A vascularized in vivo melanoma model suitable for metastasis research of different tumor stages using fundamentally different bioinks

Abstract

Although 2D cancer models have been the standard for drug development, they don't resemble in vivo properties adequately. 3D models can potentially overcome this. Bioprinting is a promising technique for more refined models to investigate central processes in tumor development such as proliferation, dormancy or metastasis. We aimed to analyze bioinks, which could mimic these different tumor stages in a cast vascularized arteriovenous loop melanoma model in vivo. It has the advantage to be a closed system with a defined microenvironment, supplied only with one vessel-ideal for metastasis research. Tested bioinks showed significant differences in composition, printability, stiffness and microscopic pore structure, which led to different tumor stages (Matrigel and Alg/HA/Gel for progression, Cellink Bioink for dormancy) and resulted in different primary tumor growth (Matrigel significantly higher than Cellink Bioink). Light-sheet fluorescence microscopy revealed differences in vascularization and hemorrhages with no additional vessels found in Cellink Bioink. Histologically, typical human melanoma with different stages was demonstrated. HMB-45-positive tumors in progression inks were infiltrated by macrophages (CD163), highly proliferative (Ki67) and metastatic (MITF/BRN2, ATX, MMP3). Stainings of lymph nodes revealed metastases even without significant primary tumor growth in Cellink Bioink. This model can be used to study tumor pathology and metastasis of different tumor stages and therapies.

Keywords: Bioink; Melanoma; Metastasis; Tissue engineering; Tumor.

Graphical abstract

Fig. 1

Arteriovenous loop surgery: (a): upper left: opened surgical site with exposed vessels (red arrow artery, blue arrow vein) vessel diameter approx. 1 mm, upper middle: lifting of the venous graft in the other leg after flushing with heparin solution, upper right: first end-to-end anastomosis (11-0 sutures, black arrow) after flushing with heparin and attaching the micro clamps between artery and vein graft and preparation of the second between vein graft and vein, lower left: patent AV loop, second end-to-end anastomosis completed (black arrow) and clamps removed, lower middle: AV loop embedded in cast Alg/HA/Gel hydrogel in PTFE chamber, entrance is closed with fibrin, green pins for stabilization, lower right: closed chamber attached to the muscle, before closing the skin; scale bar = 5 mm; (b): casting schematics with cellular and acellular layers within the chamber; the cellular layer is located around the AV loop vessels to facilitate cell survival.

Fig. 2

Printing characteristics of three inks: (a): filament fusion test and grid structure test of Matrigel, Cellink Bioink, and Alg/HA/Gel, scale bars = 5 mm; (b): DCR quantification of the grid structure test of the three inks, data shown as mean and replicates; (c): rheological evaluation of the three inks (storage modulus G’, loss modulus G’’ and value of complex viscosity of Alg/HA/Gel was published previously [9]); data shown as mean ± SD of technical replicates.

Fig. 3

Material properties of three inks: (a): storage modulus E’ and loss modulus E″ of Matrigel, Cellink Bioink, and Alg/HA/Gel (E’ of Alg/HA/Gel was published previously [7]); data shown as mean ± SD of technical replicates; (b): scanning electron microscopy characterization of the inside of the hydrogels with quantification; scale bar = 5 μm; pore size shown as box plot (Tukey) of 3 ROIs, pore count per ROI shown as mean ± SD of pores > 0.5 μm.

Fig. 4

Melanoma cells in vitro with the three different bioinks: Ki67 stainings (green, DAPI in blue) of Mel Im embedded into the hydrogels show a decreased expression after 3 days in culture in Cellink Bioink compared to the other gels Matrigel and Alg/HA/Gel and the 2D culture; isotype controls are shown in the smaller boxes; scale bars = 100 μm.

Fig. 5

Chamber explants with the three different bioinks: (a): explants and histology (HE, HMB-45 [brown], Ki67 [green]) of representative samples of the groups Matrigel, Cellink Bioink, and Alg/HA/Gel, arrows point at loop vessels, scale bars = 1 mm; (b): vascularization of representative explants visualized via light-sheet fluorescence microscopy for CD31, the chamber entrance is on the left side, arrows point at loop vessels, scale bars = 2 mm; (c): quantification of the tumor area ratio, the CD163-positive macrophages per mm² (tumor area of Alg/HA/Gel was already published [7] and added here for direct comparison) of histological sections within the chamber shown as box plot (Tukey) with mean of ≥2 sections of biological replicates, *p ≤ 0.05 (Grubb's test for outliers, Shapiro-Wilk normality test followed by Kruskal-Wallis test and Dunn's multiple comparisons test) and MIA serum levels after 4 weeks shown as individual values plus mean of the group. *p ≤ 0.05 (Shapiro-Wilk normality test followed by Kruskal-Wallis test and Dunn's multiple comparisons test), (MIA levels of Alg/HA/Gel were already published [7] and added here for direct comparison).

Fig. 6

Comparison of the primary tumor growth between the Matrigel, Cellink Bioink, and Alg/HA/Gel groups: Tumors (HMB-45-positive, brown) in Matrigel and Alg/HA/Gel formed mainly around the AV loop vessel, small metastatic colonies were seen in Cellink Bioink around the AV loop vessel (asterisks mark the lumens, pictures of two exemplary animals); arrows point at exemplary metastatic cells in the vessels and the endothelium; scale bar = 100 μm.

Fig. 7

Typical histology of primary melanoma explants as an example of the Matrigel group: (a) MITF and BRN2 immunofluorescent staining of one exemplary explant with expression of both MITF and BRN2 heterogeneously to slightly varying extents; (b) ATX and MMP3 immunohistological stainings of the tumor masses; controls without primary antibody are shown in the smaller boxes; scale bars = 50 μm.

Fig. 8

Metastases in Cellink Bioink: (a): lymph node metastases of the Cellink Bioink group in Lymphnodus subiliacus (SiLN) and Lymphnodi iliaci mediales (MILNs), scale bar = 200 μm (b): quantification of metastases in the lymph nodes shown as per animal color-coded individual values (with mean of ≥2 sections of biological replicates) plus mean of the group, *p ≤ 0.05 (paired t-test).

Fig. 9

In vivo melanoma model: three bioinks that highlight different tumor stages in the AV loop melanoma model.