From www.bloodjournal.org by guest on June 14, 2017. For personal use only.
Expression of the Blood-Clotting Factor-VI11 cDNA Is Repressed by a
Transcriptional Silencer Located in Its Coding Region
By Rob C. Hoeben, Frits J. Fallaux, Steve J. Cramer, Diana J.M. van den Wollenberg, Hans van Orrnondt,
Ernest Brilt. and Alex J. van der Eb
Hemophilia A is caused
by adeficiency offactor-Vlll procoagulant (NIII)activity. The currenttreatment by frequent infusions of plasma-derived Nlll concentrates is very effective
but has the risk of transmittanceof blood-borne viruses (human immunodeficiency virus IHIVI, hepatitis viruses). Use
of recombinant DNA-derived Nlll as well as gene therapy
could make hemophiliatreatment independent of blood-derived products. So far, the problematic production of the
Nlll protein and the low titers of the Nlll retrovirus stocks
have prevented preclinical trialsof gene therapy for hemophilia A in large-animal models. We have initiated a study
of the mechanisms that oppose efficientNlll synthesis. We
have established that Nlll cDNA contains sequences that
dominantly inhibitits own expression from retroviralas well
as from plasmid vectors. The inhibition is
not causedby
instability of the Nlll mRNA (tqI2,a 6 hours) but rather to
repression at the level of transcription. A 305-bp fragment
is identified that is involved in but not sufficient for repression.This fragment does not overlap the region recently
identified by Lynch et al (Hum Gene Ther 4259,1993) as a
dominant inhibitor of RNA accumulation. The repression is
mediated by acellular factor (or factors) and is independent
of the orientation of the element in the transcription unit,
giving the repressor element the hallmarks of a transcriptional silencer.
0 1995 by The American Society of Hematology.
H
some posttranscriptional event,8.13the intracellular transport
of W111 is very inefficient,4'I4 and, once secreted, the protein
is extremely susceptible to proteolytic degradation and needs
to be stabilized by the von Willebrand factor (vWF) ~ r o t e i n . ~
We report evidence for the presence of a dominant transcriptional silencer element in the fVIII cDNA and map the
fVIII cDNA sequences involved. Characterization of this
element will aid the development of more efficiently expressed fVIII cDNA clones. Increased expression would not
only advance the development of gene therapy protocols
but could also increase yields of recombinant-DNAderived
fVIII and, thus, will reduce the currently high price of this
product.
EMOPHILIA A is a bleeding disorder that affects approximately 8.5 in 100,OOO humans and is caused by
a deficiency of functional blood coagulation factor VnI
(fVII1). For those afflicted, hemophilia is potentially lifethreatening and crippling. Hemophilia A is treated by regular
infusions (prophylactic or on-demand) with fVIII-enriched
concentrates derived from human plasma. This protein-replacement therapy has been very effective and has resulted
in a dramatic increase in both the life expectancy and the
quality of life.' Nonetheless, even the treated patient is prone
to spontaneous hemorrhages with the associated risk of
chronic joint damage. In addition, therapy with plasma-derived NI11 has resulted in transmission of several human
viruses, such as human immunodeficiency virus (HIV) and
hepatitis virus. The ideal therapy, therefore, would be independent of blood-derived products? The recent admission
of recombinant DNA-derived fVIII for treatment of hemophilia A3 is a major step in that direction, although its extensive use is still problematic due to its high cost, which results
from the difficult production of this p r o d ~ c tSomatic
.~
gene
therapy, therefore, is deemed to be a welcome alternative
for hemophilia
Implantation of autologous cells,
genetically modified to produce fVIII, would allow a continuous production of fVIII in the patient, thereby preventing
the spontaneous bleeding episodes.
Somatic gene therapy for hemophilia A is not necessarily
restricted to the cells that normally produce fVIII. Research
has been directed at genetic modification of cells from tissue
types that are more amenable to manipulation, such as the
~ k i n ,endothelium,
~.~
bone marrow: and muscle." To date,
retrovirus-mediated gene transfer is the most tested system
in clinical trials of gene therapy.6," To accommodate factor
VI11 cDNA into the retroviral vectors, we and others have
used shortened cDNA clones from which the region encoding the nonessential internal B-domain has been deleted.7~8~'2
Such vectors have been successfully used to produce functional fVIII in primary human cell^.^.^ However, the relatively low titer of fVIII retroviruses and the disappointingly
low protein yields have hampered the initiation of preclinical
trials in large-animal models. Several mechanisms have been
identified that thwart the production of fVIII. Accumulation
of fVIII mRNAis inhibited, purportedly as the result of
Blood, Vol 85, No 9 (May l ) , 1995: pp 2447-2454
MATERIALSANDMETHODS
Cell lines and virus preparation. Cell lines were grown in highglucose (4.5 g/L) Dulbecco's modified Eagle's medium supplemented with 8% fetal calf serum in a 5% CO2 atmosphere at 37°C.
Generation of virus-producing cell lines and retrovirus harvesting,
infection, and titration were performed as described? Polyclonal cell
lines containing a single copy ofthe provirus were generated by
infection of near-confluent cell cultures with a low multiplicity of
infection (moi; <0.01) and selection in 400 pg G418 per milliliter.
The resulting G418R polyclonal cell populations were maintained
in medium containing 200 pg G418 per milliliter until use. In all
From the Laboratory of Molecular Carcinogenesis, Department
of Medical Biochemistry, University of Leiden, Leiden; and the Department of Hematology, University Hospital, Leiden, The Netherlands.
Submitted May 2, 1994; accepted December 14, 1994.
Supported in part byGrant No. 2821 780 fromthe Dutch Praeventie Fonds.
Address reprint requests to Rob C. Hoeben, PhD, Laboratory of
Molecular Carcinogenesis, Department of Medical Biochemistry,
Sylvius Laboratory, Wassenaarseweg 72, 2333AL Leiden, The Netherlands.
The publication costs of this article were defrayed in part by page
charge payment. This article must therefore be hereby marked
"advertisement" in accordance with 18 U.S.C. section 1734 solely to
indicate this fact.
0 1995 by The American Society of Hematology.
wo6-4971/95/8509-0013$3.00/0
2447
From www.bloodjournal.org by guest on June 14, 2017. For personal use only.
2448
experiments, polyclonal cultures were used at low passage number.
ThepackaginglinesPsi2andPA3
l 7 producingtheMSneo,
M5F8dB2.6: andXT-HERI5viruseshavebeendescribedpreviously. The XT-HER vector contains the human epidermal growth
factor-receptor (EGF-r) cDNA, drivenby an internal herpes simplex
virus-thymidekinase gene promoter. In addition XT-HER contains
the neomycin-resistancegene(neoR)driven
by thelongterminal
repeat (LTR) promoter. In the infected cells, mRNAs be
candetected
of 8.3 to 7.8 kb (with a neoR probe) and S kb (only with a EGF-r
probe). The M5FIX retroviral vector was constructed
by insertion
of a 1.4-kb BamHI containing the complete coding region of
the
human factor IX cDNA (provided by Dr E. Davie, University of
Washington, Seattle) into the unique BamHt site
of plasmid M5ne0.~
Polyclonalcultures
of RAT2cellswithchloramphenicol-acetyl
transferase (CAT)-reporter plasmids integrated in their genome were
generated by cotransfection of 1 pg pRSVneo,16 2 pg pMuLV-CAT
(or equimolaramountsofitsderivativeplasmidscarrying
fVIII
cDNA fragments), and 7 pg salmon-sperm DNA per 10-cm tissue
culture dish. The resulting G418R polyclonal cell population (consisting of greater than
50 transfectants) was maintained in G418containing (200 p g h L ) medium and used at low passage number.
RNA analyses. Total cellular RNA was isolated from near-confluent cell cultures by the LiCI-urea method,I7 and Northern blotting
analyses were performed on nitrocellulose. To study the stability of
the mRNA, actinomycin D (ActD; Sigma, St Louis, MO)was added
to near-confluent cultures at zero time to a final concentration of 5
pg/mL. RNA was isolated at the desired timepoints and analyzed
by Northern analysis.
Probes. The neoR-specific probe used for the Northern analysis
was the 1.3-kb
HindIII-Sma
I fragment
of
pRSVneo."
The
Sul I-Sal I fragmentfrom
NIII-specificprobewasthe4.8-kb
pUC-F8dB2.6.' The glyceraldehyde-3-phosphate (GAPDH)-specific
probe was the 1.3-kb Pst I fragment from plasmid pRGAPDH-13,"
representingtheentireratcDNA.The
c-myc-specific probe used
was the 1.3-kb Cla I-EcoRI fragment of the human
c-myc gene."'
Probeswereradiolabeledwith
[~t-~~P]deoxycytidine
triphosphate
(dCTP) with the use of random hexanucleotide primers and Klenow
Escherichia coli DNApolymerase I (Pollk). To obtainsinglestrandedprobesforthenuclearrun-onanalysis,theneoRand
GAPDH fragments were cloned into M13mp18 and MI3mp19. A
single-stranded NIII-specific probe was constructed by insertion the
658-bpHindIItfragment,encoding
part of the AIdomain, into
phage M13mp18 DNA.
CAT plasmids and assays.
Plasmid
DNA
( I O pg) was
transfected into near-confluent cultures of HeLa cells with the calcium-phosphate precipitation procedurein a 9-cm tissue-culture dish.
At 40 hours after transfection, the cells were lysed, and
SO p g of
proteinlysatewasusedfortheCATactivityassays.lhPlasmid
pMuLV-CAT was used to assess the effect of the fVIII cDNA fragments on the expression
of the marker (CAT) gene. Plasmid pMuLVCAT contains the enhancer and promoter from the Moloney murine
leukemiavirus(MO-MuLV)LTR
to drivetheCAT
gene.'" The
NI11 cDNA fragments were inserted into the unique Hpa I site, 34
and 63 bp 5' ofthetwotandemSV40polyadenylationsignals,
derived from the late transcription unit. Unless stated otherwise, O S
pg of plasmid pMuLV-CAT or equimolar amounts of the clones
canying fVIII cDNA fragments wereused in the transfection experiments. The amount of plasmid DNA was adjusted to 10 pg per dish
with plasmid pUC18 DNA. In the
Hpa I site of plasmid pMuLVCAT, the following fragments(see Fig S) have been tested for their
effect on expression (nucleotide numbering according to Wood et
al"): SX (Sal I-Xho I, -30 to 2294), XS (Xho I-Sal I, 4968 to 743%
SA (Sal I-Ace I, -30 to 1033), HH (HindIII-HindIII, 377 to 1019),
Ssp (Sal I-Spe I, -30 to S S ) , AH (Acc I-HindIII, 1033 to 2277),
HB(HindIII-Bgl 11, 1019 to1680). BH (Bgl 11-HindlII, 1680 to
HOEBEN ET AL
2277).PP (Pvu 11-Pvu 11, 1038 to 18SO), SXdXmB (Sul I-xmn I.
-30 to1375fusedto
Bgl 11-Xho I, 1680 to2294),AHdXmB
(Ace I-Xmn I, 1033 to 1375 fused to Bgl 11-Xho I, 1680 10 2294),
XmB (Xmn I-BgI 11, 1375 to 1680),XmBm (Xmn I-BtrmH1. 1375
to 1869),and AHWBm (Acc I-Bgl 11, 1033 to 1680 fused to BarrzHtHindIIt, 1869 to 2277). All sites mentioned are present i n the tVItI
cDNA, except the Sal I sites, which flank the F8dB2.6 cDNA, and
the Xhn I site, which has been created at the junctions of the
Bdomaindeletion. All fragments were isolatedfrom plasmid pUCF8dB2.6.7
Factor V/// und /X us.suys. Human f V I l I in tissue-culturemedium was quantitated via an enzyme-linked immunosorbent assay
(ELISA), as described p r e v i ~ u s l y .Human
~
AX was determined by
ELISA as described.*' In all assays, appropriate dilutions of pooled
normal human plasma were used as references. The concentrations
of N I 1 1 and flX in the pools were assumed to be 100 ng/mL and 3
pg/mL, respectively.
Nucleur run-on assays.Nucleiwereisolatedfrom
polyclonal
cell cultures by NP40 lysis as described" and stored at -80°C in a
solution containing 50 mmol/L Tris-HCI (pH 8.3). 40% (voVvol)
glycerol, 5 mmol/L MgCI2, and 0.1 mmol/L EDTA. In vitro transcription reactions with IO' nuclei per incubation were performed
for 30 minutes at 37°C in the presence of 20 mmol/L Tris-HCI (pH
8.0), 4 mmol/L MgCI2, 1 0 0 mmol/L KCI, 20% (vol/vol) glycerol,
and 250 pCi [cu-32PJuridinetriphosphate (UTP) in 200 pL. Subsequently,coldUTPwasadded
to 100 pmol/L, andtheincubation
was continued forI O minutes. The samples were treated
with RNasefree DNase (0.25 U/mL) and, subsequently, with proteinase K (100
pg/mL). The labeled RNA was precipitated and dissolved in hybridization buffer(40%[vol/vol]formamide,
20 mmol/L PIPES[pH
7.51 0.2% sodium dodecyl sulfate [SDS], 0.4 m o l L NaCI, 200 p&/
mL yeast tRNA, 1 X Denhardt's solution). Aliquots were counted,
andidenticalamounts
of radiolabeledRNAwereaddedtoeach
probe filter and hybridized for 48 hours
at 42°C. The probe filterh
were prepared by spotting 2 pg of the single-stranded phage-M I3
clones onto 0.22-pm nitrocellulose membranes. Afterthe hybridization, the filters were washed i n 2X standard sodium citrate (SSC),
0.5% SDS at 50°C and subsequently incubated in I O mmol/L TrisHCL (pH 7.6), 300 m m o l L NaCI, and I O pg/mL RNase A (DNasefree): washed in 2x SSC. 0.1% SDS at 42°C; and left toexpose
Kodak XAR-S films (Eastman-Kodak, Rochester, NY ).
RESULTS
The W111 protein consists of three types of structural domainsthat are arranged in the order Al:A2:B:A3:Cl:C2.
Theunique B domaindelimited by aminoacids 740 and
1648 is removed during the proteolytic processing and activation and isdispensable forcoagulation activity in vivo
and in vitro."." To accommodatethe NI11 cDNA into a
retroviral vector without exceeding the packaging capacity
of the vector, we used a NI11 cDNA from which we had
deleted virtually all codons for the B domain. We have previously shown that the resulting B domain-deleted NI11 is
fully f~nctional.~ The
vector used relies on differential splicing for the expression of the NI11 cDNA and the neoR gene
(Fig IA). The titers of the NI11 vectors ( 5 3 X IO' G418R
colony-forming units [CFU]/mL) are much lower than those
routinely obtained with other vectors ( IO5 to l O6 CFU/mL),
as is the amount of NI11 protein secreted. For example, the
amount of NI11 protein is at least 200-fold lower on a molar
basis than that of coagulation factor IX from cells infected
fIX cDNA (Table 1 ) . Our
with a retroviral vector carrying
attempts to increase the titer by using other vectors and by
From www.bloodjournal.org by guest on June 14, 2017. For personal use only.
TRANSCRIPTIONSILENCING BY FACTOR Vlll cDNA
2449
including an extended packaging signal""'7 into the M5neo
vector did not yield any high-titer virus producer. In contrast,
the vectors invariably lost (parts of) the N I 1 1 cDNA (data
not shown).
Sequences in thefllll cDNA repress the acc~trnulationof
its RNA. It has been reported that sequences in the fVlll
cDNA prevent accumulation of W111 mRNA,' supposedly
by destabilizing the fVIIl transcripts.I3 This could explain
Table 1. Comparison of the Clotting Factor Production of Nlll and
flX Retrovirus-Transduced Rat2 and VHlO Cells
RAT2
NI11
flX
Ratio flX:NIII
VHlO (human)
ng
fmol
ng
fmol
4
210
53x
22
4,500
205X
3
530
177X
11,350
709X
16
Production is calculated per 10' cells per day.
A
B-domain W
e
d fador Vlll
LTR
Neo
L
2
SD
SA
LTR
...........................................................................................................................
......
RNA
.........
.....
.....
#AA
' 3 . .
. . S '
B
1
2
3
2
4
8.1
2.5
neo
"
Fig l. Repression of expression b y NI11 cDNA fragments. (A) In
the retroviral vectorIM5F8dB2.6') used in this study, transcription is
driven by the myeloproliferative sarcoma virus LTR. The unspliced
mRNA, from which the B-domain-deleted NI11 can be translated,
and the spliced messenger, from which theneon is derived, are indicated: SA and SD, the splice-donor and acceptor sites, respectively;
AAA, the polylA)-tail of thetranscripts. (B) Northern blotanalysis of
the G418RRat2 cell cultures after infection
with the retroviral vectors
M5neo (lane l),
M5F8dB2.6 (lanes 2 and 41, and XT-HER (lane 3). At
24 hours before RNA isolation, the cells were fedfresh medium (lanes
1 through 3) or fresh medium with 6 mmol/L sodium butyrate (lane
4).RNA (20 p g ) was electrophoresed on a formaldehyde-agarose
gel and blotted onto a nitrocellulose membrane. Hybridization was
performed with a neon- and a GAPDH-specific probe, as indicated.
both the low virus production by the packaging cells and
the low amounts of N I 1 1 protein secreted by the f V l I I virusinfected cells. A Northern analysis of a G41SR population
of Rat2 fibroblasts that contain the M5F8dB2.6 vector confirmed a low abundance of vector-specific RNA compared
with cells infected with the parental MSneo vector (Fig 1 B),
in which both the unspliced and spliced mRNAs are distinct.
This is not due to the large size of the W111 cDNA insert,
as is shown by the amounts of 8.3- to 7.8-kb EGF-r mRNA
in cells harboring a vector carrying the human EGF-r cDNA,
which is similar in size to the B domain-deleted f V l I I cDNA.
The absence of rearrangements in the M5F8dB2.6 vector
was confirmed by a Southern blot assay on DNA extracted
from the infected Rat2 cell population (data not shown).
The integrity of the vector was further evidenced by the
appearance of W111 RNA of the expected size after treatment
of the cells with sodium butyrate (Fig. IB, compare lanes 2
and 4; in lane 4, both the unspliced 8.1-kb and the spliced
2.5-kb messengers are visible). The repression of W111
mRNA accumulation was not unique for RAT2 cells but has
been found in all cell lines tested; eg, Swiss3T3 cells, Psi-2
cells, murine endothelioma cells, and human skin fibroblasts
(data not shown).
Factor VI11 mRNA is stable in the packaging cells. TO
study the stability of the N
I11 transcripts in the packaging
cell line Psi2:M5F8dB2.6, transcription was blockedwith
ActD, an inhibitor of RNA polymerase 11. At various times,
RNA was extracted from the cell cultures and analyzed by
Northern blotting (Fig 2). No significant decrease was observed up to 6 hours after addition of the ActD, suggesting
that the W111 transcripts are rather stable in these cells.
Rehybridization of this blot with a c-myc-specific probe
showed the expected rapid decay of the c-myc transcripts
[half-life (tin), approximately 15 minutes2'), confirming the
effectivity of the ActD block. The amount of RNA loaded
was verified by reprobing the blot with a GAPDH-specific
probe, anmRNAknown
to have a longhalf-life.''From
these data we conclude that the steady-state W111 mRNA is
stable (tin, 2 6 hours). The fact that not only accumulation
of the unspliced messenger is inhibited in the MSF8dB2.6infected cells, but also that of the spliced neoR messenger,
which does not contain any fVIII, suggests. thatrepression acts before splicing occurs, possibly at the level
of transcription. The W111 viruses produced by the Psi2:
MSF8dB2.6 cell line have been used to generate the WIIIproducing rodent cell lines described here, thus excluding
the possibility that the W111 vector in this cell line contains
RNA-stabilizing mutations.
From www.bloodjournal.org by guest on June 14, 2017. For personal use only.
-=
I
HOEBEN ET AL
2450
0 30 60 90180360
Dc
l
,
-
0 30 60 90180360
emyc
b
..
! '
.
neo
.
-:.l
mmmmmo
GAPDH
Fig2. Stability of the fvlll vector mRNA in packaging cells.To
study the stability of the mRNA, ActD (final concentration,5 FglmL)
was added to near-confluentculturesof Psi2:M5F8dB2.6 cells
at zero
time. RNA was isolated at the indicated timepoints (in minutes) and
analyzed by Northern blotting. Hybridization was performed with a
neoR-specificprobe (left panel) and, after stripping of the probe, with
c-myc-and GAPDH-specific probes as indicated.The open and closed
arrowheadsdepict the unspliced and the spliced messengers, respectively.
Expression is repressed at the level qf transcription. A
nuclear run-on analysis was performed to measure the transcription rate of the integrated f V I I I provirus in both G41gR
Rat2:M5F8dB2.6 cells and G41gK Rat2:MSneo cells. As
shown in Fig 3, the neoR probe detects a signal only in
RNA isolated from the M5neo-infected cells. In the Rat2:
M5F8dB2.6 cells, neither the neoKprobe nor the NIII-specific probe shows any transcription of the provirus. The absence of signal with the antisense neo' probe confirms the
specificity of the hybridization. The equality of labeling efficiencies was verified by a probe specific for the cellular
GAPDH transcripts. Thus, insertion of the fVlll cDNA into
the M5neo vector reduces the transcription of the integrated
provirus to very low levels.
Factor VI11 cDNA fragments repress expression . f a heterologous gene. To study whether fVlll cDNA sequences
can repress expression of a heterologous gene, f v l l l cDNA
fragments were cloned in the pMuLV-CAT expression vector (see Fig 5). In this vector, expression of the CAT reporter
gene is driven by the MO-MuLV LTR.'" To maximally
mimic the natural configuration of the repressor elements,
the f V I I I cDNA fragments were introduced into the unique
Hpa I site of the expression plasmid. This site is located
within the transcription unit but 3' of the CAT-encoding
open reading frame so as not to influence translation of the
CAT message. Two fragments were tested: SX, encompassing domains A1 and A2, and XS,which encodes domains A3, Cl, and C2 (see Fig 5). Initially, both fragments
were tested in the orientation they had in the retroviral vector.Insertion of the XS fragment did not cause a lower
CAT activity than the parental pMuLV-CAT vector, whereas
insertion of the SXfragment decreased CAT activity dramatically (Fig 4, left panel). A similar decrease of CAT activity
was observed when the SX fragment was tested in the reverse
orientation. In stably transfected polyclonal cell cultures too,
the SX fragments repressed expression of the MuLV-driven
CAT-reporter gene (Fig 4, center panel). Northern blot anal-
ysis of RNA extracted from these cultures showed a vastly
reduced amount of CAT mRNA in cells carrying the
pMuLV-CAT/SX plasmid, which confirms thatthe
decreased CAT activity in the SX fragment-containing cell
population is reflecting the reduced amounts of CAT mRNA
(data not shown). From these data we conclude thatthe
repressor element that resides in the W111 cDNA inhibits the
expression of integrated as well as unintegrated heterologous
genes and that this repression is independent of the orientation of the f V I I l cDNA insert.
To map the sequences responsible for this repression in
more detail, various subfragments have been tested for their
capacity to reduce CAT activity (Fig S). The AH fragment
was a far more potent repressor than SA. The smallest fragment that could bring about repression was the PP fragment.
Smaller fragments derived from the AH fragment, eg, HB,
BH, XmB, and XmBm, did not decrease CAT activity. However, when the 305-bp XmB fragment, which by itself does
not repress expression, was deleted from AH, the resulting
clone AHdXmB was unable to repress expression. Deletion
of the XmB fragment from the SX fragment, too, disrupted
the repressor, as witnessed by the decreased CAT activity
after transient expression (Fig 5) and in stably transfected
polyclonal cell populations (Fig 4, center panel). From these
data, we conclude that sequences in the 305-bp XmB fragment are part of the repressor element, but are not sufficient
to mediate repression.
The repression can be relieved by competitor DNA. In
a competition experiment, the repression induced by the SX
and the SXdXmB fragments were compared in the absence
and in the presence of a 40-fold excess of the competitor
plasmid, pUC-AH. The SX fragment represses expression
I::I
M5F8dB2.6
M5neo
Fig 3. Inhibition of transcription in Nlll vector-infectedRat;! cells.
Transcription ofthe M5F8dB2.6retroviral vector was compared with
the parental (empty) M5neo vector in a nuclear run-on analysis. Nuclei from polyclonal,
G418-resistant,
near-confluent
cultures
of
M5neo- and M5F8dB2.6-infectedRat2 cells were isolated and incubated in the presence of labeled UTP, as described in Materials and
Methods. The labeled RNAs were hybridized with single-stranded
M13 probes, spotted onto nitrocellulosefilters, containingthe following inserts: position 1, neo probe (sense-orientation);position 2, neo
probe (antisenseorientation); position 3, GAPDH probe (sense orientation); and position 4, NIII-A1 domain (sense orientation). After hybridization, the filters were washed and subjected to autoradiography.
From www.bloodjournal.org by guest on June 14, 2017. For personal use only.
TRANSCRIPTIONSILENCING BY FACTOR Vlll cDNA
1
2
3
4
2451
5
6
7
8
9
0.1 0.1 26 18 0.0
26
25
22 0.1
10 11 12 13
"
sx, sxb xs
-
+
XS SXdXmB S&
L
% conv.
0.5 3.34.38.6
+
sx
-
+
SXdXmB
A2-Comp.
Fig 4. (Left panel) Factor Vlll cDNA fragments repress expression of a CAT reporter gene. The Nlll cDNA fragments SX and XS (encoding
the AlA2 and the
A3ClC2 domains of the
Nlll protein, respectively) were insertedinto theHpa I site of the
pMuLV-CAT reporter plasmid. The
resulting clones (0.5 p g l were transfected into RAT2 cells, and CAT activity was determined in protein lysates of the transfected cells as
described in Materials and Methods. The conversion of the substrate (96) is indicated below the figure. The transfected plasmids were as
follows: lane 1, SX fragment (orientation as in retroviral vector); lane 2, SX fragment (reverse orientation); lane 3, XS fragment (orientation
as in retroviral vector); lane 4, pMuLV-CAT without Wlll cDNA fragment; lane 5, the promoterless pBLcat5 plasmid. (Center panel) Fragments
of the factor-Vlll cDNA can repress expression of a CAT-reporter gene in stably transfected cell lines. Polyclonal cultures of Rat2 cells were
stably transfected with plasmids pMuLV-CAT (lane 6). pMuLV-CAT/XS (lane 7). pMuLV-CATISXdXmB (lane 8). or pMuLV-CATISX (lane 9).
Cells of these cultures, which contain approximatelytwo copies of the reporter plasmid
per cell (datanot shown), were harvested, and protein
panel) Repression can be relieved by cotransfection
lysates 110 pg) wereanalyzed for CAT activity as described in Materials and Methods. (Right
of excess amounts of Nlll cDNA. The reporter plasmid (0.2 pg) witheither the SX (lanes 10 and 11) or the SXdXmB fragments (lanes 12 and
13) cloned in the Hpa l site were transfected in the presence of a 40-fold molar excess of either plasmid pUC18 DNA (lanes 10 and 12) or
plasmid pUC-AH DNA, which carries the fvlll cDNA coding for the
A2 domain (lanes 11 and 13). CATactivity was determined in protein
lysates
of the transfectedcells as described in Materials and Methods. Note that the experimental conditions
used in the experiments shown in the
three panels differ and do not permitcomparison of the data.
of the CAT gene, in contrast to the SXdXmB construct (Fig
4, right panel). However, although cotransfection of the
pUC-AH construct does not increase the expression of the
reporter gene with the SXdXmB fragment inserted, the SX
fragment-induced repression was relieved, and the CAT
level was similar to the SXdXmB fragment. These data indicate that the repression is not caused by intrinsic properties
of the DNA (ie, an unfavorable conformation), but that the
repression is mediated by a cellular factor (or factors) that
can be titrated off.
DISCUSSION
This report describes the identification of sequences in the
coding region of human N I 1 1 cDNA that repress expression
at the level of transcription. This repressor downmodulates
the expression of NI11 retroviral vectors, which results in
inferior W111 protein synthesis and reduces the production
of recombinant retroviruses by at least two orders of magnitude. The scarcity of the vector-specific mRNA in the packaging lines can fully explain the low titers of the W111 retroviral vectors. This is in agreement with the observation
that amplification of the N
I11 vector in the packaging cell
lines increases virus production." Furthermore, the amount
of neo' transcripts in the target cells may be too low to allow
the cell to survive selection by G418, lowering the apparent
viral titer. Our observations confirm and extend the description by Lynch et alx of the inhibition of mRNA accumulation
by W111 cDNA sequences [nucleotides (nt) 1681 to 2277,
followed by 5002 to 55821. However, we found the repres-
sion to be dependent on the presence of the Xmrz I-&/ I1
fragment (nt 1375 to 1680), whichislocated
outside the
regionidentified by these investigators. The cause of this
discrepancy is not clear. Whereas the above mentioned study
relied on virus production as a parameter for repression, we
exploited a more direct assay based on the transient expression of a CAT gene. thereby circumventing the problems
with the instability (deletion of NI11 sequences') of the viral
vectors.
Kaufman et all3 studied RNA levels in Chinese hamster
ovary (CHO) cells containing multiple (>loo) W111 cDNA
copies and vWF cDNA expression vectors and observed 40fold lower amounts of Wlll RNA than of vWF RNA when
corrected for expression vector copy number. The transcription rates of both genes were found to be similar in a nuclear
run-on analysis, which suggested that the cause of repression
was posttranscriptional.'3 In contrast, our study suggests the
inhibition to be caused by a transcriptional mechanism. Our
conclusion is based on several lines of evidence. First, the
difference in steady-state amount of vector-specific RNA
observed in cells infectedwith the MSneo vector and its
MSF8dB2.6 derivative reflects the transcription rates as measured in a nuclear run-on assay. Second, the. subgenomic
neo' RNA, which does not contain any N I 1 1 sequences, is
repressed in cells containing the Wlll vector, suggesting a
mechanism that operates before splicing occurs. Third, the
relative stability of the W111 mRNA in the packaging lines
excludes cytoplasmic stability as a mechanism. This is further supported by the observation that insertion of the SX
From www.bloodjournal.org by guest on June 14, 2017. For personal use only.
HOEBEN ET AL
2452
I,
I
A1
A2
1
B
'.*,
I
Fragment
Repression
xs
sx
++
SA
+I-
SSP
HH
AH
PP
A1
A2
-
++
++
HB
BH
AHdXmB
XmB
XmBm
SXdXmB
SXdBBm
++
I
A3
A3
I c1 I
c2
CAT
Cl
I c2 1 Nlll
B-domain
deletion
I
Fig 5. Inhibition of heterologous gene expression by Nlll cDNA fragments. The top lines depict the domain structure of the Nlll protein
and its B domain-deleted derivative. Below this scheme, the various cDNA fragments tested for their capacity t o repress expression of the
linked CAT-reporter gene are indicated. Assays were performed as described in the legend t o Fig 4. All Wlll cDNA fragments are described in
Materials and Methods. Fragments that severely inhibited CAT expression (greater than 10-fold reduction; ++) are represented as solid bars;
the fragment that only moderately
affected CAT activity (twofoldt o threefold reduction; +/-l is represented by a shaded bar; and fragments
that affected CAT activity hardly, if at all, (less than 50% reduction; -) are hatched. All fragments were tested in the same orientation they
had in the retroviral
vector. The inset shows theCAT-reporter plasmid ~ M u L V - C A Tused
~ in this study. The Hpa I site wasused t o insert the
Wlll cDNA fragments; sd and sa, splicedonor and splice-acceptor sites, respectively; PA, polyadenylation signals.
fragment into the transcription unit of the CAT reporter does
not change the stability of the transcript (tin, >2 hours in
Rat2 cells; F.J.F., S.J.C., and R.C.H., unpublished data, February 1994). Also, the observation that treatment of the cells
with sodium butyrate, a known stimulator of transcription:'
increases fVIII RNA levels many times (Fig 1) argues
against a mechanism that operates at the level of RNA stability. Furthermore, the suspected repressor element identified
in this study functions independent of its orientation in a
transcription unit, supporting a mechanism that operates at
the DNA rather than the RNA level. Finally, the AH fragment encoding the A2 domain also markedly represses expression when placed upstream of the MuLV LTR thatdrives
the CAT marker gene (F.J.F.,S.J.C.,andR.C.H.,
unpublished data, February 1994). Collectively, these data suggest
that the repressor sequences act as a transcriptional silencer
of expression.
It is still unclear whether the observed repression is of
any physiologic relevance for the regulation of expression
of the endogenous N
I
1
1 invivo. We cannot exclude the
possibility that sequence elements that are normally separated by introns and, thus, are several kilobases apart are
adjacent in the cDNA and by coincidence gain such a function. The 305-bp Xmn I-BgI I1 fragment is derived from
exons 9 to I 1 of the W111 gene and encompasses over 9 kb
of genomic DNA.
The detailed mechanism by which the repressor acts is
still unclear. The lack of transcription of the AI domain as
observed by the absence of any hybridization signal with
the fVIII-AI probe excludes transcriptional attenuation in
the A2 domain as a mechanism. Such a mechanism has
been found to regulate expression of the c-myc gene.'" The
observation that repression can be relieved by cotransfection
of excess amounts of competitor DNA implies that the repression is not caused by the DNA sequence itself (eg. a
DNA conformation that is unfavorable to transcription), but
rather suggests that cellular components play a role in the
repression. Inspection of the sequence of the 305-bp fragment that is involved in the repression does not immediately
revealany overt consensus sites for binding of candidate
repressors. To date, however, only a few higher eukaryotic
repressor systems have been characterized in detail." In the
305-bp fragment, we did note two sequences with a 1011 1
matchwith the WTTTAYRTTTW consensus observed in
From www.bloodjournal.org by guest on June 14, 2017. For personal use only.
TRANSCRIPTIONSILENCING
2453
BY FACTOR Vlll cDNA
Kogenate Previously Untreated Patient Study Group. N Engl J Med
the autonomously replicating sequences (ARS) of y e a ~ t . ~ ’ . ~ ~
328:453, 1993
The ARS elements allow extrachromosomal maintenance of
4. Kaufman RJ, Wasley LC, Domer A J : Synthesis, processing,
plasmids in yeast and have been implicated as functioning
and secretion of recombinant human factor VIII expressed in mamas chromosomal replication origins and as nuclear-matrix
malian cells. J Biol Chem 263:6352, 1988
attachment regions (MARS),as well as modulating the activ5. Thompson AR, Palmer TD, Lynch CM, Miller AD: Gene transity of transcriptional enhancers and ~ilencers.~’”~
In MARS
fer as an approach to cure patients with hemophilia A or B. Curr
of higher eukaryotes, similar consensus sequences have been
Stud Hematol Blood Transfus 11:59, 1991
found.” Thus, an MAR or ARS in the W111 cDNA could
6. Hoeben RC, Valerio D, Van der Eb A J , Van Ormondt H: Gene
explain the repressing effects of the NI11 cDNA on trantherapy for human inherited disorders: Techniques and status. Crit
Rev Oncol Hematol 13:33, 1992
scription. In addition, the observation that deletion of the
7. Hoeben RC, Van der Jagt RC, Schoute F, van Tilburg NH,
Xmn I-BgZ I1 relieves repression while the fragment itself is
Verbeet MP, Briet E, Van Ormondt H, Van der Eb AJ: Expression
not sufficient to decrease expression fully complies with the
of functional factor VI11 in primary human skin fibroblasts after
known modular structure of MARS as well as ARS eleretrovirus-mediated gene transfer. J Biol Chem 265:7318, 1990
m e n t ~ . ” .Removal
~~
or mutation of some of the ARS ele8. Lynch CM, Israel DI, Kaufman RJ, Miller AD: Sequences in
ments decreases its function, but does not abolish it comthe coding region of clotting factor VI11 act as dominant inhibitors
pletely.32.’4 We have evidence that the fVIII-derived
ofRNA accumulation andprotein production. Hum Gene Ther
WTlTAYRTTIW elements can repress expression of heter4:259, 1993
ologous genes if inserted in the transcription
showing
9. Hoeben RC, Einerhand MPW, Briet E, Van Ormondt H, Vathe potential of these elements in the W111 cDNA to repress
lerio D, Van der Eb AJ: Toward gene therapy in haemophilia A:
Retrovirus-mediated transfer of a factor VI11 gene into murine
expression. If the expression of NI11 is inhibited by the
haematopoietic progenitor cells. Thromb Haemost 67:341, 1992
fortuitous presence of ARS or MARS in the NI11 cDNA,
10. Zatloukal K, Cotten M, Berger M, Schmidt W, Wagner E,
multiple mutations willbe required to completely abolish
Bimstiel ML:In vivo production ofhuman factor VI11 in mice
the repressor. Deletion of the 305-bp Xmn I-BgZ I1 fragment
after intrasplenic implantation of primary fibroblasts transfected by
from the expression vectors is not an option. This region of
receptor-mediated, adenovirus-augmented gene delivery. Proc Natl
the molecule is essential for NI11 function, as eight missense
Acad Sci USA 915184, 1994
mutations have been identified in this region that result in
1 1. Miller AD, Miller DG, Garcia JV, Lynch CM: Use of retrovihemophilia A.38 Currently, we are trying to elucidate the
ral vectors for gene transfer and expression. Methods Enzymol
precise mechanism of repression and to define the sequences
217:581, 1993
involved; eg, by identifying protein-binding elements in the
12. Israel DI, Kaufman RJ: Retroviral-mediated transfer and amplification of a functional human factor VI11 gene. Blood 75:1074,
305-bp Xmn I-BgZ I1fragment. Identification of the responsi1990
ble elements will permit elimination of the binding sites
13. Kaufman RJ, Wasley LC, Davies MV, Wise RT. Israel DI,
by the introduction of silent mutations. Asan alternative
Domer AJ: Effect ofvon Willebrand factor coexpression on the
approach, introns could be introduced to separate and potensynthesis and secretion of factor VI11 in Chinese hamster ovary cells.
tially inactivate various repressor elements. The generation
Mol Cell Biol 9:1233, 1989
of mutated NI11 cDNAs that are expressed more efficiently
14. Dorner AJ,Wasley LC, Kaufman RJ: Protein dissociation
will increase the amount of fVIII protein produced in the
from GRW8 and secretion are blocked by depletion of cellular ATP
cells and will improve the titer of fVIII virus stocks. This,
levels. Proc Natl Acad Sci USA 87:7429, 1990
of course, will not only improve the feasibility of gene ther15. von Ruden T, Wagner EF: Expression of functional human
apy for hemophilia A, but willalso be of considerable imporEGF receptor on murine bone marrow cells. EMBO J 7:2749, 1988
tance for the production of recombinant-DNA-derived NI11
16. Gorman C:High efficiency gene transfer into mammalian
cells, in Glover DM (ed): DNA Cloning: A Practical Approach, v01
for transfusion.
11. Washington, DC, IRL Press, 1985, p 143
ACKNOWLEDGMENT
We are grateful to Dr E. Davie (University of Washington, Seattle)
for the gift of the human flX cDNA, to Drs E. Wojcik and N.H. van
Tilburg for performing the flX and fVIII assays, andtoDrsA.
Lazaris-Karatzas and M. Notebom for many stimulating discussions.
REFERENCES
1. Rosendaal
FR.Smit C, Briet E: Hemophilia treatment in histor-
ical perspective: A review of medical and social developments. Ann
Hematol 62:5, 1991
2. Peake IR, Lillicrap DP, Boulyjenkov V, Briet E, Chan V,
Ginter EK, Kraus EM, Ljung R, Mannucci PM, Nicolaides K, Tuddenham EG: Haemophilia-Strategies for carrier detection and prenatal diagnosis. Bull World Health Organ 71:429, 1993
3. Lusher JM, Arkin S , Abildgaard CF, Schwartz RS: Recombinant factor VI11 for the treatment of previously untreated patients
with hemophilia A. Safety, efficacy, and development of inhibitors.
17. Auffray C, Rougeon F: Purification of mouse immunoglobulin heavy-chain messenger RNAs from total myeloma tumor DNA.
Eur J Biochem 107:303, 1980
18. Fort P, Marty L, Piechaczyk M, El Sabrouty S , Dani C, Jaenteur P, Blanchard JM: Various rat adult tissues express only one
majormRNA species from the glyceraldehyde-3-phosphate-dehydrogenase multigenic family. Nucleic Acids Res 13:1431, 1985
19. Adams JM, Gerondakis S , Webb E, Corcoran LM, Cory S :
Cellular myc oncogene is altered by chromosome translocation to an
immunoglobulin locus in murine plasmacytomas and is rearranged
similarly in human Burkitt lymphomas. Proc NatlAcad Sci USA
80:1982, 1983
20. Hoeben RC, Migchielsen AA, Van der Jagt RC, Van Ormondt
H, Van der Eb AJ: Inactivation of the Moloney murine leukemia
virus long terminal repeat in murine fibroblast cell lines is associated
with methylation and dependent on its chromosomal position. J Virol
65:904, 1991
21. Wood WI, Capon DJ, Simonsen CC, Eaton DL, Gitschier J,
Keyt B, Seeburg PH, Smith DH, Hollingshead P, Wion KL: Expres-
From www.bloodjournal.org by guest on June 14, 2017. For personal use only.
HOEBEN ET AL
2454
sion of active human factor VI11 from recombinant DNA clones.
Nature 312:330,1984
22. Poort SR. van der Linden IK, Krommenhoek-van Es C, Briet
E, Bertina RM: Rabbit polyclonal antibodies against the calciumdependent conformation of factor IX and their application in solid
phase immunoradiometric assays. Thromb Haemost 55:122, 1986
23. Schrier PI, Bernards R, Vaessen RTMJ, Houweling A,Van
der Eb AJ: Expression of class I major histocompatibility antigens
switched off by highly oncogenic adenovirus 12 in transformed rat
cells. Nature 305:771,1983
24. Eaton DL, Wood WI, EatonD, Hass PE, Hollingshead P,
Wion K, Mather J, Lawn RM, Vehar GA, Gorman C: Construction
and characterization of an active factor VI11 variant lacking the
central one-third of the molecule. Biochemistry 25:8343, 1986
25. Pittman DD, Alderman EM, Tomkinson KN, Wang JH, Giles
AR, Kaufman RJ: Biochemical, immunological, and in vivo functional characterization of B-domain-deleted factor VIII. Blood
81:2925,1993
26. Bender MA, Palmer TD, Gelinas RE. Miller AD: Evidence
that the packaging signal of Moloney murine leukemia virus extends
into the gag region. J Virol 61:1639, 1987
27. Armentano D, Yu SF, Kantoff PW, von Ruden T, Anderson
WF, Gilboa E: Effect of internal viral sequences onthe utility of
retroviral vectors. J Virol 61:1647, 1987
28. Dani C, Blanchard JM, Piechaczyk M, El Sabouty S, Marty
L, Jeanteur P: Extreme instability ofmycmRNAin
normal and
transformed human cells. Roc Natl Acad Sci USA 81:7046, 1984
29. Yeivin A, Tang D, Taylor MW: Sodium butyrate selectively
induces transcription of promoters adjacent to the MoMSVviral
enhancer. Gene 1 16:159, 1992
30. Spencer CA, Groudine M: Transcription elongation and eukaryotic gene regulation. Oncogene 5:777, 1990
3 I . Goodbourn S: Negative regulation of transcriptional initiation
in eukaryotes. Biochim Biophys Acta 1032:53, 1990
32. Van Houten JV, Newlon CS: Mutational analysis of the consensus sequence of a replication origin from yeast chromosome 111.
Mol Cell Biol l0:3917, 1990
33. NewlonCS, Theis JF: The structure and function of yeast
ARS elements. Curr Opin Genet Dev 3:752, 1993
34. Gasser SM, Amati BB, Cardenas ME, Hofmann JF: Studies
on scaffold attachment sites and their relation to genome function.
IntRev Cytol 11957, 1989
35. Cockerill PN, Garrard WT: Chromosomal loop anchorage of
the kappa immunoglobulin gene occurs next to the enhancer in a
region containing topoisomerase I1 sites. Cell 44273, 1986
36. Fishel BR, S p e w AO, Garrard WT: Yeast calmodulin and a
conserved nuclear protein participate in the in vivo binding of a
matrix association region. Proc Natl Acad Sci USA 90:5623, 1993
37. Boulikas T: Chromatin and nuclear matrixin development
andin carcinogenesis: A theory. Int J Oncol 1:357, 1992
37a. Fallaux FJ, Hoeben RC, Cramer SJ, Van den Wollenberg
DJM,BrietE,
Van Ormondt H,Van der Eb AJ: Repression of
transcription by coding sequences of blood clotting factor VI11
cDNA can be relieved by short chainfatty acids. MolCellBiol
(submitted)
38. Tuddenham EG, Cooper DN, Gitschier J, Higuchi M, Hoyer
LW, Yoshioka A,Peake IR, Schwaab R, Olek K,KazazianHH,
Lavergne J-M, Giannelli F, Antonarakis SE: Haemophilia A: Database of nucleotide substitutions, deletions, insertions and
rearrangements of the factor VI11 gene. Nucleic Acids Res 19:4821,
1991
From www.bloodjournal.org by guest on June 14, 2017. For personal use only.
1995 85: 2447-2454
Expression of the blood-clotting factor-VIII cDNA is repressed by a
transcriptional silencer located in its coding region
RC Hoeben, FJ Fallaux, SJ Cramer, DJ van den Wollenberg, H van Ormondt, E Briet and AJ van
der Eb
Updated information and services can be found at:
http://www.bloodjournal.org/content/85/9/2447.full.html
Articles on similar topics can be found in the following Blood collections
Information about reproducing this article in parts or in its entirety may be found online at:
http://www.bloodjournal.org/site/misc/rights.xhtml#repub_requests
Information about ordering reprints may be found online at:
http://www.bloodjournal.org/site/misc/rights.xhtml#reprints
Information about subscriptions and ASH membership may be found online at:
http://www.bloodjournal.org/site/subscriptions/index.xhtml
Blood (print ISSN 0006-4971, online ISSN 1528-0020), is published weekly by the American
Society of Hematology, 2021 L St, NW, Suite 900, Washington DC 20036.
Copyright 2011 by The American Society of Hematology; all rights reserved.