Fig 2: CXXC4 Expression Is Reduced and Negatively Correlated with PD-L1 Expression in GC(A) The expression of CXXC4 in GC samples (n = 36) and normal samples (n = 25) in GEO: GSE112369 from the GEO database. The y axis indicates the gene expression of CXXC4 in microarray GEO: GSE112369. (B) CXXC4 expression in GC tissues and adjacent normal tissues of GC patients (n = 102) detected using qRT-PCR. (C) Expression of CXXC4 normalized to GAPDH in three randomly selected pairs of GC tissue and adjacent normal tissue from patients with GC determined using western blot analysis. Images were quantified using ImageJ software after three biological repetitions. (D) CXXC4 expression identified by IHC. (E) Survival rate of clinical patients. (F) Recurrence rate of clinical patients in follow-up investigation. (G) PD-L1 expression in GC tissues and adjacent normal tissues of patients measured by qRT-PCR. (H) PD-L1 and CXXC4 expression in GC tissues and adjacent normal tissues determined using IHC. (I) Pearson’s correlation coefficient on CXXC4 and PD-L1 correlation. ∗p < 0.05 compared with the adjacent normal tissue. The measurement data are expressed as mean ± standard deviation. A paired t test was used for comparisons between the GC tissues and adjacent normal tissues. The Kaplan-Meier method was used to evaluate the patient’s survival rate with a log-rank test used to compare the survival differences. Data comparison among multiple groups was analyzed by one-way ANOVA, followed by Tukey’s post hoc test. A Pearson’s correlation coefficient was performed for analysis of the correlation between CXXC4 and PD-L1.

Fig 3: Anti‐tumor activity of PD‐L1 mAb treatment in the B16F10 xenograft mice model. A, Imaging of B16F10 tumor from xenograft mice. B, Tumor weights of the stripped tumors from xenograft mice (**P < .01, ***P < .001). C, Tumor volume changes following the treatment of IgG, PD‐L1 mAb and CTX. D, Average body weights of mice groups

Fig 4: ATP11B interacts with PD-L1 in a CMTM6-dependent manner. (A–B) ATP11B interaction with PD-L1 in vivo. Cell lysates from KPC and SW1990 cells were separately subjected to immunoprecipitation and western blotting using the indicated antibodies. (C) Lack of binding of ATP11B to PD-L1 in vitro. FLAG-tagged ATP11B and MYC-tagged PD-L1 purified from 293 T cell lysates and subjected to immunoprecipitation after co-incubation. (D) Positive correlation of ATP11B with CMTM6. Correlation of ATP11B and CMTM6 analyzed according to the pancreatic data sets of The Cancer Genome Atlas. (E–G) Representative images (E–F) and statistical results (G) of CMTM6 and ATP11B immunohistochemical staining in pancreatic cancer tissue microarray. (H) Western blotting of ATP11B, CMTM6, and PD-L1 expression in paired clinical tissue samples. (I–M) Interaction of ATP11B with PD-L1 in pancreatic cancer in a CMTM6-dependent manner. (I−J) cell lysates from KPC and SW1990 cells were separately subjected to immunoprecipitation and western blotting using the indicated antibodies. (K) Western blotting of ATP11B expression in pancreatic cancer cell lines after CMTM6 knockdown. (L–M) Cell lysates from CMTM6-KD KPC (L) and CMTM6-KD SW1990 (M) separately subjected to immunoprecipitation and western blotting using the indicated antibodies. CMTM6, CKLF-like MARVEL transmembrane domain containing 6; PD-L1 programmed cell death ligand 1.

Fig 5: Lymph flow from tumor draining lymph nodes (tdLNs) to blood circulation. (A) Presentation of tumor antigen (tAg) by antigen presenting cells (APCs) to naïve T-cells in tdLNs initiates anti-tumor response. (B) Activated T-cells can be tolerized by PD-L1 expression within an immunosuppressive tumor microenvironment (TME). (C) Lymphatic endothelial cells (LECs) may also present tAg and express high levels of PD-L1 to tolerize effector T-cells in tdLNs and efferent lymphatic vessels. TCR, T cell receptor; MHC, major histocompatibility complex; LN, lymph node. Created with Biorender.com (accessed on 21 February 2022).