Modeling Melanoma In Vitro and In Vivo

Abstract

The behavior of melanoma cells has traditionally been studied in vitro in two-dimensional cell culture with cells adhering to plastic dishes. However, in order to mimic the three-dimensional architecture of a melanoma, as well as its interactions with the tumor microenvironment, there has been the need for more physiologically relevant models. This has been achieved by designing 3D in vitro models of melanoma, such as melanoma spheroids embedded in extracellular matrix or organotypic skin reconstructs. In vivo melanoma models have typically relied on the growth of tumor xenografts in immunocompromised mice. Several genetically engineered mouse models have now been developed which allow the generation of spontaneous melanoma. Melanoma models have also been established in other species such as zebrafish, which are more conducive to imaging and high throughput studies. We will discuss these models as well as novel techniques that are relevant to the study of the molecular mechanisms underlying melanoma progression.

Keywords: 3D models; animal models; genetically engineered mouse models (GEM); melanoma; spheroid models; xenograft models; zebrafish models.

Figure 1

(A) Image showing two-dimensional cell growth in adherent cell culture in a plastic culture dish; (B) Differential interference contrast image overlaid with DAPI stain (nuclei, blue) demonstrating the morphology of melanoma cells in 2D culture, low and high magnification; (C) Deconvolved immunofluorescence image representing the highly complex internal structure of melanoma cells, labeled with markers for different endosomes or secretory vesicles (pink, red, or green) and the nucleus (blue). Scale bars: (B) 50 μm (left), 12 μm (right); (C) 6 μm.

Figure 2

Melanoma spheroid model. (A) Image demonstrating oxygen and nutrient gradient within a 3D spheroid, central necrosis and invasion of melanoma cells into the tumor stroma; (B) Growth and invasion behavior of melanoma spheroids reflects that of the original tumors: radial (RGP), vertical growth phase (VGP), metastasis (met); (C) Growth and invasion of a spheroid derived from a metastatic cell line (numbers: time in h); (D) Spheroid treated with increasing doses of the BRAF inhibitor vemurafenib (−, +, ++, +++) and stained with a live/dead assay (green: calcein-AM, live; red: ethidium bromide, dead). Note the increasing growth and invasion inhibition, as well as cell death.

Figure 3

3D skin reconstruct model. (A) Model: Keratinocytes and fibroblasts in human skin reconstructs produce a basement membrane (stained with a monoclonal antibody against collagen IV). Note the clear demarcation of the basement membrane; (B) Skin: Human fetal foreskin stained with a monoclonal antibody against α6 integrin (green) and the melanocyte marker HMB45 (red), highlighting the continuous basement membrane between epidermis including basal keratinocytes and melanocytes (co-localization, yellow) and dermis. Note the similarity between model and original; (C) Melanoma cells recapitulate original tumor phenotype in reconstructed skin. Melanoma cells were stained for S100. Radial growth phase (RGP) melanoma cells proliferate and form nests in the epidermis, but do not invade the dermis. Vertical growth phase (VGP) cells cross the basement membrane, invade and proliferate in the dermis. Cells from metastatic (met) melanoma rapidly invade deep into the dermal compartment. Figure adapted from Haass et al., 2005 [12] and Santiago-Walker et al., 2009 [11]. Scale bars: 30 μm.

Figure 4

3D neoangiogenesis model. Human microvascular endothelial cells grown in a collagen gel containing green fluorescent protein (GFP)-expressing fibroblasts (green in B) for 144 h to allow the formation of a three-dimensional vascular network, were fixed and incubated with an antibody against CD31 (A; red in B) and stained with DAPI (blue in B). Tube formation and branching frequency are used to quantify anti-angiogenic drug effects.

Figure 5

Comparison of 3D melanoma spheroid and xenograft. (A) Untreated spheroid; (B) spheroid treated with 10 μM selumetinib (MEK1/2 inhibitor AZD6244; Astra Zeneca) and stained with a live/dead assay (green: calcein-AM, live; red: ethidium bromide, dead). Note growth and invasion inhibition as well as increased cell death in the treated spheroid; (C) Untreated xenograft; (D) xenograft after treatment with 50 mg/kg selumetinib. Note the growth inhibition of the treated tumor. Sections of an untreated (E) and treated (F) xenograft stained for BrdU (green; proliferation marker), S100 (red, used here as a melanoma marker), and DAPI (blue, nuclear marker) showing a decrease in BrdU uptake after selumetinib treatment. Figure adapted from Haass et al., 2008 [24].

Figure 6

Spontaneous melanoma, which manifested on the ear of a UVB irradiated Cdkn2a^−/−, Tyr-HRAS mouse. Staining positivity for all three markers (S100, Melan A and HMB45) as well as their pattern and localization are indicative of melanoma. Scale bars are 2 mm for the low (5×) and 200 μm for the high magnification (20×) images.

Figure 7

‘Upside-down triangle approach’ (A) and ‘diamond approach’ (B).