Fig 2: (A–C) Representative images of CTCs isolated from NSCLC patients stained by antibodies against cytokeratin (red), a leukocyte marker CD45 (green), and a nuclear stain (DAPI). (A) CK (+)/PD-L1 (+) CTC, (B) CK (+)/PD-L1 (−) CTC, (C) CTC cluster with WBC. Scale bar is 10 µm. (D) CTC enumeration from blood samples of healthy donors (n = 6) and from blood samples of NSCLC patients (n = 38), (***)P-value < 0.001. (E) Number of CTCs isolated by GO chip from blood samples from different visits of 13 NSCLC patients (P1 through P13). ‘P1’ stands for patient 1. Blue bar represents the number of CK (+)/PD-L1 (−) CTCs. Red bar represents the number of CK (+)/PD-L1 (+) CTCs.

Fig 3: Kaplan–Meier life-table analysis of the PFS time in all patients. Grouping was done according to (A) CTC number more or less than 14/ml and (B) PD-L1 (+) CTC% more or less than 5%.

Fig 4: Transcriptional changes in human islets in response to IFN-γ.(A) PHATE analysis of scRNA-Seq of control and IFN-γ–treated human β cells. Human islets from 3 donors treated for 24 hours in the presence or absence of 25 ng/mL IFN-γ were sorted using TMRE and FluoZin-3 for β cell enrichment and subsequently analyzed by 10x Genomics. PHATE analysis shows distribution of islet cells from the 3 donors. (B and C) MELD revealed distinct populations of β cells in the presence and absence of IFN-γ. Separation of β cells compared with other islet cell populations in B indicates a greater impact of IFN-γ treatment on those cells. (D) PHATE analysis shows higher expression of CD274 in IFN-γ–treated β cells and (E) MELD identified a transition in β cells that is accompanied by a change from low to high CD274 (R2 = 0.87). (F) Pathway analysis of differentially expressed genes between control and IFN-γ–treated β cells reveals differences in pathways involved in IFN signaling (IFN-γ), cytokine production, antigen processing and presentation, apoptosis, and responses to TNF-α. (G) Volcano plot of differentially expressed genes between control and IFN-γ–treated β cells (2,134 genes based on P < 0.05, q < 0.05, log2fc ≤ –0.6 and ≥ 0.6), highlighting genes involved in IFN signaling, chemokines, and regulators of apoptosis. (H) Correlation of the overlapping 144 genes (among the top 1,000 differentially expressed genes in β cells) in the anti–PD-L1–treated mouse scRNA-Seq data set and the human islet IFN-γ treatment data set shown as log2fc. Highlighted genes include those involved in IFN-γ response (green), antigen processing/presentation (red), and cell death pathways (blue).

Fig 5: Increased FAS expression and apoptosis in PD-L1–positive β cells in response to IFN-γ.(A) Expression of FAS was highly correlated with CD274 expression in β cells by MELD. R2 = 0.84. Individual donor cells are shown in colors. (B) qPCR analysis demonstrated higher expression of FAS in FACS-sorted β cells with higher PD-L1 expression in the presence of IFN-γ. n = 5 for control and culture with IFN-γ. *P ≤ 0.05, **P ≤ 0.01 by 1-way ANOVA. (C) FAS protein expression was increased in PD-L1+ β cells and increased over time in culture with IFN-γ. Two-way ANOVA. (n = 4, duplicate wells.) (D) IFN-γ induced a significant reduction in live PD-L1+ β cells and a significant increase in PD-L1+ late apoptotic and dead β cells after 6 days in culture. (*P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. Two-way ANOVA with Tukey’s multiple comparison test.) (n = 4, duplicate wells.) Data for B–D shown as mean ± SEM. (E) FluoZin-3–stained β cells showing that PD-L1+ cells in the setting of IFN-γ treatment exhibited apoptotic blebs (white arrows). Acquired with a 63×/1.3 immersion objective lens and 10× magnification using a Leica SP8 STED 3× super-resolution microscope.