Tissue clearing of human iPSC-derived organ-chips enables high resolution imaging and analysis

Abstract

Engineered microfluidic organ-chips enable increased cellular diversity and function of human stem cell-derived tissues grown in vitro. These three dimensional (3D) cultures, however, are met with unique challenges in visualization and quantification of cellular proteins. Due to the dense 3D nature of cultured nervous tissue, classical methods of immunocytochemistry are complicated by sub-optimal light and antibody penetrance as well as image acquisition parameters. In addition, complex polydimethylsiloxane scaffolding surrounding the tissue of interest can prohibit high resolution microscopy and spatial analysis. Hyperhydration tissue clearing methods have been developed to mitigate similar challenges of in vivo tissue imaging. Here, we describe an adaptation of this approach to efficiently clear human pluripotent stem cell-derived neural tissues grown on organ-chips. We also describe critical imaging considerations when designing signal intensity-based approaches to complex 3D architectures inherent in organ-chips. To determine morphological and anatomical features of cells grown in organ-chips, we have developed a reliable protocol for chip sectioning and high-resolution microscopic acquisition and analysis.

Fig. 1. SC-Chips neural tissue increases in opacity overtime. (A) Schematic of SC-Chip culture. Human spNPCs and BMECs are cultured from iPSCs, with the spNPCs seeded on the top channel, and the BMECs seeded on the bottom channel, cultured under flow. (B) Tissue growth and opacity increase over time (scale bar = 200 μm). (C) Histogram of greyscale values from a single chip over time. (D) Average translucence of multiple chips over days in culture.

Fig. 2. Tissue opacity is effectively cleared using the SCALE method on chips. (A) Schematic of the SCALE process modified for organ-chip use. (B) Comparison of ICC vs. SCALE phase imaging d28 (scale bar = 200 μm). (C) Phase imaging of ICC vs. SCALE after undergoing their respective staining protocols, and quantification of opacity changes from phase images of chips processed using either ICC or SCALE (scale bar = 200 μm) (n = 10 chips for fixation, n = 5 chips for SCALE and ICC, one-way ANOVA adjusted P values are p < 0.0001 fixation vs. SCALE, p = 0.0131 fixation vs. ICC, and p < 0.0001 SCALE vs. PBS, error bars = SEM). (D) Confocal images of ICC and SCALE chips from Fig 2B, top panels indicate DAPI. White and yellow arrows highlight cells in the top channel and bottom channel, respectively. The middle panels show non-phosphorylated neurofilament heavy chain (SMI32) and islet1 (ISL1)-positive spinal motor neurons, as well as Ki67-postive proliferating cells in the top neuronal channel. Yellow box indicates area quantified for E. Bottom panel is a top-down view of the neuronal channel. All images shown in B–D are taken from the same chips. (E) Quantification of area of visible stain in the top channel (n = 12 sites per condition, 4 individual chips per condition, scale bar z-axis = 50 μm, scale bar top channel = 30 μm, p < 0.0001, unpaired t-test, error bars = SEM).

Fig. 3. SCALE of chip sections enables high resolution microscopy. (A) Representative image of the blade setup of the vibratome for PDMS chip sectioning. (B) Schematic of staining process for sections in 24 well plate. (C) ICC vs. SCALE sections showing SMI32 (white) for spinal motor neurons and nGFP (green) for BMECs, with zoomed in images of SMI32, GFP, and DAPI (scale bar = 100 μm left, 20 μm center, scale bar = 50 μm right). (D) Image at 20× shows imaging capacity for multiple cell markers in 4 wavelengths: 488, 594, 647, 405 (scale bar = 100 μm). (E) Image at 60× oil objective from section using tissue clearing protocol shows ability to decipher individual cell morphology with SMI32 (white), ISL1 (red), and DAPI (blue) (scale bar = 5 μm).