Specific Cell Targeting by Toxoplasma gondii Displaying Functional Single-Chain Variable Fragment as a Novel Strategy; A Proof of Principle

Abstract

Toxoplasma gondii holds significant therapeutic potential; however, its nonspecific invasiveness results in off-target effects. The purpose of this study is to evaluate whether T. gondii specificity can be improved by surface display of scFv directed against dendritic cells' endocytic receptor, DEC205, and immune checkpoint PD-L1. Anti-DEC205 scFv was anchored to the T. gondii surface either directly via glycosylphosphatidylinositol (GPI) or by fusion with the SAG1 protein. Both constructs were successfully expressed, but the binding results suggested that the anti-DEC-SAG1 scFv had more reliable functionality towards recombinant DEC protein and DEC205-expressing MutuDC cells. Two anti-PD-L1 scFv constructs were developed that differed in the localization of the HA tag. Both constructs were adequately expressed, but the localization of the HA tag determined the functionality by binding to PD-L1 protein. Co-incubation of T. gondii displaying anti-PD-L1 scFv with tumor cells expressing/displaying different levels of PD-L1 showed strong binding depending on the level of available biomarker. Neutralization assays confirmed that binding was due to the specific interaction between anti-PD-L1 scFv and its ligand. A mixed-cell assay showed that T. gondii expressing anti-PD-L1 scFv predominately targets the PD-L1-positive cells, with negligible off-target binding. The recombinant RH-PD-L1-C strain showed increased killing ability on PD-L1+ tumor cell lines compared to the parental strain. Moreover, a co-culture assay of target tumor cells and effector CD8+ T cells showed that our model could inhibit PD1/PD-L1 interaction and potentiate T-cell immune response. These findings highlight surface display of antibody fragments as a promising strategy of targeting replicative T. gondii strains while minimizing nonspecific binding.

Keywords: PD-L1; Toxoplasma gondii; cancer immunotherapy; immune checkpoint; scFv; surface display; targeting.

Figure 1

Engineering and characterization of recombinant T. gondii expressing a single-chain variable fragment directed against murine DEC205 (anti-DEC205 scFv). (A) Schematic representation of the anti-murine DEC205 with N-terminal HA-tagged scFv (DC2) and the anti-murine DEC205 fused to SAG1 with N-terminal HA-tagged scFv (DC2-SAG1). The locations of the Kozak sequence and start codon (ATG), signal sequences of the N-terminus surface antigen 1 (SS-SAG1) and glycosylphosphatidylinositol (GPI), human influenza hemagglutinin (HA) tag, antibody variable domains (heavy [VH] and light [VL]), peptide linkers ([Gly4Ser]3 and [Gly3AlaSer]), and surface antigen 1 sequence (SAG1) are indicated. (B) Screening based on HA expression by ELISA of different selected clones of RH-DC2 and RH-DC2-SAG1, using rabbit anti-HA followed by AP-conjugated anti-rabbit IgG. RH tachyzoites were used as a control (n = 3 replicates). (C) Western blot of parasite cell lysates of RH-DC2-SAG1 (line 1), RH wild-type (line 2), and RH-DC2 (Line 3) revealed using rabbit anti-HA followed by AP-conjugated anti-rabbit IgG. MW: molecular weight marker in kilo Dalton. (D) Binding of RH, RH-DC2, and RH-DC2-SAG1 tachyzoites to CF14 recombinant protein assessed by ELISA using T. gondii antibody from infected rabbit serum (n = 3 replicates). (E) Binding of RH (control), RH-DC2, and RH-DC2-SAG1 tachyzoites to murine dendritic cells MutuDC expressing DEC205 assessed by flow cytometer using a mouse monoclonal antibody specific for T. gondii gp23 glycoprotein, followed by APC-conjugated anti-mouse IgG. Binding was measured as the percentage of GFP-positive cells (n = 4 replicates). (F) Dot blot to compare the relative abundance of the HA-tagged proteins expressed by RH-DC2 and RH-DC2-SAG1. Serial 2-fold dilutions of tachyzoite crude lysates were loaded onto nitrocellulose membrane and probed with rabbit anti-HA followed by HRP-conjugated anti-rabbit IgG and detection with chemiluminescent HRP substrate. Statistical significance is indicated by * p < 0.05. In all ELISA experiments, optical densities were read at 405 nm. All experiments were repeated at least three times. AP: Alkaline Phosphatase. HRP: Horseradish peroxidase.

Figure 2

Engineering and characterization of recombinant T. gondii expressing a single-chain variable fragment directed against the human programmed death-ligand 1 (anti-PD-L1 scFv). (A) Schematic representation of the anti-human PD-L1 scFv with HA tag in the N-terminus (anti-PD-L1-N) and the anti-human PD-L1 scFv with HA tag in the C-terminus (anti-PD-L1-C). The locations of the Kozac sequence and start codon (ATG), signal sequences of N-terminus of surface antigen 1 (SS-SAG1) and of glycosylphosphatidylinositol (GPI), human influenza hemagglutinin (HA) tag, antibody variable domains (heavy [VH] and light [VL]), peptide linkers ([Gly4Ser]3 and [Gly3AlaSer]) and surface antigen 1 sequence (SAG1) are indicated. (B) HA expression in three selected clones of RH-PD-L1-N and RH-PD-L1-C analyzed by ELISA using rabbit HA polyclonal antibody followed by anti-rabbit IgG AP-conjugated. RH tachyzoites were used as control (n = 3 replicates). (C) Western blot of parasite cell lysates RH (Line 1), RH-DC-SAG1 (line 2), RH-PD-L1-N (line 3), and RH-PD-L1-C (line 4) revealed using rabbit anti-HA followed by anti-rabbit IgG AP-conjugated. MW: molecular weight marker in kilo Dalton. (D) Binding of histidine-tagged PD-L1 recombinant protein to RH-PD-L1-N and RH-PD-L1-C tachyzoites assessed by ELISA using anti-His tag antibody. RH and RH-DC2-SAG1 tachyzoites were used as control (n = 5 replicates). (E) Dot blot to compare the relative abundance of the HA-tagged proteins expressed by RH-PD-L1-C and RH-PD-L1-N tachyzoites. Serial 2-fold dilutions of tachyzoite crude lysates were loaded onto nitrocellulose membrane and probed with rabbit anti-HA followed by HRP-conjugated anti-rabbit IgG and detection with chemiluminescent HRP substrate. (F) Localization of anti-PD-L1 scFv display on the membrane of T. gondii. Free tachyzoites of RH-PD-L1-C recombinant strain were incubated with mouse SAG1 mAb followed by biotin anti-mouse IgG and AF-594-conjugated streptavidin (red) or rabbit anti-HA followed by anti-rabbit IgG-AF-488 (green). Nucleus was stained using Hoechst (blue). Images were captured at x600 with Olympus IX73 fluorescent microscope. Scale bars represent 5 μm. Statistical significance is indicated by * p < 0.05, ** p < 0.01. ns: not significant. In all ELISA experiments, optical densities were read at 405 nm. All experiments were repeated at least twice. AF: Alexa Fluor.

Figure 3

Binding of RH-PD-L1-C and RH-DC2-SAG1 (irrelevant scFv control) recombinant tachyzoites to human cells expressing PD-L1. (A) PD-L1 expression on HFF and MDA-MB231 cells. Cells were stained with PE-conjugated anti-human PD-L1 and analyzed by flow cytometry. Blue area depicts the profile of cells stained with isotype control IgG, and red area depicts the profile of cells stained with anti-PD-L1. (B) Quantification of binding of T. gondii recombinant tachyzoites to HFF or MDA-MB-231 cells analyzed by flow cytometry. Human cells were incubated with either RH-DC2-SAG1 or RH-PD-L1-C at MOI 5 and binding was indicated by percentages of GFP-positive cells (n = 3 replicates). (C) Binding competition with Atezolizumab. After incubating MDA-MB-231 cells with a saturating concentration of Atezolizumab, followed by incubation of the RH-PD-L1-C or RH-DC2-SAG1 strain, binding was quantified by flow cytometry (n = 5 replicates). Statistical significance is indicated by * p < 0.05, ** p < 0.01, ns: not significant. All experiments were repeated at least twice. MOI: multiplicity of infection.

Figure 4

Binding of RH-PD-L1-C and RH-DC2-SAG1 (irrelevant scFv control) recombinant tachyzoites to murine tumor cell lines expressing PD-L1. (A) Surface display level of PD-L1 on IFN-γ-stimulated B16F10 (B16F10 + IFN- γ), non-stimulated B16F10 (B16F10) and B16K1 cells. Cells were stained with APC-conjugated mouse anti-PD-L1 before flow cytometry analysis. Results are represented as flow cytometer histograms (on the left) and mean fluorescence intensity values (MFI, on the right). Black area indicates the profile of cells incubated with isotype control IgG. Profile of cells stained with anti-PD-L1 are indicated in purple area (B16K1), green area (B16F10) and red area (IFN-γ-stimulated B16F10). (B) Murine cell lines were incubated with either RH-DC2-SAG1 or RH-PD-L1-C tachyzoites at MOI 5, and the percentage of GFP-positive cells was quantified by flow cytometer (n = 3 replicates). (C) Competitive binding assay. IFN-γ- stimulated B16F10 cells were incubated with increased concentrations of Atezolizumab followed by addition of either RH-DC2-SAG1 or RH-PD-L1-C tachyzoites at MOI of 5. Binding was assessed as a percentage of GFP-positive cells by flow cytometry (n = 3 replicates). (D) Selective targeting of PD-L1 over-expressing cells by RH-PD-L1-C in mixed-cell assay. B16K1 and IFN-γ-stimulated B16F10 cells were mixed together at a ratio of 1:1, and the cell mixture was incubated with RH-PD-L1-C or the control strain, RH-DC2-SAG1, at MOI of 5. The two cell populations were discriminated by mCherry expression by gating on mCherry-positive (B16F10) and mCherry-negative (B16K1) cells. Binding was indicated by the GFP-positive cells among each group. Statistical significance is indicated by * p < 0.05, ns: not significant. All experiments were repeated at least twice.

Figure 5

In vitro biological activity. (A) In vitro oncolytic activity. MDA-MB231 and IFN-γ-stimulated B16F10 cells were left untreated or infected with either the wild-type T. gondii RH strain or RH-PD-L1-C recombinant strain at MOI of 3. Cell viability was measured 24, 48 and 72 h post infection using MTT test. Results are represented as percentages of viable cells in respect to the control uninfected cells. Optical densities were read at 490 nm. Asterisks indicate statistically significant differences between RH wild-type-treated and RH-PD-L1-C-treated cells (* p < 0.01). All experiments were repeated at least twice. (B) PD-1 expression on B3Z cells was analyzed by flow cytometer using FITC-labeled mouse anti-PD-1. Blue area indicates the profile of cells incubated with isotype control IgG, while profile of cells stained with anti-PD-1 is indicated in red area. (C,D) Activation of the MHC class I-restricted T-cell line B3Z in the context of presentation of the SIINFEKL OVA peptide or presentation of the soluble OVA peptide SIINFEKL to B3Z cells. B3Z cells were co-cultured for 20 h with IFNγ-stimulated B16F10 cells loaded with OVA peptide in presence of RH (control) or RH-PD-L1-C recombinant tachyzoites. The B3Z response was measured as β-galactosidase activity (C) by measuring absorbance at 570 nm after incubation with β-galactosidase substrate (CRPG) and IL-2 quantification (D) in co-culture supernatants by ELISA. Statistical significance is indicated by * p < 0.05. All experiments were repeated at least twice.