Temporal and spatial characteristics of tumor evolution in a mouse model of oral squamous cell carcinoma

Abstract

Objectives: We aimed to elucidate the temporal and spatial characteristics of tumor evolution in an oral squamous cell carcinoma (OSCC) mouse model with higher burden of lymphatic metastasis through high-throughput sequencing.

Methods: The OSCC model was established in 9 mice. DNA was extracted from the tumors of primary tongue lesions and disseminated tumor cells (DTCs) of submandibular gland lymph nodes and bone marrow, and then whole genome sequencing was performed. After bioinformatics analysis, somatic single-nucleotide variants (SSNVs) and copy number variations (CNVs) data were obtained. Based on SSNVs, clonal architecture and ancestor-descendant relationships among tumor cell subclones were elucidated.

Results: A total of 238 tumor-related SSNVs with 120 high-frequency mutated genes were obtained from 36 samples of 9 mice by whole-genome sequencing. The number of unique SSNVs in the primary lesion, submandibular lymph node and bone marrow was greater than the number of shared SSNVs. Furthermore, the primary lesion-originated subclones, which were identified by SSNVs, were also detected in submandibular lymph nodes in the early stage of oral carcinogenesis. Moreover, at different histopathological stages, unique subclones were also identified in DTCs isolated from lymph nodes.

Conclusion: Tumor heterogeneity is significant in primary tumor cells and disseminated tumor cells. OSCC cells probably disseminate to lymph nodes in the early stage of oral carcinogenesis. OSCC is characterized by polyclonal dissemination, and the evolutionary trajectory of DTCs is potentially dominated by the tumor microenvironment.

Keywords: Disseminated tumor cells (DTCs); Genetic heterogeneity; Oral squamous cell carcinoma (OSCC); Tumor evolution; Tumor microenvironment; Whole genome sequencing.

Fig. 1

Pathological evidence of tumor formation and local lymph enlargement induced by 4NQO. A Dysplasia; B Tongue cancer; C Lymph node metastasis; D Abnormal hyperplasia (100×); E Tongue cancer (100×); F Lymph node metastasis (100×); G Abnormal hyperplasia (200×); H Tongue cancer (200×); I Lymph node metastasis (200×)

Fig. 2

Flow cytometry results of Single cell suspension before and after MACS. A Single-cell suspension from a normal lymph node before MACS. B Single-cell suspension from normal bone marrow before MACS. C Single-cell suspension from tongue cancer before MACS. D Single cell suspension of lymph node before and after sorting with CD45/CD326 immunomagnetic beads (mouse 1). E Single-cell suspension of bone marrow monocytes before and after sorting with CD45/CD326 immunomagnetic beads (mouse 1)

Fig. 3

List of SSNV mutation patterns and SSNVs from different tissues were characterized by Venn diagrams in 9 mice. A Summary of 36 samples SSNV mutation patterns LNs: lymph nodes, BM. B Tumor SSNVs from different tissues were characterized by Venn diagrams in 9 mice. The purple circle represents SSNVs identified in the primary lesion, the blue circle represents SSNVs identified in DTCs isolated from lymph nodes, and the brown circle represents SSNVs in DTCs isolated from bone marrow. Numbers in the circle indicated the number of SSNVs. LNs: lymph nodes, BM: bone marrow, DTCs: disseminated tumor cells, Nor: Adjacent tissue, Pri: Primary tumor

Fig. 4

Heatmap of CNVs with the significantly amplified/deleted cytoband and gains and losses of copy number. A The red square indicates the cytoband with amplification/gain, and the blue square indicates the cytoband with deletion. The y-axis shows the mutation percentage in the corresponding chromosomal region. The x-axis represented sample name. The “Genes” column lists mutated genes related to human cancers located in chromosome fragments. B Gains and losses of copy number identified by GISTIC 2.0.A: Copy number gains identified in all tumor cells (primary and DTCs). B. Copy number deletions identified in all tumor cells (primary and DTCs)

Fig. 5

Average value of clustering between samples between 9 mice. Paired comparison of the mean distribution of Variant Allele Frequency subsets between primary tongue tumor cells and lymph node metastatic tumor cells in 9 mice: 1–3 dysplastic mice, 4–6 mice without lymph node metastasis, 7–9 mice with lymph node metastasis. Variant Allele Frequency subgroups were obtained by clustering the CCF values of tumor cells in different parts of each mouse. The Variant Allele Frequency subgroups of each mouse correspond to a specific color and shape

Fig. 6

Subclonal architecture of 9 mice and relationships between subclones were revealed by phylogenetic trees in three mice. A Two-dimensional scatter plots show the CCF of the mutations among different tumor cells. Each subclone was acquired by clusting of the CCF in tumor cells from different organs. B-D Phylogenetic trees of mouse1–3: Branch length was proportional to the number of substitution mutations in each subclone. The trunk branch represents subclones originating from the primary lesion, and the subbranch represents unique subclones from different tissues. Tissue resources were annotated at the end of the branch, and oncogenic genes with driver mutations were annotated on the side of the branch