Engineering cell heterogeneity into organs-on-a-chip

Abstract

Organ-on-a-chip development is an application that will benefit from advances in cell heterogeneity characterization because these culture models are intended to mimic in vivo microenvironments, which are complex and dynamic. Due in no small part to advances in microfluidic single cell analysis methods, cell-to-cell variability is an increasingly understood feature of physiological tissues, with cell types from as common as 1 out of every 2 cells to as rare as 1 out of every 100 000 cells having important roles in the biochemical and biological makeup of tissues and organs. Variability between neighboring cells can be transient or maintained, and ordered or stochastic. This review covers three areas of well-studied cell heterogeneity that are informative for organ-on-a-chip development efforts: tumors, the lung, and the intestine. Then we look at how recent single cell analysis strategies have enabled better understanding of heterogeneity within in vitro and in vivo tissues. Finally, we provide a few work-arounds for adapting current on-chip culture methods to better mimic physiological cell heterogeneity including accounting for crucial rare cell types and events.

Figure 1

Cell heterogeneity may be considered on a range of levels in physiology and disease. (A) Lung alveolar epithelial cells, or pneumocytes, are classified in two subtypes, which occur, approximately, in a 1:2 ratio in the lung. (B) Among populations of cancer cells, transcriptional heterogeneity may lead to rare drug-resistant cells as was shown by Shaffer and colleagues in a melanoma cell population. (C, D) Additionally, tumorigenic cells, when identified by particular markers, can be found across the spectrum of rarity depending on the tumor site, but also the patient. Lgr5⁺ intestinal adenoma cells and CD34⁺CD38⁻ acute myeloid leukemia cells are such examples.

Figure 2

A hierarchy of cell-to-cell variability is illustrated. While much of organ-on-a-chip development has considered tissue heterogeneity (co-culturing multiple cell types within the same organ) and body/human-on-a-chip development has considered organ-to-organ differences and scaling, a deeper level of cell-to-cell variability exists that is less-characterized by organ-on-a-chip engineers. This heterogeneity within a single cell type may arise through genetic mutations, but also specific epigenetic mechanisms which are largely unknown. Phenotypic heterogeneity is characterizable, however, with advancements in the past few years using single cell techniques, and these techniques may have application in organ-on-a-chip development by assessing cells used in the culture model to the heterogeneity found in the microenvironment to be mimicked.

Figure 3

Modeling the immune component of the small airway microenvironment on-chip. (A) Multiple immune cell types: alveolar macrophages, intraepithelial T cells and dendritic cells reside within the epithelium to defend against pathogens with innate or adaptive immune responses. Neutrophils do not generally leave the circulation at homeostasis, but may extravasate to participate in an immune response. (B) Two lung-on-a-chip adaptations of the lung microenvironment are depicted. Left is a generalized, homeostatic lung with immune component consisting of resident immune cells like alveolar macrophages, intraepithelial T cells or dendritic cells. Right is a proposed on-chip model of dynamic immunity with airway epithelial injury leading to extravasation of chemotactic neutrophils at the wound site.

Figure 4

Cell-to-cell variability has been assessed in excised tissues, organoids and in 2D cultures. General workflow for cell-to-cell heterogeneity characterization are depicted using three example cases stacked top to bottom. Setty and colleagues collected proteomic data from primary thymocytes using mass cytometry and developed an algorithm to map cells along lineage progression. Grün and colleagues evaluated cell heterogeneity in intestinal organoids using scRNA-seq as well, but proposed a more advanced clustering algorithm known as RaceID (rare cell identification) to resolve finer cell-to-cell differences, leaving cell type clusters with as few as a single cell. Even a 2D-cultured cell line was shown by Shaffer and colleagues to contain RNA copy-number variability when using RNA FISH. Cell population heterogeneity was quantifiable using the Gini index, which for certain genes correlated with chemotherapy resistance. Far fewer accounts have been published on cell-to-cell heterogeneity within organ-on-a-chip devices, and these approaches may provide key insights to advance their predictive accuracy.

Figure 5

Visual schematics of workarounds for characterizing cell-to-cell variability and rare cells or events are depicted. (A) By miniaturizing too much or by using too few cells, rare phenotypes (indicated by black arrowheads) can be lost. (B) Increasing experimental replicates will create microenvironments with rare cells, providing a range of outputs. (C) Increasing experimental duration or frequency of data acquisition may capture outputs from transient rare phenotypes (indicated by black arrowheads). (D) Beginning an organ-on-a-chip experiment with multipotent stem cells, in combination with biochemical signals and matrix will create a heterogeneous cell population including progenitors and multiple lineages.