Oncolytic viral therapy for human colorectal cancer and

liver metastases using a multi-mutated herpes

simplex virus type-1 (G207)
DAVID A. KOOBY,* JOHN F. CAREW,* MARC W. HALTERMAN, JONATHAN E. MACK,*
JOSEPH R. BERTINO, LESLIE H. BLUMGART,* HOWARD J. FEDEROFF, AND
YUMAN FONG*,1
*Department of Surgery and Department of Molecular Pharmacology, Memorial Sloan-Kettering
Cancer Center, New York, New York 10021, USA; and Department of Microbiology and
Immunology and Department of Neuroscience, University of Rochester Medical Center, Rochester,
New York 14642, USA

G207 is a multi-mutated, replicationcompetent type-1 herpes simplex virus designed to

target, infect, and lyse neurological tumors. This study
examines the feasibility of using G207 in the treatment
of human colorectal cancer and defines the biological
determinants of its antitumor efficacy. This virus was
tested on five human colorectal cancer cell lines in vitro
to determine efficacy of infection and tumor cell kill.
These results were correlated to measures of tumor
cell proliferation. In vivo testing was performed
through direct injections of G207 into xenografts of
human colorectal cancer tumors grown in flanks of
athymic rats. To evaluate an alternate method of administration, hepatic portal vein infusion of G207 was
performed in a syngeneic model of liver metastases in
Buffalo rats. Among the five cell lines tested, infection
rates ranged between 10% and 90%, which correlated
directly with S-phase fraction (8.6%36.6%) and was
proportional to response to G207 therapy in vitro
(1%93%). Direct injection of G207 into nude rat flank
tumors suppressed tumor growth significantly vs. control (0.58 6 0.60 cm3 vs. 9.16 6 3.70 cm3, P<0. 0001).
In vivo tumor suppression correlated with in vitro effect.
In the syngeneic liver tumor model, portal infusion resulted in significant reduction in number of liver nodules
(13 6 10 nodules in G207-treated livers vs. 80 6 30
nodules in control livers, P<0.05). G207 infects and kills
human colorectal cancer cells efficiently. In vitro cytotoxicity assay and tumor S-phase fraction can be used to
predict response to treatment in vivo. This antineoplastic
agent can be delivered effectively by both direct tumor
injection and regional vascular infusion. G207 should be
investigated further as therapy for colorectal cancer and
liver metastases.Kooby, D. A., Carew, J. F., Halterman,
M. W., Mack, J. E., Bertino, J. R., Blumgart, L. H.,
Federoff, H. J., Fong, Y. Oncolytic viral therapy for
human colorectal cancer and liver metastases using
a multi-mutated herpes simplex virus type-1 (G207).
FASEB J. 13, 13251334 (1999)
ABSTRACT

0892-6638/99/0013-1325/$02.25 FASEB

Key Words: attenuated virus z gene therapy z hepatobiliary

tumors z replication z competent

Colorectal cancer is the third most common malignancy and the second leading cause of cancer deaths in
the United States (1). Half of those affected with colorectal cancer will die of advanced or recurrent disease,
despite aggressive treatment (2). The liver is the most
frequent site of distant metastatic spread from malignancies of the colon and rectum (3, 4). Patients with untreated liver metastases have a median survival of 5 to 10
months (5, 6). Although many forms of therapy have
been evaluated, only surgical resection of the affected
liver offers the possibility of cure (7). Two-thirds of those
who undergo successful resection, however, experience
recurrence, presumably from microscopic residual disease (8). For patients with unresectable liver metastases,
palliative treatment, in the form of chemotherapy, yields
clinical response in only one-third of cases (9). Thus,
active investigation seeks novel therapies that may improve outcome in both resectable and unresectable cases
of this common disease entity.
Several viral-based, antineoplastic strategies are being
evaluated for treatment of colorectal cancer and liver
metastases. Most use replication-defective viruses as vectors to transfer therapeutic genes that encode protein
products such as cytokines, prodrugs, and tumor suppressors (1013). Another promising strategy involves the use
of replication-competent viruses and exploits their natural
ability to infect and lyse tumor cells (14). These oncolytic
viruses are genetically engineered to be less virulent to
normal tissues and more specific toward tumor cells.
G207 is a second-generation, multi-mutated, replication-

1
Correspondence: Department of Surgery, Memorial
Sloan-Kettering Cancer Center, 1275 York Ave., New York, NY
10021, USA. E-mail: [email protected]

1325

competent herpes simplex virus type-1 (HSV-1),2 which

has demonstrated impressive oncolytic activity in several
neurological malignancies, while sparing normal neural
tissue (15, 16).
This engineered virus, based on the wild-type
HSV-1 (F-strain), contains deletions of both copies of
the g134.5 gene, which results in attenuated neurovirulence (17, 18). Interruption of the UL39 gene
restricts replication of the G207 virus to rapidly
dividing cells (19). UL39 codes for ICP6, the large
subunit of ribonucleotide reductase, which is a key
enzyme required for DNA synthesis and HSV replication. Nondividing cells are deficient in this protein, whereas rapidly dividing cancer cells express it
freely (20, 21). Viral replication can only take place
in the presence of this enzyme; thus, replication of
G207 is limited to dividing cells. In addition, G207
was created with the marker gene lacZ, which produces a histochemically identifiable protein product.
Cells infected with G207 will turn blue when exposed
to X-gal solution. Finally, G207 has two built-in safety
mechanisms: expression of herpes thymidine kinase,
which renders the virus sensitive to ganciclovir therapy; and temperature sensitivity, which halts viral
activity in the febrile host (16).
G207 was originally designed for treatment of
neurological malignancies, because wild-type HSV is
naturally neurotropic (16). Therapeutic safety and
efficacy in animal model neurological tumors encouraged us to study the effects of G207 in solid
tumors outside the central nervous system. We report the evaluation of G207 for colorectal cancer and
liver metastases in vitro and in animal models, and
demonstrate this virus to be promising for the treatment of this disease.

Cell culture
Five human colorectal cancer cell lines were used in this
study. Four (C18, C29, C85, and C86) were isolated and
characterized at Memorial Sloan-Kettering Cancer Center.
The fifth, HCT 8, was obtained from the ATCC. Cells were
maintained in RPMI 1640 supplemented with 10% fetal calf
serum, 1% L-glutamine, 100 mg/ml penicillin, and 100 mg/ml
streptomycin at 37C in 5% CO2 humidified atmosphere and
subcultured twice a week.
In vitro assay of infection and lysis of human colorectal
cancer cells
Each of the five cell lines was plated onto flat-bottom 96-well
microtiter plates (Becton Dickinson, Franklin Lakes, N.J.) at
3 3 104 cells per well. 24 h later, medium was removed and
fresh medium with the appropriate concentration of virus was
applied in a final volume of 200 ml; MOIs of 0, 0.1, 1.0, and
2.0 were evaluated. All assays were performed in triplicate.
To evaluate infection efficiency among the five cell lines,
X-gal (5-bromo-4-chloro-3-indolyl-b-D-galactopyranoside; Fisher
Scientific, Fair Lawn, N.J.) staining was performed on identically
prepared serial plates at 24, 48, and 72 h postinfection, using a
previously described technique (22). The percent of lacZ-positive cells was calculated as a measure of infection.
At 24, 72, and 120 h, wells were exposed to 0.25% trypsin;
cells were counted on a hemocytometer using trypan blue
exclusion to assess the ability of G207 to lyse cells from the
five colorectal cancer cell lines.
Determinations of cell doubling time and S-phase fraction
Doubling time
Cells (5 3 104) were plated onto T-25 flasks (Costar Corporation, Cambridge, Mass.) in 10 cc of medium. Trypsinization
of the monolayer and counting by hemocytometer was performed at 24 h intervals. Data were plotted and doubling time
was calculated.
Cell cycle analysis

MATERIALS AND METHODS

G207 virus
G207, a gift from Dr. S. D. Rabkin, was constructed as
described previously, with deletions of both copies of the
g134.5 gene and insertion of the Escherichia coli lacZ gene into
the UL39 sequence of the R3616 mutant (16). African green
monkey kidney cells (Vero cells, ATCC, Rockville, Md.) were
maintained in Dulbeccos modified Eagles medium with 5%
fetal calf serum under standard cell culture conditions. Virus
was propagated on Vero cells at an MOI (multiplicity of
infection, number of viral particles/cell) of 0.02 at 34C,
harvested after 2 days, and subjected to freeze-thaw lysis and
sonication to release G207 from the cell fraction. Cell lysates
were clarified by centrifugation (300 3 g for 10 min at 4C)
and viral supernatants were stored at 280C. Viral titers were
determined by standard plaque assay on Vero cells.

2
Abbreviations: HSV-1, herpes simplex virus type-1; i.p., intraperitoneal; MOI, multiplicity of infection; PBS, phosphate-buffered
saline; PCR, polymerase chain reaction; PFU, plaque-forming units.

1326

Vol. 13

August 1999

Cells in log-phase growth were harvested by trypsinization,

fixed in 80% ethanol, and stored at 220C until analysis.
Subsequently, cell suspensions were digested in DNase-free
RNase (Boehringer Mannheim, Indianapolis, Ind.) for 20
min at 37C and then stained with propidium iodide solution
(50 mg/ml, Molecular Probes, Eugene, Oreg.). Cell cycle
analysis was performed on FACScan equipped with FACStation running CellQuest software (Becton Dickinson, San Jose,
Calif.). A forward angle, light scatter threshold trigger was
used to eliminate debris. Cell clumps were removed using
analysis gates on either fluorescence pulse width or height vs.
pulse area (integral). Data were analyzed for 1 3 104 to 2 3
104 cells per sample.
Viral growth curves
To demonstrate replication of virus in susceptible cells, viral
growth curves were established in two representative cell lines
(HCT 8 and C18) as described previously (23). Cells (5 3
105) were plated in 6-well plates and infected with G207 at an
MOI of 0.01 (5 3 103 PFU). Successive well contents were
collected by cell scraping at 0, 12, 24, 36, and 48 h postinfection and subjected to three cycles of freeze-thaw lysis, sonication, and centrifugation at 2000 rpm for 10 min at 4 oC.

The FASEB Journal

KOOBY ET AL.

Supernatants were plated on confluent Vero cells in 6-well

plates at varying dilutions and recovered titers were determined through standard plaque assay.

Determination of viral persistence and dissemination

Suppression of flank tumor growth in athymic rats

Tumors and organs (brain, heart, lung, liver, kidney, spleen,

testes, small bowel, and skeletal muscle) were harvested from
animals in both in vivo experiments at 1, 7, and 14 days post
G207 injection. Additional flank tumors were harvested and
sectioned at day 21 postinfection. Specimens were frozen in
Tissue-Tek embedding medium (Sakura Finetek, Torrance,
Calif.) and sectioned by cryotome. Sections were fixed with
1-% glutaraldehyde and evaluated for b-galactosidase expression by staining with X-gal solution. Slides were subsequently
counterstained with Nuclear Fast Red (Sigma, St. Louis, Mo.).
Tissues harvested from uninfected animals were used as
negative controls.

All animal work was performed under guidelines approved by

the Memorial Sloan-Kettering Institutional Animal Care and
Use Committee. Athymic rats were housed in pathogen-free
quarters in the animal facility. Animals were anesthetized with
intraperitoneal (i.p.) injections of pentobarbital (50 mg/kg).
Two injections of 2 3 106 tumor cells in 50 ml of culture
medium were administered to each side of the animal for a
total of four tumors per animal, and tumor growth was
measured with calipers three times per week. Estimates of
tumor volumes were calculated using the formula for a
prolate spheroid, 4/3(p)ab2, with a as the radius of the long
axis and b as the radius of the short axis in millimeters.
When tumor volume reached ;50 mm3, 1 3 107 plaqueforming units (PFU) of active G207 or heat-inactivated virus
(65C for 20 min) in 50 ml of culture medium was injected
directly into xenografts. Three cell lines (C85, C86, and
HCT8) were selected for the flank tumor model. Sixteen
tumors were evaluated for each cell line; eight were treated
with active G207 and eight with inactivated virus control.
Tumor measurements and animal weights were followed
regularly. All control animals (n56) had to be killed at 3 wk
due to ulceration of the skin overlying the tumor. Four of the
six G207-treated animals were maintained for 8 wk to evaluate
duration of response.
Treatment of hepatic metastases with regional vascular
infusion of G207
An established model of hepatic micrometastases was chosen
to assess the role of regional hepatic administration of G207.
Male Buffalo rats (National Cancer Institute, Bethesda, Md.)
that receive intrasplenic injections of 1 3 106 syngeneic
Morris hepatoma McA-RH7777 cells (ATCC No. CRL 1601)
will reliably develop between 60 and 100 countable liver
metastases within 3 wk of tumor challenge. Hepatoma cells
were maintained in culture and periodically implanted in
flanks to ensure tumorigenicity. Infection and lysis of hepatoma cells with G207 was confirmed in vitro.
To evaluate the ability of G207 to suppress experimental
hepatic metastases, 17 rats underwent laparotomy, splenic
tumor challenge, and splenectomy according to the model of
Lafreniere and Rosenberg (24). One week later, portal vein
infusions were performed with 1 3 108 PFU of G207 (n57) in
0.5 ml phosphate-buffered saline (PBS) or an equal volume of
PBS alone (n510).
Portal vein infusion technique
Rats were anesthetized with i.p. pentobarbital (50 mg/kg).
Midline laparotomy incisions were made and the anterior
pyloric branch of the portal vein was exposed. Under an
operating microscope, 6 0 silk suture (USSC, Norwalk
Conn.) was used to obtain proximal and distal control of the
vessel. Through a small venotomy, a polyethylene catheter
(internal diameter 5 0.28 mm, Becton Dickinson) with a
beveled tip was fed into the portal vein and temporarily
secured with a microaneurysm clip (Becton Dickinson). By
this technique, homogeneous hepatic distribution was confirmed with India ink in a few animals. Experimental animals
were infused with 1 3 108 PFU of G207 in 500 ml of PBS or an
equal volume of PBS alone. Eleven days after treatment, rats
were killed and liver nodules were counted.
ONCOLYTIC VIRAL THERAPY FOR COLORECTAL CANCER

Histochemical analysis

Quantitative PCR analysis

To further define persistence of infection and extent of
dissemination, we performed quantitative polymerase chain
reaction (PCR) analysis on genomic DNA extracted from
liver, lung, brain, kidney, testes, and serum of Buffalo rats at
day 1 (n52) and day 7 (n52) postportal vein infusion of 1 3
108 PFU of G207. Additional liver and serum samples (n52)
were analyzed at day 14. Serial femoral artery blood samples
were collected from a single rat at 0, 5, 10, 30, and 60 min
after portal vein infusion of 1 3 108 PFU of G207. Serum
samples were extracted with phenol/chloroform and precipitated using yeast tRNA. Genomic DNA extraction was performed on all tissues. Standard curves were established by
doping uninfected liver and serum with known quantities of
the G207 virus prior to DNA extraction. Real-time quantitative PCR was performed using an ABI Prism 7700 Sequence
Detector (PE Biosystems), as described previously (25, 26).
Sense (59-ATGTTTCCCGTCTGGTCCAC-39), antisense (59CCCTGTCGCCTTACGTGAA-39) primers, and a dual-labeled
fluorescent TaqMan probe (59-6FAM-CCCCGTCTCCATGTCCAGGATGG-TAMRA-39) were designed to span the 111-bp
fragment of the HSV ICP0 (immediate early gene). Additional sense (59-CGCCTACCACATCCAAGGAA-39), antisense
(59-GCTGGAATTACCGCGGCT-39) primers, and TaqMan
probe (59-JOE-TGCTGGCACCAGCTTGCCCTC-TAMRA-39)
for the 187-bp 18S rRNA coding sequence were used in the
same reaction to normalize the amount of total DNA. Samples were subjected to 40 cycles of PCR (stage 1: 50 oC, 2 min;
stage 2: 95 oC, 10 min; stage 3: 95 oC, 15 s; 60 oC, 1 min; stage
4: 25 oC, soak). The AmpliTaq Gold nuclease cleaves a
fluorescent dye (FAM or JOE) from the nonextendable
probe, liberating it from the proximity of an associated
quencher (TAMRA). Probe binding is a requisite for primer
extension. The sequence detector is coupled to a chargecoupled device camera, which records the fluorescent emission spectra from individual wells at 500 650 nm with each
cycle. Specificity for amplified product is conferred by both
probe and primer sets, obviating the need for Southern blot
analysis of the PCR product.

RESULTS
In vitro assay of infection and lysis of human
colorectal cancer cells
In vitro infection efficiency was measured by staining
for b-galactosidase with X-gal solution 24, 48, and
1327

Figure 1. LacZ expression in C85 colorectal

cancer cells after infection with G207 over
time at MOI 0.1 (original magnification:
3100). AC) Control cells infected with
heat-inactivated G207 and stained with X-gal
solution at 24, 48, and 72 h, respectively.
Cells show no evidence of G207 infection
and continue to divide appropriately. DF)
Parallel plates infected with active G207. In
panel D (24 h), only a few cells demonstrate
G207 infection (lacZ positive), which increases dramatically in panel E (48 h). F)
Most cells have been lysed by G207 and all
remaining cells show evidence of infection.

72 h after treatment with G207 at several MOIs (Fig.

1). G207 demonstrated impressive infection efficiency in three of the five cell lines tested (C85, C86,
and HCT 8) and more moderate infectivity in C18
and C29, as determined by calculating the percent of
lacZ-positive cells (Fig. 2A). Infection correlated with
cell kill. The highly susceptible cell lines (C85, C86,
and HCT8) demonstrated 76% or better cell kill at
an MOI of 0.1, 72 h after infection, whereas the
moderate responders (C18 and C29) showed 31% or
less cell kill under the same conditions (Fig. 2B). All
colorectal cancer cell lines tested demonstrated appreciable susceptibility to G207 cytotoxicity as compared with controls, even at the lowest MOI (0.1). In
all cell lines tested, infection efficiency correlated
with percent tumor cell lysis.
Determinations of cell doubling time and S-phase
fraction
In vitro doubling time and S-phase fraction were
determined for each colorectal cancer cell line. Both
parameters were determined from cell populations
in log-phase growth. The two cell lines that displayed
moderate susceptibility to G207 (C18 and C29) had
doubling times of 44 and 48 h and S-phase fractions
of 9.2% and 8.6%. By comparison, the three cell
lines that showed strong susceptibility to the virus
(C85, C86 and HCT 8) had shorter doubling times
(19 26 h) and greater S-phase fractions (20.1%
36.6%, Table 1). These results demonstrate a direct
correlation between the proliferative indices of the
cancer cell lines and their susceptibility to G207
therapy in vitro (R 5 10.8, linear regression analysis
of infection percentage at MOI 1.0 vs. S-phase fraction).
1328

Vol. 13

August 1999

Viral growth curves

In vitro viral growth curves were performed to demonstrate the ability of G207 to replicate in human
colorectal cancer cells. For this experiment, two cell
lines were evaluated: a strong responder to in vitro
cytotoxicity (HCT 8), with a short doubling time (19
h) and high S-phase fraction (37%), and a less
responsive line (C18) with a longer doubling time
(44 h) and lower S-phase fraction (9%). The results
are depicted in Fig. 3. At the 0 time point, ;50% of
viral particles are recovered from the initial infection
in both cell lines. By 12 h, there is a modest drop of
recovered viral titers by an additional 50%, coinciding with viral disassembly within the infected cell.
Subsequently, in the responsive line (HCT 8), the
titers begin to rise at 24 h and by 48 h are increased
by 1.5 logs over the number of viral particles initially
added to the well. Conversely, in the less responsive
line (C18), there is no rise in recovered titers of
G207, suggesting that this slowly dividing cell line
does not foster efficient viral replication.
Suppression of flank tumor growth in athymic rats
Direct injection of G207 suppressed xenograft tumor
growth significantly in all three cell lines tested as
compared with controls (Fig. 4A, B). HCT 8 xenografts were most susceptible, with responses in 8/8
treated tumors (4/8 partial responses and 4/8 complete responses). C86 xenografts demonstrated similar results (4/8 partial responses and 2/8 complete
responses). C85 tumors were not completely suppressed by G207 injection, but growth rate was
reduced significantly as compared with control.
As a group, G207-treated animals maintained
their health as well or better than control animals.

The FASEB Journal

KOOBY ET AL.

Figure 3. Viral growth curves for G207 in two representative

colorectal cancer cell lines. Note that the amount of virus
recovered from the HCT 8 wells (cell line doubling time of
19 h) increases 1.5 logs over 48 h. Conversely, the amount of
virus recovered from the C18 wells (cell line doubling time of
44 h) falls well below the 5 3 103 PFU that was initially plated.

G207-treated animals that showed no evidence of

ulceration were evaluated for duration of response
for an additional 8 wk. Tumors in these animals
did not continue to grow during this period, and
the animals continued to groom and gain weight
appropriately.

Figure 2. In vitro effects of G207 infection on five human

colorectal cancer cell lines. A) Summary data for all five
colorectal cancer cell lines relating percent of lacZ-positive
cells to MOI. B) Cell survival, represented as percent of
control, for the five colorectal cancer cell lines at various
times after exposure to G207 at an MOI of 0.1.

G207-treated animal gained significantly more

weight than controls during the 20 day experiment
(100 6 16 g vs. 44 6 9 g, P,0.5). All animals that
developed ulceration in the skin overlying their
tumors had to be killed. This occurred in 6/6
control animals and in 2/6 of the G207-treated
animals by day 20 posttumor inoculation. The four
TABLE 1. Correlation of cell line proliferative indices and G207
infection and cytotoxicity
Doubling
time
Cell line (hours)

C18
C29
C85
C86
HCT 8

44
48
26
24
19

S-phase fraction
(percent)a

9.2, (CV 5 5.7)

8.6, (CV 5 5.3)
20.1, (CV 5 4.8)
26.5, (CV 5 5.1)
36.6, (CV 5 4.9)

% blue cells % cell kill

MOI 1.0
MOI 0.1
(24 h)b
(72 h)c

10 6 2
40 6 5
70 6 10
90 6 5
75 6 5

165
31 6 3
86 6 3
93 6 1
76 6 6

CV represents coefficient of variance for FACScan analysis.

c
From data presented in Fig. 2A.
From data presented in
Fig. 2B.
b

ONCOLYTIC VIRAL THERAPY FOR COLORECTAL CANCER

Treatment of hepatic metastases with regional

infusion of G207
Initial studies confirmed that the Morris hepatoma
McA-RH7777 cell line is sensitive to G207 in vitro
(data not shown). In the efficacy study, animals
that received portal infusions of G207 7 days after
splenic tumor challenge recovered well, grooming
and gaining weight equal to the PBS-treated controls. At death (day 11 postinfusion), G207-treated
livers contained far fewer nodules than PBStreated controls (13 6 10 nodules vs. 80 6 30
nodules, P,0.05, Fig. 5).
Determination of viral persistence and
dissemination
Histochemical analysis
Cryosections prepared from athymic rat organs
(brain, heart, lung, liver, kidney, spleen, testes, small
bowel, and skeletal muscle) harvested at 1, 7, and 14
days post G207 flank tumor injection demonstrated
absence of b-galactosidase expression. Conversely,
strong expression was found in all tumors that
received 1 3 107 PFU via direct, single intratumoral
injection. Tumors that demonstrated only partial
response to therapy were evaluated at 21 days and
found to have persistence of lacZ positivity in a
1329

Figure 4. Results of in vivo administration of G207. A) Representative nude rats demonstrating effects of injection with
heat-inactivated virus (control) or with viable G207. Animals are shown 20 days after a single injection (1 3 107 PFU) of each
tumor. B) Summary of in vivo injection experiments for three different colorectal cancer cell lines. Tumor volume is illustrated
at various times after a single injection with heat-inactivated virus (closed circles) or active G207 (open boxes, *P,0.0001;
**P,0.01; ***P,0.05, t test).

depot-like distribution. There was less overall necrosis witnessed in these specimens, suggesting less
infectivity of the virus in neighboring cells.
In the portal vein infusion model, all organs (with
the exception of liver) were negative for b-galactosidase activity. Nontumor liver parenchyma demonstrated scant positive expression overall (less than 1
blue cell/LPF). This was found in half the sections
studied (18/36), with the remainder showing no
blue cells in the normal parenchyma. In animals with
established gross hepatic tumor, which were then
exposed to portal vein G207, X-gal staining was
impressive. Twenty-four and 48 h after exposure to
G207, ;two-thirds of existing liver nodules (;50/
80) had strong expression of b-galactosidase in
grouped arrays (Fig. 6). About half of the nodules
had a single focus, whereas the remainder had
multiple positive regions with up to five per nodule.
Peripherally located tumor-associated plaques abutted but did not cross the border into the adjacent
normal liver parenchyma.

Figure 5. Effects of hepatic portal vein infusion of G207.

Representative livers 11 days after portal vein infusion with
PBS (A) or G207 (B). Portal vein infusions were performed 7
days after splenic injection of tumor cells.
1330

Vol. 13

August 1999

Quantitative PCR analysis

Quantitative PCR analysis of tissue and serum samples was performed to further define the presence,
persistence, and extent of dissemination of G207 in
animals that were infected via portal vein infusion.
By establishing a standard curve with 1, 5, 10, 103,
and 105 PFU, we first determined our limit of
reliable detection to be 10 PFU per 50 mg of tissue.
Upon analysis of the various organs (liver, lung,
kidney, brain, and testes), only liver clearly demon-

Figure 6. Representative frozen section of hepatic tumor

nodule (original magnification: 3100) from animal with
established liver metastases 48 h after portal vein infusion of
1 3 108 PFU of G207. Section is stained with X-gal solution
and counterstained with Nuclear Fast Red. Multiple blue
plaques are evident, demonstrating impressive distribution of
G207 infection within a single nodule.

The FASEB Journal

KOOBY ET AL.

Figure 7. Summary data for quantitative PCR analysis using

primers and probe for the HSV-ICP0 DNA sequence. ICP0
copy number correlates with plaque-forming units used in
standard curve. A) Number of ICP0 copies per 50 mg of tissue
at days 1, 7, and 14 after portal vein infusion of 1 3 108 PFU
of G207 in Buffalo rats (only liver tissue was analyzed at day
14). B) Number of ICP0 copies per 0.2 ml of serum isolated
from the femoral artery of Buffalo rats at 0, 5, 10, 30, and 60
min and 1, 7, and 14 days after portal vein infusion with G207.

strated the presence of the HSV-ICP0 DNA sequence

at any time point evaluated (Fig. 7A). Minimal signal
was observed in lung, brain, and testes, but based on
our standard curve, this expression was below the
limit of reliable detection. Furthermore, livers exposed to portal venous G207 demonstrated persistence of the viral DNA sequence at 14 days. Analysis
of serum samples showed significant recovery of viral
DNA in the peripheral arterial blood at 5 min (. 600
copies/0.2 ml) postportal venous infusion, which
decreased over the first hour and was absent from
24 h and beyond (Fig. 7B).

DISCUSSION
Our study evaluates the efficacy of the multi-mutated, attenuated HSV-1, replication-restricted virus
ONCOLYTIC VIRAL THERAPY FOR COLORECTAL CANCER

G207 to treat a common nonneurological neoplasm.

We have shown that this agent can infect, replicate
within, and kill human colon cancer cells, and that
cell line characteristics such as doubling time and
S-phase fraction may play a role in determining
sensitivity to G207. We have also demonstrated two
effective methods of therapeutic delivery in direct
intratumoral injection and regional vascular infusion, and have investigated the extent of dissemination and persistence of viral protein and DNA
through assay of marker gene product and quantitative PCR.
Some of the earliest work with replicating recombinant HSV vectors was directed at development of
antiherpetic vaccines, with the goal of establishing a
lasting immune response in infected hosts (27, 28).
As better understanding of the factors responsible
for HSV virulence (18) (17) and replication (19, 29)
developed, it became clear that these mutants could
be used to treat malignancies. Neurological tumors
were first targeted due to the natural neurotropism
of herpes viruses. Selective infection and destruction
of tumors within the central nervous system, with
sparing of normal neurons, has been demonstrated
by several investigators (15, 16, 30, 31).
More recently, several groups have tested various
recombinant, replicating HSV mutants to treat nonneurological malignancies with success (20, 23, 32).
Of note is work by Kucharczuk et al. (23), who
demonstrated safety and efficacy of HSV-1716 (a
g34.5 deleted mutant) in the treatment of experimental human malignant mesothelioma cell lines in
vitro and then in an animal model of peritoneal
disease. PCR analysis was used to demonstrate that
virus was detectable in tumor, but not in normal
tissues.
G207 is a second-generation, multi-mutated HSV-1
that was developed by taking the double g34.5deleted F-strain mutant R3616 and inactivating the
DNA sequence coding for viral ribonucleotide reductase. This was done through an insertional mutation with the E. coli lacZ gene (15, 16). Thus, not
only is this virus attenuated in virulence, but its
replication is restricted to cells that presumably
produce higher levels of endogenous ribonucleotide
reductase (i.e., rapidly dividing malignant cells) and
can be detected by assaying for production of b-galactosidase. Another benefit derived from using multiple mutations to create G207 is the reduced likelihood of reversion to wild-type disease-causing parent
virus.
The precise mechanism for viral cell kill is still
being deciphered. In susceptible cell lines, characteristics of normal HSV infection such as ballooning
and formation of multinucleated giant cells are
observed. This observation and evidence of viral
replication, as demonstrated through viral growth
1331

curves, support active, productive infection with

completion of the viral lytic life cycle as one mechanism contributing to the death of infected cells.
Moreover, several authors have supported the notion
that deletion of g34.5 may promote apoptosis in
susceptible cells (17, 18, 33). The carboxyl terminus
codes for a protein product that bears homology to a
highly conserved mammalian protein known as
GADD 34. This protein is up-regulated when cells
are placed under stressful conditions (e.g., serum
starvation or viral infection) and acts to preclude
protein synthesis. In the absence of this protein,
infected susceptible cells sense the insult and respond by entering the pathway to programmed cell
death instead of reevaluating and attempting to
repair and persevere.
As for restricted replication, it is known that at
least 45 of the 84 characterized HSV genes can be
inactivated or deleted and the virus will still replicate
in cultured cells, with some altered characteristics
(34). Inactivation of the gene encoding ICP6 (or
viral ribonucleotide reductase) is well described.
This enzyme plays a key role in DNA synthesis in
prokaryotes and eukaryotes. Evidence exists to suggest that an ICP6-inactivated mutant herpes virus
may be able to compensate for its deficit by using
cellular ribonucleotide reductase (19, 29, 35, 36).
Thus, rapidly dividing cells, which presumably express higher levels of ribonucleotide reductase, may
serve as more suitable hosts for G207 replication.
Our laboratory is currently working on defining this
relationship in our models.
Our data indicate that G207 efficiently infects
experimental human colorectal cancer cells. All five
cell lines tested responded to treatment in vitro.
Three of the cell lines (C85, C86, and HCT 8) had
remarkable responses by 3 and 5 days postinfection,
even at a low MOI of 0.1. These strong responders
exhibited more rapid cell division as demonstrated
by measurements of cell doubling time and S-phase
fraction. The relationship between cellular proliferative rate and responsiveness to G207 was observed
in the viral growth curves. G207 demonstrated active
replication within the rapidly dividing HCT 8 cell
line, whereas the cell line with a slower turnover
(C18) did not support successful viral growth. Furthermore, HCT 8 cells had the highest S-phase
fraction in vitro (36.6%) and the greatest in vivo
tumor growth inhibition of the three cell lines tested
in the direct intratumoral injection model, with 4/8
tumors sustaining complete responses. The observed
correlation between cell line proliferation and susceptibility to G207 in vitro may serve as a clinically
relevant predictor of in vivo therapeutic response.
In addition to examining direct intratumoral injections, we investigated the efficacy of G207 as a
possible agent for regional antineoplastic therapy.
1332

Vol. 13

August 1999

Introducing the virus by selective intravascular infusion is appealing since it allows a diffuse distribution
of virus within the tumor. The same hypervascularity
that the tumor creates to allow its own continued
growth could be exploited to increase viral delivery.
HSV-1 is a large DNA virus that approaches 200 nm
in diameter, raising a concern that infectious particles would not cross the endothelium and enter
tumor cells. We examined this issue using a wellestablished syngeneic model of hepatic metastases,
which involves seeding the liver with tumor via portal
inoculation, a path identical to that taken by colorectal liver metastases in human beings (3). Our
results suggest that regional infusion with G207 may
be valuable for treatment of unresectable liver malignancies or may be useful as an adjuvant to surgical
resection to reduce postoperative local recurrence.
Although G207 is an attenuated virus with restricted replication, therapeutic use of replicating
viruses still raises concerns of dissemination. In our
study, we used two methods of detection. The first,
X-gal staining, requires that the virus infects the cell,
transfers its genetic material, and the transferred
marker gene functions to produce a protein (b-gal).
We evaluated various tissues (including tumor) for
the presence of the marker gene product. In the
flank tumor model, only tumor demonstrated positive results. In the portal vein infusion model, high
expression was demonstrated in established tumor
with infrequent expression in normal hepatic parenchyma. All other organs were negative for G207
infection by this detection method.
The second method, real-time quantitative PCR,
detects presence of the herpes immediate-early gene
ICP0, which plays a role in cell cycle regulation in
infected cells (34). Unlike X-gal staining, this sensitive technique does not rely on the expression of
viral genes to identify infected cells. We analyzed
liver tissue as our target organ, lung tissue because it
receives most of the hepatic venous blood flow, brain
because of natural herpes tropism, kidney because of
its function of filtration, and testes because this is an
area with inherent rapid cell division. We found
presence of the HSV-ICP0 sequence in liver of
animals that received 1 3 108 PFU of portal G207 up
to 14 days and in the serum of an animal soon after
infusion of virus, but not in any other organs examined. Moreover, no mortalities were observed, and
G207-treated rats continued to groom and gain
weight appropriately at therapeutic doses.
Primates serve as the natural host for HSV-1, which
may explain selective infection of human tumor
xenografts in rodents. Our hepatic metastases model
addresses this issue of species tropism. Morris hepatoma McA-RH7777 is a syngeneic tumor cell line
developed and passaged in Buffalo rats. In this
model, G207 demonstrates impressive tumor cell

The FASEB Journal

KOOBY ET AL.

selectivity. These results suggest that G207 therapy is

both tumor specific and effective at suppressing
tumor growth in this model.
The future of oncolytic viral therapy is evolving.
Recent advances in the literature report methods to
improve potency of attenuated herpes viruses against
malignant cells. Such advances include adding ionizing radiation as an adjuvant to enhance the viruss
ability to replicate (37) and using attenuated, replicating viruses as helper viruses for packaging immunostimulatory cytokine genes (38). Another area
worthy of investigation is cotreatment with antiviral
agents. G207 expresses HSV-thymidine kinase and is
reported to be hypersensitive to ganciclovir (16).
Future experimentation should investigate the
added efficacy of combining the natural oncolytic
properties of G207 with this suicide gene therapeutic
strategy, bearing in mind that treatment with antiviral agents could, in theory, inactivate the virus prematurely.
G207 kills experimental human colorectal cancers
efficiently and in vitro assays may predict this agents
efficacy in vivo. This antineoplastic agent can be
administered safely in experimental animals by regional infusion and direct injection and can suppress
nonneural tumor growth. These results encourage
pursuing future clinical studies of G207 in the treatment of human metastatic colorectal cancers.
We thank T. Delohery for helping with cell cycle analysis
and Yong-jia You and Rita Giuliano for technical support.
D.A.K. was supported by training grant T32 CA 09501 and
J.F.C. was supported by training grant T32 CA 09685, both
from the U.S. Public Health Service. Y.F. was supported in
part by grants CA 76416 and CA 72632 from the National
Institutes of Health.

REFERENCES
1.
2.
3.

4.
5.

6.
7.

8.

Landis, S. H., Murray, T., and Bolden, S. (1998) Cancer

statistics, 1998. CACancer J. Clin. 48, 6 29
August, D. A., Ottow, R. T., and Sugarbaker, P. H. (1984) Clin.
perspective of human colorectal cancer metastasis. Cancer Metastases Rev. 3, 303
Weiss, L., Grundmann, E., Torhorst, J., Hartveit, F., Moberg, I.,
Eder, M., Fenoglio-Preiser, C. M., Napier, J., Horne, C. H.,
Lopez, M. J., et al. (1986) Haematogenous metastatic patterns in
colonic carcinoma: an analysis of 1541 necropsies. J. Pathol. 150,
195203
Willett, C. G., Tepper, J. E., Cohen, A. M., Orlow, E., and Welch,
C. E. (1984) Failure patterns following curative resection of
colonic carcinoma. Ann. Surg. 200, 685 690
De Brauw, L. M., Van de Velde, C. J., Bouwhuis-Hoogerwerf,
M. L., and Zwaveling, A. (1987) Diagnostic evaluation and
survival analysis of colorectal cancer patients with liver metastases. J. Surg. Oncol. 34, 81 86
Wood, C. B., Gillis, C. R., and Blumgart, L. H. (1976) A
retrospective study of the natural history of patients with liver
metastases from colorectal cancer. Clin. Oncol. 2, 285288
Fong, Y., Cohen, A. M., Fortner, J. G., Enker, W. E., Turnbull,
A. D., Coit, D. G., Marrero, A. M., Prasad, M., Blumgart, L. H.,
and Brennan, M. F. (1997) Liver resection for colorectal metastases. J. Clin. Oncol. 15, 938 946
Panis, Y., Ribeiro, J., Chretien, Y., and Nordlinger, B. (1992)

ONCOLYTIC VIRAL THERAPY FOR COLORECTAL CANCER

9.
10.

11.

12.

13.

14.
15.

16.

17.
18.

19.

20.

21.
22.
23.

24.
25.
26.

27.

28.

Dormant liver metastases: an experimental study. Br. J. Surg. 79,

221223
Fong, Y., Kemeny, N., Paty, P., Blumgart, L. H., and Cohen,
A. M. (1996) Treatment of colorectal cancer: hepatic metastasis.
Semin. Surg. Oncol. 12, 219 252
Karpoff, H. M., DAngelica, M., Blair, S., Brownlee, M. D.,
Federoff, H., and Fong, Y. (1997) Prevention of hepatic tumor
metastases in rats with herpes viral vaccines and gamma-interferon. J. Clin. Invest. 99, 799 804
Tung, C., Federoff, H. J., Brownlee, M., Karpoff, H., Weigel, T.,
Brennan, M. F., and Fong, Y. (1996) Rapid production of
interleukin-2-secreting tumor cells by herpes simplex virusmediated gene transfer: implications for autologous vaccine
production. Human Gene Ther. 7, 22172224
Caruso, M., Panis, Y., Gagandeep, S., Houssin, D., Salzmann,
J. L., and Klatzmann, D. (1993) Regression of established
macroscopic liver metastases after in situ transduction of a
suicide gene. Proc. Natl. Acad. Sci. USA 90, 7024 7028
Chen, S., Chen, X. L., Wang, Y., Kosai, K., Finegold, M. J., Rich,
S. S., and Woo, S. L. (1995) Combination gene therapy for liver
metastasis of colon carcinoma in vivo. Proc. Natl. Acad. Sci. USA
92, 25772581
Kenney, S., and Pagano, J. S. (1994) Viruses as oncolytic agents:
a new age for therapeutic viruses? J Natl. Cancer Inst. 86,
11851186
Yazaki, T., Manz, H. J., Rabkin, S. D., and Martuza, R. L. (1995)
Treatment of human malignant meningiomas by G207, a replication-competent multimutated herpes simplex virus 1. Cancer
Res. 55, 4752 4756
Mineta, T., Rabkin, S. D., Yazaki, T., Hunter, W. D., and
Martuza, R. L. (1995) Attenuated multi-mutated herpes simplex
virus-1 for the treatment of malignant gliomas. Nature Med. 1,
938 943
Roizman, B., and Markovitz, N. (1997) Herpes simplex virus
virulence: the functions of the gamma (1)34.5 gene. J. Neurovirol. 3 (Suppl. 1), S1S2
Chou, J., Kern, E. R., Whitley, R. J., and Roizman, B. (1990)
Mapping of herpes simplex virus-1 neurovirulence to gamma
134.5, a gene nonessential for growth in culture. Science 250,
12621266
Goldstein, D. J., and Weller, S. K. (1988) Factor(s) present in
herpes simplex virus type 1-infected cells can compensate for
the loss of the large subunit of the viral ribonucleotide reductase: characterization of an ICP6 deletion mutant. Virology 166,
4151
Carroll, N. M., Chiocca, E. A., Takahashi, K., and Tanabe, K. K.
(1996) Enhancement of gene therapy specificity for diffuse
colon carcinoma liver metastases with recombinant herpes
simplex virus. Ann. Surg. 224, 3239
Roizman, B. (1996) The function of herpes simplex virus genes:
a primer for genetic engineering of novel vectors. Proc. Natl.
Acad. Sci. USA 93, 1130711312
Geller, A. I., and Breakefield, X. O. (1988) A defective HSV-1
vector expresses Escherichia coli b-galactosidase in cultured
peripheral neurons. Science 241, 16671669
Kucharczuk, J. C., Randazzo, B., Chang, M. Y., Amin, K. M.,
Elshami, A. A., Sterman, D. H., Rizk, N. P., Molnar-Kimber,
K. L., Brown, S. M., MacLean, A. R., Litzky, L. A., Fraser; N. W.,
Albelda, S. M., and Kaiser, L. R. (1997) Use of a replicationrestricted herpes virus to treat experimental human malignant
mesothelioma. Cancer Res. 57, 466 471
Lafreniere, R., and Rosenberg, S. A. (1986) A novel approach to
the generation and identification of experimental hepatic metastases in a murine model. J. Natl. Cancer Inst. 76, 309 322
Heid, C. A., Stevens, J., Livak, K. J., and Williams, P. M. (1996)
Real time quantitative PCR. Genome Res. 6, 986 994
Luoh, S. M., Di Marco, F., Levin, N., Armanini, M., Xie, M. H.,
Nelson, C., Bennett, G. L., Williams, M., Spencer, S. A., Gurney,
A., and de Sauvage, F. J. (1997) Cloning and characterization of
a human leptin receptor using a biologically active leptin
immunoadhesin. J. Mol. Endocrinol. 18, 77 85
Meignier, B., Longnecker, R., and Roizman, B. (1988) In vivo
behavior of genetically engineered herpes simplex viruses
R7017 and R7020: construction and evaluation in rodents.
J. Infect. Dis. 158, 602 614
Meignier, B., Martin, B., Whitley, R. J., and Roizman, B. (1990)
In vivo behavior of genetically engineered herpes simplex

1333

29.

30.

31.
32.

33.

1334

viruses R7017 and R7020. II. Studies in immunocompetent and

immunosuppressed Owl monkeys (Aotus trivirgatus). J. Infect.
Dis. 162, 313321
Goldstein, D. J., and Weller, S. K. (1988) Herpes simplex virus
type 1-induced ribonucleotide reductase activity is dispensable
for virus growth and DNA synthesis: isolation and characterization of an ICP6 lacZ insertion mutant. J. Virol. 62, 196 205
Jia, W. W., McDermott, M., Goldie, J., Cynader, M., Tan, J., and
Tufaro, F. (1994) Selective destruction of gliomas in immunocompetent rats by thymidine kinase-defective herpes simplex
virus type 1 [see comments]. J. Natl. Cancer Inst. 86, 1209 1215
Miyatake, S., Martuza, R. L., and Rabkin, S. D. (1997) Defective
herpes simplex virus vectors expressing thymidine kinase for the
treatment of malignant glioma. Cancer Gene Ther. 4, 222228
Toda, M., Rabkin, S. D., and Martuza, R. L. (1998) Treatment of
human breast cancer in a brain metastatic model by G207, a
replication-competent multimutated herpes simplex virus 1.
Human Gene Ther. 9, 21772185
Chou, J., and Roizman, B. (1992) The gamma 34.5 gene of
herpes simplex virus 1 precludes neuroblastoma cells from
triggering total shutoff of protein synthesis characteristic of
programmed cell death in neuronal cells. Proc. Natl. Acad. Sci.
USA 89, 3266 3270

Vol. 13

August 1999

34.
35.

36.

37.

38.

Whitley, R. J., Kimberlin, D. W., and Roizman, B. (1998) Herpes

simplex viruses. Clin. Infect. Dis. 26, 541553
Goldstein, D. J., and Weller, S. K. (1988) An ICP6::lacZ insertional mutagen is used to demonstrate that the UL52 gene of
herpes simplex virus type 1 is required for virus growth and
DNA synthesis. J. Virol. 62, 2970 2977
Jacobson, J. G., Leib, D. A., Goldstein, D. J., Bogard, C. L.,
Schaffer, P. A., Weller, S. K., and Coen, D. M. (1989) A herpes
simplex virus ribonucleotide reductase deletion mutant is defective for productive acute and reactivatable latent infections of
mice and for replication in mouse cells. Virology 173, 276 283
Advani, S. J., Sibley, G. S., Song, P. Y., Hallahan, D. E., Kataoka,
Y., Roizman, B., and Weichselbaum, R. R. (1998) Enhancement
of replication of genetically engineered herpes simplex viruses
by ionizing radiation: a new paradigm for destruction of therapeutically intractable tumors. Gene Ther. 5, 160 165
Toda, M., Martuza, R. L., Kojima, H., and Rabkin, S. D. (1998)
In situ cancer vaccination: an IL-12 defective vector/replicationcompetent herpes simplex virus combination induces local and
systemic antitumor activity. J. Immunol. 160, 4457 4464

The FASEB Journal

Received for publication November 19, 1998.

Revised for publication March 2, 1999.

KOOBY ET AL.