HMGB1 Expression Levels Correlate with Response to Immunotherapy in Non-Small Cell Lung Cancer

Abstract

Purpose: High-mobility group box 1 protein (HMGB1) is subject to exportin 1 (XPO1)-dependent nuclear export, and it is involved in functions implicated in resistance to immunotherapy. We investigated whether HMGB1 mRNA expression was associated with response to immune checkpoint inhibitors (ICI) in non-small cell lung cancer (NSCLC).

Patients and methods: RNA was isolated from pretreatment biopsies of patients with advanced NSCLC treated with ICI. Gene expression analysis of several genes, including HMGB1, was conducted using the NanoString Counter analysis system (PanCancer Immune Profiling Panel). Western blotting analysis and cell viability assays in EGFR and KRAS mutant cell lines were carried out. Evaluation of the antitumoral effect of ICI in combination with XPO1 blocker (selinexor) and trametinib was determined in a murine Lewis lung carcinoma model.

Results: HMGB1 mRNA levels in NSCLC patients treated with ICI correlated with progression-free survival (PFS) (median PFS 9.0 versus 18.0 months, P=0.008, hazard ratio=0.30 in high versus low HMGB1). After TNF-α stimulation, HMGB1 accumulates in the cytoplasm of PC9 cells, but this accumulation can be prevented by using selinexor or antiretroviral drugs. Erlotinib or osimertinib with selinexor in EGFR-mutant cells and trametinib plus selinexor in KRAS mutant abolish tumor cell proliferation. Selinexor with a PD-1 inhibitor with or without trametinib abrogates the tumor growth in the murine Lewis lung cancer model.

Conclusion: An in-depth exploration of the functions of HMGB1 mRNA and protein is expected to uncover new potential targets and provide a basis for treating metastatic NSCLC in combination with ICI.

Keywords: HMGB1; K-Ras mutations; Lewis lung cancer murine model; immunotherapy; non-small cell lung cancer.

Figure 1

HMGB1 mRNA expression in pre-ICI-treatment FFPE tumor tissue samples. (A) Volcano plot indicating differential expression of the genes included in the IO360 panel. Genes depicted in green indicate a log2 Fold Change > 1, genes depicted in blue indicate a nominal p-value < 0.05, genes depicted in red indicate a log2 Fold Change > 1 and a p-value < 0.05. (B) Significantly differentially expressed genes between patients with and without clinical benefit, based on the Fold Change and nominal p-value (independent t-test). P-values adjusted for multiple testing are also included.

Figure 2

Overall survival (OS) analysis according to HMGB1 mRNA expression levels (lower versus higher that the median value). (A) In the exploratory cohort (n=42), there was a significant difference in overall survival for patients with low HMGB1 versus high HMGB1 mRNA expression level (hazard ratio [HR], 0.30; 95% CI, 0.12–0.73; P=0.0076) (B) In the validation cohort (n=41), a significant difference in overall survival was confirmed for patients with low versus high HMGB1 mRNA expression level (HR, 0.39; 95% CI, 0.17–0.90; P=0.0288).

Figure 3

Effect of antiretroviral and selinexor treatment on nuclear export of HMGB1 in the PC9 cell line. (A) PC9 cells were treated with TNF-A (100ng/mL) to induce nuclear export of HMGB1, using different time intervals. Proteins from the nucleus and cytoplasm were separated and HMGB1 was detected by Western blotting. (B) PC9 cells were pre-treated with selinexor or antiretroviral drugs for 30 minutes and cells were then exposed to TNF-A (100 ng/mL) for 8 hours. Proteins from the nucleus and cytoplasm were separated and HMGB1 was detected by Western blotting. Data were analyzed using unpaired t-test comparisons.*P <0.05, **P <0.01, comparison between two groups marked with horizontal lines.

Figure 4

Effect of antiretroviral and selinexor treatment on cell viability analyzed by Cell Counting Kit-8 assay. (A) Effect of 20 uM single selinexor or antiretroviral treatments on cell viability in the H520, PC9, A549 and H1975 lung cancer cell lines, compared to the control. (B) Effect of single and combined selinexor plus osimertinib or erlotinib treatment on cell viability in the H1975 cell line, using different concentrations of selinexor. (C) Effect of single and combined selinexor plus osimertinib or erlotinib treatment on cell viability in the PC9 cell line, using different concentrations of selinexor. (D) Effect of single and combined selinexor plus MEK-inhibition treatment on cell viability in the A549 cell line, using different concentrations of selinexor. (E) Effect of single and combined selinexor plus MEK-inhibition treatment on cell viability in the H460 cell line, using different concentrations of selinexor. Data were analyzed using unpaired t-test comparisons. *P <0.05, **P <0.01, compared to control. ^#P <0.05, ^##P <0.01, comparison between two groups marked with horizontal lines.

Figure 5

Anti-tumor efficacy of PD-L1 mAb in combination with selinexor and trametinib in vivo. (A) Time schedule for the LLC subcutaneous implantation and drug treatment groups, mice were randomly grouped into 7 groups: Model group, Vehicle + Isotype group, selinexor + Isotype group, trametinib + Isotype group, PD-L1 mAb + Vehicle group, selinexor + PD-L1 mAb group, and selinexor + PD-L1 mAb + trametinib group. (B) Tumor volumes for subcutaneous LLC tumor. (C) Mice body weight in each group. (D) Subcutaneous tumor weight in each group. (E) Representative LLC tumor pictures. (n=5). Data were analyzed using one-way ANOVA comparisons. *P <0.05, compared to the selinexor + Isotype group. (Regarding the results in (B and D)), there is a certain measurement error in measuring the size of a subcutaneous tumor with caliper (B). In fact, the smaller the tumor, the greater the measurement error. The actual tumor weight (D) and pictures (C) excised from mice are more accurate and intuitive.