The emerging field of viroimmunotherapy for pediatric brain tumors

Abstract

Pediatric brain tumors are the most common solid tumors in children. Even to date, with the advances in multimodality therapeutic management, survival outcomes remain dismal in some types of tumors, such as pediatric-type diffuse high-grade gliomas or central nervous system embryonal tumors. Failure to understand the complex molecular heterogeneity and the elusive tumor and microenvironment interplay continues to undermine therapeutic efficacy. Developing a strategy that would improve survival for these fatal tumors remains unmet in pediatric neuro-oncology. Oncolytic viruses (OVs) are emerging as a feasible, safe, and promising therapy for brain tumors. The new paradigm in virotherapy implies that the direct cytopathic effect is followed, under certain circumstances, by an antitumor immune response responsible for the partial or complete debulking of the tumor mass. OVs alone or combined with other therapeutic modalities have been primarily used in adult neuro-oncology. A surge in encouraging preclinical studies in pediatric brain tumor models recently led to the clinical translation of OVs with encouraging results in these tumors. In this review, we summarize the different virotherapy tested in preclinical and clinical studies in pediatric brain tumors, and we discuss the limitations and future avenues necessary to improve the response of these tumors to this type of therapy.

Despite being rare diseases, tumors of the central nervous system (CNS) are the most common childhood cancers, with an annual incidence of nearly 6 cases per 100 000 population (age-adjusted), and are the leading cause of cancer-related morbidity and mortality.¹ In the past 2 decades, brain tumors have surpassed leukemia as the most common cause of cancer death in children.² Although the 5-year survival rates have improved to be higher than 80% for low-grade, non-metastatic tumors, the outcome for other tumors remains dismal. In addition, the standard of care only provides a short period of disease control and significant treatment-related toxicities, reaffirming the need to find novel, better-tolerated therapies with increased efficacy.^3,4

Oncolytic viral (OV) therapy is rapidly gaining importance in the field of cancer as a promising treatment strategy. OVs are naturally selected or engineered viruses (Figure 1), replicating preferentially in cancer cells and sparing normal healthy cells.⁵ In 2006, the Chinese Regulatory Agency approved the use of the oncolytic adenovirus H101 (Oncorine) for the treatment of head and neck squamous cell carcinoma (HNC).⁶ This achievement represented an important milestone in the rocky road of virotherapy, although the reception among the scientific community was controversial due to the lack of appropriate survival data. We needed to wait almost an extra decade until the US Food and Drug Administration approved 2015 an OV-based therapy for the treatment of patients with advanced melanoma after a durable response rate in a randomized, phase III trial using Talimogene laherparepvec (TVEC), a type I herpes simplex virus (HSV-1) that expresses granulocyte-macrophage colony-stimulating factor (GM-CSF) to increase tumor-antigen presentation.^7–9 TVEC was the first OV approved by the US agency, and the decision was rapidly followed by its European counterpart, the EMA (European Medicines Agency).

Figure 1.

Oncolytic viruses as unmodified or genetically modified agents. Representative scheme of the different oncolytic viruses discussed in this review. Some viruses have a natural tropism for cancer cells (left panel). Other viruses have been genetically modified to gain specific phenotypic characteristics (right panel; Ad, Adenovirus, HSV, Herpes Simplex Virus; MyV, Myxoma Virus; PSV: Poliovirus; SVV, Seneca Valley Virus; VSV, Vesicular Stomatitis Virus; VV, Vaccinia Virus).

The use of OVs in adult brain tumors is well entrenched in the clinical arena after several years of preclinical studies. Some of these clinical trials showed complete and partial responses with the presence of long-term survivors.^10,11 Clinical and preclinical studies in brain tumors have also concluded that OVs-mediated tumor lysis triggers a pro-inflammatory cascade that breaks the immunosuppressive environment and activates immune cells. This effect might modify immunologically “cold” tumors -poorly immune cell infiltrated- to “hot” tumors.^12,13 In addition, preclinical and clinical studies have shown that the administration of OVs activates an antitumoral immune response that is mainly responsible for the efficacy of this type of treatment. This immune mechanism of action of virotherapy has propelled the work of several laboratories to arm OVs with therapeutic transgenes, such as cytokines, tumor antigens, or costimulatory molecules, to augment antitumor immunity.

Of importance, the first OV (teserpaturev; G47∆; Delytact) received conditional and time-limited approval by the Japan Ministry of Health, Labor, and Welfare in 2022 for the treatment of adult patients with malignant glioma.¹⁴

While virotherapy in pediatric neuro-oncology is just emerging, a significant amount of preclinical data has been reported, and results from recent clinical trials are very encouraging. In this manuscript, we comment on the structural and immune-stimulating modifications of different OVs that have reached preclinical (Table 1) and/or clinical testing (Table 2) in pediatric brain tumors, as well as the results from representative studies (Figure 2); in addition, we will discuss aspects related to the crossroad between OVs and immune responses and the challenges and possible strategies for viroimmunotherapy for brain tumors in children.

Table 1.

Oncolytic Viruses for Pediatric Brain Tumors: Preclinical Studies

Virus	Name	Modification	Pediatric brain tumors	References
Herpes simplex virus 1	HSV1716	• Deletion of ICP34.5 gene	pHHG DIPG	15
	G207	• Deletion of ICP34.5 gene • Truncation of ICP6 gene	Medulloblastoma	16
	rRp450	• Deletion of ICP6 • Insertion of transgene encoding rat CYP2B1	Medulloblastoma, AT/RT	17
Adenovirus	Delta-24-RGD	• 24 bp deletion in the E1A region corresponding to the Rb binding site • Insertion of RGD-4C motif in the fiber knob	pHGG DIPG AT/RT CNS PNET ETMR	18, 19, 20
	ICOVIR-17K	• Inclusion of E2F responsive elements in E1A promoter • 24 bp deletion in the E1A region corresponding to the Rb binding site • Replacement of heparan sulfate proteoglycan binding motif KKTK by RDGK in fiber shaft • Insertion for expression of human PH20 hyaluronidase gene	CNS-PNET	21
	CRAd.Survivin.pK7	• Human survivin promoter • Polylysine modification (pK7) in the fiber knob	DIPG	22
Poliovirus	PVS-RIPO	• IRES region exchanged with a counterpart from rhinovirus type 2	Medulloblastoma PXA ATRT PNET	23
Measles Virus	Edmonston strain	• No modifications, Edmonston strain is naturally attenuated by mutations in HA gene	Medulloblastoma AT/RT	24, 25, 26
Reovirus	Reovirus	• No modifications	Medulloblastoma	27
Seneca Valley Virus	SVV-001	• No modifications	Medulloblastoma	28, 29
Myxoma		• No modifications	Medulloblastoma	30
Vaccinia	ddVV	• Deletion of thymidine kinase (TK) and vaccinia grow factor (VGF) genes	AT/RT	31, 32
Vesicular stomatitis virus	VSV-delta-M51	• Deletion of methionine 51 in the M protein region	AT/RT	33

Table 2.

Clinical Trials Using Oncolytic Viruses in Pediatric Brain Tumors

Oncolytic virus	Name	Tumor	NCT number	Phase	Additional therapies on study	No of patients, route (virus)	Status
Herpes simplex virus	HSV1716	Refractory or recurrent pHGG	NCT02031965	Phase I		2, intratumoral	Terminated prematurely
	G207	Recurrent pHGG	NCT02457845	Phase I	Radiation	12, intratumoral	Completed
		Recurrent pHGG	NCT04482933	Phase II	Radiation	40, intratumoral	Not yet recruiting
		Recurrent or refractory cerebellar brain tumors	NCT03911388	Phase I	Radiation	15, intratumoral	Active, recruiting
Adenovirus	Delta-24-RGD	DIPG	NCT03178032	Phase I	Radiation	12, intratumoral	Completed
Poliovirus	PVSRIPO	Grade III or IV malignant glioma	NCT03043391	Phase Ib		8, Intratumoral (CED)	Completed
Measles virus	MV-NIS	Local and disseminated Recurrent Medulloblastoma AT/RT	NCT02962167	Phase I		46, Intratumoral or in the subarachnoid space	Completed (not posted)
Reovirus	Reolysin	Refractory or recurrent pHGG	NCT02444546	Phase I	Sargramostim (GM-CSF)	6, intravenous	Completed

Figure 2.

Oncolytic viruses used in the pediatric brain tumor field. Oncolytic viruses have been tested in preclinical studies or clinical trials to treat pediatric brain tumors. (Ad, Adenovirus; HSV, Herpes Simplex Virus; MyV, Myxoma Virus; PSV, Poliovirus; SVV, Seneca Valley Virus; VSV, Vesicular Stomatitis Virus; VV, Vaccinia Virus).

Please note that although the WHO removed some nomenclature from the classification of pediatric CNS tumors in the 2016 edition, such as the umbrella term PNETs (primitive neuroectodermal tumors) and glioblastoma, it will use these terms when authors mention them in their manuscripts.³⁴

Herpes Simplex Virus

Herpes Simplex Virus (HSV) is an enveloped DNA virus with a natural neurotropic tropism that is among the most studied viruses as anticancer oncolytic agents. Different modifications of HSV have been explored to mediate tumor-specific replication of the virus, affecting nucleotide metabolism or the antiviral immune response.³⁵

HSV1716

The HSV1716 is a mutant HSV-1 virus that harbors a deletion in both copies of the neurovirulence factor ICP34.5 (RL1).^36,37 This gene encodes the protein γ₁34.5, which reverts protein translation shutoff triggered by the antiviral protein kinase RNA-activated (PKR).³⁸ Consequently, the replication of HSV1716 is severely attenuated in normal cells but not in tumor cells with an impaired PKR response. Clinical trials using HSV1716 in adult patients with different tumors, including CNS tumors, and in pediatric patients with no-CNS solid tumors demonstrated the safety of the virus.^39–41 Moreover, a therapeutic benefit was observed in a phase I trial for adults with glioblastoma multiforme.⁴¹

Preclinical experiments in pHGG and DIPG models showed that HSV1716 efficiently undermines the migration and invasion properties of these tumor cell lines. HSV1716 treatment resulted in the decrease of the invasive pattern of an orthotopic xenograft DIPG model, indicating that the OV could inhibit the typical infiltrative growing pattern of these types of tumors.¹⁵

In the clinic, a phase I dose-escalation trial was initiated for pediatric patients with refractory or recurrent high-grade glioma (NCT02031965) consisting of the intratumoral or peritumoral injection of HSV1716. The primary objective was to determine the maximum tolerated dose and dose-limiting toxicities, as well as characterization of the therapeutic effect and antitumor immune response as secondary objectives, among others. However, this trial was prematurely terminated after treating 2 patients, and results have not been published yet.

G207

Similarly to HSV1716, G207 is also an HSV-1 containing a deletion in the ICP34.5 genes, but replication of this virus was further attenuated by insertion of the Escherichia coli lacZ gene to the viral gene ICP6 (UL39), which disrupts the expression of the viral ribonucleotide reductase.⁴² The safety of G207 has been demonstrated in the HSV-sensitive New World owl monkey (Aotus nacymai) after administering up to 10⁹ PFU of the engineered virus, while only 10³ PFU of the wild-type HSV caused a fulminant encephalitis.⁴³ Of relevance, HSV G207 has been reported to efficiently infect CD111+/CD133+ glioma/cancer stem cells (GSCs/CSCs).^16,44 This receptor is expressed in pediatric brain tumors such as medulloblastoma, which renders these tumors sensitive to G207-mediated cell death and increases survival of group 3 medulloblastoma in preclinical models.¹⁶

The preclinical results and the safety profile of G207 administration bolstered the translation of G207 to the clinic for pediatric patients (NCT02457845). In 2021, Friedman et al. reported the results from a clinical trial using virotherapy for pediatric brain tumors.⁴⁵ In this phase I clinical trial, 12 patients (7–18 years old) affected by supratentorial high-grade glioma were enrolled to receive G207 HSV upon recurrence. The tumors were identified as glioblastomas (n = 10), anaplastic astrocytoma (n = 1), and pHGG (n = 1). The treatment with OV was administered as single therapy (10⁷ and 10⁸ pfu) or in combination with a single 5-Gy radiation dose 24 hours later. No serious adverse or dose-limiting toxic effects were observed. Radiographic, neuropathological, or clinical responses were seen in 11 patients, and the authors reported an OS of 12.2 months (95% CI: 8.0–16.4). Four out of eleven patients were alive 18 months after the treatment. Moreover, compared to pretreatment, an increase in immune infiltration was observed between 2 and 9 months after G207 injection.

Following the safety data of this trial and the encouraging results, a phase II clinical trial has been opened to evaluate the efficacy of 10⁸ PFU G207 HSV combined radiotherapy (5 Gy) in 40 pediatric patients of high-grade glioma (NCT04482933). In addition, a new phase I dose-escalation trial is evaluating the safety dose of G207 plus radiotherapy in children with recurrent/refractory cerebellar brain tumors following a 3 + 3 design.

rRp450

rRp450 is an HSV OV with deletion of the viral-encoded ribonucleotide reductase (ICP6), conferring selective replication in rapidly dividing cells. However, unlike G207, in rRp450, ICP6 was inactivated by the insertion of an expression cassette for the rat cytochrome P450 2B1 (CYP2B1) and maintains the ICP34.5 neurovirulence genes and ICP6.⁴⁶ The expression of CYP2B1 in infected tumor cells allows the activation of the prodrug cyclophosphamide (CPA) into the chemotherapeutically active 4-hydroxy cyclophosphamide (4-OH-CPA) to potentiate the antitumor effect.⁴⁷ Treatment of human pediatric embryonal brain tumors group 3/4 medulloblastoma and AT/RT cell lines with rRp450 showed that these cells are susceptible and permissive to virus infection and replication.¹⁷ rRp450 prolonged survival when administered intracranially into orthotopic xenografts models of medulloblastoma and AT/RT, leading to long-term survivors.

Adenovirus

Adenoviruses are nonenveloped double-stranded DNA viruses that belong to the Adenoviridae family.⁴⁸ The most widely used adenoviruses for therapeutic purposes are based on serotype 5 of the human adenovirus species C (HAdV-C5 or Ad5) due to their safety and suitability for their manufacture under GMP (Good Manufacturing Procedure conditions). The backbone of Ad5 can be easily genetically modified to confer these viruses a conditional replication restricted to malignant cells while being unable to propagate in non-tumor cells. Many strategies can be implemented to achieve such tumor selectivity, and in this review, we will focus on discussing the most relevant in pediatric brain tumors.

Delta-24-RGD

In quiescent cells, the cell cycle progression E2F transcription factor is sequestered by the tumor suppressor protein retinoblastoma (Rb) to keep the cell in cycle arrest.⁴⁹ Because the natural targets of Ad5 are quiescent epithelial cells, an important hallmark during adenovirus cycle progression is the expression of the E1A very early gene product. E1A binds to Rb to release E2F, which, in turn, activates the transcription of the adenoviral E2 genes encoding different viral proteins involved in the adenovirus genome replication.^50,51 One of the most widely used strategies to create tumor selectivity is based on this mechanism. Delta-24 is a modified adenovirus containing a 24-bp deletion in the Rb binding site in the adenoviral E1A that prevents the release of E2F in quiescent infected cells, thus halting the adenovirus replication in normal cells.^52,53 However, a differential feature typically found in tumor cells is cell cycle dysregulation by either genetic or epigenetic alterations of the Rb pathway, resulting in a persistent E2F activation. Therefore, unlike normal cells, the tumor context is permissive for Delta-24 viral cycle progression and replication, thus leading to tumor cell death. Finally, because many tumor cells do not express the Coxsackie and Adenovirus Receptor (CAR), the natural Ad5 primary receptor, an RGD motif (RGD-4C) was introduced in the adenovirus fiber knob domain to generate Delta-24-RGD (DNX-2401 or tasadenoturev), thus expanding the tropism to CAR-negative tumors through RGD-binding α_vβ₃ and α_vβ₅ integrins.⁵⁴ A first-in-human phase I clinical trial tested the use of Delta-24-RGD in adult patients with recurrent malignant glioma (NCT00805376), showing a safe profile as well as complete and partial responses that result in long-term survivors.¹¹

In the preclinical field, our group described the efficient anticancer effect of Delta-24-RGD in high-grade pediatric glioma (pHGG) and diffuse intrinsic pontine gliomas (DIPG).¹⁸ Using a frameless, guide-screw system,^55,56 Delta-24-RGD intratumoral treatment resulted in prolonged survival and the presence of long-term survivors in human xenografts and mouse immunocompetent DIPG models. In human cells in an immunodeficient background, the therapeutic effect relies almost exclusively on the oncolytic properties of the virus. On the other hand, the experiments in murine DIPG tumors (immunocompetent context) demonstrate the necessity of the immune response elicited by the virus to mediate tumor control and antitumor memory since human adenovirus does not efficiently replicate in murine cells.¹⁸ These experiments demonstrate the role of the dual mechanism of action of OVs, oncolysis, and immune activation.

Encouraging preclinical results using Delta-24-RGD to treat DIPG tumors, alone or in combination with radiotherapy^18,19 constituted the basis for the generation of a phase I trial for patients with newly diagnosed DIPG (NCT03178032).⁵⁷ The dose-escalation study enrolled 12 patients (3–18 years old) who, after confirmatory biopsy, were treated with a single intratumoral injection of Delta-24-RGD at a dose of 10¹⁰ virus particles (v.p.; n = 4) or 5 × 10¹⁰ v.p. (n = 8), followed with standard radiotherapy in 11 of these 12 subjects. The treatment was well tolerated, with asthenia, headache, vomiting, and pyrexia being the most reported adverse events. No grade 4 or 5 adverse events were observed. The investigators reported a reduction in tumor size in 9 patients, partial response in 3 patients, and stable disease in 8 patients. The median overall survival was 17.8 months, and 3 patients remained alive at 19.6, 31.4, and 33.5 months at the time of the analysis.

Treatment with Delta-24-RGD also increased survival in preclinical AT/RT and CNS-PNET human xenograft mouse models.²⁰ The survival benefit was observed both at the early stages of the disease and with established tumor masses. Of interest, Delta-24-RGD treatment of a murine model of disseminated AT/RT resulted in a reduced rate of spinal dissemination of the disease and prolonged survival. Due to the scarcity of mouse CNS-PNET and AT/RT, the immune response was evaluated using humanized mice, where a human immune system is reconstituted by infusion of hCD34+hematopoietic cells. Although these mouse models had several limitations, like the lack of HLA matching between tumor cells and hCD34 donors (among others), the results demonstrate that Delta-24-RGD can also promote an inflammatory response and T-cell infiltration in these tumors. These data underscore the potential effect of this treatment in disseminated AT/RT tumors affecting very young children.

ICOVIR17K

ICOVIR17K (or VCN-01) is an upgrade of the Ad5-based OVs derived from the ICOVIR platform.⁵⁸ Similarly to Delta-24-RGD, ICOVIR17K harbors genetic modifications that limit its replication to cells with an impaired Rb pathway, as well as the incorporation of an RGD motif into the fiber. In this virus, Rb selectivity is achieved by a combination of the 24-bp deletion in the E1A Rb binding site along with the insertion of E2F responsive elements in the E1 promoter to enhance replication selectivity in tumor cells.⁵⁹ Unlike Delta-24-RGD, the ICOVIR17K RGD motif (RGDK) is found in the fiber shaft domain, replacing the putative heparan-sulfate proteoglycan (HSPG) binding motif KKTK to facilitate infection through integrins while reducing hepatic uptake of the virus. Finally, an expression cassette encoding the human hyaluronidase PH20 was incorporated to enhance viral spreading through the extracellular matrix. ICOVIR17K has been tested in the clinic for extracranial adult tumors like HNC (NCT03799744) or pancreatic ductal adenocarcinoma (PDAC),⁶⁰ and pediatric retinoblastoma.⁶¹

In the context of pediatric neuro-oncology, the work published using ICOVIR17K consists of its preclinical evaluation for the treatment of CNS-PNETs.²¹ Orthotopic human murine models of this pediatric tumor showed prolonged survival after receiving a single dose of VCN-01. The virus was able to replicate in vivo, as well as to spread within the tumor following hyaluronic acid degradation. The in vivo experiments were carried out in the immunodeficient athymic nude mice to allow engraftment of human tumor cells, which precludes the evaluation of the adaptive immune response. Nonetheless, this mouse strain still produces myeloid immune cells, and the authors reported increased recruitment of reactive macrophages/microglia in the tumor after the viral treatment.²¹

CRAd-S-pK7

The tumor-specific killing properties of the oncolytic adenoviruses described above are based on the differences between tumor and normal cells in cellular key components for adenovirus cycle progression. An alternative strategy to generate tumor-selective adenoviruses is to replace the E1A promoter with a tumor-specific promoter. Because in adenoviruses, E1 expression is the initial event that triggers the cascade leading to virus replication, E1 regulation by tumor-specific promoters hinders adenovirus replication in non-tumor cells. Among these viruses, we can find the CRAd-Survivin-pK7 (CRAd-S-pK7).⁶² This OV construct contains, in the adenovirus serotype 5 backbone, the human survivin promoter and the incorporation of a polylysine (pK7) peptide into the C-terminal end of the fiber knob domain. The human survivin promoter, which is highly expressed in tumor cells but not in normal cells, drives the expression of the viral genes, whereas the second modification allows the adenovirus binding to heparan sulfate proteoglycans, facilitating the entry of the OV into CAR-negative tumor cells.⁶² Recent studies have reported the use of this recombinant adenovirus in preclinical models of DIPG using mesenchymal stem cells (MSCs) as vehicles to intranasally deliver CRAd-S-pK7 in DIPG-bearing immunodeficient mice. The studies showed an increase in survival when CRAd.Survivin.pK7-loaded MSCs were administered in combination with radiotherapy.²²

Poliovirus Chimeras (PVSRIPO)

Poliovirus is a nonenveloped RNA virus that belongs to the Picornaviridae family with a tropism for motor neurons that produces neurologic complications in its wild-type form. Genetic modification of the poliovirus genome has brought chimeric viruses (PVSRIPO or lerapolturev) with attenuated neurovirulence. In particular, the replacement of the neuron-specific internal ribosome entry site (IRES) found at the 5ʹ end of the poliovirus genome by the human rhinovirus 2 (HRV2) IRES.⁶³

Poliovirus tropism is determined by the presence of the cell adhesion molecule CD155, also known as Necl5 or poliovirus receptor (PVR), which overexpresses different malignant cells,⁶⁴ including adult and pediatric brain tumors.^23,65 In addition, Thompson EM et al. confirmed the high expression levels of CD155 in medulloblastoma, Pleomorphic Xanthoastrocytoma (PXA), AT/RT, and PNET samples rendered these cells sensitive to PVSRIPO infection, leading to tumor cell death.²³

The safety of this OV was confirmed in nonhuman primates⁶⁶ and in a phase 1b clinical trial for adult patients with recurrent glioblastoma.¹⁰

Recently, the results from a phase 1b clinical trial using PVSRIPO for pediatric recurrent high-grade glioma (NCT03043391) have been published.⁶⁷ Eight patients received 5 × 10e7 viral particles of the oncolytic virus via convection-enhanced delivery. The authors reported up to 26 AEs related to the virus. Although it is complicated to extract conclusions for efficacy from a phase I trial designed to evaluate toxicity in a few patients, the authors described that one patient remained alive after 22 months, and the overall median survival was 4.1 months after treatment.

Measles Virus

Measles virus (MV) is an enveloped (−) RNA virus from the Paramyxoviridae family that causes infected-cell death by the formation of non-viable syncytia. The tropism of MV is determined by the hemagglutinin (HA) protein found on its envelope that interacts with the target receptors, which in wild-type strains are the signaling lymphocyte activation molecule family member 1 (SLAMF1/CD150) and the Membrane Cofactor Protein (CD46).⁶⁸ The attenuated Edmonston strain used for MV vaccination contains mutations in HA that impede its interaction with SLAM, thus forcing the virus to infect through CD46.^69,70 CD46 is a protein expressed in the cell membrane of most human cells that protects the cell from the attack of complement in non-pathological conditions. Due to these modifications, the Edmonston MV requires high-expression levels of CD46 for efficient infection,⁷¹ thus conferring this virus a preferential tumor tropism since CD46 is often overexpressed in many cancer cells. The use of this vaccine strain during several decades has demonstrated a solid safety profile, which was further proven in clinical trials for adult patients with solid tumors, including glioblastomas.⁷²

Regarding pediatric cancer, an initial study showed the expression of CD46 in human medulloblastoma cell lines and clinical specimens.²⁴ In addition, the authors demonstrate the therapeutic effect of MV in vivo in mouse medulloblastoma xenograft by showing prolonged survival in animals receiving 5 intratumor administrations of the oncolytic virus (1 × 10⁶ TCID₅₀, every other day) and a decreased tumor burden.²⁴ Interestingly, measles treatment also improved the survival outcome of mice disseminating disease into intracranial and spinal subarachnoid spaces.²⁵

Measles was also tested in AT/RT preclinical models.²⁶ An in vivo experiment using 2 xenograft AT/RT models in athymic nude showed prolonged survival in mice treated intratumorally with 5 × 10⁵ PFU compared to those animals treated with UV-inactivated virus and also resulted in a therapeutic benefit for animals affected by disseminated disease. However, no effect was observed by intravenous administration of the virus. The preclinical data obtained in those models have led to a phase I clinical trial for local or disseminated recurrent medulloblastoma and AT/RT patients, which is currently recruiting (NCT02962167). In this trial, the MV has been genetically modified to incorporate the sodium iodide symporter (NIS) reporter to facilitate the follow-up of infected cells. The virus is administered either directly in the tumor bed for local lesions or in the subarachnoid space for disseminated disease (lumbar puncture). The trial is still active, but no results have been posted yet.

Reovirus

Reovirus is a double-stranded RNA virus that typically causes mild respiratory and enteric infections.⁷³ Among the different reovirus species, the mammalian orthoreovirus type 3 Dearing strain (reovirus T3D) has shown tumor selectiveness and is the reovirus commonly referred to as the oncolytic reovirus (Reolysin). Some studies have identified that Ras-signaling pathway hyperactivation, which usually occurs in many cancers, might be a requirement for reovirus-mediated cytotoxicity.^74,75 However, reovirus replication has also been reported in the absence of Ras activation.⁷⁶ Therefore, the mechanism underlying the tumor-specific replication of this reovirus is not completely clear.

In a preclinical study using an orthotopic immunodeficient mice model using SHH-medulloblastoma models, multiple administrations of the reovirus resulted in prolonged survival (nearly 1.5X median OS compared to control groups) as well as in a dramatic reduction in spinal and leptomeningeal metastatic dissemination.²⁷

In pediatric patients, a phase I trial in extracranial tumors (NCT01240538) tested the toxicity and clearance of Reolysin (monotherapy or combined with cyclophosphamide) after intravenous administration of the virus for 5 consecutive days in a 28-day cycle.⁷⁷ Overall, the treatment was well tolerated, although 2 patients experienced grade 5 AEs not related to the virus according to the investigator criteria. Then, a new phase I clinical trial (NCT02444546) was opened for pediatric patients with refractory or recurrent high-grade brain tumors using Reolysin (pelareorep in the clinic) in combination with GM-CSF (Sargramostim), a GM-CSF.⁷⁸ The rationale of the combination was to facilitate the mobilization of PBMCs, which served as virus carriers in the adults’ trial.⁷⁹ A total of 6 patients (2 DIPG, 3 GBM, and 1 MB) were enrolled in this trial and received. An initial cohort of 3 patients received dose level 1 of 5 × 10⁸ TCID₅₀/kg (intravenously). After 2 AEs (grades 3 and 4), a de-escalation to dose 0 was applied for 3 additional patients (3 × 10⁸ TCID₅₀/kg). Unfortunately, DLT were observed in the sixth patient, and the study was closed. No complete or partial responses were observed, and all patients experienced progression within 60 days.

Other Viruses

The viruses described above are among the most explored for the treatment of pediatric tumors in preclinical and clinical studies. Nonetheless, there is a plethora of other OVs that are showing promising results in preclinical models. The Seneca-valley virus is another picornavirus with the interesting feature of crossing the blood-brain barrier after intravenous administration that has been tested in pHGG and medulloblastoma models.^28,29 The Myxoma virus (MyV) belongs to the Poxviridae family. It causes lethal infections in the European rabbit (Oryctolagus cuniculus) and mild infections in other rabbits but is nonpathogenic in non-lagomorph vertebrates.⁸⁰ MyV has been tested in preclinical medulloblastoma models.³⁰ The Vaccinia virus (VV) is also a poxvirus, and a double-deleted modified version (ddVV) that has been evaluated as anticancer agent in pediatric tumors such as AT/RT.^31,32 The oncolytic properties of a modified vesicular stomatitis virus have been applied to AT/RT preclinical models.³³

Future Directions and Current Limitations

Current data from preclinical and clinical studies demonstrate that the mechanism of action of OVs is not based solely on their cytopathic effect but also on their intrinsic capacity to trigger antitumor immunity.^9,13 Since the OV particles will eventually be cleared, the immunostimulatory arm of OVs is essential to achieve effective, long-lasting responses. The engagement of antitumor immunity facilitates the destruction of non-infected tumor cells within the tumor mass, the invasive tumoral front, and the remote tumor sites by an abscopal effect. In addition, establishing anti-tumoral immune memory would prevent or delay recurrence (Figure 3). Due to this mechanistic immune component, oncolytic virotherapy is currently considered a legitimate class of immunotherapy termed Viroimmunotherapy.

Figure 3.

Mechanisms of local and systemic tumor-specific immune responses during viroimmunotherapy for pediatric brain tumors. Oncolytic viruses are administered to pediatric brain tumors, where they selectively replicate within tumor cells, leading to direct tumor cell lysis. This tumor-specific cell killing triggers the release of cytokines, pathogen-associated molecular patterns (PAMPs), damage-associated molecular patterns (DAMPs). Tumor-associated antigens (TAAs) are captured by professional antigen-presenting cells and presented to CD4+ and CD8+ T cells, initiating local and systemic tumor-specific immune responses. Locally, at the primary tumor site, these immune responses might induce inflammation, inhibit tumor cell invasiveness, exert cytotoxic effects, and prevent tumor relapse. Systemically, at disseminated tumor sites, circulating CD8+ T cells may recognize TAAs and mediate cytotoxic effects, contributing to the control of metastatic disease.

To further increase the anti-tumoral immune effect of virotherapy, several working groups are incorporating immunomodulators in the backbone of OVs. For instance, a variant of G207 (M002) incorporates the IL12 to activate natural killer cells and increase type 1 helper T cells (Th1) responses, eliciting a stronger antitumor effect than the parental OV.⁸¹ Delta-24 derivatives have been generated encoding co-stimulatory members of the tumor necrosis factor receptor family,^82–86 facilitating the activation of the co-stimulatory signal during antigen presentation and, in turn, generating T-cell activation and preventing T-cell anergy. Delta-24-ACT, a modified Delta-24-RGD encoding the T-cell activator 4-1BBL, showed therapeutic potential in DIPG murine models.⁸⁶ Furthermore, Delta-24-RGDOX (DNX-2440), another Delta-24 derivative expressing the T-cell activator OX40L, arrived at the clinic in a phase I trial for the treatment of adult patients with recurrent high-grade gliomas (NCT03714334). Nevertheless, clinical translation of these candidates is required to evaluate their efficacy and safety in pediatric brain tumors.

Another unmet need is a complete understanding of the OV-induced antitumor immune response mechanisms in the human context and, more specifically, in the pediatric scenario. A better comprehension of the immune system in the developing brain of children would allow designing OVs to precisely trigger antitumor immune responses to each of these stages with its intrinsic immune peculiarities. Additionally, different types of pediatric brain tumors display different mutanome and microenvironment dynamics. DIPGs are characterized by a “cold” or ‘deserted” tumor microenvironment (TME) with scarce T-cell infiltration.^87,88 In these neoplasms, macrophages/microglia are the most represented population in the TME; however, unlike what occurs in pediatric and adult HGG, M2 macrophages are not particularly abundant.⁸⁸ On the contrary, T-cell infiltration in TYR and MYC AT/RT subtypes is comparable to that observed in immunogenic adult tumors despite their low tumor mutational burden,⁸⁹ which is possibly related to endogenous retrovirus (ERV) re-expression. In addition, these tumors are vastly populated by anti-inflammatory macrophages,⁸⁹ which negatively correlates with survival in AT/RT.⁹⁰ In summary, when designing viroimmunotherapy for pediatric brain tumors, investigators should consider that these tumors consist of a heterogeneous group of diseases with distinctive aberrant genomic and TME characteristics.

To tailor the next viroimmunotherapy agents, it is also necessary to understand the interplay between OVs and the tumor intrinsic and TME responses to this type of therapy (Figure 4). The surveillance sensors of the immune system include identifying pathogen-associated molecular patterns (PAMPs) that are typically found in viruses. Toll-like receptors (TLRs) are a class of receptors that detect a broad range of PAMPs, such as virus proteins (TLR2 and TLR4) or nucleic acids (TLR7/8 and TLR9), starting a cell signaling that converges in the NF-κB pathway and ultimately triggering an antiviral response.⁹¹ In dendritic cells (DCs), TLR stimulation facilitates their activation and maturation by upregulating the secretion of type-I IFNs and the expression of membrane costimulatory molecules CD80/86 and CD40.⁹² The presence of viral DNA in the cell cytosol activates the cyclic GMP-AMP synthase (cGAS) that produces the secondary messenger cyclic GMP-AMP (2ʹ3ʹ-cGAMP). In turn, cGAMP triggers the activation of the Stimulator of IFN genes (STING), inducing the production of type-I IFNs.⁹³ In addition to the generation of PAMPs, the infection of tumor cells by the OV results in the generation of damage-associated molecular patterns, including the secretion of high mobility group box 1, translocation to the ER proteins hsp90 and calreticulin cell membrane, and the release of ATP and mitochondrial DNA from damaged cells.^11,20 These danger signals are detected by different mechanisms like TLRs and purinergic receptors in the innate immune cells (ie, DCs) to trigger the inflammatory response. Furthermore, the virus-induced tumor cell death might mediate tumor-associated antigens spreading. Overall, the antiviral inflammatory response described above generates a pro-inflammatory environment that facilitates an anti-tumor immune cell infiltration.

Figure 4.

Challenges and potential strategies to improve viroimmunotherapy for pediatric brain tumors. Described are some of the current known challenges for viroimmunotherapy for pediatric brain tumors and possible solutions. preexisting Neutralizing Antibodies (Abs): Recognition of oncolytic virus (OV) by preexisting Abs can be circumvented through the design of stealth viruses or utilizing cell carriers for delivery. Low basal tumor-infiltrating lymphocytes (TILs) frequency: OV enhances the frequency of TILs, which can be further augmented by combining with strategies such as arming viruses with T cell activators and concurrent treatment with immune checkpoint inhibitors (ICIs) or chimeric antigen receptor (CAR) T cells. Viral dominance: utilizing OVs to express specific HLA genes and tumor-associated antigens (TAA) might redirect the adaptive immunity from being antiviral to becoming mainly anti-tumoral. Virus-induced innate immunity: combination with NK cells or enhancing the activity of myeloid cells or modulation of the STING pathway might contribute to enhancing viroimmunotherapy. Tumor-intrinsic resistance: Enhanced tumor-tropism of oncolytic viruses to increase the host range, modulate cell autophagy, or unmask the expression of TAAs might overcome tumor-intrinsic resistance.

The infection of a tumor reshapes the tumor microenvironment and transforms a “cold” tumor into a “hot” neoplasm. Viral infection increased CD4+ and CD8+ T cell infiltration, as well as myeloid populations like DCs and macrophages in high-grade gliomas.⁸⁴ Paradoxically, oncolytic virotherapy also induces compensatory anti-inflammatory mechanisms that might limit their therapeutic effect. For instance, infection is followed by activation of the IDO/Kyn/AhR immunosuppressive cascade⁹⁴ and the upregulation of PD-1 in the TME.⁹⁵ Therefore, it is crucial to understand the interaction network between immune system actors and OVs in the context of brain tumors to design synergistic combination therapies. For example, the STING gene is constitutively silenced in tumor cells but not in the immune compartment in GBM,⁹⁶ which suggests the feasibility of combination therapies based on STING rescue to potentiate immunotherapy. In the context of virotherapy, STING rescue in tumor cells using methylase inhibitors such as decitabine could activate innate immune pathways and synergize with the OVs. On the other side of the coin, overactivation of STING could negatively impact the replication of the virus.

Clinical trial design incorporating window-of-opportunity components or matched biopsies might be essential to delineate the basal level of TME of pediatric brain tumors and characterize the reshaping of the TME by OVs. This is evident in recent clinical studies, such as the phase I trial using HSV-1 G207 to treat pediatric patients with recurrent high-grade glioma (NCT02457845); the injection of the OV resulted in increased CD4 and CD8 T-cell infiltration compared to pretreatment levels. Of note, T-cell infiltrates were found in areas adjacent and distal to the injection site, and they remained detectable once viral replication was no longer identified.⁴⁵ In the phase I clinical trial using Delta-24-RGD followed by standard radiotherapy to treat naïve DIPG patients (NCT03178032), pretreatment TME was characterized by a predominant myeloid population and a low frequency of CD4 and CD8 T-cells.⁵⁷ In this trial, using a longitudinal study was feasible based on the pretreatment biopsy, post-treatment biopsy at relapse, and tumor tissue from the autopsy of a single patient. This longitudinal analysis showed that myeloid cells were the predominant population before infection, that T-cells became the dominant cell population at the moment of the relapse, and that myeloid cells recovered their predominance in the autopsy specimen.

Another relevant issue for viroimmunotherapy is the route of administration of OVs. It is known that the presence of neutralizing or antiviral IgGs in the serum interferes with the effectiveness of this therapy.^45,57,97 The intratumoral administration of the virus has the advantages of resulting in a higher viral input in the tumor and the avoidance of systemic toxicity. However, the location of pediatric brain tumors, particularly those in the brainstem close to vital structure, might make this surgical strategy technically challenged. Therefore, investigators are examining the systemic delivery of viruses using cell carriers, such as neural stem cells⁹⁸ or human bone marrow-derived MSCs,^22,99 or enveloping the virus with nanoparticles.¹⁰⁰

A critical issue in the viroimmunotherapy field is the high immunogenicity of the viruses. This immunogenic capacity allows the host to rapidly erase the pathogen via the innate and adaptive immune sensors, preventing the generation of robust anti-tumoral immunity.¹⁰⁰ Elucidating whether the generated immune response after OVs administration is exclusively built against the virus or against both the virus and the tumor is the focus of several groups in viroimmunotherapy.

The use of preclinical models that allow the accurate evaluation of the dual mechanism of action of these agents, direct oncolysis and anti-tumoral immunity, is of utmost importance. Although several murine models for pediatric brain tumors have been developed during the last years, there is still a paucity of models for many tumor entities. To study the immune mechanism arm of OVs, the use of humanized mice might be a feasible option since they permit the utilization of human cell lines in the presence of a competent immune system.²⁰ However, these models are time-consuming and expensive, and they have limitations when it comes to studying the mechanism of action underlying the effectiveness of immune therapeutic approaches.

The use of orthotopic pediatric brain tumors is highly desirable since subcutaneous tumors do not properly recapitulate the tumor microenvironment. In fact, even different locations within the brain entail different immune characteristics.¹⁰¹ Direct tumor implantation on the specific brain region⁵⁵ or in situ spontaneous tumor development by genetic manipulation are being explored.^102,103 These models are probably the most accurate system; however, both have their pros and cons, whose discussion is out of the scope of this review. Another limitation results from the different interspecies sensitivity of OVs. For instance, most cells of mouse origin are non-permissive for replicating human adenovirus type 5, limiting comprehensive preclinical studies.¹⁰⁴ Other immunocompetent animal models, such as Syrian hamsters that are permissive for adenoviral replication, have been developed for adult gliomas¹⁰⁵ but have not been yet characterized for pediatric brain tumors.

Conclusions

Treatment options for pediatric brain tumors are still limited, especially at recurrence. Moreover, current therapies are very aggressive affecting not only the tumor but also the healthy tissue causing severe secondary effects that ultimately impinge on the quality of life of children and adolescents suffering from brain tumors. The safe profile shown by OVs in preclinical and clinical trials in adult and pediatric patients encourages further development of this therapeutic approach. Strategies that improve the eliciting of antitumor immune responses should be investigated to better tailor this therapy to pediatric brain tumors. In summary, the reassuring safety profile and encouraging results from preclinical and clinical studies with OVs are solid stepping stones to further developing viroimmunotherapy as a treatment for young patients with CNS tumors.

Funding

This work was supported by the Chance for Life Foundation, The Cure Starts Now Foundation, The Alliance for Cancer Gene Therapy, the National Institutes of Health (NIH)/National Cancer Institute (NCI; R01CA256006, P50CA127001), the American Legion Auxiliary Foundation (ALA), The ChadTough Defeat DIPG Foundation, the Spanish Association for Cancer Research (AECC), ChadTough Defeat DIPG Foundation, Government of Navarra (predoctoral Fellowship) and the European Research Council (ERC) under the European Union´s Horizon 2020 Research and Innovation Programme (817884 ViroPedTher). The funding bodies were not involved in the data collection, the decision to publish, or the preparation of the manuscript.

Conflict of interest statement

J.F., C.G.M., and F.F.L. report intellectual property in Delta-24 virus.

Authorship statement

Conceptualization, J.F., M.M.A., and C.G.M.; literature analysis, M.G.-M., V.L., S.G., M.M.A., C.GM. and J.F., writing-original draft preparation, M.G.-M., V.L., S.G., and A.G.G.; review and editing, M.G.-M., V.L., S.G., A.G.G., S.K., A.P., J.H., F.L., J.F., M.M.A., and C.G.-M. All authors have read and agreed to the published version of the manuscript.