Nanobody-mediated SPECT/CT imaging reveals the spatiotemporal expression of programmed death-ligand 1 in response to a CD8+ T cell and iNKT cell activating mRNA vaccine

Rationale: Although promising responses are obtained in patients treated with immune checkpoint inhibitors targeting programmed death ligand 1 (PD-L1) and its receptor programmed death-1 (PD-1), only a fraction of patients benefits from this immunotherapy. Cancer vaccination may be an effective approach to improve the response to immune checkpoint inhibitors anti-PD-L1/PD-1 therapy. However, there is a lack of research on the dynamics of PD-L1 expression in response to cancer vaccination. Methods: We performed non-invasive whole-body imaging to visualize PD-L1 expression at different timepoints after vaccination of melanoma-bearing mice. Mice bearing ovalbumin (OVA) expressing B16 tumors were i.v. injected with the Galsome mRNA vaccine: OVA encoding mRNA lipoplexes co-encapsulating a low or a high dose of the atypical adjuvant α-galactosylceramide (αGC) to activate invariant natural killer T (iNKT) cells. Serial non-invasive whole-body immune imaging was performed using a technetium-99m (99mTc)-labeled anti-PD-L1 nanobody, single-photon emission computerized tomography (SPECT) and X-ray computed tomography (CT) images were quantified. Additionally, cellular expression of PD-L1 was evaluated with flow cytometry. Results: SPECT/CT-imaging showed a rapid and systemic upregulation of PD-L1 after vaccination. PD-L1 expression could not be correlated to the αGC-dose, although we observed a dose-dependent iNKT cell activation. Dynamics of PD-L1 expression were organ-dependent and most pronounced in lungs and liver, organs to which the vaccine was distributed. PD-L1 expression in lungs increased immediately after vaccination and gradually decreased over time, whereas in liver, vaccination-induced PD-L1 upregulation was short-lived. Flow cytometric analysis of these organs further showed myeloid cells as well as non-immune cells with elevated PD-L1 expression in response to vaccination. SPECT/CT imaging of the tumor demonstrated that the expression of PD-L1 remained stable over time and was overall not affected by vaccination although flow cytometric analysis at the cellular level demonstrated changes in PD-L1 expression in various immune cell populations following vaccination. Conclusion: Repeated non-invasive whole-body imaging using 99mTc-labeled anti-PD-L1 nanobodies allows to document the dynamic nature of PD-L1 expression upon vaccination. Galsome vaccination rapidly induced systemic upregulation of PD-L1 expression with the most pronounced upregulation in lungs and liver while flow cytometry analysis showed upregulation of PD-L1 in the tumor microenvironment. This study shows that imaging using nanobodies may be useful for monitoring vaccine-mediated PD-L1 modulation in patients and could provide a rationale for combination therapy. To the best of our knowledge, this is the first report that visualizes PD-L1 expression upon cancer vaccination.

Figure S1: Flow cytometry gating strategy.Overview of the performed gating strategy.(A) Gating strategy for viable cells was performed for all flow cytometry panels.Viable cells were selected by gating all cells, single cells and viable cells.(B) Myeloid panel.Myeloid cells were selected by gating immune cells and CD11b + -cells and were subdivided into different populations: granulocytes (F4/80 lo Ly6G hi ), macrophages (F4/80 hi ) and monocytes (F4/80 lo Ly6G lo ).Dendritic cells (DCs) were selected via gating on CD11c hi and MHCII hi cells in the immune cell population.(C) Lymphoid panel.T lymphocytes were selected by gating immune cells and CD3 + -cells and were subdivided into CD4 + and CD8 + T cells.OVA-specific CD8 + T cells were selected via gating on OVA tetramer hi and CD8 hi cells in CD3 + -cell population.(D) iNKT panel.iNKT cells were selected by gating CD1d hi and TCRβ hi cells in immune cell population (E) PD-L1 expression was defined for each individual cell type.

Figure S2 :
Figure S2: B16-OVA cells present OVA in MHC-I and are recognized by OVA-specific CD8 + T cells.(A) Representative flow cytometry histograms showing expression of MHC-I or MHC-I/OVA complexes on the surface of B16 (blue), or B16-OVA (grey) cells cultured with or w/o IFN-γ (n=3).(B,C) OVA-specific CD8 + T cells were isolated from the spleen of OT-I mice and co-cultured with B16 cells, B16 cells pulsed with the MHC-I restricted peptide of OVA or B16-OVA cells for 24 h.As technical controls, OT-I cells were cultured w/o further stimulation or were stimulated with aCD3/CD28 antibody-coated beads.(B) Flow cytometry was performed to evaluate the percentage of cells that produce IFN-γ upon coculture of OT-I splenocytes and B16 variants (n=3).(C) ELISA was performed to evaluate the amount of IFNγ in the supernatants upon coculture of OT-I splenocytes and B16 variants (n=3).The bar graphs in (B,C) summarize the results as mean ± SD.Symbols represent individual data points.

Figure S4 :
Figure S4: PD-L1 expression in lung, liver and tumor assessed by measuring PD-L1 + cells shows the same trends as observed for PD-L1 MFI values.(A, D, G) Flow cytometry results represented as PD-L1 + cells in the viable cell population of (A) lung, (D) liver and (G) tumor.(B, E, H) Flow cytometry results represented as PD-L1 + cells in the non-immune cell population of (B) lung, (E) liver and (H) tumor.(C, F, I) Flow cytometry results represented as PD-L1 + cells in the immune cell population of (C) lung, (F) liver and (I) tumor.All data are shown as mean ± SD (n=1, mpc=3).

Figure S5 :
Figure S5: Proliferation of OVA-specific CD8 + T cells and iNKT cells in lung and liver of treated and untreated mice.Graphs showing percentage of OVA-specific CD8+ T cells (A) and iNKT (B) within the CD45 + population in the lung detected on days 1, 4, 7, 10 and 14 as mean ± SD.Statistical analysis was performed by two-way ANOVA followed by Tukey's post hoc test.(n=1, mpc=3).

Figure S6 :
Figure S6: PD-L1 expression at cellular level in tumor 1 day after vaccination.(A) Fraction of PD-L1+ (CD45 -) cells, DCs, macrophages (Mφ), granulocytes, monocytes, CD4 + and CD8 + T cells within viable cells 1 day after vaccination with Galsomes containing a low (blue) or high (green) αGC-dose or PBS (white) in tumor.(B) Fraction of non-immune cells and different immune cell types within viable cells 1 day after vaccination.Symbols represent individual data points and mean is indicated by a line.Statistical analysis was performed by two-way ANOVA followed by Tukey's post hoc test.(n=1, mpc=3).

Table S3 :
Model Summary of 1

Table S4 :
Pairwise comparison of PD-L1 signal in liver between treatment conditions adjusted with TUKEY.  +     +

Table S5 :
Mixed model summary of 1   based on imaging data.Contrary to the data in liver and lung the mixed model in tumor present better fitting than an only fixed effect model.

Table S6 :
Pairwise comparison of PD-L1 signal in tumor tissue between treatment conditions adjusted with TUKEY.