Name of Journal: World Journal of Gastroenterology
Manuscript Type: SYSTEMATIC REVIEWS
Role of the immune system in pancreatic cancer: focus on cellular response
Chang JH et al. Immune cells in pancreatic cancer
Jae Hyuck Chang, Yongjian Jiang, Venu G. Pillarisetty
Jae Hyuck Chang, Venu G. Pillarisetty, Department of Surgery, University of Washington Medical Center, Seattle, University of Washington, Seattle, Washington, United States of America
Jae Hyuck Chang, Department of Internal Medicine, College of Medicine, The Catholic University of Korea, Seoul, Republic of Korea
Yongjian Jiang, Department of Pancreatic Surgery, Huashan Hospital, Fudan University, Shanghai, China
Author contributions: Chang JH and Jiang Y designed the review and completed the literature search; Chang JH wrote the initial draft of the manuscript; Pillarisetty VG revised the manuscript; all authors reviewed and approved the final manuscript as submitted.
Conflict-of-interest statements: The authors have no conflict of interest of report.
Correspondence to: Venu G Pillarisetty, M.D., Department of Surgery, University of Washington Medical Center, Seattle, University of Washington, 1959 NE Pacific St. Box 356410, Seattle, WA 98195, United States of America. vgp@uw.edu
Telephone: +1-206-616-4924
Fax: +1-206-616-8136
List of abbreviations
Major histocompatibility complex: MHC
Natural killer cells: NK cells
Transforming growth factor-β: TGF-β
Indoleamine 2,3-dioxygenase: IDO
Programmed cell death ligand 1: PD-L1
Tumor-associated macrophages: TAM
Myeloid-derived suppressor cells: MDSC
Regulatory T cells: Treg
Pancreatic intraepithelial lesion: PanIN
Intraductal papillary-mucinous neoplasm: IPMN
Dendritic cell: DC
Crcinoembryonic antigen: CEA
Interferon-α: IFN-α
Granulocyte-macrophage colony-stimulating factor: GM-CSF
Signal transducer and activator of transcription 3: STAT3
Cytotoxic T lymphocyte: CTL
Natural killer group 2 member D: NKG2D
Major histocompatibility complex class I chain-related molecule A: MICA
Cytotoxic T-lymphocyte associated protein-4: CTLA-4
Vascular endothelial growth factor receptor 2: VEGFR2
Decoy receptor 3: DcR3
Genetically engineered mouse model: GEMM
Reactive oxygen species: ROS
Forkhead box P3: FOXP3
Helper T cells: Th cells
Programmed cell death-1: PD-1
Granulocyte-macrophage colony-stimulating factor secreting pancreatic cancer vaccine: GVAX
Chemokine receptor type 5: CCR5
Pancreatic stellate cell: PSC
Abstract
Pancreatic cancer remains difficult to treat, despite advanced in surgery, chemotherapy, radiotherapy, and recently developed immunotherapies. Patients with pancreatic cancer exhibit evidence of systemic immune dysfunction, and pancreatic cancer induce an immunosuppressive microenvironment. The intratumoral activation of immunity in pancreatic cancer is attenuated by inhibitory signals that limit immune effector function. Multiple types of innate and adaptive immune responses can promote an immunosuppressive microenvironment, supporting pancreatic cancer development. Key regulators of the host tumor immune response are dendritic cells, natural killer cells, macrophages, myeloid derived suppressor cells, helper T cells, cytotoxic T cells, and regulatory T cells. The function of these immune cells in pancreatic cancer is heavily influenced by the overall tumor microenvironment; therefore, understanding interactions between the immune cells and their neighboring stromal and epithelial cells forms the fundamental basis for future therapeutic development.
Key words: pancreatic cancer, innate immunity, adaptive immunity, natural killer cell, dendritic cell, macrophage, T cell
Core tip: Attenuation of immunosuppressive signaling has emerged as a key immunotherapeutic target, with remarkable outcomes in multiple cancers, most notably melanoma and non-small cell lung cancer. The promise of immunotherapy is yet to be realized for pancreatic cancer, despite the presence of an assortment of tumor-infiltrating immune cells that establish an immunosuppressive tumor microenvironment. Developing a clear understanding of the interplay of immune cells, carcinoma cells, and stromal cells in the pancreatic cancer tumor microenvironment, therefore, forms the fundamental basis to be considered for future therapeutic development. The present review provides detailed discussion on innate and adaptive immune system in pancreatic cancer, particularly focusing on cellular response and how this is influenced by the tumor microenvironment.
Introduction
Pancreatic ductal adenocarcinoma (PDA), commonly known as pancreatic cancer, is among the deadliest of human malignancies. Despite the recent advances in surgery, chemotherapy, radiotherapy, and recently developed targeted therapies, PDA continues to have less than a 10% 5-year survival rate.[1] Immunotherapy has demonstrated efficacy in the treatment of several types of solid tumors; therefore, there has been great in interest in applying various immunotherapeutic approaches to pancreatic cancer. [2,3] There are, however, several barriers which must be overcome prior to successful implementation of immunotherapy for PDA.
Although PDA is distinguished by prominent desmoplasia (fibrosis), its microenvironment is also replete with immune cells.[4] In spite of the presence of many immune cells in pancreatic cancer, immune dysfunction is observed in patients with pancreatic cancer where the tumor microenvironment is immunosuppressive thus inhibiting the activation or function of immune effectors.[5,6] Furthermore, these immune defects develop in the earliest precancerous lesions.[7]
The present review provides detailed discussion on innate and adaptive immunity in PDA, particularly focusing on the cellular immiune response and its relation to antitumor immunity.
The immunopathogenesis of PDA
Although collaborative effects such as genetic background and environmental factors are needed for development of PDA, chronic inflammation is considered as major risk factor for PDA. The general hypothesis for the pathogenesis of PDA is that when subclinical acute injuries are accumulated and become chronic, genetic alterations in the pancreatic tissue may occur. Cancer immunoediting processes surveils these alteration, and engages three processes: elemination, equilibrium, and escape.[11] Initial elimination process involves immunosurveillance in that the immuno-inflammatory cells attempt to eliminate early-staged genetically unstable or altered pancreatic cells. Host immune system and genetically altered cells which survived the elimination process enter into a dynamic equlibrium. When new variants with mutation are accumulated continuously and over the limit, immunologic elemination response becomes insufficient and the tumor cell variant acquires insensitivity to elemination. Furthermore, the immuno-inflammatory cells exhibit altered functions with subsequent production of immunosuppressive signals such as the inflammatory cytokines that promote tumor growth and invasion.[12,13] Finally, the tumor microenvironment has a highly immunosuppressive composition that contributes further to immune evasion.
The immunosuppression in PDA
PDA has many T cells, thus pancreatic cancer is classified as a T cell rich tumor like microsatellite instability colorectal tumor. Although both innate and adaptive immune responses are active against the tumors, pancreatic cancer by itself induces local and systemic immune dysfunction or immunosuppression to prevent eradication of pancreatic cancer by effector immune cells.[14,15] pancreatic cancer interferes MHC class I peptide presentation to effector T cells by downregulation of the expression of MHC class I molecules or antigen insertion into the MHC class I groove (Figure 1A).[16] Nonfunctional Fas receptors of pancreatic cancer cells render them resistant to Fas-mediated apoptosis, and the expression of functional Fas ligand on them induces apoptosis in cancer-infiltrating effector T cells and natural killer (NK) cells.[17] Pancreatic cancer secretes soluble immunosuppressive factors such as IL-10 and transforming growth factor (TGF)-β that encourage the influx of suppressive immune cells and augment suppressor cell function.[18,19] Indoleamine 2,3-dioxygenase (IDO) in pancreatic cancer, which catalyzes the breakdown tryptophan to kynurenine, suppresses antitumor T cell responses through starving T lymphocytes of tryptophan, thereby IDO induces tolerance to tumor-derived antigens (Figure 1B).[20] In addition, IDO recruits regulatory T cells (Treg).[21] Pancreatic cancer expresses immune system checkpoint ligands such as programmed cell death ligand 1 (PD-L1) to suppress effector cells (Figure 1C).[22] The immunosuppressive cells including tumor-associated macrophages (TAM), myeloid-derived suppressor cells (MDSC), and Tregs appear in the early stage of pancreatic cancer and persist through invasive pancreatic cancer.[23] In contrast, effector CD8+ T cells are dominant in the early premalignant stage, however they decrease during the progression of pancreatic intraepithelial lesions (PanINs) and intraductal papillary-mucinous neoplasm (IPMN).[7] These effector CD8+ T cells are functionally deactivated and become scarce in advanced pancreatic cancer.
Innate and adaptive immune cells in pancreatic cancer
1. Dendritic cells
Dendritic cells (DCs) are professional representative antigen-presenting cells and one of the main regulators of anti-tumor immune response (Table 1).[24] DCs facilitate antigen presentation to CD4+ and CD8+ T cells through capture, internalization and processing of tumor antigens via MHC class I and II molecules.[25] DCs migrate from damaged or invaded tissues to lymph nodes where they present antigen to T cells and hence act as messengers to link innate and adaptive immune system.
Two subtypes of DCs are myeloid DCs and plasmacytoid DCs. While myeloid DCs encounter antigens, plasmacytoid DCs produce interferon (IFN)-α without encountering with antigens.[26]Both myeloid and plasmacytoid DCs are present at immature stage and refractory to maturation signals in the tumor. Pancreatic cancer inhibits the capacity of DC resulting in compromising the recruitment, maturation, and survival of DCs (Figure 1D).[27,28] When DCs and tumor interact, levels of cytokines and chemokines responsible for DCs suppression such as IL-10, TGF-β, and granulocyte-macrophage colony-stimulating factor (GM-CSF) are increased, but those for DCs activation are inhibited mainly through activation of signal transducer and activator of transcription 3 (STAT3).[29] Korylewski et al. Demonstrated that STAT3 ablation can lead to maturation of DCs and restore DC functions in tumor-bearing mice.[30]
Pancreatic cancer tissues show few infiltrations of DCs in the tumor microenvironment, and DCs are located at the invasion edge of the tumor excluded from the tumor mass.[31] Reduction in the level of DCs in circulating blood and impairment in stimulatory function of circulating DCs has been reported in patients with pancreatic cancer.[32,33] However, presence of circulating DCs is related to prolonged survival in both resectable and unresectable pancreatic cancer.[32-34] In addition, higher preoperative circulating DCs count significantly reduced the risk of septic complication after pancreatectomy.[35] It has been demonstrated that radical resection, systemic chemotherapy, chemoradiotherapy or immune-chemotherapy in patients with pancreatic cancer can increase the numbers of circulating DCs or restore their stimulatory function.[24,33,36] These findings suggest that immune therapy targeted at increasing DCs and improving DC function would be beneficial in therapeutic approaches.
DCs-based immunotherapy in pancreatic cancer has been investigated since 1990’s, and considered as promising therapy for patients with advanced pancreatic cancer.[37] It has been reported that administration of DCs pulsed with alpha-galactosylceramide, carcinoembryonic antigen (CEA) mRNA, tumor lysate, or apoptotic tumor cells in pancreatic cancer activates antigen specific cellular components of anti-tumor response leading to expansion of IFN-γ producing natural killer T cells, activation of cytotoxic T lymphocyte (CTL), suppression of tumor growth, and prolonged survival of patients.[38-40]. Also, antigenic peptide-pulsed DCs have been recently reported to be superior to vaccine based on peptide and adjuvant.[41] Several studies showed that DC vaccine and chemotherapeutic agents can exert synergic effect on pancreatic cancer inducing tumor antigen-specific CTL. In cell and animal studies, gemcitabine-treated pancreatic cancer cell medium stimulated maturation of DCs inducing antitumor activity of CTL,[42] and concomitant gemcitabine and DC vaccine therapy increased the survival and facilitated recruitment of CD8+ T cells and CTL-mediated tumor cell lysis in a murine pancreatic carcinoma model.[43,44] In human studies, DCs were recovered and increased after cisplatin and gemcitabine chemotherapy,[45] and DC vaccine combined with chemotherapy was possibly regarded as safe and synergistically effective for inducing tumor antigen specific CTLs.[46,47] Recent multicenter clinical study suggested the possible survival benefits of peptide-pulsed DC vaccines combined with standard chemotherapy in patients with pancreatic cancer.[48]
2. Natural killer cells
NK cells are a subset of cytotoxic lymphocytes and constitute 5-20% of the mononuclear cells in blood and spleen. NK cells can lead to lysis of target cells without any prior sensitization or MHC restriction in the absence of T and B lymphocytes; thereby NK cells are considered as a first line of defense against pathogens and tumor cells.[49] NK cells play a role both in innate and adaptive immunity against tumor by secreting cytokines such as IL-3, TNF-α, GM-CSF, and IFN-γ.[50]
The amounts of circulating NK cells were reduced in patients with advanced pancreatic cancer,[45] and pre-treatment levels of peripheral NK cells positively correlated with survival in patients with pancreatic cancer.[51] These findings indicate that NK cells exert some control over tumor progression. Based on histologic evaluation, Ene-Obong et al. recently demonstrated reduction in levels of NK cells and CD8+ T cells in the juxta-tumoral area as compared with the pan-stromal area of human pancreatic cancer by sequestration with preferential migration to activated pancreatic satellite cells.[52] Moreover, human pancreatic cancer highly expresses Fas ligand that leads to apoptosis of tumor infiltrating lymphocyte including NK cells.[53]
In addition to afore-mentioned numerical suppression, significant impairment in cytotoxicity of NK cells was demonstrated in pancreatic cancer.[45,54] Funa et al. showed that both basal NK activity and in vitro responses of NK cells towards IFN-α were reduced in patients with pancreatic cancer.[54] The activity of NK cells is controlled by signals that are produced from activating receptors or inhibitory receptors. The activating receptors such as natural killer group 2 member D (NKG2D), NKp30 and NKp46 are down-regulated in patients with pancreatic cancer, and the reduced levels of them were related to tumor progression in early stage.[55] Amongst them, NKG2D is a well-studied activating receptor expressed on NK cells, T cell, and natural killer T cells. MHC class I chain-related molecule A (MICA) is the ligand of NKG2D, and is not expressed on normal tissues, but is frequently found on epithelial tumors.[56] The cells expressing MICA are susceptible towards attack by NK cells and antigen-specific T cells. Duan et al. reported that the level of preoperative soluble MICA and NKG2D expression were prognostic factors in patients who underwent resection of pancreatic cancer.[57] NK activity of hepatic non-parenchymal cells has been demonstrated to be depressed in rats with obstructive jaundice, which may enhance growth of liver metastases in pancreatic cancer with obstructive jaundice.[58] Recently, CD56+CD16- NK cells has been reported as unique immune cells developed from a regressing metastatic lesion following the treatment with anti-cytotoxic T-lymphocyte associated protein-4 (CTLA-4) in patients with pancreatic cancer.[59] Although CD56+CD16- NK cells are small subsets of overall NK cells, they could lyse pancreatic cancer cell lines targets and secreted IFN-γ when cultured in presence of high dose of IL-2. This finding supports the possibility that NK cell activation by immune check point inhibition can be a help in treatment of pancreatic cancer.
3. Macrophages
The general function of macrophages is related to infection control and wound healing. Macrophages originate from monocytes and migrate to the inflammatory sites where they are involved in the process of inflammation.[60] The macrophages in or near the tumor are particularly designated as TAMs. TAMs are increased in the tumor, and the distribution of TAMs in the tumor is related to prognosis in human cancers.[61] In human pancreatic cancer, TAMs were prominet compared to normal pancreas,[62] and TAMs infiltrated in the low grade pre-invasive pancreatic tumor lesions and persisted to invasive cancer in a mouse model of pancreatic cancer.[23] Tumor cells influence TAMs reciprocally in the tumor microenvironment. The pancreatic cancer cells induce differentiation and education of macrophages, and consequent tumor-educated macrophages enhance the progression of pancreatic cancer.[63,64] As a result, TAMs function in the tumor by facilitating tumor growth, angiogenesis, stromal remodeling, and metastasis.[63]
There are two types of macrophages in relation to inflammation, M1 and M2. M1 macrophages are classically activated with pro-inflammatory property, and M2 macrophages are alternately activated with anti-inflammatory property. For discriminating markers, CD68 is known as a pan-macrophage marker, HLA-DR and CD11c for M1 macrophage, and CD163 and CD204 for M2-macorphages.[65] In relation with tumor, M1 macrophages present anti-tumor properties, while M2 macrophages are associated with pro-tumor properties.[64] The cytokines such as IL-4, IL-10, and IL-13 from the tumor and T cells induce differentiation of macrophages to M2 phenotype in the tumor, thus M2 macrophages rather than M1 macrophages are predominant in the tumor.[64] It has been demonstrated in vitro that naïve macrophages were changed into M2 phenotype by pancreatic cancer culture supernatants.[66] The pro-tumor properties of M2 polarized macrophage include suppression of adaptive immune response and promotion of matrix remodeling and angiogenesis.[67,68] Furthermore, M2 macrophages induce epithelial–mesenchymal transition that is critically related to metastasis. Liu et al. reported that epithelial–mesenchymal transition was induced by activation of Toll-like receptor 4 on M2 macrophages through stimulating an increase in IL-10 in pancreatic cancer cells.[69] The phenotype of macrophages in the tissues has been reported to be associated with the prognosis of pancreatic cancer. Ino et al. reported M1%/M2 in patients with pancreatic cancer as an independent positive prognostic factor.[4] On the contrary, M2 macrophges are related to poor prognosis and presence of M2 macrophages in the tumor periphery has been related to large tumor size, accelerated lymphatic metastasis, local recurrence, and reduced survival.[70,71] In recent studies, infiltration of M2 macrophages at extrapancreatic nerve plexus was shown as an independent negative prognostic factor,[72] and node-infiltrating M2 macrophages promoted regional lymph node metastasis through production of vascular endothelial growth factor C.[73] Although these M1-M2 dichotomous phenotyping is useful and generally accepted, few macrophages demonstrate overlapped M1 and M2 phenotype. Recently, a new multidimensional model of macrophage activation has been suggested to decode this complexity.[74]
Macrophages are recruited into tumors and interact with the immune system following a number of mechanisms involving various cytokines. Vascular endothelial growth factor receptor 2 (VEGFR2) expressed on macrophages are representative one associated with the recruitment of macrophages. In mice orthotopic pancreatic tumor, selective inhibition of VEGFR2 inhibited infiltration of macrophages into pancreatic cancers.[75] TAMs contribute towards creation of immune suppressive tumor microenvironment through secretion or expression of cytokines and chemokines such as TGF-β, IL-10, CCL17, CCL18, CCL22, and PD-L1 (Figure 1E).[76] Interaction of macrophage inflammatory protein-3α with its chemokine receptor 6 expressed in pancreatic cancer cells had been reported to promote the invasion of pancreatic cancer cell through up-regulation of matrix metalloproteinase-9.[77,78]
Immune therapy involving strategies to modulate or ablate these TAMs may have therapeutic potential. Colony-stimulating factor 1 receptor (CFS1R) is expressed by macrophages and monocytes. Recently, blockade of CFS1R has been reported to enhance antigen presentation and antitumor T-cell immune responses in pancreatic cancer mouse model.[79] Another therapeutic candidate is decoy receptor 3 (DcR3) that is a member of the tumor necrosis factor receptor superfamily and is overexpressed in pancreatic cancer.[80] DcR3 up-regulates genes related to TAMs and down-regulates expression of MHC-II and HLA-DR on macrophages.[81] In addition, the expression level of DcR3 is inversely correlated with survival of patients with pancreatic cancer. Therefore, DcR3 would be a good target for immunotherapy with respect to modulation of TAMs with enhancing anti-tumor immune response against pancreatic cancer.
4. Myeloid derived suppressor cells
Natural suppressive cells that inhibit the antigen-specific cytolysis of allo-reactive immune responses had been introduced in spleen of mice.[82] Since these suppressive cells arise from myeloid origin, immature myeloid cells or myeloid suppressor cells also had been used for defining them. In 2007, the term MDSC was coined to reflect the origin and specify their function.[83] MDSCs are a mixture of immature myeloid cells including immature stages of macrophages, granulocytes, and dendritic cells. As the name suggests, MDSCs do not differentiate into any specific mature types.
MDSCs comprise two types of cells, polymorphonuclear granulocytic MDSCs and mononuclear monocytic MDSCs. The identification of MDSCs remains a difficult task due to their heterogeneity and lack of a singular marker. MDSCs express CD11b/Gr-1 in mice and lack the markers of mature myeloid cells.[84] In humans, IL-4Rα, CD11b, CD33, and low levels of CD 15 are expressed in both MDSCs, whereas CD14 and low level of CD 15 are expressed in monocytic MDSCs.[85,86] In presence of tumor, differentiation of immature myeloid cells is skewed towards the expansion of MDSCs. For the development and recruitment of MDSCs, tumor-derived GM-CSF has been suggested to be an important regulator in genetically engineered mouse model (GEMM) of pancreatic cancer.[87] In the tumor, MDSCs equally suppress CD4+ and CD8+ T cells, and expand immunosuppressive Tregs. Moereover, MDSCs impede innate immunity by conversion of macrophages to M2 phenotype and suppression of NK cells and NK T cell.[88,89]
MDSCs suppress T cells employing multiple mechanisms such as depletion of L-arginine, use of reactive oxygen species (ROS) or free radical peroxynitrate, and down-regulation of L-selectin.[90-92] Since both monocytic and granulocytic MDSCs use arginase that catabolizes L-arginine, MDSCs expressing high levels of arginase deplete L-arginine in tumor environment, thus hindering protein synthesis by T cell through limiting L-arginine availability to T cell.[90] MDSCs can also sequestrate cysteine that is required for T cell activation.[93] Interaction of MDSCs with antigen-specific T cells results in increase in ROS production, and MDSCs suppress CD8+ cell response using ROS in tumor.[94] Moreover, MDSCs are predominant source of free radical peroxynitrate, and the free radical can mediate tumor resistance to CTL by inhibiting binding of processed tumor peptides to MHC molecules.[91] Also, Pilon-Thomas et al. showed down regulation of src homology 2 domain-containing inositol 5’-phosphatase-1 expression in the murine pancreatic cancer cells contributed to the expansion of MDSC inducing suppression of CD8+ T cell responses.[95] In addition, MDSCs inhibit antitumor immunity empolying another mechanism that impairs T cell homing to lymph node via down-regulation of L-selectin in CD4+ and CD8+ T cells.[92]
MDSCs levels are increased in both the circulation and the microenvironment of pancreatic cancer. Clark et al. reported progressive increase of MDSCs from PanIN lesions to pancreatic cancer in GEMM, and also showed correlation between intratumoral MDSCs and the lack of tumor-infiltrating CD8+ cells.[23] These findings indicate that immune suppression by MDSCs is related to the progression from premalignant lesions to pancreatic cancer, and MDSCs follow the histologic progression of pancreatic cancer. Similar to tumor microenvironment, the MDSCs in the circulation are elevated and correlate with the stage in patients with pancreatic cancer. Gabitass et al. reported that the circulating MDSCs were significantly elevated in patients with pancreatic cancer as compared with healthy controls, and were correlated with circulating Treg levels.[13] Moreover, Diaz-Montero et al. demonstrated correlation between circulating MDSCs and clinical stage of various cancers including pancreatic cancer.[96] Recent study showed that patients with stable pancreatic cancer had lower circulating MDSCs before initiation of chemotherapy than those with progressive pancreatic cancer.[97] Therefore, the circulating MDSCs could be a predictive marker for establishing response from chemotherapy.
Inhibition of MDSC in pancreatic cancer is a potential point of cancer therapy. It has been reported that selective depletion of granulocytic MDSCs in autochthonous GEMM of pancreatic cancer enhanced apoptosis of tumor cells with increase in level of CD8+ T cells.[98] Zoledronic acid, a aminobisphosphonate and osteoclast inhibitor for osteoporotic bone disease or bone metastasis, has been studied for suppression of MDSCs. Zoledronic acid inhibited MDSC accumulation and improved host antitumor response in mice.[99] Unfortunately, related human study did not demonstrate differences in overall or progression-free survival during pre-treatment and post-treatment with zoledronic acid.[100] Further challenges employing various methods to inhibit MDSCs in patients with pancreatic cancer are demanded in future.
5. T cells
CD3+ T lymphocytes are main immune infiltrates in pancreatic cancer, and predominantly found in the stroma of both human and mouse PAC.[101] The major components of CD3+ T lymphocytes are CD4+ helper T cell, CD8+ cytotoxic/effector T cell, and CD4+CD25+Forkhead box P3 (FOXP3)+ regulatory T cell. Regulatory T cells exhibit suppressive anti-tumor immunity; thereby many researches regarding Tregs actively have been performed.
CD4+ helper T cells
CD4+ T cells activate the innate immune cells such as macrophages and assist functioning of B cells and CD8+ T cells by secreting cytokines. A decrease in the circulating level of CD4+ T cells in patients with pancreatic cancer as compared to healthy control has been reported,[45,102] and the number of CD8+ T cells in the tissue was lower in pancreatic cancer than cases of chronic pancreatitis.[103] Studies employing immunohistochemistry showed higher level of tumor-infiltrating CD4+ T cells favored better survival in pancreatic cancer.[4,104] Pancreatic cancer alters CD4+ T cell function by inhibition of CD4+ T cells proliferation and migration.[105]
CD4+ T helper (Th) cells differentiate into two subsets of cells, Th1 and Th2. Th1 cells induce cell-mediated immune responses by secreting IL-2 and IFN-γ, while Th2 cells assist humoral immune responses by secreting IL-4, IL-5, IL-9, IL-10, and IL-13.[106] With respect to the tumor, Th1 is involved in tumor killing responses, but Th2 is involved in tumor tolerating response. Th differentiation is skewed in pancreatic cancer, predominately as Th2 rather than Th1 by the influence of immunosuppressive cytokines. Tassi et al. showed that Th cells population skewed towoard Th2 cells based on immunohistochemical analysis of pancreatic cancer.[107] Also, serum cytokine levels of Th cells in patients with pancreatic cancer were shifted towards a Th2 cytokine profile.[13,18] Interestingly, this immune deviation towards Th2 in pancreatic cancer is tumor specific, and antiviral CD4+ T cell immunity in patients with pancreatic cancer showed a Th1 type rather than Th2 type.[107] Th2 skewing in pancreatic cancer is influenced by various factors such as cytokines and stromal cells. IL-10 and TGF-β, which are aberrantly produced by pancreatic cancer, had been reported to contribute towards existence of Th2 phenotype.[18] In addition, the fibroblast in the pancreatic cancer stroma is reported to be associated with Th2 shift. De Monte et al. showed that intra-tumoral Th2 cells infiltration is correlated with thymic stromal lymphopoietin from pancreatic cancer-associated fibroblast.[108] Moreover, they demonstrated tumor-infiltrating Th2/Th1 cells ratio as a positive survival marker in pancreatic cancer.
CD4+ Th cells have functional plasticity in converting Th2 into Th1. Accordingly, reversal of Th differentiation to Th1 is a potential therapeutic point. CEA-specific Th2 cells from pancreatic cancer patients could be reverted to Th1 type by combination of IL-12 and IL-27.[109] In another study, administration of immune-enhancing diets enriched with omega-3 fatty acids, arginine, and RNA before pancreaticoduodenectomy modulated Th differentiation into Th1 rather than Th2, and significantly reduced postoperative infectious complications.[106]
CD8+ cytotoxic/effector T cells
CD8+ T cells are cytotoxic effector cells, and lyse the target cells. CD8+ T cells diminish in circulation and tumor tissues in pancreatic cancer patients. Several studies had revealed decrease in circulating level of CD8+ T cells in patients with pancreatic cancer as compared with the healthy control.[36,45,110] The infiltration of CD8+ T cell is inhibited by pancreatic cancer and tumor progression. When compared with chronic pancreatitis, the number of CD8+ T cells was found to be lower in pancreatic cancer.[103] Hiraoka et al. showed that CD8+ T cells markedly infiltrated in low-grade premalignant pancreatic lesion, but were reduced during the progression of PanINs and IPMNs.[7] CD8+ T cells are reduced near the tumor, and one of suggested reasons for this is that the pancreatic stellate cells affect the migration of CD8+ T cells thus preventing their access to pancreatic cancer cells.[52] Moreover, CD8+ T cells are associated with pancreatic cancer progression and prognosis of patients with pancreatic cancer. The expression of HLA-DR on CD8+ T cells was negatively correlated with tumor size.[102] Tumor infiltrating CD8+ cells together with CD4+ cells served as a favorable prognostic factor,[104] and high densities of CD8+ T cells in the juxta-tumoral area showed better survival in patients with pancreatic cancer.[52]
Pancreatic cancer inhibits CD8+ T cells-mediated tumor cytotoxicity. Pancreatic cancer expresses TGF-β, and inhibits CD8+ T cells from expressing genes encoding cytolytic proteins such as perforin, granzyme, and cytoxin (Figure 1F).[111,112] Furthermore, pancreatic cancer cells frequently loses the expression of MHC class I, which prevent CD8+ T cells from exerting cytotoxic effect on pancreatic cancer.[16] Importantly, pancreatic cancer cells express PD-L1 that binds to programmed cell death-1 (PD-1) expressed on the surface of activated T cells, and the binding impairs T cell infiltration and function leading to T cell anergy or death.[113] Consequently, PD-1 activation blunts host immune response towards pancreatic cancer, and subsequently promotes tumor progression. Similar to PD-1, CTLA-4 is an inhibitory receptor on T cells. Inhibition of CTLA-4 results in analogous outcomes as demonstrated by inhibition of PD-1.[114] PD-1 or CTLA-4 inhibition is a promising immune therapeutic method and has been actively tested in clinical trials with respect to various cancers. PD-1 inhibition leads to increases in effector CD8+ T cells with their production of tumor-specific IFN-γ in pancreatic cancer.[115] Combination PD-1/CTLA-4 and vaccine therapy has also been studied.[116,117] PD-1 blockade as combined with GM-CSF secreting pancreatic cancer vaccine (GVAX) improved the survival in mice compared to PD-1 or GVAX monotherapy.[115] Rosiglitazone, a drug employed for the treatment of type II diabetes, has been introduced as it possesses immune modulating effects. Bunt et al. showed that rosiglitazone combined with gemcitabine increased peripheral CD8+ T cells and intratumoral CD4+ and CD8+ T cells in pancreatic cancer mice model.[118] Since TGF-β imparts potent immunosupressive signal to CD8+ T cells and is related to tumor progression, it is hypothesized that TGF-β gene silencing may restore antitumor immunity. Ellermeier et al. reported that TGF-β gene silencing with activation of retinoic acid-inducible gene I busted tumor-induced CD8+ T cell suppression leading to prolonged survival in pancreatic cancer mouse model.[119]
Regulatory T cells
Immune tolerance and suppression by T cells has been known since 1970.[120] IL-2 receptor α-chain (CD25) was identified for the CD4+CD25+ cells down-regulating immnune response in 1995,[121] and FOXP3 was introduced as a key transcription factor for development and function of regulatory T cell in 2003.[122,123] Accordingly, regulatory T cells are currently considered as CD4+CD25+FOXP3+ cells. Natural Tregs and inducible or adaptive Tregs are two types of Tregs. Natural Tregs are developed from precursor T cells in the thymus without exposure to foreign antigens and they express CD25 and FOXP3 which controls their development and function.[124] Inducible Tregs are derived from naive conventional CD4+ T cells in the periphery in response to antigen exposure.[125]
Tregs contribute towards immune suppressive activity by expression of CTLA-4 and secretion of IL-10 and TGF-β (Figure 1G). In physiological status Tregs prevent autoimmune response; whereas in the tumor, Tregs suppress anti-tumor immune response favoring tumor growth. This involves suppression of tumor-specific CD4+ and CD8+ T cells, macrophages, NK cells, and DCs in tumor microenvironment.[5,12,126,127] Tregs appears in premalignant lesion of pancreatic cancer, and a gradual increase in Tregs through the progression of PanIN and IPMN to invasive ductal carcinoma has been demonstrated.[7,128] In pancreatic cancer tissue, Tregs are increased as compared in the stroma of non-neoplastic inflammatory pancreas.[7] Interestingly, they more infiltrate adjacent to pancreatic cancer tissue.[129] These pathologic findings suggest that Tregs play a role in modulating immune response of pancreatic cancer.
The prognosis of patients with pancreatic cancer is associated with both tumor-infiltrating and circulating Treg. Hiraoka et al. demonstrated that the prevalence of Tregs was a negative prognostic factor and was related to tumor differentiation in pancreatic cancer.[7] With respect to circulating Tregs, they are elevated in patients with pancreatic cancer as compared with healthy control.[13,130] The prevalence of circulating Tregs was reported to be correlated with the tumor stage and survival.[130,131]
Migration of Tregs into pancreatic cancer is controlled by interactions between tumor chemokines and their receptor on Tregs or tumor-induced addressins on the endothelial cells and their ligands on Tregs. Tan et al. showed that pancreas cancer produced increased levels of ligands for CCR5 (chemokine receptor type 5) and Tregs expressed CCR5.[132] They also demonstrated that when ligands for CCR5/CCR5 interaction is diminished or blocked, Tregs migrated to a lower extent to the tumor, and even the tumors became smaller in size. The addressins such as vascular cell adhesion molecule-1, mucosal addressin cell adhesion molecule-1, E-selectin, and activated leukocyte cell adhesion molecule (CD166) are highly expressed on tumor-derived endothelial cells. They interact with Tregs and allow selective Tregs trans-migration from peripheral blood to pancreatic cancer.[14] Recently, it was suggested that myofibroblast in pancreatic cancer plays a role for recruitment of Tregs. Ozdemir reported that myofibroblast-depleted mouse pancreatic tumors showed increased Tregs and correlated with reduced survival rate.[133] In addition, TGF-β from pancreatic cancer is associated with recruitment of inducible Tregs in the tumor. Tregs secrete TGF-β to suppress other immune cells, while pancreatic cancer secretes TGF-β to induce Tregs.[134] In mice model, it has been reported that TGF-β converts CD4+CD25- naïve T cells into FOXP3 Tregs.[19]
Since Tregs are abundant and suppress anti-tumor immune response in pancreatic cancer, Tregs may be effective target for immune therapy. CTLA-4 on Tregs produces inhibitory signal and interacts with its ligands CD80 (B7-1) and CD86 (B7-2) on antigen presenting cells or target tissues. Therefore, anti-CTLA-4 therapy serves to reduce inhibitory signal and/or induce apoptosis of Tregs. Monoclonal antibodies to CTLA-4 have been developed to block this interaction. A previous Phase II trial with ipilimumab (anti-CTLA-4) revealed no responses in patients with advanced PDA.[135] However, a patient experienced significant delayed response with regression of the primary and hepatic metastasis lesions. In addition, CD25 is another target for inhibition of Tregs. Depletion of Tregs with anti-CD25 monoclonal antibodies alone or in combination with a whole tumor cell vaccine promotes smaller pancreatic tumor and longer survival in mouse model.[136] Denileukin diftitox, a fusion protein of IL-2 and active domain of diphtheria toxin, binds with IL-2 receptor and then is internalized into CD25+ cells, with subsequent Tregs death.[137] FDA approved denileukin diftitox for CD25+ cutaneous T‑cell lymphoma and leukemia, and its clinical trials combined with anti-cancer cancer vaccine have been performed in metastatic cancers including pancreatic cancer.[138,139] Recently, anti-glucocorticoid induced TNF receptor monoclonal antibody has been introduced to suppress Tregs. Aida et al. showed this monoclonal antibody induced suppression of Tregs infiltration in pancreatic cancer with down regulation of CCR5 and led to enhancement antitumor immunity of IFN-α gene therapy.[140]
Immune modulation by pancreatic stellate cells/fibroblast
Pancreatic stellate cell (PSC) is myofibroblast-like cell and one of the major components of pancreatic cancer stroma. Pancreatic injuries activate quiescent PSCs, which transform into activated PSCs that secrete extracellular matrix such as type I collagen. Recently, association of PSC and immune cells has been studied. Ene-Obong et al. showed that activated PSCs which secrete chemokine ligand 12 (CXCL12) reduced migration of CD8+ T cells into the juxtatumoral stroma of PDA, and knockdown of CXCL12 by all-tans retinoic acid reversed these effects (Figure 1H).[52] Similarly, Feig et al. demonstrated that fibroblast activation protein positive carcinoma-associated fibroblasts produce CXCL12, which coat the cancer cells thus preventing T cell infiltration.[141] Although anti-PD-L1 did not promote T cell function in the mice, combination with anti-PD-L1 and inhibition of CXCL12 resulted in diminishing pancreatic cancer cells. Galectin-1 secreted by PSCs also involves immune suppressive effect in pancreatic microenvironment. Tang et al. showed that galectin-1 promoted T cell apoptosis and Th2 cytokine secretion.[142] In addition to T cells, differentiation of MDSC may be promoted by PSCs. Recently, it has been reported that PSCs promoted differentiation of peripheral blood mononuclear cells into an MDSC phenotype that suppressed T cell proliferation.[143] Although the reported and ongoing studies regarding PSC and immune cells are still in the early stage, this association would be a good platform for immunotherapy in PDA.
Effect of chemotherapy on immune cells
A number of studies have reported immune-modulatory effects of chemotherapeutic reagents such as gemcitabine, 5-FU, and docetaxel. Gemcitabine has been shown to be associated with Tregs, dendritic cells, and MDSC.[42,144-146] Gemcitabine reduced Tregs accumulation in the orthotopic Panc02 murine model with increase in survival rate.[144] Gemcitabine-added pancreatic cancer cell medium stimulated dendritic cells maturation, and they induced T cell proliferation resulting in cytotoxic T cell antitumor immune response.[42] Gemcitabine directly suppressed MDSCs in mice bearing mammary carcinoma thus leading to increase in T cells and IFN-γ secretion.[145] Gemcitabine-based chemotherapy reduced the proportion and number of circulating Tregs in patients with pancreatic cancer.[146] With respect to docetaxel, it has been reported that the drug enhanced IFN-γ secretion from CD8+ T cells without inhibition of Tregs functions.[147] Immune regulatory effects of 5-FU have also been demonstrated.[148,149] The immuno-chemotherapy employing 5-FU and IFN-α induced infiltration of NK cells in mouse model of pancreatic cancer.[148] These NK cells and IFN-α treatment showed enhanced cytotoxicity on pancreatic cancer through increased expression of NKG2D ligands and MHC class I in Panc02 cells. Vincent et al. reported that 5-FU induced MDSC apoptosis and had a stronger efficacy over gemcitabine.[149]
For evaluation of the effect of chemotherapeutic agents on human pancreatic cancer, the pancreatic tissue from the patient who underwent neoadjuvant chemotherapy is a useful model. Our previous study showed that neoadjuvant treatment resulted in reduced number of MDSCs and Treg in human PDA.[150] Interestingly, the ratios of CD4+ and CD8+ cells to Tregs were higher in patients with neoadjuvant therapy, although CD8+ cells were decreased.
Clinical immune-based therapeutic implementation
Conventional therapy for the pancreatic cancer has marginal gain on the survival rate; therefore a novel therapy is required to overcome the limitation of current therapy. Immune therapy has charming advantages because it demonstrates chances for generating lifelong immune response with acceptable safety profile. Recent excellent outcomes of immune therapy with antibodies modulating immune checkpoint pathway in melanoma, renal and lung cancer have raised expectations for its application as a therapy for pancreatic cancer.[3,116,151] However, immune therapy for pancreatic cancer is still a challenging issue even though pancreatic cancer involves a number of immune cells in the tumor mircoenvironment. To date, immune therapy of pancreatic cancer includes passive immunotherapeutic approach using monoclonal antibodies or effector cells generated in vitro and active immunotherapeutic approach using vaccination to stimulate antitumor response. Monoclonal antibodies employed in passive immunotherapeutic approach block ligand-receptor signaling for growth thus leading to tumor cell death. They target tumor associated antigens, such as mucin 1, Wilms’ tumor gene 1, human telomerase reverse transcriptase, mutated K-RAS, CEA, survivin, p53, HER-2/neu, vascular endothelial growth factor or epidermal growth factor receptor, and α-enloase.[10,117] Vaccination therapy for active immunotherapeutic approach involves administering tumor associated antigens to activate tumor-specific T cells. The available kinds of vaccines are whole cancer cell-based vaccines, antigen/peptide specific vaccines, and dendritic cell-based vaccines.[10] In addition, as previously remarked, immune checkpoint inhibition to activate effector T cells is one of the most actively studied themes. The representative molecules by which checkpoint pathway negatively regulates T cell activation are PD-1 and CTLA-4. Anti-PD-L1 inhibitor (durvalumab, atezolizumab), PD-1 inhibitor (nivolumab, pembrolizumab, pidilizumab), and CTLA-4 inhibitor (ipilimumab, tremelimumab) have been employed in clinical trials.
Although many clinical trials have been performed, phase 3 trials using monoclonal antibody or vaccination combined chemotherapeutic agent have failed to improve overall survival in cases of advanced pancreatic cancer.[152,153] To increase the response, combination therapies comprising immune checkpoint inhibitors and vaccines have been attempted. In phase 1b trial, ipilimumab and GVAX were administered to metastatic patients with pancreatic cancer, and the treatment increased overall survival as compared to patients administered with only ipilimumab.[154] In phase 2 trial, GVAX and cyclophosphamide followed by Listeria monocytogenes-expressing vaccine CRS-207 extended overall survival in pancreatic cancer patients.[155] Recently, new therapeutic attempts such as depletion of fibroblast activation protein-expressing cells in vitro and agonist CD40 antibody in vivo demonstrated promising results.[156,157]
Conclusions and future prospective
Pancreatic cancer has an assortment of immune-inflammatory tumor-infiltrating cells that establish immune suppressive tumor microenvironment. Due to the presence of immune suppressive environment, pancreatic cancer once had been called as immune privilege tumor. However, the suppressive immune cells have emerged as excellent immune therapeutic targets, and immune therapy has shown remarkable outcomes in patients with melanoma and lung cancer. The immune cells in pancreatic cancer interplay in the tumor microenvironment, thereby the immune cells and their relationship with surrounding would be the fundamental basis to be considered for future therapeutic development. Numerous studies have kept elucidating the roles of immune cell subtypes and their capacity to function or dysfunction in the tumor including pancreatic cancer. Optimistically, it will not be long before immune therapy for pancreatic cancer becomes a clinical reality.