Introduction: Amidst the growing threat of antibiotic-resistant bacteria, novel approaches to combat infections are crucial. Our understanding of how human hematopoietic stem/progenitor cells (HSPC) traffic from their native bone marrow niche to the skin to undergo extramedullary granulopoiesis is limited. Using microfluidic 3D organ-on-a-chip technology, we engineered a 3D in vitrohuman skin equivalent that can emulate key features of dermal barrier function and leukocyte recruitment during simulated infections. With this model, we will establish a homeostatic circuit with a bone marrow organ-on-a-chip to evaluate the effects of soluble mediators that impact stem cell quiescence, granulopoiesis, leukocyte trafficking, HSPC trafficking, and extramedullary hematopoiesis.
Materials and Methods: We created a 3-compartment microfluidic device using 10% polydimethylsiloxane (PDMS) and standard soft lithographic methods. Compartments were separated by 30 μm pores, facilitating inter-compartment communication via diffusion. On day 0, the central compartment was embedded with human dermal fibroblasts in a fibrinogen-thrombin 3D matrix, flanked by monolayers of human cord blood-derived endothelial cells or primary human adult keratinocytes to mimic the vascular and dermal layers of the skin, respectively. The keratinocyte channel was perfused with KGM-Gold (Lonza), while the endothelial channel received EGM-2 (Lonza). On day 3, chips were fixed and labeled for CD31, Keratin 5, and Keratin 10, and imaged using confocal microscopy. Permeability of the dermal layer was assessed using 40kDa FITC Dextran. On day 3, freshly isolated human neutrophils were introduced (2E6/mL) into the vascular compartment after devices were incubated with synthetic lipopeptide Pam3CSK4 (10μg/mL).
Results, Conclusions, and Discussions: Our 3D skin model successfully mimicked key features of human skin. Confocal microscopy revealed tight PECAM-1/CD31 junctions between endothelial cells resulting in a confluent monolayer and an effective blood vessel mimic. Additionally, the distinct presence of Keratin 5 in keratinocytes confirms a stratum basale -like phenotype, which is associated with living basal-level keratinocytes residing next to fibroblasts (Thulabandu, 2017). Permeability of the keratinocyte layer was 9E-8 cm/s, tighter than what has been previously for endothelial cells in a similar model (Glaser, 2022). Peripheral blood neutrophils introduced in the vascular compartment attached, rolled, and extravasated into the dermal compartment, and this observation was augmented in the presence of Pam3CSK4.
Figure 1. Schematic of central chambers of device (A). Adult keratinocytes (top), normal human dermal fibroblasts (middle), and endothelial cells (bottom) labeled with Keratin 5 (red) and DAPI (cyan). (B). Endothelial cells labeled with CD31/PECAM-1 (magenta) and DAPI (cyan) (C). 40kDa FITC-Dextran (green) permeability through skin junctions (D). Neutrophils (green) attaching to endothelial cells and migrating up into fibroblast chamber after 30 min (E).
Conclusions: Our simple 3D skin model effectively emulates the features of the human skin, including dermal barrier function, an intact endothelial barrier, and leukocyte attachment and extravasation. Tight keratinocyte junctions simulate the barrier function of skin, making our system more physiologically relevant when eventually performing experiments pertaining to infection onset and resolution.
References: 1. Glaser, D. E. et al. Organ-on-a-chip model of vascularized human bone marrow niches. Biomaterials280, 121245 (2022).
2. Thulabandu, V., Chen, D. & Atit, R. P. Dermal fibroblast in cutaneous development and healing. Wiley Interdiscip. Rev. Dev. Biol.7, (2018).