Emerging trends in the immunotherapy of pancreatic cancer

Abstract

Pancreatic cancer (PC) is the fourth leading cause of cancer-related deaths in the U.S., claiming approximately 45,000 lives every year. Much like other solid tumors, PC evades the host immune surveillance by manipulating immune cells to establish an immunosuppressive tumor microenvironment (TME). Therefore, targeting and reinstating patient’s immune system could serve as a powerful therapeutic tool. Indeed, immunotherapy has emerged in recent years as a potential adjunct treatment for solid tumors including PC. Immunotherapy modulates the host’s immune response to tumor-associated antigens (TAAs), eradicates cancer cells by reducing host tolerance to TAAs and provides both short- and long-term protection against the disease. Passive immunotherapies like monoclonal antibodies or engineered T-cell based therapies directly target tumor cells by recognizing TAAs. Active immunotherapies, like cancer vaccines, on the other hand elicit a long-lasting immune response via activation of the patient’s immune cells against cancer cells. Several immunotherapy strategies have been tested for anti-tumor responses alone and in combination with standard care in multiple preclinical and clinical studies. In this review, we discuss various immunotherapy strategies used currently and their efficacy in abrogating self-antigen tolerance and immunosuppression, as well as their ability to eradicate PC.

Introduction

It is forecasted that by 2030, pancreatic cancer (PC) will become the second leading cause of cancer-related deaths in the United States [1]. The survival rate of Stage I PC patients is no more than 14% while the overall 5-year survival is approximately 8% [2]. The success rate of various treatment modalities for PC including surgery, chemotherapy, and radiation is limited and reoccurrence is typically inevitable [3]. Further, late diagnosis of the disease further compounds the problem leading to high mortality rate. Recently, immunotherapy has revolutionized cancer treatment especially in melanoma [4, 5]. It is increasingly being felt that immunotherapy in conjunction with the standard of care can improve the outcomes in solid tumors including PC. Studies support immunotherapy as a viable and metamorphic approach, which can boost and restore the ability of immune system to recognize and eradicate cancer cells.

The interaction of the immune system with cancer cells is comprised of the three phases: elimination, equilibrium, and escape [6–8]. During the elimination phase, the immune system can recognize and eliminate transformed cells. Transformed cells that escape the elimination phase enter the equilibrium phase, in which cancer cells undergo genomic editing and establish the tumor microenvironment (TME) that supports the growth of the early lesions. Finally, in the escape phase, cancer cells recruit immunosuppressive cells like myeloid-derived-suppressor cells (MDSCs), regulatory CD4⁺FOXP3⁺ T-cells (T_reg cells) and tumor-associated macrophages (TAMs) [9–11]. In the Kras^G12Dp53^R172H PC mouse model T_reg infiltration increases during the progression from pancreatic intraepithelial neoplasm (PanIN) to the advanced PC stage [12]. Similarly, increased numbers of CD68⁺ TAMs and MDSCs in circulation and the TME are associated with invasiveness in PC patients [9, 13]. Depletion of MDSCs in an autochthonous PC mouse model results in the unmasking of adaptive immune responses against PC, leading to cell death and remodeling of tumor stroma [12, 14]. PC and stromal cells secrete angiogenic factors that promote an immunosuppressive TME while facilitating metastasis [3]. Modulation of stromal cells and their effects are influenced by galectins, which are soluble immune-modulating glycoproteins involved in T-cell homeostasis. Gal-1 overexpressing PSCs induce apoptosis in co-cultured CD4⁺ and CD8⁺ T-cells compared to normal or quiescent PSCs [15]. Further secretion of anti-inflammatory cytokines like IL-10, IL-13 also facilitate tumor cell escape from immune surveillance by precluding immune recognition and developing an immunosuppressive TME [9, 13]. Tumor-specific cytotoxic T-cell infiltration in the tumor is crucial for T-cell mediated killing of cancer cells. However, cancer cells upregulate immune-inhibitory ligands (PD-L1, CTLA-4) leading to the inactivation of cytotoxic T-cells, which further facilitate the escape from cytotoxic cell-mediated death by CD8⁺ T-cells and NK cells [12, 16–21] (Figure 1). Additionally, exhausted CD8⁺ T-cells have chromatin-accessible-regions (ChARs) that serves as an enhancer to maintain high levels of PD-1, which further keeps CD8⁺ T-cells in an immunosuppressive state [22]. Furthermore, PC is an immunologically ‘cold’ tumor due to its low mutational load, dense desmoplasia and rigid extracellular matrix architecture, which restricts the access of effector immune cells to tumor islands, a phenomenon known as excluded infiltrate TME [23–27]. In this review, we describe the current understanding of different immunotherapeutic approaches including anti-cancer monoclonal antibodies (mAbs), T-cell-mediated immunotherapies and cancer vaccines, as powerful strategies for PC treatment. We will also discuss the clinical efficacy of immunotherapeutic strategies, challenges and assess their feasibility as next-generation treatment options, either alone or in combination with chemotherapy for PC treatment.

Figure 1. Pancreatic cancer cells establish an immunosuppressive TME.

Cancer cells secrete various anti-inflammatory cytokines like IL-10, TGF-β, IL-23, along with angiogenic chemokines (e.g., CXCL1-3, CXCL5, CXCL12, CCL2, and VEGF-A), which generate an immunosuppressive TME and facilitate cancer initiation, progression and metastasis. Upregulation of the expression of these cytokines shifts the balance in TME, which facilitates the evasion from immune surveillance during PC progression [6, 8, 19, 24]. The PC immunosuppressive microenvironment also includes crosstalk between cancer cells and various myeloid and lymphoid subsets. Tumor-associated macrophages (TAMs) and myeloid-derived suppressor cells (MDSCs) express immuno-inhibitory ligands and reactive oxygen species that inhibit infiltration and activation of T and NK cells [3, 9, 10, 12]. MDSCs and cancer cells also secrete VEGFs that promote angiogenesis, which aids in the metastasis of the cancer cells [13]. PC tumor cells and pancreatic stellate cells (desmoplasia) secrete inhibitory cytokines and chemokines, and express inhibitory surface ligands such as programmed death ligand-1 (PD-L1) and galectin-1 (Gal-1) that lead to inactivation and apoptosis of cytotoxic (CD8⁺) and helper (CD4⁺) T-cells by programmed death receptor-1(PD-1) or Gal-1 binding receptor respectively [15, 16, 20, 21]. T_reg cells suppress the functions of activated T-cells and NK cells in the TME [12, 95]. In addition, the rigid architecture of pancreatic tumor bed provides physical barrier to T-cells infiltration thereby excluding them to the edge/boundary of the tumor and thus rendering the pancreatic tumor as an immunologically ‘cold’ tumor [24–27]. All these cells are involved in maintenance of the immunosuppressive TME, and cancer progression.

2. Immunotherapy Based Approaches

The goal of immunotherapy is to induce anti-tumor responses by reprogramming and augmenting immune surveillance, and reducing immune suppression. These anti-cancer immunotherapeutic approaches can be classified into ‘passive’ and ‘active’ immunotherapies. Passive immunotherapeutic strategies involve mAbs, adoptive T-cell transfers and genetically engineered T-cells. Whereas the active immunotherapeutic approaches include vaccine-mediated immunity induced by the administration of tumor-associated antigens (TAAs) [28]. TAAs could be delivered in the form of DNA or peptide vaccines, as well as modified tumor cells or antigen-pulsed DCs. Due to genetic alterations or post-translational modifications of proteins (such as glycosylation, phosphorylation, etc.), tumor cells can express proteins that differ from their counterpart in the normal cells or are aberrantly overexpressed in tumor tissues [29]. PC cells express antigens that are either unique to the cancer or are being shared with other cancers with similar epithelial origin. The widely studied TAAs of PC that are currently being evaluated for immunotherapy in various clinical trials are listed in Table I.

Table I.

Tumor Associated Antigens (Cancer Antigens) most frequently targeted in PC immunotherapy.

Tumor Associated Antigen (Cancer Antigen)	Expression		Description
	Normal Pancreas	Pancreatic Cancer
Mucins MUC1 MUC4 MUC5AC	All three are absent	Aberrantly overexpressed and glycosylated in 100% of PC patients	Mucins are glycoproteins that are differentially overexpressed in PC but are negligibly found in normal pancreatic epithelium. These mucins (e.g., MUC1, MUC4, MUC5AC) are involved in PC pathogenesis, provide chemoresistance and enhance proliferation and survival of PC cells. Their overexpression has been correlated with poor prognosis in patients. [Ref. No. 54,64,69-70]
Telomerase	Absent	Expressed in 80–90% of PC patients	Telomerase is a ribonucleoprotein enzyme that catalyzes the synthesis of telomeric DNA. It is involved in the formation and protection of the telomere, which prevents cells from undergoing senescence. Telomerase activity has been detected in pancreatic juice samples of PC patients. hTERT expression and telomerase activity are predictors of poor outcome in pancreatic cancer. [Ref. no. 55, 71–72]
Carcinoembryonic antigen (CEA)	Absent	Expressed in 77% of PC patients and detected in patient serum.	Carcinoembryonic antigen (CEA), a glycosylated protein of MW 180 kDa, is related with tumor burden of PC due to its close association with cancer cell adhesion, metabolism, and proliferation. In clinical practice, CEA is often used to predict the outcomes of patients with resectable PC. [Ref. no. 60,73]
Mutated K-Ras (G12D)	Absent	Expressed in 89.8–94.9% of PC patients	K-ras is mutated in PC cells and inactivation of the oncogenic mutant K-ras enhances MHC I presentation. K-RAS belongs to the superfamily of small G proteins and plays crucial roles in signal transduction in cells. K-RAS mutations in PC transform and alter the biological behavior in PC cells including metabolism reprogramming, thus playing a crucial role in PC pathogenesis. [Ref. no. 77–80]
Vascular endothelial growth factor (VEGF)	Absent	Expressed in 77–93% of PC patients	VEGF, primarily VEGF-A and its receptors (VEGFR1 & VEGFR2), are primarily involved in the angiogenesis process in PC cancer. Increased vascularization of pancreatic tumors promotes their growth and metastasis by providing nutritional flow. Neovascularization also facilitates infiltration of pro-tumor immune cells (e.g., MDSCs). [Ref. no. 82–83]
Mesothelin	Absent	Expressed in ~86% of PC patients	Mesothelin (MSLN) is a glycoprotein overexpressed in various epithelial cancers like mesothelioma and pancreatic, ovarian, and lung cancers. MSLN is synthesized as a 71 kDa precursor protein, which is processed to a 30 kDa megakaryocyte-potentiating factor and a 40 kDa MSLN protein. It is attached to the plasma membrane by a glycosylphosphatidylinositol anchor and is involved in cell adhesion. MSLN serves as a marker of neoplastic transformation of pancreatic epithelial cells. [Ref. no. 86–87]

2.1. Passive Immunotherapeutic Strategies

Passive immunotherapy attacks cancer by directly targeting TAAs by the administration of diverse immune components that are engineered ex vivo. Recently, mAbs have also been used to abrogate immunosuppression by targeting signaling pathways involved in immune checkpoint. Following are the major passive immunotherapeutic strategies ongoing in preclinical studies or in clinical trials and have been summarized in Table II.

Table II.

Clinical trials testing antibody-based immunotherapies for pancreatic cancer.

Antigen & Drug	ClinicalTrials.gov Identifier	Phase	No. of Patients	Status	Median overall survival (months)	Median progression-free survival (months)	Objective response rate (%)
CD40 (CP-870,893) 0.2 mg/kg	NCT00711191 [Ref. No. 35]	I	22	Completed	7.4 (5.5 to 12.8)	5.6 (4.0 to 7.4)	7.7 (0.2 to 36.0)
CD40 (CP-870,893) + Gemcitabine	NCT01456585*	I	10	Completed	NR	NR	NR
PD-1 (CT-011) alone or in combination with Gemcitabine	NCT01313416*	II	29	Suspended	NR	NR	NR
PD-L1 (pembrolizumab)	NCT02362048*	II	73	Active	NR	NR	NR
PD-L1 (pembrolizumab)	NCT02009449*	I	350	Active	NR	NR	NR
CTLA-4 (ipilimumab)	NCT00112580	II	27	Completed	NR	NR	1 patient had Partial Response
CTLA-4 (ipilimumab) + Pancreatic Cancer Vaccine	NCT00836407*	I	30	Completed	5.7 (4.3 to 14.7)	NR	NR
HER1 (Cetuximab) + Irinotecan + Docetaxel	NCT00042939*	II	87	Completed	5.3 (4.5 to 9.4)	4.5 (2.7 to 5.6)	0.07 (0.024 to 0.198)
HER1 (Cetuximab) + Gemcitabine + Radiotherapy	NCT00225784*8	II	37	Completed	17.3 (2 to N/A)	9.1 (2 to N/A)	10 out of 37 had Partial Response
HER1 (Cetuximab) + Ixabepilone	NCT0038314*	II	54	Completed	7.6 (5.5 to 12.2)	3.9 (2.6 to 4.4)	4 patients had Partial Response
HER1 (Cetuximab) + Irinotecan + Oxaliplatin	NCT00871169*	II	58	Completed	NR	NR	6.9 (1.91 to 16.7)
HER1 (Cetuximab) + Gemcitabine + VEGF (Bevacizumab)	NCT00326911*	II	30	Terminated	5.41 (3.84 to 6.74)	3.55 (2.00 to 5.59)	4 patients had either Partial or Complete Response
HER1 (Cetuximab) + Gemcitabine + VEGF (Bevacizumab)	NCT00091026*	II	71	Completed	7.9 (5.5 to 9.5)	5.0 (3.7 to 5.5)	21 (12 to 32)
HER1 (Cetuximab) + Gemcitabine + Oxaliplatin	NCT00338039*	II	69	Completed	19.2 (14.2 to 24.2)	NR	NR
HER1 (Cetuximab) + Gemcitabine + Capecitabine + Radiation	NCT00305877*	II	65	Completed	0.38 (0.26 to 0.50)	0.17 (0.08 to 0.26)	0.30 (0.19 to 0.42)
HER2 (Trastuzumab) + Interleukin 12	NCT00004074*	I	15	Completed	NR	NR	NR
HER2 (Trastuzumab) + HER1 (Cetuximab)	NCT00923299 [Ref. No. 51]	I & II	44	Completed	4.6 (2.7–6.6)	1.8 (1.7–2.0)	NR
Mesothelin (SS1(dsFv)-PE38 immunotoxin)	NCT00006981*	I	NR	Completed	NR	NR	NR
VEGF (Bevacizumab) + Gemcitabine + accelerated Radiation Therapy	NCT00557492*	II	43	Ongoing	19.7 (16.5 to 28.2)	12.9 (7.0 to 18.7)	2.3 (0.1 to 12)
VEGF (Bevacizumab) + Octreotide Acetate + Everolimus	NCT01229943*	II	75	Ongoing	36.7 (31.8 to N/A)	16.7 (12.6 to 19.7)	31

^*Data obtained from https://clinicaltrials.gov/

2.1.1. Antibody-Mediated Passive Immunotherapy

Antibody-mediated immunotherapy involves targeting tumors using monoclonal antibodies (mAbs), antibody fragments, antibody-drug conjugates, or radio-immunotherapy conjugates to inhibit oncogenic signaling, immune suppression, or immune checkpoint blockade.

CD40 is a member of the TNF receptor superfamily and is expressed primarily on APCs such as DCs, macrophages, monocytes, B-cells and some non-immune cells like cancer cells [30]. Anti-CD40 antibodies mimic the co-stimulatory signal of the CD40 ligand (CD40L). Tumor-bearing KPC (K-ras^G12D, TP53^R172H) mice with constitutive K-ras activation and gain-of-function p53 mutation when treated with anti-CD40 antibody (clone FGK45, endotoxin-free), either alone or in combination with gemcitabine, showed a detectable tumor regression. Treatment with anti-CD40 mAb bypassed the requirement for Toll-Like Receptors (TLRs), inflammasome, Type I IFNs, and stimulator of interferon genes (STING) to effectively prime adaptive T-cell responses against PC in these animals [31]. The mechanistic role of agonistic anti-CD40 mAb is to activate host APCs (especially DCs) to induce clinically relevant antitumor T-cell responses, reverse tumor-induced immune suppression and induce T-cell-independent but macrophage-dependent tumor regression in PC patients [32]. In a clinical trial, 22 naïve patients with advanced PC were administered weekly doses of anti-CD40 mAb in combination with gemcitabine, which led to increased B-cell surface expression of co-stimulatory molecules CD86, HLA-DR, and CD54 at 24–48h post-treatment [32–35]. In a cohort of 21 chemotherapy-naïve and surgically incurable PC patients, treatment with gemcitabine and a human agonist anti-CD40 mAb (CP-870,893) for 3 weekly cycles showed enhanced overall survival of 7.4 months compared to those who received gemcitabine alone with median overall survival of 5.7 months. Upon biopsy, the tumors of anti-CD40 mAb-treated patients showed higher infiltration of macrophages, however, with relative absence of lymphocytes [35, 36].

PD-1 (CD279) is a T-cell co-inhibitory receptor expressed on the surface of activated T-cells, T_regs and monocytes that has been extensively exploited for immunotherapy. PD-1 on T-cells interacts with two B7 family ligands, PD-L1 (CD274) and PD-L2 (CD273) expressed on tumor cells that leads to T-cell anergy or death and thus leading to tumor survival [37]. Presence of infiltrating PD-1⁺ T-cells in densely or loosely desmoplastic pancreatic tumors suggests tumor antigen specific T-cell activation that correlates with increased overall survival, progression free survival and distant-metastasis free survival of PC patients [38]. Activated T lymphocytes infiltrating the TME express inflammatory cytokines like IFNγ that further stimulates PD-L1/L2 expression on the tumor cells [39, 40]. Blockade of PD-1 by a mAb abrogates the PD-1/PD-L1 axis restoring T-cell cytotoxic function [41]. In a preclinical study, combined treatment of Pan 02 tumor-bearing mice with anti-PD-L1 mAb and gemcitabine significantly reduced average tumor volume compared to gemcitabine and anti-PD-L1 mAb alone [42]. Due to the relative success of anti-PD-1 antibodies in both preclinical and clinical studies on selective solid tumors, the FDA recently approved two anti-PD-1 antibodies, pembrolizumab and nivolumab, for head and neck cancer, renal, melanoma, and non-small cell lung cancer treatment [43, 44].

CTLA-4 is another co-inhibitory molecule expressed on the surface of activated T-cells and T_reg cells. CTLA-4 present on T-cells interacts with B7-1/B7-2 ligands on APCs, resulting in depletion and suppression of CD28-mediated T-cell-activation [45]. Ipilimumab, an antagonist mAb against CTLA-4, inhibits immunosuppression by T_reg cells and enhances antitumor activity of effector T lymphocytes and innate immune cells. In a preclinical study, in-vitro treatment with ipilimumab significantly enhanced T-cell proliferation (preferentially promoting CD8⁺ T-cell expansion), Th1 cytokines release (IFN-γ, IL-2, and IL-12), and increased cytotoxicity of CD8⁺ T-cells against Colo356/FG PC cells [46]. In a Phase Ib clinical trial, patients with previously treated or histologically proven PC were given ipilimumab alone or in combination with GVAX. Post-treatment, both the single and combination treatments enhanced mesothelin (MSLN) specific CD8⁺ T-cell populations that correlated with increased survival of >4.3 months, as well as a decline in CA-19.9 levels in 7 out of 15 patients compared to patients treated with ipilimumab alone (0 out of 15 patients) [47]. Combination therapy of anti-CD40, anti-CTLA-4 and anti-PD-1 antibodies with chemotherapy/nab-paclitaxel in KPC mice resulted in tumor regression in 39% of the animals (17 out of 44 mice), along with increased CD8⁺ T-cell infiltration and reduction in T_reg cells (7-fold CD8: T_reg ratio) in the PC TME. Furthermore, PC cells implanted on the opposite flank were rejected with no additional treatment in 67-86% of mice, suggesting the development of immunological memory [48].

Several unarmed monoclonal antibodies against receptor tyrosine kinases like members of EGFR family (cetuximab, pertuzumab, and trastuzumab) [49, 50] or their ligands like VEGF (bevacizumab) [51], that are involved in tumor cell proliferation or angiogenesis respectively, have been evaluated in combination with chermotherapeutic agents with varying degree of success (summarized in Table II) [52].

MDSCs are known to secrete tumor-promoting factors, such as prokineticin 2 (PK2/Bv8). Anti-Bv8 antibody targeting the extracellular domain of Bv8 given in combination with gemcitabine reduced growth of orthotopically implanted metastatic PC cells, significantly reduced MDSCs infiltration, hypoxia and angiogenesis compared to mice treated with gemcitabine alone, indicating the significant potential of anti-Bv8 antibody as a combinatorial or post-chemotherapy treatment in PC patients [53].

2.1.2. Passive T-cell-Mediated Immunotherapy

Although monoclonal antibody based therapies can elicit direct killing of tumor cells or transiently abrogate immunosuppression, but they do not provide long-term relief to PC patients. Multiple studies are evaluating the strategies to develop passive T-cell-mediated immunotherapies including increasing the number of antigen-specific CD8⁺ T-cells, the responsiveness of the antigen-specific T-cells, or the affinity of antigen-specific T-cell receptors (TCRs). Additional summary regarding the current clinical trials utilizing these strategies is provided in Table III.

Table III.

Clinical trials testing T cell-mediated immunotherapies for pancreatic cancer

Antigen & Drug	ClinicalTrials.gov Identifier	Phase	No. of Patients	Status
Anti-CEA CAR-T cells	NCT02416466*	I	8	Ongoing
Autologous T cells transfected with chimeric anti-mesothelin immunoreceptor SS1	NCT01897415*	I	16	Ongoing
CART-meso-19 T cells + Cyclophosphamide	NCT02465983*	I	12	Ongoing
GI-4000 Vaccine + Activated T Cells	NCT00837135*	I	NR	Withdrawn
MFE23 scFv-expressing autologous anti-CEA MFEz T lymphocytes	NCT01212887*	I	14	Terminated due to safety concerns and lack of efficacy
Autologous Natural Killer/Natural Killer T Cell Immunotherapy	NCT00909558*	I	24	Suspended
Prostate Stem Cell Antigen (PSCA) Specific CAR T Cells (BPX-601) + Rimiducid	NCT02744287*	I	30	Recruiting

^*Data obtained from https://clinicaltrials.gov/

Adoptive T-cell transfer (ACT)

The primary objective of ACT therapy is to isolate and expand T-cells ex vivo and transfer these autologous lymphocytes with antitumor activity into cancer patients. This method leads to the expansion of antitumor T-cell populations in the patient resulting in increased cytokine release and tumor cell targeting.

Kawaoka et al., developed cytotoxic T-lymphocytes (CTLs) by isolating T-cells from the blood of healthy volunteers expressing human leukocyte antigen HLA-A phenotype 24/26 and stimulating them with the MUC1-expressing human PC cell line YPK-1 (HLA-A phenotype 24/02) in combination with IL-2. MUC1-specific CTLs killed five MUC1-expressing PC cell lines, irrespective of their HLA phenotype. 20 patients with resectable and 8 patients with unresectable PC were treated with MUC1-specific CTLs. Patients with non-resectable tumor did not show any improvement with median survival time (MST) of 5 months, however, 18 out of 20 patients who received MUC1-specific CTLs as an adjuvant therapy with curative surgery had MST of 17.8 months and suppressed post-surgery hepatic recurrence [54].

Murine PC cell lines have significant overexpression of telomerase activity. C57BL/6 mice were immunized with H2^b-restricted telomerase peptide emulsified with incomplete Freund’s adjuvant, in complex with macrophage-activating lipopeptide-2 (MALP-2, a Toll-like receptor 2/6 agonist) to drive the generation of telomerase-specific CTLs. Orthotopically implanted syngeneic tumor-bearing mice were treated with IL-2-expanded anti-telomerase CTLs, which significantly reduced tumor volume compared to untreated mice. In addition, anti-telomerase CTL-treated mice developed higher numbers of both CD8⁺ central memory and effector antigen-specific T-cells [55]. Furthermore, in a clinical study, 46 PC patients with non-resectable and recurrent tumors received anti-CD3-stimulated lymphokine-activated killer (CD3-LAK) therapy (25 patients) or RetroNectinVR (CH296)-induced T-cell (RIT) therapy (21 patients) at 2-week intervals. The ACT treated patients showed an increased circulating levels of IFN-γ, IL-12, and IL-2, suggesting that the combined circulatory levels of these cytokines may serve as a predictive marker of the clinical response to ACT in patients [56].

Chimeric antigen receptors (CAR) T-cells

Highly antigen-specific autologous T-cells that are genetically engineered to express tumor antigen-specific TCRs or immunoglobulin-based fusion proteins are known as chimeric antigen receptors (CAR) T-cells. These engineered CAR T-cells are then cultured and expanded ex vivo for therapeutic purposes (Figure 2).

Figure 2. CAR T-cells are genetically engineered T-cells expressing tumor antigen-specific chimeric TCR [119, 120].

The modified receptor is a chimera of a signaling domain of the TCR complex and an antigen-recognizing domain, such as a single chain fragment (scFv) of an antibody [121, 122]. CAR T-cells are not dependent on antigen presentation by MHC molecules expressed on APCs for antigen specific activation. Adoptive cell transfer of CAR T-cells involves the isolation, stimulation, expansion, transduction, and ultimately re-infusion of human T lymphocytes [123, 124]. First-generation TCRs included only the intracellular domain of the CD3ζ chain but did not show any significant in vivo efficacy in transgenic mouse model studies [125]. Second-generation CARs introduced additional co-stimulatory domains such as CD28, which significantly augmented CAR signaling, and improved cytokine production and T-cell proliferation, as well as differentiation, and survival [126, 127]. Third-generation CARs contain multiple co-stimulatory domains such as 4-1BB (CD137), and whether they have clinical benefit over second-generation CAR T-cells is still under investigation [122, 128, 129].

The differential glycosylation pattern of mucins provides a unique repertoire of antigenic epitopes that can be exploited for developing tumor specific CAR T-cells. Posey et al., designed a scFv of a high-affinity antibody (5E5) to detect truncated O-glycopeptide MUC1 epitopes that are not expressed in normal tissues. MUC1 CAR T-cells (composed of 5E5 mAb scFV on the backbone of CD8α and transmembrane domain and costimulatory domains of 4-1BB and CD3ζ) were generated that target the Tn/STn glycopeptide epitope on MUC1. Upon recognition of MUC1-9Tn, secreted high quantities of IL-2 and IFN-γ, but not in response to the non-glycosylated MUC1-60-mer. Hs766T pancreatic tumor bearing mice when treated with 5E5 CAR T-cells showed potent responses and improved survival to 113 days with 100% animals surviving compared to 40% and 33% of mice treated with non-transduced and CD19 CAR T-cells, respectively. In addition, many 5E5 CAR T-cells specifically accumulated in Hs766T tumors, in contrast to a small percentage of CD19 CAR T-cells [57, 58].

Carcinoembryonic antigen (CEA) is highly overexpressed on the surface of PC cells. Murine CEA binding domain (SCA431scFv)-containing CAR T-cells with intracellular CD28-CD3 signaling domain were adoptively transferred into Panc02 CEA⁺ tumor-bearing CEA transgenic mice. Anti-CEA CAR T-cells significantly reduced the size of pancreatic tumors and produced long-term tumor elimination in 67% of the mice without inducing autoimmune reaction. Upon re-challenge with CEA⁺ C15A3 cells, the animals rejected the cells and demonstrated increased serum levels of IL-1β and IL-5 [59]. A similar study reported eradication of CEA⁺ tumors in CEA-transgenic mice as a primary response by anti-CEA CAR T-cells with CD3ζ endo-domain and rejection of CEA⁺ PC cells upon re-challenge. Based on the CAR T-cell model, there is evidence that antigen-specific CD8⁺ T-cells can be induced to overcome self-tolerance and eliminate cancer cells while sparing normal cells [60].

Prostate stem cell antigen (PSCA) is another highly expressed TAAs in PC patients as well as in tumor-derived cell lines. In a recent study, PSCA-specific CAR T-cells showed specific targeting and lysing of PSCA-expressing PC cells (ASPC1, Capan-1) while PSCA-negative 293T-cells showed no cytotoxicity [61]. In another study, anti-PSCA CAR T-cells were engineered using antigen-recognition domains derived from mouse or human antibodies with either one (CD28) or two (CD28 + 4-1BB) T-cell co-stimulatory molecules linked to the CD3ζ endo-domain. These anti-PSCA CAR T-cells elicited anti-tumor responses in mice with established human PC-derived xenograft tumors and 2 out of 5 mice showed complete tumor eradication [62].

MSLN is highly overexpressed on PC cells compared to its negligible expression in normal pancreas. Hingorani et al., developed MSLN peptide-specific high affinity TCR₁₀₄₅ expressing CD8⁺ CAR T-cells that lysed KPC tumor cells in vitro and secreted IFN-γ upon antigen recognition. A study in KPC mice showed that TCR₁₀₄₅ CAR T-cells infiltrated the pancreatic tumors 4 days post-injection and induced apoptosis of cancer cells after 8 days of infusion. Upon second infusion, TCR₁₀₄₅ CAR T-cells showed 10-fold increased retention in pancreatic tumor compared to non-specific (TCR_gag) T-cells. However, mice in both of the treatment groups developed progressive disease. TCR1045 cell recipient mice showed less metastasis (46%) and overall survival of 96 days compared to 64% metastatic lesions and survival of 54 days in TCR_gag treated mice. Overall, these data suggest that tumor antigen-specific engineered T-cell therapies are viable options for the treatment of invasive PC [63].

2.2. Active Immunotherapeutic Strategies: Cancer Vaccines

Active immunotherapy relies on stimulation of the immune system through immunological recognition of TAAs by T and B lymphocytes. TAAs have been widely explored as cancer vaccines for treatment of PC in both in vivo mouse models and clinical trials. Cancer vaccines can be whole cancer cell-based vaccines, antigenic-peptide pulsed vaccines or DC-based vaccines. These vaccines are developed to exploit and activate both innate and active immune arms to eradicate tumor cells and evade future recurrence of the disease. Cancer vaccines currently being investigated in clinical trials in PC are summarized in Table IV.

Table IV.

Clinical trials testing cancer vaccines for pancreatic cancer.

Antigen & Drug	ClinicalTrials.gov Identifier	Phase	No. of Patients	Status
MUC1 Vaccine (Cvac vaccine)	NCT02310971*	II	0	Withdrawn
Falimarev (MUC1 PANVAC-F vaccine) + Inalimarev (MUC1 PANVAC-V vaccine) + Sargramostim (GM-CSF vaccine)	NCT00669734*	I	18	Ongoing
Telomerase vaccine (GV1001) + gemcitabine + Sargramostim + tadalafil + Radiation Therapy	NCT01342224*	I	11	Active
Telomerase vaccine (GV1001) + Sargramostim + capecitabine + gemcitabine	NCT00425360*	III	1110	Completed
CEA vaccine (ALVAC + vaccinia) + aldesleukin (IL-2) + Sargramostim	NCT00003125*	II	24	Completed
CEA vaccine (AVX701)	NCT00529984*	I & II	28	Completed
Recombinant fowlpox-CEA(6D)/TRICOM vaccine + GM-CSF vaccine + Sargramostim	NCT00028496*	I	48	Completed
Recombinant fowlpox-CEA(6D)/TRICOM vaccine + denileukin diftitox + therapeutic autologous dendritic cells	NCT00128622*	I	24	Completed
CEA vaccine (TRICOM-CEA(6D))	NCT00027534*	I	14	Completed
CEA RNA-pulsed DC cancer vaccine	NCT00004604*	I	24	Completed
CEA vaccine (carcinoembryonic antigen peptide 1-6D) + incomplete Freund’s adjuvant + sargramostim	NCT00012246*	II	7	Terminated
K-ras vaccine (TG01) + GM-CSF + Gemcitabine	NCT02261714*	I & II	32	Active
Aldesleukin + ras peptide cancer vaccine + sargramostim + DetoxPC	NCT00019331*	II	11	Completed
HLA-A*02:01 -restricted VEGFR1-derived peptide vaccination	NCT00683085*	I & II	2	Terminated
VEGFR-2 DNA vaccine VXM01	NCT01486329*	I	72	Completed
Mesothelin vaccine (CRS-207) + GVAX vaccine + gemcitabine + capecitabine + 5-FU + irinotecan or erlotinib + cyclophosphamide	NCT02004262 [Ref no. 87, 88]	II	303	Completed
GVAX Pancreas + Mesothelin vaccine (CRS-207) + Cyclophosphamide	NCT01417000 [Ref. no. 88]	II	93	Ongoing
Mesothelin vaccine (CRS-207)	NCT00585845*	I	17	Terminated
Cancer stem cell vaccine	NCT02074046*	I & II	40	Completed

^*Data obtained from https://clinicaltrials.gov/

Several mucins, which are high molecular weight glycoproteins, are differentially overexpressed in pancreatic tumors. Some of these mucins (e.g., MUC1 and MUC4) have also been demonstrated to contribute to chemoresistance, enhance proliferation and survival of PC cells [64]. Therefore, mucins are being studied as potential candidates for vaccine development for PC. Studies conducted in human MUC1-transgenic (MUC1.Tg) mice treated with MUC1 cancer vaccines failed to show any detectable responses against MUC1⁺ tumor cells despite of, MUC1-specific T-cells generating IFN-γ, IL-4 and IL-10 cytokines. The immune responses in these mice were not skewed to either type 1 or type 2 immune response thus rendering the vaccine ineffective against the B16.MUC1 tumor [65]. The CD8⁺ T-cell killing of MUC1-expressing tumor cells was found to be mediated by perforin and FasL cytolytic pathways. Also, lymphotoxin-α, but not TNF receptor-1 (TNFR-1), played a critical and non-redundant role in the cell-mediated rejection of MUC1 expressing tumor cells [66]. In a Phase I clinical trial, MUC1-peptide (PDTRPAPGSTAPPAHGVTSA)-pulsed DC vaccines were administered to 7 patients with advanced PC and 2 out of 7 patients showed significantly increased mature DCs and PBMC-mediated immune responses that were characterized by high IL-12p40 and IFN-γ secretion, respectively. However, there was no tumor rejection in these patients [67]. A similar Phase I clinical trial in 20 advanced PC patients was performed with MUC1 peptide-pulsed DCs in combination with MUC1-specific CTLs. One patient with lung metastases showed complete remission, while 5 other patients demonstrated a stable disease for at least 6 months post-therapy [68]. A study was conducted with 6 metastatic PC patient-derived DCs that were co-transfected with MUC4 and survivin mRNAs. These mRNA-loaded DCs activated CTLs against MUC4 protein. Anti-MUC4 CTLs effectively targeted a human PC cell line (Capan-2) via MHC I-restricted recognition and released IFN-γ. MUC4-mRNA-pulsed DCs stimulated more CTLs than survivin-mRNA-pulsed DCs, but comparatively elicited fewer CTLs activated by MUC4-survivin-mRNA-loaded DCs [69]. In another study, mature DCs were pulsed with a MUC4 epitope peptide (LLGVGTFVV) and co-cultured with CD8⁺ T-cells to generate MUC4-specific CTLs that could effectively kill HCT-116 colorectal cancer cells (MUC4⁺, HLA-A2+). However, intensity of MUC4 surface expression on PC cell-line HPAC proportionally increased the apoptosis of MUC4-specific T-cells suggesting a MUC4-mediated contact-dependent killing of CTLs thus rendering the therapy ineffective [70]. Based on these studies, other mucins like MUC5AC, and MUC16 that are significantly overexpressed in PC may serve as potential vaccine candidates to develop novel immunotherapies.

Due to overexpression of telomerase in PC patients, cancer vaccine containing telomerase-derived peptide (GV1001) vaccine is under clinical studies. However so far, PC patients treated with a combination of GV1001 vaccine, GM-CSF, and gemcitabine showed transitory and weak Th1-type immune response, reduced infiltration of T_reg cells, and no significant increase in median overall survival. In a related clinical trial, the GV1001 vaccine failed to enhance the effects of chemotherapy (gemcitabine and capecitabine) [71, 72].

CEA is overexpressed in >90% of PC patients making it a potential immunotherapeutic target. A Phase I clinical trial with CAP1-6D, an altered CEA peptide ligand/Montanide/GM-CSF vaccine, administered to 66 PC patients elicited robust CD8⁺ T-cell-mediated responses. In addition, 8 of the patients were found to have high IFN-γ production, of which 4 patients showed progressive disease, 3 patients had stable disease, and 1 patient showed a complete response [73].

Another cancer vaccine candidate is KIF20A, a member of the kinesin super family, which is significantly upregulated in PC. KIF20A-66, a HLA-A24-restricted epitope peptide derived from KIF20A peptide vaccine was injected into 29 PC patients in a Phase II clinical trial. The tumor size (as confirmed by CT scan) was reduced in 8 patients and 16 patients showed increased CTL responses, which correlated with the increased overall survival of vaccinated patients [74]. In a similar Phase I trial, 9 advanced PC patients who previously received chemo/radiotherapy were treated with KIF20A-10-66 (KVYLRVRPLL) peptide vaccine along with gemcitabine. These patients showed increased activation of KIF20A-specific IFN-γ-producing T-cells, had stable diseases and longer overall survival, indicating that KIF20A-based vaccines are possible immunotherapy candidates for PC [75].

K-ras is mutated in the majority of PC patients and is currently under investigation as an immunotherapy target. Inactivation of the oncogenic mutant K-ras enhances MHC I presentation [76]. In an in vivo study, mice were injected with CTLs induced by DCs pulsed with human PANC-1 cell lysate expressing mutant K-ras (12 Val) or with K-ras mutant peptide (K-ras+peptide) alone, or with cationic nanoparticles (CNP) encapsulating K-ras mutant peptide (K-ras+-peptide-CNP). The K-ras+-peptide-CNP activated CTLs upon adoptive transfer specifically killed mutated K-ras positive PANC-1 tumors and improved the survival of PANC-1 tumor-bearing nude mice, however, with no effect on SW1990 tumor bearing mice, suggesting a prospect of K-ras mutation specific PC therapy [77]. In addition, a clinical trial testing a mutant Ras peptide vaccine (administered in combination with GM-CSF) in 11 patients with advanced PC showed higher antitumor responses (92% exhibiting an immune response) with 2 patients having a complete response and an overall survival of 20.8 months [78].

The WT1 protein is another suitable vaccine target for PC due to its differential overexpression in tumor cells but not in normal pancreas [79]. In a recent study, 32 HLA-A*24:02⁺ advanced PC patients were treated with an HLA-A*24:02-restricted, modified 9-mer WT1 peptide (CYTWNQMNL) emulsified with Montanide ISA51 adjuvant (WT1 vaccine). The MST of patients who responded to the WT1 vaccine was 10.9 months. Further, these patients developed strong effector T-cell responses along with generation of WT1-specific CD8⁺ memory T-cells, whereas unresponsive patients showed MST of only 3.9 months [80]. In a separate clinical trial, 9 patients with advanced PC were vaccinated with WT1 peptide vaccine and 8 out of 9 patients demonstrated stable disease. However, no WT1-specific T-cells were observed in the circulation of these patients [81].

VEGF is another protein that is overexpressed in PC. In a trial, 9 advanced PC patients were vaccinated with 4 peptides comprising of KIF20A, cell division cycle-associated 1 (CDCA1), VEGFR1, and VEGFR2. Patients demonstrated increased anti-CDCA1 and anti-VEGFR2 CD8⁺ T-cells in circulation. Additionally, 4 out of 9 patients presented with stable disease post-vaccination [82]. In a subsequent study, vaccination with WT1 and VEGFR2 peptides generated HLA-A24-restricted CTLs, which demonstrated strong cytotoxicity towards PC cells that were HLA-A24-positive and expressed corresponding TAAs [83].

Prophylactic vaccines have recently been investigated as immunotherapy tools to target endogenous neoantigens by utilizing attenuated bacteria/virus to stimulate antitumor adaptive immune responses [84]. Listeria monocytogenes (LM) is a gram-positive bacterium that induces robust CD4⁺ and CD8⁺ T-cell responses by its selective infection of APCs (via actA virulence gene) over non-phagocytic cells (inlB gene) [85]. A LM ΔactA/ΔinlB strain engineered to express human MSLN (CRS-207 vaccine) was administered to 10 PC patients, which resulted in induction of expression of Th1 cytokines (IL-12, TNF-α). In addition, 6 out of the 10 patients developed MSLN-specific CD8⁺ T-cells [86]. Jaffee et al., conducted a clinical trial on 93 metastatic PC patients, in which 69 patients received 2 doses of cyclophosphamide with GVAX (Cy/GVAX) followed by 4 doses of CRS207 (Arm A) and 21 patients received 6 doses of Cy/GVAX (Arm B). Patients in Arm A showed an increased overall survival of 9.7 months compared to 4.6 months in Arm B, suggesting that the success of this immunotherapy might depend on the proper patient selection [87]. The Kras^G12D oncogene prophylactic vaccine (LM-K-ras vaccine) has been studied in KPC mice, either alone or in combination with T_reg depletion (anti-CD25 antibody, PC61, and low-dose Cy). KPC mice that received the vaccine at early PanIN 1 stages in combination with T_reg depletion showed prolonged survival compared to mice that received the vaccine alone suggesting the potential of T_reg depletion therapy as the prophylactic approach for PC [88].

STING is a transmembrane protein that resides on the endoplasmic reticulum, which upon activation through cyclic dinucleotides (CDNs), synthetic CDNs, or bacterial infection in the host induces IFN-β and NF-κB pro-inflammatory responses via IRF3 and Stat6 pathways [89]. It has been shown that endogenous STING activation via CDNs in the tumor microenvironment enhanced inflammatory responses, thereby inhibiting tumor progression and distant metastasis [90]. Recently, intraperitoneal injection of DMXAA, an activator of the mouse STING pathway, activated CD8⁺ T-cells that led to tumor rejection [90, 91]. Similarly, synthetic STING activators known as RR-CDGs have shown efficacy in the regression of primary pancreatic tumors, and distant metastatic lesions through T-cells recruitment in a TNF-α-dependent manner [92]. Furthermore, these novel synthetic activators of STING have demonstrated enhanced adjuvant activity to accelerate adoptive immune responses in the presence of radiation therapy [92]. The cGAMP-induced activation of endothelial cell-specific STING enhanced the antitumor responses of CD8⁺ T-cells and improved the responses of anti-CTLA-4 and anti-PD-1 immunotherapies [93].

3. Limitations of Current Strategies, Perspectives and Conclusion

The immune system has the potential to selectively target tumor cells upon strategic activation in cancer patients leading to better therapeutic outcomes. However, tumors employ extensive measures to escape immune surveillance, suggesting the necessity to develop novel counteracting strategies for the improved efficacy. Therefore, recent immunotherapeutic approaches alone or in combination with conventional treatment modalities need to be re-evaluated for successful therapeutic outcome in terms of improved patient survival.

PC presents strong immunosuppressive TME due to rigid tumor matrix architecture, T_regs infiltration along with constant antigen exposure mediated T-cell exhaustion, and upregulation of inhibitory receptors like PD-1 which collectively inhibit the infiltration and efficacy of effector T-cells and generate tolerance towards tumors [94, 95]. Thus, checkpoint inhibitors and depletion of Tregs could potentially reverse T-cell exhaustion of effector T-cells. Improper homing and inefficient infiltration of CAR T-cells to the tumor bed occur due to tumor blood vessels not responding to inflammatory stimuli. Anti-angiogenic therapy promotes normalization of tumor blood vessels, facilitating pericyte recruitment and increased tumor perfusion, which consequently increases the efficacy of CAR T-cell immunotherapy [96]. Apart from improper homing, CAR-T cells recognize TAAs that are also found at a lower level in normal tissues (which can cause toxicities in PC patients), thus if CAR T-cells survive for long periods of time in patients they have risk of developing autoimmunity in future [97]. Furthermore, activated CAR T-cells containing co-stimulatory domains like CD27, CD28 or 4-1BB release a variety of inflammatory cytokines like IL-2, IL-6, and IFN-γ after encountering tumor cells which induce macrophages to release more inflammatory cytokines thus establishing a positive cytokine-based feedback loop to enhance T-cell activity causing cytokine release syndrome (CRS), which could be fatal for patients [98, 99]. Development of short-lived CAR T-cells or combination treatment with an IL-6 receptor inhibitor like tocilizumab could be effective in reversing the effects of CRS without affecting the activity of CAR T-cells [98, 100].

Although KPC model mimics human PC, the more commonly used implantation models, do not recapitulate the complexity of TME, including the cytokine milieu present in PC patients. Further, many TAAs are expressed by normal cells as well as tumor cells and therefore cancer vaccines can potentially cause toxicities in the patients. Thus requires proper identification of TAAs that are exclusive to the cancer cells. Moreover, peptide-based cancer vaccines don’t capture all unique immunogenic epitopes present on the original antigens. Pancreas-specific transgene expression in spontaneous PC mice model and either protein fragments or intact proteins as immunogens could address these limitations faced in this field, thereby increasing the cancer vaccine efficacy. Additionally, selection of PC patient based on both tumor stage and tumor infiltrating lymphocytes (TILs) like CD8⁺ and PD-1⁺ T-cells could further increase the response to cancer vaccines [101].

Nanoparticles are capable of encapsulating multiple proteins, ligands, nucleic acids and other materials, thus increasing the epitope repertoire. Nanoparticles can also incorporate immune-stimulatory adjuvants (such as TLR agonists) or chemotherapeutic drugs to enhance the overall immunogenicity, stability, delivery and/or direct cytotoxicity of the vaccine, therefore overcoming the limitations of current cancer immunotherapies [102]. For example, mice immunized with Doxorubicin-CpG-PLGA microparticles showed a reduced tumor burden at lower drug concentrations compared to mice that received doses of soluble drug. When combined with anti-CTLA-4 antibody, the treatment successfully reduced aggressive tumor burden at both the injected and distant tumor sites in tumor-bearing mice [103]. This co-encapsulation of multiple therapeutics and immune stimulatory molecules may provide dose-sparing capabilities, reducing the cost and toxicity of cancer therapeutics [104].

Encapsulation into biodegradable nanoparticles provides protection of the payload until release [105, 106]. In addition, tuning the polymer chemistry enables sustained and controlled release of encapsulated payloads [107] and immunomodulatory capabilities [108]. Particularly, it has been demonstrated that varying chemistries of polyanhydride nanoparticles were efficiently internalized by APCs, leading to the upregulation of MHC I, MHC II and costimulatory molecules, as well as inducing the secretion of cytokines [109–111]. In addition, amphiphilic nanoparticles promoted the production of long-lived, high avidity antibody [112] with suboptimal doses of antigen [113], suggesting the development of long-lived plasma cells. Polyanhydride nanoparticles loaded with ovalbumin (OVA) induced memory CD8⁺ T-cells that were recruited and responded to subsequent challenges with OVA-secreting tumor cells [114]. Finally, many nanoparticles can be functionalized with ligands or antibodies that may increase selectivity and reduce the side effects of chemotherapeutics on healthy tissues [115]. Targeting moieties are often attached to the nanoparticle surface via a polyethylene glycol (PEG) linker [115]. This method of PEGylation allows for flexibility of the targeting moiety and may enhance interactions with cancer cell receptors [115, 116]. For example, PLGA nanoparticles covalently modified with folate via PEG demonstrated an increased association and uptake with cancer cells in vivo [117].

The limited success of immunotherapeutic studies performed in PC provides a generous room for improvement. Tailoring immunotherapy to PC patients by identifying unique tumor specific antigens through genetic screening and expression studies [29, 118]and combining it with continuous collection and screening of tumor samples in clinical trials to understand immunotherapy resistance, will further improve the response rates and survival benefits of PC immunotherapy.

Highlights.

• Pancreatic tumor microenvironment is immunosuppressive.

• Targeting certain components of tumor microenvironment abrogates immunosuppression.

• Several tumor-associated antigens have been evaluated for immunotherapy of PDAC.

• Combination therapies involving checkpoint blockade agents have been promising.

• Clinical & preclinical studies have identified challenges for immunotherapy of PDAC.