Oncolytic viruses: challenges and considerations in an evolving clinical landscape

Oncolytic viruses

Oncolytic viruses (OVs) have gained increasing attention in recent years due to their ability to selectively attack and lyse cancer cells, as well as their potential to enhance anti-tumor immunity. Interest in OVs reflects the potential benefits they offer over standard treatments for cancer, such as radiotherapy, chemotherapy, immune checkpoint inhibitors (ICIs) and targeted therapy [1,2]. These treatment strategies are not effective against certain tumor types and have a number of potential limitations, including drug resistance, recurrence of cancer and severe adverse effects [2,3]. Several clinical trials have demonstrated the ability of OVs to exploit multiple lytic mechanisms to kill cancer cells that are resistant to conventional and targeted therapies, suggesting that OVs could help to overcome some of these barriers.

Advances in genetic engineering have played a key role in allowing the development of tumor-selective OVs that have a manageable safety profile and are efficacious against a range of cancers [4]. Through rational design and genetic engineering, OVs may be tailored to an individual's tumor type and its driver mutations [5]. Viral genes or noncoding sequences can be modified to add or remove certain functions and non-viral genes or non-coding regulatory elements can be added to provide additional desirable properties [5]. Moreover, predictive tests may be developed to identify patients with a given molecular profile that makes them more likely to benefit from OV treatment; this offers the potential for a personalized medicine approach [2]. For example, constitutive IFN pathway activation was identified as a key determinant for measles virus (MV) replication where viral replication in resected human glioblastoma tumors was inversely correlated with expression of this resistance gene signature. These findings represent the development of a predictive algorithm to preselect patients for oncolytic treatment [6]. Since the oncoselectivity of vesicular stomatitis virus (VSV) is also determined by the type I IFN-associated antiviral potential of a cell, it is suggested that IFN pathway abnormalities could build up the basis for a predictive test in identifying suitable patients [7]. Another predictive test could be for the presence of RAS mutations for treatment with reovirus [8]. An important recent tactic is to combine OVs with existing cancer therapies that may act in synergy [2,5,9–11]. By combining therapeutic agents with different mechanisms of action, efficacy may be enhanced and resistance to current treatments overcome [4,11]. Therapeutic outcomes with OVs in combination with other cancer treatments are considered to have yielded the most successful results so far [4].

Multiple mechanisms of action of OVs

OVs act through multiple mechanisms to mediate anti-tumor activity (Figure 1) [12]. Firstly, they exhibit tumor selectivity and preferentially replicate within cancer cells (Figure 1A). There are various ways in which OVs can be targeted to primarily infect cancer cells in order to ensure minimal damage to healthy cells. Such strategies usually involve the exploitation of cell surface receptors and intracellular aberrations in pathways and gene expressions which are up-regulated in cancer cells, but not in healthy cells [13,14]. Some viruses, such as reovirus, have a natural preference for cancer cells whereas other viruses including adenovirus, VSV and herpes simplex virus (HSV) can be adapted to make them highly cancer specific. Although healthy cells can be infected by an OV, they are able to inhibit viral propagation via various responses which are often deficient in cancer cells [15]. For example, the oncoselectivity of VSV is determined by the type I IFN-associated antiviral potential of a cell; non-malignant cells are able to produce, secrete and respond to type I IFNs to impede viral replication whereas most cancer cells have defective type I IFN signaling and are therefore susceptible to VSV infection [7,16]. Virus replication within cancer cells then causes direct lysis of these tumor cells, termed virus-induced direct oncolysis (Figure 1B) [4].

The expression of transgenes, such as those that trigger immune host responses, activates anti-tumor immunity and recruits activated immune cells into the tumor microenvironment (Figure 1C) [4,12]. This initially triggers a local anti-tumor immune response, which contributes to local death of tumor cells. Second, the release of tumor antigens leads to a systemic anti-tumor immune response throughout the body (Figure 1D) [12]. Local OV-induced inflammation also causes destruction within the tumor microenvironment (Figure 1E). Therefore, OVs are able to indirectly mediate cell death of both infected and uninfected cancer cells and associated endothelial cells residing in the tumor vasculature to reduce tumor angiogenesis and prevent any further tumor growth (Figure 1F) [4]. Replication and amplification of OVs are major determinants of tumor eradication and lead to the infection of new tumor cells (Figure 1G) [5]. This ability of an OV to replicate and amplify is also what sets OVs apart from other non-biological cancer therapies [5]. Of note, the general mechanisms described here vary widely among OVs and are dependent upon the type of cancer cells and overall interactions between the OV, the tumor microenvironment, its state of inflammation and the host immune system [4]. These mechanisms have been described in detail in other reviews [4].

Challenges & considerations in OV development & translation to clinical practice

To date there have only been four OVs approved for use as cancer treatments, although there are many that are in the pipeline. Many OVs have discontinued development either due to lack of efficacy or unacceptable toxicity profile. There are a number of challenges in developing OVs, which may contribute to the difficulty in reaching approval. Within this chapter, we discuss the challenges and considerations of developing an OV, including identifying the most appropriate viral backbone, genetic engineering to enhance efficacy or cooperation with a combination drug and optimizing the route of administration, and how certain attributes may be advantageous [4,17].

Selection of an OV

The first consideration when selecting an appropriate OV is the viral backbone. According to the type of nucleic acid they contain, OVs can be divided into DNA and RNA viruses [18]. Several clinical trials are currently in progress for both DNA (153 trials) and RNA viruses (70 trials; Figure 2) [19]. DNA viruses have the advantage of a larger and more stable genome, making genetic engineering and the addition of multiple transgenes more straightforward [17]. However, DNA viruses have a relatively poor immunogenicity compared with some RNA viruses [4,20]. RNA viruses have attracted more attention to their oncolytic virotherapy potential due to their smaller size, more steadfast replication and potential to be more immunogenic [4,17]. DNA viruses, such as adenovirus and HSV type-1 (HSV-1), are the most frequently used platforms, followed by the RNA-type reovirus [21]. The measles vaccine virus (also an RNA virus) has recently shown promising efficacy data in early clinical trials [22]. Other viral backbones of interest include parvovirus, vaccinia virus, Newcastle disease virus, Maraba virus, VSV, poliovirus and coxsackievirus. The key properties of these viral backbones are summarized in Table 1.

Table 1. Properties of key oncolytic viruses.

	DNA				RNA
Virus	Adenovirus	Herpes simplex virus	Parvovirus H1	Vaccinia virus	Measles vaccine virus	Newcastle disease virus	Maraba virus	Reovirus	Vesicular stomatitis virus	Poliovirus	Coxsackie virus
Genome size and diameter
Capsid symmetry	Icosahedral	Icosahedral	Icosahedral	Complex	Icosahedral	Helical	Helical	Icosahedral	Helical	Icosahedral	Icosahedral
Envelope	Naked	Enveloped	Naked	Complex coats	Enveloped	Enveloped	Enveloped	Naked	Enveloped	Naked	Naked
Site of replication	Nucleus and cytoplasm	Nucleus and cytoplasm	Nucleus and cytoplasm	Cytoplasm	Cytoplasm	Cytoplasm	Cytoplasm	Cytoplasm	Cytoplasm	Cytoplasm	Cytoplasm
Transgene capacity	✓	✓	N/A	✓	✓	✓	✓	N/A	✓	✓	✓
Entry receptor	CAR	HVEM, nectin-1, nectin-2	Sialic acid residues	No specific receptor	SLAM and CD46	Sialic acid-containing proteins	LDLR	No specific receptor	LDLR	PVR or CD155	CAR and CD55
Most-treated cancer type	Brain cancer	Skin cancer	Brain cancer and pancreatic cancer	Liver cancer and solid tumors	Breast cancer and fallopian tube	Solid tumors	NSCLC	Myeloma	Solid tumors	Melanoma, glioblastoma breast cancer	Bladder cancer, lung cancer and melanoma
Ref.	[1,14,23]	[1,14,24,25,26]	[1,2]	[1,14,27]	[1,14,22]	[1,28,34]	[29,35,36]	[1,30,37]	[1,31]	[1,32]	[1,33]

N/A: Not applicable; NSCLC: Non-small-cell lung cancer.

Genetic engineering

Around two thirds of published OV clinical trials have used genetically engineered OVs rather than native viruses [21]. The most frequent modification is the deletion of non-essential viral genes to promote selective replication in tumor cells and attenuate viral pathogenicity [21]. Given increasing evidence that OVs can play an immunomodulatory role, a number of preclinical and early phase clinical trials are investigating OVs that have been engineered to induce a desired immune response [17]. The most common immunomodulating transgene addition is GM-CSF, which helps to stimulate host immune responses [21]. Many other transgene additions are also targeted toward enhancing the immune system, such as the addition of ICAM-1 [21].

An alternative reason to genetically engineer an OV is to reduce the innate toxicity of the virus. For example, VSV can cause neurotoxic side effects [38]. Therefore, attempts have been made to reduce its innate neurotropism by modifying the glycoproteins of its envelope. One strategy to develop a neuroattenuated virus was via miRNA-targeting or transductional targeting. However, this was found to be at the expense of efficacy [39]. An alternative approach involves replacing the innate glycoprotein G, which has been linked to the virus's neurotropism, with the glycoprotein GP from the lymphocytic choriomeningitis virus [39]. The resulting OV is termed VSV-GP (BI 1831169) and preferentially replicates in cancer cells (like wild-type VSV) leading to direct cell lysis, while virus propagation in normal tissues will be suppressed by the anti-viral IFN response. Preclinical studies in mice have demonstrated that VSV-GP escapes humoral immunity and demonstrates therapeutic efficacy [39,40].

Similarly, HSV has been genetically modified to reduce potential toxicity. HSV is typically associated with cold sores but can also trigger encephalitis. The neurotropic genes that contribute to this in the HSV genome are UL56 and γ34.5. Therefore, the G207 modified HSV1 OV was developed with a γ34.5 deletion [41].

OVs may also be genetically modified to increase their tumor-specificity. For example, the coxsackievirus entry receptor is not typically found on tumor cells. However, modification of its fiber protein with an RGD motif, which acts as a ligand for αβ integrins that are often upregulated on tumor cells, leads to increased viral uptake, for example, by breast cancer cells [42].

Furthermore, the anti-tumor activity of OVs can be enhanced through genetic modification, such as via inhibiting angiogenesis or altering tumor cell signaling. For example, RAMBO is a modified HSV that induces the expression of the antiangiogenic protein vasculostatin by tumor cells and leads to in vitro inhibition of glioma migration and invasion [43].

Resistance to OV therapy

A key challenge in OV therapy is the antiviral immune response of patients [44,45]. The host immune response has been shown to be detrimental in a number of preclinical OV studies, with reduced replication and earlier clearance of the OV as well as decreased anti-tumor efficacy reported in immunocompetent versus immunocompromised subjects [44,46]. The development of OV treatment strategies has therefore focused on how to manipulate the host immune system so that antiviral responses and viral clearance can be minimized while immune responses that encourage tumor destruction are promoted [44]. This issue is well documented for adenoviruses, especially those based on the adenovirus serotype 5 backbones, for which there are high rates of pre-existing immunity. This varies geographically, reportedly from around 30% in the UK population to over 90% in sub-Saharan Africa. When designing adenoviral vectors, modifications are important in order to reduce immunogenicity and bypass innate antiviral immune responses [12]. One strategy to avoid rapid viral clearance is to transiently suppress early immune responses [44]. As one example, this was demonstrated preclinically with a recombinant VSV that was engineered to express a broad-spectrum chemokine-binding protein (herpesvirus-1 glycoprotein G). The potency of the OV was enhanced through suppression of the host antiviral inflammatory response, which substantially prolonged the survival of rats with multifocal hepatocellular carcinoma [47].

OVs as combination therapy

The use of OVs in combination with other cancer therapies can improve therapeutic outcomes. There are currently 116 clinical trials that are planned, ongoing or completed which investigate this combination approach [19]. An advantage of this method is that OVs can be specifically designed to complement existing treatments, for example, to make immune-resistant tumors responsive to immunotherapy. The lytic activity of OVs leads to the release of immunomodulatory molecules which convert an immunologically ‘cold’ tumor where there is a lack of infiltrating T cells to a ‘hot’ tumor which is more susceptible to immunotherapy [17,48].

A similar approach can be used to overcome chemoresistance by genetically targeting the underlying mechanism. For example, gliomas expressing the DNA repair enzyme MGMT are resistant to the chemotherapy agent temozolomide. Δ-24-RGD oncolytic adenovirus was designed, which efficiently downregulates MGMT and the combinational approach with temozolomide induced a profound therapeutic synergy in glioma cells [49]. OVs can also be engineered to increase tumor cell sensitivity to radiation therapy by encoding the gene of the sodium-iodide symporter which is then expressed on the membrane of infected cells resulting in selective uptake of radiotherapy agents (such as iodine-131) [50].

In particular, this combinational approach has shown promise for OVs in combination with ICIs and T cell-based therapies [4]. ICIs that target immune checkpoints have become one of the key modes of treatment for cancer. However, they have several limitations, including the low number of patients that respond to treatment [4]. In combination with OVs, ICIs have shown promising results in preclinical and clinical studies [51,52]. For example, a clinical study in patients with advanced melanoma showed frequent and durable responses when the clinically licensed OV, talimogene laherparepvec (T-VEC), was combined with the ICI, ipilimumab [11]. In addition, a recent phase Ib trial demonstrated that combinatorial treatment with T-VEC and the ICI pembrolizumab led to an increased influx of T cells into tumor tissues, as well as signs of enhanced activities of both CD4⁺ and CD8⁺ T cells [53].

Cellular immunotherapy, such as T-cell therapy, is also associated with limited success [4]. This is especially true for treating solid tumors, which are difficult for T cells to infiltrate and in which they cannot survive for long. The use of OVs in combination with T-cell therapy has shown potential in enhancing the benefit of both treatments. For example, recent in vitro and in vivo data showed that an interleukin-7-loaded oncolytic adenovirus combined with B7H3-targeted CAR-T cell therapy had greater anti-tumor efficacy for glioblastoma than monotherapy with either treatment. The authors suggest that the OV improved the therapeutic efficacy of CAR-T cell therapy by providing activation signals for the T cells to infiltrate the tumor [54]. Using another strategy, several preclinical studies have provided proof-of-concept for using OV-encoded bispecific TCEs to overcome some of the challenges associated with TCE monotherapy [55]. TCEs are efficacious against certain hematologic malignancies. However, their application in clinical practice, especially against solid tumors, has thus far been restricted by their limited bioavailability and toxicity. These limitations can be overcome by encoding TCEs in OVs; OV infection at the site of a tumor induces local inflammation and attracts T cells, which can be redirected to tumor cells by TCEs. Furthermore, this mode of delivery can increase local concentrations of TCEs at the tumor site, supporting infiltration into solid tumors while reducing systemic exposure [55].

OVs have also been designed to complement traditional cancer therapies. For example, several OVs have been developed to include prodrug enzyme genes, such as yeast cytosine deaminase and the HSV-1 thymidine kinase. The result is enhanced tumor cell death when patients are treated with already licensed non-toxic prodrugs [21,56,57]. The engineering of other OVs included the addition of a transgene for HSP70, human telomerase reverse transcriptase promoter and sodium iodide symporter, which allows for radiolocalization and sensitizes cells to radiation therapy [21].

Senescence, a state of permanent cell cycle arrest, has attracted increasing attention surrounding its role in the treatment of cancer and has been considered a tumor suppressor mechanism [10,58]. Senescence can also be triggered by various anticancer treatments, including chemotherapy – a term described as therapy-induced senescence. While these senescent cells undergo permanent cell cycle arrest, they remain metabolically active and so there are concerns for tumor relapse and tumor progression. Therefore, development of treatments that target senescence should consider both sides and should ideally be combined with a second therapeutic agent that is able to eradicate senescent tumor cells, for example, an OV [10,58,59]. Preclinical studies have investigated senescence-inducing chemotherapy agents such as gemcitabine in combination with the measles vaccine virus in human pancreatic cancer cell lines. The results showed a mutual improvement in the effectiveness of each therapeutic agent where the combination of MeV and gemcitabine resulted in a tumor cell mass reduction of >50% and enhanced viral replication [10,59].

Targeting tumor cells: tumor types & routes of OV administration

OVs are so called for their ability to replicate in and to burst tumor cells. OVs do this by exploiting the modification of certain signaling pathways that tumor cells use for survival and proliferation. As a result, tumor cells within which these pathways are mostly intact would be resistant to OVs. For example, VSV will only replicate in cells with defective IFN signaling, a trait common in tumor cells. However, tumor cells with restored IFN signaling are resistant to VSV replication and oncolysis [60]. Therefore, when selecting an OV for development, it is important to consider whether the essential pathways for viral replication are active or inactive in the target tumor type.

The most common tumor types investigated for treatment with OVs to date are melanoma, gastrointestinal and other solid tumors. It is hypothesized that melanoma is the most common tumor type under investigation due to the accessibility of tumors for the local injection of OVs [21]. The most common indications under investigation in OV trials include melanoma, liver, colorectal, non-small-cell lung cancer and glioblastoma (Figure 2D). These are among the most common cancer types according to the National Cancer Institute and glioblastoma is notoriously treatment refractory, which could explain the high interest for oncolytic virotherapy [61,62]. Hematologic malignancies were among the least common indications under OV treatment investigation, perhaps as only one route of administration (intravenous [i.v.]) can be utilized. The most common route of administration is intratumoral (it.) injection and i.v. infusion [21]. Other less common routes of administration include hepatic artery infusion and intraperitoneal/intrapleural delivery [21].

It. injection has the advantage of delivering virus particles directly to the tumor site. However, a major disadvantage is that not all tumors are amenable to this, given that they may consist of multiple small nodules spread over a large area or may reside in a location inaccessible by direct it. injection [63]. This is one of the limitations of the approved OV, T-VEC, which is only licensed for it. applications [12]. Currently, T-VEC is administered via it. injection into cutaneous, subcutaneous and/or nodal melanoma lesions that are visible, palpable or detectable by ultrasound guidance [64]. However, the safety and efficacy of other methods for it. T-VEC administration are being assessed in non-cutaneous tumors such as liver metastases from solid tumors or primary hepatocellular carcinoma [64].

In contrast to it. injections, the key advantage of i.v. delivery is that OVs may be disseminated throughout the blood system. This offers the potential for effective treatment of metastases, which are responsible for the majority of deaths among cancer patients [12]. It is generally considered the preferred route of administration. However, a challenge for i.v. delivery is how effectively the immune system can remove foreign pathogens from the circulatory system. Therefore, suitable viral vectors are needed that exhibit stealthing properties. In this context, several studies have investigated the use of carrier cells for OV delivery. They aim to protect the OV from neutralization and exploit cellular features of tumor tropism; thus, in given cancer patients the OV can be primarily directed toward tumor sites [63].

A personalized-medicine approach for OV therapy

Cancers can be extremely heterogeneous, making a personalized approach to finding the best treatment strategy appealing. Recently, the idea of using human organoids to individually screen drugs for a given patient has been proposed [65]. Organoids are 3D multicellular in vitro tissue constructs that mimic the corresponding in vivo organ and have been used to study aspects of that organ in the tissue culture dish [65,66]. The use of human organoid cultures is a relatively new area, with a number of examples published in the field of oncolytic virotherapy [67,68]. Zhu and colleagues demonstrated that Zika virus selectively replicates in, and kills, patient-derived glioblastoma stem cells, while not infecting differentiated glioblastoma cells or normal neuronal cells [69]. An interesting application of patient-derived tumor organoids in oncolytic virotherapy is to create individual patient ‘virograms’ [65]. This proposal involves using tumor-tissue specimens from biopsies or tumor resections to culture tumor-derived organoids, providing a realistic personal tumor model that holds individual histological and genetic patterns. Using this model, different OVs (as well as combinations of compounds) could be tested for an individual patient ‘virogram’. The agent with the best oncolytic efficacy could then be picked for the patient's treatment [65].

Safety profile of OVs

Overall, preliminary data from randomized controlled trials indicate that OVs are generally tolerable [21,70]. In a systematic review of 119 OV publications (97 studies) from 2000 to 2020, the vast majority of common treatment-related adverse events (AEs) were reported to be low grade (grade 1–2) constitutional symptoms and local injection site reactions [21]. The most common treatment-related AEs reported were fever, chills and rigors, nausea and vomiting, flu-like symptoms, fatigue and pain, all of which are relatively transient in nature. Across the 97 studies assessed, there were 98 grade 3 and 21 grade 4 treatment-related AEs. Neutropenia (n = 11), leukopenia (n = 9), thrombocytopenia (n = 9), fatigue (n = 8) and fever (n = 8) were the most frequently reported (Table 2). However, many of these events were attributed to disease progression or to other drugs used in combination clinical trials. In another recent systematic review and meta-analysis, which pooled data from 11 studies and a total of 1452 patients, OVs were also considered to be generally safe [70]. The most common AEs of any grade with an incidence greater than 20% were fever, neutropenia, febrile neutropenia, leukopenia, diarrhea, nausea, vomiting, chills, fatigue, flu-like symptoms, decreased appetite/anorexia, extreme pain, headache, cough and thrombocytopenia [70]. The most common severe AEs with an incidence greater than 5% were neutropenia, febrile neutropenia, leukopenia, fatigue and thrombocytopenia [70]. In general, dose-escalation studies have not reported patients experiencing dose-limiting toxicities (DLTs). However, in the phase Ib/II MASTERKEY-232 study, one patient experienced a DLT, a case of fatal arterial hemorrhage, that was considered possibly related to T-VEC and led to a protocol amendment [71]. Hyperbilirubinemia was also identified as a DLT in a phase I study of recombinant vaccinia virus treatment in patients with metastatic liver cancer [72].

In case of an uncontrolled replication of OVs, licensed virostatic compounds can be used to bring down production of progeny OV particles. For example, HSV-1 thymidine kinase can convert the non-toxic prodrug ganciclovir into ganciclovir-triphosphate, which halts host cancer cell and virus DNA replication, potentially protecting patients from uncontrolled replication of T-VEC (IMLYGIC^®) [57]. Furthermore, while attenuation of VSV has significantly improved safety, anti-VSV drugs may further decrease the risk of the potential side effects caused by uncontrolled VSV replication after cancer treatment. Vesiculopolins, a new class of anti-VSV compounds, have been investigated and found to inhibit VSV-induced cytopathic effects and in vitro mRNA synthesis, demonstrating that these compounds may be useful for developing antiviral drugs against VSV [73]. Similarly, the antiviral drug cidofovir has been shown to be an effective approach to controlling a potent oncolytic adenovirus infection and may serve as a potential ‘safety valve’ for this virotherapy [74].

Current landscape of OV development

To date, only four OVs have been approved for use in patients with cancer. However, there are over 200 OVs currently in clinical trials, reflecting the considerable interest in the field at present and the promising potential of these therapies [19].

Approved OVs

The first OV to be approved was ECHO-7 (RIGVIR^®), an unmodified picornavirus that was approved in Latvia in 2004 to treat melanoma [75]. Since the approval of the first OV, the development of subsequent OVs has exploited genetic engineering to knock out viral genes and/or knock in transgenes [4]. In 2005, the engineered adenovirus H101 was approved in China for the treatment of head and neck cancer [4,76]. However, neither ECHO-7 nor H101 have been approved by the US FDA or the EMA. In 2015, T-VEC (IMLYGIC^®), a genetically engineered HSV-1, was the first OV to be approved in the USA for the treatment of advanced unresectable melanoma [77]. T-VEC was subsequently approved in Europe, Australia, Switzerland and Israel [64]. The most recent OV to be approved was a modified HSV-1, named teserpaturev or G47Δ (DELYTACT^®), which received a conditional and time-limited approval in June 2021 in Japan for malignant gliomas [4,78,79].

T-VEC

T-VEC, which is built on the HSV-1 strain JS1, has a number of genetic attributes that enable the virus to selectively replicate in tumor cells and ensure a tolerable safety profile [64,77]. First, the HSV-1 JS1 strain is, by nature, preferentially selective for tumor cells versus normal cells. Secondly, deletion of both copies of the ICP34.5 gene in T-VEC reduces the natural neurovirulence of the HSV JS1 strain and its ability to reactivate; this further enhances the preferential tumor-killing properties while also reducing the infection of healthy tissues. Third, deletion of the ICP47 gene in T-VEC allows for enhanced antigen presentation and for upregulation and earlier and increased expression of the US11 gene. The effect of this is increased replication of the genetically engineered virus in tumor cells. In addition, insertion of a human GM-CSF cassette has the effect of increasing activation of antigen-presenting cells, triggering a systemic anti-tumor immune response. In addition to melanoma, T-VEC is currently being evaluated in clinical trials of other solid tumor types [64].

T-VEC was approved based on the results of a multicenter phase III clinical trial of patients with metastatic melanoma lesions in the skin and lymph nodes. T-VEC was shown to significantly improve the rate of responses lasting continuously for at least 6 months in patients compared with GM-CSF and responses were expected to last ≥1 year [80]. The final analysis of the trial in 2019 demonstrated that T-VEC continued to result in improved longer-term efficacy compared with GM-CSF and remained well tolerated [81]. Although T-VEC has only been approved in a small number of countries, there are several studies that have assessed the safety and efficacy of T-VEC in the real-world setting. Across the studies, the percentage of patients who achieved a complete response ranged from 20 to 61.5%. Consistent with clinical trial data, T-VEC was well tolerated in these studies, with the most frequently reported AEs being influenza-like symptoms [82–88].

G47Δ

G47Δ is a triple-mutated, conditionally replicating HSV-1, which was developed from the genome of G207 (a second-generation oncolytic HSV-1) to enhance oncolytic activity and retain immunogenicity of infected cancer cells [89]. As mentioned earlier, G207 has a deletion of γ34.5, which reduces its neurotropism. It also contains an insertion of the E. coli lacZ gene in the ICP6 coding region UL39, which results in the inactivation of ribonucleotide reductase – a key enzyme for viral DNA synthesis in nondividing but not dividing cells [90]. This double mutation minimizes the chance of G207 reverting to the wild-type strain and confers properties to the virus that are favorable for treating human brain tumors [90]. G47Δ was further modified from G207 by deletion of the α47 gene, which improved efficacy [89].

OVs in clinical trials

An analysis of clinical trials documented on clinicaltrials.gov and/or EudraCT found that there are currently over 200 OVs in clinical trials (Figure 2) [19,91]. The majority of these are at phase I (n = 124), followed by phase I/II (n = 47) and phase II (n = 48). Only five phase III trials are listed (Figure 2A). There are 62 trials listed for HSV, 57 for adenovirus, 32 for vaccinia virus, 26 for reovirus, 14 for coxsackie, 11 for MV, nine for VSV, four for poliovirus, three for Maraba virus, two for Seneca Valley virus, two for a Newcastle disease virus, two for a parvovirus and one for an alphavirus (Figure 2C). An analysis of OV trials (phase I/II, II and III) by indication showed that the most common indication being trialed was skin cancer (melanoma), with 22 trials documented. This was followed by liver cancer (n = 11), non-small-cell lung cancer (n = 11), colorectal cancer (n = 11) and then brain cancer (glioblastoma; n = 8; Figure 2D). In terms of the geographical distribution of OV clinical trials, over half of documented locations were in the USA (n = 132), followed by China (n = 20), Spain (n = 18), the UK (n = 13) and Canada (n = 13; Figure 2E). Additional details on the clinical landscape of OVs have been covered extensively in another review [21].

Several OV compounds have shown promising results in phase I/II trials and are anticipated to progress into phase III trials. An adenovirus, DNX-2401, in combination with the anti-PD-1 antibody, pembrolizumab, showed encouraging activity and safety in patients with recurrent glioblastoma; a global phase III randomized controlled trial is planned [92]. A second adenovirus, CG0070, showed promising efficacy and acceptable toxicity in phase II trials in patients with non-muscle invasive bladder cancer who had previously failed Bacillus Calmette-Guérin therapy [93]; the OV had progressed into a phase III trial but this trial is currently terminated, citing ‘change in study design’ as the reasoning (NCT01438112) [94]. In a phase II trial assessing the efficacy of the oncolytic vaccinia virus GL-ONC1 for ovarian cancer (NCT02759588), the majority of patients (54%) achieved a response, despite having platinum-resistant/refractory cancer, prior bevacizumab and disease progression on their last therapy [95]. Administration was through intraperitoneal infusion at multiple doses [96].

OVs in early development & clinical studies

There are a number of OVs of interest in early clinical trials, including several RNA OVs (Table 3). VSV is a negative-sense RNA virus with a number of promising attributes for OV therapy [97,98]. These include low pathogenicity, a lack of pre-existing human immunity, human infection being usually asymptomatic, rapid replication kinetics, viral replication being cytoplasmic (meaning that there is no integration into the genome) and the genome being relatively easy to modify [98,99]. In addition, VSV can elicit a wide range of immunomodulatory responses that make the tumor microenvironment more susceptible to immune attack [99]. VSV also displays inherent tumor specificity based on its high sensitivity to type I IFNs, meaning that replication is limited to cells that are defective in their antiviral IFN signaling pathways, a property that is common among many types of cancer [39,99]. Analysis of the NCI60 panel of tumor cells has shown that over 81% of all tumor cell lines have defects in the IFN pathway [100].

Four phase I/II studies are currently ongoing for the VSV derived-OV VSV-IFNβ-NIS (Voyager V1; VV1) [19]. VV1 has been genetically engineered to enhance its efficacy. VV1 encodes human IFN-β, which increases anti-tumoral immune response and tumor specificity by triggering the mechanism of IFN-dependent inhibition of viral replication in healthy cells [97,98]. In addition, a thyroidal NIS allows for imaging of the virus [97]. In the first two parts of a three-part study (NCT02923466), VV1 exhibited an acceptable safety profile [101,102].

This offers a promising platform for further investigation in clinical trials. A phase I trial, currently recruiting patients, aims to explore the safety and early efficacy of VSV-GP alone and in combination with the checkpoint inhibitor ezabenlimab in patients with advanced, metastatic and/or relapsed/refractory solid tumors (NCT05155332) [103]. Another phase Ib trial, currently recruiting patients, aims to explore the use of VSV-GP128 in combination with the recombinant protein vaccine, ATP128, and an anti-PD1 antibody (BI 754091) in patients with stage IV colorectal cancer (NCT04046445) [104]. In addition to the modifications described for VSV-GP, VSV-GP128 has been engineered to carry a domain comprised of cancer-specific antigens, which help to induce an immune response against colorectal tumor cells [105].

Although most cancer types have defective IFN response and should therefore be sensitive to VSV-based OVs, there are some that have an intact IFN response. To sensitize this population of tumors, it may be an option to combine the VSV OV with something to modulate the IFN pathway, such as JAK inhibitors [100].

Another OV of interest is the MG1 strain of the rhabdovirus Maraba, which was selected from an intensive OV screening process and subsequently engineered to enhance tumor selectivity and potency [106]. MG1 targets tumors through multiple complementary mechanisms of action, including direct lysis of cancer cells, modulation of the tumor microenvironment and generation of tumor antigen-specific T cells [106]. Similarly to VSV, MG1 is IFN responsive and selectively targets cancer cells with a defective IFN response [106]. In addition, MG1 creates an inflammatory state that turns immunologically ‘cold’ tumors into ‘hot’ tumors that are more susceptible to immune-mediated attack. MG1 has undergone extensive preclinical evaluation and advanced into early clinical trials. Two phase I/II trials have focused on the safety and maximum feasible dose of MG1 expressing a conserved MAGE-A3 tumor [98]. The first of these studies is investigating MG1 alone or in combination with the adenovirus/MAGE-A3 (AdMA3) tumor vaccine for the treatment of MAGE-A3-positive tumors (NCT02285816). Preliminary analyses show that AdMA3 prime followed by i.v.-administered MG1 can induce a potent anti-tumor immune response [107]. The second study is investigating MG1 in combination with AdMA3 and pembrolizumab for treatment of non-small-cell lung cancer (NCT02879760) [98]. An additional phase I trial is assessing MG1 in patients with human papillomavirus-associated cancers. The study is investigating safety and maximum tolerated dose of MG1 carrying two human papillomavirus oncogenes, E6 and E7, which is hence denoted MG1-E6E7. MG1-E6E7 is being evaluated in combination with the Ad-E6E7 adenoviral vaccine and atezolizumab (NCT03618953) [99]. At the time of writing, no data are available for the MG1 trial results.

Several trials are currently ongoing for a genetically engineered measles OV, MV-NIS, all of which are being conducted in the USA (Table 3). MV-NIS is a live attenuated measles virus that has been engineered to express the human thyroidal NIS [108]. Encoding NIS provides a means for tracking the spread of the OV using noninvasive imaging techniques, such as radioiodine imaging. Another key advantage of MV-NIS is its ability to selectively target and destroy tumor cells. Whereas the wild-type strain uses CD150/signaling lymphocytic activation molecule on lymphoid cells and epithelial nectin-4 as receptors, MV-NIS primarily infects cells via CD46, a membrane regulator that is known to be overexpressed in many human cancers [109,110]. The use of MV as an oncolytic immunotherapy is described in detail in another review [22]. Phase I trial results of MV-NIS have shown promising efficacy and safety data. In a phase I trial of 16 patients with recurrent ovarian cancer who received treatment with MV-NIS administered intraperitoneally, a good safety profile was reported, as well as early evidence of anti-tumor activity (NCT00408590) [111]. A phase I trial in patients with relapsed or refractory myeloma also showed a good safety profile when MV-NIS was administered intravenously (NCT00450814) [108]. Even at the highest dose, the most significant AEs were transient in nature. A further trial in mesothelioma is listed as completed (NCT01503177). Preliminary data from 12 patients who received MV-NIS via intrapleural administration reported that there were no dose-limiting AEs and therapy was well tolerated. Stable disease was observed in 67% of patients, with the authors suggesting that MV-NIS may be associated with favorable overall survival in malignant pleural mesothelioma [112]. In addition, numerous clinical trials are ongoing (recruiting or active but not yet recruiting) across a range of cancer types, including bladder cancer (NCT03171493) [113], breast cancer (NCT04521764), ovarian cancer (NCT02068794), malignant peripheral nerve sheath tumors (NCT02700230), medulloblastoma (NCT02962167) and head and neck squamous cell carcinoma (NCT01846091) [114,115].

There are several trials investigating picornaviruses as potential OVs (Table 3). One of which is PVSRIPO, a live attenuated poliovirus type 1 vaccine with its cognate internal ribosome entry site replaced with that of human rhinovirus type 2 (HRV2) [116]. The HRV2 internal ribosome entry site mediates PVSRIPOs neural incompetence whereby it cannot translate and propagate in human neuronal cells [62,116]. PVSRIPO enters cells via the poliovirus receptor, CD155, which is ectopically expressed in virtually all solid tumors [62,116]. PVSRIPO has shown promising results in a phase I trial (NCT01491893) where it. infusion of PVSRIPO in patients with recurrent malignant glioma confirmed the absence of neurovirulence and a recommended phase II dose was established. The median overall survival among patients receiving PVSRIPO was 12.5 months, which was slightly longer than the 11.3 months in the historical control group [116]. Another phase I trial of PVSRIPO treatment in patients with unresectable melanoma demonstrated that it. PVSRIPO is well tolerated and showed promising antitumor activity [117]. There are a further two ongoing phase I clinical trials investigating PVSRIPO monotherapy via it. routes in recurrent malignant glioma in pediatric populations (NCT03043391) and in invasive breast cancer (NCT03564782). At the time of writing, no data are available for these studies.

Another type of picornavirus under investigation is coxsackievirus A21 (CVA21), a naturally occurring enterovirus (Table 3). CVA21 attaches and infects host cells via the cellular receptor intercellular adhesion molecule-1, which is often highly expressed in melanoma and other cancer cells compared with non-malignant cells, therefore allowing a tumor-selective oncolysis when applying CVA21 [118]. CVA21 has promising efficacy results in phase I trials (NCT01227551; NCT01636882; NCT02307149) in the melanoma setting and has shown to be well tolerated in patients [119,120]. Similar results were also seen in a phase I/II trial in non-muscle invasive bladder cancer (NCT02316171) [121]. In addition, other phase I/II clinical trials are ongoing (active or active but not yet recruiting) across melanoma (NCT04152863; NCT04303169), solid tumors (NCT04521621) and lung and bladder cancer (NCT2043665).

Reoviruses are another OV of interest, with pelareorep being the primary candidate under investigation (Table 3). Pelareorep is a proprietary isolate of the reovirus Type 3 Dearing strain that is currently being evaluated in several clinical trials [122]. In a phase I study (NCT00528684) in patients with glioblastoma, pelareorep was determined to be tolerable by an independent safety panel and demonstrated promising efficacy [123]. Pelareorep has also shown encouraging efficacy and safety in combination with paclitaxel and carboplatin in melanoma (NCT00984464), in combination with gemcitabine or pembrolizumab in pancreatic cancer (NCT00998322; NCT03723915) and in combination with FOLFIRI and bevacizumab in KRAS-mutant colorectal cancer (NCT01274624), although a separate investigation in colorectal cancer demonstrated inferior efficacy with pelareorep in combination with FOLFOX6 and bevacizumab versus FOLFOX6 and bevacizumab alone (NCT01622543) [124–127]. Pelareorep was also not beneficial in non-small-cell lung cancer (NCT01708993) or ovarian cancer (NCT01199263) [128,129]. However, as reoviruses preferentially replicate and induce apoptosis in cells with a constitutively activated RAS pathway, it may be that certain populations of patients will have a greater benefit, such as those with RAS pathway mutations [8]. This is currently under investigation in colorectal cancer, breast cancer and myeloma [130–132].

GL-ONC1 is a double-stranded DNA virus, with a number of features that make it a good candidate for OV therapy. These include high genetic stability with replication restricted to the cytoplasm of infected cells, a short and well-characterized life cycle with efficient cell-to-cell spreading, and an excellent safety profile. The vector also encodes the marker genes for GFP and β-glucuronidase, making it possible to track the OV and to monitor in real time the course and extent of OV-induced oncolysis [133]. As discussed earlier, GL-ONC1 has shown promising results in a phase II trial for ovarian cancer [95]. In a separate phase I trial (NCT01443260), GL-ONC1 was tested for safety and a recommended dose was established for subsequent phase II studies investigating the treatment of patients with advanced-stage peritoneal carcinomatosis. The OV was well tolerated when administered into the peritoneal cavity via infusion through an implanted catheter. In this study, fluid biopsies were used to analyze key parameters, such as efficiency of tumor colonization, in-patient virus replication and oncolysis. This novel approach highlights how patients can be individually monitored during OV treatment to help guide treatment in a personalized way [133]. Moreover, this approach may replace the need for regular biopsies, which have a high patient burden. Recent data have also shown GL-ONC1 to have promising anti-tumoral activities in cell lines originating from neuroendocrine neoplasms, which are often detected only in a progressed metastatic situation, meaning that existing therapy options are currently limited [134].

Conclusion & future perspective

OVs have so far demonstrated promising safety and efficacy in clinical trials. A great deal of progress has been made in addressing some of the challenges in their translation to clinical practice. Genetic engineering strategies have helped to enhance the safety and efficacy of OVs. OVs are now considered to act through multiple modes of action beyond direct cell lysis, with genetic engineering making the addition of immune modulators and marker genes possible. Several studies have shown that OVs and other existing cancer therapies, such as immunomodulators and ICIs, can act in synergy to optimize therapeutic efficacy and to overcome potential resistance against either individual treatment modality. A diverse range of OVs are currently in clinical development and there are many factors still under consideration, including the route of administration and optimal combinations. Most ongoing trials are at a relatively early phase and, given the flexibility and specificity that OVs offer, it is hoped that in the coming years more OVs will reach regulatory approval. This offers great potential to broaden the range of therapeutic options currently available for patients across a variety of cancers.