Engineering water exchange is a safe and effective method for magnetic resonance imaging in diverse cell types

Abstract

Aquaporin-1 (Aqp1), a water channel, has garnered significant interest for cell-based medicine and in vivo synthetic biology due to its ability to be genetically encoded to produce magnetic resonance signals by increasing the rate of water diffusion in cells. However, concerns regarding the effects of Aqp1 overexpression and increased membrane diffusivity on cell physiology have limited its widespread use as a deep-tissue reporter. In this study, we present evidence that Aqp1 generates strong diffusion-based magnetic resonance signals without adversely affecting cell viability or morphology in diverse cell lines derived from mice and humans. Our findings indicate that Aqp1 overexpression does not induce ER stress, which is frequently associated with heterologous expression of membrane proteins. Furthermore, we observed that Aqp1 expression had no detrimental effects on native biological activities, such as phagocytosis, immune response, insulin secretion, and tumor cell migration in the analyzed cell lines. These findings should serve to alleviate any lingering safety concerns regarding the utilization of Aqp1 as a genetic reporter and should foster its broader application as a noninvasive reporter for in vivo studies.

Keywords: Cell physiology; Diffusion; MRI; Reporter gene; Unfolded protein response; aquaporin-1.

Fig. 1

Diffusion-weighted imaging of Aqp1-expressing cells. a, Illustration of the Aqp1 contrast mechanism. Heterologous expression of Aqp1 increases the baseline water diffusivity () of cells by facilitating rapid water exchange across the cell membrane. The larger diffusivity of Aqp1-expressing cells () makes them observable using diffusion-weighted MRI. b, Percent increase in diffusivity () of Aqp1-expressing cells relative to control cells transduced to express GFP. Error bars represent the s.e.m. from n = 4–8 independent biological replicates. c, Diffusion maps of axial cross-sections of pellets of Aqp1- and GFP-expressing cells. Each voxel in the diffusion map represents absolute diffusivity, estimated from a voxel-wise regression of the first-order decay in signal intensity with diffusion-weighting (i.e., effective b-value). The diffusion map is denoised using a median filter and displayed using a linear 8-bit color map whose lower and upper limits denote diffusivity in µm²/ms. All MRI data were acquired at 7 T, using a diffusion time of 300 ms

Fig. 2

Effects of Aqp1 expression on cell morphology and viability. a, Representative quantitative phase microscopy images of cells engineered to express Aqp1 or GFP. Scale bar is 5 μm. Fold-change in b, dry mass, c, volume, and d, sphericity of Aqp1-expressing cells relative to GFP-cells measured using quantitative phase imaging. e, Viability of Aqp1-expressing cells relative to GFP-controls measured using the MTT assay. f, ATP levels in Aqp1-expressing cells relative to GFP. g, Caspase-3/7 activation in Aqp1-cells relative to GFP. The dashed lines denote the unit fold change representing no difference between GFP and Aqp1. Error bars in the phase imaging experiments represent the s.e.m. from n ≥ 3 images comprising 10–50 single cells. Error bars in the viability and caspase assays represent the s.e.m. from n ≥ 3 biological replicates

Fig. 3

Effect of Aqp1 expression on the unfolded protein response. a, Fold changes in key UPR-associated genes, BiP, XBP1s, and CHOP in Aqp1-expressing cells relative to GFP controls. b, Fold changes in the BiP, XBP1s, and CHOP in Aqp1-expressing cells treated with 2.5 µg/mL tunicamycin for 4 h relative to vehicle-treated cells. Error bars represent the s.e.m. from n = 3 biological replicates. * P-value < 0.05, ** P-value < 0.01, *** P-value < 0.001. GAPDH and actin were used as housekeeping genes for the mouse and human cell lines, respectively

Fig. 4

Effect of Aqp1 expression on specific cell functions. a, Representative images showing the engulfment of silica beads coated with a supported lipid bilayer (atto390, magenta) by Aqp1-IRES-GFP and GFP-expressing J774A.1 cells (GFP, green). Images were collected using a spinning disk confocal microscope with a 40 × 0.95 NA Plan Apo air objective. Yellow arrows denote beads engulfed by the cells. Scale bar is 5 μm. b, Phagocytic activity of Aqp1- and GFP-cells following exposure to lipid-coated silica beads with or without IgG1κ incorporated in the supported lipid bilayer, for 45 min. Phagocytic index was determined by measuring the average att390 fluorescence per macrophage.c, Glucose-stimulated insulin secretion in Aqp1- and GFP-expressing MIN6 cells induced by exposing cells to 20 mM external glucose for 1 h. d, Flow cytometry analysis of CD25 and CD3 surface expression in Aqp1- and e, GFP-expressing Jurkat cells, either untreated or incubated with anti-human CD3/CD28 beads at a 4:1 bead-to-cell ratio for 24 h. The numbers in each quadrant denote the fraction of cells showing surface expression of CD3 (pan T-cell marker), CD25 (activation marker), both CD3 and CD25, or neither marker. f, Percentage of CD25⁺ cells relative to the total CD3⁺ population. g, Representative images showing the migration of Aqp1- and GFP-expressing MDA-MB-231 cells beyond the Matrigel perimeter on day 6. Wide-field images covering both the Matrigel drop and the surrounding media were acquired using a scanning confocal microscope with a 10 × 10 grid of 1024 × 1024 pixel tiles at 10X magnification. h, Increase in the total area occupied by invading MDA-MB-231 cells that migrate out of the Matrigel. Error bars represent the s.e.m. from n ≥ 3 biological replicates. * P-value < 0.05, ** P-value < 0.01, *** P-value < 0.001 (2-sided, t-test)