Deciphering permissivity of human tumor ecosystems to oncolytic viruses

Abstract

Effective cancer therapy involves initiation of a tumor-specific immune response. Consequently, the interest in oncolytic viruses (OV) capable of triggering immunogenic cell death has sparked in recent years. However, the common use of pre-clinical models that fail to mirror patient tumor ecosystems (TES) hinders clinical translation. Here, we provide a condensed view on the intricate interplay between several aspects of TES and OV action and discuss these considerations in the view of recently developed pre-clinical human model systems. Given the urgent demand for innovative cancer treatments, the purpose of this review is to highlight the so-far overlooked complex impact of the tumor microenvironment (TME) on OV permissivity, with the intent to provide a foundation for future, more effective pre-clinical studies.

Introduction

Cancer immunotherapies have shown promising results in recent years [1]. Particularly treatments based on immune checkpoint blockade (ICB) dominate the immunotherapy landscape, with a total of eleven therapeutic monoclonal antibodies approved by FDA in 2024 [2]. In contrast, only four OV therapies have been approved for cancer globally, with talimogene laherparepvec (T-VEC) being the only one widely approved OV compound so far [3]. Despite the unprecedented durable responses obtained with ICB agents addressing PD-1/PD-L1, a large number of patients do not benefit from the treatment (exhibiting primary resistances to ICBs), and some responders relapse after a period of response (exhibiting acquired resistances to ICBs) [4, 5]. Therefore, alternative or synergistic therapies with a different mode of action are needed, prompting a recent increase of interest in OV therapies [3, 6, 7]. The finding that T-VEC treatment triggered an increase in CD8⁺ T cell infiltration in tumors of patients that responded to anti-PD-1 combination therapy [8], furthermore underlined the high potential of OVs: OV-mediated immune infiltration may be able to convert immunologically cold tumors into an immunologically hot state, making it an attractive combination therapy with immunotherapies like (i) ICB [9], (ii) adoptive cell therapies [7], and (iii) cancer vaccines [10] (Fig. 1). Furthermore, to unleash their full potential, OVs can be designed to express immune-modulatory cargos including cytokines and chemokines, allowing the remodeling of the TME [7], or even the expression of bispecific T cell engagers (BiTEs) leading to killing of cancer cells and tumor-associated fibroblasts [11]. The wide variety of different available OV platforms, ranging from RNA viruses like vesicular stomatitis virus (VSV) [12] to DNA viruses such as herpes simplex virus (HSV) [13], furthermore expands the flexibility of this therapy platform when it comes to cargo sizes and/or expression kinetics.

Fig. 1. The emerging intersection of oncolytic virotherapy and immunotherapies.

Activity of OVs in tumors triggers recruitment and activation of immune cells, culminating in a state known as immune-hot tumor. This condition favorably influences other immunotherapies, including cancer vaccines, immune checkpoint inhibitors, bispecific T cell engagers, and adoptive cell therapy, thereby enhancing the efficacy of combined treatments. Created in BioRender. Schoeps, B. (2025) https://BioRender.com/d03w633.

In addition to this high molecular versatility, another key characteristic of OVs is the ability to predominantly replicate in tumor tissues, leading to a local amplification and propagation of the “living” drug and finally the killing of cancer cells [7]. Key element of the mode of action that enables local drug delivery is therefore viral replication, which is closely connected to cell lysis for most viruses. Importantly, this local effect might trigger a systemic tumor-specific immune response via the release of tumor antigens into a highly inflamed microenvironment [14]. However, the advantage of local drug amplification in tumors can only be realized if important prerequisites are met: cancer cells must be permissive to all successive steps, i.e. (i) OV uptake, (ii) OV replication and thus OV enhancement and (iii) OV-induced oncolysis. In this context, permissivity of tumors to OVs also requires a tight balance between activation of the anticancer immune system and antiviral immunity which is eventually limiting virus spread and therefore the OV-mediated anti-tumoral effect [14]. Thus, it is key that the immune activation against tumor antigens is sufficiently strong to assure tumor elimination beyond the phase of OV-mediated oncolysis. Indeed, there is growing interest to modify the type of cell lysis to even further enhance immunogenic cell death [15].

High drug attrition rates in cancer—learnings for oncolytic virotherapy

Already more than 10 years ago, it was highlighted that translation of pre-clinical findings into the clinics in the cancer field is very inefficient [16]: only 5% of drugs that show anticancer activity in pre-clinical development finally obtain approval, a dismal success rate as compared to 20% for cardiovascular diseases [17]. In the meantime, this has not changed much [18] since many clinical trials are still conducted with patients that are not appropriately selected for the intended treatment regime [19].

Of note, utilization of companion diagnostic biomarkers in the approval process of small-molecule therapies and immunotherapies varies strongly, with ~85% of small-molecule therapies and ~30% of immunotherapies being licensed in a biomarker-dependent manner [20, 21]. A potential explanation for this might be that while identification of predictive biomarkers linking mode of action of a therapy to clinical efficacy is straightforward in the context of targeted therapies, it is very challenging to pin down biomarkers for immunotherapy and combination therapies.

This is especially the case for oncolytic virotherapy, where the drug itself as well as the mode of action are much more intricate than in traditional cancer therapies like chemotherapy or targeted therapies. For instance, the extent to which tumor cell lysis is required for an OV therapy to be effective in patients still remains uncertain, because oncolysis can be driven by direct killing of permissive cancer cells by replicating OVs, as well as by activation of anticancer immunity by OV-driven mechanisms [9, 22, 23]. Thus, efficacy of OV therapy is determined by two pillars, (i) tumor permissivity and (ii) anticancer immunity (Fig. 2). Although, effective anticancer immunity can be induced by OV therapy also in non-permissive tumors [22, 24], OV replication in tumors potentiates anticancer immunity even against poorly immunogenic tumors [22], rendering tumor permissivity as one important determinant of OV efficacy. The intricate mode of action combined with the high complexity of TMEs, which are often characterized by intra- as well as inter-tumor heterogeneity, aggravate successful identification of biomarkers predicting OV activity (Fig. 2). Thus far, no predictive biomarkers of clinical benefit from OV action have been established [3], impeding precise patient selection in clinical trials. The recent failure of an unstratified phase III clinical trial investigating the effect of T-VEC plus pembrolizumab in advanced melanomas [13] further highlights the need for biomarker-driven patient stratification in OV therapies.

Fig. 2. Tumor permissivity and anticancer immunity: dual pillars of OV efficacy.

The clinical efficacy of OV therapy is determined by tumor permissivity, i.e. the replication of OVs within tumors, as well as the anticancer immunity induced by OV application. Although OVs can trigger anticancer immunity also in non-permissive tumors, replication and the associated tumor cell lysis in permissive tumors can further potentiate the induced anticancer immunity. Tumor heterogeneity and the associated differences in the tumor microenvironment are key factors that impact on permissivity to OV replication. The vast range of human tumor ecosystems, together with the intricacy of how OVs operate, poses a significant hurdle in identifying clinically relevant biomarkers of OV activity. Created in BioRender. Schoeps, B. (2025) https://BioRender.com/d03w633.

A deeper understanding of OV action in tumors is a pre-requisite for the development of more efficient OV therapies and will foster the identification of biomarkers predicting OV therapy response. Thus, it is essential to study the mode of action of OVs in conditions mimicking the complexity of TESs, which comprise not only tumor cells but also cells from the TME and acellular factors [25], as closely as possible. This does not necessarily need to involve animal studies, which anyhow are often poorly predictive for OV therapy response in patients [7]. In fact, the recent change in FDA regulations supports alternative model systems to study effectiveness and safety of drugs [26].

Tumor ecosystems—complex microenvironments determining OV action

TESs consist of an intriguing network of cellular as well as acellular compartments. The cellular compartment comprises cancer cells that are enmeshed in stromal cells of the TME, including cancer-associated fibroblasts (CAFs), endothelial cells, adipocytes and immune cells [27], whereas the acellular compartment consists of extracellular matrix and local metabolic prerequisites like oxygen and nutrients [25]. The finding that interactions between different components of the TME trigger cancer progression has inspired the idea of employing OVs to reprogram the TME, thereby enhancing the anti-tumor effects [28, 29]. However, it is still widely ignored that also the TME directly impacts on OV activity within tumors.

In this chapter we therefore provide a condensed overview on currently available data linking OV activity to components of TES, including cellular factors like tumor and immune cells, but also acellular factors like oxygen and the extracellular matrix.

Cellular factors—tumor cells

Tumor cells comprise the key target of oncolytic viruses. The selectivity of the different available OV platforms towards tumor cells is driven by various molecular alterations in cancer cells affecting different steps of the viral replication cycle (Fig. 3). The key innate antiviral defense mechanism of host cells is based on the interferon (IFN) pathway. IFN production by virus-infected cells triggers an antiviral state in surrounding cells, blocking virus replication and spread [30]. This effective antiviral defense pathway is compromised in tumor cells of several cancer entities due to various molecular alterations in cancer cells [30].

Fig. 3. Unraveling tumor permissivity to oncolytic viruses: the role of tumor cell-intrinsic factors.

Molecular changes in cancer cells alter the permissivity to OVs by impacting the steps of OV uptake, OV replication and OV-induced oncolysis. Upregulation of viral entry receptors and interference of oncogenic signaling with IFN signaling (highlighted in green) increase uptake and replication of OVs, respectively. Downregulation of IFN receptors, antiviral proteins and tumor suppressors like PTEN together with defects in IFN signaling and the recognition of viral components (highlighted in red) further support viral replication in cancer cells. The factors depicted in this figure might have only been implicated in the replication of certain OVs. Created in BioRender. Schoeps, B. (2025) https://BioRender.com/d03w633.

First, mutations in the anti-proliferative and pro-apoptotic IFN genes may lead to an impaired IFN production by cancer cells [30]. In fact, it has recently been shown that half of the most prominent homozygous deletions in cancer, which affect chromosome 9p21.3 and eliminate CDKN2A/B tumor suppressors, also encompass the type I IFN gene cluster, leading to immune evasion and metastasis [31]. In addition to the fact that several tumor cells are characterized by a reduced production of IFN, they might also show a reduced sensitivity to IFN. In this context it has been shown that the highly IFN-sensitive oncolytic VSV preferentially replicates in tumor cells that are resistant to IFN [12, 32]. This resistance of cancer cells to IFN treatment could be linked to downregulation of IFN receptor expression [32], a phenomenon that can be observed in various cancers [30].

Besides alterations in the IFN pathway on the level of surface receptors, it is known that also intracellular IFN signaling can be disturbed in cancer cells. This may on the one hand be mediated by epigenetic silencing of critical interferon-responsive transcription factors like IRF5 and IRF7 [33] or multiple interferon pathway genes during immortalization [34]. In addition, it has been shown that the most frequently inactivated tumor suppressor PTEN is involved in regulation of an antiviral response [35]. PTEN controls the negative phosphorylation of interferon-regulatory factor 3 (IRF3) directly via its phosphatase activity, therefore contributing to the expression of type I IFNs and the antiviral response [35]. Furthermore, the intracellular signaling pathways that are activated in cancer cells can interfere with the IFN pathway and therefore the expression of antiviral IFN target genes. For example, it has been shown that oncogenic activation of the RAS-ERK signaling pathway interferes with expression of multiple IFN-inducible genes [36], including the antiviral factor myxovirus resistance protein 1 (MxA) [37], finally enabling efficient replication of VSV [36, 37]. As activating RAS mutations are frequently found in several cancer entities [38], these findings might have important implications for oncolytic virotherapy. This is underlined by the discovery that the antiviral state, identified by immunohistochemical staining for MxA, segregates two molecular phenotypes of pancreatic cancer [39]. The molecular alterations in cancer cells can also directly affect antiviral proteins downstream of IFN signaling. For example, mutations in antiviral proteins including interferon α-2 have been found in tumors of breast cancer patients [40]. In addition to increased permissivity mediated by deregulation of IFN effectors, it has also been shown that alterations in tumor cells can lead to resistance to OV therapy. For example, high baseline expression of antiviral factors like MxA by some pancreatic cancer cells [41] and a constitutive activation of the IFN pathway in glioblastoma cells [42] renders them resistant to OV replication.

Besides alterations in the IFN pathway, other innate immune activation pathways can be impaired in human cancer cells. For example, loss of STING signaling in melanoma cells renders them permissive to viral oncolysis [43]. Furthermore, recognition of oncolytic adenovirus by the pattern recognition receptors Toll-like receptors (TLR) 2 and 9 has been found to be impaired in breast cancer-initiating cells [44].

Molecular changes in cancer cells also lead to an altered expression pattern of surface receptors, which can be exploited to target tumor cells. For example, CD155, which serves as receptor of poliovirus, is barely expressed in normal human tissues but overexpressed by tumors [45]. In addition, by engineering OVs to express specific viral surface proteins, a re-orientation of tropism and specific targeting of cancer cells is possible [46]. Based on differences in receptor utilization, for example, the fiber knob serotype of oncolytic adenoviruses can be switched to another serotype to enhance the targeting of CD46-expressing tumor cells [46, 47].

A so-far less well studied but also important molecular change in tumor cells that is linked to the activity of OVs is the rewired tumor metabolism [48, 49]. For example, glutamine metabolism, which is upregulated by several cancers [50], has been shown to be essential for replication of an oncolytic adenovirus [48]. On the other hand, defects in arginine metabolism in cancer have been shown to attenuate the replication of oncolytic myxoma virus [51], providing another example for resistance mechanisms towards OVs.

Cellular factors—stromal and immune cells

Although the sensitivity of cancer cells to OVs is a key factor in OV therapy, the in vitro response of cancer cells to OVs is often not linked to results obtained in vivo. This is particularly true for systems that do not allow a distinction between susceptibility to the initial infection and secondary virus infection resulting from subsequently produced virus progeny. In fact, it is increasingly recognized that also the TME plays a key role in OV activity (Fig. 4).

Fig. 4. Solving the puzzle of tumor permissivity to oncolytic viruses: the key role of the tumor microenvironment.

Several aspects of the TME impact on the permissivity of tumors to OVs. These include stromal cells and the immune compartment as well as acellular factors like chemical and physical factors. The impact of local pH on OV replication warrants further investigation. Created in BioRender. Schoeps, B. (2025) https://BioRender.com/d03w633.

In a seminal study it was shown that activated cancer-associated fibroblasts (CAFs), in contrast to normal fibroblasts, are characterized by reprogrammed antiviral networks, ultimately allowing replication of several clinically relevant OVs [52]. This reprogramming is driven by transforming growth factor-β (TGF-β), which triggers fibroblast growth factor 2 (FGF2) secretion by CAFs, finally enhancing sensitivity of cancer cells to OV [52]. In contrast to this positive effect of CAFs on OV replication in tumor cells, it has also been shown that the direct cell-cell contact between tumor cells and CAFs induces production of IFN-β1 by CAFs, which in turn interferes with replication of several OVs in cancer cells [53]. The expression of IFN-β1 is initiated following transcytosis of cytoplasm of cancer cells to CAFs and subsequent activation of IRF3 in CAFs [53]. Thus, stromal cells not only serve as a cell type allowing OV replication, but also control the replication of OVs in cancer cells. In addition to CAFs, it has been shown that VSV is able to infect the tumor vasculature, leading to clot formation and vascular collapse [54]. Furthermore, infection of non-tumor cells of the TME by an oncolytic rhino:poliovirus chimera has been shown to significantly drive efficacy of OV therapy [55]. Interestingly, not only stromal cells of the TME but also cells in distant organs allow replication of OVs and thereby are able to drive anti-tumor efficacy via activation of anti-tumor immunity [22].

Another important cellular component of the TME are immune cells, which play a key role in limiting OV replication and spread in tumors [9]. The frontrunner are macrophages which directly regulate the spread of OVs in tumors. On the one hand it has been shown that macrophages prevent lymph-borne neurotropic VSV from infecting the central nervous system [56] and suppress HSV-1 replication in gliomas by phagocytosing the virus and blocking subsequent virus replication [57]. In addition to physical elimination of OVs via phagocytosis, it has been shown that macrophages induce a protective antiviral state in ovarian and breast tumors, rendering them resistant to OV therapy [58]. This antiviral state in tumor cells was mediated by macrophage-derived IFN-β [58]. In addition to IFN-β, macrophages have been shown to inhibit viral replication by production of nitric oxide [59]. In a similar manner, tumor necrosis factor, a key cytokine produced by macrophages, has been shown to trigger an antiviral state in tumor cells [60].

In addition to macrophages, OVs have also been shown to activate dendritic cells (DCs), leading to IFN-γ production followed by activation of NK cells and T cells [61]. Natural killer (NK) cells in turn are known to kill HSV-1-infected glioma cells [62] and this antiviral NK cell response has been shown to limit glioblastoma virotherapy [63]. Such an indirect effect of immune cells on OV spread is also described for T cells which increase clearance of infectious oncolytic adenovirus in tumors [64]. The role of pre-existing humoral antiviral immunity on OV replication and efficacy is less clear and might depend on the employment of distinct OVs and on the specifics of the route of administration being used [65]. For example, it has been shown that pre-existing humoral immunity to an intratumorally applied oncolytic adenovirus [66] or HSV-1 [67] slightly affected transgene expression in tumors but did not abolish gene transfer, indicating that pre-existing antibodies against these OVs might only play a minor role for virus spreading in tumors. Interestingly, it has been shown that pre-existing immunity to oncolytic Newcastle Disease Virus (NDV) did not diminish its therapeutic efficacy but rather caused superior anti-tumor effects and prolonged survival of tumor-bearing mice [68]. Similar results were obtained with an oncolytic HSV-1, where enhanced anticancer efficacy was observed in immunized mice [69]. In sum, the balance between antiviral and anti-tumor immune response has to be regarded as a key factor determining the efficacy of OVs.

Acellular factors

In addition to these cellular factors, acellular physical or chemical factors play a critical role in tumor biology [70, 71] and also in the permissivity of tumors to OVs. Physical barriers that can impact on OV spread in tumors include the extracellular matrix (ECM) [72], calcification, necrosis and a high interstitial pressure [14]. For example, the spread of oncolytic HSV within tumors is limited by fibrillar collagen in the extracellular matrix [72] and can be facilitated by increased proteolytic activity and the accompanied increased hydraulic conductivity in tumors [73]. Also, chemical factors can contribute to the sensitivity of tumors to OVs. For example, it has been shown that hypoxia reduces the replication of adenoviral OVs [74, 75]. In addition, the metabolic microenvironment of tumors has been linked to OV activity, since replication of an oncolytic adenovirus has been shown to be dependent on exogenous glutamine [48]. So far there is limited data available on the stability of OVs in the TME. But since it is known that tumors show a lower extracellular pH [76] and pH is a critical factor determining the stability of OVs [77], it is likely that the activity of OVs can be compromised in TMEs. In sum, all these studies highlight the intricate interplay between several factors of TES in the regulation of OV activity.

Deciphering tumor permissivity: models and approaches centered on human tumor ecosystems

The prominent impact of several aspects of TES on OV replication highlights the need for models retaining these ecosystems to study OV action and finally allow efficient translation of findings into the clinical context. Murine model systems allow in vivo investigations and are key for the pre-clinical evaluation of many drugs. However, they bear several important limitations that might impair clinical translatability of the findings, especially in the context of OV therapy. These limitations include differences in viral cell tropism, the immune system, and tumor heterogeneity resulting in a poorly predictive value of anticancer immune responses observed in mice as compared to patients [3, 7]. This is particularly true for the commonly used subcutaneous tumor models, which exhibit a distinct TME compared to orthotopic tumors [78]. In addition, the differences in viral cell tropism preclude the use of many mouse tumor models to study the action of OVs designed for cancer patients [3]. While humanized mice offer a promising platform to test OVs that specifically target human tumor cells, they do come with several disadvantages. These include the considerable time and effort required to establish them, the limited development of mature immune cells such as macrophages, and the need for patient blood matching [3, 79]. Besides these limitations, the most important concern is the lack of the human TME and the resulting heterogeneity of human tumors [3].

Consequently, human tumor models hold a unique significance in the advancement of numerous OV concepts. Although they lack the systemic immune response of complex in vivo systems [79], they recapitulate the heterogeneity of human tumors allowing an in-depth analysis of the impact of the TME on permissivity. The fact that patient-specific anti-tumor immunity cannot be adequately addressed in pre-clinical studies further highlights the need to study tumor permissivity as second pillar of OV efficacy.

Here, we therefore focus on currently available model systems to mimic human tumors [25, 80–82], and discuss them with regard to the above-described factors influencing tumor permissivity to OVs.

The currently available human model systems, recently also termed patient tumor avatars [80, 81], can be roughly divided into two categories:

First, patient-derived viable tumor tissues can be used in their native structure, with intact extracellular matrix, cell-cell contacts and native cellular composition; examples for this approach include patient-derived organotypic slice cultures [83] and tumor fragments [84] which can be cultivated and have already been employed to test the activity of OVs [85–88]. Second, patient-derived tumors can be further processed to obtain organoids [89], 3D bioprinted cancer models [18] or micro- and nano-engineered tumors [25].

Each of these approaches has its unique advantages and disadvantages [25, 81, 82, 88] and the major difference between the mentioned approaches is the complexity and organization of the tumor microenvironment (TME) (Fig. 5). Hence, the strategy for investigating the permissivity of tumors to OVs should be tailored to the scientific question, the OV studied, and the aspects of the TME that are intended to be addressed in terms of permissivity.

Fig. 5. Comparison of available models to study human tumor ecosystems.

The radar chart provides a visual representation of the degree (low, medium, high) to which various aspects of tumor microenvironments are captured by the specified tumor model systems. Current cancer organoid models partially resemble the immune and stromal compartment but so-far only include a few cell types and lack an extracellular matrix (ECM)-like environment [18]. Tissue fragments/slices provide a high degree of native tissue recapitulation and retain the stromal and immune cell compartment. In contrast, 3D bioprinted tumor models allow the modulation of physical factors like elasticity, plasticity and mechanical properties of ECM [18]. Micro- and nanoengineering techniques can be utilized to conduct an in-depth study of chemical factors, which include nutrient and hypoxia gradients, along with physical factors like shear stress and traction forces [25]. Created in BioRender. Schoeps, B. (2025) https://BioRender.com/d03w633.

For example, two components of the TME that regulate tumor permissivity to OVs, namely stromal cells and the immune compartment, are retained in the native structure in the slice culture and tumor fragment approach, but these systems lack the capability to study defined chemical gradients in the tissue. These parameters could instead be addressed with micro- and nano-engineered tumors, which allow a detailed analysis of the impact of several chemical factors of the TME on tumor cell permissivity to OVs, including nutrient or hypoxia gradients, shear stress and traction forces [25]. Other key features of the TME that influence the replication of OVs such as ECM stiffness, degradability, as well as intact vascularization, can be explored using the emerging 3D bioprinted cancer models, which closely recapitulate the in vivo situation [18]. Indeed, such models have already been successfully employed to study OV action [90]. The tumor-on-a-chip approach further advances 3D cell culture technologies with microfluid technology to mimic the complexity of native organs [25]. This allows real-time assessment of physiological characteristics like fluidic flow and simplifies quantitative investigations of TME parameters like cell interactions [25], both important aspects for the spread of OVs in TES. Also, human tumor organoids have been employed to successfully test the activity of OVs [91, 92]. In the past years, the variety of different human cancer organoid models further evolved, ranging from organoids that retain tumor-infiltrating lymphocytes of the primary tumor [82, 89] to cancer assembloids mimicking the native tissue architecture, including CAFs, endothelial cells, immune cells and muscle layers [93].

Conclusion and future perspectives

Although substantial evidence has accumulated in recent years that the action of OVs is critically dependent on the composition of tumors, many studies examining OV activity continue to be conducted with models that do not recapitulate human TES. This approach hinders the effective translation of pre-clinical discoveries to clinical applications, particularly in the realm of oncolytic virotherapy, where also animal models offer limited translational value.

Thus, there is a growing need to recognize the importance of alternative human tumor model systems, a concept also backed by recent changes in FDA regulations [26]. Future research should therefore include comparative studies exploring the patterns of cancer permissivity to OVs in the different available human model systems.

Although permissivity of human tumors is a pre-requisite for viral replication and subsequent oncolysis triggering an anti-tumor immune response, it is likely that permissivity alone might be not conclusive enough to evaluate the therapeutic efficacy of OVs, which also relies on the tight balance between the OV-induced anti-tumoral and anti-viral activities [14]. Therefore, research connecting (i) the permissivity of human tumor avatars to OVs with (ii) the virotherapeutic efficacy being achieved in the respective cancer patients is needed to demonstrate the true clinical value of the aforementioned model systems. Considering the limited success rate observed in phase I clinical trials [17] and the lack of reliable predictive biomarkers for OV activity [3], one potential strategy to enhance clinical outcomes could be the selective enrichment of patients bearing permissive tumors. Such enrichment could be envisioned by permissivity testing of patient-derived tumor biopsies prior to the onset of patient treatment. Additionally, the utilization of ex vivo human tumor models may offer significant insights into the varying degrees of permissivity across different tumor types. Such knowledge could be leveraged to prioritize certain tumor types in future clinical investigations. In conclusion, human ex vivo tumor models are able to serve a dual purpose: they facilitate the identification of novel molecular and cellular factors driving permissivity of human TES to OVs, and they also hold the potential to significantly contribute to clinical decision-making processes.

Outstanding questions

• What are the molecular drivers of permissivity of human tumor ecosystems to OVs in the various cancer indications and in the specific tumor settings?

• Which human tumor model systems correlate best with the virotherapeutic efficacy being achieved in the respective cancer patients and can therefore be regarded as predictive?

• Can human tumor model systems provide valuable information for identification of biomarkers being indicative of a profound cancer permissivity to OVs?

• Are there different permissivity patterns to OVs when primary tumor tissues are compared with metastasized tissues, considering the fact that both exhibit different TMEs?

• What are the major differences between human tumor ecosystems obtained (i) from treatment-naive cancer patients versus (ii) from (heavily) pretreated cancer patients and how do these differences influence the permissivity of tumors to OVs?

• How does permissivity relate to anti-cancer immunity in patients, and does this relationship vary depending on the type of OV utilized?

Author contributions

BS conceptualized the manuscript and drew the figures. BS, UML, and KE wrote the manuscript.

Competing interests

BS and KE are employees of ViraTherapeutics GmbH and Boehringer Ingelheim International GmbH.