3D Culture Modeling of Metastatic Breast Cancer Cells in Additive Manufactured Scaffolds

Cancer biology research is increasingly moving toward innovative in vitro 3D culture models, as conventional and current 2D cell cultures fail to resemble in vivo cancer biology. In the current study, porous 3D scaffolds, designed with two different porosities along with 2D tissue culture polystyrene (TCP) plates were used with a model breast cancer human cell line. The 3D engineered system was evaluated for the optimal seeding method (dynamic versus static), adhesion, and proliferation rate of MDA-MB-231 breast cancer cells. The expression profiles of proliferation-, stemness-, and dormancy-associated cancer markers, namely, ki67, lamin A/C, SOX2, Oct3/4, stanniocalcin 1 (STC1), and stanniocalcin 2 (STC2), were evaluated in the 3D cultured cells and compared to the respective profiles of the cells cultured in the conventional 2D TCP. Our data suggested that static seeding was the optimal seeding method with porosity-dependent efficiency. Moreover, cells cultured in 3D scaffolds displayed a more dormant phenotype in comparison to 2D, which was manifested by the lower proliferation rate, reduced ki67 expression, increased lamin A/C expression, and overexpression of STCs. The possible relationship between the cell affinity to different extracellular matrix (ECM) proteins and the RANK expression levels was also addressed after deriving collagen type I (COL-I) and fibronectin (FN) MDA-MB-231 filial cell lines with enhanced capacity to attach to the respective ECM proteins. The new derivatives exhibited a more mesenchymal like phenotype and higher RANK levels in relation to the parental cells, suggesting a relationship between ECM cell affinity and RANK expression. Therefore, the present 3D cell culture model shows that cancer cells on printed scaffolds can work as better representatives in cancer biology and drug screening related studies.

We are interested in the amount of the fiber parts (between two junctions or between a junction and an edge of fiber) created per pair of layers, which will be equal to We should add to that the number of fiber parts of the last column, which is .
Hence, in total: fiber parts are created in each pair of layers.
If the block consists of odd number of layers, then we should add also the cylinders of the upper layer, which are in number and their surface equals to Then, their total surface area will be We should, however, take also into account the available area of the junctions. Only the junctions of the upper and the lower layer contribute to the final available surface area. These junctions will have an area equal to the square ΑΒCD (Scheme 2) subtracted by the 4 (equal area) circular sections with each of them having an area equal to FAE.

Scheme 1
The side of the square is defined as and the radius of each circular section is Hence, .
Therefore, the area in each joining point will be "#$%&'($ = 270 ! − 4 • 25 The total number of junctions is (of the top and bottom layer) with a total surface area of Furthermore, in the 4 external sides of every pair of layers, we have ellipses derived from the fibers spreading at the joining points (until they solidify). These ellipses have totally a surface , corresponding to 1 pair of layers. For the whole block we should multiply by . So, we will have the surface If, though, the number of layers is not even, then we will have more ellipses. So, the total area will be slightly different and equal to Finally, the cylindrical surfaces at the internal sides of the joints should be also added, since they contribute to the final available surface. For each juncture, there are two surfaces (in the front and in the back).
In case of even number of layers, the number of these surfaces will be with their area being equal to 4 ! ( − 1) In case of odd number of layers, these surfaces will be in multitude of with their total area In summary, the total available surface area of the block for even number of layers will be: where is the amount of pair layers.
For prime number of layers: In case of prime number of layers: (integral number) For 20 layers, d2=250 μm and d3=150 μm: The abovementioned formulas calculate the surface corresponding to a square structure of width and length equal to d (orange square). However, this surface area is much higher than the real one.
Based on the scheme below, the lower and the upper bound of the total area of the cylinder will be the area of the green and red polygon respectively.
Determination of the lower bound of the scaffold area (green polygon).

The angle
Hence, In the bleu right triangle, we have hence, The side of the blue square, then, will be Therefore, the lower bound of the area will be given by the formula  Hence, But ΑΕ=ΑF=GB=BH. Hence, The side of the orange square is The upper bound of the scaffold area will be equal to the surface of the red polygon,

Finally,
The max value is closer to the real one because of the punching process. We consider, therefore, the final surface area of a scaffold to be equal to The aforementioned mathematical theoretical model was validated by comparing its results for 2 scaffolds of specific parameters, whose surface was measured by using μCT.

Figure S1
Figure S1. Brightfield images of methylene blue stained scaffolds (with big pores) seeded dynamically (a, c) and statically (b, d). Scaffolds were coated with 100 μg/ml COL-I. Each scaffold was seeded with 0.5•10 6 cells. Representative stereomicroscope images of the scaffolds seeded dynamically at day 1 (a) and at day 7 (c) showing that cells were less confluent in case of dynamic seeding. Representative image of the scaffold seeded statically at day 1 (b) and at day 7 (d) where cells were more efficiently attached. After 7 days of culture, scaffolds were more cell confluent.
Scale bar indicates 1 mm (a, b, c, d). A higher magnification illustrating a pore unit is also shown for both time points seeded statically with a scale bar of 500 μm. Scale bar: 100 μm. Cells appear to form a monolayer (without any clumps) throughout the whole surface. The stack of parallel fibers (in z direction) is the reason of the dense cell area appearing on the lower fiber. Scaffolds were pre-treated with Sudan Black to hinder their autofluorescence.