PMC

Inhibition of human melanoma xenograft growth, recurrence, and metastases in mice. (A) Effects of cyc treatment on s.c. MeWo/LacZ melanomas in nude mice as compared with tumors treated with cyclodextrin carrier alone. Arrow points to the disappearance of cyc-treated tumors. (B) Melanoma recurrence after sustained treatment for 3 or 10 days after tumor disappearance but not after 20 days (thick arrow). Each line represents a tumor except where noted. (C–H) Nude mice with carrier-only treated MeWo/LacZ melanomas (C and F) or treated with cyc until partial (D and G) or complete (E and H) disappearance. (I–K) Histological analyses after X-Gal-stained MeWo/LacZ cells in tumors similar to those in C–H. Arrows point to tumor sites or scars. (L–N) Effects of lentiviral-mediated SMOH silencing in MeWo cells injected s.c. (L and N) as compared with control transduced cells (L and M). Only the minus (for control, black) or the plus (for shRNA-SMOH, red) error bars are shown (L). (O–Q) Histogram (O) and images of dissected lungs after X-Gal staining (O–Q) showing the inhibition of i.v.-injected MeWo/LacZ lung metastases by systemic cyc treatment as compared with treatment with carrier alone. The size of colonies was scored (O, open bars >10 cells; filled bars <10 cells) accounting for macro- and micrometastases (arrows, P). P Inset shows a histological section of a MeWo/LacZ metastasis. (Scale bars: C–E, M, and N, 1 cm; F–H, 0.3 cm; I–K, 1 mm; O, 0.7 cm; P and Q, 1.4 mm; P Inset, 120 μm.)

RAS-MEK, AKT, and HH signaling mutually interact and regulate the subcellular nuclear localization and activity of GLI1. (A) Synergistic and additive reduction in BrdU incorporation in human melanoma cell lines after combined treatment for 48 h with cyc and the MEK inhibitor (MEK inh.) U0126 or the AKT inhibitor (AKT inh.) SH6. Values are cyc over tom (at equal concentrations) or each inhibitor over DMSO. (B) Inhibition of GLI1- or NRAS^Q61K (RAS*)-induced BrdU incorporation 48 h in SK-Mel2 cells by inhibitors of MEK/AKT or HH signaling, respectively. FK, forskolin. Only β-gal+/BrdU+ cotransfected cells were counted. Values are over lacZ transfection alone. (C and D) GLI luciferase reporter assays with a multimerized GLI-binding site reporter (blue), with mutant GLI-binding sites (GLI-BS) mut series in C, a PTCH1 reporter (gray series), or with a mutant binding site (PTCH-BSmut series in C) in SK-Mel2 (C) and COS7 (D) cells. (E and F) Subcellular localization of exogenous GLI1 in SK-Mel2. Examples (E) and histogram (F) of transfected cells. Here and in all transfection assays in each case, >500 cells were counted. (G) Inhibition by SUFUH of GLI1 reporter activity enhanced by RAS*. Transfection was with equimolar amounts except for GLI1+RAS*+SUFUH (1:1, 1:2; 1:4, 1:6). (H and I) Subcellular localization of endogenous GLI1 protein in SK-Mel2 cells treated as indicated (H) and quantification of results (I). (J) Change in gene expression 4 h after inhibition of endogenous MEK or AKT in SK-Mel2 and primary Me-3 cells by quantitative PCR. (K) Scheme of the interactions of HH-GLI signaling (in green) and peptide growth factor (PGF)-receptor tyrosine kinase (RTK)-RAS-RAF-MEK/phosphatidylinositol 3-kinase (PI3K)-AKT signaling (in blue) at the level of GLI. Negative modulators are in orange. In addition to a possible positive action on GLI1, RAS/AKT signaling may act directly on GLI1 and/or block the action of inhibitors of GLI function (INH), such as PKA or SUFUH.

Human melanocytes and melanomas harbor active SHH-GLI signaling. (A and B) H&E (Left), in situ hybridization for SHH, GLI1, and PTCH1 (three images, Center), and MITF immunolocalization (Right) in sections of human scalp hair follicles. Arrows point to melanocytes in the outer root sheath (Upper) and matrix (Lower). (C) Effects of Shh, cyc, and tom on BrdU incorporation 48 h after treatment of human foreskin melanocytes. (D and E) Effects of SHH as compared with PBS (D) and cyc as compared with tom (E) on gene expression 4–8 h after treatment determined by quantitative PCR. (F) H&E staining (Left), in situ hybridization for the indicated genes (blue in five images, Center), and immunolocalization of MITF and MELAN-A (red in two images, Right; DAPI nuclear counterstain in blue) in a human melanoma skin metastasis. Clinical data are supplied in SI Tables 1–3. Inset shows the result of hybridization with sense PTCH1 probes. See also SI Fig. 7. (G) RT-PCR analyses of five primary cultures used (Me-1–5). (H) Decrease in gene expression 4 h after 10 μM cyc treatment as compared with tom (nd, not detected). (I) Changes in BrdU incorporation in human primary melanomas 48 h after addition of different doses of cyc or tom, as indicated. (J) Recovery of viable cells after 10 or 20 day treatments with 10 μM cyc and 10 days recovery (rec) without it. Arrows point to lack of recovery. (K) Change in BrdU incorporation in primary melanomas 48 h after lipofection with siRNAs. (L) Reduction of viable cells after long-term treatment with GLI1 siRNA lipofecting anew every 2 days. (M) GLI effects on SK-Mel2 proliferation and rescue of the effects of cyc. GFP-expressing plasmids were cotransfected, and only GFP+ cells were counted. Asterisks denote significant differences from control (pCS2 empty vector). Numbers refer to number of positive cells per 1,000. ns, not significant. (N) Rescue of the inhibition of BrdU incorporation by 10 μM cyc through interference with GLI3. The arrowed brackets denote partial or total rescue. Error bars are SEM and are not provided for histograms showing ratios. Instead, asterisks denote significant values as compared with controls (P < 0.05). All results in – derive from at least three independent experiments. (Scale bars: A and B, 100 μm; F, 70 μm.)

Oncogenic RAS-induced mouse melanomas harbor and require sustained Hh-Gli signaling. (A) Quantitative PCR of the expression of the NRAS^Q61K transgene and tyrosinase (Upper) and of Gli1 and Ptch1 (Lower), in individually dissected mouse melanomas. As control, Ink4a^−/− lymph nodes were used to quantify normal gene expression: ax, axillary; cer, superficial cervical; ing, inguinal. Names of tumors: The first number denotes the mouse, and the last number denotes the melanoma sample. PM, primary skin melanoma; LNM, lymph node metastasis. (B–G) In situ hybridization of a skin melanoma for Gli1 (B), Gli1 sense control (C), Ptch1 (D), Ptch1 sense control (E), and Shh (F) probes. Specific hybridization is blue (B and D), and pigment granules are brown (B–G). (G) H&E-stained section provided as control. (H) Change in BrdU incorporation in three primary skin melanomas, 11 lymph node metastases from four different mice, and MEFs after treatment, as indicated, for 48 h. (I) Rescue of the inhibitory effects of cyc by GLI1 overexpression in 1-PM2 (Left) and 1-LNM5 (Right). The numbers are per 1,000 cells. Only doubly BrdU+/GFP+ cells were counted. ∗, significative change as compared with tom treatment; P < 0.01; ns, no significative change. (J) Reduction of the number of cells of 1-PM2 and 1-LNM-5 after treatment with cyc at 5 μM for 2, 5, and 10 days (d). Values are given as ratios of the number of cells present over those treated with tom. This effect is preceded by a reduction in BrdU incorporation and an increase (up to 5-fold) in activated Caspase-3+ apoptosis (not shown). (K) Inhibition of BrdU incorporation in two primary mouse melanoma cultures after lipofection of siRNAs against Gli1 or a control siRNA with efficiencies >70%. (L–N) Overt appearance of cutaneous melanomas (arrows) in tyr→NRAS^Q61K;Ink4a^−/− mice soon after detection (L) and 5 days (d5) after treatment with carrier alone (M and N). (O–Q) Overt appearance of cyc-treated melanomas after 5 (O and P) or 4 (Q) days of treatment showing concave depressions (arrows). Adjacent skin was unaffected. (R) Change in tumor volume in control and cyc-treated tumors. (S–V) Histological analyses of carrier-treated melanoma (S) and overt normal skin or tyr-RASQ61K; Ink4a^−/− mice (U) after H&E staining showing the abundant presence of pigmented cells. Cyc treatment (T and V) reduced tumor volume (T) and produced an overlying scar. It also decreased the number of pigmented cells in overtly normal skin adjacent to the tumor (V). (Scale bars: B–D, 20 μm; E–G, U, and V, 120 μm; L–Q, 1.3 mm; S and T, 1 mm.)