Fig 2: [68Ga]DPA PET imaging of PD-L1 overexpressed U87MG tumor. (A) U87MG cell uptake of radioactivity when incubated with [68Ga]DPA (20 µCi/mL) or [68Ga]DPA (20 µCi/mL) + BMS202 (100 µmol/L) for different periods. (B) Competitive binding of [68Ga]DPA (20 µCi/mL) to U87MG cells co-incubated with different concentrations of BMS202 for 120 min at 37 °C. Ki value is shown in the panel. (C) PET-CT images of U87MG bearing BALB/c nude mice (left panel) and Bon-1 bearing BALB/c nude mice (right panel) treated with approximately 0.5 mCi of [68Ga]DPA. Mice were imaged at 30 min and 60 min after intravenous injection of [68Ga]DPA. Representative MIP (top panel) and transverse PET-CT images (bottom panel) are shown. The white dashed circles indicate the tumor positions. (D) and (E) Whole-section immunohistochemical staining of PD-L1 in U87MG and Bon-1 tumors. (F) Enlarged view of the selected areas in (D) and (E). Scale bar, 100 µm. (G) H&E staining of U87MG and Bon-1 tumor tissues. Scale bar, 100 µm. (H) Biodistribution of [68Ga]DPA in tissue from U87MG bearing BALB/c nude mice measured by gamma counter ex vivo after 5, 30, 60, and 120 min intravenous injection of the tracer. n = 4 mice per time point. (I) Ratios of tissue uptake of [68Ga]DPA. Values were calculated based on the ex vivo biodistribution data in (G). Data are mean ± SD, n = 4.

Fig 3: Polybasic residues mediate the membrane interaction of PD-L1-CD in vitro and in cells.a Superimposed 2D 1H-15N TROSY-HSQC spectra of CD260-2903RE mutant in solution (red) and with DMPG/DH6PC bicelles (blue). The spectra were recorded at 1H frequency of 600 MHz using [15N, 1H]-labeled protein. b Wildtype and mutant CD260-290 association with liposomes (n = 3 independent experiments). The liposome binding assay was performed with (labeled as “liposome”) or without liposomes (labeled as “No liposome”) as in Fig. 1a but for CD260-290WT and CD260-2903RE mutant. The fractions after centrifugation were analyzed by SDS-PAGE, in which “S” represents the supernatant and “P” represents the pellet. c, d Localization of wildtype and mutant CD260-290 in HEK293T cells. Confocal microscopy images of HEK293T cells expressing CD260-290WT-GFP (c) or CD260-2903RE (d) protein are shown. The experiment was repeated twice with similar results. Scale bars, 5 µm. e Plasma membrane interaction of wildtype and mutant PD-L1 detected by FRET. The FRET efficiency was determined for RKO cells expressing PD-L1WT-mTFP (left top), PD-L13RE-mTFP (right top) and PD-L13RE-mTFP treated with 5 µM PS (left bottom) for 2 h. Scale bars, 10 µm. A similar whisker-dot plot shows the statistical results from n = 24, 22, or 25 WT or 3RE mutant-expressing cells, respectively (right bottom). Data are represented as mean ± SD. P values were determined by unpaired two-sided Student’s t-test. Source data are provided as a Source Data file.

Fig 4: Acidic phospholipids are required for recruitment of PD-L-CD into model and native membranes.a CD260-290 was incubated without or with DMPC liposomes supplemented with DMPE, DMPG, DMPS, or CL as indicated on the bottom. The fractions from the pellet (P) and the supernatant (S) were separated by ultracentrifugation and analyzed by SDS-PAGE. Two independent experiments were performed with similar results. b Superimposed 1H-15N TROSY-HSQC spectra of CD260-290 in a solution buffer (25 mM MES pH 6.5) (blue) and in the same solution plus DMPC/DH6PC (green) or DMPG/DH6PC (red) bicelles. The results are representative of three independent experiments. c Schematic illustration of the FRET assay for measuring the interaction between mTFP (blue) fused to the C-terminus of PD-L1 and R18 dye (red) incorporated into the plasma membrane. The membrane contains acidic (pink) and other (green) lipids. The basic patch in PD-L1-CD is indicated by + symbols. d FRET detected lipid effects on PD-L1-CD-membrane interaction. Fluorescent images of the cells treated with DMSO, 5 µM PS or 2 µM Sph were taken to detect the mTFP or R18 dye emission before or after photobleaching (BP or AP) the R18 dye, respectively. Scale bars = 10 µm. The FRET statistical results from cells treated with DMSO (n = 24), PS (n = 18) or Sph (n = 17) are shown as the mean ± standard deviation (SD). Statistical differences were determined by unpaired two-sided Student’s t-test. e Immunofluorescence detection of endogenous PD-L1 in RKO cells treated with DMSO, 5 µM PS or 2 µM Sph. Scale bars = 5 µm in all panels. The fluorescence intensities from the cells were quantified and shown as mean ± SD from n = 20 cells under each treatment. P values were determined by two-sided Student’s t-test. f Cell surface levels of PD-L1 in RKO cells analyzed by flow cytometry after treatment with DMSO, 5 µM PS or 2 µM Sph for 6 h. g PD-L1 level (left) and ubiquitination (right) in RKO cells analyzed by western blotting. RKO cells were cultured for 12 h with or without 5 µM PS or 2 µM Sph. Two independent experiments were performed with similar results. Source data are provided as a Source Data file.

Fig 5: Metformin prevents PD-L1-CD-membrane association to induce PD-L1 degradation.a Metformin reduces CD260-290 interaction with DMPG/DH6PC bicelles. Superimposed 2D 1H-15N TROSY-HSQC spectra of CD260-290 in the bicelles (red) titrated with metformin at molar ratios of [metformin]:[DMPG] = 0.05 (orange), 0.1 (green), 0.2 (light blue), 0.4 (blue), 0.6 (dark blue), 1.0 (magenta), and 2.0 (purple). The addition of metformin shifts the correlations toward that of CD260–290 in solution (black). The right panels 1–2 show the same spectral regions labeled on the full spectrum, highlighting the chemical shift changes for G273 and K270, respectively. b The comparison of chemical shift changes of CD260–290 in DMPG/DH6PC bicelles between molar ratios of [metformin]:[DMPG] = 0 and 2.0. c Metformin decreases CD260–290 insertion into DMPG/DH6PC bicelles. PRE of CD260-290 spectral signals induced by the paramagnetic probe 16-DSA (2.5 mM) localized in the bicelle membranes were measured when the sample was titrated with metformin. The intensity recovery of broadened peaks for G273 (top) and K270 (bottom) at molar ratios of [metformin]:[DMPG] = 0, 0.2, 0.4, 0.6, 0.8, 1.0, and 2.0 are shown. d FRET-detected effect of metformin on PD-L1-CD-membrane interaction. RKO cells expressing PD-L1-mTFP were treated with 5 mM metformin for 2 h. Fluorescent images of representative cells from the vehicle (DMSO) control (left) and the metformin treated (middle) samples are shown. The statistical results of FRET efficiency are represented as the mean ± SD for n = 24 or 26 cells from the control and treated samples, respectively (right). P values by unpaired two-sided Student’s t-test are indicated. Scale bars, 5 µm. e Cellular level of PD-L1WT or PD-L13RE mutant. HEK293T cells expressing PD-L1WT (top) or PD-L13RE mutant (bottom) were treated with the indicated concentrations of metformin for 24 h and analyzed by western blot. Similar results were obtained from two other independent experiments. f Surface level of PD-L1WT or PD-L13RE mutant in HEK293T cells determined by flow-cytometric analysis. After metformin (5 mM) treatment for 24 h, the HEK293T cells expressing PD-L1WT (left) or PD-L13RE mutant (right) were analyzed by flow-cytometry using anti-PD-L1 antibody with IgG, immunoglobulin G, as the control. Source data are provided as a Source Data file.