Immunotherapy in ovarian cancer

Abstract

Immunological destruction of tumors is a multistep, coordinated process that can be modulated or targeted at several critical points to elicit tumor rejection. These steps in the cancer immunity cycle include: (i) generation of sufficient numbers of effector T cells with high avidity recognition of tumor antigens in vivo; (ii) trafficking and infiltration into the tumor; (iii) overcoming inhibitory networks in the tumor microenvironment; (iv) direct recognition of tumor antigens and generation of an effector anti-tumor response; and (v) persistence of the anti-tumor T cells. In an effort to understand whether the immune system plays a role in controlling ovarian cancer, our group and others demonstrated that the presence of tumor infiltrating lymphocytes (TILs) is associated with improved clinical outcome in ovarian cancer patients. Recently, we hypothesized that the quality of infiltrating T cells could also be a critical determinant of outcome in ovarian cancer patients. In the past decade, several immune-based interventions have gained regulatory approval in many solid tumors and hematologic malignancies. These interventions include immune checkpoint blockade, cancer vaccines, and adoptive cell therapy. There are currently no approved immune therapies for ovarian cancer. Immunotherapy in ovarian cancer will have to consider the immune suppressive networks within the ovarian tumor microenvironment; therefore, a major direction is to develop biomarkers that would predict responsiveness to different types of immunotherapies, and allow for treatment selection based on the results. Moreover, such biomarkers would allow rational combination of immunotherapies, while minimizing toxicities. In this review, the current understanding of the host immune response in ovarian cancer patients will be briefly reviewed, progress in immune therapies, and future directions for exploiting immune based strategies for long lasting durable cure.

Introduction and background

Ovarian cancer is the most lethal of the gynecologic malignancies. Over 22 000 new cases of ovarian cancer are diagnosed each year in the United States resulting in greater than 14 000 deaths per year [1–3]. Despite aggressive frontline treatments with surgery and adjuvant chemotherapy, the 5-year survival rate is <25% for women diagnosed with stages III or IV disease. Although >80% of these patients will initially have a response to therapy, the majority ultimately recur, and eventually develop chemotherapy-resistant disease. The results of recent clinical trials, including trials that incorporate bevacizumab into chemotherapy regimens suggest that a plateau has been reached for conventional therapies as there is no definitive increase in overall survival. Consequently, new treatment strategies and paradigms are of great need for these patients. Among these, immunotherapy has attracted significant interest with recent improved understanding of the molecular basis of immune recognition and immune regulation of cancer cells. These interventions include immune checkpoint blockade, cancer vaccines, and adoptive cell therapy. Indeed, several immune check point inhibitors were recently approved by the Food and Drug Administration for a variety of cancers including melanoma, non-small-cell lung cancer (NSCLC), renal cell carcinomas, bladder cancer, and classical Hodgkin lymphoma. Unfortunately, there are currently no approved immune therapies for ovarian cancer. In this review, the current understanding of the host immune response in ovarian cancer patients will be briefly reviewed. Progress in immune therapies that include cancer vaccines, cell-based therapy, immune checkpoint blockade, and oncolytic virus-based therapy will be discussed, and we will briefly explore future directions for exploiting immune based strategies for long lasting durable cure.

The host immune response in ovarian cancer patients

Immunologic destruction of tumors is a multistep, coordinated process that can be modulated or targeted at several critical points to elicit tumor rejection. Although immune rejection of ovarian cancer has been demonstrated in several pre-clinical animal models, correlative human studies in patients with ovarian cancer strongly support a role for immune system involvement in patient outcome.

In an effort to understand whether the immune system plays a role in controlling ovarian cancer, our group and others demonstrated that the presence of tumor-infiltrating lymphocytes (TILs) is associated with improved clinical outcome in ovarian cancer patients [4–6]. Patients with increased frequencies of intraepithelial CD8+ TIL (55 versus 26 months, HR = 0.33, 95% CI 0.18–0.60, P = 0.0003) [4]. A meta-analysis of ten studies with 1815 ovarian cancer patients confirmed the observation that a lack of intraepithelial lymphocytes (TILs) is significantly associated with a worse survival among ovarian cancer patients [6]. Together, these studies support the notion that tumor infiltration by lymphocytes is a reflection of a tumor-related immune response. To harness the antitumor immune response, all of the steps in the cancer immunity cycle must be optimal and include [7–9]: (i) generation of sufficient numbers of effector T cells with high avidity recognition of tumor antigens in vivo, (ii) trafficking and infiltration into the tumor, (iii) overcoming inhibitory networks in the tumor microenvironment, (iv) direct recognition of tumor antigens and generation of an effector antitumor response, and (v) persistence of the antitumor T cells.

Immune suppressive networks in ovarian cancer

A major barrier to successful deployment of cancer immunotherapy for ovarian cancer patients is the immunosuppressive tumor microenvironment. Even if large numbers of tumor-specific T cells are generated in patients by immunotherapy, these T cells may not readily destroy tumor targets in vivo. Previous studies by our group and others have defined some of the dominant immune resistance mechanisms in ovarian cancer to include extrinsic suppression of CD8⁺ effector cells by regulatory T cells (Tregs) [4, 10]; metabolic deregulation via tryptophan catabolism by the immunoregulatory enzyme indoleamine-2,3-dioxygenase (IDO) [11, 12]; engagement of the inhibitory receptor PD-1 by the ligand PD-L1/B7-H1 [13–15] and development of antigen loss variants [16]; myeloid derived suppressor cells [17], and inhibitory cytokines such as TGF-β. Collectively, this redundant immunosuppressive network facilitates tumor progression by actively restraining endogenous antitumor immunity and serves as an important obstacle that must be overcome in order to implement efficacious immunotherapeutic strategies.

Evidence in several cancer systems has shown that T-cell expression of inhibitory immune checkpoint receptors is one mechanism by which tumors evade or dampen host immunity. These receptors negatively regulate T-cell function and include CTLA-4, PD-1, LAG-3, TIM-3, TIGIT, and others. Interfering with CTLA-4 and PD-1 have demonstrated clinical benefit in several human cancers. CTLA-4 regulates T-cell priming and activation, leading to expansion of auto-reactive T cells, including tumor-specific T cells. Consequently, anti-CTLA-4 therapies are associated with more significant immune-related toxicities compared with PD-1 blockade. PD-1 is a cell surface receptor that interacts with two known ligands, PD-L1 and PD-L2, resulting in inhibition of T-cell signaling and cytokine production.

Blockade of these inhibitory receptors with specific antibodies is designed to reinstate an existing antitumor response. This has been achieved through three general strategies: (i) the inhibition of the immunosuppressive receptors expressed by activated T lymphocytes, and (ii) the inhibition of the principal ligands of these receptors, such as the PD-1 ligand (PD-L1). Several antibodies directed against PD-1, PD-L1, and CTLA-4 have been developed and are being tested clinically in patients with ovarian cancer. Although these immune checkpoint blocking antibodies have shown significant promise in mediating tumor regression in melanoma and other solid tumors, the response rates in ovarian cancer have been modest. The first published data supporting checkpoint inhibitors as a potentially valuable therapeutic approach in ovarian cancer were observed in trials of the anti-PD-1 antibody nivolumab and the anti PD-L1 antibody BMS-93655 [18]. In the study reported by Hamanishi et al. [19], the best overall response rate in 20 assessable patients treated with nivolumab was 15% and ae disease control of 45%.

Two additional immune checkpoint trials using the anti-PD-L1 antibody avelumab and the anti-PD-1 antibody pembrolizumab were presented at the 2015 ASCO annual meeting [20, 21]. Of 75 heavily pre-treated ovarian cancer patients that received avelumab, 8 patients experienced partial responses, 33 patients had stable disease, and there were no complete responses, resulting in a disease control rate of 54.7%. In the KEYNOTE-028 Phase-1b study of pembrolizumab in 26 heavily pre-treated ovarian cancer patients with a PD-L1 expression level of ≥1% of tumor cells, the results showed one complete response, two partial responses, and six patients with stable disease, corresponding to a disease control rate of 34.6%. Ongoing or planned phase 3 trials in ovarian cancer with immune checkpoint inhibitors include NCT02718417 (Javelin Ovarian 100), ENGOT-ov29-GCIG (ATALANTE), NCT02580058 (Javelin Ovarian 200), and NRG-GY009. These trials are testing combinations with chemotherapy and/or bevacizumab, or the potential efficacy as maintenance therapy.

While the initial results of checkpoint inhibitors in ovarian cancer are promising, they are modest compared with the spectacular results reported for melanoma, renal cell, NSCLC, and bladder cancer. Potential mechanisms underpinning the limited antitumor efficacy of immune checkpoint inhibitors in ovarian cancer include (i) low intrinsic tumor immunogenicity and mutational burden in ovarian compared with other cancers; (ii) expression of multiple co-inhibitory receptors on T cells infiltrating ovarian cancer; (iii) compensatory upregulation of multiple immune checkpoints when one of them is blocked; and (iv) redundant immune suppressive mechanisms (e.g. TGF-β, IDO). All of these factors may counteract the beneficial effects of checkpoint blockade. In our studies, we have demonstrated that multiple inhibitory receptors are often co-expressed on tumor-antigen specific CD8⁺ T cells [13]. In human ovarian cancer, tumor antigen-specific CD8⁺ TILs co-express PD-1 and LAG-3 and are impaired in IFN-γ and TNF-α production compared with PD-1 or LAG-3 single positive cells. Simultaneous blockade of PD-1 and LAG-3 ex vivo restored effector function to a greater extent than single checkpoint blockade, thereby suggesting that monotherapy of checkpoint blockade may not be sufficient for eliciting robust antitumor responses. Importantly, we showed that blockade of PD-1, LAG-3, or CTLA-4 alone using genetic ablation or blocking antibodies conferred a compensatory upregulation of the other checkpoint pathways, potentiating their capacity for local T-cell suppression that, in turn, could be overcome through combinatorial blockade strategies [15]. Durable antitumor immunity was most strongly associated with increased numbers of CD8⁺ T cells, the frequency of cytokine-producing effector T cells, reduced frequency of Tregs and arginine-expressing monocytic myeloid-derived suppressor cells in the peritoneal TME. These data provide a basis for combinatorial checkpoint blockade in clinical intervention for ovarian cancer.

Shared tumor antigens, neoantigens, and cancer vaccines in ovarian cancer

In order to generate effector T cells with ability to recognize tumor in vivo, cancer vaccines emerged as an immunotherapeutic approach that could harness the immune system in extending remission rates and prevent further malignant growth. The biologic principle of cancer vaccines is to stimulate an immune response specifically directed against malignant cells. In this manner, cancer vaccines may be used prophylactically and therapeutically. For prophylactic vaccination, the goal is to mount an immune response that will recognize and eradicate cancer cells early enough to prevent malignant progression. As a complement to the prophylactic approach, cancer vaccines may also be used therapeutically to serve as a ‘booster’ for pre-existing antitumor immune responses or activating antitumor immunotherapies that have been actively administered to the patient. The adaptable nature of cancer vaccines is partially governed by the nature of the tumor antigen and has been leveraged to elicit anticancer immune responses in a multitude of cancer immunotherapy applications. A major question is identifying the most effective and safe vaccine targets in ovarian cancer.

Human tumor antigens defined to date can be classified into one or more of the following categories: (i) differentiation antigens that are restricted to very defined tissues, (ii) mutational antigens, (iii) amplification antigens, (iv) splice variant antigens, (v) glycolipid antigens, (vi) viral antigens, and (vii) cancer testis (CT) antigens. Although there are several options in deciding which antigen to target, the fundamental requirements of the ideal tumor antigen (TA) include: (i) limited or no expression in normal tissues, but aberrant expression at high frequencies in tumor; (ii) immunogenicity; and (iii) a role in tumor progression. While none of the current TAs completely meet all of these criteria, the family of CT antigens are closest. CT antigens are a subclass of TAs encoded by ∼140 genes. The criteria for placing antigens in this category are based on several characteristic features [22, 23]: (i) predominant expression in germ cells of the testis and generally not in other normal tissues, (ii) expression in a proportion of malignant tumors of different histological types, (iii) expression in malignancies in a lineage non-specific fashion, (iv) often mapping of the gene on the X-chromosome, and (v) often members of multigene families. Despite their poorly characterized biologic function, expression of these antigens are known to be restricted in immune privileged sites such as the testes, placenta, and fetal ovary, but not in other normal tissues. Abnormal expression of these germ-line genes in malignant tumors may reflect the activation of a silenced ‘gametogenic program’, which ultimately leads to tumor progression and broad immunogenicity [24]. The immunogenicity of CTA has led to the widespread development of cancer vaccines targeting these antigens (e.g. NY-ESO-1 and MAGE) in many solid tumors (Table 1).

Table 1.

Selected ovarian cancer/solid tumor vaccine studies targeting NY-ESO-1

Antigen	Phase	Disease	Technology	Co-therapy	Sponsor	Reference
NY-ESO-1
	I	Ovarian, fallopian tube cancer	DEC-205 Fusion protein	Poly-ICLC, IDO1 inhibitor	Roswell Park Cancer Institute	NCT02166905
	I	NY-ESO-1 expressing solid tumors	DEC-205 Fusion protein/Dendritic cell	Rapamycin	Roswell Park Cancer Institute	NCT01522820
	I/II	NY-ESO-1 expressing tumors	DEC-205 Fusion protein	Resiquimod, Poly-ICLC	Celldex Therapeutics	NCT00948961
	I	NY-ESO-1 expressing tumors	Full length protein	Montanide, Resiquimod	Mount Sinai School ofMedicine	NCT00821652
	I	Ovarian, fallopian, primary peritoneal cancer	Peptide	Decitabine, Doxorubicin, Montanide	Roswell Park Cancer Institute	NCT01673217
	I	NY-ESO-1/LAGE-1 expressing tumors	Peptide	CpG7909, Montanide	Ludwig Institute for Cancer Research	NCT00199836
	I	Ovarian, fallopian, primary peritoneal cancer	Peptide	Montanide	Memorial Sloan Kettering Cancer Center	NCT00066729
	I	Ovarian, fallopian, primary peritoneal cancer	Overlapping Long peptides (OLP4)	Montanide, Poly-ICLC	Ludwig Institute for Cancer Research	NCT00616941
	I	Ovarian, fallopian, primary peritoneal cancer	Vector (ALVAC(2)-NY-ESO-1(M) TRICOM)	GM-CSF, Rapamycin	Roswell Park Cancer Institute	NCT01536054
	I	Ovarian, fallopian, primary peritoneal cancer	Vector (ALVAC(2)-NY-ESO-1(M) TRICOM)	GM-CSF	Ludwig Institute for Cancer Research	NCT00803569
	II	Ovarian, fallopian, primary peritoneal cancer	Vector (Fowlpox-NY-ESO-1)	Recombinant Vaccinia-NY-ESO_1	Ludwig Institute for Cancer Research	NCT00112957

Among CT antigens, NY-ESO-1 is one of the most spontaneously immunogenic tumor antigens described. The NCI antigen prioritization panel has ranked NY-ESO-1 in the top 10 antigens for further development of immunotherapies. It has two characteristics that make it a viable candidate: (i) testis-restricted expression in normal tissues and (ii) its immunogenicity. The initial study of NY-ESO-1 using a combination of RT-PCR and IHC indicated aberrant expression of NY-ESO-1 in up to 43% of ovarian cancer patients [25]. This study was recently extended to more than 1000 ovarian cancer patients and the frequency was confirmed at 40% [26]. The antigen elicits both cellular and humoral immune responses in a high proportion of patients with NY-ESO-1-expressing tumors. Emerging evidence also suggest that NY-ESO-1 and some additional CT antigens, may be selectively expressed by cancer ‘stem cells’. Therefore, the development of strategies to target CT antigens in ovarian cancer could have potential therapeutic benefit. A number of NY-ESO-1-based clinical trials have been conducted in ovarian cancer patients, and additional trials are on-going. The vaccines have included long peptides [16, 27], heterologous prime boost approach with recombinant vaccinia expressing NY-ESO-1 followed by recombinant fowlpox expressing NY-ESO-1 [28], and NY-ESO-1 protein in combination with epigenetic modification [29]. While the individual clinical trials demonstrated signal of clinical benefit, a recent retrospective combined analysis of patients on these trials demonstrated a 2-year overall survival advantage for patients with NY-ESO-1 positive tumors who receive vaccination, compared with patients with NY-ESO-1 positive tumors who did not receive vaccination [26]. While these results should be interpreted with caution, the conclusion that NY-ESO-1 targeted therapy may be associated with clinical benefit has become inescapable, and would need to be tested in randomized clinical trials. In the meantime, adoptive transfer of NY-ESO-1 specific T cells is another on-going promising strategy that is discussed in greater detail below.

Advances in next-generation sequencing and epitope prediction now permit the rapid identification of mutant tumor neoantigens. This has led to efforts in utilizing these mutant tumor neoantigens for personalizing cancer immunotherapies. Indirect support for this approach comes from studies demonstrating that (i) infusion of autologous ex vivo expanded TILs can induce objective clinical responses in metastatic melanoma [30], and (ii) the relationship between pre-therapy CD8⁺ T-cell infiltrates and response to checkpoint blockade in melanoma [31]. Deep-sequencing technologies permit easy identification of the mutations present within the protein-encoding part of the genome (the exome) of an individual tumor allowing for prediction of potential neoantigens. Several pre-clinical and clinical studies have now confirmed the possibility of identifying neoantigens on the basis of cancer exome data [32–36]. Although there are limitations of probing the mutational profile of a tumor in a single biopsy [37, 38], it is evident that the vast majority of neoantigens occur within exonic sequence and do not lead to the formation of neoantigens that are recognized by autologous T cells [38, 39]. Consequently, a robust pipeline for filtering the cancer exome data is essential. Epitope presentation of neoantigens by MHC class I molecules may be predicted using previously established algorithms that analyze critical features such as the likelihood of proteasomal processing, transport into the endoplasmic reticulum, and affinity for the relevant MHC class I alleles.

In order to predict epitope abundance, gene and/or protein expression levels can also be integrated into the analysis. Based on these considerations, it becomes of interest to stimulate neoantigen-specific T-cell responses in cancer patients using two possible approaches. The first is to synthesize long-peptide vaccines that encode a set of predicted neoantigens. The second approach is to identify and expand pre-existing neoantigen-specific T-cell populations to create either bulk neoantigen-specific T-cell products or TCR engineered T cells for adoptive therapy. Although the use of neoantigen-based vaccines has not yet been reported in ovarian cancer clinical trials, it is anticipated that such clinical trials will be developed in the near future.

Adoptive T-cell therapy

Adoptive cell therapy (ACT) involves the ex vivo selection of antigen-specific T cells and their expansion to desired magnitude for achieving a targeted immune response. In contrast to vaccine strategies, ACT is free of the in vivo immune suppressive constraints that can limit the magnitude, duration, and phenotype of a desired antitumor immune response achieved by other passive immunotherapy approaches. T cells used for ACT can be derived from peripheral blood lymphocytes (PBL) or TILs. Upon modification and expansion ex vivo, the activated T cells are re-infused into patients usually after they have received a lymphodepleting pre-conditioning chemotherapy.

Initial studies demonstrating the potential of T-cell immunotherapy to eradicate solid tumors came from the NCI in studies of adoptive transfer of in vitro selected TILs. Unfortunately, methods of isolating and manufacturing TILs are labor intensive and only successful in a subset of patients. In order to improve the therapeutic potential of transferred cells, investigators have recently focused on genetic modification of PBLs to exhibit tumor antigen specificity. ACT using genetically engineered PBLs to express antitumor receptors holds promise for extending the use of ACT to patients with epithelial cancers, such as ovarian cancer.

T cells can be genetically modified to express either (i) a tumor antigen-specific T-cell receptor (TCR) encoding the α and β chains with specificity for tumor-restricted peptide expressed on a given HLA molecule or (ii) a ‘chimeric antigen receptor’ (CAR) encoding a transmembrane protein comprising the tumor antigen-binding domain of an immunoglobulin linked to one or more T-cell costimulatory molecules. CAR T cells have also demonstrated encouraging results of inducing complete response in 70%–90% of patients with relapsed or refractory B-cell acute lymphoblastic leukemia. ACT using TCR-engineered T cells have resulted in objective responses in the majority of treated patients. In ovarian cancer, there are two on-going trials evaluating the efficacy of TILs (NCT02482090, NCT01883297). Phase I studies of engineered T cells targeting MUC16 (NCT02498912), mesothelin (NCT01583686), and NY-ESO-1 (NCT01567891, NCT02457650) are also on-going. Although spectacular responses have been observed, the majority of clinical responses are short-lived with ultimate tumor relapse. A major explanation for this sub-optimal outcome is the relatively limited long-term survival and effector function due to suppression or exhaustion of infused engineered T cells.

Recent reports indicate that T cells that are expanded ex vivo to maintain more stem-like T-cell populations known as T stem cell memory (Tscm) cells, are capable of a more sustained response by replenishing effectors. A clear benefit of transferring less mature, more stem-like cells is likely due to increased persistence and replenishing capability of these cells in vivo. Conceptually, the regenerative nature of hematopoietic stem cells may provide a long-lasting, potentially life-long supply of effector T cells engineered against tumor antigens by TCR gene-modification of PBLs. This approach is currently being tested at Roswell Park Cancer Institute in a clinical trial in ovarian cancer patients.

Conclusions and future directions

Beyond the PD-1 and CTLA-4 pathways, there are additional tolerogenic mechanisms that should be targeted in ovarian cancer as part of novel combination therapies. For example, one of the most critical tolerogenic mechanisms in ovarian cancer is mediated by indole-amine-2,3,-dioxygenase (IDO), an immunoregulatory enzyme that catalyzes the rate-limiting step of tryptophan degradation along the kynurenine pathway. Both the reduction in local tryptophan levels and the production of tryptophan catabolites that are inhibitory to cells contribute to the immunosuppressive effects [40], culminating in multipronged negative effects on T lymphocytes notably on proliferation, function, and survival. IDO activity also promotes the differentiation of naïve T cells to cells with a regulatory phenotype (Treg) [41]. Since increased Treg activity has been shown to promote tumor growth and Treg depletion has been shown to allow an otherwise ineffectual antitumor immune response to occur [42], IDO expansion of Tregs may provide an additional mechanism whereby IDO could promote an immunosuppressive environment. An on-going clinical trial is testing the combination of NY-ESO-1 vaccination with an inhibitor of IDO as a strategy to counteract IDO-mediated resistance to immunotherapy (NCT02166905). This clinical trial tests a novel strategy of selectively breaking IDO-mediated immune tolerance and simultaneously promoting the generation of tumor antigen specific T cells via vaccination against NY-ESO-1. The NY-ESO-1 vaccine in this trial vaccine is composed of DEC205mAb-NY-ESO-1 fusion protein (CDX-1401) with adjuvant polyICLC. The DEC-205 receptor has been shown to be an efficient mAb-based target to enhance the induction of strong Ag-specific immune responses and cross-presentation in mice [43] and humans [44]. The trial is focused on ovarian cancer patients in first or subsequent remission. The success of the trial could lead to the development of a novel strategy to lengthen remission rates in ovarian cancer patients and minimize the risk of relapse.

Another promising approach is the development of novel synthetic biology techniques that are leading to a new generation of TCR or CAR modified T cells. Such strategies include the incorporation of cytokines such as IL-12 leading to their local delivery in the tumor microenvironment [45–48], the inclusion of decoy receptors for inhibitory molecules such as TGF-β or PD-1 [49, 50], and the inclusion of suicide gene to ablate T cells and abrogate on-going side-effects [51–53]. Hopefully, these next generation approaches would be tested in ovarian cancer in the near future, alone and in combination with other immunotherapies.

Finally, immunotherapy in ovarian cancer will have to consider the immune suppressive networks within the ovarian tumor microenvironment. Since this is likely to be shaped by the intrinsic biologic properties of the tumor, a major direction is to develop biomarkers that would predict responsiveness to different types of immunotherapies, and allow for treatment selection based on the results. Moreover, such biomarkers would allow rational combination of immunotherapies, while minimizing toxicities.

Funding

This work was support by NCI grant P30CA016056, P50CA159981, R01CA158318, and Roswell Park Alliance Foundation. The publication of this supplement and the symposium on which it is based have been supported through partnership between the Spanish Ovarian Cancer Research Group (GEICO) and the European Society for Medical Oncology (ESMO).

Disclosure

The author has declared no conflicts of interest.