Cell culture
Jurkat (clone E6.1) and MDA-MB-231, or HEK293T were cultured respectively at 37°C with 5% CO2 in RPMI media supplemented with 5% foetal calf serum (FCS; Pan Biotech), 55 µM β-mercaptoethanol, 100 U/mL penicillin and 100 µg/mL streptomycin (unless otherwise stated), or 10% FCS/DMEM media supplemented with 100 U/mL penicillin and 100 µg/mL streptomycin (unless otherwise stated). Expi293F cells were cultured at 37°C with 8% CO2 and orbital shaking (130 rpm) in Expi293F expression media (ThermoFisher).
Constructs
Anti-human TF scFv sequences were derived from the murine mAb clone TF8-5G9 and the murine-humanised mAb clone hATR-5. Plasmid insert gene constructs and DNA sequences are listed in supplementary data.
Protein expression and purification
Planktonic Expi293F cells (37.5 × 106 in 15 mL) were transfected with 15 µg of pcDNA3.1(-):TF85G9 scFvFc, hATR-5 scFvFc, or human TF-Fc (Genscript), or pSBbiGP:OKT3-TF85G9 BiTE (His-tagged; OKT3 VH-VL and TF85G9 VH-VL joined via glycine-serine linkers; IDT geneblock) using the Gibco Expifectamine293 transfection kit as per the manufacturer’s instructions (Gibco #A14524). Cells were incubated until cell viability decreased to approximately 30%, after which cell supernatant was centrifuged at 450 ×g for five minutes, followed by 2000 ×g for 20 min, and supernatant 0.2 µm filtered. Fc-fusion proteins were purified by Protein A chromatography (Pierce #20356), eluted in 500 µL fractions of 100 mM glycine (pH 3.0) and neutralised with 100 µL of 1 M Tris-HCl (pH 9.0). His-tagged proteins were purified by Protino Nickel-NTA agarose column (MN745400.25) as per the manufacturer’s instructions, prior to dialysis with phosphate buffered saline (PBS). For SPR, fusion proteins were further purified by size exclusion chromatography (Superdex 200 10/300 GL column) with PBS-EP (0.01% Tween20, 300 mM NaCl PBS, pH 7.4). In brief, proteins were centrifuged at 15,000 ×g to remove precipitate, followed by incubation with 2 mM DTT for 30 min on ice. Proteins were then injected in the column and run at a flow rate of 0.5 mL/min and 500 µL fractions collected. Boiled and reduced proteins were run on a 4-12% BOLT Bis-Tris gel, before stained with Coomassie blue.
SPR
SPR experiments were carried out using a Biacore X100 instrument at 25°C. Human TF-Fc was immobilised to the surface of a CM5 sensor chip (target of 500 response units) in 10 mM sodium acetate pH 4.0 with a flow rate of 10 µL/min. Serial dilutions (18.75 – 3.125 nM) of TF8-5G9 scFvFc were made with PBS-EP and run over the sensor chip for two cycles (flow rate 30 µL/min; contact time 180 s; dissociation time 600 s). Sensograms were globally fitted with a 1:1 binding model and binding kinetics determined with BIAevaluation software.
CAR generation
Second generation CAR constructs (scFv-c-myc tag-CD8 hinge-CD28-CD3ζ) were generated from the variable fragment sequences of anti-human tissue factor mAb TF8-5G9 and hATR-5. A sleeping beauty transposon system was utilised for transfection of the CAR cassette into Jurkat T cells. Plasmids (9 µg of pSBbiGP (or RP where stated) and 1 µg of pCMV(CAT)T7-SB100) were introduced into cells via electroporation with a neon transfection system with 100 µL gold tips (1350 V, 10 ms, three pulses) and cultured with antibiotic free media for at least 48 h, before stably transfected cells selected for with 2 µg/mL puromycin.
In vitro CAR assay
TF8-5G9 or hATR-5 CAR T cells were incubated with 2 µg/mL anti-c-myc-biotin (Biolegend #908805) in 0.1% BSA/PBS/2 mM EDTA for 30 min on ice, biotin was then detected with 4 µg/mL streptavidin-allophycocyanin (Biolegend #405207) and incubated as previously described. Cells were analysed for GFP or RFP (as stated) and c-myc expression with a BD LSR Fortessa, and subject to both FSc and SSc doublet discrimination prior to further analysis with FlowJo v10.7.2.
A 24-well plate was seeded with TF-expressing cell line MDA-MB-231 (1 × 106 cells/mL; 500 μL/well) in 5% FCS/RPMI for 24 h at 37°C with 5% CO2. Alternatively, wells were coated with human TF-Fc fusion protein (10 – 1.25 µg/mL; 300 μL/well) overnight at 4°C, wells were then rinsed twice with PBS. CAR-T or untransfected Jurkat cells (1 × 106 cells/mL; 600 μL/well) in 5% FCS/RPMI were then added to the plates and incubated at 37°C with 5% CO2 for 24 h. Positive control 25 ng/mL PMA / 0.75 µg/mL ionomycin, or negative controls as stated were also included. Cell culture supernatant was analysed by anti-human IL-2 ELISA.
In vitro BiTE assay
HEK293T cells (2 × 105) were seeded in 24-well plates in 500 μL of antibiotic-free media and incubated at 37°C with 5% CO2. At 24 h, HEK293T were transiently transfected with a total of 500 ng of plasmid DNA (pcDNA3.1(+): Human full-length TF, pSBbiGP: OKT3-TF85G9 BiTE, pSBbiGP: empty vector, or a 1:1 ratio of two plasmids where stated) using lipofectamine 3000 as described by the manufacturer. Media was changed and Jurkat cells (1 × 106 cells/mL; 500 μL) were added in 10% FCS/RPMI at 16 h post-transfection. Positive control 25 ng/mL PMA / 0.75 µg/mL ionomycin, or negative controls R10 alone. Cells were further incubated at 37°C with 5% CO2 for 24 h before culture supernatant was removed for IL-2 analysis.
IL-2 ELISA
Capture antibody anti-human IL-2 (2 µg/mL; BD #555051) in PBS was coated overnight at 4°C. Wells were washed three times with 0.05% Tween20/PBS, prior to blocking with 1% BSA/PBS for ten minutes at room temperature and blocking solution flicked out. An IL-2 standard or cell supernatant was added to the wells and incubated overnight at 4°C. Wells were washed as previously described and incubated with biotinylated anti-human IL-2 (1 µg/mL; BD #555040) diluted in 1% BSA/PBS for one hour at 37°C. After washing, wells were incubated with streptavidin horse radish peroxidase (1-in-5000 dilution; Roche #11089153001) for one hour at 37°C. Wells were washed and developed in 3,3’,5,5;-tetramethylbenzidine (Novex #00-2023), the reaction was then stopped by the addition of 2N H2SO4 and the plate read at 450 nm.
Results
A number of solid tumours express high TF levels11,13–16, we therefore aimed to test the utility of known anti-human TF monoclonal antibodies in the context of a CAR T cell or BiTE. To ensure the binding capacity of TF8-5G9 was not lost when the variable regions of this mAb were expressed as a scFvFc construct31, we expressed the TF8-5G9 scFv and its antigen human TF as Fc fusion proteins (Fig. 1A). TF8-5G9 retained a high binding affinity (Kd = 0.37 nM) as determined by SPR (Fig. 1B). We next constructed a second generation CD3/CD28 CAR with c-Myc tag by utilising the TF8-5G9 scFv. This was cloned into a sleeping beauty transposon system with SfiI restriction sites32, prior to its transfection into the Jurkat (Clone E6.1) cell line that is known to express high levels of IL-2 upon CD3/CD28 signalling. Successful transfection and expression were observed via the upregulation of the surrogate marker GFP, as well as the c-Myc tag within the CAR construct (Fig. 2A).
The TF8-5G9 CAR showed a dose-dependent response upon exposure to HuTF with increasing levels of IL-2 detected (Fig. 2B). The TF8-5G9 CAR response could be neutralised by the inclusion of TF8-5G9 scFvFc emphasising the CAR’s specificity. Furthermore, no cross-species reactivity was noted in the presence of murine TF (Fig. 2C). Triple negative breast cancer cell line MDA-MB-231 has previously been shown to express TF13, therefore we tested the CAR T cell capacity to recognise endogenous levels of TF. A clear response was observed with approximately 5 ng/mL IL-2 detected following incubation of TF8-5G9 CAR T cells with MDA-MB-231 (Fig. 2D).
In a similar manner to the TF8-5G9 CAR construct, we investigated the use of a second anti-human TF monoclonal antibody, hATR-5.33 hATR-5 scFv was found to have a high binding affinity (Kd = 0.46 nM) by SPR (Fig. 3A). The hATR-5 scFv when incorporated into a CAR construct was both surface expressed as shown by flow cytometric analysis of c-Myc expression, as well as functional, with a clear response against both solid-phase TF, as well as endogenously expressed TF on MDA-MB-231with IL-2 detected by ELISA (Fig. 3B and 3C).
To investigate the potential of utilising TF within a dual target system, we generated a BiTE that binds both human TF (TF8-5G9) and CD3 (OKT3).34 This BiTE was soluble and able to be purified from Expi293F secreting cells, however corresponding with previous studies a low yield indicated relatively low expression (data not shown). HEK293T cells were transiently transfected HEK293T to express both the full-length TF antigen, as well as the anti-TF/CD3 BiTE. In the presence of Jurkat T cells, IL-2 was released indicating that the anti-TF/CD3 BiTE was functional (Fig. 4).