Download Cationic Lipid:Bacterial DNA Complexes Elicit

[CANCER RESEARCH 60, 2955–2963, June 1, 2000]
Cationic Lipid:Bacterial DNA Complexes Elicit Adaptive Cellular Immunity in
Murine Intraperitoneal Tumor Models1
Michael Lanuti, Samantha Rudginsky, Seth D. Force, Eric S. Lambright, William M. Siders, Michael Y. Chang,
Kunjlata M. Amin, Larry R. Kaiser, Ronald K. Scheule, and Steven M. Albelda2
Department of Surgery, Thoracic Oncology Research Laboratory [M. L., S. D. F., E. S. L., M. Y. C., K. M. A., L. R. K.], and Department of Medicine, Pulmonary/Critical Care
[S. M. A.], University of Pennsylvania Medical Center, Philadelphia, Pennsylvania, 19104; Genzyme Corporation, Framingham, Massachusetts 01701 [S. R., W. M. S., R. K. S.]
ABSTRACT
Previous studies with a mycobacterial heat shock protein (hsp-65) have
demonstrated some efficacy using cationic liposome-mediated gene transfer in murine i.p. sarcoma models. To further analyze the efficacy of
hsp-65 immunotherapy in clinically relevant models of localized cancer,
immunocompetent mice bearing i.p. murine mesothelioma were treated
with four i.p. doses of a cationic lipid complexed with plasmid DNA
(pDNA) containing hsp65, LacZ, or a null plasmid. We observed >90%
long-term survival (median survival, 150 days versus ⬃25 days, treated
versus saline control, respectively) in a syngeneic, i.p. murine mesothelioma model (AC29). Long-term survivors were observed in all groups
treated with lipid complexed with any pDNA. Lipid alone or DNA alone
provided no demonstrable survival advantage. In a more aggressive i.p.
model of mesothelioma (AB12), we observed >40% long-term survival in
groups treated with lipid:pDNA complexes, again irrespective of the
transgene. To ask whether these antitumor effects had led to an adaptive
immune response against the tumor cell, we rechallenged long-term survivors in both murine models s.c. with the parental tumor cell line.
Specific, long-lasting systemic immunity against the tumor was readily
demonstrated in both models (AB12 and AC29). Consistent with these
results, splenocytes from long-term survivors specifically lysed the parental tumor cell lines. Depleting the CD8ⴙ T-cells from the splenocyte pool
eliminated this lytic activity. Lipid:pDNA treatment of athymic, SCID,
and SCID/Beige mice bearing a murine i.p. mesothelioma (AC29) resulted
in only a slight survival advantage, but there were no long-term survivors.
Treatment of immunocompetent mice depleted of specific immune effector cells demonstrated roles for CD8ⴙ and natural killer cells. Although
the exact mechanism(s) responsible for these antitumor effects is unclear,
the results are consistent with roles for both innate and adaptive immune
responses. An initial tumor cell killing stimulated by cationic lipid:pDNA
complexes appears to be translated into long-term, systemic immunity
against the tumor cell. These results are the first to demonstrate that
adaptive immunity against a tumor cell can be induced by the administration of lipid:pDNA complexes. Multiple administrations of cationic
lipid complexed with pDNA lacking an expressed transgene could provide
a promising generalized immune-mediated modality for treating cancer.
INTRODUCTION
One of the major obstacles to successful cancer gene therapy has
been the inability to deliver therapeutic transgenes effectively to
the majority of tumor cells. For this reason, gene transfer approaches that generate antitumor immune responses are especially
appealing, and many strategies have been developed using both
viral and nonviral vectors (1) in which tumor cells have been
transfected with various immunostimulatory transgenes, such as
cytokines, MHC class I molecules, and co-stimulatory molecules.
Received 10/22/99; accepted 3/28/00.
The costs of publication of this article were defrayed in part by the payment of page
charges. This article must therefore be hereby marked advertisement in accordance with
18 U.S.C. Section 1734 solely to indicate this fact.
1
This research was supported primarily by a sponsored research grant from StressGen/
Genzyme LLC. Additional support was provided by the Samuel H. Lunenfeld Charitable
Foundation and the Benjamin Shein Foundation for Humanity.
2
To whom requests for reprints should be addressed, at 856 BRBII/III, 421 Curie
Boulevard, University of Pennsylvania Medical Center, Philadelphia, PA 19104. Phone:
(215) 573-9933; Fax: (215) 573-4469.
One principle that seems to be emerging from these studies is that
repetitive immunization protocols will be needed to induce and
maintain cytotoxic T-cell responses (2). Because the administration of viral vectors has been shown to induce powerful antiviral
immune responses that impair repeat infection and expression
(3– 6), non-viral-based modalities for cancer gene therapy, such as
cationic lipid-mediated gene transfer, have particular promise in
the immunotherapy context. Potential advantages of nonviral vectors in immunotherapy applications include the simplicity of the
approach, the inherent safety of delivering a nonreplicating plasmid vector, and most importantly, the feasibility of repeatedly
delivering genes to tumors that might elicit the “danger signals”
needed to activate the immune system and lead to the destruction
of tumor cells (7).
Previous cancer gene therapy studies using cationic liposome-mediated gene transfer have examined the efficacy of specific transgenes,
including herpes simplex virus thymidine kinase (HSVtk) in pancreatic
cancer (8) and endothelial cells (9), tumor necrosis factor-␣ in a murine
model of disseminated breast cancer (10), and allogeneic MHC class I
molecules in melanoma (11, 12). Although several such studies have
demonstrated limited in vivo efficacy, one particularly promising approach used cationic lipid-mediated gene transfer of a mycobacteriumderived heat shock protein (hsp-65) in a murine i.p. sarcoma model
(13–15). In related experiments, Wells et al. (16) demonstrated that mice
immunized i.p. with B16 cells expressing hsp-65 displayed significant
resistance to a subsequent challenge with the parental B16 cells. The
underlying immune mechanisms responsible for the antitumor effects
observed with lipid:pDNA3 therapy, and more specifically with the use of
the Mycobacterium leprae-derived hsp65 gene, remain unclear. Heat
shock proteins can function as “molecular chaperones” and could potentially chaperone small antigenic peptides to MHC molecules for more
efficient antigen presentation (17–20). In addition, bacterial and mycobacterial heat shock proteins are themselves extremely immunogenic
molecules (21) and may provide an inflammatory milieu in the tumor cell
environment.
To evaluate further the utility of hsp-65 immunotherapy, we studied
the antitumor effects of cationic liposomes complexed with pDNA containing hsp65 delivered i.p. in a clinically relevant, localized model of
cancer, namely, i.p. mesothelioma. We hypothesized that cationic lipid:
pDNA complexes containing the hsp65 gene would elicit specific therapeutic efficacy compared with lipid:pDNA complexes containing the
cDNA for either a nontherapeutic bacterial protein, e.g., ␤Gal, or a
plasmid carrying no transgene, i.e., a null plasmid. Accordingly, immunocompetent mice bearing macroscopic i.p. murine mesothelioma were
treated with multiple i.p. doses of cationic lipid complexed with pDNA.
Animals were examined for survival and long-term antitumor immunity.
As postulated, we saw striking increases in survival as well as the
induction of long-term adaptive immune responses in animals treated
3
The abbreviations used are: pDNA, plasmid DNA; ␤Gal, ␤-galactosidase;
DOPE, dioleoylphosphatidylethanolamine; DMPE-PEG5000, dimyristoylphosphatidylethanolamine-polyethylene glycol 5000; DC-Chol, 3␤[N-(N⬘, N⬘⬘-dimethylaminoethane)carbamoyl]cholesterol; LLC, Lewis lung carcinoma; NK, natural killer; SCID, severe combined immunodeficient; IL, interleukin.
2955
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 2000 American Association for Cancer
Research.
CATIONIC LIPID:pDNA COMPLEXES IN CANCER THERAPY
lanta, GA), 100 units/ml penicillin G, 100 ␮g/ml streptomycin, and 2 mM
glutamine. The RENCA cells were murine (BALB/c) renal adenocarcinoma
cells of spontaneous origin obtained as a gift from Dr. Kenneth Cowan
(Medical Branch, Division of Clinical Sciences, National Cancer Institute,
Bethesda, MD). Line 1 was a murine (BALB/c) bronchoalveolar carcinoma
obtained as a gift from Dr. Steve Dubinett (UCLA, Los Angeles, CA). LLC
cells originated as a carcinoma in the lung of a C57BL/6 mouse (27) and were
purchased from the American Type Culture Collection. The murine lymphoma
cell line YAC-1 (ATCC TIB-160), which has low levels of MHC class I and
is constitutively sensitive to NK-cell-mediated lysis, has been described elsewhere (28, 29). RENCA, Line 1, LLC, and YAC-1 cells were cultured in RPMI
1640 with 10% FCS, 100 units/ml penicillin G, 100 ␮g/ml streptomycin, and
2 mM glutamine.
Mice. Female BALB/c, female CB17-SCID, and male CB17-SCID/Beige
MATERIALS AND METHODS
mice (6 – 8 weeks old; weight, ⬃25 g) were obtained from Taconic Laboratory
Cationic Lipids. Genzyme cationic lipid no. 67 (GL-67; N4-spermine (Germantown, NY). Male CBA/J mice (7– 8 weeks old; weight, ⬃25 g) were
cholesterylcarbamate) was prepared by small molecule synthesis as described obtained from the National Cancer Institute (Frederick Cancer Research &
by Lee et al. (22). The neutral colipid DOPE and the stabilizing lipid DMPE- Development Center, Frederick, MD). Homozygous NCR nude mice (5– 8
PEG5000 were obtained from Avanti Polar Lipids (Alabaster, AL). Solutions of weeks old; weight, ⬃25 g) were also obtained from Taconic Laboratory. These
GL-67, DOPE, and DMPE-PEG5000 were coformulated and lyophilized from athymic mice originally were derived from a BALB/c background. Animals
t-butanol-water (90:10, v/v) at a molar ratio shown previously to be effective, were housed either in the animal facility at the Wistar Institute (Philadelphia,
namely (GL-67:DOPE:DMPE-PEG5000, 1:2:0.05, mol/mol/mol; Ref. 23). PA) or at Genzyme. The Animal Use Committees of the Wistar Institute and
University of Pennsylvania or Genzyme approved all protocols in compliance
Prior to use, the lyophilized lipids were hydrated for 10 min with sterile water
at 4°C and then vortexed for 2 min to generate liposomes. All pDNA was with the Guide for the Care and Use of Laboratory Animals.
i.p. Murine Mesothelioma Models. Two independent intracavitary tumor
diluted in sterile water. After both lipid and pDNA were allowed to equilibrate
at 30°C for 5 min, cationic lipid:pDNA complexes were prepared by mixing models of murine mesothelioma (AC29 and AB12) were established to examthe liposomes and pDNA in a ratio of 1:4 (0.3 mM cationic lipid:1.2 mM ine the treatment efficacy of locally administered liposome-pDNA complexes.
DNA).4 The complexes then were incubated at 30°C for 15 min before dilution Murine mesothelioma (AB12 or AC29) cells were grown to 80% confluence in
1:1 (v/v) with 1.8% normal sterile saline (pH 7.0). Complexes were adminis- culture flasks, the medium was aspirated, and cells were harvested using 0.05%
tered in vivo within 15–30 min of preparation. Formulations of another cationic trypsin-Versene (Life Technologies; Grand Island, NY). Cells were then
6
lipid, DC-Chol (Avanti Polar Lipids, Alabaster, AL), and DOPE were prepared pelleted (1000 rpm, 3 min) and resuspended at 1 ⫻ 10 cells/ml in serum-free
DMEM
for
i.p.
administration.
Cell
concentrations
were
determined by countby solubilizing the lipids in CHCl3 and then drying them down to a thin film
under a stream of nitrogen as described previously (24). Residual CHCl3 was ing aliquots of the cell suspensions in a Coulter counter (Miami, FL). AB12
5
removed under high vacuum. The dried lipid film was rehydrated in sterile and AC29 cells (5 ⫻ 10 /0.5 ml) were injected i.p. into BALB/c and CBA/J
water at 4°C overnight and then sonicated for 10 min at 30°C in a bath mice, respectively, using a 26-gauge needle. Approximately 5– 8 days after
sonicator. DC-Chol was added to pDNA in a 1.6:1 ratio (1.6 mM DC-Chol:1 tumor cell injection, macroscopic (0.5–1 mm) tumor nodules could be identified on the small bowel mesentery. Later, tumor could be observed on the
mM pDNA), as described previously (24).
Plasmids. All plasmids were constructed using a backbone described diaphragm, peritoneal surface, porta hepatis, lesser sac, and retroperitoneum.
Survival studies were performed in lieu of tumor burden assessment because
previously (22, 25). The 5.6-kb pCMVhsp65 and pCMV␤Gal plasmids
contain the cDNA from the M. leprae-derived hsp65 and Escherichia of the difficulty in harvesting tumor densely adherent to the abdominal viscera.
coli-derived LacZ, respectively. The null plasmid (pNull) is identical to The GL-67:pDNA complexes were combined in a 1:4 mM ratio using 50 –100
␮g pDNA, a dose previously found to be efficacious in i.p. murine sarcoma
these plasmids except that it has no transgene cDNA, i.e., it contains the
cytomegalovirus promoter, hybrid intron, bovine growth hormone poly(A) models (13). Treatments were initiated when ⬃1 mm tumor nodules were
identified post tumor cell inoculation. Mice received four i.p. administrations
sequences, and the kanamycin resistance gene (22, 25).
Plasmid DNA was prepared by bacterial fermentation and purified by of cationic lipid:pDNA complexes at 4 day intervals.
Following treatment, long-term survivors were anesthetized (xylazine/ketultrafiltration and sequential column chromatography. Fermentation was performed in a 15-liter Chemap fermenter at 37°C for 24 h using HCD media amine; A. J. Buck & Son, Inc, Baltimore, MD) and challenged s.c. in the flank
6
(Genzyme proprietary media) containing 100 ␮g/ml kanamycin. The purified with 5 ⫻ 10 cells of the parental tumor cell line (AB12 or AC29) in 150 ␮l
preparations were predominantly supercoiled; exhibited spectrophotometric of serum-free DMEM. An additional murine bronchoalveolar lung cancer cell
A260 nm/A280 nm ratios of 1.75–2.0; were free of detectable RNA; contained line, Line 1 (syngeneic in BALB/c mice), was also used to challenge the
⬍10 ␮g of protein, 10 ␮g of chromosomal DNA, and 5 endotoxin units/mg of long-term survivors in the AB12 model to evaluate the specificity of any
plasmid DNA (LAL assay; BioWhittaker); and were free of colony-forming antitumor response. Approximately 2 weeks after the s.c. challenge, flank
tumors were excised and weighed; the animals were maintained to permit
units in a bioburden assay.
Expression of the hsp-65 gene product was confirmed in vitro by GL-67- continuation of the survival study.
Evaluation of T-Cell Responses to Lipid:pDNA Therapy. Adaptive celmediated transfection of multiple cell lines (data not shown), followed by
lular immune responses to lipid:pDNA treatment were evaluated using a CTL
immunohistochemical detection using a monoclonal antibody specific for
hsp-65 (4D6; StressGen Biotechnologies, Victoria, BC, Canada). Expression assay. Pooled spleen cells (two to four mice per group) isolated from mice
several months (see figure legends for specific time points) after i.p. treatment
from the pCMV␤Gal plasmid has been demonstrated previously (22, 25).
Cell Lines. AB12 and AC29 were murine mesothelioma cell lines (pro- with lipid:DNA complexes were cocultured in a volume of 2 ml in 24-well
6
vided by Dr. Jay K. Kolls, LSU School of Medicine, New Orleans, LA). These plates (5 ⫻ 10 cells/well in RPMI plus 10% FCS, 2-mercaptoethanol, and
5
cell lines were originally generated by Dr. Bruce Robinson (Queen Elizabeth HEPES) with stimulator cells, i.e., 2 ⫻ 10 AB12 or AC29 cells treated with
100
␮
g/ml
mitomycin
C
for
30
min
at
37°C.
Plates were incubated at 37°C in
II Medical Center, Perth, Australia) by i.p. implantation of asbestos fibers in
BALB/c and CBA/J mice, respectively, and have been well characterized (26). a 5% CO2 atmosphere for 6 days to expand tumor-specific CTLs. Effector cells
⫹
These cells were cultured and maintained in high glucose DMEM (Mediatech, were then harvested and pooled for further processing. To deplete CD8
Washington, DC) supplemented with 10% FCS (Georgia Biotechnology, At- T-cells from the expanded splenocyte population, aliquots of effector cells
were incubated for 20 min at 4°C with a CD8⫹-specific murine antibody
(LYT2; Dynal, Oslo, Norway) linked to magnetic beads. Beads and bound
4
For simplicity, complexes are referred to as GL-67:pDNA or cationic lipid:pDNA
CD8⫹ T-cells were removed from the incubation medium with a magnet.
complexes, although the liposomes contain DOPE and DMPE-PEG5000 in addition to
To measure cytotoxicity, untreated and CD8⫹-depleted effector cells were
GL-67.
2956
with cationic lipid:phsp65 complexes. Surprisingly, however, we also
found that equivalent responses could be obtained with our “control”
lipid:pDNA complexes. These experiments demonstrated that repeated
i.p. delivery of lipid:pDNA complexes, regardless of the transgene used,
induced powerful antitumor responses that included total elimination of
tumor burden for a significant proportion of animals and the development
of long-term, antitumor immunity in these animals concomitant with the
induction of tumor cell-specific cytotoxic CD8⫹ T-lymphocytes. Delivery of lipid:pDNA complexes thus could be a promising modality for
treating localized malignancies such as malignant mesothelioma, ovarian
cancer, brain tumors, or bladder cancer.
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 2000 American Association for Cancer
Research.
CATIONIC LIPID:pDNA COMPLEXES IN CANCER THERAPY
incubated with 5 ⫻ 103 51Cr-labeled target cells/well in triplicate for 5 h in a
V-bottomed 96-well microtiter plate. Target cells were incubated with murine
␥-IFN (100 units/ml; R&D Systems, Minneapolis, MN) for 24 h to up-regulate
MHC class I expression prior to 51Cr-labeling. Target cells (AB12, AC29,
RENCA, LLC, Line 1, or YAC-1) were labeled with 51Cr by incubating
3 ⫻ 105 cells in a 6-well plate with 3.7 MBq of 51Cr (New England Nuclear)
in normal growth medium for 18 h, and were then washed three times with
PBS. Varying numbers of effector cells were mixed with target cells to yield
E:T ratios of 60:1 to 7.5:1. 51Cr release was measured in wells containing
effector T cells and target cells (cpmtest), wells containing target cells in
medium alone (cpmspontaneous), and in wells containing target cells plus 1%
Triton X-100 (cpmtotal). The percentage of lysis was calculated using the
formula: 100 ⫻ (cpmtest ⫺ cpmspontaneous)/(cpmtotal ⫺ cpmspontaneous).
Antibody Depletion of Immune Effector Cells. To deplete specific immune effector cell subsets prior to treatment with lipid:pDNA complexes in the
AB12 model, BALB/c mice received i.v. injections of 100 ␮g of an isotype
control antibody (clone R35–95; PharMingen) or an antibody to deplete the
CD4⫹ (clone GK1.5; PharMingen), CD8⫹ (clone 53-6.7; PharMingen), or NK
cell populations (antiasialo GM1; Wako Bioproducts, Richmond, VA). Dose
levels were based on manufacturer’s recommendations for ⬎90% depletion of
the specific cell type. Depletion of each cell type to at least these levels was
demonstrated at these doses (data not shown). Injections were administered for
3 consecutive days prior to inoculation with AB12 cells. Thereafter, a maintenance dose of antibody was injected i.v. every 3– 4 days throughout the entire
(⬃16 day) treatment period to ensure depletion of the targeted cell type. Mice
received four i.p. injections of lipid:pNull complex or saline as described
above. At day 13 post cell (AB12) inoculation, one mouse from each group
was sacrificed, and the splenocytes were harvested for fluorescence-activated
cell-sorting analysis to quantify depletion of the specific immune effector cells;
each targeted cell type was specifically and significantly depleted by this
treatment (data not shown). The remaining mice in each group were followed
for survival to evaluate the effects of depleting these immune cell types.
Statistics. Differences in tumor weights in the animal experiments were
determined by one-way ANOVA. Post hoc comparisons of specific paired
groups were performed using Fisher’s analysis. Statistical significance was set
at P ⬍ 0.05. Kaplan-Meier survival curves were analyzed with the Mantel-Cox
Log-rank test. Results are expressed as the mean ⫾ SE.
RESULTS
The therapeutic effects of cationic lipid:pDNA treatment were
evaluated in two i.p. mesothelioma models, AC29 and AB12. The
mechanisms underlying these effects were probed as a function of
time, using several methods. The functional involvement of different
immune cells during treatment with complex was evaluated using
immunodeficient mouse strains and by depleting specific immune
cells in normal animals prior to treatment. Treatment-generated longterm memory responses were characterized for the possible involvement of NK cells and cytotoxic lymphocytes by isolating splenocytes
at times well after treatment and evaluating their ability to kill tumor
cells in vitro.
Treatment of i.p. Mesothelioma in the AC29 Tumor Model. To
assess the antitumor effects of cationic lipid:pDNA treatment, we first
studied AC29 murine malignant mesothelioma cells growing within
the peritoneal cavity of an immunocompetent host. Fig. 1A depicts
Kaplan-Meier cumulative survival in groups of CBA/J mice bearing
syngeneic i.p. mesothelioma (AC29) that received four i.p. administrations of (a) saline, (b) cationic liposomes (GL-674) alone, (c)
phsp65 alone, (d) GL-67 complexed with phsp65, or (e) GL-67
complexed with p␤Gal. The p␤Gal vector was used to delineate
specificity of the hsp65 gene product and contained the same prokaryotic plasmid backbone as the phsp65 vector. Treatments were
initiated on day 8 when macroscopic tumor nodules ⬃1 mm in size
were identified; subsequent doses were delivered at 4-day intervals.
Animals treated with saline had a median survival of 28 days; all
mice were dead by day 35. Administration of GL-67 alone or phsp65
Fig. 1. Kaplan-Meier survival curves in the AC29 model. Female CBA/J mice
(n ⫽ 10/group) bearing AC29 tumor received four i.p. doses at 4-day intervals (arrows)
of cationic lipid complexed with 50 ␮g of pDNA, lipid alone (0.3 mM), or pDNA alone
(50 ␮g). A, comparison of treatment with GL-67 alone (F), phsp65 alone (‚), and
GL-67:pDNA complexes containing the hsp65 (Œ) or ␤Gal (䡺) transgenes. B, comparison of treatment with GL-67:pDNA complexes containing either a null (‚) or hsp65 (F)
plasmid, and DC-Chol:phsp65 (E) complexes. Long-term survivors in B were challenged
with AC29 cells s.c. as described in “Material and Methods” (arrowhead). Control mice
received 0.9% NaCl (u) at each dose interval. Significant long-term survival (⬎90%) was
observed only in treated mice receiving lipid:pDNA complexes, namely, GL-67:phsp65
(A and B), GL-67:p␤Gal (A), and DC-Chol:phsp65 (B).
alone had no effect on median survival and produced no long-term
survivors. In contrast, administration of the combination of cationic
lipid and pDNA (either phsp65 or p␤Gal) resulted in a marked
survival benefit, with all of the animals alive 84 days into the study
(P ⬍ 0.0001). These treated animals also demonstrated no i.p. tumors
when sacrificed on day 94. Of note was the observation that lipid
complexed with pDNA containing either the bacterially derived LacZ
transgene (p␤Gal) or the mycobacterial hsp65 transgene were equally
effective in producing long-term, tumor-free survivors, suggesting the
lack of any transgene-specific effect.
A second survival experiment in the AC29 model was designed to
determine (a) whether the observed antitumor effects were specific to
the GL-67 liposome, and (b) whether a similar pDNA containing no
transgene at all (pNull) would have the same effect. Tumor-bearing
animals were treated with (a) saline, (b) GL-67 complexed with
pNull, (c) GL-67 complexed with phsp65, and (d) DC-Chol liposomes
complexed with phsp65 using the same dosing schedule as described
above. In this experiment, the cationic liposome GL-67 was directly
compared with DC-Chol, a commercially available cationic lipid. As
shown in Fig. 1B, the median survival of the saline-treated group was
again ⬃30 days. In contrast, all of the groups treated with liposomes
(either GL-67 or DC-Chol) plus pDNA (either phsp65 or pNull) had
marked and highly significant increases in survival (⬎90% long-term
survival at 150 days; P ⬍ 0.0001). These data confirmed that in this
localized tumor model, treating with liposomal complexes of phsp65
or pNull had marked therapeutic effects. Furthermore, these effects
were neither lipid- nor plasmid-specific.
2957
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 2000 American Association for Cancer
Research.
CATIONIC LIPID:pDNA COMPLEXES IN CANCER THERAPY
Fig. 2. Lipid:pDNA induction of specific, systemic antitumor immunity in the AB12
model. A, female BALB/c mice (n ⫽ 10/group) bearing i.p. AB12 tumors received four
i.p. doses at 4-day intervals (arrows) of cationic lipid (GL-67) complexed with 50 ␮g
pDNA containing phsp65 (‚) or pNull (E), phsp65 alone (〫; 100 ␮g), or saline (f).
Kaplan-Meier survival curves demonstrated significant (P ⬍ 0.0001) long-term survivors
(30 – 40%) in groups receiving lipid:phsp65 or lipid:pNull; no statistical significance
(P ⫽ 0.3776) was observed between the two survival curves. No survival benefit resulted
from phsp65 alone. B, surviving mice were challenged s.c. with AB12 cells (day 90;
arrowhead in A) and then with syngeneic Line 1 cells (day 119; arrowhead in A). Tumors
resulting from these challenges were harvested and weighed as described in “Materials
and Methods.” In mice treated with lipid:pDNA complexes, Line 1 tumors (hatched
columns) grew, but AB12 tumors (filled columns) did not, demonstrating the existence of
specific, long-term antitumor immunity. Bars, SE.
To determine whether a protective antitumor immune response had
been generated as a result of treatment, the long-term surviving
animals were rechallenged on day 115 (Fig. 1B, arrowhead) with a
s.c. injection of 5 ⫻ 106 AC29 cells into each flank. Tumor growth in
these long-term survivors was compared with that in naive (untreated)
CBA/J mice injected with the same number of tumor cells. After 18
days, all untreated mice had large flank tumors, whereas none of the
previously treated mice had detectable flank tumors. This result was
independent of the cationic lipid or pDNA used for treatment. Thus,
in the AC29 model, animals treated with lipid:pDNA complexes were
both “cured” of their disease and exhibited long-term protection
against a distal challenge by these same tumor cells.
Treatment of i.p. Mesothelioma in the AB12 Tumor Model. To
determine whether lipid:pDNA complexes could provide therapeutic
effects in another intracavitary tumor model in a different genetic
background, we treated immunocompetent BALB/c mice bearing a
second syngeneic i.p. murine mesothelioma, namely AB12. Animals
received four i.p. administrations of cationic liposome (GL-67) complexed with 100 ␮g of pDNA (either phsp65 or pNull) every 3 days
starting at day 5. Although the long-term survival of treated animals
in this model (Fig. 2A) was somewhat less impressive than that seen
in the AC29 model, it should be noted that AB12 tumors are more
aggressive and resistant to most chemotherapeutic and gene therapy
treatment modalities.5 Nonetheless, a significant percentage (30 –
50%) of animals survived long term (150 days) when mice were
treated with GL-67 complexed with either phsp65 or pNull; there
5
Unpublished results.
were no long-term survivors in the control, saline-treated group. There
was no statistically significant difference (P ⫽ 0.3776) between the
survival curves using phsp65 or pNull complexed with lipid. Untreated animals reproducibly died by 25–30 days (not shown) as did
the saline control group. Plasmid DNA alone (phsp65) offered a small
survival benefit compared with the saline control, but this difference
was not statistically significant.
To determine whether the long-term survivors (see Fig. 2A) had
developed specific antitumor immunity, they were challenged s.c. in
the flank on day 90 with the parental tumor AB12. On day 113, these
tumors were excised and weighed. On day 119, these same animals
were challenged s.c. with a control, syngeneic tumor cell line, Line 1
(a murine bronchoalveolar lung cancer). For each tumor cell challenge, naive mice received injections of with these same tumor cell
lines as controls. Fig. 2B shows that compared with the mass of AB12
tumor harvested from the flanks of control mice (0.27 ⫾ 0.08 g), there
was essentially no tumor growth (P ⬍ 0.01) in those mice that had
been treated previously with GL-67:phsp65 (0.02 ⫾ 0.01 g) or GL67:pNull (0.02 ⫾ 0.01 g). In contrast, Fig. 2B also shows that the flank
challenge with Line 1 cells yielded equivalent-sized tumor nodules on
the flanks of all mice tested, implying that specific and systemic
antitumor immunity to AB12 had developed as a result of the lipid:
pDNA treatment.
Lipid:pDNA Dose Response in the AB12 Tumor Model. We
examined the dose response to lipid:pDNA treatment in the AB12
model. In a treatment scheme identical to the previous AB12 experiments (see above), BALB/c mice bearing i.p. AB12 tumors were
treated with four doses of lipid complexed with p␤Gal at a GL-67:
p␤Gal ratio of 1:4 (mol/mol) and at increasing pDNA dose, i.e., 0, 10,
25, and 100 ␮g. Fig. 3 demonstrates that no therapeutic effect was
demonstrable at a dose of 10 ␮g pDNA, but that a 40% long-term
survival rate could be achieved at a dose of 100 ␮g pDNA
(P ⫽ 0.0002). Lipid complexed with 25 ␮g of p␤Gal showed an
intermediate survival advantage (20% at 70 days) that was significant
compared with saline controls (P ⫽ 0.0077) but was not significant
when compared with the 100-␮g group (P ⫽ 0.2928). A similar dose
response to lipid:pDNA complexes was observed in this model when
the phsp65 vector was used (data not shown). Thus, these data
demonstrate that i.p. AB12 tumor cells could be eliminated with
lipid:pDNA complex treatment in a dose-dependent fashion, and that
this elimination was independent of the transgene.
Fig. 3. Dose response to lipid:pDNA complexes administered i.p. in BALB/c mice
(n ⫽ 10/group) bearing i.p. murine mesothelioma (AB12). Mice received i.p. administrations of GL-67 complexed with p␤Gal (1:4 ratio) on four separate occasions (arrows).
Plasmid DNA doses were 0, 10, 25, or 100 ␮g and were complexed with appropriately
increasing amounts of lipid to maintain a 1:4 ratio. No significant therapeutic effect was
observed with 10 ␮g p␤Gal (E) compared with a saline (f) control. Improved survival
was best demonstrated in mice that received lipid complexes containing 100 ␮g p␤Gal
(䡺; P ⫽ 0.0002), with 40% survival at 70 days.
2958
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 2000 American Association for Cancer
Research.
CATIONIC LIPID:pDNA COMPLEXES IN CANCER THERAPY
Fig. 4. Effects of lipid:pDNA treatment in immunodeficient mouse strains. GL-67:
pDNA complexes (50 ␮g of pDNA) were administered i.p. in four doses (arrows) to
athymic (A), SCID (B), and SCID/Beige (C) mice (n ⫽ 12/group) bearing murine
mesothelioma (AC29) as described in “Materials and Methods.” Kaplan-Meier survival
analysis demonstrated incremental improvements in median survival in athymic (⬃10
days; A), SCID (⬃10 days; B), and SCID/Beige (⬃5 days; C). Therapeutic effects were
similar in the SCID model (B) when lipid was complexed with pDNA containing no
transgene (pNull; ‚), hsp65 (E), or ␤Gal (䡺) transgenes.
Lipid:pDNA Therapy in Immunodeficient Murine Models
bearing AC29 Tumors. To further investigate the immune mechanisms responsible for the improved survival observed in animals
treated with lipid:pDNA, survival studies were performed in immunodeficient murine models. Four i.p. administrations of GL-67:phsp65
complexes or saline were evaluated in three different immunodeficient mouse strains bearing i.p. murine mesothelioma (AC29). Fig. 4
shows the Kaplan-Meier survival graphs in athymic mice (T-cell
deficient; Fig. 4A), SCID mice (T- and B-cell deficient; Fig. 4B), and
SCID/Beige mice (T-cell, B-cell, and NK-cell deficient; Fig. 4C)
resulting from treatment. Significant (P ⬍ 0.01) increases in survival
(⬃10-day improvement in median survival) were noted in both nude
(Fig. 4A) and SCID mice (Fig. 4B) treated with GL-67:phsp65 compared with mice treated with saline. Lipid complexed with p␤Gal or
pNull yielded almost identical survival in SCID mice (Fig. 4B),
providing evidence for a mechanism of tumor lysis unrelated to the
transgene expressed by the pDNA. SCID/Beige survival (Fig. 4C) was
also improved (P ⬍ 0.01) by liposomal therapy. The data from these
three models suggest that although T, B, and NK cells may be playing
a role in the killing of tumor cells during the 3- to 4-week treatment,
that additional mechanisms exist for tumor cell cytolysis that are to
some degree independent of these immune cells. Importantly, in
contrast to the results in an immunocompetent mouse (Fig. 1) there
were no long-term survivors in any of these immunodeficient strains.
Because a lack of T cells is common to all of these immunodeficient
strains, this observation argues that at least T cells are necessary for
long-term survival.
Effects of Immune Cell Depletion on Survival. To gain additional insight into the roles played by CD4⫹ and CD8⫹ T cells and
NK cells as a result of lipid:pDNA treatment, immunocompetent mice
were depleted of these specific immune effector cells before treatment. Fig. 5 shows that in the AB12 model, lipid:pDNA complex
treatment of CD8⫹- or NK-cell-depleted animals was significantly
less effective (all animals had died by day 13) than the same treatment
in fully immune competent BALB/c mice. In fact, these CD8⫹- or
NK-cell-depleted animals died before untreated tumor-bearing animals. These results argue that both CD8⫹ and NK cells play important
roles during the induction phase of treatment, leading to a slowing of
tumor growth, the elimination of tumor cells, or both. The potential
role played by CD4⫹ cells is somewhat less clear. Fig. 5 shows that
⬃50% of the CD4⫹-depleted animals survived long term. Although
there is no statistical difference between the CD4⫹-depleted group
and the treated group, the CD4⫹-depleted group may have survived
marginally better (P ⫽ 0.16, Log-rank) than the untreated animals.
Although the quantitative results of these immune cell depletion
experiments may not be directly comparable to the results obtained in
the immunodeficient mouse strains because they used different models, i.e., AB12 and AC29, respectively, the data are consistent with
CD8⫹ and NK cells being critical to the long-term survival of the
animals.
Evaluation of Cytotoxic Antitumor Cell Responses to Lipid:
pDNA Therapy in the AC29 Tumor Model. In the AC29 model,
challenging long-term survivors with a s.c. inoculation of AC29 cells
led to the elimination of those cells (see above). We used in vitro
cytotoxicity assays to ask whether CTLs or NK cells could have
played a role in tumor cell rejection. To determine whether tumorspecific CTLs had been generated as a consequence of treatment,
spleen cells were isolated from survivors in the experiment depicted in
Fig. 5. Effects of treatment in the AB12 model in BALB/c mice depleted of specific
immune cell types. Antibodies specific for CD8⫹, CD4⫹, and NK cells were used to
deplete these specific cell types before treatment with GL-67:pNull (arrows). In fully
immunocompetent animals, the untreated group (䡺) had a median survival of ⬃20 days,
whereas the treated group (f) had ⬎65% survivors at day 57. Depletion of NK (‚) or
CD8⫹ (F) cells led to decreases in median survival (median, 9 –14 days), whereas CD4⫹
(E) depletion resulted in a median survival similar to that of untreated animals.
2959
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 2000 American Association for Cancer
Research.
CATIONIC LIPID:pDNA COMPLEXES IN CANCER THERAPY
Fig. 6. In vitro cytotoxic activity of splenocytes from surviving CBA/J mice treated with i.p. lipid:pDNA complexes for
i.p. AC29 mesothelioma. Lipid:pDNA therapy was administered i.p. at 4-day intervals on four occasions (see Fig. 1).
Spleen cells (three spleens/group) were isolated at day 158 and
stimulated for 6 days in culture with mitomycin-treated AC29
as described in “Materials and Methods.” These stimulated
cells were tested for cytotoxic activity against AC29 (A) and
LLC (B) cells. Stimulated splenocytes from untreated (f)
CBA/J mice bearing AC29 tumors showed no cytotoxicity for
the AC29 or LLC targets. Stimulated splenocytes from mice
treated with cationic lipid:pDNA complexes [GL-67:pNull (E),
GL-67:phsp65 (F), or DC-Chol:phsp65 (‰)] demonstrated activity against AC29 cells (A) but not LLC (B) cells. Stimulated
splenocytes from naive CBA/J mice elicited no cytotoxicity for
either the AC29 or LLC targets (data not shown). Data shown
are representative of two studies.
Fig. 1B at day 158 and cultured in vitro for 6 days with mitomycintreated stimulator cells (AC29 cells). Spleen cells from naive and
untreated (but tumor-bearing) CBA/J mice were processed similarly.
Splenocytes were then tested for cytotoxicity against the parental
tumor (AC29), a control murine cell line, LLC, and YAC-1 cells as
targets. The murine lymphoma cell line, YAC-1, expresses virtually
no MHC class I and is sensitive to NK-cell-mediated lysis (30).
Allogeneic LLC cells were used because no other syngeneic murine
tumor cell line was available for CBA/J mice. Fig. 6A shows that
splenocytes isolated from long-term surviving mice treated with lipid:
pDNA complexes demonstrated significant cytotoxicity for AC29
cells, whereas splenocytes from untreated mice bearing i.p. AC29 for
at least 2 weeks did not. Fig. 6B shows that, as expected, no cytotoxicity could be demonstrated against the control LLC target cells using
splenocytes from treated or untreated mice. Stimulated spleen cells
from naive CBA/J mice, i.e., bearing no tumor, also showed no target
cell lysis (AC29 or LLC; data not shown). Splenocytes from treated
animals also exhibited significant lysis of YAC-1 cells (40% at an E:T
ratio of 60:1). Together, these data indicate that cationic lipid:pDNA
therapy produced long-term surviving animals that had generated
cytotoxic lymphocytes specific for the AC29 tumor cells, that this
induction was independent of the cationic lipid and plasmid used, and
that there were significant numbers of activated NK cells present at
this time point.
Evaluation of Cytotoxic Antitumor Cell Responses to Lipid:
pDNA Therapy in the AB12 Tumor Model. Because treatment with
lipid:pDNA complexes appeared to induce specific T-cell-mediated
antitumor immunity in the AC29 model, we also investigated the
relationship between long-term protection and T-cell-mediated cytotoxicity in the AB12 model. Spleen cells from long-term survivors
that had been treated with GL-67:phsp65 in an AB12 survival study
(not shown) were harvested at day 55 and cultured in vitro for 6 days
with mitomycin-treated AB12 cells. Splenocytes were then tested for
their ability to lyse the parental tumor cell (AB12) and two additional
syngeneic tumor cell lines, namely, RENCA and Line 1. Fig. 7A
demonstrates that splenocytes isolated from the long-term survivors
Fig. 7. In vitro cytotoxic activity of splenocytes from surviving BALB/c mice in the AB12 model. GL-67:phsp65 complexes were administered i.p. at 3-day intervals on four occasions as in Fig. 2. A, spleen cells (three spleens/group) were
isolated at day 91 and stimulated for 6 days in culture with
mitomycin-treated AB12 cells and then tested for their ability to
lyse the parental AB12 cell line (u) and two additional syngeneic targets, RENCA (F) and Line 1 (‚). Only AB12 cells were
lysed. Data shown are representative of four similar experiments. B, depletion of CD8⫹ T-cells from the stimulated
splenocytes (see “Materials and Methods”) resulted in the complete elimination of cytotoxic activity against the parental
AB12 cells (u). Stimulated splenocytes from naive BALB/c
mice and from mice bearing i.p. AB12 without treatment
showed no cytotoxicity for any of the targets AB12, RENCA,
or Line 1 (data not shown).
2960
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 2000 American Association for Cancer
Research.
CATIONIC LIPID:pDNA COMPLEXES IN CANCER THERAPY
Fig. 8. Effect of inhibiting NK cells in the stimulated splenocyte
pool from long-term survivors in the AB12 model. GL-67:pDNA
complexes were administered i.p. at 3-day intervals on four occasions
as in Fig. 2. Spleen cells (three spleens/group) from long-term survivors treated with GL-67:phsp65 (‚) or GL-67:p␤gal (E), or from
untreated, tumor-bearing mice (〫), or naive mice (f) were isolated
at day 55 and stimulated for 6 days in culture with mitomycin-treated
AB12 and then tested for their ability to lyse the parental AB12 cell
line in the absence (A) or presence (B) of YAC-1 cells. The presence
of YAC-1 cells had virtually no effect on the lysis of target AB12
cells by the stimulated splenocytes, implying a minimal contribution
to direct tumor cell killing by NK cells at this time point. Data shown
in A are representative of four similar experiments.
exhibited a significant ability to lyse AB12 cells but showed virtually
no cytotoxicity against the syngeneic control target cells, RENCA and
Line 1. Specifically depleting CD8⫹ T-cells from the splenocyte pool
(see “Materials and Methods”) completely eliminated this cytotoxic
activity, as shown in Fig. 7B. Splenocytes from saline-treated mice
bearing i.p. AB12 demonstrated minimal cytotoxicity (⬍10%) toward
the parental cell line AB12 and no lysis of RENCA or Line 1 (data not
shown). As expected, splenocytes isolated from naive BALB/c mice
and stimulated with AB12 cells failed to lyse any of the targets tested
(data not shown). Taken together, these data demonstrate that treating
tumor-bearing animals with lipid:pDNA complexes resulted in the
generation of CD8⫹ T cells specific for the tumor cell present in the
peritoneal cavity during treatment. These data also suggest that CD8⫹
T cells were largely responsible for the rejection of the s.c. challenge
because their elimination resulted in essentially a complete loss of
cytotoxicity against AB12 cells (Fig. 7B).
We also evaluated the possible involvement of NK cells in animals
that survived long term as a result of treatment with lipid:pDNA.
Spleen cells were isolated from long-term surviving mice in an AB12
experiment (data not shown) at day 91 post tumor cell inoculation. In
this experiment, groups of mice had been either untreated or treated
with complexes of GL-67:phsp65 or GL-67:p␤gal. Untreated mice
were all dead by day 21, whereas the latter two groups had 50 and
20% long-term survivors, respectively. Specific NK-cell-mediated
target cell lysis was measured indirectly by incubating the cultured
splenocytes (effectors) for 45 min with an excess of unlabeled YAC-1
cells prior to incubating them with 51Cr-labeled AB12 target cells
(30). A comparison of Fig. 8A and Fig. 8B shows that in the presence
of YAC-1 cell inhibition, there was at most a marginal (⬃10%)
decrease in target cell lysis (60:1 E:T). As a positive control for the
effectiveness of YAC-1 inhibition of NK-cell-mediated lysis, the lysis
of 51Cr-labeled YAC-1 cells by these stimulated splenocytes could be
inhibited completely by first incubating the splenocytes with unlabeled YAC-1 cells (data not shown). Equivalent levels of target cell
(AB12) lysis were observed in animals receiving multiple treatments
of either GL-67:phsp65 or GL-67:p␤Gal complexes. In agreement
with the data obtained using CD8⫹ cell depletion (Fig. 7), these
results suggest that although they cannot be totally excluded, NK cells
are not likely to play a major role in the direct, cell-mediated killing
of the tumor cells in a challenge in long-term, immunized animals.
DISCUSSION
In this study, we have demonstrated that repeated i.p. administrations of a complex consisting of cationic lipid and bacterial
plasmid DNA results in powerful effects against mesothelial tumors actively growing in the peritoneal cavity. Indeed, i.p. treatment with lipid:pDNA complexes resulted in long-term surviving
animals in both the AC29 (Fig. 1) and AB12 (Figs. 2 and 3)
models. Treatment was seen to produce long-term survivors that
were not only free of tumor, but more importantly, were also
specifically protected against a later rechallenge by the tumor cell
present during treatment, i.e., a “memory” response had been
generated as a result of the treatment. Induction of this specific,
long-term protective immunity was shown to be paralleled by the
generation of tumor-specific cytolytic CD8⫹ T cells. For example,
splenocytes from long-term survivors in the AB12 model could kill
AB12 cells but not syngeneic RENCA or Line 1 cells (Fig. 7). The
splenocytes responsible for this killing appeared to be CD8⫹,
because their elimination abolished this cytolytic activity (Fig. 7).
Optimal Efficacy Is Obtained with DNA in a Complex. An
important observation from this study was that no significant therapeutic benefit was observed when DNA alone or lipid alone was
administered to the tumor-bearing animals (Figs. 1 and 2). Antitumor
activity was seen only when the DNA was first complexed with
cationic liposomes. Similar findings have been seen in a study of
i.v.-administered lipid:pDNA complexes using models of metastatic
disease (31). Thus, our observed effects, and those of Dow et al. (31),
are not simply due to the presence of bacterial DNA but seem to
require a lipid:pDNA complex. Among the potentially unique functions of such a complex (over free pDNA) are that it may (a) serve to
protect the DNA from degradation, (b) enable more effective uptake
of the DNA by the relevant cells, or (c) enhance the delivery of the
DNA to the cytoplasm or nucleus.
Treatment Activates Both Innate and Adaptive Arms of the
Immune System. These results are consistent with a biphasic immune response to cationic lipid:pDNA complexes that consists of an
immediate, innate response that then matures into an adaptive, memory-based response featuring cytotoxic lymphocytes specific for the
tumor cell present during treatment. In this view, the initial, innate
phase of the immune response is a consequence of the multiple
2961
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 2000 American Association for Cancer
Research.
CATIONIC LIPID:pDNA COMPLEXES IN CANCER THERAPY
treatments with complex, and results in the elimination of a large
proportion of tumor cells already established in the peritoneal cavity.
This cytolytic phase of the response may serve to decrease the i.p.
tumor cell load to the point where a later adaptive response can
effectively eliminate any remaining tumor cells. We would thus
interpret the results in the AC29 model as indicating that multiple
administrations of complex was a very effective regime for generating
these overlapping innate and adaptive cytolytic responses, and that
these responses were somewhat less effective in the AB12 model,
where not all animals survived long term. The immunodeficient
animal studies (Fig. 4) and the antibody depletion studies (Fig. 5)
suggest that the early, innate phase of the response is characterized by
both CD8⫹ and NK cell involvement. It is likely that neutrophils,
macrophages, and eosinophils also play a role in this early response.
This initial, cytotoxic response to treatment leads to an adaptive
immune response characterized by tumor-specific CTLs. These CTLs
are entirely capable of eliminating a s.c. challenge by tumor cells at a
time point that is well removed from treatment. For example, the
challenge time points shown in the AB12 model in Fig. 2 were at least
3 months after the final treatment. This long-lasting memory response
is systemic in nature, as demonstrated by its ability to eliminate tumor
cells implanted at a site distal to the original i.p. treatment. NK cells
did not appear to play a major direct role in this systemic immunity
because inhibiting NK-mediated killing had at most a minor effect on
tumor cell killing by splenocytes from immune animals (Fig. 8). This
result was not unexpected because these mesothelioma models are
reported to express high levels of MHC class I molecules (32), which
would make them inappropriate targets for NK cells. However, these
results do not exclude an important role for NK cells in this memory
phase of the response (i.e., by the secretion of cytokines), they simply
point to a minimal role for NK cells in the direct killing of the tumor
cells.
Our observations that lipid:DNA complexes can inhibit tumor
growth, i.e., the early innate, cytolytic phase of the response, are
consistent with several recent reports. For example, systemic administration of lipid:pDNA complexes lacking a transgene has been
shown to reduce the tumor burden in several models of lung metastases (31). This inhibition of tumor growth was found to be dependent
on NK cell activity and production of IFN-␥, and is entirely consistent
with the role found for NK cells in the early, cytotoxic phase of the
immune response in the present study (see Fig. 5). Similarly, complexes of cationic lipid, protamine, and pDNA have been seen to
inhibit tumor growth in both lung and s.c. tumor models (33, 34). In
this case, inhibition was correlated with the generation of the proinflammatory cytokines tumor necrosis factor-␣, IFN-␥, and IL-12,
which in turn were a consequence of the bacterial source of pDNA
and its immunostimulatory CpG sequences (35–39). Finally, multiple
peritumoral injections of CpG-containing oligodeoxynucleotides
themselves, i.e., in the absence of a cationic DNA-condensing agent,
have been shown to provide antitumor effects in an established s.c.
neuroblastoma model (40). These effects were also found to be
dependent on NK cells but did not appear to recruit CD8⫹ T cells into
the tumor. By contrast, in the present study, immune cell infiltration
into tumor was observed at early time points, using immunohistochemistry, in response to each treatment and consisted of macrophages, neutrophils, and T lymphocytes (data not shown). Thus, this
early phase of the response to bacterial CpG sequences, delivered with
or without a cationic, DNA-condensing agent, appears to involve
multiple components of the innate immune system and can significantly inhibit tumor growth.
It is important to note, however, that the present results extend
significantly the scope of these previous findings by demonstrating
that this innate, cytolytic phase of the response can translate into the
induction of a memory-based T-cell response that is systemic in
nature and capable of eliminating a tumor cell challenge months after
the initial treatment, i.e., a state of antitumor immunity is present. We
are aware of no other such demonstration of the generation of systemic antitumor immunity by the administration of bacteria-derived
DNA, either by itself, or as a complex with a cationic condensing
agent.
A Possible Mechanism Based on Danger Theory. The exact
mechanisms by which the initial treatment-induced inflammatory
reaction and innate immune system involvement are converted into
long-term adaptive immunity are not totally clear (41– 43). One potential scenario is that these complexes lead to the lysis of some tumor
cells by components of the innate immune system, resulting in the
release of tumor antigens, perhaps bound to heat shock proteins (44,
45). Tumor antigen release at the site of immune stimulation could
allow cross-priming of antigen-presenting cells such as dendritic cells
or macrophages, from which the long-term, tumor-specific adaptive
immune response described here could be generated (46, 47). Indeed,
a recent study noted the existence of receptors on antigen-presenting
cells for heat shock protein:peptide complexes (48) that may provide
a path for antigens into the class I presentation pathway. According to
the “danger theory” (7), tumor regression requires both T-cell activation and the presence of a second stimulus to perpetuate the T-cell
response against tumor antigen. In this danger model, the response
would proceed only as long as danger signals are present (7). We
postulate that the repeated local administration of the lipid:pDNA
complexes provides the danger signals necessary to enhance tumor
cell lysis and induce CTL-mediated antitumor immunity.
Bacterial CpG Sequences Are Likely to Play a Role in Efficacy.
It is likely that recognition of the bacterial CpG motifs in the plasmid
DNA used in this study by components of the innate immune system
are important for the generation of the subsequent immune responses.
The immunogenicity of prokaryotic, i.e., plasmid, DNA in mammals
is now well established and appears to be due in large part to the
presence of immunostimulatory motifs consisting of unmethylated
CpG dinucleotides. These CpG motifs have been found to trigger
innate immune responses in mammalian hosts, characterized by the
production of a number of cytokines, including IL-6, IL-12, and
IFN-␥ (35–39). Conversely, a significant reduction in the immunogenicity of bacterial DNA has been observed by methylating these CpG
motifs (49). Indeed, in the mesothelioma models described here,
mammalian DNA did not have the efficacy of bacterial pDNA (data
not shown).
Other, nonexclusive mechanisms for the effectiveness of this therapy are also possible. Cationic lipid:pDNA complexes have been
shown to up-regulate the expression of MHC class I molecules on
tumor cells (11), which could enhance their subsequent recognition by
CTLs. In addition, the introduction of double-stranded DNA (independent of the presence of CpG motifs) into non-immune cells has
also been shown to increase the expression of genes necessary for
antigen processing, such as proteasome proteins, transporters of antigenic peptides, and the co-stimulatory molecule B7.1 (50).
In conclusion, repeated i.p. administrations of lipid:pDNA complexes appear to generate powerful and specific antitumor immune
responses that lead to a significant cure rate. Immunologically, the
state of antitumor immunity that develops appears to be the overall
product of an initial antitumor response of the innate immune system
in response to plasmid DNA presented in the context of a cationic
liposome that in turn stimulates the generation of a specific adaptive
immune response against the tumor cell. These data are the first to
support the idea that treatment with lipid:pDNA complexes can induce a state of antitumor immunity in the treated animal, and as such
2962
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 2000 American Association for Cancer
Research.
CATIONIC LIPID:pDNA COMPLEXES IN CANCER THERAPY
have significant implications for the treatment of cancer and its
metastases.
24.
25.
ACKNOWLEDGMENTS
We acknowledge the contributions of Simon Eastman, Johanne Kaplan,
Kristin Hubley, Sirkka Kyostio-Moore, Chester Li, and Nelson Yew of Genzyme Corporation.
26.
27.
REFERENCES
1. Dranoff, G. The use of gene transfer in cancer immunotherapy. Forum, 8: 357–364,
1998.
2. Speiser, D. E., Miranda, R., Zakarian, A., Bachmann, M. F., McKall-Faienza, K.,
Odermatt, B., Hanahan, D. Zinkernagel, R. M., and Ohashi, P. S. Self antigens
expressed by solid tumors do not efficiently stimulate naive or activated T cells:
implications for immunotherapy. J. Exp. Med., 186: 645– 653, 1997.
3. Yei, S., Mittereder, N., Tank, K., O’Sullivan, C., and Trapnell, B. C. Adenovirusmediated gene transfer for cystic fibrosis: quantitative evaluation of repeated in vivo
vector administration to the lung. Gene Ther., 1: 192–200, 1994.
4. Yang, Y., Li, Q., Ertl, H. C. J., and Wilson, J. M. Cellular and humoral immune
responses to viral antigens create barriers to lung-directed gene therapy with recombinant adenoviruses. J. Virol., 69: 2004 –2015, 1995.
5. Yang, Y., Trinchieri, G., and Wilson, J. M. Recombinant IL-12 prevents formation of
blocking IgA antibodies to recombinant adenovirus and allows repeated gene therapy
to mouse lung. Nat. Med., 1: 890 – 893, 1995.
6. Cassivi, S. D., Liu, M., Boehler, A., Tanswell, A. K., Slutsky, A. S., Keshavjee, S.,
and Todd, S. T. R. J. Transgene expression after adenovirus-mediated retransfection
of rat lungs is increased and prolonged by transplant immunosuppression. J. Thorac.
Cardiovasc. Surg., 117: 1–7, 1999.
7. Fuchs, E. J., and Matzinger, P. Is cancer dangerous to the immune system? Semin.
Immunol., 8: 271–280, 1996.
8. Kazunori, A., Yoshida, T., Matsumoto, N., Ide, H., Koichi, H., Sugimura, T., and
Terada, M. Gene therapy for peritoneal dissemination of pancreatic cancer by liposome-mediated transfer of herpes simplex virus thymidine kinase gene. Hum. Gene
Ther., 8: 1105–1113, 1997.
9. Fife, K., Bower, M., Cooper, R. G., Stewart, L., Etheridge, C. J., Coombes, R. C.,
Buluwela, L., and Miller, A. D. Endothelial cell transfection with cationic liposomes
and herpes simplex-thymidine kinase mediated killing. Gene Ther., 5: 614 – 620,
1998.
10. Namiki, Y., Takahashi, T., and Ohno, T. Gene transduction for disseminated intraperitoneal tumor using cationic liposomes containing non-histone chromatin proteins:
cationic liposomal gene therapy of carcinomatosa. Gene Ther., 5: 240 –246, 1998.
11. Fox, B. A., Drury, M., Hu, H. M., Cao, Z., Huntzicker, E. G., Qie, W., and Urba, W. J.
Lipofection indirectly increases expression of endogenous major histocompatibility
complex class I molecules on tumor cells. Cancer Gene Ther., 5: 307–312, 1998.
12. Nabel, G. J., Nabel, E. G., Yang, Z. Y., Fox, B. A., Plautz, G. E., Gao, X., Huang, L.,
Shu, S., Gordon, D., and Chang, A. E. Direct gene transfer with DNA-liposome
complexes in melanoma: expression, biologic activity, and lack of toxicity in humans.
Proc. Natl. Acad. Sci. USA, 90: 11307–11311, 1993.
13. Lukacs, K. V., Nakakes, A., Atkins, C. J., Lowrie, D. B., and Colston, M. J. In vivo
gene therapy of malignant tumors with heat shock protein-65 gene. Gene Ther., 4:
345–350, 1997.
14. Lukacs, K. V., Lowrie, D. B., Stokes, R. W., and Colston, M. J. Tumor cells
transfected with bacterial heat shock gene lose tumorigenicity and induce protection
against tumors. J. Exp. Med., 178: 343–348, 1993.
15. Pardo, O. E., Colston, J. M., and Lukacs, K. V. Cancer gene therapy with a heat shock
protein gene. Adv. Exp. Med. Biol., 451: 217–223, 1998.
16. Wells, A. D., Rai, S. K., Salvato, M. S., Band, H., and Malkovsky, M. Restoration of
MHC class I surface expression and endogenous antigen presentation by a molecular
chaperone. Scand. J. Immunol., 45: 605– 612, 1997.
17. Fuller, K. J., Issels, R. D., Slosman, D. O., Guillet, J. G., Soussi, T., and Polla, B. S.
Cancer and the heat shock response. Eur. J. Cancer, 30A: 1184 –1891, 1994.
18. Suto, R., and Srivastava, P. K. A mechanism for the specific immunogenicity of heat
shock protein-chaperoned peptides. Science (Washington DC), 269: 1585–1588,
1995.
19. Macario, A. J. L. Heat-shock proteins and molecular chaperones: implications for
pathogenesis, diagnostics, and therapeutics. Int. J. Clin. Lab. Res., 25: 59 –70, 1995.
20. Tamura, Y., Peng, P., Liu, K., Daou, M., and Srivastava, P. K. Immunotherapy of
tumors with autologous tumor-derived heat shock protein preparations. Science
(Washington DC), 278: 117–120, 1997.
21. Srivastava, P. K. Heat shock proteins in immune response to cancer: the fourth
paradigm. Experientia, 50: 1054 –1060, 1994.
22. Lee, E. R., Marshall, J., Siegel, C. S., Jiang, C., Yew, N. S., Nichols, M. R., Nietupski,
J. B., Ziegler, R. J., Lane, M., Wang, K. X., Scheule, R. K., Harris, D. J., Smith, A. E.,
and Cheng, S. H. Detailed analysis of structures and formulations of cationic lipids for
efficient gene transfer to the lung. Hum. Gene Ther., 7: 1701–1717, 1996.
23. Eastman, S. J., Lukason, M. J., Tousignant, J. D., Murray, H., Lane, M. D., St.
George, J. A., Akita, G. Y., Cherry, M., Cheng, S. H., and Scheule, R. K. A
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
concentrated and stable aerosol formulation of cationic lipid:pDNA complexes giving
high-level gene expression in mouse lung. Hum. Gene Ther., 8: 765–773, 1997.
Gao, X., and Huang, L. A novel cationic liposome reagent for efficient transfection
of mammalian cells. Biochem. Biophys. Res. Commun., 179: 280 –285, 1991.
Yew, N. S., Wysokenski, D. M., Wang, K. X., Ziegler, R. J., Marshall, J., McNeilly,
D., Cherry, M., Osburn, W., and Cheng, S. H. Optimization of plasmid vectors for
high-level expression in lung epithelial cells. Hum. Gene Ther., 8: 575–584, 1997.
Davis, M. R., Manning, L. S., Whitaker, D., Garlepp, M. J., and Robinson, B. W. S.
Establishment of a murine model of malignant mesothelioma. Int. J. Cancer, 52:
881– 886, 1992.
Sugiura, K., and Strock, C. Studies in a tumor spectrum III. The effect of phosphoramides on the growth of a variety of mouse and rat tumors. Cancer Res., 15: 38 –51,
1955.
Cikes, M. Antigenic expression of a murine lymphoma during growth in vitro. Nature
(Lond.)., 225: 645– 647, 1970.
Kiessling R., Klein, E., and Wigzell, H. “Natural” killer cells in the mouse. I.
Cytotoxic cells with specificity for mouse Moloney leukemia cells. Specificity and
distribution according to genotype. Eur. J. Immunol., 5: 112–117, 1975.
Coulie, P., Somville, M., Lehmann, F., Hainaut, P., Brassuer, F., Devos, R., and
Boon, T. Precursor frequency analysis of human cytolytic T lymphocytes directed
against autologous melanoma cells. Int. J. Cancer, 50: 289 –297, 1992.
Dow, S. W., Fradkin, L. G., Liggitt, D. H., Willson, A. P., Heath, T. D., and Potter,
T. A. Lipid-DNA complexes induce potent activation of innate immune responses and
anti-tumor activity when administered intravenously. J. Immunol., 163: 1552–1561,
1999.
Christmas, T. I., Manning, L. S., Davis, M. R., Robinson, B. W. S., and Garlepp, M. J.
HLA antigen expression and malignant mesothelioma. Am. J. Respir. Cell Mol. Biol.,
5: 213–220, 1991.
Whitmore, M., Li, S., and Huang, L. LPD lipopolyplex initiates a potent cytokine
response and inhibits tumor growth. Gene Ther., 6: 1867–1875, 1999.
Li, S., Wu, S. P., Whitmore, M., Loeffert, E. J., Wang, L., Watkins, S. C., Pitt, B. R.,
and Huang, L. Effect of immune response on gene transfer to the lung via systemic
administration of cationic lipidic vectors. Am. J. Physiol., 276(5 Pt 1): L796 –L804,
1999.
Yamamoto, S., Yamamoto, T., Shimada, S., Kuramoto, E., Yano, O., Kataoka, T., and
Tokunaga, T. DNA from bacteria, but not vertebrates, induces interferons, activated
NK cells and inhibits tumor growth. Microbiol. Immunol., 36: 983–987, 1992.
Pisetsky, D. S., Reich, C., Crowley, S. D., and Halpern, M. D. Immunological
properties of bacterial DNA. Ann. NY Acad. Sci., 772: 152–163, 1995.
Krieg, A. M., Yi, A. K., Matson, S., Waldschmidt, T. J., Bishop, G. A., Teasdale, R.,
Koretzky, G. A., and Klinman, D. M. CpG motifs in bacterial DNA trigger direct
B-cell activation. Nature (Lond.), 374: 546 –549, 1995.
Klinman, D., M., Yi, A. K., Beaucage, S. L., Conover, J., and Krieg, A. M. CpG
motifs present in bacteria DNA rapidly induce lymphocytes to secrete interleukin 6,
interleukin 12, and interferon ␥. Proc. Natl. Acad. Sci. USA, 93: 2879 –2883, 1996.
Roman, M., Martin-Orozco, E., Goodman, J. S., Nguyen, M. D., Sato, Y., Ronaghy,
A., Kornbluth, R. S., Richman, D. D., Carson, D. A., and Raz, E. Immunostimulatory
DNA sequences function as T helper-1-promoting adjuvants. Nat. Med., 3: 849 – 854,
1997.
Carpentier, A. F., Chen, L., Maltonti, F., and Delattre, J-Y. Oligodeoxynucleotides
containing CpG motifs can induce rejection of a neuroblastoma in mice. Cancer Res.,
59: 5429 –5432, 1999.
Hoffmann, J. A., Kafatos, F. C., Janeway, C. A., and Ezekowitz, R. A. Phylogenetic
perspectives in innate immunity. Science (Washington DC), 284: 1313–1318, 1999.
Medzhitov R., and Janeway, C. A., Jr. Innate immune recognition and control of
adaptive immune responses. Semin. Immunol., 10: 351–353, 1998.
Janeway, C. A., Jr. Presidential Address to The American Association of Immunologists. The road less traveled by: the role of innate immunity in the adaptive immune
response. J. Immunol., 161: 539 –544, 1998.
Todryk, S., Melcher, A. A., Hardwick, N., Linardakis, E., Bateman, A., Colombo,
M. P., Stoppacciaro, A., and Vile, R. G. Heat shock protein 70 induced during tumor
cell killing induces Th1 cytokines and targets immature dendritic cell precursors to
enhance antigen uptake. J. Immunol., 163: 1398 –1408, 1999.
Melcher, A., Todryk, S., Hardwick, N., Ford, M., Jacobson, M., and Vile, R. G.
Tumor immunogenicity is determined by the mechanism of cell death via induction
of heat shock protein expression. Nat. Med., 4: 581–587, 1998.
Banchereau, J., and Steinman, R. M. Dendritic cells and the control of immunity.
Nature (Lond.), 392: 245–252, 1998.
Levitsky, H. I., Lazenby, A., Hayashi, R. J., and Pardoll, D. In vivo priming of two
distinct antitumor effector populations: the role of MHC class I expression. J. Exp.
Med., 179: 1215–1224, 1994.
Arnold-Schild, D., Hanau, D., Spehner D., Schmid C., Rammensee, H. G., de la Salle,
H., and Schild, H. Cutting edge: receptor-mediated endocytosis of heat shock proteins
by professional antigen-presenting cells. J. Immunol., 162: 3757–3760, 1999.
Klinman, D. M., Yanshchikov, G., and Ishigatsubo, Y. Contribution of CpG motifs to
the immunogenicity of DNA vaccines. J. Immunol., 158: 3635–3639, 1997.
Suzuki, K., Mori, A., Ishii, K. J., Saito, J., Singer, D. S., Klinman, D. M., Krause,
P. R., and Kohn, L. D. Activation of target-tissue immune-recognition molecules by
double stranded polynucleotides. Proc. Natl., Acad. Sci. USA, 96: 2285–2290, 1999.
2963
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 2000 American Association for Cancer
Research.
Cationic Lipid:Bacterial DNA Complexes Elicit Adaptive
Cellular Immunity in Murine Intraperitoneal Tumor Models
Michael Lanuti, Samantha Rudginsky, Seth D. Force, et al.
Cancer Res 2000;60:2955-2963.
Updated version
Cited articles
Citing articles
E-mail alerts
Reprints and
Subscriptions
Permissions
Access the most recent version of this article at:
http://cancerres.aacrjournals.org/content/60/11/2955
This article cites 48 articles, 17 of which you can access for free at:
http://cancerres.aacrjournals.org/content/60/11/2955.full.html#ref-list-1
This article has been cited by 9 HighWire-hosted articles. Access the articles at:
/content/60/11/2955.full.html#related-urls
Sign up to receive free email-alerts related to this article or journal.
To order reprints of this article or to subscribe to the journal, contact the AACR Publications
Department at [email protected].
To request permission to re-use all or part of this article, contact the AACR Publications
Department at [email protected].
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 2000 American Association for Cancer
Research.