Published ahead of print on March 11, 2005, doi:10.1164/rccm.200501-057OC
© 2005 American Thoracic Society doi: 10.1164/rccm.200501-057OC
Immunotherapy of Murine Malignant Mesothelioma Using Tumor Lysate–pulsed Dendritic CellsDepartment of Pulmonary Medicine, Erasmus MC, Rotterdam, The Netherlands Correspondence and requests for reprints should be addressed to Joost Hegmans, B.Sc., Erasmus MC, Department of Pulmonary Medicine, H-Ee2253a, P.O. Box 1738, 3000 DR Rotterdam, The Netherlands. E-mail: j.hegmans{at}erasmusmc.nl
Rationale: Exploiting the immunostimulatory capacities of dendritic cells holds great promise for cancer immunotherapy. Currently, dendritic cell–based immunotherapy is evaluated clinically in a number of malignancies, including melanoma and urogenital and lung cancer, showing variable but promising results. Objective: To evaluate if pulsed dendritic cells induce protective immunity against malignant mesothelioma in a mouse model. Methods: Malignant mesothelioma was induced in mice by intraperitoneal injection of the AB1 mesothelioma cell line, leading to death within 28 days. For immunotherapy, dendritic cells were pulsed overnight either with AB1 tumor cell line lysate, AB1-derived exosomes, or ex vivo AB1 tumor lysate, and injected either before (Days –14 and –7) at the day of (Day 0) or after (Days +1 and +8) tumor implantation. Main Results: Mice receiving tumor lysate–pulsed dendritic cells before tumor implantation demonstrated protective antitumor immunity with prolonged survival (> 3 months) and even resisted secondary tumor challenge. Tumor protection was associated with strong tumor-specific cytotoxic T-lymphocyte responses. Adoptive transfer of splenocytes or purified CD8+ T lymphocytes transferred tumor protection to unimmunized mice in vivo. When given after tumor implantation in a therapeutic setting, pulsed dendritic cells prevented mesothelioma outgrowth. With higher tumor load and delayed administration after tumor implantation, dendritic cells were no longer effective. Conclusions: We demonstrate in this murine model that immunotherapy using pulsed dendritic cells may emerge as a powerful tool to control mesothelioma outgrowth. In the future, immunotherapy using dendritic cells could be used as adjuvant to control local recurrence after multimodality treatment for malignant mesothelioma.
Key Words: cancer • DC-based vaccines • dendritic cells • immunotherapy • mesothelioma Malignant mesothelioma (MM) arises primarily from the surface serosal cells of the pleural, peritoneal, and pericardial cavities and is a highly aggressive neoplasm. MM of the pleura is most often seen in patients with a history of occupational asbestos exposure. Although the worldwide usage of asbestos has been reduced considerably, incidence and mortality related to MM continue to rise, because of the long latency period of 20 to 40 years between exposure and first symptoms (1, 2). With median survival durations of 10 to 17 months from onset of symptoms, the prognosis is poor (3, 4). To date, there is no standard curative therapy for MM. Surgical approaches such as pleurectomy and extrapleural pneumonectomy alone result in high local recurrence rates and questionable survival benefit. Additional treatments (chemotherapy, radiotherapy, gene therapy, photodynamic therapy, multimodality approaches) result in only limited improvements in response and survival (3, 5–9). The possibility to harness the potency and specificity of the immune system underlies the growing interest in cancer immunotherapy. One such approach uses dendritic cells (DCs) to present tumor-associated antigens (TAA) and thereby generate tumor-specific immunity (10–12). DCs are extremely potent antigen-presenting cells specialized for inducing activation and proliferation of CD8+ cytotoxic T lymphocytes (CTL) and helper CD4+ lymphocytes (13). This unique property has prompted their recent application as therapeutic cancer vaccines. In the design and conduct of DC-based immunotherapy trials, several important considerations influence induction of a successful protective response (14). First is the source of tumor antigen that can be loaded onto DC. In case of unknown tumor antigens, as for MM, the source of antigen is, by necessity, a tumor cell lysate, apoptotic tumor cells, whole tumor–derived RNA, or tumor-derived exosomes (15). Second is the way in which DCs are activated, because immature DCs can tolerize the antitumoral response (16). Other important variables are dose, frequency, timing, and route of administration (17–22). Taking into account these variables, most studies have shown that injection of mature tumor antigen-pulsed autologous DCs into tumor-bearing hosts induces protective and therapeutic antitumor immunity in experimental animals and for some malignancies in patients (22, 23). These promising results using DC-based immunotherapy have prompted us to test the hypothesis that autologous DC–presenting tumor antigen might also induce a protective immune response in MM. A mouse model for MM allowed us to prove this hypothesis and to study the impact of antigen source, DC maturation status, and timing of administration on outcome. Our results suggest that DC-based immunotherapy might be effective against this aggressive cancer in which TAA remain undefined. This study should pave the way for a clinical feasibility trial using autologous DCs as a therapeutic adjuvant in the treatment of patients with MM.
See the online supplement for more details regarding laboratory animals and cell lines, chromium-release assay, IFN-  enzyme-linked immunospot (ELISPOT), and the adoptive transfer of splenocytes and CD8+ cytotoxic T cells.
Animals and Cell Lines
Source of Tumor Antigen Derived from AB1 Tumor
Culture Conditions of Bone Marrow–derived DC Subtypes Used for Vaccination
DC subtype I.
DC subtype II.
DC subtype III.
Tumor Antigen Loading and Induction of Maturation
Experimental Protocols to Demonstrate Induction of Antitumoral Immunity
Protocol 1: treatment with tumor lysate–pulsed DCs before tumor implantation. On Day –14, 18 mice were vaccinated with 106 DC subtype I pulsed with AB1 tumor cell line lysate in 500 µl PBS, of which 12 mice received CpG motif–matured DCs. A corresponding group of mice was vaccinated with 106 unpulsed DC subtype I (n = 12) or PBS alone (n = 12). The vaccination procedure was repeated 1 week later (Day –7). On Day 0, the mice were inoculated into the peritoneum with 0.5 x 106 AB1 cells in 500 µl PBS. Mice were examined daily for evidence of ill health or overt tumor growth. Mice were killed if profoundly ill, according to UKCCCR regulations, and were scored as a death in survival analysis. All mice were underwent extensive autopsy. In mice that survived for a prolonged period after active DC immunotherapy, a second tumor challenge with 0.5 x 106 AB1 cells was given intraperitoneally after 3 months.
Protocol 2: treatment with tumor lysate–pulsed DCs at the day of tumor implantation.
Protocol 3: treatment with tumor lysate–pulsed DCs after tumor implantation and effect of DC subtypes on outcome.
Protocol 4: effect of high tumor load on success of treatment with tumor lysate–pulsed DCs after tumor implantation.
Protocol 5: source of antigen used to pulse DCs.
Statistical Analysis
Phenotype of Bone Marrow–derived DCs and Effects of Antigen Pulsing As previously shown (26, 27, 33), culture of lineage-negative bone marrow cells in granulocyte-macrophage colony–stimulating factor (GM-CSF) and Flt3-L leads to DC differentiation, as shown by the almost universal expression of CD11c and MHC class II. To examine the phenotype of DCs after exposure to AB1 tumor cell line lysate or ex vivo tumor lysate, we performed flow cytometry using DC maturation markers CD40, CD80, and CD86. Moreover, the effect of adding the innate immune-activating unmethylated CpG motifs on DC maturation was investigated. As shown in Figure 1, the overnight addition of AB1 tumor cell lysate to bone marrow–derived DC subtype I at Day 9 of culture induced the upregulation of CD40 and CD80 (and also CD86, data not shown), compared with unpulsed DCs, in accordance with the induction of maturation. The addition of CpG motifs during this overnight period leads to an even further mature phenotype of DCs, expressing high levels of CD40, CD80, CD86, and MHC II. This effect of CpG and tumor cell line lysate or ex vivo tumor lysate occurred irrespective of whether DCs were grown in the absence of Flt3-L (DC subtype II) or were generated from total bone marrow cells in the presence of GM-CSF (DC subtype III). DC subtype I, whether exposed or not to CpG motifs to induce further maturation, were used for experiments exploring the potential of DC immunotherapy.
Immunization with Tumor Lysate–pulsed DCs before Tumor Implantation Prevents Mesothelioma Outgrowth At Days 14 and 7 before tumor cell injection, naive mice were injected intraperitoneally with PBS, 1 x 106 untreated DC subtype I, or DC subtype I treated with AB1 tumor cell line lysate with or without CpG motifs (Figure 2). On Day 0, all mice were injected intraperitoneally with a lethal dose of 0.5 x 106 AB1 tumor cells. The first signs of terminal illness (typically formation of ascites, ruffled hair, or marked loss of condition) appeared after 10 days in the PBS-treated group receiving tumor challenge. Within 30 days, all mice from this group showed evidence of ill health or overt tumor growth. Mice were subjected to extensive autopsy, which always showed solid tumor formation within the peritoneal cavity, which was accompanied in a few cases by thick, yellow-stained ascites. The nature of the solid tumors varied from numerous small nodules spreading throughout the mesentery and peritoneal lining to a single, large mass. Strikingly, mice immunized with AB1 tumor cell line lysate–pulsed DCs showed prolonged survival, and all mice remained tumor free for more than 3 months. The protective effect occurred both in the group receiving CpG-matured antigen-pulsed DCs and in the group receiving less mature, unmanipulated antigen-pulsed DCs. In a separate group of DC-protected mice, four mice were killed at 2 months and checked for tumor growth. No tissue abnormalities or formation of tumors could be detected. Enhanced survival in 5 of 12 mice (42%) was also seen in mice injected with unpulsed DCs, suggesting that these unpulsed DCs could induce tumor protection in a nonspecific manner.
To check if DC immunization induced longlasting protective immunity, some protected mice receiving DC immunotherapy and tumor challenge were injected for a second time 3 months after the first tumor challenge with a repeated injection of 0.5 x 106 AB1 tumor cells. In these protected mice, a second tumor challenge did not lead to MM outgrowth, and mice again survived for an extended time of at least 3 months (survival curve not shown).
Immunization with Tumor Lysate–pulsed DCs at the Day of Tumor Implantation Promotes Mesothelioma Outgrowth
Immunization with Tumor Lysate–pulsed DCs after Tumor Implantation Prevents Mesothelioma Outgrowth We next examined if DC immunotherapy given after tumor challenge would inhibit MM outgrowth. When tumor lysate–pulsed CpG-matured DC subtype I was injected 1 and 8 days after a lethal dose of AB1 cells, survival was much improved. In DC-treated animals, median survival was more than 2 months (66% of animals were still alive at 2 months) compared with only 23 days in mice receiving only AB1 cells (Figure 4). We also tested whether different culture conditions used to generate DCs would lead to better success of immunotherapy and protect more mice from death. Treatment with tumor lysate–pulsed CpG-matured DC subtypes II and III enabled the complete control of MM outgrowth, with all mice surviving beyond 2 months.
We then examined the effect of prior tumor load on the success of DC immunotherapy. Because the previously described experiment showed that DC subtype III was the most effective to control MM outgrowth after tumor transplantation, we performed all succeeding experiments with this DC subtype. To allow faster tumor growth, we injected 1 x 106 instead of 0.5 x 106 AB1 tumor cells intraperitoneally, leading to death within 12 days (median survival, 10 days). Immunotherapy of MM had a better outcome when DCs were injected early in tumor development, indicating that tumor load plays an important role in survival (Figure 5). When DCs were injected 1 day after tumor cell injection, median survival was prolonged to 23 days. When DC immunotherapy was initiated 3 days after tumor implantation, median survival was 21 days, whereas when DC immunotherapy was delayed until 5 days after tumor implantation, median survival was only 13 days. In these experiments, where high tumor load was associated with delayed treatment with only one DC injection, death occurred in all animals by 35 days, irrespective of treatment.
Source of Antigen Used to Pulse DCs Very few TAA have been described in MM, and therefore a source for exogenous tumor peptides is unavailable (34). Therefore, we tested the efficacy of DC immunotherapy using DCs loaded with different "crude" sources of tumor antigens, which include AB1 cell lysate, tumor tissue lysate isolated ex vivo, and AB1-derived exosomes. As seen in Figure 6, DC subtype III pulsed with these different sources of antigen was also effective in prolonging survival when given 1 and 8 days after tumor implantation. In this experiment, AB1 cell line lysate–pulsed DCs induced the best overall survival, followed by ex vivo tumor cell lysate and AB1-derived exosome–pulsed DCs.
Successful Tumor Lysate–pulsed DC Immunotherapy Is Associated with Cytotoxic T-Cell Induction Splenocytes obtained from protected animals that had resisted a tumor challenge and from naive mice were used in a 51Cr-release assay and IFN- ELISPOT assay (Figure 7). AB1-specific CTL responses were measured in the 51Cr-release assay (Figure 7A). Corrected percentage lysis of AB1 cells by splenocytes taken from protected mice was significantly elevated compared with naive mice (mean, 24% vs. < 1%; p < 0.05). The production of IFN- by splenocytes after DC immunotherapy and tumor protection was measured using ELISPOT. The number of IFN-–producing spleen (and lymph node) cells was markedly increased in protected mice, up to 20 times higher compared with naive mice (Figure 7B). The addition of tumor cells to the ELISPOT assay did not lead to a further enhancement of number of IFN- spots, illustrating that IFN- release was spontaneous after DC immunotherapy. Thus, both 51Cr-release and IFN- ELISPOT supported an overall increase of CTL activity in DC-protected mice.
Transfer of Splenocytes or CD8+ T Cells from DC Immunotherapy–protected Mice Transfers Tumor Protection To prove that CD8+ CTL induced by DCs were mediating protection from tumor outgrowth, we evaluated the efficacy of adoptive transfer of splenocytes and CD8+ T cells in the prevention and treatment of MM (Figure 8). Intravenously injected splenocytes (10 x 106 cells) were given 7 days before and 2 days after a lethal dose of AB1 cells (Figure 8A). Splenocytes from protected mice dramatically increased the survival of mice compared with splenocytes from naive mice when injected 7 days before AB1 injection. Treatment with splenocytes 2 days after AB1 injection did not increase survival. Intravenously injected CD8+ T cells increased survival when given 7 days before or 2 days after AB1 injection (Figure 8B).
There is no widely accepted curative approach for MM and treatment is usually complicated by a high local recurrence rate, despite aggressive surgery and novel attempts to improve local control (35). Multimodality approaches, including extrapleural pneumonectomy followed by chemoradiotherapy, have been of some benefit in prolonging survival of highly selected subgroups of patients, at the expense of considerable toxicity, but they have had a relatively small impact on the majority of the patients diagnosed (8). Therefore, new therapeutic strategies are urgently needed. Alternative therapies based on the intrapleural injection of adjuvants (e.g., bacille Calmette-Guérin) or photodynamic therapy using photosensitizers remain unsatisfactory because they have shown little potential for improving local control or overall survival, and often are quite toxic. Immunotherapy approaches, such as systemic, intrapleural, or intralesional administrations of interferons (IFN-  , IFN-ß, IFN-) or interleukins (IL-2, GM-CSF, IL-12), are in an experimental stage for patients with mesothelioma (36–41). Systemic administration of cytokines has resulted in considerable toxicity in both human and murine models and intrapleural delivery produces a localized immune reaction with tumor regression in only a minority of patients (39, 42, 43). The strategy of using gene therapy to directly introduce various cytokine genes into cells has been performed (40, 44, 45). It provides an increased local concentration of cytokines while minimizing the systemic toxicities, and significant tumor reductions in animals were demonstrated. However, this research has not gone beyond the laboratory. Other studies are at the early stages of preliminary clinical trials and are not standard mesothelioma treatment. The major drawback of these strategies is that they are passive forms of immunotherapy, which will probably yield only temporary benefit, in contrast to strategies aimed at inducing an active immune response to cancer cells, such as vaccination strategies using stimulated DCs. Therefore, in this article, we have explored a new way of controlling the outgrowth of MM, which is to use the natural adjuvant capacities of DCs, the most powerful antigen-presenting cells of the immune system (13). DC-based immunotherapy is emerging as a nontoxic, efficient, and broadly useful immunotherapy strategy for the treatment of patients with cancer (46). DCs are instrumental in inducing activation and proliferation of CD8+ CTL, sometimes even bypassing the need for CD4 help. In patients with cancer and in tumor-bearing mice, DC function is suppressed through release of tumor-derived soluble factors that inhibit the differentiation, maturation, and therefore immunostimulatory function of DCs, leading to a defective induction of CTL responses (47–53). A major advantage of DC-based immunotherapy is that DCs can be generated in large amounts in vitro, in the absence of this suppressing environment, and subsequently injected in a mature state to induce CTL responses. It is now well established that tumor cells contain many antigens that can be recognized by the host immune system. DC immunotherapy was shown to be efficient as a cancer vaccine for tumors when pulsed with known TAA, such as the melanoma antigen–derived family of tumor antigens. In the case of tumors without known TAA, DCs have been pulsed with necrotic or apoptotic tumor material or with tumor-derived RNA, also leading to efficient immunity. For MM, few TAA are known like SV-40 large T antigen, a family of testis-associated antigens, Wilms' tumor-1 protein, mesothelin, calretinin, telomerase, survivin, and topoisomerase II, but these antigens are not expressed on the membranes of all tumors and are therefore not suitable as an antigen source for DC pulsing. None of these TAA have been evaluated as a source of peptides to pulse DCs or in a cancer vaccine trial (34, 54). Vaccinating against a single or a few TAA is limited by peptide restriction to a given HLA type and the induction of CTL without Th1 response. Also, tumor lysate–priming strategies are advantageous in providing the full antigenic repertoire of the tumor and, particularly, unique tumor antigens, which will theoretically decrease the ability of tumors to evade the immune response by downregulation of a single antigen. Ebstein and coworkers (34) have recently shown that human DCs pulsed with dead MM cells were able to induce a cytotoxic T-cell response in vitro directed against the tumor, particularly when DCs were loaded with apoptotic tumor material, illustrating that MM cells contain unknown TAA that can lead to an antitumoral immune response. Although this strategy was shown to be efficacious in vitro, it has not been shown that tumor antigen–pulsed DCs would have an antitumoral effect against MM in vivo.
Reliable animal models can provide useful preclinical information about DC immunotherapy and are critical for evaluating and defining tumor immunology paradigms because they provide an in vivo milieu that cannot be reproduced in vitro (55). The injection of crocidolite asbestos–induced AB1 mesothelioma cells into the peritoneal cavity of syngeneic mice provides a valid experimental model for human MM (56–60). This murine model demonstrates all the histologic and pathologic features of the human disease and, like most solid tumors, is only weakly immunogenic, without any known TAA. We used this MM model and demonstrated that DCs pulsed with lysed murine MM cells induced protective immunity to MM challenge in vivo. The first signs of terminal illness after intraperitoneal injection of 0.5 x 106 AB1 cells occurred between 2 to 4 weeks, but mice receiving tumor lysate–pulsed DCs were protected for months and even resisted a secondary challenge with tumor, illustrating the induction of long-lived immunity. In support of the induction of a systemic antitumoral immune response, we examined the lymph nodes and spleen to see if CTL activity was induced by DC immunotherapy in protected mice. The 51Cr-release and IFN- Strikingly, mice treated twice with unpulsed DCs before a lethal dose of tumor cells also showed prolonged survival in 42% of the mice, although no source of tumor antigen was provided to these cells. This phenomenon was previously described in 1985 by Knight and coworkers (63). Several explanations are possible. First, it was recently shown that DCs not exposed to tumor antigen had the capacity to strongly enhance antitumoral NK cell activity through direct activation of NK cell cytotoxicity, thus activating the innate immune response to various tumors (64). This response occurs even in the complete absence of CD8 cells and requires CD4+ T cells and DC–NK cell contact (65). DC-activated NK cells could therefore kill MM cells directly, in the absence of tumor antigen. Activated NK cells are known to kill the AB1 murine MM cells and human MM cells (66, 67). Second, because both DCs and AB1 cells were grown in fetal bovine serum–containing medium, we suspect that serum proteins might have provoked this response. Some of these proteins are antigenic and thus could induce a strong antitumoral immune response (68). To avoid this problem, we attempted to grow AB1 cells and DCs in serum-free conditions, which turned out to be troublesome. Finally, injected DCs could survive for 1 week after injection and could potentially take up tumor cell fragments, thus leading to tumor antigen presentation and to a tumor-specific immune response.
The induction of proper DC maturation is an important factor in the design of DC immunotherapy trials, because antigen presentation by immature DCs might tolerize for TAA and potentially enhance tumor growth (16). Therefore, clinical trials evaluating the potential of DC immunotherapy have included protocols to induce DC maturation. Most commonly, DCs are exposed to a cocktail of maturation cytokines or are exposed to monocyte-conditioned medium. In our system, the pulsing of DCs with lysed tumor cells already led to DC maturation, in the absence of any additional cytokines. It has been shown that necrotic tumor cells can enhance DC maturation and immunogenicity by providing so-called danger signals to DCs, one important factor being uric acid (69). To further enhance maturation of DCs, we exposed DCs to innate immune system–activating signals during the period of antigen pulsing. Synthetic, unmethylated CpG oligodeoxynucleotides (CpG) are considered potent activators of DC function and maturation in vitro and in vivo, by acting on the innate Toll-like recepter 9, expressed by myeloid DCs and plasmacytoid DCs in mice and plasmacytoid DCs in humans (70–72). Unmethylated CpG motifs have been evaluated as immunotherapeutic adjuvants in a number of preclinical cancer models, where they have led to enhanced induction of CTL and NK cell responses (73). The immunostimulatory effect of CpG motifs on DCs leads to upregulation of surface costimulatory molecules and increased cytokine production, thus further enhancing the ability of DCs to stimulate T-cell responses (74). CpG motifs are preferred in clinical trials because other potent stimulators of DC maturation, like LPS and tumor necrosis factor
Having established that DC immunotherapy had the potential to induce a protective tumor-specific immune response, we next investigated whether DCs given after tumor challenge had the capacity to eliminate or slow down tumor growth. When DC immunotherapy was given 1 and 8 days after tumor implantation, protection from tumor growth was dependent on the way in which DCs were generated. The DC subtype I (grown from lineage-negative precursors in GM-CSF and Flt3-L) was effective in 66% of mice, whereas DCs grown in GM-CSF and generated from lineage-negative cells (DC subtype II) or from whole bone marrow cells (DC subtype III) protected all mice from tumor growth. It therefore seems that addition of Flt3-L reduced the efficacy of DCs to reduce established tumor growth. Other studies demonstrated that cytokines GM-CSF and Flt3-L, and a combination of these, influence the heterogeneity of DCs generated from bone marrow cells (33, 74–76). The in vitro administration of the cytokine Flt3-L to bone marrow progenitor cultures generates immature plasmacytoid-like DCs, which could induce tolerogenic effects on T cells (77, 78). Although our phenotypic analysis of cultured DCs did not show dramatic differences in the level of expression of the costimulatory molecules CD40, CD80, or CD86, it is possible that addition of Flt3-L still induced subtle differences in the immunostimulatory potential of DCs used for immunotherapy. In particular, one possibility that we have not explored is whether different DC subsets might directly influence tumor growth by producing cytokines or chemokines with direct antitumoral activity (e.g., IL-12, interferon-inducible protein–(IP)-10, monoxine induced by In contrast to the curative effect when tumor lysate–pulsed DCs were given 1 and 8 days after tumor challenge, a poor prognosis occurred when tumor cells and DCs were injected simultaneously via the peritoneal route. The observation of a paradoxic tumor-enhancing effect of simultaneous administration of DCs and tumor cells is not without precedent and may be caused by several factors. First, high levels of cytokines or soluble mediators produced by MM cells could downregulate cellular immune responses induced by DCs. Next, tumor cells might cluster with DCs, which, through their highly motile nature, might lead to more widespread dissemination and attachment of cancer cells to the mesothelial surface. Finally, and most interesting, it was recently shown in experiments where DCs were mixed with tumor cells in vivo that DCs can transform into endothelial cells, thus enhancing tumor vasculogenesis and tumor growth (79). To make DC immunotherapy clinically applicable, an easily accessible source of tumor antigen is an absolute requisite. Therefore, we analyzed whether extracts made from in vivo established tumors would be as efficacious as the primary AB1 cell line lysate to load DCs. Furthermore, growth of the AB1 tumor in vivo under the selective pressure of the immune system might have selected loss-of-antigen variants during the immuno-editing phase of tumor growth (80). However, DCs loaded with ex vivo–isolated tumor lysate were as efficacious as AB1 cell lysate–loaded DCs in mediating protection from tumor outgrowth. In a clinical setting, tumor cells for preparing a lysate might be obtained from surgical resection specimens, from thoracoscopic biopsy material, and from cancer cells isolated from pleural effusions. Another source of tumor antigen that has received great attention lately is exosomes. Exosomes are endosomal-derived vesicles that are secreted by almost all nucleated cell types, including tumor cells, B cells, and DCs. Seminal articles by Wolfers, Chaput, and Andre and colleagues have demonstrated that tumor-derived exosomes are a rich source of shared tumor antigens that can be used for loading onto clinical grade DCs (81–83). Several early-phase vaccine trials involving exosomes as a source of tumor antigen are currently underway. We have recently shown that MM cell lines secrete exosomes, rich in heat shock proteins 70 and 90, possibly explaining why exosomes are a source of tumor-derived antigens (25). Moreover, the pleural fluid of patients with MM contains large amounts of exosomes, which could potentially serve as a source of tumor antigen in a clinical trial, similarly to what was described for ascites fluid in ovarium carcinoma (84). When DCs were loaded with AB1 cell line–derived exosomes and injected after tumor implantation, they were able to reduce tumor outgrowth, illustrating the feasibility of such an approach. In our studies evaluating the therapeutic efficacy of tumor lysate–loaded DCs given after a larger tumor challenge, MM had a better outcome when DCs were injected early in tumor development, indicating that tumor load played an important role in survival. Although the potency of immunotherapy treatment decreased when DCs were injected later, mice still showed an improved prognosis compared with no treatment, but eventually tumors escaped immune surveillance and all mice died. It is now well established that larger tumor mass is associated with an immunosuppressive milieu that has the capacity to suppress the effector arm of the antitumoral immune response (CTL response inside the tumor) and the inductive arm of the immune response (i.e., the potential of antigen-presenting DCs to induce CTL responses). To prove that CTL induced by DCs were mediating antitumor responses, the efficacy of adoptive transfer of splenocytes and CD8 cells was evaluated in the prevention and treatment of MM (Figure 8). Intravenously injected, purified CD8+ T cells from protected mice dramatically increased survival compared with CD8+ T cells from naive mice when given 7 days before and, to a lesser extent, 2 days after AB1 injection. The suppressive tumor microenvironment might explain the decreased efficacy when CD8 cells are given after tumor administration. Splenocytes from protected mice increased the survival of mice when injected 7 days before AB1 injection but not 2 days after tumor injection. In both cases, 10 x 106 splenocytes were injected. The percentage of CD8+ T cells in this splenocyte preparation was between 2 and 7%, mounting to 2 x 105 to 7 x 105 CD8 cells. Approximately seven times more purified CD8+ T cells were used in the purified CD8 fraction experiment. The lower amount of effector CD8+ T cells in the splenocyte preparation in combination with the immunosuppressive effect exerted by the tumor can explain the difference in efficacy. Another explanation why splenocytes are less effective would be the presence of T cells with regulatory capacity (regulatory T cells) within the splenocyte inoculum. We have preliminary evidence that a population of CD4+CD25+ regulatory T cells enhances mesothelioma outgrowth by suppressing the antitumoral immune response (J.P.J.J.H., unpublished observations). In our most recent studies on human MM, we have found that tumors are infiltrated with large numbers of CD4 and CD8 T cells, yet tumors escape immune destruction (unpublished observations). Preliminary data on microarray RNA expression profiles and proteomic expression suggest that high levels of several immunomodulatory substances and chemokines, which may interfere with the maturation and/or function of DCs, are also present in the supernatant of MM cell lines and the corresponding patient's pleural fluid (manuscript in preparation). As shown for other tumors, mesothelial tumors produce a number of regulatory factors (e.g., vascular endothelial growth factor, IL-6, IL-10, macrophage-colony stimulating factor, prostaglandin E2, and transforming growth factor ß) that effectively suppress the function of DCs (49–51, 85). Understanding the multiple factors that come into play at different time points of the treatment process may help to better understand and design immunotherapy protocols. In conclusion, we demonstrate in this article the potency of DC immunotherapy in the control of MM outgrowth. Our study should pave the way to a clinical trial addressing the safety and feasibility of using tumor lysate–pulsed or exosome-pulsed DCs to induce tumor-specific CTL responses in patients with MM. Although such a trial will be initiated first in patients with end-stage disease after chemotherapy, we hypothesize that DC-based immunotherapy would have its greatest effect when given at times when tumor load is minimal—for example, after extrapleural pneumonectomy. Finding the optimal conditions to deliver clinical DC immunotherapy to patients with MM will be our challenge for the future.
The authors thank Monique Willart and Thomas Soullié for their technical assistance during our mouse experiments, and Tanja Nikolic for her assistance in the statistical analysis. Joke Zuijderwijk is thanked for advice and technical assistance in ELISPOT studies and Kris Thielemans for providing GM-CSF.
Supported by the Stichting Asbestkanker Rotterdam and the Mesothelioma Applied Research Foundation. B.N.L. is supported by a VIDI grant from the Dutch Organization for Scientific Research (NWO). This article has an online supplement, which is accessible from this issue's table of contents at www.atsjournals.org Conflict of Interest Statement: J.P.J.J.H. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript; A.H. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript; J.G.A. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript; H.C.H. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript; B.N.L. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript.
* These authors contributed equally to this article. Received in original form January 13, 2005; accepted in final form February 28, 2005
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
METHODS
