Oncolytic virus for expression of immune checkpoint modulators

Abstract

The present invention provides an oncolytic virus comprising nucleotide sequence(s) encoding one or more immune checkpoint modulator(s). It also concerns a pharmaceutical composition comprising effective amount of said oncolytic virus and, eventually, a pharmaceutically acceptable vehicle and its use for treating proliferative diseases such as cancers.

Claims

1. An oncolytic virus comprising inserted in its genome a nucleic acid molecule encoding one or more immune checkpoint modulator(s), wherein said virus is a vaccinia virus defective for thymidine kinase (TK) resulting from inactivating mutations in the J2R viral gene and defective for Ribonucleotide reductase (RR) activity resulting from inactivating mutations in only the viral I4L gene, and wherein said one or more immune checkpoint modulator(s) is selected from antibodies that specifically bind to PD-L1, PD-L2, LAG3, Tim3, BTLA, or CTLA4.

2. The oncolytic virus of claim 1, wherein said vaccinia virus is selected from the group of Elstree, Wyeth, Copenhagen, and Western Reserve strains.

3. The oncolytic virus of claim 1, wherein said oncolytic virus further comprises at least one therapeutic gene inserted in the viral genome.

4. The oncolytic virus of claim 3, wherein said therapeutic gene is selected from the group consisting of genes encoding suicide gene products and genes encoding immunostimulatory proteins.

5. The oncolytic virus of claim 4, wherein said suicide gene is selected from the group consisting of genes coding protein having a cytosine deaminase (CDase) activity, a thymidine kinase activity, an uracil phosphoribosyl transferase (UPRTase) activity, a purine nucleoside phosphorylase activity, and a thymidylate kinase activity.

6. The oncolytic virus of claim 5, wherein said suicide gene product is selected from the group consisting of codA::upp, FCY1::FUR1 and FCY1::FUR1[Delta] 105 (FCU1) and FCU1-8 polypeptides.

7. The oncolytic virus of claim 4, wherein said immunostimulatory protein is an interleukin or a colony-stimulating factor.

8. The oncolytic virus of claim 7, wherein said colony-stimulating factor is human GM-CSF.

9. The oncolytic virus of claim 1, wherein said one or more immune checkpoint modulator(s) comprises an antibody that specifically binds to human PD-L1.

10. The oncolytic virus of claim 9, wherein said antibody that specifically binds to human PD-L1 is selected from the group consisting of MPDL3280A and BMS-936559.

11. The oncolytic virus of claim 1, wherein said one or more immune checkpoint modulator(s) comprises an antibody that specifically binds to human CTLA-4.

12. The oncolytic virus of claim 11, wherein said antibody that specifically binds to human CTLA-4 is selected from the group consisting of ipilimumab, tremelimumab and single chain anti-CTLA4 antibodies.

13. A pharmaceutical composition comprising an effective amount of the oncolytic virus of claim 1 and a pharmaceutical acceptable vehicle.

14. The pharmaceutical composition of claim 13 comprising from approximately 10.sup.7 pfu to approximately 5 ×10.sup.9 pfu of said oncolytic virus.

15. The pharmaceutical composition of claim 13, which is formulated for parenteral administration.

16. A method for treating a proliferative disease, comprising administering an oncolytic virus comprising inserted in its genome a nucleic acid molecule encoding one or more immune checkpoint modulator(s), wherein said virus is a vaccinia virus defective for thymidine kinase (TK) resulting from inactivating mutations in the J2R viral gene and defective for Ribonucleotide reductase (RR) activity resulting from inactivating mutations in only the viral I4L gene, and wherein said one or more immune checkpoint modulator(s) is selected from antibodies that specifically bind to PD-L1, PD-L2, LAG3, Tim3, BTLA, or CTLA4.

17. The method according to claim 16, wherein said proliferative disease is a cancer.

18. The method according to claim 17, wherein said cancer is selected from the group consisting of melanoma, renal cancer, prostate cancer, breast cancer, colorectal cancer, lung cancer, and liver cancer.

19. The method according to claim 16, wherein said oncolytic virus is administered by intravenous or intratumoral route.

20. The method according to claim 16, which comprises from 2 to 5 intravenous or intratumoral administrations of 10.sup.8 or 10.sup.9 pfu of oncolytic vaccinia virus at approximately 1 or 2 weeks interval.

21. The method according to claim 16, which further comprises administration of a prodrug and/or a substance effective in anticancer therapy.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) FIG. 1 illustrates the quantity of purified anti-PD-1 mAb, Fab and scFv produced from infected CEF.

(2) FIG. 2 illustrates an analysis by electrophoresis of the purified recombinant mAb (Vacc.mAb) and scFv (Vacc.scFv) as well as commercial J43 (BioXCell) under reduced (R) and non reduced (NR) conditions.

(3) FIG. 3: In vitro replication efficacy of viruses in LoVo infected at a MOI of 0.0001, 0.001 and 0.01 with the indicated viruses at day 5 post infection. Values are represented in mean ±SD of three individual determinations.

(4) FIG. 4 illustrates the amino acid sequences of the heavy and light chains of the anti-PD1 antibodies expressed by the oncolytic vaccinia virus described herein.

(5) FIG. 5 illustrates immunoblot analysis of the recombinant mAb and Fab secreted in cell culture supernatants obtained from CEF infected with WRTG18618, WRTG18619, WRTG18620 and WRTG18621. Commercial J43 is used as reference; M represents MW markers and C negative control.

(6) FIG. 6 illustrates the concentration of recombinant J43 produced in the serum (FIG. 6A) and in tumor homogenate (FIG. 6B) of mice treated with a single administration of WRTG18618 (10.sup.7 pfu) injected intratumorally (resulting from subcutaneous implantation of B16F10 tumors) or subcutaneously (mice without tumors) or with intratumoral administration of 1 μg or 10 μg of commercial J43 (Bioxcell). Concentration of recombinant J43 was measured by quantitative ELISA 1, 5 and 11 days after injection.

EXAMPLES

(7) These examples illustrates oncolytic vaccinia virus engineered for expressing various forms of anti-immune checkpoint inhibitors. Preclinical evidence for the beneficial effects of expressing immune checkpoint blockers from viral vectors was to be demonstrated in mouse tumor models. This implies the use of i) murine-specific anti-immune check point antibodies and ii) an oncolytic poxvirus capable of infecting murine cells with a higher efficacy.

(8) The murine-specific hamster antibody J43 was chosen to target the immune checkpoint blocker murine PD-1 (mPD-1). This antibody was shown to block the interaction of mPD1 with PD-L1 (U.S. Pat. No. 7,858,746). The antibody J43 and its isotype control (hamster IgG) are available from BioXCell. The anti mPD-1 antibodies and its isotype control were used to establish functional tests and quantitative ELISA in vitro. Further, J43 and various forms thereof were cloned in an oncolytic virus and tested in vivo in tumor animal models.

(9) The oncolytic virus chosen for these studies is a vaccinia virus (VV) Western Reserve (WR) strain defective for thymidine kinase (TK) (locus J2R) and RR.sup.− (locus I4L) rendering the virus non-replicative in healthy (non-dividing) cells. In contrast, the VV TK.sup.−RR.sup.− is supposed to selectively and efficiently replicate in tumor cells.

(10) Vectorization of anti mPD-1 Molecules in Oncolytic TK.sup.−RR.sup.− vaccinia virus.

(11) The heavy chain of J43 showed high homology with the heavy chain of anti CD79b IgG. Thus, the sequence retained for cloning of the heavy chain was the variable chain of J43 and the constant chain of anti CD79b. The light chain of J43 was cloned with signal sequence from the light chain of anti CD79b antibody. The heavy and light chains were put under the control of the viral promoters pH5R or p7.5K which have slightly different strengths. Further to “whole” antibody constructs, derivatives were generated, respectively His-tagged antigen binding fragments (Fab) constructs as well as a His-tagged single chain antibody (scFv) constructs. Two construct formats were also generated for Fab depending on the light or heavy chain portions placed under each promoter. All five constructs were inserted in the backbone of TK, RR deleted WR VV. The constructs are summarized below:

(12) WRTG18618 (or mAb1) corresponding to pH5R-HC-p7.5K-LC

(13) WRTG18619 (or mAb2) corresponding to pH5R-LC-p7.5K-HC

(14) WRTG18621 (or Fab1) corresponding to pH5R-(VH-CH1-6His)-p7.5K-LC

(15) WRTG18620 (or Fab2) corresponding to pH5R-LC-p7.5K-(VH-CH1-6His)

(16) WRTG18616 (scFv) corresponding to pH5R-VH-gs-VL-6His).

(17) HC and LC represents the abbreviation of heavy and light chains, VH and VL of heavy and light variable domains, 6His of the HIS tag (6 histidines) and gs (poly Glycine-Serine linker)

(18) Constructs 1 demonstrated the highest expression level of the recombinant mAb or Fab with the expected chain assembly profiles.

(19) Virus stocks WRTG18616 (scFv), WRTG18618 (mAb1) and WRTG18621 (Fab1) were produced in BHK-21 cells and purified by tangential flow filtration (TFF).

(20) Construction of WRTG18618 (pH5R-HC-p7.5K-LC): mAb1

(21) Fragment containing HC followed by p7.5K promoter was generated by synthetic way (Geneart: Regensburg, Germany) and was cloned in vaccinia transfer plasmid pTG18496 restricted by PstI and EcoRI to give pTG18614. Fragment containing LC was generated by synthetic way and was cloned in pTG18614 restricted by NsiI and SalI to give pTG18618.

(22) The vaccinia transfer plasmid, pTG18496, is designed to permit insertion of the nucleotide sequence to be transferred by homologous recombination in TK gene of the VV genome. It contains the flanking sequences (BRGTK and BRDTK) surrounding the J2R gene and the pH5R promoter followed by multiple cloning sites.

(23) The amino acid sequences of the anti-PD-1 HC and LC are given in SEQ ID NO: 1 and SEQ ID NO: 2, respectively and in FIG. 4.

(24) Generation of WRTG18618 was performed by homologous recombination in primary chicken embryos fibroblasts (CEF) infected with a RR-deleted WR and transfected by nucleofection with pTG18618 (according to Amaxa Nucleofector technology). Viral selection was performed by plaque purification after growth in Thymidine kinase-deficient (TK.sup.−) 143B cells cultured in the presence of bromodeoxyuridine. This selection allows only TK.sup.− rWR to remain viable. Absence of contamination by parental WR was verified by PCR.

(25) Construction of WRTG18619 (pH5R-LC-p7.5K-HC): mAb2

(26) Fragment containing LC followed by p7.5K promoter was generated by synthetic way and was cloned in vaccinia transfer plasmid pTG18496 restricted by PstI and EcoRI to give pTG18615. Fragment containing HC was generated by synthetic way and was cloned in pTG18615 restricted by NsiI and MluI to give pTG18619.

(27) Generation of WRTG18619 virus was performed in CEF by homologous recombination as described above.

(28) Construction of WRTG18621 (pH5R-(VH-CH1-6His)-p7.5K-LC): Fab1

(29) Fragment containing VH-CH1-6His followed by p7.5K promoter was generated by synthetic way and was cloned in vaccinia transfer plasmid pTG18496 restricted by PstI and EcoRI to give pTG18617. Fragment containing LC was generated by synthetic way and was cloned in pTG18617 restricted by NsiI and SalI to give pTG18621.

(30) Generation of WRTG18621 virus was performed in CEF by homologous recombination as described above.

(31) Construction of WRTG18620 (pH5R-LC-p7.5K-(VH-CH1-6His): Fab2

(32) Fragment containing LC followed by p7.5K promoter was generated by synthetic way and was cloned in vaccinia transfer plasmid pTG18496 restricted by PstI and EcoRI to give pTG18615. Fragment containing VH-CH1-6His was generated by synthetic way and was cloned in pTG18615 restricted by NsiI and MluI to give pTG18620.

(33) Generation of WRTG18620 virus was performed in CEF by homologous recombination as described above.

(34) Construction of WRTG18616 (pH5R-VH-gs-VL-6His): scFv

(35) Fragment containing VH-gs-VL-6His was generated by synthetic way and was cloned in vaccinia transfer plasmid pTG18496 restricted by PstI and EcoRI to give pTG18616.

(36) Generation of WRTG18616 virus was performed in CEF by homologous recombination as described above.

(37) In vitro characterization of the encoded anti-PD-1 antibodies

(38) The recombinant oncolytic vectors described above were tested to assess the production of the encoded anti-mPD-1 molecules. To this, permissive primary cells or cell lines were infected, and supernatant harvested. Presence of anti-mPD-1 molecules can be determined by SDS-PAGE, Western blot analysis, mass spectrometry analysis, ELISA and functional assays. Commercially available anti-mPD-1 antibodies were used as controls.

(39) Supernatant from infected CEF (moi 0.2) were analyzed (after 48 and 72 h post infection) by Western Blot under non-reducing conditions using a polyclonal anti-hamster IgG antibody for detection. WRTG18618 (mAb1) and WRTG18621 (Fab1) could be detected by ELISA. Expression profile of the product secreted by WRTG18618-infected cells has similar pattern as commercial J43 antibody and expected assembly of the product secreted by WRTG18621-infected cells was observed. ScFv expression was also detected in culture supernatant by Western Blot at the expected size.

(40) mAb Purification and in vitro Assessment of Anti PD-1 Antibodies

(41) CEF were cultured in F175 flasks and infected with 2.7×10.sup.8 pfu of anti-PD-1-expressing viruses. After 48-72 h, culture supernatants were collected and subjected to filtration on 0.2 μm filters. mAb1 was purified by Hitrap protein A (GE Healthcare) followed by Superdex 200 gel filtration (GE Healthcare) whereas Fab1 and scFv were purified by HIS-Trap (GE Healthcare) followed by Superdex 75 gel filtration (GE Healthcare). scFv eluted from gel filtration in two peaks, the first one corresponding to dimers and the second one to monomers. Significant amounts of mAbs, Fab and scFv were produced as shown in FIG. 1. The yield is between 0.5 and 2 μg/ml supernatant according to the antibody format.

(42) Recombinant purified mAb had an electrophoresis profile similar to the commercial mAbs with the correct assembly of the two light and two heavy chains under non reducing conditions and the detection of the individual light and heavy chains under reducing conditions as illustrated in FIG. 2. Moreover, purified scFv migrated on SDS-PAGE at the expected size (see FIG. 2).

(43) The recombinant products can be detected by ELISA and are functional in a competition ELISA (i.e. blocking of binding of mouse PD-L1 biot at 0.2 μg/mL to mouse PD-1 coated at 1 μg/mL), mAb being more efficient than Fab in this competition experiment. Moreover, the recombinant mAb1 is more efficient than its commercial J43 counterpart as well as VV-produced scFv dimeric fraction is more efficient than the monomeric one with a difference of approximately one log in EC50.

(44) The ability of VV-encoded mAb1, Fab1 and scFv to interact with cell-surface PD-1 was studied in flow cytometry-based binding assays using the PD-1-positive T lymphoma cell lines EL4 and RMA. 10.sup.5 EL4 or RMA cells were incubated for 45 min on ice with 100 μl of either mAb1 (5 μg/ml), commercially available anti-mPD-1 antibody J43 (5 μg/ml, BioXCell) or negative control hamster IgG (5 μg/ml, BioXCell) and washed. Binding to PD-1 was detected by incubating cells for 45 min on ice with 100 μl of 10 μg/m1 FITC-conjugated monoclonal mouse anti-Armenian+Syrian antibody cocktail (BD Pharmingen). After washing, fluorescence intensity was measured on a Navios™ flow cytometer (Beckman Coulter). Data were analysed using Kaluza 1.2 software (Beckman Coulter). WRTG18618-encoded mAb1 bound efficiently to EL4 and RMA cells.

(45) To measure the binding of Fab1 and scFv, 5×10.sup.5 EL4 or RMA cells were indirectly stained with either 5 μg/m1 Fab1 or 2.5 μg/m1 monomeric scFv (which are both His-tagged) followed by PE-conjugated monoclonal mouse anti-His tag antibody (diluted 1/10, Miltenyi Biotec) and analysed as described above. Fab1 and monomeric scFv produced from WRTG18621 and WRTG18616 were shown to bind efficiently to the surface of EL4 and RMA cells. Specific binding could be demonstrated by co-incubating EL4 cells with Fab1 or monomeric scFv and full-length J43 or negative control hamster IgG from BioXCell, followed by staining with PE-conjugated anti-His tag antibody. EL4 staining was reduced in the presence of full-length J43 but not negative control hamster IgG. Altogether, the results demonstrated that VV-encoded mAb1, Fab1 and scFv were able to recognize endogenously expressed PD-1 in two murine T lymphocyte cell lines.

(46) A flow cytometry-based competition assay was established to compare blocking activity of anti-mPD-1 molecules. The assay is based on the murine T lymphoma cell line EL4 which is highly PD-1 positive. The binding of mPD-L1-hFc (murine PD-Ligand 1 comprising human IgG Fc fragment) to the cell-surface PD1 can be detected by flow cytometry using a PE-labeled anti-hFc monoclonal antibody. Binding of mPD-L1-hFc can be competed for by anti-PD-1 antibodies leading to reduction of cell-bound PE signal.

(47) The inhibitory activity of recombinant mAb1, Fab1 and scFv on PD-L1 binding to EL4 cells was assessed and compared to commercially available J43 anti-PD-1 antibody (BioXCell). More specifically, 10.sup.5 EL4 cells were co-incubated with 2 μg/ml of mPD-L1 hFc (R&D Systems) and increasing concentrations of the various anti-PD-1 antibodies in 100 μl for 45 min on ice. After washing, cells were incubated with 100 μl of 5 μg/ml PE-conjugated mouse anti-human IgG Fc (BioLegend) for 45 min on ice. Cells were washed and the mean fluorescence intensity was measured as described above. In these conditions, the control anti mPD-1 clone J43 showed an IC.sub.50 of 1.6 μg/ml.

(48) As expected, the control hamster IgG did not show any blocking activity on PD-L1 binding to EL4 cells. The recombinant mAb1, Fab1, monomeric and dimeric scFv produced respectively from WRTG18618, WRTG18621 and WRTG18616 were able to block ligand binding to cell-surface receptor as J43 from BioXCell. Importantly, WRTG18618-encoded mAb1 exhibited enhanced blocking activity as compared to J43 from BioXCell and Fab1 and scFv versions in this assay.

(49) A quantitative ELISA was established to quantify antibody concentrations produced in mouse serum or by recombinant WR-infected cells using coated rabbit-derived anti hamster antibodies to capture J43, which could be detected in return with a goat-derived anti hamster antibody. The assay is sensitive to mouse serum and standard curves have to be generated in the presence of 50% mouse serum.

(50) In vitro cytotoxicity assay (oncolytic activity)

(51) Human LoVo colon cancer cell line was transduced in suspension at a MOI of 0.001 and 0.0001 by TK-RR deleted WR vaccinia virus without transgene (WRTG18011) and by TK-RR-deleted WR vaccinia virus expressing the different anti mPD-1 molecules (WRTG18616, WRTG18618 and WRTG8621). A total of 3×10.sup.5 cells/well were plated in 6-well culture dishes in 2 ml of medium supplemented with 10% FCS. Cells were then cultured at 37° C. and cell survival was determined 5 days later by trypan blue exclusion. As shown in FIG. 3, the oncolytic activity of the different viruses resulted in 75-95% and 99% reduction in viable cell number at a MOI of 0.0001 and 0.001, respectively. TK-RR-deleted WR virus without transgene and TK-RR deleted WR virus expressing the different anti mPD-1 molecules showed no difference in cytotoxicity. Thus the expression of the different anti mPD-1 molecules did not affect the oncolytic activity of the vaccinia virus.

(52) Optimal Design for Antibody Expression

(53) The recombinant oncolytic vectors were tested by Immunoblotting to assess expression levels and assembly of the secreted antibodies. Immunoblot analyses were performed on cell culture supernatants collected 24 h after CEF infection with mAb and Fab constructs described above. Supernatant were centrifuged 5 min at 16000 g and 25 μl were prepared in Laemmli buffer without reducing agent and loaded on PrecastGel 4-15% polyacrylamide gel (Biorad). Monoclonal commercial J43 (BioXcell) was used as reference molecule. One μg of each molecule was loaded on gel. Gel electrophoresis was performed in non-reducing conditions to preserve the assembly of light and heavy chains and allow an optimal detection (i.e. the polyclonal antibody used for detection did not recognized reduced IgG and Fab chains). Proteins were then transferred onto a PVDF membrane using the Trans-blot Turbo system (Transblot Turbo Transfer pack Biorad) with the preprogrammed protocol (High MW: 10 min; 2.5 A constant; up to 25V). Membranes were saturated overnight at 4° C. in blocking solution (8 mM NaPO.sub.4, 2 mM KPO.sub.4, 154 mM NaCl pH 7.2 (PBS) supplemented with 0.05% Tween20, 5% Nonfat dry milk Biorad). Horseradish peroxidase (HRP) conjugated goat anti Armenian Hamster IgG (Jackson Immunoresearch) at 80 ng/mL in dilution buffer (PBS, 0.05% Tween20, 0.5% Nonfat dry milk) was used for the antibody immunodetection. Development was performed with Amersham ECL Prime Western Blotting detection reagents and Molecular Imager ChemiDOC™ XRS was used to capture chemiluminescence.

(54) As illustrated in FIG. 5, both WRTG18618 and WRTG18619 secreted roughly the same quantities of an antibody product with an apparent size corresponding to two heavy and two light chains linked together comparable to that of commercial J43 mAb. However, WRTG18619 also produced a non-expected product migrating between 43 and 55 kDa markers (see the extra band indicated by an arrow in the fourth lane “18619”). Similarly, an extra band was also present in the WRTG18620-infected culture supernatant (same position and same intensity in the fifth lane “18620”) whereas WRTG18621 produced high amount of correctly assembled Fab without any detectable misassembled product (see arrow head in the sixth lane “18621”).

(55) Analyses of cell culture supernatants generated after infection of mammal cell lines BHK21 and A549 generated the same profile (data not shown).

(56) When looking at the construct design, it appears that in both WRTG18619 and WRTG18620 constructs, the antibody light chain was placed under the transcriptional control of pH5R promoter (a strong promoter). One hypothesis is that in this configuration, antibody light chains could be overexpressed with respect to heavy chains and thus can assemble in homodimers. Therefore, this extra band, migrating between 43 and 55 kDa, could correspond to overproduced light chains that have assembled into a homodimer of a theoretical mass of 47 kDa.

(57) Therefore, it is preferable to express heavy chain under the control of a stronger promoter than the one used to express light chains in order to reduce the risk of generating aberrant assembly (e.g. homodimers of light chains)

(58) All together these results confirmed the ability of the recombinant oncolytic viruses described herein to secrete mAb and Fab at detectable level. WRTG18618 and WRTG18621 were selected for the rest of experiments due to their capacity to produce antibody products that are closer to expectation than those of WRTG18619 and WRTG18620.

(59) In vivo expression of vectorised anti-PD1 antibody

(60) Expression of the vectorized anti-PD-1 J43mAb was evaluated in vivo in mice with and without subcutaneously-implanted B16F10 tumors. WRTG18618 (mAb1 construct) was injected either intra-tumorally or subcutaneously and compared to commercial J43 (intra-tumoral injection).

(61) In vivo Experiments and Sample Collection

(62) More specifically, 3×10.sup.5 B16F10 cells (murine melanoma) were injected subcutaneously (S.C.) into the right flanks of six weeks old female C57BL/6 mice. At day 0 (DO), when tumor volumes reached 100-200 mm.sup.3, 100 microliters of either WRTG18618 (10.sup.7 pfu) or commercial J43 (1 μg or 10 μg, Bioxcell) were injected intra-tumorally (I.T.).

(63) WRTG18618 was also injected S.C. in mice without tumor.

(64) Blood and tumors were collected at D1, D5 and D11. For each time point, 3 mice were anesthetized with 200 μL pentobarbital and the blood was collected by intra-cardiac puncture. The blood was stored at 4° C. during 8 hours and the serum was recovered after two centrifugations and kept at −20° C. until analysis. After blood sampling, mice were sacrificed by cervical elongation and the tumor recovered. Tumors were weighted, cut into small pieces and transferred in GentleMACS C-type tubes (Miltenyi) containing 3 ml PBS. Tumors were mechanically dissociated applying program “m-imptumor01” (GentleMacs, Miltenyi). After centrifugation at 300 g for 7 min, supernatants were recovered and kept at −20° C. until analysis.

(65) J43 concentration was measured by quantitative ELISA both in serum and in tumor homogenates at different time points after virus or commercial J43 mAb injections.

(66) Quantitative ELISA

(67) J43 concentrations was evaluated both in serum and tumor homogenates. More specifically, ninety six wells plates (Nunc immune plate Maxisorp) were incubated overnight at 4° C. with 100 μL/well of 0.8 μg/mL of goat anti-hamster IgG (Southern Biotech) in coating solution (0.05 M Na carbonate pH 9.6, Sigma). Plates were washed three times with wash buffer (300 μL/well of PBS, 0.05% Tween20) and incubated 1 h at RT with 200 μL/well of blocking solution (PBS, 0.05% Tween 20, 5% Non Fat Dry Milk). Plates were washed three times with washing buffer. A standard range of J43 (BioXcell) was prepared in 100% mouse serum (Sigma) from 1000 to 0.488 ng/mL by 2 fold serial dilutions. Each standard was then further diluted 2 fold in blocking solution (final J43 concentration from 500 to 0.244 ng/mL) to have a final concentration of serum of 50%. One hundred μL/well of each standard were added in duplicate on the plate. Samples of serum were diluted at least two fold in blocking buffer, and if necessary further diluted in 1Vol/1Vol mix of blocking buffer/serum and 100 μL were added to the plates. The plates were incubated 2 h at 37° C. After 3 washes with wash buffer, 100 μL/well of HRP conjugated goat anti Armenian Hamster IgG (Jackson Immunoresearch) at 80 ng/mL were added and plates were incubated 1 h at 37° C. After 3 washes, 100 μL/well of 3,3′,5,5′-tetramethylbenzidine (TMB, Sigma) were added and the plates incubated at RT for 30 min. The reaction was stopped with 100 μL/well of 2 M H.sub.2SO.sub.4 and the absorbance was measured at 450 nm with a plate reader (TECAN Infinite M200 PRO). The OD values obtained were transferred to the software GraphPadPrism and samples concentrations were back-calculated using the standard curve fitted with 5 parameters.

(68) As shown in FIG. 6A, when injected subcutaneously in a naive mice (no tumors implanted), production of recombinant J43 from WRTG18618 was very weak (9 ng/ml at day 1 and 4.5 ng/m1 at day 5). In contrast, high quantities of recombinant J43 were secreted in serum of mice bearing B16F10 tumors following intratumoral WRTG18618 injection, respectively 9 and 1900 times higher at D1 and D5 (83 ng/ml and 9500 ng/ml) than the concentration of recombinant J43 measured after subcutaneous virus injection. This result suggests that the oncolytic virus is multiplying preferentially in tumors rather than in subcutaneous healthy tissues. In comparison, seric concentrations found after intratumoral injection of 10 μg commercial J43 reached around 1000 ng/m1 (1306 ng/ml at D1 and 1049 ng/ml at D5) whereas the seric concentrations were much lower with 1 μg of commercial J43 (73 ng/ml at D1 and 35 ng/ml at D5).

(69) In tumor homogenates, the production of recombinant J43 followed the same trend as in serum with a peak of concentration at D5 (see FIG. 6B). The pharmacokinetics of J43 after a single IT injection of commercial mAb or WRTG18618 are very different. Following IT injection of 10 μg of commercial J43, the maximum antibody concentration was measured in serum and in tumor at D1 (1306 ng/ml and 173 ng/ml respectively) as expected, whereas IT injection of WRTG18618 resulted in maximum seric and tumor concentrations at D5 (9500 ng/m1 and 3800 ng/ml, respectively). It should be highlighted that tumor concentrations of recombinant J43 produced after virus IT injection were much higher than those measured after IT injection of 10 μg commercial J43 at all time points (410 ng/m1 versus 173 ng/ml at D1; 3800 ng/m1 versus 51 ng/m1 at D5 and 700 ng/m1 versus under detection limit at D11). Importantly, antibody production was still detected in tumors at D11 after virus treatment, which is not the case after commercial J43 injection.

(70) These data suggest that vectorization of J43 in oncolytic viruses allow a higher and longer accumulation of antibody in tumors compared to IT injection of commercial antibody.

(71) Antitumor Activity in Subcutaneous Tumor Model

(72) Antitumor activity of the various anti-PD-1-expressing vaccinia viruses described above may be tested in conventional preclinical models after implantation of tumors followed by injection of the constructs. For example, murine cancer cells are injected subcutaneously into the flanks of immunocompetent mice. When tumors reached a volume of 50-70 mm.sup.3, the mice are randomized in a blinded manner and treated with the indicated vaccinia virus at a dose of 1×10.sup.7 PFU. The vectors or control vehicle (buffer used to resuspend the virus) are directly injected into the tumor at days 10, 12 and 14 after tumor implantation. Tumor size is measured twice weekly using callipers. In particular, vectorised scFv and mAb1 antibody are at least as efficient as the co-administration of a WR vector with commercial J43 to delay tumor growth and increase survival rate.

(73) Those skilled in the art will recognize, or be able to ascertain using no more than routine experimentation, numerous equivalents to the specific method and reagents described herein, including alternatives, variants, additions, deletions, modifications and substitutions. Such equivalents are considered to be within the scope of this invention and are covered by the following claims.