RANKL/PD-1 dual blockade demonstrates survival benefit for patients with advanced lung adenocarcinoma harboring KRAS mutations

Abstract

Preclinical/clinical studies suggest that receptor activator of nuclear factor κB (NF-κB) ligand (RANKL) inhibitors combined with immune checkpoint inhibitors (RLICi) enhance anti-tumor efficacy in lung adenocarcinoma (LUAD), yet mechanisms remain unclear. Our retrospective cohort demonstrates RLICi superiority in Kirsten rat sarcoma viral oncogene homolog (KRAS)-mutant LUAD. Transcriptomics reveal that RANKL upregulation was inversely correlated with PD-L1 and CXCL9/10/11 levels, suppressing CD8⁺ T cell infiltration via phosphatidylinositol-3-kinase/AKT serine/threonine kinase-mediated PD-L1 downregulation and macrophage chemokine reduction. In murine models, RLICi outperform PD-1 monotherapy, augmenting M1 macrophage recruitment and CD8⁺ T cell influx. The prospective DEMAIN trial validates RLICi clinical efficacy. This study elucidates RANKL-driven immunosuppression in KRAS-mutant LUAD and establishes RLICi as a viable therapeutic strategy for this subset. The trial was prospectively registered in the Chinese Clinical Trial Register (registration number: ChiCTR2100047759).

Keywords: KRAS; LUAD; denosumab; efficacy; immunotherapy.

Graphical abstract

Figure 1

Retrospective NCC cohort confirms that KRAS-mut LUAD benefits more from RLICi Patients with KRAS-mut LUAD who received RLICi had a significantly better prognosis (A and B) and tumor control (C and D) than those who did not, whereas this was not the case in LUAD patients without KRAS mutations. Multivariate analysis confirmed that patients with KRAS-mut LUAD benefit more from RLICi (E). vs., versus; mut, mutant; WT, wild type; LUAD, lung adenocarcinoma; RLICi, RANKL inhibitor plus immune checkpoint inhibitor; DCB, durable clinical benefit; ECOG PS, Eastern Cooperative Oncology Group Performance Status; PD-L1, programmed cell death receptor ligand-1; Neg, negative; Pos, positive; NA, not available; yrs, years; HR, hazard ratio; CI, confidence interval. ∗, statistically significant.

Figure 2

KRAS-mut LUAD enriched RANKL expression and associated with inhibitory TIME We validated the high expression level of RANKL in KRAS-mut LUAD using transcriptomic data from different public databases (A–C). The protein expression level of RANKL has also been verified via the CICAMS cohort (D). Heatmap (E) and scatterplots (F and G) depicting the correlation of transcriptome data from the TCGA database. The TIDE score comparison of RANKL-low and RANKL-high patients within KRAS-mut LUAD subgroup in the TCGA database (H). Heatmap (I) and scatterplot (J and K) of correlation of transcriptome data from GEO and CPTAC databases. The TIDE score comparison of RANKL-low and RANKL-high patients within KRAS-mut LUAD subgroup in GEO and CPTAC databases (L). Heatmap (M) of correlation of protein expression data from the CICAMS cohort. Representative IHC shots of RANKL, PD-L1, and CD8A within patients with KRAS-mut LUAD from the CICAMS cohort (N). See Figure S2 for relevant analyses for KRAS-wt LUAD. mut, mutant; WT, wild type; TIME, tumor immune microenvironment; LUAD, lung adenocarcinoma; IHC, immunohistochemical staining; PD-L1, programmed cell death receptor ligand-1; TIDE, tumor immune dysfunction and exclusion. The significance of two sample groups was determined by the Wilcoxon test or two-tailed unpaired Student’s t test, and the correlation analysis adopted the Pearson’s method.

Figure 3

High RANKL expression reduces PD-L1 expression via the PI3K-AKT pathway Cell line screening by western blotting (A). GSEA of sequencing data from RANKL-overexpressing cell lines (B and C). To delineate RANKL-mediated signaling, we profiled baseline molecular expression (VEC vs. OE) (D), modulated PI3K activity post-RANKL overexpression (E), silenced RANKL expression (F), validated findings via exogenous ligand stimulation (G), and compared KRAS genotype-specific effects (H). See Figure S3 for relevant semi-quantitative data. GSEA, gene set enrichment analysis; PD-L1, programmed cell death receptor ligand-1; NC, negative control; VEC, vector; OE, overexpression; vs., versus; siRNA, small interfering RNA; sRANKL, soluble RANKL; FDR, false discovery rate.

Figure 4

High RANKL expression leads to weakened function of macrophages and less infiltration of CD8⁺ T cells Heatmap illustrating the correlation of transcriptome data from the TCGA database (A) and the GEO database (B). Scatterplots (C–E) depicting the correlation of transcriptome data from the TCGA database. The IHC results of samples from our hospital (CICAMS cohort) also confirmed the negative correlation between RANKL and macrophages and CXCL9–11 (F and G). Transwell assay to assess macrophage chemotactic capacity (H–K). RT-qPCR assay to assess the transcriptional changes of CXCL9–11 of macrophages (L–O). ELISA assay to assess the secretory capacity of macrophages by measuring the concentration of CXCL9–11 in the supernatant (L and P–R). See relevant Figure S4 for the identification of chemokine CXCL9–11 promoting T cell chemotaxis function. mut, mutant; WT, wild type; sRANKL, soluble RANKL. The scale in (L) represents 100 μm. ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001; ∗∗∗∗p < 0.0001. The cross in the heatmap represents no statistical significance. The data were presented as the mean ± SEM (J, K, M, N, and O–R). Comparisons were performed using one-way ANOVA with the Tukey’s test for multiple comparisons. The correlation analysis adopted the Pearson’s method.

Figure 5

Mouse experiments demonstrate the superior anti-tumor efficacy of RLICi Schematic diagram of the animal test process (A). Cross-sectional view of LLC mouse tumors (n = 5) (B) and mouse tumor growth curve (C). Cross-sectional view of CMT167 mouse tumors (n = 5) (D) and mouse tumor growth curve (E). Immunohistochemical staining results of LLC mouse tumors (F). Quantitative chart of immunohistochemical results (G–J). Immunohistochemical staining results of CMT167 mouse tumors (K). Quantitative chart of immunohistochemical results (L–O). Pem, pemetrexed; i, inhibitor; PD-1, programmed cell death receptor 1. The scale in (F) and (K) represents 100 μm. ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001; ∗∗∗∗p < 0.0001. The data are presented as the mean ± SEM (C, E, G–J, and L–O). Comparisons as to tumor growth were performed using two-way ANOVA with the Tukey’s test for multiple comparisons. Comparisons as to cell markers were performed using one-way ANOVA with the Tukey’s test for multiple comparisons.

Figure 6

The DEMAIN trial demonstrated survival benefit of RLICi with a good safety profile Flowchart of the DEMAIN study (A). Waterfall plot (B). Swimming chart (C). Labeled disease status represents the initial response evaluation. Arrows indicates patients under continued treatment, whereas their absence denotes discontinuation due to subsequent disease progression. Representative cases of DEMAIN trial (D). Tumor lesions were marked by the yellow dotted circles. Kaplan-Meier curves of progression-free survival (E) and overall survival (F). The ELISA results showed that RLICi can effectively reduce sRANKL level in the blood and collagen metabolites CTX and NTX in the urine (G–I). The Kaplan-Meier curves suggest that sRANKL in baseline blood and RANKL in baseline tissue are associated with poor prognosis. RLICi, RANKL inhibitor plus immune checkpoint inhibitor; PD-1, programmed cell death receptor 1; sRANKL, soluble RANKL; uCTX, urinary C-terminal telopeptide of type I collagen; uNTX, urinary N-terminal telopeptide of type I collagen; ELISA, enzyme-linked immunosorbent assay. ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001; ∗∗∗∗p < 0.0001. The significance of two sample groups was determined by the two-tailed unpaired Student’s t test.