P15 - Revealing Extramedullary Granulopoiesis in a 3D Microphysiological Skin Model

Introduction: : The emergence of antibiotic-resistant bacteria necessitates alternative strategies to combat infections. Traditional models of skin infection describe a loop where hematopoietic stem and progenitor cells (HSPCs) in the bone marrow produce neutrophils that then intravasate into the blood. These neutrophils, or polymorphonuclear cells (PMNs), then sense infection through cytokine and integrin signals from the skin, arrest in post-capillary venules at the site of infection and migrate into the extracellular space to clear the infection. Studies in mice, however, reveal an interesting phenomenon wherein HSPCs themselves traffic out of the bone marrow into blood, and can arrest and migrate into the site of skin infections (Falahee 2017, Kim 2011). Once there, HSPCs undergo extramedullary granulopoiesis (EMG) to produce local neutrophils capable of clearing infection. In fact, this mechanism is crucial to survival in mice following the inoculation of an S. aureus infection. A similar mechanism in humans has not been reported as it is difficult to reveal subpopulations of immune cells in infected tissue in vivo; however, if this mechanism exists, it could be leveraged to combat antibiotic resistant infections. Using a 3D in vitro organ-on-a-chip model of human skin (SkinChip) that includes key features of human skin, such as the dermal barrier and microcirculation during simulated infections, we can investigate the potential of EMG. This unique model allows for real-time imaging and high temporal and spatial resolution of PMN and HSPC attachment, rolling, and arrest along endothelium, revealing dynamic interactions between skin tissue and immune cells.

Materials and

Methods: : Our microphysiological system (MPS) utilizes a relatively simple PDMS-based microfluidic device made using standard soft lithographic methods. The model features three parallel compartments: a central dermal chamber flanked by vascular and epidermal chambers that communicate through 30 μm pores. Chambers are seeded with dermal fibroblasts in the dermal compartment (suspended in a 3D, 10 mg/mL fibrinogen gel), cord blood-derived endothelial cells lining the vascular chamber, and adult keratinocytes in the epidermis. Cells were labeled with CellTracker (A) and seeded in the SkinChip, revealing a compartmentalized skin architecture after three days. Keratinocytes crowding pores create tight junctions that dramatically limit free diffusion of mediators from the top chamber to the rest of the chip. The bottom vascular chamber creates a confluent endothelial barrier that models a perfusable blood vessel. Chips were stimulated with three conditions modeling inflammation: 10ng/mL TNF-α/IL-1β (“Hi AB”), 1 μg/mL Pam3CSK4 (“Hi P”), or 100pg/mL TNF-α/IL-1β & 10ng/mL Pam3CSK4 (“Lo PAB”) for 24 hours before labeling with primary anti-ICAM-1 antibody (B). Relative ICAM-1 surface expression was quantified in ImageJ/Fiji (C). SkinChips were also stimulated with Hi P and Lo PAB for 4h before perfusion of both human peripheral blood isolated HSPCs and PMNs through the vascular chamber, and attachment of CellTracker-labeled immune cells were counted (D). SkinChips were then fed with an HSPC granulopoietic expansion media containing Stem Cell Factor (SCF), Granulocyte-Colony Stimulating Factor (GCSF), IL-6, and flt3 ligand suspended in StemSpan SFEM II and EGM2 (Lonza) (E).

Results, Conclusions, and Discussions:: Our SkinChip reveals an immune-reactive and perfusable model of human skin that mimics critical features of the response to inflammation including attachment, extravasation, and expansion of HSPCs. First, our model can respond to artificial stimulation of the skin. In human skin, the introduction of foreign bacteria, like S. aureus, exposes the skin to pathogen-associate molecular patterns (PAMPs) like LPS where resident skin immune cells produce damage-associated molecular patterns (DAMPs, TNF-α, IL-1β, IL-6) which then upregulate integrins through downstream signaling. Our SkinChip predictably upregulates the expression of ICAM-1 in response to direct DAMP stimulation but also upregulates these integrins in response to PAMPs (Pam3CSK4, an artificial PAMP) via Toll-like receptor 2 (TLR2) signaling. Direct stimulation with human-generated DAMPs is a robust positive control to measure integrin upregulation, but Pam3CSK4 stimulation through the TLR2 pathway reveals how the SkinChip can respond to infection by generating enough of its own DAMPs to stimulate downstream immune responses without exogenous intervention. Given this immune reactiveness, the SkinChip after 4h of exogenous DAMP and PAMP stimulation recruits both HSPCs and PMNs to differing degrees. The SkinChip recruits the most HSPCs and PMNs after high exposure directly to DAMPs (Hi AB), although the SkinChip still recruits relatively more PMNs after PAMP exposure alone, revealing slightly different mechanisms in total recruitment. No other study has captured HSPC arrest and extravasation with such temporal resolution under conditions simulation inflammation. After arrest though, many HSPCs remain relegated to the vascular niche, and do not migrate as deep or as quickly into the dermal layer as PMNs (not shown). However, these peripheral blood HSPCs, when the SkinChip media is changed from EGM2 to a granulopoietic expansion media, can expand and differentiate into PMNs in both the dermal compartment and marginated pool over 3 to 7 days. Further work is planned to examine differing rates of HSPC expansion with additional TLR2 stimulation, and these expanded PMNs will undergo surface marker characterization and functionality assays such as NETosis.

Acknowledgements and/or References (Optional)::