Cancer immune escape: implications for immunotherapy, Granada, Spain, October 3–5, 2011

Since 1985, a series of international conferences have taken place in Granada on the role of MHC class I antigens in cancer immune escape. This latest meetings, sponsored by the Ramon Areces Foundation, University of Granada, and University Hospital Virgen de las Nieves, were focused on research to improve our understanding of the molecular mechanisms of tumor evasion from immune surveillance and their implications for cancer immunotherapy. The outcome of complex interactions between the tumor and host immune system leads to either tumor progression or rejection. According to the theory of cancer immunosurveillance, if the immune system is able to keep cancer growth under control, the tumor can eventually be rejected. However, tumor cells can escape if the immunologic equilibrium is not maintained, leading to tumor growth and metastatic progression. It is becoming evident that malignant cells can develop sophisticated mechanisms of immune escape that limit the effectiveness of immunomodulatory cancer treatments. Cancer immunotherapies designed to stimulate a powerful anti-tumor immune response by augmenting tumor immunogenicity and decreasing tumor-induced immunosuppression are being very actively introduced into the clinic setting. However, although the clinical responses observed with these immunotherapies have been encouraging, they remain far from satisfactory. The success of immunotherapy depends on overcoming cancer immune escape caused by a dysfunction of the host immune system and/or by the immunosuppressive effect of the tumor microenvironment. One of the central mechanisms of immune evasion is associated with immune selection that favors the outgrowth of human leukocyte antigen (HLA) class I-negative tumor cells. In addition to intrinsic immune escape mechanisms, malignant cells induce activation of various immunosuppressive mechanisms in the tumor microenvironment to protect against immune reactivity. It is, therefore, very important and timely to explore new ways of overcoming cancer resistance to therapy and to improve our understanding of the molecular pathways involved in tumor-host interactions and tumor-induced suppression.

This report summarizes the main topics addressed at the conference, including: the role of tumor cell immunogenicity and the tumor microenvironment; novel concepts and strategies in cancer immunotherapy; and the optimization of existing cancer vaccines and immunotherapy protocols.

Targeting the tumor microenvironment, pre-clinical studies

Suppressive factors produced by tumors or tumor-associated cells pose a major obstacle to immune-mediated tumor eradication and immunotherapy. It is essential to target cancer cells, but this may select for escape variants and is usually inadequate to eradicate the tumor. Furthermore, the elimination of tumor-induced immunosuppression alone may not be sufficient to inhibit malignant progression. Several speakers at the meeting described the need to combine therapies for effective tumor eradication, based on pre-clinical studies. All cancer cells depend for their growth and survival on the development of a supportive stroma framework, including fibroblasts, extracellular matrix, blood vessels, macrophages, and immune cells, among others. The negative role of the altered tumor vasculature in the success of cancer immunotherapy and the importance of overcoming tumor endothelial barriers was addressed by Günter Hämmerling [German Cancer Center (DKFZ), Heidelberg, Germany]. He suggested that successful clinical cancer immunotherapy requires a combination of vaccination or adoptive T-cell transfer with approaches that modulate the tumor microenvironment, thereby enhancing T-cell infiltration. Experiments with Rip.Tag mice (autochthonous insulinomas) revealed that the aberrant tumor endothelium forms a barrier against infiltration by tumor-specific lymphocytes. However, activation of the endothelia and efficient infiltration of effector cells into the tumor tissue, leading to tumor eradication, has been achieved by inducing an inflammatory environment in the tumor by irradiation or by immunostimulatory CpG oligonucleotides. Gene expression profiling identified RGS5 (regulator of G protein signaling 5), which is overexpressed in the tumor vasculature, as a master gene involved in this barrier function. Tumors developing in RGS5 knockout mice displayed a normalized tumor vasculature with high T-cell infiltration resulting in tumor eradication. On the other hand, increased T-cell infiltration also resulted in higher numbers of infiltrating regulatory T cells. It was demonstrated that efficient (90%) Treg depletion in Foxp3.Luci.DTR mice led to tumor-specific T-cell activation, strong T-cell infiltration and tumor rejection, and normalization of the tumor vasculature.

It has become increasingly evident that cancer immunotherapy is less effective in established tumors, in both patients and experimental models. Hans Schreiber (University of Chicago, USA) underscored the importance of targeting tumor stroma in the tumor microenvironment as a means of preventing immune escape. The targeting or elimination of cancer cells may not be sufficient for tumor eradication. Stromal cells pick up and present antigens released from the cancer cells and thereby also become targets for T cells. Interestingly, the exclusive targeting of stromal cells achieves long-term growth equilibrium in large established tumors and does not select for heritable escape variants; however, cancers are not eradicated under these conditions because an extremely thin rim of cancer cells survives at the tumor margins, oxygenated by the pre-existing vasculature of adjacent normal tissues. It was observed that the host immune system can defeat a challenge from lethal numbers of highly aggressive cancer cells solely through stromal targeting. However, successful elimination required the cooperation of CD4+ and CD8+ T cells. Therefore, the targeting of tumor stroma with T cells is a key to prevent immune escape in large established tumors and to achieve a cure. In contrast, experimental data presented by Thomas Blankenstein [Max-Delbrück-Center for Molecular Medicine (MDC), Berlin, Germany] suggested that tumor-infiltrating lymphocytes may reflect cancer-induced inflammation rather than immunosurveillance. Results obtained from transgenic mouse models with spontaneous colorectal tumor development (Cre × LoxP-Tag) showed that sporadic immunogenic cancer at the time of initial recognition induces an aberrant rather than a protective T-cell response, resulting in tolerance at the pre-malignant stage, long before tumors become apparent. Hence, it has been proposed that cancer cells do not need to escape in a clinically relevant model and that general immune suppression is a late event that probably involves immature myeloid cells, requires antigenic tumors, and appears to be a symptom not a cause of tumor growth.

Michael Shurin (University of Pittsburgh Medical Center, USA) reported new data showing that tolerogenic regulatory dendritic cells (DCs), among other immune regulatory cells in the tumor microenvironment, could also be targeted in an attempt to overcome cancer-associated immunosuppression and improve the efficacy of cancer therapy at late, clinically relevant stages. He first demonstrated that certain antineoplastic chemotherapeutic agents directly upregulate the development, maturation, and functional activation of conventional DCs both in vitro and in vivo if used at ultra-low non-cytotoxic concentrations that do not induce cell death or inhibit the cell cycle. This new phenomenon has been called “chemomodulation”. Importantly, this treatment also decreases tumor-induced polarization of conventional DCs into immunosuppressive regulatory DCs in the tumor microenvironment. Chemomodulation also downregulates the formation and activity of myeloid-derived suppressor cells (MDSCs) and regulatory T cells. Interestingly, chemomodulation upregulates the expression of components of the antigen presenting machinery (APM) in tumor cells, thereby increasing their immunogenicity and recognition by tumor-specific T cells. Viktor Umansky [German Cancer Research Center (DKFZ), Heidelberg, Germany] demonstrated that inhibition of tumor MDSCs increases immunotherapy-induced anti-tumor immune responses in the Ret transgenic spontaneous murine melanoma model. In this study, the phosphodiesterase-5 inhibitor sildenafil reduced the production of inflammatory mediators and immunosuppressive factors in association with lower MDSC levels. It also restored ζ-chain expression in T cells and significantly improved the survival of treated mice. Treatment with very low non-cytotoxic and non-cytostatic doses of paclitaxel or with the proton pump inhibitor omeprazole also prolonged survival and partially normalized T-cell functions, suggesting that inhibitors of the immunosuppressive tumor microenvironment should be applied in conjunction with low dose chemo- and immunotherapies to increase their efficacy. The benefits of combined cancer therapy were also described by Esteban Celis (H. Lee Moffitt Cancer Center, Tampa, USA), who provided new insights from pre-clinical cancer vaccination studies. Powerful long-lasting T-cell responses were induced in mice by the administration of TriVax, a vaccine containing the following three therapeutic components: CD8 T cell-specific synthetic peptides, toll-like receptor (TLR) agonists and anti-CD40 antibody. Surprisingly, however, the success of peptide vaccination in mice was inhibited by interferon-gamma (IFNγ), which was explained by the induction of tumor expression of inhibitory PDL1 (B7H1) and by the increased tumor expression of non-cognate (irrelevant) major histocompatibility complex (MHC)—class I molecules, which interfere with the costimulatory activity of the CD8 molecules and limit the activation and effector functions of the cytotoxic T cells.

Gloria González-Aseguinolaza (CIMA, University of Navarra, Pamplona, Spain) reported that the long-term controlled expression of interleukin-12 (IL-12) using gutless adenovirus vector in combination with oxaliplatin has proven effective and clinically applicable for the treatment of hepatic colon cancer metastases. Oxaliplatin showed synergism with IL-12 in the stimulation of an efficient immune response against cancer cells, with the complete eradication of pre-established colorectal cancer liver metastases and the prevention of experimental relapses in a syngeneic model. The mechanism of this synergism involves a reduction in the immunosuppressive microenvironment of tumors and an increase in the ratio of CD8+/T regulatory cells to CD8+/myeloid-derived suppressor cells. However, recent data obtained in a clinically relevant animal model of chronic hepatitis B infection showed that long-term expression of IL-12 in an already immunosuppressive microenvironment significantly augments the expression of immunosuppressive molecules, increasing the risk of developing a liver tumor.

MHC class I loss and cancer immune escape: negative impact on the outcome of cancer immunotherapy

Several speakers highlighted the role of tumor HLA class I expression in immune-mediated cancer rejection and the negative effect of HLA loss on the clinical outcome of cancer immunotherapy. Federico Garrido (University of Granada and University Hospital Virgen de las Nieves, Spain) discussed the key role of MHC expression in tumor rejection, emphasizing the negative impact of altered HLA class I expression on cancer immunotherapy. Immune selection leads to the escape of HLA class I-negative tumor variants and to malignant growth and metastatic colonization. Dr. Garrido presented experimental and clinical evidence suggesting that cancer immunotherapy changes the tumor microenvironment, which then leads to the immunoselection of HLA class I-negative metastases that do not respond to therapy. A correlation was found between tumor HLA class I alterations and metastatic cancer progression in melanoma patients with autologous tumor vaccine and cytokines and in bladder carcinoma patients undergoing Bacillus Calmette-Guèrin (BCG) immunotherapy. It was observed in all of these cases that immunotherapy leads to the selection and outgrowth of HLA class I-negative tumors, mostly with irreversible alterations. Interestingly, whole genome transcriptional analysis of the melanoma metastases clearly attributed their regression to an acute immune rejection mediated by the upregulation of genes involved in antigen presentation and interferon-mediated responses (STAT-1/IRF-1). In contrast, progressing melanoma metastases showed low transcription levels of genes involved in these pathways. This can be explained by the heterogeneity of tumors with HLA class I loss, which can be divided into two groups, those that respond to therapy by upregulating HLA expression (“soft” lesions) and those with irreversible/structural defects (“hard” lesions), which are resistant to antigen-specific stimulation and will eventually progress. Therefore, the nature of pre-existing HLA class I lesions may be an important determinant of the final outcome of cancer immunotherapy. Natalia Aptsiauri (University Hospital Virgen de las Nieves, Granada, Spain) presented an example of the emergence of HLA class I-negative tumor escape variants in a patient with metastatic melanoma undergoing immunotherapy with tumor mRNA-transfected autologous DCs. Total loss of class I expression in a cell line established from a post-therapy metastatic lesion was caused by a mutation in the beta-2-microglobulin (B2M) gene and loss of heterozygosity (LOH) in chromosome 15. Adenoviral vector-mediated transfer of the wild-type human B2M gene recovered normal HLA class I expression, and a cytotoxic T-lymphocyte (CTL) test and enzyme-linked immunosorbent spot (ELISPOT) assay confirmed the functional activity of the restored cell surface HLA class I complex. Strikingly, the same B2M mutation and LOH in chromosome 15 were found in a HLA class-negative tumor nest microdissected from a metastatic lesion with a heterogeneous HLA class I immunolabeling pattern. Immunohistochemical analysis of the tumor and metastases showed mostly positive or heterogeneous patterns of HLA class I expression with the presence of HLA class I-negative nodules in the pre-vaccination lesion and the emergence of totally HLA class I-negative cells after vaccination. An increased presence of LOH in chromosomes 6 and 15 was observed in the successive metastases. Interestingly, LOH in chromosome 15 was an early event in metastatic progression. These data verify that the class I-negative tumor phenotype confers an advantage for malignant growth and dissemination, and they suggest that immunotherapy might lead to the generation of resistant metastatic lesions with irreversible defects in HLA class I antigens that will eventually progress. This case also demonstrates that it is possible to miss the consequences of immune evasion developed by the tumor before immunotherapy. This should be taken into account when immunologic treatment protocols are designed. The evaluation of HLA class I expression in primary tumors and metastases should be considered before the selection of cancer immunotherapy as a treatment option. Importantly, making tumors visible to the immune system represents a considerable challenge, which in many instances may require restoration of tumor HLA class I expression. Structural alterations underlying development of HLA class I-negative tumor phenotypes can be restored by gene therapy. In this context, we successfully used an adenoviral vector carrying the human B2M gene to recover normal HLA class I expression in B2M-negative tumor cells in vitro and in an in vivo model of human tumor xenogenic transplantation. Thorbald van Hall (Leiden University Medical Center, Leiden, the Netherlands) proposed an alternative approach for recruiting immune responses to tumor cells with MHC-I-defects and described a different possible target for CTLs, that is, novel tumor antigens that are selectively presented by tumor escape variants (TEIPP or T-cell epitopes associated with impaired peptide processing). These peptides are not presented by MHC I under normal conditions, but they emerge on the surface of tumor cells with alterations in HLA class I expression, notably those caused by defects in the TAP (transporter associated with antigen presenting) intracellular peptide transporter. A broad cytotoxic CD8+ T-cell population recognizes these as neo-epitopes in a complex with classical MHC-I molecules. The surprising absence of TEIPP peptides from normal cells was not related to weak binding or to the limited availability of MHC-I heavy chains but rather to expression levels of the antigen. Interestingly, a major fraction of TEIPP T cells are restricted by the non-classical MHC Qa-1b, which is the mouse homolog of HLA-E. This low-polymorphic molecule normally accommodates monomorphic signal peptides and functions as a ligand for the germ-line receptors CD94/NKG2 on NK cells. Finally, TEIPP T-cell responses can easily be detected in humans and help to reveal the identity of the first TEIPP detected in humans. Hence, the therapeutic application of classical and non-classical TEIPP antigens might enable the CTL-mediated elimination of tumors with MHC-I presentation defects by preventing this type of immune escape. Barbara Seliger (Martin Luther University Halle-Wittenberg, Germany) demonstrated that MHC downregulation caused by APM alterations plays a significant role in cancer immune escape by preventing the display of tumor-associated antigenic peptides. The downregulation or absence of different APM components may take place at different steps of the antigen-processing and presentation pathway and can be caused by structural alterations or by dysregulation, both at transcriptional and post-transcriptional levels. Some APM components, for example, tapasin, were shown to be dysregulated due to the altered binding of transcription factors. The post-transcriptional downregulation may be mediated by a higher proteasomal turnover of APM components and by protein instability. Experiments designed to identify TAP1-regulated genes suggested that the IFN-gamma transcription signal transduction pathway may be important for the proper expression of MHC class I APM components. Reduced HLA class I and APM component expression was demonstrated in JAK2-deficient IFN-gamma-resistant cells. Using oncogenic transformants as a model for MHC class I abnormalities, a transcriptional down-regulation of various APM promoters was found in HER-2/neu-overexpressing cells. Dr. Seliger also showed that the transcription factors E2F1 and CREB play an important regulatory role in APM component expression in tumors. Soldano Ferrone (University of Pittsburgh, USA) further discussed the importance of changes in HLA class I antigen expression by malignant cells. The malignant transformation of cells is often associated with changes in the expression of classical and non-classical HLA class I antigen, HLA class II antigen, and NK cell-activating ligand (NKCAL). Based on clinical and experimental evidence, it was previously assumed that the changed expression profile of HLA antigens on malignant cells represented loss of classical HLA class I antigen and induction of HLA class II antigen and non-classical HLA class I antigen expression. However, more recent findings have revealed that classical HLA class I antigens have a more restricted distribution and HLA class II antigens a broader distribution in normal tissues than was originally believed. Importantly, in some cases, pre-malignant or malignant cells can acquire classical HLA class I antigen expression and/or lose HLA class II antigen expression. This information may contribute to our understanding of the interactions of tumor cells with the host immune system and may assist the optimization of immunotherapeutic strategies for cancer treatment. Soldano Ferrone also presented novel data on the in vitro upregulation of melanoma antigens and HLA class I/class II molecules on tumor cells induced by a combination of mutant BRAF inhibitor and IFN-alpha. Annette Paschen (University Hospital Essen, Germany) gave an overview of tumor HLA class I loss associated with B2M mutations in melanoma cells and reported the case of a patient with progressive metastatic melanoma and variable HLA class I expression profile in different cell lines obtained from several lesions. Two HLA class I-negative cell lines that showed total HLA class I loss associated with B2M gene deletion were lysed by NKL cells more effectively than were HLA-positive cells. Dr. Paschen also demonstrated that these melanoma cell lines express MICA and ULBP2, which are ligands of the activating NK cell receptor NKG2D, providing experimental evidence of their importance in tumor cell killing by NK cells. She also found a downregulation of NKG2DL expression in the presence of IFN-gamma, protecting the tumor cells from NK cell cytotoxicity. Tumor cells may also escape from NKG2D-dependent immune responses by proteolytic ligand shedding. Increased levels of soluble MICA (sMICA) and sULBP2 were measured in sera from melanoma patients in comparison to healthy controls, and elevated sULBP2 was found to be an independent predictor of poor prognosis, even in early-stage melanoma. Regarding the regulation of NKG2DL in melanoma cells, Dr. Paschen demonstrated that the tumor-suppressive microRNAs miR-34a and miR-34c as well as p53 are involved in the control of ULBP2 expression, which strengthens the role of this specific NKG2DL in anti-tumor immune responses. Matthias Kloor (University Hospital Heidelberg, Germany) addressed the immunoediting of microsatellite-unstable colorectal cancer (CRC). A subset of about 15% of colorectal cancers have a deficient DNA mismatch repair system and accumulate small mutations at repetitive DNA sequences, a phenotype termed as “high level microsatellite instability” (MSI-H). Mutations of the B2M gene are frequent in MSI-H CRCs. However, in MSI-H CRC, unlike other types of cancer such as melanoma, B2M alterations are associated with the absence of distant metastases and a favorable outcome. However, MSI-H CRC patients with disease stage II and III frequently show resistance to standard chemotherapy. MSI-H colorectal cancers have a distinct histopathological appearance, characterized by poor differentiation and a locally advanced tumor stage. A typical feature of MSI-H colorectal cancer is dense infiltration with intratumoral lymphocytes, suggesting that the host antitumoral immune response plays an important role in this disease. It has been suggested that the immunogenicity of MSH-I CRC may be caused by MMR deficiency-induced Frameshift Peptides (FSPs). A systematic identification of FSP antigens, based on the human genome database, has been done to determine whether FSPs trigger an antigen-specific immune response. Tumor-infiltrating lymphocytes (TILs) recognize predicted FSP antigens in vitro, and FSP-specific T-cell responses are also detectable in peripheral blood drawn from CRC patients. Based on this information, candidate peptides have been selected for the vaccination of MSI-H CRC patients. A clinical trial was recently started to evaluate FSP-based vaccination in the clinical setting.

Tumor immune escape and cancer immunotherapy/vaccination in patients

Insights from pre-clinical studies combined with immunomonitoring data obtained from patients undergoing cancer immunotherapy are expected to improve and optimize current immunotherapy protocols. Pierre Coulie (University of Louvain, Brussels, Belgium) discussed the clinical significance of the presence and activity of T lymphocytes in melanoma metastases in patients with clinical responses after vaccination (phase I/II, using tumor-specific antigens administered as peptides, proteins, DCs pulsed with peptides, or recombinant poxviruses). T-cell response analyses indicated that the blood levels of anti-vaccine CTLs were very low, even in patients displaying tumor regression. Tumor regression was associated with activation of CTLs directed against tumor antigens not present in the vaccine. These T cells could be detected even before the vaccination. In addition, new CTLs appeared after vaccination as a result of clonal and antigen spreading. Dr. Coulie suggested that melanoma patients spontaneously mount anti-tumor CTL responses that fail to reject the tumor due to local immunosuppression or reduced tumor antigen expression. Vaccination may activate a small number of anti-vaccine CTLs, which may relieve suppression after infiltrating the tumor and promote activation of many other anti-tumor CTLs that contribute to tumor regression. Comparative gene expression analysis showed no difference between pre-vaccination cutaneous metastases from melanoma patients who showed either complete tumor regression or no regression following vaccination with MAGE tumor antigens (MAGE-A3 peptides administered alone, or recombinant canarypoxviruses encoding MAGE-A1 and MAGE-A3 antigenic peptides). However, a difference in T-cell and macrophage signatures was observed between these groups, including IFNγ target genes and the IFNγ transcript itself. Using immunohistology on consecutive tumor sections, they established that this inflammatory signature correlates with the degree of immune cell infiltration. Therefore, melanoma metastases host various degrees of active Th1 inflammation, and it was concluded that the immunosuppressive environment in these tumors does not result in a complete inhibition of T-cell activation.

One of the mechanisms of tumor escape from immunity may be a process of tolerization caused by exposure of the specific reactive T cells to the chronic antigenic stress represented by tumor antigens. This tolerization or anergy is also observed in elderly people without cancer but with latent infections, especially with β Herpesvirus HHV5 (cytomegalovirus, CMV), which is persistent and cannot be eliminated by the immune system. Decreased immunity with age may contribute to increased cancer incidence in the elderly, because they are immunocompromised with few naïve cells and dysfunctional (exhausted) memory cells, due to chronic antigenic stress and thymic involution. Graham Pawelec (University of Tübingen, Germany) addressed the role of immunosenescence in tumor immune escape and presented immunomonitoring data from a study of melanoma patients treated with anti-CTLA-4 antibody and from a pilot RNA vaccination trial (RNA vaccine encoding the antigens NY-ESO1, Survivin, MAGE-A3 and Melan-A). In this study, long-term survival in stage IV melanoma was associated with the presence of pro-inflammatory NY-ESO-1-specific CD4 and CD8 T-cell responses. Extended survival was only associated with CD8 T-cell responses to Melan-A, and survival was not associated with responses to MAGE-A3 or Survivin. In addition, a correlation was observed between NY-ESO1 response and CMV status in the long-term survival group (>18 months). This suggests that the measurement of CMV seropositivity should be included in future immunomonitoring. Cornelis J.M. Melief (Leiden University and ISA Pharmaceuticals, Leiden, the Netherlands) outlined different types of therapeutic vaccines for HPV6 and discussed the immunotherapy of established cancer caused by high-risk human papillomavirus. It was shown that short-peptide vaccines are not effective in therapeutic vaccines against HPV16-induced cancer, while long-peptide vaccines that harbor both CD4 and CD8 T-cell epitopes and require DC processing are effective. Further improvements are possible by adding TLR ligands or by conjugating TLR ligands to the long peptides. Clinical data demonstrated that treatment with the HPV16 SLP (synthetic long peptide) vaccine is effective in patients with established vulvar intraepithelial neoplasia (VIN). In patients with cancer, it is attractive to combine this type of vaccination with immunogenic forms of cancer chemotherapy and with immunomodulatory drugs. Dr. Melief suggested developing combination treatments with long-peptide vaccination, immunogenic chemotherapy, and checkpoint control (e.g., CTLA-4 blocker, PD-1, PD-L1 blockers) for maximally effective cancer treatment. He also suggested the local delivery of mAb treatments in slow-release formulations close to tumor-draining lymph nodes in order to reduce toxicity. Isabel Poschke (Cancer Center Karolinska, Stockholm, Sweden) showed that tumor-dependent phenotypic and functional alterations in lymphocytes can be detected even in early-stage breast cancer patients, including the pronounced differentiation of circulating T cells paralleled by modified homing capacity in peripheral blood. Intra-tumor T cells exhibited signs of both activation and exhaustion. ζ-chain loss was highest in the tumor and decreased with greater distance from it. T-cell ζ-chain expression in peripheral blood of breast cancer patients was downregulated in comparison to healthy controls but was normalized after surgical tumor resection. ζ-chain expression in peripheral blood T cells correlated with lymphocytic proliferative activity and with the expression of pro-inflammatory cytokines in CD8+ T cells. Rolf Kiessling (Karolinska Institutet, Stockholm, Sweden) reported that oxidative stress plays an important role in tumor-induced immune suppression. Cancer cells can produce significant amounts of ROS and recruit or induce ROS-producing cells (MDSCs). Increased Treg accumulation in cancer may result from aberrant proliferation and trafficking as well as a greater resilience to oxidative stress compared with conventional T cells. This enhanced antioxidative capacity of Tregs possibly serves as feedback inhibition during inflammation and prevents uncontrolled immune reactions by favoring survival of suppressor rather than effector cells. Human Tregs express and secrete higher levels of thioredoxin-1, a major antioxidant molecule. Thioredoxin-1 has an essential role in maintaining surface thiol density as the first line of antioxidative defense mechanisms and is sensitive to proinflammatory stimuli, mainly tumor necrosis factor-α, in a nuclear factor-κB-dependent fashion. The antiapoptotic and oncogenic potential of (secreted) Trx-1 suggests that it may exert effects on Tregs beyond redox regulation. T and NK cells are more sensitive to oxidative stress. ROS have an immunosuppressive effect on T and NK cells (inhibit NF-kB nuclear translocation, lead to loss of the CD3ζ chain, decrease proliferation, cytotoxicity, and cytokine production, and induce apoptosis) and impair DC maturation. The number of circulating CD14 (+) HLA-DR (−/low) MDSC was increased in a group of melanoma patients in active stage III/IV. These cells have strong suppressive effects on T-cell proliferation and IFN-gamma production. Arginase 1 and oxidative stress contribute to suppression. These MDSCs overexpress activated STAT3, and their activity is suppressed by STAT3 inhibition. All these data suggest the importance of developing treatments to counteract MDSCs or their products or to design ROS-resistant T cells, including transfer of the catalase gene to T cells or the use of high-dose antioxidant (vitamin E) therapy in patients with colorectal carcinoma. A combination of antioxidative treatment and immunotherapy represents a promising approach. Michele Maio (University Hospital of Siena, Italy) described the epigenetic remodeling of cancer cells as a novel strategy to counteract tumor immune escape. Epigenetic alterations are associated with various routes of cancer immune escape through changes in the expression of tumor-associated antigens (TAAs) currently utilized as therapeutic targets and alterations in the expression of HLA antigens and accessory/co-stimulatory molecules, which are all required for the effective immunologic recognition of cancer cells. Investigation of the immunomodulatory activities of epigenetic drugs (e.g., DNA hypomethylating agents, DHA) demonstrated that in vitro and in vivo exposure of different neoplastic cells to the prototype DHA 5-aza-2′-deoxycytidine (5-AZA-CdR) modulated their constitutive expression of HLA class I antigens, accessory/co-stimulatory molecules, and different TAAs, especially cancer/testis antigens (CTA). These changes significantly improved the recognition of cancer cells by antigen-specific CTLs. Impressively, a single course of 5-AZA-CdR induced persistent de novo expression of CTA MAGE-A1, SSX 1-5, and NY-ESO-1 in mononuclear neoplastic cells from the peripheral blood and bone marrow of patients with acute myeloid leukemia or myelodysplastic syndrome. “Epigenetic chemoimmunotherapy” was suggested as a future approach that will combine DHA immunomodulation and other types of therapy, such as immunomodulatory antibodies, CTA-based vaccines, or BRAF inhibitors. Francesco Marincola (National Institutes of Health, Bethesda, USA) presented an interesting concept of the “immunologic constant of rejection”, based on the similarity of transcriptional signatures observed in allograft rejection and tumor destruction, in graft-versus-host disease, in acute clearance of pathogens, and in autoimmunity. This “constant” includes the coordinated activation of interferon-stimulated genes (ISGs) and immune effector functions (IEFs). Dr. Marincola discussed the significance of transcriptional signatures observed in pre-treatment biopsies as predictive of responsiveness to biological therapy. He also illustrated how the transcriptional signatures observable during and after therapy document the switch from chronic to acute inflammation that leads to tumor rejection. A better understanding of the mechanisms underlying the change from chronic inflammatory processes to a tumor-destroying acute reaction may guide the development of novel therapies.

Concluding remarks

This international conference in Granada featured dynamic presentations focused on the impact of tumor immune escape on the clinical outcome of cancer immunotherapy. In particular, it addressed the importance of defining tumor escape variants and the need to improve understanding of cancer immune escape mechanisms and to monitor key biomarkers of cancer rejection or resistance to therapy.

The main obstacle to successful immunotherapy is the ability of tumors to evolve into resistant escape variants under the pressure of immune selection. Exciting new studies reported at the conference led the speakers to propose a variety of novel and promising approaches to the optimization of cancer treatment. These include the exploration of new targets for recognition by CTLs (frameshift peptides or T-cell epitopes associated with impaired peptide processing selectively presented by tumor escape variants), the targeting of tumor stroma, the application of angiogenesis inhibitors, the use of combination therapy using cytokines together with inhibitors of oncogenic signaling pathways (e.g., mutant BRAF inhibitors), the use of antioxidants, and the combination of chemo-immunomodulation with epigenetic remodeling. The significance of a T-cell dependent antigen-specific HLA class-I-restricted anti-tumor immure response is undeniable. Hence, all attempts to increase CTL-induced tumor immunogenicity and rejection remain of special interest. An attractive idea in this context is the application of a gene therapy aimed at recovery of HLA class I expression in tumor cells with “hard” alterations, which are believed to be responsible for metastatic progression in patients with a mixed response to therapy.

Tumor rejection is known to involve multiple mechanisms and factors that are responsible for cancer regression. However, the key factor(s) and common immune signatures responsible for coordinating tumor rejection are not yet well defined. Accumulating evidence from the molecular analysis of cancer metastases before and after immunotherapy has helped to identify molecular pathways activated during immune-mediated tumor rejection. These pathways appear similar to those observed in other forms of immune-mediated tissue-specific rejection, such as allograft rejection, pathogen clearance, graft-versus-host disease, and autoimmune disease, supporting the existence of the so-called immunologic constant of rejection, which includes the coordinated activation of interferon-stimulated genes and immune effector functions. Identification and characterization of the main driving force in tumor rejection responsible for switching from tolerance to tumor regression will be increasingly important in future efforts to develop novel strategies to counter tumor resistance mechanisms. Finally, clinical outcomes can be improved by the early detection of cancer and the combined application of chemotherapy, immunotherapy, and gene therapy to augment tumor immunogenicity and reduce tumor-induced immunosuppression.