Next Generation Therapeutic Strateg‐Es: Evolving cancer immunotherapy through agents that Engage, Expand and Enable the anti‐tumor immune response

The development and FDA approval of immune checkpoint blocking antibodies have brought new light to cancer immunotherapy. While immune checkpoint blockade (ICB) has demonstrated clinical benefit in certain tumors as monotherapy, effective therapy of established tumors necessitates a combination of multiple immuno‐oncology agents targeting diverse functions of the immune system. These combination strategies should be tactically designed to Engage the immune system by inducing a tumor‐antigen specific T‐cell population, Expand the number of antigen‐specific cytotoxic T cells and increase their migration to the tumor microenvironment, and once there, Enable prolonged and persistent effector function. Although viral therapeutic cancer vaccines have demonstrated little efficacy as monotherapies, they have substantial potential to Engage the immune system as one branch of a multipronged treatment strategy. This review will summarize prior and ongoing Phase II and III clinical trials built upon the foundation of viral therapeutic cancer vaccines. We examine their efficacy as a monotherapy, and more importantly, when combined with additional agents that Expand and Enable the immune system. It is clear that the future of cancer immunotherapy will include evolving treatment strategies made up of multiple agents, and we are optimistic that in this context viral therapeutic cancer vaccines will emerge as an important part of next generation effective therapeutic strategies.


| INTRODUC TI ON
Over the past century, cancer treatment evolved from monotherapy using basic chemotherapy drugs to strategies combining advanced chemotherapy with radiation, surgery, and targeted molecular therapies. 1 It has already become clear that cancer treatment in the 21st century will be defined by immunotherapy. Determining the amount and activity of antigen-specific T cells in the tumor and tumor microenvironment (TME) is a key determining factor for immunotherapy efficacy. The relationship between immune function and tumor immune context was first described in colorectal cancer, where the presence of T cells was shown to be highly important. 2 Based on these and related findings, 3 the immune context of various tumors is described as a spectrum ranging from "cold" tumors that have little to no immune effector infiltration and high levels of suppressive immune cells such as myeloid-derived suppressor cells (MDSCs) to "hot" tumors that are highly inflamed and have significant amounts of effector immune infiltrate, especially T cells and cytotoxic T cells, and low MDSC infiltration. 4,5 Because ICB is dependent in part on the presence of intratumoral tumor antigen-specific T cells, it is most effective in "hot" tumors such as bladder, head and neck, melanoma, and non-small cell lung cancer. In immunologically "cold" tumors that lack T-cell infiltration there are no targets for immune checkpoint blockade drugs, and these treatment strategies are ineffective. In order to treat patients presenting with "cold" tumors, alternative strategies must be employed to engage the immune system by bolstering immune surveillance and elimination. 5

| Engage: Priming the immune system
One of the most deeply researched strategies for engaging the immune system is through vaccination. Since their invention in the late 18th century, prophylactic vaccines have been incredibly efficacious in preventing the spread of infectious disease. While certain cancers are caused by infectious diseases (human papillomavirus, hepatitis B virus) and effective preventative vaccines have been developed for them, 6 the majority of cancers are the result of spontaneous mutation or environmental factors, and therefore cannot be prevented in this way. However, the mechanism of prophylactic vaccines is one that can be co-opted for the treatment of already existing tumors.
Therapeutic vaccines generate a tumor antigen-specific T-cell response by stimulating a patient's immune system with a tumor antigen in an immunogenic formulation that activates an immune response. Cancer vaccines can thus create a population of tumor antigen-specific cytotoxic T cells, resulting in an increased number of cancer-specific T cells, and turning "cold" tumors "hot." There have been several cancer vaccine success stories, 7,8 but decades of research into therapeutic cancer vaccines have demonstrated that despite being well tolerated by patients, 9 pursuing therapeutic cancer vaccines as a monotherapy or even in combination with cytokine to induce expansion is not a fruitful treatment strategy. 10-12

| Targeting diverse immune functions for effective immunotherapy
These results and others have shown that cancer immunotherapy in the 21st century is already beginning to follow the same trend as chemotherapy in the 20th century. The failures of monotherapy strategies for multiple classes of immuno-oncology (IO) agents have led researchers to the conclusion that a tactically designed combined approach, simultaneously targeting diverse immune-tumor interactions, is necessary for the optimal immunotherapeutic treatment strategy. Interactions between growing tumors and the immune system were concisely described by Dunn et al as the Three E's of Immunoediting. 13 Building off of this foundation, rationally selected combination immunotherapy should include agents capable of enabling at least three aspects of the immune system ( Figure 1 In this review, we will focus on efforts to Engage the immune system through viral-based tumor vaccines. Although they have demonstrated little efficacy as monotherapy, these vaccines have enormous potential as one leg of a multipronged treatment strategy.
We will discuss current clinical trials combining viral-based tumor vaccines with agents that Expand and Enable the immune system, as well as novel agents that will Evolve the current clinical strategies for cancer treatment.

| VIR AL VEC TOR BA S ED C AN CER VACCI N E S
There are two initial choices in therapeutic vaccine design, the vaccine target and the platform. Ideal targets should be present on cancer cells but not normal cells, highly immunogenic and necessary for tumor survival. 14 Vaccine targets are grouped into two major categories, tumor-associated antigens (TAA), which are selfantigens upregulated in tumor cells compared to normal tissue, and tumor-specific antigens, which are only expressed in tumor tissue.
Tumor-associated antigens have the advantage of being common to multiple patients and tumor types. However, because they are selfantigens, cancer vaccines must combat immune tolerance against them to stimulate a rare population of T cells. Additionally, some tumor-associated antigens are still expressed on normal tissues, which can lead to off-target effects and toxicity. Tumor-specific antigens such as oncoviral antigens, oncogenes, and neoantigens are specific to tumor tissue. While oncoviruses associated with cancer etiology have led to the development of prophylactic cancer vaccines, 15 20 and more than a dozen additional gene therapy-based drugs are on the market worldwide. 21 Oncolytic viruses are viruses that selectively infect tumor cells. These viruses then undergo a lytic life cycle, resulting in tumor cell death, spreading the virus and releasing tumor-associated antigens. 22 This immunogenic cell death enhances the immune response and boosts immune-mediated cell killing. 23 Oncolytic virus-based therapies have been developed across multiple viral platforms, including adenoviruses, coxsackie virus, herpes simplex virus, measles virus, vaccinia and others. 22 Currently, one oncolytic viral therapy, talimogene laherparepvec (T-VEC), has been approved for use in melanoma, 8 with clinical trials underway in most solid tumors. 24 T-VEC is a herpesvirus that was genetically modified to increase its immunogenicity by removing the genes ICP34.5 and ICP47 and replacing them with the gene for GM-CSF, an immunostimulatory cytokine. 25 Treatment with T-VEC plus GM-CSF resulted in an overall survival (OS) of 23.3 months compared to 18.9 in the GM-CSF alone arm, with 88.5% of patients estimated to survive at 5 years post-treatment. 8 There are currently approximately 100 ongoing clinical trials utilizing oncolytic viruses, and it is clear that this will be an ongoing direction of research and treatment for years to come.

| ENGAGING THE IMMUNE SYS TEM
We have identified three broad clinical modalities that are effective at engaging the immune system, resulting in increased numbers of antigen-specific T cells: 1. Chemotherapy-and radiation-induced immunogenic cell death, 2. Adoptive immune therapy, 3.

F I G U R E 1
Four Es of effective anti-tumor combination immunotherapy. IO, immuno-oncology; TME, tumor microenvironment Therapeutic vaccines (Figure 1). Substantial preclinical and clinical data have demonstrated that certain doses of standard-of-care chemotherapy and radiation cause immunogenic cell death. 26 In immunogenic cell death, tumor cell death results in changes to cell surface markers and the release of soluble factors that stimulate the presentation of tumor antigens to T cells, increasing the proportion of antigen-specific T cells. 27 Adoptive immune therapy utilizes exogenous expansion of a patient's own immune cells, often including genetic engineering to make the cells specific for tumor antigen (CAR-T). 28 Three CAR-T therapies are currently FDA approved, targeting the CD19 antigen in acute lymphoblastic leukemia, diffuse large B-cell lymphoma, 29 and mantle cell lymphoma. 30,31 Despite these successes, CAR-T and adoptive immune therapy face challenges such as antigen selection, the cost and labor associated with personalized therapy, and antigen escape.
It is likely that, similar to therapeutic vaccines, successful application of CAR-T will involve combination with additional agents that Enable and Expand the immune system. 32 The earliest research into therapeutic vaccines focused solely on the vaccine itself, targeting TAAs to promote immune destruction of the tumor. Phase I studies consistently demonstrated the safety of cancer vaccines, 9 and several viral vaccine platforms have moved from Phase I to Phase II studies in the monotherapy form (Table 1).

| Adenoviral vaccines
Two different adenovirus-based vaccines have entered Phase II studies. Patients with hormone-refractory and recurrent prostate cancers are being treated with an adenoviral vaccine utilizing the TAA prostate-specific antigen (PSA), a highly prevalent antigen in prostate cancer that is also utilized as a serum biomarker for cancer progression. 40 High proportions of both populations exhibited anti-PSA T-cell immune responses, with the majority of patients experiencing a decrease in serum PSA or an increase in PSA doubling time. 33 A separate study treated patients with colon, lung, or breast cancer with ETBX-011, an adenovirus-5 vaccine targeting carcinoembryonic antigen (CEA), a common TAA across many solid tumors. 35,41,42 The 19 patients in the Phase II cohort experienced a 12-month survival probability of 48%, and 10/19 (53%) had a positive CEA-directed cell-mediated immune response by ELISPOT. Importantly, this trial demonstrated that pre-existing immunity to Adenovirus subtype-5 did not significantly impact survival outcomes. 41

| Poxviral vaccines
The majority of monotherapy therapeutic vaccine trials have been performed using poxviruses. Several trials have utilized the vaccine TroVax, a modified vaccinia Ankara-based vaccine that targets the oncofetal antigen 5T4. 5T4 expression is associated with a tumor-initiating phenotype and highly expressed in tumor cells compared to normal tissue in multiple solid tumors. 37 38 and Elkord et al demonstrated that patients with a greater than average 5T4-specific T-cell proliferative response had a significant survival advantage compared to those who did not, indicating that immune response to the vaccine and density of CD3 cells was likely driving increased survival. 34 Following these data, a double-blind Phase II study examining TroVax in ovarian, fallopian tube, and peritoneal cancer was initiated and is currently ongoing with a primary endpoint of RECIST-defined progression at 25 weeks. 43

| Prime/boost
Two monotherapy Phase II trials have been conducted that utilized a "diversified prime/boost" vaccination method. In this strategy patients are "primed" with a recombinant vaccinia vaccine, followed by subsequent vaccinations (boosts) with avipox vaccines. This approach is commonly utilized when treating with vaccinia-based vaccines, as vaccinia-immune patients are able to mount an immune response to the vaccine antigen after the first vaccination but not to the second. 18 This strategy was tested in a Phase II trial utilizing rV-PSA (recombinant vaccinia) followed by rF-PSA (recombinant fowlpox) in prostate cancer. 45.3% of patients were free of PSA progression at 19.1 months, and 46% of patients had an increase in PSA-reactive T cells. 39 This trial also reported a trend favoring the treatment group receiving vaccinia prime, which was verified in a later trial that also demonstrated the improved efficacy of prime-boost as opposed to vaccination with fowlpox alone. 18,39,44 Monotherapy prime-boost has also been investigated targeting the cancer-testis antigen NY-ESO-1 in fallopian tube cancer, ovarian cancer, peritoneal cancer, and melanoma. This recent Phase II trial found that melanoma patients had an objective response rate of 14% with a median progression-free survival (PFS) of 9 months

| Proinflammatory cytokines
The cytokines most frequently combined with immunotherapy in the clinic are Interleukin-2 (IL-2) and granulocyte-macrophage results are yet reported 51 (Table 2). While few trials examine the efficacy of vaccine and cytokine alone, cytokines are a common agent in the multi-combination trials found in Tables 3 and 4.   (Table 2). While no Phase III trials have been conducted utilizing PROSTVAC-V/F alone, one was conducted combining PROSTVAC-V/F with GM-CSF which will be discussed below. 11

| Endocrine deprivation
Preclinical evidence has demonstrated that endocrine depriva- The data from these trials identify two primary points of con-  agents are the primary mechanism for enabling the immune system, reversing the mechanisms of T-cell exhaustion and allowing infiltrating T cells to have greater and longer efficacy. There are two additional modes of enabling the immune system, metabolic support and radiation/chemotherapy induced immunogenic modulation (Figure 1).
Two additional currently recruiting Phase II trials are also applying vaccines built on the modified vaccinia Ankara virus. The first is treating patients with PD-L1 positive fallopian tube, ovarian or peritoneal cancers, and is treating with an MVA-based vaccine expressing the TAA p53 in combination with pembrolizumab with a primary outcome of response rate. 63

| Immunogenic modulation
Instead of directly enabling the immune system, it is also possible to promote increased immune efficacy by sensitizing tumor cells directly.   While cytolytic cancer therapies aim to completely remove the tumor, most are not capable of killing all cancer cells. However, when patients who recurred following chemotherapy and radiation therapy were put on immunotherapy, it was observed that they had a higher clinical benefit than those who had not been previously treated. 103  In addition to these completed trials, one currently ongoing study in colorectal cancer is combining checkpoint inhibition with other enabling agents. This trial is utilizing the adenoviral vaccine Ad-CEA and additionally treating with the anti-PD-L1 antibodies avelumab, FOLFOX, and bevacizumab. 68 One therapeutic viral vaccine has entered Phase III in combina-

| Metabolic support
Similar to checkpoint inhibition, metabolic support targets the immune system. Depending on the make-up of the tumor microenvironment, cancer cells can deprive the TME of glucose and increase levels of lactate in the TME, both of which have been shown to inhibit the functions of TILs. 113 It is possible that these could be addressed in part through "metabolic tuning" of T cells ex-vivo prior to re-introduction during adoptive immune therapy. Alternatively, metformin, a drug that changes mitochondrial respiration, has been shown to promote TIL function, indicating its potential as an immunometabolism modulating therapeutic agent in the clinic. 114 Another metabolic target is the cytosolic enzyme indoleamine 2,3-dioxygenase-1 (IDO1), which has been shown to have immunosuppressive effects. Widely expressed in human tumors, IDO1 catalyzes tryptophan, a necessary metabolite for T-cell function.
In addition to depriving T cells of tryptophan, the catabolites of tryptophan produced by IDO1 induce T-cell apoptosis, block T-cell activation and trigger differentiation of immunosuppressive T-regulatory cells. 115 (Table 3).
With several Phase II clinical trials already completed using enabling agents, and more ongoing trials capitalizing on the newest iteration in the form of immune checkpoint antibodies, it is clear that combining agents that Engage with those that Enable will continue to be a fruitful clinical strategy moving forward. Moreover, many of the trials described above also include agents to Expand the immune system, either cytokines, costimulatory molecules or both. However, there are concurrent trials that are not hesitating to move past combining one or two IO agents toward true combination immunotherapy.

| INTEG R ATED S TR ATEG IE S FOR IMMUNOTHER APY
Similar to the path followed by chemotherapy in the 20th century, the endgame of immunotherapy will be tactically designed multi- There are seven additional trials in colorectal cancer, pancreatic cancer, triple negative breast cancer, and squamous cell carcinoma.
These trials all utilize different combinations of the same extensive group of standard-of-care and IO agents (Table 4) and are attempting high-number combinations in patients. [76][77][78][79][80][81][82] Of note, patients will be vaccinated with combinations of adenoviral (ETBX, previously utilized in monotherapy trials listed in Table 1)  that combine multiples of these functions together in order to make high-level combinatorial therapy more manageable and improve patient welfare.
In addition to the continued evolution of IO agents, it is clear that immunotherapies need to induce an evolution of the anti-cancer immune response within the patient. It has been clear for decades that immunotherapy is in part effective due to the process of antigen cascade (also known as antigen spreading). This occurs when an Engaged and Enabled immune system results in tumor cell death and tumor antigen uptake and presentation, either mediated by T-cell cytolysis or through radiation/chemotherapy-induced immunogenic cell death. These innate tumor antigens may be more efficacious than those delivered through engaging agents and also allow for a broad, evolving immune response within a patient beyond the response induced by treatment with IO agents. Antigen cascade has been observed in both the preclinical and clinical setting, and further research into the most effective methods of inducing an evolving immune response within individual patients is clearly necessary. 123 The broad applicability of the most commonly used therapeutic viral vaccine platforms and the ability to design them with the appropriate TAAs to target diverse cancers make them a highly attractive therapeutic strategy for engaging the immune system. With decades of mono-and minimal-combination vaccine clinical trials completed and a rapidly evolving slate of IO agents in development, it is clear that the coming decades of immunotherapy will see tactically designed multicombination clinical trials utilizing agents that Engage, Expand and Enable the immune system ( Figure 1). With therapeutic viral vaccines as a foundation, we are optimistic that these evolving clinical strategies will result in effective therapy of established tumors.