Loss of STK11 Suppresses Lipid Metabolism and Attenuates KRAS-Induced Immunogenicity in Patients with Non-Small Cell Lung Cancer

Abstract

As many as 30% of the patients with non-small cell lung cancer harbor oncogenic KRAS mutations, which leads to extensive remodeling of the tumor immune microenvironment. Although co-mutations in several genes have prognostic relevance in KRAS-mutated patients, their effect on tumor immunogenicity are poorly understood. In the present study, a total of 189 patients with non-small cell lung cancer underwent a standardized analysis including IHC, whole-exome DNA sequencing, and whole-transcriptome RNA sequencing. Patients with activating KRAS mutations demonstrated a significant increase in PDL1 expression and CD8+ T-cell infiltration. Both were increased in the presence of a co-occurring TP53 mutation and lost with STK11 co-mutation. Subsequent genomic analysis demonstrated that KRAS/TP53 co-mutated tumors had a significant decrease in the expression of glycolysis-associated genes and an increase in several genes involved in lipid metabolism, notably lipoprotein lipase, low-density lipoprotein receptor, and LDLRAD4. Conversely, in the immune-excluded KRAS/STK11 co-mutated group, we observed diminished lipid metabolism and no change in anaerobic glycolysis. Interestingly, in patients with low expression of lipoprotein lipase, low-density lipoprotein receptor, or LDLRAD4, KRAS mutations had no effect on tumor immunogenicity. However, in patients with robust expression of these genes, KRAS mutations were associated with increased immunogenicity and associated with improved overall survival. Our data further suggest that the loss of STK11 may function as a metabolic switch, suppressing lipid metabolism in favor of glycolysis, thereby negating KRAS-induced immunogenicity. Hence, this concept warrants continued exploration, both as a predictive biomarker and potential target for therapy in patients receiving ICI-based immunotherapy.

Significance: In patients with lung cancer, we demonstrate that KRAS mutations increase tumor immunogenicity; however, KRAS/STK11 co-mutated patients display an immune-excluded phenotype. KRAS/STK11 co-mutated patients also demonstrated significant downregulation of several key lipid metabolism genes, many of which were associated with increased immunogenicity and improved overall survival in KRAS-mutated patients. Hence, alteration to lipid metabolism warrants further study as a potential biomarker and target for therapy in patients with KRAS-mutated lung cancer.

Figure 1

Genetic landscape of cohort of patients with NSCLC. A, Tissue biopsies from 189 treatment-naïve patients with NSCLC, 114 of which were from the primary lung tumor, 31 from an affected lymph node, and 44 from a metastatic site, were sectioned and stained with either hematoxylin and eosin or via IHC for PDL1. Tissues then underwent a standardized genomic analysis consisting of whole-exome DNA sequencing and whole-transcriptome RNA-seq. Pie charts are shown representing the most frequently observed mutations across (B) all 189 patients, (C) the 132 patients with LUAD, and (D) the 57 patients with SCC. CNA, copy-number amplification; CNL, copy-number loss; MT, mutant.

Figure 2

KRAS-mutated NSCLC tumors display increased PDL1 expression. A and B, Tissue biopsies from 189 treatment-naïve patients with NSCLC were stained with either H&E or via IHC for PDL1. Representative images are shown for low and high expression in LUAD (N = 132) and SCC (N = 57) histologies. C, PDL1 TPS arranged by tumor histology. D, PDL1 TPS plotted against percentile CD274 mRNA expression. E, PDL1 TPS expression for tumor specimens obtained from the primary lung tumor (lung, N = 114), lymph node (LN, N = 31), or distant metastatic site (Met, N = 4). F and G, PDL1 TPS or CD274 mRNA expression for either KRAS-WT (N = 144) or KRAS mutated patients (MT, N = 45). H and I, PDL1 TPS or CD274 mRNA for either KRAS-WT or KRAS-MT patients arranged by tumor histology. J, PDL1 TPS or CD274 mRNA for either KRAS-WT or KRAS-MT patients arranged by tumor site. H&E, hematoxylin and eosin; NS, nonsignificant.

Figure 3

STK11 co-mutation negates KRAS-induced immunogenicity. A, PDL1 TPS for all patients with NSCLC arranged by TP53 mutation status [N = 44 WT; N = 145 TP53-mutated (MT)]. B and C,CD274 mRNA expression or PDL1 TPS arranged by combined KRAS and TP53 mutation statuses (N = 26 KRAS-WT/TP53-WT, N = 18 KRAS-MT/TP53-WT, N = 118 KRAS-WT/TP53-MT, N = 27 KRAS-MT/TP53-MT). D, PDL1 TPS arranged by STK11 mutation status (N = 166 wild-type or WT, N = 23 STK11-mutated or MT). E and F,CD274 mRNA expression or PDL1 TPS arranged by combined KRAS and STK11 mutation statuses (N = 131 KRAS-WT/STK11-WT; N = 35 KRAS-MT/STK11-WT; N = 13 KRAS-WT/STK11-MT; N = 10 KRAS-MT/STK11-MT). G, Percent CD8⁺ T-cell infiltration arranged by KRAS mutation status. H, Percent CD8⁺ T-cell infiltration arranged by combined KRAS and TP53 mutation statuses. I, Percent CD8⁺ T-cell infiltration arranged by combined KRAS and STK11 mutation statuses.

Figure 4

Mutant KRAS-induced immunogenicity is associated with increased lipid metabolism. A, Tissue biopsies from treatment-naïve patients with NSCLC were subjected to whole-transcriptome RNA-seq. Patients were grouped based on KRAS and TP53 mutation statuses and gene set enrichment analysis (GSEA) performed. Significantly altered gene sets involved in tumor immunity, apoptosis, or metabolism are shown using a Z-score cutoff of 2.5. Focused heatmaps are shown for select, significantly altered genes in the (B) adaptive immune response gene set or (C) glycerophospholipid metabolic process gene sets, both using a FDR-adjusted P value of <0.05. D, Patients were grouped based on KRAS and STK11 mutation statuses, and GSEA performed as previously and significantly altered gene sets involved in the adaptive immune response shown. Focused heatmaps are shown for select, significantly altered genes in the (E) lymphocyte-mediated immunity gene set or (F) cell activation involved in immune response gene sets, both using a FDR-adjusted P value of <0.05. Patients were again grouped based on KRAS and STK11 mutation statuses and GSEA performed. Significantly altered gene sets are shown for pathways involved in G innate or humoral immunity, (H) apoptosis/cell death, or (I) lipid metabolic processes. Focused heatmaps are shown for select, significantly altered genes in the (J) glycerolipid biosynthetic process gene set or (K) glycerophospholipid metabolic process gene set, both using a FDR-adjusted P value of <0.05.

Figure 5

Mutant KRAS-induced immunogenicity requires the transcriptional upregulation of lipid metabolism. A, Tissue biopsies from treatment-naïve patients with NSCLC were subjected to whole-transcriptome RNA-seq as described and representative heat maps shown for select genes involved in lipid metabolism. B and C, Focused heatmaps are shown for select, significantly altered genes involved in lipid metabolic processes using a FDR-adjusted P value of <0.05. The percent CD8⁺ T-cell infiltration for patients arranged by combined KRAS mutation and (D) LPL mRNA expression, (E) LDLR mRNA expression, or (F) LDLRAD4 mRNA expression. Kaplan–Meier plots indicating months of OS for patients with NSCLC arranged by (G) CD8 mRNA expression, (H) KRAS mutation status alone, (I) combined KRAS and STK11 mutation statuses, (J) KRAS mutation status and LPL mRNA expression, (K) KRAS mutation status and LDLR mRNA expression, and (L) KRAS mutation status and LDLRAD4 mRNA expression. High, above median; low, below median; MT, mutated; NS, nonsignificant.

Figure 6

Schema describing the presumptive mechanism through which STK11 loss attenuates KRAS-induced immunogenicity in NSCLC. In patients with NSCLC, gain-of-function KRAS mutations seem to enhance tumor immunogenicity, increasing both PDL1 expression and CD8⁺ T-cell infiltration. However, KRAS/STK11 co-mutated tumors displayed an immune-excluded phenotype, with diminished expression of PDL1 and limited CD8⁺ T-cell infiltration. This immune-excluded phenotype was paralleled by reductions in lipid metabolism, suggesting that the loss of STK11 may function as a metabolic switch, suppressing lipid metabolism in favor of glycolysis to attenuate KRAS-induced immunogenicity.