| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Experimental Therapeutics |
Calydon, Inc., Sunnyvale, California 94089
 |
ABSTRACT |
|---|
 Top ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
-Fetoprotein (AFP), a tumor marker currently used for the
diagnosis and management of HCC, is an oncofetal protein expressed in a
majority of HCCs but rarely in normal hepatocytes. Because AFP
gene expression is tightly regulated at the level of
transcription, AFP transcriptional regulatory elements (TRE) are
excellent candidates for generating HCC-specific oncolytic
adenoviruses. We devised a new strategy for the AFP TRE to control an
artificial E1A-IRES-E1B bicistronic cassette in an adenovirus 5 vector
(Ad5) and constructed an HCC-specific oncolytic virus, CV890. In
vitro, CV890 expression of the E1A and
E1B genes, virus replication, and cytopathic effects
were examined by Northern blot, Western blot, virus yield assay, and
3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay in
AFP-producing cell lines (HepG2, Huh7, Hep3B, PLC/PRF/5, and SNU449),
non-AFP-producing cell lines (Sk-Hep-1, Chang liver cell, LNCaP,
HBL-100, PA-1, UM-UC-3, SW 780, Colo 201, and U118 MG), and
non-AFP-producing human primary cells (lung fibroblast, bladder smooth
muscle, and mammary epithelial). CV890 efficiently replicates in and
destroys AFP-producing HCC cells as well as wild-type Ad5, but
replication is highly attenuated in non-AFP-producing HCC cells or
non-HCC cells. CV890 produced 5,000–100,000-fold less virus than
wild-type Ad5 in non-AFP-producing cells. CV890 was attenuated 100-fold
more than CV732, a virus containing the AFP TRE driving the
E1A gene alone, in non-AFP-producing cells. These
studies demonstrated that expression of both E1A and
E1B genes under the control of a bicistronic
AFP-E1A-IRES-E1B cassette yielded improvements in virus specificity
equivalent to driving the E1A and E1B
genes with two independent TREs yet requires only one TRE thereby
conserving genomic space within the virus. Significantly, CV890
produced nearly the same yield of virus in cells that produced AFP over
a 75-fold range, from a low of 60 ng AFP/106 cells/10 days
to as high as 4585 ng AFP/106 cells/10 days. In
vivo, antitumor efficacy of CV890 was examined in
BALB/c-nu/nu mice containing large s.c. HepG2 or Hep3B
tumor xenografts. Tumor volume of distant xenografts dropped below
baseline 4 weeks after a single i.v. injection. Combination of CV890
with doxorubicin demonstrated synergistic antitumor efficacy, yielding
complete elimination of distant Hep3B tumors 4 weeks after a single
i.v. administration of both compounds. Our results support the clinical
development of CV890 as an antineoplastic agent for the treatment of
localized or metastatic HCC. |
INTRODUCTION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
To target the oncolytic adenoviruses to specific tumors, we and others
have used tumor- or tissue-specific promoters to restrict the virus
replication to target tumor cells (8, 9, 10
, 15)
. In the
first 2.5 h after adenovirus infection, the E1A gene is
the only viral gene transcribed. The E1A proteins transactivate both
viral and cellular genes critical for productive viral infection
including E1B, E2, and E4 (16)
. Adenovirus E1B
genes are also expressed relatively early in viral infection. The E1B
proteins inhibit E1A-induced p53-dependent apoptosis, aid in the
cytoplasmic accumulation of late viral mRNA, and promote shutoff of
host cell protein synthesis (17)
. Previous studies have
shown that E1A expression controlled by a specific TRE can initiate the
selective replication of the recombinant adenoviruses in target cells
while restricting replication in nontarget cells (8, 9, 10
, 15)
. However, the leakiness of foreign TREs in E1A control
yielding low levels of E1A may result in loss of specificity (18
, 19)
. To control the viral replication more stringently,
we explored previously use of a second transcriptional control to
control E1B (9
, 10)
. We demonstrated that separated TRE
control of both the E1A and E1B genes
significantly improved the specificity and the in vitro
therapeutic index of CV764 and CV787, two prostate cancer-specific
adenovirus variants (9
, 10)
. However, the two TRE
sequences must contain 
20% heterologous sequences to prevent
homologous recombination that deletes the intervening E1A sequence
during replication (9
, 10)
. Primary liver cancer currently
has only a single well-characterized TRE that demonstrates tumor cell
specificity. Therefore, we devised an alternate strategy to retain the
high level of specificity for target cells obtained by controlling
expression of the E1A and E1B genes with a single
TRE. In this communication, we demonstrated that an AFP-E1A-IRES-E1B
bicistronic expression cassette fulfilled the necessary requirements
and created an AFP-producing hepatoma-specific adenovirus variant,
CV890, for additional clinical development.
Of concern clinically is whether oncolytic viruses replicate in and destroy tumor cells expressing varying levels and, in particular, low levels of the tumor marker. This question is difficult to ask of the prostate-specific viruses CV706 and CV787 in the prostate cancer system because of the lack of different prostate cancer cells lines expressing varying levels of PSA. This is not the case in the liver cancer system, because many cell lines expressing AFP are available. We show that CV890 replication in AFP+ cells is nearly independent of AFP production levels between 60 ng and 4500 ng AFP/106 cells/10 days, a 75-fold range. Also clinically, once tumor specificity is attained, the focus shifts to efficacy. With CV890 we extend our previous observations that reintroduction of the adenovirus E3 region into these constructs and synergy with chemotherapy both independently increase the efficacy of oncolytic adenoviruses (9 , 20) .
|
MATERIALS AND METHODS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Plasmid Construction.
A composite AFP TRE that includes an AFP enhancer and promoter element
was assembled by overlap PCR and cloned into adenovirus. The enhancer
element (-3756 bp--3167 bp) was amplified from human genomic DNA
(Clontech, Palo Alto, California) with primers 39.055B
(5'-gtgaccggtgcattgctgtgaactcacgt) and 39.055D
(5'-ataagtggcctggataaagctgatgca). The promoter (-123 bp-+14 bp) was
amplified with primers 39.055J (5'-gtcaccggtctttgttattggcagaatt) and
39.055M (5'-atccaggccacttatgagctctgtgtggaa). The two products were
annealed and used as the template for PCR with primers 39.055B and
39.055J. A 0.8 kb PinAI (isoschizomer of AgeI)
fragment from this overlap product was used as the AFP TRE.
To control both the Ad5 E1A and E1B expression from a single TRE, an IRES from EMCV was introduced into the junction of these two genes. A 0.5 kb EMCV IRES fragment was placed into the E1A-E1B modified intergenic region to form the bicistronic E1A-IRES-E1B cassette. The final plasmid, CP686, was generated by placing a 0.8 kb PinAI fragment that contains the AFP TRE in front of the bicistronic E1A-IRES-E1B cassette (detail structure will be reported elsewhere). Another plasmid, CP219, containing a single AFP-control E1A expression was derived from CP124 (10) by placing the PinAI fragment containing the AFP TRE to the PinAI site in front of the E1A coding region. All of the DNA manipulations were performed according to established protocols (9 , 10) .
Virus Construction.
CV890 was constructed by cotransfection of the 293 low-passage cells
with CP686 and pBHGE3 (9
, 10)
. Briefly, CP686 and pBHGE3
DNA were cut by appropriate endonuclease digestion and transfected
using Lipofectin (Life Technologies, Inc., Grand Island, New York) in a
4:1 molar ratio of lipid to DNA. Transfections were incubated for 7–14
days, harvested, freeze-thawed, and plaqued on the 293 cells. Potential
recombinant viruses were picked and expanded by infecting the 293
cells. To screen for potential recombinants, viral DNA was purified
from crude plate lysates using the QIAamp Blood kit as suggested by the
manufacturer (Qiagen, Santa Clarita, CA). PCR was performed using
several primer pairs specific for the adenovirus E1 sequence that
spanned the desired AFP inserts, and the PCR products were digested
with restriction endonucleases to confirm the structure.
CV840, derived from CP686 and pBHG10 transfection, and CV732, derived from CP219 and pBHG10 transfection, were also generated in this study for the comparison with CV890. CV732 was designed to be substantially equivalent to AvE1ao4i containing a single AFP driving E1A in an E3- adenovirus as described previously (15) . CV730, an E1A-deleted adenovirus mutant;3 CV702, an E3 mutant; and CV802, a wt virus, were used as control in this study. All of the viruses were plaque purified, and structures were verified before additional characterization.
Virus Yield and One-Step Growth Curves.
Six-well dishes (Falcon) were seeded with 5 x 105 cells/well for 24 h before infection.
Cells were infected at a MOI of 2 pfu/cell for 3 h in serum-free
medium. After 3 h, the virus-containing medium was removed,
monolayers were washed three times with PBS, and 4 ml of complete
medium (RPMI 1640 with 10% fetal bovine serum) was added to each well.
After infection (72 h), cells were scraped into the culture medium and
lysed by three cycles of freeze-thaw. The one-step growth-curve time
points were the time when samples were harvested after infection. Two
independent infections of each virus-cell combination were titered in
duplicate on 293 cells (10)
.
Northern Blot Analysis.
Hep3B or HBL100 cells were infected at an MOI of 20 pfu/cell with
either CV802 or CV890 and harvested 24 h after infection. Total
cellular RNA was purified using the RNeasy protocol (Qiagen). The
Northern blot was conducted using NorthernMax Plus reagents (Ambion,
Austin, TX). RNA (5 µg) was fractionated on a 1% agarose,
formaldehyde-based denaturing gel and transferred to a BrightStar-Plus
(Ambion) positively charged membrane by capillary transfer. The
antisense RNA probes for E1A (adenovirus genome 501 bp-1141 bp) or E1B
(1540 bp-3910 bp) were PCR products cloned in pGEM-T easy (Promega) and
transcription labeled with [32P] UTP. Blots
were hybridized at 68°C for 14 h with ZipHyb solution and washed
using standard methods (Ambion). Membranes were exposed to BioMax
film (Kodak).
Western Blot Analysis.
Hep3B or HBL100 cells were infected at MOI of 20 pfu/cell with either
CV802 or CV890 and harvested 24 h after infection. Cells were
washed with cold PBS and lysed for 30 min on ice in a NP40 equivalent
[50 mM Tris (pH 8.0), 150 mM NaCl, and 1%
IGEPAL CA360] (Sigma Chemical Co., St. Louis, MO), 0.5% sodium
deoxycholate, and protease inhibitor mixture (Roche, Palo Alto, CA).
After 30 min of centrifugation at 4°C, the supernatant was harvested
and the protein concentration determined with the protein assay
ESL kit (Roche). Fifty µg of protein/lane were separated on
8–16% SDS-PAGE and electroblotted onto Hybond ECL membrane
(Amersham Pharmacia, Piscataway, New Jersey). The membrane was
blocked overnight in PBS with 0.1% Tween 20 supplemented with 5%
nonfat dry milk. Primary antibody incubation was performed at room
temperature for 2–3 h with PBS with 0.1% Tween 20/1% milk diluted
antibody followed by wash and 1-h incubation with diluted horseradish
peroxidase-conjugated secondary antibody (Santa Cruz Biotechnology
Inc., Santa Cruz, CA). ECL (Amersham Pharmacia) was used for signal
detection. E1A antibody (clone M58) was from NeoMarkers (Fremont, CA),
E1B-21kD antibody was from Oncogene (Cambridge, MA). All of the
antibodies were used according the manufacturer’s instruction.
Cell Viability Assay and Statistical Analysis.
To determine the cell killing effect of CV890 and doxorubicin in
combination treatment, a cell viability assay was conducted as
described previously with modifications (21)
. On 96-well
plates, cells of interest were seeded at 10,000 cells/well 48 h
before infection. Cells were treated with virus alone, drug alone, or
in combination. Cell viability was measured at different time points by
removing the medium, adding 50 µl of 1mg/ml solution of MTT (Sigma
Chemical Co.), and incubating for 3 h at 37°C. After removing
the MTT solution, the crystals remaining in the wells were solubilized
by the addition of 50 µl of isopropanol followed by 30°C incubation
for 30 min. The absorbency was determined on a microplate reader
(Molecular Dynamics) at 560 nm (test wavelength) and 690 nm (reference
wavelength). The percentage of surviving cells was estimated by
dividing the OD550 - OD650 of virus- or drug-treated cells by the
OD550 - OD650 of
control cells. Six replica samples were taken for each time point, and
each experiment was repeated at least three times.
For statistical analysis of synergy, CurveExpert (shareware by Daniel Hyams, version 1.34) was used to plot the dose-response curves for virus and drugs. Based upon the dose-response curves, the isobolograms were made according to the original theory of Steel and Peckham (22) and as modified by Aoe et al. (23) .
Animal Studies.
Six to eight-week-old athymic BALB/C nu/nu mice were
obtained from Simonson Laboratories (Gilroy, CA) and acclimated to
laboratory conditions 1 week before tumor implantation. Xenografts were
established by injecting s.c. 1 x 106 Hep3B, HepG2, or LNCaP cells suspended in 100
µl of RPMI 1640. When tumors reached between 200 and 300
mm3
, mice were randomized and dosed with 100 µl
of test article via intratumoral or tail vein injection. Tumors were
measured in two dimensions by external caliper, and volume was
estimated by the formula [length (mm) x width
(mm)2
]/2. Animals were humanely killed when
their tumor burden became excessive. Serum was harvested weekly by
retro-orbital bleed. The level of AFP in the serum was determined by
AFP Immunoassay kit (Genzyme Diagnostics, San Carlos, CA). The
difference in mean tumor volume and mean serum AFP concentration
between treatment groups was compared for statistical significance
using the unpaired, two-tailed t test. Immunohistochemical
analysis of tumors for adenovirus replication was conducted by
following a procedure described previously (9)
.
|
RESULTS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
To control both E1A and E1B genes with single
AFP, an EMCV IRES structure was applied to build an artificial
E1A-IRES-E1B bicistronic expression cassette. EMCV IRES has been well
characterized and applied for mediating secondary gene expression
(24)
. This artificial E1A-IRES-E1B bicistronic cassette
enables both E1A and E1B to form a single unit for transcription,
whereas the IRES structure gives the internal translation initiation
for E1B translation. In the current study, E1A-IRES-E1B bicistronic
cassette was placed under the transcriptional control of the AFP TRE to
target the new ARCAs to AFP-producing cells. Fig. 1
illustrates the schematic structure of resultant viruses CV890 and
CV840 that have intact E3 or deleted E3, respectively. CV702 and CV802
have wt E1 region and are used as wt Ad5 controls with or without E3
deletion, respectively (10)
. Another ARCA, CV732,
containing single AFP to control only E1A expression, was used to
compare with CV890 and CV840. CV787, a prostate cancer-specific
adenovirus variant (9)
, was used as a negative control for
in vivo liver cancer xenograft animal studies. CV730 was an
E1A-deleted mutant and was also used as a control in the current study.
|
shows the virus yield assay controlled by wt CV702 and
CV802 in a panel of different tumor cell lines including
AFP+ cells (HepG2 and Hep3B; see below for AFP
status) and AFP- cells (Chang liver cell,
UM-UC-3, SW 780, CoLo 201, and U118 MG). Our results showed that all of
the CV890, CV840, and CV732 have clear cell selective patterns. Whereas
wt viruses CV702 and CV802 replicate well in all of the cell lines, the
AFP viruses CV732, CV840, and CV890 are attenuated by 100- 100,000-fold
when compared with wt controls in AFP- cells.
From the parallel comparison of three variants, both CV890 and CV840
have a better replication specificity than CV732. In
AFP- cells, CV890 showed 100,000-fold
attenuation, whereas CV732 only showed 100-fold attenuation. In
addition, CV890 and CV840 showed a 5-fold increase in replication
efficiency in AFP+ cells compared with CV732.
Finally, in all of the AFP- cells, although the
initial infection was at MOI of 2, the burst size for both CV890 and
CV840 was 1 pfu/cell, suggesting that the virus replicates poorly. For
CV732, however, there is definitely a clear replication in the
AFP- cells with an average burst size of 100
pfu/cell. This increased virus replication or leakiness is attributable
to reduced transcriptional control.
|
indicated that CV802 was toxic to Hep3B cells and Chang liver cells
regardless of AFP status, whereas CV890 and CV732 were only toxic to
AFP-producing Hep3B cells but not to AFP- Chang
liver cells. Interestingly, when both CV890 and CV732 were toxic to
AFP-producing Hep3B cells, CV890 had a superior cytotoxicity. For
example, there were no surviving CV890-infected Hep3B cells 9 days
after infection, whereas the viability of CV732-infected Hep3B cells
was 50%. The in vitro comparison in Fig. 2
suggests that
CV890 has a better replication specificity and killing ability to the
target cells than CV732. Therefore, the strategy of using single AFP to
control the E1A-IRES-E1B bicistronic cassette is better than the
strategy of controlling E1A alone. On the basis of the comparison in
different AFP-controlled variants, we focused additional preclinical
work on CV890.
Transcription and Translation of the E1A-IRES-E1B
Bicistronic Cassette of CV890 in Different Cells.
In wt adenovirus infection, E1A and E1B genes
produce a family of alternatively spliced products. It has been found
that there are five E1A mRNAs, but among them the 12S (880 nts) and 13S
(1018 nts) mRNAs are the dominant ones at both early and late times
after infection. The 12S and 13S mRNAs encode the gene products of 243
amino acids (243R) and 289 amino acids (289R) respectively
(25)
. The two major E1B transcripts are 12S (1031 nts) and
22S (2287 nts) mRNAs that encode the
Mr 19,000 and
Mr 55,000 E1B proteins respectively.
E1B 12S mRNA only encodes the Mr
19,000 product, whereas the 22S mRNA encodes both the
Mr 19,000 and
Mr 55,000 products because of
different initiation sites during translation. In the current study,
the generation of E1A-IRES-E1B bicistronic cassette was expected to
change the pattern of E1A and E1B transcripts in viral infection.
Therefore, Northern blot analysis was conducted to evaluate the
steady-state level of E1A and E1B transcripts. First, CV802 or CV890
was infected to Hep3B (AFP+) or HBL100
(AFP-) cells for 24 h. The total RNA
samples were separated on agarose gels and processed for Northern blot
by hybridizing to antisense RNA probes. The Northern blot with E1A
probe visualized the 12S and 13S mRNAs in both Hep3B and HBL100 cells
infected with CV802 (Fig. 3A)
. For CV890, E1A transcripts can only be seen in Hep3B
cells but not in HBL100 cells, indicating the conditional transcription
of E1A. It is of interest to find that in CV890, there is only one
large transcript (about 3.5 kb marked by an arrow in Fig. 3A
), whereas the 12S and 13S mRNAs are no longer present.
This large transcript indicates the continuous transcription of
E1A-IRES-E1B bicistronic cassette, suggesting an alteration of viral
E1A splicing pattern in CV890. In the bottom panel of Fig. 3
A, transcription of E1B from CV890 is also AFP-dependent.
It is clear that both 12S and 22S mRNAs of E1B are present in wt CV802
samples, whereas the 12S mRNA and an enlarged 22S mRNA (3.5 kb) are
present in CV890-infected Hep3B cells. The identity of this enlarged
mRNA is the same 3.5-kb transcript as visualized in the E1A blot, which
is from the transcription of E1A-IRES-E1B bicistronic cassette.
Therefore, the E1B mRNA is tagged after E1A mRNA in this large
transcript. This large transcript contains all of the coding
information for E1A and E1B, Mr
19,000, and E1B, Mr 55,000. The mRNA
splice pattern shown in CV802-infected Hep3B cells is not shown in
CV890-infected Hep3B cells for the 12S mRNA where the E1B probe is
missing. Meanwhile, in the E1B Northern blot, because of the selection
of our E1B probe (1540 bp-3910 bp) and mRNA of the adenovirus gene IX
(3580 bp-4070 bp), the hexon-associated protein was also detected. In
CV890-infected Hep3B cells, gene IX expression is equivalent to that of
CV802, whereas in CV890-infected HBL100, its expression was also
completely shut down. This result also demonstrated that the
AFP-controlled E1A-IRES-E1B expression is the key for late gene
expression as well as viral replication.
|
.
Under our experimental condition, E1A expression level of CV890 in
Hep3B cells is similar to that of CV802 (Fig. 3B)
. However,
when E1B Mr 19,000 protein was
detected, it was found that the expression level was much lower than
CV802 E1A (Fig. 3B)
. Previously, it has been addressed that
IRES-mediated second gene has less expression (26)
. Taken
together, CV890 infection in permissive Hep3B cells can produce normal
amount of E1A and lesser amounts of E1B proteins capable of initiating
a normal productive infection. In AFP- cells,
however, this process was attenuated because of a lack of E1A and E1B
gene transcription and translation. These data demonstrated that the
expression of both E1A and E1B genes are under
the control of AFP TRE, and the artificial E1A-IRES-E1B bicistronic
cassette is functioning properly in CV890.
In Vitro Replication Specificity of CV890 in Tumor
Cells and Primary Cells.
From in vitro comparison of virus yield, CV890 has a better
specificity profile than CV732 (Fig. 2)
. To gain additional insights
about using CV890 in liver cancer therapy, more tumor cell lines and
primary cells were tested to characterize in vitro virus
replication. First, all of the cells in the study were analyzed for
their AFP status by AFP immune assay. On the basis of AFP produced in
the cells and medium, all of the cells were divided into three groups:
high (>2500 ng/106 cells/10 days), low (<600
ng/106 cells/10 days), and none (undetectable in
our study; Table 1
). It was confirmed that replication of CV890 in different cell lines
correlates well with the AFP status of the host cell. In Fig. 4
, among the group of liver cell lines, CV890 only replicates well in
AFP+ cells including Hep3B, HepG2, Huh7, SNU449,
and PLC/PRF/5. The amount of AFP required for the promoter activity
seems very low, because one of the hepatoma cell line, SNU449, a
previously reported AFP- cell (27)
,
produces very low AFP (
60 ng/106 cells/10
days) compared with other cells. Nevertheless, even with very low
amounts of AFP, SNU449 cells can still support CV890 replication
comparable or only slightly less than cells producing significantly
higher levels of AFP such as HepG2. Compared with CV802, CV890 is
attenuated 5,000–100,000-fold in cells that do not produce any AFP,
including the hepatoma cell Sk-Hep1, the Chang liver cell, other tumor
cells, and primary cells (Fig. 4)
. Taking the results from Figs. 2
and 4
together, CV890 has shown a good specificity profile from a broad
spectrum of tumor cells. Among them, only the
AFP+ liver cells, regardless of AFP production
level, are permissive for CV890.
|
|
additionally demonstrated that CV890 has similar growth kinetics to wt
CV802 in AFP+ cells except that virus yields are
slightly lower (2–8-fold) in low AFP-producing cells. In consideration
of experimental error, there is no significant difference in the
replication of CV890 and CV802 in AFP+ hepatoma
cells. However, the growth curves of CV890 in
AFP- cells showed clear attenuation. During a
5-day experiment, CV890 failed to replicate in
AFP- cells including liver cell (Chang liver),
ovarian cancer cell (PA-1), and primary cells (normal human lung
fibroblast cells; Fig. 5
). From all of the in vitro virus
replication studies, it is clear that replication of CV890 is under the
tight control of AFP TRE, and this adenovirus variant has an excellent
specificity profile of preferentially targeting AFP-producing HCC
cells.
|
1200
mm3
in 6 weeks, similar to the group treated with
vehicle (data not shown). This study indicates that CV890 does not
attack the LNCaP xenograft and maintains the desired specificity
profile under in vivo conditions.
To evaluate in vivo antitumor efficacy of CV732
(E3-) and CV890 (E3+),
different studies were carried out in the nude mouse model harboring
human hepatoma xenografts. We have shown previously that the addition
of the E3 region of adenovirus significantly increased the antitumor
activity of the prostate-specific adenovirus CV787 using
PSA+ LNCaP mouse xenografts (9)
.
BALB/c-nu/nu mice with Hep3B xenografts were established.
Tumor-bearing animals were challenged with single or multiple doses of
CV730, CV732, and CV890 via tail vein i.v. administration. Tumor volume
and serum levels of AFP were monitored weekly. Animals harboring 300
mm3
Hep3B xenografts were grouped
(n = 7) and injected with vehicle alone
(control group), CV890 (1 x 1011
particles/dose), CV730 (1 x 1011
particles/dose), or CV732 (1 x 1011 particles/dose). In Fig. 6
, the Hep3B xenograft was shown to be very aggressive, and tumors grew
10-fold in 5 weeks. No difference in tumor volume was observed between
vehicle group and CV730 group (data not shown). In contrast, a single
i.v. administration of CV890 showed significant tumor growth
inhibition. In the CV732 group, single-dose i.v. injection also reduced
the tumor growth as compared with the control group; however, CV732 was
much less effective compared with CV890. For example, the average tumor
volume of the CV890-treated group remained unchanged or slightly
decreased in the course of the 6-week study, whereas tumors treated
with CV732 increased from 274 mm3
(relative tumor
volume of 100%) to 1038 mm3
(relative tumor
volume 544%). Both the control group and CV732 group were terminated
at week 6 because of excessive tumor size. Previously, CV732 has been
demonstrated that it could restrict the hepatoma tumor from growth
after 5 doses of i.v. administration (data not shown). Similar efficacy
can be achieved with just a single i.v. administration of CV890,
indicating that under in vivo condition, CV890 has better
efficacy than CV732 to hepatoma xenografts. In this experiment, it was
of interest that three of seven CV890 treated mice were tumor-free 3
weeks after treatment. However, in the CV732 group, xenografts in two
mice stopped growing, but none of treated animals were tumor-free
through the 6-week experiment. There was no tumor reduction in this
group or the control group of animals. By statistical analysis, the
differences in mean relative tumor volumes and serum AFP concentrations
between CV890-treated and CV732-treated or vehicle-treated tumors are
significant (P < 0.01; results not shown).
Taken together, these studies show that CV890 has a significant
antitumor activity, and its oncolytic efficacy is significantly better
than CV732, an adenovirus variant similar to AvE1a04 I, in which a
single AFP TRE was used to drive the E1A gene
(15)
. As described previously and shown clearly in Fig. 1A
comparing CV840 and CV890, the increased antitumor
activity of CV890 compared with CV732 and AvE1a04 i can be attributed
to the addition of the E3 region of adenovirus (9)
.
|
Synergistic Antitumor Efficacy of CV890 in Combination with
Chemotherapeutic Agents.
Previously, we have demonstrated in vitro and in
vivo that the combination of CV787 with paclitaxel or docetaxel
leads to synergistic antitumor (20)
. In the current study,
different chemotherapeutic agents were tested in combination with CV890
for their in vitro cell killing effect in Hep3B cells. We
optimized drug concentration to a certain extent that it would not
generate extensive cytotoxic effect by itself. Under such conditions,
some agents had shown higher cell killing effect in combination with
CV890. Among them, doxorubicin, a drug currently used in the treatment
of HCC (6)
, showed synergistic cytotoxicity with CV890.
Fig. 7A
is an example of using doxorubicin together with CV890 on
Hep3B cells. Doxorubicin at 10 ng/ml did not generate cytotoxicity,
whereas CV890 at an MOI of 0.01 (pfu/cell) destroyed 35% of the cells
by day 9. However, when both doxorubicin and CV890 were applied
together, 90% of the cells were killed 9 days after treatment. To
determine the potential synergistic effect from the combination
treatment, the MTT cell viability data were subjected to additional
statistical analysis. Fig. 7B
shows a representative 50%
cell growth inhibition isobologram of doxorubicin and CV890 on Hep3B
cells at day 5. First, the dose-response curves of doxorubicin alone or
CV890 alone were made (result not shown). On the basis of the original
theory of Steel and Peckham (22)
and method by Aoe
et al. (23)
, three isoeffect curves (modes I,
IIa, and IIb) were constructed. A combination that gives data points to
the left of the envelope of mode IIb line can be described as
supra-additive (synergism; Ref. 20
). From this
isobologram, several data points were in the synergy area, indicating
that the combination of CV890 and doxorubicin provides synergistic
effects on the killing of Hep3B cells.
|
shows weekly changes in the relative tumor size normalized to 100% at
day 1. In this experiment, by week 6, animals in the control group
experienced tumor growth of 700% of baseline, whereas in the CV890
group and the combination group, animals experienced tumor reduction or
were tumor-free. Of the eight Hep3B xenografts treated with CV890
alone, three animals (37.5%) had no palpable tumor at week 5, and
another three animals experienced tumor regression of >60%. In the
combination group, four of eight animals were tumor-free from week 5,
and the other four animals experienced a tumor reduction of 95%. All
of the animals in the CV890 and combination groups were alive, and the
tumor was suppressed even 10 weeks after treatment, whereas the control
animals were sacrificed for excessive tumor burden after week 6 (result
not shown). CV890 also caused a drop in the serum AFP concentration in
these mice (data not shown). Statistical analysis shows that
differences in mean relative tumor volumes and serum AFP concentrations
between CV890 and vehicle-treated group or combination and
doxorubicin-treated group are significant at different times
(P < 0.005; results not shown). The strong
efficacy in the combination treatment shows that a single i.v.
injection of CV890 in combination with doxorubicin can eradicate
aggressive Hep3B xenografts in most animals.
|
|
DISCUSSION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
In the ARCA technology, the TREs placed to control E1A or E1B are the
switches to turn on adenovirus replication. After adenovirus
penetration, only the cells that provide the requisite transcription
factors to turn on the TRE switches will support productive viral
replication. In the current study, this rational was additionally
verified in CV890, an adenovirus variant created with a new strategy.
To control both E1A and E1B genes with single AFP
TRE, we constructed an E1A-IRES-E1B bicistronic cassette by replacing
the E1A-E1B junction sequence with the ECMV IRES fragment. This
bicistronic cassette was placed under the control of an AFP TRE
generating CV890. Transcription of both E1A and
E1B genes of CV890 was found to be under the tight control
of AFP TRE and translation of the E1B mediated by EMCV-IRES, as
designed. Transcription of the E1A-IRES-E1B bicistronic cassette forms
an enlarged 22S mRNA that has all of the coding genes for E1A and E1B.
When this cassette was driven by the AFP TRE, E1 transcripts were only
found in AFP+ cells (Fig. 3A)
. E1A and
E1B proteins were detected only in the CV890-infected Hep3B cells
(AFP+) but not in AFP-
HBL-100 cells (Fig. 3B)
. These data demonstrated that both
the E1A-IRES-E1B transcription and translation in CV890 are
AFP-dependent. This is the first effort to alter the adenovirus E1
transcription pattern to control both E1A and E1B expression by a
single tumor-specific TRE.
We also found the resulting virus, CV890, has a better specificity and efficacy profile than CV732, a replicating adenovirus containing a single AFP TRE driving E1A in an E3- virus background. CV732 was designed to be functionally identical to AvE1a04 I as described previously (15) .
For specificity, we compared CV890 with CV732 in vitro in
various cell lines. Preliminary studies demonstrated that the
E1A gene of CV732 is tightly controlled by the AFP TRE, and
the E1B gene is driven by the endogenous adenovirus E1B
promoter (data not shown). Burst size data showed that CV732 produced
100–1,000 fewer infectious virus particles in
AFP- cells as compared with
AFP+ cells. Burst size data also showed that
CV890 actually had a higher replication efficiency (2–5 times higher)
in AFP+ cells than CV732. In contrast, in
AFP- cells, CV890 gave additional 100-fold
attenuation when compared with CV732 (Fig. 2A)
. The overall
replication difference between AFP+ and
AFP- cells were 100-1000 fold for CV732 and
5,000–100,000-fold for CV890, demonstrating the marked improvement in
replication selectivity of CV890 over CV732. Thus, the AFP TRE
controlling the E1A-IRES-E1B bicistronic cassette clearly leads to a
better restriction of virus replication to AFP-producing cells than a
single AFP TRE driving E1A alone. For cytotoxicity comparison between
CV890 and CV732, we conducted virus-killing kinetics on Hep3B
(AFP+) and the control Chang liver cells
(AFP-). The MTT assay demonstrated that CV890
has a significantly better cytotoxicity on the target Hep3B cells than
CV732 (Fig. 2B)
.
To evaluate cancer drugs, therapeutic index is a ratio of the doses at
which therapeutic effect and toxicity occur. Both virus burst size
assay and growth curve assay data indicated that, at a given dose of
CV890, target AFP+ cells can produce
5,000–100,000 more viruses than nontarget AFP-
cells, demonstrating an in vitro therapeutic index of
5,000–100,000 to 1. The high specificity of CV890 likes is not
achievable by other types therapy. For example, chemotherapeutics, such
as alkylating agents and DNA synthesis inhibitors, target both
malignant and normal cells, resulting in therapeutic indices of 2:1 to
6:1 (29)
. However, as shown in Table 1
and Fig. 4
, even
low levels of expression of AFP (60 ng/106
cells/10 days) can lead to activation of the AFP TRE within CV890 and
support replication of adenovirus. Indeed, a 75-fold difference in AFP
levels in different cells only slightly reduced virus yield. This point
has significant implications in restricting the TREs suitable for the
design of replicating oncolytic adenoviruses to treat cancer. Namely,
proposed tumor markers that are "enriched" in or
"overexpressed" in tumor cells may not offer a TRE with
sufficiently tight regulation to yield the desired specificity. Rather,
TREs that are expressed "only" in the target tumor cells or in
medically unimportant tissues from a clinical side-effect profile may
be suitable.
CV890 also showed much better antitumor efficacy than CV732 in the
human liver cancer Hep3B mouse xenograft model (Fig. 6)
. Tumor
reduction was achieved in all of the animals after receiving a single
i.v. administration of CV890, whereas CV732 could only slow down or
stop the tumor growth in a low percentage of animals, but had no tumor
reduction during the entire course of the experiment. In most animals
that received a single dose of CV732 via tail vein i.v. administration,
tumors kept growing and caused animal death from excessive tumor burden
after 6 weeks. Similar animal results have been reported in an AvE1a04
i study that showed 20–50% of animal loss after 7–13 weeks
(15)
. In another study, AFP-controlled E1A was used to
generate a pair of complementary adenoviral vectors for HuH7 xenograft
treatment (19)
. After 3 weeks, HuH7 tumor growth was
inhibited by intratumoral injection. However, there was no clear tumor
reduction, and the dual vector system was concluded to have less
oncolytic potency (19)
. Taken together, our in
vitro and in vivo results presented in the current
study demonstrated that CV890 not only has a significantly higher
specificity (100-fold better), but also exhibits much better antitumor
efficacy than adenovirus variants including CV732 and AvE1a04 i. This
increased efficacy is attributable to the reincorporation of the
adenovirus E3 region as described previously (9)
.
Current treatment for primary liver cancer often fails when the cancer
has developed multiple foci or spread beyond the liver
(3, 4, 5)
. Our data suggest that virotherapy with CV890 may
provide a potential good treatment for AFP-producing HCC. Previous work
has demonstrated that it is possible to achieve a high transduction
efficiency of hepatocytes by i.v. administration of adenovirus
(30, 31, 32)
. Because the majority of liver tumors derive
their blood supply from the hepatic artery, whereas normal liver is fed
by the portal vein, it might be possible to direct the virus to the
tumor tissue by injecting it into the hepatic artery. Furthermore, the
replicating virus-infected cells produce progeny that can infect and
kill neighboring target cells. This is a significant advantage over
nonreplicating vectors of which the utility has been limited thus far
by inefficient tumor cell transduction. This was supported by our
animal study that i.v.-administered CV890 was able to cause tumor
regression or eradication in 90% of the treated animals without
toxicity (Fig. 8)
, and extensive virus-infected cells were observed
within the CV890-treated Hep3B xenografts.
To support clinical development, we also explored the possibility of adjuvant therapy by combining CV890 with chemotherapeutic agents. A synergistic decrease of cell viability of AFP-producing cells (HepG2 and Hep3B) was observed when CV890 was combined with doxorubicin. Hep3B cells cultured with doxorubicin for 24 h before or after infection with CV890 had significantly decreased viability compared with cells treated with either agent alone. Hep3B cells treated with doxorubicin exhibited a greater burst size of CV890, whereas cytolytic selectivity of CV890 was fully remained. A synergistic antitumor efficacy of combination therapy with CV890 and doxorubicin was also observed in vivo in Hep3B xenografts. Tumor growth was inhibited by a single i.v. administration of CV890 at a dose of 1 x 1011 particles, whereas combination treatment of CV890 with doxorubicin eliminated tumors within 6 weeks. Statistical analysis of the in vivo studies indicated that the CV890 and doxorubicin combination group showed a significant synergy with a 5.3-fold higher inhibition of tumor growth over additive effect (data not shown). Our previous study with CV787, a prostate cancer-specific adenovirus variant (9 , 20) , and this current study demonstrate that tumor-specific viruses together with chemotherapeutic agents provide synergistic antitumor effects. Therefore, CV890 in combination with doxorubicin has strong preclinical support for clinical development.
In summary, we have successfully extended our technology platform by creating CV890, a tumor-specific adenovirus by linking two essential viral genes, E1A and E1B, with an IRES. Use of an AFP TRE-E1A-IRES-E1B cassette yields a virus of very high specificity for target cells (5,000–100,000:1) with only a single tumor-specific TRE. We also show that CV890 replication in AFP+ cells is independent of AFP production levels between 60 ng and 4500 ng AFP/106 cells/10 days. The TRE-E1A-IRES-E1B bicistronic cassette strategy saves space within the virus genome allowing the reincorporation of the adenovirus E3 region adding much needed antitumor efficacy. A single dose of CV890 via i.v. administration inhibits tumor growth of the very aggressive human Hep3B xenograft in BALB/c-nu/nu mice. Synergistic antitumor efficacy was observed both in vitro and in vivo when CV890 was combined with a chemotherapeutic agent, doxorubicin, while retaining the same specificity of destroying AFP+ cells and leaving AFP- cells intact. A single i.v. administration of CV890 in combination with doxorubicin eliminated distant tumors within 4 weeks after treatment. These results support the clinical development of CV890 as an antineoplastic agent for the treatment of localized or metastatic HCC tumors.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
1 To whom requests for reprints should be
addressed, at Calydon, Inc., 1324 Chesapeake Terrace, Sunnyvale, CA
94089. Phone: (408) 734-0733; Fax: (408) 734-2808; E-mail: dhenderson{at}calydon.com 
2 The abbreviations used are: HCC, hepatocellular
carcinoma; ARCA, attenuated replication competent adenovirus; ECL,
enhanced chemiluminescence; EMCV, encephelomycarditis virus; nt,
nucleotide; Ad5, adenovirus type 5; E3, E3 region of Ad5; wt,
wild-type; TRE, transcription response element; AFP, -fetoprotein;
IRES, internal ribosome entry site; PSA, prostate-specific antigen;
MTT, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide; MOI,
multiplicity of infection(s); pfu, plaque-forming units; CV, Calydon
virus.
3 D. R. Henderson and D-C. Yu, unpublished data.
Received 2/28/01. Accepted 6/21/01.
|
REFERENCES |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
-fetoprotein gene in unicentric and multicentric human hepatocellular carcinomas.. Hepatology, 17: 35-41, 1993.[Medline]
-fetoprotein.. Crit. Rev. Eukaryot. Gene Expr., 7: 11-41, 1997.[Medline]
-fetoprotein expression in human hepatoma.. Mol. Cell. Biol., 11: 5885-5893, 1991.-fetoprotein gene.. J. Biol. Chem., 262: 4812-4818, 1987.This article has been cited by other articles:
|
|
 |
|
  E. G. Cafferata, D. R. Maccio, M. V. Lopez, D. L. Viale, C. Carbone, G. Mazzolini, and O. L. Podhajcer A Novel A33 Promoter-Based Conditionally Replicative Adenovirus Suppresses Tumor Growth and Eradicates Hepatic Metastases in Human Colon Cancer Models Clin. Cancer Res., May 1, 2009; 15(9): 3037 - 3049. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 O. Nakajima, A. Matsunaga, D. Ichimaru, Y. Urata, T. Fujiwara, and K. Kawakami Telomerase-specific virotherapy in an animal model of human head and neck cancer Mol. Cancer Ther., January 1, 2009; 8(1): 171 - 177. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Leja, H. Dzojic, E. Gustafson, K. Oberg, V. Giandomenico, and M. Essand A Novel Chromogranin-A Promoter-Driven Oncolytic Adenovirus for Midgut Carcinoid Therapy Clin. Cancer Res., April 15, 2007; 13(8): 2455 - 2462. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. Hoffmann and O. Wildner Restriction of adenoviral replication to the transcriptional intersection of two different promoters for colorectal and pancreatic cancer treatment. Mol. Cancer Ther., February 1, 2006; 5(2): 374 - 381. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. Ramesh, Y. Ge, D. L. Ennist, M. Zhu, M. Mina, S. Ganesh, P. S. Reddy, and D.-C. Yu CG0070, a Conditionally Replicating Granulocyte Macrophage Colony-Stimulating Factor-Armed Oncolytic Adenovirus for the Treatment of Bladder Cancer Clin. Cancer Res., January 1, 2006; 12(1): 305 - 313. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. Li, N. Idamakanti, T. Arroyo, S. Thorne, T. Reid, S. Nichols, M. VanRoey, G. Colbern, N. Nguyen, O. Tam, et al. Dual Promoter-Controlled Oncolytic Adenovirus CG5757 Has Strong Tumor Selectivity and Significant Antitumor Efficacy in Preclinical Models Clin. Cancer Res., December 15, 2005; 11(24): 8845 - 8855. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Wirth, F. Kuhnel, B. Fleischmann-Mundt, N. Woller, M. Djojosubroto, K. L. Rudolph, M. Manns, L. Zender, and S. Kubicka Telomerase-Dependent Virotherapy Overcomes Resistance of Hepatocellular Carcinomas against Chemotherapy and Tumor Necrosis Factor-Related Apoptosis-Inducing Ligand by Elimination of Mcl-1 Cancer Res., August 15, 2005; 65(16): 7393 - 7402. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. L. Chu, D. E. Post, F. R. Khuri, and E. G. Van Meir Use of Replicating Oncolytic Adenoviruses in Combination Therapy for Cancer Clin. Cancer Res., August 15, 2004; 10(16): 5299 - 5312. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. Kasuya, T. M. Pawlik, J. T. Mullen, J. M. Donahue, H. Nakamura, S. Chandrasekhar, H. Kawasaki, E. Choi, and K. K. Tanabe Selectivity of an Oncolytic Herpes Simplex Virus for Cells Expressing the DF3/MUC1 Antigen Cancer Res., April 1, 2004; 64(7): 2561 - 2567. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. S. Banerjee, A. A. Rivera, M. Wang, L. T. Chow, T. R. Broker, D. T. Curiel, and D. M. Nettelbeck Analyses of melanoma-targeted oncolytic adenoviruses with tyrosinase enhancer/promoter-driven E1A, E4, or both in submerged cells and organotypic cultures Mol. Cancer Ther., April 1, 2004; 3(4): 437 - 449. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
W.-P. Lee, D.-I. Tai, S.-L. Tsai, C.-T. Yeh, Y. Chao, S.-D. Lee, and M.-C. Hung Adenovirus Type 5 E1A Sensitizes Hepatocellular Carcinoma Cells to Gemcitabine Cancer Res., October 1, 2003; 63(19): 6229 - 6236. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. Li, Y. Chen, J. Dilley, T. Arroyo, D. Ko, P. Working, and D.-C. Yu Carcinoembryonic antigen-producing cell-specific oncolytic adenovirus, OV798, for colorectal cancer therapy Mol. Cancer Ther., October 1, 2003; 2(10): 1003 - 1009. [Abstract] [Full Text]
|
|
|
|
|
|
G. W. Demers, D. E. Johnson, V. Tsai, S.-F. Wen, E. Quijano, T. Machemer, J. Philopena, M. Ramachandra, J. A. Howe, P. Shabram, et al. Pharmacologic Indicators of Antitumor Efficacy for Oncolytic Virotherapy Cancer Res., July 15, 2003; 63(14): 4003 - 4008. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Wirth, L. Zender, B. Schulte, B. Mundt, R. Plentz, K. L. Rudolph, M. Manns, S. Kubicka, and F. Kuhnel A Telomerase-dependent Conditionally Replicating Adenovirus for Selective Treatment of Cancer Cancer Res., June 15, 2003; 63(12): 3181 - 3188. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Hamada, S. Kohno, M. Iwamoto, H. Yokota, M. Okada, M. Tagawa, S. Hirose, K. Yamasaki, Y. Shirakata, K. Hashimoto, et al. Identification of the Human IAI.3B Promoter Element and Its Use in the Construction of a Replication-selective Adenovirus for Ovarian Cancer Therapy Cancer Res., May 15, 2003; 63(10): 2506 - 2512. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. M. Bernt, D. S. Steinwaerder, S. Ni, Z.-Y. Li, S. R. Roffler, and A. Lieber Enzyme-activated Prodrug Therapy Enhances Tumor-specific Replication of Adenovirus Vectors Cancer Res., November 1, 2002; 62(21): 6089 - 6098. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. Bernt, M. Liang, X. Ye, S. Ni, Z.-Y. Li, S. L. Ye, F. Hu, and A. Lieber A New Type of Adenovirus Vector That Utilizes Homologous Recombination To Achieve Tumor-Specific Replication J. Virol., October 2, 2002; 76(21): 10994 - 11002. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. M. Nettelbeck, A. A. Rivera, C. Balague, R. Alemany, and D. T. Curiel Novel Oncolytic Adenoviruses Targeted to Melanoma: Specific Viral Replication and Cytolysis by Expression of E1A Mutants from the Tyrosinase Enhancer/Promoter Cancer Res., August 15, 2002; 62(16): 4663 - 4670. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Zhang, N. Ramesh, Y. Chen, Y. Li, J. Dilley, P. Working, and D.-C. Yu Identification of Human Uroplakin II Promoter and Its Use in the Construction of CG8840, a Urothelium-specific Adenovirus Variant That Eliminates Established Bladder Tumors in Combination with Docetaxel Cancer Res., July 1, 2002; 62(13): 3743 - 3750. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Tsukuda, R. Wiewrodt, K. Molnar-Kimber, V. P. Jovanovic, and K. M. Amin An E2F-responsive Replication-selective Adenovirus Targeted to the Defective Cell Cycle in Cancer Cells: Potent Antitumoral Efficacy but No Toxicity to Normal Cell Cancer Res., June 1, 2002; 62(12): 3438 - 3447. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Cancer Research | Clinical Cancer Research |
| Cancer Epidemiology Biomarkers & Prevention | Molecular Cancer Therapeutics |
| Molecular Cancer Research | Cancer Prevention Research |
| Cancer Prevention Journals Portal | Cancer Reviews Online |
| Annual Meeting Education Book | Meeting Abstracts Online |
