Immune suppressive microenvironment in liver metastases contributes to organ-specific response of immunotherapy in advanced non-small cell lung cancer

Abstract

Background: The liver is a frequent site of metastases and liver metastases (LM) correlate with diminished immunotherapy efficacy in non-small cell lung cancer (NSCLC). This study aimed to analyze whether tumor response to immunotherapy differs between pulmonary lesions (PL) and LM in NSCLC and to explore potential mechanisms through multiomics analysis.

Methods: This observational longitudinal clinical cohort study included patients with NSCLC with LM receiving immunotherapy was conducted to evaluate organ-specific tumor response of PL and LM. We collected paired PL and LM tumor samples to analyze the organ-specific difference using whole-exome sequencing, RNA sequencing, and multiplex immunohistochemistry.

Results: A total of 52 patients with NSCLC with LM were enrolled to evaluate the organ-specific response of immunotherapy. The objective response rate (21.1% vs 32.7%) and disease control rate of LM were lower than that of PL (67.3% vs 86.5%). One-third of patients showed mixed response, among whom 88.2% (15/17) presented with LM increasing, but PL decreasing, while the others had the opposite pattern (p=0.002). In another independent cohort, 27 pairs of matched PL and LM tumor samples from the same individuals, including six simultaneously collected pairs, were included in the translational part. Genomic landscapes profiling revealed similar somatic mutations, tumor mutational burden, and neoantigen number between PL and LM. Bulk-RNA sequencing showed immune activation-related genes including CD8A, LCK, and ICOS were downregulated in LM. The antigen processing and presentation, natural killer (NK) cell-mediated cytotoxicity and T-cell receptor signaling pathway were enriched in PL compared with LM. Multiplex immunohistochemistry detected significantly lower fractions of CD8⁺ cells (p=0.036) and CD56^dim+ cells (p=0.016) in LM compared with PL. Single-cell RNA sequencing also characterized lower effector CD8⁺ T cells activation and NK cells cytotoxicity in LM.

Conclusions: Compared with PL, LM presents an inferior organ-specific tumor response to immunotherapy. PL and LM showed limited heterogeneity in the genomic landscape, while the LM tumor microenvironment displayed lower levels of immune activation and infiltration than PL, which might contribute to developing precise immunotherapy strategies for patients with NSCLC with LM.

Keywords: Immunotherapy; Liver Neoplasms; Lung Neoplasms; Non-Small Cell Lung Cancer; Tumor Microenvironment.

Figure 1

Organ-specific response to immunotherapy differs between PL and LM in patients with advanced NSCLC. (A) Best changes from baseline in target lesions of PL and LM in per patients. (B) Organ-specific response rate and disease control rate in PL and LM. (C) Swimming plot showing progression-free survival and treatment strategy. Pathological type, PD-L1 expression level, gender, and EGFR/ALK mutation status are indicated for each patient. (D) The constituent ratio of different response patterns. (F) Kaplan-Meier estimates of progression-free survival stratified by organ-specific response patterns. ALK, anaplastic lymphoma kinase; ECOG PS, Eastern Cooperative Oncology Group performance status; EGFR, epidermal growth factor receptor; LM, liver metastases; NSCLC, non-small cell lung cancer; OSDCR, organ-specific disease control rate; OSRR, organ-specific response rate; PD, progressive diseases; PD-L1, programmed death-ligand 1; PL, pulmonary lesions; PR, partial response; SD, stable diseases.

Figure 2

Genomic landscapes profiling in paired PL and LM. (A) The workflow chart and samples grouping of the translational part. (B) Venn diagram of genetic variation, SNV, and CNV in the PL and LM groups. (C) The comparison of CNV all, CNV loss, and CNV gain burden between the PL and LM groups. (D) The comparison of TMB between the PL and LM groups. (E) The comparison of neoantigens number predicted by SNV and indel(-type) variants between the PL and LM groups. (F) The most common mutational genes are depicted in the heat map. Note: *, p<0.05. CNV, copy number variation; IHC, immunohistochemistry; LM, liver metastases; LUAD, lung adenocarcinoma; LUSC, lung squamous carcinoma; mIHC, multiplex IHC; muts/Mb, mutations per megabase; NSCLC, non-small cell lung cancer; PL, pulmonary lesions; SNV, single nucleotide variant; TMB, tumor mutational burden; WES, whole-exome sequencing.

Figure 3

LM shows an immunosuppressive tumor microenvironment. (A) A volcano plot of differentially expressed genes between PL and LM with four immune-related genes downregulated in LM marked in a frame. (B) Gene set enrichment analysis between PL and LM among three immunotherapy efficacy-related pathways: the antigen processing and presentation, NK cell-mediated cytotoxicity, and T-cell receptor signaling pathway. (C) Gene Ontology analysis of differentially enriched biological processes upregulated (red) or downregulated (blue) in LM. (D) The comparison of 22 immune cells proportion estimated by the CIBERSORT approach between the PL and LM groups. Note: *, p<0.05. DC, dendritic cell; KEGG, Kyoto encyclopedia of genes and genomes; LM, liver metastases; NK, natural killer; PL, pulmonary lesions.

Figure 4

Lower fractions of CD8⁺ cell and CD56^dim+ cell infiltrated in LM. (A) The positive rate of CD8⁺ cells, CD56^dim+ cells, CD56^bright+ cells, CD68⁺HLA-DR⁺ cells, and CD68⁺HLA-DR⁻ cells in the total tumor area, tumor parenchyma area and tumor stroma area in the PL and LM groups. (B) The relative infiltration proportion of CD8⁺ cells, CD56^dim+ cells, CD56^bright+ cells, CD68⁺HLA-DR⁺ cells, and CD68⁺HLA-DR⁻ cells in the tumor parenchyma area and tumor stroma area in the PL and LM groups. (C) The changes of CD8⁺ cells, CD56^dim+ cells, CD56^bright+ cells, CD68⁺HLA-DR⁺ cells, and CD68⁺HLA-DR⁻ cells infiltration from PL to LM in the total population (24 pairs) and group A (8 pairs). (D) The changes of CD8⁺ cells and CD56^dim+ cells in each patient of group A. Note: *, p<0.05. HLA-DR, human leukocyte antigen-DR; LM, liver metastases; PL, pulmonary lesions.

Figure 5

The scRNA-seq profiling of T/NK cells in paired PL and LM. (A) Uniform manifold approximation and projection (UMAP) plot of T/NK cells profiled in the present study colored by subcluster. (B) Feature plots showing the normalized expression of canonical marker genes in each subcluster of T/NK cells. (C) Proportions of T/NK cell subtypes in PL and LM, respectively. (D) UMAP plot of the representative gene for CD8Teff, HelperT, Treg, and CD8Tem. (E) GSEA between PL and LM among the T-cell receptor signaling pathway and NK cell-mediated cytotoxicity in isolated T and NK cells. (F) UMAP plot showing the T-cell activation and cytotoxicity scores in PL and LM. (G) Column chart showing the T-cell activation and cytotoxicity scores in each subcluster of T/NK cells. (H) Bubble diagram depicting the significant chemokines and (I) ligand–receptor interactions in PL and LM. GSEA, gene set enrichment analysis; LM, liver metastases; NK, natural killer; PL, pulmonary lesions; scRNA-seq, single-cell RNA sequencing.