| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Experimental Therapeutics |
Thoracic Oncology Research Laboratory, Division of Pulmonary, Allergy, and Critical Care Medicine [D. H. S., R. W., M. K., S. M. A.], Department of Surgery [M. O., K. M. A., L. R. K.], Institute for Human Gene Therapy [Y. Z., G-P. G., J. M. W.], University of Pennsylvania Medical Center, Philadelphia, Pennsylvania 19104, and Biogen, Inc., Cambridge, Massachusetts 02142 [J. B.]
 |
ABSTRACT |
|---|
 Top ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
10,000 patients worldwide (1
, 2)
. The incidence, especially in Europe, is rising (3)
. Malignant pleural mesothelioma is unresponsive to chemotherapy and radiotherapy regimens and typically recurs even after the most aggressive attempts at surgical resection (4)
. Multimodality approaches have been of some benefit in prolonging survival in very highly selected subgroups of patients but have had a relatively small impact upon the majority of the patients diagnosed with this disease (5)
. As the incidence of pleural mesothelioma peaks over the next 10–20 years, new therapeutic measures will be necessary.
Despite repeated observations that mesothelioma is well known to inhibit host cellular and humoral antitumor immune responses, there may be a role for intrapleural immunotherapy in the treatment of this refractory neoplasm (6
, 7)
. Using animals models that closely mimic the human disease (8)
, positive therapeutic results have been seen with IFN-
(9)
, interleukin 2 (10)
, interleukin 12 (11)
, and antisense TGF3
-ß treatment (12)
. Some of these findings in animal models have been translated into clinical studies. In several Phase I/II clinical trials conducted over the past decade, some antitumor efficacy has been demonstrated with systemic and local administration of recombinant human interleukin 2 (13
, 14)
and granulocyte/macrophage-colony stimulating factor (15)
.
An area of especially active interest and promise in the treatment of mesothelioma has been the administration of IFNs. IFNs have immunoregulatory effects upon antibody production, NK and T-cell activation, macrophage function, delayed-type hypersensitivity, and MHC antigen expression (16, 17, 18, 19)
. They also have antiangiogenic properties (20
, 21)
, as well as direct antiproliferative effects (19
, 22)
. IFNs have been demonstrated to directly inhibit the cellular proliferation of several mesothelioma cell lines in vitro (23
, 24)
, and antitumor activity has been shown in animal models (9)
. Accordingly, both type I IFNs ( and ß) and type II IFN (IFN-
) have been studied in clinical trials, although the results have not proved as promising as hoped. For example, s.c. administration of IFN-2a to patients with mesothelioma yielded overall response rates of <20% (25
, 26) . The Southwest Oncology Study Group reported no clinical responses after 6 weeks of systemic IFN-ß treatment in 14 patients with pleural mesothelioma and but observed significant toxic side effects (27)
. Improved responses were observed when intrapleural natural IFN-ß was studied in 29 patients with malignant pleural effusions, 11 of whom had "complete remission of pleural effusion," including one patient with mesothelioma (28)
. IFN- has also been studied in the treatment of mesothelioma, especially in the context of intrapleural immunotherapy for mesothelioma. Some positive responses have been seen, especially in patients with very early disease (29
, 30)
.
One hypothesis for the relatively low response rates of IFNs in antitumor therapy is that systemic delivery does not engender the sustained intratumoral cytokine levels necessary for inducing antitumor responses. For example, pharmacokinetic studies after i.v. injection have shown that the half-life of recombinant IFN-ß is only 5 min (31)
. A potential method for increasing local tumor cell exposure to IFNs with minimal systemic side effects could be tumor cell transduction with IFN genes. A number of investigators have tested this hypothesis for IFN-ß by stably transducing tumor cells with IFN-ß and showing marked decreases in tumorigenicity (32, 33, 34)
. This idea has been extended recently to show that in vivo delivery of the IFN-ß gene to flank or brain tumors using various vector approaches resulted in good therapeutic efficacy (35, 36, 37)
.
A number of characteristics make pleural malignant mesothelioma an especially attractive target for gene therapy (38) . The location in the potential space of the thoracic cavity makes the tumor highly accessible, facilitating directed administration of gene therapy vectors and subsequent analysis of treatment effects. In addition, local extension of disease, rather than the development of widespread distant metastases, is responsible for the majority of the morbidity and mortality associated with this neoplasm. Thus, small increments of improvement in local control could lead to significant improvement in survival or palliation. Accordingly, our group has successfully completed several Phase I clinical trials of intrapleural gene therapy for mesothelioma involving intrapleural instillation of recombinant adenoviral vectors containing the herpes simplex virus tk (HSVtk) gene in combination with systemic ganciclovir (39, 40, 41, 42) .
Because of the lack of effective therapy, the potential for response to IFN immunotherapy, and the proven feasibility of local Ad gene therapy, malignant mesothelioma represents an ideal tumor to test IFN gene therapy. On the basis of recent successes using an adenovirus expressing the human and murine IFN-ß gene (35 , 36) , the present study was designed to assess the potential of using a recombinant, replication-deficient adenovirus carrying the murine IFN-ß gene (Ad.muIFN-ß) to treat established i.p. tumors and to explore the mechanisms of the observed therapeutic effects.
We show that Ad.muIFN-ß efficiently transfects murine mesothelioma cells both in vitro and in vivo, and that a single i.p. administration of Ad.muIFN-ß can effectively treat established malignant mesothelioma tumors in syngeneic, immune-competent mice leading to significantly prolonged survival. The effects were attributable to immunological mechanisms because: (a) a single i.p. administration of Ad.muIFN-ß conferred long-lasting, systemic antitumor immunity associated with tumor-specific CTL reactivity; and (b) minimal antitumor efficacy was noted with i.p administration of Ad.muIFN-ß in tumor-bearing immune-deficient mice. This systemic immune response was adequate to inhibit growth of a distant flank tumor upon treatment of i.p. disease. Thus, intracavitary delivery of the IFN-ß gene via an Ad vector provides a remarkably effective treatment for locally established murine mesothelioma, as well as generating distant antitumor effects. The mechanism of action in this model appears to be primarily dependent on CD8+ T lymphocytes rather than on innate immune mechanisms, antiangiogenesis or direct antiproliferative effects.
|
MATERIALS AND METHODS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
REN cells are a human mesothelioma cell line isolated in our laboratory from a patient’s tumor specimen (38) . I-45 is a human mesothelioma cell line obtained as a gift from Dr. Joseph Testa (Fox Chase Cancer Institute, Philadelphia, PA). These cells were cultured and maintained in RPMI 1640 with 10% FBS, 100 units/ml penicillin G, 100 µg/ml streptomycin, and 2 mM glutamine.
Recombinant Ad Vectors.
Ad vectors encoding the murine and human IFN-ß genes were provided by the Vector Core of the Institute for Human Gene Therapy (University of Pennsylvania Medical Center). An adenovirus transfer vector encoding the murine IFN-ß gene driven by the CMV early promoter, pAdCMV-muIFN-ß, was constructed by ligating a cDNA insert encoding murine IFN-ß1a into the plasmid pAd.CMVlink1 (45)
. pAd.CMV-muIFN-ß was cotransfected with the ClaI-restricted H5.010CMVEGP viral backbone into 293 cells for homologous recombination. The white recombinant plaques were isolated by a green/white selection process (46)
. The white plaques were expanded for confirmation in restriction enzyme analysis of the virus genome and muIFN-ß expression. A similar virus was prepared by ligating a cDNA insert containing the human IFN-ß1a (35)
into the same shuttle vector.
As a control vector, replication-deficient Ad vector encoding the herpes simplex type I tk gene driven by the RSV promoter (Ad.RSVtk) was used as described previously (47) . Virus preparations were produced in 293 cells and purified on CsCl gradients after three rounds of plaque isolation. Vector preparations were shown to be negative for the presence of wild-type adenovirus. The particle:pfu ratio of each preparation was determined in 293 cells and ranged between 50:1 and 150:1.
Recombinant muIFN-ß.
Recombinant muIFN-ß (41 units/ng) was provided by Biogen Inc. The protein was kept at -70°C and did not undergo more than one freeze-thaw.
Quantitation of muIFN-ß Protein Levels by ELISA.
One million AB12 cells in six-well plates were infected with Ad.muIFN-ß at a MOI of 50. After 24 h, the wells were washed and refed with 1 ml of medium containing 1% FBS. Twenty-four h later, supernatants were collected, and IFN-ß concentration was quantified by ELISA. Cell counts per well at the end of the collection period were 4 million.
To determine the levels and duration of expression of muIFN-ß after administration of Ad.muIFN-ß, naïve or tumor-bearing mice received injections i.p. with 10 9 pfu of Ad.muIFN-ß. At various time points, animals (n = 3/group) were sacrificed, bled, and subjected to peritoneal lavage with 3 ml of PBS. muIFN-ß levels in blood and peritoneal lavage fluid were then determined by ELISA (see below).
Murine IFN-ß Assay.
Ninety-six-well plates were coated overnight at 4°C with a monoclonal antibody to mouse IFN-ß (muIFN-ß MB-7; Yamasa Shoyu Co., Ltd., Tokyo, Japan). The antibody was used at a concentration of 10 µg/ml in the coating buffer containing 50 mM sodium bicarbonate/carbonate, 0.2 mM MgCl2, and 0.2 mM CaCl2 (pH 9.6). After the plates were blocked with 0.5% nonfat dry milk in PBS for 1 h at room temperature, medium from infected cells or muIFN-ß protein standards (Biogen, Inc.) diluted in 0.5% nonfat dry milk and 0.05% Tween 20 in PBS were added. The plates were then successively incubated at room temperature for 1 h with an anti-muIFN-ß rabbit sera (produced at Biogen; 1:2000 dilution), 1 h with horseradish peroxidase-conjugated donkey antirabbit antibody (Jackson ImmunoResearch; 1:5000 dilution), and the substrate solution (4.2 mM tetramethylbenzidine and 0.1 M sodium acetate-citric acid, pH 4.9). After the reaction was stopped with 2N H2SO4, absorbance was measured at 450 nm.
In Vitro Cytotoxicity Assay.
To evaluate the direct antiproliferative effect of Ad.muIFN-ß, AB12 and AC29 cells were infected with Ad.muIFN-ß or Ad.RSVtk at various concentrations (MOIs of 0.01, 0.1, 1, 10, and 50) and seeded in 96-well plates at a density of 2 x 10 3 cells/well. Similar experiments were conducted with Ad.huIFN-ß using REN or I-45 cells. Survival of infected cells compared with uninfected cells was measured 3 days after infection using the MTS assay, which is a colorimetric test for the quantification of cell viability and proliferation (Cell Titer 96 Aqueous Non-Radioactive MTS Cell Proliferation Assay; Promega Corp., Madison, WI). The percentage of inhibition of cell growth was calculated according to the following formula:
 |
Experimental Animals.
Female BALB/c, female CB17-SCID, and female CB17-SCID/beige mice (6–8 weeks of age; weight, approximately 20–25 g) were obtained from Taconic Laboratory (Germantown, NY). Female CBA/J mice (6–8 weeks of age; weight, 20 g) were obtained from The Jackson Laboratory (Bar Harbor, ME). Animals were housed in the animal facility at the Wistar Institute (Philadelphia, PA). All mice were maintained in a pathogen-free animal facility for at least 1 week before each experiment. The Animal Use Committees of the Wistar Institute and University of Pennsylvania approved all protocols in compliance with the Guide for the Care and Use of Laboratory Animals.
Treatment of Tumor-bearing Mice with Ad Vectors.
Two independent intracavitary tumor models of murine mesothelioma (AB12 and AC29) were established to examine the treatment efficacy of locally administered Ad.muIFN-ß. Murine mesothelioma (AB12 and AC29) cells were grown to 80% confluence in culture flasks, the medium was aspirated, and cells were harvested using 0.05% trypsin-versene. Cells were then pelleted (1000 rpm for 3 min) and resuspended at 1 x 105 cells/ml in serum-free DMEM for i.p. administration. AB12 and AC29 cells (5 x 105 in 0.5 ml) were injected i.p. into BALB/c and CBA/J mice, respectively, using 25-gauge needles. Four to 7 days after tumor cell injection, macroscopic (3–4 mm) tumor nodules could be identified on the small bowel mesentery. Later, tumor masses could be observed on the diaphragm, peritoneal surface, porta hepatis, lesser sac, and retroperitoneum. Survival studies were performed in lieu of tumor burden assessment because of the difficulty in harvesting tumor densely adherent to the abdominal viscera. Animals were sacrificed when they met predetermined criteria established for minimizing pain and suffering. Treatment was initiated when approximately 3–4-mm tumor nodules were identified after tumor cell inoculation. This was usually 4 days after injection of AB12 cells and 7 days after injection of AC29 cells.
For protein experiments, three groups (n = 9) of AB12 tumor-bearing mice were treated i.p. with saline or with recombinant murine IFN-ß at a dose of 3000 units/dose three times weekly ("low dose") or 30,000 units/dose ("high dose") three times weekly for as long as the mice survived.
In gene therapy experiments, unless otherwise noted, mice received a single i.p. administration of Ad.muIFN-ß at a dose of 1 x 109 pfu. In control mice, Ad.RSVtk was injected at an identical dose of 1 x 109 pfu without administration of ganciclovir to control for nonspecific effects of Ad transduction.
To study the effects of Ad.muIFN-ß in tumor-bearing immunodeficient mice, we used both SCID mice and SCID/beige mice, which underwent i.p injection of murine mesothelioma cell lines (AB12 and AC29) as described above.
For the in vivo dose-escalation experiments, BALB/c mice received an i.p. injection of 5 x 10 5 tumor cells diluted in 500 µl of serum-free DMEM. Four days after tumor implantation, mice received a single i.p. injection of 500 µl of DMEM containing varying doses of Ad.muIFN-ß (1 x 10 6, 1 x 107, 1 x 10 8, and 1 x 10 9 pfu). Animals were then observed closely, and survival studies were performed.
"Bulky" Tumor Model.
To examine the effect of Ad.muIFN-ß instillation against bulkier i.p. tumors, we staggered the time of vector instillation (4, 7, and 10 days after tumor inoculation) to allow some of the animals to develop larger (5–7 mm) i.p. nodules, as well as allow more tumor involvement in the porta hepatis and small bowel mesentery. BALB/c mice received i.p injections of 5 x 10 5 tumor cells in 500 µl of DMEM on day 0; subsequently on day 4, 7, or 10, test animals underwent single i.p. instillation of 500 µl of DMEM containing 1 x 10 9 pfu of Ad.muIFN-ß. Survival studies were then performed.
Tumor Rechallenge of Long-Term Survivors.
Animals who were long-term survivors from the above experiments (survival >2 months after primary tumor instillation and single Ad.muIFN-ß treatment) were rechallenged in one flank with s.c. flank injections of 5 x 10 6 cells of the parental tumor cell line (AB12 and AC29) in 100 µl of serum-free DMEM. Each group contained 10 animals. Long-term survivors in the AB12 model were also challenged s.c. with 5 x 10 6 cells in the contralateral flank with an additional murine bronchoalveolar lung cancer cell line, L1C2, syngeneic in BALB/c to evaluate the specificity of the antitumor response. Approximately 11 days after the s.c. challenge, flank tumors were excised and weighed.
To determine the role of T-cell subsets in rechallenge immunity, long-term survivors from AB12/Ad.muIFN-ß experiments were rechallenged with AB12 cells as above. Three groups (each group with 5 mice and 10 flank tumors) were studied. One group received i.p. injections of saline, one group was depleted of CD4+ T cells before and during tumor growth using the schedule described below, and one group was depleted of CD8+ T cells before and during tumor growth. After 11 days, animals were sacrificed, and tumor weights were determined.
In Vivo Depletion of CD4+ and CD8+ T Cells.
To deplete specific immune effector cells subsets prior to and during treatment with Ad.muIFN-ß in the AB12 model, BALB/c mice received i.p. injections of 200 µg of purified monoclonal antibodies purified from the anti-CD4+ hybridoma GK1.5 and the anti-CD8+ hybridoma 53-6.7 (obtained from the American Type Culture Collection). Injections were administered 3 days and 1 day prior to inoculation with AB12 cells. Thereafter, a maintenance dose of antibody was injected i.p. every 7 days throughout the entire experimental period to ensure depletion of the targeted cell type. CD4+ and CD8+ T-cell depletion was confirmed by flow cytometry of splenic suspensions at the time of tumor injection and weekly afterward. BALB/c mice received i.p. injections of 5 x 10 5 AB12 cells in 500 µl of DMEM on day 0. Four days after tumor inoculation, mice received a single i.p. injection of 500 µl of DMEM containing 1 x 10 9 pfu Ad.muIFN-ß. Survival studies were then performed as described above.
Effects of Treatment of i.p. Tumor on Tumors at Distant Sites.
To examine the effect of treatment of i.p. treatment of tumor with Ad.muIFN-ß on the growth of tumor at distant sites, four tumor models of s.c. and/or i.p. AB12 tumors were established. In group 1, 10 mice were injected with 1 x 106 AB12 cells in 100 µl of serum-free DMEM in each flank. In Group 2, 10 mice were injected with 1 x 106 AB12 cells in each flank plus 5 x 105 AB12 cells and 500 µl of DMEM i.p. In group 3, 10 mice were injected with 1 x 106 AB12 cells in each flank and 4 days later received a single i.p. injection of 500 µl of DMEM containing 1 x 109 pfu of Ad.muIFN-ß. This group served as a control to see whether systemic circulation of IFN-ß would affect tumor growth. In Group 4, 10 mice were injected with 1 x 106 AB12 cells in each flank plus 5 x 105 AB12 cells i.p. Four days later, the animals were treated with 1 x 109 pfu Ad.muIFN-ß i.p. This group served as the experimental group to test whether treatment of i.p. tumor would affect the growth of the flank tumors. On day 19 after tumor injection (15 days after treatment with Ad.muIFN-ß), the flank tumors were harvested and weighed.
To determine the importance of the B- and T-lymphocytes in distant response, a similar experiment was performed in three groups of SCID mice (n = 10) using the conditions described above for groups 2, 3, and 4.
CTL Assay.
We performed standard chromium-release assays on murine splenocytes to evaluate for the development of tumor-specific CTLs resulting from Ad.muIFN-ß treatment of mice bearing i.p. tumors. Splenocytes (two mice/group) were harvested 10 days after i.p. treatment with Ad.muIFN-ß and used as effector cells. AB12 and EJ62 cells were used as target cells. Importantly, the effectors (murine splenocytes from animals post-Ad.muIFN-ß treatment) were not stimulated in vitro with target cells prior to testing in the chromium-release assay. To evaluate NK cell activity, effector cells were reacted with YAC-1 cells. Target cells (AB12, EJ62, and YAC-1 cells) were labeled with 51Cr by incubating 1 x 104 cells in a 96-well plate with 3.7 MBq of 51Cr (New England Nuclear Life Science Products, Boston, MA) in normal growth medium for 1 h and then washed five times with PBS. Varying numbers of effector cells were mixed with target cells to yield E:T ratios of 100:1 to 6.25:1 and were incubated for 5 h in a V-bottomed, 96-well microtiter plate. 51Cr release was measured in wells containing effector cells and target cells (cpm test), wells containing target cells in medium alone (cpm spontaneous), and wells containing target cells plus 10% SDS (cpm total). The percentage of lysis was calculated using the following formula:
 |
To deplete CD8+ T-cells from the splenocyte population, aliquots of effector cells were incubated for 30 min at 4°C with magnetic MicroBeads conjugated to antimouse CD8+ T-cell antibodies (Ly-2; Miltenyi Biotec, Inc., Auburn, CA). Beads and bound CD8+ T-cells were magnetically removed from the incubation medium. The CD8+ T-cell-depleted splenocytes were then tested for CTL activity as described above.
Statistics.
The significance of the in vitro inhibition results were determined by Student’s t test (two-tailed) comparing growth inhibition after administration of Ad.IFN versus Ad.RSVtk. Differences in flank tumor weights in the animal experiments were determined by one-way ANOVA adapted for multiple comparisons (Tukey-Kramer test; JMP; SAS Institute Inc., Cary, NC). 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 mean ± SEM.
|
RESULTS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, doses of Ad.huIFN-ß as small as 0.01 MOI induced significant (30–40%) growth-inhibitory effects. At MOIs of 10, 60–70% growth inhibition was seen. There was minimal cytotoxicity with the control adenovirus. At every dose, the percentage of inhibition induced by Ad.huIFN-ß was significantly greater than the corresponding dose of Ad.RSVtk control virus (P < 0.001).
|
32,000 units of IFN-ß per 106 cells/24 h.
The effects of Ad.muIFN-ß on the growth of two murine mesothelioma cells (AB12 and AC29) were studied as above. As shown in Fig. 1, C and D
, growth was also inhibited by Ad.muIFN-ß. At every dose, the percentage of inhibition of Ad.muIFN-ß was significantly greater than the corresponding dose of Ad.RSVtk control virus (P < 0.001). However, these cells were less sensitive to the antiproliferative effects of the virus, with maximum inhibition being only 50–60% of control virus at the highest dose of 50 MOI. This could represent the fact that these murine lines are less infectable by adenovirus than are the human lines, or that murine protein has an inherently lower antiproliferative activity than human protein.
These studies show that Ad.IFN-ß has direct antiproliferative effects on human and murine mesothelioma cells in vitro. Although the human mesothelioma cells are quite sensitive to the direct toxic effects of IFN-ß, mouse studies using these lines would have to be conducted in immunodeficient animals. Because many of the antitumor effects of IFN-ß are thought to be attributable to its effects on the immune system (36) , we reasoned that such studies would not model the clinical situation in which we would actually use this vector. Accordingly, we conducted the animal studies using the two murine mesothelioma cell lines.
Treatment of Established i.p. Mesothelioma with Recombinant Murine IFN-ß or a Single Administration of Ad.muIFN-ß
To study a clinically relevant model of mesothelioma, we used two models of murine mesothelioma, both based on injection of mouse mesothelioma cells injected into the peritoneal cavity of the appropriate strain of mice. Cell line AB12 grows in BALB/C mice. Cell line AC29 grows in CBA/J mice. Both lines can be injected i.p., where they form diffuse tumors throughout the peritoneal cavity (43)
, similar to the presentation of peritoneal human mesothelioma.
As a point of comparison, we first tested the efficacy of injection of IFN-ß protein in the AB12 model. After visible tumor nodules were confirmed, three groups (n = 9) of tumor-bearing mice were treated i.p. with saline or with recombinant murine IFN-ß at a dose of 3000 units/dose three times weekly ("low dose") or 30,000 units/dose ("high dose") three times weekly for as long as the mice survived. The low dose of 3000 units of IFN-ß corresponds approximately to the maximally tolerated doses used in a humans (10 million units or 143,000 units/kg body weight). As shown in Fig. 2
, administration of recombinant IFN-ß at either dose led to significant (P = 0.02 by log-rank test) increases in median survival from 25 days (control) to 37 days; however, there was only one long-term (>60 days) survivor. There was no difference between the low- and high-dose groups.
|
shows the Kaplan-Meier survival curves of groups of tumor-bearing mice (10/group) treated with one dose of Ad.muIFN-ß (109 pfu), a control adenoviral vector Ad.RSVtk (109 pfu) given without ganciclovir, or vehicle alone. Vector was injected after confirmation of the presence of visible tumor nodules (3 mm in size) in the peritoneum. This occurred either 4 days (AB12) or 7 days (AC29) after injection of tumor cells.
|
, all control animals (saline and Ad.RSVtk) died by day 31 with a median survival of 22 days. In contrast, animals injected with one dose of Ad.muIFN-ß had a 100% survival rate (P < 0.0001). In the AC29-injected animals (Fig. 3B)
, the vehicle-injected control animals were all dead by 60 days with a median survival of 40 days. Animals injected with the control adenovirus had a slight prolongation of median survival (40 days versus 54 days), and there were two long-term survivors. This difference did not quite reach significance (P = 0.065 via log-rank test) compared with controls. However, animals injected with one dose of Ad.muIFN-ß had a 100% survival rate (P < 0.0001), similar to the results with the AB12 cell line. These experiments have been repeated multiple times with virtually identical results (see below). These studies demonstrate that although multiple doses of recombinant muIFN-ß result in a small survival advantage, a single dose of Ad.muIFN-ß is remarkably effective in treating two separate models of i.p. established murine mesothelioma. In both experiments, all of the animals survived after Ad.muIFN-ß treatment.
Evaluation of Serum and i.p. Levels of IFN-ß after i.p. Injection of muIFN-ß and Ad.muIFN-ß
To determine the local and systemic levels of IFN-ß after i.p. injection of muIFN-ß or Ad.muIFN-ß, animals were given a single dose of 30,000 units of recombinant protein or one dose of 10 9 pfu of Ad.muIFN-ß via an i.p. injection. At various time points after injection, animals were sacrificed, and the levels of IFN-ß in serum or peritoneal lavage fluid were determined by ELISA. Levels in serum and after a 3-ml peritoneal lavage at baseline were below detection levels (<0.4 ng/ml). After one injection of recombinant muIFN-ß, we were not able to detect activity in the serum at any time point. Peritoneal lavage levels were 43.2 ± 14 ng/ml at 5 min after injection, 15.5 ± 2.4 ng/ml at 30 min after injection, and undetectable after 4 h. After one injection of Ad.muIFN-ß, 1 day after i.p. injection, serum levels averaged 2.7 ± 0.7 ng/ml. At the same time, peritoneal lavage levels averaged 12.5 ± 4.5 ng/ml. By days 4 and 14, however, muIFN-ß levels were below the level of detection in both serum and lavage fluid. Similar results were seen when 10 9 pfu of Ad.muIFN-ß were injected into tumor-bearing animals; clearly detectable levels were present in peritoneal lavage fluid 1 and 2 days after injection but were undetectable by 4 days. Thus, as expected, peritoneal levels were very short lived after injection of recombinant protein (<4 h), whereas detectable levels of IFN-ß were present up to 2 days after the single injection of Ad.muIFN-ß.
Dose-Response Experiments
To examine the dose-response characteristics of Ad.muIFN-ß, two sets of experiments were performed using the AB12 model. In the first experiment, groups of 10 mice were injected with 5 x 105 AB12 tumor cells on day 0 and, after confirmation of visible tumor nodules on day 4, were injected with one dose of virus or vehicle. Doses of Ad.muIFN-ß were 109, 108, 107, or 106 pfu. Animals were then followed for survival. As shown in the Kaplan-Meier survival curve (Fig. 4A)
, all control animals died by day 27. Compared with controls, a clear dose-response effect of vector was seen with long-term survival rates of 90% in the animals injected with 109 pfu (P < 0.0001) and 30% in the animals injected with 108 pfu (P < 0.0001). Survival was slightly but significantly increased in the 107 pfu animals (P < 0.03). There was no increase on the survival rate in animals treated with 106 pfu of Ad.muIFN-ß.
|
As shown in Fig. 4B
, control animals died rapidly with a median survival of 17 days and with all animals dead by 28 days. All treated groups had significant increases in survival compared with control animals (P < 0.0001) with a clear dose-response effect noted. Animals injected on day 7 had a 33% long-term survival rate with an increase in median survival to 44 days (P < 0.0001). Even those animals with large tumors treated on day 10 had a significant (P = 0.0003) increase in median survival (31 days versus 17 days) with one long-term survivor. Consistent with our previous study (Fig. 3)
, animals injected at day 4 had 100% survival.
Thus, Ad.muIFN-ß successfully treats tumors in a dose-responsive fashion. The dose inducing complete survival was 109 pfu. Significant therapeutic effects were also seen in animals with large tumor burdens.
Induction of Long-Term Immunity after Ad.muIFN-ß Therapy
To determine whether treatment with Ad.muIFN-ß therapy induced an acquired immune response against the tumor cells, treated animals "cured" of their tumors were rechallenged with a s.c. injection of tumor. Groups of 10 mice "cured" of AB12 tumors (>60 days after tumor injection and treatment with Ad.muIFN-ß) or naïve, control BALB/c mice were injected in both flanks with either 5 x 106 AB12 cells or L1C2 cells. L1C2 cells are a mouse lung adenocarcinoma cell line that also grows in BALB/c mice.
After 11 days, the flank tumors were removed and weighed. As shown in Fig. 5
, there were statistically significant differences (P < 0.0001) in the size of the tumors induced by injection of AB12 cells into "cured" mice versus previously untreated mice. Specifically, tumors in previously treated mice did not grow, in contrast to the marked growth of the same cells injected into naïve BALB/C mice. Interestingly, there appeared to be some "cross-sensitization" to the L1C2 cell line, because these cells did not grow well in the "cured" mice compared with control mice (P < 0.0001). Similar results were seen with the AC29 model, where none (0 of 10) of the "cured" mice injected with AC29 cells developed flank tumors, whereas all of the control CBA/J mice (10 of 10 tumors) developed large flank tumors (data not shown).
|
Effects of Ad.muIFN-ß in Immunodeficient Animals
Although our rechallenge experiments indicated development of long-term immunity to tumor cells, these studies did not define how tumor cells were eliminated after i.p. injection. IFN-ß has been shown to work by a number of mechanisms, including direct cytotoxicity, inhibition of angiogenesis, NK and macrophage activation, as well as induction of an acquired (T-cell-mediated) immune response (see "Discussion"). To determine which of these mechanisms were operative in our model, we tested the efficacy of Ad.muIFN-ß after injection into tumor-bearing immunodeficient mice.
Groups of 10 SCID/beige mice were injected with 5 x 105 AB12 tumor cells on day 0. On day 4, two mice were sacrificed to confirm the presence of visible tumor nodules. After confirmation, animals were injected with one dose of Ad.muIFN-ß (109 pfu) or vehicle. Animals were then followed for survival. As shown in Fig. 6A
, in contrast to the immunocompetent BALB/c mice injected with AB12 cells (see Fig. 3
), the therapeutic efficacy of Ad.muIFN-ß was almost completely lost, with no increase in median survival and no survival advantage. Because SCID/beige mice are deficient in B cells, T cells, and NK cells, we conducted a similar experiment in SCID mice that lack B and T cells but have intact NK function. As shown in Fig. 6B
, virtually identical results were obtained; the therapeutic efficacy of Ad.muIFN-ß was almost completely lost, with virtually no increase in median survival and no survival advantage.
|
, in this model, there was a slight increase in median survival from 26 to 32 days (P < 0.0001); however, all of the treated animals were dead by day 39. These data show that in the AB12 model, almost all of the therapeutic effects of Ad.muIFN-ß are attributable to factors related to acquired immunity. In the AC29 model, there appears to be a small increase in survival (which may be attributable to Ad or IFN-ß effects), but these effects do not lead to long-term survival.
Determination of the Immune Mechanism of Tumor Killing
Depletion of CD4 and CD8 Cells.
The experiments described above indicate that acquired immune responses are the primary mechanism of the effects of Ad.muIFN-ß in our models. To more precisely define this mechanism, groups of mice were treated with anti-CD4 or anti-CD8 antibodies to deplete these populations of cells. After adequacy of depletion was confirmed by performing fluorescence-activated cell sorter analysis of spleen cells (data not shown), groups of 10 BALB/c mice received injections of 5 x 105 AB12 tumor cells on day 0. On day 4, after confirmation of the presence of visible tumor nodules, animals were injected with one dose of Ad.muIFN-ß (109 pfu) or vehicle and followed for survival. Depletion of the appropriate T-cell population was confirmed weekly by analysis of spleen cells by flow cytometry (data not shown).
As shown in Fig. 7
, there were no significant differences in survival between the untreated animals and the CD8+ T-cell-depleted animals treated with Ad.muIFN-ß. In contrast, the animals depleted of CD4+ T cells showed a marked increase in survival, although not to the level of intact animals (P < 0.0004 versus the saline group). This experiment indicates that, in this model, CD8+ T cells are central to the therapeutic effects of Ad.muIFN-ß. CD4+ T cells also play a role in the generation of antitumor immunity, but even in their absence, significant antitumor effects are still observed.
|
. In one group, CD4+ T cells were depleted using antibody injections at the time of tumor rechallenge of "cured" animals, and in a second group, CD8+ T cells were depleted by antibody injections. Tumor growth was then assessed at 11 days. Consistent with previous data (Fig. 5)
, AB12 cells injected into the flanks of naïve animals grew to large size, whereas tumor growth was completely inhibited in previously treated, "cured" mice (Fig. 8)
. Depletion of CD8+ T cells completely blocked this antitumor effect, allowing tumors to grow to near control size in these animals. However, depletion of CD4+ T cells had very little effect on tumor growth (tumor size was not significantly different from that seen in nondepleted "cured" mice). This experiment demonstrates that CD8+ T cells are absolutely required for antitumor immunity. CD4+ T cells not directly involved in the recall antitumor effects but seem to be more important in helping to generate CD8+ T cells.
|
shows that there was significant cytolytic activity of splenocytes from the Ad.muIFN-ß-treated animals against AB12 targets. In contrast, there was no cytolytic activity seen in control splenocytes or splenocytes from the animals treated with the control Ad.RSVtk vector against the AB12 targets. No cytolytic activity was seen against EJ62 targets using splenocytes from the Ad.muIFN-ß-treated animals or any treatment group (Fig. 9B)
.
|
. Splenocytes from control or Ad.muIFN-ß-treated animals were also tested for cytolytic effects against labeled YAC-1 target cells. No significant lysis was seen at E:T ratios of up to 100 (data not shown), indicating minimal NK cell activity.
These experiments demonstrate that treatment of i.p. AB12 tumor with Ad.muIFN-ß generates tumor-specific CD8+ cytotoxic T cells and that these cells account for almost all of the in vitro antitumor activity. No evidence for active NK cell lysis was seen.
Effects of Treatment of i.p. Tumor on Tumors at Distant Sites
The demonstration of systemic cytotoxic CD8+ T lymphocytes suggested that antitumor effects might not be limited to the local (i.p.) tumor site. To determine whether successful treatment of local (i.p.) disease could lead to antitumor effects at distant (flank) sites, the following four groups of animals were established. In group 1, 10 mice were injected with 1 x 106 AB12 cells in each flank. This group served as the control group for the growth of flank tumors. In group 2, 10 mice were injected with 1 x 106 AB12 cells in each flank plus 5 x 105 AB12 cells i.p. This group served as a control group to rule out inhibitory effects on flank tumor growth by the concurrent growth of i.p. tumor. In group 3, 10 mice were injected with 1 x 106 AB12 cells in each flank and 4 days later were treated i.p. with 1 x 109 pfu of Ad.muIFN-ß. This group served as a control to see whether systemic circulation of IFN-ß after i.p. injection would affect tumor growth. In group 4, 10 mice were injected with 1 x 106 AB12 cells in each flank plus 5 x 105 AB12 cells i.p. Four days later, the animals were treated with 1 x 109 pfu of Ad.muIFN-ß i.p. This group served as the experimental group to test whether treatment of i.p. tumor would affect the growth of the flank tumors.
On day 19 after tumor injection (15 days after Ad.muIFN-ß), the flank tumors were harvested and weighed. As shown in Fig. 10A
, AB12 flank tumors grew to large size in the control animals (group 1), and this growth was not affected by the presence of tumor growing i.p. (group 2) or by the injection of Ad.muIFN-ß into the peritoneal cavity (group 3; P > 0.05 for all comparisons). However, flank tumor growth was markedly inhibited in the group 4 animals (P < 0.05, all comparisons).
|
. These experiments show that treatment of i.p. tumor also results in "distant" antitumor effects that require an intact immune system. |
DISCUSSION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
or AC29 (Fig. 3B)
i.p. mesotheliomas tumors showed complete tumor regression. A control adenovirus vector had minimal antitumor effect in vivo. In contrast, multiple i.p. injections of high doses of recombinant mu.IFN-ß had only minimal therapeutic effects (Fig. 2)
. Significant therapeutic effects were seen in animals treated with various Ad.muIFN-ß doses (Fig. 4A)
and with very large tumor burdens (Fig. 4B)
. A dose of 1 x 10 9 pfu of Ad.muIFN-ß led to almost 100% survival. Taking into account size differences (3000-fold), this dose would be the human "equivalent" of 3 x 10 12 pfu of virus, a dose similar to the highest dose of Ad.HSVtk (1 x 10 12 pfu) that we administered in our previous mesothelioma clinical trial (40)
. Even a single dose of vector one log lower, 1 x 10 8 pfu, significantly prolonged survival compared with control vector and resulted in complete tumor regression in 30% of the treated animals. A single i.p injection of Ad.muIFN-ß in mice bearing bulky intra-abdominal mesothelioma tumors (7 or 10 days after instillation) also resulted in prolongation of survival compared with controls, although with less antitumor efficacy than mice treated at earlier tumor stages (day 4). These results suggest that treatment of mesothelioma at an early stage or in association with a surgical "debulking" procedure to remove gross disease may be successful strategies for Ad.muIFN-ß gene therapy in future human clinical trials. In addition to effective treatment of local i.p. tumors, we were also able to demonstrate that treatment of i.p. tumor led to inhibition of growth of tumor cells at a distant (flank) site (Fig. 10)
. This finding raises the exciting potential of using i.p. (or intrapleural) therapy as a way to treat systemic disease.
The type I IFNs, IFN- and IFN-ß, are a family of closely related cytokines that possess potent antiviral and immunoregulatory activities (17
, 18
, 48)
. IFN- and IFN-ß are produced by both immune and nonimmune cell types and can be induced to high systemic levels in response to viral infections. Type I IFNs can inhibit tumor growth directly by suppressing cell replication and inducing differentiation or apoptosis (19
, 22
, 49)
and indirectly by activating NK-cell mediated lysis (48)
and tumoricidal properties of macrophages (32
, 33
, 50)
. Type I IFNs can also promote tumor regression by suppression of angiogenesis (21)
and stimulation of specific cellular immune responses. Additionally, IFN-ß also has the potential to induce NK cell proliferation (51)
and to promote antitumor T-cell responses by inducing proliferation of memory phenotype subsets and prolonging survival of activated populations (52
, 53)
. Finally, type I IFNs can up-regulate MHC class I expression (54)
.
Taking advantage of these multiple potential mechanisms for antitumor effects, a number of groups have shown that expression of IFN-ß by tumor cells (through ex vivo or in vivo transfection) results in impressive antitumor effects in animal models. In 1994, Yagi et al. (55) showed antitumor effects of cationic liposome encapsulated huIFN-ß in human gliomas that had been transplanted into nude mice. More recent work by the same group has shown antitumor effects and the generation of cellular immunity by liposomal-mediated transfer of murine IFN-ß into mouse gliomas (37 , 56) . Studies performed by Xie et al. (32) and Xu et al. (33) of injecting human and murine cancer cells transfected with murine IFN-ß into nude mice demonstrated that cell growth of transduced tumor cells was markedly inhibited, with the antitumor effects being attributed to up-regulation of inducible nitric oxide synthase in murine macrophages. Qin et al. (35) showed that an adenovirus expressing the human IFN-ß gene (Ad.huIFN-ß) could be injected into established human tumors (breast, colon, cervical, and hepatoma) growing as flank tumors in SCID or nude mice and induce significant inhibition of tumor growth. Because these latter studies were done in immunodeficient animals, effects were likely attributable primarily to innate immune responses, direct antiproliferative effects, or antiangiogenic activity.
In experiments with perhaps more relevance to an actual clinical scenario, Lu et al. (36) showed that an Ad.muIFN-ß vector multiply injected into a murine sarcoma line growing in the flanks of immunocompetent C3H mice led to tumor growth inhibition. In these studies, an antitumor immune response was postulated because: (a) injection of Ad.muIFN-ß induced a prominent T-cell infiltration into the tumors; (b) reinjected tumors did not grow in previously treated animals; and (c) the vector was ineffective when injected into tumors growing in immunodeficient mice.
Given this multiplicity of effects, it was thus of interest to define the mechanisms by which i.p. injection of Ad.muIFN-ß led to its antitumor effects in our mesothelioma models. Our studies show that acquired, CD8+ T-cell-mediated immune responses are primarily responsible for the therapeutic efficacy observed. There appear to be relatively small contributions from other mechanisms, such as macrophage or NK cell activation, direct antiproliferative effects, or antiangiogenesis. This conclusion is based on the following data:
(a) Successful treatment led to antitumor immunity, evidenced by the failure of AB12 cells to grow in the flanks of previously treated animals (Fig. 5)
. This immunity was completely abolished by depletion of CD8+ T cells at the time of rechallenge (Fig. 8)
. Interestingly (see below), this antitumor immunity extended to a second, unrelated lung cancer tumor.
(b) The therapeutic effects of Ad.muIFN-ß were almost completely attenuated in SCID/beige and SCID mice (Fig. 6)
. The complete lack of efficacy in SCID/beige mice, who lack B, T, and NK cells, shows that the antiproliferative activity of IFN-ß on the tumor cells and the antiangiogenic properties were minimal. This is in marked contrast to the effects of IFN-ß injected into human tumors growing on SCID/beige mice reported by Qin et al. (35)
. The experiments in SCID mice, which lack B and T cells but have functional macrophages, neutrophils, and NK cells, were particularly useful in showing that the IFN-ß-dependent macrophage and NK cell activation that appear prominent in other models were of only minor significance important in our system. It is interesting that in at least one other syngeneic murine tumor model (the growth of liver metastases after injection of the CT26 colon carcinoma line), NK cell activation appears to be very important in the antitumor effects.4
(c) We were able to demonstrate that splenocytes isolated from Ad.muIFN-ß-treated animals (but not control animals or animals treated with control adenoviral vector) had high specific cytolytic activity against AB12 tumor cells (Fig. 9)
. This lytic activity was completely abolished by depletion of CD8+ T cells (Fig. 9C)
showing that CD8+ T cells, rather than NK cells, were of primary importance. Of note is the fact that CTL activity was elicited in cells without the need for in vitro stimulation, suggesting very strong CTL responses.
(d) We were able to abrogate completely the antitumor effects of Ad.muIFN-ß by depleting animals of CD8+ T cells using injection of anti-CD8 antibodies (Fig. 7)
. A significant, but lesser, inhibitory effect was seen after depletion of CD4+ T cells, suggesting that both arms of the acquired immune system play important roles in generating antitumor immunity. However, depletion of CD4+ T cells did not block the ability of "cured" mice to reject a rechallenge of tumor cells (Fig. 8)
.
These findings share similarities and differences with the previous studies of Natusme et al. (56) and Lu et al. (36) . Similar to our findings, Natusme et al. (56) , who injected liposomally encapsulated IFN-ß into murine brain tumors, found immunity to re-injected tumor cells and detected splenic CTLs, the specific antitumor cytolytic activity of which could be abrogated completely by CD8+ T-cell depletion. They were also able to block antitumor effects in animals by administration of anti-CD8 monoclonal antibodies. In contrast to our results, however, detection of CTL activity required two to three rounds of in vitro stimulation, and no cross-sensitization responses were noted. Although a more limited immunological analysis was performed, the study of Lu et al. (36) was similar to ours in that immunity to re-injection of tumor cells was observed and that the therapeutic effects of Ad.muIFN-ß were largely lost in SCID mice, thus suggesting minimal effects of NK cells. Important differences between the studies included: (a) that multiple doses of vector were administered to flank tumors by Lu et al. (Ref. 36 ; compared with a single i.p. dose in our study); (b) no cross sensitization was observed; and (c) an "immunogenic" tumor (UV-2237m) versus "non-immunogenic" tumor (AC29; Ref. 57 ) was used. Experiments showing that local treatment could lead to distant therapeutic effects were not performed in either of the other studies.
It is interesting to speculate on why administration of Ad.muIFN-ß was so effective in our models of mesothelioma. An exceptionally powerful T-cell response was generated because treatment with only a single dose of Ad.muIFN-ß not only produced a high percentage of long-term tumor-free survivors but also resulted in inhibitory effects on tumors cells growing at a distant site. In addition, we saw CTL activity in splenic cells without the need for in vitro stimulation. In previous immunotherapy experiments using liposomes containing DNA and using the same cell lines, but with multiple treatments, we found that at least one round of in vitro stimulation was needed to detect CTL activity (43) . Treatment in the current study led to the generation of systemic immunity to challenge with parental tumor cells; however, somewhat surprisingly, we also observed immunity to another syngeneic, nonmesothelioma tumor, the L1C2 murine lung adenocarcinoma. The appearance of such cross-sensitization has been reported (58) but is unusual.
There are a number of possible explanations for the strength of these effects. For example, the route of administration may be important. The peritoneal cavity is an anatomical site where immune responses are vigorous, and thus the i.p. route has been used often for in vivo immunization (59 , 60) . One reason for this is the large surface area available. Another advantage could be the presence of large numbers of free floating macrophages and dendritic cells in this cavity (61) that have free access to the tumors growing at the surface of the peritoneum. In addition, it has been recognized recently that normal mesothelial cells also have antigen-presenting capability and may be participating in the generation of antitumor immune responses (60 , 62 , 63) . Experiments are ongoing to test the effects of Ad.muIFN-ß injected into flank tumors compared with i.p. tumors. The extent of gene transfer after vector administration may also be important. We have shown previously that large numbers of tumor cells and normal mesothelial cells are transduced after i.p. injection of adenovirus (38) . There may also be gene transfer into leukocytes within the peritoneal cavity, although we have not specifically examined this issue. Future studies are aimed at exploring some of these questions.
There are also some factors specific to mesothelioma that may also be important. Unlike many tumors, mesotheliomas express high levels of MHC class I antigens (64)
that may make them more resistant to NK cell activity and more susceptible to CD8+ T-cell lysis. Both AB12 and AC29 cells express high levels of MHC class I molecules (data not shown). Mesothelioma cells make large amounts of TGF-ß that may induce an immunosuppressive environment (12)
. Bielefeldt-Ohmann et al. (9)
showed that IFN- suppressed TGF-ß mRNA expression in mesothelioma tumors. Although not examined in our study, it is possible that IFN-ß inhibited the production of TGF-ß in a similar fashion, thus improving immune responses. Finally, the fact that mesothelioma cells are sensitive to the antiproliferative and apoptotic effects of Ad.muIFN-ß (Fig. 1)
could play a key role in generating appropriate "danger signals" needed to drive effective antitumor responses (65
, 66)
. The importance of having the tumor cells transduced is suggested in preliminary experiments, where i.p. injection of Ad.muIFN-ß followed by injection of tumor cells is ineffective whereas ex vivo transduction of tumor cells prior to i.p. injection leads to complete cure in immunocompetent mice but not in SCID mice.
In summary, we report that adenovirus-mediated muIFN-ß gene therapy can exert potent antitumor activities in in vivo models of murine malignant mesothelioma. A single i.p. administration of Ad.muIFN-ß was sufficient to cause complete regression of i.p. tumors in immunocompetent animals and induce long-lasting immune responses that protected them from s.c. challenge with parental tumor cells and at least one other syngeneic tumor. Effects on distant tumor were also seen. This poten
) or with recombinant murine IFN-ß at a dose of 3000 units/dose three times weekly ("low dose"; •) or 30,000 units/dose ("high dose"; 
) three times weekly for as long as the mice survived. Administration of recombinant IFN-ß at either dose led to significant (P = 0.02 by log-rank test) increases in median survival from 25 days (control) to 37 days; however, there was only one long-term (>60 days) survivor. There was no difference between the low- and high-dose groups.
) was administered to BALB/c mice bearing AB12 tumors 4 days after tumor injection (A) or to CBA/J mice bearing AC29 tumors 7 days after injection (B). Control mice received 500 µl of saline i.p. (
), 106 (
) pfu at day 4 after tumor inoculation. Control mice received 500 µl saline i.p. (
) by the use of magnetic beads completely lost their cytotoxic activity against the parental AB12 cells when compared with nondepleted splenocytes (