Abstract
Blockade of immune checkpoints in cancer immunotherapy have astonished the oncology field with durable clinical responses in different cancers. Yet, some patients do not respond to currently available checkpoint inhibitors and, therefore, identification of new potential immunotherapeutic targets that might apply in these patients remains important. The inhibitory receptor ILT4 may represent such a promising target. Thus far, various roles of ILT4 have been described, nevertheless its role in tumour immunology remains unknown. In this review we will speculate on its role by reviewing currently available literature. We suggest, that in the tumour microenvironment, where in addition to tolerogenic DC also tumour cells themselves can express ILT4 as well as its ligand HLA-G and can secrete high levels of IL-10, this could lead to suppressive pressure on the anti-tumour immune response, providing incentive to study the ILT4/HLA-G axis in immunosuppression in cancer.
Introduction
Cancer is emerging as one the leading causes of morbidity and mortality worldwide, with approximately 14 million newly diagnosed patients and 8.2 million cancer related deaths in 20121. While the human immune system can mount cytotoxic immune responses that can eradicate tumour cells2, immune tolerance induced by cancers can limit the effectiveness of the immune response3. Immune tolerance describes a process in which a cancer changes its phenotype to protect itself from being killed by tumour-specific lymphocytes2. Tumour cells thrive in immunosuppressed environments. More precisely, tumours themselves may create localized microenvironments in which immune function is compromised. In many cancer patients, there is evidence of reduced levels of some types of circulating effector lymphocytes, indicating a systemic defect in immune function. Human cancer cells can release either transforming growth factor (TGF)- or interleukin (IL)-10, both potently immunosuppressive, i.e. these proteins inhibit T cell growth and division, avoid dendritic cell (DC) maturation, and suppress expression of major histocompatibility complex (MHC) type II molecules. This immunoevasive strategy of cancer cells might modify the mix of immune cells around them as well, for instance the tumour environment can be dominated by regulatory T (Treg) cells. The induction of an immunosuppressive tumour microenvironment may frustrate antigen-presenting cells (APC) in inducing an anti-tumour response4.
Cancer immunotherapy aims to break immune tolerance and has demonstrated promising results in cancer patients5,6. According to the point where the intervention takes place, immune-based therapies can be classified into (1) adoptive transfer of effectors, (2) vaccination, (3) and immunomodulatory therapy (3). Firstly, tumour-specific antibodies and T cells can be expanded to large numbers ex vivo, thereby bypassing the need for antigen presentation and proliferation in vivo, also known as passive immunotherapy, i.e. no immunological memory will be induced. Secondly, the patient’s immune system can be supplied with either whole-tumour cells, tumour antigens, or peptides, either injected directly into the patient, or loaded on ex vivo cultured DC. However, the efficacy of cancer vaccines can be limited by the process of immune editing, suppression of antigen processing and presentation, and difficulty maintaining a robust immune response at the site of the tumour. Lastly, immunomodulatory therapy aims to fight cell populations in the tumour microenvironment that lead to local immunosuppression, e.g. myeloid-derived suppressor cells (MDSC) and Treg cells, or aims to re-activate tumour-suppressed effector cells as will be further discussed below3.
Checkpoint inhibitors
Aforementioned immunomodulatory therapy also includes checkpoint blockade, as immune checkpoints do not only tightly regulate T cell responses to avoid autoimmunity, but can also limit effectiveness of anti-tumour immune responses3. Antibodies directed against the immune checkpoints cytotoxic T-lymphocyte-associated antigen-4 (CTLA-4) and programmed death (ligand)-1 (PD-(L)1) are the most well known checkpoint blockade-based therapies6.
CTLA-4 (also known as CD152) belongs to the immunoglobulin superfamily (IgSF). Whereas its expression is transiently upregulated on activated T cells, CTLA4 is constitutively expressed on Treg cells, thereby playing a crucial role in immune control and immune tolerance7. CTLA-4 competes to bind to its cognate ligands CD80/CD86 on APC or tumour cells, thereby preventing ligation of CD80/CD86 with CD28, which is a co-stimulatory activating receptor. This leads to T cell motility attenuating T cell activation6. Also, CTLA-4 expressing Treg cells can inhibit T cell activation by downregulating CD80/CD86 on APC8,9. It is believed that CTLA-4 plays a major role during the priming phase of tumour-reactive T cells by APC6.
Similarly to CTLA-4, PD-1 (also known as CD279) is constitutively expressed on Treg cells and can be, upon activation, expressed on effector T cells, natural killer (NK) cells, NK T (NKT) cells, and B cells7,10. In addition, PD-1 is highly expressed on exhausted T cells7 and blockade of PD-1 was found to restore effector T cell functions11. PD-L1 (also known as B7-H1 or CD274) is the major PD1 ligand, and is expressed on lymphocytes, DC, and macrophages10. Moreover, PD-L1 has been found on several tumour types, but also on myeloid cells present in the tumour microenvironment6, with expression associated with poor clinical outcome for melanoma12, breast cancer13, and ovarian cancer14. Expression of the minor ligand PD-L2 (also known as B7-H2 or CD273) is restricted to DC and macrophages, but is also present in leukaemia and B cell lymphomas10. Ligation of PD-L1/2 with PD-1 rather coinhibits preactivated T cells by inhibiting signalling via the T cell receptor (TCR) and thereby creates immune resistance by functionally inactivating the tumour-infiltrating T cells2,6.
Ipilimumab (anti-CTLA-4 mAb) and nivolumab (anti-PD-1 mAb) have astonished the oncology field with durable clinical responses in different types of cancer15,16. Firstly, Hodi et al. (2010) showed improved overall survival in metastatic melanoma patients after ipilimumab treatment17. Similarly, long-term survival was observed in patients with advanced melanoma receiving nivolumab18. Likewise, Brahmer and Pardoll (2013) reported durable and significant responses in patients with non-small cell lung cancer (NSCLC) induced by therapeutic treatment with ipilimumab19, which provided a gateway to the development of therapies for different cancer types which were previously believed not to respond to immunomodulatory therapies. Ipilimumab also caused immune-mediated tumour regression in metastatic renal cell cancer patients20. Yet, some patients do not benefit from currently available checkpoint inhibitors, e.g. due to reduced infiltration of effector T cells within the tumour to be unleashed upon anti-CTLA-4 or anti-PD-1 therapy. Therefore, identification of additional immune checkpoints that might apply in these patients and that are able to reduce not only the levels of tumour-released suppressive cytokines, but also the tumour-induced development of suppressive T cell subsets, is a promising line of investigation.
In this review, we discuss a new promising immune checkpoint pathway, i.e. the immunoglobulin-like transcript 4 (ILT4)-Human Leukocyte Antigen G (HLA-G) pathway. Expression of ILT4 is not restricted to tumour cells merely, but can be found on APC and tumour-infiltrating lymphocytes (TIL) as well21,22. Within the tumour microenvironment, this could lead to an abundance of suppressive pressure on the anti-tumour immune response. Therefore, ILT4 might be of clinical relevance, as blocking it can not only inhibit tumour cells, but tolerogenic DC as well, similarly to anti-PD-(L)1 treatment. This provides incentive to study the role of ILT4/HLA-G pathway in immunosuppression in cancer.
ILT receptors
The immunoglobulin-like transcript (ILT) family constitutes transmembrane, cell-surface protein receptors, preferentially expressed on myeloid lineage cells, and encompasses activating receptors (ILT1, ILT6, ILT7, and ILT9) and inhibitory receptors (ILT2, ILT3, ILT4, and ILT5)23. The latter group of receptors have longer cytoplasmic domains, containing immunoreceptor tyrosine based inhibitory motif (ITIM)-related sequences. The group of activating receptors, however, possess short cytoplasmic domains and lack aforementioned signalling motifs. Yet, ILT6 has been described to lack a transmembrane region and might be a secreted protein24. While ligation of activating ILT receptors by their ligand results in transduction of activation signals to cells, the immune inhibitory ILT receptors are potent inhibitors of cell activation, for instance in DC, and play a role in modulating T cell responses24,25. Thereby, these receptors exert critical functions in balancing between immune activation and immune inhibition. Whereas silencing of ILT inhibitory receptor expression leads to inappropriate activity of effector immune cells, overexpression may create an immunosuppressive environment26.
ILT4/HLA-G axis
The inhibitory receptor ILT4 has received less attention thus far compared to CTLA-4 and PD-L1, but deserves exploration as a therapeutic target in cancer immunotherapy. Similarly to PD-L1, ILT4 can be expressed on tolerogenic DC21 as well as on tumour cells22 and its ligands HLA-G and MHC-I/CD8 can be expressed on tumour cells, DC and T cells25,27. Thus far, various roles of ILT4 have been described, yet its precise role in tumour immunology remains unknown. Here we will speculate on its role by reviewing the currently available literature. Gregori et al. (2010) studied the interaction between ILT4 and its ligand HLA-G and showed this to be required for the IL-10 dependent differentiation of type 1 T regulatory (Tr1) cells25. ILT4 has also been described to be involved in CD8 T cell priming. It was shown that dermal CD14+ DC primed naïve CD8+ T cells into type 2 cytokinesecreting CD8+ T cells, presumably by binding of ILT4 to its ligand CD8, thereby preventing binding of the TCR/CD8 complex to MHC-I/peptide complexes. Blocking ILT4 on dermal CD14+ DC enhanced the generation of cytotoxic T lymphocytes and inhibited the production of type 2 cytokines. Adding either soluble ILT2 or ILT4 receptors to these cultures again caused inhibition of CTL priming28. This indicates ILT inhibitory receptors influence CD8+ T cell responses and might be of clinical relevance for the treatment of malignancies, as effective anti-tumour immune responses require proper CD8 T cell priming.
ILT4’s main ligand HLA-G represents a tolerogenic and nonclassical MHC class 1 molecule with immunomodulatory properties, such as inducing tolerance and inhibiting immune responses29,30. As mentioned above, the expression of HLA-G is not restricted to tumour-infiltrating cells merely, but can be found on tumour cells as well25,27,30. HLA-G has been described to be of clinical relevance, as HLA-G expression is associated with malignant transformation and is not expressed on adjacent normal cells31,32. Moreover, in solid tumours its expression increases with progressing levels of dysplasia33,34. Finally, HLA-G has been proposed as a prognostic marker, since high expression has been significantly correlated with poor prognosis33–37. In this context, expression of tolerogenic HLA-G molecules represents a mechanism that may favour tumour survival through interaction with inhibitory receptors. HLA-G is expressed in different isoforms: membrane-bound HLA-G1, G2, G3, and G4, and soluble HLA-G5, G6, and G7. HLA-G1 and G5 are the most common isoforms of HLA-G that have been implicated in immune suppression. Alongside ILT4, HLA-G is also able to interact with ILT4’s family member ILT2, which is often expressed on immune effector cells such as T cells, B cells and NK cells. HLA-G expressing tumour cells can have detrimental effects on the development, activation status and functions of various immune effector cells, e.g. HLA-G can inhibit T and B cell proliferation, cytotoxic activity of cytotoxic T cells and NK cells, the potential of neutrophils to phagocytose, and the function of DC, via binding to its receptor ILT2/4. More tolerogenic consequences of HLA-G expressing tumour cells are depicted in figure 1, taken from a review by Rouas-Freiss et al (2014)30,38–51.
Upon ligation of HLA-G by ILT4, signal transducer and activator of transcription 3 (STAT3) becomes phosphorylated, leading to an enhanced production of the immune suppressive cytokines IL-6 and IL-10 by tumour cells52. Both phosphorylated STAT3 and IL-10 hamper the maturation and activation of DC. More importantly, these promote the development of tolerogenic CD14+ DC with low antigen-specific CD8+ T cell priming capacity and a propensity for expansion of suppressive FoxP3+ Treg cells53. Silencing of STAT3 prevented the generation of tolerogenic CD14+ DC, allowing for conventional DC development.
ILT4 can be expressed by tumour cells and is associated with poor clinical survival in several cancers.
As mentioned previously, ILT4 can be expressed on tumour cells22 and overexpression might indicate an immune suppressive microenvironment, favourable for the tumour to progress, which makes it an interesting target for anti-cancer therapy. Additionally, HLA-G has been proposed as a prognostic marker in correlating HLA-G expression to a clinical prognosis33–37. Its receptor ILT4, however, has received less attention thus far in clinical correlation studies, linking ILT4 expression in the tumour and in tumour-draining lymph nodes to poor clinical outcome. However, recent studies demonstrate a high expression of the inhibitory receptor ILT4 in breast cancers22, NSCLC23,54,55, and leukaemia56. Liu et al. (2014) assessed ILT4 and IL-10 expression in ductal and lobular breast cancer carcinomas. Both were expressed in breast cancer tissues and appeared to be absent in healthy tissues. Furthermore, ILT4 levels were significantly correlated with IL-10 expression and a poor clinical prognosis, because of less infiltration of lymphocytes into the tumour environment and more lymph node metastasis22. Similar results were observed by Sun and colleagues (2008), whom reported ILT4 to be highly expressed in both tumour as well as stromal cells of NSCLC, and ILT4-expressing tumours to have less TIL55. More recently, a study was published demonstrating that ILT4-overexpressing NSCLC cells are responsible for an increase in B7-H3 expression via the PI3K/AKT/mTOR signalling pathway. B7-H3 has been associated with an inhibition in the number of lymphocytes infiltrating the tumour environment, similar to ILT4, and ILT4/B7-H3 co-expression was significantly correlated with a reduction in TIL as well as more lymph node metastasis and advanced tumour stage. Most importantly, co-expression of ILT4 and B7-H3 indicated a poor overall survival54. Another very recent study demonstrates that ILT4 drives tumour growth and metastasis in NSCLC. Manipulating ILT4 expression in vitro showed that ILT4 enhances proliferation, migration and invasion. More importantly, also in vivo assays in mice showed that ILT4 drives NSCLC tumour growth and metastasis. Next to that, ILT4 markedly activated the ERK signalling pathway and upregulated VEGF-C expression, of which the latter enhances proliferation, metastasis and invasion of cancer cells. Zhang et al. (2015) therefore suggested that targeting ILT4 might be an effective approach for NSCLC therapy23. Yet another study demonstrated the correlation between ILT4 and another ligand angiopoietin-like (ANGPTL) protein 2 expression, which may also be important for NSCLC progression57. ANGPTL proteins can exhibit multibiological properties, involving lipid metabolism, inflammation, and the invasive potential of cancer cells56,58. In this study, ILT4-positive tumours had more lymph node metastasis and ILT4 expression was, again, significantly correlated with a shorter overall survival57. This indicates that ILT4, next to its main ligand HLA-G, might be proposed as a prognostic marker as well and that HLA-G independent activation of ILT4 might also result in enhanced suppression of the local tumour environment.
Tolerogenic CD14+ DC expressing ILT4 may contribute to a tolerogenic environment, favourable for organ transplantation, but possibly detrimental for anti-tumour immunity.
DC are potent inducers of either immune tolerance or immune activation. Proper regulation between tolerogenic and immunogenic DC is critical, since aberrant activation is closely related to the development of immune tolerance or autoimmune diseases. Tolerogenic DC can induce tolerance by modulating T cell activation. Moreover, they can contribute to tolerance by the induction of Treg cells59. Tolerogenic DC have mostly been studied by the need to develop strategies to avoid autoimmunity or transplant rejection. Progress in our understanding of immune regulation might be equally important for treatment of immunologic deficiencies and malignancies. Several lines of evidence indicate that ILT4 expression on DC drives a tolerogenic environment. Firstly, Manavalan et al. (2003) reported tolerogenic DC to exhibit a high expression of ILT3 and ILT4 on the cell surface. Allospecific T cells, rendered anergic by the exposure to tolerogenic DC, acquired suppressive activity, mediated by ILT3 and ILT4 and could therefore be favourable in preventing the development of Graft versus Host Disease21. This is consistent with data by Biedroń and colleagues (2015), whom reported high HLA-G serum levels and expression of its receptors on NK cells and monocytes to be correlated with better tolerance to the graft in patients following allotransplantations. The protective role of HLA-G was ascribed to immunomodulatory activity by interacting with inhibitory receptors, i.e. ILT2 and ILT460. As mentioned before, Gregori and colleagues (2010) demonstrated ILT4 to be highly expressed in IL-10 producing DC, termed DC-10, which are potent inducers of regulatory Tr1 cells in vitro. More importantly, these cells also expressed high HLA-G and induction of Tr1 cells by DC-10 was found to be IL-10 dependent and required signalling via the ILT4/HLA-G axis, allowing for the suppression of other immune cells. Blocking either IL-10R, ILT4, or HLA-G avoided the induction of Tr1 cells. Based on these results the authors suggested the following model for tolerance induction: IL-10 production by DC-10 inhibits proliferation and cytokine production of T cells, induces HLA-G expression on T cells, promotes T cell anergy and enhances the expression of the receptors ILT2, ILT3, ILT4 and ILT4’s main ligand HLA-G on DC. Whereas, signalling via ILT4/HLA-G on T cells contributes to the induction of Tr1 cells, this interaction on DC enhances IL-10 production by DC-10, and this ultimately amplifies the “tolerogenic” loop by enhancing de novo ILT4 and HLA-G expression on other DC25. Previously, Liang et al (2008) also characterized tolerogenic DC by the expression of ILT inhibitory receptors52. Being cells that naturally secrete IL-10 and express HLA-G and ILT4, tolerogenic CD14+ DC can contribute to sustain a tolerogenic environment favourable for organ transplantation. However, in tumours the presence or induction of these CD14+ DC can possibly aid tumour development and frustrate anti-tumour immunity.
Proposed immunosuppressive function of ILT4 in cancer
Currently, oncology research has been focussing on expression of ILT4 or HLA-G on tumour cells, while expression on APC and lymphocytes within the tumour environment might be of equal importance. Given the high IL-10 secretion by many tumours, as well as the induction of phosphorylated STAT3, accompanied by IL-6 and IL-10 secretion, one can imagine that within the tumour microenvironment ILT4 might be induced on APC and infiltrating T cells in tumours. When in addition to tumour cells, APC and TIL also express ILT4 as well as its ligand HLA-G and secrete high levels of IL-10, this could lead to an abundance of suppressive pressure on the anti-tumour immune response. In figure 2, we propose a model of how the ILT4/HLA-G axis can continuously promote the development and maintenance of an immunosuppressive tumour microenvironment. Firstly, tumour cells themselves can express ILT4, which upon binding to its ligand HLA-G on APC can induce IL-10 production and enhanced ILT4 and/or HLA-G expression on both tumour cells as well as on APC. Additionally, HLA-G on tumour cells can bind to ILT4 on APC, again inducing IL-10 and up-regulation of ILT4 and/or HLA-G. Secondly, ILT4 expressed on DC can bind to HLA-G on naive CD4+ T cells, inducing either immunosuppressive FoxP3+ Treg cells or IL-10 secreting Tr1 cells. Likewise, upon binding of the TCR/CD8 complex to ILT4 on APC, CD8 suppressor cells can be induced and priming of tumour-specific CD8 cytotoxic T lymphocytes by properly activated DC can be hampered. This indicates how the ILT4/HLA-G axis can upon binding create an immunosuppressive tumour microenvironment, favourable for the tumour to progress and metastasize. Considering ILT4 and HLA-G expression not only on tumour cells, but on APC and infiltrating T cells as well, blocking either one of these would inhibit a tumour from progressing, but also APC from creating an immunosuppressive environment. Therefore, anti-ILT4 therapy, either or not combined with other checkpoint blockade inhibitors such as anti-CTLA-4 and/or anti-PD-1, might be a new promising type of immunomodulatory therapy and deserves further investigation.
Future directions and perspectives
As indicated, cancer immunotherapy has shown promising results in cancer patients. Whereas durable responses have been seen in patients treated with anti-CTLA4 or anti-PD-(L)1, some cancer patients have not shown to respond to currently available therapies. Therefore, it remains important to identify new potential immunotherapeutic targets that may be critical components of frustrating anti-tumour immunity. ILT4 may represent such a promising target. Its expression is not only found to be enhanced on tumour cells, but on tolerogenic CD14+ DC as well. We suggest that expression of ILT4 on tumour cells, as well as on tolerogenic DC and infiltrating lymphocytes, and the secretion of IL-10 might lead to an abundance of suppressive pressure on the anti-tumour immune response. Therefore, we propose a model of how the ILT4/HLA-G axis can continuously promote the development and maintenance of an immunosuppressed tumour microenvironment. While many unknowns and questions remain, studying the effect of inhibition of ILT4 as cancer immunotherapy might be a promising line of investigation.