving. Prior to see-lin~, the bioreactor
and Fibercell discs were washed twice in medium c- ntzlining 5% FBS.
The reactor was seeded with 2.4 to 3.6xl o8 producer cells (sarne for ProPak andPA317) in a total volume of 500 ml of mediurn (i.e. 2 to 3xl o4 cells/cm2) (DMEM with 5%
to 10% FBS). The agitation rate upon seeding was set at a~ xilllately 80 rpm. A
10 minimum of 3 hours was required for all of the cells to become attached to the Fibercell
dlscs.
Approximately 4 to 18 hours after seeding, 50% of the medium was exchanged,
and 5ml of Pluronic F68 (Sigma cat# P5556) was added to the bioreactor to a final
concentration of 0.1 %. At this time aeration was initiated by gently sparging medical
grade air/5% CO2 mix (Altair cat#39222) into the reactor through a 0.2 mm filter.
Agitation was increased to approximately 250 rpm. Cells were allowed to grow for three
days at 37~C and then the temperature was lowered to 32~C. Medium exchanges should be
perforrned on the cells every 4 to 12 hours, or the reactor operated in perfusion-mode. The
perfusion rate was 2 to 6 reactor volumes per day. Perfused supernatant harvested from the
20 bioreactor was collected into 2L glass vessels (vented cap) kept on ice. The harvested
sUpern~t~nt was filtered through a 0.45,um filter, aliquoted, and snap frozen in MeOH/dry
ice. Snap-freezing can be perforrned in other ways such as in liquid nitrogen. Thc
bioreactor and set-up of the continuous perfusion operation are shown in Figures 11 and
12.
Sup~ P~ Lion from Roller Rottles
Producer cells were seeded at 3x104 cells/cm2 in 0.25 ml of DMEM plus 5% FBS
per cm2 of surface area. The rotational speed was set at 0.6 rpm. Cells were grown at 37~C
for two days. On day three the temperature was decreased to 32~C. The medium wasaspirated and replaced with fresh DMEM medium plus 5% FCS (0.2 ml/cm2 of surface30 area). Medium was ad~ed with a pipette to the bottom of the flask to prevent cells from
CA 02238434 1998-06-08
W O 97/21824 PCT~US96/19904
59
peeling off and to avoid bubble formation and cell contact. The serum level may be
lowered to 2% FBS.
For PA3 1 7-based producers, two batches of medium were collected every day
(every 12 hours) and pooled to make a single lot. For ProPak-A based producers,
S supernatants were collected ever.,v 12 to 24 hours. Each batch was filtered through a
0.45~1m filter immediately, and kept at 4~C until batches were pooled (preferably no longer
than about 12 to 24 hours). The pooled batches were aliquoted, snap frozen, and stored at -
80~C. Supernatant collection can continue as long as producer cells are healthy.Figure 13 shows the comparison of vector production from ProPak-A cells culture
10 in T-flasks and in the packed-bed bioreactor with aeration as assayed on 293 cells. The
supernatant produced using the packed bed bio-reactor mediates considerably higher
kansduction than that produced in T-flasks. Figure 14 shows that ProPak-X cells also
produced higher tr~n~d~lction efficiency supern:~t~nt when cultured in the packed-bed
bioreactor compared to T-flasks. Dilution of vector sup~ t:~nt produced by ProPak-A
15 under different culture conditions is shown in Figure 15. Again, the aerated packed-bed
bioreactor produced higher quality vector supern~nt~ as (let~rmin~d by transduction
efficiency.
F.x~n~rle 5: Prim~ry Cell Transductions
Cell selection ~3n~1 ~n~lysis
Apheresed samples were obtained with informed consent from multiple myeloma
or breast cancer patients. Stem cells were mobilized into the peripheral blood by treatment
with cyclophosphamide (Cytoxan) and GM-CSF ~multiple myeloma), G-CSF above
(breast cancer), or Cytoxan + VP-16 + CDDP + G-CSF (breast cancer). Apheresis for total
white cells was started when the peripheral blood white cell count was greater than 500
cells/ml and the platelet count was greater than 50,000 cells/ml. Patients were apheresed
daily until 6 x 108 mononuclear cells (MNC) were collected.
The cells were washed twice in PBS (partially depleting platelets), and CD34+ cells
were positively selected using a Baxter IsolexTM cell selector. The recovered cells
averaged 85% to 99% CD34+ purity as determined by FACS analysis.
CA 02238434 1998-06-08
WO 97/21824 PCT~US96/19904
Tr~n~ tion
CD34+ cells ~0.5 x 1 o6) were suspended in 0.5 ml of freshly thawed retroviral
supernatant and diluted 1:1 in Whitlock/Witte medium (50% IMDM, 50% RPMI 1640,
10% FCS, 4x l O-s M 2-mercaptoethanol, 1 OmM HEPES, 100 ,u~ml penicillin, 100 mg/ml
streptomycin, and 4mM glutamine) conrz1ining cytokines at the following final
concentrations: c-kit ligand (Amgen) 100 ng/ml or leukemia inhibiting factor (LIF,
Sandoz), IL-3 (Sandoz) 20 ng/ml; IL-6 (Sandoz) 20 ng/ml. Protamine sulfate was added at
a final concentration of 4 ,ug/ml or polybrene was added at a final concentration of 8 ,ug/ml.
The cells and vector were centrifuged at 2800 x g, 33~C to 35~C, for three hours. The cells
were resuspended in medium with cytokines and cultured for three days. After three days,
cells were harvested and approximately 4 x l 05 cells used to determine bulk transduction
eff1ciency, and plated in methylcellulose to determine progenitor cell transduction
efficiency (see below).
For the LLySN and LMily ~ectors, which contain the murine Lyt-2 marker gene,
bulk transduction efficiency was determined by staining with APC or PE conjugated anti-
Lyt-2 (Ph~nningen) and sulfurhodarnine conjugated anti-CD34. The results are shown in
Figures 1 6A and 1 6B. Vector supernatant from either ProPak-A or ProPak-X transduced
primary human cells with greater efficiency than that from PA3 17 cells. Furthermore, the
combination of amphotropic and xenotropic vector can be useful to deliver two vectors to
the same cell since they do not compete for the same receptor.
Methylce]lulose assay
Cells from each transduction (2.5 x 103 to 10 x 103) were added to 4 ml of
methylcellulose medium (Stem Cell Technologies) plus 1 ml IMDM con~inin~; the
following cytokines (final concentration): c-kit ligand 100 ng/ml; GM-CSF (Amgen) 10
ng/ml; IL-3 10 ng/ml; IL-6 10 ng/ml; rhEPO (Amgen) 2 units/ml. 1.0 ml of the
cell/cytokine methylcellulose mixture was plated onto five 35 mm plates using a 5 ml
syringe and 16 gauge needle, and the plates were placed in a 37~C incubator for 2 weeks.
After 14 days, single methylcellulose colonies were picked and analyzed for the
presence of neo or RevM 10 by PCR.
-
CA 02238434 1998-06-08
W O 97/21824 PCTrUS96/19904
61
An~lysis
Individual colonies were picked by aspiration in 5111 and transferred to 25 ,ul to 5
,ul of PCR Iysis buffer. PCR lysis buffer is a 1 :1 mixture of buffer A ( 100 mM KCl, 1 ()
mM Tris pH 8.2, 2.5 mM MgCl2) and buffer B (10 mM Tris pH 8.3, 2.5 mM MgCl2, 1%
Tween 20, 1% NP40, 100 ~ll/ml proteinase K). The mixture was allowed to incubateovernight at 37~C or for 2 h at 56~C. The proteinase K was inactivated by heating at 94~C
for 10 to 30 minutes, and 5-10, Ll of the Iysate was used for the PCR reaction.
The PCR reaction arnplified a 1 OObp fragment of the ~-globin gene and either a
240bp fragment of the neo gene or a 1 80bp fragment of RevM10, depending on the vector
used. The primers used were as follows.
Rev: 5' TCgATTAgTgAACggATCCTT 3' (SEQ ID NO: 5)
5' CTCCtgACTCCA~T~TTgCAg 3' (SEQ ID NO: 6)
Neo: 5' TCgACgTTgTCACTgAAgCg 3' (SEQ ID NO: 7)
5' gCTCTTCgTCCAgATCATCC 3' (SEQ ID NO: 8)
Beta-globin: 5' ACACAACTgTgTTCACTAgC 3' (SEQ ID NO: 9)
5' CAACTTCATCCACgTTCACC 3' (SEQ ID NO: 10)
The reactions were performed in a 40 ~ul final volume in a Perkin Elmer thermal
cycler 9600 as follows: 5 min denaturation at 94~C; 40 cycles of 30 sec at 94~C, 30 sec at
62~C, 1 min at 72~C; and 10 min at 72~C. PCR products are visualized by ethidiumbromide agarose gel electrophoresis (Sambrook et al. (1989) supra) and the PCR products
confirmed by Southern blot hybridization. A sample was considered positive if both the
Rev or neo band and the ~B-globin band were present. The results, summarized in Table 7,
show that for two different tissues (MPB and ABM), the ProPak supern~t~nt~ performed
better than the PA3 17 supern~t~nt~.
The results in Figure 16A show that both the xenotropic ProPak-X (X.36) and the
arnphotropic ProPak-A.6 were able to produce retroviral vector ~ dla~ions which
tr~ns~ ce~ human hematopoietic stem/progenitor cells with higher efficiency than that
produced from PA3 17 cells.
202962003740
o
Table 7
S
METHYL CELLULOSE DATA--CD34~ CELLS
(NumberofColonies) D
TISSUE LABELPACKAGING VECTOR TESTED POSITIVE TRANSDUCTION PCR TARGET
CELLS EFFICIENCY(%) SEQUENCE ~r
MPB Ampho PA317 LLySN 24 4 17 neo o, r
MPB Xeno PP-X.36 LLlySN 24 7 29 neo N
MPB Ampho PP-A.6 LLySN 24 11 46 neoABM Ampho PP-A~6 LMiLy 56 13 23 rev ~
ABM Ampho PA 317 LMiLy 56 4 7 rev o
Key: MPB: Mobilized penI her~l blood
ABM: Adult bone marrow from a cadaver
neo: Neomycin
rev: Rev Ml0
c
CA 02238434 1998-06-08
W O 97/21824 63 PCT~US96/19904
The following series of examples describe the generation of high quality vector
supern~nts with MLV-xenotropic, MLV-amphotropic or GaLV envelope and the abilityof these different enveloped vectors from human or murine-based p~ck~ging cells to
tr~n~ ce primary human hematopoietic cells. The complementary tropisms and safety of
5 the ProPak-A and ProPak-X cell lines were also exploited to produce vector supern:~t~nt~
by co-culture. It was demonstrated that after inoculation of hematopoietic stem and
progenitor cells (HSPC) with co-culture supern~t~nt~ conts~inin~ both amphotropic and
xenotropic vector particles, 100% of the colony-forming progeny contained transgene.
Expression of transgene in cells with a phenotype characteristic of hematopoietic
10 progenitor cells was also demonstrated.
Fxz~mple 6: Derivation of p~ck~ Cells
To facilitate derivation of packaging cell lines with a variety of envelopes, a cell
line expressing the MLV Gag-Pol functions was first constructed. The gag-pol gene was
expressed from the MMLV-LTR promoter, the CMV-IE promoter, or the Rous Sarcoma
Virus LTR (Figure 1). The gag-pol open reading frame (ORF) (~igg et al., (1996) supra)
was sub-cloned into the expression vector pMLV*. This plasmid was derived from
pCMV* (Rigg et al. (1996) supra) by replacement of the CMV promoter with the MLVLTR (M-LTR) promoter. The envelope protein ORF's were sub-cloned into the restriction
sites shown in plasmid pCI (Promega, Madison, WI) which contains the cytomegalovirus
immediate early promoter (CMV-IE), a chimeric intron (SD/SA), and the simian virus 40
late polyadenylation se~uence (pA).
The xenotropic envelope (Ex) gene was obtained by PCR from a linear template
derived from plasmid pXeno (O'Neill et al. (1985) J. Virol. S3:100-106) using the
following oligonucleotide primer pair:
5'-ACCTCGAGCCGCCAGCC~GAAGGTTCAGCGTTCTC-3' (SEQ ID NO: 11) a~d
5'-AATCTAGACttaTTCACGCGATTCTACTTC-3' (SEQ ID NO: 12). The underlined
ATG corresponds to nucleotides 291 to 293, and lower case letters denote the stop codon,
nucleotides 223 to 225 (O'Neill et al. (1985) supra).
The chimeric envelope (Eax) gene was constructed by replacing the sequence
between the Apa I and Bgl II sites of the amphotropic envelope gene (nucleotides 772 to
963; on et al. (1990) J Virol. 64: 757-766) with a synthetic DNA corresponding to the
CA 02238434 1998-06-08
WO 97~1824 PCT~US96/1990
64
sequence of the xenotropic envelope gene from 1 OAl (nucleotides 801 to 965; Ott et al.,
(1990) supra).
Plasmids encoding Gag-Pol were cotransfected into 293 cells which are free of
MLV-like sequences, which are free of endogenous MLV-like sequences (Rigg et al.5 (1996) supra). Transient Gag expression was detected by ELISA in all cases, but it was
only possible to isolate clones stably expressing Gag at levels equivalent to the ProPak-A.6
cells (Rigg et al. (1996) supra) with the MLV-LTR driven Gag. Transient cotransfection
with vector and envelope plasmids of the 293/Gag-Pol cells (called ProGag) yielded higher
titer supernatants than cotransfection of Anjou cells (Pear et al. (1993) Proc. Natl. ~4cad.
Sci. USA 90: 8392-8396).
Next, expression plasmids bearing the genes for the xenotropic (Ex, Figure 1) oramphotropic (Rigg et al., (1996) supra) envelope proteins were stably introduced into
ProGag cells, and clones ~ e~ing either xenotropic (called ProPak-X) or amphotropic
(ProPak-A.52) envelope proteins were isolated by flow cytometry. In turn, the vectors
15 LLySN and LMiLy (Figure 2) were transduced into these cells, and the resulting producer
cells were used to generate the supern~t~ntc used in this study.
Retroviral vectors LMiLy and LLySN have been described (Rigg et al., (1995)
supra and Rigg et al, (1996) ~7-ol. 218: 290-295). LMiLy and LLySN encode the Lyt2
surface antigen (Ly) (Tagawa et al., (1986) Proc. Natl. ~lcad. Sci. USA 83: 3422-3426)
20 expressed either directly from the retroviral LTR (L) promoter in LLySN, or via an internal
ribosomal entry site (i) in LMiLy. The p~ck~ging~psi) sequence in LMiLy is to nucleotide
566 (Shinnick et al., (1981) Nature 293: 543-548), and LLySN has the longer psi sequence
and non-functional ATG (Bender et al., (1987) J T~irol. 64: 1639-1646, Miller and
Rosman, (1989) Biotech~ziques 7: 980-990), and also contains the neomycin
25 phosphotransferase gene (N) expressed from an internal SV40 promoter (S). Unless
indicated other~,vise, LLySN supern~t~nt~ were prepared in T-flasks from G418 resistant
cell populations, and LMiLy supernatants were prepared from producer cell clones in the
sparged packed-bed bioreactor.
In contrast to Ex and Ea envelopes (Figure 1), no cells stably expressing either the
30 chimeric Eax (Figure 1) or GaLV envelope proteins could be isolated in repeated attempts
since cell lysis resulted While transient supernatants could be prepared by cotransfection
of ProGag cells with the Eax and vector expression plasmids, no stable expression of the
CA 02238434 1998-06-08
W O 97/21824 PC~US96/19904
6S
Eax envelope was achieved in 293, ProGag or HT1080 cells. Most dramatic was
transfection of the GALV envelope-encoding plasmid which resulted in total Iysis of
ProGag or 293 cell cultures within 24 hours. Transient GaLV supernz~t~nt~ could not be
produced with cotransfected ProGag cells although the GaLV envelope gene was
functional, since vector particles were released if murine cells were cotransfected.
Fx~rr~le 7: Dete~ ion of Tropism
To confirm the phenotype of vector particles bearing the amphotropic, xenotropicor chimeric arnpho/xeno envelope~ supernatants were inoculated onto cell lines from
different species chosen to distinguish between the various envelope tropisms. ~ells were
1 û inoculated with vector particles packagcd in the new ProPak-X (xenotropic) or
ProPak-A.52 (amphotropic) cell lines, or with vector supernatants from cotransfection of
vector and the expression plasmid encoding the Eax envelope. For comparison, MLVparticles pseudotyped with the GaLV (from PG13 cells) glycoprotein or the G protein of
vesicular stomatitis virus (VSV-G, Yee et al., (1994) Proc. Na~ cad Sci. USA 91: 9564-
9568) were also included. The tropism observed for the xenotropic, amphotropic and
GaLV envelopes (Fig. 20A) corresponded with those established for retroviruses (Teich,
(1984) Cold Spring Harbor Laboratory).
By analogy to the genotype and receptor tropism of MLV-lOAl (Ott et al., (1990)
supra; Wilson et al., (1994) J. Virol. 68: 7697-7703; and Wilson et al. (1995) J. Yirol. 69:
534-537), the Eax envelope should mediate retroviral vector binding to both the
amphotropic and xenotropic receptors. Vector particles bearing the Eax envelope
transduced quail cells, which are resistant to amphotropic particles but susceptible to
xenotropic vector (Figure 17B), and also transduced murine NIH/3T3 cells which are
resistant to xenotropic vector (Figure 17A). Therefore, this modification of thearnphotropic envelope protein conferred specificity for a cell line that cannot be transduced
with arnphotropic vector~ establishing the extended tropism. As expected, all cell types
were transduced with the pantropic MLV(VSV-G) vector particles. Lower transduction
efficiencies were achieved with this pseudotype, possibly a result of having to prepare
supern~t~nt~ by transient cotransfection because of the toxicity of the VSV-G protein.
CA 02238434 1998-06-08
W O 97/21824 66 PCT/US96/19904
Fx~mrle 8: ~esistance to Human Serum
It was shown above that amphotropic and ecotropic particles generated from 293
cells were not inactivated by treatrnent with human serum under conditions whichinactivate amphotropic vector packaged in murine PA317 cells. The xenotropic vector was
5 similarly tested for resi~t~nce and the xenotropic particles were found to not be inactivated
by treatment with human serum (Figure 18). Vector particles from the ProPak-A.52packaging clone were also resistant, while the PA317-packaged vector was almost
completely inactivated (Figure 18), as has been previously demonstrated (Takeuchi et al.,
(1994) ~ Virol. 68: 8001-8007; Rigg et al., (1996~ supra).
~0 Fx~mrle 9: RCR Generation Test
The Gag-Pol p~ck~gin~ functions in ProPak-X and ProPak-A.52 cell lines are
expressed from the MLV-LTR promoter and these cells therefore carry more MLV-derived
sequences than the ProPak-A.6 cells. However, generation of RCR due to recombination
of MLV-derived sequences is unlikely since a minimum of three recombination events
15 would still be required. Only 32 nucleotides of the R sequence are retained in the LTR
promoter used, a relatively short sequence homology with vectors. Nevertheless, cultures
and supern~t~nt~ were stringently tested for RCR. C o-cultures of ProPak-A and ProPak-X
cells carrying the vector BC140revM10 were m~int~ined. This vector, as before (Rigg et
al., (1996) supra~, rapidly generated RCR in PA317 cells (Table 8). Although cultures
20 were m~int~ined for up to 3 months, in no case was RCR detectable either in sup~ at~lls
or cells from cultures cont~ining ProPak cells (Table 8).
CA 02238434 1998-06-08
W O 97/21824 PCT~US96/19904
67
Table 8. Tests for RCR in producer cell cultures or co-cultures
p~kzl~ing Cell Supernzlt~nt Co-culture
Expt. Line(s) Vector Carried RCR (wk) RCR (wk 12)
PA317 LMTNL (> 12) ND
PA317 BC140revM10 6 ND
Pro-Pak-A.52 LMTNL (>12) ND
ProPak-A.52 BC140revM10(>12) ND
ProPak-X LMTNL (>12) ND
ProPak-X BC140revM10(>12) ND
PP-A.52/PP-X* LMTNL (>12) negative
PP-A.52/PP-X* BC 140revM 10(> 12) negative
2 PA317 BC 140revM10 6 ND
ProPak-A.52 BC140revM10(>12) negative
ProPak-X BC140revM10(>12) ND
ProPak-A.6 BC140revM10(>12) negative
PP-A.52/PP-X* BC140revM10(>12) negative
ND: not ~letermined
~>12): no RCR detected 12 weeks after G418-resistant pools established.
*Co-culture of PP-X and PP-A producer cell populations.
5 Vectors were introduced into the p~qck~;ing cells by spinoculation at 1400g with MLV(VSV-
G) pseudotype supern~t~nt~ produced by transient transfection of ProGag cells with vector
and VSV-G expression plasmids (Yee et al., (1994) supra). Populations of producer cells
were selected for G418 resistance and passaged every 3 or 4 days. RCR was ~letected by
S+L- assay on PG4 cells (ATCC CRL 2032) by inoculation with supern~t~nt from producer
cell cultures, or after 3 passages of co-culture with Mus dunni cells (Forestell et al., (1995)
supra; Printz et al., (1995) Gene Iher. 2: 143-150) in the presence of 2,ug/ml polybrene.
F~mple 10: Improved Retroviral Veçtor Supernatant ~roduction in a Bioreactor
Applying principles of retroviral vector production defined with PA317-based
producer cell lines (Forestell et al., (1995) supra). vector production from ProPak cell lines
in a packed-bed bioreactor was investigated. Using the PP-A.52.LMiLy producer cell linc,
the packed-bed bioreactor operated in fed-batch mode was compared with a roller bottle
and T-flask for production of vector supernatant. The effect of supplemental sparging of
_
-
CA 02238434 l998-06-08
WO 97/21824 PCT~US96/19904
68
an air/CO2 mix into the bioreactor was also examined to deterrnine if oxygen transfer from
the bioreactor head-space was limiting cell growth and vector production. Supplçmentz-l
aeration was achieved by direct micro-sparging of a medical grade mix of air with 5%
CO2, and 0.01% Pluronic F-68 (Sigma, St. Louis, MO) was added to prevent cell damage
from shear. Vector supernzlt~nt~ harvested from the different production vessels were
compared for their ability to transduce NIH/3T3 cells (Figure 19). Supernatants harvested
from the T-flask and roller bottle cultures yielded transduction efficiencies of 62% and
63% respectively, while supernatant from the non-sparged packed-bed bioreactor yielded a
transduction efficiency of 78%. Gene transfer was greatest (100%) with vector produced
in the sparged packed-bed bioreactor i~licating that head-space oxygen transfer alone was
limiting vector production and cell grow~h in the non-sparged bioreactor. Measurements
of the glucose and glutamine concentrations indicated that these key nutrients were not
limiting in any of the cultures. In this particular experiment, vector production was also
analyzed by measurement of the viral envelope and Gag proteins by ELI~ to determine if
the improved transduction efficiency was due to increased vector production. Although
the final volumetric cell density in the sparged packed-bed bioreactor was slightly lower
than in the T-flask (4.0x106 and 4.7x106 cells/ml respectively), the vector supernatants
from the sparged packed-bed bioreactor yielded higher levels of viral protein as well as
higher transduction than T-flask supernatants. These results indicated that the improved
production in the sparged packed-bed bioreactor was due to increased cell-specific vector
productivity.
PA317,PG13, ProPak-A.6, ProPak-A.52, and ProPak-X based produced cell lines
were cultured in the packed-bed bioreactor under sparged conditions. In every case, a 2- to
20-fold improvement in vector production was achieved in comparison to T-flask cultures.
In particular, it was determined if the increased gene transfer into cell lines could
also be achieved within primary cells. In addition to the higher transduction of NIH/3T3
cells (Figure 19), ProPak supernatants produced in the sparged packed-bed bioreactor
yielded approximately 3-fold higher transduction efficiencies with CD34-positive cells
than vector preparations from T-flask cultures (Table 9). Using two different assays (Lyt2
3 0 expression in bulk cultures or PCR analysis of CFU-C), different ~stim~t~s for gene
transfer efficiency were obtained; however, the relative numbers were consistent.
CA 02238434 1998-06-08
W O 97/21824 PCT~US96/199~4
69
Table 9. Tr~n~ t~on of CD34-positive cells from MPB(normal donor): comparison of T-
flask and packed-bed bioreactor supernatants.
Transduction Efficiency (%)
T-Flask Packed-Bed
Lyt2 CFU-C Lyt2 CFU-C
ProducerCellLine Expression Marking Expression Marking
ProPak-A.LMiLy 1.3 5.3 (7/132)3.g 15.0 (22/147)
ProPak-X.LMiLy 3.8 18.1 (27/149) 9.9 43.3 (49/113)
Cells were inoculated once, and Lyt2 e2S~res~iion was measured 3 days later. Gene m~rking
of CFU-C, determined by P~l~, is expressed as a percentage and values in brackets are the
nurnber of RevM10-positive colonies divided by the total number that yielded ~-globin
signals, as described above.
Exz~mple 11: Pr;mRry Cell Transduction with Different Vector Tropisms
This ~xample compared tr~n~dTl~tion of CD34-positive hematopoietic progenitor
cells or CD4-positive PBL with vectors of different tropisms (amphotropic, xenotropic,
GaLV) prepared from stable producer cell lines. Cells were exposed to vector once to
directly compare transduction efficiencies which are presented as Lyt2 surface marker
expression relative to the expression achieved with PA317-packaged vector preparations
(Table 10). Regardless of tropism, all vector types successfully trz~ncd~ e-1 CD34-positive
cells isolated from MPB or ABM, or CD4-positive PBL (Table 10). This implies that the
receptors for all three vector types are expressed on these cells. While no single vector
tropism appeared to mediate significantly higher levels of transduction than any other, the
highest transduction efficiencies were achieved with ProPak-A or ProPak-X supernatants
~ (Table 10). Although derived by different means, comparable transduction was achieved
with vector from either ProPak-A.6 or ProPak-A.52 clones (Table 10). Also in these
comparisons, amphotropic vector supern~ts~nt~ from the human ProPak-A cell linesconsistently transduced a higher proportion of target cells than amphotropic vector
prepared from PA317-based producer cells (Table 10).
CA 02238434 l998-06-08
WO 97~1824 PCT~US96/19904
The phenotype of the transduced progenitor cells was determined by flow
cytometry. The majority of cells expressing the I,yt 2 gene exhibited a phenotype
characteristic of early hematopoietic progenitor cells, that is, CD34 antigen was expressed
(Figure 20, paneI B) in the absence of antigens specific for differenti~t~d hematopoietic
S cell lineages (Figure 20, panel D).
Since no single tropism appeared to preferentially tr~n~duce primary cells, cells
were also inoculated with a mixture of vector packaged separately in ProPak-A orProPak-X cells, to test whether higher transduction could be achieved by simultaneous
inoculation with vectors targeting distinct receptors. In all but one of rlve cases ~PP-A &
10 PP-X mix; Table 3) slightly higher transduction was achieved with the mixture (Table 10).
It is possible that in these experiments the concentration of vector in the supernatants may
be limiting, since individual supernatants were diluted 4-fold in the mixture, compared
with 2-fold dilutions where single inocula were used.
CA 02238434 1998-06-08
W O 97/21824 71 PCTnUS96/19904
Table 10. Transduction of primary human hematopoietic cells with vector supernatants of
different tropisms from T-flasks (A, LLySN vector) or the sparged packed-bed bioreactor
(B, LMiLy vector).
A. T-Flask/LLySN Relative Lyt2 Expression
CD4+ CD34+ CD34+ CD34+
Producer Cell Line Base PBL MPB (a) MPB (b) MPB (c)
PA317 1.0 1.0 1.0 1.0
ProPak-A.52. 2.4 1.6 1.8 1.6
ProPak-A.6 2.7 2.0 2.4 1.7
ProPak-X 2.7 1.5 1.9 1.2
PG13 2.6 1.1 1.3 ND
ProPak-A.52 & X mix ND 2.0 ND ND
B. Bioreactor/LMiLy Relative Lyt2 Expression
CD4+ CD34+ CD34+ 34+/Thy-1+
Producer Cell Line Base PBL ABM MPB (b) MPB (d)
PA317 1.0 1.0 1.0 1.0
ProPak-A.5.2 3.9 ND 2.5 ND
ProPak-A.6 3.0 2.5 3.2 1.5
ProPak-X 5.8 2.1 2.0 1.9
PG13 1.2 0.7 0.9 ND
ProPak-A.52 & X mix 8.1 2.8 3.7 2.8
s
Transduction of different primary cell types achieved after a single inoculation with
supernatants cont~ining the LLySN vector (A), or the LMiLy vector (B). Tr~ncd~lction
efficiencies were mcasured as the proportion of Lyt2-expressing cells, and values have
been norm~li7P-l to that achieved using PA317-based supernatants. Hematopoietic cell
10 populations isolated from six different tissues were inoculated, specifically, CD4+ PBL, or
- CD34+ cells selected from: ABM or MPB from 2 breast cancer patients (a and c), a
multiple myeloma patient (b), or a norrnal donor (d). ND=not determined.
CA 02238434 1998-06-08
W O 97/21824 PCTAJS96/19904
72
F~c~mple 12: Vector Production in Producer Cell Co-Cultures
In attempts to increase the transduction efficiency, vector supernz-t~nt~ were
produced from a co-culture of the complementary ProPak-X.LMiLy and
S ProPak-A.52.LMiLy producer cells in the packed-bed bioreactor. This technique, known
as ping-pong amplification, results in higher titers with murine and avian producer cell
lines, probably as a result of increased vector copy number (Bodine et al., (1990) Proc.
Natl. Acad. Sci. USA 87: 3738-3742, Cosset et al., (1993) Yirol. 193: 385-395, Hoatlin et
al., (1995) ~ Mol. Med. 73: 113-12Q). The problem that has arisen in the past with
10 co-culture production is the generation of RCR (Bodine et al., (1990) supra; Cosset et al.,
(1993) supra; Muenchau et al., (1990) Virol. 176: 262-265). However, it was already
shown herein that no RCR arises during extended co-culture of ProPak-A and ProPak-
~cells carrying the BCl40revM10 vector (Table 8), an event that is even less likely with the
LMiLy vector which lacks sequences that overlap with the ProPak p~(~k~ging constructs.
15 Nevertheless, supernatants and end-of-production cells were exhaustively tested and found
to be free of RCR by skingent assays (incubation or co-culture with Mus d2~nni and S+L-
assay on PG-4 cells). In addition, the presence of both amphotropic and xenotropic vector
particles was confirmed by transduction of permissive and restrictive cell lines.
The ability of PA317, ProPak-A, ProPak-~ or ProPak-A/X.LMiLy vector
20 superns-t~nt.~ to tr:~n.~ .e CD34-positive cells upon a single spinoculation was first
compared. Cell samples were analyzed for Lyt2 expression, and the presence of the
revM10 transgene in comrnitted progenitor cells was determined by PCR. Consistent with
Lyt2 expression assays ~Table 10), supern~t~nt~ from the human-based ProPak-A or -X
producers tr~n~ ce~l cells more efficiently than PA317-packaged vector, and the highest
25 level of transduction was achieved with the ProPak-A/X supernatant (Table 11).
Using the ProPak-A/X.LMiLy supern~t~3nt, different cytokine combinations and
multiple rounds of inoculation were investigated in an effort to m~Yimi7~ gene transfer
efficiency. CD34-positive cells were spinoculated with PP-A/X.LMiLy vector once, twice
on the same day, or once a day on two consecutive days in the presence of IL-3~ IL-6, and
30 either LIF or SCF. Prior to the first spinoculation, cells were incubated for 1 day in
20ng/ml each of IL-3 and IL-6, and 50ng/ml SCF or LIF as described above. The results
in Table 12 show that while gene transfer was higher in cultures treated with SCF after a
-
CA 02238434 1998-06-08
W O 97/21824 PCT~US96/19904
73
single inoculation, gene transfer into CFU-C was 100% for cultures inoculated twice on
consecutive days in either LIF or SCF. F~urtherrnore, in both cytokine cocktails the Lyt2
transgene was expressed in up to 40% of the total cell population, and in 35% of the cells
which were also CD34-positive.
S
Table 11. Transduction of CD34+/Thy+ cells from MPB: comparison of vector
supc, .,t.l~.~t~ from different producer cell lines.
Transduction Efficiency (%)
Producer Cell LineLyt2 Expression CFU-C
PA317.LMiLy 2.8 12 (27/219)
ProPak-A.6.LMiLy 4.1 24 (49/206)
ProPak-X.LMiLy 5.2 25 (35/141)
ProPak-A.52/X.LMiLy 7.9 36 (65/182)
10 Transduction efficiencies achieved with vector supernatants prepared in the sparged
packed-bed bioreactor from the producer cell lines shown. CD34+/Thy+ cells from normal
MPB were inoculated at unit gravity for 1 hour followed by a 3 hour spinoculation.
Tran~cl~ tion efficiency is given as the ~xpression of the Lyt2 surface antigen or the
m~rkin~ frequency of individual methyl cellulose colonies (C~U-C; Table 9).
CA 02238434 1998-06-08
O 97/21824 PCTnJS96/19904
74
Table 12. Tr~n~ ction of CD34+ cells from MPB: comparison of inoculation protocol and
cytokine cocktail.
Transduction Efficiency (~/0)
IL-3, IL-6, LIF IL-3, IL-6, SCF
Inoculation Lyt2 CFU-C Lyt2 CFU-C
Protocol Expression PCR Expression PCR
once 13.3 51 (47/92) 14.5 74(68/92)
twice, sarne day 22.7 86 (79/92) 22.3 100(92/92)
twice, c~n~ec-ltive days 40.4 100(92/92) 40.4 100(92/92)
Transduction of CD34+ cells (MPB) with packed-bed bioreactor supernatant from a co-
culture of ProPak amphotropic and xenotropic LMiLy producer cells. Cells were
spinoculated and incubated as shown in the presence of IL-3, IL-6, and LIF or SCF.
Expression of Lyt2 was determined 2 days after the last inoculation, and the mzlrkin~
frequency of individual methyl cellulose colonies (CFU-C) was deterrnined (Table 9).
Table 13. Characteristics of the ProPak Cell Lines
Packaging Tr~nc-luctiorl Human Expression PlasmidsCell Line Tropism Efficiency Serum RCR Gag-Pol Env
ProPak-X Xt;~ iC 2PG13; PA317 Resistant Negative pMLV*gp pCI-Ex
ProPak-A.52 amphotropic >PA317 Resistant Negativc pMLV*gp pCMV*Ea
ProPak-A.6~ amphotropic ?PA3 17 Resistant Negative pCMV-gp pCMV*Ea
Rigg et al., (1996) supra.
CA 02238434 1998-06-08
W O 97/21824 PCTAJS96/19904
Described above are safe, new retroviral vector packaging cell lines which can target
distinct receptors on human cells and provide efficient gene transfer to human cells for gene
therapy applications. The new lines described in Example 6 differ from the ProPak-A
S clone 6 of Example 1 since the Gag-Pol ORFs were expressed from the MLV-LTR promoter
(Table 13) instead of the CMV promoter in ProPak A.6. Despite carrying additional
MLV-specific sequences, no RCR was detectable in up~rn~t~ntc or cells using a stringent
test, namely extended co-culture of ProPak-X and ProPak-A.52 cells carrying the
BC140revM10 vector. These ProPak vector particles are resistant to human serum, and
should therefore remain functional if ~lmini~tcred in vivo to hllm:~n~
The ProPak-X and ProPak-A cell lines produce particles which utilize distinct
receptors on hurnan cells, and the tropisms were confirrned by d~ liniLlg the ability to
tr~n~d~lre cell lines from various species. Both receptors can be targeted by producing
sup~rn~t~nt~ from a co-culture of ProPak-A and ProPak-X producer cell lines. The safety of
the ProPak cell lines al}ows for eff1cient ping-pong amplification of the vector without the
danger of RCR formation. Using ProPak-A/X vector, 100% mslrkin~ of CFU-C derivedfrom CD34-positive cells purified from MPB was achieved. Significantly, the present
experiments demonstrated expression of the Lyt2 transgene in 4()% of the inoculated
CD34-positive cells derived from MPB, 2 days post-inoculation.
The present exp~rim~nt~ also indicated that high-eff1ciency gene transfer is best
achieved with supern~tslnt~ prepared from stable producer cell cultures. The approach
described above is applicable to production from stable cells, existing and future vector
systems.
As is a~ l to those of skill in the art, various modifications and alterations to the
above can be made without departing from the spirit and scope of the invention disclosed
herein. Accordingly, the invention is limited only by the following claims.
CA 02238434 1998-06-08
WO 97/21824 PCT~US96/19904
76
S~U~.~N~ LISTING
(1) GENERAL INFORMATION:
(i) APPLICANT: RIGG, RICHARD J.
DANDO, JONAT~AN S.
CHEN, JINGYI
FORESTELL, SEAN P
PLAVEC, IVAN
BOHNLEIN, ERNST
(ii) TITLE OF INVENTION: METHOD FOR OBTAINING RETROVIRAL
PACKAGING CELL LINES PR~U~1NG HIGH TRAN~U~1N~ EFFICIENCY
RETROVIRAL ~U~NATANT
(iii) NUMBER OF SE~U~N~S: 12
(iV) CORRESPON~ ADDRESS:
(A) ADDRESSEB: MORRISON & FOERSTER
2~ (B) STREET 755 PAGE MILL ROAD
( C ) CITY: PALO ALTO
(D) S TATE CA
(E) COUNTRY: USA
(F) ZIP. 93304-1018
(V) COMPUTER READABLE FORM:
(A) MEDIUM TYPE: F1OPPY disk
(B) COMPUTER: IBM PC compatible
(C) OPERATING SYSTEM PC_DOS/MS-DOS
(D) SOFTWARE PatentIn Re1eaSe #1.0, VerSiOn ~1.30
(Vi) CURRENT APPLICATION DATA:
(A) APPLICATION NUMBER: PCT/US96/
(B) FILING DATE:
(C) CLASSIFICATION:
(Viii) ATTORNEY/AGENT INFORMATION:
(A) NAME: KONSKI, ANTOINETTE F.
(B) REGISTRATION NUMBER 34,202
(C) REFERENCE/DOCKET NU.MBER- 20296-20036.40
(iX) TELECOMMUNICATION INFORMATION:
(A) TELEPHONE: (415) 813-5600
(B) TELEFAX: (415) 494-0792
(C) TELEX: 706141
(2) INFORMATION FOR SEQ ID N0: 1:
(i) SEQUENCE CHARACTERISTICS
(A) LENGTH: 45 baSe PairS
(B) TYPE nUC1eiC aCid
(C) STRANDEDNESS Sing1e
(D) TOPOLOGY: 1inear
(xi) SEQUENCE DESCRIPTION: SEQ ID NO:1:
AAA~U~GC GGCCGCGCCG CCACCATGGG CCAGACTGTT ACCAC 45
CA 02238434 1998-06-08
w o 97nl824 PCTrUS96/19904
77
(2) INFORMATION FOR SEQ ID NO:2:
(i) SEQUENCE CHARACTERISTICS:
(A) LENGTH: 36 base pair~
(B) TYPE: nucleic acid
(C) sTR~Nn~n~s single
(D) TOPOLOGY: linear
(xi) SEQUENCE DESCRIPTION: SEQ ID NO:2:
AR~AGC GGCCGCTCAT TAGGGGGCCT CGCGGG 36
(2) INFORMATION FOR SEQ ID NO:3:
(i) SEQUENCE CHARACTERISTICS:
(A) LENGTH: 38 base pairs
(B) TYPE: nucleic acid
(C) STRANDEDNESS: single
(D) TOPOLOGY: linear
(xi) SEQUENCE DESCRIPTION: SEQ ID NO:3:
TAATCTACGC GGCCGCCACC ATGGCGCGTT CAACGCTC 38
(2) INFORMATION FOR SEQ ID NO:4:
(i) SEQUENCE C~ARACTERISTICS:
(A) LENGTH: 36 ba~e pairs
(B) TYPE: nucleic acid
(C) STRANDEDNESS: single
(D) TOPOLOGY: linear
(xi) SEQUENCE DESCRIPTION: SEQ ID NO:4:
A~TGTGATGC GGCCGCTCAT GGCTCGTACT CTATGG 36
(2) INFORMATION FOR SEQ ID NO:5:
(i~ SEQUENCE CHARACTBRISTICS:
(A) LENGTH: 21 base pairs
(B) TYPE: nucleic acid
(C) STRANDEDNESS: R ingle
(D) TOPOLOGY: linear
(xi) SEQUENCE DESCRIPTION: SEQ ID NO:5:
TCGATTAGTG AACGGATCCT T 2l
50 (2) INFORMATION FOR SEQ ID NO:6:
(i) SEQUENCE CHARACTERISTICS:
(A) LENGTH: 21 ba~e pairs
(B) TYPE: nucleic acid
(c) STR~NDEDNESS: ~ingle
(D) TOPOLOGY: linear
CA 02238434 1998-06-08
WO 97/21824 PCT~US96/19904
78
(xi) SEQUENCE DESCRIPTION: SEQ ID NO:6:
CTCCTGACTC CAATATTGCA G 2l
(2) INFORMATION FOR SEQ ID NO:7:
(i) SEQUENCE CXARACTERISTICS:
(A) LENGTH: 20 base pairs
(B) TYPE: nucleic acid
~C) STR~Nn~nN~.~S: single
(D) TOPOLOGY: linear
(Xi) S~QU~N~ DESCRIPTION: SEQ ID NO:7:
TCGACGTTGT CACTGA~GCG . . . . . 20
(2) INFORMATION FOR SEQ ID NO:8:
(i) SEQUENCE CHARACTERISTICS:
(A) LENGTH: 20 base pairs
tB) TYPE: nucleic acid
(C) STR~NDEDNESS: sinyle
(D) TOPOLOGY: linear
(xi) SEQUENCE DESCRIPTION: SEQ ID NO:8:
GCTCTTCGTC CAGATCATCC 20
(2) INFORMATION FOR SEQ ID NO:9:
(i) SEQUENCE CHARACTERISTICS:
(A) LENGTH: 20 base pairs
(B) TYPE: nucleic acid
(C) sTR~n~nN~s single
(D) TOPOLOGY: linear
(xi) SEQUENCE DESCRIPTION: SEQ ID NO:9:
ACACAACTGT GTTCACTAGC 20
(2) INFORMATION FOR SEQ ID NO:lO:
(i) SEQUENCE C~ARACTERISTICS:
(A) LENGTX: 20 base pairs
(B) TYPE: nucleic acid
(C) STR~NDEDNESS: single
(D) TOPOLOGY: linear
(xi) SEQUENCE DESCRIPTION: SEQ ID NO:lO:
CAACTTCATC CACGTTCACC 20
(2) INFORMATION FOR SEQ ID NO:ll:
(i) ~U~N~ CHARACTERISTICS:
(A) LENGTH: 37 base pairs
(B) TYPE: nucleic acid
CA 02238434 1998-06-08
WO 97/21824 PCT~US96/19904
79
(C) STRANDEDNESS: single
(D) TOPOLOGY: linear
(xi) ~ ~C~ DESCRIPTION: SEQ ID NO:ll:
s
ACCTCGAGCC GCCAGCCATG GAAGGTTCAG C~ll~lC 37
(2) INFORMATION FOR SEQ ID NO:12:
(i) SEQUENCE CHARACTERISTICS:
(A) LENGTH: 30 base pairs
(B) TYPE: nucleic acid
(C) STRANDEDNESS: single
(D) TOPOLOGY: linear
(xi) SEQUENCE DESCRIPTION: SEQ ID NO:12:
AATCTAGACT TATTCACGCG ATTCTACTTC 30
