SVR Angiosarcomas can be Rejected by CD4 Costimulation Dependent and CD8 Costimulation Independent Pathways
- 21 Downloads
We wished to determine whether virally-induced endothelial tumors are rejected by CD4 and CD8 lymphocytes, and whether there are differences in requirements for costimulation in the rejection of these tumors by lymphocyte subsets.
We have developed a model of endothelial tumorigenesis through the sequential introduction of SV40 large T antigen and oncogenic H-ras into endothelial cells. These cells (SVR cells) form highly aggressive angiosarcomas in immunocompromised mice, but do not grow in syngeneic C57BL/6 mice. Using both acute blockade with systemic administration of antibodies and mice genetically deficient in the costimulatory molecules CD28, CD40, and CD40L, we have delineated the requirements of costimulation required to reject this virally-induced endothelial tumor.
Control of SVR angiosarcoma is mediated through T lymphocytes, and both CD4 and CD8 lymphocytes are capable of controlling SVR angiosarcoma growth in vivo. Mice genetically deficient in CD28, CD40, and CD40L were able to reject SVR tumors, but depletion of these mice of CD8, but not CD4 cells led to rapid tumor growth. This data suggests that CD4 mediated rejection has a greater dependence of costimulation than CD8 mediated rejection. Surprisingly, acute depletion of costimulatory molecules in immunocompetent C57BL/6 mice led to rapid tumor growth.
Significant differences exist in the immune status of mice acutely depleted of costimulatory molecules versus genetically deficient mice. Our results suggest that acute depletion is more immunosuppressive than genetic depletion. Humans who undergo costimulatory blockade may require periodic surveillance for virally-induced tumors.
The emergence of AIDS, as well as an increasing population of patients receiving immunosuppressive therapies for transplant regimens and inflammatory disorders, has led to an increased number of virally-induced tumors. Patients infected with HIV have been demonstrated to have an increased frequency of virally induced tumors, such as Kaposi’s sarcoma, HHV-8-induced body cavity lymphoma (1, 2, 3), Epstein-Barr induced lymphoma, and anal squamous cell carcinoma (4,5). These tumors are due in part to viral-specific oncogenes, such as human papillomavirus E6 and E7 (6), Epstein Barr LMP-1 (7), SV40 large T antigen in human mesothelioma (8,9), HHV-8 specific G proteins, and IL-6 homologs (10, 11, 12). The role of the immune system in controlling the development of these tumors is demonstrated by the appearance of these tumors only when lymphocyte counts are severely depleted in AIDS patients, and regression of these tumors when immunity is partially restored through combination antiretroviral therapy (HAART) (13). Similarly, patients on immunosuppressive regimens including cyclosporine and prednisone demonstrate a high incidence of virally induced tumors, of which regression can be induced upon reversal of immunosuppression. More recently, novel immunosuppressive molecules such as CTLA4-Ig and anti-CD40L/anti-CD40 based therapies are undergoing clinical trials as immunosuppressive agents for transplantation and severe inflammatory diseases such as psoriasis and graft-versus-host disease (14,15). The long-term consequences of these novel methods of immunosuppression are not known.
We have developed a model of endothelial tumorigenesis by sequential introduction of a temperature-sensitive SV40 large T antigen and oncogenic H-ras into murine endothelial cells (16). In immunocompromised mice, these cells (SVR cells) form progressively growing tumors that lead to death of the host in 4 weeks, through invasive growth. However, in syngeneic C57BL/6 mice, only slight tumor growth is observed, followed by tumor regression. In this study we describe the role of lymphocyte subsets and costimulatory molecules in mediating rejection of this tumor. This model may be useful in rapidly establishing the efficacy of immunosuppressive regimens in preclinical studies, and may help predict whether novel forms of immunosuppression will lead to an increased incidence of virally induced tumors.
Materials and Methods
Adult male 8–12 week old wild type (C57BL/6), C57BL/6 SCID, and C57BL/6 Nude mice were obtained from Jackson Laboratories (Bar Harbor, ME) and housed in specific pathogen free conditions. Similarly aged male CD40L−/−, CD40−/−, and CD28−/− mice (all on C57BL/6 background) and RAG1 −/− were obtained from Jackson Laboratories and bred as homozygotes under sterile conditions at Emory University.
Creation of Tumor Line and Administration of Tumor
SVR (ATCC 2280) cells were created by introducing temperature sensitive SV40 large T antigen (58-3 allele) and H-ras oncogene into C57BL/6 microvascular endothelial cells. 1 × 106/300 µL SVR cells in cell culture media were injected subcutaneously into the lateral thoracic area using a total volume of 300 µL. Tumor volume was measured using the formula (width2 × length) ×0.52, where width represents the shortest dimension (16).
T Cell Depletion
GK 1.5 (anti-CD4) (100 µg) and TIB105 (anti-CD8) (100 µg) were given as intraperitoneal injections of 100 µg on days −3, −2, −1 prior to SVR administration and weekly thereafter to deplete CD4+ and CD8 + T cells respectively. The antibodies were purified from ascites generated from the GK 1.5 and TIB105 hybridoma cell lines, originally obtained from the American Type Culture Collection (Manassas, VA). Depletion of T cell subsets was confirmed by flow cytometry using anti-CD3, anti-CD4 (L3T4 clone) and anti-CD8 antibodies (Pharmingen, San Diego, CA and/or Caltag Laboratories, Burlingame, CA).
Splenic Cell Preparation
Spleen tissue was harvested from animals and made into a single cell suspension using a wire mesh and RPMI supplemented with 10% FBS. Red cell lysis was performed using a proprietary lysis buffer solution from R + D Systems (Minneapolis, MN). Cells were counted and 2 × 107 splenocytes were then injected intravenously via the penile vein concurrently with SVR inoculation.
Positive Selection of CD8+ T Cells
C57BL/6 spleen and lymph nodes cells were made into single cell suspensions through a wire mesh, washed with RPMI supplemented with 10% FBS and then placed over nylon wool columns. The enriched T cells were then resuspended in 80 µl MACS buffer (PBS + 0.5% BSA + 2 mM EDTA) per 2 × 107 cells. 20 µl MACS CD8 microbeads (Miltenyi Biotec, Bergisch Gladbach, Germany) per 107 total cells were added and incubated at 4°C for 15 min. Cells were then positively selected on the auto-MACS (Miltenyi Biotec). Pre- and post-samples were analyzed via flow cytometry (anti-CD4, anti-CD8a, anti-CD3, Pharmingen, San Diego, CA) to ensure purity.
500 µg of both hamster anti-mouse anti-CD40L mAb (MR1, Bioexpress, Lebanon, NJ) and CTLA4-Ig (provided by Diane Hollenbaugh, Bristol Myers-Squibb) were administered intraperitoneally on days 0, 2, 4, 6 following SVR injection.
Statistics were calculated with an unpaired Student t test, and significance was assigned if the p value was less than 0.05.
Transfer of Naive Syngeneic Lymphocytes from C57BL/6 Mice to SCID Mice Bearing Tumor Leads to Rejection of Tumor
SVR cells do not produce progressive tumors in syngeneic C57BL/6 mice, but form rapidly growing tumors in nude and SCID mice. In order to determine whether this difference was immune mediated, we transferred splenocytes from naïve adult C57BL/6 mice to C57BL/6/SCID mice bearing SVR tumors. Mice receiving splenocytes on the same day or at day 7 after inoculation of 1 × 106 SVR cells were able to control and eliminate tumors, while mice receiving splenocytes on day 14 or unreconstituted mice showed rapid tumor growth, ultimately necessitating sacrifice. Tumor size was assessed at one month after SVR inoculation, at which point animals were sacrificed due to large tumors.
CD4+ and CD8+ T lymphocytes are Capable of Rejecting the Tumor
Transfer of Naïve CD8 Cells is Sufficient to Cause Rejection of SVR Cells
CD4 Mediated Rejection of Tumor is Dependent on CD28, CD40, and CD40L
Effect of Single Blockade of CTLA4-Ig and Anti-CD40L in Syngeneic Mice
In order to determine whether either CTLA4-Ig or anti-CD40L was sufficient alone to allow SVR tumor growth in syngeneic C57BL6 mice, mice were injected with 5 × 105 SVR cells, and immunosuppressed with each agent individually. Mice were injected with immunosuppressive agent on the same day as injection with SVR (day 0), and on days 2,4, and 6 after injection with SVR cells. All animals treated with either CTLA4-Ig or anti-CD40L experienced large tumor growth and eventually succumbed to their disease. Combined therapy with CTLA4-Ig and anti-CD40L also lead to large tumor growth, but was not significantly different from individual therapy with CTLA4-Ig or anti-CD40L (data not shown).
T lymphocytes are required for the control of virally induced tumors of both animals and humans. This is clinically evident in the increased incidence of virally induced cancers, such as lymphomas due to EBV, Kaposi’s sarcoma due to HHV8, and cervical/anal carcinomas due to HPV (17,18). The functional requirements for control of virally induced tumors in humans is difficult to study, as reconstitution of the human immune system in the clinical setting of virally induced tumors is often difficult, especially in the setting of acquired immune deficiency syndrome (AIDS). Regression of virally induced tumors and lymphoproliferative disorders has been seen in transplant patients in which iatrogenic immunosuppression has been reversed by discontinuation of drugs (19,20). Regression of Kaposi’s sarcoma has been occasionally observed in AIDS patients receiving highly effective antiretroviral therapy (HAART) (21, 22, 23). However, knowledge of the precise subsets of cells responsible for rejection of virally induced tumors has not been well studied, partially due to the difficulties in sampling regressing tumors in humans. We have developed a model of a virally induced tumor through the sequential introduction of a temperature sensitive large T antigen and oncogenic H-ras into C57BL/6 microvascular endothelial cells. The resulting angiosarcomas do not grow in adult syngeneic C57BL6 mice, but grow well in allogeneic nude or SCID mice (16). We used this model to study the immune requirements for growth of these tumors. Infusion of naïve T lymphocytes into mice bearing tumor led to rejection of tumor growth, and lymphocytes were observed in the tumor, indicating that rejection of the tumor was immune mediated. Tumors failed to grow continuously in C57BL6 mice depleted of either CD4+ or CD8+ T cell subsets, or in mice homozygous for CD28, CD40, or CD40L. However, vigorous tumor growth was observed in mice homozygous for CD28, CD40, and CD40L mice depleted of CD8+ but not CD4+ cells. This indicates that CD4+ and CD8+ cells can reject SVR tumors independently, but CD4+ mediated rejection may have a greater requirement for costimulatory pathways.
Rejection of SV40-induced tumors has been studied in transgenic models in which large T antigen is targeted to an organ using a specific promoter (24, 25, 26). In these cases, tolerance may develop to large T antigen because the animal is exposed to this protein at an early age. The mode of presentation of antigen may also play a role, and differences in cytotoxic T cell responses to SV40 large T antigen have been observed depending on whether the antigen is introduced in the form of virus, recombinant DNA, or tumor cell expressing T antigen (27, 28, 29, 30). We believe our model has relevance to adult tumors, as many tumor virus infections may occur during adulthood, including EBV, HPV, and HHV8 infections. Viral induced cancers have also been recently demonstrated to have mutations in ras family oncogenes as a second hit, similar to our model (31,32).
Our findings show significant differences with other established models of transplant rejection. CD4+ blockade is sufficient to cause long-term cardiac allograft acceptance in mice (33,34). In the case of acceptance of allogeneic skin, combined blockade of the CD40 and CD28 pathways is insufficient to cause long-term acceptance, and the rejection of skin may be mediated by a novel class of activated T cells which express an NK like marker, asialo-GM1 (35). Acceptance of our tumor shows an intermediate phenotype, in that depletion of CD4 is insufficient to promote in vivo tumor growth, but blockade of CD28 and CD40 with CLTA4-Ig and MR1 antibodies allows vigorous tumor growth.
Immune responses to murine tumors have been demonstrated in several nonendothelial models (36, 37, 38). In several of these studies, CD8 mediated rejection of tumor was costimulation dependent. In our study, we found that CD8 lymphocytes were able to eliminate tumor in the absence of CD28, CD40, and CD40L, but these costimulatory molecules were required for CD4 mediated rejection. Thus, the requirements for rejection via costimulatory molecules differ between tumor types.
Interestingly, tumor growth was more vigorous in the CD28 −/− mice depleted of CD8 lymphocytes than in the more immunocompromised RAG1 −/− knockout mice. This implies that under certain conditions, partial immunosuppression may be increase or decrease susceptibility to tumor growth compared with severe immunosuppression. This phenomenon has been observed in both mice and humans. Vaccination of tumor prone mice with tumor antigens has been shown to enhance tumor growth, and this tumor growth is CD4 mediated (39). Mice partially immunosuppressed by expressing a transgene for CTLA4-Ig showed decreased ultraviolet-induced tumor formation compared with wild type mice (40). Humans with concurrent chronic lymphocytic leukemia and squamous cell carcinoma have skin cancers highly infiltrated with lymphocytes and an aggressive course (41,42). These data suggest that certain lymphocyte populations may contribute to tumor growth.
We compared our results of growth of SVR tumor cells in mice deficient in the costimulatory molecules CD28, CD40, and CD40L with syngeneic immuno-competent C57BL6 mice treated with CTLA4-Ig and anti-CD40L antibodies alone. Surprisingly, tumor growth was vigorous in immunocompetent mice treated with either CTLA4Ig or MR1 alone, compared with the corresponding knockout mice. Thus, in our system, antibody mediated immunosuppression is more effective than genetic knockout of costimulatory molecules. We are uncertain of the reasons for this difference, but compensation may exist in the knockout mice that are not present in the antibody treated mice. Current knowledge of the immune system suggests several possibilities for the differences we observed between acute costimulation blockade versus genetic costimulation blockade. These possibilities include a strong inhibitory effect on gamma delta T lymphocytes due to acute costimulation blockade, versus preservation of gamma delta T cell function in genetic knockout mice (43,44). Gamma delta T cells have been demonstrated to play a crucial role in the defense against cutaneous squamous cell carcinoma and may be important in defense against other malignancies (45). Other possibilities include upregulation of alternative costimulatory molecules, such as ICOS, in genetically deficient mice (46,47). These possibilities are currently under investigation in our laboratory. Finally, our data suggests that patients receiving costimulatory blockade may be at increased risk for virally induced malignancy.
JLA was supported in part by the American Skin Association and grants AR02030, AR44947, and RO1AR47901 from the National Institutes of Health. AB, MD, SC, and CL were supported in part by grants DK50762, DK/AI 40519, AI44644, and P30AR42687 (Emory Skin Disease Research Core Center Grant) from the National Institutes of Health, EEC Award 9731643 from the National Science Foundation, and the Carlos and Marguerite Mason Trust.
- 2.Offermann MK. (1999) Consideration of host-viral interactions in the pathogenesis of Kaposi’s sarcoma. J. Acquir. Immune Defic. Syndr. 21 Suppl. 1: S58–S65.Google Scholar