Tumour-associated macrophages (TAMs) are often implicated in cancer progression but can also exert anti-tumour activities. Selective eradication of cancer-promoting (M2-like) TAM subsets is a highly sought-after goal. Here, we have devised a novel strategy to achieve selective TAM depletion, involving the use of T cell engagers to direct endogenous T cell cytotoxicity towards specific M2-like TAMs. To avoid "on-target off-tumour" toxicities, we have explored localising expression of the T cell engagers to the tumour with enadenotucirev (EnAd), an oncolytic adenovirus in Phase I/II clinical trials. A panel of bi- and tri-valent T cell engagers (BiTEs/TriTEs) was constructed, recognising CD3ε on T cells and CD206 or folate receptor β (FRβ) on M2-like macrophages. Initial characterisation of BiTE/TriTE activity and specificity was performed with M1- and M2-polarised monocyte-derived macrophages and autologous lymphocytes from healthy human peripheral blood donors. T cell engagers were inserted into the genome of EnAd, and oncolytic activity and BiTE secretion assessed with DLD-1 tumour cells. Clinically-relevant ex vivo models (whole malignant ascites from cancer patients) were employed to assess the efficacies of the free- and virally-encoded T cell engagers. T cells activated by the CD206- and FRβ-targeting BiTEs/TriTEs preferentially killed M2- over M1-polarised autologous macrophages, with EC50 values in the nanomolar range. A TriTE with bivalent CD3ε binding - the first of its kind - demonstrated enhanced potency whilst retaining target cell selectivity, whereas a CD28-containing TriTE elicited non-specific T cell activation. In immunosuppressive malignant ascites, both free and EnAd-encoded T cell engagers triggered endogenous T cell activation and IFN-γ production, leading to increased T cell numbers and depletion of CD11b+CD64+ ascites macrophages. Strikingly, surviving macrophages exhibited a general increase in M1 marker expression, suggesting microenvironmental repolarisation towards a pro-inflammatory state. This study is the first to achieve selective depletion of specific M2-like macrophage subsets, opening the possibility of eradicating cancer-supporting TAMs whilst sparing those with anti-tumour potential. Targeted TAM depletion with T cell engager-armed EnAd offers a powerful therapeutic approach combining direct cancer cell cytotoxicity with reversal of immune suppression.

Background Programmed death-ligand 1 (PD-L1) is an important immune checkpoint protein that can be regarded as a pan-cancer antigen expressed by multiple different cell types within the tumor. While antagonizing PD-L1 is well known to relieve PD-1/PD-L1-mediated T cell suppression, here we have combined this approach with an immunotherapy strategy to target T cell cytotoxicity directly toward PD-L1-expressing cells. We developed a bi-specific T cell engager (BiTE) crosslinking PD-L1 and CD3ε and demonstrated targeted cytotoxicity using a clinically relevant patient-derived ascites model. This approach represents an immunological ‘volte-face’ whereby a tumor immunological defense mechanism can be instantly transformed into an Achilles’ heel for targeted immunotherapy. Methods The PD-L1 targeting BiTE comprises an anti-PD-L1 single-chain variable fragment (scFv) or nanobody (NB) domain and an anti-CD3 scFv domain in a tandem repeat. The ability to activate T cell cytotoxicity toward PD-L1-expressing cells was established using human carcinoma cells and PD-L1-expressing human (‘M2’) macrophages in the presence of autologous T cells. Furthermore, we armed oncolytic herpes simplex virus-1 (oHSV-1) with PD-L1 BiTE and demonstrated successful delivery and targeted cytotoxicity in unpurified cultures of malignant ascites derived from different cancer patients. Results PD-L1 BiTE crosslinks PD-L1-positive cells and CD3ε on T cells in a ‘pseudo-synapse’ and triggers T cell activation and release of proinflammatory cytokines such as interferon-gamma (IFN-γ), interferon gamma-induced protein 10 (IP-10) and tumour necrosis factor-α (TNF-α). Activation of endogenous T cells within ascites samples led to significant lysis of tumor cells and M2-like macrophages (CD11b+CD64+ and CD206+/CD163+). The survival of CD3+ T cells (which can also express PD-L1) was unaffected. Intriguingly, ascites fluid that appeared particularly immunosuppressive led to higher expression of PD-L1 on tumor cells, resulting in improved BiTE-mediated T cell activation. Conclusions The study reveals that PD-L1 BiTE is an effective immunotherapeutic approach to kill PD-L1-positive tumor cells and macrophages while leaving T cells unharmed. This approach activates endogenous T cells within malignant ascites, generates a proinflammatory response and eliminates cells promoting tumor progression. Using an oncolytic virus for local expression of PD-L1 BiTE also prevents ‘on-target off-tumor’ systemic toxicities and harnesses immunosuppressive protumor conditions to augment immunotherapy in immunologically ‘cold’ clinical cancers.

Tumor cells exhibiting the Warburg effect rely on aerobic glycolysis for ATP production and have a notable addiction to anaplerotic use of glutamine for macromolecular synthesis. This strategy maximizes cellular biosynthetic potential while avoiding excessive depletion of NAD+ and provides an attractive anabolic environment for viral infection. Here, we evaluate infection of highly permissive and poorly permissive cancer cells with wild-type adenoviruses and the oncolytic chimeric adenovirus enadenotucirev (EnAd). All adenoviruses caused an increase in glucose and glutamine uptake along with increased lactic acid secretion. Counterintuitively, restricting glycolysis using 2-deoxyglucose or by limiting glucose supply strongly improved virus activity in both cell types. Antagonism of glycolysis also boosted EnAd replication and transgene expression within human tumor biopsies and in xenografted tumors in vivo. In contrast, the virus life cycle was critically dependent on exogenous glutamine. Virus activity in glutamine-free cells was rescued with exogenous membrane-permeable α-ketoglutarate, but not pyruvate or oxaloacetate, suggesting an important role for reductive carboxylation in glutamine usage, perhaps for production of biosynthetic intermediates. This overlap between the metabolic phenotypes of adenovirus infection and transformed tumor cells may provide insight into how oncolytic adenoviruses exploit metabolic transformation to augment their selectivity for cancer cells. Significance: This study describes changes in glucose and glutamine metabolism induced by oncolytic and wild-type adenoviruses in cancer cells, which will be important to consider in the preclinical evaluation of oncolytic viruses.

The interplay between oncolytic virus infection and tumour hypoxia is particularly unexplored in vivo, although hypoxia is present in virtually all solid carcinomas. In this study, oncolytic adenovirus infection foci were found within pimonidazole-reactive, oxygen-poor areas in a colorectal xenograft tumour, where the expression of VEGF, a target gene of the hypoxia-inducible factor (HIF), was attenuated. We hypothesised that adenovirus infection interferes with the HIF-signalling axis in the hypoxic tumour niche, possibly modifying the local vascular supply. In vitro, enadenotucirev (EnAd), adenovirus 11p and adenovirus 5 decreased the protein expression of HIF-1α only during the late phase of the viral life cycle by transcriptional down-regulation and not post-translational regulation. The decreasing HIF levels resulted in the down-regulation of angiogenic factors such as VEGF, coinciding with reduced endothelial tube formation but also increased T-cell activation in conditioned media transfer experiments. Using intravital microscopy, a decreased perfused vessel volume was observed in infected tumour nodules upon systemic delivery of EnAd, encoding the oxygen-independent fluorescent reporter UnaG to a tumour xenograft grown under an abdominal window chamber. We conclude that the attenuation of the HIF pathway upon adenoviral infection may contribute to anti-vascular and immunostimulatory effects in the periphery of established infection foci in vivo.

Ionising radiation causes cell death through the induction of DNA damage, particularly double-stranded DNA (dsDNA) breaks. Evidence suggests that adenoviruses inhibit proteins involved in the DNA damage response (DDR) to prevent recognition of double-stranded viral DNA genomes as cellular dsDNA breaks. We hypothesise that combining adenovirus treatment with radiotherapy has the potential for enhancing tumour-specific cytotoxicity through inhibition of the DDR and augmentation of virus production. We show that EnAd, an Ad3/Ad11p chimeric oncolytic adenovirus currently being trialled in colorectal and other cancers, targets the DDR pathway at a number of junctures. Infection is associated with a decrease in irradiation-induced 53BP1 and Rad51 foci formation, and in total DNA ligase IV levels. We also demonstrate a radiation-associated increase in EnAd production in vitro and in a pilot in vivo experiment. Given the current limitations of in vitro techniques in assessing for synergy between these treatments, we adapted the plaque assay to allow monitoring of viral plaque size and growth and utilised the xCELLigence cell adhesion assay to measure cytotoxicity. Our study provides further evidence on the interaction between adenovirus and radiation in vitro and in vivo and suggests these have at least an additive, and possibly a synergistic, impact on cytotoxicity.