Introduction
Despite advances in targeted therapies (small molecules and antibodies) and, more recently, IO drugs such as immune checkpoint inhibitors, cancer remains a leading cause of mortality [
1]. A significant proportion of patients whose disease remains unresponsive or develops resistance to current IO therapies underscores the need for novel therapeutic approaches. Oncolytic viruses, a diverse class of viruses broadly characterized historically by their tumor cell-killing potential, have re-emerged as a promising option for cancer therapy with the recent clinical approval of a modified herpesvirus, talimogene laherparepvec (TVEC), for malignant melanoma [
2]. Driving this is the realization that OVs demonstrate an ability to activate the immune system coincident with tumor cell killing. As such, OVs are increasingly considered to be multimodal IO agents whose antitumor activity could be further enhanced by combination with other IO or cancer therapeutics. Advances in virus engineering and understanding of tumor biology have enabled the development of multiple OVs exploiting different cellular mechanisms to drive tumor selectivity, oncolytic efficacy, and immunogenicity. Indeed, several other candidates are currently being evaluated in the clinic, including the reovirus, Reolysin [
3] and coxsackievirus A21, CVA21 (Cavatak) [
4], for which interim results suggest some antitumor efficacy with manageable safety profiles. The approval of TVEC in patients with metastatic melanoma and the recent demonstration that its use could potentiate current IO therapy provide a further rationale for development of other OV candidates [
5]. However, limitations of the current OV strategies, such as their restriction to intratumoral delivery, indicate the need to develop OVs with broader applicability.
In parallel with the development of new-generation OV approaches, our understanding of why some patients respond to IO-based therapies and others remain resistant has also evolved. Responding tumors are often associated with an immunologically ‘hot’ phenotype, characterized by a high mutational burden and high density T-cell infiltrate. Conversely, ‘cold’ tumors contain increased suppressive immune-cell populations (e.g., myeloid-derived suppressive cells and T regulatory cells) and a low density of T-cell infiltrate and exhibit a low mutational burden and/or antigen expression [
6]. The multimodal mechanisms of OVs make them well suited to transform the tumor microenvironment (TME) of non-responding patients into ‘hot’ tumors; the liberation of tumor antigens via immunogenic cell lysis is expected to increase T-cell infiltration and, by way of their immunogenic properties, may also potentially reverse established suppressive elements [
7]. Additionally, these same properties are expected to enhance responses in those patients who already benefited from IO therapy and, as such, provide rationale for MEDI5395 and IO combination approaches.
Newcastle disease virus (NDV) is a negative-strand RNA avian paramyxovirus that exhibits both potent oncolytic activity and immunostimulatory properties [
8]. We previously described MEDI5395, an attenuated form of the wild-type 73-T strain of NDV [
9]. NDV elicits a strong type I interferon (IFN) response via engagement of cytosolic retinoic acid-inducible gene I (RIG-I)-like RNA innate immune sensors [
10,
11]. Through virus engineering, MEDI5395 immune modulatory potential has been further enhanced by insertion of a transgene encoding granulocyte–macrophage colony-stimulating factor (GM-CSF) [
9]. The resulting therapeutic candidate, MEDI5395, maintains NDV’s potent oncolytic activity against a wide range of human tumor cells (Harper et al., manuscript in preparation); however, its immunostimulatory properties have not been fully explored. It was hypothesized that, coincident with virus-induced tumor killing, MEDI5395 would elicit a type I IFN response together with other proinflammatory cytokines in the TME and initiate the generation of a robust systemic adaptive antitumor immune response.
The effect of MEDI5395 on primary human immune-cell populations that may impact the antitumor immune response is described here. Exposure of human immune cells to MEDI5395 led to preferential, but non-productive, infection of cells within the myeloid lineage, including monocytes, macrophages, and dendritic cells (DCs). However, unlike tumor cell infection, this did not result in efficient virus replication, but to type I IFN and pro-inflammatory cytokine production and enhanced antigen presentation and T-cell activation. Furthermore, infected myeloid cells could transfer MEDI5395 to uninfected tumor cells, leading to tumor cell lysis. Tumor-derived antigens released during NDV infection were efficiently taken up by DCs and were presented in a stimulatory context to tumor antigen-specific T cells. These data support the hypothesis that, in addition to its direct tumor-killing activity, MEDI5395 has potent immunostimulatory effects that have the potential to transform the TME (e.g., from ‘cold’ to ‘hot’) and enhance antitumor immunity.
Materials and methods
Virus infections and measurement of infectious particles
NDV was prepared and purified as described previously [
9]. Virus infections were initiated by incubating cells at the indicated multiplicity of infection (MOI) at 37 °C and 5% CO
2. In some cases, virus was incubated with the cells, followed by three washes of complete medium before the cells were returned to the incubator. Infectious virus titers were measured with a plaque assay. Briefly, tenfold serial dilutions of virus stock or assay samples were placed on Vero cells (American Type Culture Collection, Manassas, VA, USA) and incubated in 2% methylcellulose in complete medium for 6 days. Chicken anti-NDV antibody (Abcam, Cambridge, UK) was added in blocking solution for 1 h at room temperature. Plates were washed, the secondary antibody was applied [rabbit anti-chicken Y (H + L)-horseradish peroxidase] and incubated for 1 h, plates were washed again, and 3-amino-9-ethylcarbazole substrate was added for 45 min. Plates were washed and then allowed to dry before plaques were counted and virus titers were calculated.
Cell lines
Tumor cell lines used in this study were cultured according to standard procedures. HT10180 fibrosarcoma cells were maintained in minimum essential medium plus 10% fetal calf serum (FCS; Thermo Fisher Scientific, Waltham, MA, USA), and MEL624 melanoma and MDA-MB-231 breast cancer cells lines were maintained in Dulbecco’s modified Eagle medium plus 10% FCS. For production of MEL624 and MDA-MB-231 cells expressing the pp65 protein of human cytomegalovirus, cells were transduced with lentivirus expressing the full-length coding sequence of pp65 (Atum, Newark, CA, USA) and a puromycin resistance cassette. Successfully transduced cells were selected in the presence of 0.5-mg/mL puromycin (Thermo Fisher Scientific) before use in the indicated assays.
PBMCs and cell isolation
Peripheral blood mononuclear cells (PBMCs) from leukocyte cones from healthy donors were isolated with standard Ficoll-Paque Plus (GE Healthcare, Chicago, IL, USA) density gradient centrifugation. PBMCs from donors with known human leukocyte antigen (HLA) type and antigen reactivity were purchased from Cellular Technology Limited (OH, USA). Whole PBMCs were resuspended in culture medium (RPMI-1640, 10% FCS, 1% penicillin–streptomycin solution; all from Thermo Fisher Scientific) and cultured at 37 °C and 5% CO2 for the indicated time. CD14-expressing (CD14+) cells, plasmacytoid DCs (pDCs), and T cells from PBMCs were isolated by using CD14+ positive selection, pDC isolation, and CD3+ negative selection kits (StemCell Technologies, Cambridge, MA, USA), respectively, with a RoboSep automated cell separator (StemCell Technologies).
Monocyte-derived macrophage and moDC differentiation
For differentiation of macrophages and monocyte-derived DCs (moDCs), CD14
+ monocytes were resuspended in culture medium supplemented with 100-ng/mL macrophage colony-stimulating factor (M-CSF) (PeproTech, Rocky Hill, NJ, USA) and cultured at 37 °C and 5% CO
2 for 6 days. Macrophage cultures were supplemented on day 3 with another 100-ng/mL M-CSF. Day 6 macrophages were gently scraped from the culture plates after brief treatment with Accutase (Thermo Fisher Scientific), washed with phosphate-buffered saline (PBS), counted, and re-plated in fresh medium to adhere overnight in appropriate assay plates. Polarization of macrophages was conducted according to standard protocols, using IFN-γ (M1) or interleukin-4 (IL-4) and IL-13 (M2) (PeproTech, Rocky Hill, NJ, USA) [
12]. Immature moDCs were derived from CD14
+ monocytes by culturing at 37 °C and 5% CO
2 for 6 days in medium supplemented with 100-ng/mL GM-CSF and 100-ng/mL IL-4 (both from PeproTech). moDC cultures were supplemented on day 3 with another 100-ng/mL GM-CSF and 100-ng/mL IL-4. Floating moDCs were collected and the loosely adherent cells were recovered with gentle scraping. Both fractions were pooled and plated in appropriate assay plates. In some cases, differentiated cells were infected with virus at the indicated MOI and washed before re-plating.
T-cell assays
For Staphylococcal enterotoxin B (SEB) stimulations, flat-bottomed plates were coated with 0.5-µg/mL anti-CD3 (OKT3; BioLegend, San Diego, CA, USA), and SEB (Sigma Chemical Company, St. Louis, MO, USA) was added at a concentration of 450 ng/mL. Whole PBMCs were added with the indicated dose of virus and incubated for 3–5 days before the supernatant was harvested. For DC–T-cell mixed leukocyte reactions and DC presentation assays, isolated T cells and differentiated moDCs were co-cultured in U-bottomed plates at the indicated ratio and incubated at 37 °C and 5% CO
2. For DC presentation assays, immature moDCs were, where indicated, first matured by the addition of lipopolysaccharide (LPS; Sigma Chemical Company) (10 ng/mL) and IFN-γ (100 IU/mL) [
13] and cultured overnight. Mature moDCs were then washed thoroughly and co-cultured with T cells. In the indicated experiments, 2.5 μM HLA-A*0201-restricted CMV-specific peptide, pp65 363–373 (NLVPMVATV), was added. The cultures were maintained for 5 days before the supernatant was harvested.
Cell staining and flow cytometry
After collection, cells were washed in PBS and resuspended in ultraviolet (UV) live/dead Zombie dye (BioLegend) and labeled according to the manufacturer’s instructions. Cells were then washed with fluorescent activated cell sorting (FACS) buffer (PBS, 1% bovine serum albumin, 0.1% NaN3) and resuspended in TruStain FcX (BioLegend) for 15 min to block Fc receptors before the addition of specific antibodies at prevalidated concentrations. Antibodies were obtained from BioLegend, Thermo Fisher Scientific, or Abcam (Cambridge, UK). Cells were incubated with antibodies for a further 30 min and then washed twice with FACS buffer. After staining, cells were fixed with 4% formalin and resuspended in FACS buffer. Samples were acquired with a LSRFortessa flow cytometer (BD Biosciences, Franklin Lakes, NJ, USA), and data were analyzed with FlowJo software (FlowJo, Ashland, OR, USA). Gating was determined with fluorescence-minus-one, unstained, and untreated controls. For labeling of macrophages before co-culture with tumor cells, CellTracker Orange CMTMR dye (Thermo Fisher Scientific) was used according to the manufacturer’s instructions. Apoptosis of tumor cell targets was measured by the addition of IncuCyte caspase 3/7 reagent (Essen BioScience, Ann Arbor, MI, USA) to the cell culture before imaging with IncuCyte ZOOM (Essen BioScience).
Cytokine measurement
The secreted cytokines were measured in cell-free cell culture supernatants, using the IFN-α2a kit, GM-CSF kit, and Human Proinflammatory panel Meso Scale Discovery platform (Meso Scale Diagnostics, Rockville, MD. USA), or validated IFN-γ (BD BioSciences), IL-2 and CCL13 (R&D Systems, Minneapolis, MN, USA) enzyme-linked immunosorbent assay (ELISA) protocols according to the manufacturers’ directions. Standard curves and results were generated with Discovery workbench software (Meso Scale Diagnostics) for the Meso Scale Discovery kits and Prism software program, version 7.04 (GraphPad Software, La Jolla, CA, USA), for ELISAs. Enzyme-linked immunosorbent spot assays (ELISPOTs) were performed with 96-well ELISPOT plates (Millipore Sigma, Billerica, MA, USA) and human IFN-γ ELISPOT antibody kits (Mabtech, Cincinnati, OH, USA). Spots were enumerated with an ImmunoSpot plate reader (Cellular Technology Limited Europe, Bonn, Germany).
Statistics
Prism software, version 7.04, was used for determination of statistical significance. A two-way analysis of variance was applied to the data in the original scale, considering treatment and dose as two different factors, and the P value of the interaction was assessed. Dunnett’s test was applied for multiple comparisons in case of a significant P value. Significance was reported as *P < 0.05; **P < 0.01; and ***P < 0.001.
Discussion
OV therapy has the potential to transform treatment in many cancer settings. Previously, we have described efforts to develop an OV with a multimodal mechanism of action, including broad oncolytic activity and an ability to enhance the antitumor immune response [
9]. Here, the impact of MEDI5395 infection on human immune-cell populations was examined. It has been demonstrated that NDV induces activation of immune-cell populations, including human PBMCs [
18], NK cells [
19,
20], and mouse macrophages [
21-
24]. These findings are extended here with a comprehensive analysis of human PBMC populations acting as surrogates for subsets of immune cells present within the TME and the peripheral circulation. MEDI5395 infection of PBMCs elicited a robust proinflammatory cytokine response, including bioactive GM-CSF transgene expression, self-limiting infection of myeloid populations, and concomitant activation of specific immune-cell populations, including NK cells, macrophages, and DCs. These observations support the hypothesis that MEDI5395 drives an immune phenotype consistent with the positive transformation of the TME [
6].
MEDI5395 is engineered to express a GM-CSF transgene, a cytokine with pleiotropic immunomodulatory properties, including the ability to enhance aspects of antigen presentation within the TME [
25]. In whole-PBMC preparations, MEDI5395 was found to induce human immune-cell activation independently of GM-CSF transgene expression, whereas GM-CSF produced from NDV replication was found to further enhance monocyte activation and IL-8 production. These results suggest that the effect of the GM-CSF transgene is dependent on the cellular context and may act directly on monocyte-derived cells such as DCs and macrophages [
25]. These observations demonstrate the potential of including immune-modulating transgenes in the development of NDV-based therapeutics to further enhance antitumor activity.
The data presented here inform our understanding of the possible impact of MEDI5395 infection on the TME. A potent type I IFN and proinflammatory cytokine response was observed after NDV infection, consistent with engagement of the cytosolic virus sensing proteins RIG-I and toll-like receptor 3 (TLR-3) and TLR-7/8 [
26]. Although RIG-I is expressed in most cell types, TLR-3 and TLR-7 expression is restricted to DCs and other antigen-presenting cells that are present in the TME and are believed to play a critical role in the recognition of cancer by the immune system [
27]. Immune cells in the TME would probably be activated through these pathways. In many cancers, the tumor cell-intrinsic type I IFN response is believed to be deficient [
4]. In such cases, triggering of a type I IFN response via the immune component of the TME may benefit the antitumor immune response. IO therapies with an immune-activating profile, such as demonstrated here with the phenotypic and functional activation of innate immune cells, are proposed to contribute to the favorable transformation of so-called “cold” tumors (those with an immune composition associated with poor prognosis, including low T-cell infiltrate) to “hot” tumors [
6,
28].
The finding that moDCs could endocytose MEDI5395-infected tumor cells and present tumor-derived antigens to specific T cells supports a mechanistic link between the noted oncolytic activity of NDV and its ability to promote antitumor immune responses. NDV-induced tumor cell lysis, which may involve direct induction of proapoptotic signals via RIG-I engagement [
29], would release antigens and cellular debris from susceptible tumors, thus contributing to the initiation of tumor-specific immune responses. Although a model antigen (pp65) was used in the present study, it is conceivable that therapy-induced tumor destruction in vivo may allow for the presentation of novel tumor-specific antigens, thus broadening the antitumor T-cell response, through the process of antigen spreading [
30]. Moreover, presentation of tumor antigens by local DCs and some macrophage populations would probably be rendered immunogenic by the type I IFN-driven inflammation associated with infection. Indeed, IFN-α itself is known to augment DC cross-presentation [
31]. The observation that direct infection of DCs by MEDI5395, a possible by-product of therapy, enhanced their function as antigen-presenting cells suggests a further mechanism by which MEDI5395 may impact the priming of antitumor immune T-cell response. A similar DC-maturing effect has been observed using an NDV-GFP virus [
32]. Concomitant to this, release of IFN-γ by activated T cells, as observed in vitro here, may upregulate PD-L1 expression within the TME. Indeed, the combination of OVs and PD-1 blocking antibodies are already being testing in the clinic [
28].
Several cancer types are characterized by different tumor-associated macrophage populations [
33]. M2-polarized macrophages, though not a direct surrogate of tumor-associated macrophages or myeloid-derived suppressive populations, are clearly distinguished from classically activated antitumoral macrophages. MEDI5395 selectively infected M2- but not M1-polarized macrophages and augmented their cytokine output, suggesting a mechanism through which MEDI5395 may reduce the immunosuppressive potential within the TME. In support of this hypothesis, others have demonstrated that type I IFN signaling is linked to the reprogramming of myeloid-derived suppressor cells [
34,
35]. This finding may help guide patient or indication selection for MEDI5395 therapy to those where strong M2 signatures within the TME are observed.
The ability for any OV to remain in the circulation long enough to achieve its maximal therapeutic potential before neutralization by antiviral antibodies remains a key challenge for OV therapy development, particularly where there is pre-existing immunity to the virus vector. Consequently, many OV strains are limited to intratumoral delivery. NDV has advantages over most OV platforms in this regard because humans are not natural hosts and, being rarely exposed, have little pre-existing immunity [
8]. The wild-type NDV strains PV701 and HUJ have been shown in phase 1/2 clinical trials to be amenable to intravenous and systemic delivery, to be well tolerated, and to elicit some objective responses [
36,
37]. The findings presented here showing that NDV differentially infected cells of the myeloid lineage and that macrophages could act as a vector to mediate infection of tumor cells suggest a potential mechanism by which intravenous MEDI5395 may transit to tumor lesions. Circulating monocytes that become infected with MEDI5395 may contribute to this effect upon egress from the blood and differentiation into macrophages or DCs within the TME. These observations, together with our previous data demonstrating efficient tumor targeting after intravenous delivery of MEDI5395 in preclinical mouse models [
9], provide a strong rationale for the clinical development of intravenously delivered, NDV-based OV therapies.
The emerging portfolio of cancer drugs, such as immune checkpoint blockade monoclonal antibodies and antibody–drug conjugates, has been providing much-needed therapeutic options and has afforded a small proportion of patients with unprecedented levels of survival compared with the current standards of care. However, as a large proportion of treated patients have disease that remains unresponsive or later develops resistance, new therapeutic strategies are urgently required. Guided by the hypothesis that successful treatment of diverse cancer indications will require an agent exhibiting multiple mechanisms of action, we have developed MEDI5395, a recombinant NDV exhibiting several beneficial attributes. The results presented here provide evidence that MEDI5395 would promote immune-cell activation and proinflammatory responses and/or reduce the immunosuppressive milieu dominant in many cancers. Translated to the human setting, these observations have the potential to provide a new benchmark in the standards of care across a range of cancer indications and to improve overall patient survival.
Compliance with ethical standards
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.