| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
* Division of Surgical Oncology, Department of Surgery, and
Neurosurgery Service, Massachusetts General Hospital, Boston, Massachusetts 02114, USA
1Correspondence: Division of Surgical Oncology, Massachusetts General Hospital, Cox 626, 100 Blossom St., Boston, MA 02114 USA.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: HSV1 • gene therapy • ribonucleotide reductase
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, 2)
. Most gene therapy research has been devoted to the
development of strategies that allow viruses to deliver their genetic
payload without subsequent viral replication. It has been a long-held
belief that viruses used for delivery of genes into human cells for
therapeutic purposes should be rendered incapable of replication based
on concerns that viral replication may lead to cellular transformation
or produce significant illness. The overwhelming majority of cancer
gene therapy strategies reported to date use replication-incompetent
viruses (3)
. These viruses have served principally as
vehicles for delivery of therapeutic genes, such as cytokines, suicide
genes, costimulatory molecules, or tumor suppressor genes. However,
viruses engineered to remain replication-competent may be exploited for
cancer therapy because viral replication within cancer cells results in
oncolysis and produces progeny virion that can infect adjacent cancer
cells.
There has been growing interest in oncolytic viruses, which have
recently been introduced into clinical trials (4)
. One
particular replication-competent adenovirus that has received
considerable attention, Onyx-015, is defective in viral E1B 55 kDa
protein expression (5)
. It has been proposed that the
principal mechanism of tumor destruction by Onyx-015 results from
adenoviral replication specifically within p53-deficient cells,
although this concept has recently been challenged (6
7
8)
.
Reovirus has also been examined as a replication-competent oncolytic
virus (9)
because of its preferential replication in cells
with an activated ras pathway (10)
.
HSV1 is a double-stranded DNA virus that has been adapted for cancer
therapy. HSV1-based vectors were initially examined as potential
vehicles for gene transfer into the central nervous system. Unwanted
cytopathic effects resulting from viral infection plagued these early
experiments (11
, 12)
. However, these cytopathic effects
have subsequently been exploited to treat cancer. Entry of wild-type
HSV1 into cancer cells leads to a sequential cascade of viral gene
expression that ultimately results in production of multiple progeny
virions and cell death (13)
. Progeny virion can then
infect adjacent cancer cells to enhance the anti-tumor effects. Unlike
most cancer gene therapy strategies, those that rely on oncolysis
induced by viral replication do not require prolonged transgene
expression. HSV1 mutants have been engineered that are defective in
expression of genes that are important but not essential for viral
replication, such as thymidine kinase (14)
, ribonucleotide
reductase (15)
, and gamma 1 34.5 (16)
. These
mutants appear to replicate more robustly in tumor cells than in normal
cells, and have been directly inoculated into brain tumors in animals
to achieve anti-tumor effects (14
, 16
, 17)
. Unfortunately,
treatment strategies that require direct viral inoculation into each
and every tumor nodule are neither feasible nor effective for treatment
of solid tumor metastases, which typically present as multiple and
diffuse tumor nodules. Accordingly, we have examined a cancer therapy
approach designed to target an HSV1 mutant for diffuse liver metastases
after intravascular delivery. Here we show that an oncolytic HSV1
mutant defective in ribonucleotide reductase expression replicates
preferentially in tumors and not in surrounding normal liver after
intravascular delivery into mice bearing diffuse liver metastases.
Moreover, this viral replication produces significant anti-neoplastic
effects after a single intravascular administration, whereas
replication-incompetent HSV1 mutants do not produce any measurable
anti-tumor activity against liver metastases. Inhibition of tumor
growth is equivalent in immune-competent and immune-incompetent mice.
We also demonstrate that HSV1-mediated oncolysis of diffuse liver
metastases is effective in mice preimmunized against HSV1.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, 19)
. Primary mouse hepatocytes were prepared from BALB/c mouse
fresh liver specimens in a similar manner. Primary hepatocytes were
maintained in William’s medium E containing bovine serum albumin,
insulin, transferrin, selenium, trace elements, dexamethasone, linoleic
acid, linolenic acid, glucagon, penicillin, streptomycin, and fungizone
on collagen-coated plates. The HSV 1 vectors hrR3 (15)
(kindly provided by S. K. Weller, University of Connecticut),
d120 (20)
, and d27 (21)
(both kindly provided by D. M. Knipe, Harvard Medical School) were
derived from the parental wild-type strain KOS (kindly provided by
D. M. Coen, Harvard Medical School). hrR3 and KOS were propagated
and titered on Vero cells, and d120 and d27 were
propagated and titered on E5 and V27 cells, respectively. Heat
inactivation of hrR3 was performed as described (22)
.
Viral replication and cytotoxicity assays
Viral replication assays were performed as described
(21)
. Briefly, 3 x 106 cells
were infected with 6 x 106 plaque forming
units (pfu) of virus for 2 h, at which time unadsorbed virus was
removed by washing with a glycine-saline solution (pH 3.0). Forty hours
after infection the supernatant and cells were harvested, exposed to
three freeze/thaw cycles to release virions, and titered on Vero cells.
The results are the combination of three independent experiments. Viral
cytotoxicity assays were performed as described (23)
.
Briefly, cells were plated onto 96-well plates at 5000 cells per well
for 36 h. Virus was added at multiplicity of infection (moi)
values ranging from 0.0001 to 10 and incubated for 6 days. The number
of surviving cells was quantitated using a colorimetric MTT assay.
Tests were performed in quadruplicate.
Western blot analysis
Cell and liver tissue lysates containing equal amounts of
protein as determined by BCA protein assay reagent (Pierce, Rockford,
Ill.) were resolved by gel electrophoresis on 4–20% Tris-glycine
polyacrylamide gels (Novex, San Diego, Calif.), and proteins were
transferred onto nitrocellulose membranes (Schleicher & Schuell, Keene,
N.H.). Membranes were blocked with 5% nonfat milk overnight at 4°C.
Membranes were incubated with 5 µg/ml mouse anti-ribonucleotide
reductase mAb (MAB3033, Chemicon International, Inc., Temecula, Calif.)
for 1 h or 0.44 µg/ml mouse anti-ß-actin mAb (A-5441, Sigma,
St. Louis, Mo.) for 1 h. After washing three times, membranes were
incubated with peroxidase linked anti-mouse immunoglobulin (NA 931,
Amersham Life Sciences, Arlington Heights, Ill.) diluted 1:4000 for
1 h. Specific proteins were detected using an enhanced
chemiluminescence (ECL) system following the manufacturer’s
instructions (Amersham Life Sciences, Inc.).
Animal studies
BALB/c and athymic BALB/c (nu/nu) mice were obtained from
Charles River Labs (Wilmington, Mass.). Animal studies were performed
in accordance with policies of the Massachusetts General Hospital
Subcommittee on Research Animal Care. To examine sites of hrR3
replication in livers, MC26 liver metastases were generated by
injection of a single-cell suspension of 1 x
105 MC26 cells in 100 µl HBSS without
Ca2+ or Mg2+ into the
spleens of BALB/c mice. Eight days later, 1 x
107 or 1 x 108 pfu
hrR3 in 100 µl media was injected into a non-tumor-bearing portion of
the spleen (n=3 per group). The presence of tumor in the
spleen does not affect delivery of virus to the liver after inoculation
of virus into the spleen (data not shown). Mice were killed 12 days
after tumor inoculation.
For partial hepatectomy experiments, mice underwent a 75% hepatectomy
or sham laparotomy as described (24)
. Mice were killed at
specified times, and livers were harvested to assay for ribonucleotide
reductase expression (n=2 per group). To determine whether
hrR3 replicates in regenerating mouse livers, 1 x
108 pfu hrR3 in 100 µl media was injected into
the spleen 4 days after partial hepatectomy (n=3 per group).
Mice were killed 8 days later.
To assess the therapeutic efficacy of HSV1 against early liver metastases, a single-cell suspension consisting of 1 x 105 MC26 cells in 100 µl HBSS without Ca2+ or Mg2+ was injected into spleens of BALB/c mice, followed 3 days later by 5 x 107 pfu hrR3, heat-inactivated hrR3, d27, d120 in 100 µl media or media alone (n=4 per group). Mice were killed 14 days after tumor implantation. To assess therapeutic efficacy against late liver metastases, a single-cell suspension of MC26 cells was injected into the spleen, and mice were then treated with either hrR3 or heat-inactivated hrR3 7 days after tumor implantation (n=5 per group). Again, mice were killed 14 days after tumor implantation.
To examine the therapeutic efficacy of hrR3 in preimmuned mice with
liver metastases, 1 x 107 pfu KOS in 100
µl media was injected into the subcutaneous (s.c.) flanks of mice.
This had previously been demonstrated to protect mice from subsequent
lethal challenge with HSV1 (25)
. Control mice were
vaccinated with media instead. Two mice in each group were killed after
28 days to collect serum for measurement of the presence of antibodies
capable of neutralizing hrR3 cytotoxicity against MC26 cells. This
cytotoxicity assay was performed as described above, except in this
case hrR3 was incubated with one of the four mouse serum samples for 30
min prior to dilution and application to MC26 cells. In a separate
experiment, mice preimmunized with either KOS or media were injected
with MC26 cells into the spleen after 25 days and then treated 5 x 107 pfu hrR3 into the spleen after 3 more days
(n=5 per group). Mice were killed 14 days after tumor
implantation.
Subcutaneous tumors were generated by implantation of 2.5 x 106 MC26 cells in 100 µl HBSS s.c. into the flanks of immune-competent BALB/c mice or immune-incompetent athymic BALB/c (nu/nu) mice. Eight and 11 days later, 1 x 108 pfu hrR3 in 50 µl media or media alone was injected into the center of the developing tumor (n=5 per group). Flank tumor length (L) and width (W) were determined every 3–4 days; tumor volume (TV) was determined by the following formula: TV = (L x W2)/2. In a separate experiment, only 1 x 104 MC26 cells in 100 µl HBSS were injected into the spleens of BALB/c mice, followed 3 days later by either 5 x 107 pfu hrR3 or media alone (n=5 per group). Eight days after intrasplenic tumor inoculation, 2.5 x 106 MC26 cells were implanted s.c. into the flank and tumor volume was determined every 3–4 days. In this experiment, 1 x 104 MC26 cells were initially implanted into the spleen rather than 1 x 105 MC26 cells in order to allow the mice to live long enough to monitor their flank tumor growth.
Histochemistry
Histochemical staining to detect ß-galactosidase expression
was performed as described (19)
. Briefly livers were
perfused in vivo with phosphate-buffered saline (PBS),
followed by 4% paraformaldehyde in PBS, harvested, soaked in 4%
paraformaldehyde in PBS for 2 h, soaked in 30% sucrose in PBS for
24–48 h, and snap frozen in OCT compound in liquid nitrogen. Five
micron frozen sections were prepared and incubated in X-gal solution
(Fisher Scientific, Pittsburgh, Pa.) overnight. Sections were
counterstained with neutral red (Sigma).
Statistical analysis
Mean liver and spleen weights and flank tumor volumes were
compared using an unpaired two-tailed t test (InStat,
Graphpad Software, New York, N.Y.).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. To exploit this difference in ribonucleotide
reductase expression, we selected an HSV1 mutant, hrR3, that is
defective in infected cell protein 6 (ICP6) expression because of an
insertion of the Escherichia coli
ß-galactosidase gene into this locus (15)
. ICP6 serves
as the large subunit of viral ribonucleotide reductase. We reasoned
that hrR3 replication will proceed more robustly in cells with high
levels of cellular ribonuceotide reductase and high intracellular pools
of deoxyribonucleotides. To test this hypothesis, we compared
replication of hrR3 with that of a wild-type HSV1 strain, KOS, in
primary cultures of human hepatocytes and in HT29 human colon carcinoma
cells (Fig. 1A
). hrR3 replicated nearly as efficiently as KOS in the human
colon carcinoma cells; however, as expected hrR3 replication was three
log orders less than that of KOS in human hepatocytes. Levels of
cellular ribonucleotide reductase were significantly higher in HT29
cells than in human hepatocytes as assessed by Western blot (Fig. 1A
, inset). We performed a similar analysis in a mouse
model, comparing replication of both viruses in primary cultures of
mouse hepatocytes and in MC26 mouse colon carcinoma cells (Fig. 1B
). As was observed in the human tissue system, hrR3
replication was similar to that of KOS in the colon carcinoma cells,
but nearly three log orders less than that of KOS in the mouse
hepatocytes. HSV1 displays tropism for human and primate cells relative
to mouse cells (13)
, and the overall efficiency of
replication observed in the mouse system was less than in the human
system. Nonetheless, a significant differential in replication
efficiency was still observed. And similar to the human cells, mouse
colon carcinoma cells expressed significantly higher levels of
ribonucleotide reductase compared to mouse hepatocytes (Fig. 1B
, inset).
|
We have previously demonstrated that hrR3 replicates preferentially in
human colon carcinoma liver metastases compared to normal liver cells
after portal venous administration in a nude mouse xenograft model
(23)
. However, the tropism of HSV1 for human cells
relative to mouse cells made interpretation of these results difficult.
Accordingly, we examined hrR3 replication in BALB/c mice bearing
syngeneic MC26 colon carcinoma liver metastases. To establish
experimental liver metastases, we implanted 1 x
105 MC26 cells into the spleens of BALB/c mice,
and 8 days later we inoculated either 1 x
107 or 1 x 108 pfu of
hrR3 into the spleen. Four days after inoculation with the lower dose
of hrR3, ~50% of tumor nodules stained blue when examined for
ß-galactosidase expression (Fig. 2A
). However, only a minority of cells within any tumor
deposit stained blue. After inoculation with the higher dose of hrR3,
more than 90% of tumor nodules harbored blue-stained cells, and the
overall extent of ß-galactosidase expression within each nodule was
generally greater (Fig. 2B
). Notably, the surrounding normal
liver parenchyma did not contain any cells that expressed
ß-galactosidase, including hepatic endothelial cells. We did not
observe any ß-galactosidase expression in brain or lung specimens
either (data not shown).
|
To confirm that the preferential hrR3 replication in metastatic tumor
deposits resulted from the general absence of cell division in the
normal liver, we experimentally increased the level of cell division in
the normal liver by performing a 75% hepatectomy in BALB/c mice. This
procedure has been demonstrated to produce hypertrophy of the remaining
liver (24
, 26)
. We confirmed that ribonucleotide reductase
levels are low in mouse livers, but quickly rise in response to partial
hepatectomy (Fig. 3A
). This increase in ribonucleotide reductase levels persists
for at least 10 days, and corresponds to hepatic regeneration. In a
separate set of experiments, BALB/c mice were subjected to 75%
hepatectomy or sham laparotomy; 4 days later, 1 x
108 pfu of hrR3 was inoculated into their
spleens. Livers were harvested 8 days later and examined for
ß-galactosidase expression. There were numerous islands of
blue-staining cells in the mice subjected to hepatectomy and no
blue-staining cells in the livers from the mice subjected to sham
laparotomy (Fig. 3B, C
) or in livers from mice
subjected to hepatectomy without subsequent administration of virus
(data not shown). The blue-stained cells were generally observed in
clusters rather than as individual cells, which suggests that the
progeny virions produced from hrR3 replication infected adjacent cells.
Thus, hrR3 does not appear to replicate in quiescent liver cells, but
does replicate in a regenerating liver.
|
HSV1 replication produces significant oncolytic effects
Because the ß-galactosidase gene in hrR3 is under the control of
the promoter for the HSV1 early gene ICP6, expression of this marker
gene occurs early in the course of viral infection and is not
necessarily associated with completion of the lytic replicative cycle.
Accordingly, we directly examined the oncolytic effects associated with
infection of tumor cells by replication-conditional and
replication-incompetent HSV1 mutants. d120 and
d27 are HSV1 mutants defective in expression of the
immediate-early proteins ICP4 and ICP27, and are therefore capable of
replication only in cells transformed with ICP4 or ICP27, respectively
(20
, 21)
. HT29 human colon carcinoma cells and MC26 murine
colon carcinoma cells were infected with hrR3, heat-inactivated hrR3,
d120, or d27 in increasing moi values. HT29 cell
destruction was nearly complete 6 days after hrR3 infection when using
titers as low as one viral pfu per 100 tumor cells (Fig. 4A
). In contrast, very little cytotoxicity was observed with
heat-inactivated hrR3, d120, or d27. The minimal
cytotoxicity observed after infection with d120 or
d27 using the highest moi value is probably secondary to
HSV1 virus-induced host shutoff protein rather than HSV1 replication
(27)
. As expected, mouse colon carcinoma cells were
slightly less susceptible to the cytopathic effects of hrR3 infection
(Fig. 4B
). The moi values that produce nearly complete cell
destruction are one to two log orders lower than those required of
replication-competent reovirus (9)
. We observed similar
oncolytic effects of hrR3 infection in a panel of several different
human and mouse colon carcinoma cell lines (data not shown; ref
19
).
|
Although we observed evidence of hrR3 replication within tumor nodules
and not in normal liver after intraportal delivery, the therapeutic
efficacy of this replication has never been demonstrated. Therefore, we
examined the oncolytic efficacy of hrR3 when administered
intravascularly to treat diffuse liver metastases. We used a
well-described model of experimental liver metastases involving tumor
cell implantation into the spleen (28)
rather than the
portal vein, because portal vein implantation in BALB/c mice frequently
leads to thrombosis and death. Also, we chose to examine the
therapeutic efficacy of administration of hrR3 regionally into the
portal vein rather than into the tail vein, because we have observed
that portal venous inoculation produces significantly higher levels of
hrR3 replication in liver metastases compared to tail vein inoculation
(data not shown). Diffuse colon carcinoma liver metastases were
established in syngeneic BALB/c mice by intrasplenic inoculation of
1 x 105 MC26 tumor cells, and mice were
then randomized to treatment with a single intrasplenic inoculation of
5 x 107 pfu hrR3, 5 x
107 pfu heat-inactivated hrR3, or media alone.
When mice were killed 14 days after initial tumor cell implantation,
all of the mice treated with heat-inactivated hrR3 or media alone had
distended abdomens with bloody ascites and some were moribund. In
contrast, the hrR3-treated animals all appeared healthy. The number of
tumor nodules in the hrR3-treated group ranged from one to five per
animal, whereas the nodules in the mice treated with heat-inactivated
virus or media alone were too numerous to count (Fig. 4C
and
Table 1A
). The striking difference in liver tumor burden that
resulted from this single administration of virus also yielded a
difference in liver tumor weights that was statistically significant.
Two representative livers from each group were sectioned and stained
with hematoxylin and eosin. Both hrR3-treated livers contained tumor
nodules, although the size of the nodules was markedly smaller than in
livers of animals treated with HBSS or heat-inactivated virus. In
livers treated with heat-inactivated hrR3 or media, much of the liver
parenchyma was replaced by tumor. A significant inflammatory response
was not evident in any of the liver sections and no histopathological
evidence of hepatotoxicity was observed in areas of normal liver (data
not shown).
|
In this treatment model, virus was inoculated into the spleens just 3
days after tumor cell implantation, when liver metastases are
microscopic and not visible by gross inspection. Although the size and
weight of splenic tumors did not differ significantly between control
mice and hrR3-treated mice, it is possible that much of the anti-tumor
effect resulted from hrR3 oncolysis of tumor cells within the spleen
prior to their metastasis to the liver. The inability to reproducibly
inoculate tumor cells or virus directly into the hepatic artery or
portal vein in these mice necessitated a different strategy to address
this issue. We repeated the same experiment; however, the inoculation
of hrR3 or heat-inactivated hrR3 was performed 7 days after tumor cell
implantation, at which time tumors were 1–3 mm in size. Significant
anti-tumor activity was still observed after a single administration of
hrR3 despite the larger initial hepatic tumor burden (Table 1B)
. These
results suggest that hrR3 exerted its anti-tumor effects in the liver
after diffuse hepatic metastases had already been established.
Host immune responses do not contribute significantly to hrR3
oncolysis
The striking anti-tumor effects observed as a result of a single
intravascular administration of hrR3 could theoretically be a
consequence of viral replication and oncolysis or, alternatively, could
be a result of anti-tumor immune responses. To examine whether viral
replication is required to produce the anti-tumor effects, we
established diffuse MC26 liver metastases in another cohort of mice and
treated them with an intrasplenic inoculation of either d120
or d27. Tumor-bearing mice treated with either of these
replication-incompetent HSV1 mutants appeared similar to those treated
with media or heat-inactivated virus, and many were moribund when
killed 14 days after tumor implantation. Livers from these animals
appeared identical to those of the mice treated with either media or
heat-inactivated virus (Fig. 4C
). These results strongly
suggest that viral replication is required to achieve the anti-tumor
effects observed in vivo.
We also compared the oncolytic effects of hrR3 against MC26 cells
growing in immune-competent, syngeneic BALB/c mice and in
immune-incompetent, athymic BALB/c (nu/nu) mice. We chose a flank tumor
growth model in order to allow serial tumor size measurements to look
for subtle differences in anti-neoplastic efficacy between the two
models. Subcutaneous MC26 tumors were established on the flanks of both
the immune-competent and immune-incompetent mice, and the tumors were
treated with direct intralesional inoculations of either hrR3 or media
8 and 11 days later. The MC26 tumors treated with media grew equally
well in both types of mice (Fig. 5A
). Intralesional inoculation of hrR3 produced significant
anti-tumor effects in both types of mice. These anti-tumor effects were
nearly identical, with a 54% and 61% reduction in tumor size 23 days
after initial tumor implantation in BALB/c and BALB/c (nu/nu) mice,
respectively. These results suggest that the intact immune system in
BALB/c mice neither enhances nor attenuates the anti-neoplastic effects
of hrR3 infection of MC26 tumors compared with that observed in BALB/c
nude mice.
|
As another measure of the host immune response’s role in the observed
anti-tumor effects, we examined whether hrR3 treatment of diffuse liver
metastases results in a measurable effect against uninfected tumors
growing remotely on the flank. We again treated diffuse MC26 liver
metastases with a single intrasplenic inoculation of hrR3. Five days
later, 2.5 x 106 MC26 cells were implanted
s.c. over the flank, and flank tumor size was determined serially.
Control mice had MC26 cells implanted into the flank, but did receive
hrR3 treatment of liver metastases. None of the flank tumors in either
group were rejected and the rate of flank tumor growth was no different
in mice treated with hrR3 compared to control mice (Fig. 5B
). These results again suggest that the principal
mechanism by which hrR3 achieves its anti-tumor effects is oncolysis
resulting from viral replication rather than induction of host immune
responses against the tumor.
HSV1-mediated oncolysis of diffuse liver metastases is effective in
mice immunized against HSV1
The prevalence of preexisting antibodies to HSV1 in some
populations in the U.S. is as high as 80% (29)
, which may
reduce the efficacy of HSV1 therapy (30)
. To examine the
effect of immunity to HSV1 on hrR3 oncolysis of liver metastases, we
vaccinated BALB/c mice with either wild-type HSV1 (KOS strain) or
media. The presence of neutralizing antibodies in two mice from each
group was confirmed 28 days later (Fig. 6
). Experimental MC26 liver metastases were established, followed by
treatment with hrR3. The presence of neutralizing antibodies in the
KOS-vaccinated mice neither enhanced nor reduced the efficacy of hrR3
treatment of the liver metastases (Table 1C)
.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
). First, tumor cells
infected with HSV are destroyed during viral replication; transgene
expression is not required for anti-tumor activity (23
, 32
, 33)
. Second, these oncolytic effects are present even at
extremely low multiplicity of infection (moi) values compared to
adenovirus, vaccinia virus, and reovirus (23
, 32)
. Third,
HSV’s intrinsic thymidine kinase is expressed during viral
replication, thereby permitting enhancement of the anti-tumor effect by
treatment with prodrugs such as ganciclovir and bromovinyldeoxyuridine
(23
, 32
, 34)
. Fourth, HSV1 vectors can carry up to 50 kb
of transgene sequence, thereby allowing delivery of additional
therapeutic genes to increase oncolytic activity (35)
. And
fifth, the virus is a common pathogen in humans, yet very rarely causes
serious medical illness (36)
.
One of the earliest replication-competent HSV1 mutants examined for its
anti-tumor efficacy is dlsptk, which is defective in
expression of viral thymidine kinase (14)
. It was proposed
that replication of dlsptk proceeds selectively in actively
dividing cells, which can complement the absence of viral thymidine
kinase. Direct inoculation of this vector into intracranial gliomas
produces anti-tumor effects in mice (14)
. The vector that
we have investigated in our studies, hrR3, is defective in viral
ribonucleotide reductase expression. When inoculated directly into
intracranial gliosarcomas, this vector also produces anti-tumor effects
(17)
. Another HSV1 mutant, G207, which is defective in
both gamma 1 34.5 and ribonucleotide reductase, has also been
inoculated directly into intracranial tumors to produce anti-tumor
effects (16)
. None of these studies have demonstrated that
1) intravascular delivery of HSV1 can selectively target
multiple and diffuse tumors; 2) HSV1 replication is required
to achieve the observed anti-tumor effects; 3) oncolysis
results from viral replication rather than induction of host immune
responses against the tumor; 4) oncolysis is effective in
mice immunized against HSV1; and 5) HSV1-induced oncolysis
is an effective therapy for tumors of non-CNS origin. We have provided
experimental data that support each of these five important points.
In previous work we demonstrated that hrR3 replicates specifically in
liver metastases rather than normal liver, but we did not demonstrate
the therapeutic efficacy of this replication (19
, 23)
. In
the present study we demonstrate the therapeutic efficacy of hrR3
replication in liver metastases, and have examined the role of host
immunity. HSV1 antigens expressed on tumor cells may serve as potent
immunogens and modulate the anti-neoplastic activity observed after
tumor infection with HSV1. Previous in vivo studies
examining HSV1 oncolysis have used highly immunogenic tumor models
(17)
, and in some of these models the host immune response
contributed significantly to the anti-tumor effect (37
, 38)
. We have used several controls in our experiments and
concluded that the host immune response contributed minimally (if at
all) to the observed anti-tumor effects in the liver. First, the
observed anti-tumor effects on MC26 cells infected with hrR3 were
similar in immunocompetent mice and congenitally athymic mice. Second,
we could not detect a vaccination response against MC26 tumor cells
implanted into the flank after treatment of MC26 liver metastases with
hrR3. Third, treatment of liver metastases with replication-incompetent
HSV1 mutants d120 and d27 did not produce any
measurable reduction in tumor growth. ß-galactosidase is an immunogen
expressed by hrR3 but not by d120 or d27. However, it is
unlikely that this E. coli protein accounted for
the differences between these viruses in their anti-tumor activity,
since it is only one of many immunogens expressed by HSV1.
In an overwhelming majority of patients with solid tumor metastases,
most deposits of metastatic tumor are diffuse and not detectable by
radiological techniques. This clinical scenario precludes therapeutic
approaches that require inoculation of each and every metastatic tumor
deposit with an anti-neoplastic agent. Accordingly, we have focused on
strategies involving intravascular delivery to target diffuse
metastatic deposits. Unlike other solid tumors, colorectal carcinoma
commonly spreads to the liver without simultaneous spread to other
organs (39)
. Because of this unique tumor biology,
therapies directed specifically against liver metastases have enhanced
survival (40)
.
Much of the safety of oncolytic viral therapy depends on the selectivity of viral replication for tumor cells compared to normal cells. We demonstrate that levels of ribonucleotide reductase are high in colon carcinoma cells and low in hepatocytes and that hrR3 selectively replicates in cells with high levels of ribonucleotide reductase. However, even when administered into the portal vein, hrR3 that passes into the systemic circulation could replicate in actively dividing cell populations outside of the liver. We examined lung and brain sections and did not detect ß-galactosidase staining or cytopathic effects (data not shown). It remains to be determined whether hrR3 can be detected in other organs by more sensitive techniques such as polymerase chain reaction, in situ hybridization, or immunohistochemical staining for HSV1 proteins.
Although a single administration of hrR3 produced significant
anti-tumor activity against macroscopic liver metastases 7 days after
implantation, the tumor burden we treated was limited relative to that
observed in many clinical scenarios. Most patients with liver
metastases with whom this approach can be examined have larger tumor
burdens. Because HSV1 replication is several orders of magnitude more
robust in human cancer cells than in murine cancer cells, HSV1
oncolytic therapy may be more effective in patients than in mice
despite their higher liver tumor burdens. Also, multiple dose
administrations may enhance efficacy in patients. Although antibodies
to HSV1 are relatively common in humans, our data suggest that this may
not adversely affect therapeutic efficacy. Finally, colon carcinoma
liver metastases in humans are supplied principally by the hepatic
artery, whereas normal liver parenchyma is supplied principally by the
portal vein (41)
. Accordingly, administration of hrR3 into
the hepatic artery would be expected to result in higher levels of HSV1
replication in liver metastases compared to administration into the
portal vein. Though this can be readily accomplished in patients, it is
not possible in BALB/c mice.
HSV1 thymidine kinase is expressed during viral replication, which
enhances cellular susceptibility to ganciclovir (42)
. In
the present study, we have focused our investigation on the
anti-neoplastic activity specifically attributable to HSV1 replication
and did not examine the effects of ganciclovir administration. We
previously demonstrated that exposure of tumor cells to ganciclovir
after administration of hrR3 enhances tumor cell destruction in some
cell lines, but not in others (17
, 43)
.
We have demonstrated the therapeutic potential of hrR3 against liver
metastases; however, we anticipate that modifications to address three
issues will enhance the safety and efficacy of HSV1-based oncolytic
therapy. First, ICP6 in hrR3 is inactivated solely by an insertion of
the ß-galactosidase gene (15)
. Because expulsion of this
inserted gene could reconstitute a wild-type virus, we recently
constructed another HSV mutant with a significant portion of ICP6
deleted, which makes it a safer virus to examine in clinical trials
(44)
. Second, hrR3 is most effective against dividing
cells; accordingly, G0 cancer cells may be less
susceptible to productive hrR3 infection. Third, actively dividing cell
populations in the body other than tumor cells may theoretically be
susceptible to hrR3. All three of these issues can be addressed, for
example, by construction of an HSV1 vector in which a viral gene
required for replication (e.g., ICP4) is regulated by a promoter for a
tumor-associated antigen such as carcinoembryonic antigen (CEA). This
type of HSV1 vector would not be as susceptible to reconstitution of
wild-type virus, could infect nondividing cancer cells, and should not
replicate in cells that do not express CEA. Unlike intralesion
inoculation, intravascular administration of hrR3 targets widespread
disease; therefore, HSV1 mutants engineered to restrict productive
infection to tumor cells hold promise as cancer therapeutic
agents.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  K. K. Tanabe, A. Lemoine, D. M. Finkelstein, H. Kawasaki, T. Fujii, R. T. Chung, G. Y. Lauwers, Y. Kulu, A. Muzikansky, D. Kuruppu, et al. Epidermal Growth Factor Gene Functional Polymorphism and the Risk of Hepatocellular Carcinoma in Patients With Cirrhosis JAMA, January 2, 2008; 299(1): 53 - 60. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Kuruppu, A.-L. Brownell, A. Zhu, M. Yu, X. Wang, Y. Kulu, B. C. Fuchs, H. Kawasaki, and K. K. Tanabe Positron Emission Tomography of Herpes Simplex Virus 1 Oncolysis Cancer Res., April 1, 2007; 67(7): 3295 - 3300. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Cross and J. K. Burmester Gene therapy for cancer treatment: past, present and future. Clin. Med. Res., September 1, 2006; 4(3): 218 - 227. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. Kasuya, D. K. Kuruppu, J. M. Donahue, E. W. Choi, H. Kawasaki, and K. K. Tanabe Mouse Models of Subcutaneous Spleen Reservoir for Multiple Portal Venous Injections to Treat Liver Malignancies Cancer Res., May 1, 2005; 65(9): 3823 - 3827. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. F. Stanziale, H. Petrowsky, P. S. Adusumilli, L. Ben-Porat, M. Gonen, and Y. Fong Infection with Oncolytic Herpes Simplex Virus-1 Induces Apoptosis in Neighboring Human Cancer Cells: A Potential Target to Increase Anticancer Activity Clin. Cancer Res., May 1, 2004; 10(9): 3225 - 3232. [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]
|
|
|
|
|
|
R. J. Wong, M.-K. Chan, Z. Yu, T. H. Kim, A. Bhargava, B. M. Stiles, B. C. Horsburgh, J. P. Shah, R. A. Ghossein, B. Singh, et al. Effective Intravenous Therapy of Murine Pulmonary Metastases with an Oncolytic Herpes Virus Expressing Interleukin 12 Clin. Cancer Res., January 1, 2004; 10(1): 251 - 259. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. T. Mullen and K. K. Tanabe Viral Oncolysis for Malignant Liver Tumors Ann. Surg. Oncol., July 1, 2003; 10(6): 596 - 605. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. H Palmer, M.-J. Chen, and D. J Kerr Gene therapy for colorectal cancer Br. Med. Bull., December 1, 2002; 64(1): 201 - 225. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
X. Fu and X. Zhang Potent Systemic Antitumor Activity from an Oncolytic Herpes Simplex Virus of Syncytial Phenotype Cancer Res., April 1, 2002; 62(8): 2306 - 2312. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. T. Mullen and K. K. Tanabe Viral Oncolysis Oncologist, April 1, 2002; 7(2): 106 - 119. |
