Introduction

Cancer is still a serious threat to human health and a major cause of death worldwide, even among adolescents and young adults [1, 2]. The scientists have been pursuing the ideal tumor prevention and treatment strategies all the time. Numerous promising tactics have been well developed, such as immunotherapy, photodynamic therapy and oncolytic virotherapy (OVT) [3,4,5,6].

OVT has its unique advantages and prospects, because oncolytic viruses (OVs) preferentially infect and replicate in tumor cells and destroy them, while leaving healthy cells largely untouched [7]. With increasingly high therapeutic efficacy being achieved recent years and owing to the unique features such as specific tumor tropism, low cytotoxicity against normal cells, OVT has been inviting a great attention as an ideal weapon against cancers.

OVT has a long development history. Originally, viruses were known as the cause of human diseases, including some cancers [8]. It was not until early 1950s that the potential of viruses as anti-cancer agents had been recognized and applied [7, 9]. At that time, the application of tumor treatment with the spontaneous viruses or wild type viruses which quite often being scavenged by immune system, merely induces a subtle inhibition to tumor progression in patients. Meanwhile, these non-engineered viruses sometimes inevitably infect and spread to normal tissues, indiscriminately killed both tumor and normal cells, causing a series of unpredicted side effects. Therefore, safety and efficacy were the greatest challenges for the development of OVT. With the leap of gene cloning in the molecular virology, the scientists focus on improvement of their antitumor specificity and efficiency by manipulating the viral genomes. As shown in Fig. 1, we propose that OVT can be divided into the following four phases of development. The viruses originally used for treatment are usually spontaneous viruses. The first generation (G1) of engineered OVs mainly focus on manipulating within virus genome. By the genetic recombination the viruses were conferred with high specificity against tumor cells without targeting normal tissues. The first application of virotherapy with the engineered thymidine kinase (TK)-deficient herpes simplex viruses (HSV) was initiated in 1991 [10]. The second generation (G2) of engineered OVs armed with viral and/or non-viral genes. A series of chimeric viruses strategies, such as transductional targeting, transcriptional targeting, micro-RNA targeting and DNA shuffling approaches have been developed for restricting virus infection and toxicity in off-target tissues [11,12,13]. For example, Myb34.5, a second-generation replication-conditional HSV-1, has been exploited to target and dampen the pancreatic tumors [13]. Moreover, HSV engineered in gH of a scFv targeting the cancer-specific HER2 receptor, scFv-HER2-gH chimera, can enter, replicate and kill cancer cells efficiently [14]. The  third generation (G3) OVs were engineered with multiple coordinated viral and non-viral genes for tumor immunotherapy. Rivadeneira et al. demonstrates that intratumoral delivery of leptin by a VV can metabolically enhance tumor-infiltrating lymphocytes (TILs) effector and memory functions through improved mitochondrial oxidative phosphorylation, thereby potentiating therapeutic efficacy [15]. Anthony et al. engineered the vaccinia virus to express a nonsignaling, truncated CD19 (CD19t) protein for tumor-selective delivery, enabling targeting by CD19-CAR T cells [16]. Keeping stringency on tumor specificity and normal tissues safety usually hampers replicative fitness of viruses in target tissues. Thus, scientists keep pursuing ideal OVs that are highly tumor-specific without an attenuated clinical efficacy. In the first place, OVs have been designed to eliminate infected cancer cells by taking advantage of some of the most important properties of viruses or immune responses, including direct oncolysis, antitumor immunity, vascular-disrupting effects and bystander killing effect [17]. Secondly, along with the improvement of the tumor specificity, scientists are also constantly boosting the potency of OVT via prodrug activation, radiosensitization, immunostimulation and so on [18,19,20,21,22,23]. Worth to be noted, among these designs a second-generation oncolytic HSVs expressing TNF-α are being developed for cancer therapy and exerting its high efficacy for cancer therapy [24].

Fig. 1
figure 1

A timeline of milestones in the evolution of OVT

Full size image

Until now, more than twenty different virus families have been engineered for cancer therapy, including but not limited to HSV, adenovirus (Ad), measles virus (MV), Newcastle disease virus (NDV), vaccinia virus (VV), reovirus, myxomavirus, poliovirus, poxviruses, vesicular stomatitis virus (VSV) [7, 19, 25,26,27]. These engineered viruses usually focused on targeting replication of OVs in the tumor bed, initiation of an immune-stimulating or immune-recruiting inflammatory response and exposure of tumor-associated antigens that can be targeted by the immune system [28]. Moreover, the safety and efficacy of OVs in combination with other treatments have been explored continuously [28,29,30]. Arming strategies that combine chemo-, radio- and immuno-therapies with OVT will be strengthened by greater viruses replication and spread [19, 30,31,32,33,34,35,36]. In this review, the summary of the knowledge on the OVT, including the development history, the applications of preclinical studies, the mechanism of enhancing the safety and efficacy, and clinical trials were provided. In addition, the most important attractive schemes of genetic modifications and combinatorial regimens with OVs were highlighted.

OVs in preclinical development

As a promising cancer therapy strategy, OVT has immeasurable application potential, bringing a bright future to cancer patients. Many natural and genetically engineered OVs have been developed and underwent pre-clinical research stages (Table 1). Although the idea of using viruses to treat cancer originated in 1950s and has been around for more than 70 years, the modern era of OVT can be traced back to a 1991 cornerstone study, in which a TK gene was deleted in HSV with attenuated neurovirulence was shown to be active in a murine glioblastoma model [10]. Subsequently, the OVT upsurged globally and made great advance. The researchers began immersing themselves in manipulating various modifications with different types of viruses and testing them in animal models.

Table 1 The preclinical studies of OVs
Full size table

Recently, Lin et al. developed a novel immunotherapeutic HSV-1 (OVH-aMPD-1) expressing a scFv against PD-1, which releases damage-associated molecular patterns (DAMPs), promoting antigen cross-presentation by DCs, and enhancing the infiltration of activated T cells; these modifications resulted in activation of antitumor T-cell that led to reduced tumor burdens in a mouse model of liver cancer [29]. In addition to awaken T cell response, activating other types of immune cells is also a wise option. The combination of EGFR-CAR NK-92 cells with oHSV-1 resulted in more efficient killing of MDA-MB-231 breast tumor cells and significantly longer survival of tumor-bearing mice when compared to monotherapies [36]. A UV light-inactivated HSV-1 (UV-HSV-1) potently activates human peripheral blood mononuclear cells (PBMCs) to lyse leukemic cell lines and primary AML samples, but not healthy allogeneic lymphocytes. The data suggested that UV-HSV-1 synergizes with IL-15 and IL-2 in inducing activation and cytolytic activity of NK cells [37]. Moreover, to reduce toxicity and enhance oncolysis to destroy glioma, Delwar et al. replaced the HSV ICP4 promoter with the survivin promoter and introduced the 5’UTR of rat FGF-2, and 5 copies of the miRNA 124 target sequence 3’UTR into the ICP4 gene. The intratumorally injected oHSV-1 was demonstrated to be effective in mice bearing human glioma U87 tumors, whereas viral DNA was almost undetectable in normal organs [38]. To evade antiviral defense response, arming oHSV with antiangiogenic N-terminal cleavage fragment of brain angiogenesis inhibitor (Vstat120) shields oHSV-Vstat120 from inflammatory macrophage antiviral response, without reducing safety [39]. oHSV-Vstat120 treated mice harboring renal adenocarcinoma and melanoma tumors presented increased infiltration of tumor-associated macrophages (TAMs), NK cells, and tumor-infiltrating lymphocytes [40].

Activating the host immune system seems to be a popular route for potentiating anti-tumor effect of OVs. Polysialic acid (polySia) is expressed on several malignant tumors of neuroendocrine origin. PolySia-dependent systemic infection in vivo facilitated effective uptake of viruses in subcutaneous polySia-expressing human tumors, whereas hepatic viral load and hepatotoxicity were significantly reduced. Enhanced tumor regression and prolonged survival was only observed in immunocompetent mice, but not in T-cell-deficient mice, suggesting that a polySia-retargeted oAd elicits an effective tumor-directed T-cell response after systemic virus delivery and facilitates therapy of disseminated lung cancer [41]. DNX-2401 (Delta-24-RGD; tasadenoturev) is a tumor-selective, replication-competent oAds, which is proven to be safe in mice and results in a pronounced increase in survival in immunodeficient and immunocompetent models of high-grade pediatric glioma and diffuse intrinsic pontine gliomas [42]. The Ad was engineered to express an EGFR-targeting BiTA (cBiTA) antibody under the control of the major late promoter, leading to generation of ICOVIR-15 K-cBiTA, which bound specifically to both CD3 + and EGFR + cells. Intra-tumor (IT) injection of this cBiTA-expressing Ad increased the accumulation and persistence of tumor-infiltrating T cells and the antitumor efficacy in vivo [43]. Actually, as MSCs present tropism for tumors, the use of MSCs to transport OVs to tumor sites is a promising alternative to IT administration [40]. The data suggested that treatment with oAd-MSCs significantly reduced tumor volumes by 50% and induced a pro-inflammatory TME. In a veterinary dog trial with dCelyvir (canine MSCs infected with an oAd ICOCAV17) in 27 canine patients, Cejalvo et al. observed an excellent toxicity profile as well as a clinical benefit in 74% of patients, including 14.8% showing complete remissions [44]. Actually, it is a very promising attempt to arouse T cells by designing BiTAs OVs [45, 46]. Particularly, together with T cells a VV encoding a secretory BiTA consisting of two scFvs specific for CD3 and EphA2 (EphA2-TEA-VV) had potent antitumor activity in comparison with control VVs plus T cells in a lung cancer xenograft model [47]. In vivo, the therapeutic efficacy of MVs targeted to HER2/neu and EpCAM by designing ankyrin repeat proteins (DARPins), was confirmed in an orthotopic ovarian carcinoma model revealing an effective reduction of tumor mass [48]. Overall, these successful preclinical results have made a decisive contribution to further investigation in the clinics.

Safety of oncolytic virotherapy

Therapeutic safety remains a paramount concern during OVT while the tumor targeting/tropism is a highly desirable characteristic for OVs. Generally, tumor-specific and natural receptors were responsible for tumor selectivity and cell entry. To achieve cancer cell specificity in different OVs, a few viruses, e.g., parvovirus and NDVs, own a naturally tumor tendency. Many, if not most, such as MVs, Ads, VSVs, VVs and HSVs exhibit no preference for cancer cells. Thus, the viruses from these families need to be designed to preferentially target cancers rather than normal tissues.. Genetically engineered viruses can be exploited in several aspects, such as tumor cell receptor targeting, driving the expression of certain viral replication genes by promoters and enhancers, translational targeting, engineered microRNA target sequences, immunogenic tumor-associated antigen targeting, etc. (Table 2 and Fig. 2) [19, 49]. Taking HSV, one of the most widespread and widely used OVs, as an example, to improve its safety, various engineering and modifications have been carried out on its genome [24]. Mutants of HSV-1 with deletion of ICP34.5 and ICP47 genes (such as T-VEC) have been successfully harnessed as attenuated oncolytic vectors [50, 51]. For HSV-based OVT, the detargeting-retargeting strategies so far were based on genetic manipulations of glycoprotein (g) D, gB and/or gH [52]. In particular, to enhance the tumor tropism and safety of HSV, a novel ligand in gH was designed to confer tumor cells entry [14]. To re-target the virus tropism to the HER2- and GCN4R-positive cells, the HER2 binding peptide was inserted in gB and GCN4 peptide in gD or gB [53, 54]. A safe and effective therapeutic oncolytic HSV-2 (deletion of ICP47 and ICP35.4) was also be used in combination with doxorubicin for breast cancer treatment [55]. Similarly, arming the miR-122a complimentary sequences to HSVs have shown high specificity to target hepatocellular carcinoma cells [12]. Engineering miRNA target sequences into viruses’ genomes was thereby inhibiting spread in tissues expressing cognate miRNAs. Tumor-specific translational regulation presents an attractive possibility for generating oncoselective therapies. Villanueva et al. reported the insertion of CPE regulatory sequences in the 3’-UTR of the E1A gene that confers translational E1A expression regulation, resulted in tumor-specific AdCPE viruses [56]. It is demonstrated that neurotoxicity was most profoundly reduced in a virus carrying four tandem copies of a neuronal miR-125 target sequence inserted in the 3′-UTR of the VSV polymerase gene [57]. Alexander Muik et al. have engineered a chimeric VSV, an oncolytic virus called rVSV (GP) devoid of natural neurotoxicity with undetectable immunogenicity and enhanced oncolytic potency [58].

Table 2 Tumor specificity mechanisms of OVs
Full size table
Fig. 2
figure 2

The tumor specificity of oncolytic virotherapy. IA: Deletion of the required genes for virus replication in normal cells. IB: Deletion or inactivation of the required genes for virus replication and insertion of killer genes. IC: Transcriptional targeting. IIA: Retargeting strategies based on genetic manipulations of glycoproteins. IIB: T-cell activator. III: Translational targeting. IV: Hypoxic or drug induction. V: Intracellular restrictions by miRNA targeting

Full size image

Here we summarized the virulence and tumor specificity mechanisms of different virus families in recent years (Table 2). Among them, the selection of tumor-specific antigens is a leader in increasing the safety of OVs. The detargeting-retargeting strategies were based on genetic manipulations of glycoprotein of different types of viruses, such as antigens of HER2, EGFR, GCN4, EpCAM have been sucessfully applied in HSV, VSV and MV etc. To date, OVT have been evaluated for safety by both localized and systemic administration. The most common adverse effects are fever and general flu-like symptoms. Moreover, no transmission of OVs from treated patients to others has been reported [19]. However, therapeutic safety concerns must be scrupulously addressed to ensure the safety of patients and other people who may have contact with the patients. The development of OVT were greatly benefited from the studies on structures and characteristics of virus particles [59,60,61,62]. More engineered OVs for particular tumor treatment will be safely applied in clinical trials and approved protocols.

Efficacy of oncolytic virotherapy

Although safety concern is a paramount priority, high efficacy to eliminate tumors is the goal of OVT. OVs can destruct cancer cells in many ways, including direct oncolysis, antitumor immunity, vascular-disrupting effect, bystander killing effect [17]. Therefore, to pursue the ideal therapeutic effect, we may start from following aspects. First, the importance of tumor targeting in improving therapeutic effect is out of question. Due to the rapid replication and cell lysis properties of some virus families, with a wide range of tissue tendencies, it is necessary to continue rational optimization of these viruses to efficient kill specific types of cancer. For example, the natural neurotropism of HSVs has made it attractive as vectors for the development of OVs for application in the nervous system [63, 64]. Moreover, retargeted OVs infected only cells that expressed the targeted TAAs, such as EGFR, HER-2, PSMA, GCN4R (Fig. 3A and Table 3). Second, suitable doses and delivery system of OVs in administration, such as intratumor (I.T.), intra-vein (I.V.) and intra-muscle (I.M.) injection, are required [65,66,67]. Third, to elicit the bystander immune response is a preeminent blueprint [68, 69]. Fourth, arming the viruses with destruction/immunostimulatory genes, innovative combination with other therapies are promising strategies gaining momentum [70]. Herein, the arming mechanisms of OVs were summarized (Table 3 and Fig. 3) and discussed below.

Fig. 3
figure 3

The therapeutic efficacy of oncolytic virotherapy

Full size image
Table 3 The therapeutic efficacy mechanisms of OVs
Full size table

The optimization on virus spread and delivery of OVs play a crucial role directing therapeutic efficacy. There are several host barriers hampering the potency of OVT in patients. If the OVs is not administrated I.T., I.V. and I.M. injection of OVs was usually hindered by antibodies and complements in the blood stream. Thus, it is essential to develop strategies to escape antibody and complement neutralization in the blood stream. To limit the neutralization of OVs, there are several classical oncolytic vector shielding strategies, including envelope protein exchange within a virus species or families, multiple epitope replacements, devising cell carriers, and chemical modifications [19, 49, 71, 72].

To restrict antibody-mediated HSV neutralization, the antibodies targeting functional epitopes on HSV glycoproteins can mediate neutralization directly. For example, epitopes modification on HSV have been well-defined and characterized in humans [73,74,75]. MVs-based shielded oncolytic vectors to circumvent antibody neutralization have been developed by exchanging the envelope glycoproteins, hemagglutinin (H) and fusion (F) protein, with those from the non-cross-reactive Tupaia paramyxovirus [76]. In genital disease, HSV-2 vaccination with human papillomavirus vectors expressing HSV glycoprotein antigens was developed successfully for eliciting anti-viral response [77]. Cristian et al. demonstrated that Ads coated with MHC-I tumor epitopes increase the antitumor immunity and efficacy against melanoma [78]. Cell carriers, such as cytokine-induced killer cells, mesenchymal stem cells (MSCs), neural stem cells (NSCs), and stromal vascular fraction cells (SVFs), are capable of accelerating the OVs delivery to tumors and in the same time protecting OVs from antibody neutralization [7, 79,80,81]. Multiple studies have demonstrated that MSCs or NSCs allow for safe and efficient ferrying of OVs to tumor foci to trigger immune response [65, 71, 79, 82,83,84]. Specifically, a TK-positive oVV ACAM2000, delivered by autologous adipose SVF cells, fostered such treatment in the patients with advanced solid tumors or acute myelocytic leukemia (AML) in a great safety and accessibility. The clinical data revealed that the viral DNA could be readily detected in all patients’ blood samples immediately after treatment [80]. Certainly, chemical or other modifications are also good OVs shielding option. Reoviruses and HSVs have been armed with cyclophosphamide, an immune modulator, to combat the antibody neutralization, thereby enhancing the virus infection [85]. Since copper in serum prevents replication of HSV-1, when armed the oHSV with a copper chelator ATN-224, significantly enhanced its therapeutic efficacy by increasing serum stability and systemic delivery of oHSV [86]. Rommelfanger et al. have demonstrated that the combination of VSVs and LPS generated significantly enhanced therapy of melanoma B16ova tumors upon direct I.T. administration [87]. Besides, the modification of the fiber knob and an arginine-grafted biodecomposible polymer arming were proved to be a feasible strategy to dodge antibody neutralization during systemic administration [88]. When measured just before the second treatment cycle, serum neutralizing antibodies titers differed in 83% of patients, suggesting that even minor changes in the fiber knob would able to circumvent host antibody neutralization [88]. Another example of modification is that the NDVs armed with regulators of complement activity CD46 and CD55 could enhance the efficient complement evasion [89]. Some complement inhibitors, such as CP40, have been shown to abolish host antibody neutralization and augment the dose of infectious oVVs ferried to tumor sites [90].

Once high doses of the viruses were maintained in the tumor microenvironment (TME), the therapeutic efficacy will be ultimately determined by the potency of OVs. As shown in Figs. 3 and 4, to reinforce the antitumor activity of OVs, eliciting bystander cell killing, introduction of pro-apoptotic or toxin genes and innovative combination therapy strategies were developed. OVs could use oncolysis to kill the infected tumor cells directly in TME. Except tumor cells, OVs can target several other components including cancer-associated fibroblasts (CAFs) and vascular endothelial cells (ECs). Then OV infection and the lysed cells causes the release of cytokines or neo-antigens, as well as the OV-armed immuno-stimulation genes, including GM-CSF, INF-γ, to initiate anti-viral immune priming by stimulating immune cells, including T cells, NK cells. The recruitment and maturation of innate immune cells which can cross-present TAAs to CD8 T cells, thus generating populations of TAA-specific CTLs. The generation of an OV infection-mediated anti-tumor immune response also counteracts the immunosuppression associated with myeloid derived suppressor cells and Tregs. In addition, the various destructive genes (such as pro-apoptotic genes, toxin genes) that are engineered within the OVs will take effects in TME. It is effective to mediate T and/or NK cell bystander killing of uninfected tumor cells in TME by engineering BiTA, CiTA, TriKA etc. (Fig. 4). Thus, OV infection acts on both the innate and adaptive immune system, which work together to kill cancers. The promising methods to create the bystander killing were prodrug activations, radiosensitization and immunostimulation [19]. For example, the purine nucleoside phosphorylase (PNP), one of convertase enzymes expressed in infected cells could convert prodrugs within the TME into toxic metabolites which eventually diffuse into and destruct adjacent uninfected tumor cells [19]. The sodium-iodide symporter (NIS) concentrates radioactive ions in infected cells, which triggers radiation poisoning of uninfected bystander tumor cells [19, 91, 92]. The clinical study demonstrated that oMV therapy can function as an antigen agnostic vaccine, increasing cytotoxic T-lymphocyte responses against TAAs in patients with multiple myelomas [92]. Of course, the most exciting strategy is the clinical application of OVs immunotherapy. The successful introduction of the granulocyte macrophage colony-stimulating factor (GM-CSF) gene into oHSVs represents a great breakthrough of immunostimulation. Such oHSVs, including T-VEC, CG0070, JX594, JX963, etc., have been shown in clinical trials to stimulate granulocytes and monocytes to elicit impressive anti-tumor immunity [21, 30, 93,94,95]. T-VEC, which produce GM-CSF, can efficiently treat the patients with metastatic melanoma, pancreatic carcinoma etc. [18, 21, 30, 96,97,98]. The phase III trial proved that local intralesional injections with T-VEC in advanced malignant melanoma patients can not only suppress the growth of injected tumors but also act systemically and prolong overall survival (OS) [30, 99]. Besides of immune stimulatory cytokines GM-CSF, IFNα, IL-12, IL-15 etc., immune checkpoint inhibitors (ICIs), bispecific T-cell activators (BiTA), some pro-apoptotic or toxin genes and shRNAs (targeting Bcl-2, Survivin, COX-2 or STAT3) were also engaged in OVT [22, 29, 100,101,102,103,104]. The redirecting of T cells to the tumor by arming oVVs with BiTA (EphA2-TEA-VV) has the potential to boost the antitumor activity of oncolytic VVs [47]. An HSV-2 based OV can actively recruit T effector cells to the site of infection, suggesting that oHSV-2-based virotherapy can be armed with adoptive T-cell therapy to advance its therapeutic effect against solid tumors [105]. Expression of cytokines together with BiTAs has shown to mediate T cell bystander killing of uninfected tumor cells not only in vitro, but also in vivo [47, 100, 101, 106]. A combination of trans-genes encoding BiTAs, ICIs and APC enhancers will remove suppressive hurdles in the TME and allow for optimal antitumor efficacy of armed OVs [22]. The antibodies against immune checkpoint receptors have been exploited to conquer cancer by inducing T cell response, such as the antibodies against CTLA4, PD-1, PDL-1 and some alternative antibody formats (scFvs, Fabs, scAbs and VHHs) [22, 29, 107]. Zamarin et al. boosted the efficacy of systemic immune checkpoint blockade and avoided additional systemic toxicity by engineering a recombinant ICOS ligand-expressing NDV (NDV-ICOSL) [108]. Antibodies against immune checkpoint receptors, such as anti-CTLA4 and anti-PD-1, has clearly proven the therapeutic potential of antigen presentation and T-cell response against cancer [22, 29]. Moreover, the larger natural antibodies are not easy to eliminate and penetrate into solid tumors, the alternate antibody forms such as scFvs, Fabs, scAbs and VHHs have been increasingly exploited and applied [22, 29].

Fig. 4
figure 4

OVT in the tumor microenvironment (TME)

Full size image

Maria et al. engineered a specific oncolytic Ads expressing a scFv of an antibody against PD-L1 to combine blockage of PD-1/PD-L1interaction with the antitumoral activity of Ad5 [109, 110]. They also armed Ads expression of an Anti-PD-L1-scFv improves anti-tumoral efficacy in a melanoma mouse model [109] Anthony et al. engineered the OVs to express a nonsignaling truncated CD19 (CD19t) protein tumor-selectively, enabling CD19-CAR T cells to target, and showing effective anti-tumor effect [16]. A recent report by Rivadeneira et al. demonstrated that OVs engineered to express the adipokine leptin boosted T cell metabolic function in the TME, and thereby allowed a superior antitumor response [15]. Dendritic cells played important role in oncolytic virotherapy. Cytopathogenic infection of neoplastic cells releases the proteome and exposes pathogen- and damage-associated molecular patterns. At the same time, sublethal infection of antigen-presenting cells, such as dendritic cells and macrophages, yields potent, sustained type I interferon-dominant activation in an immunosuppressed microenvironment and promotes the development of tumor antigen-specific T cell responses in vitro and antitumor immunity in vivo [111]. The recombinant poliovirus/rhinovirus chimera oncolytic virus PVSRIPO’s immune adjuvancy stimulates canonical innate anti-pathogen inflammatory responses within the TME that culminate in dendritic cell and T cell infiltration. The findings provide mechanistic evidence that PVSRIPO functions as a potent intratumor immune adjuvant and generates tumor antigen-specific cytotoxic T lymphocyte responses [111]. T-Vec results in a rapid eradication of malignant cells and leads to interferon pathway activation and early influx of natural killer cells, monocytes, and dendritic cells. These events are followed by enrichment in cytotoxic T cells and a decrease of regulatory T cells in injected and noninjected lesions [112]. High-grade serous ovarian cancers (HGSOCs) exhibit limited response to immune checkpoint blockade. In a new study in Cancer Cell, Duraiswamy et al. highlighted that intratumoral CD28 co-stimulation by myeloid-antigen-presenting cells as a key mechanism was required for activation of programmed cell death receptor 1 (PD-1)+ tumor-infiltrating T lymphocytes during PD-1 blockade in HGSOC [113, 114].

The destructive genes, e.g. pro-apoptotic and toxin genes, have been engineered with OVs successfully. For example, arming OVs with a secretable and self-multimerizing apoptosis inducer is a approachable strategy to enhance the potency of OVT. Loya et al. armed HSV with a secreted form of an Her2 single chain antibody linked to the Fas ligand extracellular domain (Her2-COL-sFasL), which improved the bystander effect of OVT effectively [17]. Arming human MSCs with oHSV and its pro-apoptotic variant, oHSV-TRAIL, proved to be efficient in treatment for malignant glioblastoma multiforme [115]. Therapy of experimentally induced lung melanoma in mice with IL-15-carrying myxomavirus delivered by MSCs led to marked regression of lesions and with increased animal survival, suggesting that it allowed for safe and effective delivery of OVs to pulmonary melanoma lesions triggering immune responses [83]. HSV1716 administration led to marked tumor shrinkage in primary mammary tumors and a decrease in metastases by reprograming tumor-associated macrophage to a less immunosuppressive phenotype. This was associated with a significant increase in the recruitment/activation of cytotoxic T cells [116]. A pro-apoptotic gene p53 has been engineered in Ads to treat hepatocellular carcinoma (HCC) and could prolong the survival time of the patients [117]. Dual silencing of Bcl-2 and Survivin with oHSV-1 was also a promising tool for improving the antitumor efficacy [118]. A toxin gene, staphylococcal enterotoxin A, is also a potential useful anti-tumor agent in arming Ads [119]. A virulence factor, helicobacter pylori neutrophil-activating protein (HP-NAP), can mediate antitumor effects by recruiting neutrophils and inducing Th1-type differentiation in the TME. Thus, Ads armed with HP-NAP gene provoked antitumor immune response and enhanced the therapeutic effect against neuroendocrine tumors [120]. The study demonstrated that the cancer-associated fibroblasts (CAFs) induced high levels of fibroblast growth factor 2 (FGF2), which enhanced the susceptibility of the cancer cells to OV infection and improved therapeutic efficacy [121]. Telomelysin, a telomerase-specific replication-competent Ads with hTERT promoter, has been proven to have a strong antitumor effect on a variety of cancers and applied in combination treatment for head and neck squamous cell carcinoma [122]. The control of exogenous gene expression can also improve OVT. Jochen Stritzker et al. has characterized a doxycycline-inducible promoter system in oVVs, which was proven to be beneficial to OVT [123]. Therefore, determination of the structure and characteristics of various viruses and tumor cells will be greatly beneficial for the development of efficient OVT.

Overall, in addition to edit the viruses and exogenous genes, to excavate the reasonable combinatorial modalities are regarded as an excellent strategy to improve efficiency, especially ICIs [124,125,126,127] (Tables 3 and 4). For example, T-VEC with ipilimumab (a CTLA-4 inhibitor) had a tolerable safety profile, and the combination appeared to have greater efficacy than either T-VEC or ipilimumab monotherapy [30, 35]. The combination of intratumoral G47Δ and systemic anti-CTLA-4 antibody was shown to recruit effector T cells into the tumor efficiently while decreasing regulatory T cells [128]. Viral replication and the creation of new T-cell clones have been detected during treatment with reovirus pelareorep combined with a PD-1 inhibitor pembrolizumab [129]. While anti-PD-1 antibody monotherapy moderately improved tumor survival, when co-administered with oncolytic Zika virus (ZIKV), survival extended [27].

Table 4 The clinic trials of OVs
Full size table

OVs in clinical trials

Although the pre-clinical trials so far have established the safety and efficacy of those approaches, the challenge now is to achieve safety and efficacy in clinics. Many promising OVs, such as oHSVs, oAds, and oVVs, have been applied in clinic trials successfully (Table 4).

T-VEC, a recombinant oHSV, which is administered by direct I.T. injection to patients with metastatic malignant melanoma led to lesion regressions of [30, 34, 35, 51, 96, 99, 130]. As an example, the biodistribution, shedding, and potential transmission of T-VEC have been systematically evaluated during and after completion of therapy in adults with advanced melanoma [131]. The data demonstrated that T-VEC improved longer-term efficacy versus GM-CSF and maintained well tolerated. The final planned OPTiM analysis suggested that the median OS was 23.3 months (95% confidence interval [CI] 19.5–29.6) and 18.9 months (95% CI 16.0–23.7) in the T-VEC and GM-CSF arms, respectively [130]. A phase II study evaluated patients with unresectable stage IIIB-IVM1c malignant melanoma who received T-VEC plus ipilimumab or ipilimumab alone. The results showed that 39% (n = 38/98) in the combination arm and 18% (n = 18/100) in the ipilimumab arm had an objective response. Eight responders (combination, n = 7 [18.4%]; ipilimumab, n = 1 [5.6%]) had pseudo-progression; most occurred by week 12 and were caused by an increase in existing lesions [30]. In addition, to determine the safety of administering HSV1716 (Seprehvir) systemically, Streby et al. conducted the phase I trial of intravenous (I.V.) injection in young patients with relapsed or refractory extra-cranial solid cancers [132]. They did not observe any dose-limiting toxicities. All five HSV-1 seronegative patients seroconverted by day 28. Four out of nine patients had detectable HSV-1 genomes in peripheral blood on day + 4, which is consistent with de novo virus replication. A phase I/IIa trial of intrapleural administration of HSV1716 with malignant pleural mesothelioma patients demonstrated that viral replication/persistence in pleural fluid in seven of the twelve patients. Induction of Th1 cytokine responses to HSV1716 treatment was achieved in eight patients and four patients developed novel anti-tumor IgG [133]. However, it is also suggested that the efficacy of T-VEC therapy in patients with in-transit melanoma metastasis diminished with increasing lesion size [134]. Of 27 patients, an objective response was observed in 11 (40.7%), including one patient with partial response (3.7%) and 10 with complete response (37.0%). Logistic regression demonstrated each millimeter increase in maximum lesion diameter predicted decreased ORR (odds ratio [OR] 0.866, 95% CI 0.753–0.995; p = 0.04) [134]. Todo et al. have been reported the results of a phase I/II trial using triple-mutated oHSV-1 G47Δ in Japanese patients with recurrent or progressive glioblastoma [135, 136]. G47Δ caused immediate infiltration of lymphocytes that seemingly directed towards tumor cells, which was reflected on image studies with features characteristic to G47Δ therapy. Long-term survival (> 46 months) was observed in 3 of 13 patients, which may be due to the delayed effect of G47Δ via antitumor immunity [136].

Since that first approve of the human p53 adenovirus (Gendicine), a steady stream of new oAds entering the clinical arena [137, 138]. Clinical studies demonstrated that DNX-2401 is safe and tolerable after injection into the cerebellar peduncle in pediatric patients with diffuse intrinsic pontine gliomas and can induce a direct oncolytic effect followed by an antitumor immune response [68]. ICOVIR5 was derived from the oAd DNX-2401. The clinical results in 12 patients treated with a single dose up to 1 × 1013 viral particles showed that ICOVIR5 was able to reach melanoma metastatic lesions after infusion but failed to induce tumor regressions [139]. The homing capacity of MSCs to tumors makes them excellent carriers of anticancer therapeutics [40, 44]. Autologous MSCs may allow an increasing amount of ICOVIR5 by repeated administration, avoiding or minimizing emergent toxicities [82]. Evidence have been reported that MSCs successfully delivered an oAd CRAd-S-pK7 with fiber modification of seven lysine residues to diffuse intrinsic pontine glioma [71]. Similarly, it is shown to protect CRAd-S-pK7 from neutralizing antibodies within patient ascites fluid and to enhance delivery of CRAd-S-pK7 by NSCs for treatment of metastatic ovarian cancer [84]. Recently, the safety and feasibility of NSC-CRAd-S-pk7 in patients with newly diagnosed high-grade glioma have been examined, and the results showed that the median progression-free survival was 91 months (95% CI 85-not reached) and median OS was 184 months [65]. In addition, Pascual-Pasto et al. confirmed that the oAd VCN-01 provided targeted therapeutic activity against even chemo- resistant retinoblastoma. The phase I data in patients showed the feasibility of the administration of intravitreous VCN-01 and resulted in antitumor activity in retinoblastoma vitreous seeds and evidence of viral replication markers in tumor cells [140]. In another phase I study of gene-mediated cytotoxic immunotherapy using aglatimagene besadenovec (AdV-tk), an adenoviral vector expressing the HSV-tk gene, followed by valacyclovir, 3 patients in a dose of level 2 (3 × 1011 vp) survived more than 24 months after treatment, and 2 remain alive without progression at 37.3 and 47.7 months after AdV-tk injection [141]. Enadenotucirev is a tumor selective oAd, which can be administrated intravenously in patients undergoing primary tumor resection [142]. Additionally, the EVOLVE (Evaluating Oncolytic Vaccine Efficacy) study of the enadenotucirev, administered intravenously to patients with epithelial solid tumors, showed that enadenotucirev monotherapy can be administered in a single cycle or repeated cycles with manageable tolerability [67]. Recent clinic trial confirmed that enadenotucirev is a radiosensitizer in chemoradiation therapy of locally advanced rectal cancers [143]. Intravenously dosed enadenotucirev plus paclitaxel demonstrated manageable tolerability and increased tumor immune-cell infiltration in phase 1 studies in platinum-resistant ovarian cancer [144].

An oVV, Pexa-Vec (pexastimogene devacirepvec, JX-594), engineered to express GM-CSF, was administered IT and IV to patients with HCC and colorectal cancer, respectively [94, 145, 146]. No dose-limiting toxicity (DLT) was reported, and the maximum tolerated dose was not reached in phase Ib trial of biweekly IV of Pexa-Vec. Moreover, the most common adverse events were grade 1/2 flu-like symptoms, generally lasting less than 24 h [146]. TG4023 is a modified vaccinia virus Ankara (MVA), the first-in-human study demonstrated that IT injections of TG4023 were feasible and well tolerated, and the maximum tolerated dose (MTD) was defined as 4 × 108 p.f.u. [147]. The safety of oVV GL-ONC1 have been determined when delivered intravenously with chemoradiotherapy to patients with primary, nonmetastatic head and neck cancer [148]. Moreover, the study showed that GL-ONC1 was well tolerated when administered into the peritoneal cavity of patients with advanced stage peritoneal carcinomatosis. Importantly, in 8 of 9 studied patients, effective peritoneal infections, in-patient replication of GL-ONC1, and subsequent oncolysis were detected [149]. ACAM2000, a TK-positive strain of oVV, is the current smallpox vaccine in the US. The phase I clinical trial confirmed that ACAM2000/SVF can safely be administered in patients with advanced metastatic solid tumors or advanced AML [80].

In addition to the above described oHSVs, oAds, and oVVs, an oMV engineered to express the human thyroidal natrium iodine symporter (MV-NIS) monitors localization of viral gene expression and successfully used in clinical trials against multiple myelomas and ovarian cancers [92, 150, 151]. Packiriswamy et al. conformed that MV-NIS treatment significantly (P < 0.05) increased cytotoxic T-lymphocyte responses against TAAs in patients with MM [92]. An oncolytic parvovirus ParvOryx containing native parvovirus H-1 (H-1PV) have been shown to be a promising candidate for treatment of patients with recurrent glioblastoma and metastatic, inoperable pancreatic cancers [152, 153]. Pelareorep, an oncolytic reovirus, in combination with chemotherapy and pembrolizumab in patients with advanced, pre-treated pancreatic ductal adenocarcinoma (PDAC) was well-tolerated and showed prolonged efficacy in 3 of 11 patients (27.3%) [129].

Despite the confirmed safety and antitumor efficacy of OVs, additional challenges have been gained from the ongoing and completed clinical trials. A first insight is that the predictive values including safety and efficacy profile are limited by the relatively small sample size of patients and short follow-up. A second awareness is that the antigenic specificity of the T cell response to these OVs has not been determined, and whether the treatment expands the appearance of new antigen specific T cell lineages; further research is required to monitor/determine any relationship between virus persistence and the TME. Third, the role of adaptive immunity in restricting the benefits of repeated administrations of viruses is unknown. In addition, it is not clear which administrations of OVs is better, injecting the tumor intratumorally, intravenously, or orally, which may vary depending on the individual tumors, viruses, patients, and combination therapy regimen.

Significantly, Gendicine is the first OV approved for clinical OVT in the world in 2003 [137, 138], which was approved for head and neck carcinoma by China FDA and T-VEC is the second OV approved for clinical OVT in the world in 2015 [154], which was approved for melanoma by the US FDA. Many promising OVT clinical trials are under way but there is still a long way off to improve their safety and efficiency.

Conclusions

OVT is an amazingly versatile and malleable class of cancer therapy, which has the unique advantages when compared with that in conventional therapies. OVs can attack tumor cells selectively, and then trigger the cell death by multiple approaches, including direct oncolytic effects, targeting blood vessel endothelial cells, delivery of the therapeutic genes within tumors, synergistic effects with traditional and immunotherapies, resulting in systemic anticancer effects. The toxicity of OVs has been self-limiting flu-like illness and fever etc. Until now, OVT has become a realistic therapeutic candidate, and has been evaluated for safety by both localized and systemic administration in clinics. From the previous studies, we conclude that the status of OVs potencies including: (i) induces systemic tumor-specific immunity, (ii) synergistic effects with other therapies, (iii) different tumor sites and patients showed varying response to different viruses, (iv) neutralizing antibody is not a barrier to successful therapy; and (v) anti-tumor T cell (BiTAs, checkpoint inhibitory T-cell-activators/CiTAs) or NK cell (trispecific killer activators, TriKAs) responses augment antitumor efficacy by OVTs.

Oral, I.V., I.T., intrapleural, intraperitoneal (IP), aerosol and limb injections are the common delivery routes for OVs. However, these methods still have their own disadvantages. To be specific, oral administration is most convenient and most unavailable. I.V. and I.T. injections are easy to be neutralized in blood stream of patients. Besides, not all patients can be adapted to I.T. injection. Intrapleural injection should be utilized by using an indwelling intrapleural catheter. To avoid uncontrolled adverse events and long-term complications of OVs, the patients need to orchestrate the appropriate time and delivery routes in clinics.

We believe OVT has a bright future and requires continue efforts working for its safety and efficiency. It is wise to explore the key factors affecting the efficacy of OVs from three aspects: virus, tumor and patient. This include reconstructing the viruses for better efficiency with more safety, utilizing intrinsic tumor-associated genes for target specificity, invoking immune responses from host for enhanced tumoricidal effect. To further avoid host immunity to viruses or enhance tumor specific immunity induced by OVs in the future, the potential novel investigations should be focusing on the following aspects: (i) sequential harness of two different OVs, (ii) choreographed combination of OVs and antibody therapies (anti-PD-1/PDL-1, anti-CTLA-4), or cell therapies (adoptive cell transfer therapy, DC, Car-T), and (iii) improve the efficacy of administration and delivery by excellent cell carriers (MSCs, NSCs, etc.).