The idea of using viruses to treat cancer cells has been around nearly as long as the discovery of viruses themselves. In 1892, Dmitri Ivanovsky [1] described an agent which was “non-bacterial” affecting the tobacco plant, naming this the tobacco mosaic virus. Six years later, “foot and mouth virus” was not only filtered out by scientists but was directly implicated in causing animal disease [2]. Three years later, in 1901, the yellow fever virus became the first virus discovered by scientists causing disease in humans [3]. Finally, in 1931, Max Knoll and Ernst Ruska obtained the first images of viruses with the invention of electron microscopy [4].

During the forty years between Ivanovsky’s discovery and Knoll and Ruska’s images, several case reports were published detailing patients with cancers who magically improved after becoming infected with viral illnesses. In 1897, roughly 5 years after the Ivanovsky’s discovery of the tobacco mosaic virus, George Dock [5] published a case report of a 42-year-old woman with leukemia who became afflicted with influenza. Notably, this was roughly 4 decades before it was known that a virus was responsible for influenza. Dock reported that after his patient developed “tonsillitis and coryza” as well as severe fatigue and weakness, her white blood cell count decreased roughly 70-fold with a concordant reduction in her hepatosplenomegaly as well. She was able to live for roughly one more year after this infection, a remarkable achievement for a patient with acute leukemia at that time. In 1951, there was another case report involving a four year-old boy with profound hepatosplenomegaly and lymphocytic leukemia with a white blood cell count of 300,000, hemoglobin of 5.6 grams per deciliter, and a platelet count of 150,000 [6]. He was treated with intravenous iron for 33 days with no effect and discharged home. About 12 days later, he had a fever of 38.6 degrees Celsius with new, enlarged cervical lymph nodes and vesicles on his face, scalp, and neck consistent with varicella. After seven days, his white blood cell count decreased to 7500 and he no longer had any hepatosplenomegaly. This “remission” lasted roughly 1 month, and he unfortunately died shortly afterwards from “acute lymphatic leukemia”. While the ultimate outcome of these case reports of patients with mostly hematologic malignancies typically ended in death after a few months, they offered a possible window into new cancer therapeutics. This window could be summarized with the following principle: In mostly hematologic malignancies, which tended to have the most rapid cell growth and produce the most depleted immune systems in patients, certain viruses under the right contexts could induce short remissions.

In 1947, Sidney Farber [7] published the results of his study utilizing aminopterin—often considered the first successful chemotherapy agent—to induce remissions in children with leukemia from his basement lab in Boston. Two years later, alongside the infancy of chemotherapy, a clinical trial was presented at the American Association for Cancer Research in which 22 patients with Hodgkin’s lymphoma were “treated” with the hepatitis B virus [8]. In this study, 7 out of 22 patients had a demonstrable clinical response with 4 out of 22 patients experiencing a reduction in tumor size. Unfortunately, 14 out of 22 patients developed hepatitis. Over the next 30 years, there were at least three other clinical trials in which cancer patients were infected with viruses. In 1952, 34 patients with various cancers were infected with Egypt Virus 101 in which only 4 patients demonstrated tumor regression [9]. In 1956, 30 patients with cervical cancer were infected with adenovirus with no tumor responses seen [10]. In 1974, 90 patients with various cancers were treated with the mumps virus in which 37 patients experienced either a complete regression or tumor shrinkage of $>$50% [11].

Owing to both the lack of a positive signal and improved ethical guidelines in clinical trials in the 1970s, there was a pause on treating cancer patients with non-attenuated viruses. Instead, the ability to develop ex vivo culture of human cells allowed researchers to use rodent models to further test oncolytic viral therapies. One study utilized various cancer rodent models and infected them with very large doses of the Russian Far East encephalitis virus producing complete regressions [12]. Unfortunately, in each case, all rodents died of encephalitis shortly afterwards. Clearly, viruses showed great potential although their unpredictability and pathogenicity continued to prevent any meaningful progress until viral engineering became available in the 1990s.

Historically, oncolytic viral therapy (OVT) focused on direct entry and replication of viruses within rapidly dividing tumor cells causing their destruction. However, this paradigm has changed, and more promising potential lies in OVT transforming the tumor microenvironment from an immunosuppressed environment into an immunostimulatory environment in combination with checkpoint inhibitor therapies (see Fig. 1). In normal cells, protein fragments are expressed on major histocompatibility (MHC) class I molecules on the surface of the cell [13]. Cytotoxic T cells (CTL) then recognize either healthy or viral/tumoral antigens via an interaction between T cell receptors (TCRs) and MHC class I molecules. This interaction causes the CTL to either spare the cell (if healthy) or kill the cell (if viral/tumoral). However, tumor cells manipulate this interaction in many ways including by evading recognition by the CTL or suppressing antigen presentation [14]. Either through viral engineering or naturally occurring mechanisms, OVT offers a therapeutic strategy that circumvents this problem.

Fig. 1.

Oncolytic viruses mediate antitumor response synergistically with checkpoint inhibitor activity through multiple mechanisms of action. Mechanism-A: viral replication within tumor cells. Mechanism-B: Induction of tumor cell death via Mechanism-A. Mechanism-C: release of cytokines and initiation of innate immunity.

The first OVT approved by the FDA in 2015 was Talimogene laherparepvec (T-VEC), an intralesional therapy for unresectable melanoma with limited visceral disease [15]. T-VEC is an attenuated herpes simplex virus I (HSV-1) that is genetically modified to specifically infect and lyse cancer cells and preserve normal cells. Normal cells block viral replication via the enzyme protein kinase R (PKR) which phosphorylates eukaryotic initiation factor 2 (eIF2). Activated eIF2 causes the disruption of viral protein synthesis [16]. HSV-1 contains infected cell protein 34.5 (ICP34.5) which dephosphorylates eIF2 and allows viral protein synthesis and subsequent viral replication to succeed [17]. Infected cell protein 47 (ICP47) is another protein produced by herpes viruses that inhibits antigen presentation to CTSs and allow for further viral proliferation [18]. In T-VEC, the genes for both ICP34.5 and ICP47 are deleted and replaced with human granulocyte-macrophage colony stimulating factor (GM-CSF) [19]. In cancer cells, the PKR-eIF2 defense mechanism against viruses is defective [16]. Therefore, T-VEC may safely infect normal cells without replicating within and destroying them and instead target cancer cells causing their destruction. This feat of viral engineering ultimately led to a phase III clinical trial of 436 patients with Stage IIIb–IV unresectable melanoma in which participants were enrolled in a two-to-one ratio to T-VEC and GM-CSF [15]. While T-VEC did not demonstrate an overall survival benefit, patients who were treated with T-VEC had a higher durable response rate compared to CM-CSF, which led to FDA approval.

While T-VEC may be limited as a monotherapy, several trials are evaluating the immunostimulatory action of T-VEC within the tumor microenvironment and combining it with immune checkpoint inhibitors including pembrolizumab and atezolizumab for hepatocellular carcinoma, breast cancer, and other gastrointestinal cancers (see Table 1). As checkpoint inhibitors specifically act to restore normal immune function by blocking immunoinhibitory proteins, combining these therapies with T-VEC may create a powerful synergistic treatment strategy. In 2016, a clinical trial combining T-VEC with the CTLA-4 inhibitor, ipilimumab, in previously untreated and resectable Stage IIIb–IV melanoma revealed a response rate of 50% [20]. This is an improvement from previous response rates with ipilimumab monotherapy (11%) and T-VEC monotherapy (26%) [20]. Other possible combinatorial therapies with OVT include PD-L1 and PD-1 inhibitors [21]. As OVTs like T-VEC specifically attract T-cells to the tumor microenvironment and upregulate tumor expression of PD-L1, for example, combining T-VEC with pembrolizumab offers a mechanism to enhance treatment response of an already effective therapy [22]. One of the limitations of anti-PD-1 therapy is that it is most effective against tumors which express a high amount of PD-L1. If PD-L1 levels were artificially enhanced by OVTs like T-VEC, then deeper responses may be observed [22]. Currently, there are several clinical trials evaluating the efficacy of T-VEC monotherapy or in combination with immune checkpoint inhibitors against melanoma, non-melanoma skin cancer, pancreatic cancer, unresectable gastrointestinal or ovarian cancer, angiosarcoma, sarcoma, rectal cancer, liver cancer, and breast cancer (see Table 1).

While T-VEC is the only FDA-approved OVT, there are several other herpes-based OVTs that are under investigation as monotherapies, in combination with immune checkpoint inhibitors, or in combination with chemotherapies. HF-10 is a naturally occurring HSV-1 OVT which has deletions in UL56 and LAT and increased expression of UL53 [23]. While the ultimate purpose of these natural genetic mutations is unknown, HF-10 has demonstrated antitumor responses in murine models of melanoma [24]. M032 is another herpes-based OVT that has been reengineered with a deletion of the $\gamma$ 34.5 subunit, which causes its neurovirulence, and instead expressed interleukin-12 (IL-12) [25]. Once tumor cell destruction occurs, IL-12 is released into the tumor microenvironment and causing T-cell recruitment and response [25]. There are currently 2 clinical trials evaluating the efficacy of M032 in glioblastoma multiforme. Similarly, G207 is another HSV-1-based OVT which, like T-VEC, has a deletion in ICP34.5 which eliminates its virulence [26]. Uniquely, G207 substitutes infected cell protein 6 (ICP6) with LacZ—this allows it to target tumor cells with p16 tumor suppressor defects [26]. G207 is currently being evaluated in clinical trials for glioblastoma multiforme (see Table 1). One newer HSV2-based OVT, OH2, which also has ICP34.5 removed. ICP34.5, along with infected cell protein 47 (ICP47), allows the virus to present antigens within tumor cells to further promote oncolytic activity [27]. OH2 is being evaluated in several trials against melanoma, and GI cancers—including in combination with checkpoint inhibitor therapies (see Table 1).

Like herpes-based OVTs, adenovirus-based OVTs are modified to include deletions in genes that promote virulence in healthy cells. The first adenovirus-based OVT was H101, which includes E1B and E3 gene partial deletions. H101 was approved by the China Food and Drug Administration Department in 2005 for nasopharyngeal carcinoma [28]. One of the more successful trials of H101 was a study of 165 patients with unresectable hepatocellular carcinoma who were treated with TACE and with or without H101. TACE with H101 showed an overall survival benefit of 12.8 months (TACE with H101) versus 10.49 months (TACE without H101) (p = 0.046) [29]. H101, along with other adenovirus-based OVTs, is currently under investigation for several cancers including intratumoral injections for gynecologic malignancies, hepatocellular carcinoma, and cholangiocarcinoma (see Table 2).

Another adenovirus-based OVT is DNX-2401, which is currently being investigated in clinical trials against recurrent glioma [30]. DNX-2401 specifically deletes E1A, a gene responsible for its virulence, and inserts an RGD-4C motif which allows it to specifically infect tumor cells [31]. Further trials are underway investigating its efficacy against recurrent gliomas in combination with other therapies, including checkpoint inhibitors (see Table 2). In one active phase II trial, DNX-2401 is injected directly into a recurrent glioblastoma or gliosarcoma followed by pembrolizumab every 3 weeks for up to 2 years or until disease progression (NCT02798406).

One last adenovirus-based OVT currently being utilized alongside checkpoint inhibitors is ONCOS-102, which is like H101 and DNX-2401 but is also armed with GM-CSF [32]. The first indication of a positive synergistic effect of ONCOS-102 and checkpoint inhibitor therapy was in a melanoma huNOG mouse model [33]. This paved the way for a phase I trial in which patients with checkpoint inhibitor-refractory, unresectable melanoma were given ONCOS-102 in combination with pembrolizumab [34]. Objective responses were observed in 35% of patients. The FDA has now fast tracked this combination to allow for phase II/III studies for checkpoint inhibitor-refractory, unresectable melanoma. ONCOS-102 is also being utilized in combination with another checkpoint inhibitor, durvalumab, in a current Phase I/II study in patients with peritoneal disease with epithelial ovarian or metastatic colorectal cancer [35] (see Table 2).

Newcastle disease virus (NDV) is often thought of as an ideal candidate for OVT as it does not cause serious illness in humans [36]. MEDI9253, an NDV-based OVT, has emerged as a safe anti-cancer therapy currently being studied with checkpoint inhibitor therapies for several metastatic solid tumors (see Table 3). Like ONCOS-102, MEDI9253 carries a gene for GM-CSF, however, it also encodes an IL-12 gene, which induces the production of IFN-$\gamma$ and attracts natural killer and T-cells to the tumor microenvironment [37]. There is currently a large phase I trial investigating MEDI9253 in combination with durvalumab in patients with advanced/metastatic solid tumors (NCT04613492). JX-594 is a vaccinia-based OVT that also carries a gene for GM-CSF and a deletion for thymidine kinase, which allows it to exclusively replicate in cells with high levels of thymidine kinase, namely, tumor cells [38]. JX-594 in combination with sorafenib had promising phase I/II trial results for hepatocellular cancer, however, the phase III PHOCUS trial was discontinued in 2019 after the study was unlikely to meet its primary objective. Other trials are underway including a phase I/II study combining JX-594 with the checkpoint inhibitor, cemiplimab, in metastatic renal cell carcinoma (NCT03294083) (see Table 3).

There is an abundance of trials utilizing OVT, with most of the promising studies utilizing a synergistic effect with checkpoint inhibitors against solid tumors. One common barrier is the limitation of drug delivery and the challenges of evading immune responses to OVT [39]. While there are trials that are trying to utilize novel viral engineering to deliver OVT intravenously, almost all have poor bioavailability and must be injected intratumorally or locoregionally. This presents a challenge for metastatic cancers with a large tumor burden. One strategy that is being explored involves coating OVTs with biocompatible polymers to block antibodies from binding [40]. Another strategy is to pre-treat patients with a complement-inhibiting peptide [40]. Moreover, due to the intricate engineering process involved with OVT, there is a high cost associated with production and utilization of OVT that may not be practical in real world clinical settings. The estimated cost of T-VEC plus ipilimumab for advanced unresectable melanoma, the most promising OVT thus far, was $\$$494,983 per patient compared to $\$$132,950 for ipilimumab monotherapy [41]. If certain OVTs can show a significant PFS or OS benefit in phase III trials, it will be imperative to create sustainable production mechanisms that allow for improved costs, especially when combined with checkpoint inhibitors.

Perhaps the most promising trend with OVT lies in the combination therapy with checkpoint inhibitors, specifically in “cold tumors” with low expressions of PD-L1 on the surface. OVTs promote immunologic responses including type I interferons which increase PD-L1 expression on cancer cells [42]. This would allow for tumors with previously low levels of PD-L1 expression which would have been poor targets for checkpoint inhibition to potentially be ideal targets for checkpoint inhibition.

JR — drafting of the manuscript; SH — drafting of the manuscript, creation of figure; AS — manuscript revision; All — final approval of the manuscript submitted.

Not applicable.

Not applicable.

This research received no external funding.

AS reports research grants (to institution) from AstraZeneca, Bristol Myers Squibb, Merck, Exelixis, Clovis, KAHR medical, Actuate therapeutics, Incyte corporation, Daiichi sankyo, and advisory board fees from AstraZeneca, Daiichi Sankyo, Bristol Myers Squibb, Pfizer and Exelixis. The remaining authors report no relevant COIs.