Skip to main content
. 2023 Jun 30;115(11):1392–1403. doi: 10.1093/jnci/djad126

Figure 6.

Figure 6.

Anti–programmed death ligand 1 (PD-L1) and/or anti–signal transducer and activator of transcription 3(STAT3) in humanized mice (HM). Tumor volume decreases upon anti–PD-L1 in the HM model compared with the HM control in (A) CUHN036 and (B) CUHN022. No changes were observed in control vs anti–PD-L1–treated NOD-scid-gamma (NSG) mice. Treatment with anti–STAT3 or the combination of anti–PD-L1 and anti–STAT3 (combo) ablated tumor growth in the NSG and HM models. In CUHN022, the anti–STAT3 group relapsed following end of treatment (EOT) in the HM group, while the combo maintained response. Arrows indicate date of treatment, with both immunoglobulin G and dimethylsulfoxide control, anti–PD-L1, and anti–STAT3 given simultaneously. C) In CUHN022, anti–PD-L1, anti–STAT3, and combo increase CD3+CD4+ and CD3+CD8+ tumor-infiltrating cells. Anti–PD-L1 and combo increase CD3+granzyme B+ cells. Quantification to the right. Arrows indicate double-positive or single-positive signal (ie, CD3+granzyme B+ vs CD3+granzyme B-). Human T-cells (CD45+CD3+) analyzed from (D) HM spleen or (E) peripheral blood, from control, anti–PD-L1, anti–STAT3 EOT, combo EOT, anti–STAT3 (4 weeks post-EOT), and combo (4 weeks post-EOT) from the CUHN022 HM mice. Cells were additionally assayed for CD4, CD8, and PD-1 expression (statistical analysis of total CD3+CD4+ cells is indicated above, analysis of CD3+CD4+PD-1+ is indicated below). The combination released group exhibited a decrease in peripheral CD4 T-cells at 4 weeks post-EOT compared with EOT, whereas the anti–STAT3 release mice had increased peripheral CD4 and CD8 subpopulations. F) Immunohistochemistry images (left) and quantification (right) of PD-L1, interleukin enhancer binding factor 2 (ILF2), interleukin enhancer binding factor 3 (ILF3), phospho–STAT3, and STAT3 show an increase in the HM model that is blocked upon anti–PD-L1, anti–STAT3, or combo treatment. Anti–PD-L1 had no effect on the NSG model. Anti–STAT3 reduced PD-L1, interleukin ehancer binding factor 2 (ILF2), ILF3, and phospho–STAT3 in NSG and HM tumors, with deeper effect in HM particularly in the combination arm. G) Multicolor proximity ligation assay shows an increase in interaction between PD-L1–ILF2, PD-L1–ILF3, and PD-L1–STAT3 in the HM compared with NSG control. Anti–PD-L1 and combo decreased the interaction of PD-L1–STAT3. In the HM model, targeting PD-L1 reduced the interaction between PD-L1 and STAT3, with the combination of anti–PD-L1 and anti–STAT3 being even more effective. The PD-L1–ILF2 and PD-L1–ILF3 interactions remained stable, with neither therapy effectively impacting this association. H) Proposed conceptual model of PD-1– PD-L1 interaction leading to immune escape on one side, and PD-L1 recruitment of ILF2–ILF3 and subsequent activation of STAT3, leading to tumor progression. *P ≤ .05, **P ≤ .01, ***P ≤ .001, ****P ≤ .0001. Error bars represent standard deviation. Scale bar = 25μm (B), scale bar = 50μm (E). α-PD-L1 = anti–PD-L1; α-STAT3 = anti–STAT3; α-STAT3* and combo* = samples collected 4 weeks after EOT; HM = humanized mice; NSG = NOD-scid-gamma; PLA = proximity ligation assay.