Fig 2: Characterization of anti-4-1BB trimerbodies. a Sensorgrams (black curves) and fitting curves for 1D8 antibodies (2 and 4 nM), obtained by biolayer interferometry with surface-immobilized m4-1BB. b The binding to 4-1BB on the cell surface of stimulated mouse CD8a+ T cells measured by FACS. c–e Costimulatory activity of anti-4-1BB antibodies. Mouse CD8a+ T cells were stimulated with immobilized anti-CD3 mAb in the presence of m4-1BBL, 1D8N5, 1D8N18, or 1D8 IgG, and proliferation (c) and secretion of IFN-γ (d) were measured after 48 h and cell viability (e) after 72 h. Data are reported as fold change relative to the values obtained from anti-CD3 mAb-stimulated CD8a+ T cells. Rat IgG2a and MFE-23N18 were used as controls. Data are mean ± SD (n = 3), *P ≤ 0.05, **P ≤ 0.01, Student's t test. f RICS analyses performed in living HEK293m4-1BB-S cells at regions containing clusters formed upon 1D8 IgG or 1D8N18 addition and at regions where clusters where not present (insert and zoomed-in regions ii, and i, respectively). Representative maximum intensity projection maps showing the RICS analyzed regions of interest. Values in the zoomed-in regions show the diffusion coefficient of bound antibody. The color heat map indicates in blueish tones the lower intensity and in redder tones the higher intensity. g Statistical analysis of the quantified diffusion coefficient obtained from 5 to 7 independent live cell experiments and 3–5 different regions of interest per cell (N-CR non-clustered region, CR clustered region). Data are presented as median (center line), upper and lower quartiles (boxes), and minimum-maximum (whiskers). h Arrangement of 1D8N18 trimerbody in solution, as determined by SAXS. Rigid-body overlaying of the ab initio-determined SAXS envelope for 1D8N18. The generated model (where each chain is colored in blue, magenta, or cyan) fits into the envelope (colored in pale gray)

Fig 3: Modulation of the immune compartment in GL261 tumor-bearing mice after treatment with Delta-24-ACT in combination with PD-L1 antibody. Analysis of tumor-infiltrating lymphocyte populations in the brains of mice bearing GL261 tumors treated with PBS (control), aPD-L1, Delta-24-ACT or Delta-24-ACT/aPD-L1. The studied populations were CD45+ cells (A), myeloid cell populations, including macrophages, DCs and microglia (B), and lymphoid cell populations, including CD8+, CD4+, Tregs and NK cells (C). The analyses of the functional status of the myeloid cells described above included analyses of PD-L1 (D), CD86+MHCI+I (E), CD137L (F) markers and the percentage of M1 macrophages (G). Percentage of DCs present in the lymph nodes. (H) DC functional status. The CD40, CD86 and PD-L1 markers were analyzed (I). Phenotype of CD8+ and CD4+ T cells (J) based on studies of the CD137, GITR, PD-1 and GrzB markers. (K) Quantification of splenocyte IFN-gamma production by ELISPOT after the indicated treatment. Top panel: mean counted spots. Lower panel: mean spot size (one-way ANOVA).

Fig 4: Pharmacokinetic properties and tumor imaging of the EGFR-targeted 4-1BB-agonistic trimerbody. Pharmacokinetic studies after a single i.v. dose (1 mg/Kg) of 1D8N18 or 1D8N/CEGa1 in CD-1mice (a) or of 3H3 IgG or 1D8N/CEGa1 in BALB/c mice (b). Pharmacokinetic study after a single i.p. dose (1 mg/Kg) of 3H3 IgG or 1D8N/CEGa1 in BALB/c mice (c). d In vivo fluorescence imaging of A431 tumor-bearing nude mice 24 h after i.v. injection of PBS or 100 µg of Cy5-labeled cetuximab, CF647-labeled 3H3 IgG or CF647-labeled 1D8N/CEGa1. e Tumor to normal tissue (T/N) ratios. The color scale bar represents the fluorescence intensity recorded as photons per second per centimeter squared per steradian (p/s/cm2/sr). Data are mean ± SD (n = 3), **P ≤ 0.01, ***P ≤ 0.001, Student's t test

Fig 5: Characterization of the EGFR-targeted 4-1BB-agonistic trimerbody. a Arrangement of 1D8N/CEGa1 trimerbody in solution by SAXS. Rigid-body fitting of the model corresponding to 1D8NCEGa1 inside the SAXS envelope (colored in pale gray). Each chain has been colored in blue, magenta, or cyan. b Sensorgrams (black curves) and the results of fitting to a 1:1 model (red curves) obtained using biolayer interferometry for the interaction of 1D8N/CEGa1 (2 and 4 nM) with immobilized m4-1BB and the interaction of 1D8N/CEGa1 (0.5 and 1 nM) with immobilized hEGFR. c Simultaneous binding to both m4-1BB and hEGFR was demonstrated for 1D8N/CEGa1 but not 1D8N18. Biosensors were coated with m4-1BB, after which 4 nM of 1D8N/CEGa1 (black curves) or 1D8N18 (blue curves) were added. d The binding of anti-4-1BB antibodies to m4-1BB on the cell surface of stimulated mouse CD8a+ T cells measured by FACS. e, –f Mouse CD8a+ T cells were plated with immobilized anti-CD3 mAb and hEGFR or BSA in the presence of 1D8N18, 1D8N/CEGa1, or 1D8 IgG, and proliferation (e) and IFN-γ secretion (f) were determined after 48 h. EGFR-negative 3T3 cells or EGFR-positive 3T3hEGFR cells were co-cultured with mouse CD8a+ T cells in the presence of anti-CD3 mAb and 1D8N18, 1D8N/CEGa1, or 1D8 IgG. IFN-γ secretion analyzed after 48 h (g) and cell viability after 72 h (h). Data are represented as fold change relative to the values obtained from anti-CD3 mAb stimulated cells. Rat IgG2a and MFE-23N18 were used as negative controls. Data are mean ± SD (n = 3), *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, Student's t test