Figure 1: Cardiac glycosides are preferentially toxic to melanomas by inhibiting the ATP1A1 Na⁺/K⁺ ATPase.

(a) Number of hUCB cells, primary melanocytes (hMEL1) or melanoma cells (derived from patients M481, M491 and M214) after 3 days in cultures supplemented with the cardiac glycosides gitoxin (10 μM), digitoxin (2.5 μM) or digoxin (2.5 μM; n=8 replicate cultures for each cell type from two independent experiments). (b) Number of primary melanoma cells or primary melanocytes or immortalized melanocytes (hiMEL) in cultures supplemented with 25 or 50 nM digitoxin for 3 days (n=6 replicate cultures for each cell type from two independent experiments). (c) IC₅₀ values for digitoxin in cultures of melanoma cells derived from 15 different patients as well as the A375 melanoma cell line and hUCB cells (n=1–4 independent experiments per cell type). (d) Plasma membrane potential of primary melanoma cells or primary melanocytes or immortalized melanocytes treated with 25 or 50 nM digitoxin for 16 h in culture (n=6 replicate cultures for each cell type from 2 independent experiments). (e) ATP1A1 transcript levels by microarray analysis of 33 patient-derived melanomas and normal melanocytes (**P<0.01; two-tailed Student’s t-test). Microarray data can be found at the NCBI GEO database under the accession code GSE83583. (f,g) Rescue of cell number (f) and plasma membrane potential (g) in digitoxin-treated melanoma cells after expression of cardiac glycoside-insensitive mouse Atp1a1, but not human ATP1A1 (n=3–10 replicate cultures per melanoma; each melanoma tested in 1–3 independent experiments). All data represent mean±s.d. Statistical significance was assessed by one-way (b,d) or two-way (f,g) analysis of variance followed by Dunnett’s multiple comparisons test. For all panels, each treatment was compared with controls (NS, not significant; *P<0.05; **P<0.01; ***P<0.001).