Environmentally Controlled Microfluidic System Enabling Immune Cell Flow and Activation in an Endothelialised Skin-On-Chip

Abstract

Integration of reconstructed human skin (RhS) into organ-on-chip (OoC) platforms addresses current limitations imposed by static culturing. This innovation, however, is not without challenges. Microfluidic devices, while powerful, often encounter usability, robustness, and gas bubble issues that hinder large-scale high-throughput setups. This study aims to develop a novel re-usable multi-well microfluidic adaptor (MMA) with the objective to provide a flexible tool for biologists implementing complex 3D biological models (e.g., skin) while enabling simultaneous user control over temperature, medium flow, oxigen (O₂), nitrogen (N₂), and carbon dioxide (CO₂) without the need for an incubator. The presented MMA device is designed to be compatible with standard, commercially available 6-well multi-well plates (6MWPs) and 12-well transwells. This MMA-6MWP setup is employed to generate a skin-on-chip (SoC). RhS viability is maintained under flow for three days and their morphology closely resembles that of native human skin. A proof-of-concept study demonstrates the system's potential in toxicology applications by combining endothelialised RhS with flowing immune cells. This dynamic setting activates the monocyte-like MUTZ-3 cells (CD83 and CD86 upregulation) upon topical exposure of RhS to a sensitizer, revealing the MMA-6MWP's unique capabilities compared to static culturing, where such activation is absent.

Keywords: immune cell activation; immune cell flow; microfluidics; organ‐on‐chip; reconstructed human skin; skin‐on‐chip.

Figure 1

Schematic diagram of the microfluidic multi‐well adaptor (MMA) construct. A) Main manifold connecting (a2) external fluidic routing tubes for medium perfusion and allowing light transmission through (a1) the traversing round apertures. B) Biocompatible double‐side adhesive tape layer (142 µm‐thick) with through holes allowing (b2) fluidic routing and (b1) light transmission from (A) to (C). C) High transparency and auto‐fluorescence‐free (188 µm‐thick) COP layer with through holes allowing fluidic routing (c2) from (B) to (D). D) Biocompatible double‐side adhesive tape layer (142 µm‐thick) with patterned microfluidic channels (500 µm‐wide) allowing fluidic routing to the manifold (E) and between different wells (d2). (D) contains through holes for light transmission (d1) from (C) to (E). E) Manifold with through holes (e2) and nozzles (e4) allowing fluidic routing from (A) to a standard 6‐well plate (6MWP). Part (E) presents optical apertures (e1) allowing the transmission of light from a microscope to the biological sample once routed through (A), (B), (C), and (D). Part (E) has also assembled toroidal O‐rings (e5) guaranteeing the sealing, while assembled to the 6MWP, of the overall structure to external factors, such as contamination or gas environment. All parts have some extra features (b3, c3, d3, e3) to allow alignment of the multiple layers and to ease the assembling of the MMA.

Figure 2

Assembly of the complete skin‐on‐chip (SoC) microfluidic device and flow characterization during perfusion. A) Expanded view of the MMA‐6MWP assembly including the MMA, the transwell cell culture insert, containing the RhS, and the 6MWP. B) Complete internal fluidic sealed structure of the myeloid cell‐complemented SoC model. The MMA connects three wells in a series. The direction of flow is indicated with dashed arrows. It is designed to maintain a very low volume of media in the entering well of the 6MWP holding circulating immune cells (“immune cell reservoir”). The second well connected to the previous one contains the RhS (“tissue reservoir”) and is designed to allow the flowed medium to contact only the EC layer at the bottom of the transwell insert. The third well works as a medium collector (“collection reservoir”). Excess medium is collected into an Erlenmeyer flask (“collection flask”). At the end of each experiment, the RhS and the flowed media can be recovered by opening the assembly. Created with BioRender.com. C) Modelled WSS at the transwell membrane: when applying a flow of 150 µL min⁻¹, WSS values range between 2.95 × 10⁻⁴ Pa and 1.63 × 10⁻³ Pa, lower than those reported for human blood vessels in literature. D) Modelled Reynolds number at 1 µm under the transwell membrane when applying a 150 µL min⁻¹ flow. Values range between 3.85 × 10⁻⁶ and 2.36 × 10⁻⁵ (laminarity regime under membrane). E) Heating holder of the assembled SoC MMA‐6WMP. F) Cross‐section of the heating holder that shows the MMA‐6MWP‐holder ensemble. A custom‐made polyamide heater integrated into the base of an aluminum plate warms the MMA‐6MWP ensemble, which has been designed to be compatible with real‐time imaging using a Leica DMi8 Inverted stage. G) Temperature calibration of the system was carried out by placing temperature probes (JTs) in three coaxial regions of three different wells near to the transwell membrane. The results allowed to assess H) zonal and I) mean weighted temperature of the culture medium to guarantee appropriate calibration of the temperatures set by the controlling unit.

Figure 3

Complete MPS platform prototype (CubiX MVP2C) controlling the gaseous environment (percentages of CO₂, N₂, and O₂), the perfusion (flow rate), and the heating (temperature) of the myeloid cell‐complemented SoC without the need for an external incubator. A detail of the constructed multi‐well microfluidic adaptor (MMA) is presented in the top‐right of the figure. Medium circulates from the pressurized medium bottle to the “collection flask” via the MMA as depicted by the black dashed arrows. The main components of the CubiX‐MMA‐6MWP‐heater system described in Figures 1 and 2 are labeled in white boxes.

Figure 4

Histological analysis of reconstructed human skin (RhS) in static and dynamic conditions resembles native human skin. Histology of normal human skin and RhS sections by A) hematoxylin and eosin (H&E) staining and immunohistochemical stainings using antibodies directed against B) the basal epidermal layer marker keratin 15 (K15), C) the proliferation marker Ki67, the intermediate epidermal differentiation markers D) involucrin and E) K10, F) the late differentiation marker loricrin, and G) the cornified envelope protein filaggrin. Representative pictures of stained 5‐µm‐thick paraffin‐embedded RhS sections derived from n = 3 independent experiments performed in triplicate for static and duplicate for dynamic conditions are shown. Scale bar = 100 µm.

Figure 5

Exposure of NiSO₄ marginally affects endothelialised RhS morphology and viability in static and dynamic conditions. A) CD31 staining to visualize the ECs underneath the RhS in static or dynamic conditions. Magnifications of CD31⁺ cells are shown in black rectangles. Scale bar = 100 µm. B) Hematoxylin and eosin (H&E) staining of endothelialised RhS exposed to either H₂O or NiSO₄ (190 mm) for 24 h in static or dynamic conditions (scale bar = 100 µm) and C) MTT assay of endothelialised RhS exposed to either H₂O or NiSO₄ (190 mm) for 24 h in static conditions. In these experiments, MUTZ‐3 cells were co‐cultured with the endothelialised RhS models. Data from n = 3 independent experiments are shown as mean ± SD. Each experiment is represented by a different symbol.

Figure 6

Dynamic flow affects epidermal cell proliferation. Proliferating epidermal cells were identified in endothelialised RhS exposed to either H₂O or NiSO₄ (190 mm) for 24 h in static or dynamic conditions A) via Ki67 (red) and DAPI (blue) immunofluorescent staining (scale bar = 50 µm). B) Ki67⁺ nuclei in the stratum basale were calculated and expressed as proliferation index. Data from n = 3 independent experiments are shown as mean ± SD. Each experiment is represented by a different symbol.

Figure 7

Exposure of endothelialised RhS to NiSO₄ promotes the maturation of MUTZ‐3 cells in dynamic conditions. MUTZ‐3 cells were cultured in the presence of endothelialised RhS exposed to H₂O (grey bars) or 190 mm NiSO₄ (green bars) for 24 h in static or dynamic conditions and were assessed in terms of A) cell viability. Data from n = 3 independent experiments are shown as mean ± SD. Each experiment is represented by a different symbol. B) shows a simplified overview of the maturation process of the MUTZ‐3 cells. Created with BioRender.com and adapted from Santegoets et al.^{[ 32 ]} After co‐culture with the endothelialised RhS, expression of the surface markers CD83 and CD86 was measured within C) the live or D) the more differentiated CD14⁺HLA‐DR⁺ cell populations. CD83 and CD86 expression levels were normalized to the control + H₂O (set as 1). Data from n = 3 independent experiments are shown as mean ± SD (^* p < 0.05; unpaired t‐test). Each experiment is represented by a different symbol.

Figure 8

Dynamic flow does not influence MUTZ‐3 cell maturation in the presence of endothelialised RhS exposed to H₂O. MUTZ‐3 cells were cultured in the presence of endothelialised RhS exposed to H₂O in either static (grey bars) or dynamic conditions (green bars) for 24 h and were assessed in terms of expression of the surface markers CD83 and CD86. Data from n = 3 independent experiments are displayed as percentage (%) of live cells. Results are shown as mean ± SD. Each experiment is represented by a different symbol.