Endogenous retroviruses and their receptors: development of new targeted virotherapy tools – EndoRetroTarget

In the course of our work, several approaches have been developed to specifically target Env-K, Env-W, HEMO and ASCT2:

1) Identification of ligands such as nanobodies (S. Moutel and F. Perez) or alphaReps (P. Minard) by phage display. AlphaReps are a class of small protein ligands consisting of repeated alpha-helices. Phages were counter-selected on control cells and selected on cells transfected to express the target protein. Ligand binding is finally analyzed by flow cytometry.

2) ADC — Anti-HEMO monoclonal antibodies were obtained in mice. ScFv were prepared from these antibodies, then fused to human Fc domains (human IgG2, pFUSE, InvivoGen), and coupled to MMAE (Mono-Methyl Auristatin E) via a maleimide moiety, a drug/antibody ratio of 8-10 was obtained. The cytolytic activity of ADCs was evaluated in vitro in HEMO-transfected and non-transfected cells and in tumor cell lines.

3) Application of ligands to the preparation of oncolytic viruses (OVs) — Generation of viral pseudotypes and replicative VSVs redirected via substitution of VSV-G by:

a) phage display ligands and vaccine-derived ScFv fused with measles H protein and expressed in the presence of F protein; ligands/ScFv fused with the transmembrane domain of the human PDGFR protein. Evaluation by 2 types of assay: (i) cell fusion experiments via co-culture of cells expressing the target (Env or ASCT2) and a fragment of beta-Gal with cells expressing the ligand/ScFv fused with H + F protein or with PDGFR and the complementary fragment of beta-Gal, (ii) pseudotyping experiments, GFP monitoring by epifluorescence microscopy or flow cytometry.

b) endogenous human Env-K receptor (R) (VSV-R-Env-K) to target Env-K-positive cells (“reverse virus”). In vitro and in vivo evaluation (see below).

c) Env-W (syncytin-1) to target ASCT2-positive cells. In vitro evaluation of VSV-Syn infection (GFP) and induced-lysis (measurement of intracellular ATP, Promega) in transfected cells, tumor cell lines, spheroids, organoids, primary tumor/PDX slices (300-micrometer-thick slices, vibratome). In vivo evaluation of VSV-Syn in PDX.

4) Application of ligands to CAR preparation: preparation of a retroviral vector consisting of an anti-Env-K nanobody fused with CD8 hinge and transmembrane domains, the 4-1BB co-stimulatory domain and the CD3z signaling domain (2d generation CAR). Anti-Env-K CARs were transduced in CD8 T lymphocytes from healthy donors, then their cytolytic activity was assessed in vitro in the presence of Env-K-transduced and control cells.

Targeting 3 ERV envelop proteins: HEMO, Env-K and Env-W

>10 anti-HEMO ScFv obtained in mice were used to prepare (i) ADCs and (ii) redirected viral pseudotypes via fusion with measles H protein in the presence of F protein. Anti-HEMO ADCs induced cell death in HEMO-transfected cells but not in the tumor cell lines evaluated. Redirected pseudotypes infect cells transfected with HEMO but not the tumor cell lines tested.

Phage display screenings led to the identification of >10 anti-Env-K nanobodies, 3 anti-Env-K alphaReps, and 2 anti-Env-W alphaReps. They bind cells transfected with the targeted Env, but do not recognize the tumor cell lines evaluated. These ligands were used to prepare redirected viral pseudotypes via fusion with measles H protein + F or fusion with PDGFR. Anti-Env-K pseudotypes infected stable Env-K lines, but not the tumor cell lines tested. We did not obtain infectious pseudotypes targeting Env-W (alphaRep-H + F). A CAR was also prepared from an anti-Env-K nanobody, but no cytotoxic activity was observed when CAR-T cells were co-cultured with Env-K-positive cells.

We have obtained “reverse” replicating viruses (VSV-R-Env-K) to target Env-K. These viruses specifically infect stable Env-K cell lines. When evaluated in vitro in tumor cell lines and in vivo in PDX, low infection rates of between 1% and 2% of cells were measured.

Overall targeting can in several cases be obtained, but targeting efficacy seems to be limited by a not sufficient level of expression of the cell membrane proteins to be targeted in bona fide tumor cells, whereas they are in transduced cell lines.

Targeting ASCT2, the endogenous Env-W (syncytin-1) receptor:

The VSV-G protein was replaced by Env-W to prepare viral pseudotypes and replicative oncolytic viruses (VSV-Syn). Efficient infection, fusion and/or lysis of ASTC2-positive cells was observed in vitro in tumor cell lines, spheroids, organoids derived from liver metastases (pancreatic cancers) and thick slices of tumors (primary tumors and PDX). While dendritic cells and syncytiotrophoblasts are the healthy cells in which ASCT2 is most highly expressed, no VSV-Syn infection was observed in PBMCs and placental normal cells. A patent has been filed (October 2023) and a paper is in preparation (Tran et al.).

VSV-Syn OVs are currently being tested in vivo in immunodeficient mice bearing patient-derived tumor xenografts (PDX). Several PDXs with the highest expression levels of ASCT2 (as well as HERV-K envelope proteins, see above) and from several cancer types have been selected and are being evaluated in immunodeficient mice (NSG). Mice are treated by repeated intra-tumoral or intra-veinous injections with the oncolytic viruses VSV-Syn. Our initial results indicate infection of the PDX with VSV-Syn and viral replication with syncytia formation.

This research program has led to the identification of protein ligands against 3 endogenous retrovirus envelope proteins: Env-K, Env-W and HEMO. These ligands recognize their targets in their native conformation on the surface of transfected cells. To detect and eliminate tumor cells that express Env at a lower level, the ligands will be multimerized into homo- or hetero-multimers (i.e. ligands that recognize different domains of the same target protein).

The Env-K protein was also targeted via its endogenous receptor (R) expressed in place of VSV-G in pseudotypes and replicative viruses, VSV-R-Env-K. A pediatric cancer-derived PDX in which Env-K is expressed was transplanted into mice. VSV-R-Env-K did not induce tumor regression, infecting between 1 and 2.5% of grafted tumor cells. The infectivity of the VSV-R-Env-K replicative virus will continue to be evaluated, notably in vitro via the preparation of PDX-derived cells and organoids, and in vivo in immunocompetent murine models.

Finally, the use of Env-W (syncytin-1) to target its ASCT2 receptor has led to the development of a VSV-Syn redirected oncolytic virus whose efficacy has been demonstrated on a wide panel of tumor cell lines, PDX, spheroids and organoids expressing ASCT2. As expected, due to its sensitivity to interferon responses, VSV-Syn selectively infects cancer cells. Expression of Env-W on the surface of VSV-Syn-infected cells leads to their fusion with neighboring non-infected cells expressing ASCT2, resulting in the formation of multinucleated cell structures (syncytial) known to generate particularly immunogenic death. VSV-Syn viruses will next be evaluated in immunocompetent mice grafted with syngeneic murine tumor cells, in order to assess the VSV-Syn-induced immune responses required for tumor elimination.

Given the high expression of its target protein in certain cancers, its sensitivity to interferon, and its fusion properties, we aim to pursue the development of VSV-Syn to the clinical stage. As VSV-Syn recognizes both canine and human ASCT2, it will next be evaluated in dogs for mammary and subcutaneous tumors (OncoVet, veterinary clinic in Lille). Dogs will be treated by repeated intra-tumoral injections of VSV-Syn. This veterinary medicine project, carried out jointly by UMR9196 and Viroxis, has received ANR funding (ANR2023 CE18-OVICAN).

Elements of retroviral origin, or HERVs (human endogenous retroviruses), represent a significant proportion of the human genome. Globally repressed because of their infectious origin, they can be expressed in normal contexts, notably placentation, and/or pathological contexts. In cancer, epigenetic heterogeneity can lead to the activation of a set of HERVs, which in turn can be associated with aberrant expression of oncogenes via the use of their promoter regions (LTRs, long terminal repeats). Although derived from integrative infections that occurred tens of millions of years ago, some HERV genes still encode functional envelope proteins, whose ability to bind to an endogenous receptor - or even to perform cell fusion - has been maintained. In the course of our work on HERV envelope proteins, we have identified several receptors and obtained specific antibodies. These ligands are being used to develop new therapeutic tools that target the tumor-specific HERV envelope proteins, HERV-K and HEMO. Pseudotypes have been "retargeted" through the expression of two measles proteins, wild-type F protein and H protein fused to the HERVs ligands. In partnership with VIROxIS, we propose to produce preclinical batches of oncolytic measles virus or vesicular stomatitis virus (VSV) re-targeted against HERV-K and HEMO and to evaluate their activity against human tumors derived from digestive, gynecological and lung cancers that can be transplanted in mice (PDX, patient derived xenograft). Retargeting of oncolytic viruses aims to reduce the risk of infection/elimination of non-cancerous cells. In addition, the efficacy of retargeted oncolytic viruses will be tentatively enhanced by bispecific approaches, combining a receptor and an antibody binding the envelope of the targeted HERV.
In our studies, we have achieved infection of up to 50% of HERV-K envelope positive cells with pseudotypes expressing H protein fused to the HERV-K receptor and lesser percentages with H pseudotypes fused to anti-HERV-K or anti-HEMO scfv. Therefore, this research program will start with the preparation of pre-clinical batches of HERV-K fused oncolytic viruses, while new anti-HERV-K and anti-HEMO antibodies will be selected from humanized antibody libraries. We currently have more than ten antibodies against the two envelope proteins targeted in this program and a new screening of antibody libraries has already started in collaboration with the Curie Institute. Validation of the anti-HERVs antibodies will be performed following their evaluation in PDX mice in the drug-conjugated antibody (DCA) format. The antibodies with the best anti-tumor activity will be fused to the H protein for application to oncolytic viruses. In parallel, their anti-tumor activity will also be evaluated in the form of chimeric antigen receptor (CAR) T cells. This research program aims at the development of new cancer therapies based on the use of HERV receptors and specific antibodies, in particular mono- or bispecific re-targeted oncolytic viruses never developed so far, which due to their replicative property could efficiently eliminate solid tumors.