CHAPTER 3
Application of phiC31 for Gene Therapy in Mouse
Hematopoietic Stem Cells
Portions of this study were published as
Jason J. Hoyt, Annahita Keravala, Jennifer Duda, Christopher H. Contag and Michele P. Calos.
Non-viral, site-specific gene transfer into murine hematopoietic stem cells
2005. Molecular Therapy 11:S76-7
I wrote up all results and shared equal participation in this study with Annahita Keravala.

51
ABSTRACT
Autologous gene therapy for human hematopoietic stem cells (HSC) would
alleviate many problems associated with the current restricted supply of matching bone
marrow donors. The current gene therapy paradigm is to use retroviruses capable of
transducing HSC, however, this is fraught with numerous challenges. And while non-
viral integration into human HSC has been demonstrated, it has yet to be shown in
mouse HSC. Use of mouse HSC is important when transplantation of human HSC into
NOD/SCID mice is not practical or desired. We undertook a study to determine the
effectiveness of phiC31 integrase, a site-specific serine recombinase, in integrating a
luciferase donor gene into mouse HSC. As previous studies had used viral vectors
capable of transducing quiescent HSC, we first had to determine the best delivery
method for the non-viral phiC31 integrase. We demonstrate high levels of expression
from a GFP marker gene after enriching mouse whole bone marrow for the Lineage
depleted (Lin
-
) fractions followed by electroporation with the Amaxa nucleofector. We
also show that phiC31 integrase is capable of providing stable and long-term in vitro
expression of a luciferase-bearing plasmid containing the phiC31 attB recognition
sequence. Furthermore, we repopulated lethally irradiated BALB/C mice with
nucleofected Lin
-
HSC co-transfected with a luciferase plasmid and a phiC31 encoding
plasmid. A 164-day time-course experiment demonstrated short-term expression of the
donor luciferase gene. Evidence for long-term expression was not confirmed. These
results suggest several conclusions: possibly both short and long-term progenitors had
integration events, but the electroporation toxicity was too great for cells that received
DNA. The Lin- fraction may need to be purified further for less ‘mature’ HSC in order
to ensure long-term engraftment and expression, or electroporation conditions may need
more optimization. Our results demonstrate the utility of phiC31 in mouse HSC and
provide additional evidence for alternatives to a virus-based gene therapy regime.

52
INTRODUCTION
Inherited disorders of the hemoglobin genes are the most common genetic
diseases in the world. They account for approximately 7% of all genetic disorders (1).
Patients faced with either a genetic disease or cancers of the blood have few curative
options. For those with even a mild genetic disease, such as mild beta-thalassemia, life-
long blood transfusions must be taken bi-weekly. Those unfortunate enough to have a
more severe form of beta-thalassemia will not live past their teens, barring a bone
marrow transplant from a matched donor. Unfortunately, matched donors are rare, and
the procedure is lengthy, expensive, and invasive.
An alternative is to use the patient's own bone marrow stem cells. In one option,
the long-term progenitors that are capable of repopulating the entire blood system can
be fractionated and expanded ex vivo. The expanded population can then be tested and,
in the case of cancer, any cells testing positive for cancer phenotypes or genotypes can
be removed and the remaining normal cells introduced back into the patient (2). This
procedure has a very low probability of success, as methods to expand normal long-
term progenitors have not seen high success rates to date. In addition, this method is
difficult in the case of genetic diseases, since most progenitors will have the same
genetic abnormality. This creates the additional need to carefully screen cells prior to
autologous transplantation. The other option is to deliver the correct gene into the
progenitor population and then introduce the cells back into the patient where
progenitors will engraft into the host bone marrow. Along with a drug selection gene or
a curative gene that confers a natural selective advantage, cells will engraft and expand
in vivo.
To date, the most effective means of delivering transgenes into the non-dividing
fraction of bone marrow cells has been the use of a lentiviral vector (reviewed in 3).
The innate ability of the lentivirus to transduce non-dividing cells makes it an attractive
choice when targeting quiescent cells such as HSC. In this procedure, a lentivirus is

53
gutted to render it replication-defective, and the donor gene of choice is packaged
inside. This method of gene delivery has had mixed success. While 10 boys in France
were treated with retroviruses to cure SCID-X1, four went on to develop T-cell
leukemia, resulting in one death (4). The details are still being sorted out, but it appears
that the strong viral promoter used with the donor gene integrated upstream of the
LMO2 oncogene in three of the boys, leading to its activation and over-expression.
Lentivirus normally has a somewhat random integration pattern, albeit a propensity
toward gene active regions. Follow-up studies involving new X-SCID mouse models
suggest that this might be attributable to a unique combination of the donor gene and
the LTR viral promoter (5).
Adding to the difficulty in using retroviruses, short-term expression of donor
genes due to gene silencing and position effects (6) is a common occurrence. To
overcome position effect silencing, insulator sequences flanking the donor gene have
been used with limited success (7). Additional attempts to clone and expand a high
expressing transduced cell suffer the same difficulties as previously mentioned when
expanding cells ex vivo. In addition, it is thought that several types of long-term
progenitors co-exist and that all, or most, are needed for a properly functioning and
diverse blood system (8). Cloning and expanding a single long-term progenitor then
reduces the epigenetic diversity, even within an autologous transplant.
An alternative to the viral system of delivery is to use one of several non-viral
methods that have been developed (reviewed in 8 and 9). Here, we report the results of
a study using the phage integrase, phiC31. Phage integrases are capable of recombining
two attachment sites, termed attB and attP, within the bacterial and phage genomes,
respectively. It was later discovered that recognition sites with similar homology to
phiC31 attP reside throughout the human genome; these sites are termed “pseudo attP.”
It has been shown that a plasmid carrying a donor gene of choice and the attB
recognition sequence can be co-delivered with a plasmid encoding phiC31 integrase for

54
unidirectional integration into the human genome.
With any non-viral system, one of the most difficult hurdles to overcome is gene
delivery. This is especially true when delivering the therapeutic gene to primary
quiescent cells, such as HSC. Methods of delivery include magnetofection, lipid based
complexes, gene guns, and electroporation. Delivery into human CD34+ cells, an early
hematopoietic progenitor, has seen some success using electroporation techniques
(10,11,12,13). To date, no one has demonstrated effective non-viral gene delivery to the
mouse equivalent of human CD34+ cells. There is some debate as to what constitutes
the equivalent, but in general, mouse whole bone marrow is obtained and fractionated
for at least the Lin
-
population. This population contains a heterogeneous population of
multipotent progenitors, containing both long-term and short-term HSC. These exhibit
various levels of repopulation potential and self-renewal capacity.
In this study, we evaluated several different electroporation devices, controlling
for dozens of variables in an effort to deliver naked DNA plasmids to lineage-depleted
hematopoietic cells (Lin
-
HSC). We found that the Amaxa nucleofector (Cologne,
Germany) was the most effective. Long-term expression of integrated donor genes is
essential for a therapeutic effect. Therefore, we also evaluated integration stability in
culture. We found that there was indeed an integrase effect, in which a donor luciferase
plasmid required phiC31 for long-term expression. We also studied the repopulation
potential of Amaxa-nucleofected Lin
-
HSC in vivo with BALB/c mice. Here,
repopulation and engraftment occurred at modest levels. However, we were unable to
confirm long-term gene expression in vivo, which was more likely due to unoptimized
experimental conditions for electroporation, than with the ability of phiC31 to integrate
a donor gene.

55
MATERIALS AND METHODS
Plasmids for eGFP and Luciferase assays. The plasmids used in this study
were constructed as follows. Plasmid pCMVInt (Fig. 3-1 A) as described previously
(14). Plasmid pBLB (Fig. 3-1 D) contains attB and a CAGGS promoter driving a
luciferase gene. It was constructed by digesting plasmid pBeta-actin-phiC31-IRES-
eGFP with AflIII to obtain the beta-actin insert. This was further digested with SmaI
and blunted with Klenow DNA polymerase I. Plasmid pNBL2 (15) carries the firefly
luciferase gene driven by the CMV promoter, an SV40/bacterial promoter-driven
neomycin/kanamycin resistance gene, and the phiC31 attB site. This was digested with
XhoI to remove the CMV promoter and blunted with Klenow in order to isolate the attB
fragment and luciferase gene. This was then ligated to the AflIII/SmaI blunted beta-
actin insert to obtain pBLB.
Plasmid pDB2 was derived from pEGFP-C1 (Inivtrogen, Carlsbad, CA) by
placing ~300 b.p. of attB into the MluI site. Plasmid pBEB (Fig. 3-1 B) contains attB
and a CAGGS promoter driving eGFP expression. It was constructed by digesting
pDB2 with AflIII/AgeI (to excise CMV) and digesting the beta-actin-phiC31-IRES-
EGFP plasmid with AflIII/SmaI to release the CAGGS promoter. The CAGGS
promoter was then ligated into the digest pDB2. Plasmid pBGLB (Fig. 3-1 C) used the
same CAGGS promoter cloned into the NheI/BglII sites of plasmid pGL3 (Promega,
San Luis Obispo, CA, USA). Plasmid pTA-attB (14) was then digested with
BamHI/EcoRV to obtain a 307-bp attB fragment, which was then ligated into the
BamHI/PshAI sites of pGL3 to produce pBGLB containing attB and firefly luciferase
driven by the CAGGS promoter and an SV40 polyA.
Plasmid pCLB (Fig. 3-1 E) is similar to pBLB, except that it contains the rabbit
beta-globin polyA rather than SV40 polyA downstream of firefly luciferase. It was
constructed by digesting plasmid pCAGGS (16) with EcoRI, blunting and inserting
firefly luciferase obtained from pGL3 with SmaI/NaeI digestion. The attB fragment

56
came from pTA-attB. The plasmid was BamHI/EcoRV digested, blunted and placed
into the PstI site of the pCAGGS-Luc to create pCLB. Not diagramed are plasmids
pMND-eGFP-1 (gift from Don Kohn, Children’s Hospital, Los Angeles, CA) and
pMND-eGFP-2, which is a second prep of pMND-eGFP-1. Both plasmids have an
MND promoter driving eGFP expression, an SV40 polyA, and kanamycin resistance.
Plasmid pMax (Amaxa, Cologne, Germany) is an expression vector encoding green
fluorescent protein, a GFP from the copepod Pontellina plumata.
Mouse Whole Bone Marrow Isolation and Fractionation. All animals were
housed under pathogen-free conditions, and experimental protocols were submitted to
and approved by the Administrative Panel on Laboratory Animal Care (APLAC) at
Stanford University. Five days prior to harvest, 6-8 week old BALB/c mice (Jackson
Labs, Bar Harbor, Maine) were injected with 200 µL of 25 mg/mL 5-fluoruracil to
stimulate blood progenitors. Whole bone marrow (WBM) was then harvested from the
femurs and tibias. Initially, cells were collected as a single cell suspension after flushing
the femur and tibia with PBS using a half inch 29 gauge needle into a 10 cm collection
dish. Approximately 20-30 million WBM cells per individual mouse were obtained with
this method.
Later, isolation was via a bone crushing method, in which the tibias and femurs
of all mice were collected, crushed and homogenized with a mortar and pestle. PBS +
2% fetal bovine serum was used to continually bathe the crushed material to prevent
over drying. Cells were then collected and filtered using a 70 µm cell strainer (Becton
Dickinson 352350). This preferred WBM isolation technique produced between 100-
150 million cells per individual mouse. These numbers reflect post red blood cell (RBC)
lysis. Briefly, cells were spun down for 5 minutes at 1500 rpm and 4°C, after which the
supernatant was discarded. Ten mL of 1X RBC lysis buffer (eBioscience, San Diego,
CA) was added and left at room temperature for 2-3 minutes to lyse the RBCs. Cells
were then washed with PBS to remove excess ammonium chloride. This constituted the

57
WBM fraction. Cells were then either electroporated directly, fractionated further
followed by electroporation, or allowed to incubate in a culture overnight (See cell
culture methods).
Fractionation and enrichment for the Lin
-
or Sca
+
populations was performed via
magnetic cell sorting with MACS (Miltenyi Biotec, Sunnyvale, CA). Briefly, WBM
was resuspended in 10 mL PBS + 2% FBS (FACS buffer). PE-conjugated antibody
(Ebioscience, San Diego, CA) cocktail was added for a final 1:50 dilution. The antibody
cocktail consisted of 200 µL/antibody of GR-1 + Mac1 to deplete the myeloid lineage,
B220 for B lymphocyte depletion, CD3 for T lymphocyte depletion, and TER119 for
erythroid depletion. Cells were incubated on ice in the dark for 15 minutes. 200 µL of
Fc block, an antibody to the Fc-delta receptor, was then added, and incubation on ice in
the dark continued for another 10 minutes. The solution was then washed with 20 mL
PBS + 2% FBS and spun down at 1200 rpm for 5 minutes. Cells were resuspended in 8
mL FACS buffer. 2 mL of anti-PE magnetic beads (MACS) were added and incubated
in the dark on ice for 15 minutes and washed with 20 mL FACS buffer. This was spun
down at 1200 rpm for 5 minutes and resuspended in 5 mL FACS buffer for addition to
the MACS magnetic columns.
LS columns (Miltenyi Biotec, Sunnyvale, CA) were equilibrated with 3 mL
FACS buffer. Cells were strained in a 20 µM nylon mesh strainer to remove clumps. 0.5
mL of the cell solution was added to any one column and washed 3x with 3 mL FACS
buffer to collect the enriched cell population.
Electroporation. Prior to electroporation, cells were either incubated overnight
after isolation, pelleted and washed 2x with FACS buffer, or electroporated
immediately after isolation. BALB/c whole bone marrow cells, fractionated Lin-, or
fractionated Lin
-
Sca
+
cells were electroporated in 100 µL volumes at various cell
concentrations using the Amaxa Nucleoporator (AMAXA, Cologne, Germany)
according to the manufacturer’s instructions with program U-08 and cell kits CD34,

58
K562, BALB/c T-cell, mouse neural stem cell, or KG1. Plasmid pEGFP-C1
(Stratagene, La Jolla, CA) was used at concentrations of 1 µg, 5 µg or 20 µg on WBM.
For fractionated cells, plasmids pMax, pBEB, pBLB, pBGL3, and pCLB were all tested
at various concentrations. To increase electroporation efficiency, immediately after
electroporation, cells were spun down in Stem Span SFEM media (Stem Cell
Technologies, Vancouver, B.C., Canada) at 1200 rpm for 5 minutes. The supernatant
was discarded and the pellet allowed to incubate for 10 minutes at 37°C. Cells were
then resuspended and cultured (see method on cell culture of enriched murine cells).
The BTX ECM 830 square wave porator (Genetronics, San Diego, CA) was
used on WBM with one of four manufacturer recommended protocols: dog bone
marrow, human K562, mouse ESC, myeloid progenitor, or human bone marrow. For
the dog protocol, the setup was the following: mode = HV, voltage = 1.8 kV, pulse
length = 15 µsec, number of pulses = 1, cuvette = 4 mm gap, volume = 800 µL PBS,
and temperature = 4°C. Human K562 protocol was: mode = LV, voltage = 300 V, pulse
length = 15 msec, number of pulses = 1, cuvette = 4 mm gap, volume = 800 µL PBS,
and temperature = 4°C. The mouse ESC protocol as follows: mode = LV, voltage =
500 V, pulse length = 99 µsec, number of pulses = 1, cuvette = 2 mm gap, volume =
400 µL PBS, and temperature = 25°C. The myeloid progenitor protocol as follows:
mode = LV, voltage = 300 V, capacitance = 1000 µF, pulse length = 15 msec, number
of pulses = 1, cuvette = 4 mm gap, volume = 800 µL PBS, and temperature = 25°C.
Human bone marrow protocol as follows: mode = LV, voltage = 300 V, capacitance =
250 µF, pulse length = 15 msec, number of pulses = 1, cuvette = 4 mm gap, volume =
400 µL PBS, and temperature = 25°C. In all four protocols, the cells were immediately
transferred to 2 mL RPMI media and incubated for 48 hours at 37°C. Various cell
concentrations and plasmid concentrations of plasmid pEGFP-C1 were used.
The Bio-Rad Gene Pulser (Bio-Rad, Hercules, CA, USA) was also used with

59
WBM at 4 x 10
6
cells per reaction. Voltage was kept constant at 150 V, cells
resuspended in 400 µL PBS, placed in a 2 mm cuvette. Resistance at 100 Ohms and
capacitance was varied using either 3 µF, 25 µF, 125 µF, 500 µF, or 960 µF.
Additionally, either 5 µg or 20 µg of plasmid pEGFP-C1 was used. Controls consisted
of WBM cells without DNA and electroporated at 25 µF, 500 µF, or 960 µF. After
electroporation cells were immediately transferred to 2 mL RPMI media and incubated
at 37°C for 48 hours before analysis.
For the Bio-Rad xCell Gene Pulser (Bio-Rad, Hercules, CA) the following
conditions were used on 5 x 10
5
WBM cells: 320 V, 1550 µF, 4 mm cuvette at a volume
of 500 µL serum-free RPMI media. After electroporation cells were immediately
transferred to 2 mL Stem Span SFEM media and incubated at 37°C for 48 hours before
analysis. Plasmids pBEB, pMND-eGFP-1 and pMND-eGFP-2 were used at
concentrations of 5 µg, 10 µg, or 20 µg.
Cell Culture of enriched murine cells. Fractionated and enriched Lin
-
or Lin
-
Sca
+
cells were incubated overnight in 10 mL Stem Span SFEM media with murine IL-
3 (10 ng/mL), murine IL-6 (50 ng/mL), and murine SCF (100 ng/mL) all from
PeproTech (Rocky Hill, NJ, USA). Electroporated cells were transferred to 10 cm
dishes containing Stem pro-34 SFM media and the interleukin supplements previously
listed. Cells were incubated at 37°C for at least 24 hours prior to gene expression
analysis via flow cytometry for GFP expression or bioluminescence imaging for
luciferase expression.
BALB/c irradiation and repopulation. For each in vivo experiment, and just
prior to injection, six week-old BALB/c mice were lethally irradiated with a 200-kV
irradiator (RT250; Phillips, Shelton, CT) with a dose of 800 cGy given in two doses
five minutes apart to destroy all blood progenitors. Either retroorbital plexus vein or
lateral tail-vein injection method with a 1 cc syringe and 27
1/2
gauge needle was used to
inject modified cells resuspended in 100 µL PBS. For tail-vein injections, a heat lamp

60
was used for five minutes prior to injection to increase blood flow and visualize the vein
more readily. Mice were placed in autoclaved cages with acidified water containing
200mg/L sulfamethoxazole and 8 mg/mL trimethoprim to prevent opportunistic
infections. Mice were housed in the Research Animal Facility at Stanford University.
Bioluminescence imaging and flow cytometry. To measure luciferase
expression in culture or in vivo, the IVIS 200 machine (Xenogen, Alameda, CA) was
used. The IVIS 200 consists of a mounted and cooled integrated CCD camera in a dark
specimen chamber to capture luciferase-emitted photons measured in photons per
second. Photon emissions are measured as selectable regions of interest (ROI) and
quantified using the LivingImage 2.5 software (Xenogen) and Igor imaging software
(WaveMetrics, Lake Oswego, OR). Luciferase expression is shown as pseudocolor
images of photon intensity with red as the most intense and blue as the least intense
(17,18). Cultured cells received 5 µL/mL D-luciferin (Xenogen) at 30 mg/mL
immediately prior to imaging. Cells were then exposed for 30 seconds with medium
binning and an f-stop of 1. For in vivo imaging of live mice, 200 µL at 30 mg/mL of D-
luciferin was injected intraperitoneally15 minutes prior to imaging. Mice were then
anesthetized using isofluorene and imaged dorsally and ventrally with five-minute
exposures, small binning, and an f-stop of 1.
Flow cytometry for propidium iodide (PI) and green fluorescent protein (GFP)
was carried out using the Scanford flow cytometer (Becton–Dickinson, Mountain View,
CA). Typically, 10 µL PI (0.1 mg/mL) was added to cell samples for cell viability
analysis. Data was analyzed with FloJo software (Treestar, San Carlos, CA). Cultured
cells were spun down at 1200 rpm and washed 2x with cold PBS + 2% FBS prior to
analysis. Single stained (PI or GFP) samples were used for compensation controls.
MpsL1 integration site analysis. For integration analysis of in vitro cultures,
genomic DNA was prepared through a DNeasy tissue kit (Qiagen, Germany). For in
vivo studies, mice were bled via the retroorbital vein and genomic DNA prepared with

61
the blood and cell culture DNA mini kit (Qiagen, Germany). Nested PCR was
performed using primers specific to the attB core and the mpsL1 genomic DNA as
previously described (19).

62
RESULTS
Eletroporation of murine whole bone marrow. To be considered as a viable
alternative to viral vectors, non-viral approaches must first overcome the barriers
associated with transgenic delivery into the normally quiescent hematopoietic stem
cells. Previous studies had successfully used electroporation to deliver naked plasmid
DNA into human CD34+ cells (11,12,13), mouse embryonic stem cells (21), rat neural
progenitors (22), and human mesenchymal stem cells (23). We tested four
electroporation devices for their ability to efficiently transfect naked plasmid DNA into
mouse whole bone marrow obtained from six week-old BALB/c mice. The devices used
were the BTX ECM 830 square wave porator, Bio-Rad xCell Pulser, Bio-Rad Gene
Pulser, and Amaxa nucleofector. Initial tests were carried out with pEGFP-C1, which
has an eGFP gene under the control of the CMV promoter. GFP expression was
quantified with flow cytometry 48 hours post-electroporation.
The first device tested was the BTX ECM 830 utilizing a square wave pulse,
which has been shown to have higher transfection efficiency than exponentially
decaying pulses in some cell types (24). Whole bone marrow was harvested from seven
BALB/c mice. Experiments were performed in duplicate with either 5 µg or 20 µg of
pEGFP-C1. Cell densities per cuvette were varied using 1x10
6
, 5x10
6
, 15x10
6
, 20x10
6
,
or 30x10
6
cells. Finally, we tested five different manufacturer supplied protocols:
Human hematopoietic K562, dog bone marrow, mouse embryonic stem cell, human
myeloid progenitor, or human bone marrow. Negative controls were used in each
protocol (excluding the human bone marrow due to lack of cells), at a cell concentration
of 5x10
6
cells without any DNA plasmids. Flow cytometry analysis indicated that the
human K562 protocol with 20x10
6
WBM cells and 20 µg of pEGFP-C1 was the most
efficient protocol, with 1.47% of the cells expressing GFP (Table 3-1).
We then tested the Bio-Rad Gene Pulser that uses an exponential decay for
electroporation. In this trial, total bone marrow from three BALB/c mice was harvested

63
and depleted of red blood cells to produce WBM. All variables were performed in
duplicate at cell densities of 4x10
6
cells per cuvette. A total of 10 variables were tested
with pEGFP-C1, and GFP expression was analyzed with flow cytometry. Controls
consisted of similar capacitance without the addition of DNA. The best result came
from using a capacitance of 960 µF and 20 µg of DNA (Table 3-2). The total GFP
positive was 3.47%, but when normalized against the 960 µF control reaction, that
number dropped to 2.55% positive.
Additionally, we tested the Bio-Rad xCell Pulser, which is capable of both
exponential decay and square wave pulses. For this study, we used the square wave
pulse. WBM was harvested from one BALB/c mouse and used at a density of 5x10
5
cells per cuvette. The plasmids used were two different preparations of pMND-eGFP,
which has an MND promoter driving eGFP expression and pBEB, which contains eGFP
driven by the CAGGS promoter (Fig 3-1). Only a single device variable was used, and
the amount of DNA was varied. All experiments were performed in duplicate. Controls
consisted of naïve WBM cells that received no DNA and no pulse, WBM + pulse and
no DNA, 5 µg MND-eGFP-1 and no pulse, and finally 5 µg pBEB and no pulse. GFP
expression was measured 48 hours later. None of the test variables produced values
exceeding the background controls (Table 3-3).
Finally, we used the Amaxa nucleofector, which is a device with proprietary,
unknown voltage and solution conditions. Five kits were tested, K562, human CD34,
mouse BALB/c T-cell, mouse neural stem cell, and KG-1. Two mice were harvested for
WBM and each experiment consisted of either 1x10
6
or 2x10
6
cells with 1 µg, 5 µg, or
20 µg pEGFP-C1. Due to a lack of cells, each experiment was performed only once.
Flow cytometry analysis demonstrated that the CD34 kit with 20 µg DNA and 2x10
6
WBM cells was the best condition, with results as high as 4.77% GFP positive.
Corrected for background with the CD34 control, this result dropped to 4.58% GFP
positive (Table 3-4).

64
Lin
-
HSC enrichment results in higher DNA uptake. With a high of only
4.58% electroporation efficiency in WBM (Table 3-4), successful in vivo
experimentation was unlikely. As WBM is a heterogeneous mixture consisting of many
different blood lineages, it was possible that unwanted cells were acting as ‘fillers’ and
reducing DNA uptake. To test this theory, we decided to fractionate the WBM for a
more purified population enriched for the Lin
-
population. The Lin
-
HSC population is
depleted of erythroid, T lymphocyte, B lymphocyte, and myeloid lineages. Lin
-
is also
part of an important set of markers (c-Kit
+
, Lin
-
, Sca
+
, Thy
lo
) for both short-term (ST)
and long-term (LT) HSC progenitors, in which just a few LT-HSC are required to
reconstitute the long-term hematopoietic multilineage in mice (25).
For this study, 20 BALB/c mice were harvested for WBM and a total of 1.3 x
10
8
cells were obtained. We then performed magnetic cell sorting using a MACS kit and
five antibodies (GR-1, Mac1, B220, CD3, and TER119) that recognize cell surface
markers of cells we desired to deplete. This left an enriched Lin
-
population of
approximately 1.3 x 10
7
cells. Cells were incubated overnight in Stem Span media and
interleukins IL-3, IL-6 and SCF prior to Amaxa nucleofection. We tested three eGFP
plasmids, the Amaxa supplied pMax (CMV promoter driving eGFP), pBEB, and
pMND-eGFP-2. All sample plasmids were tested in duplicate. A single ‘No DNA’
control was tested as well. Program U-08 and the CD34 kit were used. “Nucleofected”
cells were then incubated in stem span media with interleukin supplements and
analyzed for GFP expression and cell viability 24 hours later (Table 3-5). Cells were
gated for the side-scatter view. The pMax plasmid gave the two highest results with
37.4% and 34.6% GFP positive. When only viable cells (propidium iodide exclusion)
were counted, GFP positive cells rose to 59.2% and 34.7% for the two pMax replicates,
suggesting continued eGFP expression in live cells. While pBEB and pMND-eGFP-2
did not perform as well, at 10.6% and 7.05% GFP positive respectively, those numbers
rose dramatically when only viable cells were counted, to 40.1% and 30% respectively.

65
Noticeably, electroporation caused severe cell death. 43.5% death was observed with
the ‘No DNA’ control, and the addition of DNA increased cell death further. The most
viable sample, pMND-eGFP-2, only had a viability of 21.4%. These results indicate that
lineage depletion is a necessary step to achieving efficient transient plasmid
transfection.
PhiC31 integrase mediates stable expression in Lin
-
HSC. For therapeutic non-
viral gene therapy to be a viable alternative to viral methods, stable long-term
integration and expression need to be demonstrated. Therefore, we initially tested the
utility of a phiC31 integrase plasmid to be co-electroporated with several luciferase
constructs bearing an attB site. We chose to deliver luciferase, rather than GFP, to
eventually utilize in vivo bioluminescence imaging that could be performed on live
mice. Three constructs were built for this assay (Fig. 3-1), plasmids pBGLB, pCLB and
pBLB. All three constructs bear the attB recognition site for future integration studies.
Plasmid pMax, which does not bear luciferase, was used as a transfection control. Lin
-
HSC was enriched and harvested from 19 BALB/c mice to give a final cell count of
8x10
6
cells. The day after harvest, four groups, with 2x10
6
cells each, were mixed and
nucleofected with 5µg pMax, 5µg pBLB, 5µg pCLB, or 5µg pGL3. Samples with
pBLB, pCLB, and pBGLB also received 10µg of pCMVInt, which encodes for the
phiC31 integrase. Figure 3-2 shows bioluminescence readings at 24 and 48 hours post
nucleofection. pCLB was nearly 2-fold higher than pBGLB and nearly 6-fold higher
than pBLB, while the control pMax was more than a log below all luciferase constructs.
Cell viability, as measured by propidium iodide exclusion and flow cytometry, was
15.5% for both pMax and pBGLB. Plasmids pCLB and pBLB had cell viabilities of
25.9% and 26.9%, respectively.
We next sought to test if phiC31 integrase was necessary for continued
expression of luciferase in cultured Lin
-
HSC. WBM cells were harvested from 20
BALB/c mice, followed by Lin
-
HSC enrichment and stimulation overnight.

66
Experimental samples performed in triplicate consisted of pCLB (5µg) + pCMVInt
(10µg) and pCLB (5µg) + pmInt (10µg), a phiC31 integrase mutated at amino acid 12
(S12A) to render it catalytically inactive (26). Inactive phiC31 is still capable of binding
attB and pseudo attP recognition sequences, but is unable to complete the catalytic
recombination and therefore cannot integrate a donor plasmid. Additionally, 5µg of
pCLB alone was nucleofected in duplicate. 5µg of pMax and Lin
-
HSC that were
nucleofected but received no DNA were used as negative controls. All 10 experimentals
were nucleofected, using 2x10
6
Lin
-
HSC. Bioluminescence and cell viability were
tested at 24 and 80 hours post nucleofection with the IVIS 200 and flow cytometer. At
24 hours, cells receiving pCLB alone were expressing luciferase nearly a log higher
than cells receiving pCLB and phiC31 integrase (Fig 3-3A,B). Cell viability showed
that pCLB had an average that was 12% higher than pCLB plus pCMVInt (Fig 3-4). In
contrast, after 80 hours pCLB + pCMVInt was a full log higher in luciferase expression
than both pCLB alone and pCLB + pmInt (Fig 3-3B). This difference was significantly
different with a P-value of 0.003. To test whether this difference was due to increased
cell death in the pCLB and pCLB + pmInt groups, viability was assessed and found to
be statistically similar across groups (Fig 3-4, p-value = 0.354). At 100 hours, the pCLB
+ pCMVInt group was still a log higher (average of 3.19x10
5
photons/second) than
pCLB + pmInt (average of 4.93x10
4
photons/second), which was slightly above
background controls (1.6x10
4
photons/second). Cells were unable to be cultured past
100 hours for continued analysis. We performed a PCR analysis for mpsL1 integration,
a known mouse “hotspot” for phiC31-mediated integration (19), but were unable to
detect any signal. This suggests that mpsL1 is not the preferred integration site in mouse
HSC. Despite a lack of confirmation for integration, our results demonstrated an
integrase effect in mediating stable expression in the Lin
-
HSC fraction.
PhiC31 nucleofected Lin
-
HSC repopulate lethally irradiated BALB/c. Prior to
luciferase construct optimization, in which plasmid pCLB was chosen for future

67
experiments, we performed a repopulation longevity experiment with Lin
-
HSC and
pBLB. We transplanted 2x10
6
Lin
-
HSC that had received the following treatments into
four cohorts of mice: 5µg pBLB + 10µg pCMVInt (n=6), 5µg pBLB + 10µg pCS (n=6,
pCS is pCMVInt with the CMV promoter and phiC31 gene removed), cells without
nucleofection and without DNA (n=2), and a sham injection of 100 µL PBS (n=1). All
mice were lethally irradiated BALB/c and retroorbitally injected. Immediately
following injection, all mice convulsed, leading to two deaths in the pCMVInt + pBLB
group and one in the pCS + pBLB group. The cause of this convulsion is not known,
however cardiac arrest in animals has been reported following injection (27). Four days
post-transplantation, the sham-injected control died, indicating successful lethal
irradiation of total bone marrow. Mice underwent live in vivo bioluminescence imagine
at nine days post-transplantation. The pCMVInt group showed no significant difference
compared to the control groups. Mice were kept alive up to 164 days and then
euthanized. Of the remaining pCMVInt group, 3/4 survived, while 3/5 remaining of the
pCS group survived. At no time point did the pCMVInt + pBLB show higher luciferase
expression than the controls. A check for integration at mpsL1 in the genomic DNA
showed no obvious integration of the donor, pBLB. Importantly, we demonstrated that
Lin
-
HSC, nucleofected with a donor plasmid and pCMVInt, were capable of
repopulating the hematopoietic system in mice.
Having demonstrated repopulation potential and with the optimized luciferase
plasmid pCLB, we undertook another in vivo study. To avoid mortality following
retroorbital injection, we injected nucleofected cells via the lateral tail vein.
Additionally, cells were carefully washed and resuspended to avoid clumping. The
experimental treatments and results for this study are shown in Table 3-6. To reduce
cell death during nucleofection, and in an attempt to improve mouse survival rates, we
lowered the amount of DNA nucelofected per cuvette from 15µg (5µg pCLB + 10µg
pCMVInt) to a maximum of 2µg (1µg each of pCLB and pCMVInt). While this

68
reduction in DNA increased the number of lethally irradiated mice surviving beyond
two weeks (data not shown), it probably had a negative affect on overall integration
efficiency of pCLB. Subsequent in vitro optimization studies suggested that the optimal
amount of DNA to achieve both high luciferase expression and high viability was 5µg
pCLB + 5µg pCMVInt (Fig 3-6). Comparing luciferase expression of the seven
surviving mice that received pCLB + pCMVInt to two controls that received cells, but
no DNA, showed a modest significant difference (p-value = 0.06 from student t-test) at
21 days post-transplantation (Fig 3-5, Table 3-6). By day 31, the pCMVInt group had
fallen to background levels. This suggests that marking of the short-term, but not the
long-term, HSC progenitors may have occurred. At day 38, blood was drawn from each
of the groups to check for presence of pCLB integration at mpsL1. PCR analysis of
isolated genomic DNA in the pCLB + pCMVInt group did not show conclusive
integration.

69
DISCUSSION
This study demonstrates that non-viral phiC31-mediated gene delivery into
lineage-depleted mouse HSC is possible. While there are several examples of non-viral
delivery into human hematopoietic progenitor cells (11, 12, 13), this is the first example
in mouse blood progenitors. This is an important advancement since the transplantation
of human HSC into NOD/SCID mice is not always practical, or desired. The Amaxa
nucleofector, which is effectively a modified cell electroporator, was found to have the
greatest gene delivery efficiency in this study. It was observed that there was a DNA
concentration-dependent toxicity associated with electroporation and that
electroporation alone caused a baseline value of cell death at about 43.5%. This toxicity
has been observed with human CD34+ cells as well (11, 28). We ameliorated this effect
by building and testing several different reporter constructs. We were able to improve
efficiency with just a minor adjustment in DNA and cell concentration. To date, the
most successful conditions appear to use no more than 10 µg total DNA and at least
2x10
6
cells. Further optimization of DNA and cell concentration could be carried out to
improve upon these results.
An early finding was that mouse whole bone marrow was highly intractable to
gene delivery via electroporation. Efficient gene delivery demanded enrichment for less
mature bone marrow cells known as lineage-depleted hematopoietic stem cells (Lin
-
HSC). While enrichment for the Lin
-
HSC population within WBM resulted in higher
DNA uptake, it is still unknown which WBM population was preventing further uptake.
It could have been a single population, for example B-lymphocytes, or a mixture of the
four populations that were depleted. A FACS experiment involving Amaxa
nucleofection of B-lymphocytes versus granulocytes, for example, could resolve this
question. We theorize that gene delivery could be higher with enrichment for the Lin
-
fraction into the Lin
-
Sca-1
+
population, or even enrichment for the most primitive
mouse long-term HSC population, c-Kit
+
Thy1.1
lo
Lin
-/lo
Sca-1
+
Flk-2
-
(29). The main

70
limitation to nucleofection, or any other electroporation technique, is that large numbers
of cells are required (at least 10
6
for Amaxa nucleofection). Presumably this is to
‘buffer’ the voltage that travels through the cuvette. Since enriching for just a few
thousand LT-HSC requires a massive starting amount of WBM, there may need to be a
tradeoff that takes place between efficiency of gene delivery and cell enrichment for
progenitors. One last option is to perform nucleofection on the Lin
-
HSC first and then
FACS sort for the more primitive cells.
PhiC31 integrase has previously been shown to be an effective tool in achieving
integration, leading to long-term expression of donor genes in many cell types and
tissues. In all of these studies, in order to observe phiC31-mediated long-term
expression, cells had to be cultured for at least 10 days. Therefore, it came as a surprise
that stable expression was seen within 80 hours post-nucleofection into Lin
-
HSC.
Normally, transient expression of marker genes is still evident for at least one week.
These results suggest that integration occurred in one or more subtypes within the Lin
-
population that had been co-nucleofected with pCLB and pCMVInt. MpsL1 is a site in
the mouse genome that has been observed to be a preferred phiC31-mediated
integration site in several mouse tissue types such as liver (19), muscle (30), lung (31),
embryonic tissue (32) and NIH3T3 cells (32). However, we were unable to detect
mpsL1 integration in either cultured Lin
-
HSC or mouse blood draws after engraftment
and repopulation. There are several possible explanations for this observation. If
integration occurred in cells other than the repopulating progenitor cells, from which
genomic DNA was later isolated in in vivo studies, then a lack of mpsL1 integration was
to be expected. In the case of Lin
-
HSC cultured up to 100 hours, it could have been that
mpsL1 is just not a preferred integration site. To determine new genomic integration
sites, a method for expanding and maintaining cell viability in culture would need to be
worked out first, followed by plasmid rescue (33). Alternatively, plasmid rescue could
be performed on mouse genomic DNA derived from blood of transgenic Lin
-
HSC

71
recipients. Another possible reason mpsL1 was not detected is that integration may have
occurred through a non-phiC31-mediated mechanism such as nonhomologous end
joining (NHEJ). PCR for a portion of the luciferase gene was performed to detect such
an event, but the evidence was inconclusive (data not shown). Interestingly, phiC31-
mediated integration sites in human cord blood and bone marrow CD34+ progenitor
cells were found to differ from those reported in other human primary cells (10). Such
data would support the altered integration site preference theory.
We observed a somewhat high host mortality rate after transplantation of
nucleofected cells, using either the retroorbital or lateral tail-vein injection method. This
toxicity was observed even when cells were carefully washed and clumps and bubbles
were avoided in the injection syringe. We surmise that this mortality was due to cellular
necrosis, which introduced high loads of toxins into the animals. In the future, by
sorting for live cells via FACS, it may be possible to avoid such high host death rates.
A live cell sort, combined with a sort for primitive LT-HSC mentioned above, might be
the best strategy in future studies.
When human CD34+ cells were nucleofected with Sleeping Beauty, another
non-viral gene delivery system, stable gene transfer was only 1-6% for over four weeks
in vivo (11). Those levels eventually fell to background from either lack of long-term
progenitor marking or cell death in the long-term population due to electroporation
toxicity. Since phiC31 integration efficiency is 5-10% in cells that are traditionally
easier to transfect than HSC, our initial results are comparable to those seen when using
Sleeping Beauty. Because phiC31 integrase has the advantage of being more site-
specific than Sleeping Beauty, it may be the preferable non-viral strategy for gene
therapy in HSC. Coupled with an in vivo selection strategy where certain cellular
properties confer an expansion advantage for transplanted cells, phiC31 may very well
be a suitable alternative to the current viral standards. SCIDS and Fanconi’s anemia are
two genetic diseases in which the therapeutic gene confers a selective advantage to the

72
host. For other genetic diseases, the co-delivery of a mutant MGMT gene on the donor
plasmid can provide a drug inducible in vivo selection alternative (34, 35).

73
Table 3-1. BTX ECM 830 electroporation efficiency on BALB/c WBM
Controls Percent GFP Positive
K562 0.63
Dog bone marrow 0.76
Mouse ESC 0.71
Myeloid progenitor 0.31
Samples Percent GFP Positive
K562 20m-20µg 1.47
K562 20m-5µg 1.34
dog 10m-20µg 1.31
myeloid 10m-20µg 1.13
mouse 10m-20µg 1.09
bone 30m-20µg 1.06
mouse 10m-5µg 1.04
mouse 15m-20µg 1.00
bone 10m-20µg 0.96
dog 5m-20µg 0.95
mouse 15m-5µg 0.93
myeloid 10m-5µg 0.92
K562 5m-20µg 0.91
K562 5m-5µg 0.90
K562 10m-20µg 0.87
mouse 5m-20µg 0.85
myeloid 5m-5µg 0.83
myeloid 5m-20µg 0.80
dog 1m-20µg 0.74
dog 1m-5µg 0.72
mouse 5m-5µg 0.67
myeloid 1m-20µg 0.62
myeloid 1m-5µg 0.58

Percent GFP positive BALB/c whole bone marrow cells after electroporation with the BTX
ECM 830 device and five different protocols (Human K562, dog bone marrow, mouse ESC,
myeloid progenitor, or human bone marrow). Either 5µg or 20µg plasmid pEGFP-C1 was used
on 1x10
6
(1m), 5x10
6
(5m), 15x10
6
(15m), 20x10
6
(20m) or 30x10
6
(30m) WBM cells. Controls
received 5x10
6
cells without DNA. All trials were performed in duplicate.

74
Table 3-2. Bio-Rad Gene Pulser electroporation efficiency on BALB/c WBM
Controls Percent GFP Positive
25 µF 2.48
500 µF 1.32
960 µF 0.92
Samples Percent GFP Positive
960 µF - 20µg 3.47
500 µF - 5µg 2.65
960 µF - 5µg 2.53
3 µF - 20µg 2.45
25 µF - 20µg 2.27
500 µF - 20µg 2.22
125 µF - 20µg 2.16
3 µF - 5µg 2.14
125 µF - 5µg 2.13
25 µF - 5µg 1.87
Percent GFP positive BALB/c whole bone marrow (WBM) cells after electroporation with the
Bio-Rad Gene Pulser and various amounts of plasmid pEGFP-C1. Each sample contained 4x10
6
cells and was performed in duplicate, shown is the average percent GFP positive as quantified
via flow cytometry. Capacitance is shown by µF and plasmid concentration measured in
micrograms (µg).

75
Table 3-3. Bio-Rad xCell Pulser electroporation efficiency on BALB/c WBM
Controls Percent GFP Positive
WBM, no DNA, no pulse 0.18
WBM, no DNA, pulse 0.28
pBEB 5 µg, no pulse 0.37
pMND-eGFP-1 5 µg, no pulse 0.35
Samples Percent GFP Positive
pBEB 10 µg 0.34
pBEB 20 µg 0.34
pMND-eGFP-2 5 µg 0.33
pMND-eGFP-2 20 µg 0.33
pMND-eGFP-1 20 µg 0.31
pMND-eGFP-2 10 µg 0.31
pMND-eGFP-1 5 µg 0.27
pBEB 5 µg 0.27
pMND-eGFP-1 10 µg 0.23
Percent GFP positive BALB/c whole bone marrow (WBM) cells after electroporation with the
Bio-Rad xCell Pulser. Two different preps of plasmid pMND-eGFP were used. Each sample
contained 5x10
6
cells and was performed in duplicate, shown is the average percent GFP
positive as quantified via flow cytometry. Plasmid concentration is measured in micrograms
(µg).

76
Table 3-4. Amaxa nucleofection efficiency on BALB/c WBM
Controls Percent GFP Positive
K562 2m 0.25
T 1m 0.22
N 2m 0.19
CD34 2m 0.19
K562 1m 0.14
CD34 1m 0.12
T 2m 0.11
KG1 1m 0.10
N 1m 0.06
Samples Percent GFP Positive
CD34 2m 20µg 4.77
CD34 1m 5µg 2.27
CD34 2m 5µg 1.38
KG1 1m 20µg 1.23
T 2m 5µg 0.93
T 2m 1µg 0.71
CD34 2m µg 0.60
CD34 1m 20µg 0.58
K562 2m 5µg 0.52
K562 1m 5µg 0.44
T 1m 20µg 0.44
KG1 1m 5µg 0.41
T 1m 5µg 0.38
T 1m 1µg 0.36
K562 1m 20µg 0.31
N 1m 1µg 0.29
N 2m 20µg 0.28
K562 2m 1µg 0.25
N 2m 1µg 0.25
K562 2m 20µg 0.22
T 2m 20µg 0.22
N 2m 5µg 0.20
N 1m 20µg 0.20
N 1m 5µg 0.16
K562 1m 1µg 0.14
CD34 1m 1µg 0.02
Percent GFP positive BALB/c whole bone marrow (WBM) cells after Amaxa nucleofection.
Percent GFP positive quantified via flow cytometry. Five kits human CD34, Kg1, mouse
BALB/c T-cell (T), human K562, and mouse neural stem cell (N) were used with either 1x10
6
or 2x10
6
. Plasmid pEGFP-C1 was used at concentrations of 1 µg, 5 µg, or 20 µg.

77
Table 3-5. Lin
-
HSC enrichment results in higher DNA uptake
Control % GFP+ Total Pop. % Viable %GFP+ of Viable Pop.
Lin- HSC, no DNA 0.39 43.5 0.72
Samples % GFP+ Total Pop. % Viable %GFP+ of Viable Pop.
pMax 8µg 37.40 19.1 59.2
pMax 8µg 34.60 16.4 34.7
pBEB 10µg 10.60 18.3 40.1
pBEB 10µg 9.79 20.9 33.9
pMND-eGFP-2 10µg 7.05 19.8 30
pMND-eGFP-2 10µg 6.15 21.4 24.5
Percent GFP positive BALB/c Lin
-
HSC 24 hours post Amaxa nucleofection. Green fluorescence and
cell viability was quantified via flow cytometry. Amaxa program U-08 and the CD34 Kit were used.
Sample pBEB contained 3x10
6
cells, pMND-eGFP-2 3x10
6
cells and pMax 2.25x10
6
cells. Thc
control reaction had 2.25x10
6
cells and did not receive any DNA.

78
Table 3-6. In vivo bioluminescence 21 days post Lin
-
HSC transplantation
n Cells Treatment n day 21 Avg Photons/Sec SE
7 2x10
6
Lin
-
1 µg pCLB + 1µg pCMVInt 6 4.033 x 10
5
4.883 x 10
4
4 2x10
6
Lin
-
+ 1.6x10
5
WBM 1 µg pCLB + 1µg pCMVInt 1 5.2381 x 10
5
0
2 2x10
6
Lin
-
no DNA 0 - -
2 2x10
6
Lin
-
+ 1.6x10
5
WBM no DNA 0 - -
1 5.0x10
5
WBM no Amaxa, no DNA 1 3.3045 x 10
5
0
2 1.6x10
5
WBM no Amaxa, no DNA 1 2.806 x 10
5
0
1 0 Naïve, no irradiation 1 3.7522 x 10
5
0
1 0 Irradiation Control 0 - -
Eight different treatments and control were tested for the ability of pCMVInt, which bears PhiC31, to
integrate pCLB. Bioluminescence imaging at 21 days post transplantation shows a slight significant
difference in the DNA groups compared to the ‘Cells, no DNA’ group (p-value: 0.06, Student t-test with
unequal variance). Legend: n, number of BALB/c mice per group; SE, standard error.

79
Figure 3-1. Schematic diagrams of primary plasmid constructs used for this study.
A. pCMVInt, expresses phiC31 integrase in mammalian cells. B. pBEB, eGFP driven
by CAGGS promoter and an SV40 polyA. C. pBGLB, luciferase driven by the CAGGS
promoter with an SV40 polyA and the attB recognition site for integration. D. pBLB,
similar to pBGLB, but with KanR resistance instead of AmpR and th attB site upstream
of the CAGGS promoter. E. pCLB, highest expressing luciferase plasmid measured in
this study, with a rabbit beta-globin polyA rather than SV40.

80
0.00E+00
1.00E+06
2.00E+06
3.00E+06
4.00E+06
5.00E+06
6.00E+06
7.00E+06
pBLB +
pCMVInt
pCLB +
pCMVInt
pBGLB +
pCMVInt
pMax
L
u
c

e
x
p
r
e
s
s
i
o
n

q
u
a
n
t
i
f
i
e
d

a
s

p
h
o
t
o
n
s
/
S
e
c
24 hrs
48 hrs
Figure 3-2. Co-delivery of plasmids encoding phiC31 integrase and luciferase into
Lin
-
HSC using the Amaxa nucleofector. Graph shows bioluminescence for three
different luciferase expressing constructs co-delivered with pCMVInt, which encodes
the phiC31 integrase. Plasmid pCLB was nearly two-fold higher than the next best
construct, pBGLB. Plasmid pMax, which only expresses eGFP, was delivered as a
negative control.

81


0.0E+00
5.0E+06
1.0E+07
1.5E+07
2.0E+07
no DNA +
Amaxa
Naïve pCLB pCLB +
pmInt
pCLB +
pCMVInt
pMax
L
u
c
i
f
e
r
a
s
e

m
e
a
s
u
r
e
d

i
n

p
h
o
t
o
n
s
/
s
e
c
o
n
d
24 Hours
80 Hours
Figure 3-3. Bioluminescence of Lin
-
HSC with and without active phiC31. A)
Enriched 2x10
6
Lin
-
HSC were nucleofected with pCLB (5 µg) + pCMVInt (10 µg)
(n=3), pCLB (5 µg) + pmInt (10 µg) (n=3), pCLB alone (5 µg) (n=2), pMax (5 µg)
(n=1), or no DNA (n=1).

Luciferase expression was quantified with an IVIS-200
bioluminescence imager 24 hours after Amaxa nucleofection. Red indicates regions of
highly intense luciferase expression while violet/blue indicates least intense. Expression
is measured in photons/second x 10
6
. B) Graphical representation of average
Naïve
pCLB
pCLB
+ pmInt
pCLB
+ pCMVInt
A.
B.
pMax
No DNA

82
bioluminescence measurements at 24 hours (lined) versus 80 hours (solid) with
appropriate standard error bars. At 80 hours post nucleofection, pCLB + pCMVInt is
one log higher (2.03x10
6
) than without phiC31 integrase (pCLB, 1.17x10
5
) or with an
inactive PhiC31, pmInt (1.82x10
5
). This suggests a PhiC31 effect in maintaining
luciferase expression.

83
0
10
20
30
40
50
60
70
80
90
no DNA +
Amaxa
Naïve pCLB pCLB +
pmInt
pCLB +
pCMVInt
pMax
P
e
r
c
e
n
t

V
i
a
b
l
e

(
P
I

e
x
c
l
u
s
i
o
n
)
24 Hours
80 Hours
Figure 3-4. Viability of Lin
-
HSC co-delivered via Amaxa nucleofection with
plasmids encoding luciferase and phiC31. Lin
-
HSC were assessed for cell viability at
24 hours (lined) and 80 hours (solid) after nucleofection. The five groups each had cell
densities of 2x10
6
Lin
-
HSC and were nucleofected with pCLB (5 µg) + pCMVInt (10
µg) (n=3), pCLB (5 µg) + pmInt (10 µg) (n=3), pCLB alone (5 µg) (n=2), pMax (5 µg)
(n=1), or no DNA at all (n=1).

Cells were stained with propidium iodide for viability
and analyzed via flow cytometry. By 80 hours, there was no significant difference in
viability for the three groups that received pCLB (p-value = 0.354).
Naïve

84
Figure 3-5. Bioluminescence imaging of BALB/c mice 21 days post transplantation
of Lin
-
HSC. Ventral view of four mice that had received 2x10
6
Lin
-
HSC nucleofected
with pCLB and pCMVInt (left most four). A fifth control mouse, which received 5 x
10
5
WBM cells, but no DNA, is shown on the far right. A modest difference in
expression was seen between the group receiving pCMVInt and controls (p-value:
0.06). Luciferase expression is shown as pseudo color images of photon intensity. Blue
represents less bioluminescence intensity, while read indicates higher bioluminescence.
The scale is in photons of bioluminescence per second. See table 2-6 for detailed photon
measurements.

85
0
20
40
60
80
100
120
5ug
pCLB
,5ug Int
5ug
pCLB
,5ug Int
5ug
pCLB
,1ug Int
5ug
pCLB
,1ug Int
3ug
pCLB
,3ug Int
3ug
pCLB
,3ug Int
3ug
pCLB
,1ug Int
3ug
pCLB
,1ug Int
Neg + PI Neg no
PI
5ug
pCLB
,5ug mInt
5ug
pCLB
,1ug mInt
5ug
pCLB
,1ug mInt
3ug
pCLB
,3ug mInt
3ug
pCLB
,1ug mInt
P
e
r
c
e
n
t

V
i
a
b
l
e
0.00E+00
2.00E+05
4.00E+05
6.00E+05
8.00E+05
1.00E+06
1.20E+06
1.40E+06
1.60E+06
1.80E+06
P
h
o
t
o
n
s
/
S
e
c
o
n
d
% Viable 72 hrs Expression
Figure 3-6. Effect of DNA concentration on Lin
-
HSC viability and luciferase
activity in culture. In this graph, luciferase expression is represented with lighter bars
overlayed on the dark bars, which represent percent viability. The left-hand Y-axis
corresponds to cell viability and the right Y-axis to the luciferase expression. We
studied the effect that varying total DNA concentration would have on cell viability. We
also studied the ratio of phiC31 (Int) to donor pCLB that was needed to achieve the
highest donor gene expression. Results indicate that there was no significant difference
in cell viability across groups, but the ratio and amount of DNA plasmids were
important. Five micrograms each construct, pCLB + Int (pCMVInt), produced optimal
results. After 72 hours, the catalytically active phiC31 (Int) maintained donor gene
expression at higher levels than inactive phiC31 (mInt). Viability was assessed through
flow cytometry and luciferase expression with quantified by bioluminescence imaging.
Each nucleofected treatment had 4.0x10
6
Lineage depleted HSC, including the negative
DNA control (Neg), which was nucleofected. PI = propidium iodide.

86
REFERENCES
1. Weatherall DJ, Clegg and JB. 2001. Inherited haemoglobin disorders: an
increasing global health problem. Bull. World Health Organ. 79:704-12.
2. Yanovich S, Mitsky P, Cornetta K, Maziarz RT, Rosenfeld C, Krause DS, Lotz
JP, Bitran JD, Williams S, Preti R, Somlo G, Burtness B, and Mills B. 2000.
Transplantation of CD34+ peripheral blood cells selected using a fully automated
immunomagnetic system in patients with high-risk breast cancer: results of a
prospective randomized multicenter clinical trial. Bone Marrow Transplant
25:1165-74.
3. Biffi A and Cesani M. 2008. Human hematopoietic stem cells in gene therapy: pre-
clinical and clinical issues. Curr. Gene Ther. 8:135-46.
4. Pike-Overzet K, van der Burg M, Wagemaker G, van Dongen J, and Staal F.
2007. New insights and unresolved issues regarding insertional mutagenesis in X-
linked SCID gene therapy. Mol. Therapy 15:1910-6.
5. Shou Y, Ma Z, Lu T, and Sorrentino BP. 2006. Unique risk factors for insertional
mutagenesis in a mouse model of XSCID gene therapy. Proc. Natl. Acad. Sci. USA
103:11730-5.
6. Persons DA, Hargrove PW, Allay ER, Hanawa H, and Nienhuis AW. 2003. The
degree of phenotypic correction of murine beta -thalassemia intermedia following
lentiviral-mediated transfer of a human gamma-globin gene is influenced by
chromosomal position effects and vector copy number. Blood 101:2175-83.

87
7. Arumugam PI, Scholes J, Perelman N, Xia P, Yee JK, and Malik P. 2007.
Improved Human beta-globin Expression from Self-inactivating Lentiviral Vectors
Carrying the Chicken Hypersensitive Site-4 (cHS4) Insulator Element. Mol.
Therapy 15:1863–71.
8. Sieburg HB, Cho RH, Dykstra B, Uchida N, Eaves CJ, and Muller-Sieburg CE.
2006. The hematopoietic stem compartment consists of a limited number of discrete
stem cell subsets. Blood 107:2311-6.
9. Li SD and Huang L. 2006. Gene therapy progress and prospects: non-viral gene
therapy by systemic delivery. Gene Therapy 13:1313-19.
10. Hollis RP, Nightingale S, Calos MP, Crooks GM, and Kohn DB. 2004. Site-
specific genomic integration in hematopoietic cells. Mol. Therapy 9:S397.
11. Hollis RP, Nightingale SJ, Wang X, Pepper KA, Yu XJ, Barsky L, Crooks GM,
and Kohn DB. 2006. Stable gene transfer to human CD34(+) hematopoietic cells
using the Sleeping Beauty transposon. Exp. Hematol. 34:1333-43.
12. Gresch O et al. 2004. New non-viral method for gene transfer into primary cells.
Methods 33:151-63.
13. Von Levetzow G, Spanholtz J, Beckman J, Fischer J, Kogler G, Wernet P,
Punzel M, nad Giebel B. 2006. Nucleofection, an efficient nonviral m ethod to
transfer genes into human hematopoietic stem and progenitor cells. Stem Cells and
Development 15:278-85.
14. Groth AC, Olivares EC, Thyagarajan B, and Calos MP. 2000. A phage integrase
directs efficient site-specific integration in human cells. Proc. Natl. Acad. Sci. USA
97:5995-6000.

88
15. Thyagarajan B and Calos MP. 2005 Site-specific integration mediated by phiC31
integrase for high level protein production. In Methods in Molecular Biology,
Therapeutic Proteins: Methods and Protocols. Smales CM , James DC (eds).
Humana Press: Totowa, NJ, 308:99-106.
16. Niwa H, Yamamura K, and Miyazaki J. 1991. Efficient selection for high-
expression transfectants with a novel eukaryotic vector. Gene 108(2):193-9.
17. Shinde R, Perkins J, and Contag CH. 2006. Luciferin derivatives for enhanced in
vitro and in vivo bioluminescence assays. Biochemistry 45:11103-12.
18. Rice BW, Cable MD, and Nelson MB. 2001. In vivo imaging of light-emitting
probes. J. Biomed. Opt. 6:432-40.
19. Olivares EC, Hollis RP, Chalberg TW, Meuse L, Kay MA, and Calos MP.
2002. Site-specific genomic integration produces therapeutic factor IX levels in
mice. Nature Biotechnology 20:1124-8.
20. Gresch O et al. 2004. New non-viral method for gene transfer into primary cells.
Methods 33:151-63.
21. Lakshmipathy U et al. 2004. Efficient transfection of embryonic and adult stem
cells. Stem Cells 22:531-43.
22. Gärtner A, Collin L, and Lalli G. 2006. Nucleofection of primary neurons.
Methods Enzymol. 406:374-88.
23. Aluigi M, Fogli M, Curti A, Isidori A, Gruppioni E, Chiodoni C, Colombo MP,
Versura P, D'Errico-Grigioni A, Ferri E, Baccarani M, and Lemoli RM. 2006.
Nucleofection is an efficient nonviral transfection technique for human bone
marrow-derived mesenchymal stem cells. Stem Cells 24:454-61.

89
24. Andreason GL and Evans GA. 1989. Optimization of electroporation for
transfection of mammalian cell lines. Anal. Biochem. 180:269-75.
25. Christensen JL and Weissman IL. 2001. Flk-2 is a marker in hematopoietic stem
cell differentiation: a simple method to isolate long-term stem cells. Proc. Natl.
Acad. Sci. USA 98:14541-6.
26. Smith MC, Till R, Brady K, Soultanas P, Thorpe H, and Smith MC. 2004.
Synapsis and DNA cleavage in phiC31 integrase-mediated site-specific
recombination. Nucleic Acids Res. 32:2607-17.
27. Schneider, K. 2005. Anaesthesia (inhalative or i.v. narcosis) of laboratory animals.
Practical Methods in Cardiovascular Research. Springer Berlin: Heidelberg,
Germany. Chapter 1.1.
28. Wu MH, Smith SL, and Dolan ME. 2001. High efiiciency electroporation of
human umbilical cord blood CD34+ hematopoietic precursor cells. Stem Cells
19:492-9.
29. Wagers AJ, Sherwood RI, Chrisensen JL, and Weissman, IL. 2002. Little
evidence for developmental plasticity of adult hematopoietic stem cells. Science
297:2256-9.
30. Portlock JL, Keravala A, Bertoni C, Lee S, Rando TA, and Calos MP. 2006.
Long-term increase in mVEGF164 in mouse hindlimb musle mediated by phage
phiC31 integrase after nonviral DNA delivery. Human Gene Therapy 17:871-6.
31. Aneja MK, Imker R, and Rudolph C. 2007. Phage phiC31 integrase-mediated
genomic integration and long-term gene expression in the lung after nonviral gene
delivery. J. Gene Med. 9:967-75.

90
32. Hollis RP, Stoll SM, Sclimenti CR, Lin J, Chen-Tsai Y, and Calos MP. 2003.
Phage integrases for the construction and manipulation of transgenic mammals.
Reprod. Biol. Endocrinol. 1:79-91.
33. Chalberg TW, Portlock JL, Olivares EC, Thyagarajan B, Kirby PJ, Hillman
RT, Hoelters J, and Calos MP. 2006. Integration specificity of phage phiC31
integrase in the human genome. J. Mol. Biol. 357:28-48.
34. Zielske SP, Reese JS, Lingas KT, Donze JR and Gerson SL. 2003. In vivo
selection of MGMT(P140K) lentivirus–transduced human NOD/SCID repopulating
cells without pretransplant irradiation conditioning. J. Clin. Invest. 112:1561-70.
35. Ball CR, Pilz IH, Schmidt M, Fessler S, Williams DA, von Kalle C, and Glimm
H. 2007. Stable differentiation and clonality of murine long-term hematopoiesis
after extended reduced-intensity selection for MGMT P140K transgene expression.
Blood 110:1779-87.