An Engineered Paper‐Based 3D Coculture Model of Pancreatic Cancer to Study the Impact of Tissue Architecture and Microenvironmental Gradients on Cell Phenotype

Abstract

The spatial configuration of cells in the tumor microenvironment (TME) affects both cancer and fibroblast cell phenotypes contributing to the clinical challenge of tumor heterogeneity and therapeutic resistance. This is a particular challenge in stroma‐rich pancreatic ductal adenocarcinoma (PDAC). Here, a versatile system is described to study the impact of tissue architecture on cell phenotype using PDAC as a model system. This fully human system encompassing both primary pancreatic stellate cells and primary organoid cells using the TRACER platform to allow the creation of user‐defined TME architectures that have been inferred from clinical PDAC samples and are analyzed by CyTOF to characterize cells extracted from the system. High dimensional characterization using CyTOF demonstrates that tissue architecture leads to distinct hypoxia and proliferation gradients. Furthermore, phenotypic markers for both cell types are also graded in ways that cannot be explained by either hypoxia or coculture alone. This demonstrates the importance of using complex models encompassing cancer cells, stromal cells, and allowing control over architecture to explore the impact of tissue architecture on cell phenotype. It is anticipated that this model will help decipher how tissue architecture and cell interactions regulate cell phenotype and hence cellular and tissue heterogeneity.

An Automation Workflow for High‐Throughput Manufacturing and Analysis of Scaffold‐Supported 3D Tissue Arrays

“…As PSCs are more contractive than cancer cells, we used 6 mg mL −1 of type I bovine collagen for tissue manufacturing to ensure a good distribution of PSC cells after 3 days of culture. [ 43 ] Although 6 mg mL −1 type I bovine collagen is slightly more viscous than the 3 mg mL −1 type I bovine collagen that we optimized the OT2 protocols with, no adaptation of the seeding sequences was needed to achieve consistent deposition. In addition to KP4 cell monoculture and PSCs monoculture, we used two coculture ratios to create distinct microenvironments: 70% KP4 cells with 30% PSCs and 50% KP4 cells with 50% PSCs.…”

“…The quality of the confocal images achievable from SPOT cultures however will provide a useful dataset as such segmentation algorithms become available. Given previous work from our lab showing the compatibility of our digestion protocols for enabling flow‐cytometry and CyTOF, [ 43 ] we anticipate these workflows will enable single‐cell organoid analysis.…”

“…Although no single-cell analyses were performed with PDOs seeded in a paper scaffold in this study, it will be relatively simple to adapt and perform personalized single-cell analysis using CyTOF, as similar studies have been performed on organoids but at a smaller scale. [43] Further, the described workflow takes advantage of an open-source and inexpensive OT2 liquid handler, to allow for easy implementation confirming SPOT platforms coupling with OT2 seeding support patient-derived organoid cell culture. d) The health/polarization of organoids on day 12 was assessed based on CK19 and ZO-1 immunostaining.…”

“…Using CytoexploreR package (https://dillonhammill.github.io/CytoExploreR/), live single‐cell population was gated using Gaussian parameter discrimination as described previously. [ 43,59 ]…”

“…Harvested cells were fixed, permeabilized, barcoded, and pooled as previously mentioned. After pooling, samples were incubated with a cocktail of metal‐tag conjugated primary antibodies (Table S1, Supporting Information, conjugation was performed as previously described) [ 43 ] in 3% BSA/PBS for 30 min at RT, sample were agitated after 15 min to prevent cell sedimentation that could lead to uneven staining. After staining, samples were washed twice with 3% BSA/PBS.…”