Intra-Patient Heterogeneity of Circulating Tumor Cells and Circulating Tumor DNA in Blood of Melanoma Patients

Abstract

Despite remarkable progress in melanoma therapy, the exceptional heterogeneity of the disease has prevented the development of reliable companion biomarkers for the prediction or monitoring of therapy responses. Here, we show that difficulties in detecting blood-based markers, like circulating tumor cells (CTC), might arise from the translation of the mutational heterogeneity of melanoma cells towards their surface marker expression. We provide a unique method, which enables the molecular characterization of clinically relevant CTC subsets, as well as circulating tumor DNA (ctDNA), from a single blood sample. The study demonstrates the benefit of a combined analysis of ctDNA and CTC counts in melanoma patients, revealing that CTC subsets and ctDNA provide synergistic real-time information on the mutational status, RNA and protein expression of melanoma cells in individual patients, in relation to clinical outcome.

Keywords: CTC; ctDNA; liquid biopsy; melanoma.

Figure 1

BRAF/NRAS mutations determine cellular marker expression in malignant melanoma. (A) Volcano plot of differentially regulated genes in the TCGA data set, comparing melanomas containing BRAF/NRAS mutations (RAS/RAF activating) and not BRAF/NRAS mutated melanomas (other). Significantly reduced genes (<lgFC − 1) are depicted in violet; significantly increased genes (>lgFC 1) in yellow (B) t-SNE plot, based on melanoma markers expressed on single cells (Tirosh et al.), derived from BRAF/NRAS mutated tumors and not BRAF/NRAS mutated tumors. (C) Venn diagram of all differentially expressed genes between RAS/RAF activating and other tumors of the TCGA and Tirosh cohorts and specific melanoma marker genes. (D) Significantly differentially expressed melanoma marker genes in the Tirosh data set. * = p < 0.05.

Figure 2

Method to analyze melanoma cell subsets. (A) Tested isolation methods for the membrane marker dependent and independent enrichment of melanoma cells. (B) Recovery rate of 50 melanoma cells in 7.5 mL of blood and subsequent histological slides needed for analysis. (C) Amount of melanoma cells positive for membrane markers CSPG4/MCAM and intracellular S100. N = 3–4.

Figure 3

Combined enrichment method for the analysis of cellular subpopulations of circulating melanoma cells in patients. (A) Overview of patients with cutaneous, acral, mucosal and uveal melanoma of different disease stages, included in the study. (B) Time course of CTCs, enriched by positive selection (orange circles) and Parsortix™ (green circles) and their detected mutations, as well as LDH and S100 serum levels in patient #1. (C) Mass spectrum plot of BRAFV600E mutation and (D) Sanger sequencing of the same sample. (G) Mass spectrum plot of NRASQ61RL mutation. (H) Percentage of patients positive for circulating melanoma cells enriched by positive selection (MACS, M+), Parsortix™ or both. (I) Distribution of enriched CTCs in stage 4 patients prior to treatment. (J) CTCs enriched by positive selection, and Parsortix™ in NRAS/BRAF mutated patients, and not BRAF/NRAS mutated patients.

Figure 3

Combined enrichment method for the analysis of cellular subpopulations of circulating melanoma cells in patients. (A) Overview of patients with cutaneous, acral, mucosal and uveal melanoma of different disease stages, included in the study. (B) Time course of CTCs, enriched by positive selection (orange circles) and Parsortix™ (green circles) and their detected mutations, as well as LDH and S100 serum levels in patient #1. (C) Mass spectrum plot of BRAFV600E mutation and (D) Sanger sequencing of the same sample. (G) Mass spectrum plot of NRASQ61RL mutation. (H) Percentage of patients positive for circulating melanoma cells enriched by positive selection (MACS, M+), Parsortix™ or both. (I) Distribution of enriched CTCs in stage 4 patients prior to treatment. (J) CTCs enriched by positive selection, and Parsortix™ in NRAS/BRAF mutated patients, and not BRAF/NRAS mutated patients.

Figure 4

Comparison of CTC and ctDNA detection (A–C) CTCs, total ctDNA and ctDNA < 150 bp detected in patients with primary tumors (Breslow depth), lymph node metastasis (LN) and distant metastasis (M) (D,E) Time course of CTC detection, ctDNA, mutations detected in both CTCs and ctDNA, LDH and S100 serum levels in patients 2 and 3. (F) Mutational status for each detection method compared to the primary melanoma.

Figure 5

Prognostic value of liquid biopsy markers (A–E) Kaplan–Meier curves for each marker in the tested melanoma patient cohort. (F) Hazard ratio.