A complex 3D human tissue culture system based on mammary stromal cells and silk scaffolds for modeling breast morphogenesis and function

Abstract

Epithelial-stromal interactions play a crucial role in normal embryonic development and carcinogenesis of the human breast while the underlying mechanisms of these events remain poorly understood. To address this issue, we constructed a physiologically relevant, three-dimensional (3D) culture surrogate of complex human breast tissue that included a tri-culture system made up of human mammary epithelial cells (MCF10A), human fibroblasts and adipocytes, i.e., the two dominant breast stromal cell types, in a Matrigel/collagen mixture on porous silk protein scaffolds. The presence of stromal cells inhibited MCF10A cell proliferation and induced both alveolar and ductal morphogenesis and enhanced casein expression. In contrast to the immature polarity exhibited by co-cultures with either fibroblasts or adipocytes, the alveolar structures formed by the tri-cultures exhibited proper polarity similar to that observed in breast tissue in vivo. Only alveolar structures with reverted polarity were observed in MCF10A monocultures. Consistent with their phenotypic appearance, more functional differentiation of epithelial cells was also observed in the tri-cultures, where casein alpha- and -beta mRNA expression was significantly increased. This in vitro tri-culture breast tissue system sustained on silk scaffold effectively represents a more physiologically relevant 3D microenvironment for mammary epithelial cells and stromal cells than either co-cultures or monocultures. This experimental model provides an important first step for bioengineering an informative human breast tissue system, with which to study normal breast morphogenesis and neoplastic transformation.

Fig. 1

(A) DNA quantification analysis by PicoGreen^™ DNA assay showed cells proliferating over time in different experimental groups. (B) Quantitative analysis of Ki67⁺ staining MCF10A cells counted in different microscopic fields. An inhibitory effect of stromal cells on epithelial cell growth was observed in the tri-cultures and co-culture during the first week. A significant difference between tri-culture and co-cultures was observed (*p < 0.05), while no difference exists between the two co-culture groups (^#p > 0.05). The percentage of positive staining cells in each group decreased over time.

Fig. 2

Growth profile and viability of MCF10A cells in different cultures developed on silk scaffolds (at day 6). CLSM images showed both alveolar and duct-like structures formed in co-cultures (Top row, a₁-c₁), while only alveolar structures were observed in the monocultures (d₁). (Red, MCF10A cells labeled with Di I^™; Green, stromal cells labeled with CMFDA). Viability of MCF10A cells in different groups detected by viability staining (middle row, a₂-d₂)). H&E staining showed morphological characteristics of the epithelial structures formed by MCF10A cells in different groups (bottom row, a₃-d₃). (Arrowhead notes duct-like structure, asterisk notes alveolar structure).

Fig. 3

Quantitative analysis of the percentage of duct-like epithelial structures generated by MCF10A cells in the different groups. Compared with the co-cultures, the highest percentage of ductal structures (N_duct/N_{(acini +duct)} ×100%) was found in the tri-culture (*p < 0.05). No significant difference exists between the two co-culture groups (^#p > 0.05). The percentage of ductal structures in each group increased over time.

Fig. 4

Morphological characteristics of the epithelial structures formed by MCF10A cells on silk scaffold: phalloidin-FITC staining (A) showed lumen formation in the alveolar structure. Positive E-cadherin and collagen IV staining indicated the presence of tight junctions and basement membrane and thus the integrity of the tissue-like structures (B). TEM images showed the ultrastructures of epithelial MCF10 cells in the tri-cultures (D_1–2). Cell-cell junctions, such as tight junctions (white arrows), desmosomes (white asterisk) and hemi-desmosomes (white arrowhead) were observed (scale bar=300nm).

Fig. 5

Immunostaining with sialomucin (red) and GM130 (green) showed the polarity of the alveolar structures formed by tri-cultured MCF10A cells on silk scaffold (b₁-b₃, day 6), which is similar to the positive control of human native breast tissue (a₁-a₃) in comparison to the co-cultures (c₁-c₃, d₁-d₃). The cell nucleus was counter-stained by DAPI (blue). Reversed polarity was also observed in the monoculture (e₁-e₃, day 6).

Fig 6

Transcript expression levels of α-casein (A) and β-casein (B) by real-time RT-PCR at the indicated time points. A significant increase in the expression of both α and β-casein was detected in MCF10A cells with both stromal cells. (n=3, *p < 0.05, ^#p > 0.05). Consistent with the result of RT-PCR, a more intensely stained casein protein was observed in the tri-culture group (C₁) in comparison with co-culture group with HMF (C₂) and monoculture group (C₃) when detected by immunostaining.