Engineering Tumor Stroma Morphogenesis Using Dynamic Cell-Matrix Spheroid Assembly

The tumor microenvironment consists of resident tumor cells organized within a compositionally diverse, three-dimensional (3D) extracellular matrix (ECM) network that cannot be replicated in vitro using bottom-up synthesis. We report a new self-assembly system to engineer ECM-rich 3D MatriSpheres wherein tumor cells actively organize and concentrate microgram quantities of decellularized ECM dispersions which modulate cell phenotype. 3D colorectal cancer (CRC) MatriSpheres were created using decellularized small intestine submucosa (SIS) as an orthotopic ECM source that had greater proteomic homology to CRC tumor ECM than traditional ECM formulations such as Matrigel. SIS ECM was rapidly concentrated from its environment and assembled into ECM-rich 3D stroma-like regions by mouse and human CRC cell lines within 4–5 days via a mechanism that was rheologically distinct from bulk hydrogel formation. Both ECM organization and transcriptional regulation by 3D ECM cues affected programs of malignancy, lipid metabolism, and immunoregulation that corresponded with an in vivo MC38 tumor cell subpopulation identified via single cell RNA sequencing. This 3D modeling approach stimulates tumor specific tissue morphogenesis that incorporates the complexities of both cancer cell and ECM compartments in a scalable, spontaneous assembly process that may further facilitate precision medicine.


Spon1
Figure S1: a. Top 10 CRC core matrisome genes based upon bulk RNA Seq of cells alone spheroids at Day 7 of culture.b.Heatmap of core matrisome genes and their relative expression patterns.Expression data was normalized within each cell line column.c.Bar chart showing the overlap between core matrisome proteins identified within ECM biomaterials (LC-MS) vs. the number of CRC core matrisome genes identified from the cells alone spheroids.d.Percentage similarity between the ECM biomaterials and CRC spheroids.e. Core matrisome expression by cell type from in vivo MC38 tumors determined by scRNA Seq.f. tSNE and violin plots of relative expression of the top 5 matrisome genes predominantly expressed by non-cancer cells within in vivo MC38 tumors.g.Heatmap of CRC ECM biomarkers found within ECM biomaterials.

Figure S6 :
Figure S6:Bubble plots of the top 10 IPA predicted activation and top 2 genes identified within the assigned categories by cell line a-b.MC38 c-d.CT26 e-f.HT-29 g.All conserved upstream regulators and their downstream targets across CRC cell lines predicted by IPA.h.Only direct interactions between upstream regulators and target genes i. Direct interactions between gene targets with multiple shared regulators j.IPA tumor microenvironment pathway showing an example of interaction between target genes with 3 identified upstream regulators k.Interaction network between the identified upstream regulators and their most common targets within the TME pathways.

Figure S7 :Figure S8 :
Figure S7: Representative images of HT-29 cells alone spheroids and MatriSpheres cultured in 384-well plates for 7 days.

MC38 CT26 HT-29 HT-29 + Collagen I Do Not Form Single Spheroids S2c S2d S2e MC38 CT26 HT-29 Cells Alone SIS ECM Collagen I Matrigel Figure S2: a
. Masson's Trichrome stained spheroid sections show varying concentrations of SIS ECM digest and Type I Collagen integrate within MC38 MatriSpheres.b.SIS ECM particles were combined with MC38 cells to form spheroids via hanging drop or ultra-low attachment round-bottom 96-well plates.c.Quantification of equivalent diameters at days 3, 5, and 7 of culture, n >= 37. Statistics are one-way ANOVA followed by Sídák's multiple comparisons test.# = Matrigel statistically different from cells alone on the same day, * = SIS ECM statistically different from cells alone on the same day.d.Quantification of metabolic activity via CellTiter-Glo 3D assay at day 7. Statistics were calculated by a two-way ANOVA followed by a Tukey's multiple comparisons test.e.Average number of cells per spheroid counted from manual dissociation of spheroids at day 7. f.Representative images of Masson's Trichrome stained spheroid sections at day 7. Scale bars are indicated on the images.Statistical significance where p < 0.05 is denoted with *, ≤ 0.01 with **, ≤ 0.001 with ***, and ≤ 0.0001 with ****.

Day 5 Day 3 Cells Alone 125 μg/mL SIS ECM 25 μg/mL Collagen I Day 10 Day 7 Day 12 Day 14 Day 16 Day 18 Day 21 S3b S3c Figure S3:
A. Representative images of HT-29 spheroids grown for 21 days.b.Quantification of HT-29 MatriSphere diameters at days 14 and 21.Statistics were calculated by a two-way ANOVA followed by a Sídák's multiple comparisons test.c.Quantification of metabolic activity of spheroids via CellTiter-Glo 3D.d.Rheological characterization of cell culture media and ECM biomaterials at experimental and traditional hydrogel formation concentrations.e. Gelation kinetics curves showing turbidity of ECM biomaterials at working ECM concentrations in comparison to control RPMI media.f.Gelation kinetics curves at varying ECM concentration by ECM biomaterial.Statistical significance where p < 0.05 is denoted with *, ≤ 0.01 with **, ≤ 0.001 with ***, and ≤ 0.0001 with ****. (1)

S4a S4b CA IX MC38 CT26 HT-29 MC38 + SIS ECM CT26 + SIS ECM HT-29 + SIS ECM
Figure S4: a. Representative images of CT26 + SIS ECM & HT-29 + SIS ECM spheroid sections stained for E-CAD, N-CAD, CHP, and DAPI.b.Quantification of CA IX staining from representative spheroid sections.Statistics were calculated by a one-way ANOVA followed by a Tukey's multiple comparisons test.c.Heatmaps of SIS ECM-incorporated spheroid sections stained with CHP to visualize denatured fibrillar collagen regions.d.

S5a S5b S5c S5d Normalized Expression S5e S5f # of Genes Figure S5:
a. Heatmap showing normalized gene expression and hierarchical clustering of mouse CRC lines (MC38 and CT26) by sample.b.PCA plot, frequency histogram and number of mouse genes across different significant thresholds.c.Venn diagram of significantly overlapping genes between MC38 and CT26 d. table of corresponding fold change and adj.p-value of each gene.e. Heatmap showing normalized gene expression and hierarchical clustering of human CRC line HT-29 by sample.f.PCA plot, frequency histogram and number of human genes across different significant thresholds g.GSEA plots showing the top gene set within each cell line based upon enrichment score (ES).ES and gene score plots showing a heatmap of the leading-edge genes between treatments (SIS ECM vs. cells alone spheroids).