Integrated spatial analysis of gene mutation and gene expression for understanding tumor diversity in formalin-fixed paraffin-embedded lung adenocarcinoma

Abstract

Introduction: A deeper understanding of intratumoral heterogeneity is essential for prognosis prediction or accurate treatment plan decisions in clinical practice. However, due to the cross-links and degradation of biomolecules within formalin-fixed paraffin-embedded (FFPE) specimens, it is challenging to analyze them. In this study, we aimed to optimize the simultaneous extraction of mRNA and DNA from microdissected FFPE tissues (φ = 100 µm) and apply the method to analyze tumor diversity in lung adenocarcinoma before and after erlotinib administration.

Method: Two magnetic beads were used for the simultaneous extraction of mRNA and DNA. The decross-linking conditions were evaluated for gene mutation and gene expression analyses of microdissected FFPE tissues. Lung lymph nodes before treatment and lung adenocarcinoma after erlotinib administration were collected from the same patient and were preserved as FFPE specimens for 4 years. Gene expression and gene mutations between histologically classified regions of lung adenocarcinoma (pre-treatment tumor in lung lymph node biopsies and post-treatment tumor, normal lung, tumor stroma, and remission stroma, in resected lung tissue) were compared in a microdissection-based approach.

Results: Using the optimized simultaneous extraction of DNA and mRNA and whole-genome amplification, we detected approximately 4,000-10,000 expressed genes and the epidermal growth factor receptor (EGFR) driver gene mutations from microdissected FFPE tissues. We found the differences in the highly expressed cancer-associated genes and the positive rate of EGFR exon 19 deletions among the tumor before and after treatment and tumor stroma, even though they were collected from tumors of the same patient or close regions of the same specimen.

Conclusion: Our integrated spatial analysis method would be applied to various FFPE pathology specimens providing area-specific gene expression and gene mutation information.

Keywords: cancer therapy; drug resistance; formalin-fixed paraffin-embedded specimens; intratumoral heterogeneity (ITH); non-small cell lung cancer (NSCLC); spatial transcriptome; tumor microenvironment; tyrosine-kinase inhibitors (TKIs).

Figure 1

Overview of spatial transcriptome and gene mutation analysis with lung adenocarcinoma before and after erlotinib administration. (A) The medical history of the patients analyzed in this study. By punch biopsy, lung lymph nodes were collected before erlotinib administration and used for gene mutation testing. The patient was treated with erlotinib 150 mg/day for 26 days, withdrawn due to side effects such as anemia, and then started on erlotinib 100 mg/day after symptoms improved. The lung tumor was surgically removed about 70 days after starting the medication. After that, the disease recurred, and the patient was treated with gefitinib 250 mg/day, but EGFR ex19del and T790M were detected, so the dosage was changed to osimertinib 80 mg/day on 9 October 2019, and he is alive now. (B) Workflow of simultaneous analysis of transcriptome and gene mutation from tissue microdissections in lung adenocarcinoma. 1: Morphological classification of adenocarcinoma by the experienced pathologist. 2: Micro-area with a diameter of 100 µm of the tumor, tumor stroma, remission stroma, and normal lung are collected from lung adenocarcinoma after erlotinib administration by microdissection punching system. In the pre-treatment lung lymph node, the micro-area in the tumor is collected. 3: The tissue microdissections lysed with proteinase K followed by the mRNA extraction with oligo dT magnetic beads. Extracted mRNA is decross-linked by incubation at 85 °C for 5 min and used for RNA-seq. After mRNA purification, DNA was extracted with carboxyl-coated magnetic beads from the supernatant. Extracted DNA was decross-linked by incubation at 90°C for 5 min and used for the detection of EGFR driver gene mutation.

Figure 2

RNA-seq and EGFR exon 19 deletion detection from microdissected FFPE tissues of lung adenocarcinoma. (A) Spatial distribution of sampling locations. Left: tumor in pre-treatment lung lymph node; middle: tumor, tumor stroma, and remission stroma in post-treatment lung adenocarcinoma; right: normal lung in post-treatment lung adenocarcinoma. (B) Venn diagram of the genes detected from tissue microdissections or the whole tissue section of post-treatment lung adenocarcinoma. The genes detected from tissue microdissections were merged by histological classification (post-treatment tumor, normal lung, tumor stroma, and stroma in remission). (C) The number of genes detected from tissue microdissections. Comparison of the number of detected genes between areas with different histological classifications. Student’s t-test, ***P < 0.001. (D) Results of EGFR ex19del detection. The EGFR ex19del region was PCR amplified, and Bioanalyzer DNA1000 measured the length of the PCR amplified product. The 15 bp and 1500 bp peaks indicate lower and upper markers, respectively. The peak at 86–89 bp is deletion positive, and the peak at 103 bp is wild type (The peaks at 131–170 bp represent the by-products generated by PCR amplification).

Figure 3

Spatial comparison of transcriptome and EGFR driver gene mutation in lung adenocarcinoma. (A) The specific distribution of tissue microdissections in the classification of treatment or histology in PCA. Each shape represents a result of hierarchical clustering using the top 40 genes in average expression level. (B) The spatial distribution of site-specific genes in lung adenocarcinoma after erlotinib administration (NAPSA, tumor site-specific gene; SGK1, tumor stroma site-specific gene; CSRP1, stroma in remission site-specific gene). (C) Heatmap of marker genes within cell types of lung adenocarcinoma (AT1 cell: PDPN, AGER, and CAV1; AT2 cell: SETPD, SETPB, SETPC, and SETPA1; Clara: SCGB3A1, and SCGB3A2; CAFs: THY1, and COL1A1, and CAL1A2; Fibroblasts: DCN; Myofibroblasts: TAGLN, ACTA2, ACTG2, MYH11, and MYLK; Ciliated: FOXJ1 and CAPS; B cell: CD19, CD79A, and CD79B; DC: CCL17, CCL22, CD207, CD1C, and CD1A; T cell: CD2, CD27, CD28, CD3D, CD3E, CD3G, and CD69; Mast: CPA3, CLU, TPSAB1, TPSB2, and MS4A2; Monocytes/Macrophages: ECGR1A, CCR2, MCFMP1, MARCO, CX3CR1, ITGAX, ITGAM, FCGR3A, CD163, CSF1R, MSR1, CD14, MRC1, and APOE; Endothelial: CLDN5 and PECAM1).

Figure 4

Gene mutations and gene expression in lung adenocarcinoma. (A) The genes highly expressed in post-treatment tumors compared with pre-treatment tumors. The oncogenic transcription factor FOS was highly expressed in post-treatment tumors, tumor stroma, and remission stroma. EGR1, a tumor suppressor gene, was highly expressed only in post-treatment tumors. (B) The spatial distribution of the EGFR ex19del. Left: tumor in pre-treatment lung lymph node; Right: tumor in post-treatment lung adenocarcinoma. Bottom: normal lung in post-treatment lung adenocarcinoma. The area surrounded by the black line is the tumor. (C) The enriched pathways in EGFR ex19del- negative or EGFR ex19del-positive tissue microdissections. Asterisk(*) related to ECM.

Figure 5

Interaction between tumor and tumor stroma in the tumor microenvironment after erlotinib administration. (A) The spatial distribution of the tumor marker gene. KLF6 is more highly expressed in stroma than in the tumor on the tissue section. (B) Interaction scores between ligands expressed in tumor stroma and receptors expressed in other regions, showing pairs of tumor stroma and post-treatment tumor interaction scores greater than 35. (C) Interaction scores between ligands expressed on post-treatment tumor and receptors expressed in other sites, showing pairs of post-treatment tumor and tumor stroma interaction scores greater than 35. (D) The spatial distribution of ligand-receptor pairs related to ECM cell–cell interaction on the tissue sections. The ECM ligand COL3A1 is highly expressed in the tumor stroma, and the ECM receptor ITGA2 is highly expressed in the tumor.