In vitro model of human mammary gland microbial colonization (MAGIC) demonstrates distinctive cytokine response to imbalanced human milk microbiota

ABSTRACT Despite the established concept of the human mammary gland (MG) as a habitat with its own microbiota, the exact mechanism of MG colonization is still elusive and a well-characterized in vitro model would reinforce studies of the MG microbiota development. We aimed to establish and characterize an in vitro cell model for studying MAmmary Gland mIcrobial Colonization (MAGIC) model. We used the immortalized cell line MCF10A, which expresses the strong polarized phenotype similar to MG ductal epithelium when cultured on a permeable support (Transwell). We analyzed the surface properties of the MAGIC model by gene expression analysis of E-cadherin, tight junction proteins, and mucins and by scanning electron microscopy. To demonstrate the applicability of the model, we tested the adhesion capability of the whole human milk (HM) microbial community and the cellular response of the model when challenged directly with raw HM samples. MCF10A on permeable supports differentiated and formed a tight barrier, by upregulation of CLDN8, MUC1, MUC4, and MUC20 genes. The surface of the model was covered with mucins and morphologically diverse with at least two cell types and two types of microvilli. Cells in the MAGIC model withstood the challenge with heat-treated HM samples and responded differently to the imbalanced HM microbiota by distinctive cytokine response. The microbial profile of the bacteria adhered on the MAGIC model reflected the microbiological profile of the input HM samples. The well-studied MAGIC model could be useful for studies of bacterial attachment to the MG and for in vitro studies of biofilm formation and microbiota development. IMPORTANCE The MAGIC model may be particularly useful for studies of bacterial attachment to the surface of the mammary ducts and for in vitro studies of biofilm formation and the development of the human mammary gland (MG) microbiota. The model is also useful for immunological studies of the interaction between bacteria and MG cells. We obtained pioneering information on which of the bacteria present in the raw human milk (HM) were able to attach to the epithelium treated directly with raw HM, as well as on the effects of bacteria on the MG epithelial cells. The MAGIC cell model also offers new opportunities for research in other areas of MG physiology, such as the effects of bioactive milk components on microbial colonization of the MG, mastitis prevention, and studies of probiotic development. Since resident MG bacteria may be an important factor in breast cancer development, the MAGIC in vitro tool also offers new opportunities for cancer research.

List of Supplementary file 1: List of supplementary tables: Table S1: Spearman rank order correlations between number of cells seeded on TransWells and TEER development parameters Table S2: Primers used in this study Table S3: Initial screening for expression of mucin genes in 20 × diluted pooled cDNA sample.
Table S4: Pre-experiment on growth of bacteria in the human milk sample ML 10 at 37 °C and their adhesion to differentiated MCF10A Table S5: Number of bacteria (colony forming units (CFU)), grown on BA, TSA, WCA-M and Rogosa agar in human milk samples and in culturepositive raw human milk samples and in trypsinised MCF10A cells after the 1-h adhesion assay.Table S6: Prevalence of genus-positive samples of raw human milk and of trypsinised MCF10A cells after the adhesion assay.

Supplementary table S7: Prevalence of species-positive samples of raw human milk and of trypsinised MCF10A cells after the adhesion assay.
Species of the colonies were determined with MALDI-TOF MS.

Species
Prevalence Transepithelial electrical resistance (TEER) development in relation to frequency of media change (every day-1-DAY vs every two days-2-DAY) and in relation to cell manipulation at room temperature (RT) or at 37 °C by using heating pad (HP).(B) Relation between sodium fluorescein flux and TEER during cell growth on PCCM.Cells reached the highest TEER in 14-17 days after seeding on PCCM (the experiment was performed without heating pad).The experiments were performed in 4 biological replicates.Results are shown as mean ± standard error.Relative gene expression for E-cadherin and tight junction proteins during the cell growth on porous cell culture membranes (PCCM).Results are shown as normalised gene expression relative to the expression in cells confluent on 48-well cell culture plates (CCP).Statistical significance was analyzed with non-parametric one-way ANOVA (Kruskal-Wallis statistic) and Dunn's multiple comparison test with adjustments of p-values to account for multiple comparisons; *p<0.05,**p<0.01.The experiment was performed in three independent experiments with two biological and two technical replicates.Dotted line marks two-fold change in gene expression.Alcian blue/PAS staining of differentiated MCF10A cells grown on porous cell culture membranes (PCCM) for 19 days.Cells grown on PCCM were washed once with HBSS and fixed in 4% paraformaldehyde.After dehydration in series of ethanol concentrations, the inserts were embedded in paraffin, sectioned, mounted on glass slides, and allowed to dry.Two μm thick paraffin tissue sections were stained with Alcian blue at pH 2.5 (A and B) and Periodic-Acid Schiff (C and D) according to the protocols of the Institute of Patology, Wild Animals, Fish and Bees of the Veterinary Faculty University of Ljubljana, covered and examined by light microscopy.

Supplementary table S1: Spearman rank order correlations between number of cells seeded on TransWells and TEER development parameters Supplementary table S2: Primers used in this study
Only mucin genes with Ct lower than 30 were analyzed for differential gene expression.

Table S5 : Number of bacteria (colony forming units (CFU)), grown on BA, TSA, WCA-M and Rogosa agar in human milk samples and in culture-positive raw human milk samples and in trypsinised MCF10A cells after the 1-h adhesion assay.
Genus of the colonies were determined with MALDI-TOF MS.NRID -no reliable edintification.